eBooks

Ancient Greek and Latin in the linguistic context of the Ancient Mediterranean

0923
2024
978-3-8233-9585-0
978-3-8233-8585-1
Gunter Narr Verlag 
Carlotta Viti
10.24053/9783823395850

Latein und Griechisch werden in diesem Sammelband unter dem Aspekt des Sprachkontakts untersucht, ein Thema, das in unserer globalen und multiethnischen Gesellschaft besonders aktuell ist. Spezialist:innen verschiedener Universitäten und Länder nehmen in Ihren Beiträgen unter anderem die linguistische Variation der griechischen Dialekte, den griechisch-lateinischen Bilinguismus, den Sprachkontakt im alten Italien, Mittleren Osten und Mittelmeer sowie Übersetzungen und Glossen in den Blick. Landkarten und Bilder alter Inschriften und Manuskripte bereichern die Diskussion. Aus interdisziplinärer Perspektive wird außerdem die Linguistik des Lateinischen und des Griechischen in ihrem Zusammenhang mit Epigraphik, Philologie, Textkritik und grammatischer Theorie untersucht. Neben Latein und Griechisch werden Daten zahlreicher alter und moderner Sprachen mit einbezogen.

<?page no="0"?> ISBN 978-3-8233-8585-1 This collective volume examines Latin and Ancient Greek from the perspective of language contact, a topic that is particularly relevant in our globalized and multi-ethnic society. Specialists from various universities and countries investigate, among other things, the linguistic variation of the Greek dialects, Greek-Latin bilingualism, language contact in ancient Italy, in the Near East and in the Mediterranean, as well as problems of translations and glosses. Maps and images of old inscriptions and manuscripts enrich the discussion. From an interdisciplinary point of view, Greek and Latin linguistics is also discussed in relation to epigraphy, philology, textual criticism and grammatical theory. In addition to Latin and Greek, data from numerous ancient and modern languages are presented. Sprachvergleich Studien zur synchronen und diachronen Sprachwissenschaft Band 5 Ancient Greek and Latin in the linguistic context of the Ancient Mediterranean Viti (ed.) Carlotta Viti (ed.) Ancient Greek and Latin in the linguistic context of the Ancient Mediterranean <?page no="1"?> Ancient Greek and Latin in the linguistic context of the Ancient Mediterranean <?page no="2"?> herausgegeben von Paola Cotticelli-Kurras (Verona) Katrin Schmitz (Wuppertal) Joachim Theisen (Athen) Carlotta Viti (Lorraine) wissenschaftlicher Beirat Daniel Petit (Paris) Georges-Jean Pinault (Paris) Sabine Ziegler (Berlin) Sprachvergleich Studien zur synchronen und diachronen Sprachwissenschaft Band 5 <?page no="3"?> Carlotta Viti (ed.) Ancient Greek and Latin in the linguistic context of the Ancient Mediterranean <?page no="4"?> DOI: https: / / doi.org/ 10.24053/ 9783823395850 © 2024 · Narr Francke Attempto Verlag GmbH + Co. KG Dischingerweg 5 · D-72070 Tübingen Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlages unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Überset‐ zungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Alle Informationen in diesem Buch wurden mit großer Sorgfalt erstellt. Fehler können dennoch nicht völlig ausgeschlossen werden. Weder Verlag noch Autor: innen oder Herausgeber: innen übernehmen deshalb eine Gewährleistung für die Korrektheit des Inhaltes und haften nicht für fehlerhafte Angaben und deren Folgen. Diese Publikation enthält gegebenenfalls Links zu externen Inhalten Dritter, auf die weder Verlag noch Autor: innen oder Herausgeber: innen Einfluss haben. Für die Inhalte der verlinkten Seiten sind stets die jeweiligen Anbieter oder Betreibenden der Seiten verantwortlich. Internet: www.narr.de eMail: info@narr.de Elanders Waiblingen GmbH ISSN 2569-2275 ISBN 978-3-8233-8585-1 (Print) ISBN 978-3-8233-9585-0 (ePDF) ISBN 978-3-8233-0521-7 (ePub) Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http: / / dnb.dnb.de abrufbar. <?page no="5"?> 7 11 61 91 97 137 161 217 243 253 271 Contents Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carlotta Viti Introduction: Language contact, comparative linguistics, and comparative literature in their historical and cultural context . . . . . . . . . . . . . . . . . . . . . . Monique Bile, René Hodot, Guy Vottéro Le grec ancien : une réalité multiforme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Emmanuel Weiss La signature des hiéromnémons (À propos des tablettes de Locres n° 15 et 26) Georgios K. Giannakis The Magna Graecia Tablets in the Dodona Corpus . . . . . . . . . . . . . . . . . . . . . José Luis García Ramón From the Files of a new Bechtel, Die historischen Personennamen des Griechischen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Valerio Pisaniello & Stella Merlin Ancient Greek and late Anatolian languages: from coexistence on the territory to survival in the scholarly tradition . . . . . . . . . . . . . . . . . . . . . . . . . Filip De Decker The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? An analysis of the aspect use in the Epic-Ionic -σκ-iteratives in the Odyssey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rex Wallace Comments on an Old Latin Inscription from San Giuliano (CIL I2, 2780) . . Luca Rigobianco Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gualtiero Calboli Nausistrata (Ter. Phorm. 784-1055) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . <?page no="6"?> 295 317 345 389 413 439 456 Michael Weiss & Ryan Windhearn Syntactic Theory and Textual Criticism in Plautus: Camerarius’ Emendation of Amphitruo 260 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Paola Cotticelli-Kurras The creation of linguistic metalanguage in Antiquity and Middle Ages as result of translational processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anna Orlandini & Paolo Poccetti La négation comme moteur de l’évolution linguistique . . . . . . . . . . . . . . . . . Vit Bubenik Language Contact in Antiquity. Participial Constructions in Hellenistic Greek, Hebrew/ Aramaic and Old Church Slavonic . . . . . . . . . . . . . . . . . . . . . Mauro Aresu & Silvia Cabriolu Observations on the variation of word-order: a comparison of Preclassical, Classical and Biblical Latin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Francesca Cotugno The Vandalic language in the light of Latin medieval manuscripts . . . . . . . Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Contents <?page no="7"?> Contributors Mauro ARESU: Research fellow in Latin Language and Literature at the University of Cagliari (Italy), Department of Literature, Languages and Cultural Heritage. Email: mauro.aresu2@unica.it Monique BILE: Associate Professor (Maître de Conférences) Emer. of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages. Email: monique.bile@univ-lorraine.fr Vit BUBENIK: Professor Emer. of Linguistics at the Memorial University of Newfoundlands (Canada), Department of Linguistics. Email: vbubenik@mun.ca Silvia CABRIOLU: Post-Doc in Historical and General Linguistics at the University of Cagliari (Italy), Department of Literature, Languages and Cultural Heritage. Email: silvia.cabriolu@unica.it Gualtiero CALBOLI: Professor Emer. of Latin Language and Literature at the University of Bologna Alma Mater Studiorum (Italy), Department of Classical and Italian Philology. Email: gualtiero.calboli@unibo.it Paola COTTICELLI-KURRAS: Professor of Historical and General Linguistics at the University of Verona (Italy), Department of Cultures and Civilizations. Email: paola.cotticelli@univr.it Francesca COTUGNO: Post-Doc in Historical and General Linguistics at the University of Verona (Italy), Department of Cultures and Civilizations. Email: francesca.cotugno@univr.it <?page no="8"?> Filip DE DECKER: Post-Doc holder of a Marie Skłodowska Curie Actions, European Fellowship, and research fellow at the University of Verona (Italy), Department of Cultures and Civilizations. Email: filipdedecker9@gmail.com José Luis GARCĺA RAMÓN: Universität zu Köln / Università Cattolica del Sacro Cuore di Milano (Italy), Dipartimento di Filologia Classica, Papirologia e Linguistica Storica. Email: garcia.ramon@uni-koeln.de Georgios K. GIANNAKIS: Professor of Greek Linguistics at the Aristotle University of Thessaloniki & Center for the Greek Language (Thessaloniki, Greece), Department of Linguis‐ tics. Email: ggianak@lit.auth.gr René HODOT: Professor Emer. of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages Email: rene.hodot@univ-lorraine.fr Stella MERLIN: Post-Doc in Historical and General Linguistics at the University of Naples Federico II (Italy), Department of Human Studies. Email: stella.merlin@unina.it Anna ORLANDINI: Professor Emer. of Latin and Greek Linguistics, “Centre A. Ernout”, Paris-Sor‐ bonne (France). Email: orlandinianna@libero.it Valerio PISANIELLO: Associate Professor of Historical and General Linguistics at the University of Chieti-Pescara (Italy), Department of Languages, Literatures and Modern Cultures. Email: valerio.pisaniello@unich.it Paolo POCCETTI: Professor Emer. of Historical and General Linguistics at the University of Rome 2, Tor Vergata (Italy), Department of Literary and Philosophical Studies and of History of Art. Email: paolopoccetti@tiscali.it Luca RIGOBIANCO: Researcher of Historical and General Linguistics at the University Ca’ Foscari of Venice (Italy), Department of Humanities. Email: luca.rigobianco@unive.it 8 Contributors <?page no="9"?> Carlotta VITI: Professor of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages. Email: carlotta.viti@univ-lorraine.fr Guy VOTTÉRO: Professor Emer. of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages. Email: guy.vottero@univ-lorraine.fr Rex WALLACE: Professor Emer. of Classics at the University of Massachusetts Amherst (United States of America), Department of Classics. Email: rwallace@umass.edu Emmanuel WEISS: Associate Professor (Maître de Conférences) of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages. Email: emmanuel.weiss@univ-lorraine.fr Michael WEISS: Professor of Classics and Indo-European Linguistics at Cornell University, Ithaca, NY (United States of America), Department of Linguistics. Email: mlw36@cornell.edu Ryan WINDHEARN: Research fellow in Indo-European Linguistics at Cornell University, Ithaca, NY (United States of America), Department of Linguistics. Email: ryan.windhearn@gmail.com Contributors 9 <?page no="11"?> 1 I would like to thank Martin Worthington and Geoffrey Khan for useful information about Akkadian, Hebrew, and Semitic in general. I also thank Michael Weiss, Anna Orlandini, Paolo Poccetti, Paola Cotticelli Kurras, Gualtiero Calboli, José Luis García Ramón and Rex Wallace for their kind feedbacks on this introduction. The arguments here presented and their potential shortcomings are purely my responsibility. Introduction: Language contact, comparative linguistics, and comparative literature in their historical and cultural context 1 Carlotta Viti (University of Lorraine) 1. Possible dialogues between language contact and historical linguistics The popularity of linguistic research traditions often depends on factors external to language itself. In our global, multilingual and multicultural societies, a particular interest has recently developed in academia around the subject of language contact. In general linguistics, since Weinreich’s (1977) pioneering studies on this topic, language contact has been the subject of publications with data drawn from many different languages (cf. Thomason & Kaufman 1988; Aikhenvald & Dixon 2001a; Thomason 2001; Matras 2009; Ansaldo 2013; Bakker & Matras 2013; Grant 2020, etc.). For example, it has been often observed that nouns (especially nouns other than kinship terms, body part nouns, personal pronouns and low numerals) are usually borrowed more easily than verbs or other parts of speech, and that lexical items in general are usually borrowed more easily than phonemes and morphemes (cf. Moravcsik 1978). But this also depends on the intensity of language contact. It is true that borrowing particularly prevails in situations of casual or superficial contact. In cases of in‐ tense contact such as substrate interference, however, structural transfer has no constraints, and usually occurs earlier and more extensively than lexical transfer (cf. Thomason & Kaufman 1988: 83 ff). For example, the Asia Minor Greek dialects studied by Dawkins (1916) adopt many features of Turkish grammar, <?page no="12"?> ranging from the loss of gender and adjective-noun agreement to word order and even vowel harmony. In general, it is now recognized that manifestations of language contact depend more on the sociocultural settings of a speech community than on purely linguistic factors. More specific predictions can be made by considering social variables such as the presence of an indigenous superordinate group, a migrant superordinate group, an indigenous subordinate group or a migrant subordinate group (cf. Thomason 2001: 23), as well as prestige factors, since the same linguistic phenomenon may have different outcomes according to different variables of language contact. There are various theories of language contact, which is also acknowledged as a major mechanism of language change. For example, Harris & Campbell (1995) consider language contact as one of the three main mechanisms of syntactic change, alongside reanalysis and extension. By contrast, in the past, language contact was neglected. This neglect is even more pronounced in the field of historical linguistics and Indo-European (IE) studies. The reconstruction of Proto-Indo-European (PIE) is achieved by observing mechanisms of internal language change. This is not only evident in Internal Reconstruction, by definition, but also in the Comparative Method. It is traditionally assumed that two or more structures are cognates only once language contact, as well as other possible factors of formal correspond‐ ence such as chance and independent drift, are ruled out (cf. Meillet 1925). Accordingly, the establishment of a genetic relationship is “an argument by elimination” (Harrison 2003: 215). The Neogrammarians, among the founders of the Comparative Method, invoked language contact only to explain exceptions to sound laws. Latin rūfus “red, reddish; redhaired” (< PIE *h 1 rowd h -ó-), for example, presents a radical fricative f, instead of a plosive as in ruber “red” (< PIE *h 1 rud h -ró-, cf. gr. ἐρυθρóς), because it is borrowed from a Sabellic dialect, as the change PIE *d h > f is unconditioned in Sabellic. Exceptions to sound laws are usually unsystematic (unless they can be subsumed under another more specific sound law). Still in recent times, Watkins (2001) considers the Comparative Method to be the best heuristic tool to describe language relatedness, and dismisses alternative models which also take into account areal diffusion (e.g., Punctuated Equilibrium, cf. Dixon 1997) as unsatisfactory. Contact is still currently often seen as “a confounding factor” (Walkden 2013: 96) in the establishment of genetic inheritance. According to Dunn (2015: 190), for example, “all historical linguistics is phylogenetic, since phylogenetics encompasses the scientific investigation of the descent of organisms in general”, which in principle excludes the possibility of historical language contact. There have been studies of language contact in IE scholarship (e.g., Meid’s 2012 12 Carlotta Viti <?page no="13"?> writings on the contacts between Celtic and Germanic, as well as on Cimbrian, an Upper Germanic variety spoken in some northeastern regions of Italy which has been strongly influenced by Italian). These studies, however, have set out to discuss specific cases of language contact, rather than to establish a theory of language contact, as we have in general linguistics. Moreover, they are rare in comparison with IE studies on other linguistic topics such as historical phonology, morphology and (to a lesser extent) syntax. As the IE languages are the most studied languages of the world, this approach has also affected the diachronic study of other language families. Anybody trained in IE linguistics, myself included, agrees with Watkins (2001) that the Comparative Method remains the most valuable scientific instru‐ ment of linguistic reconstruction in IE and beyond (cf. also Baldi 1990; Polomé & Winter 1992; Weiss 2015). As Weiss (2015) rightly points out, one of the great merits of the Comparative Method is that it enables the reconstruction of much deeper linguistic stages than those of our earliest written records. Being the best method, however, or being a correct method, does not mean being perfect or complete - no scientific method can have such pretensions, as it always implies a certain simplification of reality. The Comparative Method devotes more attention to form than to function, since changes affecting forms are more regular and therefore more predictable than changes affecting meanings. Owing to this, linguistic reconstruction is performed by simplification, since there is apparently no need for two or more forms for (what seems to be) the same function: given a form x and a cognate form y, the Comparative Method reconstructs one single form for the proto-language, which may be x or y or a different form from which both x and y descend - it does not reconstruct both forms. As such, the Comparative Method cannot always adequately deal with linguistic variation. This reduction of structural diversity is also a reason why language contact is traditionally excluded from the domain of reconstruction: language contact intrinsically implies language variation. Co-occurring struc‐ tures, however, often show a functional competition, if function is not limited to the basic lexical-semantic level but also includes pragmatic, sociolinguistic, and dialectal factors. This functional competition, which may be captured by the framework of Construction Grammar (Goldberg 1995; 2006), for example, emerges even more clearly in situations of language contact. Moreover, if one accepts the uniformitarian assumption that languages of the past behaved in the same way as contemporary languages, functional competition and structural variation may be also reconstructed for proto-languages. The reconstructed PIE, for example, was neither simpler nor more regular than the attested daughter languages in morpho-syntax (cf. Viti 2015). Although a functional Introduction 13 <?page no="14"?> competition can be more easily identified in different morphological and syntactic structures than in different phonetic representations, phonetics is also amenable to a functional analysis. For example, in Kiezdeutsch (the variety of German spoken by young generations in multicultural suburban environments, originally in Berlin-Kreuzberg), Weirich et al. (2020) have observed that the phonetic alternation between the voiceless palatal fricative [ç] (ich-laut) and its coronal variant [ɕ] may express distinct sociolinguistic connotations. Although the pronunciation [ç] as in Hochdeutsch predictably has a higher prestige than its Kiezdeutsch realization [ɕ], not all listener groups behave alike (e.g., the non-canonical pronunciation is more stigmatized by older listeners than by young listeners, by in-groups as opposed to out-groups, etc.). Similar considerations may be applied to the Family Tree model. The results of the Comparative Method make it possible to represent the relationship among the various IE languages in a Family Tree, which does not allow cross-branching. That is, after two or more branches are separated, they do not intersect anymore. In this, sub-branching is typically established on the basis of shared innovations. The Family Tree is also a cultural product of its time. Schleicher (1861), one of the creators of the Stammbaum, was inspired by the newly developed method of the stemma codicum of the philological research tradition. The latter aims to establish the relationship among different manuscripts of a certain text, and reduces the lectio transmitted by two or more codices to one original version. The stemma codicum is also presented as a cladogram (cf. Hoenigswald 1963; Fisiak 1990). We have to consider, however, that the Family Tree may not be equally satisfactory to describe cognateness in all language families. It works well for the IE languages, for which it was originally introduced, because these languages have spread across an extremely large area, ranging from Iceland and Ireland to the West to the Tarim Basin (in modern-day Xinjiang, China) to the East, and some of them have been isolated from each other for centuries. When, however, a language family has remained in a relatively more restricted area, the structure of a Family Tree is more controversial. In Semitic, for example, which has remained substantially limited to the Near East until the Muslim conquests in the early Middle Ages, there are different models of classification (cf. Huehnergard & Rubin 2012), and sub-branching is much more debatable than in IE. Similarly, a Family Tree does not adapt so well to situations of extensive bilingualism, as in the languages of South East Asia, where cross-branching may occur. At their time, the concepts of sound laws and of the Family Tree were also challenged, especially by scholars working on dialectology and modern languages (cf. Schuchardt 1885), whose complex developments were not easily amenable to the regularity of the Neogrammarians’ Lautgesetze. Schmidt’s 14 Carlotta Viti <?page no="15"?> (1872) Wellentheorie, implying a gradual diffusion of linguistic features from its region of origin, was posited as an alternative to the Stammbaum to explain branching within the IE language family (cf. also Porzig 1954 for sub-branching and areal diffusion in IE). De Saussure’s (1916) attention to the social factors underlying linguistic change was also critical of the Neogrammarians’ recon‐ struction methodology. These alternative models (which are not incompatible with the Comparative Method in principle) were adopted by a minority of scholars. Otherwise, the Comparative Method has been tested by means of computational and statistical approaches, which are quite popular at the present time (cf. Ringe et al. 2002; McMahon & McMahon 2005, etc.; more recently, cf. Heggarty et al. 2023). The earliest statistical methods applied to historical linguistics, going back to Swadesh’s Glottochronology in the 1950s, assumed a constant rate of language change (which is clearly not true). Later models, e.g., the “character-based models of change”, continued to consider parsimony to be one of the most reliable criteria to account for branching, which is also controversial (cf. Dunn 2015: 196 ff). The newest computational models, such as “likelihood methods”, accept a variable rate of change, but still assume that this variability can be measured by means of mathematical algorithms. Some models evaluate the statistical likelihood of branching by means of the Bayesian Monte Carlo Markov chain. Different computational methods exist, whose reconstructions do not always overlap (cf. Widmer 2018). Most historical linguists, however, are sceptical of computational approaches to language change. This is not only due to the fact that most historical linguists are unfamiliar with complex mathematical methods (this problem would be easily solved by collaborating with a computer scientist), but also and especially because they reject a basic assumption of these methods, that is, the idea that language change and language affiliation can be statistically predicted. In historical linguistics, most scholars hold that language change is regular a posteriori but unpredictable a priori (or at least not predictable by mathematical methods), and characterized by variation between different outputs in the intervening time: A > A/ B > B, where there is no way to predict that A will change to B. For instance, in sound change, we commonly have [s] > [h] > zero, but some languages retain [s] (e.g., Indic), others develop [h] (e.g., Iranian) and others proceed to the development of zero (e.g., Greek psilotic dialects). If we have to put forward hypotheses as regards the more or less likely occurrence of change, these hypotheses depend on a multitude of linguistic and extra-linguistic factors, which by definition defy considerations of parsimony, and which cannot be captured by a mathematical algorithm (Ockham’s razor usually does not work in language change, see below). Rural spaces, for example, Introduction 15 <?page no="16"?> 2 “Ciò che, dunque, determina il ‘tempo’ del mutamento linguistico è la debolezza o la crisi di una tradizione, che lascia spazio alle varianti contrarie; e ciò che lo rallenta è, invece, la saldezza della medesima tradizione. Fatti culturali, dunque: il tempo della lingua non è il tempo della natura, ma il tempo della cultura.” (Lazzeroni 1987: 32) I fully agree with this statement. are usually more conservative than urban spaces (cf. Janda & Joseph 2003: 62-63). In periods of political, economic and social stability, language change is slower than during periods of turmoil or war (but we cannot predict based on mere linguistic features whether a speech community will undergo social instability). There is a difference between variation in social sciences, which can be measured statistically, and the implementation of a language change. See Lazzeroni’s (1987) dated but still instructive observations against computational approaches to language change. 2 Especially for lexical features, matters of a speaker’s choice play a crucial role (cf. Hagège 1993: 9 ff). For ancient languages, we also have to take into consideration that the interpretation of a language change can also depend on text transmission, as some language varieties (especially those associated with socially higher registers) may have been better transmitted than others and may therefore not accurately reflect the language as it was really used. Typologists who know languages prima manu by fieldwork also realize that language change has no predictable rate and mainly depends on social and cultural factors. Consequently, they also do not appreciate computational approaches (cf. Aikhenvald & Dixon 2001b: 7 et passim). The matter is not settled, however. As can be seen, the Comparative Method, as well as the Family Tree, are not a tenet established once and for all, but rather the result of a lively debate that has developed across two centuries (Hymes 1974; Fox 1995; Durie & Ross 1996). Although they are substantially correct, new findings from other language families or other geographic areas, as well as from dialectology and language contact, may be useful to elucidate certain linguistic phenomena concerning language variation which are not adequately captured by these theoretical tools. The reluctance to contemplate contact factors in historical linguistics is based on two main assumptions, related to each other, which have been proven to be wrong in more recent literature. Firstly, it was assumed that internal and external language change are mutually exclusive, and secondly, that an internal explanation to language change is always to be preferred to an external one. Lass (1997: 201 ff), for example, argues that explaining language change in terms of language contact must be the ultima ratio to be used only when an internal explanation (by means of analogy, reanalysis, etc.) is not available. This is because internal change always takes place, even without an external 16 Carlotta Viti <?page no="17"?> influence, while external change is not necessary. An explanation in terms of inheritance therefore seemed to be more parsimonious. This argument is substantially similar to that used by the Neogrammarians to account for sound laws. Terms such as “simple”, “economic”, “consistent” or “harmonious” also abound in structuralist studies, and are evaluated positively in comparison to complex, heavy or inconsistent systems. Although these beliefs continue to appear, especially in the generative research tradition, nowadays historical linguists commonly admit the possibility of multiple causes interacting in language change. Indeed, as Joseph (2013) observes, it is good practice in historical linguistics to take all possible factors into consideration, since a historical explanation has to be correct and complete, and not necessarily simple. Note that Joseph has worked intensively on Balkan languages - he understands the importance of language contact very well. Dorian (1993) also contests previous assumptions whereby divergence and convergence in language change are simplistically associated to internal and external factors, respectively. She shows that language contact may bring about both divergence and convergence - it must be studied on a case-by-case basis. This may be seen as less economic, but it is more reliable. Moreover, a linguistic feature may develop by internal language change and at the same time be reinforced by the pressure of language contact. This is the logic of Heine & Kuteva’s (2003) “contact-induced grammaticalization”, that is, a kind of grammaticalization (i.e., an internal language change by which content words tend to become more grammatical with increasing formal erosion, fixed position, semantic bleaching, etc.) which also occurs in situations of language contact. Heine & Kuteva (p. 71-73) report the case of some Western Slavic languages, such as Sorbian, Czech, and Slovenian, which at least in the spoken language have developed patterns of marking (in)definiteness in a similar way to the article. While Czech and Slovenian use some demonstrative elements to mark definiteness, Sorbian has grammaticalized the use of the numeral “one” as an indefinite marker. On the one hand, these represent the typical paths leading to the formation of articles in languages. On the other, the influence of German is undeniable. German, which possesses both definite and indefinite articles, is spoken in geographically close areas which have always had an intense commercial and cultural interaction with West Slavic speech communities. Also in this case, contact-induced grammaticalization is especially observable in morpho-syntactic change, but phonology is by no means excluded. In both cases, the change may affect both marked and unmarked features. For example, retroflex consonants, which represent phonetically marked segments, could develop in the Satəm group of IE languages by internal language change Introduction 17 <?page no="18"?> 3 On the basis of current findings in contact linguistics, Hock’s (1975) arguments that Ancient Indic retroflex consonants may have developed independently from Dravidian can no longer be retained. Hock observed that retroflex consonants may emerge in languages without external influences, that they are attested in varieties of Early Vedic with very few Dravidian lexical borrowings etc. Firstly, the fact that Indic retroflex consonants may have been influenced by Dravidian by no means excludes that in other languages (e.g., in some Southern Italian dialects, Sardinian and Norwegian) retroflex consonants may have developed spontaneously, without contact. Retroflex consonants may develop in different ways, and probably no change can be motivated only by contact, as probably no change can be only internally motivated - we have to examine them case by case. The argument that, since retroflex sounds may develop secondarily by internal language change, a contact explanation is “unnecessary”, is a wrong argument, as it is founded again on an assumption of economy (that is, on the idea that a language change may have only one motivation) which is now surpassed in linguistic theory. Secondly, as we have seen, not only are internal and external factors not incompatible, but substrate interference starts with phonological, as well as syntactic features, rather than with borrowings. Cf. the discussion of the Indic case study in Thomason & Kaufman (1988: 139 ff), who are also in agreement with most scholars of South Asian languages, arguing for an early influence of Dravidian over Indic in retroflex articulation (cf. Emeneau 1956; 1980; Masica 1976, etc.). In general, the influence between Indic and Dravidian has been mutual, and does not exclude substrate and adstrate factors from Munda and other language families spoken in the area. The result of this complex interaction is South Asia as a Sprachbund. already according to the so-called “Ruki” sound law. In Indic, however, retroflex consonants have a much wider extension than in Armenian or Balto-Slavic. In Indic, retroflex consonants are presumably reinforced by a substrate effect of the Dravidian languages, for which a retroflex articulation is especially characteristic. Retroflex sounds can be reconstructed for Proto-Dravidian as well. 3 Contact-induced grammaticalization may change past assumptions con‐ cerning language contact, as contact is not invoked anymore to explain the development of a seemingly unnatural change or of a language feature that appears to be inconsistent with other features of the system, but rather interacts with internal language change in the same direction. Note that the capacity to assess a directionality in language change is actually a strong point of the Comparative Method: a [p] changes into a [f] much more often than the other way round, so that, in the presence of a [p] in one language and of a corresponding [f] in another related language, we can also reconstruct [p] for their proto-language ceteris paribus. This directionality has been also extended from the domain of phonology, for which it was originally postulated, to morphology and syntax. For example, it has been demonstrated that the change from postpositions to case markers is much more frequent than the opposite change (cf. Hagège 2010; Givón 2021, etc.), so that the proto-language 18 Carlotta Viti <?page no="19"?> of a language with postpositions and of another language with cognate case markers also has to be assigned postpositions. The same directionality may be identified in case of language contact. I have never had difficulties in accepting this concept as I saw it in operation in my own experience. The reader shall forgive this personal anecdote as it is functional to my argument. When I was working at the University of Zurich, German was my working language as well as the language of my daily life. When I had to express likes, I always said ich mag X “I like X” with canonical subject marking, instead of X gefällt mir with non-canonical subject marking. In most contexts both constructions are possible in German (apart from specific cases such as preferences concerning food, where we commonly use non-canonical subject marking with the verb schmecken “to taste”). Still, I consistently used mögen and not gefallen. Note that my mother language, Italian, only admits non-canonical subject marking with likes: mi piace has the same syntactic pattern as German es gefällt mir. (Italian is not like French, where the structure X me plaît exists but is less frequent than j’aime X, j’adore X. The latter options are not available in Italian for this function.) The same applies to my native dialect, Tuscan, which has a different verbal lexeme but the same syntax: mi garba (to.me it.pleases) “I like”. Thus, when speaking in German, I used the pattern which was different from the one used in my mother language. I asked my Italian friends who were living in Zurich, and they told me it was the same for them. This is understandable when we consider that diachronically, in internal language change, experiential predicates tend to acquire canonical subject marking. In Old English, the verb lician “to like” required a dative experiencer, listan “to desire” (German gelüsten) an accusative experiencer, and so on. With time, the English language has lost many impersonal constructions of experiential predicates (cf. van der Gaaf 1904; Allen 1995, etc.). The pattern where the experiencer is also the subject turns out to be preferred in internal language change as well as in situations of language contact. All this confirms that the exclusion of language contact from the practice of linguistic reconstruction may seriously impinge upon an adequate under‐ standing of language change, since no language has evolved in isolation. As LaPolla (2009: 227) pointed out, “language contact is a part of the development of all languages, and so we cannot treat internal language change independ‐ ently from changes influenced by language contact”. Given the fundamentally communicative function of language, contact seems rather to be a natural condition of language, at both a microlevel (as in dialects and sociolects) and a macrolevel (involving different languages), and factors of multilingualism often have a profound effect on the development of a language. A scientific dialogue Introduction 19 <?page no="20"?> 4 Alessandro Manzoni (1785-1873), one of the most influential Italian authors, defined “nation” or “fatherland” (nazione, patria) as “Una d’arme, di lingua, d’altare, / di mem‐ orie, di sangue e di cor” (a unity of army, language, altar, memories, blood, and heart). These verses are drawn from his poem Marzo 1821, the date of an important battle of Italian patriotic movements against the Austro-Hungarian Empire, which had occupied the North of the Italic peninsula. is therefore needed between scholars of historical linguistics and scholars of language contact, who so far have been usually working independently from each other, to compare and combine their techniques, so that some findings of language contact may be incorporated into the methodologies of linguistic reconstruction. This implies an interdisciplinary relationship of historical lin‐ guistics in general, and IE studies in particular, with other disciplines focusing on language interaction and on its extra-linguistic context, such as dialectology, sociolinguistics, anthropological linguistics, pragmatics, discourse analysis, contrastive linguistics, and translation studies. 2. Intellectual history of language contact in Indo-European studies If in principle there is no incompatibility between language contact and the Comparative Method, one may wonder why language contact has been so long neglected in IE studies as compared to other linguistic topics. This is probably also due to ideological reasons. In the past, it was long assumed that language was directly connected to “race”, and this in turn discouraged the study of non-genetic linguistic relationships. We have to put the founding writings of IE linguistics in their historical context. Apart from more or less impressionistic statements about the possible relationships among various IE languages, the earliest serious endeavours to reconstruct the history of IE originated in a series of scholars mainly operating in 19 th century’s Germany - a period in which issues of race and nation, in various forms, played a prominent role in scientific and literary discussions. The first generations among these scholars, that is, Wilhelm von Humboldt (1767-1835), Jacob Grimm (1785-1863) and Wilhelm Grimm (1786-1859), Franz Bopp (1791-1867) etc. - immense scholars - were deeply influenced by Romanticism. This movement, which developed between the end of the 18 th century and the first half of the 19 th century, was rooted in the idea of a “nation”, considered as a unity of language, religion, traditional customs etc. ultimately to be attributed to an ethnic group. 4 Later, in the second half of the 19 th century, Europe was rather influenced by Positivism as well as Darwinism, which are often presented as a reaction of rational thinking against 20 Carlotta Viti <?page no="21"?> Romantic ideas of feelings and subjective interpretations, but still shared the interest in genetic connections. Darwin’s On the origin of species, published in 1859, postulated a common descent of species through a branching pattern of evolution. In addition to philology, as we have seen in §1, Schleicher (1821-1868) was also deeply influenced by Darwin’s theory of evolution (cf. Schleicher 1863). Note that the basic terminology of the family tree (ancestor language, mother / daughter / sister languages, language family, cognate, etc.) is based on metaphors of genetic relations. The same metaphors recur in philology, which speaks about families of manuscripts, genealogy, generation, spurious etc. Latin spurius, for example, originally meant “illegitimate” - it was used in legal language to denote the child of an unknown father or a child born from incest (cf. OLD s.v. spurius; EM 645). The language of metaphors is based on experiences and ideas that are commonly shared in a speech community. All this fits in with the pervasive climate of nationalism, that is, the idea that a state must coincide with (what is considered to be) a nation, an idea which was well established in the 19 th century. Several movements and wars nurtured by nationalist ideas developed in this period. A long hostility existed between Germany and France, for example, as both wanted to extend their control over continental Europe. After the Franco-Prussian War (1870-1871), the Second French Empire lost Alsace and Lorraine, which were annexed to the North German Confederation led by the king of Prussia. The “young” generations of IE scholars, the Neogrammarians (Leskien, Brugmann, Osthoff, Delbrück, Braune, Behagel, Paul, Sievers, etc.) - also immense scholars - working between the second half of the 19 th century and the first decades of the 20 th century, lived in this political and social climate. It is understandable that they were more interested in studying genetic inheritance than contact in languages. The interests of linguists, as of any other people, are influenced by the ideas of their time. We all know how these dangerous nationalist ideas ended up in the subse‐ quent decades and there is no need to dwell further on those dramatic events. Under Nazism, which ruled Germany from 1933 to 1945, the very concept of Indogermanisch was associated to ideas of a pure “Aryan” race. Even the writings of early Indo-Europeanists were reinterpreted in the light of Nationalsozialismus - note the emphasis on “national” in this term. It is clear that language contact could not be an ideal topic of research at that time. Scholars interested in linguistics therefore continued to stress the indigenous material in a language rather than manifestations of language contact. The latter was considered as a sort of contamination of language. The fact that Ancient Egypt - in Africa - and the Ancient Near East - populated by Semitic people, among others - had Introduction 21 <?page no="22"?> advanced cultures which exerted a strong influence on the Greek world since Mycenaean times (see below) was virtually a taboo under the Nazis. But recall that these ideas circulated well beyond Germany and their allies. France and Great Britain had maintained vast colonial empires for centuries, again justified by ideas of racial superiority. In France, the diplomat Arthur de Gobineau, author of the Essai sur l’inégalité des races humaines (1853-1855), asserted the superiority of the Aryan race. In the British Empire, scholars of Sanskrit typically ignored or downplayed the linguistic and cultural contribution of Dravidian and other indigenous languages of India to the development of Indic. It is enough to mention Beames (1872-1879), which remained one of the standard books on New Indic languages at least until Masica (1993). Beames often expressed the argument that the Indo-Aryans were more donors than recipients of linguistic features as they were, in his view, “superior morally as well as physically to the aborigines”, p. 10 et passim). The Jewel in the Crown did not obtain its independence until 1947. Again, language contact was hardly compatible with imperialist and racial arguments. A few words are needed at this point on the Black Athena debate. As is well-known, Bernal (1987; 1991; 2006) considered the reception of Graeco-Roman civilization in Western tradition to be conditioned by racialist ideas, and proposed an alternative model in which certain Afro-Asiatic civiliza‐ tions, notably Egyptians and Phoenicians, played a much more important role in the development of Greek language and culture. Relying on ancient Greek mythological and literary sources, he hypothesized that the ancient Egyptians and Phoenicians colonized parts of ancient Greece, which, in his view, has to be interpreted as a sort of periphery of the Levant. Moreover, he proposed numerous new explanations for Greek words with no etymology, or with a controversial etymology, which he traces back to Afro-Asiatic expressions. In order to avoid misunderstandings, I wish to state clearly that I do not agree with Bernal’s linguistic reconstruction. Bernal is sympathetic with the idea of macro-families, such as Nostratic, and consequently does not follow the assumption of regular IE sound laws. In contrast, I do not subscribe to the concept of macro-families and follow the traditional procedures of the Comparative Method. In this, I agree with Jasanoff & Nussbaum (1996), who recall the method of finding correct etymologies, that is, by looking at regular sound changes, rather than superficial resemblance, etc. However, while we share the same method of linguistic reconstruction, I disagree from Jasanoff & Nussbaum’s (1996) paper with regard to a couple 22 Carlotta Viti <?page no="23"?> 5 - apart from their “strong” language. They say that Bernal will earn a place “in the crank linguistic literature” (p. 202) together with people who derived, for instance, Hungarian from Sumerian or Vietnamese from Ancient Egyptian. They say there is no methodological difference between works of this kind. I do not think we can compare complete fantasy works like these (mainly written a long time ago, by the way) and somebody who proposes borrowings between ancient Greece and the Ancient Near East - even though Bernal’s etymologies are also wrong, in my opinion, at least there is geographic proximity and plenty of evidence of historical contact (see §3). Instead, I think that Bernal will earn a place in intellectual history, as he managed to document via a broad range of historical sources the racial and antisemitic ideas that occupied large portions of Western academia until the end of the Second World War. This is not what he aimed at, but it is still significant, as intellectual history is now an established discipline with its own research tradition. of points. 5 They maintain that Afro-Asiatic borrowings in Ancient Greek are “relatively few in number and - with some exceptions on the Semitic side - late in date” (p. 201). Relying on Masson (1967), a study on Semitic borrowings in Ancient Greek, they assign these terms to the domains of fabrics and items of clothing, commercial terms, vessels, and plants, e.g., Ancient Greek κύμινον n. “cumin”, attested since the Mycenaean age (ku-mi-no, ku-mi-na in a list of spices in Mycenae). But, in fact, Masson is much more open and cautious on this matter. Besides “sure” borrowings belonging to these semantic domains, she also discusses at length a series of “possible” borrowings, which are semantically much more heterogeneous. Among these “mots dont l’origine sémitique est possible” (Masson 1967: 77 ff), we find the name of certain animals, such as λέων / λĩς m. “lion” (p. 85-87). The former variant, λέων, is attested since Mycenaean (re-wo-pi, instr.pl.; re-wo-te-jo adj.). It is considered to be a borrowing from Semitic (cf. Akkadian lābu “lion”, Ugaritic lb’ “id.”, Hebrew labī’ “id.”) or from an unknown Mediterranean language. The rarer variant λĩς has been connected with Hebrew laīš “lion”, but in this case as well the evidence is not conclusive, and it is possible that we are dealing with a mot voyageur of the Mediterranean. Cf. also DELG 635; GEW 113; EDG 854 (for which a “Semitic origin is probable” especially for λέων - note that Beekes is not particularly keen to admit Semitisms). The lion is not native to Europe while it lived in various Near Eastern desertic regions for centuries (in the Near East, the last exemplar of an Asiatic lion was killed in Iraq in 1918). As such, the lion was commonly portrayed in Near Eastern art since antiquity, when it was a symbolic image of the king. Despite the uncertainties about the ultimate source of Ancient Greek λέων and λĩς (in my view, they are more probably borrowed from an unknown Mediterranean language than from Semitic), what is sure is that these words are not IE. Still, some IE scholars have thought the contrary in the past. I briefly Introduction 23 <?page no="24"?> discuss this case not to revisit the etymology of λέων and λĩς, as both Masson (1967) and all lexica recognize that they are not of IE origin, but rather as an exercise of intellectual history of language contact. According to Thieme (1954: 32-37), λέων is etymologically connected with Vedic ruváti “lows” (sound of cattle) and with Homeric Greek βουλυτός m. “evening” (the latter interpreted as “time when the cows come back from the pasture lowing”, rather than “time for unyoking oxen” (βου-λυ-τός), as it is commonly thought). Λĩς, instead, is connected by Thieme to Sanskrit līna- “lying or resting on, lurking, hiding’ as in Kālidāsa’s expression kuñjalīnān-… siṃhān “lions hiding in the underbrush”. Clearly, all this is pure fantasy. But we may better understand the ideological ground behind this hypothesis if we know that it comes from a study, entitled Die Heimat der indogermanischen Gemeinsprache, where the author posits the original homeland of the Indo-Europeans in Northeastern Europe by means of lexical examples such as the name of the salmon (now obsolete for this argument). If then we know that Paul Thieme (1905-2001) was working in Germany during the Third Reich (he also served in the German army during the Second World War), then the picture is much clearer. With this, I do not want to imply that Thieme, one of the greatest Indologists, was sympathetic towards Nazism. On the contrary, we have evidence that he was hostile to the Nazis. For example, in his study Fremdling im Veda, published in 1938, Thieme studied the derivation chain of Vedic arí- “foreigner; enemy”, aryá- “related to the foreigner; kind, favorable”, and ā́rya-, the latter being a common endoethnonym of the Indians of IE origin, opposed to dásyuor dāsá-). Although his interests rely rather on the morphology of these forms, he also explains that ā́ryais to be interpreted in the sense of “hospitable” (zu den Gastlichen gehörig, wirtlich, p. 145), protector of foreigners. He argues that the meaning “lord” as well as the ethnic sense are only secondarily derived from a denotation of a master that is generous to his guests. This argument was dangerous at that time and yet corrected the Nazi’s abuse of the word “Arier”. Thieme later defected during the Cold War from the German Democratic Republic to West Germany (where he became Professor of IE studies at the University of Frankfurt), then moved to the US etc. Still, as everybody else, he reflected the interests of his own time. My point is therefore that, when we study problems of language contact in the ancient IE languages, we must be very cautious because most authors of our sources lived in periods when language contact, as well as cultural contact, was refused or downplayed for ideological reasons related to racial and nationalist arguments - not necessarily their own personal ideas, but certainly ones that were widespread both in society and in academia. The relationship between linguistics and history is not limited to language change, 24 Carlotta Viti <?page no="25"?> 6 Cf. https: / / www.ebl.lmu.de/ corpus/ L/ 1/ 2/ SB/ I 7 Thetis, in her turn, has been approached to Šiduri, the alewife in the Epic of Gilgamesh, in a scene in Il. 24, 120 ff. Here she tries to comfort his son Achilles, who is mourning Patroclus’ death, and says that he should instead enjoy the pleasures of life. Similarly, as Gilgamesh is grieving over the death of Enkidu, Šiduri encourages him to search as in historical linguistics, but also concerns the history of ideas - intellectual history. Ideas affect research interests and the interpretation of data. We have also to pay attention to the fact that etymological dictionaries and the secondary literature in general often repeat former sources without further elaborating the argument. For example, (apart from the Pre-Greek argument, see below), EDG is often a translated copy of GEW - which is understandable, as unfortunately Beekes died before having the time to complete his work. It is also possible (albeit not certain) that the list of early Semitisms in Ancient Greek has to be enlarged with respect to what is commonly assumed to include certain theonyms. One finds insights on this in the works of Burkert, a great scholar of Greek religion and literature who had additionally been trained in IE linguistics and also studied Akkadian. Burkert (2009: 36-37) suggests that the name Τηθύς, -ύος f. “Tethys”, denoting the wife of Oceanus and the mother-god of the river-gods and Oceanides, may be a borrowing from a variant of the name of the Akkadian deity Tiamat, which is also written as Tiamtu, Tâmtu, Tawatu etc. This is not argued just on the base of sound similarity, but rather because the Homeric passage at issue (Hom. Il. 14,200-201) has precise analogies in the Enūma Eliš, one of the most important creation myths of the Babylonian tradition. In the Iliad, Hera says to Aphrodite: εἶμι γὰρ ὀψομένη πολυφόρβου πείρατα γαίης, / Ὠκεανόν τε θεῶν γένεσιν καὶ μητέρα Τηθύν “For I am going to see the boundaries of the all-nurturing earth, and Oceanus, the origin of the gods, and mother Tethys”. At the beginning of the Enūma Eliš epic, we read: “when on high no words was used for heaven / nor below was firm ground called by name, / Primeval Apsu was their progenitor, (Akkadian apsûm-(ma) rēštû zārûšun) / Mother Tiamat was she who bore them all (mummu tiāmtu muʾallidat gimrīšun)”. 6 As Burkert (2009) observes, the similar formulation (Ὠκεανός / Apsu are described as progenitors, and Τηθύς / Tiamtu as mothers) suggests that the concepts of these gods are related. In the Akkadian theogony, Tiamtu plays an important role; for example, she is engaged in the decisive battle against Marduk (cf. Jacobsen 1968). Of Tethys, instead, we have no further information - note that we are speaking about Τηθύς “Tethys”, and not about Θέτις “Thetis”, the mother of Achilles (although some overlaps between these two marine goddesses is plausible). 7 In Homer, Tethys is mentioned only in this incidental passage, within the story of the Διὸς ἀπάτη, as Hera has Introduction 25 <?page no="26"?> love and pleasure (cf. Sironi & Viano 2015). See below for further parallels between Gilgamesh and the Homeric poems. 8 The claim of Paul Kretschmer (1916) that Ἄδωνις (which he reconstructs as originally aspirated) is IE and etymologically connected to ἁδεῖν, ἁνδάνω “to please” is minori‐ tarian (e.g., DELG s.v. Ἄδωνις does not even take it into consideration) and currently obsolete (cf. discussion in Marcovich 1996). P. Kretschmer (1866-1956), incidentally, is usually reticent to admit Semitisms. to find an excuse to obtain Aphrodite’s girdle in order to seduce Zeus. It is therefore more probable that the representation of this Greek goddess derives from the Akkadian one rather than the other way round. The name of Aphrodite herself is possibly borrowed from Semitic, although not directly from the name of her mythological Northwest Semitic correspondence Aštoret / Astarte (the relationship with ἀφρός “foam” is secondary by folk etymology), cf. GEW 196-197; EDG 179. This is because the cult of the goddess of love, sexuality, and fertility, as well as war and power, is certainly of Mesopotamian origin (cf. Sumerian Inanna, Akkadian Ištar, etc.), and not IE. An oriental origin is also probable for Adonis, who in Greek mythology is consistently associated with Aphrodite. On the one hand, from the content point of view, the cult of Aphrodite and Adonis may be a continuation, adapted to the Greek pantheon, of the ancient Sumerian couple of Inanna and Dumuzid, which had been variously replicated in other religions of the ancient Near East, as in the Semitic figures of Baal and Tammuz (cf. West 1997: 57). On the other hand, from the formal point of view, the name of Ἄδωνις has been usually explained as a borrowing from the Canaanite form ’ādōn, which means “lord” (cf. DELG 21; GEW 22; EDG 23). 8 The right track to identify new possible loanwords, in my view, is to search appropriate functional correspondences in mythology and literature. That is, 1) when a Greek word is without a plausible IE etymology (this is the conditio sine qua non to start the research), and 2) this word has a similar form and meaning with respect to a word of an ancient Semitic language (as we are dealing with different language families, sound laws do not help in this case), then 3) we should search for similar contexts in which the forms at issue appear. Words are not transmitted in isolation, but rather in a context, and therefore may maintain their connotations after being borrowed (as the association of Tethys with water). This implies interdisciplinary connections between IE linguistics and ancient literatures, religions, and cultures. A proviso in the use of the term “similar” is needed. As we have seen, the Comparative Method is based on regular correspondence and not on superficial similarity. We have therefore to ascertain that the rendition of a certain foreign phoneme is consistent with that of other borrowings presenting the same phoneme (although the same foreign 26 Carlotta Viti <?page no="27"?> 9 PIE roots with three or more consonants usually present consonant nexus distributed in one syllable (CCvC or CvCC). So-called dysillabic roots, such as Ancient Greek γενεin γένεσις f. “origin”, derive from PIE roots with a final laryngeal, in this case *ĝenh 1 - “generate”. Since laryngeals were originally consonants, these roots were originally monosyllabic with a complex coda. In Semitic, by contrast, the typical root has three radical slots in which three or four consonants are inserted. Most roots present one consonant in each radical slot (C-C-C). A few roots have four consonants and put one consonant in the first and third slot and a cluster of two consonants in the second slot (C-CC-C). Thus, the radical pattern in PIE and in Semitic differ, although both language families also attest minor alternative root patterns. phoneme may be rendered in different ways in different periods in which it enters the target language. According to whether the borrowing is more or less ancient, its phonetic structure will also be more or less integrated in the phonology of the borrowing language). The hypothesis of a Semitic borrowing becomes more plausible when the form has a triconsonantal and disyllabic root, as this is the typical root structure in Semitic, while in IE the typical root structure is biconsonantal and monosyllabic (CvC). 9 But the matter is much more complicated, as we have to consider different kinds of borrowing. Borrowing is not just a simple transfer of lexical material from a source language A to a target language B. Sometimes, the integration of an originally foreign source may bear some analogy to other lexemes of the target language. For example, Ancient Greek ἀδάμας, -αντος m., attested as a common name meaning “hardest metal, steel” since Hesiod (later also “diamond” or “fixed, unalterable”), is often considered, especially for semantic reasons, to be originally “a loanword that was adapted by folk etymology” to δάμνημι (EDG 19; cf. also GEW 19; Ayil 2024: §18 s.v. שׁי ִ מ ָ לּ ַ ח - ḥallāmīs̆ ). A possible comparison has been suggested in this sense with Akkadian adamu “valuable stone” (He‐ brew), but this is not certain (DELG 18 considers ἀδάμας a native term). Thus, when we do etymologies, we often move in the realm of possibility, rather than of certainty, as regular sound laws may interfere with analogy, as the Neogram‐ marians had already stated. Interferences are even more complex to detect when we deal with calques, or semantic loanwords, i.e., when the phonetic material is indigenous but the semantic pattern is influenced by the meaning of a foreign source. Being possible does not mean being wrong, and various degrees of prob‐ ability exist, which must be judged case by case. Introduction 27 <?page no="28"?> 3. Ancient Greek in its contact context 3.1. Mediterranean substrates Greek borrowings from the East are not only from Semitic, of course - borrowings from Semitic are simply more interesting because in such cases we can often make textual parallels. Some Semitic languages, such as Akkadian (Eastern Semitic), Ugaritic, Biblical Hebrew, and Aramaic (all three belonging to Northwest Semitic), had great literary traditions which were widely influential in the ancient Near East. Akkadian, in particular, is attested from the middle of the 3 rd millennium BC until the 1 st century AD (although in the last centuries it was only used for liturgical or academic purposes). By contrast, other linguistic sources are not attested and often not even identifiable. Still, it is especially to these unknown Mediterranean substrate languages that Ancient Greek owes the largest part of its non-inherited vocabulary. Greek forms provided with suffixes such as -νθος (e.g., ὄλυνθος m. “edible fruit of the wild fig”) and -(σ)σος (e.g., κυπάρισσος f. “cypress”) are commonly related to a non-specified Anatolian language (cf. Meillet 1948: 65 et passim). Forms such as Greek μίνθη f. “mint” and Latin menta “id.”, or Greek ῥόδον n. “rose”, Latin rosa f. “id.”, Persian gul “id.” (< *wṛdi-), the latter borrowed into Armenian ward, are too similar to be due to chance but cannot be related to any regular sound correspondence between these languages. There are plenty of these forms, especially among names of plants, animals, and concrete instruments, as well as toponyms, which the Greeks had found in their migratory path through the Balkans to the Aegean. García Ramón (2011) observes that we have to assume more than one substrate languages, and identifies three of them, that is, a heterogeneous Med‐ iterranean substrate language, a Minoan substrate language, and an Anatolian substrate language (Hittite or Luwian). In the latter case, borrowings can also proceed in the other direction, that is, from Greek to Anatolian, or can represent independent developments of previous sources used in the area. Beekes (2014) calls one of these non-attested Mediterranean substrate languages “Pre-Greek”. According to Beekes, a word deprived of a PIE etymology probably has a Pre-Greek origin 1) when it presents an anomalous phono-morphology from an IE point of view (marked consonant clusters such as κχ, e.g., Βάκχος, or τθ, e.g., Ἀτθίς, -ίδος, and rare suffixes such as -αμβ-, -ανδ-, -ανθ-, -αγγ-, -ινδ-, -ινθ-, -ιγγ-, -υμβ-, -υνδ-, -υγγ-), or 2) when the root presents an anomalous alternance with other words, an alternance which is not found in inherited PIE words. The most recurrent patterns are π / πτ (e.g., πόλεμος / πτόλεμος “war”), ξ / σσ (e.g., Ὀδυσσεύς (also Ὀλυττευς) / Οὐλίξης, Οὐλιξεύς (further borrowed through Western Greek dialects into Lat. Ulixes), ττ / σσ (e.g., θάλασσα / θάλαττα “sea” (cf. also δαλάγχαν· θάλασσαν Hesychius.), σ(σ) / στ (e.g., φαῦσιγξ / φαῦστιγξ 28 Carlotta Viti <?page no="29"?> “blister”), plosive / zero (e.g., κάνδαρος / ἄνραξ “charcoal”) etc. Along these lines, Beekes (2014) reconstructs the phonology and the morphology of Pre-Greek. On the other hand, it may seem that Beekes (2010; 2014) connects words with different places and manners of articulation, as well as with zero, when it is convenient, as some forms have been explained differently in the literature and may also receive an IE etymology. Eg. ὀφθαλμός “eye” (originally Pre-Greek according to Beekes 2014: 100-101) is commonly derived from the PIE root *h 3 ek w - “see, behold”, similarly to Latin oculus “eye” and Vedic ákṣi- “id.”, whose radical labiovelar is differently simplified in Ancient Greek, sometimes with irregular results (e.g., Boeotic ὄκταλλος “eye”), for taboo reasons related to the fear of the evil eye (cf. DELG 811-813; NIL 370-383). Moreover, not all alternations indicated by Beekes are clearly developed from the same form, especially when isolated glosses are put together with other more frequently attested forms with disparate meanings. E.g., μάργος “mad” and Hesychius’ glosses such as μαρικᾶς˙ κίναιδος “catamite” and ἄβαρκνα˙ λιμός “hunger, famine” (cf. Beekes 2010: 905) may also be etymologically separate. (Cf. also Colvin 2016 for a critical assessment of Beekes’ hypotheses). More study is needed to elucidate the complex substrate vocabulary of Ancient Greek. 3.2. Eastern influences on Minoan and Mycenaean civilizations While the linguistic non-PIE influence on Greek vocabulary mainly comes from non-attested Mediterranean languages, from the cultural point of view the massive influence of the ancient Near East is undeniable since the beginning of Greek civilization. Without going in details, I limit myself to recall here some of the most salient manifestations of this influence in art and material culture, as they have been reconstructed from archaeological findings and texts. Firstly, when the speakers of the IE variant that would become Greek descended into the Greek peninsula and the surrounding area of the Aegean, at the end of the 3 rd millennium (ca. 2100 BC), they encountered a more advanced civilization, the Minoans, centred in Crete, which was at the crossroads of an intense commercial and cultural exchange with other palace societies of the Bronze Age. In Ancient Egypt, Mesopotamia, Anatolia, and the Levant, we also have in this period various centralized organizations where the palace, dominated by an absolute monarch, is the head of political, administrative, and economic powers. The monumental architecture symbolically represents the importance of the palace for the community. Artistic motifs are also shared among these civilizations. The bull, for example, often appears in Minoan paintings, and is the object of the story of the Minotaur and of several ancient Near Eastern myths and cults, Introduction 29 <?page no="30"?> 10 Bulls and cows are also found in IE imagery, of course - it is enough to think at the importance of the cow (Vedic gó-) in the Indic tradition. Indra, the most important god in Early Vedic, is commonly represented as a bull (vṛ́ṣan-). In later Hinduism, the bull Nandi is the vahana (i.e., animal vehicle) of Shiva etc. The spread of these images is understandable in ancient societies, whose economies were especially based on farming. As in the case of linguistic contact, in cultural contact as well, external factors and internal factors are not mutually exclusive, as an inherited cultural pattern may be entrenched by the presence of similar motifs in a proximate culture. I will come back to this point. such as the Egyptian Apis bull (ḥjpw), the Mesopotamian bull of heaven killed by Gilgamesh, the representation of Marduk as the bull of Utu (the sun god in Sumerian religion), etc. 10 The Greeks took control of Crete and of the Minoan sphere of influence but continued this system of palace economy. Their Mycenaean civilization (ca. 1600-1200 BC) is pervaded by Minoan cultural motifs. The Linear B, for example, used to record the Mycenaean language, is a syllabic writing system written on clay tablets similarly to the Linear A used for the Minoan language (although the Linear A is still undeciphered, and the Minoan language, still unknown, is certainly not IE). Different syllabic and cuneiform writing systems are also attested by other Mesopotamian and Anatolian languages of the 2 nd millennium (within IE, Hittite, (Cuneiform) Luwian, and Palaic). As the ancient societies of the bronze age had such a close interaction, it is understandable that they also came to a similar end. At the end of the 2 nd millennium (from ca. 1250 to 1150 BC), a series of destructions occurred in the whole Eastern Mediterranean. Troy VIIa is destroyed in ca. 1220 BC, the Hittite Empire in ca. 1200 BC. Syria and Palestine experienced periods of instability. Ancient Egypt was attacked - ancient Egyptian sources speak of the “sea people”. The Mycenaean civilization was definitively destroyed in ca. 1050 BC - according to Greek mythology, this was due to the Dorian invasion and to the return of the Heraclids. The Mycenaean clay tablets were burned in the palaces, and have been preserved precisely because they were baked in these arson attacks. The same occurred e.g., to Hittite clay tablets. A type of Aegean syllabary will be longer in use in Cypriote, a dialect closely related to Mycenaean which has been deeply studied by Egetmeyer (2010). 3.3. Eastern influences on Archaic Greece We do not know much about the subsequent Greek “dark ages”, usually posited between 1100 BC, at the beginning of the iron age, and 750 BC. The crisis of the Mediterranean may have temporarily reduced the contact between the Greeks and the East. This contact, however, resumed at the end of the dark ages with the 30 Carlotta Viti <?page no="31"?> migrations (starting in the second quarter of the 8 th century BC and lasting until the end of the 7 th century BC), which led to the foundations of Greek emporia and colonies around the whole Mediterranean. Such migrations had different causes - the need for arable land, the search for necessary products for the survival of the community, rivalry between aristocratic families with consequent exile of one of the two parties etc. They were also enabled by previous experiences of the Phoenicians (from Tyre, Sidon, Byblos, etc.), who in antiquity were considered to be the merchants par excellence. At the beginning of the 10 th century BC already, the Phoenicians had settled down in Kition, in southern Cyprus, and in the subsequent centuries they had founded colonies in the Mediterranean up to Tartessos, in Southern Spain (8 th century BC). Although Homer does not speak about the Greek migrations, he makes several mentions of the Phoenician people, described as clever but untruthful (e.g., πολυπαίπαλος “exceeding crafty” in Od.15,419; ἀπατήλια εἰδώς “skilled in wiles” in Od. 14,288, etc.). Subsequent myths represent the Greek migrations as the journeys of Homeric heroes after the Trojan war. A “colony” (ἀποικία) and an “emporium” (ἐμπόριον) have different ties to the founding city. Although the colony is substantially an independent state (e.g., the city does not interfere with the political affairs of the colony), a linguistic and cultural relationship is always maintained between the two. Instead, an emporium has no political independence, it is just a place of exchange around a harbor, where different Greek and foreign groups may interact. The contact with non-Greek people is therefore more intensive in emporia. We can mention the emporium of Al Mina, in Syria (8 th -7 th century BC), and that of Naucratis in Egypt (7 th -6 th century BC). When we talk of exchange, we imply that the Greeks were both the receiver and the donor of material products and ideas. As anticipated, however, Near Eastern civilizations (especially Ancient Egypt and Mesopotamia) were more ancient and advanced than Greek civilization, as the Greeks always acknowledged (see below). The direction of cultural influence was therefore more from the Near East to Greece than the other way round at least until the Classical age. The most evident gift received from the East is alphabetic writing. It is probably not by chance that the earliest Greek emporium, Πιθηκοῦσαι “Ischia” in the Gulf of Naples (founded in 770 BC by Chalcis and Eretria of Euboea and home of a mixed population of Greeks, Phoenicians, and Etruscans) is also the place where we have one of the most ancient documentations of the Greek alphabetic writing: Nestor’s cup (Νεστορος … ποτεριον). Since the 8 th century BC, Greek alphabetic inscriptions increase and are attested in the whole Mediterranean. In this period, the Greek alphabet is also adopted (in modified variants) by Phrygian (spoken in the Introduction 31 <?page no="32"?> west-central part of Anatolia) and by some Anatolian languages of the 1 st millennium BC, such as Lycian, Lydian, and Carian. In the 7 th century BC, a Western variety of the Greek alphabet (through an Etruscan mediation) also led to the Latin alphabet. Contact with the East also contributed to the end of the Greek Dark Ages; in fact, a subsequent period (from the second half of the 8 th century BC to the end of the 7 th century BC) is commonly called the “Orientalizing epoch”. In this period, cultural Near Eastern motifs (especially from Syria, Assyria, Phoenicia, and Egypt) spread across the whole Mediterranean (also in Ancient Italy, e.g., in Etruria) and stimulated a resurgence of artistic activities. Exchanges of oriental objects could occur on various occasions: sports competitions (the Ὀλυμπιακοὶ Ἀγῶνες were founded in 776 BC according to tradition), votive offerings in temples (visited by both Greeks and Asians. Recall the relationship between Croesus, the king of Lydia, and the oracle of Delphi), banquets, donations, marriages, festivities etc. The aristocratic elites appreciated the acquisition of expensive objects, which could display their high social status and facilitated the establishement of new relationships with peers. With respect to the simple patterns of the proto-geometric and geometric periods, the Orientalizing style (introduced in Corinth at first) is much more elaborated. It presents complex scenes consisting of humans, animals, plants, as well as abstract motifs, with curves, palmettes, and volutes. Some of their referents (palms, lions or panthers) were clearly located in the Near East. Besides art and techniques (e.g., the use of metallurgy for building weapons, as well as engraved gems), the exchanges also concerned less tangible cultural products and ideas. We have seen in §2 that the cults of Aphrodite and Adonis come from the East. Similarly, Artemis had been influenced by the Πότνια Θηρῶν (cf. Hom. Il. 21,470-471 πότνια θηρῶν / Ἄρτεμις ἀγροτέρη), which is related to the Phrygian cult of the goddess Cybele, the mother of living beings, the symbol of wild nature (later borrowed in Rome as the Magna Mater). However, while disciplines such as archaeology and history readily admit Eastern influences on Greek art, Greek literary scholarship is not so keen to discuss external sources. This may be due to certain inhibiting ideological factors, as we have seen in §2. However, the studies of Burkert (1984; 2004; 2009, etc.) and West (1997; 2011; 2014, etc.), among others, have proven that the Eastern influence on Greek literature can no longer be underestimated. For example, the parallels between the Homeric poems and the Gilgamesh epic (cf. George 2003) have been recognized at least since Ungnad (1923), and concern both the description of specific characters and the representation of events, as well as some formulations or the structure of a story. In the Iliad, 32 Carlotta Viti <?page no="33"?> 11 In addition to the above-mentioned masterpieces of Burkert and West, cf. also Gresseth (1975); Beye (1984); Burgess (1999); Foley (2005); Rollinger (2011); Currie (2012); Dar‐ dano (2012); Zgoll (2012); Haubold (2013); Musiał (2013); Bachvarova (2016); López-Ruis et al. (2018); Clarke (2019) etc. the friendship between Achilles and Patroclus evokes that between Gilgamesh and Enkidu. The motif of Gilgamesh’s journey and adventures after the death of his friend - notably a journey to the realm of death - recurs in the Odyssey. In the κατάβασις (Od. 11), Odysseus meets Achilles, among other heroes. Like Gilgamesh, Achilles is often represented as a death and underworld figure. The Gorgon (appearing since Il. 11,36, where she rages in the battle together with Deimos “Terror” and Phobos “Fear”) evokes Humbaba, the monstrous enemy of Gilgamesh. Odysseus’s fight against the Cyclops (Od. 9) also recalls the battle of Gilgamesh against the bull of heaven. Diomedes’ aggression against Aphrodite, who goes to complain to her mother Dione (Il. 5), is reminiscent of Gilgamesh’ aggression against Ishtar, the goddess of love. The encounter between Odysseus and Nausikaa has been compared to that between Enkidu and Šamḫat - the latter brings wild Enkidu in contact with civilization. King Alcinous is similar to Utnapishtim, who lives at the mouth of the rivers, and so on. The literature on the motifs shared between the Homeric poems and the Gilgamesh epos, as well as other Near Eastern legends, is very rich. 11 Similarly, Near Eastern motifs resonate in Hesiod. Hesiod’s Theogony presents similar topics as the Babylonian creation myths of the Enūma Eliš, as well as of the epics of Atrahasis and of the song of Kumarbi and Ullikummi (a Hittite composition which goes further back to a Hurrite model). Hesiod’s Works and Days is an example of wisdom literature, which has numerous illustrious antecedents in Anatolia and in Mesopotamia since the 3 rd millennium BC (e.g., the Instructions of Shuruppak addressed to his son, in Sumerian). Proverbs and fables also have their literary models in the Near East. Lyric poets have experienced journeys to the East. For example, Alcaeus (7 th -6 th century BC) visited Egypt, and his brother Antimenidas served in the Babylonian army (cf. Strabo 1.2.30; 13.2.3). Pre-Socratic philosophers, previously understood to represent the first instances of Greek rational thinking (on the basis of Aristotle’s interpretation), were in fact mainly “shamans”, who presented ideas already found in Eastern traditions about ancient magic, science, and religion. West (1983) has amply discussed the shamanic background of Early Greek religion, as can be seen in Pythagorean and Orphic traditions. In the same vein, the works on nature of Parmenides (6 th -5 th century BC) and of Empedocles (5 th century BC) are not scientific reflections on natural elements, but rather sapiential and mystical poems which, according to Kingsley (1995; 1999; 2003), also present meditative Introduction 33 <?page no="34"?> experiences of incubation (ἐγκοίμησις). The latter was different from sleeping or dreaming; it rather aimed at a state of consciousness in which the seer could enter into contact with the gods and with the underworld. These figures of seers and healers, depositaries of an ancient holistic wisdom, were widespread in the ancient Near East and in Central Asia. Later on, the resumption of Eastern traditions is even more evident in the Hellenistic age, when Alexander the Great (356-323 BC) was called “King of Asia” (βασιλεὺς-δὲ-τῆς-Ἀσίας, Plutarch, Alexander 34,1) and founded Greek cities with mixed populations across Egypt, the Near East and Central Asia. Some dissenting conservative voices exist, but nowadays they are distinctly minoritarian. They contest the tight relationship between Ancient Greece and the Ancient Near East on the basis of two main hermeneutic misunderstandings. Firstly, the identification of a model may be wrongly implied as an undervalua‐ tion of an author or text. According to Dowden (2001: 173), “the influence of the East should not be overstated and it should not be used as a determinist tool to undervalue the level of culture the Greeks ‘arrived’ with, which they found and shaped in Greece, or developed for themselves.” But acknowledging the presence of a model (instead of presenting Greek texts as originated ex nihilo, as in the past) in no way undermines the importance of a work of art. We do not diminish the art of Milton when we recognize an example of his models in Dante, and we do not diminish the art of Dante when we recognize one of his models in Vergil, and of Vergil in Homer. Having models is a natural condition for literary texts, as for any other product of art or science and for human behavior itself. Models provide explanations for intellectual and social entities, and are not to be meant in an evaluative sense. On the contrary, the Ancients themselves enounced their models and were proud of them. When Virgil connects his epic to Homer, he wants to reflect the nobility ascribed to the Homeric poems. See below for further considerations on the right way to interpret ancient imitation. When we say that something is a model, this is to mean as temporally precedent, and not as artistically superior. Sometimes the Ancients appreciated more the target than the source of an influence. We only have fragments of Ennius, for example, because Roman grammarians considered him to be surpassed by Vergil. Secondly, the identification of an external model may be wrongly considered to be mutually exclusive of an inherited pattern. Parsimony is again the motivation for this assumption. For example, Zsolt (forthcoming: 5) excludes the possibility that certain phraseological motifs (e.g., the expression “black earth”) may have been transmitted from the Anatolian languages (Hittite or Luwian) into Homeric Greek when they have parallels in other IE languages. In this case, an explanation in terms of contact is considered to be unnecessary, and the 34 Carlotta Viti <?page no="35"?> patterns at issue are ascribed to the inherited PIE tradition. An author, however, may have different types of sources simultaneously, and an inherited model may be reinforced by the presence of an external one. For example, the Homeric hero does not only reflect the representation of some Near Eastern heroes such as Gilgamesh, but is also rooted in IE poetic motifs which reverberate in Arjuna in the Mahābhārata, in Beowulf etc. (cf. Nagy 2005). A certain poetic formulation may be attested in other early IE languages, and therefore be considered to be inherited from PIE, and at the same time sustained by contact with an external source. In other words, among the vast inherited poetic repertoire, some expres‐ sions may be favoured as they find correspondences in some outside traditions to which a poet may be exposed. I do not imply that an explanation in terms of contact is always appropriate - this must be evaluated case by case. I rather suggest that in principle we cannot rule out the possibility of multiple literary and cultural models. As with any historical explanation, the interpretation of a literary motif does not have to be necessarily simple or economic. It does have to be necessarily precise and exhaustive, which is compatible with the presence of multiple poetic impulses. This is something similar to what we have observed in §1 as regards contact-induced grammaticalization, where internal and external motivations interact, and where a construction may be the result of an internal language change and at the same time may be reinforced by language contact. I see striking analogies between contact-induced grammaticalization and a comparative literature where inherited motifs are often also combined with external literary sources. 3.4. The Classical ideological revolution If we browse through a handbook of Greek history and civilization (e.g., Burckhardt 1898-1902 or Murray 1993) in the light of the relationship with the East, we notice that the Classical age (5 th -4 th century BC) is the only period in which this Eastern influence is less apparent - I say “less” because contact never stopped, but it was less recorded in our sources. This is essentially because of the Persian Wars (499-449 BC). At the end of the 6 th century BC, the increasing power of the Achaemenid Empire and its occupation of the Greek colonies in Asia Minor had rendered the trade of Eastern products more difficult (this was also a determinant of the end of the Orientalizing epoch). During the Persian Wars, as many Greek cities were destroyed or ravaged (Athens was set on fire by the Persians in 480 BC) and the Greek population was decimated, Greece saw the emergence of new ethnic ideas, which may be compared, mutatis mutandis, with certain nationalist thoughts of the 19 th century illustrated above. The Persian Wars were now reinterpreted as a cultural opposition between a free Greece Introduction 35 <?page no="36"?> and a tyrannical Asia. Actually, this was an untrue picture. On the one hand, numerous Greek poleis, Athens included, had experienced tyrannies as well in the recent past, during the 7 th and 6 th century BC (in Athens, the tyranny of Hippias, Pisistratus’s son, ended in 510 BC, just eleven years before the beginning of the Persian Wars. Sparta was rather the exception for not having had tyrants in this period). Indeed, the figure of the τύραννος m. “absolute ruler” (a borrowing from Pre-Greek according to EDG 1519) was not connotated negatively in origin - he was often a defender of the people against the abuses of the aristocracy, and some tyrants, such as Pittacus of Mytilene (c. 640-568 BC) and Periander of Corinth (before 634 - c. 585 BC) were even included among the Seven Sages. The negative image of the tyrant rather emerges in Classical Greece, from which it has been transmitted to our modern political lexicon. On the other hand, the Persian empire granted considerable freedom to its subjects to express their customs and religions (the Persian empire mainly expected taxes from them), and was culturally extremely advanced (cf. Briant 1996). A notable outcome of the Persian Wars was the changed image of the Asian and of the foreign in general. Besides the representation of the tyrant, we can identify this ideological change if we follow the semantic change of particular keywords denoting foreigners and enemies on the basis of their etymology and their earliest occurrences in texts. Originally, the word βάρβαρος is an onomatopoetic formation which simply means “foreign” in the sense of “non-Greek speaking” and has no negative connotations. This word is not attested in Homer but is clearly presupposed by the Homeric expression βαρβαρό-φωνος “speaking a non-Greek language”, referred to the Carians in the catalogue of ships (Il. 2,867). Later, it often describes the Medes and the Persians (cf. LSJ s.v.), and after the Persian Wars it is commonly used in the sense of “brutal, rude, savage”. This negative connotation eventually also affects the linguistic meaning, and the substantivized use τὸ βάρβαρον may denote a “bad Greek”. The image of the enemy itself changed. In Homer, the Trojans and their Asiatic allies speak the same language as the Greeks, believe in the same gods, have the same manners of conducting a fight or a council, the same habits in celebrations, funerals, clothing etc. Their behaviors, their feelings are represented exactly like the behaviors and feelings of the Greeks, and the reader is profoundly moved while reading of Hector who bids farewell to his wife and his only child on the Trojan walls, conscious of his impending death. Indeed, Homer represents Hector as more human than Achilles (cf. Schadewaldt 1965: 177), and in general seems to be more sympathetic towards the Trojans than the Greeks, whose leader, Agamemnon, is especially arrogant and unjust. In the same 6 th book of the Iliad, before the 36 Carlotta Viti <?page no="37"?> 12 Besides, in only one passage of the Odyssey (14,102), ξένος means “hireling” and refers to a foreign laborer who takes care of the cattle. 13 A similar path of change is attested for Latin hostis m., which originally meant “stranger, foreigner, guest”, with no negative connotations, as its IE cognates (Gothic gasts, Old Church Slavonic gostĭ, etc.). Only later, as the sense of “guest” had been expressed by hospes, -itis m. “guest, visitor; host, entertainer” (a compound from hostis + potis “having power”, cf. Russian gospód’ “the Lord, God”), hostis acquired the value of stranger in the negative sense “enemy” (cf. LS, OLD s.v. hospes, hostis; EM 301). The Romans were aware of this semantic change, cf. Varro L.L. 5,3 et multa verba aliud nunc ostendunt, aliud ante significabant, ut hostis: nam tum eo verbo dicebant peregrinum qui suis legibus uteretur, nunc dicunt eum quem tum dicebant perduellem. ὁμιλία between Hector and Andromache, we have another meeting, between the Greek champion Diomedes and the Lycian soldier Glaucus. Although they fight on opposite sides, they realize that they share ancient ties of hospitality (ἦ-ῥά-νύ μοι ξεῖνος πατρώϊός ἐσσι παλαιός “But really you are an ancient guest-friend of my father’s house”, Diomedes says). Instead of fighting, they therefore exchange their sets of armour. Note the word ξεῖνος (epic and Ionic form of ξένος) in this passage. In the Iliad, this form only means “guest-friend”, with positive connotations of alliance and loyalty (to the point that Paris’ infringement of a tie of hospitality is presented as the cause of the Trojan war). In the Odyssey, we additionally find the meaning “stranger, wanderer, refugee” who has to be protected. This meaning may be equally ancient - its use in the Odyssey (often considered to be more recent than the Iliad) may be due to contextual factors, as the journey is the main topic of the second epic. For example, in the 9 th book of the Odyssey, the Cyclops asked: Od. 9,252 ὦ ξεῖνοι, τίνες ἐστέ; πόθεν πλεῖθ᾽ ὑγρὰ κέλευθα; “Strangers, who are you? Where do you sail from? ”, and Odysseus replies that they are Greeks and begs for hospitality and protection. Od. 9,270-271 Ζεὺς δ᾽ ἐπιτιμήτωρ ἱκετάων τε ξείνων τε, / ξείνιος, ὃς ξείνοισιν ἅμ᾽ αἰδοίοισιν ὀπηδεῖ “Zeus is the avenger of suppliants and wanderers, (Zeus) protector of strangers, who goes together with the reverend wanderers”. 12 Only after Homer, in the lyrics and especially in the Classical age, ξένος is used in the sense of “foreigner” as opposed to natives and citizens with negative connotations (cf. LSJ s.v. ξένος). 13 Even later, since Classical Greek, we find the adjectival use as “strange, unusual” or “ignorant of something” (i.e., strange to something). After the Persian Wars, the enemy was evil, and Asia was the enemy. Now, if the Classical era was the least touched by Eastern influence, as compared to the archaic period on the one hand and to the Hellenistic period on the other, then the very term “classical” probably needs to be interpreted Introduction 37 <?page no="38"?> 14 Cf. e.g., Gellius 19.8.15, who recommend to check the best authors to find evidence for correct grammatical usages (in this case, for the use of quādrīgae f. “set of four horses” in the plural and of harēna f. “sand” in the singular): ite ergo nunc et, quando forte erit otium, quaerite an ‘quadrigam’ et ‘harenas’ dixerit e cohorte illa dumtaxat antiquiore vel oratorum aliquis vel poetarum, id est classicus adsiduusque aliquis scriptor, non proletarius. Note the opposition between classicus “of the highest rank” and proletarius “low, common” to refer to literary models. not as “superior”, as it was commonplace in Roman reception, 14 but as “purer, polished” from foreign influence. That is, a period may have been selected to represent the model of Greek civilization not for its intrinsic value in language, literature, art or culture in general, but for its more homogenous, native element, freer from variation implied in linguistic and cultural contact. This may lead to a drastic revision of the Graeco-Roman reception. Concepts such as reception and classics entail considerable ideological implications. Not by chance, it is in the Classical age, in Herodotus’ Histories of the Persian Wars, that we have the first explicit statement of what it is to be “Greek” (τὸ Ἑλληνικόν): “having the same blood and the same language” (ἐὸν ὅμαιμόν τε καὶ ὁμόγλωσσον), “as well as common shrines of the gods and sacrifices” (καὶ θεῶν ἱδρύματά τε κοινὰ καὶ θυσίαι), “and similar ways of life” (ἤθεά τε ὁμότροπα, Herodotus 8,144). Note the connection between language and blood - this is something very similar to the concept of “nation” which we have seen in §2. This passage appears at the end of the 8 th book of the Histories, in one of the most celebrated political speeches in Greek literature, on which it is appropriate to dwell a moment. We are at a crucial point of the Persian Wars, with the Persians seeking to divide the Athenians from the Spartans. They therefore send ambassadors to Athens to propose a cessation of the fighting with the Persians, as the latter had a much more powerful army and would certainly win the war, and advised them to accept an alliance. As the Spartans are informed of this proposition, they also send ambassadors to the Athenians to warn them not to join the Persians. On the one hand, the Athenians reply to the Persians that they recognize their military superiority but refuse the proposal of an alliance (8,143 ἀλλ᾽ ὅμως ἐλευθερίης γλιχόμενοι ἀμυνεύμεθα οὕτω ὅκως ἂν καὶ δυνώμεθα. ὁμολογῆσαι δὲ τῷ βαρβάρῳ μήτε σὺ ἡμέας πειρῶ ἀναπείθειν οὔτε ἡμεῖς πεισόμεθα “Nevertheless, longing for freedom, we will defend ourselves the best we can. But as to come to an agreement with the barbarian, do not try to persuade us, nor will we consent”). On the other hand, the Athenians reply to the Spartans that it is a shame even to think that they could abandon their common intent, firstly because they still have to avenge the Persian sack of Athens, and secondly because they share kinship in blood and speech with the other Greeks etc. The 38 Carlotta Viti <?page no="39"?> 15 This is not the only reason, of course, for the consideration of ancient authors in the subsequent reception. In the case of Empedocles, the term “pre-Socratic” or “early phi‐ losopher” mainly serves to set him apart from later Classical philosophers such as Plato and Aristotle. Note, however, that Pre-Socratic philosophers also include Democritus (c. 460 c. 370 BC), although he was younger than Socrates himself (c. 470-399 BC). Thus, factors other than chronology are here at stake. A further motivation may be seen in the geographical setting of certain pre-Socratic philosophers afar from Athens (e.g., Parmenides and Empedocles were both operating in Magna Graecia). These factors are not mutually exclusive.- whole passage is centred on the opposition between expressions of slavery related to the Persians (here called tout court “Barbarians” - βάρβαρος, in the singular) and expressions of freedom concerning the Greeks. The idea of an alliance with the Persians is presented as equivalent to reducing Greece itself in slavery: μηδίσαντες καταδουλῶσαι τὴν Ἑλλάδα (there is nothing on earth for which) “we would take the Persian part and enslave Greece” (8,144). I find it also significant that some Greek authors who chronologically belong to the Classical age are traditionally treated as archaic or pre-Classical. We have just seen an example in Empedocles, who was born in c. 494 and died in c. 434 BC, that is, during the Persian Wars. He was even younger than Aeschylus (c. 525/ 524 - c. 456/ 455 BC). Still, Empedocles is conventionally included in the heterogeneous group of the “Pre-Socratics” (a term going back to the late 18 th century but popularized by Diels) or “Early Greek philosophers”, while Aeschylus is considered to be Classical. This may have also 15 to do with the fact that Aeschylus was much more engaged, in both his personal life and his works, with the new nationalistic topics brought about by the Persian Wars. In his tragedy Πέρσαι (performed in 472 BC), Atossa, the Persian queen mother, is informed by a messenger of the defeat of her son, the emperor Xerxes, in the naval battle of Salamis, which is described in details - Aeschylus himself had fought against the Persians at Salamis (480 BC), as well as in Marathon (490 BC), where his brother fell in the field of battle, and in Plataea (479 BC). The tragedy polarizes the difference between Xerxes’ ὕβρις and the Greeks’ quest for freedom. For the Ancients (at least for those not living during the Persian Wars) having foreign, eastern influences was not a problem. This is because they had a completely different approach to their sources. Nowadays, authors and artists strive to show that they are innovative. External influences must be recognized for intellectual honesty but they must be limited, as any piece of art, literature, scholarship etc. is assumed to say something new. This attitude is a relatively late phenomenon, which may ultimately go back to the search for the πρῶτος εὑρετής “first discoverer, inventor” in the Classical age. The noun εὑρετής is Introduction 39 <?page no="40"?> 16 “Die Empfindung, daß die Dinge als solche εὑρήματα seien, kommt erst nach Homer durch Kombination mit erschlossenen εὑρεταί zum Ausdruck. Betrachten wir z. B. die Ilias und die Odyssee, so können wir nur eine völlige Ignorierung von πρῶτοι εὑρεταί feststellen.” (Kleingünther 1933: 4). attested since Plato and Isocrates (cf. LSJ s.v.). Apart from isolated previous statements about who may have found something (οἳ πρῶτοι … εὗρον), the whole issue of identifying the discoverer or inventor of something appears as a literary τόπος at the beginning of the 4 th century. As Kleingünther (1933) observes, there are no traces of this enquiry in Homer, as the earliest authors usually ascribed the creation of things and ideas to the gods. 16 This new inquiry, in my view, may be also considered as one of the intellectual products of the Greek victory over the Persians - as we have seen, this event had an enormous impact on Greek material life and ideological repertoire. As civilized Greece had overcome the barbarian Persian (so the argument was), then the Greeks could be emancipated from external influences and search in their own history the ingenious inventors and discoverers of techniques (in case they attributed a discovery or invention to a non-Greek, they presented that innovation as a myth somehow related to their Greek ancestors). This was also compatible with the new contrast between the main victors of the Persian Wars, Athens and Sparta, which led to the Peloponnesian War (431-404 BC). Athens’ democracy was considered to be “the school of Greece” (cf. Pericles’ boastful funeral oration pronounced for the war dead at the end of the first year of the Peloponnesian War: τήν τε πᾶσαν πόλιν τῆς Ἑλλάδος παίδευσιν εἶναι, Thucydides 2.41). Athens could therefore assume a prime advantage in choosing its artistic and literary models and in identifying the first discoverers and inventors. It is understandable that the query about the πρῶτος εὑρετής was so vital in Athens (particularly in the school in the Sophists) rather than in other poleis. Previously, instead, authors preferred to attribute their ideas to earlier recognized models in order to give them more authority - to the point that often a celebrated author is ascribed works that he did not actually compose. This is precisely the opposite of plagiarism. The Homeric hymns, for example, are traditionally ascribed to Homer but in reality go back to different authors living in different epochs, from the archaic age to Hellenistic times. The name of Homer himself comprises generations of ἀοιδοί, and much of Alexandrian philology was devoted to distinguish what is original (in the sense of genuine) and what is spurious in Homer, what lectio has to be considered to be more authentic, precisely because at that time there was a wealth of manuscripts transmitting different versions of the Iliad and of the Odyssey which were all 40 Carlotta Viti <?page no="41"?> 17 The verb μιμέομαι is attested once in the Hymn to Apollo, which is however traced back only to the 6 th century BC (cf. Càssola 1975: 101). The poet describes the “big marvel” (μέγα-θαῦμα) of girls of Delos who praise Apollo and sing. “Recalling the men and the women of the past, they sing a hymn, and charm the humans’ tribes. And they can imitate the tongues of all humans and their clattering speech: each would say that he himself were singing, so their beautiful song is conforming.” (μνησάμεναι ἀνδρῶν τε παλαιῶν ἠδὲ γυναικῶν / ὕμνον ἀείδουσιν, θέλγουσι δὲ φῦλ᾽ ἀνθρώπων. / πάντων δ᾽ ἀνθρώπων φωνὰς καὶ βαμβαλιαστὺν / μιμεῖσθ᾽ ἴσασιν: φαίη δέ κεν αὐτὸς ἕκαστος / φθέγγεσθ᾽: οὕτω σφιν καλὴ συνάρηρεν ἀοιδή). Note the onomatopoetic hapax βαμβαλιαστύς f. “chattering, stammering”, which may allude to the incomprehensible sounds of foreign languages, similarly to βάρβαρος. ascribed to Homer. In Rome, one of the earliest philological exercises can be seen in Varro’s attempt to identify 21 genuine comedies by Plautus, as many more spurious versions existed, and so on. It is true that authorship is an ancient phenomenon - the first authored composition, to my knowledge, is that of Enheduanna, the daughter of the Akkadian king Sargon I (second half of the 3 rd millennium BC) in her hymn to the goddess Inanna. In Greece, the σφραγίς “seal, signet” has been used since the archaic age. But the matter of signalling the author has not been always equally important. Σφραγῖδες are especially common in the Hellenistic and Roman periods. The very concept of imitation may be misleading when related to antiquity. The Ancient Greek name of imitation is μίμησις, an action noun derived from the verb μιμέομαι “to imitate”, which in its turn is a denominative verb from μῖμος m. “mime” (a word with no convincing IE etymology. EDG 954-955 suggests a Pre-Greek origin). Neither μίμησις nor μιμέομαι or μῖμος are attested in the Homeric poems or in Hesiod. 17 They are Classical words, mainly occurring in the theatre and in philosophical works (cf. LSJ s.v.). Secondly, even in Classical Greek, they are not used in the sense of “imitation of previous authors”, but rather in that of “representation of nature”, as in Plato and in Aristotle. Auerbach’s (1946) book Mimesis has therefore the appropriate subtitle Dargestellte (and not nachgeahmte) Wirklichkeit in der abendländischen Literatur. Plato (Republic, 10 th book) presents the example of a bed, which may have three types of increasingly indirect representations, that is, the idea of a bed created by the god, the material bed created by the carpenter on the basis of that idea, and the picture of a bed created by a painter who imitates the concrete object. Like a painter, the poet’s representation is far removed from the truth. The use of μίμησις as imitation of previous authors is relatively late in antiquity - it mainly goes back to the Περὶ μιμήσεως by Dionysius of Halicarnassus, in the Augustan period, a fragmentary treatise discussing the best literary models to be imitated. Note that Dionysius belonged to Atticism and therefore aimed to Introduction 41 <?page no="42"?> 18 In the Indic tradition, the śrútiis opposed to the smṛ́tilit. “remembrance, reminiscence” (from the root smr̥ “remember”), referring to the body of knowledge that is transmitted memoriter by human teachers (Vedāṅgas, Sūtras, Itihāsas, Purāṇas, etc.). As smr̥ denotes a mental process, and not the writing, the more profane literature of the smṛ́tialso implies an oral transmission, similarly to the śrúti-. Within the IE domain, Indic is probably the branch where the orality has been longer maintained as a main mode of transmission. Although the Brahmi (from which the Devanāgarī developed) has been established at least since the 3 rd century BC, the writing has been usually considered as an aide-mémoire and has not replaced the traditional teaching of the Veda based on repetition and memorizing. imitate the elegant style of Classical Greek authors, far from the varietas of Asianism as well as from more vulgar phrasing of the Koiné (cf. Jonge 2008: §1.3). This meaning of μίμησις therefore presupposes an established philological practice and a concept of literature and classics. Thirdly, owing to the original connection of μίμησις with mimes and theatre, in this word I see the connotation of a visual replication. The same holds true for its Latin functional correspondent imitātiō f. “imitation”, a nomen actionis from the verb imitor, -āri “to imitate”, the latter being a frequentative formation (from a non-attested verb *imō, -āre) etymologically cognate with the Latin noun imāgō, -inis f. “picture, image”. Now, a visual representation implies reading and writing. In Early Greek, however, before the Classical age, literary works mainly had an oral transmission. This is better reflected in the Vedic term śrútif. “hearing, listening” (from the root śru “hear, listen”), referred to the Veda (cf. MW s.v. śrúti- “that which has been heard or communicated from the beginning, sacred knowledge orally transmitted by the Brahmans from generation to generation”). 18 An oral transmission is much more fluid than a written text, and is not considered as an imitation or reproduction, but rather as the new performance of a song. When an ἀοιδός of the archaic age sings about the Trojan war or Odysseus’ journeys, he is not copying Homer - he is singing Homer. In fact, he can well insert some variation into his traditional repertoire (this also depends on the audience), and can even improvise. The Doloneia in the 10 th book of the Iliad, for example, is a later addition. The seminal studies of Parry (1971) and Lord (1960; 1991) on South Slavic guslar have elucidated the strategies behind an oral composition. Nagy (1996), who has elaborated Lord’s theories and is the author of numerous interesting studies on Homer and on Greek poetry in general, also stresses the importance of an oral composition as a dynamic process. He compares it to a bird which is continuously changing its song on the basis of its limited repertoire. There is tradition and also innovation in this, and creativity always proceeds within the lines of the tradition. In this case as well, a parallel from Indic is in order. Vedic poets often say that they 42 Carlotta Viti <?page no="43"?> 19 Cf. LS and OLD s.v. littera as well as von Kamptz’ entry in Th.L.L. 7,2 (especially 1522,18-1524,34; 1524,83-1525,52). 20 Although Latin littera “letter” presents semantic calques of Ancient Greek γράμμα in its various meanings of “written sign”, “epistle”, etc., neither Latin pl. litterae has the same use as Ancient Greek pl. γράμματα nor Latin litterātūra has the same use as Ancient Greek γραμματική (as implied instead in EM 363). In Latin, the meaning of “literature” is common and attested early for these terms. Instead, Ancient Greek γράμματα mainly means “set of written characters”, “documents, public records”, “laws, rules” or “book, treatise” (the latter since Plato), and only marginally indicates “letters, learning” (again in Plato), cf. LSJ s.v. Similarly, the meaning “literature” is definitely not typical of Ancient Greek γραμματική (attested since Classical Greek, e.g., in Plato), which rather means “grammar” (often with τέχνη) or, less commonly, “alphabet, script”. In English-Ancient Greek dictionaries (cf. Woodhouse 1910: 495), at the entry “literature” we find μουσική f., which however is properly quite another thing, that is, “anything presided by the Muses”, and especially poetry sung to music (cf. LSJ s.v.). In the specific sense of “art or letters”, μουσική is only used since the Classical period, especially in Plato, and remains rare. It predictably decays in later varieties of Greek - in Modern Greek, “literature” is βιβλιογραφία, λογοτεχνία or φιλολογία. 21 Although the extremely rich Old Indic vocabulary has various possibilities to denote literary texts, none corresponds to the modern sense of “literature”. Vidyā́f. (from the root vid “to know”) rather means “body of knowledge” in general, including philosophy praise the gods “with new songs” (návyasībhir gīrbhír, Rig-Veda 6,49,1) and that they are aided “by an ancient prayer” (pratnéna mánmanā, Rig-Veda 8,6,11). This apparent contradiction is perfectly intelligible in the context of an oral transmission, where the old content revives at every new performance. In the context of an oral transmission, many concepts of modern text criticism, such as “classic’, “imitation”, “plagiarism”, “genre” or “literature” appear anachronistic when related to the Early Greek linguistic and cultural tradition. I think these concepts also presuppose the writing as a main mode of transmission, as well as an established philological practice. These conditions are met in Hellenistic schools of philology and especially later in Rome. I find it significant that Latin litterae f.pl. “literature, written monuments, records” (attested in this sense since Cicero already, and also meaning “sciences or learning in general, liberal education, scholarship”, as well as “letters in the sense of epistles; edict, ordinance”, etc.), 19 is transparently based on the noun of the written sign, littera. The derivation litterātūra f. “writing formed of letters; science of language, grammar, philology” passed by inheritance (e.g., French littérature) or borrowing (e.g., English literature) to the various modern languages of Europe. Similarly to reception, literature as well implies ideological considerations. Ancient Greek does not have a word for “literature” as we understand it today, as a collection of written works that represent a form of art. 20 Neither is there a word for it in Sanskrit. 21 Where the highest models Introduction 43 <?page no="44"?> and science, and is preferable referred to the knowledge of the Vedas. Śāstrán. means “order, command, precept, rule” in Vedic, and in Classical Sanskrit indicates any instrument of teaching or advice - it is therefore mainly connotated in didactic sense, and can also refer to the Veda. Akṣáran. (lit. imperishable) properly denotes the “syllable” (which in Indian thought has a mystical value), and later also means “vowel, sound, word, or letter”. Sāhityan. (vr̥ddhi from sahita-, from sa- “together” and dhā “to place, set”) lit. means “association, combination”, as well as “society”, and only marginally may have the meaning of literary composition. Áṅgameans “limb of the body” and metaphorically “division of science”; it usually has a technical sense. We are therefore dealing with highly polysemic lexemes, which only in some marked contexts may indicate a body of literature. Alternatively, to express more specific meanings, Classical Sanskrit may recur to compounds such as kāvyaśāstran. N. of a short work on poetics. - 22 Cf. Hom. Il. 6,168-169: πόρεν δ᾽ ὅ γε σήματα λυγρὰ / γράψας ἐν πίνακι πτυκτῷ θυμοφθόρα πολλά “And he (scil. king Proetus) gave him (scil. to Bellorophon) baneful signs, having written many life-destroying (words) on a folded tablet”. Proetus had asked Bellerophon to bring these written signs to his father-in-law, who was supposed to kill Bellerophon. The passage is inserted in Glaucus’ speech in the above-mentioned encounter with Diomedes. (Homer, the Veda) are orally transmitted, there is no need for that. Although the Greek alphabetic writing has been attested since the early archaic age (8 th century BC), and although one single vague hint at its use also appears in Homer, 22 it was not widespread until the Classical age. Even many Classical Greek texts, such as drama (tragedies and comedies) and oratory, as well as Herodotus’ histories, continue to be intended for an oral performance. Even Plato, where we have often found new technical senses for old words, chooses to write διάλογοι “conversations”. While oral transmission permits a continuous possibility of variation and change, the writing definitively fixes one version. Establishing the interrelationship of the different sources of oral transmission is an extremely complicated philological issue, as can be seen in the long diatribes of the Homeric question (Analyse, Unitarismus, etc.). Oral traditions are resistant to being captured in the stemma codicum, as they imply a variation of the sources, similarly to what we have seen above for the variants occurring in language contact that are not easily amenable to the schematism of the Family Tree model. Similar patterns of variation and change may run through comparative literature and comparative linguistics. 4. Presentation of this volume After having set the historical and cultural context, I would like to say few words on the present volume, which mainly discusses language variation and language contact in Latin and Greek and related topics. This aims to be a contribution to 44 Carlotta Viti <?page no="45"?> the existing literature on historical language contact in general linguistics, as well as in some recent interesting works on language variation in the Greek and Roman world (cf. e.g., Clackson et al. 2020; Bianconi et al. 2022). In addition to this introduction, the volume consists of fifteen papers, presented by scholars from different universities and different countries. Most of them are written in English, but French and German are also represented, which is consistent with the subject of multilingualism here investigated. They are arranged in thematic order. The first four papers analyze aspects of Greek dialects and of ancient Greek inscriptions. Monique Bile, René Hodot and Guy Vottéro (three leading specialists in Greek dialectology working at the University of Lorraine) present a joint paper entitled Le grec ancien: une réalité multiforme, in which they investigate the variety of Greek dialects in their historical and geographical framework in the ancient Mediterranean. They take into consideration the various archaeological, epigraphic, and papyrological sources, and discuss the most relevant phonetic, morphological, syntactic, and lexical features of the Greek dialects. They rightly point out that, despite their dialectal variety, the Greeks were conscious of speaking the same language and of belonging to the same ethnic group. This dialectal variety was also exploited for literary purposes, as different dialects tended to be associated to different literary genres. Even after the Roman conquest, the authors observe, Greek continued to be used in different sociolinguistic contexts, as the language of an illustrious body of literature, as a political and administrative language in the Eastern part of the Roman Empire, as the language of the Christian Church (and after the schism, of the Orthodox Church), as a vernacular language etc. A knowledge of the Greek dialects is therefore essential to obtain a profound knowledge of the Greek language itself, as well as of PIE, owing to the importance of Ancient Greek in linguistic reconstruction. A number of new epigraphic findings also expand our knowledge of Ancient Greek society. For example, the authors mention the word ἀρχερμηνεύς “chief translator” (from Kolossai in Phrygia). The following paper, La signature des hiéromnémons, is written by Emmanuel Weiss, a French scholar of Greek and Latin linguistics who has also recently organized a conference on the Doric dialect at the University of Lorraine in October 2022. E. Weiss analyzes some bronze tablets from Epizephyrian Locri, a colony founded by Locris (North-West Greece) in Magna Graecia (today’s Calabria) in the middle of the 7 th century BC. The corpus of these tablets (dated to the end of the 4 th century BC) is linguistically heterogeneous, as its main Doric component attests interference with the Greek Koiné. The author studies in particular the variant forms of the names of the “hieromnemons”, the sacred secretaries Introduction 45 <?page no="46"?> (from ἱερός “sacred, holy” and μνήμων “remembering, reminding; counsellor, recorder”) who were in charge of administrating the treasure of the temple of Zeus. He observes that the different epigraphic and linguistic representations of their names indicate different degrees of penetration of the Koiné in Locri - an interesting example of the manner through which linguistic variants can provide information on the sociolinguistic context of the speech community. A variety of North-West Greek was spoken at Dodona, in Epirus, the home of one of the most important Panhellenic sanctuaries. The language attested in the tablets of this sanctuary is precisely the subject of Georgios K. Giannakis’ paper The Magna Graecia tablets in the Dodona corpus. Giannakis is a Greek scholar working at Aristotle University of Thessaloniki, especially interested in the Greek dialects and in IE linguistics. He is also the Editor-in-Chief of the Encyclopedia of Ancient Greek language and linguistics. Giannakis observes that, owing to its Panhellenic character, the sanctuary of Dodona attracted visitors from many other places of mainland Greece as well as of Magna Graecia. Accordingly, the oracular texts, written on lead tablets, often present phenomena of interference between two or more dialects. He analyzes the form (in both its linguistic and epigraphic aspects) and the content of these tablets, which are both private and public, and which deal with a variety of issues ranging from financial or professional business to family problems, marriage, childbirth, migration etc. The author observes that, despite the geographically peripheral position of Epirus, the sanctuary of Dodona attests a vibrant multilingual environment. This shows how areal factors have to be integrated alongside sociolinguistic factors in order to account for a language variety. The next paper, From the files of a new Bechtel, Die historischen Personennamen des Griechischen, is written by José Luis García Ramón, a scholar of IE studies who has recently worked at the Università del Sacro Cuore in Milan. Among his many research projects, García Ramón is working on a revised version of Bechtel’s study of Greek personal names. Bechtel (1855-1924), who was also a specialist of Greek dialects, wrote before the discovery of Mycenaean (deciphered by Ventris only in 1952). Several analyses of Bechtel therefore need a revision. In this paper, García Ramón analyses in particular the variants of the name Ἀλεξιτέλης, which is interpreted as “the one who guards his company” or “the one who wards off the fatal doom”. This interpretation is achieved by discussing IE etymology in the light of Mycenaean as well as of poetic diction attested in Homer and in the lyrics. In addition to the elucidation of this specific case study, this is also relevant from a methodological point of view, as Greek personal names, and Greek lexemes in general, may be etymologically explained by reference to phraseology. 46 Carlotta Viti <?page no="47"?> The two subsequent papers discuss certain aspects of the relationship be‐ tween Ancient Greek and the Anatolian languages. The hypothesis of an Anatolian-Greek language area is quite controversial. Cotticelli-Kurras (2021) has observed that we can talk of some linguistic interference between Western Anatolian and Ancient Greek but not of an authentic language league, as some defining features of this putative language area are typologically common, and some others are shared only by part of the languages of the area. With this background, Valerio Pisaniello (from the University of Chieti-Pescara) and Stella Merlin (from the University of Naples) present an enriching paper entitled Ancient Greek and late Anatolian languages: from coexistence on the territory to survival in the scholarly tradition. Pisaniello is specialized in IE studies and in the linguistics and philology of the ancient Anatolian languages. Merlin is a specialist of historical linguistics and comparative philology. The two authors investigate various phenomena of linguistic interference between Ancient Greek and the Anatolian languages of the 1 st millennium BC, such as Lycian and Lydian. They analyze the language as well as the material aspects of both monolingual and bilingual inscriptions (funerary texts, votive texts, decrees, etc.), which are often difficult to interpret, as Lycian and Lydian are Restsprachen. For this, they also avail themselves with information attested in the glosses of later Greek literary and lexicographic sources, where “Lydia” was often generically used to denote the Near East, often with connotation of luxury and opulence. Their analysis of the epigraphic evidence (e.g., of onomastics) in Greek, Lycian, and Lydian casts insights on the highly multilingual texture of I millennium’s Anatolia. On the other hand, Filip De Decker (a Belgian scholar, from the University of Ghent and the University of Verona, specialized in Greek and Latin linguistics) presents a case study where an explanation traditionally suggested in terms of language contact between Ancient Greek and Anatolian does not hold, in his paper entitled The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? An analysis of the aspect use in the Epic-Ionic -σκ-iteratives in the Odyssey. Although this volume is especially concerned with contact, one has also to identify cases in which contact did not occur. For Epic-Ionic -σκ-iteratives, De Decker argues that an internal explan‐ ation is preferable. While the Hittite -šksuffix is essentially imperfective, the Epic-Ionic -σκsuffix often also has different aspectual functions of perfectivity. He therefore analyzes a large corpus from Homer’s Odyssey in the light of the linguistic literature on tense, aspect and modality. The next two papers discuss inscriptions in Early Latin. Rex Wallace, an American scholar from the University of Massachusetts Amherst, is the author of the paper Comments on an Old Latin inscription from San Giuliano (CIL I2, Introduction 47 <?page no="48"?> 2780). Wallace has published widely on historical and comparative linguistics, especially on the languages of Ancient Italy (see, for instance, his 2007 book The Sabellic Languages of Ancient Italy and the 2018 book of Anthony Tuck and Rex Wallace entitled The archaeology of language at Poggio Civitate). In this paper, he studies a Latin inscription of the 3 rd century BC found in San Giuliano, in the Etruscan territory. He analyzes the epigraphic and linguistic characteristics of the inscription, which indicate the presence of Latin speakers in the area at that period. Luca Rigobianco, an Italian scholar working at the University of Venice, is the author of the paper entitled Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? Rigobianco is especially interested in Latin and in the fragmentary languages of Ancient Italy as well as in their linguistic contact. In this paper, he studies the development of vowel reduction and deletion in non-initial syllables in Early Latin. Rigobianco considers both the position of the stress and the rhythmic tendency, and makes comparisons with data from Sabellic and from Etruscan in order to establish the directionality of the phonetic change. The following two papers are devoted to Latin comedy, especially developed in Archaic Latin between the second half of the 3 rd century BC and the first half of the 2 nd century BC. Latin comedy intrinsically presupposes a relationship with Greek comedy, as the plots (and often also the names of characters and places) of the palliata are mostly drawn from the Νέα κωμῳδία of the Hellenistic age. Since, however, the Greek original is often lost or fragmentary, it is difficult to establish to what extent the Latin authors followed their source or innovated. This problem is discussed in Gualtiero Calboli’s paper Nausistrata (Ter. Phorm. 784-1055). Calboli, an Italian scholar from the University of Bologna, is a classicist and a linguist; he is the editor of the Journal of Latin Linguistics, and treated the philological transmission of Cato and of the Rhetorica ad Herennium. In this paper, he focuses on Nausistrata from Lemnos, a character of Terence’s Phormio, which is based on the Ἐπιδικαζόμενος, a lost comedy of Apollodorus of Carystus, a pupil of Menander whose floruit is posited in Athens in the first quarter of the 3 rd century BC. On the one hand, Calboli observes that the figure of Nausistrata and her story must already have been in the Greek model. This is because Nausistrata presents specific pieces of information about the economic situation of Lemnos, a colony of Athens, which would have been impossible to access for somebody living in Rome one century later. A detail on a Greek island turns out therefore to reveal the right interpretation in the quaestio of Terence’s approach to his Greek models. On the other hand, Calboli continues, Terence innovates by adding certain traits of the motif of the uxor saeua (a wife that severely reproaches her husband), this time taken from Plautus. Accordingly, different Latin and Greek sources overlap in the same text 48 Carlotta Viti <?page no="49"?> and in the same character. A different perspective on Latin comedy is provided by Michael Weiss and Ryan Windhearn, two American scholars (both from Cornell University, Ithaca, NY), authors of the paper Syntactic theory and textual criticism in Plautus: Camerarius’ emendation of Amphitruo 260. Ryan Windhearn is a scholar of IE studies specialized in diachronic syntax. Michael Weiss is a specialist in historical linguistics and IE studies, and has published numerous works especially on the historical phonology and morphology of Greek, Latin, and the Sabellic languages (cf. his 2020 2 Outline of the historical and comparative grammar of Latin, which is now considered a classic of Latin linguistics). In this volume, Weiss and Windhearn search the correct lectio in a corrupt passage of Plautus’ Amphitruo (v. 260-261) by considering not only metrical and philological factors, but also findings of syntactic theory and of corpus linguistics. On the one hand, recent generative literature has shown that the syntactic constituency of auxiliary and modal constructions represents a specific relationship between Tense (expressed by the auxiliary) and the lower verbal domain (expressed by the participle), which may have different manifestations in different languages. On the other, Windhearn’s previous corpus analysis conducted on Plautus, as well as on Cato, demonstrates that Latin attests a right-headed structure, with a sequence participle-auxiliary, in the vast majority of cases. On the basis of these considerations, the two authors discuss different word orders attested in Plautus in different syntactic environments, and compare them to analogous constituent structures of English and other modern languages. They therefore propose an original method of using syntactic considerations in order to solve a crux of textual criticism. As anticipated, not all papers of the volume discuss language contact. In this case, a topic commonly addressed in the light of contact, such as Latin comedy with respect to its Greek original, is examined instead on the basis of internal factors. Paola Cotticelli-Kurras, an Italian scholar from the University of Verona, is the author of the paper entitled The creation of linguistic metalanguage in Antiquity and Middle Ages as result of translational processes. It is difficult to resume the breadth of Cotticelli-Kurras’ research interests, ranging from comparative morphology and syntax to language contact, historical and general linguistics, Hittitology and IE studies in general, the study of the grammatical tradition and metalanguage, writing systems etc. In this paper, she studies the development of the language of grammar in Late Antiquity (approximately from the 3 rd to the 7 th century AD) and in the Middle Ages. She shows how the dominant use of Latin and Greek in the Western and Eastern part, respectively, of the Roman Empire interacted with several other languages, which adopted expressions of the main cultural languages in grammatical borrowings and calques. She also discusses the most important Latin texts of grammar of this Introduction 49 <?page no="50"?> period, whose terms and concepts have exerted a profound impact on linguistic theory in the subsequent centuries up to contemporary times. For example, her paper shows how Latin drew from Greek key terms concerning all domains of grammar (phonology, morphology, and syntax), and at the same time how these terms were reinterpreted, and not simply translated, to adapt to the grammar of a different language as well as to new ideas developing in different times. This offers significant insights not only into the emergence of language contact in a specific linguistic domain, such as the study of grammar, which is under-investigated in contact linguistics, but also into the establishment of a scholarly tradition which has translation practice and Greek-Latin bilingualism at its very roots. Anna Orlandini (Centre Ernout, Paris-Sorbonne) and Paolo Poccetti (Univer‐ sity of Roma 2, Tor Vergata) are the authors of the following paper, La négation comme moteur de l’évolution linguistique. Both Orlandini and Poccetti are specialists of historical linguistics and IE studies. Paolo Poccetti is particularly interested in the linguistics of Greek and of the ancient languages of Italy (Latin, the Sabellic languages, etc.). Anna Orlandini conjugates the scholarship of classical philology and literature with the findings of modern linguistics and theory of language. In this paper, Anna Orlandini and Paolo Poccetti examine the effects of negation in different contexts which have never been scrutinized as a whole so far. They also identify a consistent change leading from constructions reflecting the language of logic to constructions that are more typical of natural languages. In the former type of languages, we have a structure with “double negation” (DN), as in Classical Latin, where two negations cancel each other. In the latter, we have a structure with “negative concordance” (NC), featured by cumulative negation, where only one element has a negative force, as in several Romance languages. In the change from DN to NC, the thematic role of negative indefinite pronouns plays a major role. These pronouns undergo a process of semantic weakening to the point of acquiring a positive value (especially that of “free choice” pronouns) in roles different from that of the subject, which is signalled by requiring a further negation in the clause. Moreover, the evolution of the negative structures has an impact on the weakening of the disjunctive particle aut. In the scope of a negative operator, aut switches from the exclusive disjunctive function, proper to the language of logic, to an inclusive one proper to natural languages. Negation also affects the evolution of the derivates with the suffix -teroexpressing alterity, especially in the pronoun system, which is related to the decay of the dual number and to the shift to multiple indefinite markers typical of natural languages. 50 Carlotta Viti <?page no="51"?> The next two papers discuss certain Latin syntactic constructions in compar‐ ison to Biblical Hebrew. Vit Bubenik, a Czech scholar from Memorial University of Newfoundland, Canada, is the author of Language contact in Antiquity. Participial constructions in Hellenistic Greek, Hebrew / Aramaic and Old Church Slavonic. Bubenik is specialized in both IE and Semitic historical linguistics, and has published numerous studies on morpho-syntactic change in Ancient Greek and Medieval Greek as well as in South Slavic, Middle Indo-Aryan etc. In this paper, he shows how Hellenistic Greek presents an increase in periphrastic constructions, in absolute constructions, and in nominalized infinitives with the article with respect to Classical Greek. He therefore searches an explanation for these morpho-syntactic phenomena by comparing constructions of the Greek New Testament with their correspondent sentences in Hebrew and in Aramaic, and evaluates to what extent we may have internally or externally conditioned changes. He also compares the relevant constructions in Hebrew and Aramaic with translations in Old Church Slavonic, with special focus on participles. Mauro Aresu and Silvia Cabriolu (both from the University of Cagliari) are the authors of the paper Observations on the variation of word-order: a comparison of Pre-classical, Classical and Biblical Latin. Mauro Aresu is a scholar working on classical philology and linguistics whose research interests also include Biblical Hebrew. Silvia Cabriolu is especially interested in Greek and Latin linguistics and in the Sabellic languages. They study word order variation in different syntactic constituents and in different varieties of Latin, such as Pre-Classical Latin, Classical Latin, and Biblical Latin. On the one hand, they compare corpora based on different authors and genres, which may also bring about word order variation in the analyzed constituents. On the other, they compare Biblical Latin with its Hebrew original and observe similarities and differences in translation. They also show how the same construction may be translated differently in different contexts, as prophetical texts, for instance, demand a higher fidelity to the original with respect to narrative texts. Francesca Cotugno, from the University of Verona, is the author of the paper entitled The Vandalic language in the light of Latin medieval manuscripts. Cotugno is a scholar of IE comparative linguistics, and is particularly interested in Vulgar Latin spoken in Roman Britannia, as well as in the Germanic and Celtic languages which interacted with Latin in Late Antiquity and in the Early Middle Ages. Her paper is placed last in the current volume just from a thematic point of view, as it addresses one of the languages of the Barbarian invaders of the Roman Empire and its relationship with Latin. As the Western Roman Empire was experiencing profound political instability, between the 5 th and the 6 th century AD, the Vandals founded a kingdom in North Africa Introduction 51 <?page no="52"?> and on various Mediterranean islands. The author discusses the main events of this period and the main linguistic features of the Vandalic language as they can be assessed on the basis of onomastics, glosses, and manuscripts, as well as indirect Latin testimonia. It is a difficult task, since there is a limited literature on Vandalic (therefore, this paper is also a contribution to Germanic linguistics), and since the interpretation of the fragmentary sources is often controversial. Moreover, Vandalic as well as other Early Germanic languages are often generically indicated as “Gothic” in the sources, and therefore their traits are not easy to disentangle. After a brief presentation of the papers and of their authors, I hope the reader may find our material of interest. I would like to express my gratitude to the authors for having kindly accepted to contribute to the volume. I also thank the dear colleagues and friends that have facilitated this task with exchanges of information and of points of view, or simply with their kind presence. In our field, as in any job, it is very important to work in a friendly and serene intellectual environment, which makes a scientific dialogue more fruitful. I would therefore like to thank my colleagues at the University of Lorraine, who have helped to make my stay in Nancy a very pleasant one. The wonderful library in the Department of “Histoire et Cultures de l’Antiquité et du Moyen Âge” (HISCANT-MA) has made it possible to find precious bibliographical resources. I also thank the colleagues of my former university, the Research Centre for History and Culture of Beijing Normal University, Beishida, and the Centre of Foreign Languages and Cultures of UIC (Beijing Normal University - Hong Kong Baptist University United International College) in China. Both the HISCANT-MA and UIC have financially supported the publication of the present volume. I am very grateful also for their contribution in this regard. Many thanks also to Narr Publishing House for having helped me during the whole publication process. Nancy, 20 th August 2023 Bibliographical references Aikhenvald, A. & R. Dixon (2001a) (eds.) Areal diffusion and genetic inheritance. Problems in comparative linguistics, Oxford, Oxford University Press. Aikhenvald, A. & R. Dixon (2001b) “Introduction”, in Aikhenvald, A. & R. Dixon (2001a), 1-26. Allen, C. (1995) Case marking and reanalysis. Grammatical relations from Old to Early Modern English, Oxford, Oxford University Press. Ansaldo, U. (2013) Contact languages, Cambridge, Cambridge University Press. 52 Carlotta Viti <?page no="53"?> Ayil, E. (2024) Identifying the stones of Classical Hebrew: a modern philological approach, Leiden, Brill. Bachvarova, M. (2016) From Hittite to Homer: the Anatolian background of Ancient Greek epic, Cambridge, Cambridge University Press. Bakker, P. & Y. Matras (2013) (eds.) Contact languages, Boston & Berlin, de Gruyter. Baldi, Ph. (1990) (ed.) Linguistic change and reconstruction methodology, Berlin & New York, Mouton de Gruyter. Beames, J. (1872-1879) A comparative grammar of the modern Aryan languages of India, I-III, New Delhi, Munshiram Manoharlal. Beekes, R. S. P. (2010) Etymological dictionary of Greek, Leiden, Brill. Beekes, R. S. P. (2014) Pre-Greek: phonology, morphology, lexicon, Leiden, Brill. Bernal, M. (1987; 1991; 2006) Black Athena: Afroasiatic roots of classical civilization, I-III, Philadelphia,-Rutgers University Press. Beye, C. (1984) “The epic of Gilgamesh, the Bible, and Homer: some narrative parallels”, in Evjen, H. (ed.) Mnemai: Classical studies in memory of K. Hulley, Chico (CA), Scholars Press, 7-19. Bianconi, M.; M. Capano; D. Romagno; F. Rovai (2022) (eds.) Ancient Indo-European languages between linguistics and philology. Contact, variation, and reconstruction, Leiden, Brill. Bowern, C. & B. Evans (2015) (eds.) The Routledge handbook of historical linguistics, London & New York, Routledge. Briant, P. (1996) Histoire de l’empire perse-: de Cyrus à Alexandre, Paris, Fayard. Burckhardt, J. (1898-1902) Griechische Kulturgeschichte, Berlin & Stuttgart, Spemann. Burgess, J. (1999) “Gilgamesh and Odysseus in the otherworld”, Echos du monde classique 43: 171-210. Burkert, W. (1984) Die orientalisierende Epoche in der griechischen Religion und Literatur, Heidelberg, Winter. Burkert, W. (2004) Babylon, Memphis, Persepolis: eastern context of Greek culture, Harvard University Press. Burkert, W. (2009) 3 Die Griechen und der Orient: von Homer bis zu den Magiern, München, Beck. Càssola, F. (1975) Inni omerici, Fondazione Lorenzo Valla, Mondadori. Clackson, J.; P. James; K. McDonald; L. Tagliapietra; N. Zair (2020) (eds.) Migration, mobility and language contact in and around the ancient Mediterranean, Cambridge, Cambridge University Press. Clarke, M. (2019) Achilles beside Gilgamesh: mortality and wisdom in early epic poetry, Cambridge, Cambridge University Press. Colvin, S. (2016) “Review of Beekes (2014) Pre-Greek: Phonology, Morphology, Lexicon”, The Journal of Hellenic Studies 136: 309-310. Introduction 53 <?page no="54"?> Cotticelli-Kurras, P. (2021) “On the concept of an Anatolian-Greek language area”, Old World: Journal of Ancient Africa and Eurasia 1, https: / / tinyurl.com/ 5n7r67db Currie, B. (2012) “The Iliad, Gilgamesh, and Neoanalysis”, in Montanari, F.; A. Rengakos; Ch. Tsagalis (eds.) Homeric contexts: Neoanalysis and the interpretation of oral poetry, Berlin, De Gruyter, 543-580. Dardano, P. (2012) “Il vento, i piedi e i calzari. I messaggeri degli dei nei miti ittiti e nei poemi omerici”, in Bolatti-Guzzo, N.; S. Festuccia; M. Marazzi (eds.) Centro mediterraneo preclassico. Studi e ricerche III, Roma, Herder, 53-87. Dawkins, R. M. (1916) Modern Greek in Asia Minor, Cambridge, Cambridge University Press. DELG = Chantraine, P. (1968-1980) Dictionnaire étymologique de la langue grecque: his‐ toire des mots, rev. by Taillardat, J.; O. Masson; J.-L. Perpillou (2009), Paris, Klincksieck. De Saussure, F. (1916) Cours de linguistique générale, ed. by Bally, C. & A. Sechehaye, Lausanne & Paris, Payot. Dixon, R. M. W. (1997) The rise and fall of languages, Cambridge, Cambridge University Press. Dorian, N. (1993) “Internally and externally motivated change in language contact settings: doubts about dichotomy”, in Jones, Ch. (ed.) Historical linguistics: problems and perspectives, London & New York, Longman, 131-155. Dowden, K. (2001) “West on East: Martin West’s East Face of Helicon and its forerunners, The Journal of Hellenic Studies 121: 167-175. Dunn, M. (2015) “Language phylogenies”, in Bowern, C. & B. Evans (eds.), 190-211. Durie, M. & M. Ross (1996) (eds.) The comparative method reviewed. Regularity and irregularity in language change, New York & Oxford, Oxford University Press. EDG cf. Beekes, R. S. P. (2010) Egetmeyer, M. (2010) Le dialecte grec ancien de Chypre, Berlin, de Gruyter. EM = Ernout, A. & A. Meillet (1959) 4 Dictionnaire étymologique de la langue latine, rev. ed. by André, J. (2001), Paris, Klicksieck. Emeneau, M. (1956) “India as a linguistic area”, in Hymes, D. (ed.) Languages in culture and society, New York, Harper & Row, 642-653. Emeneau, M. (1980) “Linguistic prehistory of India”, in Dil, A. (ed.) Language and linguistic area, Stanford, Stanford University Press, 85-104. Fisiak, J. (1990) (ed.) Historical linguistics and philology, Berlin & New York, Mouton de Gruyter. Foley, J. (2005) (ed.) A companion to ancient epic, Oxford, Blackwell. Fox, A. (1995) Linguistic reconstruction. An introduction to theory and method, Oxford, Oxford University Press. García Ramón, J. L. (2011) “Sprachen in Kontakt in Griechenland und Kleinasien im 2. Jahrtausend v. Chr.”, in Boschung, D. & C. M. Riehl (eds.) Historische Mehrsprachig‐ 54 Carlotta Viti <?page no="55"?> keit: Workshop des Zentrums für Antike Kulturen des Mittelmeerraumes (ZaKMiRa) und des Zentrums Sprachenvielfalt und Mehrsprachigkeit (ZSM) an der Universität Köln ( Juli 2008), Aachen, Shaker, 23-45. George, A. (2003) The Babylonian Gilgamesh epic: introduction, critical edition and cuneiform texts, I-II, Oxford, Oxford University Press. GEW = Frisk, H. (1960-1972) Griechisches etymologisches Wörterbuch, I-III, Heidelberg, Winter. Givón, T. (2021) The life cycle of adpositions, Amsterdam & Philadelphia, Benjamins. Goldberg, A. (1995) Constructions: a Construction Grammar approach to argument struc‐ ture, Chicago, University of Chicago Press. Goldberg, A. (2006) Constructions at work: the nature of generalization in language, Oxford, Oxford University Press. Grant, A. (2020) (ed.) The Oxford handbook of language contact, Oxford, Oxford University Press. Gresseth, G. (1975) “The Gilgamesh epic and Homer”, The Classical Journal 70: 1-18. Hagège, C. (1993) The language builder, Amsterdam & Philadelphia, Benjamins. Hagège, C. (2010) Adpositions, Oxford, Oxford University Press. Harris, A. & L. Campbell (1995) Historical syntax in cross-linguistic perspective, Cambridge, Cambridge University Press. Harrison, S. (2003) “On the limits of the comparative method”, in Joseph, B. & R. Janda (eds.), 343-368. Haubold, J. (2013) Greece and Mesopotamia: dialogues in literature, Cambridge, Cambridge University Press. Heggarty, P., C. Anderson, M. Scarborough, B. King, R. Bouckaert, L. Jocz, M. Kümmel, Th. Jügel, B. Irslinger, R. Pooth, Henrik L., R. F Strand, G. Haig, M. Macák, R. Kim, E. Anonby, T. Pronk, O. Belyaev, T. Dewey-Findell, M. Boutilier, C. Freiberg, R. Tegethoff, M. Serangeli, N. Liosis, K. Stroński, K. Schulte, G. Gupta, W. Haak, J. Krause, Q. Atkinson, S. Greenhill, D. Kühnert, R. Gray (2023) “Language trees with sampled ancestors support a hybrid model for the origin of the Indo-European languages”, Science 381, DOI: 10.1126/ science.abg0818 Heine, B. & T. Kuteva (2003) “Contact-induced grammaticalization”, Studies in Language 27: 529-572. Hock, H. H. (1975) “Substratum influence on (Rig-Vedic) Sanskrit? ”, Studies in the Linguistic Sciences 5: 76-125. Hoenigswald, H. (1963) “On the history of the comparative method”, Anthropological Linguistics 5: 1-11. Huehnergard, J. & A. Rubin (2012) “Phyla and waves: models of classification of the Semitic languages”, in Weninger, S. (ed.) The Semitic languages: an international handbook, Boston, De Gruyter, 259-278. Introduction 55 <?page no="56"?> Hymes, D. (1974) (ed.) Studies in the history of linguistics. Traditions and paradigms, Bloomington & London, Indiana University Press. Jacobsen, Th. (1968) “The battle between Marduk and Tiamat”, Journal of the American Oriental Society 88: 104-108. Janda, R. & B. Joseph (2003) “On language, change, and language change - Or, of history, linguistics, and historical linguistics”, in Joseph, B. & R. Janda (eds.), 3-180. Jasanoff, J. & A. Nussbaum (1996) “Word games: the linguistic evidence in Black Athena”, in Lefkowitz, M. & G. MacLean Rogers (eds.) Black Athena revisited, Chapel Hill & London, The University of North Carolina Press, 177-205. Jonge, C. (2008) Between grammar and rhetoric: Dionysius of Halicarnassus on language, linguistics, and literature, Leiden & Boston, Brill. Joseph, B. (2013) “Multiple sources and multiple causes multiply explored”, in De Smet, H.; L. Ghesquière; F. van der Velde (eds.) On multiple source constructions in language change, Amsterdam & Philadelphia, Benjamins, 675-691. Joseph, B. & R. Janda (2003) (eds.) The handbook of historical linguistics, Oxford, Blackwell. Kingsley, P. (1995) Ancient philosophy, mystery and magic. Empedocles and Pythagorean Tradition,-Oxford, Oxford University Press. Kingsley, P. (1999) In the dark places of wisdom, Point Reyes, Golden Sufi Center Publishing. Kingsley, P. (2003) Reality, Inverness, Golden Sufi Center Publishing. Kleingünther, A. (1933) ΠΡΩΤΟΣ ΕΥΡΕΤΗΣ. Untersuchungen zur Geschichte einer Frages‐ tellung, Leipzig, Dieterich’sche Verlagsbuchhandlung. Kretschmer, P. (1916) “Mythische Namen. 4. Adonis”, Glotta 7: 29-39. LaPolla, R. (2009) “Causes and effects of substratum, superstratum and adstratum influence, with reference to Tibeto-Burman languages”, Senri ethnological studies 75: 227-237. Lass, R. (1997) Historical linguistics and language change, Cambridge, Cambridge Univer‐ sity Press. Lazzeroni, R. (1987) “Il mutamento linguistico”, in Campanile, E.; E. De Felice; R. Gusmani; R. Lazzeroni; D. Silvestri (eds.) Linguistica storica, Roma, La Nuova Italia Scientifica, 13-54. López-Ruis, C.; F. Karahashi; M. Ziemann (2018) “They who saw the deep: Achilles, Gilgamesh and the Netherworld”, Kaskal 15: 85-108. Lord, A. (1960) The singer of tales (reprinted in 2000 with an introduction of S. Mitchell and G. Nagy), Cambridge (MA), Harvard University Press. Lord, A. (1991) Epic singers and oral tradition, Ithaca, Cornell University Press. LS = Lewis, Ch. & Ch. Short (1879) A new Latin dictionary founded on the translation of Freund’s Latin-German lexicon edited by E. A. Andrews, Rev. 1963, Oxford, Clarendon. 56 Carlotta Viti <?page no="57"?> LSJ = Liddell, H. & R. Scott (1843) A Greek-English lexicon. Revised and augmented throughout by H. Jones, with a revised supplement, Ninth edition 1996, Oxford, Clarendon. Marcovich, M. (1996) “From Ishtar to Aphrodite”, The Journal of Aesthetic Education 30: 43-59. Masica, C. (1976) Defining a linguistic area: South Asia, Chicago, University of Chicago Press. Masica, C. (1993) The Indo-Aryan languages, Cambridge, Cambridge University Press. Masson, É. (1967) Recherches sur les plus anciens emprunts sémitiques en grec, Paris, Klincksieck. Matras, Y. (2009) Language contact, Cambridge, Cambridge University Press. McMahon, A. & R. McMahon (2005) Language classification by numbers, Oxford Univer‐ sity Press. Meid, W. (2012) Ausgewählte Schriften zum Indogermanischen, Keltischen und Germani‐ schen, Innsbruck, Innsbrucker Beiträge zur Sprachwissenschaft. Meillet, A. (1925) La méthode comparative en linguistique historique, Oslo, Aschehoug & Co. Meillet, A. (1948) 6 Aperçu d’une histoire de la langue grecque, Paris, Hachette. Moravcsik, E. (1978) “Language contact”, in Greenberg, J. (ed.) and Ch. Ferguson & E. Moravcsik (ass. eds.) Universals of Human Language, I, Stanford, Stanford University Press, 93-122. Murray, O. (1993) 2 Early Greece, Cambridge (MA), Harvard University Press. Musiał, A. (2013) “The Closer East: the epic of Gilgamesh and Homer”, Scripta Classica 10: 33-51. MW = Monier-Williams, M. (1899) A Sanskrit-English Dictionary, Oxford, Clarendon. Nagy, G. (1996) Poetry as performance, Homer and beyond, Cambridge, Cambridge University Press. Nagy, G. (2005) “The epic hero”, in Foley, J. (ed.), 71-89. NIL = Wodtko, D.; B. Irslinger; C. Schneider (2008) Nomina im indogermanischen Lexicon, Heidelberg, Winter. OLD = Glare, P.G.W. (2012) 2 (ed.) Oxford Latin dictionary, Oxford, Oxford University Press. Parry, M. (1971) The making of Homeric verse: the collected writings of Milman Parry, ed. by Parry, A., Oxford, Oxford University Press. Polomé, E. & W. Winter (1992) (eds.) Reconstructing languages and cultures, Berlin & New York, Mouton de Gruyter. Porzig, W. (1954) Die Gliederung des indogermanischen Sprachgebiets, Heidelberg, Winter. Ringe, D.; T. Warnow; A. Taylor (2002) “Indo-European and computational cladistic”, Transactions of the Philological Society 100: 59-129. Introduction 57 <?page no="58"?> Rollinger R. (2011) “Altorientalische Einflüsse auf die homerischen Epen”, in Rengakos, A. & B. Zimmermann (eds.) Homer-Handbuch: Leben-Werk-Wirkung, Stuttgart & Wiemar, 213-227. Schadewaldt, W. (1965) 4 Von Homers Welt und Werk. Aufsätze und Auslegungen zur homerischen Frage, Stuttgart, Koehler. Schleicher, A. (1861) Compendium der vergleichenden Grammatik der indogermanischen Sprachen, I-II, Weimar, Böhlau. Schleicher, A. (1863) Die Darwinsche Theorie und die Sprachwissenschaft - offenes Sends‐ chreiben an Herrn Dr. Ernst Haeckel,-Weimar, Böhlau. Schmidt, J. (1872) Die Verwandtschaftsverhältnisse der indogermanischen Sprachen, Weimar, Böhlau. Schuchardt, H. (1885) Über die Lautgesetze. Gegen die Junggrammatiker, Berlin, Oppen‐ heim. Sironi, F. & M. Viano (2015) “Šiduri and Thetis: a note on two parallel scenes (Gilgameš OB VA + BM III and Hom. Il. XXIV 120-132)”, Kaskal 19: 203-211. Thieme, P. (1938) Fremdling im Veda, Leipzig, Deutsche Morgenländische Gesellschaft. Thieme, P. (1954) Die Heimat der indogermanischen Gemeinsprache, Wiesbaden, Steiner. Th.L.L. = Thesaurus Linguae Latinae, https: / / publikationen.badw.de/ de/ thesaurus/ lemm ata Thomason, S. (2001) Language contact: an introduction, Washington DC, Georgetown University Press. Thomason, S. & T. Kaufman (1988) Language contact, creolization, and genetic linguistics, Berkeley & Los Angeles, University of California Press. Ungnad, A. (1923) Gilgamesch-Epos und Odyssee, Kulturfragen, Breslau, im Selbstverlag des Herausgebers. van der Gaaf, W. (1904) The transition from the impersonal to the personal construction in Middle English, Heidelberg, Winter. Viti, C. (2015) Variation und Wandel in der Syntax der alten indogermanischen Sprachen, Tübingen, Narr. Walkden, G. (2013) “The correspondence problem in syntactic reconstruction”, Dia‐ chronica 30: 95-122. Watkins, C. (2001) “An Indo-European linguistic area and its characteristics: Ancient Anatolia. Areal diffusion as a challenge to the Comparative Method? ”, in A. Aikhen‐ vald & R. M. W. Dixon (eds.), 44-63. Weinreich, U. (1977) Sprachen in Kontakt. Ergebnisse und Probleme der Zweisprachigkeits‐ forschung, München, Beck. Weirich, M.; S. Jannedy; G. Schüppenhauer (2020) “The social meaning of contextualized sibilant alternations in Berlin German”, Frontiers in Psychology 11: 1-18, https: / / tinyu rl.com/ 2bccsd3x 58 Carlotta Viti <?page no="59"?> Weiss, M. (2015) “The Comparative Method”, in Bowern, C. & B. Evans (eds.), 127-145. West, M. (1983) The Orphic poems, Oxford, Clarendon. West, M. (1997) The East face of Helicon: West Asiatic elements in Greek poetry and myth, Oxford, Clarendon. West, M. (2011) The making of the Iliad: disquisition and analytical commentary, Oxford, Oxford University Press. West, M. (2014) The making of the Odyssey, Oxford, Oxford University Press. Widmer, P. (2018) “Indogermanische Stammbäume. Datentypen und Methoden”, in Rieken, E.; U. Geupel; Th. Roth (eds.) 100 Jahre Entzifferung des Hethitischen, Wies‐ baden, Reichert, 373-388. Woodhouse, S. C. (1910) English - Greek dictionary. A vocabulary of the Attic language, London, Routledge & Sons. Zgoll, C. (2012) “From wild being to human to friend. Reflections on anthropology in the Gilgamesh epic and in Homer’s Odyssey”, Kaskal 9: 137-155. Zsolt, S. (forthcoming) “Alleged Anatolian phraseological borrowings in Homer’s lan‐ guage. A reconsideration”, in van Beek, L. (ed.) Language change in Epic Greek and other oral traditions, Leiden, Brill. Introduction 59 <?page no="61"?> 1 Cf. Paul Veyne, L’Empire gréco-romain, 2005. 2 « Le monde grec n’était pas européen », Lucien Febvre, cité par Emmanuelle Loyer, L’impitoyable aujourd’hui, 2022, p.-315. Le grec ancien : une réalité multiforme Monique Bile, René Hodot, Guy Vottéro (Université de Lorraine) Abstract: Knowledge of the ancient Greek language cannot be limited to the study of literature, despite its richness and variation. It is also worth considering the extension of the ancient Greek world in space and time, in the Mediterranean and beyond, from the 15 th century BC to the 5 th century AD, as well as a variety of archaeological, epigraphic, papyrological sources, and diverse linguistic practices (in their dialectal, social, and pragmatic variations). All this may bring benefits on widening and deepening our perception of Greek. Keywords: dialects, epigraphy, geography, history, inter-comprehension, Greek language, Mycenaean, sociolinguistics. A---Cadre et sources (1) L’histoire du monde grec antique s’étend sur une vingtaine de siècles, de l’époque mycénienne à l’Antiquité tardive. L’extension spatiale de ce monde a varié selon les époques ; au total, il a touché d’ouest en est des territoires s’étendant de l’Ibérie à l’Inde, et du nord au sud, de la Mer Noire à la Nubie. Complétée par les apports de Rome 1 , la civilisation grecque s’est donc répandue bien au-delà de l’Europe 2 . Déjà, dans le petit pays qu’était la Grèce continentale, prévalait un morcelle‐ ment géographique et linguistique, chaque région se distinguant des autres par une façon propre de parler « le grec ». Ensuite la multiplication des colonies au peuplement très varié, présentant chacune un cas particulier, a constitué un <?page no="62"?> 3 Cf. pour la Grande Grèce, S. Mimbrera Olarte, EAGLL, § 1.d et 2.g. 4 F. Gaignerot-Driessen, «La formation des cités grecques (XIV e s. -VII e s. av. n.è.) : approches archéologiques à partir de quelques cas crétois”, REG 131 (2018), 49-73. 5 Références chez A. Bernand, Alexandrie des Ptolémées, Paris, 1995, 61 et 65. 6 D’Ercole-Zurbach 2019, 608-609. 7 Richer 2018, 67-68 et 222-223. 8 Richer 2018, 81-83. 9 M. Sartre, L’Asie Mineure et l’Anatolie d’Alexandre à Dioclétien, Paris, 1995, 297-298. des premiers creusets linguistiques grecs 3 . Inversement la conquête de l’Italie par Rome a contribué à imposer très tôt le latin comme unique langue officielle pour toute la péninsule. (2) L’héritage grec du monde occidental L’héritage grec du monde occidental ne se limite pas à la littérature et à la philosophie. En effet, les sources dites « littéraires » (parmi lesquelles les his‐ toriens anciens tiennent une place de premier choix) ne constituent pas toute la documentation disponible. Ainsi les conditions historiques du monde politique grec peuvent être appréhendées par l’archéologie : comment comprendre, par exemple, le passage, en Crète, de la société palatiale du II e millénaire à la cité-Etat sans tenir compte des données archéologiques (localisation et type de l’habitat, existence de monuments tels que temples 4 ) ? En outre, les fouilles permettent souvent de confirmer des données fournies par la littérature, comme celles d’Alexandrie, à l’architecture décrite par plusieurs écrivains 5 , de Naucratis, emporion d’Egypte dont la fondation a été célébrée par Hérodote (II, 178-179) 6 ou du Ménélaion, temple mentionné par le même historien VI, 61 7 . Le vaste domaine de l’archéologie possède plusieurs ramifications permettant d’aboutir à une datation assez précise et renseignant sur des usages quotidiens. La céramologie est l’une d’elles : à côté des exemples célèbres des céramiques à figures noires et à figures rouges (Corinthe, Athènes) exportées en Sicile et en Italie du sud, les vases laconiens du VI e siècle trouvés à Samos, en Cyrénaïque et en Etrurie, révèlent que Sparte a été une cité d’artistes appréciés en dehors du Péloponnèse, avant de devenir la cité militaire que l’on sait 8 . La production de céramique, de type usuel ou plus raffiné (« la sigillée orientale »), demeura pendant une longue période une activité artisanale, attestée par Pline le jeune, en Asie Mineure et en Anatolie, où plusieurs ateliers (à Ephèse, Pergame et sans doute plus loin) envoyèrent leurs produits en Mer Noire et en mer Egée 9 . Les représentations figurant sur toutes ces céramiques, mythologiques ou non, rendent compte de l’environnement culturel d’une communauté. Un autre volet fondamental de l’archéologie est la numismatique : les mines, comme celles du Laurion qui ont procuré l’argent à Athènes pendant 62 Monique Bile, René Hodot, Guy Vottéro <?page no="63"?> 10 La datation exacte et le lieu d’apparition de la monnaie sont sujets à discussion parmi les numismates, cf. Rebuffat 1996, 23-29.- 11 G. Le Rider, Monnaies crétoises du V e au I er siècle av. J.-C., Paris, 1966, 191-194. Cette hypothèse est acceptée par Rebuffat 1996, 88. 12 E. Babelon, Les monnaies grecques, aperçu historique, Paris 1921, 84-85. 13 Cf. E. Will, Historica graeco-hellenistica. Choix d'écrits 1953-1993, Paris 1998. 14 Le Dyscolos, La Samienne, Le Bouclier. 15 Cf. Goukowsky - Feyel, 2019-; Goukowsky, 2020. 16 Sauf indication contraire, les documents sont datés avant notre ère. des siècles, et les métaux monétaires utilisés, sont des indices déterminants de la puissance d’une cité. La monnaie est une réalité grecque à partir de l’époque archaïque 10 : Platon, qui refuse au citoyen de sa cité idéale le droit de posséder de l’or ou de l’argent (Lois V 742 a), admet une monnaie locale pour payer les artisans, les mercenaires, les esclaves, et une monnaie grecque pour le commerce, le voyage. Si les différents types de monnaie et de poids ont pu freiner un certain temps l’essor du commerce international, celui-ci put enfin se développer sans doute grâce aux mercenaires ; c’est ainsi que les soldats revenus en Crète porteurs de numéraires étrangers permirent un monnayage local dans des ateliers crétois 11 . De façon différente de la littérature, mais tout aussi prégnante, les légendes monétaires renseignent sur les divinités honorées d’une cité (en particulier les divinités poliades), sur ses magistrats, sur les engouements d’une époque, comme les types de médailles olympiques (thème de la Victoire, aigle de Zeus) 12 . En plus de son importance dans les échanges commerciaux (terrestres ou maritimes), la monnaie acquiert ainsi un rôle didactique de premier plan 13 . La papyrologie est une autre discipline permettant de découvrir les modes de vie grecque, surtout celle de l’Egypte depuis les Ptolémées jusqu’à l’invasion du pays par les Arabes, donc pendant les siècles où le grec était la langue au moins écrite de l’élite. Parmi les papyrus les plus riches d’enseignements, il faut citer celui de La Constitution des Athéniens (découvert en 1890), les 1800 papyrus d’Herculanum, contenant des œuvres philosophiques grecques (en particulier celles du philosophe Philodème de Gadara), le papyrus de Derveni en Macédoine, transcrivant un poème orphique, les papyrus d’Oxyrhynchos, avec de nombreux textes littéraires redécouverts (fragments de l’Iliade et de l’Odyssée, d’Hypéride, d’Hérondas, de Sappho, de Sophocle…, comédies de Ménandre 14 , Helléniques d’Oxyrhynchos  15 ). Très instructifs également sont les papyrus de Zénon, du nom de ce Grec d’Asie Mineure venu à Alexandrie au III e siècle 16 , car ils sont un bon indicateur de la vie quotidienne, des structures sociales et aussi des préjugés ethniques que Zénon partageait sans doute avec certains Le grec ancien : une réalité multiforme 63 <?page no="64"?> 17 C. Orrieux, Les papyrus de Zenon. L’horizon d’un grec en Egypte au III e siècle avant J.C., Paris 1983. 18 Textes cités dans l’ouvrage de J.-M. Bertrand, Inscriptions historiques grecques, Paris, 1992, 17-22, 29-30, 61-64, 112-113, 176-178, 226-227. 19 Pouilloux 2003, n° 9, 47-50. 20 Cf. Pouilloux 2003, n° 15 et 16, 67-70, et E. Samama, Les médecins dans le monde grec, Genève, 2003, n° 126 et 127, 231-234 : si le texte de Gortyne indique de façon lapidaire le contexte social et politique de la cité, celui de Cnossos semble prendre plaisir à détailler les malheurs de la cité souvent rivale, et alors son alliée. 21 Cf. F. Sokolowsky, Lois sacrées d’Asie Mineure, Paris, 1955, et Lois sacrées des cités grecques, Paris, 1969. de ses compatriotes 17 . Bien d’autres papyrus mettent en lumière des sujets économiques (baux de location, comptes d’une exploitation agricole), artistiques (musique, avec Mésomède de Crète), scientifiques (astronomie). Leur apport est donc considérable pour le droit, la religion (oracles) ou le domaine profane (magie, lettres privées). La société égyptienne, très largement hellénophone à l’époque ptolémaïque, puis romaine, s’y dévoile sous ses différents aspects (politiques, culturels, sociétaux). Enfin, tout aussi importante que la littérature, sinon plus, parce que la palette des sujets abordés est plus large, est l’épigraphie aux multiples supports (tablettes d’argile mycéniennes, pierres - matériau qui a le mieux survécu -, métaux, céramique, cuirasses…). Il y a lieu de distinguer au moins deux grandes catégories d’inscriptions, les officielles et les privées. Les textes officiels, qui émanent des cités, sont divers : décrets, traités d’alliance, lois restituent des moments de première importance, et en disent parfois plus que les historiens grecs de l’Antiquité. Qu’on songe au « Marbre de Paros », à la loi de Dréros, assignable à 650, sur la non-itération du cosmat (magistrature suprême en Crète), où apparaît la première attestation du mot πόλις « cité », au décret athénien pour les Chalcidiens (en 446/ 445), au traité d’alliance entre Athènes et des rois thraces (en 357), aux documents relatifs au roi indien Aśoka (après 268), ou au serment du roi du Pont Pharnace envers les Grecs de Chersonèse (en 179) : toutes ces inscriptions confirment ou révèlent des positions politiques 18 . Les décrets honor‐ ifiques apportent des précisions sur les stratégies des cités, tels le décret athénien en faveur des trois fils de Leucôn, roi du Bosphore, région essentielle pour l’alimentation d’Athènes en blé (en 347/ 346) 19 , ou les lettres-décrets de Cnossos et de Gortyne (en 218) honorant Hermias médecin de Cos, et mentionnant la stasis dont a été affectée Gortyne lors de la guerre de Lyttos (en 221-219) 20 . Les textes législatifs, autre volet des documents officiels, sont de deux sortes : il y a les lois sacrées  21 et les lois civiles, essentiellement fournies par la Crète, déjà célèbre dans l’Antiquité pour ses lois. L’inscription la plus représentative est celle qui 64 Monique Bile, René Hodot, Guy Vottéro <?page no="65"?> 22 Les éditions les plus complètes, avec commentaire détaillé, sont celles de M. Guarducci, Inscriptiones Creticae IV, Rome 1950, 123-171 n° 72, de R.F. Willetts, The Law Code of Gortyn, Berlin 1967, et de M. Gagarin - P. Perlman, The Laws of Ancient Crete, c. 650-400 BCE, Oxford2016, 334-428. Le texte et une traduction en français sont fournis par H. van Effenterre - F. Ruzé, Nomima II, Rome 1995, avec commentaires des colonnes dans plusieurs parties de l’ouvrage. 23 Le mur des LG devait faire partie d’un édifice circulaire dont le diamètre est estimé à 33, 30-m; sa longueur est de 8-m 71, sa hauteur 1-m 705. 24 Sauf la douzième qui n’en a que 19. 25 Pour J. Goody, Pouvoirs et savoirs de l’écrit, Paris, 2007, 199, « On peut également considérer que la gravure des codes, comme le code crétois des lois de Gortyne, ou le code encore plus ancien de Hammurabi, était une façon d’en faire l’annonce aux citoyens-». 26 Lysias est à cet égard l’auteur le plus utile. 27 Dans la très grande loi sacrée des mystères d’Andanie, datée de 92/ 91, un passage concerne le traitement à tenir envers les esclaves fugitifs réfugiés dans un temple, cf. Schwyzer, 1923, 315, l. 80-84. Le cas est évoqué dans les Lois de Gortyne, colonne I 39-46 et dans un autre texte gortynien un peu antérieur (IC IV 41, IV 6-18 = Bile 2016, n° 20, 89). 28 E.g. Migeotte 1986 ; D. Knoepfler (éd.), Comptes et inventaires dans la cité grecque, Neufchâtel-Genève, 1988. 29 Il est possible que la vente d’esclaves ait fait l’objet d’une mention sur des registres municipaux, cf. P. Ismard, La cité et ses esclaves. Institution, fictions, expériences, Paris, 2019, 43-48. est communément appelée Lois ou Code de Gortyne (LG ou CdG), découverte en 1884 par deux épigraphistes, l’Italien Federico Halbherr et l’Allemand Ernst Fabricius, et qui a fait l’objet de nombreuses traductions en diverses langues 22 . La « Reine des inscriptions », aux dimensions impressionnantes 23 et comportant 12 colonnes de plus de 50 lignes chacune 24 , constitue un monument imposant, destiné à une publicité maximale 25 . Ce document, daté de 450 selon la plupart des commentateurs, est unique pour la connaissance du droit grec, car les autres témoignages, par exemple les plaidoyers attiques, ne fournissent que des allusions aux lois athéniennes 26 , tandis que quelques inscriptions s’intéressent à des points particuliers du droit 27 . Il s’agit, à Gortyne, de droit privé, — divorce, héritage et institution de la patrooikos, qui rappelle l’épiclérat (athénien surtout) mais avec de notables différences. Les comptes des cités et les inventaires sont une autre source fondamentale de renseignements, parce qu’ils traitent de sujets pratiques souvent absents de la littérature (cours des monnaies, contrats de prêt) 28 : la cité se montre dans son activité économique, envers les sanctuaires, envers ses citoyens 29 et, dans le cas d’Athènes, envers ses «-alliés-» de la Ligue de Délos, auquel le phoros est Le grec ancien : une réalité multiforme 65 <?page no="66"?> 30 Cf. M.-Chr. Hellmann, Choix d’inscriptions architecturales grecques, traduites et com‐ mentées, Lyon, 1999, 63-66. 31 D’où leur nom béotien de ϝικατιϝετιεες, cf. infra § 5 « lexique-». 32 Sur tous ces aspects très divers, mais riches d’enseignement, malgré les difficultés pratiques qu’on peut rencontrer, on consultera avec profit les ouvrages suivants : Hyettos 1976, p.-271-281 ; Kalliontzis 2020, chap. I, II, IV, VIII, IX. 33 Parmi l’ample documentation, on citera L. Darmezin, Les affranchissements par consé‐ cration en Béotie et dans le monde grec hellénistique, Nancy-Paris 1999, et D. Mulliez, Corpus des inscriptions de Delphes V.1 Les actes d’affranchissement : prêtrises I à IX (n° 1- 722), Athènes, 2019, ouvrage qui commence une série sur les affranchissements delphiques. 34 P. Ismard, « Les affranchis d’Apollon. Delphes, II e - I er siècles avant notre ère », Les mondes de l’esclavage, une histoire comparée, sous la direction de P. Ismard, Paris, 2021, 75-79. réclamé. D’autres textes, comme les comptes de l’Erechteion (en 408/ 407) font voir que travaillaient côte à côte citoyens pauvres, métèques et esclaves 30 . Là où ils sont attestés, comme en Béotie, les catalogues militaires apportent un éclairage encore différent : ces textes comptabilisent chaque année les jeunes hommes âgés de 20 ans 31 , qui quittent l’éphébie pour être versés dans le corps des hoplites et des cavaliers ; ils présentent plusieurs intérêts : militaire (formation reçue, constitution de l’armée), démographique (possibilité de suivre l’évolution démographique d’une cité, quand on dispose de séries à peu près assurées), onomastique (bon aperçu de l’onomastique locale et/ ou régionale, chaque nom étant accompagné de son patronyme), chronologique et social (avec la possibilité dans certains cas de reconstituer une généalogie) 32 . A mi-chemin entre documents officiels et privés s’insèrent les actes d’affran‐ chissement, groupe particulièrement bien représenté, puisqu’à partir du III e s. ces textes ont été produits en très grand nombre ; certains émanent des cités, qui « offrent » par consécration à une divinité un ancien esclave, devenu affranchi, alors que des particuliers adoptent le même schéma 33 . L’intérêt immense de ces milliers de textes est, outre l’énoncé de modalités distinctes selon les lieux, les époques et les individus, de souligner l’originalité de l’affranchissement grec par rapport à l’affranchissement romain 34 . Ainsi les caractéristiques le plus fréquemment rencontrées sont la paramoné — l’affranchi(e) doit rester auprès de son ancien maitre pendant une durée déterminée (souvent jusqu’au décès de ce dernier) — et le statut des enfants de l’affranchi(e) vivants ou à naître, qui dépend de l’arbitraire du maître. Souvent esquissées dans la littérature grecque, les relations entre maître et esclaves sont explicites dans les actes d’affranchissement des particuliers et s’ajoutent aux témoignages de la vie sociale et affective des Grecs. 66 Monique Bile, René Hodot, Guy Vottéro <?page no="67"?> 35 Il suffit de consulter le LGPN en ligne pour s’en convaincre. 36 Pour la société du nord du Ghana qu’il a étudiée, J. Goody met l’accent sur les conséquences sociales de l’écriture qui affectent ceux qui la pratiquent et les illettrés : « Seuls ceux qui savaient lire et écrire étaient habilités à participer aux délibérations. Les autres se retrouvaient désormais exclus », L’homme, l’écriture et la mort, Paris, 1996, p.-229. 37 Exemples chez Guarducci 1978, qui souligne que les citoyens appelés à prêter serment sont soumis aux mêmes malédictions (ἀραί) en cas de violation du serment, comme l’indique le traité entre la cité crétoise d’Hiérapytna et ses colons établis près d’Arcades, IC III, III, 5, l. 22-26. 38 M. P. Nilsson, La religion populaire dans la Grèce classique, Paris, 1955, 190-199, et Les croyances religieuses de la Grèce antique, Paris, 1955, 185-201 : l’auteur écrit qu’à l’époque de Platon et de Démosthène, les tablettes incantatoires avaient de nombreux partisans de leur efficacité. Un bon exemple en est fourni par une tablette de plomb crétoise, cf. Cl. Brixhe-A. Panayotou, « Le plomb magique de Phalasarna IC II-XIX 7 », Hellènika Symmikta. Histoire, linguistique, épigraphie II, Nancy-Paris 1995, 23-68. Celle-ci peut se manifester aussi dans les inscriptions privées, autre dimen‐ sion de l’épigraphie. Ces documents sont de plusieurs sortes : le volet le mieux attesté réside dans les épitaphes, dont les plus savantes, en vers, s’inspirent souvent de la langue homérique et de l’idéologie épique. Les stèles funéraires, répandues dans tout le monde grec, à toutes les époques, donnent des indications variées sur l’onomastique 35 , les conditions sociales. En effet, écrire ou faire écrire suppose une position dans la société, un accès à la culture, une certaine maîtrise de l’écrit ou la conscience que le recours à l’écrit est un moyen sûr de se démarquer socialement 36 : une épitaphe est, outre le témoignage d’une relation privée entre deux ou plusieurs personnes, le reflet d’un statut privilégié, car, même si l’origine sociale des personnes est passée sous silence, elle se déduit de l’acte d’écriture. Assez concis dans les épitaphes d’époque archaïque, le texte gagne en ampleur par la suite, avec indications sur la vie de la personne défunte, les circonstances de sa mort, malgré le caractère stéréotypé des expressions. En rapport avec la mort, les tablettes de plombs gravées d’imprécations, souvent enfouies avec le défunt, participent d’une croyance très archaïque au pouvoir performatif de la malédiction (la parole crée l’acte) 37 : dans les formules censées agir magiquement contre leur destinataire, sont employés des termes souvent inconnus des textes littéraires et l’exposé des griefs par l’auteur de la malédiction est une source de renseignements sur la superstition et les mentalités. Car, en imputant la mort du bétail à un voisin, le thème du mauvais œil est souvent invoqué ; la religion grecque et la magie pouvaient faire bon ménage 38 , et cette coexistence est un élément qui intéresse aussi bien les philosophes que les spécialistes des religions. Le grec ancien : une réalité multiforme 67 <?page no="68"?> 39 E.g. les Iamata d’Epidaure, 48 inscriptions gravées sur quatre stèles (IG IV 2 1.121-124), datées de la deuxième moitié du IV e s., cf. C. Prêtre - P. Charlier, « Stigmata et grammata dans les récits de guérisons miraculeuses d’Epidaure. Une nouvelle analyse sémantique et clinique », BCH 138 (2014), 185-199. Les textes crétois de Lébéna, publiés dans IC I, XVII, 152-178, décrivent des processus identiques, analysés par P. Sineux, « Le dieu ordonne. Remarques sur les ordres d’Asklépios dans les inscriptions de Lébéna », Kentron 20 (2004), 137-146. 40 A Eleutherna par exemple, IC II, XII, 31, 168-170, texte daté du II e siècle, mais aussi en Grande Grèce, en Thessalie, à Hipponion (la plus ancienne) et à Rome. 41 C’est ainsi qu’Athénaios, sculpteur de Paros, signe son œuvre réalisée à Gortyne au I er siècle (IC IV 343). 42 Cf. pour la Crète, P. Perlman, « Reading and writing Archaic Cretan Society », Cultural Practices and Material Culture in Archaic and Classical Crete, Berlin-Boston (2014), 177-206, et J. Whitley, « Cretan Laws and Cretan Literacy », American Journal of Archaeology 101 (1997), 635-661. 43 Cependant les textes écrits par des affranchis sont le signe que certains esclaves, dont les esclaves publics, pratiquaient l’écrit, cf. P. Ismard, La démocratie contre les experts. Les esclaves publics en Grèce ancienne, Paris, 2015, 63-73. En outre, dans sa pratique de la religion officielle, le Grec s’adressait aux dieux de la cité, qui pouvaient lui délivrer un oracle (message interprété comme tel, ainsi à Dodone), ou auxquels il demandait la guérison, comme en attestent les ex-voto pour le culte d’Asklépios (Epidaure, Lébéna en Crète) : les textes des fidèles expliquent comment, par les rites de l’incubation, la divinité s’était manifestée 39 . Ces corpus, qui mêlent descriptions de maladies et recettes pour des guérisons, contribuent, ainsi que les tablettes d’exécration, à définir le recours au merveilleux. Dans le domaine mystique, les textes « orphiques » 40 , attestés de la fin du V e siècle à la moitié du III e siècle de notre ère, manifestent un autre versant du sentiment religieux grec. Si les textes sont le plus souvent des demandes aux dieux, il y a aussi des dédicaces de vainqueurs aux jeux panhelléniques ou de guerriers. Enfin, les signatures d’artistes, que ce soit sur des monuments publics ou privés, constituent des renseignements utiles sur ces techniciens, dont la littérature n’a pas retenu les noms, et sur leur lieu d’activité, parfois différente de leur cité 41 . Il est aisé de conclure que l’épigraphie est une source fondamentale pour la compréhension du monde grec depuis l’époque mycénienne. Les producteurs des textes, cités ou particuliers, tenaient à publier leurs décisions, à faire connaitre leur vie professionnelle ou privée ; cette activité rédactionnelle, qui met en lumière le rôle de l’écrit dans une société antique et pose la question de son lectorat 42 , est, pour tous les antiquisants, le domaine le plus assuré sur lequel ils doivent baser leurs recherches. Car peu de domaines sont occultés, et même si certaines couches de la population ne dominaient pas cette technique 43 , la civilisation grecque s’exprime avec force dans les écrits qui ont subsisté. 68 Monique Bile, René Hodot, Guy Vottéro <?page no="69"?> 44 Cf. Hdt, VIII, 144-: τὸ Ἑλληνικόν, ἐὸν ὅμαιμόν τε καὶ ὁμόγλωσσον, καὶ θεῶν ἱδρύματά τε κοινὰ καὶ θυσίαι, ἤθεά τε ὁμότροπα «-Ce qui unit tous les Grecs, — même sang et même langue, sanctuaires et sacrifices communs, semblables moeurs et coutumes-» (trad. Ph.-E. Legrand, CUF). Mais, comme on le sait, la réaction des différentes cités face à l’invasion perse de 481 a été très variable, cf. le même Hdt. VII, 132, 138. 45 Pour le grec mycénien et le linéaire B, on a utilisé principalement ici les ouvrages suivants : Bartoněk 2003, Duhoux-Morpurgo Davies (2008-2014), Duhoux 2013, ainsi que les articles de l’EAGLL (« Mycenaean Script and Language » d’Y. Duhoux et «-Variation in Mycenaean Greek de C. Milani). 46 La quasi-totalité des textes mycéniens retrouvés actuellement (plus de 6000) sont écrits sur argile : principalement des tablettes (« feuille » ou « page »), mais aussi des « étiquettes » apposées sur des récipients (paniers en osier, etc.) et des nodules, sortes de boulettes d’argile frappées d’un sceau et destinées à recevoir un lien pour marquer animaux ou objets. On dispose aussi d’environ 170 textes peints sur vases, la plupart fabriqués en Crète et exportés. 47 Grâce à la paléographie et à l’étude des empreintes digitales laissées sur l’argile, il est possible d’attribuer à tel ou tel scribe (désigné par un numéro d’ordre) des lots plus ou moins importants de textes (e.g. le scribe 1 de Pylos est l’auteur de plus de 200 tablettes et «-étiquettes-»), mais aussi de déterminer des usages individuels en matière d’écriture, d’orthographe et de langue. En outre on a découvert que les scribes étaient assistés par des aides (enfants ou adultes âgés) pour la fabrication des tablettes. Enfin, en établissant des liens entre scribes et/ ou sites, on a pu montrer l’existence de plusieurs écoles scribales. (3) Les Grecs étaient dialectophones Dans les faits le grec ancien des époques archaïque et classique est un ensemble de dialectes. Cette situation s’explique géographiquement, historiquement et culturellement (cf. infra § 4), si bien que la Grèce d’avant la période hellénistique ne constitue pas une nation au sens moderne du terme et, si le sentiment d’appartenir à une même entité existe bel et bien chez certains, tout particulièrement chez ceux qui sont au contact direct de « barbares » puissants, comme les Ioniens 44 , cette entité grecque n’est pas uniforme. On pourrait objecter le cas du grec mycénien, la forme de grec la plus ancienne actuellement connue 45 . En effet il présente toutes les caractéristiques d’une koinè : a) il est documenté 46 sur un territoire important recouvrant la Crète et une grande partie de la Grèce continentale : Attique, Béotie, Péloponnèse (notamment l’Argolide, la Laconie, la Messénie), Phocide et Thessalie, et sur une longue période, puisque les documents en linéaire B s’échelonnent entre le XIV e et le XII e siècle ; b) on observe une remarquable tendance à l’uniformité dans l’écriture proprement dite (malgré le tracé complexe de nombreux signes et l’existence de variantes), mais aussi dans l’orthographe, le vocabulaire, la phraséologie, etc. Cependant, grâce aux nombreuses études menées sur les pratiques matérielles (fabrication des tablettes, techniques d’écriture 47 …), il est possible de préciser Le grec ancien : une réalité multiforme 69 <?page no="70"?> 48 Il faut voir dans la plupart des tablettes des documents « provisoires » destinés à la rédaction de textes synthétiques, écrits tantôt sur tablettes (notamment des tablettes « page »), tantôt sur matériaux périssables (parchemin, tablettes de cire…), dont on n’a découvert aucun exemplaire à ce jour, mais qui ont très vraisemblablement existé. A propos des nodules et de leur utilisation, cf. l’étude détaillée d’Olivier - Melena - Piteros 1990. 49 Cf. CoMIK, tablette KN V[1] 114. 50 Cf. Duhoux---Morpurgo Davies 2008, p.-135-136. 51 On s’accorde aujourd’hui pour considérer que le syllabaire chypriote du premier millénaire ne dérive pas du linéaire B, mais des écritures « chypro-minoennes », qui n’ont qu’un rapport lointain avec le linéaire B. 52 Il va se soi qu’en dehors de leurs fonctions leurs usages quotidiens pouvaient être différents et variables d’une région à l’autre. les caractéristiques de cette forme linguistique : a) elle est attestée presque exclusivement dans des textes officiels purement administratifs à usage interne 48 et écrits par des « fonctionnaires », en charge non seulement d’écrire, mais aussi de contrôler et de gérer ; b) même si l’on relève un certain nombre de variations locales ou individuelles, ces « scribes » devaient de toute évidence recevoir une même formation, mais aussi la transmettre (une tablette au moins semble avoir eu un usage scolaire 49 ) ; c) le «-mycénien-» est donc un parler socialement typé-: c’est la langue de l’administration des royaumes achéens, et, comme le suggère la nature des documents et leur localisation, elle n’était vraisemblablement pas répandue en dehors des milieux du pouvoir ; les identifications de « scribes » donnent d’ailleurs des informations plus précises : les deux sites les plus importants, Cnossos et Pylos, n’ont livré respectivement qu’environ 70 et 30 scribes identifiés ; or, comme à cette époque la population de la Crète a pu être estimée entre 80 000 et 140 000 personnes, et celle de la Messénie entre 40 000 et 50 000 personnes 50 , il est évident que les «-scribes-» mycéniens ne formaient qu’un minuscule noyau de lettrés ; d) les textes mycéniens les plus récents sont antérieurs d’au moins quatre cents ans aux premières inscriptions alphabétiques grecques connues, qui datent du VIII e siècle avant J.-C., et il n’existe aucune trace de linéaire B à date post-mycénienne 51 ; ceci donne à penser que cette écriture, pur produit des palais mycéniens, a disparu en même temps que ceux-ci. Cela dit, cette forme de langue n’était pas artificielle, c’était une langue réellement employée dans le milieu des scribes  52 : on observe d’un site à l’autre, ou dans le même, des variations faibles, mais régulières, dans les usages graphi‐ ques, orthographiques, morphologiques ou syntaxiques ; ainsi le scribe 128 de Cnossos, à qui il arrive régulièrement d’écrire au même cas, genre et nombre un adjectif signifiant « en cuir », emploie trois formes différentes : wi-ri-ne-jo (ϝρῑνείοις [LB 14] : suffixe -ειο-) ~ wi-ri-ne-o (ϝρῑνέhοις [LB 13] : suffixe -ειο- > -εho-) ~ wi-ri-ni-jo (ϝρῑνίοις : suffixe -ιο-) ; ces formes constituent des 70 Monique Bile, René Hodot, Guy Vottéro <?page no="71"?> 53 Cf. Bartoněk 2003, chap. IX, p.-446-497. 54 Cf. les différentes théories énumérées par C. Milani, EAGLL. 55 Cf. Y. Duhoux, EAGLL, § 9. 56 La question de l’intercompréhension comme critère de différenciation entre langue et dialecte est très discutée en linguistique des langues modernes, car ce critère est à la fois relatif (il varie selon les connaissances linguistiques des individus, cf. infra) et subjectif (car il est souvent accompagné de considérations politiques, sociétales…), d’autant plus qu’il s’agit, quand on étudie les dialectes grecs anciens, de travailler sur des textes écrits. C’est ainsi que, pour répondre à cette question, on se réfère habituellement aux sources littéraires, notamment Hérodote (VIII, 144, cf. supra n. 43) qui évoque clairement l’unité de la langue grecque, et du monde grec, face aux Barbares. Mais est-on sûr que tout Grec de l’Antiquité était capable de comprendre Hérodote ? Dans sa présentation de « The Greek Notion of Dialects » (Verbum 1987), A. Morpurgo Davies mettait en garde contre cette notion d’intercompréhension (p.-8, 2 e §) et faisait part dans la note 4 de sa propre expérience : « as a native speaker of Italian, I can read Spanish, which I have never studied, but I cannot read Sicilian or Milanese, two Italian dialects, without the help of a translation ». Et il arrivait aux Grecs de l’Antiquité de se poser la question variantes morphologiques et phonétiques caractérisées, or elles sont écrites par une seule et même personne, dans des contextes rigoureusement identiques. Elles relèvent donc de la variation individuelle, observable dans toute forme linguistique vivante. Enfin, quand on compare le mycénien aux dialectes grecs du premier millén‐ aire, on constate qu’il présente certaines caractéristiques : a) étant donné la date de ces textes, des traits linguistiques plus archaïques que dans n’importe quel état de langue ultérieur hellénique, e.g. l’existence de labiovélaires, ou le suffixe comparatif en -josà tous les cas) ; b) des variations qui relèvent de la dialectologie, la plus évidente étant que le mycénien est un parler assibilant où *ti > si, et donc qu’il diffère d’une grande partie des dialectes connus ultérieurement ; il devait donc nécessairement exister entre 1400 et 1200 au moins deux dialectes ou groupes dialectaux, l’un assibilant comme le mycénien écrit des «-scribes-», l’autre non assibilant ; d’autres variations sont assurées 53 , mais elles sont plus difficiles à préciser étant donné le décalage chronologique entre le mycénien et les autres dialectes, mais aussi les imprécisions de l’écriture linéaire B, où les graphies ambiguës sont nombreuses ; il est donc difficile d’arriver à une caractérisation dialectale assurée du mycénien 54 . On peut même envisager l’hypothèse de sa disparition pure et simple sans descendance 55 . Il faut attendre le VIII e s. et surtout les trois siècles suivants pour rencontrer des formes linguistiques grecques impliquant un nombre bien plus élevé de personnes accédant à l’écrit. Et ce qu’on observe alors, c’est une multiplicité de variantes régionales, impliquant de nombreux locuteurs : les dialectes. Ces dialectes, plus ou moins inter-compréhensibles  56 et plus ou moins proches, Le grec ancien : une réalité multiforme 71 <?page no="72"?> du caractère plus ou moins barbare, c-à-d. incompréhensible, de tel ou tel dialecte, e.g. Athénée, Deipnosophistes, VIII, 42 (= 350 a), évoquant un célèbre cithariste athénien du IV e s., Stratonicos : πάλιν δ’ ἐπερωτηθείς [Στρατόνικος ὁ κιθαριστὴς], ὥς φησιν Ἡγήσανδρος [historien du II e s. av. J-C.], πότερα Βοιωτοὶ βαρβαρώτεροι τυγχάνουσιν ὄντες ἢ Θετταλοί, Ἠλείους ἔφησεν « comme on lui demandait qui des Béotiens ou des Thessaliens étaient les plus barbarophones, il répondit : les Eléens » ; en effet les Eléens, sur le territoire desquels se trouvait le célèbre sanctuaire d’Olympie, passaient pour barbarophones, principalement en raison du passage de *-s > -r en finale dans leur dialecte (cf. Eustathe, Comm. ad Il. B [v. 536-537], 279, 35-38). 57 Cf. Dialectes 1984. 58 Cf. Duhoux 1983, tableau p.-38. 59 Et certainement incompréhensible dans sa forme orale lors des représentations publiques : la grande majorité des citoyens athéniens, syracusains… qui assistaient aux représentations des tragédies attiques ou des odes pindariques, était certainement dans la même situation que l’Européen « moyen » actuel qui suit un opéra sans avoir à sa disposition un livret ou un prompteur ; ils étaient captivés par les décors, les jeux de scène, les chants, la musique ! nécessitent une étude spécifique 57 , car souvent ils ne sont pas attestés à la même époque 58 et n’ont pas eu la même importance ou le même destin au sein de leur famille linguistique et sont donc plus ou moins bien représentés épigraphiquement. Cette situation dialectale ne paraît pas avoir choqué les Grecs des époques archaïque et classique. On le constate à travers la littérature : les écrivains « grecs » jouent en effet très souvent des différences dialectales, e.g. Homère, Pindare, Hérodote… sans oublier les poètes tragiques athéniens dans les chœurs. On assiste donc à la création de langues artificielles (ne correspondant pas à la langue parlée au quotidien 59 ), les « langues littéraires », et présentant des traits principalement « épiques », « doriens », « éoliens », « ioniens », spécifiques à un ou plusieurs genres. Ainsi la « langue homérique » est utilisée pour la majorité des épigrammes et pour les diverses «-suites d’Homère-», l’«-ionien-» principalement pour les élégies, l’histoire et les sciences, le « dorien » pour la lyrique chorale ; au V e s. l’attique acquiert un statut de langue littéraire et vient perturber, sans la faire disparaître, cette distribution : il est utilisé pour la philosophie, l’histoire, l’éloquence, le théâtre (sauf les chœurs). Ce morcellement linguistique du monde grec surprend quand on le compare à la situation de l’Italie, mais en réalité c’est la situation de l’Italie qui est particulière : la réalité linguistique y est masquée par l’organisation centralisée de l’imperium romanum et la tradition littéraire ; seules les inscriptions lais‐ sent filtrer, à condition d’être très attentif, tel ou tel fait dialectal, comme le 72 Monique Bile, René Hodot, Guy Vottéro <?page no="73"?> 60 E.g. dans le SCBacch. (CIL Ι 2 , 581 / X, 104) les Gn. sg. nominus, senatuos (l. 7, 8, 17…), la préposition aruorsum (= adversum ; l. 24) ; dans la Lex Iulia (CIL Ι, 206 / ΙI 2 , 593), la séquence arieis (corrigée en ab ieis ; l. 148). 61 Musée Archéologique National de Florence, salle des grands bronzes. Inscription CIE 4196. 62 Préparé par le vaste mouvement de colonisation qui, à partir du VIII e s., a élargi considérablement le monde grec. Le contraste est d’ailleurs frappant entre Hésiode dont, selon ses propres dires, le plus long voyage a été la traversée des 5 kilomètres qui séparent Aulis et Chalcis, Epaminondas qui a conduit sa flotte jusqu’à Byzance et Cnide, et leurs descendants qui ont vu Philippe de Macédoine faire de l’Europe une des trois «-entraves de la Grèce-», que les Romains ont ensuite reprise à leur compte. 63 Elle se manifeste sous de multiples aspects, e.g. bipolarisation politique autour des régimes démocratique athénien et oligarchique spartiate, développement de structures fédératives, institutions communes comme la proxénie (et ses décrets aux formulaires fort semblables), exaltation des concours panhelléniques (ceux de la « période », qui acquièrent au V e s. un grand renom, cf. les Odes de Pindare, puis d’autres qui s’y ajoutent progressivement). 64 Cet alphabet reprend le tracé des lettres ioniennes et certaines évolutions, comme l’utilisation de H pour noter un e: , mais il est adapté à la phonologie de l’attique, cf. la distinction ε: ~ e: , ɔ: ~ o: . La date officielle de l’adoption de cet alphabet à Athènes est 403, sous l’archontat d’Euclide. 65 Cf. infra n. 71. A la fin du IV e s., le grec va donc s’écrire partout de la même manière. Sénatus-consulte des Bacchanales ou la Lex Iulia  60 , ou permettent d’imaginer la situation linguistique à Rome même, comme la statue dite « l’Orateur » (ou « Arringatore »), qui représente Aulus Metellus, un notable étrusque du début du I er s., vraisemblablement sénateur romain (toge à large bande, anneau à l’annulaire de la main gauche, bottines hautes de type mullei calcei), prêt à prendre la parole en public et affirmant sans complexe son « étrusquitude » dans l’inscription gravée sur la bande au bas de la toge 61 . Au fil des siècles, avec le développement des échanges 62 , la montée en puissance des « impérialismes » de Sparte, et surtout d’Athènes (qui se manifeste dans tous les domaines : politique, géographique, culturel et artistique), le monde des cités grecques s’ouvre progressivement et on voit se développer une uniformisation de plus en plus grande 63 . Sur le plan linguistique cela se traduit par a) l’adoption d’un même alphabet (dit « ionien-attique », car d’origine ionienne, mais adapté pour l’attique 64 ), qui remplace rapidement dans le courant du IV e s. tous les anciens alphabets épichoriques 65 , b) l’introduction dans les textes dialectaux de formules ou formulaires empruntés et l’apparition d’usages linguistiques nouveaux (e.g. décrets de proxénie, arbitrages entre cités avec recours à des juges grecs, mais extérieurs aux deux parties…). C’est ainsi qu’une partie de plus en plus importante des Grecs s’est mise à pratiquer plusieurs grecs, non seulement de façon passive en tant qu’auditeurs Le grec ancien : une réalité multiforme 73 <?page no="74"?> 66 Cf. La koinè grecque antique, vol. I-V ; Torallas Tovar S., EAGLL ; Teodorsson Sv.-T., EAGLL. 67 Sous le roi Archélaos 1 er la cour de Pella devient un important centre culturel grec. Mais l’« atticisation » de la Macédoine avait commencé plus tôt, cf. Brixhe - Panayotou 1988. 68 Cf. La koinè grecque antique, vol. I, p.-59-82. 69 Ainsi la «-démocratie-» athénienne qui, durant plus d’un siècle, avait été conquérante, a cédé la place à Athènes même à une oligarchie «-modérée-», d’abord imposée par les Macédoniens (avec Démétrios de Phalère), puis par les Romains. des épopées ou spectateurs des tragédies …, mais pour beaucoup activement dans leur vie quotidienne, par le commerce, les voyages, les relations entre cités… Et au courant du V e s., avec la création de l’empire maritime athénien on assiste au développement d’une forme linguistique supra-dialectale, une koinè à base d’attique et d’ionien, qui va se répandre dans les milieux politiques et intellectuels, mais aussi économiques 66 . Les premières régions touchées sont naturellement celles qui sont soumises à Athènes et qui relèvent du monde ionien (Eubée, Chalcidique, Propontide, îles de l’Egée). Puis, grâce au prestige culturel d’Athènes, cette koinè sera adoptée par la cour de Macédoine 67 et une partie de l’Asie Mineure (Lydie, Carie notamment 68 ). A côté de cette variété linguistique « haute » on peut supposer que s’est développée une forme populaire en vigueur dans le milieu maritime et commerçant, qui va jouer un rôle de plus en plus important avec l’évolution générale du monde grec à partir du IV e s. En effet, à la suite de la conquête macédonienne, l’uniformisation politique et sociale 69 de la Grèce va accélérer le processus de remplacement des dialectes par cette langue standardisée, forme simplifiée de l’attique, expurgé de ses particularités les plus marquantes (phonétique, cf. les séquences -sset -rsqui ont remplacé -ttet -rrdans θάλασσα, ἄρσην ; morphologie, cf. les composés en -λᾱος, -ου au lieu de la déclinaison « attique » -ληως, -ω, la flexion de type Σωκράτης, -ην, -ου au lieu de -η, -ους, les aor. en -α au lieu de la formation thématique, e.g. pour λέγω : εἶπα au lieu d’ εἶπον…). Et la conquête romaine achèvera ce processus. Sous l’empire romain elle est en effet devenue « la langue grecque » : langue littéraire, celle de toute la littérature grecque, langue politique, celle de toute la partie orientale de l’empire romain, langue éclésiastique, celle des Évangiles chrétiens, puis, après le Grand Schisme d’Orient, celle de l’Église orthodoxe, langue vernaculaire qui est la variante basse des formes précédentes. En effet, comme toute langue vivante, cette koinè s’est diversifiée dans le temps et dans l’espace, car elle était soumise à des variations locales influencées par le dialecte anciennement parlé, à l’évolution chronologique, aux distinctions sociales. 74 Monique Bile, René Hodot, Guy Vottéro <?page no="75"?> 70 Sans inclure les colonies, supposées utiliser le dialecte de leur métropole. 71 La syntaxe paraît moins touchée, car les inscriptions ont très souvent un caractère stéréotypé et formulaire (e.g. décrets, catalogues de noms, voire épigrammes…), voir cependant infra § 8. 72 Alphabets dits « épichoriques ». L’ouvrage de Jeffery, LSAG, présente 22 tableaux alphabétiques, avec des variantes ; Guarducci 1967 répertorie en deux tableaux 34 variétés (p. 381 sq.). Si, dans bien des cas, les variations n’empêchent pas la lecture, d’autres devaient être pour beaucoup de lecteurs extérieurs et ultérieurs une gêne (e.g. une partie du Péloponnèse), voire une énigme (Corinthe et sa zone d’influence, Gortyne). 73 L’emploi du terme «-isoglosse-» emprunté à la dialectologie des langues modernes est souvent abusif à cet égard, car les points où s’observe le fait en cause en grec ancien sont trop rares pour permettre le tracé d’une ligne. Il conviendrait de se borner modestement au terme de «-trait en commun-». C’est cette koinè qui, après une nouvelle période « dialectale » durant le Moyen Âge jusqu’au XIX e s., est à l’origine du grec moderne dans ses variantes haute, la katharevousa, et basse, la dimotiki. (4) Les dialectes grecs La multiplicité des dialectes grecs s’explique géographiquement, historiquement et culturellement: géographiquement le territoire de la Grèce est « émietté » en une série de massifs montagneux (80 % du territoire continental) et de plaines aux dimensions réduites, avec une péninsule, le Péloponnèse, plusieurs archi‐ pels dispersés (îles Ioniennes en mer Ionienne, Sporades, Cyclades, Lesbos et Dodécanèse en mer Égée) et trois grandes îles (la Crète, Chypre et Corcyre), sans compter les nombreuses colonies. Ces caractéristiques contribuent à l’isolement des populations et les conséquences de ces modes de vie ont été nombreuses, tant sur le plan politique que sur le plan linguistique ou culturel : Duhoux (1983, p. 38) identifie ainsi 23 variétés dialectales 70 , dont une qualifiée indistinctement de « dialectes du nord-ouest », présentant des particularismes affectant la phonétique, la morphologie, le lexique 71 , souvent même écrites dans un alphabet différent d’une région à l’autre 72 . La dialectologie du grec ancien ne se réduit pas à la classification des dialectes les uns par rapport aux autres, comme pourraient encore le donner à croire certains travaux récents. L’utilité de cette approche est limitée, car elle repose sur des données nécessairement restreintes, en raison des aléas de la documentation disponible, très inégalement répartie 73 , et sur des hypothèses et projections dans le passé antérieur à la documentation qui sont difficilement vérifiables. Dans son Einführung in die Griechischen Dialekte (Darmstadt, 1977), Rüdiger Schmitt s’était abstenu, pour sa description des dialectes, de les présenter par groupes ; il s’est contenté d’étudier successivement les grandes variétés, en Le grec ancien : une réalité multiforme 75 <?page no="76"?> 74 1. Nordwestgriechisch, 2. Saronisch, 3. Westargolisch, 4. Inseldorisch, 5. Kretisch, 6. Lakono-Messenisch, 7. Achaiisch, 8. Elisch, 9. Boiotisch, 10. Thessalisch, 11. Lesbisch, 12. Arkadisch, 13. Kyprisch, 14. Pamphylisch, 15. Ionisch, 16. Attisch, 17. Mykenisch : un ordonnancement qui reproduit grosso modo la classification généralement admise entre dorien éolien---arcado-chypriote et ionien-attique, sans la figer. 75 Cf. Dialectes 1984, p. 171. Cet article dérangeant a reçu un accueil … mitigé ; depuis sa publication, les dialectes grecs ont été mieux connus, par l’organisation de rencontres internationales et par la parution de plusieurs monographies importantes, de recueils collectifs dédiés à des thèmes spécifiques (v. ci-dessous) et de recueils critiques d’in‐ scriptions pour des dialectes particuliers (collection Paradeigmata en cours). 76 Cf. ἱμιτία «-moitié-», et ἡμίλαιον «-moitié de la récolte-» en Arcadie, V e s. les numérotant de 1 à 17 74 (p. 26 à 117) ; ce n’est qu’ensuite qu’il abordait la « Gruppierung der Dialekte » (p. 118 à 133). Cette dernière section du livre est une revue de la bibliographie remontant à la fin du XIX e siècle, revue au terme de laquelle l’auteur présente, sous forme de schémas, les cinq ou six façons différentes dont ont été rangés les groupes dialectaux, selon les points cardinaux : grec du l’ouest, grec du nord, grec de l’est, grec du sud, positions intermédiaires et combinaisons (p. 129-132). La multiplicité de ces arrangements signale à elle seule la fragilité des hypothèses sur lesquelles ils reposent. « Avant de comparer les dialectes, il importe de les connaître » 75 . De bien les connaître, de les étudier chacun pour lui-même ; cette démarche est alors profitable à la communauté des antiquisants dans son ensemble. B---Bénéfices de l’étude des dialectes (5) Lexique La prise en compte de l’épigraphie, dialectale ou non, enrichit considérablement le lexique grec. Le « Bulletin épigraphique » de la REG apporte chaque année son lot de « mots nouveaux et mots rares » qui appartiennent à toute sorte de domaines, à toutes les régions du monde hellénisé et souvent à une haute époque. Certains sont particulièrement pittoresques, tels ἱμιτολαχία « demi-part de terrain cultivée en légumes verts » (Larisa) 76 , φρυνοποπειον « réchaud-crapaud” (Béotie IV e ) et hομοσέπυοι « qui a la même huche » (Sélinonte V e s., première attestation du terme, cf. Aristote, Pol. I 2, 5 ὁμοσίπυοι « membres d’un même oikos » en Sicile) ; ou encore ces deux mots symétriques : ἄμναμμος « petit-fils » en cyrénéen et en crétois et πρόμαια «-arrière-grand-mère-» en Carie. D’autres renvoient aux institutions ou témoignent de faits de société : ἀρχερμηνεύς « traducteur en chef » ou « interprète » (Kolossai de Phrygie, date ? ), φασικός (νόμος) « (loi) sur la dénonciation » (Priène III e s.), 76 Monique Bile, René Hodot, Guy Vottéro <?page no="77"?> 77 Ils quittent la catégorie des éphèbes après leur formation militaire de base, pour être versés dans le corps des hoplites ou cavaliers. 78 Cf. διφθέρα mot utilisé par les Ioniens selon Hdt V 58. 79 Avec comme premier élément τήβεννα, nom étrusque de la toge (cf. DELG). 80 Le masculin κονχυλευπλυτής « teinturier en pourpre » (en Phénicie également) montre une jolie variante du nom de la «-pourpre-». 81 Cf. P. Chantraine, La formation des noms en grec ancien, Paris 1933, 356. 82 Ce surcomposé présente avec -ιρla forme lesbienne du radical de ἱερός (Hodot 1990, 235-237). hοδο̅τε̅ρες hapax « magistrats chargés des routes » (Cumes, époque archaïque), ἀργυραμοιβήιον hapax, « bureau de change » (Thasos V e s), ὁρκιστήρ hapax « celui qui fait jurer » (Naupacte (V e s.), ϝικατιϝετιεες « ceux qui sont âgés de 20 ans » 77 (Béotie, III e -II e s. ; Crète, Eleuterna, III e -II e s.), parfois liés à une époque bien précise : δικαστοφυλακία hapax, « garde des juges étrangers » (Cos, II e s. av.), τυμβοκλέπτης hapax, « pilleur de tombes » (Cyzique, I er s. ap.), ληστοφυλακικός hapax « -(navire) spécialisé dans la protection contre les pirates-» (Samothrace, I er s. av./ ap.). Noter encore, pour l’histoire de l’écriture, la présence à Olbia au V e s. av.-de l’hapax διφθέριον « peau servant de registre de commerce » 78 ; et pour les rela‐ tions entre Grecs et Romains, ces exemples de mots hybrides : συνορία «-limite de propriété » (Naples, II e s.ap.), sans doute un calque de confinium, σελλοφόρος hapax, « porteur de litière » (Aphrodisias), τηβεννοφόροι « porteurs de toges » (Larisa, vers 86/ 85) pour désigner les Romains 79 . La formation du féminin pour des noms de métiers ou d’êtres vivants se révèle diverse : κιθαρισ-τίς hapax, « cithariste femme » (Aspendos, date ? ), πλοκ-ίστρα « friseuse » (Tanagra, II e s. av.), πορφυροπώλ-ισσα « marchande de produits faits avec de la pourpre » 80 (Phénicie) ; dans la langue de l’élevage, ἀρνήαδες « brebis » et χιμαίραδες « chèvres » (traité frontalier à Aigai d’Eolide, IV e / III e s.) s’ajoutent à la série des noms de femelles pourvus du suffixe -άδ- 81 -: le premier est la ‘féminisation’ de ἀρνηός « bêlier » (DELG) et a servi de modèle pour former le second, qui désigne certainement lui aussi la femelle productive, sur χίμαιρα «-jeune chèvre née à la fin de l’hiver précédent-» (DELG). Le vocabulaire religieux est prolifique : δο̅ματοποιϝοί hapax « magistrats chargés de la construction du temple » (Argos, début IV e s), hεροποιϝοί hapax « magistrats chargés de la statue d’Héra » (id.), θυμελοποιοί « (collège de) fabricants d’autel » (Epidaure, vers 360), συν-ίρ-ωργοι 82 « co-desservants du culte » (Erésos IV e s.) ; ἀντικαταδεσμεύω « se livrer à un contre-envoûtement », hapax (Athènes, IV e s. av)-; σπονδιστήριον hapax, «-local à libation-» (Pisidie). Le grec ancien : une réalité multiforme 77 <?page no="78"?> 83 Le cœur de chacun de ces deux mots est un thème verbal. 84 De βάναυσος « artisan » (et en grec moderne « grossier, vulgaire » : v. DELG). Plutarque utilise de son côté la forme βαναυσουργία, Marc. 14. 85 Uniquement après un radical monosyllabique à voyelle brève, du type ἕμ-μεναι, δό-μεναι, Hodot 1990, 157-8. 86 Hodot 2008, 129-138. 87 Voire par « j’y jetterai un regard pour trouver », L. Bardollet, Homère. L’Iliade. L’Odyssée, Bouquins, 1995, 397. (6) Morphologie L’enrichissement apporté au lexique est aussi source de gains pour notre connaissance de la formation des mots. Le § 5 met en évidence l’inventivité des composés et de la suffixation. Extraits du « Bulletin épigraphique », on peut encore citer deux noms d’activité 83 : βαναυσ-εργασ-ἰα « activité artisanale » à Cyrène (IV e s.), qui allie composition et suffixation 84 , et μετα-σκλούτω-σις « renouvellement du pavement » à Aphrodisias, qui combine un préfixe et un suffixe. Le système verbal grec n’a pas cessé de se restructurer tout au long de son histoire, entre deux tendances contraires : • autonomie des thèmes aspecto-tem‐ porels les uns par rapport aux autres (ἔρχομαι ~ ἦλθον, λέγω ~ εἶπον ~ εἴρηκα, ὁράω ~ εἶδον ~ ὄπωπα) et prolifération des formations concurrentes pour un thème donné, dans un même dialecte ou d’un dialecte à l’autre (λείπω ~ λιμπάνω, ἔνεγκον ~ ἔνικα, ὄπωπα ~ ἑόρακα) ; • alignements divers qui modèlent un thème sur un autre (δείκνυμι remplacé par δεικνύω ; ἔνεγκα-εἶπα). Les conditions d’emploi de la finale d’infinitif actif athématique -μεναι en lesbien 85 rendent improbable, voire exclu, que les formes homériques du type ὀρνύμεναι, εἰπέμεναι, soient héritées : on peut les qualifier d’extensions artificielles à fonction métrique (Chantraine 2013, 490). De même, au moyen, la variété des désinences observées d’un dialecte à l’autre et à l’intérieur d’un même dialecte suggère qu’à l’aube du I er millénaire, le grec n’avait pas encore de forme fixée pour la 3 e personne du pluriel de l’impératif moyen (Hodot 1990, 162-163). Un cas particulier : comment des inscriptions en dialecte attique aident à ré-interpréter deux passages d’Homère 86 . En Iliade IX, 167, et en Odyssée 2, 294, apparaît le futur ἐπιόψομαι, généralement rapporté à la conjugaison de ὁράω et traduit par « voir, poser les yeux », d’où « choisir » 87 . Pourtant, si on avait vraiment là le composé d’ὄψομαι « je verrai », on n’expliquerait pas le maintien de la voyelle finale du préverbe : la forme attendue serait ἐπ-όψομαι, bien attesté par ailleurs. En fait, le « choix », sens que retient maint traducteur, implique l’intervention non de la vue, mais de la voix. Cela ressort de trois inscriptions attiques : 78 Monique Bile, René Hodot, Guy Vottéro <?page no="79"?> 88 Cf., chez A. Blanc---E. Dupraz éd., Procédés synchroniques de la langue poétique en grec et en latin, Bruxelles 2007, les contributions de C. Le Feuvre, « La reprise décalée, un procédé de renouvellement formulaire dans la poésie lyrique et épique grecque » (spéc. p. 128-35), et de F. Létoublon, « A propos de quelques répétitions non formulaires dans l’Iliade-». 89 À traduire donc par « j’interpellerai ». Un schéma comparable se retrouve en Od. 7, 322-4, cette fois avec le verbe « voir » : οἵ μιν ἴδοντο … ὅτε τε ξανθὸν Ῥαδάμανθυν ἦγον, ἐπoψόμενον Τιτυόν « ceux de nos gens qui l’ont vu, quand ils emmenèrent le blond Rhadamanthe voir Tityos » (ἐπoψόμενον présente l’élision régulière du préverbe devant l’initiale vocalique du radical). • IG II 2 , 1933 (Eleusis, ca. 323-320) : Τούσδε ἐπιώψ[ατ]ο ὁ ἱεροφάντης « Voici ceux <qu’>a désignés le hiérophante-», l. 1 • IG II 2 , 1934 (fin IV e s.) : [Ἱεροφάν]της Νουφράδου Περιθοίδης ἀνέ|[γρ]α[ψεν] τοὺς ἐπιοφθ[έ]ντας ὑφ᾽ ἑαυτοῦ « [Hiérophan]tès fils de Nouphradès du dème des Périthoïdes a enregistré les hommes désignés par lui-», l. 1-2 • IG I 1 3 , 3 (sur les fêtes d’Héraklès à Marathon, vers 490-480) : τριάκοντα ἄνδρ[ας ἐς] | τὸν ἀγο̃να ἐπιόφσασθ[αι ἐκ] | το̃ν ἐπιδέμο̅μ « (les agonothètes) désigneront pour le concours trente hommes parmi les visiteurs », l. 3-5. Il convient de reconnaitre dans ces trois textes la présence de la racine *wek w - / *wok w - / *uk w - « parler », celle d’où vient εἶπον ; c’est la présence ancienne du / w/ qui a maintenu l’hiatus entre la voyelle du préverbe et la voyelle du radical. Le texte même d’Homère le confirme : on sait qu’il aime jouer des mots et que les faits d’écho sont un des principes de composition de l’épopée 88 . Or, dans les ‘couplets’ que constitue chacun des deux passages cités ci-dessus, et qui expriment l’un et l’autre des commandements en vue d’un départ, ἐπιόψομαι est précédé, deux vers plus haut, d’un terme signifiant un enrôlement intimé oralement : κλητοὺς ὀτρύνομεν en Il. IX 167, ἑταίρους ἐθελοντῆρας συλλέξομαι en Od. 2, 294. Ces deux expressions préparent l’énoncé de ἐπιόψομαι, qui doit donc appartenir au même champ sémantique 89 . (7) Phonétique Les textes épigraphiques présentent dans ce domaine un net avantage sur les textes littéraires : ils sont parvenus jusqu’à nous sans intermédiaire, les graphies que l’on y trouve sont donc celles de leur(s) auteur(s) et de l’époque où ils vivaient. Nous avons là un témoignage pris sur le vif et la confrontation entre textes, épigraphiques comme littéraires, de la même période permet de tracer un état des lieux de la langue, à une époque donnée, dans une région donnée et dans un milieu social donné. Le grec ancien : une réalité multiforme 79 <?page no="80"?> 90 N.B. : pour les siècles sont utilisées les abréviations suivantes : d = début ; am = 1 ère moitié ; pm = 2 e moitié ; f = fin. Les graphies pour une même période sont classées par ordre de fréquence ; la multiplicité des graphies à l’époque tardive traduit non seulement le poids des traditions orthographiques, mais aussi l’existence vraisemblable de variations socio-linguistiques. Les exemples apparaissent dès la période mycénienne avec l’identification des différents «-scribes-» et de leurs variations (cf. supra § 3). Avec l’apparition de l’alphabet et les nouveaux supports utilisés (pierre, céramique cuite, métal), mais aussi l’élargissement de l’écriture à une partie bien plus importante de la population, il faut recourir à d’autres approches d’autant que les perspectives changent : il est devenu très difficile de définir des styles individuels (excepté pour la céramique qui peut être produite facilement en plusieurs exemplaires), on a très souvent affaire à de multiples intervenants (e.g. pour les décrets : initiateur —> discussion et décision collective —> secrétaire —> graveur —> éventuellement correcteur), mais en même temps il est possible de préciser des évolutions linguistiques plus larges et donc plus représentatives de la société. C’est ainsi que, dans une région comme la Béotie où la tradition graphique était ouverte à toutes les innovations, on peut très souvent noter avec une certaine précision les évolutions phonétiques. Les graphies pour la diphtongue ai et le *o: / en sont une bonne illustration 90 : - VI e am VI e - V e IV e III e III e f-II e am - ΑΙ, ΑΕ ΑΙ, ΑΕ, Α ΑΙ, Η Η, ΑΙ , Ε Η, ΑΙ, ΕΙ, Ε *ai ae æ: e: e: i: / e: - Ο Ο Ω Ω, ΟΥ Ω, ΟΥ, Υ *o: o: o: o: o: / u: u: Pour illustrer le stade évolutif atteint à partir du millieu du III e s., on peut citer des formes comme : a) (pour *ai) ἐπι τη ἱαρη γη (dat. sg. = *ται ἱαραι γαι, décret économique d’Acraiphia, Migeotte 1986 n° 16 B), εἱρεθεντας (= *αἱρεθέντας ; IThesp 41A, l. 9), φηνητη à côté de φηνειτη (= att. φαίνεται; décret religieux de Tanagra, Roesch in Roller, n° 87 A, l. 10 et 22), Δαμενετω (= *-αινετος acte d’affranchissement, IThesp 213, l. 22, III e f/ II e d) ; b) (pour *o: ) Σουκρατιος à côté de Σωκρατιος (= att. *Σωκράτους, actes d’affranchissement d’Orchomène IG VII 3200/ 3201, III e f), ἁυς (= ἁως, att. ἕως ; acte d’affranchissement de Chéronée, ZPE 29, 1978, 123-137, n° 7, III e f/ II e d) ou encore αὐτυς κη ἐκγονως (Acc. pl. = 80 Monique Bile, René Hodot, Guy Vottéro <?page no="81"?> 91 Cf. l’acte d’affranchissement d’Orchomène, IG VII 3201 cité supra ; mais dans l’acte IG VII 3200, les noms des mêmes personnages sont écrits Ἀντιγενιος Σωκρατιος, ἱαραρχιοντων Ἀγεισινικω Σουκρατιος, Σωσιβιω Πουθιλλιος. 92 Cf. A.J. Heisserer - R. Hodot, « The Mytilenean Decree on Concord », ZPE 63 (1986), p.-109-128 + pl. II-III . 93 Cf. Hodot 1990, p.-94 pour le a: , p.-83-84 + Tab. V, 4 (p.-111) pour *t+i. αὐτως, att. αὐτούς ; décret de proxénie, IThesp 19, l. 5, III e pm). Les variations graphiques étaient donc très importantes et touchaient tous les types de textes, notamment les textes civiques. Mais on constate également que les évolutions phonétiques du dialecte local préparent la koinè. En effet béotien et koinè se sont rejoints sur plusieurs points, comme l’évolution *ai > e: > i: , et la fermeture de o(: ) > u: . Ainsi la flexion des thèmes en -a(: ) en béotien, si éloignée a priori de celle de la koinè, a fini par présenter, sinon le même jeu d’alternance a(: ) ~ *e: (> i: ), du moins une situation qui pouvait rappeler celle de δόξα/ -ης, la flexion béotienne étant -Α/ -ΑΝ/ -ΑΣ/ -Η. De même la flexion des thématiques, caractérisée par le Gén. -ω et l’Acc. pl. -ως, rejoignait celle de la koiné -ου, -ους, et cela peut expliquer l’apparition de « formes de koinè » dans des textes de bonne facture dialectale par ailleurs, e.g. la séquence Ἀντιγενιος Σουκρατιος, ἱαραρχιοντων Σωσιβιου Πουθιλλιος, Ἀγεισινικω Σουκρατιος… 91 . Il suffisait donc parfois de changer de graphie pour obtenir de la koinè ! Et cette dernière est devenue peu à peu suffisamment familière à l’oreille de beaucoup de Béotiens pour que son adoption comme moyen de communication courant n’ait pas provoqué de bouleversement. Mais cette mutation a pris un certain temps, excepté dans les classes dirigeantes, pour qui la koinè représentait un moyen de pouvoir, en Béotie et hors de Béotie. Une évolution phonétique peut aussi avoir une autre signification ; ainsi l’hapax δαμοκρασία que l’on rencontre dans le décret de Mytilène Sur la concorde  92 est intéressant à plusieurs titres, car il montre à la fois l’influence des idées politiques athéniennes qui se concrétise par l’emprunt du terme tout à fait symbolique de δημοκρατία, et en même temps la réaction linguistique des Mytiléniens qui adaptent cet emprunt lexical à la phonétique de leur dialecte : a: et assibilation de *t+i  93 . (8) Syntaxe En syntaxe, les différences dialectales sont plus ténues, moins faciles à illustrer. L’une des plus patentes est la variation, d’un dialecte à l’autre, des conjonctions Le grec ancien : une réalité multiforme 81 <?page no="82"?> 94 Noter la variation de l’ordre des mots. 95 Hodot 2021. 96 Il en va de même couramment en béotien à date ancienne, mais le suffixe n’a pas exactement la même forme : -ios et non -eios : e.g. Καλιαια ἐμι [ἁ κυλιξ]. 97 Récemment est apparu un premier exemple en lesbien d’adjectif andronymique, au sud de la Troade, à Assos, dans une dédicace à Artémis (IV e s.), qui de plus enchaine deux AP : Φαΐλα Μεγιστιαία τῶ Ἐχερτιδαίω « Phaïla (fille de) Mégistos, le (fils) d Echertidas » (T. Özhan, ZPE 221, 2022, 83-86). 98 Syntaxe des inscriptions dialectales d’Argolide (VII e s. av. J.-C.- II e s. ap. J.-C.), Etudes anciennes 79, Nancy-Paris. 99 Elle est fondée sur les espèces de mots : chapitre I, Nom, adjectif et pronom ; ch. 2, Prépositions-; ch. 3, Verbe-; ch. 4, Conjonctions-; ch. 5, Adverbes et particules. et particules modales : αἴ δέ κε/ κα ~ ἐὰν δέ 94 , les deux tournures étant l’une et l’autre suivies du subjonctif. Pour noter à côté d’un nom de personne le nom du père, le grec disposait de deux procédés : le nom du père au génitif (GA : génitif adnominal) ou un adjectif dérivé du nom du père (AP : adjectif patronymique). Comme l’usage des AP apparaît répandu en thessalien, béotien et lesbien, cette particularité syntaxique est souvent invoquée pour établir le rattachement du thessalien, du béotien et du lesbien à un groupe « éolien ». Pourtant, la diversité des formes dans les poèmes homériques montre déjà à elle seule que l’héritage est moins simple qu’il ne peut paraître : Ζεὺς Κρονίδης ~ Δία Κρονίωνα, Achille Πηληιάδης ~ Πηλεΐδης ~ Πηλεΐων, Νέστωρ Νηλήιος ~ Νέστορ Νηληιάδη 95 . Plus tard les formes et les conditions d’emploi des AP varient d’un des trois dialectes « éoliens » aux deux autres. En thessalien (cf. Morpurgo 1968), en plus de leur fonction patronymique, les adjectifs dérivés de noms de personne ont conservé l’emploi d’adjectifs d’appartenance, du type Ἀνφιονεία ἀ στάλα « la stèle appartenant à Amphion » (Bechtel 1921, 199) 96 . En béotien, G. Vottéro (1987) a montré que l’usage généralisé du patronyme, et donc de l’AP, est relativement tardif et lié à des conditions politiques particulières. En lesbien, l’AP fonctionne comme « une catégorie lexicale dotée d’une formation propre et ayant (…) une sphère d’emploi très délimitée-» (Hodot 1990, 222 97 ). Une première somme sur la syntaxe d’un dialecte particulier a été publiée en 2021 par E. Nieto Izquierdo 98 . D’une organisation traditionnelle 99 , l’ouvrage se conclut par un chapitre 6 intitulé « Apport de l’analyse syntaxique des inscriptions à la classification dialectale de l’Argolide » ; dans la section « Traits pertinents pour la comparaison », a) « emploi systématique de l’article avec les noms de divinité à l’époque archaïque », avec cette prise de position : «-l’article [étant] de création postmycénienne, […] il serait difficile de faire remonter un tel usage à l’époque de la communauté dorienne. Pour résoudre la question, 82 Monique Bile, René Hodot, Guy Vottéro <?page no="83"?> 100 Klincksieck, 3 ème éd., 1960, p.-117. 101 Italiques de J. Humbert. il faudra étudier minutieusement les données archaïques, et faire des relevés exhaustifs pour trouver des dialectes qui pourraient aussi partager ce trait. » Dans l’étude attentive du participe qui occupe les pages 302 à 324, E. Nieto consacre « quelques mots » à « la prétendue expression du temps relatif accordée par de nombreux savants au participe grec » ; en fait, le participe (« sauf futur ») et l’infinitif « expriment seulement l’aspect […], la temporalité relative entre les différents évènements ne ressortant que d’inférences qui n’ont pas à voir avec l’opposition présent ~ aoriste-». Deux recueils d’articles portant sur les dialectes ont été consacrés naguère à des questions de syntaxe-: • Les modes dans les dialectes grecs anciens, R. Hodot éd., Verbum 23-3, 2001. Mérite une mention particulière la contribution de C. Dobias-Lalou, « De l’optatif comme mode de luxe dans l’usage épigraphique » (p.-269-280). On sait que l’optatif au cours du temps a perdu du terrain par rapport au subjonctif ; comme l’indiquait Jean Humbert dans sa Syntaxe grecque  100 , déjà chez Homère, l’optatif a servi à «-exprimer une étroite subordination entre un verbe dépendant et un verbe principal qui se situent tous les deux dans le passé », extension d’emploi qui, tendant « à vider l’optatif de son sens propre pour en faire un signe de subordination, a beaucoup contribué a précipiter le déclin de ce mode, qui s’est fondu dans le subjonctif 101 ». Quant à elle, C. Dobias-Lalou relève « l’extrême rareté, voire l’absence totale de l’optatif dans certains corpus épigraphiques dialectaux » (p. 269). Si la prépondérance du subjonctif est généralement liée à la nature des documents, « la prise en compte des situations d’énonciation » permet d’expliquer certains des emplois de l’optatif. « De portée essentiellement pragmatique, mis au service de la sémantique pour le renforcement des contenus lexicaux, au service de la stylistique pour des effets de décalage expressif, voire au service de la rhétorique dans la levée de l’équivoque ou la recherche de la uariatio, il peut être absent des inscriptions de tel dialecte pour la simple raison que la documentation conservée ne se prêtait pas à son emploi. Dès lors, c’est plus exactement la double modalisation virtuelle qui apparaît comme un luxe. Mais la langue n’en a pas fait l’économie aussi vite qu’on l’a dit parfois. » (conclusion, p.-279-280). • Dialectes grecs et aspect verbal, R. Hodot et G. Vottéro éd., Etudes anciennes 35, 2008. Pour le traitement statistique du texte et la portée générale des conclusions, nous distinguerons la contribution d’Yves Duhoux, « L’aspect verbal en messénien : le règlement des Mystères d’Andanie (92-91 avant Le grec ancien : une réalité multiforme 83 <?page no="84"?> 102 De ce point de vue la perte de la quasi-totalité des études d’Aristote sur les constitutions des cités grecques est un désastre. 103 Comme le montrent clairement les modèles géographiques utilisés pour la Thessalie, cf. Y. Auda - L. Darmezin - J.-C. Decourt - Br. Helly - G. Lucas, « Espace géographique et géographie historique en Thessalie », Actes des Xèmes Rencontres internationales d’archéologie et d’histoire, Juan-Les-Pins, 1990. 104 Cf. P. Cabanes, « Cité et ethnos dans la Grèce ancienne », Mélanges Pierre Lévêque II, p.-63-82, Besançon, 1989. 105 Cf. Br. Helly, L’Etat thessalien. Aleuas le Roux, les tétrades et les tagoi, Lyon, 1995. J.-C. : Schwyzer 74) », p. 57-90. « Une même action, celle d’écrire, dans des contextes strictement comparables (l’organisation des Mystères), peut être et est effectivement exprimée par les trois aspects verbaux. […C]hacun d’entre eux est utilisé à bon escient. […] L’acte d’«-écrire-» aboutit forcément toujours au résultat qu’est le document écrit. Mais le parfait s’oppose à l’aoriste et au présent en ce que c’est lui, et lui seul, qui indique explicitement que l’action verbale a débouché sur un résultat […]. La structure du système aspectuel est rigoureusement conforme à celle de l’époque classique. […C]’est l’aspect d’état qui est le terme marqué du système aspectuel et s’oppose au bloc formé par les aspects progressif (thème de présent) et ponctuel (aoriste). […Cf.] la manière dont le présent et l’aoriste s’opposent : le présent présente l’action verbale de façon plus concrète ou palpable que l’aoriste, qui la rend de manière plus abstraite. Ceci montre bien que c’est le présent qui est le terme marqué du couple aspectuel présent ~ aoriste.-» (p.-86-87). (9) Les emplois Aux variations dialectales correspondent souvent des situations politiques particulières 102 . En fait, du IX e au V e s. et au-delà, l’histoire « grecque » est la juxtaposition de nombreuses histoires « locales ». Si le modèle politique dominant en Grèce est celui de la cité, entité vivant en relative autarcie et organisée autour d’un centre urbain 103 , dans certaines régions (e.g. N et N-O. de la Grèce) ce sont les ethnè qui dominent, entités bien plus larges et adaptées à un mode de vie différent 104 , et il existe d’autres modèles. Par exemple, au V e s., l’unité de l’Attique ou de la Laconie contraste avec, en Béotie, un système confédéral laissant à chaque cité suffisamment d’autonomie, mais permettant aussi de faire face collectivement aux ambitions des voisins ; en Thessalie la situation est encore différente, avec son organisation en tétrades, poleis et kléroi  105 . Et, comme on l’a vu supra (§ 3) l’uniformisation linguistique apportée par la koinè va de pair avec l’uniformisation politique et sociale de la Grèce. Les dialectes avaient à l’origine et ont eu pendant longtemps une fonction de communication, comme toute forme de langage. Mais, avec l’apparition, puis 84 Monique Bile, René Hodot, Guy Vottéro <?page no="85"?> 106 C’est la forme d’alphabet épichorique le plus pauvre en caractères (18 lettres) et on ne le rencontre que dans ce texte, cf. Bile 1988, p.-76-77. 107 Cf. E. Weiss, Les Tables d’Héraclée. Εtude historique et linguistique, Nancy, 2016, notamment p.-77-78 et 109-110. 108 Cf. La koinè grecque antique vol. V. 109 Nouvelle édition : G. Biard - Y. Kalliontzis - A. Charami, « La base des Muses au sanctuaire de l’Hélicon-», BCH 141 (2017), p.-697-752. 110 Cf. Vottéro, « Koinas et koinès en Béotie », La koinè grecque antique II, p. 80-81, Nancy, 1996. 111 Cf. R. Van Bremen R., « The date and context of the Kymaian decrees for Archippe (SEG 33, 1035-1041) », REA 110 (2008), p. 357-382 et R. Hodot, « Dialecte, koinè, latin… Etat l’extension de la koinè, qui a concurrencé de plus en plus les différents dialectes, on voit apparaître d’autres emplois. Ainsi les dialectes peuvent avoir une valeur symbolique, voire identitaire. Car, étant la forme linguistique la plus ancienne, ils sont souvent liés à des pratiques sociales anciennes, qui, confrontées aux évolutions des sociétés, peuvent aboutir à des crispations internes et se traduire dans des textes « réactionnaires », comme les Lois de Gortyne gravées au V e s. sur un pan de mur monumental, dans un alphabet archaïque et spécifique 106 . Le choix d’écrire en dialecte peut aussi apparaître comme la volonté de revenir à une situation ancienne de référence : ainsi les Tables d’Héraclée, rédigées dans une langue qui ressemble au dialecte de la métropole, Tarente, comportent des références claires au passé (καθὼς τὸ ἀρχαῖον / κὰτ τὰ ἀρχαῖα), mais les nombreux koinéismes qu’on y trouve montrent que le dialecte ancien n’était plus parlé 107 . On observe également des cas intéressants de changement de code à l’intérieur d’un même texte : le dialecte est employé parallèlement à la koinè, mais on constate que généralement la koinè est utilisée dans les passages essentiels, comme le contrat liant une cité à sa créancière, la titulature des magistrats présidant à des rites 108 … Les dialectes peuvent enfin avoir une fonction « décorative » : ainsi deux textes béotiens du Ι er s. av. J-C, IG 3195 (palmarès des Charitésia d’Orchomène) et IThesp 288-297 109 (bases des Muses), cherchent à rappeler la notoriété ancienne de la région en faisant revivre le dialecte disparu depuis deux siècles 110 . A Kymè d’Eolide au II e s., sur le monument qui célébrait la bienfaitrice Arkhippè, on a gravé six décrets relatifs aux divers engagements qu’elle avait pris : rédigés en koinè, ils témoignent que cet idiome était parfaitement maitrisé par la classe dirigeante de la cité ; deux autres décrets occupent la place la plus en vue : postérieurs à la réalisation de ces promesses, on a eu recours pour eux au dialecte, comme une distinction supplémentaire — mais ce n’est qu’un habillage manifeste de textes en koinè 111 . A Kymè encore, sous Auguste, deux longs Le grec ancien : une réalité multiforme 85 <?page no="86"?> des lieux-», Actes du colloque sur «-L’Eolide dans l’ombre de Pergame-», Topoi, Suppl. 14, 2016, 36-37. 112 IK 5, 19. 113 SEG XXXII, 1243, l. 21 et 26-27. décrets honorent chacun un prytane de la cité : dans l’un, le Romain L. Vaccius Labeo est gratifié du titre de Philokymaios  112 , dans l’autre, est glorifiée la lignée de Kléanax, à la fois qualifié de kalliteknos (il a donné à son fils une bonne éducation) et autorisé officiellement à ajouter Philopatôr à son nom 113 ; la langue des deux textes qui voudrait être le dialecte s’en écarte à plusieurs reprises, c’est l’intention qui comptait. Conclusion L’attico-athénocentrisme est souvent une forme de myopie qui fausse le juge‐ ment : par ex. on a longtemps rangé Pindare (518-438) dans la poésie archaïque, alors que, né 7 ans après Eschyle, toute son œuvre appartient au V ème siècle : les odes se sont échelonnées de 498 jusqu’en 446, 22 ans après la première victoire de Sophocle. Et les aristocrates qu’il célèbre (dont l’Athénien Mégaclès, Pythiques 7, en 486) ne sont pas différents de ceux qui, au temps de la démocratie triomphante, continuent à jouer dans la politique athénienne un rôle longtemps discret, puis déterminant, manifesté par la réaction oligarchique des années 411 et 404 ; leurs descendants seront aux siècles suivants les acteurs de l’évergétisme. Cette évolution aboutit à l’inclusion du monde grec, depuis le domaine continental jusqu’aux colonies les plus éloignées, en passant par l’Asie Mineure, dans un vaste ensemble méditerranéen, l’empire romain — où les questions linguistiques restent cruciales, avec l’Orient grec et l’Occident romain (et les ‘bureaux’ impériaux correspondants), avec les réactions ‘localistes’ (carrières romaines de notables provinciaux qui renchérissent sur leurs origines). Et quand la koinè a été généralisée, le souci de distinction sociale pousse à s’en démarquer : c’est ainsi que peuvent s’expliquer, dans la continuité des alternances codiques relevées dès le IV e s., les résurgences de certains dialectes, et en partie également le phénomène littéraire de l’atticisme qui se développe aux premiers siècles de l’Empire. Références Bibliographiques Sources antiques Auteurs anciens (Homère, Nonnos de Panopolis, Pausanias, Strabon, Thucydide, Tite-Live) : éditions habituelles des collections CUF, Loeb, Teubner. 86 Monique Bile, René Hodot, Guy Vottéro <?page no="87"?> CoMIK : J. Chadwick - L. Godart - J.T. Killen - J.-P. Olivier - A. Sacconi - I.A. Sakellarakis, Corpus of Mycenaean Inscriptions of Knossos, Rome-Cambridge, 5 vol. 1986-1998. Eustathe, Comm. ad Il. : G. Stallbaum, Commentarii ad Homeri Iliadem, Leipzig, 1927. IG VII : G. Dittengerger, Inscriptiones Graecae, VII Inscriptiones Megaridis et Boeotiae, Berlin 1892. IThesp. : P. Roesch, Inscriptions de Thespies (13 fasc.), édition électronique récolée, contrôlée, préparée et mise en forme par G. A. Argoud - A. Schachter - G. Vottero, MOM-Lyon 2007 (rév. 2009). LGPN : Lexicon of Greek Personal Names en ligne, <https: / / www.lgpn.ox.ac.uk >, Oxford TLG : Thesaurus Linguae Graecae en ligne, <http: / / stephanus.tlg.uci.edu>, Univ. of California, Irvine. Sources modernes Bartoněk 2003 : A. Bartoněk, Handbuch des mykenischen Griechisch, Heidelberg. Bechtel 1921-: F. Bechtel, Die griechischen Dialekte I ( 2 Berlin, 1963). Bile 2016 : M. Bile, Paradeigmata. Recueil d’inscriptions grecques dialectales VI, 1. La Crète, Nancy. Brixhe---Panayotou 1988 : Cl. Brixhe---A. Panayotou, «-L’atticisation de la Macédoine : l’une des sources de la koiné-», Verbum XI 3-4, 245-260. Chantraine 2013-: P. Chantraine, Grammaire homérique I- Phonétique et morphologie, Paris 1942, nouvelle édition revue et corrigée par M. Casevitz. D’Ercole-Zurbach 2019-: M.-C. D’Ercole---J. Zurbach, Naissance de la Grèce. De Minos à Solon, 3200 à 510 avant notre ère, Paris. Dialectes 1984 : M. Bile, C. Brixhe, R. Hodot, «-Les dialectes grecs, ces inconnus-», BSL 79, 155-203. Duhoux 1983 : Y. Duhoux, Introduction aux dialectes grecs anciens. Problèmes et méthode. Recueil de textes traduits, Louvain-la-Neuve. Duhoux 2013 : Y. Duhoux, Le mycénien. Paradeigmata, recueil d’inscriptions grecques dialectales I, Nancy. Duhoux, EAGLL : Y. Duhoux, «-Mycenaean Script and Language-». Duhoux-Morpurgo Davies : Y. Duhoux - A. Morpurgo Davies, A Companion to Linear B, 3 vol. 2008-2014, Louvain-la-Neuve. EAGLL : G. K. Giannakis éd., Encyclopedia of Ancient Greek Language and Linguistics, Leyde, 2014. Goukowsky---Feyel 2019 : P. Goukowsky---Chr. Feyel, Le profil d’une ombre. Etudes sur les Helléniques d’Oxyrhynchos, Nancy [texte, traduction, commentaires philologique et historique]. Goukowsky 2020 : P. Goukowsky, Les Helléniques d’Oxyrhynchos. Complément d’enquête, Nancy. Le grec ancien : une réalité multiforme 87 <?page no="88"?> Guarducci 1967-: M. Guarducci, Epigrafia greca I, Rome. Hodot 1990 : R. Hodot, Le dialecte éolien d’Asie. La langue des inscriptions VII e s. a.C. - IV e s. p.C., Paris. Hodot 2001 : Les modes dans les dialectes grecs anciens, R. Hodot éd., Verbum 23-3, Nancy. Hodot 2005 : R. Hodot, «-Sur la dialectologie du grec ancien-: quelques tautologies-», in F. Poli - G. Vottéro éd., De Cyrène à Catherine : Trois mille ans de Lybiennes (Mélanges C. Dobias-Lalou), EA 30, Nancy, 111-18. Hodot 2008-: R. Hodot, «-Aspect et morphologie. Le cas de ἐπιόψασθαι-», in Hodot-- -Vottéro 2008. Hodot 2021-: R. Hodot, «-Langue homérique et dialectologie : à propos de l’expression des patronymes dans l’Iliade-», in M. Bile---R. Hodot---G. Vottéro éd., Questions de dialectologie grecque, EA 77, 133-169. Hodot - Vottéro 2008 : R. Hodot et G.Vottéro (éd.), Dialectes grecs et aspect verbal, Nancy. Hyettos 1976 : R. Etienne - D. Knoepfler, Hyettos de Béotie et la chronologie des archontes fédéraux entre 250 et 171 av. J.-C., Athènes-Paris. Jeffery, LSAG : J.H. Jeffery, The Local Scripts of Archaic Greece 2 , Oxford 1989. Kalliontzis 2020 : Y. Kalliontzis, Contribution à l’épigraphie et à l’histoire de la Béotie hellénistique : de la destruction de Thèbes à la bataille de Pydna, Athènes-Paris. Koiné grecque antique : Cl. Brixhe éd., La koiné grecque antique, I- Une langue introuvable ? EA 10, 1993 - II- La concurrence, EA 14, 1996 - III- Les contacts, EA 17, 1998 ; R. Hodot éd. IV- Les koinés littéraires, EA 24, 2001---V- Alternances codiques et changements de code, EA 29, 2004, Nancy. Lupu 2009 : E. Lupu, Greek Sacred Law. A Collection of New Documents, Leyde-Boston. Migeotte 1986, L'emprunt public dans les cités grecques, Québec-Paris. Milani, EAGLL : C. Milani, «-Variation in Mycenaean Greek-». S. Mimbrera Olarte, EAGLL : S. Mimbrera Olarte, «-Magna Graecia, dialects-». Olivier-Melena-Piteros 1990 : J.-P. Olivier - J. L. Melena - Chr. Piteros, « Les inscriptions en linéaire B des nodules de Thèbes (1982) : la fouille, les documents, les possibilités d’interprétation-», BCH 114, 101-184. Morpurgo 1968-: A. Morpurgo Davies, «-Thessalian Patronymic Adjectives-», Glotta 46, 85-106. Pouilloux 2003-: Choix d’inscriptions grecques, sous la direction de Jean Pouilloux, Paris. Rebuffat 1996-: F. Rebuffat, La monnaie dans l’Antiquité, Paris. Richer 2018-: N. Richer, Sparte, cité des arts, des armes et des lois, Paris. Schwyzer 1923-: E. Schwyzer, Dialectorum graecarum exempla epigraphica potiora, Leipzig. Sokolowsky 1955 : F. Sokolowsky, Lois sacrées de l’Asie Mineure, Paris. Sokolowsky 1962 : F. Sokolowsky, Lois sacrées des cités grecques. Supplément, Paris. Sokolowsky 1969 : F. Sokolowsky, Lois sacrées des cités grecques. 88 Monique Bile, René Hodot, Guy Vottéro <?page no="89"?> Torallas Tovar, EAGLL : S. Torallas Tovar, «-Koine, Features of-». Teodorsson, EAGLL : Sv. T. Teodorsson, «-Attic-», § 4-6. Verbum 1987 : Cl. Brixhe (dir.), Actes de la 1ère rencontre internationale de dialectologie grecque, Nancy. Vottéro 1987 : G. Vottéro, «-L’expression de la filiation en béotien-», Verbum 10 (1987), p.-211-231. Le grec ancien : une réalité multiforme 89 <?page no="91"?> 1 L. Del Monaco 1997. La signature des hiéromnémons (À propos des tablettes de Locres n° 15 et 26) Emmanuel Weiss (Université de Lorraine) Abstract: The tablets of Locris are a group of accountancy tablets from Epizephyrian Locris going back to the end of the 4 th century BC. Two of them (tablets n°15 and 26) present ligatures which have been variously interpreted. This paper proposes a new decoding of them. They could correspond to a form of abbreviation corresponding to the title of the magistrates who are responsible for accounting operations, that is, the hieromnemons. Keywords: Epigraphy; Tablets of Locris; Epizephyrian Locris; hieromne‐ mons; Ligature. 0. Les Tablettes de Locres sont un ensemble de 39 tablettes de bronze (37 si on accepte la reconstitution de L. Del Monaco 1 , qui a proposé de voir dans les tablettes n° 35, 36 et 37 les fragments d’un même objet - tablette désignée ci-dessous par le sigle DM) découvertes en 1958 sur le site de Casa Marfiotti. Elles sont datées de la fin du quatrième siècle av. J.-C. et transcrivent une série d’opérations comptables impliquant la cité de Locres Épizéphyrienne et le temple de Zeus de cette même cité. De nombreux magistrats sont nommés dans les tablettes, dont entre autres les hiéromnémons, qui gèrent le trésor du temple. Dans le présent article nous nous intéressons aux différentes variantes du nom des hiéromnémons présente dans le corpus des tablettes, qui témoignent par leur diversité de la pénétration de la koinè à Locres Épizéphyrienne. Au passage cet examen nous permettra de résoudre un petit mystère épigraphique concernant les tablettes n° 15 et 26. <?page no="92"?> 2 J. Blomqvist 1975 et 1978. I Le nom des hiéromnémons Le dialecte de Locres Épizéphyrienne a été étudié de manière magistrale par J. Blomqvist dans deux articles parus dans les années 1970 2 . Il y a démontré que ce dialecte relève du dorien sévère, alors qu’on aurait pu s’attendre à une forme de dorien doux, en raison de l’origine locrienne des fondateurs de la cité. Pour expliquer cette anomalie, il a été amené à l’hypothèse que la langue des Tablettes de Locres relève en fait d’une koina dorienne sévère qui se serait substituée dans le sud de la Grande Grèce aux dialectes initiaux de chaque cité, en raison peut-être de l’influence politique de Tarente. Quoi qu’il en soit, les différentes formes du nom des hiéromnémons dans le corpus des tablettes s’explique aisément par la concurrence entre une forme dialectale (Ͱιαρομνάμονες) et la forme de la koinè (ἱερομνήμονες). La forme dialectale se caractérise-: - par la conservation de l’aspiration initiale (notée Ͱ)-; - par le vocalisme [a] du thème de l’adjectif « sacré » attesté dans le corpus des tablettes au génitif pluriel féminin Ͱιαρᾶν (n° 23, ligne 10 ; n° 30, ligne 10) ou ἱαρᾶμ, sans notation de l’aspiration initiale, et avec assimilation de la consonne finale à la consonne initiale du mot suivant μίστωμα (n°31, ligne 9)-; - par l’absence du passage de [a-: ] à [ε-: ]. Cette forme est attestée une fois dans le corpus des tablettes, au nominatif pluriel (n°14, ligne 1). La forme de la koinè se caractérise par-: - sans doute la perte de l’aspiration initiale, qui n’est pas notée-; - le vocalisme [e] du thème de l’adjectif «-sacré-» (ἱέρος)-; - le passage de [a-: ] à [ε-: ]. Cette forme n’est pas attestée directement dans le corpus des tablettes. La majorité des formes attestées relève en fait de formations «-mixtes-». Tout d’abord la forme ἱαρομνάμονες, sans notation de l’aspiration initiale mais avec maintien des autres caractéristiques dialectales, est attestée six fois dans le corpus au nominatif pluriel (n° 13, ligne 6 ; n° 22, ligne 6 ; n° 25, ligne 4 ; n° 32, ligne 3 ; n° 33, ligne 4 ; n° 38, ligne 1 ; DM, ligne 5), deux fois à l’accusatif pluriel (n° 4, ligne 7 ; n° 32, ligne 1), et trois fois indirectement, par le biais du verbe dérivé ἱαρομναμονέω (n° 8, ligne 3-; n° 23, ligne 1-; n° 30, ligne 5). 92 Emmanuel Weiss <?page no="93"?> 3 A. De Franciscis 1972, p.-129. 4 F. Costabile 1992, pp. 259 et 281. La variante Ͱιερομνάμονες, comportant à la fois le signe de l’aspiration initiale et le vocalisme de la koinè, apparaît trois fois au nominatif pluriel (n° 12, lignes 1 et 2-; n° 15, ligne 1-; n° 26, ligne1). Enfin, la variante ἱερομνάμονες, reprenant dans la première syllabe le vocal‐ isme de la koinè, sans notation de l’aspiration initiale, est attestée dix fois au nominatif pluriel (n°2, ligne1 ; n° 3, ligne 6 ; n° 7, ligne 1 ; n° 7, ligne 9 ; n° 16, ligne 4 ; n° 17, ligne 1 ; n° 24, ligne 5 ; n° 27, ligne 6 ; n° 28, lignes 4 et 5 ; n° 29, ligne 6), une fois à l’accusatif pluriel (n°5, lignes 7 et 8), une fois au génitif pluriel (n° 5, ligne 4), une fois au datif pluriel (n° 20, ligne 17), et six fois indirectement, par le biais du verbe dérivé ἱερομναμονέω (n° 6, ligne 2 ; n° 10, lignes 1 et 2 ; n° 11, lignes 1 et 2-; n° 18, ligne 1-; n° 19, ligne 1-; n° 29, lignes 1 et 2). II Les ligatures des tablettes n° 15 et 26 Au bas des tablettes n° 15 et 26 apparaissent deux «-ligatures-»-: pour la tablette n° 15, ligne 16, pour la tablette n° 26, ligne 12. Dans son édition de 1972 (la première édition complète des Tablettes), A. De Franciscis 3 a interprété ces deux ligatures comme le début de deux noms de mois, servant à la datation de ces deux tablettes, respectivement Πατρ… (tablette n° 15) et ᾽Αγρ… (tablette n° 26). Cette interprétation est reprise par F. Costabile dans son édition de 1992 4 , mais soulève plusieurs problèmes. Tout d’abord les tablettes sont le plus souvent non datées, et lorsqu’elles le sont (tablettes n° 20 et 33), la date figure en début de texte et non à la fin. Ensuite, si un mois d’ ᾽Αγρήιος est bien attesté dans le corpus des Tablettes (n° 13, ligne 11 ; n°22, ligne 17 ; n° 31, ligne 7 ; n° 34, ligne 10), il n’y a aucune autre trace d’un nom de mois commençant par Πατρ… La signature des hiéromnémons 93 <?page no="94"?> 5 F. Ghinatti 1998. Une autre interprétation a été proposée par F. Ghinatti en 1998 5 , proposant de voir dans les deux ligatures les trois premières lettres d’un nom de magistrature, les πράκτορες, des «-collecteurs-». Le problème est que ce titre ne figure nulle part ailleurs dans le corpus des tablettes. Nous proposons de notre côté de voir dans ces deux ligatures les trois premières lettres (ΙΑΡ) du nom dialectal des hiéromnémons (Ͱιαρομνάμονες), la ligne horizontale joignant ces trois lettres provenant peut-être du signe Ͱ notant l’aspiration initiale. En effet, dans la tablette n° 15 la ligature est précédée, sur la même ligne, par les trois lettres Ζ, Ν et Ξ, qui se trouvent coïncider avec les initiales des trois hiéromnémons mentionnés aux lignes 1 à 3, Zoïlos, Nikias et Xénon. De la même manière, dans la tablette n° 26, la ligature est précédée, sur la même ligne, par les trois lettres Σ, Σ et Μ, qui coïncident avec les initiales des trois hiéromnémons mentionnés aux lignes 1 à 3, Sosistratos, Sosippos et Ménipos. De plus, à la ligne précédente figure, isolée, la lettre Φ, qui coïncide avec l’initiale du magistrat éponyme Phaikionos, mentionné à la ligne 1. Il paraît naturel de voir dans cette double collection d’initiales le « visa » des magistrats responsables des opérations comptables retranscrites dans les tablettes n° 15 et 26. Il ne serait donc pas étonnant que ces initiales soient accompagnées d’une espèce de «-sigle-» rappelant leur titre de hiéromnémons. Si notre interprétation est exacte, les tablettes n° 15 et 26 attesteraient une forme de « code switching » dans la mesure où les ligatures étudiées refléteraient le nom dialectal des hiéromnémons (Ͱιαρομνάμονες) alors que, dans le texte même de ces deux tablettes figure une forme mixte Ͱιερομναμόνων influencée par la koinè (n° 15, ligne 1-; n° 26, ligne1). Références bibliographiques Blomqvist, J. (1975) «-The Dialect of Epizephyrion Locri-», Opuscula Atheninensia XI (1975) 2 : 17-35. Blomqvist, J. (1978) «-Additional Remarks on the Locrina Bronze Tablets-», Opuscula Atheniensia XII (1978), 7 : 17-132. Costabile, F. (1992) (a cura di) Polis ed Olympieion a Locri Epizefiri, Costituzione Economia e Finanze di una Città della Magna Grecia, Editio altera e Traduzione delle Tabelle Locrese, Catanzaro. De Franciscis, A. (1972) Stato e Società in Locri Epizefiri (l’archivio dell’Olympieion locrese), Naples. 94 Emmanuel Weiss <?page no="95"?> Ghinatti, F. (1998) « Cronologia e rotazione dei fatarchi nelle tabelle di Locri Epizefiri », Minima Epigraphica et Papyrologica I (1998)-: 55-77. Del Monaco, L. (1997) «-Le Tavole di Locri sono 37. Un nuovo attacco tra le tabb. 35, 36, 37-», Rivista di Filologia e di istruzione classica 125-: 129-149. La signature des hiéromnémons 95 <?page no="97"?> 1 The bibliography on the topic is rich and beyond the scope of this study. The following works are mentioned here as a basic reference framework and within the purview of the study. For an early detailed account of the relations of Greece and South Italy and Sicily with respect to the Greek colonies, see Dunbabin (1948), and a series of conference The Magna Graecia Tablets in the Dodona Corpus Georgios K. Giannakis (Aristotle University of Thessaloniki & Center for the Greek Language) Abstract: The publication in 2013 of the complete corpus of the lead tablets discovered at the sanctuary of Dodona stimulated an intensive and systematic effort to study the documents, adding an important corpus to the documentary wealth of the epigraphic body of Greek. The present study aims to investigate the relations of the Dodona oracle with the Greek world of the West, mainly the Greek colonies of South Italy and Sicily (Magna Graecia). To this end, the sub-corpus of tablets that document these relations of the oracle has been systematically studied and commented upon from the point of view of language, writing, historical references and other relevant cultural information. The conclusion reached is that, despite the relative isolation of the area at the fringes of the Greek world, the oracle played an important role as a sacred center and affected the lives of the inquirers (individuals or organized groups and cities) by means of its consultatory and advisory function to the pilgrims who sought its advice on a variety of matters. Keywords: Dodona, Dodona lead tablets, Epirus, Greek dialects, Greek epigraphy, Greek language, Magna Graecia, oracle of Dodona. 1 Introduction The relations between ancient Greece and Italy go as far back as the oldest written attestation and even further back if we consider archaeological and mythological evidence. The Ionian Sea, or better the Adriatic strait, was and still is the main route of communication between the two areas. 1 This strait can function as <?page no="98"?> proceedings on Magna Graecia held annually since 1961; for the linguistic situation in Sicily, see Willi (2008) and Tribulato (2012); for the linguistic situation in ancient Mediterranean, see Clackson (2015), Clackson et al. (2020), Molinelli (2017). Of course, the best reference works for our study are those with the primary sources, the corpora that contain the written documents unearthed at the Dodona sanctuary, namely Lhôte (2006) (conventionally referred to as LOD) with 167 tablets discussed in a detailed commentary, but mainly Dakaris / Vokotopoulou / Christidis (2013) with the complete corpus available thus far of 4216 tablets edited and annotated in detail (for short referred to as DVC). For a recent critical evaluation of the issues related to the nature, character and other problems of the tablets, see the essays in Soueref (2017). 2 This fact may also explain (in addition to its relative isolation at the fringes of ancient Greece) why the Dodona oracle remained at the backdrop of the Greek world (cf. also Dillon 1997: 97). both a barrier of separation but also and more frequently as a bridge facilitating contacts and relations (cf. Lombardo 2017: 113). In our case, as shown by the close relations between the two worlds since antiquity, it rather served more the second function, forming a common ground for collaboration and contact. Admittedly, a turning point in the investigation of the relation between the Dodona oracle and the territories of central and western Mediterranean, especially the Greek cities of Magna Graecia was the study by Ioulia Vokotopoulou in the thirty-first convention on Magna Graecia that took place in Taranto in October 1991 (see Vokotopoulou 1992). It was the first study in which, based on epigraphical evidence drawn from the Dodona lead tablet corpus known at the time (in addition to other evidence), a clear and documented connection was made between the activities of the oracle and the Greek cities of South Italy. Although Vokotopoulou did not discuss the epigraphical evidence in detail, it had become clear to the participants of the meeting that one can speak of these connections of the oracle on real arguments deriving from written sources, some of which may be quite old, with most of them dating to the 5 th and the 4 th centuries BC. 2 As is already known, after the 2013 publication of DVC, our corpus of the lead tablets coming from Dodona consists of 4216 lamellae of various length and/ or completeness. The overwhelming majority of these tablets deal with different issues that relate to different areas of mainland Greece, with scant evidence coming from the islands and from Macedonia or Thrace. However, in our corpus there are also several tablets that directly or indirectly speak of the involvement of the Dodona oracle in maritime activity of the inquirers who consult it, but perhaps only about 25 or so tablets are directly associated with Magna Graecia, mainly Sicily and the lower part of the boot of the Italian peninsula. These tablets testify to the close relations between the sanctuary of Dodona and the South Italian Greek colonies, almost all of which are of Doric descent, and in this sense of the same phylum as the people inhabiting the wider Dodona and Epirus territory. 98 Georgios K. Giannakis <?page no="99"?> 3 Some parts of this study are drawn from other studies of the author, such as Giannakis (forthcoming), and Crespo / Giannakis (2019). 2 The Dodona site Dodona is located in Epirus, in the northwest of Greece, on the plain near the present-day city of Ioannina, and is surrounded by a number of mountains of the Pindus range to the east, the Ionian Sea to the west, the Illyrians to the north, and the Gulf of Ambracia to the south, and with the ethnic group known as Chaones inhabiting the northern part, the Molossoi the central and eastern zone, and the Thesprotoi the southern area (see Map 1). 3 The sanctuary itself is believed to be (one of) the oldest of the oracles of ancient Greece, dedicated to Zeus (Zeus Naios) and secondarily to Dione (in a sense Zeus’ consort). References to it are made by ancient authors, such as Homer Il. 16.233-235, Od. 14.327-330, Herodotus 2.52.2, Strabo 7, fr. 3, and others, see below: Map 1. Ancient Epirus and environs. (Source: Creative Commons Attribution-Share Alike 4.0 International license). The Magna Graecia Tablets in the Dodona Corpus 99 <?page no="100"?> Il. 16.233-235 Ζεῦ ἄνα Δωδωναῖε Πελασγικὲ τηλόϑι ναίων Δωδώνης μεδέων δυσχειμέρου, ἀμφὶ δὲ Σελλοὶ σοὶ ναίουσ’ ὑποφῆται ἀνιπτόποδες χαμαιεῦναι ‘High Zeus, lord of Dodona, Pelasgian, living afar off, brooding over wintry Dodona, your prophets about you living, the Selloi who sleep on the ground with feet unwashed’. (tr. Lattimore) Od. 14.327-329 (cf. 19.296-298) τὸνδ’ ἐς Δωδώνην φάτο βήμεναι, ὄφρα ϑεοῖο ἐκ δρυὸς ὑψικόμοιο Διὸς βουλὴν ἐπακούσαι, ὅππως νοστήσει’ Ἰϑάκης ἐς πίονα δῆμον ‘But he said Odysseus had gone to Dodona, to listen to the will of Zeus, out of the holy deep-leaved oak tree, for how he could come back to the rich countryside of Ithaka’. (tr. Lattimore) Hdt. 2.52.2 καὶ μετὰ χρόνον ἐχρηστηριάζοντο περὶ τῶν οὐνομάτων ἐν Δωδώνῃ· τὸ γὰρ δὴ μαντήιον τοῦτον ενόμισται ἀρχαιότατον τῶν ἐν Ἕλλησι χρηστηρίων εἶναι, καὶ ἦν τὸν χρόνον τοῦτον μοῦνον. ‘And after some time, they asked the oracle at Dodona about the names; for this place of divination was believed to be the most ancient oracle in Hellas, and at that time it was the only one’. Strab. 7, fr. 3 ὅτι ἡ παροιμία ‘τὸ ἐν Δωδώνῃ χαλκίον’ ἐντεῦϑεν ὠνομάσϑη· χαλκίον ἦν ἐν τῷ ἱερῷ ἔχον ὑπερκείμενον ἀνδριάντα κρατοῦντα μάστιγα χαλκῆν, ἀνάϑημα Κορκυραίων· ἡ δὲ μάστιξ ἦν τριπλῆ ἁλυσιδωτὴ ἀπηρτημένους ἔχουσα ἐξ αὑτῆς ἀστραγάλους, οἳ πλήττοντες τὸ χαλκίον συνεχῶς, ὁπότε αἰωροῖντο ὑπὸ τῶν ἀνέμων, μακροὺς ἤχους ἀπειργάζοντο, ἕως ὁ μετρῶν τὸν χρόνον ἀπὸ τῆς ἀρχῆς τοῦ ἤχου μέχρι τέλους καὶ ἐπὶ τετρακόσια προέλϑοι· ὅϑεν καὶ ἡ παροιμία ἐλέχϑη ‘ἡ Κερκυραίων μάστιξ.’ ‘The proverbial phrase, ‘the copper vessel in Dodona’, originated thus: In the temple was a copper vessel with a statue of a man situated above it and holding a copper scourge, dedicated by the Corcyraeans; the scourge was three-fold and wrought in chain fashion, with bones strung from it; and these bones, striking the copper vessel continuously when they were swung by the winds, would produce tones so long that anyone who measured the time from the beginning of the tone to the end could count to four hundred. Whence, also, the origin of the proverbial term, ‘the scourge of the Corcyraeans’. 100 Georgios K. Giannakis <?page no="101"?> 4 Cf. Hdt. 2.52-57, and for the three priestesses (Promeneia, Timarete and Nikandre) 2.55.3; also, Strabo 7.7.12. For the general historical context and other aspects of the Dodona oracle, see the recent work by Piccinini (2017). Like so many other sacred places in antiquity and throughout history, Dodona too has its etiological foundation myth that goes back to the depths of prehistory. One such myth tells how the idea of founding an oracle at Dodona started. Two black doves (δύο πελειάδας μελαίνας) were sent off from Thebes in Egypt flying one to the site of Siwah and the other one to Dodona. Upon arriving to Dodona, the dove landed in an oak whereupon in human voice instructed the people for the founding of an oracle of Zeus (Hdt. 2.55.2: ἱζομένην δέ νιν ἐπὶ φηγόν αὐδάξασθαι φωνῇ ἀνθρωπηίῃ ὡς χρεὸν εἴη μαντήιον αὐτόθι Διὸς γενέσθαι). Dillon (1997: 96) says that “this is probably an aetiological myth dating to the time of change from the Selloi to the three priestesses”, who had replaced the interpreters in the oracle and are mentioned by Herodotus by order of age, “perhaps indicating a sequence of succession and apprenticeship” (loc. cit.). 4 3 The queries The general linguistic makeup of the area is Northwest Greek, a sub-branch of Doric, with some influence from other dialects or languages of the area, such as Illyrian. However, the cultural influence of the oracle was panhellenic, drawing visitors from most of mainland Greece and to some extent from Magna Graecia. This fact anticipates a ‘colorful’ and mixed dialectal character of the oracular texts, a fact that is documented by the textual evidence of the corpus. The tablets contain the inquiries (and some answers to them by the oracle officials) by visitors or by communities from all over the ancient world on a variety of everyday matters of individuals or communities. Although many inquiries were written by the consultants themselves on narrow lead plates, many others were dictated to professional scribes employed by the oracle and so written on the spot, a fact that may have its significance for the linguistic character of many of these inquiries. Also, in some tablets one side seems to contain the question and the other side the oracle’s response. Classical authors offer some information The Magna Graecia Tablets in the Dodona Corpus 101 <?page no="102"?> 5 Different means and mechanisms were used for issuing the oracle’s replies to the inquirers. According to the passage from the Odyssey above, one way was interpreting by the priests of the noise made by the leaves of the sacred oak of Zeus; other sources speak of other artificial procedures that produced sounds, such as the copper vessel mentioned by Strabo in the passage above (an offering to the oracle by the inhabitants of the island of Corcyra), the cooing of the wood pigeons (πελειαί), the noise of running water from a nearby natural spring, and other devices (see Crespo / Giannakis 2019: 47). about the procedure followed in the query of the oracle and the way in which the oracle’s replies were issued. 5 Most of the queries are brief statements, indicating the specific inquiry that is addressed to Zeus Naios and Dione, often mentioning the name of the inquirer or inquirers. Most often the inquiries are expressed through a direct or indirect question or through an independent sentence, whereas a small number of them are in the form of an appeal. In a small number of tablets, we seem to have the oracular responses, which take the form of simple or conditioned commands, prohibitions or warnings, and statements (see also Eidinov 2007, Piccinini 2013). In general, the content is plain and unambiguous, using a stereotypical opening formula like θεὸς τύχα ‘God, good luck! ’, followed by the receiving god (normally Zeus and his consort Dione), and then the statement of the inquiry itself in straightforward language and simple style, often mixing dialectal features and official and more popular styles. In this sense, the texts of the Dodona corpus are close in style and structure to similar texts from other ancient sanctuaries, e.g. Delphi or Didyma, although there are also significant differences between them (see Fontenrose 1978: 11-57 and 417-429). There are two main categories of inquiries, private and public. As expected, the majority of them belong to the first category, i.e. they are private inquiries by individuals who seek an answer by the oracle on a variety of issues, such as family affairs, property, financial matters, marriage, childbirth, work, travel, migration, health issues etc. The public inquiries (a total of 32, see Parker 2016: 75; according to Bonnechere 2017: 75, the percentage of the public tablets may be ca 2 % of the total corpus) concern questions by groups of people, such as communities, cities and towns, or ethnic groups, and deal with matters of action to be taken with respect to relations to other groups and cities or concern colonization and other such issues. In public inquiries the name of the group is explicitly mentioned in the tablets, with most prominent (and expectedly so) being that of the people of Dodona (Δωδωναῖοι), and secondarily by other groups, e.g. the nearby Molossians and the Chaones, but also others. 102 Georgios K. Giannakis <?page no="103"?> 6 This map should be read in conjunction with Map 1 which shows the circum-Adriatic regions where different Greek cities had established colonies and had a special significance to the activities of the Dodona oracle, as documented among others also in the lead tablets that interest us in this study. 4 Ancient Greek dialects Since the beginning of its literate history with the oldest inscriptions of 8 th c. BC and thereafter in the literary production, Ancient Greek is characterized by its many different local varieties with each having distinct and distinctive linguistic features that produce a colorful picture of a dialectal map of the language. With the repeated waves of colonization by different Greek cities this linguistic situation is transmitted to the new establishments both along the coast of Asia Minor and the circum-Pontic regions, as well as to North Africa and mainly to Central and Western Mediterranean territories, principally in Southern Italy and Sicily, and across the Adriatic coastal area. Reasonably the dialects in the new territories reflect the dialects of the founding mother cities. The following Map 2 shows the rough areal distribution of the Greek dialects in the Classical Period, both in the mainland and in its colonies. In the area that is of immediate interest for our study, the two main dialectal branches are Doric in a general sense (i.e. different subvarieties of Doric, e.g. Doric, Northwest Greek and Achaean) and secondarily Ionic. 6 It is also true that often some features either of the writing conventions or of the dialects themselves developed in the colonies of the West certain special features that distinguish these varieties from the corresponding varieties of the mother cities of mainland Greece, forming thus a sociolinguistic area of special interest for study. If one adds to all this the possible influence from neighboring Greek dialects or from other local languages, the situation gains still further interest. Although the Dodona material consists of tablets of diverse and/ or mixed dialectal character, the main body of evidence is the local Doric idiom of dialect which has some special features of its own that are consonant with the dialects of the Northwestern branch of Greek. In the Dodona lead tablet corpus, there are inquiries that directly or indirectly relate to Magna Graecia, and this is an important source of information for the study not only of the language history but also of historical and cultural developments in the ancient world. The Magna Graecia Tablets in the Dodona Corpus 103 <?page no="104"?> Map 2. The Distribution of Ancient Greek Dialects (Source combined maps based on: https: / / en.wikipedia.org/ wiki/ Ancient_Greek_dialect s#/ media/ File: Magna_Graecia_ancient_colonies_and_dialects-en.svg and https: / / en.wikipedia.org/ wiki/ Ancient_Greek_dialects#/ media/ File: AncientGreekDialec ts_(Woodard)_en.svg). 5 Epirus, Dodona and the Greek world For the period before the 6 th c. BC the literary evidence for Dodona is very scant and consists of references by ancient authors, like Homer. However, archaeological and mythological information is relatively better, and this offers a ground for building a picture of the relations of Dodona and Epirus with the rest of the Greek world and of the surrounding areas in and outside mainland Greece (see also Domínguez 2018). Chapinal-Heras discusses the early relations and connections of Dodona with the rest of the Greek world saying that, “While colonization must have motivated many of these connections with lands abroad, another element, the sanctuary of Dodona, surely played an active role in the process” (2021: 24). In the spirit of Castiglioni and other researchers, he further says that, based on this evidence, we may have to reevaluate our ‘isolation theory’ of Epirus and see it under the light of new information that derives from recent excavations in the area and other relevant studies. As further pointed out by Filos (2018a: 241-242), Epirus has been in contact with neighboring peoples, tribes and languages, sometimes even within a bilingual environment, and it may thus be a good case of a language contact area. This scholarly activity could also yield interesting results and useful information on other aspects of the history and cultural life of Epirus. Despite its marginal status in terms of 104 Georgios K. Giannakis <?page no="105"?> 7 DVC 228B, 283B, 366A, 1244, 1453B, 1483A, 1772A, 1809, 2054B, 2261A, 2363B, 2546A, 2610A, 3030A, 3656B, 3742B, 3946B, 4154B; 1711A, 3685B; 957B. 8 For ναῦς (acc. νᾶϝα) and the rest of the terms of this group: DVC 167A, 272A, 298A, 302B, 323A, 371A, 843B, 1005A, 1182A, 1210B, 1363A, 1521B, 1554A, 1687A, 2046, 2365A, 2626B, 2629B, 2641A, 2762A, 2810A, 3024A, 3214A, 3389, 4097A. 9 See Casewitz (1985) for the ‘language’ of colonization; cf. further n. 19. the degree of integration with the rest of the Greek world, Epirus seems to have been in contacts, directly or indirectly, with the outside world, first and foremost with the Greek poleis of the Northwest, i.e. from part of Acarnania and the nearby coastal towns (Ambracia, Corcyra and others), as well as to the colonies to the North (Apollonia, Epidamnos, etc.). From the 6 th c. onwards and mainly during the 5 th -4 th c. BC, these connections are documented by literary evidence deriving from the Dodona lead tablets. In this direction, Euboea’s role may have been especially significant, since as argued by Piccinini (2017), in their journeys to Central and Western Mediterranean the Euboeans most likely used the ports in Epirus as intermediate stations, and thus also offering the opportunity to the Dodona oracle to play some role in all this, also seen in some tablets that present interesting evidence for these connections, e.g. DVC 35A (see discussion in 6.2.2 below), although it is not entirely clear which Cumae is meant, the town in Euboea, its Magna Graecia colony or the town in Asia Minor. 5.1 The ‘language’ of maritime and trade activity In the DVC corpus this maritime activity is documented in a good number of tablets with relevant terminology, such as the frequently used verb πλέω or the nouns πλοῖον or πλοῦς. 7 In addition, other relevant vocabulary, such as ναυκληρέω or more commonly its corresponding Doric form ναυκλᾱρέω, ναύκληρος, ναῦς, ναυτιλλέω, ναύσταθμος, ναυπαγέω, 8 indicate the relation of the oracle with the world of the sea, ship-building and ship-using, not necessarily its direct implication with such activities but its involvement in one way or another with people who busy themselves with seafaring or enquire the oracle about such a possibility (see the list of tablets using this terminology in DVC II: 559). Furthermore, other such terms include οἰκέω, ϝοικίζω, οἴκησις and ἐποικία that refer to the move to a new place or to the establishment of a new colony, although not necessarily or always involving seafaring (it can be a place or town reachable by land), and not always to Magna Graecia (it can be some other place, coastal or inland), but the oracle has a role to play in this area (for the loci of these terms, see DVC II: 525). 9 The following two tablets are characteristic examples of the connection of the oracle to people whose concern is some kind of maritime activities. The first The Magna Graecia Tablets in the Dodona Corpus 105 <?page no="106"?> 10 Pharos (or in other instances also Paros, cf. Strab. 7.5.5 Φάρος, Πάρος λεγομένη πρότερον ‘Pharos, in the past also called Paros’; see also Hammond 1967: 605) was a colony of Paros from the early fourth century BC. Other references to Pharos in the DVC corpus are: 463A, 1137B, 2762A, 3030A, 3146A, 3517A, with relevant bibliography in the DVC commentary on 228B (2013 I: 85). is an interesting tablet (of the 4 th c. BC) written in koine (according to the DVC editors), although there are mostly clear Doric features (see below) of the inquiry of someone named Aristodamos who consults the oracle if it would be good and proper to sail to Pharos 10 during the summer of this year, most likely for trade reasons. The text is given here as interpreted by the editors (DVC 228B): (NB: The texts are reproduced from DVC and/ or LOD as indicated each time) θεὸς τύχα ἀγαθά· Ἀριστόδαμος επ̣[ικ]οινῆται Διὶ Ναΐωι καὶ Διώναι καὶ θεῶν <ἦ> ἦ λώϊον καὶ ἄμεινον ΣΑΤΕΙ πλέοντι [κατὰ θάλασσ]αν καὶ ἐς Φάρον τοῦ θέρεος ‘God, Good Luck! Aristodamos enquires Zeus Naios and Dione and the other gods if it would be preferable and better for me to sail to Pharos this year during the summer’. The name of the inquirer must be Doric seen in the -ᾱof the -δᾱμος part; of extreme interest is the form ΣΑΤΕΙ which raised a lot of discussion. It seems that Méndez Dosuna’s suggestion to see in the word an otherwise unattested adverb ΣΑΤΕΙ in the meaning ‘this year’ is to my mind a correct suggestion, cf. the Doric form σᾶτες from Sicily, also Dor. τᾶτες, Att. τῆτες, and the gloss of Hesychius σᾶτε[ι]ς· τὸ ἐπ’ ἔτος. Δωριεῖς. *τῆτες· ἐν τῷδε τῷ ἔτει. οἱ δὲ Δωριεῖς σᾶτές φασιν. If this is the case, the form of the word agrees with the Doric origin of the inquirer’s name. Unlike the statement in DVC that the dialect of the tablet is koine, perhaps we have a case of a Doric tablet written in mixed 106 Georgios K. Giannakis <?page no="107"?> dialect, and this may be part of the effort by the oracle officials to accommodate the language to that of the inquirer (on which see Crespo / Giannakis 2019, Giannakis forthcoming). The second tablet to be mentioned here is DVC 366A, a mid-4 th c. tablet, with a private inquiry referring to the Greek colony of Apollonia on the Adriatic coast: [εἰς] Ἀπολλωνίαν πλεύσας ἦι ἀλάστων τῆ- [ιδ]ε ὄντων πυνθάνοιτο; John Chadwick’s suggestion to translate “on sailing to Apollonia, whether he will get news of the miserable wretches there” makes good sense in that the inquiry perhaps deals with a common situation in ancient times concerning people being trapped in a place and not being able to return home, or even it refers to fugitives (not an uncommon theme in our corpus) whose life turned out to be worse in the new place than before, or something of this sort, and the inquiry relates to the possibility of recovering these people safely from a hostile situation. The alternative suggestion by Cassio (2007), repeated by Tselikas in the DVC commentary on the tablet (but without mentioning Cassio’s analysis), to read [εἰς] Ἀπολλωνίαν πλεύσας ἢ ἰάλας τῶν τῆ[δ’] ἐόντων πυνθάνοιτο is also a reasonable reading of the text, especially in view of the reading ἰάλας which makes it a participle coordinated with πλεύσας, and thus referring to a trip to Apollonia either by sea (πλεύσας) or by land (ἰάλας). The mission is ἰάλας τῶν τῆ[δ’] ἐόντων πυνθάνοιτο, i.e. to seek information from those who are in Apollonia either sailing there himself or sending someone for this purpose (cf. also commentary in DVC). In DOL the reading is ἢ ἱάλας τῶν … with aspirated initial ἱ-, also referring to CIOD 159A, and to the verb ἐφ-ιάλλω which is well attested in Theocritus and elsewhere. Alternatively, if we follow Lhôte (2006: 208-211; see LOD 97), the word ἀλάστων could be the Doric form of ἀλήτης ‘the wandering one’, i.e. referring to people who are lost or trapped there for some reason and have no way out from their wanderings (although the -σis a problem with such a reading). Curiously enough, Lhôte’s discussion of this tablet (LOD 97) is ignored by the DVC editors, so much so since Lhôte makes a rather interesting suggestion regarding the interpretation of the difficult word ἀλάστων: his proposal is to read ἀλα᷄ιστων, i.e. a Doric form of ληιστός with the privative prefix ἀ-, thus ἄληιστος ‘not (to The Magna Graecia Tablets in the Dodona Corpus 107 <?page no="108"?> 11 In fact, summing up his discussion of this problem, Lhôte says that in the corpus of the Dodona tablets that he studied “this reduction (viz. omission of the i-subscript) is a characteristic of the popular Greek language much earlier than what is usually supposed, and can be verified in both Attic and Doric” (2006: 390). Cf. further Schwyzer (1953: 349), Buck (1955: 34-36), Threatte (1980: 358 ff., who characterizes such omissions “careless”), Bartoněk (1966: 44), Méndez Dosuna (1985: 456), Lejeune (1987: 226-227), Sihler (1995: 58). 12 However, DVC 228B could be another tablet with reference to this town if the lacuna in l. 5 is restored as [ἐς Ἀπολλωνί]αν as suggested in the DVC commentary. 13 86A, 228B, 279A, 430, 580A, 1005A, 1782B, 1809, 2564B, 2650B, 2859B, 2945B, 3001A, 3147B, 3195B, 3364A. 14 1156B, 2361B, 2784A, 3109A, 3221B, 3239A, 4072A. 15 1537A, 2782A, 3220A, 3695B, 4088B, 4162B. be) plundered’ (cf. verb ληίζω), with the omission of the subscript ι which is common in inscriptions (on which see Lhôte 2006: 388-390). 11 Strangely enough, the town of Apollonia is mentioned only in this tablet, although geographically this is one of the closest to Dodona coastal towns of Chaonia in northern Epirus. 12 In several tablets there is a question to the oracle if it would be good for the inquirer to take a sea route for a journey, also leaving room to be understood that for some people the sea is not a familiar way of traveling, and thus some anxiety is possible to be detected in the linguistic features of the inquiry. The next example (DVC 1809) is such a case. Despite its extremely fragmentary state, we may have one such example in this tablet with the inquiry by a group of people (adduced from the plural of the participle [πλέ]οντες and the verb [πράσσοιμες] as reconstructed by the editors) and concerns an excursion by sea (θ[άλασσαν] [πλέ]οντες): there are such dilemmas as to whether one should take a sea route or some other way, apparently by land, to reach the destination: DVC 1809 (second half of the 5 th c. BC): [ἐ̄] κ̣ατὰ θ̣[άλασσαν] [πλέ]οντες λṓϊ[στά κα καὶ] 3 [βέ]λτισ̣τα [πράσσοιμες]; ‘If sailing by sea we would do the most preferable and the best thing’. Furthermore, traveling and trade may also be assumed for tablets in which the language points to such activity, e.g. the word πόντος in DVC 654A, the word θάλασσα in several tablets, 13 or the verbs πορεύομαι 14 and πωλέω 15 (only 3220A is relevant here, the rest refer to some selling and involve no journey) 108 Georgios K. Giannakis <?page no="109"?> 16 In DOL we read the comment: “…Persée, roi de Macédoine, qui, compte tenu du dialecte, a envoyé un émissaire épirote consulter à Dodone au moment de la troisième guerre de Macédoine (171-167 av.)”. 17 In this sense, the word πόρος is matched in other cases by πόντος which functions in similar ways, i.e. as a passage to the other side, its etymological meaning, cf. Lat. pons (pont-) ‘bridge’, Skt. pathás ‘road’, OCS pǫtŭ ‘road’, etc. which indicate movement and/ or commercial activity for which the inquirers thought that it would be good to first ask the oracle before undertaking such an enterprise. Such is DVC 3109A dating to the mid-fourth c. BC (but according to DOL it is much later dating to 170-169 BC, cf. also n. 16, and the reference is to the town of Pharos, a tablet written in the local Epirote dialect). In this tablet the terms ἐποικ[ία]ν πορευομέν[ωι] καὶ <α̣> ἀμβαίνοντ(ι) signal an expedition to move to a new establishment, a decision for which the consultation of the oracle is being sought ([ἐπ]ερωτῆι … [ἦ]) by the inquirer (Πεσ̣εύας or perhaps Περσεύας 16 ), a common practice in the ancient world. However, this inquiry gives no indication as to whether the trip is by sea or by land (most likely the former, especially if it is about Pharos), the participle ἀμβαίνοντ(ι) could refer to both, take an uphill inland course or board on a ship for a journey. [ἐπ]ερωτῆι Πεσ̣εύας Δία Νᾶον καὶ Διώ̣[ναν] [ἦ] λώϊον καὶ ἄμεινον ἐσσεῖται αὐτῶι [καὶ] 3 ὀ̣χυρὸν εἰς τὰν ἐποικ[ία]ν πορευομέν[ωι] καὶ <α̣> ἀμβαίνοντ(ι) τά(χ)ιστα Similarly, the participle πορευομέν[ωι] can refer to a course either by land or by sea. After all, it is a derivative of πόρος which is the passageway to the other side by crossing a natural obstacle, mainly sea, river or like. 17 Interestingly, the narrowest part of the Ionian Sea, i.e. the strait of Otranto, is often referred to as Ἰόνιος πόρος (e.g. Diod. Sic. 15.13 and 16.5; Polyb. Hist. 2.14.4, etc.). If we take the linguistic evidence seriously, then most likely this tablet belongs in the same group as others that deal with journeys by sea. In several other tablets there are also direct or indirect references to travel for reasons of commerce, trade and the like with the use of relevant terms such as the verb ἐμπορεύομαι and its cognates ἐμπόριον and ἐμπορία (also with initial The Magna Graecia Tablets in the Dodona Corpus 109 <?page no="110"?> 18 See DVC 105B, 279A, 430, 466A, 516B, 580A, 597A, 703B, 723A, 1194A, 1259B, 1363A, 1695A, 1715B, 1854A, 2048B, 2053A, 2054B, 2082B, 2108B, 2154B, 2366B, 2624A, 2634A, 2652A, 2715B, 2802B, 2810A, 2929B, 2945B, 3001A, 3364A, 3442A, 3497A, 3590A, 3651A, 3653B, 3831A, 4065A, 4193A, 4199B; 3, 418A(? ), 510A,1095B, 1628A, 1673B, 1685, 1765B, 1770B, 2184, 2235B, 2464B, 2483, 2491B, 2776B, 2993B, 2999B, 3185, 3247B, 3545B. 19 This category of tablets is signaled by terms like πεδαϝοικέω and μετοικέω ‘change home’, μένω and καταμένω ‘stay’ (i.e. whether one should stay or leave home), κατὰ χώραν ‘(leave) to a place’, ἄπειμι ‘go away’, ἀποίχομαι ‘depart’, ἐξοικέω ‘depart from one’s home’, ἐκδημέω ‘leave the city’, ἀποδημέω ‘change city’, πλάνα ‘wandering’ (i.e. whether one should take the risk of making the journey); see Chaniotis’ detailed list of tablets on p.-64, n. 54. ἐνfor ἐμ-), a fact that shows that the oracle played a significant role in advising people on this activity (see index in DVC II: 515-516). Interestingly, these terms are etymologically related to πόρος and πορεύομαι discussed above, referring to trade of goods in sea journeys, therefore the previous comments apply here as well. It should also be noted that not all these tablets speak of a journey to the West and to the cities of South Italy and Sicily. 18 6 The ‘Magna Graecia’ Dodona tablets As already said, the number of tablets of clear South Italian provenance is rather small relative to the large volume of the corpus; there are only 25 or 26 such tablets amounting to a slim 0.5% of the total corpus. The numbers may not be impressive, but sometimes what they tell us is of special significance as historical evidence, and this seems to be the case here. These tablets are the following (adding to Lombardo’s 2017 list which contains 16 tablets): (a) public: DVC 3111, 3079A, 3134A; LOD 5, 17, 132, 154, 155; (b) private: DVC 24-26, 959A, 1432, 3111, 2333B / 35A-37B, 280A, 2367, 4154B; LOD 146, 75, 132, 114, 111, 103, 133, 102, 106. There may be a few more tablets that could have a relevance to our topic, but they are fragmentary, e.g. DVC 562A in which we have only the part ἐν Βαρί[ωι---- -], a reference to Βάριον (Bari) or some other name? In the tablets we have strong evidence for movements of people or more commonly of individuals in search of a new residence or better future and employment, as we can see in a good number of tablets that deal with travel and change of residence or emigration to various places. According to Bonnechere (2017: 74), these tablets amount to 155 (= 12 % of the total of the DVC corpus), not a small sub-corpus. Also, Chaniotis (2017: 64) refers to a total of 63 trade tablets (see DVC I, no. 3 for details), and another total of 87 migration tablets. 19 All this activity of groups of people or individuals moving to other places for trade or permanent residence is one issue that has special significance for establishing 110 Georgios K. Giannakis <?page no="111"?> 20 The significance of the oracles in general and particularly of that of Dodona is stressed with a rhetorical question by Cicero, who in his De Divinatione I.3 states rather emphatically that no quest for establishing a new colony was ever possible without first consulting one of the ancient oracles. He says: Quam vero Graecia coloniam misit in Aeoliam, Ioniam, Asiam, Siciliam, Italiam sine Pythio aut Dodonaeo aut Hammonis oraculo? ‘But really, what colony did Greece undertake to Aeolia, Ionia, Asia, Sicily and Italy without a prophecy from either Pythia or Dodona or Ammon? ’. The three most prominent oracles of antiquity are quoted together as a triad, following an old tradition of using in such references triads of sanctuaries or divinities (on which see Pease 1920: 490, Parke 1967: 211, and others). the possible connections and/ or associations of the Dodona oracle with the people who live abroad (presumably in Greek colonies of Southern Italy and Sicily) or others who were thinking of traveling or emigrating: they all ask the oracle for advice as to what to do. 20 This means that the oracle’s purview of influence went far beyond the narrow confines of Epirus and its immediate surroundings, extending from (much of) mainland Greece to the Greek colonies of the circum-Adriatic area and of Magna Graecia, perhaps even to some foreign peoples of the area. However, the high number of tablets on the general category of travel should be no surprise as this is part of the traditional activity among the peoples of the Mediterranean who travel abroad κατ’ ἐμπορίαν καὶ κατὰ θεωρίαν (Isoc.17.4), i.e. for business and for enjoying the visit to other places. We should add of course also ἐποίκησιν, colonization, and this seems to be an old practice of all peoples of the area, not only of the Greeks. The tablets indicate that the inquirers that concern us in this study are of three types, namely (i) people who ask the oracle’s advice about leaving their hometown and migrating to some other place, (ii) people who inquire on whether they should undertake a business trip, or (iii) people of colonies of the West who consult the oracle on various matters that concern them. As in the rest of the corpus, these tablets are either public (made by poleis and other public establishments) or private (i.e. made by single individuals, families or other small groups of people). This common activity creates favorable conditions also for the activities of the Dodona oracle, and this is what we will concentrate on in the following. Ancient sources speak of the Epirotes as having little or no acquaintance with the sea and with maritime life and activities of this sort, although this could be one side of the story but not the whole story. Nevertheless, the idea of the oracle’s marginality and isolation was based already in the ancient world on Homer’s Odyssey with reference to Odysseus’ journey among the Epirotes who are not acquainted with the sea and are ignorant of the use of salt, as prophesied The Magna Graecia Tablets in the Dodona Corpus 111 <?page no="112"?> 21 See further Pausanias 1.12.5 who repeats the Homeric passage in his explanation of the naval defeat of the Epirotes under Pyrrhus in Sicily in 278 BC: Πύρρον δὲ ἐς Σικελίαν ἀπήγαγε πρεσβεία Συρακουσίων· Καρχηδόνιοι γὰρ διαβά‐ ντες-τὰς Ἑλληνίδας τῶν πόλεων ἐποίουν ἀναστάτους, ἣ δὲ ἦν λοιπή, Συρακούσαις πολιορκοῦ‐ ντες προσεκάθηντο. ἃ τῶν πρέσβεων Πύρρος ἀκούων Τάραντα μὲν εἴα καὶ τοὺς τὴν ἀκτὴν ἔχο‐ ντας Ἰταλιωτῶν, ἐς δὲ τὴν Σικελίαν διαβὰς Καρχηδονίους ἠνάγκασεν ἀπαναστῆναι Συρα‐ κουσῶν. φρονήσας δὲ ἐφʼ αὑτῷ Καρχηδονίων, οἳ θαλάσσης τῶν τότε βαρβάρων μά‐ λιστα-εἶχον-ἐμπείρως- Τύριοι Φοίνικες τὸ ἀρχαῖον ὄντες, τούτων ἐναντία ἐπήρθη ναυμαχῆσαι τοῖς Ἠπειρώ‐ ταις-χρώμενο, οἳ μηδὲ ἁλούσης Ἰλίου θάλασσαν οἱ πολλοὶ μηδὲἁλσὶν ἠπίσταντό πω χρῆσθαι. μαρ‐ τυρεῖ-δέ-μοι καὶ Ὁμήρου ἔπος ἐν Ὀδυσσείᾳ·οἳ οὐκ ἴσασιθάλασσανἀνέρες, οὐδέ θʼ ἅλεσσι μεμιγ‐ μένον-εἶδαρ- ἔδουσιν. Hom. Od. 11.122 ‘Pyrrhus was brought over to Sicily by an embassy of the Syracusans. The Carthaginians had crossed over and were destroying the Greek cities, and had sat down to invest Syracuse, the only one now remaining. When Pyrrhus heard this from the envoys he abandoned Tarentum and the Italiots on the coast, and crossing into Sicily forced the Carthaginians to raise the siege of Syracuse. In his self-conceit, although the Carthaginians, being Phoenicians of Tyre by ancient descent, were more experienced sea men than any other non-Greek people of that day, Pyrrhus was nevertheless encouraged to meet them in a naval battle, employing the Epeirots, the majority of whom, even after the capture of Troy, knew no thing of the sea nor even as yet how to use salt. Witness the words of Homer in the Odyssey: —Nothing they know of ocean, and mix not salt with their victuals’. (Text and translation from Perseus: https: / / anast rophe.uchicago.edu/ cgi-bin/ perseus/ citequery3.pl? dbname=GreekNov21&query=Paus .%201.12.5&getid=1); accessed 25/ 01/ 2023. by the soothsayer Teiresias (Od. 11.118-137; Od. 23.266-84; 21 cf. Castiglioni 2016: 114). Yet, Castiglioni and other scholars (e.g. Prestianni Giallombardo 2002: 125) argue that the sanctuary of Dodona was not in such an isolation as often believed but constituted a center by the crossroads of high mobility from different areas either to consult the oracle for future action or to initiate newcomers in the area. The cumulative evidence from the oracular tablets, as well as mythological traditions and information from other sources call for a reevaluation of the importance of the connections between the Dodonean oracle and navigation in general, especially in the western Mediterranean territories. To appreciate the relations of the Dodona oracle with the world of the West the entire picture of what comes west of Epirus should be considered, i.e. the Adriatic coast with the Greek colonies and other commercial stations as well as 112 Georgios K. Giannakis <?page no="113"?> the central and western Mediterranean, especially the southern Italian peninsula and Sicily, where Greek colonization thrived since early times. Thus, we must consider different types of tablets, depending on whether they show direct relevance of the oracle to the colonies of the west, i.e. whether the inquirers come from the colonies of South Italy and Sicily. A second group consists of inquiries made by mainlanders, i.e. by people who live in different parts of mainland Greece who enquire the oracle for advice as to what they should do with regards to a possible journey. Finally, there is a group of tablets that show the relevance of the oracle to the wider area of the Adriatic Sea and the Greek colonies there, in which advice is sought for future action. The reason for making this distinction is that the information drawn from these tablets shows different types of activity and influence of the oracle on a Panhellenic level as well as its possible international connections. This fact may also explain the multidialectalism observed in the language of the tablets of our corpus. As put by Dawson (2016: 325), “[…] connectivity can lead to hybridization, emulation, competition, and changes in group identity, to name but a few possibilities.” All these processes are observed in different tablets of the corpus, yielding information about historical and social changes in the area during antiquity. 6.1 Why and how to travel? As explained by Eidinov (2007: 73) and documented by the textual evidence of the DVC corpus, the reasons for traveling vary. Some inquiries simply raise the question of the journey, with some of them also naming the destination, e.g. LOD 100 mentions the Adriatic city of Adria inhabited by the Tisates: πλε͂ν ἐς Ἀδρίαν ἐς Τισατε͂ς ‘to sail to Adria, to the Tisates’, (for a detailed discussion, see Castiglioni 2016: 114 ff.), whereas in other tablets the destination is left blank, e.g. LOD 160 (5 th -4 th c. BC) περὶ τᾶς ὁρμᾶς ‘regarding the departure’, while in other occasions the exact destination is not directly mentioned but a general reference is made to larger regions and in a somewhat vague way. For instance, in one tablet (Eidinov no. 18) there is reference to Elis, but this seems not to be the town but rather the general area of Elea; similarly, in Eidinov no. 21, Caria is mentioned without being sure which Caria is meant, on which see LOD 129 for an interpretation of the toponym; the same with other toponyms, like Messene in inquiry no. 13, and Sicily in no. 9 in Eidinov’s list. In other cases, the exact way of the journey is the object of the inquiry, i.e. whether it should be conducted by sea or by land. The latter of course is meant to be to some nearby destination reachable by land, e.g. Illyria or some other place in the inland, as e.g. in DVC 3109A (discussed in 5.1) in which The Magna Graecia Tablets in the Dodona Corpus 113 <?page no="114"?> 22 All numbers here refer to the LOD corpus. there is a general reference to a move to a colony but with no specification of the way the journey will be made. Whatever the way one plans to make the trip, among the central concerns are safety and whether the entire enterprise is worthwhile, concerns that are reflected in many inquiries. These concerns are more pronounced with journeys to be held by sea, something that may hint at some fear of the outcome since for most people of this area of Greece seafaring was not a common activity. In any case, as adduced from the textual evidence, most people are interested in traveling for trade and commerce, for temporary work, or for establishing a new permanent residence. Other travelers plan to move to mainland Greece, for example to Alyzeia in Acarnania or to Hermion in Argolis, and the just mentioned Elis in the Peloponnese, as well as the coast of Ambracia, to a less known group of the Πενέσται (a group of serfs(? )) in Thessaly, and then Apollonia, Chimaera (a colony of Corcyra in the Chaonian territory, today’s southern Albania), Epidamnos, and Orikos to the north, and of course the western colonies in Magna Graecia. Castiglioni (2016: 113-114) records the destinations of the inquirers as indicated in the tablets, dating mostly to the fourth century, and these are relatively close to Dodona, such as Alyzeia (no. 63), 22 a Corinthian settlement on the Acarnanian coast, Apollonia (no. 97), and Epidamnos (nos. 98, 99), both Corcyraean-Corinthian colonies of southern Illyria. In other documents reference is made to more distant destinations such as Syracuse (no. 103), Hergetium (no. 75), Messina (no. 106), or Sicily (no. 102), or to other areas of Magna Graecia such as Taras (nos. 5, 132), Thurii (no. 111), Metapontum (no. 146), Heraclea (nos. 7, 16, 132), Hipponion (no. 156), Croton (no. 114), Rhegium (nos. 55, 154), as well as to the midand upper Adriatic Sea region, as in Pharos (nos. 130, 6B) and Adria (no. 100). In addition to the Greeks, the oracle was then also frequented by non-Greek peoples, not only from the neighboring Illyria (e.g. no. 7), and, as it seems based on a small plate, also northern Picene (no. 164). If nothing else, this is an impressive list of places that in one way or another relate to the oracle of Dodona as destination for travelers, besides all other areas of mainland Greece or oversees that appear in the corpus (see section 7 below for more placenames). It seems that of all categories of tablets, those that in one way or another refer to travel are among the most numerous in the corpus. Bonnechere (2017: 74) gives a comparative table of the distribution of themes among the tablets before the publication of DVC and after it (Table 1 below). If we also consider that many tablets that are grouped under some other category, particularly the 114 Georgios K. Giannakis <?page no="115"?> 23 As noted by Bonnechere (2017: 74, n. 36), these figures are based on counts made by Duval (2016: 162 ff.), where more details are given. categories of ‘Work’, ‘Money’, and ‘Success in business’, deal with matters of travel or change of residence in search for a better life, the category ‘Travel’ becomes the most numerous of all tablets. As stated by Eidinov (2007: 72), “This concentration of questions around problems and concerns relating to travel supports an emerging historical picture of the peoples of the Mediterranean being consistently on the move: trading-or stealing-skills and goods, making religious pilgrimage, or migrating in order to find work”. Table 1. The Dodona inquiries in numbers (Source: Bonnechere 2017: 74). This table shows interesting results which may prove important for the study not only of the language but also of the historical, cultural and religious life of the people who surrounded the Dodona oracle or consulted it. 23 This is a fertile ground for investigation, especially for an area of the ancient Greek world that for various reasons has been poorly investigated, remained in the backdrop of history and had not had the proper attention in the past, a fact stressed by Chaniotis (and others in the Soueref 2017 collection) in the concluding statement of his contribution, saying: The Magna Graecia Tablets in the Dodona Corpus 115 <?page no="116"?> The corpus of the Dodona tablets is the result of work that lasted for decades. And I would not be surprised if one or two decades of further intensive work and research will be necessary to fully understand these texts and exploit the information that they contain about dialects, the origin of the worshippers, the divinatory practices, social, economic, and religious phenomena, political institutions, and historical events. The publication of the existing photographs and the taking of new ones, using new technologies, are the most urgent tasks. (Chaniotis 2017: 65) 6.2 The testimonies of the tablets As already said, the tablets that provide information about the western connec‐ tions of the Dodona oracle are of two types, public and private. Both provide important information, but the public tablets are rated as more significant in that they demonstrate the official relations and upgraded connections of the oracle with organized communities and the official governments of the cities of the West. Furthermore, as stated by Lombardo (2017: 117), of still higher significance for our topic are rated the tablets that represent the inquiries to the oracle by cities or individuals from Magna Graecia than those in which a cursory mention is made of some toponym or personal name of western provenience. Thus, following the account given by Lombardo (2017: 117-119), the picture is as follows. 6.2.1 Public (or possibly public) inquiries A clear case is the following inquiry by the people of Taras as given by Lhôte (LOD 5, with useful historical and linguistic commentary; not included in DVC): Facsimile: Autopsy of 1998-decipherment: 116 Georgios K. Giannakis <?page no="117"?> Θεό[ς]. Τύχαι ἀγαθᾶι. [Ἐπερωτῆι] hα πόλις - hα τῶν Ταραν[τίνων] τὸν Δία τὸν Ναῖον καὶ τ[ὰν Διώναν] περὶ παντυχίας καὶ π[ερὶ χωρίων ? ] τὰ χη̣ρῶι (? ) καὶ περὶ τῶν […] Lhôte’s translation is as follows: ‘God. For good fortune. The city, that of the people of Taras asks Zeus Naios and Dione about general prosperity, about the territories under its control and about the-…’ As far as the language (= dialect) and the writing are concerned, the following points are of special interest: The dialect is Doric with some characteristics of the Doric Severior, e.g. the form χηρῶι (Att. χειροῖ); aspiration is indicated with the sign ᛀ of the original tablet which must represent the sign Ͱ which was commonly used in Magna Graecia at the time of the introduction to Athens of the Euclidean alphabet (403/ 2 BC), where the sign H was used for long ē instead (see Jeffery 1990: 183); furthermore, the omega is pointed, much smaller and a little elevated relative to the rest of the letters; finally, pi has a shorter second vertical line  , typical of Laconian and of the Doric colonies of South Italy, especially Taras. In this tablet we have punctuation (the dash ( ̵) in the second line) which may indicate a prepared tablet in advance with the opening formula pre-recorded to which the text of the inquiry was added, a convention that occurs in several tablets. The name of the inquiring city is reconstructed from the part ΤΩΝΤΑΡΑΝ[ , a reading by Lhôte that makes good sense in this context and shows that the inquiring group is the city of Taras. Lhôte suggests that this tablet refers to one of the expeditions to Italy by the Molossian kings, thus dating to the period between 350 to 280 BC. This date differs by a margin of two or more decades from an earlier dating by Vokotopoulou in the last quarter of the 4 th century (1992: 78). Whatever the historical reference may be, the fact remains that the text makes a clear reference to a consultation of the oracle of Dodona by the people of Taras, and this cannot be ignored but be considered as an important piece of information for the connections of the kings of Epirus and of the oracle itself with the Greek cities of the West, a fact also known from historical documents (for a discussion of the historical and cultural context, see The Magna Graecia Tablets in the Dodona Corpus 117 <?page no="118"?> 24 For a more general overview of the topic of mobility, international relations and language contacts in ancient Mediterranean, see Clackson et al. (2020); interestingly enough, despite the subject matter of the volume edited by Clackson et al. (2015), the important corpus of the Dodona lead tablets is absent from any discussion. 25 One of the meanings of the word ὁδός is journey by sea or river, waterway, etc.; cf. similar use of Mod. Gk. δρόμος as in expressions like the farewell wish καλό δρόμο! ‘have a good trip! ’, etc. Parke 1967, Dillon 1997, Hammond 1997, Dieterle 2007, Georgoudi et al. 2012, Meyer 2013, Piccinini 2017, Soueref 2017). 24 A second example is provided by DVC 2766A (= LOD 154; first half of the 5 th c. BC), a tablet that seems to be the inquiry by a certain Rheginos and concerns a journey assumed from hοδο[ν in the second line. 25 It is not entirely clear if it is an individual with this personal name (which is attested in S. Italy and especially in Rhegium) or someone coming from Rhegium, or even the city of Rhegium (see commentary in DVC). Lhôte reconstructs ‘Ρε̄γῖν̣ [οι rather than ‘Ρε̄γῖν̣[ος opted for by DVC; the DVC solution is supported by the participle τυχο̄ ́ν (if the reconstruction is correct) that should refer to the masculine name of the inquirer. In any case, the reference of the name clearly points to S. Italy. Ῥε̄γῖν̣[ος ἐπερōτε͂ι-------] hοδὸ[ν-------τυ]-- 3 χṓν Although Lhôte’s restitution of the tablet is in some ways different from DVC’s, its general validity for our argument does not change (see LOD 154). His reading is: [Αὐτοὶ] ‘Ρε̄γῖν̣[οι-------] [τίνα] hοδὸ̣[ν-----------] [---ἔ]χον[τες-----------] 118 Georgios K. Giannakis <?page no="119"?> 26 In fact, this tablet served as the basis for the reconstruction by Lhôte of the preceding tablet (LOD 154) just discussed. The meaning of the tablet in its DVC edition is: ‘a Rheginos (i.e. a person from Rhegium) asks-… if he should make the journey’. Although problematic with regards to its interpretation (see Lombardo 2017: 117), the next fragmentary tablet is of some interest for our topic. It concerns LOD 155 = DVC 3256 (versions (a) and (b) respectively below). 26 (a) (b) Αὐτοὶ ‘Ρ[ε̄γῖν̣οι Αὐτοκ̣ρ[.][-----] The interpretation is different in the two corpora: DVC gives Αὐτοκ̣ρ[.][- - -], i.e. a person’s name (Αὐτοκρ[άτης] or Αὐτοκρ[ατίδας] vel sim., Lhôte reads Αὐτοὶ ‘Ρ[ε̄γῖνοι - - -], i.e. ‘the people of Rhegium themselves…’, which is completely different from the DVC reading. The alphabet is Chalcidian. Lhôte uses several arguments in favor of his reading, such as (a) the fact that this tablet is next to the preceding one (LOD 154) where again the same name appears, with a distinctive shape of the initial consonant Ρ- (  ). This is the occasional writing of <ρ> in the area in the fifth century BC, as also noted by Jeffery (1990: 183); (b) the dimensions of the two fragments are identical; (c) both fragments belong to the same excavation project; (d) also, note the shape of the gamma (  ). In addition, both tablets are dated to the 5 th c. BC. If Lhôte’s reading is correct, the tablet belongs in the documents that support our thesis of the relations of the Dodona oracle with the S. Italian city of Rhegium. The name of the city of Heraclea and its people appears in a number of tablets, e.g. DVC 515A, 517B, 1852B, 3003A, 3079A, 3111, 3134A, some of which are not sure that they refer to the place or to an individual named Herakleios or a group of people, as they are very fragmentary. It is also possible for some of the tablets to refer to another homonymous city and not the Greek colony in Magna Graecia. For instance, in the following rather fragmentary tablet DVC 3079A (first half of the 4 th c. BC, perhaps slightly later) The Magna Graecia Tablets in the Dodona Corpus 119 <?page no="120"?> [τίν]ι θύον[τες θεῶν] [κα]ὶ τελέ[οντες------] 3 [. .]ν̣ται χώρ̣[αν------] [Ἡρ]ακληΐων [-----]; 2-3 [σώσο]νται; we have the name [Ἡρ]ακληΐων referring to the city of the Heracleans, most likely that of Magna Graecia (cf. Lombardo 2016: 137-138, 2017: 117), but Dakaris has some doubts about the actual location, i.e. which of the thirty different towns of the same name this tablet refers to (see DVC commentary). However, the Lucanian Heraclea is mentioned in a few more tablets (e.g. LOD 6, 132 = DVC 3111 and 3134A, see below). Perhaps this is an inquiry to the oracle for protection of their city as can be assumed from the reconstructed verb [σώσο]νται. The meaning is something like the following: ‘(so and so ask) which of the gods should they sacrifice and make offerings to … so they protect the country/ city of the Heracleans’. Another possibly relevant case is the following mid-4 th c. BC tablet DVC 3134A (= LOD 17) with the name Herakleios mentioned in it perhaps as the inquirer, but not much more can be said because of the extremely fragmentary character of the tablet. Note the reading by Lhôte as Ἡρακλείας in version (b) below, the name of the city rather than a person’s name; also, some fluctuation concerning the writing of rho, in the first case an  (on which see our earlier comment) and in the second case a regular Ρ; also mark the conical shape of the omega ( again see our earlier comment on LOD 5), as well as the almost lunar shape of the sigma in the word [- - -]ώρας. Irrespective of the reading, the fact remains that the name either as an ethnikon or as a city name appears in the tablet. (a) DVC 3134A: 120 Georgios K. Giannakis <?page no="121"?> 27 This is a technical meaning of the middle verb ἀπογράφομαι referring to one’s enlisting or registering in the catalogs of citizens of a city (see LSJ s.v. ἀπογράφω), still present in the language today (cf. also noun ἀπογραφή ‘registration of people, census’). [-----] Ἡράκλειος πὲρ [-----] [-----]ώρας τίνι κα θ̣ε̣ῶν θ̣[ύων-----] (b) LOD 17: [περὶ ? ] Ἡρακλείας ΑΡ------ ------ ΩΙΑ τίνι κα <θεῶν ? >------ 6.2.2 Private inquiries A straightforward case and in many ways interesting tablet is DVC 3111 (= LOD 132) which records a private inquiry: Νικόμαχος ἐρωτ̣ῆι τὸν Δία τὸν Νάϊον ἦ ἀπογραψάμ[ε]νό[ς] κα ἐς Τ[ά]ραντα ἐξ Ἡρακληΐα[ς] ἄμεινον πράσσοι ‘Nikomachos asks Zeus Naios if it would be better for him to register as citizen in Taras moving from Heraclea’. This is an inquiry by Nikomachos, in which two cities are mentioned, Taras and Heraclea. The inquirer consults the oracle whether it would be better for him to change citizenship (ἀπογραψάμενος) 27 and perhaps also residence moving from Heraclea to Taras (ἐς Τ[ά]ραντα ἐξ Ἡρακληΐα[ς]), a fact that is important enough to ask the help of the oracle. After all, as suggested by Lhôte in his commentary, there may be an allusion here to historical facts and developments in the area after the Epirote king Alexandros campaigned against Italy, where he met a terrible fate. Many residents of Heraclea that came under the influence of the victorious Lucanians may have decided to flee the city and seek refuge in other Greek cities, with Nikomachos being one of them. The involvement of the oracle in such a private matter transcends the private sphere and impacts the public life as well The Magna Graecia Tablets in the Dodona Corpus 121 <?page no="122"?> 28 43 tablets of the corpus have land cultivation as subject of inquiry (see Chaniotis 2017: 63, n. 52). of the areas affected by such events, like population shifts and movements which also affect both the politics and the demographics of the area. Although private inquiries tend to be of lesser significance than the public inquiries, yet they provide important information about the relations of the Dodona oracle with the world of Magna Graecia, either through their consulta‐ tion with the oracle on a possible action decision or on some other personal or family matter. Lombardo (2017: 118-119) discusses three more such cases, admittedly a small number but still present in the corpus. The first is DVC 2333B-2334B (= LOD 146), a tablet dating to the second half of the 4 th c. BC, in which we have just two words, the name Archias and his place of origin (ethnikon), Metapontium, an important coastal town by the mouth of the gulf of Taras. The rest of the tablet with the inquiry itself has not been recovered. The name Ἀρχίας is common in the Doric areas of both mainland Greece and of the Doric colonies of Magna Graecia (see LGPN, vol. IIIA). Μεταποντῖνος Ἀρχίας Another tablet (DVC 1432 = LOD 75), dating to the second half of the 4 th c. BC, comes from Hergetium (Greek Ἑργέτιον, also without initial aspiration Ἐργέτιον, or even Σεργέντιον), a small town in the interior of Sicily, north of Hybla Heraea, also mentioned in the Delphian list of the theōrodókoi (col. IV 106; see New Pauly s.v.). This tablet offers the full inquiry by Agelochos who comes from this town (ἐξ Ηεργετίω hορμημένωι) asking the oracle if it would be preferable for him to cultivate the land, a simple concern of a simple person. 28 Among the typical Doric features (some of the strict Doric variety) in the language of the tablet are: the Doric Severior ω and η for ου and ει vocalism as in the name of the town Hεργετίω, the infinitive in -ην in γαοργῆ(ν), and the Northwest form in -είμενος/ -ήμενος of the participle hορμημένωι of the present ὁρμέω (with a shift from ὁρμάω; cf. Buck 1955: 125, Méndez Dosuna 1985: 489 ff.; but see also the comments by Lhôte in LOD 75 who wonders whether we have a perfect form here without reduplication, not a likely possibility), aspiration 122 Georgios K. Giannakis <?page no="123"?> (h) is indicated with the sign Ͱ, all reflecting the dialect of the Sicilian colony of the origin of the inquiry. See further comments in DVC. Ἀγελόχωι ἐξ hεργετίω hο- 3 ρμημένωι ἄμεινόν ἐστι γαοργῆ(ν); -‘Would it be better for Agelochos who comes from Hergetium to work the land? ’ DVC 24A-25A (= LOD 114) is a late 5 th c. BC tablet written on both sides (ἀμφίγραφον) presenting interesting features both in terms of language and for the clear reference to the place it comes from. The subject of the inquiry is double, property (πανπασίαν, a common theme in the Dodona corpus, cf. the index of DVC II: 567 where more than 70 tablets are mentioned) and change of residence to the city of Croton for the entire family, an important issue on which the oracle is asked for advice. 24A. θεὸς τύχα ἀγαθά· περὶ πανπασίας καὶ περὶ Ϝοικε̄ ́σιος 3 ἰς Κρ(ό)το̄να ἐ̃ βέλτιων καὶ ἄμεινο(ν) 25Α. αὐτο̃ι καὶ γενεᾶι : καὶ γυναι- 6 κί; The Magna Graecia Tablets in the Dodona Corpus 123 <?page no="124"?> ‘God, Good Fortune! Regarding his property and his residence in Croton if it would be preferable and better for him, his family and his wife.’ There are several linguistic points to be made here. Firstly, an archaic form of alphabet is used, e.g. the shape of gamma 〱 in γενεᾶι and γυναικί (cf. also Jeffery 1990, Table of Letters). Most likely the inquirer comes from Epirus and wishes to emigrate to Croton (the intention to move is expressed by the syntax εἰς + acc. Κρότωνα). In the second side of the tablet we have the use of punctuation (: ), not an uncommon albeit not very frequent practice in our corpus. Then, the readings ϝοικε ̄ σιος for ϝοικήσιος (with retention of the digamma ϝ), ἐ̃ for ἦ ‘if ’, Κρ(ό)τōνα for Κρ(ό)τωνα, and the spelling βέλτιων for neuter βέλτιον may suggest insufficient competence by the scribe in the Milesian alphabet (or a case of hypercorrection according to the DOL comment), perhaps due to its recent spread to the Doric territory. Nevertheless, this is a good enough condition to produce a tablet with a mixture of features that could be understood as due to dialectal mixture, although this may not be the actual case (see Giannakis forthcoming). In the spelling ἰς in the phrase ἰς Κρ(ό)τōνα we may have an early instance of the raising of <ει> to <ι> (elsewhere cf. also ἰς τὸν ἐπίλο(ι)πον χρόνον, etc.), which is characteristic of North-West Doric and of other dialects, i.e. the local dialect (see Méndez Dosuna 1985: 49-54, Lhôte 2006: 385-387; see also Buck 1955: 31). It is not clear whether this is the case here or we deal with an instance of the infiltration of a local feature into the language of the tablet. As we note elsewhere (Crespo / Giannakis 2019), in this case we possibly have some type of accommodation in the dialect idiom used by each speaker, so that effective communication is secured with no misunderstanding. To our mind, this process is a major reason for mixed-dialect queries. This dialectal accommodation is proof of cooperation in communication, a process which, along with other factors, could explain how the Greek dialects of the classical period were progressively replaced by the Attic-Ionic koine (or how koine was adopted for official state business as in the case of ancient Macedonia; see Crespo forthcoming). This is not different from the processes described by sociolinguists with respect to dialect changes that have taken place in other parts of the world even in the 20 th century, which received speakers of mutually intelligible varieties of the same language (see, for instance, Trudgill 1986; Kerswill 2002; Kerswill / Trudgill 2005). The next tablet (DVC 35A, 350-300 BC, 355/ 4 as per DOL) is of a somewhat ambiguous status as far as its origin is concerned as it is not beyond all doubt what city exactly is meant by the origin of the inquirer, Κυμαῖος ‘the Cumaean’, in addition to its linguistic consistency. 124 Georgios K. Giannakis <?page no="125"?> 29 This may also be one more way of bolstering the panhellenic position of the oracle by its officials, i.e. to give full identification to the inquirers whose origin is from distant lands. Euboean Kyme is such a place, considering also that the nearby Delphic oracle is being overlooked in this case and a Cumaean chose to consult the Dodona oracle rather than that of Apollo. But see also the next tablet that may come from a yet more distant region. θεὸς τύχαι ἀγαθᾶι καὶ Διὶ Προ̣ναΐωι καὶ Διώναι· Πορῖνος Κυμαῖος Εὐάνδρου εἰρωτᾶ τὸν θεὸν εἰ τὸν ξατράπην 3 καὶ ὕπαρχον θεραπεύοντ[ι] λώϊον καὶ ἄμεινον ἔσται ‘God. Good Fortune! To Zeus Pronaios and Dione: The Cumaean Porinos, son of Euandros, asks the God whether it would be better for him to court the satrap and the governor.’ This is a typical example of a hybrid tablet with a mixture of dialects, Doric and Ionic, where we can clearly see at least the opening of the query written in Doric and the rest in Ionic. Here of course we also have the rather rare case of full identification of the inquirer with name, city of origin and patronymic (Πορῖνος Κυμαῖος Εὐάνδρου); the origin of the inquirer may also explain the Ionic character of the tablet. The rarity of the full identification of the inquirer is explained by the DVC editors with the comment that this happens mainly with tablets concerning inquiries from distant regions or queries by eponymous people (e.g. 41A, 93B, 1075A, 1088A, 1536A, 2333B, 2470A, 2510, 2631, 2648A, 2666B, 2895, 3212B, 4020A). 29 The Ionic features are clear from the bulk of the features of the text; we also observe a kind of ‘Ionization’ even of Zeus’ epithet: instead of the common formulaic use of Νάιος, the more familiar to the (perhaps) Ionic-speaking inquirer form Προνάωι is used (possibly a blend with Ἀθηνᾶ Προναΐα from Delphi; see Tselikas 2018: 255). As for the Doric features, these include the long ᾱ in τύχαι ἀγαθᾶι (although this may be a fossilized phrase used as an opening formula in almost all tablets irrespective of dialect) and Διώναι (the same as the preceding may be true for this form). This is a query by Πορῖνος Κυμαῖος (from Euboean Kyme (? ) or from Cumae, the Euboean colony in Italy (? ), or even from the similarly named city in Asia Minor, a solution opted for by Lhôte 2017: 43-44; cf. also the comment in DOL with a good discussion of the historical background). It is written in Ionic and addressed to τύχαι ἀγαθᾶι καὶ Διὶ Προναΐωι καὶ Διώναι in Doric. The inquirer The Magna Graecia Tablets in the Dodona Corpus 125 <?page no="126"?> accommodates his writing (including the addition of the odd Προ-, since in Doric this is simply Νάϊωι) to that of his addressees perhaps(? ) as a means of gaining their favorable consideration, an instance of accommodation(? ) (see Méndez Dosuna 2018: 266 and 279; see further Giannakis forthcoming). Irrespective of the final identification of the city of origin of the inquirer, the prestige of the Dodona oracle is evident in this tablet, even if it is the Euboean city (most people opt for this interpretation). 6.2.3 Other tablets In DVC 4154B (end 5 th -beg. 4 th c. BC), someone named Archon or Archon(i)das or some other name starting with Ἀρχωhas the anxiety whether he should sail to Sicily, an understandable concern for many people, especially for those who were not so familiar with such an enterprise. This nervousness is felt in the shift from the third person in the first verb (ἱστορε[ῖ) to the first person in the second verb which expresses the inquiry itself in direct speech (πλέω), not an uncommon phenomenon in our corpus. [θε]ὸς τύ[χα] ἀγαθ[ά]· Ἀρχω[-----] ἱστορ̣ε̣[ῖ τὸ]ν θεὸν 3 πότερον πλέω ἐς Σικ[ελίαν] It seems that the prestige and the respect of Dodona’s oracle went well into the eastern parts of the Greek world, at least as far as Rhodes. If LOD 89 is interpreted correctly by Lhôte, then this is one of the few examples reporting that islands of the Aegean get into contact with our oracle for a possible excursion to the West or perhaps on behalf of one of their colonies of the West. This is a rather complex and difficult tablet as it is a palimpsest, with layers of different inquiries scratched on the lead sheet on top of each other. One of them (Side A) is of special interest for our topic, and this is what we will briefly consider below. The question concerns the chances of the inquirer to be successful in conducting some trading in addition and in parallel to his normal occupation: Τύχα ἀγαθά. Ἦ τυγχάνοιμί κα ἐμπορευόμενος ὅπυς καὶ δοκῆι σύμφορον ἔμειν καὶ ἄγων τῆι(? ) κα δοκῆι ἅμα τᾶι τέχναι χρεύμενος. 126 Georgios K. Giannakis <?page no="127"?> 30 The spelling ΕΜΕΙΝ in the tablet surely represents ΗΜΕΙΝ, as in insular Doric (Rhodian, Cretan, etc.) the E stands for secondary η, usually symbolized with [ẹ̄] (cf. also Buck 1955: 31 and 128). ‘Good fortune. Would I be successful trading where I judge profitable, and if I import and export as it seems to me good, while continuing to carry out my craft activity? ’ Lhôte gives an extensive commentary on the text both on linguistic and historical details of its composition; some characteristic points that concern the identification of the tablet are the following: (a) the infinitive form ἔμειν points to the Rhodian dialect, 30 a sub-branch of insular Doric (see Bechtel 1923 II: 646-647). According to Dubois (1989: 299, Index), in the Rhodian colonies of Sicily we have similar infinitival formations: ἀναθέμειν (4x), εἴμειν (8x), δόμειν (1x), ἀποδιδόμειν (1x), ἀποδόμειν (1x). The form δόμειν is found in Dodona in a copy of an original document from Acragas, which is an offer by the Molossians to a group proxeny to Acragantines, cf. Dubois (1989 no. 184, 4 th -3 rd c. BC): (…) ἔδοξε τοῖς Μολοσσοῖς προξενίαν δόμειν τοῖς Ἀκραγαντίνοις ‘the Molossians decided to send a mission to the people of Acragas’. (b) the adverb ὅπῡς has a rather complex prehistory, but the -ῡς seems to be the result of the phonetic contraction from an older form in -υις (see Buck 1955: 103). (c) there is a problem with the adverbial form τῆι, since the Doric forms in -ῆ never have an ι subscript (see Lejeune 1939: 283-284). This could be an error by the original transcriber (Carapanos). Lhôte concludes his comments on the tablet by saying: “We are therefore dealing with a craftsman and merchant who expresses himself in the Rhodian dialect: rather than thinking of a Rhodian proper, we could imagine that we have a native of a Rhodian colony of Sicily, for example Acragas or Gela” (2006: 188). Of course, the involvement of the Dodona oracle is an undisputed fact. It seems that DVC 2367 (second half of the 4 th c. BC) also belongs to this group of inquiries: θεὸς τύχα ἀγαθά· Ἐπίλυτος ἐπερωτῆι τὸν Δία τὸν Νάϊον καὶ τὰν Διώναν τί κα ποιῶν εὐτυχιοῖ καὶ τίνι θεῶν θύσας 3 καὶ πότερα τὰν τέχναν hὰν ἐπαιδεύθην ἐργάζωμαι ἢ ποτ’ ἄλ- The Magna Graecia Tablets in the Dodona Corpus 127 <?page no="128"?> λο τι hορμάσω καὶ ἦ λαμψῶμαι αἴ κ̣’ ἐπιχηρῆι καὶ πότερα τὰν Φαινομέναν γυναῖκα λάβω ἢ ἄλλαν καὶ πότερα καὶ δὴ 6 λάβω ἢ ποτιμένω ‘God, good fortune; Epilytos inquires Zeus Naios and Dione as to what he should do in order to be prosperous in life, and which of the gods he should offer sacrifices, and whether I should continue practicing the trade I have been trained for or if I should try another one, and in case I accept it, if I will be successful, and whether I should take Phaenomena as my wife or some other woman, and whether I should marry (now) or should wait’. The dialect is clearly Doric with certain items that point to the South Italian origin of the tablet, e.g. the inquirer’s name Ἐπίλυτος, which is a common name in this area (see LGPN IIIA, s.v.), as is the use of the sign h- (in the actual writing the symbol <Ͱ> is used) for aspiration (hάν, hορμάσω), and the shape of the omega (on which see our earlier comments). This is a composite private inquiry of general nature comprising many requests, such as well-being (τί κα ποιῶν εὐτυχιοῖ), whether he should continue with his occupation (τὰν τέχναν hὰν ἐπαιδεύθην ἐργάζωμαι) or he should try another one (ἄλλο τι hορμάσω) and if he will be successful in his new profession in the latter case (καὶ ἦ λαμψῶμαι αἴ κ’ ἐπιχηρῆι), as well as whether he should marry a woman called Φαινομένα or some other woman (πότερα τὰν Φαινομέναν γυναῖκα λάβω ἢ ἄλλαν), or whether the marriage should be postponed for later (καὶ δὴ λάβω ἢ ποτιμένω; note the Doric form ποτί for πρός). Furthermore, note the shift from third to first person in the verbs used in the tablet, a common phenomenon in the corpus, perhaps a sign of a relaxed syntax and of a more popular style of composition of the tablet. 7 The role of onomastics In addition, the study of onomastics is another way of establishing the inter‐ national connections of the Dodona oracle with specific geographical areas and places. Names are often associated with specific regions and their speech, serving as identity markers either of the place or of the individual or the group they stand for. In the Dodona corpus we do have such evidence, especially the mention of geographical and place names, e.g. cities and towns or other localities, some ethnic names or names of the people of towns and cities, and 128 Georgios K. Giannakis <?page no="129"?> 31 For the value of the study of Greek personal names as linguistic material, see Morpurgo Davies (2000); for a more general study, see Hornblower / Matthews (2000), for ethnic names, see Fraser (2009), and of course the full account in the LGPN, with vol. IIIA containing 3330 different names from Epirus, although the names from the DVC corpus are not included. personal names that identify as deriving from or tied to Magna Graecia, a fact that became clear from the examples discussed earlier. The number of geographical names that are identified with certainty with Magna Graecia, the Ionian Sea and the Adriatic is relatively small, amounting to about 26 names, some of which may be ethnika. Thus, Lhôte (2006: 442-443) lists such names (with numbers referring to the tablets discussed in the LOD corpus): along the Adriatic coast, mostly the area of Chaonia: Ἀπολλωνία (97), Ἀπολλωνιάτας (50), Βυλλίονες (7), Ἐπίδαμνος (98, 99), Φάρος (130), Χάονες (11), Χεμάριοι (ethnikon; 131), Ὠρίκιοι (ethnikon) and Ὠρικός (2, 54); Magna Graecia: Ἀδρία (100), several forms of the following: Ἡρακλέα (6), Ἡρακλεία (17), Ἡρακληία (132), and the ethnikon Ἡρακλεῶται (6); hεργέτιον (75), Θούριος/ Θούριοι (111), Ἱππώνιον (156) (It. Vibo Valentia), Κρότων (114), Μεσσήνη (106), ethnikon Μεταποντῖνος (146), Ὀριγγαῖοι (ethnikon? ; 126), the ethnikon(? ) ‘Ρηγῖνοι (154, 155), Σικελία (102), Σύβαρις (133), Συράκοσαι (103), ethnikon Ταραντῖνοι (5) and toponym Τάρας (132), Τισατε̃ς (100). In addition, the DVC corpus contains a few more names that relate to the areas mentioned in the tablets, among them (possibly) Βάριον (562A) (modern Bari), Δεξαιρεᾶται (1070A) (a Chaonian phylum), Καρχᾱδών (1363A), Λίσ(σ)ον (3200A) (in the Dalmatian coast; see DVC II: 474-476 for the relevant tablets). This may not be an impressive list of cities that find a place in our corpus but is long enough to point to the significance of the oracle in the public and private life of the inquirers deriving from these places, also pointing to the sanctuary’s international status in the ancient world. A full and detailed study of the onomatological material (not only place names but ethnika, personal names and others) of the Dodona corpus would be worthwhile, since many relations and associations of the oracle can be extracted from onomastics, along with other useful information for the overall study of the area (cf. Filos 2018b with bibliography; see also Crespo / Giannakis forthcoming). 31 8 Concluding note In the introductory remarks in their recent book Migration, mobility and language contact in and around the Ancient Mediterranean, Clackson et al. state: “One spur for much recent research has been the programmatic work The Magna Graecia Tablets in the Dodona Corpus 129 <?page no="130"?> of Peregrine Horden and Nicholas Purcell (2000), The Corrupting Sea, which sets forth a model of a ‘connected’, ‘networked’ Mediterranean in ancient and medieval history”. They continue adding that “In these and other works, ‘connectivity’ has come to be used as a loose term describing the different ways in which individuals, goods and ideas have moved around the Mediterranean and even further afield, and how geographically distant regions came into close contact through an aggregation of many short journeys” (2020: 3). In the present study, we applied the same spirit focusing on the concept of ‘connectivity’ and on the ‘many short journeys’ that the ancient sanctuary of Dodona helped create in its consultatory role and its ‘wise’ replies to the inquiries by simple people or by communities of the Greek colonies of Magna Graecia. The picture emerging from this study is that the sanctuary of Dodona played an important role in the private and public life of the people of the Greek colonies of the West, i.e. of the circum-Adriatic region and of South Italy and Sicily. Compared to the bulk of the full corpus of this epigraphic material, the relevant tablets may not form a large part, but their significance is high as they establish this connection of Dodona with other cities of the Greek colonies of the West having maritime and other activities. It is worth noting here that some tablets may not be directly relevant to the relations and connections of the Dodona oracle to the South Italian Greek colonies, e.g. DVC 24A-25B (whose inquirer may be an Epirote), or the ambiguous DVC 35A. However, as put by Lombardo, they “can be appreciated from our point of view only as testimonies of the behaviour and of the trust placed on the Oracle by people coming from ‘this side of the Ionian poros’, that is from Aegean and continental Greece, like Athenians, Ionians, B[o]eotians, or even Epirotes” (2017: 119). But now we see that many people enquiring the Dodona oracle come from ‘the other side of the Ionian poros’ as well, i.e. from the Greek colonies in South Italy and Sicily. As a general conclusion then, we can concur with the conclusion reached by Vokotopoulou over thirty years ago about the overall position of the oracle in the ancient world: Dodona was the oldest and the most venerable oracle of Greece, but its role in Panhellenic affairs was secondary. On the other hand, it must have been the closest to simple people, who were concerned about their everyday problems, always ready to help them, without extravagant demands, both humble and intimate like its pilgrims” (1992: 90), a fact turning the Dodona oracle into a nuclear sacred center in a wider ‘connected’ and ‘networked’ Mediterranean world in ancient times. 130 Georgios K. Giannakis <?page no="131"?> Bibliographical references Bartoněk, Α. (1966) Development of the long-vowel system in Ancient Greek dialects, Prague, Státní Pedagogické Nakladatelství. Bechtel, F. (1923) Die griechischen Dialekte, II, Die westgriechischen Dialekte, Berlin, Weidmann. Bonnechere, P. (2017) “Oracles and politics in Ancient Greece, in regard to the new lamellae of Dodona: a needed palinody”, in Soueref, K. (ed.), 67-78. Buck, C.D. (1955) The Greek dialects. Grammar, selected inscriptions, glossary. Chicago & London, University of Chicago Press. Cabanes, P. (1976) L’Epire de la mort de Pyrrhos à la conquête romaine, 272-167 a.C., Paris, Pu De Franche Comte. Cabanes, P. (1988) “Les concours des Naia de Dodone”, Nikephoros 1: 49-84. Cabanes, P. (1999) “La tradition de la migration troyenne en Épire et en Illyrie méridio‐ nale”, in Braccesi, L. & M. Luni (eds.) I Greci in Adriatico, 1, Hesperia 15, Rome, L’Erma di Bretschneider, 61-66. Casevitz, M. (1985) Le vocabulaire de la colonisation en grec ancien. Étude lexicoligique: les familles de κτίζω et de οἰκέω-οἰκίζω, Paris, Klincksieck. Cassio, A.C. (2007) “Enquiries and responses: two lead tablets from Dodona”, in Hatzo‐ poulos, M.B. & V. Psilakakou (eds.) Φωνῆς χαρακτὴρ ἐθνικός (Actes du V e Congrès International de Dialectologie Grecque, Athènes 28-30 Septembre 2006), Athens, Diffu‐ sion de Boccard, 29-34. Cassio, A.C. & D. Musti (1991)-(eds.) Tra Sicilia e Magna Grecia. Aspetti di interazione culturale nel IV sec. a. C.-(Atti del convegno Napoli 19-20 marzo 1987), Pisa, Edizioni Dell’Ateneo. Castiglioni, M.P. (2016) “Dodona e il commercio nell’Adriatico: a proposito della lamella oracolare sui Tisates”, in Struffolino, S. (ed.) Ἡμέτερα γράμματα. Scritti di epigrafia greca offerti a Teresa Alfieri Tonini, Milan, Ledizioni, 113-130. Chaniotis, A. (2007) “The historical significance of the Dodona’s tablets”, in Soueref, K. (ed.), 51-65. Chapinal-Heras, D. (2019) “Territory and boundaries. The role of Molossian sanctuaries”, Dialogues d’histoire ancienne 45.2: 147-165. Chapinal-Heras, D. (2021) Experiencing Dodona. The development of the Epirote sanctuary from Archaic to Hellenistic times, Berlin & Boston, De Gruyter. Clackson, J. (2015) Language and society in the Greek and Roman worlds, Cambridge, Cambridge University Press. Clackson, J. et al. (eds.). (2020) Migration, mobility and language contact in and around the Ancient Mediterranean, Cambridge, Cambridge University Press. Crespo, E. (2023) “Dialects in contact in the ancient kingdom of Macedon”, in Cassio, A.C. & S. Kaczko (eds.) Alloglōssoi. Multilingualism and Minority Languages in Ancient The Magna Graecia Tablets in the Dodona Corpus 131 <?page no="132"?> Europe (Trends in Classics - Greek and Latin Linguistics, no. 2), Berlin & Boston, De Gruyter, 59-81. Crespo, E. & G.K. Giannakis (2019) “Dialectally hybrid inquiries in the Dodona tablets”, Lingvarvm Varietas 8: 43-61. Crespo, E. & G.K. Giannakis (forthcoming) “The provenance of the Dodona Oracle consultants”, Paper presented at Multilingualism and Minority Languages in Ancient Europe MuMiL-EU Final Congress HERA (Humanities in the European Research Area) - Joint Research Project - “Uses of the Past”, Università di Roma 26 th -27 th June 2019. Dakaris, S.; I. Vokotopoulou; A.-F. Christidis (2013) Τα χρηστήρια ελάσματα της Δωδώνης των ανασκαφών Δ. Ευαγγελίδη, 2 vols., S. Tselikas (ed.), Athens, Βιβλιοθήκη της εν Αθήναις Αρχαιολογικής Εταιρείας, nos 285 and 286. Dawson, H. (2016) “‘Brave new worlds’: Islands, place-making and connectivity in the Bronze Age Mediterranean”, in Molloy, B.P.C. (ed.) Of Odysseys and oddities. Scales and modes of interaction between prehistoric Aegean societies and their neighbours, Oxford & Philadelphia, Oxbow Books, 323-341. De Sensi Sestito, G. & M. Intrieri (2011) (eds.) Sulla rotta per la Sicilia: l’Epiro, Corcira e l’Occidente (= Diabaseis 2), Pisa, ETS. Dieterle, M. (2007) Dodona. Religionsgeschichtliche und historische Untersuchungen zur Entstehung und Entwicklung des Zeus-Heiligtums, Hildesheim; Zürich; New York, Olms. Dillon, Μ. (1997) Pilgrims and pilgrimage in Ancient Greece, London & New York, Routledge. DOL = Dodona Online. Domínguez, A.J. (2018) (ed.) Politics, territory and identity in Ancient Epirus. Pisa, ETS. Dubois, L. (1989) Inscriptions grecques dialectales de Sicile. Contribution à l’étude du vocabulaire grec colonial, Rome, École Française de Rome. Dunbabin, T.J. (1948) The Western Greeks. The history of Sicily and South Italy from the foundation of the Greek colonies to 480 B.C., Oxford, Clarendon. Duval, N. (2016) La divination par les sorts dans le monde oriental méditerranéen du II e au VI e siècle après J.-C. : Étude comparative des sortes Homericae, sortes Astrampsychi et tables d’astragalomancie en Asie mineure, PhD Diss. University of Montreal. DVC: see Dakaris; Vokotopoulou; Christidis 2013. Eidinov, E. (2007) Oracles, curses, and risk among the Ancient Greeks, Oxford & New York, Oxford University Press. Filos, P. (2018a) “The dialectal variety of Epirus”, in Giannakis, G.K.; E. Crespo; P. Filos (eds.), 215-247. Filos, P. (2018b) “Linguistic Aspects of Epirote Ethnics”, in Domínguez, A.J. (ed.), 283-301. Fontenrose, J.E. (1978) The Delphic oracle, its responses and operations, with a catalogue of responses, Berkeley; Los Angeles; London, University of California Press. 132 Georgios K. Giannakis <?page no="133"?> Fraser, P.M. (2009) Greek ethnic terminology, Oxford & New York, Oxford University Press. Garland, R. (2014) Wandering Greeks. The Ancient Greek diaspora from the Age of Homer to the death of Alexander the Great, Princeton & Oxford, Princeton University Press. Georgoudi, S.; R.K. Piettre; F. Schmidt (2012) (eds.) La Raison des signes. Présages, rites, destin dans les sociétés de la Méditerranée ancienne, Leiden & Boston, Brill. Giacomelli, R. (1988)-Achaea Magno-Graeca. Le iscrizioni arcaiche in alfabeto acheo di Magna Grecia, Brescia, Paideia. Giannakis, G.K. (ed.) (2015) Ο γλωσσικός χάρτης της κεντρικής και βόρειας Ελλάδας κατά την αρχαιότητα, Thessaloniki, Centre for the Greek Language. Giannakis, G.K. (forthcoming) “Dialect convergence and linguistic change: the Dodona tablets corpus and its significance for the study of the history of the Greek language”, in di Bartolo, G. & D. Kölligan (eds.) Postclassical Greek: problems and perspectives (Trends in Classics - Greek and Latin Linguistics, no. 4), Berlin & Boston, De Gruyter. Giannakis, G.K.; E. Crespo; P. Filos (2018) (eds.) Studies in Ancient Greek dialects: from central Greece to the Black Sea (Trends in Classics Supplementary Vol. 49), Berlin & Boston, De Gruyter. Guzzo, P.G. (1996) La Magna Grecia. Italici e italioti, Turin, Electa & Gallimard. Hammond, N.G.L. (1967) Epirus. The geography, the ancient remains, the history and the topography of Epirus and adjacent areas, Oxford, Clarendon. Hammond, N.G.L. (1997) “Αρχαία Ήπειρος”, in Sakellariou, M.B. (ed.) Ήπειρος. 4000 χρόνια ελληνικής ιστορίας και πολιτισμού, Athens, 34-62. Hornblower, S. & E. Matthews (2000) Greek personal names. Their value as evidence, Oxford & New York, Oxford University Press. Jeffery, L.H. (1990) The local scripts of Archaic Greece: a study of the origin of the Greek alphabet and its development from the eighth to the fifth centuries B.C., 2 nd ed. with a supplement by A.W. Johnston, Oxford, Oxford University Press. Kerswill, P. (2002) “Koineization and accommodation”, in Chambers, J.K.; P. Trudgill; N. Schilling-Estes (eds.) The handbook of language variation and change, Malden; Oxford; Victoria, Wiley, 669-702. Kerswill, P. & P. Trudgill (2005) “The birth of new dialects”, in Auer, P.; F. Hinskens; P. Kerswill (eds.) Dialect change: convergence and divergence in European languages, Cambridge, Cambridge University Press, 196-220. Lejeune, M. (1939) Les adverbes grecs en -θεν, Bordeaux, Delmas. Lejeune, M. (1987 [1972]) Phonétique historique du mycénien et du grec ancient, Paris, Klincksieck. Lhôte, É. (2006) Les lamelles oraculaires de Dodone (Hautes études du monde gréco-romain 36), Geneva, Droz. Lhôte, É. (2017) “La datation des textes oraculaires de Dodone”, in Soueref, K. (ed.), 41-49. The Magna Graecia Tablets in the Dodona Corpus 133 <?page no="134"?> LOD: see Lhôte, É. 2006. Lombardo, M. (2016) “Eraclea ed Eracleoti nelle laminette oracolari di Dodona”, in Struffolino, S. (ed.), Ἡμέτερα γράμματα. Scritti di epigrafia greca offerti a Teresa Alfieri Tonini, «Aristonothos» 12: 131-150. Lombardo, M. (2017) “The oracle of Dodona: a view from the other side of the Ionios Poros”, in Soueref, K. (ed.), 113-122. Méndez Dosuna, J. (1985) Los dialectos dorios del noroeste. Gramática y estudio dialectal, Salamanca, Ediciones Universitad de Salamanca. Méndez Dosuna, J. (2007a)-“Notes de lecture sur les lamelles oraculaires de Dodone”, Zeitschrift für Papyrologie und Epigraphik 161: 137-144. Méndez Dosuna, J. (2007b)-“Παρατηρήσεις στις νέες μαντειακές πινακίδες της Δωδώνης”, in Studies in Greek Linguistics 27. Proceedings of the Annual Meeting of the Department of Linguistics, School of Philology, Faculty of Philosophy, Aristotle University of Thessaloniki, May 6-7, 2006. In Memoriam A.-F. Christidis, Thessaloniki, 277-285. Méndez Dosuna, J. (2014) “Northwest Greek”, in Giannakis, G.K. et al (eds.) Encyclopedia of Ancient Greek Language and Linguistics, Leiden & Boston, Brill, vol. 2: 518-524. Méndez Dosuna, J. (2016) “Some critical notes on the new Dodona lead plates”. Zeitschrift für Papyrologie und Epigraphik 197: 119-139. Méndez Dosuna, J. (2018a) “The language of the Dodona oracular tablets: The non-Doric inquiries”, in Giannakis, G.K.; E. Crespo; P. Filos (eds.), 265-296. Méndez Dosuna, J. (2018b) “West by Northwest: a linguistic survey of the Doric oracular inquiries of Dodona”, in Kalaitzi, M.; P. Paschidis; C. Antonetti; A.-M. Guimier-Sorbets (eds.) Βορειοελλαδικά. Tales from the Lands of the ethne. Essays in Honour of Miltiades B. Hatzopoulos (MELETEMATA 78), Athens, 33-52. Meyer, E.A. (2013) The inscriptions of Dodona and a new history of Molossia. Stuttgart, Steiner. Molinelli, P. (ed.) (2017) Language and identity in multilingual Mediterranean settings, Berlin & Boston, De Gruyter. Morpurgo Davies, A. (2000) “Greek personal names and linguistic continuity”. in Pro‐ ceedings of the British Academy 104: 15-39. Parke, H.W. (1967) The oracles of Zeus: Dodona, Olympia, Ammon, Oxford, Blackwell. Parker, R. (2016) “Seeking advice from Zeus at Dodona”, Greece & Rome 63: 69-90. Pease, A.S. (1920) M. Tulli Ciceronis De Divinatione, Liber primus, Urbana, University of Illinois Press. Piccinini, J. (2013) “Beyond prophecy. The oracular tablets of Dodona as memories of consultation”, Incidenza dell’Antico 11: 63-76. Piccinini, J. (2017) The shrine of Dodona in the Archaic and Classical Ages. A history, Macerata, Edizioni Università di Macerata. 134 Georgios K. Giannakis <?page no="135"?> Poccetti, P. (2017) “Santuari oracolari e pratiche della scrittura: Grecia e Italia a con‐ fronto”, in Soueref, K. (ed.), 79-94. Poccetti, P. (2019) “Forme linguistiche e non linguistiche dell’ oracolarità di Dodona”, in Malacrino, C.; K.I. Soueref; L. Vecchio (eds.) Dodonaios. L’oracolo di Zeus e la Magna Grecia, Reggio Calabria, Laruffa, 115-131. Prestianni Giallombardo, A.M. (2002) “L’oracolo di Dodona e le navigazioni adriatiche nei secoli VI-IV a.C.”, in Braccesi, L. & M. Luni (eds.) I Greci in Adriatico, 1, «Hesperìa» 15: 123-136. Rinaldi, E. (2021) “The meeting halls of the Hellenistic Epirus”. «Thiasos»: Rivista di archeologia e architettura antica 10.1: 371-400. Rohlfs, G. (1977)-Grammatica storica dei dialetti italogreci (Calabria, Salento), Munich, Beck. Schwyzer, E. (1953) Griechische Grammatik, vol. I, Munich, Beck. Sihler, A.L. (1995) New comparative grammar of Greek and Latin, New York & Oxford, Oxford University Press. Soueref, K. (ed.) (2017) Dodona. The omen’s questions. New approaches in the oracular tablets, Ioannina, Ephorate of Antiquities of Ioannina. Threatte, L. (1980) The grammar of Attic inscriptions, vol. 1, Phonology, Berlin & New York, De Gruyter. Tribulato, O. (ed.) (2012) Language and linguistic contact in Ancient Sicily, Cambridge, Cambridge University Press. Trudgill, P. (1986) Dialects in contact, Oxford & New York, Blackwell. Tselikas, S. (2015) “Οι δωρικές διάλεκτοι στο corpus των χρηστηρίων ελασμάτων της Δωδώνης”, in Giannakis, G.K. (ed.), 61-73. Tselikas, S. (2017) “Χρηστήρια ελάσματα της Δωδώνης. Εκδοτικά και ερμηνευτικά ζητήματα”, in Soueref, K. (ed.), 35-39. Tselikas, S. (2018) “The Doric dialects in the corpus of the oracular tablets from Dodona”, in Giannakis, G.K.; E. Crespo; P. Filos (eds.), 249-264. Vecchio, L. (2017) “Dodona and Pharos”, in Soueref, K. (ed.), 123-130. Vokotopoulou, I. (1992) “Dodone et les villes de la Grande Grece et de la Sicile”, in Stazio, A. & S. Ceccoli (eds.) La Magna Grecia e i grandi santuari della madrepatria (Atti del trentunesimo convegno di studi sulla Magna Grecia, Taranto, 4-8 ottobre 1991), Taras, Istituto per la storia e l’archeologia della Magna Grecia, 63-90. Willi, A. (2008) Sikelismos. Sprache, Literatur und Gesellschaft im griechischen Sizilien (8.-5. Jh. v.-Chr.), Basel, Schwabe. The Magna Graecia Tablets in the Dodona Corpus 135 <?page no="137"?> From the Files of a new Bechtel, Die historischen Personennamen des Griechischen: 1. Ἀλεξιτέλης ‘the one who guards his company’ or ‘the one who wards off the fatal doom’ (Ηοm. τέλος θανάτοιο) and the “short” forms Ἀλεξιτώ, Ἀλεξιτίδης, Myc. a-re-ki-si-to / Aleksito-/ , and Myc. a-ra-ki-to / Alalkito-/ * José Luis García Ramón (Universität zu Köln / Università Cattolica del Sacro Cuore, Milano) Abstract: The MN Ἀλεξι-τέλης (Ceos, iv/ iii), which underlies the trun‐ cated forms WN Ἀλεξι-τώ (Amorgos, ca. 300), MN Ἀλεξι-τίδης (Νaxos, ca. 600-550), as well as Myc. a-re-ki-si-to / Aleksi-to-/ (Cnossos), may be interpreted as ‘the one who guards his comrades’ (ἀλέξο/ ε-, τέλος ‘military unit’) or, most probably, ‘the one who wards off the fatal doom (τέλος θανάτοιο)’. This is the reflex of the collocation [ WA R D O F F - A S S I G NM E NT of D E ATH ] (τέλος ‘doom’ , cf. ἐπιτέλλo/ ε- ‘to impose’, θάνατον ἐπιτέλλειν Pind.), which is richly attested in Greek in synonymous expressions of both constitutive elements. Μyc. a-ra-ki-to / Alalki-to-/ (Thebes) is a truncated form of *Ἀλαλκι-τέλης (: aor. ἀλαλκo/ ε-, instead of *Ἀλαλκε-τέλης) with fluctuation / -i°/ ~ / -e°/ , as in Myc. a-ke° ~ a-ki° (Alph. Ἀγε° ~ Ἀγι° et al.) or by proportional analogy ἀλέξ-ο/ ε- : Myc. / Aleksi-to-/ : : ἀλαλκ-o/ ε- : Myc. / Alalki-to-/ . Keywords: Greek personal names; Greek lexique; Greek phraseology; Mycenaean; Mycenaean continuity. §1. The personal names formed from ἀλέξο/ ε- : : ἀλαλκo/ ε- (and med. ἀλέξασθαι) ‘to ward off, keep off ’ and ‘to guard, defend’, have been set forth under Ἀλεξε-, Ἀλεξι- (pp. 33-4) and Ἀλξι- (p. 38) in Friedrich Bechtel’s epoch-making Die historischen Personennamen des Griechischen bis zur Kaiserzeit (Halle, 1917: HPNG). The dossier of names available at this time has been <?page no="138"?> * The final version of this paper has benefited from the observations by Riccardo Ginevra (Milano, UCSC), Alexander Nikolaev (Boston), Alan J. Nussbaum (Cornell), and Ana Vegas Sansalvador (Köln). José L. Melena (Vitoria) has kindly shared with me his editions of the texts from Knossos (2019) and Pylos (2021), as well as his new edition of TITHEMY (forthcoming). Warm thanks are due to all of them. The onomastic data are from Fr. Bechtel, HPNG and/ or from (extremely helpful) P.M. Frazer - E.Matthews, LGPN (Lexicon of the Greek Personal Names, Oxford, 1988+: link http: / / clas-lgpn2.classics.ox.ac.uk/ cgi-bin/ lgpn_search.cgi): each name is mentioned by its earliest attestation, all of them dated BCE). For the Mycenean forms cf. the doxography by F. Aura Jorro, DGMic (Diccionario Griego Micénico I-II, Madrid, 1985+), F. Aura Jorro et alii, DMic. Supl. (Madrid 2020). Standard works are mentioned according to the common conventions. Vedic translations are taken from Jamison & Brereton 2014. Abbreviations: ΚΝ(ossos), PY(los), TH(ebes). MN: man’s name, WN: woman’s name; “short” form: truncated form of a compounded name. 1 These lemmata match Bechtel’s , κος and κε-, κο- (HPNG, 35), κεσ-, κης (HPNG, 36) and κι- (HPNG, 37-38). enlarged by new evidence from Alphabetical Greek (and Cypriot) and, since 1953, from Mycenaean, which should be integrated under the lemma ἀλέξο/ ε- : ἀλαλκo/ ε- (and ἀλκ-, ἀλκή *1 as a separate lemma) in a reelaboration of Bechtel’s Personennamen (in preparation), which will systematically take into account the phraseological patterns underlying the compounded names (García Ramón 2021). The present contribution takes as its starting point the essen‐ tials about semantics and collocations of ἀλέξο/ ε- : : ἀλαλκo/ ε- (§2) and the possibility to interpret names with Ἀλεξ(ι)° as reflecting poetic (or non poetic) phraseology (§3), and will focus on three concrete issues: on the one hand, three truncated variants of the compounded man’s name (MN) Ἀλεξι-τέλης (Ceos, iv/ iii), which were not known at the time of Bechtel’s Personennamen, namely the woman’s name (WN) Ἀλεξιτώ (Amorgos, ca. 300), and its masculine coun‐ terpart (MN) a-re-ki-si-to / Aleksito-/ (Knossos), as well as the MN Ἀλεξιτίδης (Naxos, ca. 600-550), all of which presuppose the existence of an unattested Alph. Gk. *Ἀλέξιτος (§4); on the other hand, the semantics of Ἀλεξι-τέλης, which may be elucidated in the light of the different meanings of τέλος (§5) and of their occurrences and their synonyms in the Greek poetic phraseology (§6-7); finally the enigmatic MN a-ra-ki-to in the new Mycenaean tablets from Thebes, which, in my opinion, conceals / Alalki-to-/ , formed from aor. ἀλαλκo/ ε-, like a-re-ki-si-to / Aleksi-to-/ from pres. ἀλέξο/ ε- (§8-10). §2. The semantics and constructional pattern of ἀλέξο/ ε- : : ἀλαλκo/ ε- (only active, beside med. ἀλέξασθαι) is a twofold one, namely (a) ‘to ward off, keep off ’(‘someone / something’ [accusative]), ‘from / in favour of someone / some‐ thing’ [dative]) and (b) ‘to guard, stand by’ (someone’ [dative], secondarily also 138 José Luis García Ramón <?page no="139"?> 2 The synonymy of ἀλέξο/ εwith ἀμύνο/ εand βοηθέο/ εwas explicitly mentioned by ancient scholars, cf. e.g. the Hesychian glosses ἀλεξήσουσαν · βοηθοῦσαν (ad Il. 8.365), ἤλαλκε · ἤμυνεν, ἐβοήθησεν, and ἀλάλκῃ · ἀπαλεξήσῃ. βοηθήσῃ, also ἀλεξομένους · βοηθουμένους, ἀμυνομένους (ad Hdt 1.211.2 … τούς τε λειφθέντας τῆς Κύρου στρατιῆς ἐφόνευε ἀλεξομένους). 3 Cf. Haudry 1977: 202-4; Hill 2012 (*h 2 lekso/ e- “ab-, fern-halten”). Οn the specific shift to ‘to imped, prohibite’ in Vedic (ŚBr), cf. Oertel 1936: 79-80; Griepentrogg 1995: 53-4. [accusative]). Both senses are shared by synonymous ἀλεείνο/ ε-, ἀλεύασθαι, ἀρκέο/ ε- (Hom.+), and especially ἀμύνo/ ε- 2 (s. above and §6). Some instances: (a) with κακὸν ἦμαρ ‘fatal day’: Ιl. 9.251 φράζευ ὅπως Δαναοῖσιν ἀλεξήσεις κακὸν ἦμαρ ‘consider how you may ward off from the Danaans the fatal day’, Od. 10.288 ἔρχευ, ὅ κέν τοι κρατὸς ἀλάλκῃσιν κακὸν ἦμαρ ‘it shall ward off you from your head the evil day’. Cf. also the formula / ἀμύνετο νηλεὲς ἦμαρ # (Il. 11.484=13.514), Od. 8.525 with κακὸν ἦμαρ: ἄστεϊ καὶ τεκέεσσιν ἀμύνων νηλεὲς ἦμαρ ‘seeking to ward off from the city and his children the pitiless day’). - With acc. κύνας ‘dogs’: Ιl. 23.185 ἀλλὰ κύνας μὲν ἄλαλκε Διὸς θυγάτηρ Ἀφροδίτη ‘but kept dogs (sc. from Hector) … Aphrodite’, Ιl. 13.475 ἀλέξασθαι μεμαὼς κύνας ἠδὲ καὶ ἄνδρας ‘eager to ward off dogs and men’ (sc. Idomeneus, compared with a boar which whets his tusks). (b) with an animate (dative): Il. 5.779 (αἱ δὲ βάτην…) ἀνδράσιν Ἀργείοισιν ἀλεξέμεναι μεμαυῖαι ‘(they [sc. the goddesses] went …) in their eagerness to stand by the men of Argos’, Ιl. 22.196 εἴ πως οἷ καθύπερθεν ἀλάλκοιεν βελέεσσι ‘hoping if they from above might somehow defend him with missiles …’, Il. 15.565 Ὣς ἔφαθ’, οἳ δὲ καὶ αὐτοὶ ἀλέξασθαι μενέαινον ‘and they were eager to ward off themselves’. Cf. also, with ἀμύνο/ ε-, Il. 6.262 ὡς τύνη κέκμηκας ἀμύνων σοῖσιν ἔτῃσι ‘just as you are weary with defending your fellows’. - With accusative and genitive: Od. 17.364 ἀλλ’ οὐδ’ ὧς τιν’ ἔμελλ’ ἀπαλεξήσειν κακότητος ‘not even so would she save one of them from ruin’. The prehistory of the verb is of no major help for the interpretation of the names dealt with in the present paper, but some interesting points may be shortly remembered at this point. Gk. ἀλέξο/ εand Ved. rákṣ-ami , practically with the same semantics and constructions, 3 clearily point to an inherited *h 2 léks-o/ e-. Contrarily, the reduplicated aor. ἀλαλκo/ ε- (*h 2 (e)l-h 2 lk-ó/ é-, ac‐ tually a Transponat) has no correspondence in Vedic, where rakṣ has developed to a neo-root, with aor. rakṣ-iṣ-, as well as a perf. rarákṣand an “intensive” rārakṣ- [hapax ptc. rārakṣāṇá-, §8], which do not formally match ἀλαλκo/ ε-. Whether IE *h 2 leks-o/ eis to be analyzed as a -s-present, or as an originally desiderative formation (Aktionsart! ) or as a root-extension of IE *h 2 lek- (or From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 139 <?page no="140"?> 4 “Desiderativmorphem -s-” (Leukart 1994: 227 n.251); “-s-Bildung (Desiderativ)” (M. Kümmel, LIV 2 s.v. *h 2 leks- “fernhalten”; “eine alte Bildung auf -se/ o-” (Hill 2002: 261). 5 Hill 2012: 250-9, esp. 255-6 (“negativ sprechen”). The putative Vedic root rakṣ- 2 ‘injure, harm’ (Av. ras ‘id.’), Ved. rákṣas- ‘daming, evil genius’ (and rakṣás- ‘fiend, injurer, evil genius’): OAv. rašah- ‘id.’ is, in my opinion, easier to be connected semantically with PGm. *lahe/ aeasier than with ἀλέξο/ ε- : Ved. rákṣa-. 6 Probably formed from (nominal) *alk-ti- (Benveniste 1948: 35, 44; Heubeck, 1957: 274 = Kl.Schr. 487) not from an “aor. zu ἄλκω” (pace Bechtel, HPNG). The MN Ἀλξιάδας (Lindopolitai, ca. 325), Ἀλξιάδης (Οropos, iii) relies on Ἀλξι°, most probably on Ἀλξ-α᷄ νωρ. 7 An opposition between “celui qui a mission de repousser (l’ennemi)” and ἀλεξήτωρ (Ζεῦ ἀλεξήτορ S. OC 143) “qui repousse le danger” (Benveniste 1948: 31, 45, 49) can hardly be stated in the light of the texts. *h 2 lek̂-) 4 is irrelevant for our purposes, as is the possibility of a non-enlarged variant *h 2 lekor *h 2 lek̂-) underlying PGm. *lahe/ a- ‘to insult’, ‘to prohibit’ (OE (be)léan ‘to scold, rebulk, reproach’, “uituperare”, OSax. lahan ‘to revile, vilify’, OHG lahan ‘to keep from something’). 5 §3. For each name with first member Ἀλεξ(ι)° and exceptional Ἀλεξε° (cf. Ἀλεξέ-βιος beside Ἀλεξί-βιος, cf. §10.3), or Ἀλξ(ι)°), 6 as for each compounded name in Greek, the question arises whether (1) it gives a recognisable meaning, i.e. it reflects a phrase or collocation, as is the case of common compounds, e.g. ἀλεξάνεμον · ἐσθῆτα παχεῖαν, καὶ εὐπαγῆ [ad Od.14.529 χλαῖναν … ἀλεξάνεμον ‘which wards off the wind’], or ἀλεξιάρης · ἐπίκουρος, ἀλεξίκακος [ad Hes. Op. 464 νειὸς ἀλεξιάρη … ‘fallow land (is) an averter of ruin …’] respectively (both Hsch.), or (2) it consists of a mere juxtaposition of two lexemes which make no sense, as many other “composés irrationnels” (in Olivier Masson’s formulation). (1) Assured instances of names reflecting recognisable, well-attested syn‐ tagms are, among others, those with °μαχος or with °ανδρος / °ᾱ́νωρ (and °μβροτος). As to Ἀλεξί-μαχος, cf. Il. 9.605 οὐκέθ’ ὁμῶς τιμῆς ἔσεαι πόλεμόν περ ἀλαλκών ‘you will not be in like honour’ even though you may ward off the war’), or 20.396 ἐσθλὸν ἀλεξητῆρα μάχης ‘a mighty warder off of battle’), 7 and synonymous Ἀμυνό-μαχος (an Acharnian, Αttica, ii). The type Ἀλέξ-αvδρoς, with Ἀλεξ-άvδρα (: Myc. a-re-ka-sa-da-ra / Aleksandrā/ MY) and the variants Ἀλεξ-ήνωρ and Ἀλεξί-μβροτος, Ἀλεξι-μβροτίδας and, with Ἀλξ(ι)°, Ἀλξήνωρ : Myc. a-ka-sa-no / Alksānōr/ , is especially interesting, as it may reflect the two meanings of ἀλέξο/ ε- : : ἀλαλκο/ ε-, med. ἀλεξα-), namely ‘the one who wards off the men’ (a) ‘as enemies or (b) ‘guards the men (from a danger)’, e.g. (a) Il. 13.475 ἀλέξασθαι μεμαὼς κύνας ἠδὲ καὶ ἄνδρας (cf. §2), also Od.14.531 κυνῶν ἀλκτῆρα καὶ ἀνδρῶν. (b) Il. 5.779 ἀνδράσιν Ἀργείοισιν ἀλεξέμεναι μεμαυῖαι ‘in their eagerness to stand by the men of Argos’. 140 José Luis García Ramón <?page no="141"?> 8 Alexander Nikolaev points out to me that the root rakṣin vratā́ni … arakṣata may be a replacement of *peh 2 -s-, cf. Hitt. paḫš- ‘to protect’ used with lingai-, pl. lingauš ‘oath(s)’, as well as the use of Old Slovene pasti ‘to preserve’ with rota ‘oath’, an exact cognate of Ved. vratá- (Freising fragments, II.23-24 ese roti, choise ih ne pazem-“oaths which we do not keep”). One can only look forward to further development of this brilliant idea. 9 A phraseological correspondence between Ved. árakṣat … gáyam ‘guarded his patri‐ mony/ provisions’ (RV I 74.2c) and MN Ἀλεξί-βιος, Ἀλεξέ-βιος can be assumed on the strength of the aequabile Ved. gáya- : Gk. βίος (βιοτός). Whether the same may apply to Ved. vratā́ni … arakṣata (VI 8. 2b vratā́ny agnír vratapā́ arakṣata “Agni, as protector of commandments, guarded the commandments”) and ΜΝ Ἀλεξί-δικος, WN Ἀλεξι-δίκᾱ must remain open at this point, as Ved. vráta- (n.) ‘commandment’ hardly matches δίκη. 10 Equally obscure is the semantics of WM Ἀλευαγόρα (Pherai, aet. hellen.). 11 Among others, ‘to have’ (τιμὴν δ´ αὐτὸς ἔχοι Od. 1.117), or ‘to seek’ (μὴ φεύγειν τοὺς πόνους, ἢ μηδὲ τὰς τιμὰς διώκειν Th. 2.63.1), or to ‘find’ or ‘receive’ (τιμὴν εὑρίσκεσθαι, δέκεσθαι Pi. P. 1.48, 8.5) or ‘to lead/ put in’ (ἐν τιμῇ ἐν τιμῇ ἄγειν, or ἄγεσθαι, τίθεσθαι Ηdt.+), ‘to bestow’ or ‘receive’ e.g. θεῶν ἒξ ἔμμορε τιμῆς ‘has received a share of honour from the gods’ (Od. 5.335), ἔμπης τιμὴ καὶ τοῖσιν ὀπηδεῖ (Hes. Op. 142), τοῖς φίλοις τιμὰς νέμειν ‘to pay due regard’ (S. Aj. 1351,+) An instance of Ved. rákṣ nr̥ ̄ ́n perfectly matches ἀλέξο/ εἄνδρας: RV I 174.1ab t u vàṃ rā́jendra yé ca devā́ rákṣā nr̥ ̄ ́n pāhy àsura tvám asmā́n “You are king, Indra: guard those who are gods and the superior men; protect us, lordly one”. 8 The perfect correspondence is actually exceptional. 9 Τhe same collocation underlies compounded names with synonymous ἀμύνο/ ε- (Ἀμύν-ανδρος, Attica, iv+) or ἀλεύο/ ε-, cf. Od. 16.72/ 3 αὐτὸς μὲν νέος εἰμὶ καὶ οὔ πω χερσὶ πέποιθα / ἄνδρ’ ἀπαμύνασθαι, … ‘nor have I yet trust in my strength to defend me against a man’, Il.12.155 βάλλον ἀμυνόμενοι σφῶν τ’ αὐτῶν … ‘men kept hurling stones in defense of their own lives’, also / δηΐων ἀνδρῶν ἀλεωρήν # ‘a defense against the hostile men’ (Βechtel, HPNG, 33) in ἐς πόλεμον φορέειν δηΐων ἀνδρῶν ἀλεωρήν Il. 12.57 (of κρημνοί ‘stakes’ .54), 15.533 (of θώρηξ ‘corselet’ .529). (2) Other names with Ἀλεξ(ι)° do not allow recognizing a precise sense at all. This is, among others, the case of names with °αγóρᾱς (and quasi-synonymous °βουλος, cf. βουλή ‘counsel, council’), or with °φάης and °φάνεια (cf. φάος ‘light’, °φανής ‘shining’): Ἀλεξ-αγόρας (Thera, a. 480-450) 10 or Ἀλεξί-βουλος (Delphi, ca. 208), Ἀλεξιφάης (Pieria, a. 200-166 +), Ἀλεξι-φάνεια (Siphnos, s. d.). Equally opaque is the MN Ἀλεξί-τιμος (frequent in Athens, a. 414/ 3+; also Αmorgos, iv/ iii), which presupposes a “short” form *Ἀλεξί-τος: τιμή, actually ‘honour’, is attested as the object of verbs with different semantics, 11 but not with ‘to ward off, keep off ’, even less of ‘to guard, stand by, keep safe’. To interpret Ἀλεξί-τιμος as ‘the one who averts a punishment’ is hardly conceivable: τιμή means ‘honour, position of honour’, ‘reward’ (e.g. Il. 1.159 τιμὴν ἀρνύμενοι Μενελάῳ σοί τε … πρὸς Τρώων ‘seeking to win honour [not ‘recompense’] for Menelaos and for From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 141 <?page no="142"?> 12 “um Ehre zu gewinnen dem Menelaos” (Schadewaldt), cf. also Il. 5.552 τιμὴν Ἀτρείδῃς … ἀρνύμενοι ‘to win honour for the Atrides‘ (“um den Atreus-Söhnen … Ehre zu gewinnen”). 13 The same applies to Od. 22.57 τιμὴν ἀμφὶς ἄγοντες ἐεικοσάβοιον ἕκαστος ‘bringing in requital the worth of 20 oxen’ (“indem wir zur Buße, jeder für sich, den Wert von 20 Rindern bringen”). you from the Trojans’), 12 and the supposed sense ‘penalty, compensation’ in some few instances is merely contextual (by association to τίνο/ ε- ‘to pay’ in “synchronic” etymology) and ultimately not a real one, cf. Il. 3.288/ 9 εἰ δ’ ἂν ἐμοὶ τιμὴν Πρίαμος Πριάμοιό τε παῖδες / τίνειν οὐκ ἐθέλωσιν Ἀλεξάνδροιο πεσόντος ‘and if they are not minded to pay me the penalty’. 13 In the following an attempt will be made to interpret Ἀλεξι-τέλης as a meaningful compound, i.e. as the reflex of a phraseological pattern which may be elucidated in the light of Greek poetics. A first task is thus to determine the real sense or senses of τέλος underlying °τέλης in Ἀλεξι-τέλης. §4. The name Ἀλεξι-τέλης (Bechtel, HPNG, 34) occurs in three texts from Ceos (Karthaia, iv/ iii: IG XII 5), and refers to two different persons: an official of a dual institution, teamed up with a certain Cleomelos, in honourific decrets for two Athenian citizens (Αλεξιτελης 528.7, 538.5), and an archon in a register of offerings (επι αρχοντος Αλεξιτελους 544. A2,1.36, Αλεξιτελης B2,1.54). Two names (still unknown at the time of Bechtel’s HPNG), namely the WN Ἀλεξιτώ (Amorgos, ca. 300) and the MN Ἀ̣λεξιτίδης (Naxos, ca. 600-550), are surely connected with Ἀλεξι-τέλης on purely formal grounds, regardless of how the semantics of the latter may be interpreted: Aλεξιτω Λαμψου (Minoa, epitaph: SEG 33: 698). το̣ι— — — — — —ν ⋮ και ⋮ Α̣λεξιτιδης ⋮ αν[ε]θ̣ε̣σ̣α̣ν̣ ․․․ ιοι ⋮ Α̣πολōνι (Naxos, dedication: SEG 16: 477). Both names can safely be understood as variants of a MN *Ἀλέξιτος, which is not directly attested in 1 st millennium Greek, but surely underlies Mycenaean a-re-ki-si-to / Aleksito-/ (KN, PY), gen. a-re-ki-si-to-jo (KN): PY Vn(1) 865.1 na-u-do-mo - (2/ x; H 48) - … - - - - .5 ka-ro-qo 1 a-re-ki-si-to 1 The tablet, a list of na-u-do-mo / nāudomoi/ ‘ship-builders’ (line .1) includes a series of man’s names. Most of them are rather obscure, but at least two of them are surely common in Greek: ka-ro-qo (also in line .5) is transparently Greek, i.e. / K h arōk w s/ or / K h arok w os/ (: χαροπός ‘with greedy eyes’ Hom.+, χάροψ Opp.), and has an IE etymology (*ĝ h er(H)- ‘to enjoy, desire’, *h 3 ek u̯ - ‘to look’); u-re-u .9 / h Ul(l)ēus/ (: Ὑλ(λ)εύς), formed from the Boeotian place-name Ὕλη (Hom.), Ὕλαι (Str.+), is 142 José Luis García Ramón <?page no="143"?> 14 Actually *su̯ól-h 2 -/ *sul-éh 2 -, IE *su̯el- ‘to smoulder’, cf Vine 1999: 573, and the discussion by Beekes, EDG s.v., with reference to an alternative derivation from *h 2 us-l-éh 2 beside *h 2 éu̯s-l-h 2 - (ON usli ‘glowing ashes’, IE*h 2 eu̯s-, Gk. εὕο/ ε-, Lat. ūro/ e- ‘to burn’) which goes back to Jakob Wackernagel. 15 The transcriptions of i-qi-ja, a-na-ta and a-na-mo-to imply the existence of psilosis in Mycenaean Cnossos as against etymological */ ( h )ikwijā-/ (cf. ἱππίᾱ ‘equine’, also e-pi-qo-i for */ ep- (h) ikwo-/ in Thebes), */ anh aitai/ (°*sai̯-), */ an-( h )armostoi/ (cf. Hom. ἅρμα). 16 On the possibility of a restoration a]re-ki-si<-to> in KN X 7712.b ]re-ki-si , / a[ cf. Melena 1983: 271-3 (aliter Melena 1976: 426 / A]leksis/ : Ἄλεξις). 17 That the referent of te-]ṃị-ḍẉẹ-ta (pl.neut. / ter]midwenta/ or / te]rmidwentas/ ? ) remains unclear is irrelevant at this point. 18 In the formula [MN GEN - o-pa] the appurtenance relation between its two members is more than evident, and makes the genitive ending -(C)o-jo unnecessary (albeit used sometimes): this may explain the occurrence of endingless -(C)o in this of formula, as well as in that with [ M O N T H N A M EGEN - me-no] and the onomastic formula with [ F A T H E R N A M EGEN - i-jo] and variants.These and other supposed instances of thematic genitive in -(C)o in Mycenaean are in fact a mirage, as I have tried to show (García Ramón 2024, especially 425-6 on the formulas with o-pa]). based on ὕλη ‘wood, forest’, ‘fire wood’ (Hom.+), which may well have a Greek (i.e. Indo-European) etymology 14 as well. The occurrence of two Greek names in the tablet speaks in favour of the possibility of an interpretatio Graeca for a-re-ki-si-to. In Cnossos a MN a-re-ki-si-to is well attested in tablets of the series Sf (registers of chariots) and So (registers of wheels): KN Sf(2) 4420.a - a-re-ki-si-to-jo , o-pa (L; 129? ) - .b ]i-qi-ja / a-na-ta , a-na-mo-to , C A P S 80 The mention of gen. a-re-ki-si-to-jo after o-pa / h opā-/ ‘work (to be performed)’ (Melena 1983, esp. 272-3) reflects the standard formula [ NAM EGEN - o-pa]: the ‘work to be performed by a-re-ki-si-to’ refers to 80 undercarriages of cha‐ riots (/ ikwiai/ or / ikk w iai/ : PGk. * (h) iku̯-i ā-) which are registered as ‘not inlaid’ (/ an-aitai/ probably an error for *a-na-to- / -toi/ ) and ‘not fitted out’ (*/ an-armostoi/ ). 15 A MN a-re-ki-si-to , obviously a cartwright, may be assumed also in two other texts 16 with a less transparent syntax, namely So 1053 and 4433 (L; 131), both with the ideogram R O TA and the indication / termid-went-/ ‘with border(s)’: KN So 1053 .a - ]a, a-re-ki-si-to-jo [ (13; —) - -.b te-]ṃị-ḍẉẹ-ta R O T A [ KN So(2) 4433 .a ] te-mi-dwe-te (L; 131) - - .b ] , a-re-ki-si-to , wo-zo-me-no , R O T A ZE 1 [ The lacunary character of So 1053 lets the possibility of a reading o-p]a after a-re-ki-si-to-jo open. 17 In So(2) 4433.b, a-re-ki-si-to may be an endingless form instead of †a-re-ki-si-to-jo (albeit without mention of o-pa), 18 rather than a From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 143 <?page no="144"?> 19 The possibility of a restoration as pa-ro , a-re-ki-si-to] , wo-zo-me-no as a variant of a-re-ki-si-to-jo , o-pa (dubitanter Lejeune 1968: 25 = MPhM 3.301) is hardly convincing: wo-zo-me-no and o-pa are not necessarily synonyms (Melena 1983: 270, 272), and a syntagm of the type pa-ro , a-re-ki-si-to is not attested as a way of expressing the agent of a passive verb. 20 Among others Ruijgh 1967: 264; Lejeune 1968: 25 (= MPhM 3.301); Leukart 1994: 228 n.251, 2000: 206-7 n.19; Ilievski 1996: 59. 21 Cf. Ruijgh 1983: 396 n.20, who admits the possibility of Ἀλέξι-τος (: Ἀλεξι-τέλης) as well. 22 The fact that Ἀλεξί-τιμος is an “irrational” compound (cf. §3) is obviously irrelevant at this point, 23 Earliest attestations: Φιλιτώ (Athens, iv+; Thasos, iv+). Myc. (gen.) pi-ri-to-jo (ΚΝ C(2) 911.1) could conceal a MN / P h ilito-/ (: Φίλιτος), but a superlative / P h ilisto-/ (: Φίλιστος; also Φiλιστο-κράτης, Thera s. d.) is supported by the WN pi-ri-ta (PY Vn 34.4) which can only be / P h ilistā-/ (: Φιλίστᾱ), not / P h ilitā/ . 24 Earliest attestations: Μενιτος (Crannon, a. 60; also Μενίτας, Atrax, iv; Μενίτης, Athens, v+); WN Μενιτώ (Cos, ca. 200), Μενιττώ (Miletos, a. 269). 25 The names Κάλλιτος (Olympia, i), WN Καλλιτώ (Thera, vi? ) or Χάριτος (Attica, iv), WN Χαριτώ (Attica. ca. 317/ 6) , Χαρίτᾱ (Karpasia, v/ iv) are truncated forms of Kαλλι-τέλης nominative of rubric: 19 whatever the case, a-re-ki-si-to can only be the name of the wheelright who is at work with a pair of wheels ( R O TA ZE 1) with borders (du. / termid-wente/ ) which are ‘being worked out’ (du. / wr̥d z omenō/ , cf. (*u̯r̥ĝ-i̯o/ ecf. Av. vərəziia-, Goth. waurkjan “ἐργάζεσθαι”). Myc. a-re-ki-si-to conceals / Aleksi-to-/ , a truncated form of */ Aleksi-telēs/ (: Ἀλεξι-τέλης), as is commonly assumed. 20 Also / Aleksi-st h o-/ (cf. *Ἀλεξι-σθένης) 21 or / Aleksi-t h o-/ (cf. *Ἀλεξί-θεος? ) could be possible, but none of these forms are actually attested: / Aleksito-/ is the only transliteration supported by (and fitting into) the evidence of 1 st millennium, namely Ἀλεξιτώ (and Ἀλεξιτίδης) and the full name Ἀλεξι-τέλης (iv-iii). It must be stressed that Alph. -ιτος in Ἀλέξ-ιτος* (-ιτώ, -ιτ-ίδης) could still point to a MN Ἀλεξί-τιμος, 22 but this does not apply to Myc. a-re-ki-si-to / Aleksito-/ : the “short” form of °τῑ μος (cf. τῑμή, τίο/ ε- ‘to revere, honour’) would have been noted with °qo-, i.e. †a-re-ki-si-qo, from *°qi-mo / °k w īmo-/ (cf. Μyc. qi-si-ja-ko / k w īsi-ark h o-/ as per Vegas Sansalvador 1999). We can therefore conclude that Ἀλεξιτώ (and Ἀλεξιτίδης) are variants of an *Ἀλέξιτος, which is indeed attested in Mycenaean a-re-ki-si-to / Aleksito-/ , a truncated form of Ἀλεξι-τέλης. An alternative interpretation of a-re-ki-si-to and its alphabetic variants as / Aleks-ito-/ , Ἀλεξ-ιτώ, Ἀλεξ-ιτίδης, i.e. as an onomastic derivative in -ιτος (fem. -ιτώ) turns out to be unnecessary, and in fact wrong: the derivative names in -ιτος (and -ιτώ et sim.) are denominal, formed both from substantives or adjectives (also as members of a compounded name), e.g. Φίλιτος (frequent) and WN Φιλιτώ 23 (cf. φίλος), or Μένιτος and WN Μενιτώ, Μενιττώ 24 (cf. μένος). 25 144 José Luis García Ramón <?page no="145"?> (Attica, iv+), Χαρι-τέλης (Attica, iii), or denominals from κάλλος (in -ιτος), χάρι(τ)- (in -τος, not in -ιτος). 26 For a complete account of the data (also of τελέο/ ε-) cf. Waanders 1983: 21-169. The author tries to trace back all forms to *telh 1 - (not *telh 2 -) ‘to charge’, of which τελέω would be a primary verb, not a denominative of τέλος, and assumes that the existence of a meaning ‘end’ “finds little support in the data” (pp. 235-6). This is surely wrong, but the dossiers and discussion remain very helpful. 27 Gk. τέλος (a) ‘end, fin(al)’, and ‘fulfillment, result’ may well be traced back to *k u̯ él(h 1 )-es- “completion of a circle (cf. *k u̯ ek u̯ (h 1 )lo- ‘wheel’: Ved. cakrá- ‘circle’), completion”, whence ‘result’ (Watkins, AHDΙΕR, s.v. k w el(ə)- ‘to turn’): Hom. πέλομαι ‘to be’ (lexicalized, but περιπλομένων ἐνιαυτῶν ‘during the circling of the years’ (Od. 1.116) beside περιτελλομένων ἐνιαυτῶν ‘id.’ (Il. 2.551): Ved. cár-ati ‘go forward’, also ‘commit, complete’, fin(al)’; Ved. caramá- ‘last’, Welsh pellaf ‘id.’). Cf. also Oscan pupluna- “cyclōna” (deiv]ại pupluna[i] Sa 61 Mefete, near Aquinum), which lives on in Lat. Populōna, epithet of Iuno in Campania (García Ramón 2016: 358-62), Umbr. pupřiko- “κυκλικός” (Prosdocimi 1996: 543), epithet of Puemunthe partner of Vesuna- ‘Lady of the Year’ (*u̯et-s-ōnā- Weiss 2010: 236-44 ‘who goes in a (sc. yearly) circle’), the Umbrian Iuno (García Ramón 2016: 356-8). Whether τῆλε ‘far’, Tηλε° (Hom. +): Lesb. πήλυι ‘id.’ (Thess.Boeot. Πειλε°): PGk. *k u̯ ēle reflect locative *k u̯ ḗlh 1 -Ø “am Wendepunkt, am Ende” of the same lexeme (Harðarson 1995) may remain open. The alternative interpretation of τέλος (a) as the outcome of *k u̯ él(h 1 )-es- ‘accomplishment’ (*k u̯ él(h 1 )- ‘to carry out, perform, execute’ (Waanders 1983: 2-20; Vine 1998: 59-60; Leukart 2000: 203-4) is phonetically possible. 28 Cf. Waanders 1983: 234; LSJ s.v. “something done or ordered to be done, task, service, duty”. 29 It may remain open at this point whether Lat. tolerō, -āre ‘to endure, bear’ is a denominative *teles-āof *tél(h 2 )-esor belongs to a class of presents in -sā- (of the type errāre, vexāre, and others), with -sthat reflects the primary desiderative/ inchoative As to Λάμπιτος (and WN Λαμπιτώ) beside Λάμπετος cf. §10.2 à propos Myc. a-ra-ki-to. §5. How to understand is °τέλης in the MN Ἀλεξι-τέλης is a difficult question, which can be answered on the basis of the different meanings of τέλος (and its collocations) 26 more easily than starting from the etymology of the (at least two) terms which have formally merged in τέλος. For our purposes, two basic meanings of τέλος, with their respective specializations (all attested since Homer) may be retained as worth considering, namely, (a) ‘end, fin(al)’, ‘result’, and ‘goal’, also ‘fulfillment, performance’ (cf. τελέο/ ε-, τελείο/ ε- ‘to fulfill, achieve’: IE *k u̯ el(h 1 )better than IE *tel(h 1 )-, i.e. τέλος : *k u̯ él(h 1 )-esbetter than *tél(h 1 )-es-, 27 (b) ‘duty, charge, task’, 28 cf. Myc. te-re-ta / teles-tā-/ ‘official’ (: τελεστής, El. τελεστα), te-re-ja-e / teleiāh en/ ‘to perform a task’ (*telesā-), cf. Hom. τλῆναι ‘to take on oneself ’, ‘to bear, endure’, ἀνατέλλο/ ε- ‘to make to raise up’ (IE *telh 2 - ‘to lift, support, take on oneself, bear’, whence τέλος: *tél(h 2 )-es-). 29 The term From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 145 <?page no="146"?> *-(h 1 )s- (Nussbaum 2021: 16 with n. 46, also 2007 handout). As Alan J. Nussbaum points out to me (per litteras), the interpretation of tolerāre as denominative of *télh 2 -es- (not attested in Latin, this being exceptional, cf. genus beside generāre, onus beside ōnerāre) seems to be undermined by the contrast between the “active” reading of tolerāre ‘to bear’ and “passive” τέλος ‘that which is borne’. 30 Cf. also Hes.Th. 638 ἶσον δὲ τέλος τέτατο πτολέμοιο ‘the outcome of the war was evenly balanced’ (cf. schol.vet. ad loc. οὐ γὰρ ἡ νίκη τῶν θείων πρὸς ὁμογενῆ τὰ ῥευστὰ γίνεται, ἀλλ’ ἀνόμοια τὰ μαχόμενα). 31 Cf. also A. Sept. 367 ἐλπίς ἐστι νύκτερον τέλος μολεῖν, Pl. Leg. 717e … τῶν ἤδη τέλος ἐχόντων. underlies the lexicalizations as ‘military post, duty’ , later ‘military unit’ (Hdt., Thuc.+) and as ‘imposition, assignment’, whence ‘doom’ (cf. ἐπι-τέλλο/ ε- ‘to enjoin, impose’, ἐν-τέλλο/ ε- ‘id.’ : ‘to make endure’). On the assumption that the synchronic polysemy of τέλος results from the formal merging of two different lexemes, °τέλης in Ἀλεξι-τέλης may a priori reflect two values: on the one hand, τέλος πολέμοιο ‘result, outcome of the war’ (5.1) or τέλος ‘death’ (i.e. βίου τέλος ‘end of life‘, 5.2), which imply τέλος (a) ‘end’; on the other hand, τέλος ‘military unit’ (5.3) or the phraseme τέλος θανάτοιο / θανάτου ‘assignment of death’, i.e. ‘fatal doom’ (5.4), which are both specializations of τέλος (b) ‘duty’, ‘imposition’. 5.1. On τέλος as reflex of [τέλος - πολέμοιο/ πολέμου] ‘end of war’, i.e. ‘result/ outcome of the war’ (Hom.+), cf. Il. 3.291 αὖθι μένων, ἧός κε τέλος πολέμοιο κιχείω ‘here remaining until I reach an end of war’, 20.101 … εἰ δὲ θεός περ / ἶσον τείνειεν πολέμου τέλος, οὔ κε μάλα ῥέα / νικήσει(ε) … ‘if a god should stretch the outcome of war with even hand, then not lightly would he vanquish, …’. The collocation was actually felt by ancient scholars to mean ‘the victory’, e.g. Il. 16. 630 ἐν γὰρ χερσὶ τέλος πολέμου, ἐπέων δ’ ἐνὶ βουλῇ· … ‘for in our hands is the outcome of war, that of words is in the council’ (cf. Eust., Comm., p. 907 τέλος δὲ πολέμου τὴν νίκην φησίν). 30 The interpretation of Ἀλεξι-τέλης as ‘the one who wards off the outcome (of war: [τέλος πολέμοιο]) sc. ‘the victory’ makes hardly sense, and has no phraseological support, as [τέλος πολέμοιο] occurs as the object of verbs with different semantics, but never ‘to ward off ‘. 5.2. For τέλος ‘end’ as metaphor for ‘death’ (Hdt., Attic), and as shortened reflex of [βίου τέλος] ‘end of life’ and variants (Attic-Ionic, cf. D. 57.27 πᾶσίν ἐστιν ἀνθρώποις τέλος τοῦ βίου θάνατος, S. OC 1530 εἰς τέλος τοῦ ζῆν ἀφικνῇ), cf. e.g. Hdt.1.31 οἱ νεηνίαι οὐκέτι ἀνέστησαν, ἀλλ’ ἐν τέλεϊ τούτῳ ἔσχοντο ‘the youths did not yet rise again, but were held by death there’, X.Cyr. 7.3.11 ἔχει τὸ κάλλιστον τέλος ‘he reached the most beautiful end of life’ 31 (beside ibid. 8.7.6 ἐμοὶ μὲν τοῦ βίου τὸ τέλος ἤδη πάρεστιν) and the explicit gloss ἔχει τέλος · τετελεύτηκε [Λάκωνες] (Ηsch.). The interpretation of Ἀλεξι-τέλης as ‘the one 146 José Luis García Ramón <?page no="147"?> 32 Cf. Il. 18.298 νῦν μὲν δόρπον ἕλεσθε κατὰ στρατὸν ἐν τελέεσσι ‘now take your supper throughout the host by companies’ (also Il. 11.730 δόρπον ἔπειθ’ ἑλόμεσθα κατὰ στρατὸν ἐν τελέεσσιν), Hdt. 1.103.1 καὶ πρῶτός τε ἐλόχισε κατὰ τέλεα τοὺς ἐν τῇ Ἀσίῃ, Th. 6.42.2 κατὰ τέλη στρατηγῷ προστεταγμένοι. 33 The “passive” reading of the action noun is recognizable in Engl. ‘command’ (General Eisenhower’s command ‘those commanded by Eisenhower’ [Alan J. Nussbaum, p.c.]), also Span. comando.- 34 On possible instances of τέλος (c) in compounded names with °τέλης, cf. Leukart 2000 (on Ἀριστο-τέλης as “celui qui a la meilleure troupe”. The connection between land tenure and military obligations in the case of the Pylian te-re-ta (Chadwick 1987: 82-3) is not relevant for the military sense of τέλος (c), pace Leukart 2000: 206. who wards off the end’ or ‘… the end of life’ (i.e. ‘death’) is entirely possible. However, the fact that the meaning ‘death’ and the syntagm βίου τέλος are attested first in Classical Greek is not easily compatible with the attestation of MN a-re-ki-si-to / Aleksito-/ (: Ἀλεξι-τέλης) already in Mycenaean. In any event, the interpretation semantically matches (and is complementary with) the alternative explanation of the name which relies on the poetic collocation [τέλος - θανάτοιο] ‘doom of death’ (s. below 5.4). 5.3. For τέλος ‘company, military unit’, cf. Il. 10.470 αἶψα δ’ ἐπὶ Θρῃκῶν ἀνδρῶν τέλος ἷξον ἰόντες ‘and they came swiftly to the unit of Thracian men’ (“zur Schar” Schadewaldt), Th. 2.22.2 … τέλει ἑνὶ τῶν ἱππέων ‘with one squadron of knights’, or the expression ἐν τελέεσσι ‘by companies’ (Hom.), κατὰ τέλεα (Hdt.), Att. κατὰ τέλη ‘in companies’. 32 The original sense was ‘(military) duty’ (Hom. ἐπιτέλλο/ ε- ‘to enjoin, impose? ), whence the lexicaliza‐ tion as ‘men on (military) duty’ and ‘military unit’ (i.e. ‘those charged with a duty-…’): 33 the starting point for the semantic shift is recognizable in Il. 10.55/ 6 αἴ κ’ ἐθέλῃσιν / ἐλθεῖν ἐς φυλάκων ἱερὸν τέλος ἠδ’ ἐπιτεῖλαι ‘if he wishes to approach the sacred duty/ company of the guards, or give orders to them’ (φυλάκων τέλος ‘the duty of the guards’, whence ‘the guard’s team’ (recte Hainsworth 1985: 183). The semantic shift A B S T R A C T → IN DIVID U AL S of τέλος ‘duty’ to ‘military unit’ is of the same type as that of λόχος from ‘ambush’ (Od. 14.217 ὁπότε κρίνοιμι λόχονδε ἄνδρας ἀριστῆας ‘whenever I picked the best men for an ambush’) to ‘men making an ambush’ (Ιl. 8.522 μὴ λόχος εἰσέλθῃσι πόλιν ‘lest the men of an ambush enter the city … ’) and to ‘military unit’ (Οd. 20.49/ 50 εἴ περ πεντήκοντα λόχοι μερόπων ἀνθρώπων / νῶϊ περισταῖεν ‘if fifty troops of μ. men should stand about us’. The interpretation of Ἀλεξι-τέλης as ‘the one who guards stand by the (i.e. his) company’ or ‘his comrades-in-arms’, as already suggested by Leukart (2000: 207 n.19), is quite possible, 34 and would have a striking semantic correspondence in the MN Ἀλεξί-λοχος (Kleitor, ca. 230-200) ‘‘the one who guards stand by From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 147 <?page no="148"?> 35 Cf. the discussion by Waanders 1983: 48-51 (puts τέλος θανάτοιο beside τελέω κῆρα; also informative about semantic analyses from the Ancient scholarship), 232-4. 36 Cf. also Il. 5.818 ἀλλ’ ἔτι σέων μέμνημαι ἐφετμέων ἃς ἐπέτειλας ‘I am still mindful of your charges which you laid on me’, Od. 11.622 ὁ δέ μοι χαλεποὺς ἐπετέλλετ’ ἀέθλους ‘he laid on me hard labors’. the (i.e. his) unit’. The internal syntax of both compounded names reflects the construction of ἀλέξο/ ε- : : ἀλαλκo/ ε- (b) with an animate (dative), as in ἀνδράσιν Ἀργείοισιν ἀλεξέμεναι (Il. 5.779), εἴ πως οἷ … ἀλάλκοιεν (Ιl. 22.196, §2a) as well as that of synonymous ἀμύνο/ ε- (κέκμηκας ἀμύνων σοῖσιν ἔτῃσι ‘you are weary with defending your fellows’ (Il. 6.262). Furthermore, the sense of Ἀλεξι-τέλης (and Ἀλεξί-λοχος ) as ‘the one who guards / stand by his company’ matches that of other names with synonyms of Ἀλεξ(ι)° (like Ἁλευ(ε)°, Ἀμυν(ε)°) and of °τέλης, °λόχος (like °λᾱος, °στρατος, also sensu lato °ανδρος: Ἀλεξί-λεως (Lampsakos, ca. 440-420), Ἀλεξί-λᾱς (Epidauros, iv/ iii); Ἀλεξί-στρατος (Rhodos, s. d.; cf. ἠμύνοντο τὴν Δαρείου στρατιήν Hdt. 3.158.1), as well as Ἀλέξ-αvδρoς and Ἀμύν-ανδρος (§3). The interpretation in terms of °τέλης ‘military unit, company’, which relies on a common phraseme, remains fully satisfactory: any attempt to relativize its plausibility would be nonsensical. This does not, of course, exclude the possibility of an alternative interpretation, which is, moreover, strongly supported by Greek poetic phraseology (cf. §6). 5.4. Τhe phraseme τέλος θανάτοιο/ θανάτου (Hom.+) ‘assignment of death’, or ‘assignment amounting to, consisting in doom’, i.e. ‘fatal doom’ (not neces‐ sarily ‘the end in death’, “das böse Endes des Todes” [Schadewaldt] , albeit syn‐ chronically equivalent, s. above) 35 is, in my opinion, the best candidate for °τέλης in Ἀλεξι-τέλης. The τέλος θανάτοιο may be understood as a command, i.e. a decision that one is assigned (cf. ἐπιτέλλο/ ε- ‘to enjoin, prescribe’, ἐντέλλο/ ε- ‘id.’: *telh 2 -) and must endure (cf. τλῆναι: *telh 2 -). The conceptualization of ‘death’ as a charge (like ἐφετμαί ‘commands’, ἄεθλον ‘labour’) enjoined by someone is evident in the expression θάνατον ἐπιτέλλειν (implying in fact a τέλος θανάτου) in Pi. N. 10.77 καὶ ἐμοὶ θάνατον σὺν τῷδ’ ἐπίτειλον, ἄναξ ‘enjoin death to me along with him, Lord! ’. 36 The occurrence of τέλος θανάτοιο and τελευτήσει κακὸν ἦμαρ in the very same context, referring to the same state of affairs, points in the same direction, cf. Οd. 17.476 Ἀντίνοον πρὸ γάμοιο τέλος θανάτοιο κιχείη ‘may the doom of death come upon Antinoos before his marriage’ beside 15.524 … εἴ κέ σφιν πρὸ γάμοιο τελευτήσει κακὸν ἦμαρ ‘whether before the wedding he (sc. Eurymachos) will bring upon then the day of reckoning’ (“Den Unheilstag vollenden wird! ”, Schadewaldt). The collocation [ A S S I G NM E N T (τέλος ‘doom’) - of D E ATHGEN ], with some formulaic variants, evokes a fatal doom which ‘veils’ or ‘covers’ (καλύπτο/ ε-) or 148 José Luis García Ramón <?page no="149"?> 37 Cf. also Il. 16.502 … τέλος θανάτοιο κάλυψεν / ὀφθαλμοὺς ῥῖνάς θ’(ε) (= 16.855, 22.361). 38 Cf. also Mimn. fr. 2.6 Κῆρες δὲ παρεστήκασι … ἡ μὲν ἔχουσα τέλος γήραος ἀργαλέου, ἡ δ’ ἑτέρη θανάτοιο ‘but the dark spirit of doom stand beside us, one holding generous old age as the outcome, the other death’. 39 Cf. also Il. 9.416 οὐδέ κέ μ’ ὦκα τέλος θανάτοιο κιχείη ‘and the doom of death would not come on me’. 40 Cf. also Ηes. fr. 25.2 δ[εξ]αμένῳ δέ ο[ἱ αἶψα τέλος θανάτοι]ο παρέστη, A. Sept. 903/ 5 (lyr.): … κτέανα †τ’ ἐπιγόνοις + / δι’ ὧν αἰνομόροις, / δι’ ὧν νεῖκος ἔβα / καὶ θανάτου τέλος ‘… that property … over which a dreadful fate came to them, over which came strife and doom of death’. 41 Αlso λοιγός ‘misfortune, calamity’ belong here, cf. Ιl. 21.138 Τρώεσσι δὲ λοιγὸν ἀλάλκοι, 21.539 Τρώων ἵνα λοιγὸν ἀλάλκοι), and, with ἀμύνο/ ε-, Il. 1.341… ἀεικέα λοιγὸν ἀμῦναι, 1.456 Δαναοῖσιν λοιγὸν ἄμυνον, 1.456 Δαναοῖσιν λοιγὸν ἄμυνον, 5.603 … εἷς γε θεῶν, ὃς λοιγὸν ἀμύνει. ‘comes upon’ (κίχη-) humans, οr ‘happens’ (ἐτύχθη, παρέστη) οr ‘is ordained’ (πεπρωμένον ἐστίν et al.) to them, or which one tries ‘to ward off ’ (ἀλεείνo/ ε-) or ‘to escape’ (aor. ὑπαλυσσα-, [ἐκ]φυγο/ ε-]) (§6). Τhe phraseme τέλος θανάτοιο is explicitely glossed as periphrasis for ‘dead’ by the Ancient scholars, cf. Il. 5.553 -… ἀρνυμένω· τὼ δ’ αὖθι τέλος θανάτοιο κάλυψεν ‘seeking both to win (sc. respect for the Atrides) but both of them there did the doom of death veil’, 37 cf. Eust. (v. 553) Ὅτι τὸ “τοὺς δὲ τέλος θανάτοιο κάλυψε” περίφρασίς ἐστι τοῦ ἔθανον. Importantly, fates (Κῆρες, μοῖρα) are explicitly said to lead to (: assign, enjoin) the fatal doom. cf. Il. 9.411 μήτηρ … μέ φησι διχθαδίας Κῆρας φερέμεν θανάτοιο τέλοσδε ‘… my mother tells me that twofold fates are bearing me towards the doom of death’, 38 Il. 13.602 τὸν δ’ ἄγε μοῖρα κακὴ θανάτοιο τέλοσδε ‘an evil fate was leading him to the doom of death-…’. Other instances: Il. 11.451 φθῆ σε τέλος θανάτοιο κιχήμενον, οὐδ’ ὑπάλυξας ‘the doom of death has been too quick in catching up with you, and you did not escape’, 39 Od. 24.124 ἡμετέρου θανάτοιο κακὸν τέλος, οἷον ἐτύχθη ‘(Ι will tell) our bitter doom of death, and how it happened’), 40 Il. 3.309 Ζεὺς . . οἶδε, ὁπποτέρῳ θανάτοιο τέλος πεπρωμένον ἐστίν. ‘… for which of the two the doom of death is ordained’. §6. The occurrence of τέλος θανάτου and its equivalents (acc. κῆρα ‘fate’, also sensu lato κακὸν ἦμαρ ‘fatal day’ and κακότην ‘evil case’, μόρον ‘death’ and τὸ μόρσιμον [Pi.], νηλεὲς ἦμαρ ‘day one can not escape’, ὄλεθρον ‘death’ …) 41 as the object of ‘warding off ’ (ἀλέξο/ ε- : : ἀλαλκο/ ε-, and its synonyms ἀλεείνο/ ε-, ἀλεύασθαι , ἀμύνo/ ε-, also ἀρκέο/ ε-), as well as, significantly, of aor. φυγο/ ε- ‘to escape, avoid’ (not durative ‘to flee’: Hom. ἔφυγε, which overlaps with aor. From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 149 <?page no="150"?> 42 Cf. e.g. Il. 13.436 οὔτε γὰρ ἐξοπίσω φυγέειν δύνατ’ οὔτ’ ἀλέασθαι ‘so that he could neither run backward nor evade him’ (sc. Poseidon). 43 Also ἀλέξο/ εwith gen. κακότητος, cf. Od. 17.364 ἀλλ’ οὐδ’ ὧς τιν’ ἔμελλ’ ἀπαλεξήσειν κακότητος ‘not even so would she save one of them from ruin’. 44 Compounds with the second member ἀρή ‘ruin’ (cf. Ἄρης) or κακός, i.e. something [ E V I L ] one wards off, may fill the position of (and refer to) D E A T H , e.g. ἀλεξι-άρη ‘she that guards from death and ruin’ (Hes. Op. 464 νειὸς ἀλεξιάρη ‘fallow land is an averter of ruin’, cf. ἀλεξιάρης · ἐπίκουρος, ἀλεξίκακος Hsch. ad loc., also Nic.Τher. 861), ἀλεξί-κακος ‘keeping off ill or mischief ’ (sc. μῆτις Il. 10.20). 45 Also Il. 11.360, 14.462. Aor. ἠλεύατο is interpreted by the scholiasts and glossographers as ἀπέφυγεν and as ἐξέκλινε (with an explicit indication of the type of movement), cf. ἠλεύατο · ἐξέκλινεν, ἀπέφυγεν, also ἀλευάμενος · ἐκκλίνας. φυλαξάμενος (Hsch.). ἠλεύατο) 42 yield strong support to the existence of a complexe phraseme [ WA R D O F F - A S S I G NM E NT (: τέλος ‘doom’) of D E ATH ]. The evidence in Homer and in lyric poetry may be set forth, in a rather schematic way, as follows. (1) τέλος θανάτου as the object of synonyms ἀλεείνο/ εand of aor. ἐκφυγo/ ε-: - - Οd. 5.326 ἐν μέσσῃ δὲ καθῖζε τέλος θανάτου ἀλεείνων - - ‘… and sat down in the middle, seeking to escape the doom of death’ Cf. also Archil. fr. 6.3 αὐτὸς δ’ ἐξέφυγον θανάτου τέλος· ἀσπὶς ἐκείνη / ἐρρέτω (cf. Plut. Apophthegmata Laconica 239b) ‘but I escaped from the doom of death; let this shield go forth! ’, CThgn. 767/ 8 …· τηλοῦ δὲ κακὰς ἀπὸ κῆρας ἀμῦναι / γῆράς τ’ οὐλόμενον καὶ θανάτοιο τέλος ‘… and keep the malevolent spirits, accursed old age, and doom of death far away! ’. (2) ἀλέξο/ ε- : : ἀλαλκεῖν with κακὸν ἦμαρ as object: 43 - Od. 10.288 ἔρχευ, ὅ κέν τοι κρατὸς ἀλάλκῃσιν κακὸν ἦμαρ. - ‘go-…, this shall ward off from your head the evil day’. Cf. also ἀλεξί-μορος ‘warding off death’ (τρισσοὶ ἀλεξίμορος, οf Athena, Artemis, Phoibos): S. OT. 159-164 ἄμβροτ’ Ἀθάνα, / γαιάοχόν τ’ ἀδελφεὰν / Ἄρτεμιν, ἃ κυκλόεντ’ ἀγορᾶς θρόνον / εὐκλέα θάσσει, / καὶ Φοῖβον ἑκαβόλον, ἰώ, / τρισσοὶ ἀλεξίμοροι προφάνητέ μοι). 44 (3) the collocation [ἀλέξο/ ε- - τέλος θανάτοιο] is expressed by means of synonyms of its two components: With ἀλεείνο/ ε-: / κῆρ’ ἀλεείνων # (formula Il. 3.32 et al.), κῆρ’ ἀλεείνοντες (Hes. fr. 96.83). With aor. ἀλευα- (cf. ἀλεύετο· ἐξέφευγεν Suda): Od. 12.157 ἤ κεν ἀλευάμενοι θάνατον καὶ κῆρα φύγοιμεν ‘(either die,) or turn aside from death and escape destruction’, also the formulaic segment / ἀλεύατο κῆρα μέλαιναν # after the trochaic caesura (Il. 3.360 … ἔγχος· ὃ δ’ ἐκλίνθη καὶ ἀλεύατο κῆρα μέλαιναν ‘… the spearhead, yet he (sc. Alexander) bent away to one side and avoided black death’ (=Il. 7.253-4), 45 Il. 20.302 μόριμον δέ οἵ ἐστ’ ἀλέασθαι. 150 José Luis García Ramón <?page no="151"?> 46 Cf. the variant with gen. παιδóς instead of dat. οἱ, Il. 15.534 ὅς οἱ καὶ τότε παιδὸς ἀπὸ χροὸς ἤρκεσ’ ὄλεθρον. 47 Riccardo Ginevra points out to me the parallelism between τέλος θανάτοιο and ON ban(a)-orð ‘the killer’s word’, i.e. ‘the death’s word’ in the phraseme banaorð bera af X ‘to bring the killer’s word from X’ (i.e. the ‘death message’) for ‘to kill’ (cf. Watkins 1995: 421-2 with a selection of texts). 48 Instead of unmetrical †κακότητα βριθεῖαν #. With ἀμύνo/ ε-: Od. 8.525 ἄστεϊ καὶ τεκέεσσιν ἀμύνων νηλεὲς ἦμαρ (and Il.13.514 ἀμύνετο νηλεὲς ἦμαρ); CThgn. 767/ 8 …· τηλοῦ δὲ κακὰς ἀπὸ κῆρας ἀμῦναι; A. Ag. 1381 ὡς μήτε φεύγειν μήτ’ ἀμύνεσθαι μόρον. With ἀρκέο/ ε-: Il. 20.289 σάκος τό οἱ ἤρκεσε λυγρὸν ὄλεθρον ‘either on the helmet or on the shield that had warded from him woeful dead’ [= 6.16]; Il. 13.440 ὅς οἱ ἀπὸ χροὸς ἤρκει ὄλεθρον (sc. the coat of bronze). 46 With aor. φυγο/ ε- (cf. Hom. ἐξέφυγον θανάτου τέλος, oben [1]): Il. 11.362 ἐξ αὖ νῦν ἔφυγες θάνατον, κύον (= 20.449), 17.714 θάνατον καὶ κῆρα φύγωμεν, 18.117 φύγε κῆρα et al. (cf. κῆρ’ ἀλεείνων Il. 3.32, ἀλεύατο κῆρα μέλαιναν 3.360); Od. 23.238 ἀσπάσιοι δ’ ἐπέβαν γαίης, κακότητα φυγόντες (cf. ἀπαλεξήσειν κακότητος Od. 17.364); Od. 9.61 … οἱ δ’ ἄλλοι φύγομεν θάνατόν τε μόρον τε ‘but the rest of us escaped death and doom’ (cf. μόριμον … ἀλέασθαι (Il. 20.302; ἀμύνεσθαι μόρον A. Ag.1381). The collocation τέλος θανάτοιο ‘the assignment of death’ (ordered by the fate) for ‘death’ 47 is compatible with the conceptualization of ‘death’ as some‐ thing ‘evil’ and ‘heavy’, which one tries to ward off, as found in Il. 21.548 ὄπως θανάτοιο βαρείας χεῖρας ἀλάλκοι ‘so as to ward off from him the heavy hands of death’. Let us remember the occurrences of terms for ‘death’ in nominal phrases with βαρύς (κακότητα βαρεῖαν Il. 10.71) 48 or ἐμβριθής (A. Pers. 693 τί ἐστὶ Πέρσαις νεοχμὁν ἐμβριθὲς κακόν …; ) or the Mycenaean MN pu 2 -ke-qi-ri / P h uge-g w rīn-/ , which mirrors the collocation [ E S C A P E - H E AV Y ] and designates its bearer as ‘the one who escaped (: ἔφυγε) that what is heavy (: βρῑ, i.e. Hom. βριθύς, or βαρύς, or ὄβριμος)’: H E AV Y is a transferred epithet (in Calvert Watkins’s terminology) of the spear (of the enemy), of one’s evil misfortune (death), of an evil enemy, or of the stone thrown by an enemy (García Ramón 2009). The death is also referred to as ‘heavy’, at least, in Old Norse acc. bana þungan ‘heavy death’ (Riccardo Ginevra, p. c.), cf. Snorri Sturluson, Háttatal 64.5/ 6 … en ræsir Þrœnda / Ribbungum skóp bana þungan “… the ruler of the Þrœndir created a harsh death for the Ribbungar” (Gadde 2017). §7. Interim conclusion: The MN Ἀλεξι-τέλης can be interpreted in two ways, each of which is perfectly compatible with one of the senses of ἀλέξο/ ε- : : ἀλαλκο/ εand of τέλος (*télh 2 -es-). From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 151 <?page no="152"?> 49 Cf. lastly Aravantinos-Godard-Sacconi 2001: 293. On the one hand, as ‘the one who guards / stands by (ἀλέξο/ ε- [b]) his company’, i.e. ‘his comrades’, cf. ἀνδράσιν Ἀργείοισιν ἀλεξέμεναι ‘to stand by the men of Argos’ (Il. 5.779), εἴ πως οἷ … ἀλάλκοιεν ‘hoping if they … might somehow defend him’ (Ιl. 22.196), and with synonym ἀμύνο/ ε- (κέκμηκας ἀμύνων σοῖσιν ἔτῃσι ‘you are weary with defending your fellows’ Il. 6.262): Ἀλεξι-τέλης, like Ἀλεξί-λοχος, fits into the semantics of Ἀλεξί-λεως, Ἀλεξί-λᾱς and Ἀλεξί-στρατος, and even of Ἀλέξ-αvδρoς and variants. On the other hand, as ‘the one who wards off (ἀλέξο/ ε- [a]) the assignment (sc. of death)’, i.e. ‘the fatal doom’, with °τέλης as the reflex of the phraseme τέλος θανάτοιο / θανάτου (Hom.+), i.e. something which one has received as a fatal prescription (cf. Pi. N. 10.77 καὶ ἐμοὶ θάνατον-… ἐπίτειλον, ἄναξ ‘enjoin death to me, Lord! ’) and has to endure (cf. τλῆναι: *telh 2 -) as one’s share: τέλος θανάτοιο / θανάτου matches thus semantically ‘the end in death’ of current translations in Homeric “synchrony”, but does not necessarily reflect the sense of τέλος (a) ‘end, fin(al)’ (§5). The expression is indeed constitutive of a complex phraseological collocation [ WA R D O F F - A S S I G NM E N T (: τέλος ‘doom’) - of D E ATH G E N ], for ‘ward off the death’, which is expressed by means of ἀλέξο/ ε- : : ἀλαλκο/ εand τέλος θανάτοιο and their variants and synonyms, namely ἀλεείνο/ ε-, ἀλεύασθαι, ἀμύνo/ ε-, ἀρκέο/ ε-, also aor. φυγο/ ε- ‘to escape’, and acc. κῆρα, κακὸν ἦμαρ, κακότην, μόρον and τὸ μόρσιμον, νηλεὲς ἦμαρ, ὄλεθρον). The interpretation of the MN Ἀλεξι-τέλης as ‘the one who wards off the doom of death’ is strongly supported by the fact that its underlying collocation is richly attested since Homer. §8. Let us finally turn to the obscure MN a-ra-ki-to in the new tablets from Thebes, which has defied any attempt of interpretation: TH Gp 179.1 u-re-we VIN S 1 ro-[ Pelop. str. ( — ) - .2 a-ra-ki-to Ṣ 1 pi-ri[ - A-ra-ki-to is surely a recipient of wine rates (in dative), as is u-re-we .1 (dat. / h Ul(l)ewei/ , cf. nom. u-re-u / h Ul(l)ēus/ PΥ Vn 865.9: Ὑλ(λ)εύς, 49 cf. ὕλη ‘wood, forest’), which may well be interpreted as Greek (§4). As in the case of a-re-ki-si-to, the occurrence beside a Greek name is no decisive argument in favour of the Greek character of a-ra-ki-to, but it still makes it conceivable. If we examine other tablets of the same series, which record rates of wine (VIN: 152 José Luis García Ramón <?page no="153"?> 50 The indications of animals (e-]p̣ẹ-to-i / h erpetoi h i / ‘for animals’, cf. Lesb. ὄρπετον), workers (to-pa-po-ro-i / X-p h oroi h i/ ‘for X-bearers, carriers’ (Gp 181) or place names (in locative or allative, e.g. *63-te-ra-de) are of no help at this point. 51 Cf. García Ramón 2006: 45-50; 2021: 37-9. Less clear are ]ạ-me-ro (dat.) / (h) Āmerōi/ ? (cf. ἡμέρα ‘day’, Att. ἡμέριος ‘living one day, daily’, ‘for one day’ [: MN Ἡμέριος], or “short” form of a compound with ° (h) āmero-, e.g. di-wi-ja-me-ro ‘period of two days’ [TH], Att. ἐφ-ήμερος [ἐπάμερος Pind.], Ἀγαθ-ᾱ́μερος, Εὐ-ήμερος), and qe-da-do-ro (prob. man’s name dat. / °andrōi/ or / °dōrōi/ ) and 5̣6̣-ru-we / -ruwei/ or / -luwei/ . 52 Ἀλαλκομενηίς (fem. *-ēu̯-ídformed from a derivative in *-ēu̯-, namely Ἀλαλκομενεύς (recte Ηerod. 3.1, p. 84 Ἀλαλκομενηΐς, ὅπερ οὐ παρὰ τὸ ἀλαλκεῖν, ὡς Ἀρίσταρχος, ἦν γὰρ Ἀλαλκηΐς, ἀλλ’ ὥσπερ Θησεύς Θησηΐς, οὕτως Ἀλαλκομενεύς Ἀλαλκομενηΐς), which is actually quoted as epithet of Ζeus (EM). The epithet Ἀλαλκομενηίς was misundertood as formed from a compound with °μένης by some ancient scholars, cf. Apoll. 22.7 ἀλαλκομενηΐς· ἡ ἀλάλκουσα τῷ ἑαυτῆς μένει, ὅ ἐστι βοηθοῦσα (also Hsch. ad Ε 908, as well as Ευst. 1. p. 63 … (with mention of some imaginary forms) ἀπὸ τοῦ ἀλαλκεῖν ἤτοι βοηθεῖν διὰ μένος, ἢ ἀπό τινος Ἀλαλκομενέως ἥρωος, ἢ ἀπὸ Ἀλαλκομενίου ὄρους, οἱ Ἐτυμολόγοι φασί. Gp 181, and 184) and flour (FAR: Gp 215), we find further human recipients (in dative) 50 who also bear Greek names: TH Gp(1) 181 .1 e-]p̣ẹ-to-i S 1 (306) - - - .2 ]1̣ - TH Gp(1) 184 .1 Ṿ 3 to-pa-po-ro-i , VIN s 1 (306) - - - .2 ]* 5̣6̣-ru-we v 2 e-pe-to-i v 1 - TH Gp 215 .1 ]ạ-me-ro , qe-da-do-ro FAR V 3 - - - - .2 a-]ko-ro-da-mo FAR v 1 - Surely recognizable as Greek is, at least, the name a-ko-ro-da-mo (dat.) / Akro-dāmōi/ (: Ἀκρόδᾱμος) ‘the top of the community’ (Gp 215). 51 The possi‐ bility that a-ra-ki-to is a Greek name, as mentioned above, turns out thus to be plausible and legitimizes the attempt at an interpretatio Graeca for it. §9. The best candidate for a-ra-ki-to is, to my mind, / Alalkito-/ , namely dat. / Alalkitōi/ , a formation relying on aor. ἀλαλκo/ ε-, i.e. strictly parallel to the MN a-re-ki-si-to / Aleksito-/ (: pres. ἀλέξο/ ε-, §3). A MN / Alalkito-/ , regardless of how it should be accounted for, is a striking novelty, and indeed an unicum, since until now there has been no incontestable evidence for personal names relying on aor. ἀλαλκo/ ε-. It is true that an (otherwise not attested) ptc. *ἀλαλκόμενος seems to underlie the place name Ἀλαλκομεναί in Boeotia (Str. 9.2.36 Ἀλαλκομενῶν τοίνυν μέμνηται ὁ ποιητής, ἀλλ’ οὐκ ἐν καταλόγῳ, Paus.+) and the divine epithet Ἀλαλκομένη (Ἀθηνᾶι Ἀλαλκομένηι Chios, 46.2 Forrest, aet. hellen.), as well as the derivatives Ἀλαλκομενηίς, epithet of Athena (Ἥρη τ’ Ἀργείη καὶ Ἀλαλκομενηῒς Ἀθήνη Il. 4.8 = 5.908) 52 and the Boeotian month name Ἀλαλκομένιος, Ἀλαλκομενεῖος (aet.hellen.; also Ἀλαλκομεναῖος, aet.imp.). This From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 153 <?page no="154"?> 53 The possibility of an old, obsolete middle form might remain open, as pointed out by Alexander Nikolaev (p. c.), but cannot be proved or disproved. 54 On the synonymity with act. ἀλέξοντες cf. Schol. Vet. ad loc. ἀλεξόμενοι] τουτέστιν ἀλέξοντες, ἀλλήλοις προσβοηθοῦντες. The same applies to participles used as names, Ἀλεξομενός (Athens, a. 550+), Ἀλεξαμενός (Teos, v/ iv) and Ἀλέξων (Delphi, a. 337? +), which formally may be a truncated form of Ἀλεξ(ι)°. 55 Cf. Grasberger 1888: 246 (“da sie nicht sich, sondern den Menschen zu liebe schützt”, as per Friedrich G. Welker, with reference to previous proposals). 56 Being formally an intensive, rārakṣāṇáis , as other forms, not easy to identify as such or as a perfect ( Jamison 1983: 41 “disputed grey area”; Schaefer 1994: 27; Kümmel 2000: 412 assumes a perfect “lange Reduplikationssilbe …, die bei der mit Laryngal anlautenden Wurzel ererbt sein kann”). does not mean, however, that Ἀλαλκομεναί was originally a Greek form: aor. ἀλαλκo/ ε- (unlike pres. ἀλέξo/ ε-) is attested exclusively in the active voice, and this strongly suggests that Ἀλαλκομεναί, Ἀλαλκομένη et al. conceal non-Greek proper names which have been secondarily adapted to Greek. Their exact meaning cannot be elucidated in the light of the middle participle of ἀλαλκο/ ε-, which simply does not exist or, at least, is not attested at all in Greek. 53 At any case, the transitive reading ‘to ward off (someone from oneself)’ may be assumed also for the middle participle (cf. Od. 9.57 … τόφρα δ’ ἀλεξόμενοι μένομεν πλέονάς περ ἐόντας ‘so long we held our ground and warded them off, though they were more than we’), 54 not only for active ἀλέξων (P. Pae. 52f.9/ 11 … ἦλθον / ἔταις ἀμαχανίαν ἀ̣[λ]έξων / τεοῖσιν ἐμαῖς τε τιμ̣[α]ῖς ‘I have come to ward off helplessness from your kinsmen and from your own honours’), aor. ἀλαλκών (Ιl. 9.605 πόλεμόν περ ἀλαλκών ‘even warding off the war’). Accordingly, the place name Ἀλαλκομεναί may have been (re)interpreted within Greek as ‘the one who is warded off ’ as well as ‘… who wards off ’ (i.e. “the warder town”, cf. Steph.Byz. Ἀκεσαμεναί, ὡς Ἀλαλκομεναί Κλαζομεναί, πόλις Μακεδονίας), 55 which surely applies to Athena Ἀλαλκομένη. A striking coincidence with Ved. rakṣ (§2), albeit not directly relevant for the interpretation of Ἀλαλκομεναί et al., may be pointed out: the middle participle forms, namely pres. rákṣamānaand rārakṣāṇá- (“intensive”: hapax), have both a transitive meaning ‘who protects, guarding’, and so does act. rákṣant- ‘id.’, cf. RV I 96.6c amr̥tatváṃ rákṣamāṇāsa enam ‘protecting their immortality, protecting him, …’, IV 3.14ab rákṣā ṇo agne táva rákṣaṇebhī rārakṣāṇáḥ ‘guard us, Agni, with your guards, ever guarding …’ (“der du (schon immer) geschützt hast / (immer) schützst” Kümmel 2000: 412), 56 V 44.7c ghraṃsáṃ rákṣantam pári viśváto gáyam ‘… shelter for us, which protects (our) patrimony on all sides’. §10. On the assumption that the MN a-ra-ki-to conceals / Alalkito-/ , three possible interpretations may be envisaged: 154 José Luis García Ramón <?page no="155"?> 57 The meaning would match that of Ved. rakṣitá- ‘guarded’ (with secondary -i-, frequent in Vedic), cf. RV X 85.4ab-āchádvidhānair gupitó bā́rhataiḥ soma rakṣitáḥ ‘protected by those whose regulation is sheltering, guarded by those belonging to heights, o Soma, …’. 58 Cf. Ἁγχέ-μαχος (Hom. ἀγχέ-μαχος ‘fighting in close combat’) instead of regular Ἁγχί-μολος (Hom. ἀγχί-μολος ‘coming close’). (1) as / Alalk-ito-/ , a verbal adjective formed from the aorist stem ἀλαλκο/ ε-, i.e. ‘warded off ’, ‘guarded’. 57 The formation would be semantically in order, but has no formal parallel within Greek, where *-ετος (*/ Αlalketos/ ), not *-ιτος, would be expected. A fluctuation / e/ ~ / i/ is indeed attested in other cases, cf. infra. (2) as / Alalk-ito-/ , an onomastic derivative in / -ito-/ : -ιτος. This possibility is contradicted by the fact that this type of derivatives is usually formed from nouns or adjectives, not from verbal stems, as it is the case with aor. ἀλαλκο/ ε-. It is true that Λάμπ-ιτος (Miletos, a. 360-350), WN Λαμπ-ιτώ (Sparta, a. 411; Samos, iv +) and Λάμπ-ετος (Magnesia, iii/ ii) may be formed from λάμπο/ ε- ‘to shine’, which would make them similar to Myc. a-ra-ki-to / Alalkito-/ , but a formation from °λαμπος, °λάμπᾱς remains most plausible. (3) as / Alalki-to-/ , a truncated form of */ Alalki-telēs/ (: *Ἀλαλκι-τέλης), i.e. */ Alalki-t-/ (from aor. ἀλαλκο/ ε-, with a second member with initial / °t-/ ), exactly like Ἀλεξιτώ, Myc. a-re-ki-si-to / Aleksito-/ from Ἀλεξι-τέλης, its coun‐ terpart formed from pres. ἀλέξο/ ε- (§3). This proposal faces an obvious difficulty: for a thematic stem like ἀλαλκο/ εone expects a first member ἀλαλκ-ε°, not ἀλαλκ-ι°. Τhe difficulty is not unsurmountable, given that a fluctuation / e/ ~ / i/ both for expected / e°/ and for expected / i°/ (e.g. Ἀγι-λάϊος beside regular Ἀγε-λάϊος, and Ἁγχέ-πολις beside regular Ἀγχί-πολις : ἄγχι°) 58 is well attested in onomastic compounds (and in simplicia, cf. Λάμπ-ετος ~ Λάμπ-ιτος above). A more important piece of evidence for the interpretation of a-ra-ki-to may be seen in compounded names in which expected / -e°/ coexists with secondary / -i°/ (most probably due to analogical extension from the types ἀργί-πους, or λυσι-μελής): this is the case of names with regular first member Ἀγε°, γε° (ἡγέο/ ε-), Ἀρχε°, Ἐχε°, Φαινε°, Φειδε°, Χαιρε° that coexist with variants with -ι° (Ἀγι°, γι°, Ἁρχι°, Ἑχι°, Φαινι°, Φειδι°, Χαιρι°) and, in some of them, with -o° (Ἀγο°, Ἁρχο°, Φαινο°, Φειδο°, Χαιρο°). Some instances with second member °δᾱμος, °λᾱος (°λέως) and °λόχος (all taken from Bechtel, HPNG and/ or from LGPN, whence without indication of place and date) speak for themselves: From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 155 <?page no="156"?> 59 Cf. García Ramón 1992, esp. 333, 339-40, with reference to continuity of the semantics by means of synonyms; Hajnal 1997: 128-9; aliter Leukart 1994: 34 n.18 (“fürs Mykenische sehr unwahrscheinlich”). 60 The type Ἀλεξ(ε)° may be attested already in Mycenaean, if [.]ke-re-sa-wo KN As(2) 1516.13 conceals / [Au]kselāwo-/ (Leukart 1994: 227 n. 251), but / [De]kse-lāwo-/ is also possible. Ἀγε/ ι° Ἀγε-λάϊος Ἀγι-λάϊος (also Ἀγó-λᾱος) γε/ ι° Ἡγέ-λοχος Ἁγί-λοχος - Ἀρχε/ ι° Ἀρχέ-δημος Ἀρχί-δημος (also Ἀρχó-δᾱμος, Myc. a-ko-da-mo / Ark h o-dāmo-/ ) Ἐχε/ ι° Ἐχέ-λᾱος Ἐχί-λᾱος - Φαινε/ ι° Φαινέ-λᾱος Φαινί-λᾱς (also Φανó-λᾱς) Φειδε/ ι° Φειδέ-λᾱος Φειδί-λεως (also Φειδó-λᾱς) Χαιρε/ ι° Χαιρέ-δᾱμος Χαιρί-δημος (also Χαιρó-λᾱς) The same fluctuation may be observed in Myc. MN a-ke-wa-ta (and a-ke-wa-to, both PY) beside a-ki-wa-ta (KN), with / °wastās/ (variant of / °wastos/ : *u̯ast(u̯)-o-) ‘inhabitant of the / wastu/ ’, for which it is impossible to determine whether a-ke° / a-ki° conceal / Age°, / Agi°/ , or / Ark h e°/ , / Ark h i°/ or even / h Āge°/ , / h Āgi°/ . 59 A striking case is the coexistence of Ἀλεξε° and Ἀλεξι° with °βιος as the second member, namely Ἀλεξέ-βιος (Delos, iii) and Ἀλεξί-βιος (Megalopolis, iv/ iii; Attica, iv+), which are indeed both regular as to word-formation. Οn the one hand, Ἀλεξέ-βιος (: pres. ἀλέξο/ ε-) 60 fits into the pattern of the type Ἀρχέ-βιος, Ἑλκέ-βιος, Χαιρέ-βιος (: thematic stems ἄρχο/ ε-, ἕλκο/ ε-, χαίρο/ ε-). On the other hand, Ἀλεξί-βιος fits into the τερψίμβροτος-type, like e.g. Δεξί-βιος, Κτησί-βιος. We can therefore interpret a-ra-ki-to / Alalki-to-/ as a truncated form of a compounded name */ Alalki-telēs/ or the like (/ Alalki-T-/ ), formed from the aorist stem ἀλαλκο/ ε-, and assume that / Alalki°/ instead of / Alalk-e°/ reflects a secondary fluctuation as in the case of Myc. a-ke° ~ a-ki° or Alph. Ἀγε° ~ Ἀγι° , Ἀρχε° ~ Ἀρχι° , Ἐχε°~ Ἐχι° et sim. The collocation [ WA R D O F F - A S S I G NM E NT (of D E ATH )], with aor. ἀλαλκο/ ε-, instead of pres. ἀλέξο/ ε- (the standard type) is actually not attested, but its existence is not a condicio sine qua non for the creation of / Aleksi-to-/ , since the appurtenance of the first members of a-ra-ki-to / Alalki-to-/ (: aor. ἀλαλκο/ ε-) and of a-re-ki-si-to / Aleksi-to-/ , fem. Ἀλεξιτώ and Ἀλεξιτέλης (: pres. ἀλέξο/ ε-) to the same verb allows for the assumption of a proportional analogy, namely ἀλέξ-ο/ ε- (Ἀλεξι°) : : / Aleksi-to-/ (*Ἀλέξ-ιτος) ἀλαλκ-o/ e- (*Ἀλαλκε°) : : X , whence X: → / Alalki-to-/ (*Ἀλάλκ-ι-τος) 156 José Luis García Ramón <?page no="157"?> §11. To sum up: 1. Τhe WN Ἀλεξιτώ (Amorgos, ca. 300) and MN a-re-ki-si-to / Aleksi-to-/ (Cnossos), Ἀλεξιτίδης (Νaxos, ca. 600-550) are truncated forms of the compounded MN Ἀλεξι-τέλης (Ceos, iv/ iii), not οf a putative *Ἀλεξί-τιμος, the truncated form of which would have been spelled as *a-re-ki-si-qo in Mycenaean (with °qo / °k w o-/ for *°qi-mo / °k w īmo-/ : °τῑμος). 2. Ἀλεξι-τέλης may be interpreted as either ‘the one who guards / stands by his company’, i.e. ‘his comrades’ (cf. ἀλέξο/ ε- [b], τέλος ‘military unit’) or, most probably, ‘the one who wards off the fatal doom’, as a hypostasis of the phraseological collocation [ WA R D O F F (: ἀλέξο/ ε- [a]) - A S S I G NM E NT ( τέλος ‘doom’ , of D E ATH : τέλος θανάτοιο)]’, which is widely attested in Greek, expressed by synonyms of both constitutive elements, e.g. Hom. τέλος θανάτου ἀλεείνων, or ἀλάλκῃσιν κακὸν ἦμαρ, and others. 3. The ΜΝ a-ra-ki-to (Thebes) should be analyzed as / Alalki-to-/ , formed from aor. ἀλαλκo/ ε- (like Ἀλεξιτώ, a-re-ki-si-to from ἀλέξο/ ε-) and may be either a truncated form of *Ἀλαλκι-τέλης (instead of expected *Ἀλαλκε-τέλης, with a fluctuation of the same type as Myc. a-ke° ~ a-ki° or Alph. Ἀγε° ~ Ἀγι° et al.) or have been created by proportional analogy ἀλέξ-ο/ ε- : Myc. / Aleksi-to-/ (*Ἀλέξι-τος) : : ἀλαλκ-o/ e- : X , whence X: → Myc. / Alalki-to-/ (*Ἀλάλκ-ιτος). Bibliographical references Aravantinos, V. L.; L. Godart; A. Sacconi (2001) Thèbes. Fouilles de la Cadmée I, Pisa - Rome, Istituti editoriali poligrafici internazionali. Bechtel, F. (1917) HPNG = Die historischen Personennamen des Griechischen bis zur Kaiserzeit, Halle, Niemayer. Beekes, R. S. P. (2009) EDG = Etymological Dictionary of Greek, Leiden, Brill. Chadwick, J. (1987) “The Muster of the Pylian Fleet”, in Ilievski, P.H. & L. Crepajac (eds.) Tractata Mycenaea. Eighth International Colloquium (Ohrid 15.-20.9.1985), Skopje, Macedonian Academy of Sciences and Arts, 75-84. Gade, K. E. (2017) ‘Snorri Sturluson, Háttatal 64’, in Gade, K. E. & E. Marold (eds.) Poetry from Treatises on Poetics. Skaldic Poetry of the Scandinavian Middle Ages 3. Turnhout: Brepols. García Ramón, J. L. (1990-1991) “Los antropónimos micénicos a-pi-wa-to y a-ke-wa-to, a-ke-wa-ta, a-ki-wa-ta : / °wastos/ (/ °wastās/ ) ‘habitante del / wastu/ ’”, Minos 25/ 6: 331-341. García Ramón, J. L. (2006) “Zur Onomastik der neuen Texten aus Theben“, in Deger-Jal‐ kotzy, S. & O. Panagl (eds.) Die neuen Linear B-Texte aus Theben: Ihr Aufschlusswert für From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 157 <?page no="158"?> die mykenische Sprache und Kultur (Internationales Kolloquium Wien, 5.-6.12.2002), Wien, Verlag der Österreichischen Akademie der Wissenschaften, 37-52. García Ramón, J. L. (2009) “Mycenaean Onomastics, poetic phraseology and Indo-Europ ean comparison: the man’s name pu 2 -ke-qi-ri”, in Yoshida, K. & B. Vine (eds.) East and West (Papers in Indo-European Studies, Conference Kyoto University, 11.-12.07.2007), Bremen, Hempen, 1-26. García Ramón, J. L. (2016) “In search of Iuno in the Sabellic domain: Umbrian, Marsian Vesuna-, Oscan Pupluna-”, in Ancillotti, A.; A. Calderini; R. Massarelli (eds.) Forme e strutture della religione nell’Italia mediana antica (Terzo Convegno Internazionale Insti‐ tuto di Ricerche e Documentazione sugli antichi Umbri, Perugia-Gubbio 21.-25.9.2011), Roma, “L’Erma” di Bretschneider, 353-365. García Ramón, J. L. (2020) “Onomástica griega, fraseología, lengua poética indoeuropea: para una reelaboración de Fr. Bechtel, Die historischen Personennamen des Griechi‐ schen.” Atti del Sodalizio Glottologico Milanese 15, 2020, 27-59. García Ramón, J. L. (2024) “The supposed thematic genitive in -(C)o in Mycenaean: a Mirage”, in Bennet, J.; A. Karnava; T. Meissner (eds.) Ko-ro-no-we-sa (In memory of John Bennett) 15 th International Colloquium on Mycenaean Studies (Athens, September 2021), 417-432. Grasberger, L. (1888) Studien zu den griechischen Ortsnamen: Mit einem Nachtrag zu den griechischen Stichnamen, Würzburg, Stahel. Griepentrog, W. (1995) Die Wurzelnomina des Germanischen und ihre Vorgeschichte, Innsbruck, Innsbrucker Beiträge zur Sprachwissenschaft. Ηainsworth, B. (1985) The Iliad: A Commentary III: books 9-12, Cambridge, Cambridge University Press. Hajnal, I. (1997) Sprachschichten des Mykenischen Griechisch. Suplemento a Minos nº 14. Salamanca, Servicio de Publicaciones. Harðarson, J. A. (1995) “Griechisch τῆλε”, Historische Sprachwissenschaft 108: 205-206. Haudry, J. (1977) L´emploi des cas en védique. Introduction à l’étude des cas en indo-eu‐ ropéen, Lyon, L’Hermès. Heubeck, A. (1957) “Weitere Bemerkungen zu den griechischen Personennamen auf den Linear B-Tafeln”, Beiträge zur Namenforschung 8: 268-278 (= Kl. Schr., 1974, 481-491). Hill, E. (2002) “Ved. rákṣati ‘schützen, hüten’ und gr. ἀλέξω ‘abwehren’: Versuch der syntaktischen und semantischen Rekonstruktion eines grundsprachlichen Verbs”, Historische Sprachwissenschaft 115: 239-264. Ilievski, P. H. (1996) “The Mycenaean personal names in -to with regard to the derivatives from verbal adjectives”, in de Miro, E.; L. Godard; A. Sacconi (eds.) Atti del II Congresso Internazionale di Micenologia vol. I (Roma & Napoli, 14-20.10.1991), Roma, Edizioni dell’Ateneo, 51-69. 158 José Luis García Ramón <?page no="159"?> Jamison, S. W. (1983) “Two problems in the inflection of the Vedic intensive” Münchener Studien zur Sprachwissenschaft 42: 41-73. Jamison, S. W. & J. P. Brereton (2014) The Rigveda: The earliest religious poetry of India, New York, Oxford University Press. Kümmel, M. J. (2000) Das Perfekt im Indoiranischen, Wiesbaden, Reichert. Lejeune, M. (1968) “Chars et roues à Cnossos: structure d’un inventaire”, Minos 9: 9-61 (MPhM 3. 285-330). Leukart, A. (1994) Die frühgriechischen Nomina auf -tās und -ās. Untersuchungen zu ihrer Herkunft und Ausbreitung (unter Vergleich mit den Nomina auf -eus), Wien, Verlag der Österreichischen Akademie der Wissenschaften. Leukart, A. (2000) “Le nom d’Aristote”, Philokypros (Mélanges de philologie et d’Anti‐ quités grecques et proche-orientales dédiés à la mémoire d’Olivier Masson), Supple‐ mentos a Minos 16: 201-208. Melena, J.L. (1976) “El testimonio del micénico a propósito de los nombres de las distintas fuerzas en Homero”, Emerita 44: 421-436. Melena, J.L. (1983) “Further thoughts on Mycenaean o-pa”, in Heubeck, A. & G. Neumann (eds.) Res Mycenaeae (Akten des VII. Internationalen Mykenologischen Colloquiums in Nürnberg vom 6.-10. April 1981), Göttingen, Vandenhoeck & Rupprecht, 258-286. Nussbaum, A. J. (2021) “Spēs exploration”, in Tarsi, M. (ed.) Studies in general and historical linguistics offered to Jón Axel Harðarson on the occasion of his 65 th Birthday, Insbruck, Innsbrucker Beiträge zur Sprachwissenschaft, 1-27. Nussbaum, A. J. (2007) “Latin present stems in -sā-: A possibly not so minor type”, handout, East and West Indo-European Conference (Kyoto, 11-12. September 2007). Narten, J. (1964) Die sigmatischen Aoriste im Veda, Wiesbaden, Harrassowitz. Oertel, H. (1926) The syntax of cases in the narrative and descriptive prose of the Brāhmaṇas, I, Heidlberg, Winter. Prosdocimi, A. L. (1982) “La triade Gabovia di Gubbio”, Religioni e civiltà: scritti in memoria di Angelo Brelich, I, Bari, Dedalo, 452-478. Ruijgh, C. J . (1967) Études sur la grammaire et le vocabulaire du grec mycénien, Am‐ sterdam, Hakkert. Ruijgh, C. J . (1983) “Observations sur les neutres en -s/ h-”, in Heubeck, A. & G. Neumann (eds.) Res Mycenaeae (Akten des VII. Internationalen Mykenologischen Colloquiums in Nürnberg vom 6.-10. April 1981), Göttingen, Vandenhoeck & Rupprecht, 390-407. Schaefer, Ch. (1994) Das Intensivum im Veda, Göttingen, Vandenhoeck & Ruprecht. Vine, B. (1998) Aeolic ὄρπετον and deverbative *-etóin Greek and Indo-European, Innsbruck, Innsbrucker Beiträge zur Sprachwissenschaft. Vine, B. (1999) “On ‘Cowgill’s law’ in Greek”, in Eichner, H.; H. Ch. Luschützky; V. Sadovski (eds.) Compositiones Indogermanicae. In Memoriam Jochem Schindler, Praha, Enigma, 556-600. From the Files of a new Bechtel, Die historischen Personennamen des Griechischen 159 <?page no="160"?> Waanders F. M. J. (1983) The history of τέλος and τελέω in Ancient Greek, Amsterdam, Hakkert. Watkins, C. (1995) How to kill a dragon in Indo-European, New York & Oxford, Oxford University Press. Watkins, C. (2011) 3 AHDIER: The American Heritage Dictionary of Indo-European Roots, Boston & New York, Houghton Mifflin Harcourt. Weiss, M. (2010) Language and ritual in Sabellic Italy, Leiden, Brill. 160 José Luis García Ramón <?page no="161"?> 1 This paper is a result of the project PALaC, that has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement n° 757299). Valerio Pisaniello wrote sections 2.1, 2.1.1., 2.1.2, and 3.1; Stella Merlin wrote sections 2.2 and 3.2. Both authors are responsible for sections 1 and 4. Ancient Greek and late Anatolian languages: from coexistence on the territory to survival in the scholarly tradition 1 Valerio Pisaniello & Stella Merlin (University of Chieti-Pescara & University of Naples) Abstract: During the 1 st millennium BCE, Greek language spread throughout Asia Minor, and its diffusion, which is reflected in a large number of local inscriptions found in different Anatolian regions, inevi‐ tably led to interactions, at different levels, between Greek and the epichoric languages of Anatolia. Such language contact resulted in several interference phenomena, both unidirectional and bidirectional, that can be observed not only in multilingual inscriptions, where they are most expected, but also in monolingual ones. Reflexes of such interactions are also found in Greek literary works, as well as in later scholarly sources, especially lexicographic ones of Byzantine times, among which a prominent role is played by Stephanus of Byzantium and Hesychius. This kind of evidence, traditionally referred to as glosses, is of special importance for the knowledge of the Anatolian languages themselves, because most of time they are languages of fragmentary attestation. Through a philological and linguistic analysis of such materials and a discussion on selected cases of language interference, this chapter will provide an overview of the role of the Greek language during the 1 st millennium BCE in two particular regions of Anatolia, namely Lycia and <?page no="162"?> 2 An updated collection and edition of all of the texts was published by Beckman, Bryce, and Cline (2011). 3 See already Forrer 1924: 9. Also note that the place name Akhaiwijais attested in the tablet C 914 from Knossos (a-ka-wi-ja-de). Note, however, that Egetmeyer (2022) recently suggested that Ahhiya(wa) was not related to Ἀχαιοί, but should rather be identified with the Chios island. 4 See, e.g., Hawkins 1998, Niemeier 1998, Kelder 2004-2005, Steiner 2007, Fischer 2010, Kelder 2012. For a recent overview of the whole issue, see Giusfredi (2021-2022), with further references. 5 While some inscriptions in Linear A were found at Miletus (see Niemeier 2005), as far as we know, no evidence of Linear B script comes from any site of Asia Minor. 6 Cf. Wachter 2001: 210-223. Lydia, its interactions with the two Anatolian languages, together with the metalinguistic information retrieved from late Greek sources. Keywords: Lycian, Lydian, Greek language in Anatolia, language contact, glosses. 1. Introduction It is difficult to precisely date the Greek presence in Asia Minor. As is well-known, a number of Hittite texts of the 2 nd millennium BCE mention the Ahhiya(wa)-people in relation to western Anatolia. 2 However, while there can be little doubt that such a designation corresponded to the Greek Ἀχαιοί, 3 it is highly debated both to what kind of political entity it referred - whether a kingdom, a federation, or something else - and where it was actually located, i.e., whether Ahhiya(wa) referred to the Mycenaean kingdoms of the Greek mainland or to peripheral centres located on the coastal areas of Asia Minor and the Aegean islands. 4 In any case, no documents in a Greek language are currently known from the 2 nd millennium BCE Anatolia: 5 if we except sporadic Greek personal names in Hittite sources, the history of the Greek language in Asia Minor only begins in the 1 st millennium BCE for us, as a result of the Greek colonization - traditionally dated to the 11 th century BCE - and after the development of the Greek alphabet between the 9 th and 8 th centuries BCE. The earliest Greek documents from Asia Minor are vase inscriptions dated to the early and mid-7 th century BCE. 6 During the 1 st millennium BCE, Anatolia was populated by several local states and kingdoms with which Greek peoples came into contact. In each community, a number of epichoric and foreign languages were employed and interacted with each other, and Ancient Greek became part of this complex picture. Memories and echoes of such interactions are also found in the Greek literature and 162 Valerio Pisaniello & Stella Merlin <?page no="163"?> 7 On the position of Lycian within the Anatolian branch, see, e.g., Melchert 2003 and Rieken 2017. On Luwian varieties, see Mouton and Yakubovich 2021. 8 Lycian inscriptions on stone and pottery were first published by Kalinka (1901; siglum: TL), and later republished by Friedrich (1932: 52-90), who also included the coin legends. Those discovered after Kalinka’s publication have been collected by Neumann (1979; siglum: N = nova, up to N 323), and the whole corpus has been re-edited with some corrections by Melchert (2001a). An updated edition of the whole corpus, also adding some newly discovered texts, can be found in Réveilhac’s 2018 PhD dissertation (note that texts published after Neumann 1979 keep the siglum N). Finally, an edition of all the text not included in Melchert 2001a has been provided by Christiansen (2019). An online edition of the Lycian corpus is also available on the website of the eDiAna project (ht tps: / / www.ediana.gwi.uni-muenchen.de/ index.php). Lycian coins have been published by Mørkholm and Neumann (1978), Vismara (1989a, 1989b, 1996), and Müseler (2016). 9 There are two graffiti on ceramics from Rhodes predating the 5 th century BCE: N 300b, perhaps of the second half of the 7 th century BCE, and N 300a, ca. 580-550 BCE. However, they only include a few signs, and the Lycian status of N 300b is far from assured (see Neumann 1979: 10-11). So far, the earliest unquestionably Lycian inscriptions are found on pottery sherds dated to ca. 500 BCE: N 313a, from Xanthos, only includes the personal name Pinike, probably explained as the Lycian rendering of the Greek name Ἐπίνικος (see Neumann 1979: 30), and a newly discovered pithos fragment from the vicinity of Patara only contains the title [pd]ḍẽnehm̃mi ‘ruler’ (see Dündar and Réveilhac 2021). On the other hand, the earliest coin legends in Lycian date to the reign of Kuprlli (ca. 480-440 BCE); see Bryce 1986: 51. As for stone inscriptions, they are very difficult to date, except in the not so frequent cases in which an explicit reference to a Lycian ruler is found. Generally, paleographic criteria are followed to date the inscriptions. persist over time until the Byzantine period, many centuries after the Anatolian communities abandoned the use of the epichoric languages. In this contribution, we will limit our research to two major areas, Lycia and Lydia, which, due to the amount of documentation, allow for a thorough analysis of the complex interactions between Ancient Greek and the local Anatolian languages, both on the Anatolian territory and in the Greek literary and scholarly traditions. 2. Ancient Greek and Lycian 2.1. Ancient Greek in Lycia and its interactions with the Lycian language Lycian was an Indo-European language of the Anatolian branch belonging to the Luwic subgroup together with the different varieties of Luwian and the Carian language. 7 The Lycian language is attested by ca. 200 inscriptions roughly covering the whole Lycian territory, the great majority of which are engraved on stone, a limited number on pottery, and a couple on silver containers. To this corpus, one should add several coin legends. 8 Most of documents are dated between the 5 th and 4 th centuries BCE. 9 While ceramic inscriptions and Ancient Greek and late Anatolian languages 163 <?page no="164"?> Linguistic features and the occurrence of some formulae can also provide an indication, as well as architectonic features of the monuments on which the inscriptions are found. However, no criterion is unproblematic (on the whole issue, see Rix 2016: 65-117 and Christiansen 2021, with references). So far, the latest Lycian inscription seems to be TL 35, perhaps to be dated to the reign of Ptolemy II (see Rix 2016: 83). 10 See, e.g., Kalinka 1901: 5; Houwink ten Cate 1961: 3; Bryce 1986: 54-60. 11 See Quadrio 2012 for the history of this glottonym and, more generally, on the relationship between Milyan and Lycian. coin legends include very short texts, mostly consisting of isolated names and abbreviations, stone inscriptions provide longer and more complex texts, which have been crucial to the understanding of the language. All of the texts are written in an epichoric alphabet related to the Greek one, but the exact nature of such a relationship is not entirely clear, although the assumption that it was an adaptation of a Greek alphabet from Rhodes seems to be quite plausible. 10 The great majority of stone inscriptions include funerary inscriptions gener‐ ally providing information about who built the monument, his genealogy, and who was allowed to be buried there. However, a limited number of inscriptions belonging to different typologies exist, including votive texts, dedications, de‐ crees, and a long but unfortunately poorly-understood historical-propagandistic inscription. Based on the typology of documents, it appears clear that Lycian was the official language of the region, employed in both the public and private spheres. The inscriptions attest to two different Lycian varieties: Lycian A, also simply referred to as Lycian, which is found on the vast majority of documents, and a linguistically more conservative variety called Lycian B or, conventionally, Milyan, 11 only attested by two poetic inscriptions. Within the corpus of Lycian documents, some multilingual texts can be found, belonging to different typologies, in which Greek is the language that most frequently occurs alongside Lycian, pointing to the important role that this language already played in Lycia before the Hellenistic period. A complete list of multilingual Lycian-Greek documents found in Lycia can be found in Table 1 below. We decided to also include multilingual documents featuring texts not related to each other, clearly pointing to a later reuse of the monument. Although they are not relevant for the current discussion on the relationship between Lycian and Greek, they may provide some insight into how the linguistic landscape of Lycia may have appeared in later times, when the epichoric language had probably entirely been ousted by Greek both from the public and private monumental inscriptions. 164 Valerio Pisaniello & Stella Merlin <?page no="165"?> Publication Provenance Typology Remarks TL 5 Telmessos funerary later and unrelated Greek inscription TL 6 Karmylessos funerary corresponding texts TL 7 Karmylessos funerary two later and unrelated Greek inscrip‐ tions TL 8 Karmylessos funerary two later and unrelated Greek inscrip‐ tions TL 10 Pinara funerary later and unrelated Greek inscription TL 23 Tlos funerary corresponding texts TL 25a Tlos votive corresponding texts, followed by a sig‐ nature N 343 Tlos dedication (? ) corresponding texts TL 32 Kadyanda funerary corresponding texts (legends identi‐ fying depicted figures) TL 44 Xanthos fu‐ nerary / his‐ torical-prop‐ agandistic a long Lycian (A) text, followed by a celebrative Greek epigram and an unclear Milyan text. TL 45 Xanthos decree corresponding texts TL 50 Xanthos funerary later and unrelated Greek inscription (2 nd c. CE) N 324 / N 325 Xanthos votive Lycian text with related Greek epi‐ gram and short poems N 311 Xanthos votive short Lycian text with related longer Greek epigram N 312 Xanthos votive corresponding texts N 320 Xanthos decree corresponding texts. The stele also fea‐ tures an Aramaic version of the text. N 341 Xanthos funerary later and unrelated Greek inscription (Roman Imperial period) TL 56 Antiphellos funerary corresponding texts TL 65 Isinda decree texts perhaps related, but too fragmen‐ tary to be evaluated TL 70 Kyana funerary related Greek text only including the name of the builder Ancient Greek and late Anatolian languages 165 <?page no="166"?> 12 But see also, among others, Gusmani 2007: 17-19. Publication Provenance Typology Remarks TL 72 Kyana funerary corresponding texts. The sarcophagus also includes a second, unrelated Greek inscription TL 73 Kyana funerary two later and unrelated Greek inscrip‐ tions TL 115 Limyra funerary unrelated Greek inscription (unless it is a signature) TL 117 Limyra funerary corresponding texts TL 121 Limyra funerary later and unrelated Greek inscription (2 nd c. BCE) TL 134 Limyra funerary related Greek text only including the names of the builder and his wife TL 139 Limyra funerary Greek text on a different tomb perhaps corresponding to the last line of the Lycian text TL 143 Limyra funerary related Greek text only including the name of the builder and his genealogy N 302 Korydalla funerary related Greek text only including the name of the builder and his genealogy Tab. 1. A full list of multilingual documents from Lycia, ordered according to their provenance. Documents with related texts are emphasised in bold. In the following sections, we will show the interactions between Ancient Greek and Lycian on the Lycian territory by discussing the relevant documentation - mostly (but not exclusively) including the multilingual inscriptions - arranged, as far as possible, in a chronological order. Indeed, as recently remarked by Réveilhac (2021), 12 the history of the Greek language in pre-Hellenistic and early Hellenistic Lycia can be divided into two stages. Firstly, between the late 5 th and early 4 th centuries BCE, the Greek language is strictly tied to the poetic genre and public/ official use by the Lycian rulers. Then, roughly from mid-4 th century BCE, probably as a consequence of the Hecatomnids’ establishment in Lycia, it detaches from the poetic genre and spread to the private dimension, being increasingly used, e.g., in funerary inscriptions, finally completely replacing the Lycian language during the 3 rd century BCE. 166 Valerio Pisaniello & Stella Merlin <?page no="167"?> 13 See Keen 1998: 66, with references. 14 An exception to this scenario could be represented by a votive inscription found at Phaselis, in Eastern Lycian, also dated before the 4 th century BCE, which contains the dedication of a tithe to Athena Polias by two men (TAM II 1184: Νίκ̣ανδρος ὁ Νικίο̄νος | καὶ Πολυαίνετος | ὁ Πολυκάρτεος | τἀθαναίαι τᾶι Πολιάδι | ἀπὸ ναυτιλίας | δεκάταν ἀνέθε̄καν, ‘Nikandros, son of Nikion, and Polyainetos, son of Polykarteos, dedicated a tithe to Athena Polias from a voyage’). However, this is only an apparent exception, because Phaselis was not a Lycian city but a fully Greek polis, founded by the Rhodian city of Lindos in 691/ 690 BCE and culturally independent from Lycia (See Keen 1998: 233-235; Keen and Hansen 2004: 1140-1141). 15 The name of the dynast is lost in the Lycian text (we only read that he was the son of Harpagus) and only partly readable in the Greek epigram (l. 5/ 24 […]ις). The possible candidates are Xerẽi (cf., e.g., Imbert 1894: 459-461, Childs 1979, Bryce 1986: 97 fn. 126) and, more likely, Xeriga (cf., e.g., Laroche 1974: 145-146, Bousquet 1975, Schürr 2007: 32). In any case, the monument was probably erected by Erbbina (see Réveilhac 2021: 69). 16 Two hexameters (ll. 1-2), two pentameters (ll. 3-4), two elegiac couplets (ll. 5-8), and four hexameters (ll. 9-12). 2.1.1. Ancient Greek and Lycian between the late 5 th and the first half of the 4 th century BCE Contacts between Lycians and Greeks probably occurred from at least the Archaic period, as is shown by the ceramic findings at Xanthos. 13 The earliest Greek inscriptions from Lycia date to the 5 th century BCE and are not only coeval with the Lycian ones, but are mostly in connection with them. As pointed out by Réveilhac (2021), in the earliest phase the use of the Greek language in the Lycian cities was restricted to the public inscriptions of the local dynasts and closely linked to the poetic genre. 14 This is clearly shown by the Greek inscriptions from Xanthos found on monuments dated to the reign of Erbbina/ Arbinas (ca. 395-380 BCE), son of Xeriga/ Gergis, including: 1. The so-called “Xanthos Pillar” (TL 44), originally supporting the tomb of a Lycian dynast, probably Xeriga/ Gergis, on which one can find a long historical-propagandistic inscription in Lycian A celebrating the military exploits of Xeriga, 15 very difficult to interpret (from 44a, 1 to 44c, 19), a celebrative Greek epigram with a very peculiar metrical structure (44c, 20-31), 16 and a rather unclear text in Milyan (44c, 32-44d, 71); 2. The base of a statue dedicated to Leto, with a long Greek hexametric epigram celebrating Arbinas’ exploits on face A, followed by an elegiac couplet with the poet’s signature, and probably five different short poems Ancient Greek and late Anatolian languages 167 <?page no="168"?> 17 See Réveilhac 2021: 75-76. 18 For the Greek texts and a commentary, see Bousquet 1975, Savalli 1988, and Hansen 1989: 281-284 (n. 888), with references. For the Lycian inscriptions, one should refer to the editions mentioned in fn. 8 above. 19 For the Greek text, see Bousquet 1975 and Hansen 1989: 284-285 (n. 889), with references. 20 For a thorough analysis, see Cassio 2019: 17-37. See also Podestà 2017: 70-71; Facella 2018. 21 See Cassio 2019: 24, with references. 22 Cf. Thuc. 2.69. On the relationship between Lycia and Athens, see also Keen 1998 and Thonemann 2009. 23 The Lycian text reads: xbane: ese: trbbẽnimi: tebete: terñ se milas-ñtr-: pddẽneke: xb-nije: izredi ehbijedi: h-tahe. Its interpretation is not entirely clear, but see the tentative translation by Schürr (2010: 152): “Bei Kyaneai nieder Trbbenimi warf das Heer und den Melesandros, vor den eki, den kyaneischen, eigenhändig h-tahe.” in hexameters and elegiac couplets on face B. 17 Faces C and D are occupied by two fragmentary Lycian inscriptions (N 324 and N 325); 18 3. The base of a statue dedicated to Artemis, with a fragmentary Greek epigram in elegiac couplets on the front side, followed by the poet’s signature, and a short Lycian inscription on the left side (N 311). 19 Some observations on the Greek language of these poetic inscriptions are in order. The epigram on the Xanthos Pillar is written in the Attic dialect, and had a specific model, because the first line - [ἐ]ξ οὗ τ’ Εὐρώπην [Ἀ]σίας δίχα πόν[τ]ος ἔνε̄μ[ε]ν̣, ‘Since the sea divided Europe asunder from Asia’ - was a literal quotation of the incipit of an epigram ascribed to Simonides celebrating an Athenian victory against the Persians (FGE 45), transmitted by several literary sources. The rest of the text was also clearly inspired by the same model, which may also justify its peculiar metrical structure. The text also featured some unusual Homerisms that were not found in Simonides’ text but instead reflect an independent knowledge and re-elaboration of the Homeric text. 20 We do not know who the poet was, because he did not leave his signature at the end of the epigram. Based on the amount of Attic material in the epigram, which is even larger than in Simonides’ one, it has been suggested that it was the work of an Attic poet, although one cannot exclude a local poet with a good knowledge of the Attic dialect, 21 nor a Greek poet of different origin, as is the case of the Greek epigram on the base of a statue dedicated to Leto discussed below. Note that the choice of the Attic dialect may appear to be unexpected, especially after the defeat and death of the Athenian general Melesander in Lycia (430/ 429 BCE), 22 an episode that was also mentioned in the Lycian inscription on the same monument (TL 44a, 44-46). 23 If this choice 168 Valerio Pisaniello & Stella Merlin <?page no="169"?> 24 See Réveilhac 2021: 77. 25 Σύμμαχος Εὐμήδεος Πελλανεὺς μάντις ἀ[μύμων] | δῶρον ἔτευξε ἐλεγῆια Ἀρβιναι εὐσυνέ[τω]ς, ‘Symmachos, son of Eumedes, from Pellana, excellent diviner, capably composed the elegy as a gift for Arbinas.’ 26 Note the self-designation of Symmachos as μάντις ἀμύμων. 27 See Bousquet 1975: 146 fn. 20: “Dorisme fortuit, ou influence du datif pluriel courant, qui ne pouvait trouver place dans le vers? ”. Also note the unexpected form κεχαριμμένα for κεχαρισμένα in l. 17, perhaps a mistake, showing the expected outcome of Lesbian and Thessalian (vs. the Laconian outcome with long vowel and simple consonant; see Mitchell 1984: 734-735). 28 For possible Iranian influences on this epigram, see Herrenschmidt 1986. 29 The Lycian text reads eb[e]ija gar-i: zeusi: , but a reading eb[e]ij-agar-i: zeusi: is also possible. In any case, it is highly likely that the dative (a)gar-i Zeusi adapted the Greek did not depend on an Athenian poet, it could be explained as either a mere reflex of the high prestige of the Attic culture in Lycia or as a specific political choice to celebrate Xeriga’s exploits against the Athenians by employing both the language and the stylistic modules with which the now-defeated had celebrated their former victories. 24 The Greek epigram found on face A of the base of the statue dedicated to Leto bears the signature of the poet, an otherwise-unknown Symmachos, son of Eumedes, who was from Pellana, in Laconia. 25 However, despite the Doric background of the poet, the language of the epigram is, again, the Attic dialect, with some Homerisms. 26 The only possible Doricism would be φραδᾶι in l. 8 instead of the expected φραδῇ. 27 Sporadic Doricisms can also be found in the poetic compositions found on face B, for which we do not have poets’ signatures (cf. Λατοῖ in l. 26 and Λατὼ in l. 51 vs. Λητῶι in l. 9 of Symmachos’ epigram). 28 The same observations are valid for the short Greek epigram on the base of a statue dedicated to Artemis, although the text is very fragmentary: The Greek dialect employed is again the Attic one, with sporadic Doricisms (cf. θηροφόνα in l. 2 and παιδοτρίβας in l. 7). The name of the poet is unfortunately broken (Ἐπ[…] in l. 7). In these Greek poetic texts of the first phase, clearly belonging to the public and official dimension, no influence of the local Lycian language can be detected, nor it would really be expected, because native Greek poets were entrusted with their composition, and there were strong and prestigious Greek literary models behind them. One finds some evidence for Greek influence on Lycian, although limited to sporadic loanwords (all technical terms), in some cases - seemingly - from a Doric dialect (Rhodian? ). Besides personal names (e.g., Milas-ñtra- < Μελήσανδρος in TL 44a, 45), consider krzz-nase- ‘peninsula’ < χερσόνᾱσος (TL 44a, 53), trijere- < τριήρης (TL 44b, 22, 23), broken sttrat[…] < στρατηγός (TL 44b, 18), *(a)gara- < ἀγορά (cf. (a)gar-i in TL 44b, 62), 29 and perhaps sttala- ‘stele’ Ancient Greek and late Anatolian languages 169 <?page no="170"?> Ζεύς ἀγοραῖος. Note that the same theonym is also calqued in Lycian as nelezẽ: Trqqñ[t° in N 324, 8 from Xanthos. 30 On these words, see Melchert 2004 and Neumann 2007, with references. 31 Consider the likely existence of a genuine Lycian verb stta-, on which see Jasanoff 2010, with references. However, see also Schürr 2014. 32 See Keen 1998: 148-170. 33 Oreshko’s (2021: 134) reservations and suggestion of an Anatolian etymology seem to be quite unfounded. 34 See Wörrle 1991: 203-217; Keen 1998: 161; Réveilhac 2021: 78. 35 Cf. Lyc. ne=we me=i: kumaza … ‘if the priest(ess) do not … to him’ (ll. 22-23) vs. Gr. ἐὰν δὲ ἡ ἱέρεια μὴ ὑπα[κού]σηι ‘if the priestess will not obey’ (l. 40); see Neumann 2007: 175 s.v. kumaza. 36 See Domingo Gygax 2001: 88-89. < στα᷄λα (TL 44a, 1; 44c, 5, 7), 30 although a genuine Lycian etymology is not excluded. 31 Another Greek poetic inscription was found at Limyra in Eastern Lycia, a city ruled by the king Perikle (370-360 BCE), 32 whose name was, of course, the Lycian adaptation of the Greek name Περικλῆς. 33 Not unexpectedly, the text (SEG 41 1382), 34 a dedication to Zeus only consisting of two hexameters, is written in the Attic dialect with some Homerisms. The altar on which the inscription occurs is fragmentary, so that we do not know whether a Lycian inscription accompanied the Greek text. A possible exception to the exclusively poetic use of the Greek language in this phase could be represented by the bilingual inscription TL 65 from Isinda, which is very difficult to evaluate because of its fragmentary status. Both the Lycian and the Greek texts seem to concern a religious regulation, but only one parallel passage can perhaps be identified between the two versions, 35 so that it is not entirely assured that it actually was a bilingual document with corresponding texts. It is equally possible that the Greek text was a later addition concerning analogous matters to those dealt with in the Lycian one. In the Lycian text, a certain Xezixa is mentioned, which poses a chronological problem. Indeed, it is debated whether he should be identified as the Xeziga mentioned in the Xanthos Pillar (TL 44), who belonged to the Xanthos dynasty and was Kuprlli’s son and Xeriga’s uncle. If this were the case, one could suggest that the inscription should be dated to the late 5 th or early 4 th century BCE. However, this cannot be obvious, because either he could have been a later homonymous (the different spelling could also be significant in this respect) or, if he actually was Kuprlli’s son, the text might simply have included references to past events. 36 Note that an early date for the Lycian text would also be prevented by the occurrence of a late variant of the letter ẽ, as pointed out by Adiego (2014: 254). Furthermore, the Lycian 170 Valerio Pisaniello & Stella Merlin <?page no="171"?> 37 See Rix 2016: 306-310, with references. 38 See also the discussion in Réveilhac 2021: 82-83, with references. 39 See Bousquet 1986. 40 One might wonder whether one of the two small Aramaic fragments found at Xanthos - n. 5627 (fr. A) and n. 5743 (fr. B) - edited by Dupont-Sommer (1979) could actually belong to an Aramaic version of this decree (see Merlin and Pisaniello 2020: 192 fn. 32). 41 See Keen 1998: 10: “Given the Carian nature of the cult and the fact that by this time Lycia was under direct control of the Hekatomnid rulers of Caria, it seems plausible that, though the decree is couched in terms to indicate that the Xanthians had requested the setting up of this cult and Pixodaros, the Carian satrap, had merely approved it, in fact what was happening was that a Carian cult was being forced upon the Lycians, as part of a process of Carianization.” text includes a second personal name, Pttule: if he should be identified with Ptolemy I Soter, or even Ptolemy II Philadelphus, a dating of the inscription to the 3 rd century BCE would be likely. 37 All in all, it is very unlikely that TL 65 could attest to a public and non-poetic use of the Greek language in a very early phase, but the extremely fragmentary status of the document prevents any conclusive judgement. 38 2.1.2. Ancient Greek and Lycian from the mid-4 th century BCE onwards During the 4 th century, the situation of the Greek language in Lycia seems to be changing in two directions: on one hand, Greek detaches from the poetic genre and begins to be employed for other kinds of composition; on the other hand, it widens its domains of use to also include the private sphere, alongside a continuing public use. Such a progressive expansion of the Greek language should probably be connected to the establishment in Lycia of the Carian dynasty of the Hecatomnids, closely related to the Achaemenid Empire. A public use of the Greek language is attested by some documents from Xanthos dated to the reign of the Carian satrap Pixodaros (340-334 BCE), one of Hecatomnus’ sons. The bilingual Lycian-Greek inscription TL 45 is a fragmentary decree including a tax regulation for Xanthos, Tlos, Pinara, and Kadyanda. 39 It is also possible that the document was trilingual in origin, also including an Aramaic version of the decree, 40 as is the case of another decree issued under Pixodaros’ rule (and probably by Pixodaros himself, despite its incipit), 41 the famous Letoon trilingual inscription (N 320), concerning the introduction in Lycia of the Carian cult of the ‘King of Kaunos’ and related issues. The decree, probably dating to the first year of the reign of Artaxerxes IV (337/ 336 BCE), included a Lycian and a Greek version that more or less correspond to each other, and a more concise version in Aramaic - one of the Ancient Greek and late Anatolian languages 171 <?page no="172"?> 42 Cf. Molina Valero 2004. administrative languages of the Achaemenid Empire. Here some information provided by the Lycian and Greek text is lacking (mostly related to matters of local administration, such as the names of the local rulers), while some other relevant details not found in the other versions are added (e.g., the date of the decree and the designation of Pixodaros as ‘satrap of Caria and Lycia’, vs. only ‘satrap of Lycia’ in the Lycian and Greek versions). Therefore, the Aramaic text should be regarded as quite distinct from the Lycian and Greek ones: it probably was the official document from the point of view of the Achaemenid administration, only recording those data that were relevant for the administrative apparatus of the Empire. Matters of local government in Lycia, less important to the Achaemenid administration, were only recorded in the other versions. For this reason, it is also likely that the Aramaic version was more accurate than the other two in some respect. Particularly, the Lycian and the Greek versions seem to concern the establishment of the cult of two different deities, the “King of Kaunos” (Lyc. Xñtawat(i)- Xbidẽñne/ i- = Gr. Βασιλεὺς Καύνιος) and an otherwise unknown deity called ArKKazuma in Lycian and Ἀρκεσιμα in Greek. Conversely, the Aramaic text has a different formulation, kndwṣ ’lh’ kbydšy wknwth, ‘the King, god of Kaunos, and his colleagues,’ which, despite seeming less accurate because it replaces a theonym with a generic expression, was probably more precise. Indeed, in the Anatolian religion, minor deities were commonly organised in circles (the Hittite kaluti-) clustered around a main god. Therefore, it is likely that Pixodaros’ decree concerned the establishment of the cult of the ‘King of Kaunos’ and his circle of deities, rendered, in the Aramaic version, by the expression wknwth ‘and his colleagues’. The puzzling ArKKazuma/ Ἀρκεσιμα could thus be explained as the Lycian and Greek adaptation of a foreign expression meaning the same. Because Pixodaros belonged to a Carian dynasty, the ‘King of Kaunos’ was a Carian deity, and Carian influences can be detected in the Aramaic text, 42 it is likely that ArKKazuma/ Ἀρκεσιμα also reflected an originally-Carian form, perhaps a 172 Valerio Pisaniello & Stella Merlin <?page no="173"?> 43 In his analysis of the Letoon trilingual, Melchert (2018) regarded Carruba’s explanation as problematic. He accepted the hypothesis of a derivative built with the Carian suffix -umabut regarded the Carian toponym Αργασα as a better base. Of course, this could be possible, but it would not fit with the Aramaic expression, which should thus be regarded as a quite inaccurate formulation. An alternative solution - although also not unproblematic - has been recently suggested by Vernet (2023), who explains ArKKazuma as the Lycian adaptation of a Carian name built with *ara/ i- ‘companion’ + an unclear element perhaps comparable with HLuw. hasam(a/ i)- ‘mighty’. As for the Greek adaptation, which displays some mismatches compared to the Lycian one, one might think of a folk-etymological association with the Greek adjective ἀρκέσιμος ‘helpful’ (vel sim.), from ἄρκεσις ‘help’ (see Rasmussen 1974: 69), attested as personal name, e.g., in a brief message on an amphora sherd from the Athenian agora, dated to ca. 475-450 BCE (Agora XXI B 7: Εὐμελὶς ἧκ[ε] | ὁς τάχος. | Ἀρκέσιμος, ‘Eumelis, come as quick as possible! Arkesimos’), in an inscription found at Kamiros (Rhodes), to be dated to ca. 268 BCE (Tit. Cam. 17, l. 12: Ἀρκέσιμος Δάμωνος), and in a long inscription from Eretria of the early 3 rd century BCE (IG XII,9 245, l. 222: Πολυκλῆς Ἀρκεσίμου). The alleged Ἀρκεσίμου in CIG 9899+ from Frikya (northern Syria) is currently read Ἀρκεσιλάου (see, e.g., Jalabert and Mouterde 1955: 120; Hajjar 1990: 2283; Altheeb 2015: 245). 44 See Wörrle 1991: 218-224. 45 See Wörrle 1991: 224-234 for a thorough analysis of the text. Cf. also Keen 1998: 159-160. derivative noun built on the Carian cognate of the Hittite base arḫa-/ irḫa- (cf. Hitt. irḫatt- ‘row, series, circle’), as suggested by Carruba (1999). 43 Besides Pixodaros’ decrees at Xanthos, other public documents in Greek were found at Limyra. Particularly, we could mention a brief honourary decree (SEG 41 1379), probably dated to the second half of the 4 th century BCE, that was issued by the ‘community’ (κοινόν) of the Pernitai to Purimatis, son of Masas (Πυριματις Μασατος), because of his valour and righteousness (καλοκἀγαθίας καὶ δικαιοσύνης ἕνεκεν) towards the city of Pernis, although the text did not include a complete picture of the events that led to the decree. This individual clearly was Purihimeti, son of Masasa, who also was the “author” of a Lycian funerary inscription found in the nearby (TL 99). Apparently, the decree did not have a Lycian version. 44 Another, unfortunately very fragmentary, Greek inscription from Limyra (SEG 41 1380), dated to the late 4 th or early 3 rd century BCE, recorded a letter written to an unknown addressee in response to a former petition from a group of people, perhaps the polis of Limyra. As far as can be read, it dealt with issues that were standing ‘from the time of Perikle until now’ (ἀπὸ Περ[ικ]λέους ἕως τοῦ νῦν, l. 8), perhaps related to religious matters (cf. τόν τε ἱεροκήρυκα καὶ τὸν ἱερέα in l. 4). 45 Again, a Lycian version of the document has currently not been found. This document also mentioned the Pernitai, but the context is extremely Ancient Greek and late Anatolian languages 173 <?page no="174"?> 46 Cf. Keen 1998: 160-161. 47 See Christiansen 2020 for a thorough analysis. 48 Although the private dimension of funerary inscriptions is beyond doubt, one should notice that they sometimes included administrative information, such as the penalties imposed on those found guilty of any violation or the agreements made with the local authority (the Lycian miñti-) and the fees for burial concessions (cf., e.g., Bryce 1986: 71-72). 49 See Adiego 2014: 259-260. 50 For the interpretation of this clause, see Oreshko 2021: 132. unclear, and, in general, Wörrle’s (1991) explanation of the Pernitai as perioeci of the Limyreans involves some problems. 46 The Lycian-Greek bilingual religious regulation of Isinda discussed above (TL 65) might roughly belong to the same period, or even later, but the text is much too fragmentary to be properly evaluated. Finally, the bilingual inscription N 343 from Tlos, probably to be dated to the second half of the 4 th century BCE, could have been an honourary inscription issued by the city of Tlos to an individual named Pteuna, but the text is extremely fragmentary and very little can be read, so it is also possible that it was Pteuna that dedicated a statue to the city of Tlos. 47 Besides the public use, as mentioned, during the 4 th century BCE the Greek language also spread to a more “private” dimension, beginning to be employed in funerary and private votive inscriptions. 48 Among these inscriptions, one finds some Lycian-Greek bilingual texts, most of which include roughly corre‐ sponding texts as far as their content is concerned (besides TL 32, which only includes Lycian and Greek personal names identifying the individuals on the relief, 49 see TL 6, 23, 25a, 56, 72, 117, and N 312). In one case, TL 139 from Limyra, one faces a more complex situation, once the Lycian text is compared to the Greek inscription occurring on a different tomb located a short distance away. Here follow the two texts: (1) 1. ebẽñnẽ: xupa m=ẽ=ti prñna[wat]ẽ tilume: zizah-- 2. mah tideimi hrppi ladi ehbi: xuxuneje 3. se tideime ehbij[e] se uwe: ti hrppi t-ti tike m=ẽne 4. tubidi h[p]pñter[us] m-h-i: se marazija miñtaha: 5. hlm̃midewe: mlejeusi: murñna ‘This tomb, Tilume, son of Zizah-ma, built it for his wife Xuxune and his sons, and if anyone will put somebody in addition (lit. ‘upon’), 50 the hppñterus-deities and the court of the miñti will strike him. For ? Hlm̃midewe, Mlejeusi, Murñna.’ (2) 1. Αρμαπιας ἠργάσατο ἑαυτῶι καὶ τῆι γυναικὶ καὶ τοῖς υἱοῖς καὶ τοῖς οἰκέοις 2. Ελμιδαυαι καὶ Μλααυσει καὶ Μορναι 174 Valerio Pisaniello & Stella Merlin <?page no="175"?> 51 See, e.g., Molina Valero 2007: 137-138. Note that lines 1-4 of the Lycian text are written on the beam above the door, while line 5 is written on the door frame just under the beam, and not aligned to the preceding text. 52 See the discussion in Wörrle 1995: 391-392. ‘Armapias made (it) for himself, (his) wife, (his) sons, and (his) household members, Elmidawa, Mlause, and Morna.’ The last line of the Lycian text includes the names Hlm̃midewe, Mlejeusi, and Murñna, probably in dative case, but with an unclear relationship with the rest of the text (one could think that they were the sons of Tilume, but this is not explicit). The second line of the Greek text includes the same names, seemingly the personal names of the οἰκεῖοι found in line 1. Therefore, it is likely that this Greek text was somehow related to the Lycian inscription TL 139, although they clearly were two different texts, as shown by the fact that the name of the builder mentioned in TL 139 is Tilume, while the Greek text mentions a certain Armapias. It has been suggested that line 5 of the Lycian text was a later addition, pointing to a reuse of the tomb, 51 but different solutions are also possible. 52 In three bilingual inscriptions, the informative content is weighted heavily to‐ wards the Lycian side. Thus, TL 70 from Kyana displays a three-line Lycian text stating that the tomb was built by Sbikaza, son of Xñtanube and nephew of Temusemuta, while the Greek text only reads Σπιγασα. Similarly, TL 134 from Limyra includes a four-line Lycian funerary inscription stating that the tomb was built by Masasi for his wife Xuwata and providing other information, whereas the Greek text written below the Lycian inscription consists only of the two personal names mentioned in the Lycian text, Μασα Κοατα, whose inflection is unclear. According to Kalinka (1901: 87), based on the Lycian text, Μασα should be explained as a nominative and Κοατα as a dative. Conversely, Wörrle (1991: 221, 1995: 397) regarded Μασα as a genitive and Κοατα as a nominative, while Réveilhac (2021: 87-88) understood both names as datives, Μασᾳ Κοατᾳ. In our view, it is also possible to merely regard them as Greek uninflected transcriptions of the Lycian names. It is difficult to evaluate these inscriptions, because we do not have much information about who was responsible for the actual drafting of the funerary inscriptions and, more generally, on the sociolinguistic situation of Lycia in the 4 th century BCE. One could perhaps speculate that these Greek texts, which cut down to the essentials the information contained in the Lycian text - i.e., to personal names, and in a non-adapted form - reflected the will of the clients to add a Greek epigraph to the Lycian funerary inscription, even when they did not have much Ancient Greek and late Anatolian languages 175 <?page no="176"?> 53 It is unclear whether the short Greek inscription Φοίνικος Τυριω - which can be morphologically analyzed in different ways (cf. Wörrle 1991: 221, with references) - following the Lycian inscription TL 115 from Limyra had any relation with the Lycian text (see Wörrle 2012: 419). See also Oreshko (2020: 24-25), who tentatively regarded it as evidence for the existence of a Phoenician community at Limyra. 54 Followed by Neumann (1979: 15). 55 See Réveilhac 2018: 100-102 for a thorough analysis. Note that the situation was different at Phaselis, in eastern Lycia, where a number of Greek inscriptions in the Doric dialect were found (see Réveilhac 2018: 103-109). However, as mentioned, Phaselis was not a Lycian city, but an independent Greek polis. 56 The bases of statue on which the Lycian-Greek bilingual inscription TL 25 (Tlos) is found also bear the inscription Θεόδωρος Ἀθηναῖος ἐπόησε ‘Theodoros, the Athenian, made (it),’ but it probably referred to the statues dedicated by Xssbezẽ, the “author” of the Lycian inscription. 57 The only possible exception is the unique ξατράπης found in the Letoon trilingual inscription (N 320, 1) vs. usual σατράπης in Greek. Although, of course, the noun had an Iranian origin, the Greek form occurring in N 320 could depend on the Lycian adaptation of the Iranian word (the Lycian text has the verb xssaθrapaza- ‘rule as satrap’, from a noun *xssaθrapaza-, morphologically adapted through the suffix -azaof the agent competence in Greek or the commissioned funeral company did not have scribes with the requested skills. 53 TL 143, also from Limyra, is different because the Greek text, although only reading Κοδαρας Οσαιμιος (both with Greek endings), precedes the Lycian one and includes the patronymic not found in the Lycian inscription. Also in N 302 from Korydalla, the Greek text precedes the Lycian one, which is longer (four lines vs. only one in Greek) and more informative. Indeed, only Σαπια Μαναπιμ[ι]ος can be read in the Greek line, but the stele is broken on the right side. Since the inscription is written stoichedon, and, based on the Lycian text, lines of 28 letters should be reconstructed, if this also applied to the Greek text, one could restore [κατεσκευάζατο] after the patronymic, as per Tritsch (1976). 54 The Greek language of these public and private inscriptions is quite different from the language of the poetic texts of the previous stage. First of all, while the latter were written in an Attic dialect with several Homerisms and sporadic Doricisms probably due to the native language of the poets, the later non-poetic inscriptions display the koinè diálektos, basically consisting of an Attic variety with some external influences, mostly from the Ionic dialect. 55 Also, no scribal signatures are found, so we do not have any information about the writers. 56 The crucial difference to the earlier poetic texts lies in the Lycian influences that can be detected in these Greek documents. Besides the Greek transcriptions of Lycian personal and place names (in‐ cluding ethnonyms), no Lycian lexical material is found in the Greek versions occurring on multilingual documents. 57 However, sporadic Lycian loanwords in 176 Valerio Pisaniello & Stella Merlin <?page no="177"?> nouns). However, it is perhaps inappropriate to speak of a proper loanword in this case; it could be an example of phonetic interference. 58 The etymology and meaning of this Lycian term is not entirely clear: a comparison with Hitt. miyant- ‘grown’, (LÚ) mayant- ‘young adult’ (both from mai-/ mi- ‘grow’) has been convincingly suggested. See the discussion in Oreshko 2019a: 105-117, with references. 59 See, e.g., TAM II 611, 17 (Tlos); TAM II 385, 6 (Xanthos); TAM II 870, 6 (Idebessos), all of them dated to the Roman Imperial period. 60 See Schürr 1999 and eDiAna, s.v. *pii̯etEr- (https: / / www.ediana.gwi.uni-muenchen.de/ dictionary.php? lemma=1708). Greek do exist, although they are currently only found in monolingual Greek inscriptions. Thus, in the following funerary inscriptions, the local authority is referred to as μενδῖται / μίνδις, patently a morphologically adapted borrowing from Lycian miñti-, which probably denoted the local community: 58 (3) TAM II 40 (Telmessos; late 4 th century BCE) Μοσχίωνος τοῦ Πεδετέριος Λιμ[υ]ρέως. ταγὴν δὲ ἔταξαν οἱ μενδῖται τοῖς ἀνοίγουσιν τò μνῆμα Ἀλεξανδρείου δραχ<μ>ῶν ἕξ. ‘(Tomb) of Moschion, son of Pedeteris, from Limyra. The menditai fixed as fine for those who open the tomb six Alexandrian drachmas.’ (4) Petersen-Luschan, Reisen II 22,27 (Kyana; 3 rd century BCE) τὸν τάφον τοῦτον κατεσκεύασεν τόν τε ἄνω καὶ τὸν κάτω Περπένηνις […] Ἀππάδιος ἑαυτῶι καὶ τῆι γυναικί· καὶ μη[θ]ενὶ ἐξέστω ἀνοῖξαι τὴν σορὸν οὗ η[…] ἐστὶν, τοῖς δὲ λοιποῖς τάφοις τοῖς τε ἄνω καὶ τοῖς κάτω χρήσονται πάν[τες] οἱ συνγενεῖς· μὴ ἐξέστω δὲ ἀνοίγειν μηθενὶ ἄνευ τῆς μίνδιος, ἀλλὰ συνπαρα[ι]νέτωσαν αὐτοὺς, εἰ δὲ μὴ, κύριοι ἔστωσαν κωλύοντες καὶ ζημιοῦντες αὐτούς. ‘This tomb, the one above and the one below, Perpenenis […] son of Appadis built (it) for himself and (his) wife: no one should be allowed to open the coffin where […] is; all the relatives will use the remaining tombs, both those above and those below: no one should be allowed to open (anything) without the mindis, but they should join in approving them, otherwise, they should have the authority to hinder and punish them.’ However, as far as one can see, the use of this loanword was occasional: in most of Greek inscriptions from Lycia, the local authority is simply referred to as δῆμος or, sometimes, πόλις. Another Lycian term probably entering the Greek language of Lycia and occasionally surfacing in Greek monolingual inscriptions is πιάτρα ‘daughter-in-law’, which seemingly had a longer life in the local variety of Greek. 59 Although the Lycian form is currently not attested, one may reconstruct a derivative noun *pijatrafrom the verb pije- ‘give’ (cf. Luw. piyattara/ i- ‘gift, reward’). 60 Ancient Greek and late Anatolian languages 177 <?page no="178"?> 61 See Neumann 1974: 110-111, not unanimously accepted (see, e.g., Oreshko 2020: 32). 62 See Neumann 1974: 109-110, but Starke (1990: 297-298) suggested ‘burden, tax’ for the Lycian word, followed by Melchert (2004: 3). 63 See Melchert 2004: 6. On xla(i)-, see Sasseville 2021a: 103-104. 64 The preverbs epe and ewẽ perfectly match the Hittite āppa and awan, while the verbal root tl(e)is traced back by Adiego (1993) to PIE *k w el- ‘turn’ (cf. Lat. colō ‘dwell’). 65 A third example, Lyc. ahñta- (N 320, 17), formerly explained as the plural present participle of ah- ‘be’ meaning ‘possessions’, and thus regarded as a calque on Gr. ὄντα, οὐσία (see, e.g., Melchert 2014a: 68, with references), should possibly be ruled out, because the sequence setahñt-i, previously segmented as se=t(e)=ahñt-i, is currently analysed as se=tahñt-i, with tahñta- (= Mil. tasñta-, ultimately < PIE *d h eh 1 ) meaning ‘dwelling for cattle’ and matching Gr. οἰκήματα. See eDiAna, s.v. *tasn̥tā- (https: / / ww w.ediana.gwi.uni-muenchen.de/ dictionary.php? lemma=1904), with references. 66 See Oettinger 1981: 11-12; Gusmani 2007: 19; Melchert 2014a: 68. Conversely, Oreshko (2020: 33) suggested the possibility of a trivial development independent from Greek. 67 Note that kumehe/ ialso occurs in other Lycian inscriptions (TL 84 and TL 149), but it is not clear whether it could mean ‘sacrificial sheep’ also there. 68 See, e.g., Gusmani 2007: 19: “non si riscontra alcuna connessione formale tra la terminologia amministrativa licia rappresentata da pddênehm̃i, asaxlaza, epewêtlm̃mêi - designanti rispettivamente dei sovrintendenti regionali, il governatore di Xanthos e gli abitanti del relativo contado - e le traduzioni greche (ἄρχοντες, ἐπιμελητής, περίοικοι), che per la loro genericità semantica fanno pensare a corrispondenze approssimative, nient’affatto istituzionalizzate.” As for the opposite path of transmission, i.e., from Greek to Lycian, one essentially finds the same situation observed for the previous stage, with only sporadic (and not assured) Greek loanwords in Lycian texts, such as sttala- (provided that it was a borrowing), perhaps the title(? ) manaxine < μονογενής (only in TL 40a and 40b), 61 and the very uncertain -m͂m-ma- < Gr. ἄμωμος ‘blameless’. 62 Besides loanwords, some lexical calques - both structural and semantic - can be identified, but the issue is quite problematic. Consider the following forms occurring in the trilingual Letoon inscription: Lyc. asaxlaza- (N 320, 5), an agent noun in -azafrom a base verb perhaps composed of the preverb ese- + xla(i)- ‘control’, 63 matching ἐπιμελητής ‘curator, manager’ in the Greek version; Lyc. epewẽtlm͂me- (N 320, 6, 13-14, 32), convincingly explained by Adiego (1993) as a participle of a verb epe=(e)wẽ=tl(e)meaning ‘move around behind’ (vel sim.), 64 corresponding to Gr. περίοικος; 65 and Lyc. kumehe/ i- ‘sacred’ with the specific meaning of ‘sacrificial sheep’ (N 320, 27), like ἱερεῖον found in the corresponding Greek passage. 66 Given the nature of the Lycian corpus, the fact that these Lycian forms only occur in N 320 so far 67 may suggest that they were calques on the corresponding Greek forms, which appear to semantically roughly match them. However, such an assumption - which is by no means assured 68 - poses some problems, because several hints suggest 178 Valerio Pisaniello & Stella Merlin <?page no="179"?> that it was the Greek version of the decree that depended on the Lycian one (see below). Therefore, despite their unique occurrence in N 320, one should perhaps assume that these calques originated independently from the specific context of the multilingual document in which we find them attested. However, once one disengages them from the Greek version of N 320, the possibility opens that the Greek models of these alleged calques could have been different from the forms attested in the Greek text of N 320. In other words, if asaxlazawas not an occasional formation calqued on ἐπιμελητής in the context of the Letoon trilingual, but had an independent life in the Lycian lexicon, its model could have been a different Greek word, and ἐπιμελητής may simply have been chosen as one of the possible corresponding expressions by the translator of the text. Needless to say, in such a scenario the assumption that they were calques would be quite aprioristic. However, the case of kumehe/ iseems to be hardly dismissible, which raises the following question: does the dependency of the Greek version on Lycian actually exclude the possibility of Lycian calques on Greek? In our opinion, the answer is no. The trilingual stele of the Letoon, indeed, is just the final output of what we can imagine was a complex process of elaboration, as shown by the fact that it included, as mentioned, two different texts of the decree: the more concise official version addressed to the Achaemenid administration, in Aramaic language, and the official version for the local administration, in Lycian and Greek. Conceiving the work of the scribes as a mechanical and straightforward process, first involving the drafting of the Lycian version and then its translation into Greek, is hardly plausible. One should probably understand it as a more dynamic process consisting of a number of stages in which, for instance, scribes with different skills worked together to produce a unified text in two different languages, which opens the possibility of Greek influences on the Lycian version. One could also imagine that some Greek designations for official titles or positions could have reflected specific choices made by the strongly Hellenised Hecatomnid administration, which influenced the Lycian text when the decree was being drafted. One cannot either rule out the possibility that the first draft of the decree - perhaps a concise text only containing general instructions to be implemented by the local administration - was issued in Greek by Pixodaros’ office, and the Lycian text we read represented a thorough reworking of such a document carried out after the implementation of the required measures (e.g., the appointment of cult officials). The definitive form of the Lycian text thus produced could then have been translated into the Greek version we read, which could also retain some of the choices found in the first draft. Of course, this is nothing more than a very speculative scenario of how things might have been, but it would allow us to Ancient Greek and late Anatolian languages 179 <?page no="180"?> 69 On the discrepancies between the Lycian and Greek versions and possible analyses, see, e.g., Storme 2014 and Melchert 2021: 349-359. According to Melchert (2021: 358), the Lycian and Greek texts we read can be the product of “a native speaker of Lycian with an excellent knowledge of Greek,” who also “may have had to invent suitable phraseology for which Greek had ready-made collocations and vocabulary that were familiar to him” (p. 359). Such a scenario, quite plausible, is not necessary in contrast with the one tentatively suggested here, because it only concerns the final steps of the composition of the document, and can be easily accommodated with the assumption of a longer redactional phase including preliminary drafts. explain both some phenomena of mutual influence and those cases of significant divergence between the two versions. 69 Outside the Letoon trilingual, a possible calque is found in the funerary inscription TL 6 from Karmylessos: (5) 1. ebẽñnẽ ñtat- m=ene=prñnaw-tẽ pulenjda mullijeseh se=dapara pulenjdah puri- 2. himetehe pr[ñ]n[e]zijehi hrppi lada epttehe se=tideime se=ije ti=(e)seri ta- 3. di tike ñtat[a] ebehi me=ije [httẽm]i punamaθθi aladahali: ada ∠ - 4. τοῦτο τὸ μνῆμα ἐργάσαντο Ἀπολλωνίδης Μολλίσιος καὶ Λαπάρας 5. Ἀπολλ[ω]νίδου Πυριμάτιος οἰκεῖοι ἐπὶ ταῖς γυναιξὶν ταῖς ἑαοτῶν 6. [κα]ὶ τοῖ[ς] ἐγγόνοις· καὶ ἄν τις ἀδικήσηι τὸ μνῆμα τοῦτο 7. ἐξώλεα [κ]αὶ πανώλεα εἴη ἀοτῶι πάντων Lycian: ‘This burial chamber, Pulenjda, (son) of Mullijese, and Dapara, (son) of Pulenida, household members of Purihimete, built it for their wives and sons; and who lays someone (else) in this burial chamber, the anger of all (be) upon him. For the transfer (of the corpse): 5 adas.’ Greek: ‘This tomb, Apollonides, (son) of Mollisis, and Laparas, (son) of Apollo‐ nides, household members of Purimatis, built (it) for their wives and descend‐ ants: if someone will damage this tomb, the complete destruction of everything be on him.’ In this bilingual inscription, in which the Lycian and Greek versions mostly correspond, Lyc. prñnezi(je)- ‘household member’ (ultimately built on an unat‐ tested Lycian noun cognate of Hitt. per/ parn-, Luw. parna- ‘house’) is matched by Gr. οἰκεῖος. The latter form also occurs in the Greek inscription dealt with above together with TL 139 (ex. 2). According to Rutherford (2002: 205-206), the Lycian word could be a calque on the Greek form, unless one assumes an independent development. However, as noted by Brixhe (2007: 933), Gr. οἰκεῖος did not have the meaning of Lyc. prñnezi(je)in the rest of the Greek world, so that it should probably be regarded as a semantic calque on Lycian. A likely Lycian semantic calque on Greek could be the Lycian preverb epi, which, although being cognate of Luw. appi ‘back’ (cf. also Hitt. āppa ‘id.’), in some 180 Valerio Pisaniello & Stella Merlin <?page no="181"?> 70 See, e.g., Melchert 2004: 14 and Melchert 2014a: 68. Such a solution is not accepted in eDiAna, s.v. Proto-Luwic *ópi (https: / / www.ediana.gwi.uni-muenchen.de/ dictionary.php? lemma= 3087). 71 See Rutherford 2002: 206, with references. 72 However, according to Daues (2009: 56-59), one could also defend the opposite scenario, i.e., Lyc. hrppi + dative as a calque on Greek. 73 See also the bilingual inscriptions TL 23 (Tlos) and TL 117 (Limyra), in which the Greek version has ἑαυτῶι and the Lycian one accordingly has (hrppi) atli ehbi. 74 For a broader discussion on this formula and its possible relevance for the relative chronology of the Lycian inscriptions, see Rix 2016: 108-113. instances seems to mean ‘upon’, like Gr. ἐπί, 70 although its interpretation in most of the contexts remains unclear. Vice versa, it has been suggested that the sporadic use of ἐπί + dative in some Greek funerary inscriptions to indicate the people for whom the tomb was built, generally expressed by a simple dative elsewhere, was a calque on the Lycian formula hrppi ‘for’ (< hr(i) ‘upon’ + epi) + dative. See particularly the bilingual inscription TL 6 from Karmylessos quoted above (ex. 5), in which Lyc. hrppi lada epttehe se=tideime is matched by Gr. ἐπὶ ταῖς γυναιξὶν ταῖς ἑαοτῶν [κα]ὶ τοῖ[ς] ἐγγόνοις, 71 vs., e.g., TL 56 from Antiphellos, in which Lyc. hrppi ladi: ehbi se tideime: ehbije: ‘for his wife and his sons’ is matched by a simple dative in the Greek version, αὐτῶ[ι] τε καὶ γυναικὶ καὶ τέκνοις ‘for himself, and (his) wife, and (his) sons.’ 72 The latter example also provides evidence for another possible example of Greek influence on Lycian: the Greek list of beneficiaries of the monument starts with the builder himself, expressed by a reflexive pronoun (αὐτῶ[ι]), information that is not explicitly found in the Lycian version, although, of course, the builder of the monument was probably also one of its beneficiaries. In a limited number of Lycian inscriptions, the builder is also explicitly mentioned among the beneficiaries with the formula (hrppi) atli (ehbi) ‘for himself’ (lit. ‘for his person; for his own self’), which, due to its sporadic occurrence, is regarded as modelled on the Greek formula. 73 Of course, one cannot exclude an independent development; however, even in this case, language contact may have favoured its diffusion. 74 An assured example of Lycian influence on Greek is the occurrence of the word υἱός ‘son’ after the father’s name in genitive case in the filiation formulae of Greek inscriptions, which is probably determined by the almost-consistent presence of the word tideimi ‘son’ in Lycian filiation formulae. Of course, the ex‐ plicit occurrence of υἱός in filiation formulae was a possible and independently attested strategy in Greek, but it was far less common than the occurrence of the mere father’s name in genitive case. In this case, interference simply promoted a lesser but fully grammatical use in Greek. Indeed, its probable interpretation as a calque on Lycian relies on the fact that υἱός mostly seems to be confined Ancient Greek and late Anatolian languages 181 <?page no="182"?> 75 See Merlin and Pisaniello 2019: 97-98 for a broader discussion and further references. 76 See Wörrle 2012: 423. 77 For a broader analysis and further references, see Merlin and Pisaniello 2019: 98-100. 78 See, e.g., Rutherford 2002: 208-209 and Dardano 2015: 221. Sciancalepore (2017) provided the same explanation for Greek filiation formulae occurring in inscriptions from Lydia and Caria. to bilingual texts for which a direct Lycian model was available. 75 Outside these documents (i.e., in monolingual Greek inscriptions from Lycia), we only found a single example, the inscription H II 34 from Limyra (4 th -3 rd century BCE), 76 which structurally perfectly corresponded to a Lycian funerary inscription: (6) 1. Τοῦτο τὸ μνῆμα κατεσκευάσατο Κενδας Ασσ̣α̣ υἱ[ὸς] ἑαυτῷ <τ>ε 2. καὶ τῆι γυναικὶ αὐτοῦ καὶ το̣ῖ̣[ς] τέκνοις ‘This tomb, Kendas, son of Assa, built (it) for himself, his wife, and (his) children.’ Less straightforward is whether the sporadic absence of tideimi in Lycian filiation formulae could depend on the common Greek pattern without υἱός. Such an explanation seems likely in at least one case, the bilingual inscription N 312 from Xanthos, in which the Lycian version, occurring after the Greek one, seems to be secondary as far as the content is concerned: (7) 1. Δεμοκλ[εί]δης Θε[ρ]βεσιος 2. Λιμυρεύς ἀγαϑῆι τύχηι 3. Ἀρτέμιδι ἀνέϑηκεν 4. ñtemuxlida krbbe[s]eh 5. zemuris ertemi 6. xruwata Greek: ‘Demokl[i]des, (son) of The[r]besis, of Limyra, for good luck, dedicated to Artemis.’ Lycian: ‘Ñtemuxlida, (son) of Krbbe[s]e, of Limyra, votive offerings for Artemis.’ Again, filiation formulae without tideimi also occur sporadically in monolingual Lycian inscriptions, so that one should probably assume that it was a possible although lesser strategy in Lycian. A direct Greek model may have determined the choice of this use rather than the more common one. 77 Conversely, what cannot be regarded as an interference phenomenon is the absence of the definite article before the father’s name in the Greek filiation formulae. Although this absence is sometimes blamed on Lycian, which did not have the definite article, 78 one should instead notice that it was regular in Greek. Indeed, as shown by Merlin and Pisaniello (2019: 101-103), the situation of 182 Valerio Pisaniello & Stella Merlin <?page no="183"?> 79 See, e.g., Meisterhans 1900: 223-224, Gildersleeve 1911: 266, and Miller 1916, with further references. 80 Further examples can be found in Merlin and Pisaniello 2019: 101-103. 81 See Merlin 2022 for a thorough discussion and examples. 82 See Schürr 2010. 83 Another example of formal match between Lycian and Greek in a formulaic context concerns “no one else” prohibitions expressed in Greek by ἕτερος/ ἄλλος οὐδείς/ μηδείς or οὐδείς/ μηδεὶς ἕτερος/ ἄλλος, corresponding to kbi tike or tike kbi in Lycian (see Macedo 2021 for a thorough analysis). 84 According to Kloekhorst (2011: 17-18), Lycian unmarked word order was SVO. Greek inscriptions from Lycia, as far as definite articles in filiation formulae are concerned, perfectly corresponds to the rest of the Greek world: before a father’s name in genitive case, definite articles are consistently lacking when the son’s name is not in genitive case, while they consistently occur when the son’s name is in genitive case. 79 To provide just one example showing both cases, in TAM I 5 from Telmessos we read Διογένην Διογένου τοῦ Διογένου τοῦ Σωσικλέου, ‘Diogenes (son) of Diogenes, (son) of Diogenes, (son) of Sosikles,’ in which the article is only lacking between the first name in accusative case and the following name in genitive case. 80 Furthermore, as remarked by Merlin (2022), in general, the issue of the presence/ absence of determiners cannot be merely reduced to the definite article, because all the definiteness strategies should be taken into account, including, e.g., the use of demonstratives and possessives, which, unlike a proper definite article, are present in Lycian. Indeed, once one considers other determiners, some interesting structures can be identified that seem to be characteristic of Asia Minor Greek, which may possibly find an explanation in a language contact perspective. 81 Turning to another type of formulae, some translation phenomena are clearly visible in the curse formulae appended to Lycian and Greek inscriptions. Besides various examples of more or less accurate correspondence between Lycian and Greek curse formulae, one should particularly mention the phrase ἁμαρτωλὸς-ἔστω ‘be him guilty! ’ (against deities, usually θεοῖς (κατα)χθονίοις, θεῶν πάντων, etc.) that frequently occurs in the apodosis of the curse formulae in Greek inscriptions from Lycia (both monolingual and bilingual), because such an expression is specific to this region and does not occur elsewhere. Therefore, it is likely that it depended on Lycian curse formulae with sm̃mati ‘guilty’, 82 which is corroborated by the curse formulae of the Letoon trilingual (N 320), actually providing the suggested correspondence. 83 Interference also affected word order in several inscriptions. We will limit the discussion here to two major examples: first, Lycian funerary inscriptions mostly show a marked OVS order, 84 with a topicalized object in first position Ancient Greek and late Anatolian languages 183 <?page no="184"?> 85 On Lycian nasalized preterits, see Adiego 2015, with references. 86 The examples from N 320 are also discussed in Melchert 2021: 357-358, with the suggestion that Greek word order actually depended on Lycian. 87 See Melchert 2018 and Melchert 2021 for the analysis and translation of the Lycian text of N 320. anaphorically resumed by clitic pronouns, as, e.g., in the inscription TL 6 quoted above (ex. 5), in which one finds both the clitic pronouns =ene attached to the conjunction m(e) and the nasalized preterit form prñnaw-tẽ, originating from pronoun cliticization: 85 (8) ebẽñnẽ ñtat- m= ene= prñnaw-te-˜ - this. A C C . S G burial chamber. A C C . S G C O N J it build. P S T .3 S G -it - - - - - - - pulenjda mullijeseh … - - - - PN. N O M . S G PN. G E N . S G - - - - - - - - - - ‘This burial chamber, Pulenjda, (son) of Mullijese, … built it.’ The same word order is frequently found in Greek funerary inscriptions from Lycia, both bilingual (cf., e.g., the Greek version of TL 6: τοῦτο τὸ μνῆμα ἐργάσαντο Ἀπολλωνίδης Μολλίσιος …) and monolingual (see, e.g., H II 34 mentioned above; ex. 6). Second, in the following examples from Pixodaros’ decrees from Xanthos, one finds very uncommon Greek verb-initial clauses matching Lycian sentences showing the same marked word order, not counting clitics: 86 (9) N 320, Lyc. 12-15 / Gr. 12-14 s=ed(e)=eliñ=t-tẽ: teteri: se(j)=epewẽtlmẽi: hrmada: ttaraha: me=xbaitẽ: z-: ese=xesñtedi: qñtati<=ti>: se=pigrẽi: ‘And the city and the perioeci transferred ? land sections of the city and bound over ? a plot <that> Xesñtedi and Pigrẽi shall till ? .’ 87 -καὶ ἔδωκαν ἡ πόλις ἀγρὸν ὃγ Κεσινδηλις καὶ Πιγρης κατηργάσατο ‘And the city gave (pl.! ) the field that Kesindelis and Pigres worked at.’ 184 Valerio Pisaniello & Stella Merlin <?page no="185"?> 88 One could also suspect that a(i)- ‘do, make’ in the Lycian version may have triggered the choice of ποιέω ‘make’ vs. other possible verbs occurring elsewhere with ὅρκος (e.g., the more technical ὄμνυμι ‘swear’). One should notice that ποιέω sometimes occurs with ὅρκος (cf., e.g., Xenoph. De Rep. Laced. 15.7), so this would be a further example in which a direct Lycian model may have had a role in the choice of one of the possible means of expression of a given concept. 89 See, e.g., Blomqvist 1982: 17-19 and Rutherford 2002: 207. 90 The bibliography on Lycian onomastics, especially in relation to Greek, is massive: one may refer, among others, to Zgusta 1964, Colvin 2004, Cau 2003, 2005, 2006, Réveilhac 2018, and Adiego 2020. (10) N 320, Lyc. 30-32 / Gr. 26-28 se=ije=hri(j)=aitẽ: tasa: mere: ebette: teteri: arñnas: se(j)=epewẽtlmẽi: arñn-i: ‘And the city of Xanthos and the Xanthian perioeci made oaths for these laws.’ -καὶ ἐποιήσαντο ὅρκους Ξάνθιοι καὶ οἱ περίοικοι ‘And the Xanthians and perioeci made oaths.’ 88 (11) TL 45 A, Lyc. 1-3 / Gr. 1-3 [pijet]e=ñn=ẽ pixe[s]ere kat[amla]h arñna se tlawa se p[inale] se xadaw-ti … ‘Pixesere, (son) of Katamla, gave it to them, to Xanthos and Tlos and Pinara and Kadyanda …’ -ἔδωκεν Πιξώδαρος Ἑκατό[μνου Ξα]νθίοις Τλωίτοις [Πι]ναρέοι[ς Κανδα]ϋδέοις … ‘Pixodaros, (son) of Hecatomnos, gave to the Xanthians, Tloeans, Pinareans, Kadyandeans …’ Note that, in ex. 9, the dependency of the Greek sentence on the Lycian one (or, at least, on a text such as the Lycian one) would be granted by the Greek plural form ἔδωκαν occurring with a singular subject, whereas the Lycian version provides the probably original model with the expected plural subject. As a last phenomenon of syntactic interference, it has been observed that the frequent occurrence of καί in N 320 probably calqued the frequent Lycian use of the conjunction se in sentence-initial position. 89 We conclude our overview of language and cultural contact between Greek and Lycian in Asia Minor with some remarks on onomastics. The relevance of onomastic data is twofold: on one hand, they had - and still have - crucial relevance for the study of Lycian phonetics and phonology; much of our knowledge of these aspects relies on adaptation strategies of both Lycian names in foreign sources and foreign names in Lycian; 90 on the other hand, they provide insights into the sociolinguistic situation of the multicultural and multilingual context of the region. Indeed, as far as Greek and Lycian are concerned, the study of personal names provides a fairly parallel picture to the one presented so Ancient Greek and late Anatolian languages 185 <?page no="186"?> 91 Although it is quite a different phenomenon, one should mention that double names are sometimes also attested for place names in Lycia. Despite some examples of unexpected adaptations of Lycian toponyms in Greek (e.g., Pinale vs. Πίναρα and Zemurẽ vs. Λίμυρα), consider particularly those cases in which Lycian and Greek employ two radically different toponyms for the same place, such as Arñna vs. Ξάνθος, Wehñte/ ivs. Φέλλος, and, perhaps, Ερευα(τη/ ις) vs. Ἐλευθεραί (on which, see Pisaniello and Merlin forthcoming). See also Schürr 2007: 33-34. 92 For a general overview, see, e.g., Schuler 2019. 93 See, e.g., Melchert 2013: 48 and Melchert 2014a: 68. 94 See Schürr 2007: 36. far based on epigraphic documentation: from the 5 th to the late 3 rd century BCE, epichoric names were prevalent, although Greek names also occurred, and a couple of examples of individuals seemingly bearing both a Lycian and a Greek name were attested (see below); 91 later, the frequency of Greek personal names started to increase, eventually outnumbering the Lycian ones, which, however, did not entirely disappear. 92 Although there are several examples of adaptation of names among the different languages more or less directly attested in Lycia (besides Lycian and Greek, one could mention, e.g., Aramaic, Old Persian, and Carian), we will limit the discussion here to only Lycian and Greek, focusing on just a few of the many examples relevant to the issue of language contact. First of all, a number of Lycian personal names are built with a past participle as second member (e.g., Mahanepijemi ‘given by the gods’ < mahana- ‘god’ + pijemi, past participle of pije- ‘give’), a word formation strategy that, in the broader Anatolian perspective, is quite uncommon, being only attested in Lycia, Caria, and Pisidia. Conversely, such a strategy is very frequent in Greek, so that it is generally assumed that its occurrence in some 1 st millennium BCE Anatolian languages probably is due to Greek influence at the structural level. 93 Such a hypothesis can be corroborated by a couple of examples of direct translation: (1) in the Letoon trilingual inscription (N 320), one of the rulers installed by Pixodaros in Lycia bears the name Natrbbijẽmi ‘given by Natr’ in the Lycian version, but in the Greek text one finds Ἀπολλόδοτος, and it is possible that the Lycian name was calqued on the Greek one, rather than the other way round; (2) in the late Greek inscription TAM II 650 from Kadyanda (1 st century CE), an individual bears the Lycian name Ορνεπειμις (= *Urnepijẽmi), which is probably the Lycian translation of his father’s name Μεγιστόδοτος (Lyc. *urneis related to Luwian uranna/ i- < ura- ‘great’). 94 A different possible case of translation is found in the bilingual inscription TL 25a from Tlos, in which an individual named Xssbezẽ in the Lycian version bears the name Πόρπαξ in the Greek one. In this case, it is likely that the Greek 186 Valerio Pisaniello & Stella Merlin <?page no="187"?> 95 In the whole Greek onomastic corpus, it is only found as a dog name in Xenoph. Cyn. 7.5. 96 Also note the name Οσσυβας, attested in Lycia in some Greek inscriptions of the Roman period, which could be an adaptation of Xssbezẽ. 97 Hajnal (1995: 35-36 n. 28) tentatively traced it back to the PIE root *sah 2 (i̯)- ‘to bind’. 98 See Colvin 2004: 66-67 and Melchert 2014a: 68. One could also add that the name of Xssbezẽ’s father, Kruppse in the Lycian version, is unexpectedly rendered as Θρύψις in Greek (gen. Θρύψιος). Adiego (2007a: 766) regarded it as a Greek name, but, as far as we have seen, it only occurs here in the whole of the Greek onomastic corpus, so that it is likely that it was an original Lycian name. Schürr (2007: 34-35) explained the mismatch invoking a sound change, also providing further examples, but one could also suggest that this may be a further attempt to Hellenize the name, given that a noun θρύψις (but feminine and with genitive in -εως) actually exists in Greek, meaning ‘softness, weakness’. 99 See Molina Valero 2004: 1016 and Adiego 2020: 45. 100 On the role of Milyan in Greek adaptations of Lycian names, see also Martínez Rodríguez 2021. name was secondary to the Lycian one, because Πόρπαξ only occurs here as a personal name, 95 while Xssbezẽ (and variants) is attested elsewhere in the Lycian corpus. 96 However, it is unclear whether it was a translation, because πόρπαξ means ‘handle (of a shield)’, but the meaning of Lyc. xssbezẽ is unknown. 97 Note that in the same inscription TL 25a, one possibly also finds a partial Hellenization of a Lycian name: Lyc. Purihimete is rendered as Πυριβάτης in the Greek version, instead of the expected (and elsewhere attested) Πυριματις or Ποριματις. 98 The latter example leads us to the last issue we will briefly address in this section. In the Greek inscription JHS 15 (1895): 108,18, the same Lycian name Purihimete occurs as Πορασιματις, showing the preservation of intervo‐ calic / s/ that in Lycian consistently yielded / h/ . An analogous phenomenon is found in the personal name Κονδορασις vs. Lyc. Qñturahe-, both occurring in the Letoon inscription N 320, and in some toponyms, such as Telebehivs. Τελμησσός, Tuburehivs. Τυβερισσός, and Tuminehivs. Τυμνησσός. Such a situation can be accounted for through multiple solutions; one does not need to find a unique explanation valid for all of these cases. Thus, e.g., Κονδορασις in N 320, like its Aramaic corresponding form KDWRS, may depend on a Carian preform, and Qñturahecould be an independent Lycian adaptation of the same preform. 99 Instead, in other cases one may suspect that the model for the Greek form was not Lycian but rather Milyan, which preserved intervocalic / s/ (thus, e.g., the Milyan name for Tymnessos was Tuminesi). 100 During the 3 rd century BCE, Lycian gradually lost ground to Greek, eventually disappearing from epigraphic use. All that remained of the Lycian language in Lycia was restricted to onomastics and sporadic borrowings, along with some Ancient Greek and late Anatolian languages 187 <?page no="188"?> 101 The passages of the Iliad are the following: 2.877; 5.105; 5.173; 5.479; 5.645; 6.168; 6.171-173; 6.188; 6.210; 6.225; 12.312; 12.318; 16.437; 16.455; 16.514; 16.542; 16.673; 16.683; 17.172. 102 Σαρπηδὼν δ’ ἦρχεν Λυκίων καὶ Γλαῦκος ἀμύμων | τηλόθεν ἐκ Λυκίης, Ξάνθου ἄπο δινήεντος, ‘Sarpedon was lord of the Lycians, and blameless Glaucus, far away, from Lycia, from the whirling waters of Xanthos.’ 103 See Podestà (2019) on Hecataeus, two-thirds of whose fragments are transmitted in the Ethniká by Stephanus of Byzantium. formulae and stylistic modules that continue to characterize the monolingual Greek inscriptions of the later phase. However, traces of Lycian survived in the Greek literature contemporary with and following the time of its epigraphic attestation, also surfacing in the later scholarly tradition. 2.2. The Lycians and Lycian language in literary and scholarly tradition The Lycians have been present in Greek literature since the very beginning. The region of Lycia (Λυκία) is often mentioned in the Iliad (but never in the Odyssey). 101 Lycians are the last contingent of the Trojan army listed in the so-called Catalogue of the Trojans (Il. 2.876-877), a shorter counterpart of the previous Catalogue of ships (Il. 2.494-759): the region, said to be ‘far away’, settled around the Xanthos river, is ruled by the kings Glaucus and Sarpedon. 102 As Podestà (2022: 40-41) observed, Lycia in Homer is frequently defined as wide and fruitful, in seeming contradiction with the region of Xanthos river, and this could be evidence of the intersection of two traditions: on one side, the memory of Lukka, which extended across Western Anatolia in the territories of Lydia and Caria, together with, on the other side, the power established in the Xanthos valley during the 8 th -7 th centuries BCE. Moreover, Lycia is mentioned once in the Homeric Hymns (Homeric Hymn to Apollo 3.179), together with Maeonia, namely the Homeric name for Lydia (see section 3.2 below). It also appears a few times in Hesiod and the tragedians, and in the texts of the historians (Hecataeus, 103 Herodotus, and Thucydides). As a matter of fact, as the history of the region goes on, the reference to the political, social, and linguistic entity represented by the name of Lycia changes: the ancient Anatolian nationality (probably corresponding to that indicated by the name Lukka in Hittite texts) became a satrapy of the Achaemenid Empire during the reign of Darius I (522-486 BCE), then it was Hellenized following the conquest in 334 BCE by Alexander the Great, and finally (43-44 CE) it became 188 Valerio Pisaniello & Stella Merlin <?page no="189"?> 104 See Keen (1998) for the relationships between Lycians and foreign rulers. Some reference to Lycia is also attested, beyond Lycian-Greek bilinguals, by monolingual Greek inscriptions from different regions: however, several of them date from the first centuries CE and thus refer to the Roman province of Lycia or Lycia and Pamphylia, not to mention the fact that, meanwhile, Lycia had also become a personal name. 105 Bryce (1986) is one of the most comprehensive assessments of Greek and Roman sources about Lycia: see, in particular, Bryce 1986: 11-23, on the traditions about the participation of Lycians in the Trojan war, the Bellerophon tradition, and the reference to Τρεμίλαι, and Bryce 1986: 99-114, devoted to the historical background. More recently, Podestà 2022 offers a deep analysis of the ancient Lycian historians, from Menecrates of Xanthos (5 th -4 th centuries BCE) to Capito of Lycia (5 th -6 th centuries CE). 106 On Menecrates of Xanthos, see Podestà 2022: 91-141. For the reference to the Lykiakà, see also Bryce 1986: 208. 107 According to the myth, Apollo’s epithets include Λύκειος and Λυκηγενής, which could refer either to the wolf as a sacred animal to him or to the land of Lycia, the region in which he was allegedly born. the Roman province of Lycia (united with Pamphylia under the rule of some emperors). 104 According to Bryce (1986: 215): “the majority of our Greek and Roman writers knew only of a Hellenised Lycia, or a Helleno-Romanised Lycia, and apparently believed that the Lycian people had never been any more than a refugee offshoot of the early Greek or Aegean civilisations. Even the fact that the Lycians originally had a distinctive language of their own seems to have aroused little or no interest, to judge from the absence of any reference to it in our literary sources.” 105 Nonetheless, memories and echoes of the ancient Anatolian Lycia, its history, and myths, remain in the ancient literary and scholarly tradition even long after its disappearance. This is precisely the concept of “survivals” transmitted by indirect tradition, mediated by both time and language. In this sense, Greek sources can be examined in search of linguistic as well as cultural and historical information about Lycia, its inhabitants, and their aboriginal language. Ancient monographs entirely devoted to the history of Lycia, known as Λυκιακά, started to be written between the 6 th and 5 th centuries BCE, Menecrates of Xanthos plausibly being the first Lycian historian. 106 One of the fragments, transmitted by Antoninus Liberalis (2 nd century CE? ) deals with the etymology of the name for Lycians, presenting a tradition according to which it derives from λύκιοι ‘wolfs’. According to the legend, the goddess Leto with her children Artemis and Apollo, after being driven by the inhabitants from the Melite spring, was accompanied by wolves to the river Xanthos. There the infants were washed, the Xanthos became sacred to Apollo, and the region, formerly named Tremilis, was named Lycia in memory of the wolves who cared for the goddess. 107 Ancient Greek and late Anatolian languages 189 <?page no="190"?> 108 See Podestà 2022: 267-272; cf. also Merlin and Pisaniello 2020: 190 with fn. 27. 109 See Melchert 2004: 70, with details on occurrences, related forms (e.g., Trm̃ milis), and further references. 110 See, e.g., Hesychius τ 1299: Τρεμ[ε]ιλία· ἡ Λυκία. 111 This clarification is made necessary by the possible conflict with other definitions, particularly in the context of the Anatolian scholarship (cf., e.g., Pisaniello 2020, with references). The reference to such an original name is also found in the Ethniká by Stephanus of Byzantium under the entry Τρεμίλη: it is first said, reporting a fragment of the 5 th century BCE poet Panyassis, that the name of Τρεμίλη is derived from the eponym hero Tremilus; then, according to Alexander Polyhistor, that Bellerophon, having defeated the Τρεμιλεῖς, changed their name in Lycians. 108 In these types of texts, there are surely many elements that can lead to a clear and more complete representation of the mythological, historical, and geographical aspects of Lycia. What we are interested in, in this contribution, is purely linguistic information and its metalinguistic scope: in the passage of Stephanus mentioned above, we have evidence of a name change for the people. This is a particularly fortunate case, because we have the actual correspondence of the Greek form with the Lycian form Trm̃ mile/ i-, attested in the corpus of Lycian inscriptions, which means ‘Lycian’ both as an adjective and a noun. 109 Thence, Greek shows an adapted loanword, which is attested in the two morphological variants of the first and third declension: Τρεμίλαι, -ῶν, οἱ as well as Τρεμιλεῖς, -έων, οἱ. 110 The forms with metathesis are also attested: Τερμίλαι, -ῶν, αἱ (Hero‐ dotus, Strabo) to indicate the region (cf. the variant Τρεμιλίς, -ίδος attested by Menecrates of Xanthos apud Antoninus Liberalis, see above) and Τερμιλεῖς, -έων, οἱ, the latter only attested four times, twice in the grammarian Herodian (2 nd century CE) and twice in Stephanus of Byzantium (6 th century CE). The linguistic information we are looking for are usually called “glosses”. Before discussing some examples of Lycian (and then Lydian) glosses, it is worth briefly defining the concept of gloss in itself. 111 In our research, we consider any metalinguistic information about a given language as gloss. In the particular situation we are dealing with, an Anatolian (here Lycian or Lydian) gloss in Greek is essentially a Greek text providing some more or less extended linguistic information about an Anatolian language under consideration. The typical occurrence is a quotation for which a translation is generally provided or can be easily recovered by the context, e.g., “the Lycians called x so-and-so”, “this is the name of x in Lycia”, et sim. This is the case of the “overt glosses”: the metalinguistic information about the foreign language is clearly given by the 190 Valerio Pisaniello & Stella Merlin <?page no="191"?> 112 See, e.g., Schmitt 2014: 255. 113 See the discussion in Merlin, Pisaniello, and Rizza (forthcoming-a), with references. 114 See, e.g., Willi 2004, with references. 115 Cf. Photius, s.v.-Ἰδαρναῖοι-(ι-15). 116 See Adiego 2007b: 7-12, 455. Greek text, being separated from the purely linguistic flow of the text itself. But there are also “covert glosses”, which are more difficult to detect because, as we will see in some examples, they are implicit, and both the co-text and context should be taken into consideration in order to detect them. Glosses are found, as expected, in all the scholarly literature as glossaries, lexica, commentaries on literary texts, whenever a foreign term or expression needs an explanation. However, glosses are also found in literary texts, of different genres, both in form of overt and covert information. Strabo, for example, provides interesting cases of overt glosses of different languages, such as Geogr. 5.4.12, in which it is said that hirpos means ‘wolf ’ in the language of Samnites (and this is confirmed by Phaestus: hirpini appellati nomine lupi quem hirpum dicunt Samnites), or Geogr. 11.14.8, where the author reports that tigris means ‘arrow’ in the Median language. 112 It is clear from these examples that the reliability of the gloss as a true indication of a foreign word cannot be completely entrusted to the metalinguistic awareness of the ancient authors, but, on the contrary, must be verified from the point of view of modern research and knowledge of the model languages. Covert glosses may be such by a twofold path: they can be either for‐ eign expressions employed without any explanation within a literary text or opaque and implicit metalinguistic information. The possible Lydian for‐ mula pronounced by a Lydian woman in a parodic composition by Hipponax (fr. 95 Degani) 113 or the (pseudo-)Iranian expressions employed by the Per‐ sian ambassador Pseudartabas in Aristophanes’ Acharnians (100) 114 belong to the first case. An example of an opaque and implicit gloss is given by the entry Ἰδάρνας (Hsch. ι 168) for which different explanations are provided: it can mean ‘castrated’ or ‘barbarian’; some say that it is the name of a diviner; others that there is a city in Caria named Idarne and the diviners are called by that name. 115 The text is highly elliptical and the reader may imagine that a Carian name for the diviners was somehow connected with the city name. 116 The difference between the two categories of covert glosses resides in the aim: they could be linguistic expressions inserted in order to reach some stylistic effect, Ancient Greek and late Anatolian languages 191 <?page no="192"?> 117 From the point of view of the linguistic description, covert glosses could also be represented as phenomena of code-switching: however, because the written dimension always precedes the oral one in the ancient documentation, we conventionally also use the term glosses in reference to possible code-switching phenomena for which we have only written records. 118 To this list, one could add καδρεμα ‘grain parching’ and ερευα ‘free(dom)’, on which see Pisaniello 2021a and Pisaniello and Merlin forthcoming, respectively. 119 See, e.g., EDHIL: 659-660. 120 On this toponym, see Merlin and Pisaniello 2020. as beauty, elegance, exoticism, but also humor and parody; very differently, a metalinguistic gloss aims to describe a linguistic expression in itself. 117 Lycian glosses for which an explicit mention of the Lycian origin is given are, in fact, very few. Molina Valero (2010) selected the following: πίναρα ‘round’, πάταρα ‘basket’, τύμνια ‘wand’ by Xanthians, in Lycia; σιφλός ‘a passive man’, a word transmitted by Eustathius, but it is possibly a Greek word used in Lycia during the 12 th century CE, Τζέλυμοι, a local variant for Σόλυμοι, which seems to point to a phonetic variation rather than a genuine gloss; finally Ὕλαμοι, the name of fruits, according to Stephanus of Byzantium, who reported a previous tradition about a Lycian city. This last one could be an example of covert gloss, for which an explicit metalinguistic reference to Lycian is not given by the ancient commentator, but some textual reference is present in the same text and could direct the attention to a possible Lycian word. 118 Unfortunately, a clear correspondence as that observed for Τρεμιλεῖς (see above) is not always available for these lemmas, and careful assessment of all the available evidence is needed. Thus, e.g., Patara and Pinara are two city names, which Stephanus traces back to the alleged Lycian words πάταρα ‘basket’ and πίναρα ‘round’, respectively. The former can actually be connected with an Anatolian word having the same meaning, i.e., the Hittite pattar ‘basket’. 119 Although, of course, nothing can guarantee that this was the true etymology of the Lycian toponym Pttaraand did not reflect, e.g., a folk etymological association, the combination between Stephanus’ gloss and an independent Hittite evidence makes it likely that Lycian actually had a word pttara- (vel sim.) meaning ‘basket’. Conversely, in the case of Pinara, no comparable form is currently attested in Anatolian, and the fact that the Lycian city name was Pinalefurther complicates the picture. 120 The case of τύμνια ‘stick’ is of partic‐ ular relevance because it appears in Stephanus, under the entry Τυμνησσός, a city of Caria, named as such because of the “wand of Tymnessos”, since τύμνια is the name given by the Xanthians to the wand. Molina Valero (2010: 461-462) suggests that this word could be related to the Lycian verbal root tub(e)i- ‘strike’, 192 Valerio Pisaniello & Stella Merlin <?page no="193"?> 121 See Melchert 2004: 72. 122 See, e.g., Melchert 2003, Rieken 2017, Yakubovich 2022. 123 51 Lydian inscriptions were published by Buckler (1924), which became 53 in Friedrich’s edition (Friedrich 1932: 108-123). The reference edition is Gusmani 1964 (= LW), with Ergänzungsbände, including 61 inscriptions. Those found in the following years were published separately in different volumes and journals (a full list can be found in LW, Erg. 1: 8). The whole corpus has also been published by Melchert (2001b), and an online edition is currently available on the eDiAna project website (https: / / www.ediana.gwi. uni-muenchen.de/ index.php). 124 See LW 49, a graffito from Silsilis (Egypt), and LW 52, a legend found on different coins. Some inscriptions are dated to the 6 th century BCE (the pottery inscriptions LW 30, 31, 32, the stone inscription LW 48, and the seals LW 55 and 56), while others date to the late 6 th or early 5 th century BCE (the pottery inscriptions LW 57, 60, 61, and the graffito LW 58). See Gusmani 1964: 17-18. 125 See Gusmani 1964: 19-20 and Gérard 2005: 21-22. 126 The kingdom of Lydia flourished between the 7 th and the 6 th centuries, until 546 BCE, when Croesus was defeated by the Persians and Lydia became a Persian satrapy until the conquest by Alexander the Great in 334 BCE. For the history of Lydia in the 1 st millennium BCE, see especially Högemann and Oettinger 2018, Payne 2023, and Klinkott 2023: 599-611. 127 On the Lydian alphabet, see, e.g., Gérard 2005: 22-28, with references. matching the Luw. dup(a)i- ‘id.’, 121 although the required change / bn/ > / mn/ is not attested for Lycian, but only for Greek. Furthermore, the corresponding Lycian toponym is Tuminehi-, whose relation to the verb tub(e)iis by no means straightforward. All in all, if one accepted an ultimate derivation from tub(e)i-, bringing all these forms together in a consistent picture is a far from easy task. 3. Ancient Greek and Lydian 3.1. Ancient Greek in Lydia and its interactions with the Lydian language Lydian was an Anatolian language whose exact position within the Anatolian group is still not entirely clear, because of our limited knowledge of the language. 122 The Lydian corpus is currently composed of little more than 100 inscriptions (many of them very fragmentary) - also including metric texts - and coin legends. 123 Graffiti and coin legends are attested since the second half of the 7 th century BCE, 124 but most of the inscriptions are dated to the 5 th and 4 th centuries BCE. Lydian inscriptions mostly come from Sardis, 125 the capital of the Lydian kingdom, 126 and are written in an epichoric alphabet that probably was an adaptation of a Greek eastern alphabet. 127 As far as the history of Ancient Greek in Lydia is concerned, we will outline a general overview based on the inscriptions discovered so far, although only focusing the discussion on the period of coexistence with the Lydian language Ancient Greek and late Anatolian languages 193 <?page no="194"?> 128 See, e.g., Cook 1987: 17-18. 129 Hdt. 1.92: Κροίσῳ δὲ ἐστὶ ἄλλα ἀναθήματα ἐν τῇ Ἑλλάδι πολλὰ […] ἐν δὲ Ἐφέσῳ αἵ τε βόες αἱ χρύσεαι καὶ τῶν κιόνων αἱ πολλαί.’ 130 See Wankel 1979: 1-5 (n. 1). 131 See Cook 1987: 40 (n. 33). 132 See Börker and Merkelbach 1979: 1-2 (nn. 101 and 104). 133 See Buckler and Robinson 1914: 58-61 (n. 20) and Buckler and Robinson 1932: 103 (n. 102). An inscription of the same, quite atypical type is the Pytharchos inscription found in the vicinity of Persepolis, also dated to the 5 th century BCE and only reading Πυθάρχο εἰμί (see Benvenuto and Pompeo 2017: 27-28, with references). - i.e., roughly between the 5 th and 2 nd centuries BCE - focusing mainly on the earliest inscriptions. Currently, the earliest Greek documents from Lydia seem to be some very fragmentary inscriptions found on a number of fragments of column bases in the temple of Artemis at Ephesus, all probably bearing the same text, which can be reconstructed as Βασιλεὺς Κροῖσος ἀνέθηκεν, ‘The king Croesus dedicated (it).’ 128 (12) [Βασιλεὺς] Κρ[οῖσος ἀνέθηκεν] [Βασιλεὺς Κροῖσος ἀνέ]θηκ[εν] Βα[σιλεὺς Κροῖσος] ἀν[έθηκεν] [Βασιλεὺς Κροῖσος ἀνέθηκ]εν These inscriptions refer to the dedication of the columns of the temple by king Croesus, thus can be dated to the period of his reign, between 560 and 546 BCE, and they seem to confirm Herodotus’ statement that “there are many offerings of Croesus in Hellas […] at Ephesus, the golden oxen and the greater part of the columns.” 129 Plausibly also related to the temple of Artemis is another inscription from Ephesus dated to the 6 th century BCE, written on a silver plate in Ionic alphabet and with a partly boustrophedon orientation, which records amounts of money probably allocated to the construction of the temple (Ephesos 239). 130 Other archaic documents of the 6 th century BCE from Ephesus include an inscription containing rules for the interpretation of signs through ornithomancy (Ephesos 1) 131 and a sacred law concerning sacrifices (Ephesos 2), whereas a couple of inscriptions identifying sanctuaries (Ephesos 899 and 900) 132 and another very fragmentary sacred law (Ephesos 3) are dated to the 5 th century BCE. Among the oldest Greek documents found in Lydia, one also finds inscriptions belonging to the private sphere, such as the funerary inscription on a marble stele from Sardis, dated to the mid-5 th century BCE (perhaps even before) on a paleographic basis, which only reads Λεωμάνδρο : εἰμί, ‘I am of Leomandros’ (Sardis 7,1 102). 133 Possibly later but still dated to the 5 th century BCE is 194 Valerio Pisaniello & Stella Merlin <?page no="195"?> 134 See Buckler and Robinson 1932: 103-104 (n. 103). 135 First edited by Cousin and Deschamps (1889). Among the numerous studies, cf., e.g., van den Hout 1949, Hansen 1986, Briant 2003, and Lombardi 2010, with further references. 136 See Cousin and Deschamps 1889: 541-542. 137 See lines 22-28 of the letter: φυτουργοὺς γὰρ [ἱ]εροὺς Ἀπόλλ[ω]νος φόρον ἔπρασσες καὶ χώραν [σ]καπανεύειν βέβηλον ἐπ[έ]τασσες ἀγνοῶν ἐμῶν προγόνων εἰς τὸν θεὸν [ν]οῦν, ‘you exacted tribute from the sacred gardeners of Apollo and ordered them to dig profane land, ignoring the disposition of my ancestors towards the god.’ 138 See especially van den Hout 1949: 144-152 (probably a forgery of the Roman Imperial period to prove the antiquity of the temple of Apollo) and Hansen 1986 (a “political manifesto by the Ionian Greeks against the Persians and their king” forged between 494 and 490 BCE, after the destruction of the temple of Apollo at Didyma by Xerxes, and originally written in Persian). another inscription on a marble stele from Sardis (Sardis 7,1 103), unfortunately fragmentary, which could have been a metric inscription (two iambic verses), based on the current restoration of the text: 134 (13) [μνῆμ’] Ὀξύλο [το͂ - ⏑ ] [- πρὸ]ς τὴν [ὁδὸν] | [Στράτ]ηγος Ἴ[λου(? ) τῆιδ’] εὐαυ[γὲς] εἵ[σατο] | ‘[The monument] of Oxylos, [(son) of …, agai]nst the [road, Strat]egos, (son) of I[los(? ),] pl[aced here] in full vie[w].’ While these are the earliest original Greek documents from Lydia, one should mention that there are some considerably later official inscriptions, dated to the Roman Imperial period, whose texts (or part of them) could be traced back to a much earlier date, i.e., to the period of Achaemenid control in Lydia. The first one is the much discussed letter of Darius I to Gadatas, 135 found in the vicinity of Magnesia on the Meander, on the Carian border: although the inscription clearly dates to the 2 nd century CE, the text appears as a letter sent by Darius I (550-486 BCE) to his “servant” (δούλος) Gadatas, an official of the Persian king in Asia Minor, containing a reprimand for failing to properly execute the king’s orders. The second column of the stele included a very fragmentary inscription of the Roman Imperial period concerning Apollo, 136 thus being seemingly related to Darius’ letter. 137 The style, terminology, and phraseology of the letter are patently non-Greek, so it is generally regarded as the translation of an original Aramaic document, possibly made already in the 5 th century BCE together with the original document. However, some scholars suggested that it is a forgery, made either in the time of Darius I or later in the Roman Imperial period. 138 Ancient Greek and late Anatolian languages 195 <?page no="196"?> 139 On the meaning of ὕπαρχος, see Robert 1975: 312-313. 140 See, e.g., Robert 1975, de Hoz 1999: 130 (n. 2.1), Canali De Rossi 2004: 136 (n. 235), Petzl 2019: 106-108 (n. 434). 141 It is possible that the original document was written in Aramaic, but an Ionic version may also have existed (see Robert 1975: 310). 142 See the genitives in -εω and the forms ἐτέων, τριήκοντα, and Λυδίης. 143 Cf. Robert 1975: 310 fn. 11: “les parties II et III ne semblent pas susceptibles de contenir des formes ioniennes caractérisées, différentes de la koinè.” 144 Cf. lines 1-4: ἔτεος ΙΙΙΙΙΙΙ, μηνὸς ἐβδόμω· βασιλέοντος Ἀρταξέσσεω· ἐξσατραπεύοντος Ἰδριέως. 145 See Sokolowski 1955: 173-174 (n. 75) and Canali De Rossi 2004: 140-141 (n. 240). 146 The name derived from Old Persian *R̥taxšaça-; cf., e.g., Aram.’rtḥšsš, Elam. I/ Artak‐ šašša (with several variants), and Lyd. Artakśassa- (see Schmitt 1987). The second document includes the dedication of a statue to Zeus Baradates by Droaphernes, “satrap” (ὕπαρχος) 139 of Lydia. The inscription, found at Sardis, can be dated to the mid-2 nd century CE, but the first two lines place the dedication in the 39 th year of the reign of a king Artaxerxes, probably the second, thus 365 BCE (although Artaxerxes I is not entirely excluded, thus 426 BCE). 140 Particularly, lines 1-5 would include a copy of the original document, 141 as would be shown by the large amount of Ionicisms: 142 (14) ἐτέων τριήκοντα ἐννέα Ἀρταξέρξεω βασιλεύοντος, τὸν ἀνδριάντα Δροαφέρνης Βαρ<ά>κεω, Λυδίης ὕπαρχος Βαραδατεω Διί. ‘In the 39 th year of Artaxerxes’ reign, Droaphernes, (son) of Barakes, satrap of Lydia, (dedicated) the statue to Zeus Baradates.’ This section is followed - separated by a leaf - by two injunctions, dated to the Roman Imperial period, 143 but also clearly related to the dedication of the statue, perhaps summarizing materials from the original decree. A third document one could add is an inscription from Tralles - on the border between Lydia and Caria - that can be dated to the 1 st century CE, but the text seems to be a copy of an earlier decree dated to the 7 th year of reign of an Artaxerxes, probably the third, given the mention of the Carian satrap Idrieus (thus 352 BCE). 144 The text is a decree of asylia for the sanctuary of Dionysus, and it is suspected to be a later forgery, in connection with the general re-evaluation of the rights of asylia that took place under Tiberius. 145 From a linguistic point of view, besides Ionicisms (e.g., ἱκετηρίην and the genitive in -εω), one could note some peculiarities, such as the unique Greek rendering of Artaxerxes’ name as Ἀρταξέσσεω (gen.), which is closer to the original form compared to the more common Greek form Ἀρταξέρξης, 146 and 196 Valerio Pisaniello & Stella Merlin <?page no="197"?> 147 The name of the satrap in the non-Iranian traditions is generally closer to the Old Iranian *xšaθra-pāthan its Old Persian reflex xšaçapāvan-, probably pointing to Median as the direct source (see Schmitt 2014: 284-285). For the prosthesis in ἐξσατραπεύω, cf. perhaps Aram. ’ḥšdrpn (besides ḥštrpn’) and Akk. aḫšadrapannu. See also ἐξαιθραπεύον[τ]ος in some inscriptions from Mylasa in Caria of the 4 th century BCE (see Canali De Rossi 2004: 136-140, ns. 236, 237, and 238). 148 See Buckler 1924: 38 (n. 20). 149 Buckler (1924: 38) also noted that “[a]bout a metre to the westward and aligned with this pedestal stands an exact duplicate, on which a 2-line inscription has been carefully erased,” later (Buckler and Robinson 1932: 91) adding that, based on its length and position, it could have been a duplicate of the bilingual inscription. 150 See Gusmani 1964: 74-75. 151 See also Buckler 1924: 57-58 (n. 40). For the Lydian text and analysis, we follow Payne and Sasseville 2016, with adjustments by Sasseville 2021a: 524. the verb ἐξσατραπεύω, which is also closer to the Iranian name of the satrap than the common Greek σατραπεύω. 147 The bilingual inscription LW 20 from the temple of Artemis at Sardis 148 can also be dated to the mid-4 th century BCE, one of the only two Lydian-Greek bilingual inscriptions found so far. The very short inscription is found on the base of a statue and consists of a dedication to Artemis: 149 (15) nannas pakiwališ artimuλ - ‘Nannas, (son) of Pakiwas, for Artemis.’ Νάννας Διονυσικλέος Ἀρτέμιδι ‘Nannas, (son) of Dionysikles, for Artemis.’ The only noteworthy element is the replacement of the Lydian patronymic pakiwališ with Διονυσικλέος, probably a translation, because Lyd. *pakiwa-, in all likelihood, matched Gr. Βάκχος. 150 Probably not much later, perhaps dated to 330-325 BCE, is the other - also very short - Lydian-Greek bilingual inscription, LW 40, found in the temple of Athena at Pergamon: 151 (16) ẹšν̣ tasẽν̣ f̣ ? acνil partaras maλiλ - ‘Partaras dedicated (vel sim.) this column to Malis.’ Παρτάρας Ἀθηναίηι ‘Partaras to Athena.’ In this case, one may notice that the Greek text is clearly secondary, because it only includes the dedicator’s name and the theonym Athena in dative case, while the Lydian one consists of a full sentence. Quite similarly to LW 20 (ex. 15 Ancient Greek and late Anatolian languages 197 <?page no="198"?> 152 See also, e.g., Serangeli 2015 and Warbinek and Giusfredi 2023. 153 See Robinson 1923. 154 See Wankel 1979: 6-9 (n. 2). 155 See Buckler and Robinson 1912 for a thorough analysis of the text. 156 For the date of the document, see, e.g., Debord 1982: 244-251 (for previous different suggestions, see Buckler and Robinson 1912: 22-26 and Buckler and Robinson 1932: 5-6). For the reasons that may have prompted the copying of the original document almost a century later, see Descat 1986. 157 See Buckler and Robinson 1912: 28-52 and Cuny 1913. 158 Among the earliest documents, we could mention, e.g., a law about debt (ca. 297/ 296 BCE) and a lease tender (ca. 290 BCE) from Ephesus (Wankel 1979: 8-27, nn. 3 and 4; ), a boundary stone from Philadelphia (de Hoz 1999: 157, n. 5.12; 279-267 BCE), and a couple of ex-voto inscriptions from Sardis (Sardis 7,1 87; first half of the 3 rd century above), the Greek version includes a translation, this time of the divine name: the Lydian goddess Malis (cf. Hitt. Maliya, Lyc. Malija) is equated to Greek Athena. 152 Besides such documents, a parallel “pure” Greek tradition also existed, as is attested, e.g., by the poetic epitaph of Elpis in elegiac couplets (Sardis 7,1 104), probably dated to the end of the 4 th or beginning of the 3 rd century BCE, structured in the form of a monologue with an address to a passer-by and displaying some epicisms (e.g., εἰν and ξένε). 153 Another poetic epitaph in elegiac couplets is that of Matis (Petzl 2019: 226-227, n. 688), also from Sardis, whose suggested date spans from the 4 th to the mid-3 rd centuries BCE. A long inscription from Ephesus (Ephesos 572), dated to the last quarter of the 4 th century BCE, decrees the death sentence of more than 40 Sardians who profaned the sacred objects and assaulted the theoroi from Ephesus. 154 This inscription includes several non-Greek personal names, which show some interesting phenomena of adaptation for which Lydian interference has been invoked (see below). Much debated is the date of an important public document, the inscription Sardis 7,1 1, which is the copy of a mortgage deed through which Mnesimachus conveyed to Artemis the estate awarded to him by an Antigonus, in order to return a loan that the temple made to him. 155 The inscription is probably dated around the (second) half of the 3 rd century BCE, but, if the Antigonus mentioned should be identified as Antigonus I Monophtalmus (382-301 BCE), the original deed can be traced back to the end of the 4 th century. 156 What is of interest from our point of view is the large amount of non-Greek, probably Lydian, proper names that the document includes. 157 Apart from these, however, no clear examples of Anatolian influence on the Greek text can be found. For the 3 rd and 2 nd centuries BCE, several monolingual Greek inscriptions, both private and public, are attested from different areas in Lydia. 158 However, 198 Valerio Pisaniello & Stella Merlin <?page no="199"?> BCE) and Thyateira (de Hoz 1999: 160, n. 5.25; 276/ 275 BCE). Other texts of the 3 rd and 2 nd centuries can be found, e.g., in de Hoz 1999 and Petzl 2019. 159 On this word, see eDiAna, s.v. kawe- (https: / / www.ediana.gwi.uni-muenchen.de/ dictio nary.php? lemma=212). 160 Cf. Sardis 7,1 51 I; Sardis 52 I-II (126-127 CE); Sardis 53 (1 st -2 nd century CE); Sardis 54 II (1 st -2 nd century CE). 161 See, e.g., Buckler and Robinson 1913: 361-368, Masson 1950, Hawkins 2013: 183-187. For another possible Anatolian loanword surfacing in later Greek inscriptions from Lydia, see Pisaniello 2021b. 162 See eDiAna, s.v. *Héu̯h-eh 2 -/ *HuH-éh 2 - (https: / / www.ediana.gwi.uni-muenchen.de/ dict ionary.php? lemma=1500). 163 To the objection presented by Obrador-Cursach, one could also add that the phonetic value of the Lydian letter usually transliterated as <d> is not entirely assured. According to Oreshko (2019b), e.g., it should be interpreted as / j/ . contrary to the Greek texts found in Lycia, the amount of Anatolian interference that can be observed on the Greek language of these inscriptions is extremely limited, with the Anatolian material being restricted to onomastics and sporadic loanwords, only one of which is attested so far in the epigraphic documentation. Indeed, only the Lydian word kawe- ‘priest(ess)’, 159 also occurring in Hipponax as καύης (fr. 3 Degani), has been recognized in some later honourific inscriptions of the Roman Imperial period, 160 in which it occurs as καυειν (acc.sg.), probably / kawin/ . 161 All of the other assured or alleged Lydian words in Greek - either true loanwords or merely glosses - only occur in the literary and scholarly tradition, with a marked preponderance of forms in the works by Hipponax of Ephesus (6 th century BCE). As far as grammatical interference is concerned, the very few alleged cases that have been suggested so far do not always seem to hold water. Thus, at the phonological level, Oettinger’s (2002) suggestion that psilosis in Ionic and Aeolic could depend on the loss of the initial “laryngeal” in Lydian can no longer be accepted, because Oettinger (2021) himself has shown that the Lydian outcome of an initial “laryngeal” was k-, reflected as κor γin Greek (cf., e.g., PIE *h 2 éu̯h 2 eh 2 - ‘grandfather’ > Lyd. Kuka- > Gr. Γύγης). 162 As for morphology, Dardano’s (2011) hypothesis that the innovative Greek relational suffix -ίδᾱς/ -ίδης, also employed to build patronymics, could derive from the Lydian suffix -da- (< *-ida- < PIE *-iyo-) forming relational adjectives (cf., e.g., Lyd. sfarda- ‘Sardian’ < *sfar-ida-) involves a number of problems at different levels, thoroughly discussed by Obrador-Cursach (2021: 61-66), 163 who also has shown that an explanation in terms of inner-Greek development is available and is probably preferred. Perhaps a more plausible example of morphological interference could be the genitive in -δος in Greek adaptations of Lydian personal names found in an Ancient Greek and late Anatolian languages 199 <?page no="200"?> 164 Such an alternation should probably be explained in the framework of the so-called “rhotacism” that affected Luwian, Lydian, and Pamphylian in the 1 st Millennium BCE, on which see Sasseville 2021b, with references. 165 See Gérard 2005: 84. 166 See, e.g., Högemann and Oettinger 2018: 74-80. 167 See Gérard 2005: 105, with references, for the explanation of the form. 168 See Melchert 2014b: 71, with references. 169 See especially Rojas 2019: 35-43. Cf. also Pisaniello 2021b. inscription from Ephesus (Ephesos 572; cf., e.g., Ἀταδος, Καρουδος, Σαπλαδος, etc.). As suggested by Kearns (1994), if one considers the alternation between δ and λ occasionally found in Greek adaptation of Lydian names (cf., e.g., Ἀλυάττης/ Ἀδυάττης), 164 such an ending could reflect a Hellenization of the Ly‐ dian suffix -libuilding relational adjectives that replaced the genitive singular (cf., e.g., pakiwališ ‘of Pakiwas’ in the ex. 15 above). 165 However, an inner-Greek explanation also remains available, as suggested by Obrador-Cursach (2021: 65 fn. 28). Little can be said for the opposite path of interference, i.e., from Greek to Lydian. Lydian texts include some divine names surely borrowed from Greek (Lyd. Lew-/ Lef- < Gr. Δεύς [dialectal] and Lyd. Lamẽtru- < Gr. Δημήτηρ), which, however, can only attest for relevant cultural interactions between Greeks and Lydians, which were mutual, if Gr. Βάκχος and Ἄρτεμις were actually derived from Lyd. Pakiwaand Artimu-. 166 Outside onomastics, no Greek lexical material has been identified in Lydian inscriptions so far. At the textual level, it is possible, as per Melchert (2014b: 71), that Lydian “talking inscriptions” on seals and objects such as manelim, ‘I am of Manes’ (LW 56), 167 reflected a Greek model (cf. Leomandros’ inscription above), while the possible Greek influences on the metrical structure of the Lydian poetic inscriptions are more uncertain. 168 As is the case with Lycian, Lydian also gradually gives way to Greek and disappears from epigraphic usage, its only survival in the region in later periods being, as far as we can see, personal names and very sporadic loanwords - e.g., καυειν ‘priest(ess)’, a technical term - in the local variety of Greek. Nevertheless, the memory of the Lydian past remained in the region even in later periods, as witnessed by the re-erection of monuments, statues, and stelae with Lydian inscriptions in Roman Sardis between the 2 nd and 4 th centuries CE, as well as the possible recovery of local cults, which have convincingly been explained as a reaction to the destruction of pagan holy places that had accompanied the process of Christianization of the region. 169 200 Valerio Pisaniello & Stella Merlin <?page no="201"?> 170 Maeonians are mentioned in, e.g., Il. 2.864-866; 5.43. According to some ancient traditions, Homer himself comes from Maeonia, a tradition that continued until the 19 th century (e.g., colui per lo cui verso | il meonio cantor non è più solo ‘who for whose verse the Maeonian cantor is no longer alone’ in G. Leopardi, Canti, 2.21-22, in reference to Dante who stands beside Homer). 171 This is by no means a rigid scheme: Pindarus, e.g., mentioned both Lycia and Lydia. 172 See also Melchert 2008: 154. However, Oreshko’s (2019b) hypothesis that the Lydian letter <d> represented / j/ would change this scenario. 3.2. The Lydians and Lydian language in literary and scholarly tradition The case of Lydia is very different from that of Lycia: first of all, direct and indirect evidence are inversely balanced. If for Lycian we have a significant number of inscriptions (although it is still a language of fragmentary attestation) together with some bilinguals able to illuminate the relations between Lycian and other languages, in particular Greek, for Lydian we have an even smaller number of texts, some almost incomprehensible. Therefore any “external” information is most valuable in understanding the language. The Lydian glosses, part of this portion of the evidence, are in fact far more numerous than the Lycian ones. Nonetheless special caution must be exerted, as we shall see, when faced with the term ‘Lydia’. To begin with, the name Lydia is not present in the Homeric poems, in which the name Maeonia is found instead. 170 In contrast, the reference to the name of Lydia (Λυδία) is well attested in the Greek lyric poets, including Sappho, Alcaeus etc. thus showing a sort of complementary distribution of mentioning Lycia and Lydia in the epic and lyric traditions. 171 According to Herodotus (1.7) the name of Λυδοί has been given to the people, previously known as Μαίονες, by an eponym king named Lydos, son of Atys. Van den Hout (2003) has attempted an Anatolian reading of the Greek word Μαίονες, proposing an etymology that may link it to the Lydian toponym Maddunnašša attested in Hittite sources: the Lydian basis *madunwould be derived from *may-un, showing the Lydian change *y > d, in turn from *mai-won (with the same suffix corresponding to the Luwian -wann(i)-). This would be a possible model form for the Greek ethnonym Μαίονες. 172 Gusmani’s Lydisches Wörterbuch includes a list of 46 glosses (Gusmani 1964: 271-278), for which a connection to the Lydian language has been proposed by ancient scholarship. Of these, 24 are regarded as assured Lydian glosses, 5 are falsely attributed to the Lydian language by the ancient sources, and 17 are signaled by a question mark indicating the possibility that a given word is a Lydian gloss, although the text is highly corrupted and philologically uncertain. Ancient Greek and late Anatolian languages 201 <?page no="202"?> 173 See Gusmani 1964: 275-277, with details on sources and further references. 174 See Gusmani 1964: 273. 175 See, e.g., Chantraine 1933 and Hawkins 2010, with references. 176 †κοαλδδεῖν·Λυδοὶ τὸν βασιλέα. For a detailed discussion of the philological and inter‐ pretative problems, see Merlin, Pisaniello, and Rizza forthcoming-b, with references. 177 This is probably true even when his text appears as metalinguistic at first sight, such as in the fr. 2 Degani: Ἑρμῆ κυνάγχα, Μηιονιστὶ Κανδαῦλα, ‘Hermes dogthrottler, in Maeonian Candaula.’ On this topic, see See Merlin, Pisaniello, and Rizza forthcoming-a. Among the assured Lydian glosses, Gusmani reported, e.g., λαβρυς ‘double axe’, λαιλας ‘tyrant’, μωλαξ ‘wine’, τεγουν ‘stealer’. 173 However, it should be stressed that even the value of these glosses is not completely equal: we have, for example, an Anatolian counterpart of τεγουν, visible in the Hittite verbal root taya- ‘to steal’, hence the Greek form could easily be explained by the Byzantine reading of <γ> as the approximant [j]. On the contrary, there is no attested Lydian word that could have been the model for μῶλαξ (Hsch. μ 2030): of course, this is not to deny that μῶλαξ is a true gloss, but to record that its reliability cannot be verified. The word ζακυνθιδες is among the pseudo-glosses: this word, supposed to be the Lydian name for turnip or, according a different tradition, round gourd, is in fact derived from the toponym Ζακύνθος. 174 However, the basic word does show indeed a possible “foreignness”, because it is formed with the suffix -νθος, for which, since Kretschmer (1896), an Anatolian origin, or an Anatolian parallel, has been invoked. 175 One of the most corrupted gloss is κοαλδδειν (Hsch. κ 3169): in addition to some difficulties in the transmission of the text, for the interpretation of the gloss two have been the Lydian candidates proposed by scholars, namely qaλm(λ)u- ‘king’ (Gr. πάλμυς), and qλd-n-, a proper name of a god or a man. 176 The definitive answer is not self-evident and cannot be easily retrieved from the testimony of Hesychius. Hawkins (2013: 155-193), in a monograph devoted to the language of Hipponax is its entirety, analyzed a dozen of possibly Lydian words detectable in his iambic compositions, e.g., ἀρφύτνον ‘discus’, καύης ‘priest’, πάλμυς ‘king’. Hipponax of Ephesus (6 th century BCE) probably represents the most frequent reference in the Hesychian Lydian glosses: this means that the Lexicon of Hesychius comments first and foremost on the Hipponax’ text and, because the only manuscript we have for this work is layered, it cannot be considered as the faithful mirror of the Lydian language. Diversely, Hipponax lived in a time coeval with the Lydian attestations and may in fact report some linguistic forms in use by the Lydian language, although his intent is probably never metalinguistic, but always literary. 177 202 Valerio Pisaniello & Stella Merlin <?page no="203"?> 178 For further details, see Merlin 2020: 490-492; 501-503. To further complicate the picture of the analysis of the Lydian glosses, the very meaning of ‘Lydian’ in the scholarly tradition is not obvious, as mentioned at the beginning. A complementary observation comes from the adverb λυδιστί which is in fact attested very rarely with the meaning of ‘in the Lydian language’ but more often with a more generic ‘in the Lydian manner or style’. Moreover, the reference to Lydia found in some glosses seems primarily to demand a cultural and geographic reading, rather than an exact linguistic one: in other terms, when Hesychius reports that, e.g., βάκκαρις ‘unguent from asarum’ (β 107) and καρύκη (κ 915) ‘spicy elaborated sauce’ are, respectively, a Lydian perfume and a Lydian food, the reference is first and foremost to the product and not to the word that denotes it. This does not preclude the possibility that such products also concealed a Lydian name, but we have no certainty of this. 178 The general impression is that the reference to Lydia, very different from that to Lycia, appears in the Greek sources as a sort of hyperonym, with the general and collective meaning of Near East, highly characterized by those features of richness, opulence, and sophistication with which Greeks, from the Persian wars onward, used to represent in a unified way the variegated world of the Near Eastern people and civilizations. 4. Conclusion To conclude this overview on the relationships between Ancient Greek and late Anatolian languages, limited here to Lycian and Lydian, a summary of the most relevant issues can be sketched. In Lycia, the use of Greek probably started with native Greek professional poets who composed celebrative poems for the local rulers, to become then widespread and expand its scope to include the private sphere, as shown, e.g., by the large number of funerary inscriptions. In these texts, one can see a local variety of Greek with several Lycian interferences involving different levels of analysis. Bilingual documents show interesting contact phenomena: although some of them were perhaps confined to single documents and should probably be explained as occasional translation strategies, others also occur in monolingual Greek inscriptions from Lycia, and can possibly be regarded as structural features of the local variety of Greek. What we face in Lycia is a play of mutual interference between Lycian and Greek, which were actors of a dynamic dialogue without unidirectional relations. The linguistic output of this interaction was almost never the creation Ancient Greek and late Anatolian languages 203 <?page no="204"?> of agrammatical forms, but rather the promotion of fully grammatical structures - although rare and marginal - that could better match the ones of the other language. At the end of this process, however, Lycian gave way to Greek, at least as far as the epigraphic use is concerned: what remains of this Anatolian language on Lycian territory was mostly onomastics and a few borrowings. In the late scholarly tradition, the Lycian language is very little represented, being essentially limited to a small number of glosses - not always explicit and not necessarily reliable - mostly explaining toponyms in Stephanus’ Ethniká. As for Lydia, Ancient Greek was employed in both the public and the private spheres from the very beginning. Differently from Lycia, the interaction between Greek and Lydian is far more limited: only two very short bilingual documents are found, and contact phenomena are essentially restricted to a few borrowings and onomastics; no assured examples of grammatical interference can be detected. On the other hand, one can see a major lexical interference, although often for parodistic purposes, in a local Greek author, Hipponax of Ephesus, who is responsible for the bulk of the Lydian glosses in the later scholarly tradition, especially Hesychius. The analysis of such glosses reveals that “Lydia” also was a generic designa‐ tion for “Near East”, representing, in both the literary and scholarly tradition, a set of features involving fashion, cosmetics, and music. Our perspective in this paper has essentially been limited to linguistic data, which needs to be implemented with those from other disciplines, such as his‐ tory and archaeology, and should also be supplemented with a comprehensive analysis of the use of Greek in the entire Anatolian environment, which was an interconnected, borderless world. Bibliographical references Adiego, I.-X. (1993) “Licio epewẽtlm͂mẽi”, Aula Orientalis 11: 139-145. Adiego, I.-X. (2007a) “Greek and Lycian”, in Christidis, A.-F. (ed.), 763-767. Adiego, I.-X. (2007b) The Carian language, Leiden & Boston, Brill. Adiego, I.-X. (2014) “Las inscripciones plurilingües en Asia Menor: hacia una clasificación tipológica y un análisis funcional”, in Eck, W. & P. Funke (eds.) Öffentlichkeit - Monument - Text. XIV Congressus Internationalis Epigraphiae Graecae et Latinae. 27. - 31. Augusti MMXII. Akten, Berlin & Boston, de Gruyter, 231-269. Adiego, I.-X. (2015) “Lycian nasalized preterites revisited”, Indogermanische Forschungen 120: 1-30. 204 Valerio Pisaniello & Stella Merlin <?page no="205"?> Adiego, I.-X. (2020) “Adaptación griega de nombres licios, adaptación licia de nombres griegos: aspectos fonológicos”, in Ruiz Darasse, C. (ed.) Comment s’écrit l’autre? Sources épigraphiques et papyrologiques dans le monde méditerranéen antique, Bordeaux, Ausonius, 43-60. Altheeb, D. (2015) “Funerary sculptures in the tomb of Abedrapsas at Frikya”, Zeitschrift für Orient-Archäologie 8: 236-249. Álvarez-Pedrosa, J.A.; M.C. Benvenuto; F. Pompeo (2017) (eds.) Del Indo al Egeo. Relaciones culturales y lingüísticas en el interior del Imperio aqueménida, Madrid, Escolar y Mayo. Beckman, G.; T. Bryce; E. Cline (2011) The Ahhiyawa texts, Atlanta, Society of Biblical Literature. Benvenuto, M.C. & F. Pompeo (2017) “Linguistic representations of identity in the Achaemenid world. Case studies”, in Álvarez-Pedrosa, J.A.; M.C. Benvenuto; F. Pompeo (eds.), 17-41. Blomqvist, J. (1982) “Translation Greek in the trilingual inscription of Xanthus”, Opuscula Atheniensia 14/ 2: 11-20. Börker, Ch. & R. Merkelbach (1979) Die Inschriften von Ephesos. Teil II. Nr.-101-599 (Repertorium), Bonn, Habelt. Bousquet, J. (1975) “Arbinas, fils de Gergis, dynaste de Xanthos”, Comptes rendus des séances de l’Académie des Inscriptions et Belles-Lettres 119: 138-150. Bousquet, J. (1986) “Una nouvelle inscription trilingue à Xanthos? ”, Revue Archéologique 1: 101-106. Briant, P. (2003) “Histoire et archéologie d’un texte. La Lettre de Darius à Gadatas entre Perses, Grecs et Romains”, in Giorgieri, M.; M. Salvini; M.-C. Trémouille; P. Vannicelli (eds.), 107-144. Brixhe, C. (2007) “Greek Translations of Lycian”, in Christidis, A.-F. (ed.), 924-934. Bryce, T. (1986) The Lycians - Volume I. The Lycians in literary and epigraphic sources, Copenhagen, Museum Tusculanum. Buckler, W. (1924) Sardis. Volume VI. Lydian inscriptions. Part II. A collection of the texts in Lydian script found at Sardis and elsewhere, Leiden, Brill. Buckler, W. & D. Robinson (1912) “Greek inscriptions from Sardes I”, American Journal of Archaeology 16/ 1: 11-82. Buckler, W. & D. Robinson (1913) “Greek inscriptions from Sardes III”, American Journal of Archaeology 17/ 3: 353-370. Buckler, W. & D. Robinson (1914) “Greek inscriptions from Sardes IV”, American Journal of Archaeology 18/ 1: 35-74. Buckler, W. & D. Robinson (1932) Sardis. Volume VII. Greek and Latin inscriptions. Part I, Leiden, Brill. Canali De Rossi, F. (2004) Iscrizioni dello Estremo Oriente Greco. Un repertorio, Bonn, Habelt. Ancient Greek and late Anatolian languages 205 <?page no="206"?> Carruba, O. (1999) “Ar/ w / w azuma”, Kadmos 38: 50-58. Cassio, A.C. (2019) “Metamorfosi della lingua epica tra oriente e occidente. Da Omero alle laminette orfiche e alla celebrazione poetica dei dinasti della Licia”, in Willi, A. & P. Derron (eds.) Formes et fonctions des langues littéraires en Grèce ancienne, Vandœuvres, Fondation Hardt, 7-41. Cau, N. (2003) “Nuovi antroponimi indigeni nelle iscrizioni greche della Licia di età ellenistico-romana”, Studi Ellenistici 15: 297-340. Cau, N. (2005) “Nuovi antroponimi indigeni nelle iscrizioni greche della Licia di età ellenistico-romana”, II, Studi Ellenistici 16: 377-421. Cau, N. (2006) “Onomastica licia”, Studi Ellenistici 18: 345-376. Chantraine, P. (1933) La formation des noms en grec ancien, Paris, Klincksieck. Childs, W. (1979) “The authorship of the inscribed pillar of Xanthos”, Anatolian Studies 29: 97-102. Christiansen, B. (2019) “Editions of Lycian inscriptions not included in Melchert’s Corpus from 2001”, in Adiego, I.-X.; J.V. García Trabazo; M. Vernet; B. Obrador-Cursach; E. Martínez Rodríguez (eds.) Luwic dialects and Anatolian. Inheritance and diffusion, Barcelona, Edicions de la Universitat de Barcelona, 65-134. Christiansen, B. (2020) “Eine neue lykisch-griechische Bilingue aus Tlos: eine Dedikation oder Ehreninschrift der Polis von Tlos? ”, in Zimmermann, M. (ed.) Das Xanthostal Lykiens in archaisch-klassischer Zeit. Eine archäologisch-historische Bestandsaufnahme, Göttingen, Vandenhoeck & Ruprecht, 262-272. Christiansen, B. (2021) “Palaeographic dating of Lycian inscriptions. A critical review of former studies and a new approach”, Hungarian Assyriological Review 2: 27-63. Christidis, A.-F. (2007) (ed.) A history of Ancient Greek. From the beginnings to Late Antiquity, Cambridge, Cambridge University Press. Colvin, S. (2004) “Names in Hellenistic and Roman Lycia”, in Colvin, S. (ed.) The Greco- Roman East. Politics, culture, society, Cambridge, Cambridge University Press, 44-84. Cook, B. (1987) Greek inscriptions, Berkeley & Los Angeles, University of California Press. Cousin, G. & G. Deschamps (1889) “Lettre de Darius fils d’Hystaspes”, Bulletin de Correspondance Hellénique 13: 529-542. Cuny, A. (1913) “Questions gréco-orientales”, Revue des Études Anciennes 15: 399-404. Dardano, P. (2011) “I patronimici in -ίδᾱς del greco antico tra conservazione e innova‐ zione”, Res Antiquae 8: 41-62. Dardano, P. (2015) “Le iscrizioni bilingui licio-greche nel loro contesto socio-storico: tipi e funzioni a confronto”, in Marotta, G. & F. Rovai (eds.) Ancient languages between variation and norm, Pisa, ETS, 207-226. Daues, A. (2009) “Form und Funktion - die Wortstellung in den lykischen Grabins‐ chriften”, in Rieken, E. & P. Widmer (eds.) Pragmatische Kategorien. Form, Funktion 206 Valerio Pisaniello & Stella Merlin <?page no="207"?> und Diachronie. (Akten der Arbeitstagung der Indogermanischen Gesellschaft vom 24. bis 26. September 2007 in Marburg), Wiesbaden, Reichert, 53-63. Debord, P. (1982) Aspects sociaux et économiques de la vie religieuse dans l’Anatolie gréco-romaine, Leiden, Brill. Descat, R. (1986) “Mnésimachos, Hérodote et le système tributaire achéménide”, Revue des Études Anciennes 87: 97-112. Domingo Gygax, M. (2001) Untersuchungen zu den lykischen Gemeinwesen in klassischer und hellenistischer Zeit, Bonn, Habelt. Dündar, E. & F. Réveilhac (2021) “A pithos with Lycian inscription from Patara”, Kadmos 60: 157-172. Dupont-Sommer, A. (1979) “Deux fragments araméens recueillis dans le Létôon de Xan‐ thos”, in Fouilles de Xanthos. Tome VI. La stèle trilingue du Létôon, Paris, Klincksieck, 170-175. EDHIL = Kloekhorst, A. (2008) Etymological dictionary of the Hittite inherited lexicon, Leiden & Boston, Brill. Egetmeyer, M. (2022) “Woher kam der Mann von Ahhiya? ”, Kadmos 61: 1-36. Facella, M. (2018) “Epigramma della ‘Stele di Xanthos’”, Axon 2: 91-98. FGE = Page, D. (1982) Further Greek epigrams. Epigrams before A.D. 50 from the Greek anthology and other sources, not included in Hellenistic Epigrams or The Garland of Philip, Cambridge, Cambridge University Press. Fischer, R. (2010) Die Aḫḫijawa-Frage. Mit einer kommentierten Bibliographie, Wiesbaden, Harrassowitz. Forrer, E. (1924) “Vorhomerische Griechen in den Keilschrifttexten von Boghazköi”, Mitteilungen der Deutschen Orient-Gesellschaft zu Berlin 63, 1-22. Friedrich, J. (1932) Kleinasiatische Sprachdenkmäler, Berlin, de Gruyter. Gérard, R. (2005) Phonétique et morphologie de la langue lydienne, Louvain-la-Neuve, Peeters. Giannakis, G.K. et al. (2014) (ed.) Encyclopedia of Ancient Greek language and linguistics, Leiden & Boston, Brill. Gildersleeve, B. (1911) Syntax of Classical Greek from Homer to Demosthenes. Second Part. The syntax of the simple sentence continued embracing the doctrine of the article, with the cooperation of Ch. Miller, New York; Cincinnati; Chicago, American Book Company. Giorgieri, M.; M. Salvini; M.-C. Trémouille; P. Vannicelli (2003) (eds.) Licia e Lidia prima dell’ellenizzazione. Atti del Convegno internazionale. Roma, 11-12 ottobre 1999, Roma, Consiglio Nazionale delle Ricerche. Giusfredi, F. (2021-2022) “On the Aegean-Anatolian historical and linguistic interface in the Final Bronze Age”, News from the Lands of the Hittites 5-6: 1-15. Gusmani, R. (1964) Lydisches Wörterbuch. Mit grammatischer Skizze und Inschriftensamm‐ lung, Heidelberg, Winter (Erg. 1. 1980; Erg. 2. 1982). Ancient Greek and late Anatolian languages 207 <?page no="208"?> Gusmani, R. (2007) “Continuità, fratture e processi di osmosi nel panorama linguistico dell’Asia Minore del I millennio a.C.”, in Urso, G. (ed.) Tra Oriente e Occidente. Indigeni, Greci e Romani in Asia Minore. Atti del convegno internazionale. Cividale del Friuli, 28-30 settembre 2006, Pisa, ETS, 11-21. Hajjar, Y. (1990) “Divinités oraculaires et rites divinatoires en Syrie et en Phénicie à l’époque gréco-romaine”, in Haase, W. (ed.) Aufstieg und Niedergang der römischen Welt (ANRW). Geschichte und Kultur Roms im Spiegel der neueren Forschung. Teil II: Principat.-Band-18: Religion (Heidentum: Die religiösen Verhältnisse in den Provinzen [Forts.]). 4. Teilband, Berlin & Boston, de Gruyter, 2236-2320. Hajnal, I. (1995) Der lykische Vokalismus. Methode und Erkenntnisse der vergleichenden anatolischen Sprachwissenschaft, angewandt auf das Vokalsystem einer Kleincorpusspr‐ ache, Graz, Leykam. Hansen, O. (1986) “The purported letter of Darius to Gadates”, Rheinisches Museum für Philologie 129/ 1: 95-96. Hansen, P. (1989) Carmina Epigraphica Graeca Saeculi IV a. Chr. n. (CEG 2), Berlin & New York, de Gruyter. Hawkins, J. D. (1998) “Tarkasnawa King of Mira ‘Tarkondemos’, Bogazköy sealings and Karabel”, Anatolian studies 48: 1-31. Hawkins, S. (2010) “Greek and the languages of Asia Minor to the classical period”, in Bakker, E.J. (ed.) A companion to the Ancient Greek language, Chichester, Wiley-Black‐ well, 213-227. Hawkins, S. (2013) Studies in the language of Hipponax, Bremen, Hempen. Herrenschmidt, C. (1986) “Une lecture iranisante du poème de Symmachos dedié à Arbinas, dynaste de Xanthos”, Revue des Études Anciennes 87-: 125-135. Högemann, P. & N. Oettinger (2018) Lydien. Ein altanatolischer Staat zwischen Griechen‐ land und dem Vorderen Orient, Berlin & Boston, de Gruyter. Hout, M. van den (1949) “Studies in Early Greek letter-writing II”, Mnemosyne 2/ 2: 138-153. Hout, Th. van den (2003) “Maeonien und Maddunnašša: zur Frühgeschichte des Lydi‐ schen”, in Giorgieri, M.; M. Salvini; M.-C. Trémouille; P. Vannicelli (eds.), 301-310. Houwink ten Cate, Ph. (1961) The Luwian population groups of Lycia and Cilicia Aspera during the Hellenistic period, Leiden, Brill. Hoz, M. P. de (1999) Die lydischen Kulte im Lichte der griechischen Inschriften, Bonn, Habelt. Imbert, J. (1894) “Les termes de parenté dans les inscriptions lyciennes”, Mémoires de la Société de Linguistique de Paris 8: 449-472. Jalabert, L. & R. Mouterde (1955) Inscriptions grecques et latines de la Syrie. Tome IV: Laodicée. Apamène. N os 1243-1997. Chronologie des inscriptions datée des Tomes I-IV, Paris, Geuthner. 208 Valerio Pisaniello & Stella Merlin <?page no="209"?> Jasanoff, J. (2010) “Lycian statti ‘stands’”, in Klinger, J.; E. Rieken; Ch. Rüster (eds.) Investigationes Anatolicae. Gedenkschrift für Erich Neu, Wiesbaden, Harrassowitz, 143-152. Kalinka, E. (1901) Tituli Lyciae lingua Lycia conscripti, Wien, Österreichische Akademie der Wissenschaften. Kearns, J.M. (1994) “A Greek genitive from Lydia”, Glotta 72: 5-14. Keen, A. (1998) Dynastic Lycia. A political history of the Lycians and their relations with foreign powers. C. 545-362 B.C., Leiden; Boston; Köln, Brill. Keen, A. & M.H. Hansen (2004) “Lykia”, in Hansen, M.H. & Th.H. Nielsen (eds.) An inventory of archaic and classical poleis, Oxford, Oxford University Press, 1138-1143. Kelder, J. (2004-2005) “Mycenaeans in Western Anatolia”, Talanta 36-37: 49-88. Kelder, J. (2012) “Ahhiyawa and the world of the Great Kings. A re-evaluation of Mycenaean political structures, Talanta 44: 41-52. Klinkott, H. (2023) “The satrapies of the Persian Empire in Asia Minor. Lydia, Caria, Lycia, Phrygia, and Cappadocia”, in Radner, K.; N. Moeller; D.T. Potts (eds.), 592-648. Kloekhorst, A. (2011) “The opening formula of Lycian funerary inscriptions: mẽti vs. mẽne”, Journal of Near Eastern Studies 70: 13-23. Kretschmer, P. (1896) Einleitung in die Geschichte der griechischen Sprache, Göttingen, Vandenhoeck & Ruprecht. Laroche, E. (1974) “Les épitaphes lyciennes”, in Fouilles de Xanthos V, Paris, Klincksieck, 123-149. Lombardi, P. (2010) “La ‘colpa’ di Gadatas. Osservazioni lessicali sulla ‘Lettera di Dario a Gadatas’”, Incidenza dell’Antico 8: 163-195. Macedo, J.M. (2021) “‘Nobody else’: word order in Greek funerary inscriptions from Asia Minor and in Lycian”, Glotta 97: 158-177. Martínez Rodríguez, E. (2021) “Milyan dialectal divergence and its traces in onomastics”, Kadmos 60: 137-156. Masson, O. (1950) “Lydien kaveś (καύης)”, Jahrbuch für kleinasiatische Forschung 1: 182-188. Meisterhans, K. (1900) Grammatik der attischen Inschriften, Berlin, Weidmannsche Buch‐ handlung. Melchert, C. (2001a) Lycian Corpus, available online: https: / / linguistics.ucla.edu/ people/ Melchert/ lyciancorpus.pdf. Melchert, C. (2001b) Lydian Corpus, available online: https: / / linguistics.ucla.edu/ people/ Melchert/ lydiancorpus.pdf. Melchert, C. (2003) “The dialectal position of Lydian and Lycian within Anatolian”, in M. Giorgieri; M. Salvini; M.-C. Trémouille; P. Vannicelli (eds.), 265-272. Melchert, C. (2004) A dictionary of the Lycian language, Ann Arbor & New York, Beech Stave Press. Ancient Greek and late Anatolian languages 209 <?page no="210"?> Melchert, C. (2008) “Greek mólybdos as a loanword from Lydian”, in Collins, B.J.; M.R. Bachvarova; I.C. Rutherford (eds.) Anatolian interfaces. Hittites, Greeks and their neighbours. Proceedings of an international conference on cross-cultural interaction, September 17-19, 2004, Emory University, Atlanta, GA, Oxford, Oxbow Books, 153-157. Melchert, C. (2013) “Naming practices in secondand first-millennium Western Anatolia”, in Parker, R. (ed.) Personal names in Ancient Anatolia, Oxford, Oxford University Press, 31-49. Melchert, C. (2014a) “Greek and Lycian”, in Giannakis, G.K. et al. (ed.), vol. 2, 67-70. Melchert, C. (2014b) “Greek and Lydian”, in Giannakis, G.K. et al. (ed.), vol. 2, 70-71. Melchert, C. (2018) “The trilingual inscription of the Létôon. Lycian version”, available online on Achemenet (http: / / www.achemenet.com/ pdf/ lycian/ lyletoon.pdf). Melchert, C. (2021) “Bilingual texts in first-millennium Anatolia”, in Payne, A.; Š. Velhartická; J. Wintjes (eds.), 349-378. Merlin, S. (2020) “‘Pre-Greek’ between theories and linguistic data. Examples from the Anatolian area”, in Repanšek, L.; H. Bichlmeier; V. Sadovski (eds.) Vácāmsi miśrā krṇavāmahai. Proceedings of the international conference of the Society for Indo-Euro‐ pean Studies and IWoBA XII, Ljubljana 4-7 June 2019, Hamburg, Baar, 487-506. Merlin, S. (2022) “Remarques sur la détermination du nom entre lycien et grec d’Asie Mineure”, Nouvelles Assyriologiques Brèves et Utilitaires 67: 153-155. Merlin, S. & V. Pisaniello (2019) “Linguistic strategies in filiation formulas: data from Lycian-Greek bilingual texts”, Incontri Linguistici 42: 89-108. Merlin, S. & V. Pisaniello (2020) “On the interchange between l and r in Lycian and the case of Pinara”, in M.C. Benvenuto; J.E. Braarvig; F. Pompeo; A.V. Rossi; V. Sadovski (eds.) Word formation, grammar and lexicology. Contrastive, multilingual and comparative-historical perspectives, Hamburg, Baar, 183-198. Merlin, S.; V. Pisaniello; A. Rizza (forthcoming-a) “‘Restsprachen’ in ancient Anatolia: di‐ rect and indirect sources, transmission and reconstruction”, to appear in a forthcoming volume. Merlin, S.; V. Pisaniello; A. Rizza (forthcoming-b) “Esichio come fonte per lo studio di forme anatoliche in greco”, forthcoming in the Proceedings of the XLIV Convegno della Società Italiana di Glottologia, Linguistica e filologia tra oriente e occidente. Napoli 24-26 ottobre 2019. Miller, Ch. (1916) “Note on the use of the article before the genitive of the father’s name in Greek papyri”, The American Journal of Philology 37/ 3: 341-348. Mitchell, E. (1984) The Laconian dialect, Ph.D. Dissertation, University of Edinburgh. Molina Valero, C. (2004) “Usos onomásticos no semíticos en la versión aramea de la inscriptión trilingüe de Janto”, Interlingüística 15: 1011-1019. Molina Valero, C. (2007) “Contactos griego-licio: las inscripciones bilingües”, in Justel, J.J.; B.E. Solans; J.P. Vita; J.Á. Zamora (eds.) Las aguas primigenias. El Próximo Oriente 210 Valerio Pisaniello & Stella Merlin <?page no="211"?> Antiguo como fuente de civilización. Actas del iv Congreso Español de Antiguo Oriente Próximo (Zaragoza, 17 a 21 de Octubre de 2006), Zaragoza, Instituto de Estudios Islámicos y del Oriente Próximo, 127-141. Molina Valero, C. (2010) “Las glosas licias en fuentes griegas”, in Cortés, F. Gabaudan & J.V. Méndez Dosuna (eds.) Dic mihi, Musa, virum. Homenaje al profesor Antonio López Eire, Salamanca, Ediciones Universidad de Salamanca, 459-464. Mørkholm, O. & G. Neumann (1978) Die lykischen Münzlegenden, Göttingen, Vanden‐ hoeck & Ruprecht. Mouton, A. & I. Yakubovich (2021) “Where did one speak luwili? Geographic and linguistic diversity of Luwian cuneiform texts”, Journal of Language Relationship 19: 25-53. Müseler, W. (2016) Lykische Münzen in europäischen Privatsammlungen, İstanbul, Arkeo‐ loji ve Sanat Yayınları. Neumann, G. (1974) “Beiträge zum Lykischen V”, Die Sprache 20: 109-114. Neumann, G. (1979) Neufunde lykischer Inschriften seit 1901, Wien, Verlag der Österrei‐ chischen Akademie der Wissenschaften. Neumann, G. (2007) Glossar des Lykischen. Überarbeitet und zum Druck gebracht von Johann Tischler, Wiesbaden, Harrassowitz. Niemeier, W.D. (1998) “The Mycenaeans in Western Anatolia and the problem of the origins of the Sea Peoples”, in Gitin, S.; A. Mazar; E. Stern (eds.) Mediterranean peoples in transition. Thirteenth to early tenth centuries B.C.E., Jerusalem, Israel Exploration Society, 17-65. Niemeier, W.D. (2005) “Minoans, Mycenaeans, Hittites and Ionians in Western Asia Minor: new excavations in Bronze Age Miletus-Millawanda”, in Villing, A. (ed.) The Greeks in the East, London,-British Museum, 1-36. Obrador-Cursach, B. (2021) “Hypotheses of interference between Greek and the lan‐ guages of Ancient Anatolia: the case of patronymics”, Journal of Language Relationship 19/ 1: 54-70. Oettinger, N. (1981) “Probleme phraseologischer Interferenzen zwischen orientalischen und klassischen Sprachen”, Glotta 59: 1-12. Oettinger, N. (2002) “Die griechische Psilose als Kontaktphänomen”, Münchener Studien zur Sprachwissenschaft 62, pp. 95-101. Oettinger, N. (2021) “Language contact between Lydian and Greek or the origin of Lydian k”, in Bianconi, M. (ed.) Linguistic and cultural interactions between Greece and Anatolia. In search of the Golden Fleece, Leiden & Boston, Brill. Oreshko, R. (2019a) “Luwian word for ‘place, plot of land’ and Lycian alaha-, aladehal(i), *aladehxxaand miñti-”, Münchener Studien zur Sprachwissenschaft 73/ 1: 81-120. Oreshko, R. (2019b) “Phonetic value of Lydian letter <d> revisited and development of PIE dentals in Lydian”, Wekwos 4: 191-262. Ancient Greek and late Anatolian languages 211 <?page no="212"?> Oreshko, R. (2020) “Ethnic groups and language contact in Lycia (I): the ‘Maritime Interface’”, Journal of Language Relationship 18/ 1: 13-40. Oreshko, R. (2021) “Observations on the Xanthos trilingual: syntactic structure of TL 44a, 41‒55 and the Lycian terminology of art and war”, Hungarian Assyriological Review 2: 95-144. Payne, A. (2023) “The Kingdom of Lydia”, in Radner, K.; N. Moeller; D.T. Potts (eds.), 174-230. Payne, A. & D. Sasseville (2016) “Die lydische Athene: eine neue Edition von LW 40”, Historische Sprachforschung 129, 66-82. Payne, A.; Š. Velhartická; J. Wintjes (2021) (eds.) Beyond all boundaries. Anatolia in the first millennium BC, Leuven; Paris; Bristol, Peeters. Petzl, G. (2019) Sardis: Greek and Latin inscriptions. Part II: finds from 1958 to 2017, Cambridge & London, Harvard University Press. Pisaniello, V. (2020) “Glossenkeil and indentation on Hittite tablets”, Altorientalische Forschungen 47: 123-142. Pisaniello, V. (2021a) “The Lycian toponym Κάδρεμα and the Anatolian wheat”, in Giusfredi, F.; Zs. Simon; E. Martínez Rodríguez (eds.) Studies in the languages and language contact in Pre-Hellenistic Anatolia, Barcelona, Universitat de Barcelona, 67-80. Pisaniello, V. (2021b) “The epithet Τιαμου of the Moon-god in Lydia”, Kadmos 60: 117-136. Pisaniello, V. & S. Merlin (forthcoming) “Lycian Ἐρευάτης in Stephanus of Byzantium”. Podestà, S. (2017) “L’origine contesa delle città licie occidentali: dal mito alla colonia”, Erga-Logoi 5: 61-80. Podestà, S. (2019) “Il mondo di Ecateo di Mileto: la Licia nei frammenti della Periegesi”, Geographia Antiqua 28: 71-86. Podestà, S. (2022) Lykiaka. Frammenti, Besançon, Presses universitaires de Franche-Comté. Quadrio, T. (2012) “Il glottonimo ‘miliaco’”, in Orioles, V. (ed.) Per Roberto Gusmani. Tomo 2. Linguistica storica e teorica. Studi in ricordo, Udine, Forum, 355-378. Radner, K.; N. Moeller; D.T. Potts (2023) (eds.) The Oxford history of the Ancient Near East. Volume V: the Age of Persia, New York, Oxford University Press. Rasmussen, J.E. (1974) Haeretica Indogermanica. A selection of Indo-European and Pre-Indo-European studies, København, Munksgaard. Réveilhac, F. (2018) Contact linguistique et emprunts onomastiques entre grec et lycien: apports à la phonétique et à la morphologie, Ph.D. Dissertation, Sorbonne Université. Réveilhac, F. (2021) “Le statut du lycien et du grec dans les inscriptions de Lycie”, in L’Anatolie: de l’époque archaïque à Byzance, Besançon, Presses universitaires de Franche-Comté, 67-96. 212 Valerio Pisaniello & Stella Merlin <?page no="213"?> Rieken, E. (2017) “The dialectology of Anatolian”, in Klein, J.; B. Joseph; M. Fritz (eds.) Handbook of comparative and historical Indo-European linguistics, vol. 1, Berlin & Boston, de Gruyter, 298-308. Rix, E. (2016) Tombs and territories: the epigraphic culture of Lycia, c. 450-197 BC, Ph.D. Dissertation, University of Oxford. Robert, L. (1975) “Une nouvelle inscription grecque de Sardes: règlement de l’autorité perse relatif à un culte de Zeus”, Comptes rendus des séances de l’Académie des Inscriptions et Belles-Lettres 119/ 2-: 306-330. Robinson, D. (1923) “Two new epitaphs from Sardis”, in Buckler, W.H. & W.M. Calder (eds.) Anatolian studies presented to Sir William Mitchell Ramsay, Manchester, Univer‐ sity Press, 341-353. Rojas, F. (2019) The pasts of Roman Anatolia. Interpreters, traces, horizons, Cambridge, Cambridge University Press. Rutherford, I. (2002) “Interference or translationese? Some patterns in Lycian-Greek bilingualism”, in Adams, J.N.; M. Janse; S. Swain (eds.) Bilingualism in ancient society, Oxford, Oxford University Press, 197-219. Sasseville, D. (2021a) Anatolian verbal stem formation. Luwian, Lycian and Lydian, Leiden & Boston, Brill. Sasseville, D. (2021b) “Rhotacism in 1 st -Millennium BC Anatolia. Comparative Luwian and Lydian phonology”, in Payne, A.; Š. Velhartická; J. Wintjes (eds.), 636-650. Savalli, I. (1988) “L’idéologie dynastique des poèmes grecs de Xanthos”, L’Antiquité Classique 57: 102-123. Schmitt, R. (1987) “Artaxerxes”, in Yarshater, E. (ed.) Encyclopaedia Iranica. Volume II, London & New York, Routledge & Kegan, 654-655. Schmitt, R. (2014) Wörterbuch der altpersischen Königsinschriften, Wiesbaden, Reichert. Schuler, Ch. (2019) “Lycian, Persian, Greek, Roman: chronological layers and structural developments in the onomastics of Lycia”, in Parker, R. (ed.) Changing names. Tradition and innovation in Ancient Greek onomastics, Oxford, Oxford University Press, 195-216. Schürr, D. (1999) “Gräko-lykisch πιατρα”, Die Sprache 41: 24-38. Schürr, D. (2007) “Formen der Akkulturation in Lykien: Griechisch-lykische Sprachbe‐ ziehungen”, in Schuler, Ch. (ed.) Griechische Epigraphik in Lykien. Eine Zwischenbilanz. Akten des int. Kolloquiums. München, 24.-26. Februar 2005, Wien, Österreichische Akademie der Wissenschaften, 27-40. Schürr, D. (2010) “Eine lykische Fluchformel mit Zukunft”, Epigraphica Antolica 43: 149-158. Schürr, D. (2014) “Ist lykisch sttaein Erb- oder ein Lehnwort? ”, Historische Sprachfor‐ schung 127: 141-149. Ancient Greek and late Anatolian languages 213 <?page no="214"?> Sciancalepore, E. (2017) “La relazione sintattica determinante-determinato in alcune epigrafi greche dell’Iran achemenide”, in Álvarez-Pedrosa, J.A.; M.C. Benvenuto; F. Pompeo (eds.), 223-234. Serangeli, M. (2015) “Heth. Maliya, lyk. Malija und griech. Athena”, in Zinko, Ch. & M. Zinko (eds.) 1. Grazer Symposium zur indogermanischen Altertumskunde. “Der antike Mensch im Spannungsfeld zwischen Ritual und Magie”. Graz, 14. - 15. November 2013, Graz, Leykam, 376-388. Sokolowski, F. (1955) Lois sacrées de l’Asie Mineure, Paris, De Boccard. Starke, F. (1990) Untersuchung zur Stammbildung des keilschrift-luwischen Nomens, Wiesbaden, Harrassowitz. Steiner, G. (2007) “The case of Wiluša and Ahhiyawa”, Bibliotheca Orientalis 64: 590-611. Storme, B. (2014) “The beginning of the Lycian and Greek versions of the Létôon trilingual: syntax and semantics”, Historische Sprachforschung 127: 125-140. Thonemann, P. (2009) “Lycia, Athens and Amorges”, in Ma, J.; R. Parker; N. Papazarkadas (eds.) Interpreting the Athenian empire, London, Duckworth, 167-194. Tritsch, F.J. (1976) “The Lycian bilingual in Stoichedon from Korydalla”, Kadmos 15/ 2: 158-167. Vernet, M. (2023) “Who was ArKKazuma? Some considerations that point to a Carian origin of this divinity”, News from the Lands of the Hittites 7: 235-248. Vismara, N. (1989a) Monetazione arcaica della Lycia. I. Il dinasta Wekhssere I, Milano, Ennerre. Vismara, N. (1989b) Monetazione arcaica della Lycia. II. La collezione Winsemann Falghera, Milano, Ennerre. Vismara, N. (1996) Monetazione arcaica della Lycia. III. Le prime emissioni del Wedri. Le serie di Θa, delle città di Θib-nuw- (Simena), di Zagaba (Lagbe), di Zẽmuri (Limyra), di Prl (Aperlai) e le emissioni federali di Ite e di Te, Milano, Ennerre. Wachter, R. (2001) Non-Attic Greek vase inscriptions, Oxford, Oxford University Press. Wankel, H. (1979) Die Inschriften von Ephesos. Teil Ia. Nr.-1-47 (Texte), Bonn, Habelt. Warbinek, L. & F. Giusfredi (2023) “Maliya, Malija, Malis, Athena. From Kizzuwatna to the Aegean: borrowings, translations, or syncretisms? ”, Asia Anteriore Antica 5: 139-151. Willi, A. (2004) “Old Persian in Athens revisited (Ar. Ach. 100)”, Mnemosyne 57: 657-681. Wörrle, M. (1991) “Epigraphische Forschungen zur Geschichte Lykiens IV. Drei griechi‐ sche Inschriften aus Limyra”, Chiron 21: 203-239. Wörrle, M. (1995) “Epigraphische Forschungen zur Geschichte Lykiens V. Die griechi‐ schen Inschriften der Nekropolen von Limyra”, Chiron 25: 387-418. Wörrle, M. (2012) “Die griechischen Sepulkralinschriften von Limyra”, in Borchhardt, J. & A. Pekridou-Gorecki (eds.) Limyra. Studien zu Kunst und Epigraphik in den Nekropolen der Antike, Wien, Phoibos, 411-457. 214 Valerio Pisaniello & Stella Merlin <?page no="215"?> Yakubovich, I. (2022) “The place of Lydian in the Anatolian family through the lens of recent research”, Journal of Language Relationship 20/ 3: 191-221. Zgusta, L. (1964) Kleinasiatische Personennamen, Prag, Verlag der Tschechoslowakischen Akademie der Wissenschaften. Ancient Greek and late Anatolian languages 215 <?page no="217"?> The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? An analysis of the aspect use in the Epic-Ionic -σκ-iteratives in the Odyssey Filip De Decker (Marie Skłodowska Curie Actions---European Fellowship - Individual Fellow & Università degli Studi di Verona) Abstract: In this article I investigate a specific case of possible lan‐ guage contact, namely whether the (alleged) imperfective meaning of the Epic-Ionic -σκ-iteratives can be considered a Graeco-Anatolian isogloss. To assess this, I discuss the 111 instances of the Odyssey to check whether the differences between aorist and imperfect, valid for the non-iterative forms, also apply to the iterative forms. I first briefly describe the problem (without delving into the origin of the suffix) and discuss aspect in (Homeric) Greek. Then I start my analysis of the forms by determining my corpus (providing facts and figures and explaining why I leave out certain forms) and then proceed to the actual aspectual analysis, using the traditional approach (sc. the distinction punctual vs. durative) and combine this with Vendler’s verbal classifications applied to active verbs in Homer by Napoli and the distinction perfective - imperfective as defined by Comrie, Bertinetto, Bache and Bybee, Perkins and Pagliuca. This study will show that the “epic-Ionic iteratives” are subject to the same aspectual distinctions as the “normal” verb forms. They do not have imperfective meaning and cannot therefore be used to support language contact be‐ tween Greek and Anatolian (if the forms were indeed imperfective, the question would be why they are also found in the aorist and not simply in the imperfect). Thus, the use of the Epic-Ionic -σκ-iterative suffix can better receive an internal explanation than to be related to language contact. Of <?page no="218"?> 1 This research was conducted at the Università degli Studi di Verona during the project Particles in Greek and Hittite as Expression of Mood and Modality (PaGHEMMo) with Professor Paola Cotticelli-Kurras as Principal Investigator and Dr. Filip De Decker as researcher, which has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie Grant Agree‐ ment Number 101018097. This article has greatly benefitted from the feedback by Paola Cotticelli-Kurras, Francesca Cotugno, Anna Dentella, Federico Giusfredi, Elena Martínez-Rodríguez, Stella Merlin-Defanti, Valerio Pisaniello, Alfredo Rizza, Jelena Živojinović (all Università degli Studi di Verona), and from the observations by the audience of the Linguistisches Kolloquium at the Ludwig Maximilians Universität München, and the audience and participants of the Workshop Languages and Cultures in Contact in the Ancient Mediterranean at the Università degli Studi di Verona and the International Conference Delbrück Colloquium on Historical and Comparative Syntax of Indo-European, held at the Università degli Studi di Verona. Finally, I warmly thank the editor, Professor Carlotta Viti, for her feedback and observations and for allowing me to submit this article to the volume she has edited. All shortcomings, inconsistencies and/ or errors are mine and mine alone. 2 Pisani (1959: 176-177) was the first to state that the suffix was an isogloss of the languages in Asia Minor (Ionic Greek, Hittite and Armenian), while Puhvel (1991) was the first to explicitly state that the suffix was an isogloss between Hittite and Ionic Greek and that it proved the existence of a Sprachbund between these two languages (but he left out Armenian, contrary to Pisani). course, in order to correctly investigate language contact, we should also be able to determine the instances in which contact did not occur. 1 Keywords: tense and aspect, Graeco-Anatolian language contact, “epic-Ionic iteratives”, epic Greek. 1 Introduction: The suffix as part of the Graeco-Anatolian Sprachbund One of the peculiarities of Greek epic diction is the existence of so-called iterative forms. These are forms in which a suffix -σκis added to the imperfect or aorist stem, while it is missing in all other tenses and moods. It has been argued that these Greek forms, often called “epic-Ionic iteratives”, have imperfective meaning and that the creation of this type of forms was the result of contact with the Hittite forms in -šk- (Pisani 1959: 176-177, Puhvel 1991). 2 Puhvel argued that the Greek forms with the suffix were imperfective and that the use of the suffix to mark this imperfectivity was a trait that Ionic Greek had adopted from Hittite, which used the suffix -škto mark imperfectivity as well (Puhvel 1991: 13-20, Miller (2014: 133-134, 182, 321, 352, 356); see however the discussion in Bianconi 2015: 154-160 who doubted the 218 Filip De Decker <?page no="219"?> 3 Bianconi (2015: 160) “Insomma, nonostante in queste forme, la cui classificazione specifica risulta ancora non del tutto chiara, paia abbastanza presente una qualche influenza areale, al momento esse non possono essere considerate prova decisiva per una area linguistica greco-anatolica.”(underlining is mine).This isogloss is also quoted in Watkins (2001: 58), but he himself did not argue for such a Sprachbund). For a skeptical view on the Sprachbund see Hajnal (2014, 2018). This feature was not listed as a possible isogloss in Romagno (2015). 4 In line with the requirements by the editors, I use the Latin form of the Greek names (contrary to my usual practice). 5 I refer to De Decker (2022: 67-73, ftc: §2.2) for a more detailed discussion and more references on tense and aspect in general. validity of this isogloss 3 ). As the use of this suffix is not found outside epic and Ionic Greek (except when imitating Homer) and the suffix is only productive in Homer (and epic) and Herodotus, 4 Puhvel argued that its use was a feature of a Sprachbund in Anatolia in the 2 nd and 1 st Millennium BC: from the Anatolian languages, the suffix would have been expanded towards the Greek of Ionia. In what follows, for reasons of space I will only address the alleged imperfective meaning of the suffix (and not the issue of the Sprachbund, for which see most recently Cotticelli-Kurras 2021). The imperfective meaning seems established for Hittite (Bechtel 1936, Puhvel 1991, Cambi 2007, Daues 2009, pointing at the differences between Greek and Hittite, and recently Pisaniello 2020: 34-40, but see the doubts by Melchert & Hoffner 2008: 317-329 and Mattiola & Inglese 2018, who follow Dressler 1968 in explaining the suffix as conveying verbale Pluralität). However, if it can be shown that the forms do not have imperfective meaning in epic Greek, it becomes less probable that this feature was the result of such a Sprachbund. As the origin of the suffix in the “epic-Ionic iteratives” has been discussed elsewhere (cf. De Decker 2023: 10-16, where I have also put forward some more general objections against the assumption that the suffix use in the “epic-Ionic iteratives” was an isogloss between Ionic Greek and Hittite), I do not repeat those observations here and only focus on the iterative meaning. 2 Previous scholarship on [(Homeric) Greek] aspect 2.1 Perfective versus imperfective aspect The number of studies on tense and aspect is almost infinite and a detailed discussion is impossible. 5 For perfective and imperfective I will use the following definitions by Comrie (which in my opinion are still valid today): “perfectivity indicates the view of a situation as a single whole, without distinction of the various separate phases that make up that situation; while the imperfective pays The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 219 <?page no="220"?> 6 I refer to De Decker (2022: 67-73, ftc: §2.2) for a more detailed discussion and more references on tense and aspect in general. 7 For a detailed analysis with more references and an analysis of the theories by Hollenbaugh (2018, 2021), I refer to De Decker (2022: 74-833, ftc: §2.2). essential attention to the internal structure of the situation” (Comrie 1976: 16, accepted in Bybee & Perkins & Pagliuca 1994: 125-126, Bhat 1999: 45-49, 58); “(perfective) will typically denote a single event, seen as an unanalysed whole, with a well-defined result or end-state, located in the past. More often than not the event will be punctual, or at least, it will be seen as a single transition from one state to its opposite, the duration of which can be disregarded” (Dahl 1985: 78, Bybee & Dahl 1989: 55, Bybee & Perkins & Pagliuca 1994: 54, Bache 1997: 304, Bhat 1999: 45-49, 58). Actions described by the imperfective can be (in the description by Comrie 1976: 24-40, followed in Deo 2006: 48-98, Dahl 2010: 69-73) habitual and continuous (progressive or non-progressive), or (as Bybee & Perkins & Pagliuca 1994: 125-127 state in more in detail) “progressive (ongoing at reference time, with dynamic verbs), continuous (progressive, ongoing at reference time, with static and dynamic verbs), habitual (customarily repeated on different occasions), iterative (repeated on a particular occasion), frequentative (habitual, ongoing and frequent) and continuative (ongoing, with the intent of the agent to keep the action going)”. I want to add that the distinction between frequentative and iterative is also an important one as the latter refers to “events repeated on the same occasion”, while the former refers to “events repeated on different occasions” (Bybee & Perkins & Pagliuca 1994: 127, Bhat 1999: 53). In his analysis of the Italian indicative, Bertinetto (1986 passim, but especially the schema on page 119, see also Pisaniello 2020: 15-29, especially 22, 2022: slide 4) distinguished between imperfective with abituale, continuo and progressivo and perfective with compiuto and aoristico -> ingressivo (I leave the terms in Italian, as an accurate one-to-one translation is not entirely possible). For my purposes here, I follow these analyses and say that imperfectivity refers to habitual, ongoing and repeated actions (I add a category “repeated”, because not all repeated actions are ongoing and/ or habitual, but I will not distinguish so sharply between continuous, continuative and progressive, as Bertinetto and Bybee & Perkins & Pagliuca would do). 2.2 Homeric aspect A detailed study on (Homeric) Greek aspect cannot be performed here, 6 but I will follow my earlier observations and start from a combination of what I called the “Vendler-Napoli”-hypothesis and the analyses by Comrie and others, quoted in the previous subchapter. 7 Generally speaking, the distinction between 220 Filip De Decker <?page no="221"?> 8 Pott’s comparison of point to line was quoted, almost verbatim, in Aken (1861: 5, Dauer - Vollendung - Punkt and especially 15 Der Aorist bezeichnet häufig den Punkt, auf den sich die Bed. (sc. Bedeutung) der ganzen Handlung concentrirt, die ἀκμή, Spitze derselben, 1865: 11), Mutzbauer (1893: 11) and Delbrück (1897: 230, quoting Mutzbauer), without mentioning Pott’s name, however. See also Bornemann & Risch (1973: 214), also without mentioning Pott’s name. 9 The list of recent treatments of tense and aspect in Greek is obviously much longer. 10 Vendler (1957). Vendler’s schema has since then become the standard starting point for verbal analyses, and in spite of some of the problems that could appear with it (as discussed in Matthews 1990: 57-60, Bertinetto & Squartini 1995 and Depraetere 1995a---the list of criticisms and suggested improvements is much longer, as can be seen in Napoli 2006: 24, footnote 1), this is still a reliable comparative tool (in my opinion). 11 For a more detailed discussion of her theory, see De Decker (ftc: §2.2). aorist and imperfect is described in terms of momentaneous and punctual versus durative (see already Buttmann 1810: 488-490), but it has never been worded as accurately as by Pott (1833: 57): der griech. Aorist verhält sich zum Impf. (und Praes.) wie Punct zur Linie (hence the description “punctual”). He was neither the first nor the last to observe this, however, 8 and nowadays the difference between present and aorist stem is more described in terms of duration and completion or boundedness (as e.g. in Jacquinod 2017: 686, but see already Buttmann 1810: 488-490), or imperfective versus perfective (as e.g. in Van Emde Boas & Huitink 2010: 140-142 and Hettrich 2016: §1.1). 9 In her recent treatment of Homeric aspect, Napoli started from Vendler’s (1957) classification of verbs as being states, activities, accomplishments and achievements, 10 and the description of states as non-dynamic, durative and telic, activities as dynamic, durative and non-telic, accomplishments as dynamic, durative and telic and achievements as dynamic, non-durative and telic (Napoli 2006: 32-47). She applied these classes to the verbs in Homeric Greek and was also able to show that there was a clear correlation between perfectivity (the aorist in this case), telicity and the use of countable, defined and animated direct objects, so that “he killed X” would appear in the aorist while “he killed enemies” would be in the imperfect. To my knowledge, she was one of the first to do so and her analysis has provided important insights, but there is one important observation to make. 11 She considered iterativity as an individual aspectual category as well (2006: 31-32), arguing that iterativity was often derivative and not an inflectional category, marked by reduplication or affixation. Accordingly, a speaker can depict the same situation as semelfactive (i.e. occurring only once) or iterative (i.e. continuously recurring in the same circumstance), adding that if one did not interpret iterativity as an aspectual category, it should be considered part of The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 221 <?page no="222"?> 12 I owe valuable feedback on this point to Valerio Pisaniello (Università degli Studi di Verona). 13 Also García-Ramón (2012) argued that the metre was not the deciding factor. 14 Kühner (1835: 77), Stolpe (1849, especially page 45), Curtius (1852: 29, 1880: 406), Týn (1860), Giacalone-Ramat (1967: 115-116), Wathelet (1973: 401-403), Zerdin (1999: 301, 325-326), Kimball (2014), Allan (2016: 93-94. The iteratives were not included in the analyses of Napoli (2006) nor in Hollenbaugh (2018). Hollenbaugh (2021) treated the iterative forms of aorist and imperfect, and discussed passages in which both forms occurred, but did not explicitly state that an aspectual difference could be found between them. Daues (2009) and Miller (2014: 133-134, 334, 352) do not discuss the aspect of the iterative forms, although she (Daues 2009: 87) stated that the interaction of meaning, suffix and aspect is much more complex as one would expect from her analysis (Dass das Zusammenspiel von Lexembedeutung, Verbalaspekt und Suffix im Detail viel komplexer ist, als im Rahmen dieses Beitrags dargestellt werden kann, wird freilich eingeräumt.). the imperfective aspect. 12 As will be shown here, the iterative forms are subject to aspectual distinctions as well. Moreover, I believe that one should distinguish between frequentative and iterative (as argued above). In what follows, I will apply these findings to address the tense and aspect use in the iterative forms, and I will show that the metre played no significant role at all. 13 It has been noted before that the iterative forms are subject to the same aspectual distinctions as the other verbal forms, 14 and the current analysis will follow closely earlier studies into the matter (especially De Decker ftc.). Before proceeding to the actual analysis, I will first outline how I determined the corpus and discuss some debated issues, and then provide facts and figures. 3 The facts and figures of the iterative forms in the Odyssey 3.1 The forms ἔσκε(ν), ἔφασκε(ν)/ φάσκε(ν) and ἔβασκεν/ βάσκε(ν), and the forms in -ασκε Before we proceed to the actual analysis, we first need to determine the corpus and decide which forms can be considered iteratives and which ones cannot, and which tense certain forms have. As definition for an “epic-Ionic iterative”, I use the commonly accepted one, namely that it involves the aorist or imperfect form in -σκof a verb that does not have this suffix in the other tenses or moods and of which there are no relics of this suffix in other Indo-European languages. As such, the forms ἔσκε and ἔφασκε / φάσκε cannot be considered as “genuine” iteratives, as the former has a connection in Old-Latin escit and the latter has present forms in later Greek, with φάσκω gradually replacing the present forms of φημί. The same applies to the form παρέβασκε. In this 222 Filip De Decker <?page no="223"?> 15 In spite of this, Risch (1974: 277) considered the form to be an iterative after all. 16 Already Buttmann (1830: 384-385) considered the from to be unexplainable and did not understand why κρύψασκε was not attested. 17 West (2000: 278) referred to Chantraine in his apparatus to defend the transmitted reading. 18 It was not addressed in Faesi (1858a: 29, 1862: 182), von Doederlein (1863: 175, 1864: 314), Düntzer (1864: 99, 1866: 253, 1873: 270, 1878: 283), La Roche (1870a, nor in 1870b: 116, 139, 1871: 128), Peppmüller (1876: 289), Monro (1901: 177). In their com‐ mentaries, Ameis (1868c: 23) and Ameis & Hentze did not address the form (1871: 35), but in the Anhang they (Ameis & Hentze 1887: 98) mentioned Fick’s emendation (1885: 340, 379). Russo & Fernández-Galiano & Heubeck (1992: 105) discussed the habitual meaning of the iterative forms in this passage, but did not state anything on the forms themselves. 19 Fick (1885: 340, 379 - Lies entweder κρύπτεσκε (vom imperfect) oder κρύψασκε (vom aorist); κρύπτασκε ist nichts). Kirchhoff (1879: 433, but in 1859: 88 he still read διαρρίπτασκεν). Ameis & Hentze (1887: 98) and Vogrinz (1889: 117) mentioned both options, but did not state which option was the best. 20 Actually, they read (’)ῥίπτεσκε, as according to them every form in epic Greek was augmented, and where the metre did not allow it, the augment was removed by apocope. case we are not dealing with a real epic-Ionic iterative, because the Homeric imperative βάσκε (mostly in the phrase βάσκ’ ἴθι) as well as the Sanskrit form gacchati “s/ he goes” seem to imply that the root *g w em had an -sḱ-formation in PIE already, or at least that the suffix -σκexisted outside the past indicative forms for that root. 15 In principle, one could argue that the Greek form is derived from *g w eh 2 , as the zero grade of that form with the suffix would also yield βάσκ-, but in that case the connection with gacchati has to be abandoned. I do not count ἔσκε(ν), ἔφασκε(ν)/ φάσκε(ν) nor ἔβασκεν/ βάσκε(ν) as iterative forms, and their uses deserve a separate analysis, although there is one instance of ἔφασκε(ν)/ φάσκε(ν) with a clear iterative meaning. Second, some forms, such as κρύπτασκε, ῥίπτασκε and ἰσάσκετο, pose problems as to their exact formal determination. 16 These forms have been explained as an aorist of a verb in -άζω (Curtius 1880: 410-411, Ebeling 1885a: 602, Brugmann 1900: 294, 1902: 274, Chantraine 1953: 323, Kimball 2014). 17 While some commentaries did not address the issue, 18 others considered them to be imperfect forms from verbs in -άω (Monro 1891: 47, Schwyzer 1939: 711, explicitly rejecting Brugmann’s analysis of ἰσάσκετο), to be mixtures of the imperfect κρύπτεσκε and the aorist κρύψασκε (Fick 1885: 340, 379), or they remained agnostic (as Risch 1974: 277). Fick (1885: 340, 379) therefore suggested to change κρύπτασκε into κρύψασκε or κρύπτεσκε, and Kirchhoff (1879: 433) changed διαρρίπτασκεν (Odyssey 19,575) into διαρρίψασκεν. 19 Van Leeuwen and Mendes da Costa (1895: 648, 1898: 92-93), on the other hand, suggested to read ῥίπτεσκε. 20 Forms The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 223 <?page no="224"?> 21 This form was not discussed in the commentaries on this passage (Faesi 1862: 182, Düntzer 1864: 99, Ameis 1868c: 23, Ameis & Hentze 1871: 35, Monro 1901: 177, Russo & Fernández-Galiano & Heubeck 1992: 105). Russo & Fernández-Galiano & Heubeck 1992: 105 stated that the iterative form proved that this was a habit of Odysseus. The Homeric grammars by Vogrinz (1889: 116), Chantraine (1948: 322) mentioned the form, but they did not state that this form was special or unique. Krüger (1853: 83, 94-95) and Monro (1891: 47) did not mention this form in their discussion of the iteratives. 22 The instances are Odyssey 14,220, 17,316, 17,420, 19,76, 19,575. 23 The instances are Odyssey 1,264, 2,104, 2,105, 3,264, 7,260, 8,225, 11,512, 11,515, 11,619, 12,380, 13,350, 15,385, 16,140, 19,149, 19,150, 19,229, 19,574, 22,46, 22,209, 22,315, 22,358, 22,414, 22,427, 23,9, 23,65, 23,353, 24,139, 24,140. 24 The instances are Odyssey 5,331, 5,332, 8,88, 8,89, 8,374, 8,376, 23,94, 23,95. 25 The instances are Odyssey 1,435, 4,627, 5,84, 5,154, 5,158, 5,480, 6,6, 6,95, 7,99, 7,138, 7,171, 8,92, 8,259, 13,194, 14,17, 17,169, 17,211, 17,257, 17,294, 17,331, 18,2, 18,7, 18,325, 18,325, 19,442, 20,3, 20,7, 20,290, 21,41, 22,185, 24,209, 24,212, 24,390. 26 The Apologoi refer to Books 9---12 in which Odysseus explains to the Phaeacians how his men died due to their own stupidity and how he tried to save them, but did not succeed. As he tries to defend his actions, these books are called Apologoi. In those books, there are also “genuine” speeches, e.g. when Odysseus speaks to his men or to the gods. 27 The instances are Odyssey 11,208, 11,586, 11,587, 11,587, 11,592, 11,597, 11,599, 12,241, 12,242. 28 The instances are Odyssey 9,184, 9,188, 9,429, 10,108, 11,240, 11,596, 12,238, 12,330, 12,355. where an -άω present is attested, such as τρωπάσκετο from τρωπάω, are obviously catalogued as imperfects and not as aorists of an -άζω verb. Third, there is the form ἵστασχ’ (Odyssey 19,574), occurring in the same passage as διαρρίπτασκεν. It is the only active and transitive iterative form of the root *steh 2 in Homer and is metrically equivalent to the aorist στήσασχ’ (which is never attested). 21 3.2 Facts and figures on the iteratives in the Odyssey Below I provide the figures of the iterative forms in the Odyssey. Odyssey Aorist Imperfect Overall Speech 5 22 28 23 34 Narrative 8 24 33 25 41 Apologoi 26 9 27 9 28 18 224 Filip De Decker <?page no="225"?> 29 In Book 4, Menelaus was asked why he did not act and avenge the death of his brother Agamemnon. In a very long speech, he tried to defend himself and explain why he could not have done anything to prevent the murder from happening. I consider this to be a special case between speech and narrative, as this cannot be conceived as a simple speech, but contains long narrative descriptions and speeches. 30 Odyssey 4,368. 31 The instances are Odyssey 2,324, 2,331, 4,769, 8,328, 10,37, 13,167, 17,482, 18,72, 18,111a (disputed), 18,400, 20,375, 21,361, 21,396, 21,401, 23,148. 32 The instances are Odyssey 4,772, 13,170, 23,152. 33 The instances are Odyssey 2,324, 2,331, 4,769, 4,772, 5,331, 5,332, 8,88, 8,89, 8,328, 8,376, 10,37, 11,208, 11,586, 11,587, 11,587, 11,592, 11,597, 11,599, 12,241, 12,242,13,167, 13,170, 14,220, 17,316, 17,420,17,482, 18,72, 18,111a (disputed), 18,400, 19,76, 19,575,20,375, 21,361, 21,396, 21,401, 23,94, 23,95,23,148, 23,152. 34 The instances are Odyssey 1,264, 1,435, 2,104, 2,105, 3,264, 4,368, 4,627, 5,84, 5,154, 5,158, 5,480, 7,99, 7,260, 8,92, 9,184, 9,429, 10,108, 11,512, 11,515, 11,596, 12,238, 12,355, 13,194, 13,350, 14,17, 17,257, 17,294, 17,331, 18,325, 18,325, 19,149, 19,150, 19,229, 19,442, 21,41, 22,315, 22,414, 22,427, 23,65, 23,353, 24,139, 24,140. 35 Odyssey 8,374. 36 The instances are Odyssey 6,6, 6,95, 7,138, 7,171, 8,225, 8,259, 9,188, 11,240, 11,619, 12,330, 12,380, 15,385, 16,140, 17,169, 17,211, 18,2, 18,7, 19,574, 20,3, 20,7, 20,290, 22,46, 22,185, 22,209, 22,358, 23,9, 24,209, 24,212, 24,390. Odyssey Aorist Imperfect Overall Menelaus’ speech 29 0 1 30 1 Speech introductions 15 31 0 15 Speech Conclusions 3 32 0 3 Overall 40 71 111 Table 1: the figures in the Odyssey of the iterative forms per tense and per type of text. - Aorist Imperfect Overall Main clauses 39 33 42 34 81 Relative clauses 1 35 29 36 30 Overall 40 71 111 Table 2: the figures in the Odyssey of the iterative forms per tense and per type of clause. A comparison with the data of the Iliad shows that there are no significant differences: The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 225 <?page no="226"?> Aorist Imperfect Overall Speech 7 41 48 Narrative 21 47 68 Speech introductions 12 2 14 Speech conclusions 4 0 4 Total 44 90 134 Table 3: the data of the iterative forms in the Iliad per type of speech. - Aorist Imperfect Total Main clauses 41 44 85 Relative clauses 3 40 43 Causal clauses 0 2 2 Temporal clauses 0 4 4 Table 4: the data of the iterative forms in the Iliad per type of clause. 3.3 Preliminary observations From the tables and the data about the iterative forms in the Odyssey, we note the following: a. among the iterative forms, there are more imperfects than aorist forms; b. the iterative forms used to introduce or conclude direct speech appear only in the aorist; c. the iterative forms in the aorist appear almost exclusively in the main clause; d. the iterative forms used in relative clauses are de facto only attested in the imperfect. As there are less aorists than imperfects, I will start with analysing the former. 226 Filip De Decker <?page no="227"?> 37 The instances are Odyssey 2,324, 2,331, 4,769, 4,772, 8,328, 10,37, 13,167, 13,170, 17,482, 18,72, 18,111a (disputed), 18,400, 20,375, 21,361, 21,396, 21,401, 23,148, 23,152. 38 This specification was first made by Stolpe (1849: 38-39), Týn (1860: 682) and later by Zerdin (1999: 298) and Pagniello (2007), and is also visible in speech introductions of the so-called τις-Reden, see De Decker (2015a: 64-65). Kluge (1911: 35-36) also mentioned this analysis, but he explicitly denied its correctness. As this use occurs especially with undefined subjects such as τις or ἄλλος, I use the term τις-actions, based on the coinage τις-Reden by Fingerle (1939: 283-293) for speech introductions by undefined subjects. 39 An example is Iliad 22,375, where Homer described with two iterative forms how anonymous Greek soldiers voiced their relief that Hektor had died (εἴπεσκεν) and cowardly stabbed the hero’s dead body (οὐτήσασκε). 4 The aorist 4.1 The τις-actions Of the 40 aorist forms, 18 occur in what can best be called τις-actions: 37 the use of the iterative forms to mark not a repeated action by one single person, but the repetition of a single action by a different persons. 38 This approach explains many iterative instances with an undefined subject and is to be preferred over Schulze’s (1892: 63) and Kluge’s (1911: 35-36) explanation that the iterative form was a combination of an unreal or potential indicative merged with a modal particle. As these forms describe a single action that is completed by the actor as soon as the other starts doing it, the aorist is used. These actions are almost a textbook case of a perfective action. This is particularly common in speech introductions and conclusions, and in the Odyssey all instances are in fact attested in either a speech introduction or a speech conclusion, contrary to the Iliad where these forms sometimes described other actions as well. 39 I give an example from an introduction and one from conclusion. (1) ὧδε δέ τις εἴπεσκε δόμων ἔκτοσθεν ἀκούων: (Odyssey 23,148). “And so someone would say, hearing it from outside the palace: ” In this specific speech conclusion, Homer describes how anonymous Ithacans or foreigners passing by Odysseus’ passage might think that the noise and music mean that queen Penelope finally decided to marry and would state that she had not waited until Odysseus had returned. The example means that any by-passer would say this once, not that just one person would pass by and make this statement several times. (2) ὣς ἄρα τις εἴπεσκε, τὰ δ’ οὐκ ἴσαν ὡς ἐτέτυκτο (Odyssey 4,772, 13,170, 23,152). “So one would speak, but they did not know how it had happened.” The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 227 <?page no="228"?> 40 I analysed this in more detail for the speech introductions in De Decker (2022: 176-182, 306-312). This speech conclusion is used when a group of anonymous people finished commenting on an event of which they did not know the complete result. That we are dealing with a large group, is clear from the plural form ἴσαν. 4.2 Other aorist forms Now, I will address another example with iterative forms in the aorist. (3) ὡς δ’ ὅτ’ ὀπωρινὸς Βορέης φορέῃσιν ἀκάνθας ἂμ πεδίον, πυκιναὶ δὲ πρὸς ἀλλήλῃσιν ἔχονται, ὣς τὴν ἂμ πέλαγος ἄνεμοι φέρον ἔνθα καὶ ἔνθα: ἄλλοτε μέν τε Νότος Βορέῃ προβάλεσκε φέρεσθαι, ἄλλοτε δ’ αὖτ’ Εὖρος Ζεφύρῳ εἴξασκε διώκειν. (Odyssey 5, 328-332). “As when the Northern wind in late summer carries thorny plants over the field, densely they are held against each other, so the winds carried it (sc. Odysseus’ ship) over the sea. Now the southern wind would throw it to the northern to be carried, and then the east wind would let it to the west wind to drive it forth.” After Odysseus left Calypso and built a raft to sail home, Poseidon decided to render his journey much more difficult by sending a fierce storm and turning the sea into a violent whirlpool. The verbs describe how Odysseus and his raft were being swept away by the different winds. The iteratives προβάλεσκε and εἴξασκε are in the aorist, because they describe a single passage, and because the action of προβάλεσκε is completed when that of εἴξασκε starts and vice versa. These are repeated but completed actions and therefore the iterative forms are in the aorist. The verb that describes the overall action of the winds and also refers to the entire event, φέρον, is in the imperfect, because it describes a continuous action. The distinction between a non-iterative verb in the plural and singular verb forms in the aorist with an iterative suffix is a construction that is attested elsewhere in Homer as well. The non-iterative plural verb form in the imperfect describes the overall verbal action and has no iterative suffix because it simply describes the action of the group. Instead, the iterative forms are in the aorist and have the iterative suffix because they describe how each individual member of the group performed his action repeatedly. 40 5 The imperfect As has been stated already, the imperfect is much more common among the iterative forms. This is in line with the iterative-habitual meaning of the suffix. 228 Filip De Decker <?page no="229"?> 41 The instances are Odyssey 1,264, 1,435, 7,171, 17,257, 18,325. 42 Odyssey 15,385. In several instances, the imperfect is used to describe a habit, a continuous event (e.g. the verb φιλέεσκε 41 ), an action spreading over a longer period of time (e.g. the verb ναιετάασκε 42 ) or an iterative action (by one person or by many) that reoccurred on multiple occasions and not one that was repeated in one single instance. Two examples below will illustrate that. The most famous passage as illustration for the iterative form in the imperfect as indication for a repeated action is without any doubt the following. (4) ὣς ἔφαθ’, ἡμῖν δ’ αὖτ’ ἐπεπείθετο θυμὸς ἀγήνωρ. ἔνθα καὶ ἠματίη μὲν ὑφαίνεσκεν μέγαν ἱστόν, νύκτας δ’ ἀλλύεσκεν, ἐπὴν δαΐδας παραθεῖτο. ὣς τρίετες μὲν ἔληθε δόλῳ καὶ ἔπειθεν Ἀχαιούς: ἀλλ’ ὅτε τέτρατον ἦλθεν ἔτος καὶ ἐπήλυθον ὧραι, καὶ τότε δή τις ἔειπε γυναικῶν, ἣ σάφα ᾔδη, καὶ τήν γ’ ἀλλύουσαν ἐφεύρομεν ἀγλαὸν ἱστόν. ὣς τὸ μὲν ἐξετέλεσσε καὶ οὐκ ἐθέλουσ’ ὑπ’ ἀνάγκης (Odyssey 2,103-110, cf. 19,148-155 and 24,138-146). “So she spoke and her manly spirit convinced us all. During the day she was weaving the big shroud, but at night she would unravel it, after she had placed her torches besides her. By her trick she misled the Achaeans for three years and made them believe (that she was finishing the shroud). But when the fourth year came and the seasons had passed, someone from the housewives, who knew everything, told us about it and we caught her unravelling the shining shroud. So then she finished it, unwilling and by force.” When pressed by the suitors to choose one of them as her new husband, Penelope answered that she would indeed make her choice, but not before she had finished the shroud in which Laertes was to be buried. During the day, she would weave the cloth but she would unravel it again at night, thus never finishing it. By doing so, she kept the suitors waiting. One day, however, an unfaithful servant noticed what she was doing and informed the suitors. They lay in ambush, surprised her and forced her to finish the shroud. The suitors, and Antinous in particular, related this to Telemachus to counter his complaints about their outrageous behaviour in the palace. In this passage, we have 2 iterative forms, ὑφαίνεσκεν and ἀλλύεσκεν, which could theoretically also have appeared in the aorist, but are in the imperfect because they refer to an action that was repeated for a period of more than three years. The tenses of the other verb forms agree with the traditional aspectual distinctions as The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 229 <?page no="230"?> well. The imperfect and present-stem forms, ἔφαθ’, ἐπεπείθετο, ἔληθε, ἔπειθεν, ἀλλύουσαν and ἐθέλουσ’, refer to ongoing actions and/ or to actions that have a lasting effect (such as “speaking”), whereas the aorist forms, ἦλθεν, ἐπήλυθον, ἔειπε and ἐφεύρομεν, are perfective, as all of them refer to completed actions and some of them also to single actions (ἔειπε and ἐφεύρομεν). The optative παραθεῖτο is also perfective, as bringing a torch is a single action, but given the fact that this action also belongs to the ongoing trickery and deceit, one might argue that a present optative would have been more suited. This is only apparent, however, as in the description the placing of the torch is an action that has to be completed before the two others can start and as the form has perfective meaning, the aorist is expected after all. The imperfect is also used to describe objects, institutions, and peoples that have always existed and or lived in a certain place, or individual persons who have always acted in certain way. One example is the one below. (5) ἦλθε δ’ ἐπὶ πτωχὸς πανδήμιος, ὃς κατὰ ἄστυ πτωχεύεσκ’ Ἰθάκης, μετὰ δ’ ἔπρεπε γαστέρι μάργῃ ἀζηχὲς φαγέμεν καὶ πιέμεν: οὐδέ οἱ ἦν ἲς οὐδὲ βίη, εἶδος δὲ μάλα μέγας ἦν ὁράασθαι. Ἀρναῖος δ’ ὄνομ’ ἔσκε: τὸ γὰρ θέτο πότνια μήτηρ ἐκ γενετῆς: Ἶρον δὲ νέοι κίκλησκον ἅπαντες, οὕνεκ’ ἀπαγγέλλεσκε κιών, ὅτε πού τις ἀνώγοι ὅς ῥ’ ἐλθὼν Ὀδυσῆα διώκετο οἷο δόμοιο, καί μιν νεικείων ἔπεα πτερόεντα προσηύδα (Odyssey 18,1-9). “Then came also a public beggar, who used to beg throughout the city of Ithaca and stood out for his gluttonous stomach to eat and drink continuously. He had no force nor power, but a large posture to look at. His name was Arnaeus. That was the name his queenly mother had given (him) at birth, but all the youngsters called him Irus, because he would go and pass on messages, whenever someone ordered him (to go) somewhere. He came to him and tried to chase Odysseus from his home, scolded him and spoke winged words: ” This passage describes the beggar Irus, states that he walked around in the city asking for food and was called Irus because he would go around and bring messages, whenever he was asked to do. The two iterative forms, πτωχεύεσκ’ and ἀπαγγέλλεσκε, describe Irus’ habits of begging throughout Ithaca and passing on messages. The imperfect is used because these actions were not a single event, but a habit and something Irus had always been doing. The imperfect κίκλησκον (italicised in the fragment) is not an epic-Ionic iterative, as it also has a reduplication and the suffix is attested in the non-imperfect tenses and in the non-indicative moods as well. As was stated above, I consider the form ἔσκε (italicised in the fragment) to be of a special type and not an epic-Ionic iterative. 230 Filip De Decker <?page no="231"?> 6 Imperfect and aorist in contrast Now, I will discuss two passages in which iteratives in the aorist and the imperfect co-occur, but nevertheless show a difference in meaning. (6) ταῦτ’ ἄρ’ ἀοιδὸς ἄειδε περικλυτός: αὐτὰρ Ὀδυσσεὺς πορφύρεον μέγα φᾶρος ἑλὼν χερσὶ στιβαρῇσι κὰκ κεφαλῆς εἴρυσσε, κάλυψε δὲ καλὰ πρόσωπα: αἴδετο γὰρ Φαίηκας ὑπ’ ὀφρύσι δάκρυα λείβων. ἤτοι ὅτε λήξειεν ἀείδων θεῖος ἀοιδός, δάκρυ ὀμορξάμενος κεφαλῆς ἄπο φᾶρος ἕλεσκε καὶ δέπας ἀμφικύπελλον ἑλὼν σπείσασκε θεοῖσιν: αὐτὰρ ὅτ’ ἂψ ἄρχοιτο καὶ ὀτρύνειαν ἀείδειν Φαιήκων οἱ ἄριστοι, ἐπεὶ τέρποντ’ ἐπέεσσιν, ἂψ Ὀδυσεὺς κατὰ κρᾶτα καλυψάμενος γοάασκεν. (Odyssey 8,83-92). “The very famous singer was singing all these stories. But Odysseus took his large purple cloak with his strong hands and put it over his head and hid his beautiful face (with it). He was embarrassed to shed tears before the eyes of the Phaeacians. Whenever the divine singer stopped singing, Odysseus wiped the tears away, took the cloak from his head, took a cup with two goblets and made a libation to the gods. But when he started singing again and the most noble Phaeacians incited him to do so, as they enjoyed his words, Odysseus would hide his head again and continued to cry.” In these lines, Homer describes how Odysseus tried to hide his tears when he was in the palace of the Alcinous and heard the bard Demodocus sing about the fall of Troy and the pivotal role Odysseus played in this. The aorist forms, ἕλεσκε and σπείσασκε, refer to single and completed actions that are repeated on several occasions (with the former being an “achievement-verb” and the latter an “accomplishment-verb”). By contrast, the imperfect γοάασκεν describes the durative process of wailing and lamenting, which is uninterrupted, continues throughout the passages and can be described as an “activity-verb”. The non-iterative verb forms in this passage show a similar aspectual distinction, as the aorists (the participles ἑλών, ὀμορξάμενος, ἑλών and καλυψάμενος, the indicative εἴρυσσε, the injunctive κάλυψε and the optative λήξειεν) refer to completed actions, while the forms in the present stem (the participle λείβων, the injunctives ἄειδε, αἴδετο, and τέρποντ’) describe ongoing or durative actions. The only exceptions are ἄρχοιτο and ὀτρύνειαν, and it is unclear why the former is in the imperfect and the latter in the present. (7) καὶ μὴν Σίσυφον εἰσεῖδον κρατέρ’ ἄλγε’ ἔχοντα λᾶαν βαστάζοντα πελώριον ἀμφοτέρῃσιν. ἤτοι ὁ μὲν σκηριπτόμενος χερσίν τε ποσίν τε λᾶαν ἄνω ὤθεσκε ποτὶ λόφον: ἀλλ’ ὅτε μέλλοι ἄκρον ὑπερβαλέειν, τότ’ ἀποστρέψασκε κραταιΐς: The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 231 <?page no="232"?> αὖτις ἔπειτα πέδονδε κυλίνδετο λᾶας ἀναιδής. αὐτὰρ ὅ γ’ ἂψ ὤσασκε τιταινόμενος, κατὰ δ’ ἱδρὼς ἔρρεεν ἐκ μελέων, κονίη δ’ ἐκ κρατὸς ὀρώρει. (Odyssey 11,593-600). “And I saw Sisyphus having to undergo strong suffering, handling an enormous stone with both hands. And then he pushed the stone upwards to the hill pressing with his hands and feet, but whenever he was about to push it over the hill, a mighty force would it turn around and then the shameless stone rolled back down. But he then started pushing it again and from his limbs the sweat was pouring down, and dust rose out of his head.” In this passage, taken from the description of the eternal punishment of the sinner Sisyphus in the Nekyia, a clear distinction in the aspect use of the iterative forms can be discerned. Sisyphus is condemned to eternally push a rock up a hill, only to see it roll down that same hill, once he has almost reached the top. There are 3 iterative forms in this passage, the aorists ἀποστρέψασκε and ὤσασκε, and the imperfect ὤθεσκε. The use of different tenses is surprising, since we would either expect these forms to be all in the imperfect, given the fact that the passage describes an eternal punishment, or all in the aorist, because Odysseus saw this punishment only once. The aorist ἀποστρέψασκε is relatively straightforward, as it refers to a completed action, namely that of a god to push the rock down the hill. The distinction between ὤθεσκε and ὤσασκε is more complicated, and as the forms are metrically equivalent, the metre cannot have played any role. In my opinion, the first one is an imperfectum de conatu whereas the second one is an inchoative aorist. In the case of the second example, Sisyphus has to start over again, because the rock has rolled from the hill again, whereas in the first instance, Odysseus described the punishment. It is a de conatu because Sisyphus will never succeed in pushing it up until the top. In short, I hope to have shown that the tense use of the iterative forms is determined by aspectual factors and that these forms are not imperfective. In the next section, I will briefly discuss the findings that were established based on the facts and figures. 7 Conclusion Especially since Puhvel (1991), it has been argued that the so-called “epic-Ionic” iteratives in -σκhave imperfective meaning and that the suffix with its imperfective meaning is an isogloss of a Graeco-Anatolian Sprachbund. In order to assess the value of this isogloss, I addressed the aspect of these “epic-Ionic” iteratives in -σκin the Odyssey, investigating whether or not the suffix had indeed imperfective meaning. I first explained what forms I would (not) include and how I would parse them, then provided facts and figures of the iterative 232 Filip De Decker <?page no="233"?> forms and then proceeded to the actual analysis. As there were less aorist than imperfect forms, I started with the aorist. Using the definitions of perfective and imperfective by Comrie, Bertinetto, Bache, Bybee and others, and starting from the aspectual analyses by Napoli (who based her research on Vendler but did not address the iterative forms herself), I was able to show that the distinction between the aorist and the imperfect for the iterative forms was the same as those for the other verb forms in Homer. The higher frequency of imperfects compared to the aorists may be explained by the fact that most of these iterative forms refer to a habit, a durative or continuous action, or an action repeated on several occasions - in these cases the imperfect was preferred. The aorist was used instead to denote a completed action which was repeated on one occasion by many different characters (the so-called τις-actions) or which was repeated by one character on many occasions (e.g. many soldiers speaking to each other on one specific moment). As these verb forms refer to one action that is completed and does not reoccur anymore after this instance, the aorist is used. The imperfect, on the other hand, was used to describe habits, continuous actions, durative events and single actions repeated on many occasions. This was especially clear in the description of the eternal punishments of the sinners in Odyssey 11 and in the description of Penelope’s “shroud-deceit”. In almost all of the passages discussed, the non-iterative verb forms conformed to the traditional aspectual distinctions as well. This also explains why the aorist is almost exclusively used in the main clause, as the subordinate clauses, and especially the relative clauses, describe habits, continuous actions and states in the past and do not usually refer to completed actions, which makes the use of the imperfect more expected. As such, the conclusions and observations made here were similar to those of the Iliad (as had been argued for already in De Decker ftc). Given the fact that these iterative forms do not simply have imperfective meaning in epic Greek, the alleged imperfective meaning of the “epic-Ionic” iteratives in -σκcannot considered to be a Graeco-Anatolian isogloss. 8 Bibliographical references Online resources Thesaurus Linguae Graecae: http: / / stephanus.tlg.uci.edu/ Iris/ inst/ tsearch.jsp Chicago Homer: https: / / homer.library.northwestern.edu/ html/ application.html Perseus: http: / / www.perseus.tufts.edu/ hopper/ collection? collection=Perseus: collection: Greco-Roman The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 233 <?page no="234"?> Other references Aken, A.F. (1861) Die Grundzüge der Lehre von Tempus und Modus im Griechischen: historisch und vergleichend aufgestellt, Rostock, Stiller. Aken, A.F. (1865) Die Hauptdata der griechischen Tempus- und Modulehre, historisch und vergleichend, Berlin, Enslin. Allan, R. (2016) “Tense and Aspect in Greek: two historical developments; Augment and Perfect”, in Runge, S. & Ch. Fresch (eds.) The Greek verb revisited, Bellingham, Lexham, 81-121. Ameis, K.F. (1868a) Ilias. Für den Schulgebrauch erklärt. Erster Band. Erstes Heft. Gesang I-III, Leipzig, Teubner. Ameis, K.F. (1868b) Anhang zu Homers Ilias Schulausgabe. I. Heft. Erläuterungen zu Gesang I-III, Leipzig, Teubner. Ameis, K.F. (1868c) Anhang zu Homers Odyssee Schulausgabe. Erläuterungen zu Gesang XIX - XXIV, Leipzig, Teubner. Ameis, K.F. (1870a) Ilias. Für den Schulgebrauch erklärt. Erster Band. II. Heft. Gesang IV-VI, Leipzig, Teubner. Ameis, K.F. (1870b) Anhang zu Homers Ilias Schulausgabe. II. Heft. Erläuterungen zu Gesang IV-VI, Leipzig, Teubner. Ameis, K.F. & C. Hentze (1871) Homers Odyssee. Für den Schulgebrauch erklärt. Gesang XIX-XXIV, Leipzig, Teubner. Ameis, K.F. & C. Hentze (1887) Anhang zu Homers Ilias Schulausgabe. Erläuterungen zu Gesang VII-IX, Leipzig, Teubner. Ameis, K.F. & C. Hentze (1888a) Ilias. Für den Schulgebrauch erklärt. Gesang X-XII, Leipzig, Teubner. Ameis, K.F. & C. Hentze (1888b) Anhang zu Homers Ilias Schulausgabe. Erläuterungen zu Gesang X-XII, Leipzig, Teubner. Ameis, K.F. & C. Hentze (1906) Ilias. Für den Schulgebrauch erklärt. Gesang XXII-XXIV, Leipzig, Teubner. Bache, C. (1997) The study of Aspect, Tense and Action. Towards a theory of the semantics of grammatical categories, 2 nd revised edition, Frankfurt, Lang. Bechtel, G. (1936) Hittite verbs in -SK-: a study of verbal aspect, Ann Arbor, Edwards. Bertinetto, P.M. (1986) Tempo, Aspetto e Azione nel verbo italiano. Il sistema dell’indicativo, Firenze, Accademia della Crusca. Bertinetto, P.M. & M. Squartini (1995) “An attempt at defining the class of ‘gradual completion’ verbs”, in Bertinetto, P.M.; V. Bianchi; J. Higginbotham; M. Squartini (eds.), vol. 1, 11-26. Bertinetto, P.M.; V. Bianchi; J. Higginbotham; M. Squartini (1995) (eds.) Temporal reference. Aspect and Actionality. Vol.1: Semantic and syntactic perspectives. Vol. 2: Typological perspectives, Torino, Rosenberg & Sellier. 234 Filip De Decker <?page no="235"?> Bhat, D. N. S. (1999) The Prominence of Tense, Aspect, and Mood. Amsterdam, Benjamins. Bianconi, M. (2015) “Contatti Greco-Anatolici e Sprachbund Egeo-Micrasiatico. Stato della ricerca e nuove prospettive”, Archivio Glottologico Italiano 100: 129-178. Bornemann, E. & E. Risch (1973) Griechische Grammatik, Frankfurt, Diesterweg. Bottin, L. (1969) “Studio dell’aumento in Omero”, Studi Micenei ed Egeo-Anatolici 10: 69-145. Brügger, C. (2017) Homer Iliad. The Basel Commentary. Book XXIV, Translated by S. Douglas Olson, Berlin, De Gruyter. Brugmann, K. (1890) Griechische Grammatik, München, Beck. Brugmann, K. (1892) Grundriß der vergleichenden Grammatik der indogermanischen Sprachen. Zweiter Band, Lehre von den Wortformen und ihrem Gebrauch. Zweiter Teil, Erste Lieferung, Strassburg, Trübner. Brugmann, K. (1900) Griechische Grammatik, München, Beck. Brugmann, K. (1902) “Die ionischen Iterativpräterita auf σκον”, Indogermanische For‐ schungen 13: 267-277. Brugmann, K. (1904) Kurze vergleichende Grammatik der indogermanischen Sprachen, Strassburg, Trübner. Brugmann, K. (1916) Grundriß der vergleichenden Grammatik der indogermanischen Sprachen, II.3, Strassburg, Trübner. Buttmann, Ph. (1810) Griechische Grammatik, fünfte, vermehrte und verbesserte Ausgabe, Berlin, Myliussische Buchhandlung. Buttmann, Ph. (1830) Ausführliche griechische Sprachlehre, Erster Band, Berlin, Dümmler. Bybee, J. & Ö. Dahl (1989) “The creation of Tense and Aspect systems in the languages of the world”, Studies in Language 13: 51-103. Bybee, J.; R. Perkins; W. Pagliuca (1994) The evolution of grammar. Tense, Aspect and Modality in the languages of the world, Chicago, Chicago University Press. Cambi, V. (2007) Tempo e Aspetto in ittito con particolare riferimento al suffisso -ske/ a-, Alessandria, Edizioni dell’Orso. Chantraine, P. (1948) Grammaire homérique, Paris, Klincksieck. Chantraine, P. (1953) Grammaire homérique. Tome II: Syntaxe, Paris, Klincksieck. Comrie, B. (1976) Aspect, Cambridge, Cambridge University Press. Cotticelli-Kurras, P. (2021) “On the concept of an Anatolian-Greek language area”, Old World: Journal of Ancient Africa and Eurasia 1: 1-19. Cotticelli-Kurras, P. & F. Giusfredi (2018) “Ancient Anatolian languages and cultures in contact: some methodological observations”, Journal of Language Relationship 16: 172-193. Curtius, G. (1846) Die Bildung der Tempora und Modi im Griechischen und Lateinischen sprachvergleichend dargestellt, Berlin, Besser. The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 235 <?page no="236"?> Curtius, G. (1852) “Vermischte Etymologien”, Kuhns Zeitschrift für vergleichende Sprach‐ forschung 1: 25-36. Curtius, G. (1863) Erläuterungen zu meiner griechischen Schulgrammatik, Prag, Tempsky. Curtius, G. (1864) Griechische Schulgrammatik. Zweite Auflage, Prag, Tempsky. Curtius, G. (1873) Das Verbum der griechischen Sprache, seinem Baue nach dargestellt, Erster Band, Leipzig, Hirzel. Curtius, G. (1880) Das Verbum der griechischen Sprache, seinem Baue nach dargestellt, Zweiter Band, Leipzig, Hirzel. Dahl, E. (2010) Time, Tense and Aspect in Early Vedic grammar, Leiden, Brill. Dahl, Ö. (1985) Tense and Aspect systems, Oxford, Blackwell. Dahl, Ö. (1995) “Areal tendencies in tense-aspect systems”, in Bertinetto, P.M.; V. Bianchi; Ö. Dahl; M. Squartini (eds.) Temporal reference. Aspect and Actionality. Vol.2: Typological Perspectives, Torino, Rosenberg & Sellier, 11-28. Daues, A. (2009) “Zum Funktionsbereich des Suffixes *-sḱe/ o im Junghethitischen und Homerischen”, in Lühr, R. & S. Ziegler (eds.) Protolanguage and prehistory. Akten der XII. Fachtagung der Indogermanischen Gesellschaft. Krakau, 11. bis 15. Oktober 2004, Wiesbaden, Reichert, 82-99. De Decker, F. (2015a) “The augment in Homer, with special attention to speech intro‐ ductions and conclusions”, in De Decker, F.; K. Jakob; M. Klumm; M. Kunzmann; I. Lindbüchl; T. Stoll (eds) JournaLIPP 4 Proceedings of the 21 st LIPP Symposium, 53-71, online publication: https: / / lipp.ub.lmu.de/ index.php/ lipp/ article/ view/ 4841/ 2723. De Decker, F. (2015b) A morphosyntactic analysis of speech introductions and conclusions in Homer, PhD Thesis LMU München, online publication: https : / / edoc.ub.uni-muenc hen.de/ 17995/ De Decker, F. (2016) “A contrastive analysis of the Homeric and Hesiodic augment, with special focus on Hesiod”, International Journal of Diachronic Linguistics and Linguistic Reconstruction 13: 33-128. De Decker, F. (2022) Studies in Homeric speech introductions and conclusions: Tense, Aspect, Augment, Mood, Modality, and Modal Particle, Alessandria, Edizioni dell’Orso. De Decker, F. (2023) “A contrastive study of the co-occurrence of the augmented, unaug‐ mented and iterative forms in -σκin Iliad 24”, (Kuhns) Zeitschrift für vergleichende Sprachforschung 134: 9-95. De Decker, F. (forthcoming) “An analysis of the aspect use in the Epic-Ionic -σκ-Iteratives in the Iliad”, to appear in International Journal of Diachronic Linguistics and Linguistic Reconstruction. Delbrück, B. (1879) Syntaktische Forschungen IV. Die Grundlagen der griechischen Syntax, Halle, Buchhandlung des Waisenhauses. Delbrück, B. (1893) Vergleichende Syntax der indogermanischen Sprachen, I, Strassburg, Trübner. 236 Filip De Decker <?page no="237"?> Delbrück, B. (1897) Vergleichende Syntax der indogermanischen Sprachen, II, Strassburg, Trübner. Delbrück, B. (1900) Vergleichende Syntax der indogermanischen Sprachen, III, Strassburg, Trübner. Deo, A. (2006) Tense and Aspect in Indo-Aryan languages: variation and diachrony, PhD Thesis Stanford. Depraetere, I. (1995a) “The effect of temporal adverbials on (a)telicity and (un)bounded‐ ness”, in Bertinetto, P.M.; V. Bianchi; J. Higginbotham; M. Squartini (eds.), vol. 1, 43-54. Depraetere, I. (1995b) “On the necessity of distinguishing between (Un)Boundedness and (A)Telicity”, Linguistics and Philosophy 18: 1-19. Dressler, W. (1968) Studien zur verbalen Pluralität: Iterativum, Distributivum, Durativum, Intensivum in der allgemeinen Grammatik, im Lateinischen und Hethitischen, Wien, Böhlau. Düntzer, H. (1864) Homers Odyssee. Erklärende Schulausgabe. Drittes Heft. Buch XVII-- -XXIV. Nebst Register, Paderborn, Schöningh. Düntzer, H. (1866) Homers Ilias. Erklärende Schulausgabe. Erstes Heft. Buch I-VIII, Pader‐ born, Schöningh. Düntzer, H. (1873) Homers Ilias. Erklärende Schulausgabe. Erstes Heft. Buch I-VIII, Pader‐ born, Schöningh. Düntzer, H. (1878) Homers Ilias. Erklärende Schulausgabe. Drittes Heft. Buch XVII-XXIV. Nebst Register, Zweite, neu bearbeitete, Auflage, Paderborn, Schöningh. Ebeling, H. (1885a) Lexicon Homericum. I. Α-Ξ, Leipzig, Teubner. Ebeling, H. (1885b) Lexicon Homericum. II. Ο-Ω, Leipzig, Teubner. Faesi, J.U. (1858a) Homers Iliade, erklärt von J. U. Faesi. Erster Band, Berlin, Weidmann. Faesi, J.U. (1858b) Homers Iliade, erklärt von J. U. Faesi. Zweiter Band, Berlin, Weidmann. Faesi, J.U. (1860) Homers Odyssee, erklärt von J. U. Faesi. Erster Band. Berlin, Weidmann. Faesi, J.U. (1862) Homers Odyssee, erklärt von J. U. Faesi. Zweiter Band, Berlin: Weidmann. Fick, A. (1885) Die Homerische Ilias nach ihrer entstehung betrachtet und in der ursprün‐ glichen sprachform wiederhergestellt, Göttingen, Vandenhoeck & Ruprecht. Fingerle, A. (1939) Typik der homerischen Reden, München, unpublished PhD thesis. García-Ramón, J.-L. (2012) “Aspect and Mood in Indo-European reconstruction”, in Melchert, C. (ed) The Indo-European verb, Bremen, Hempen, 73-85. Giacalone-Ramat, A. (1967) “La funzione del suffiso -ΣΚnel sistema verbale greco”, Archivio Glottologico Italiano 52: 105-123. Hajnal, I. (2014) “Die griechisch-anatolischen Sprachkontakte zur Bronzezeit - Sprach‐ bund oder loser Sprachkontakt? ” Linguarum Varietas 3. 105-116. Hajnal, I. (2018) “Graeco-Anatolian contacts in the Mycenaean period”, in Fritz, M.; B. Joseph; J. Klein (eds.) Handbook of comparative and historical Indo-European linguistics III, Berlin, De Gruyter, 2037-2055. The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 237 <?page no="238"?> Herbig, G. (1896) “Aktionsart und Zeitstufe”, Indogermanische Forschungen 6: 157-270. Hettrich, H. (2016) “Überlegungen zum Gebrauch der verbalen Aspektstämme bei Homer”, International Journal of Diachronic Linguistics and Linguistic Reconstruction 13: 183-204. Hollenbaugh, I. (2018) “Aspects of the Indo-European aorist and imperfect. Re-evaluating the evidence of the R̥gveda and Homer and its implications for PIE”, Indo-European Linguistics 6: 1-68. Hollenbaugh, I. (2021) Tense and Aspect in Indo-European: a usage-based approach to the verbal systems of the Rigveda and Homer, PhD Dissertation, UCLA. Jacquinod, B. (2017) “The syntax of Greek”, in M. Fritz; B. Joseph; J. Klein (eds.) Handbook of Comparative and Historical Indo-European Linguistics I, Berlin, De Gruyter, 682-695. Kimball, S. (1980) “A Homeric Note”, Glotta 58: 44-46. Kimball, S. (2014) “Homeric κρύπτασκε, ῥίπτασκε, ἰσάσκετο”, Glotta 90: 163-173. Kirchhoff, A. (1859) Die Composition der Odyssee. Gesammelte Aufsätze, Berlin, Hertz. Kirchhoff, A. (1879) Die homerische Odyssee (zweite umgearbeitete Fassung von Die homerische Odyssee und ihre Entstehung und Die Composition der Odyssee), Berlin, Hertz. Kluge, H. (1911) Syntaxis Graecae quaestiones selectae, Inauguraldissertation Berlin. Kölligan, D. (2007) Suppletion und Defektivität im griechischen Verbum, Bremen, Hempen. Krüger, K.W. (1846) Griechische Sprachlehre für Schulen. Erster Theil: über die gewöhnliche, vorzügsweise die attische Prosa. Zweites Heft: Syntax, Berlin, Krügers Verlagsbuchhand‐ lung. Krüger, K.W. (1853) Griechische Sprachlehre für Schulen. Zweiter Theil: Ueber die Dialekte, vorzugsweise den epischen und ionischen. Erstes Heft: Formlehre, Berlin, Krügers Verlagsbuchhandlung. Krüger, K.W. (1859) Griechische Sprachlehre für Schulen. Zweiter Theil: Ueber die Dialekte, vorzugsweise den epischen und ionischen. Zweites Heft: Poetisch-dialektische Syntax, Berlin, Krügers Verlagsbuchhandlung. Kühner, R. (1835) Ausführliche Grammatik der griechischen Sprache. Wissenschaftlich und mit Rücksicht auf dem Schulgebrauch. Zweiter Theil, Hannover, im Verlage der Hahnschen Hofbuchhandlung. Kühner, R & F. Blass (1890) Griechische Grammatik. Formenlehre. Erster Band, Hannover, Hahn. Kühner, R. & F. Blass (1892) Griechische Grammatik. Formenlehre. Zweiter Band, Hann‐ over, Hahn. Kühner, R. & B. Gerth (1898) Ausführliche Grammatik der griechischen Sprache. Zweiter Theil. Satzlehre. Erster Band, Hannover, Hahn. Kühner, R. & B. Gerth (1904) Ausführliche Grammatik der griechischen Sprache. Zweiter Theil. Satzlehre. Zweiter Band, Hannover, Hahn. 238 Filip De Decker <?page no="239"?> La Roche, J. (1870a) Homer für den Schulgebrauch erklärt. Gesang I-IV, Berlin, Ebeling & Plahn. La Roche, J. (1870b) Homer für den Schulgebrauch erklärt. Gesang V-VIII, Berlin, Ebeling & Plahn. La Roche, J. (1871) Homer für den Schulgebrauch erklärt. Gesang XXI-XXIV, Berlin, Ebeling & Plahn. Lazzeroni, R. (1977) “Fra glottogonia e storia: ingiuntivo, aumento e lingua poetica indoeuropea”, Studi e Saggi Linguistici 17: 1-30. Lazzeroni, R. (1984) “La formazione del sistema dei tempi e degli aspetti”, Atti del Sodalizio Glottologico Milanese 24: 55-63. Lazzeroni, R. (2017) “Divagazioni sull’aumento in Omero”, in Marotta, G. & F. Strik Lievers (eds.) Strutture linguistiche e dati empirici in diacronia e sincronia, Pisa, Pisa University Press, 33-56. Leaf, W. (1886) Homer: The Iliad, I, Books I-XII, London, Macmillan. Leaf, W. (1888) Homer: The Iliad, II, Books XIII-XXIV, London, Macmillan. LIV 2 = Lexikon der indogermanischen Verben. Die Wurzeln und ihre Primärstammbil‐ dungen. Unter Leitung von H. Rix und der Mitarbeit vieler anderer, bearbeitet von M. Kümmel, Th. Zehnder, R. Lipp, B. Schirmer. Zweite, erweiterte und verbesserte Auflage, bearbeitet von H. Rix and M. Kümmel (2001), Wiesbaden, Reichert. LSJ = Liddell, H.; R. Scott; H. Jones; R. McKenzie (1996) Greek-English Lexicon, with a revised Supplement, with the help of many scholars, Oxford, Oxford University Press. Ludwich, A. (1892) Homeri Carmina recensuit et selecta lectionis varietate instruxit Arthurus Ludwich. Pars prima. Homeri Ilias. Volumen prius. Leipzig, Teubner. Ludwich, A. (1897) Homeri Carmina recensuit et selecta lectionis varietate instruxit Arthurus Ludwich. Pars prima. Homeri Ilias. Volumen alterum Leipzig, Teubner. MacLeod, C. (1982) Homer. Iliad XXIV, Cambridge Greek and Latin Classics, Cambridge, Cambridge University Press. Matthews, S.T. (1990) A cognitive approach to the typology of verbal aspect, Ph.D. Dissertation, University of Southern California. Mattiola, S. & G. Inglese (2018) The Hittite suffix -ške/ abetween verbal aspect and pluractionality: a typological approach, Handout of the lecture presented at the Conference Syntax of the World’s Languages 8 (INALCO, Paris, September 3-5 2018). Melchert, C. & H. Hoffner (2008) A grammar of the Hittite language, Winona Lakes, Eisenbrauns. Miller, D.G. (2014) Ancient Greek dialects and early authors. Introduction to the dialect mixture in Homer, with notes on lyric and Herodotus, Berlin, De Gruyter. Monro, D. (1884) Homer. Iliad Books I-XII, Oxford, Oxford University Press. Monro, D. (1891) A grammar of the Homeric dialect, Oxford, Oxford University Press. Monro, D. (1901) Homer’s Odyssey XIII-XXIV, Oxford, Clarendon. The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 239 <?page no="240"?> Monro, D. & Th. Allen (1908) Homeri Opera: Iliadis libros I-XII continens. Oxford, Oxford University Press. Mutzbauer, C. (1893) Die Grundlagen der griechischen Tempuslehre und der homerische Tempusgebrauch, Strassburg, Trübner. Mutzbauer, C. (1909) Die Grundlagen der griechischen Tempuslehre und der homerische Tempusgebrauch, Strassburg, Trübner. Napoli, M. (2006) Aspect and Actionality in Homeric Greek. A contrastive analysis, Milan, Angeli. Nauck, A. (1877) Homeri Ilias cum potiore lectionis varietate edidit Augustus Nauck. Pars prior, Berlin, Weidmannsche Buchhandlung. Negri, M. (1976) “Studi sul verbo greco II”, Acme 29: 233-250. Pagniello, F. (2007) “The past-iterative and the augment in Homer”, Indogermanische Forschungen 112: 105-123. Peppmüller, R. (1876) Commentar des vierundzwanzigsten Buches der Ilias. Mit Einleitung. Als Beitrag zur homerischen Frage bearbeitet, Berlin, Weidmannsche Buchhandlung. Pisani, V. (1959) “Obiter scripta”, Paideia 10: 168-181. Pisaniello, V. (2020) I suffissi verbali ittiti -anna/ ie -šš(a)-. Studio sul sistema aspettuale ittita, Pisa, Pisa University Press. Pisaniello, V. (2022) “Verbal aspect in Anatolian: the Luwian suffixes -s(s)aand -z(z)aand the verbal reduplication”, presentation held at the Delbrück Colloquium on Historical and Comparative Syntax of Indo-European, at the University of Verona on 9-12 November, 2022. Pott, A.F. (1833) Etymologische Forschungen auf dem Gebiete der Indogermanischen Sprachen, mit besonderem Bezug auf die Lautumwandlung im Sanskrit, Griechischen, Lateinischen, Litauischen und Gothischen, Erster Theil: Präpositionen, Detmold & Lemgo, Meyer’sche Hofbuchhandlung. Puhvel, J. (1991) Hittite and Homer, Innsbruck, Innsbrucker Beiträge zur Sprachwissen‐ schaft. Richardson, N. (1993) The Iliad: a commentary. 6: 22-24, Cambridge, Cambridge University Press. Risch, E. (1974) Wortbildung der homerischen Sprache, Berlin, De Gruyter. Rix, H. (1976) Historische Grammatik des Griechischen. Laut- und Formenlehre, Darmstadt, Wissenschaftliche Buchgesellschaft. Rix, H. (1992) Historische Grammatik des Griechischen. Laut- und Formenlehre, 2. erwei‐ terte und verbesserte Auflage, Darmstadt, Wissenschaftliche Buchgesellschaft. Romagno, D. (2015) “The Greek-Anatolian area in the 2 nd millennium B.C.: between language contact, Indo-European inheritance and typologically natural tendencies”, Studi e Saggi Linguistici 53: 429-446. 240 Filip De Decker <?page no="241"?> Russo, J.; M. Fernández-Galiano; A. Heubeck (1992) A commentary on Homer’s Odyssey. Volume III: Books XVII-XXIV, Oxford, Clarendon. Schulze, W. (1892) Quaestiones Epicae, Gütersloh, Bertelsmann. Schwyzer, E. (1939) Griechische Grammatik auf der Grundlage K. Brugmanns Griechischer Grammatik, München, Beck. Schwyzer, E & A. Debrunner (1950) Griechische Grammatik. Teil II. Syntax. München, Beck. Stolpe, A. (1849) Iterativorum Graecorum vis ac natura ex usu Homeri atque Herodoti demonstrata, Bratislava, Klein. Tichy, E. (2009) Indogermanistisches Grundwissen für Studierende sprachwissenschaftlicher Disziplinen, Bremen, Hempen. Týn, E. (1860) “Über den Gebrauch und die Bedeutung der iterativen Imperfecta und Aoriste im Griechischen”, Zeitschrift für die österreichischen Gymnasien 10: 677-695. Van Leeuwen, J. & M.B. Mendes da Costa (1886) Der Dialekt der Homerischen Gedichte: für Gymnasien und angehende Philologen, aus dem Holländischen von E. Mehler übersetzt, Leipzig, Teubner. Van Leeuwen, J. & M.B. Mendes da Costa (1895) Homeri Iliadis Carmina cum apparatu critico ediderunt J. Van Leeuwen J.F. et M.B. Mendes da Costa, Leiden, Sijthoff. Van Leeuwen, J. & M.B. Mendes da Costa (1898) Het Taaleigen der Homerische gedichten, Leiden, Sijthoff. Van Thiel, H. (1996) Homeri Ilias, Hildesheim, Olms. Van Thiel, H. (2011) Homeri Ilias, second edition, Hildesheim, Olms. Van Emde Boas, E. & L. Huitink (2010) “Syntax”, in Bakker, E. (ed.) A companion to the Ancient Greek language, Malden, MA, Blackwell, 134-150. Van Emde Boas, E.; M. De Bakker; L. Huitink; A. Rijksbaron (2019) Cambridge grammar of Classical Greek, Cambridge, Cambridge University Press. Von Thiersch, F. (1826) Griechische Grammatik, vorzüglich des homerischen Dialektes, Leipzig, Fleischer. Vendler, Z. (1957) “Verbs and Times”, The Philosophic Review 66: 143-160. Vogrinz, G. (1889) Grammatik des homerischen Dialektes, Paderborn, Schöningh. Von Doederlein, L. (initially L. Doederlein) (1863) Homeri Ilias. Emendavit et illustravit D. Ludovicus Doederlein. Pars prior I- XII, Leipzig, Dörffling & Franke. Von Doederlein, L. (initially L. Doederlein) (1864) Homeri Ilias. Emendavit et illustravit D. Ludovicus Doederlein. Pars posterior XIII-XXIV, Leipzig, Dörffling & Franke. Wathelet, P. (1973) “Études de linguistique homérique”, Antiquité Classique 42: 379-405. Watkins, C. (2001) “An Indo-European linguistic area and its characteristics: Ancient Anatolia. Areal diffusion as a challenge to the Comparative Method? ”, in Aikhenvald, A. & R. Dixon (eds.) Areal diffusion and genetic inheritance. Problems in comparative linguistic, Oxford, Oxford University Press, 44-63. The imperfective meaning of the Epic-Ionic -σκ-iteratives as Graeco-Anatolian isogloss? 241 <?page no="242"?> West, M. (1998) Homerus Ilias. Volumen I: Rhapsodiae I-XI, Berlin, De Gruyter. West, M. (2000) Homerus Ilias. Volumen II: Rhapsodiae XIII-XXIV, Berlin, De Gruyter. Zerdin, J. (1999) Studies in the Ancient Greek verbs in -SKŌ, DPhil Thesis, Oxford. Zerdin, J. (2002) “The ‘Iterative-Intensivesin -σκ”, Oxford Working Papers in Linguistics 7: 103-130. 242 Filip De Decker <?page no="243"?> 1 For the findspot and for other artifacts recovered from the deposit, see Villa D’Amelio 1963: 65-69 and Rallo 2005: 232. 2 See Rallo 2005: 232-233. 3 See Sofredi and Susini 1966: 188, no. 9; L’Année Epigraphique 1966: 85 and 1984: 93-94, no. 374; and CIL I 2 , 2870. Comments on an Old Latin Inscription from San Giuliano (CIL I 2 , 2780) Rex Wallace (The University of Massachusetts Amherst) Abstract: This paper is a discussion of an Old Latin inscription incised on a statue base recovered from a votive deposit near the settlement of San Giuliano in Etruscan territory. The inscription, which dates to the 3 rd c. BCE, is dedicated to the god Apollo. The inscribed statue base documents the worship of Apollo in southern Etruria and is witness to the presence of Latin-speaking followers of the god in the area in the 3 rd c. BCE. Keywords: Old Latin; San Giuliano; Apollo; ‘open’ O ; votive dedication. 1. Introduction In 1957, an inscribed trapezoidal statue base cut from grey volcanic tuff, measuring 20 cm in height, was recovered among the scattered remains of a votive deposit in the environs of San Giuliano, a small settlement located approximately 30 km east of Tarquinia in the Ager Tarquiniensis (fig. 1). 1 The date of the votive deposit is the 3 rd c. BCE. 2 A 3 rd c. date, perhaps even the 1 st half of the 3 rd c., for the inscribed base is supported by the inscription’s paleographical features. After the inscription’s initial publication by Villa D’Amelio in 1963, it appeared in several epigraphic venues, including the Corpus Inscriptionum Latinarum, although with little additional discussion. 3 In this article, I contribute notes on the inscription’s language, layout, and paleography. <?page no="244"?> 4 Old Latin inscriptions in which every constituent is abbreviated are rare. See the dedication to Victory set up by Saluius and Statius Flauius (CIL I 2 , 387). 5 Villa D’Amelio 1963: 68. Word-forms in Latin inscriptions are printed in small capitals. 6 See footnote 3 and add Rallo 2005: 233 and Steingräber 2018: 400, both of whom erroneously transliterate the nomen as S M O N . 7 Gasparini 2009: 374, fn. 15. 8 Quite apart from the impossibility of Gasperini’s proposal, the inscription is too early for a consonant-stem dative case ending to be marked by the inflectional suffix I = / iː/ . 9 In CIL I 2 , 2780, Rallo 2005: 233, and Steingräber 2018: 400, the abbreviation D in line 2 is expanded as D ( O N O M ). This is possible, but less likely than D ( E D E T ), given that the name of the donor is present in line 1. Old Latin votive inscriptions with the structure theonym (dat.) + D ( O N O M ) ‘an offering for X’ are typically found when the name of the 2. Inscription The inscription was incised in two lines a few centimeters from the upper edge of the face of the statue base (fig. 2). Line 1 is the onomastic phrase of the base’s donor. The phrase consists of the praenomen, nomen, and patronymic, all in abbreviated forms. 4 Line 2 is a verb phrase comprising the name of the divinity to whom the votive was dedicated and a verb of dedication. Both words are abbreviated. The constituents of both phrases were separated by punctuation in the form of a period placed at mid-line level. The inscription is cited in (1) as per the reading of A. Degrassi, which was published by Villa D’Amelio. 5 (1) A . SMO . TI (F) - APOLO . D The text in (1) is accepted by almost all commentators. 6 The only dissenting voice is Gasparini, who suggested that the final letter of line one may better be read as F , an attractive proposal to be sure, but one that is paleographically problematic. 7 The last two letters of this line were squeezed together at line-end, without punctuation separating the letters. And though the stone is weather-beaten, pitted, and damaged because of the friability of the tuff, making the reading of the line’s final letter challenging, F is unlikely given that horizontal bars are not visible, even under magnification. Gasperini also proposed A P O L O NI for line two, claiming that there was no trace of a line-final D . However, punctuation is visible following the final O of A P O L O . The last letter in the line is clearly visible too; it is D . 8 Following Old Latin orthography, the abbreviations in the text may be expanded as A ( V L O ) . S ( E ) M O ( NI O ) . TI ( B E R I ) | A P O L O ( N E ) . D ( E D E T ). 9 Assuming 244 Rex Wallace <?page no="245"?> donor is not overtly expressed. See, for example, P A R C I A . M O U R T I A | D O N O (CIL I 2 , 2844) and N E U N A . D O N O (CIL I 2 , 2845), both inscriptions belonging to the 3 rd c. (Wachter 1987: 373). This structure is also found when the donor is indicated by a syntactic role other than the subject, e.g., I U N O N E . L O C I N A | D O N O . P R O | C . R U T I L I O . P . F (CIL I 2 , 359). 10 On the loss of word-final - S in Old Latin, see Meiser 1998: 96 and Weiss 2020: 66-67. 11 See Weiss 2020: 110-111 for developments of original *ey in Old Latin. 12 For the spelling, compare Faliscan apolonos (gen.), which was incised on a black glazed cup dating to the first quarter of the 5 th c. See Bakkum 2008: 416, no. 10 (Old Faliscan), and Giacomelli 1963: 59, no. 31. 13 For abbreviations of theonyms in Old Latin votive inscriptions, see, for example, D E I V ( E I S ) N O [ V ] E S E D E ( B O S ) (CIL I 2 , 375); I U N O N E R E C̣ ( I N A I ) (CIL I 2 , 378,); V I C ( T O R I A E ) (CIL I 2 , 387); N ( U M I S I O ) M A R T ( I O ) (CIL I 2 , 2435); D I A N A I O P I F E R ( E I ) | N E M O R E N S E I (CIL I 2 , 1480); and I U N O ( N E ) P A L O S C A R I A (CIL I 2 , 2439). The abbreviation I U N O ( N E ) provides the perfect parallel for our A P O L O ( N E ). 14 The abbreviated form A P O L S (gen.) = A P O L ( O N E ) S (CIL 1 2 2387; ST VM 7; Imag. Ital. Marsi 1, pg. 332 [400-350 BCE]), which Rix includes in ST as Marsian in language, is, if not a Latin form, then a borrowing from Latin. The form A P L O N E (CIL I 2 2378a; ST VM 6; Imag. Ital. Supinum 1, pp. 336-37 [c. 200 BCE]), also recovered in Marsian territory, stands for A P ( E ) L O N E by syllabic notation (see Vine 1993: 241, fn. 2, and 338). A P ( E ) L O N E , with E in the antepenultimate syllable, reflects the Oscan form of the theonym, cf., Oscan appelluneís (gen.), ST Po 14, and appelluneí (dat.), ST Si 3. 15 See Vine (1993: 332-333) for syllabic notation in word-initial syllables. that the inscription belongs to the first half of the 3 rd c. BCE, the nominative singular of the praenomen and nomen may be expanded without the case-ending - S . 10 The dative ending of A P O L O should be expanded as - N E , or possibly even - N E I , to provide a better fit with the chronology of the inscription. 11 For the spelling of the theonym with a single medial L and with an O as the ultimate vowel of the stem, compare other attestations in Old Latin, e.g., A P O L O N E S (gen.), incised on a bronze plate, origin unknown but possibly Latium (CIL I 2 , 37); A P O L OṆ [ E ] (dat.), incised on a stone pedestal base (? ) recovered from Praeneste (CIL I 2 , 59), and A P O L O N E (dat.), incised on a small terracotta box recovered from Cales (CIL I 2 , 399), all dating to the 3 rd c. 12 The abbreviation of the theonym is unusual. Theonyms were abbreviated infrequently in Old Latin votive dedications. 13 A P O L O ( N E ) may be the only abbreviation of this theonym in Old Latin. 14 Presumably, the dative form was readily recoverable given the context. The form of the patronymic, which was expressed by the praenomen of the father but without the abbreviation of its head noun filius, is also unusual. In this instance, the absence of the abbreviation F is almost certainly to be attributed to lack of space at line-end. The missing vowel letter in the initial syllable of the nomen S ( E ) M O is best explained as a case of syllabic notation. 15 This form of abbreviated writing was Comments on an Old Latin Inscription from San Giuliano (CIL I 2 , 2780) 245 <?page no="246"?> 16 For discussion of Latin letter names, see Gordon 1973, Vine 1993: 332-333, Wachter 1987: 50-53, Wallace 2011: 23-24, and Weiss 2020: 33. 17 Etruscan semnies and its feminine counterpart, semni, are attested in funerary inscrip‐ tions from Tarquinia (ET Ta 1.168, ET Ta 1.173). The nomen σemna is attested in a few inscriptions from Castellina in Chianti (ET Vt. 1.61-64). The origin of the nomen is probably Etruscan, but whether it is northern Etruscan, as Rallo (2005: 233-234) and Steingräber (2018: 401) propose, is impossible to say. 18 Schulze 1904: 228. 19 Villa D’Amelio (1963: 68, fn. 4) suggests a possible connection between Semonius and the theonym Semo. 20 See Cencetti (1957) for Old Latin letter forms. See CIL I 2 , 358 (end of 4 th c.). motivated by the pronunciation of Latin letter names. The name of the letter m, like that of the other continuants, had the structure eR and was pronounced as / em/ . 16 In this case, the convention may have been employed as a space-saving device. The abbreviation S ( E ) M O is expanded as Semonius in all publications. The name is not attested in Republican Latin; it makes its first appearance in an imperial Latin inscription recovered from Rome (CIL VI 7465). Semonius is generally connected to the family of Etruscan nomina represented by semnies (fem. semni) and semna. 17 However, the correspondence between the Etruscan and Latin names is an imperfect one. Rather, it is Latin Semnius, which appears in a funerary inscription recovered at Colonia Saturnia (CIL XI, 2653), also of imperial date, that corresponds most closely to Etruscan semnies. 18 Even so, this does not rule out an onomastic relationship between Semonius and semnies because the transposition of onomastic forms in the two languages allowed for differences in suffixal morphology. See, for example, the Etruscan nomen perśu / perʃu/ and its probable Latin counterpart Persius. 19 3. Layout and Paleography Each line of the inscription was incised with six letters, but the stonecutter made no attempt to align the letters vertically, apart from those positioned at the beginning of the lines (fig. 3). The letters are not evenly spaced, and their size varies. For example, the letter S , which stands upright ( ), is taller than the other letters in line 1. The letter D is elevated with respect to the other letters in line 2. The letter L , in contrast, sits lower. The paleography of the letters on the statue base is consistent with a 3 rd c. date. The three tokens of the vowel O are open at the bottom; they have the shape ᴒ. 20 They are the same height as other letters. P has a rounded open head that extends 246 Rex Wallace <?page no="247"?> 21 Compare the forms of P on CIL I 2 , 558, 554, 547, 569, all incised on Praenestine mirrors of the late 4 th / early 3 rd century BCE. 22 See, for example, the form of the letter A in captions on Praenestine mirrors, e.g., CIL I 2 547, 548, 549, 552, 555. See also Wachter 1987: 175-177. 23 Steingräber (2018: 400-401) compares the letterforms in CIL I 2 , 2843, but the forms, including L , are not comparable in all respects. 24 See, for example, CIL I 2 , 37, CIL I 2 , 59 and CIL I 2 , 399, all cited in section 2. 25 For discussion of the replacement of the oldest form of the Latin oblique stem apollōn- ( A P O L O N -) by the short-vowel stem apollen- ( A P O L E N -), and ultimately apollinvia vowel weakening, see Vine 1993: 241-246. For the idea that the stem / apollen-/ is an urban Latin innovation, see Wachter 1987: 397-398. This proposal is perhaps supported by the fact that the innovative forms, albeit with the later change of e to i by vowel weakening, appear in the Old Latin literary record, e.g., Plautus (Men. 871, Apollinis), Ennius (sc. 41 V, Apollinem). 26 So Cencetti (1957: 194). downward but does not bend toward the letter’s upright bar. 21 The transverse bars of both tokens of A are horizontal; the top of the letter is rounded. 22 M is squat and broad; its four bars are oblique. The most unusual letterform is L . Its bars are disarticulated. The left bar is not quite vertical; the right bar slants upward at an acute angle giving the letter something of a u-like appearance. Both bars are roughly the same length. The form of L here has been compared to the L in CIL I 2 , 2843, an inscription on a stone cippus belonging to the end of the 4 th / the early part of the 3 rd c. BCE, whose bars are also disarticulated. 23 However, the oblique bar of the L of CIL I 2 , 2843 is shorter than the oblique bar of the L in CIL I 2 , 2870, and its angle is less acute. Apart from the disarticulation of its bars, the L of CIL I 2 , 2843 resembles closely the canonical form of Old Latin . 4. Concluding remarks The stem of the theonym, apollōn-, which maintained the long vowel of the Attic/ Ionic Greek original in its final syllable, is typical of Old Latin inscriptions dated to the 4 th -3 rd c. BCE. 24 Forms with a short e in the final syllable of the stem, apollen-, are the result of a chronologically later innovation and are found primarily, but not exclusively, in votive dedications recovered at colonial sites established in the late 3 rd - early 2 nd c. BCE. 25 Paleographically, this inscription is noteworthy for its open O , squat M , open P , and disarticulated L . It is not clear that the form of L , which is like no other in the Old Latin corpus, is to be viewed as possibly reflecting a tendency toward cursive writing, which is attested in the development of other Old Latin letterforms, most prominently E and F . 26 Comments on an Old Latin Inscription from San Giuliano (CIL I 2 , 2780) 247 <?page no="248"?> 27 In addition to the references cited in §2, add the following dedications to Apollo: CIL I 2 , 384 [c. 150]), A P O L E N E (Mosciano in Picenum; CIL I 2 , 368 [c. 150], A P O L E N E I (Pisaurum); CIL I 2 , 1928 [c. 150], A P O L I N E (Morrovalle in Picenum); CIL I 2 , 2628 [c. 150], A P O L I N E (Veii,), CIL I 2 , 2894 [2 nd c.], [ A P ] O L E N [ (Ariminium), CIL I 2 , 2895 [2 nd c.] [ A P O L ] E N I (Ariminium). 28 See Rallo 2005 and Steingräber 2018: 401. 29 For the cult of Apollo medicus in Etruria, see Haack 2007 and Steingräber 2018. The cult of Apollo was introduced to central Italy at the beginning of the 5 th c. BCE. The theonym first appears in Faliscan (apolonos gen.) on ceramic vase found in the area of the Tempio Maggiore on Colle di Vignale, the oldest cult center at Falerii Veteres. Etruscan apalus, which was incised beneath the foot of a bucchero cup dating to the first half of the fifth century BCE, may be the genitive of the theonym (Maras 2009a: 287-288 and Maras 2009b: 246-247). The find comes from a sanctuary located outside of the settlement Narce in the Ager Faliscus. The earliest attestation in Latin appears to be on a gemstone dated to the first part of the 4 th c. BCE (Allen and Wallace 2021: 310-311). Votive dedications to the god Apollo are well-documented in Old Latin, both in Latium and in areas populated by Latin and Roman colonists. 27 The context in which CIL I 2 , 2780 was found—a votive deposit filled with anatomical offerings 28 —documents the worship of Apollo in his therapeutic aspect in southern Etruria and is witness to the presence of Latin-speaking adherents of the god in this area in the 3 rd c. BCE. 29 Abbreviations CIL I 2 = Degrassi and Krummery 1986. CIL VI = Bormann and Henzen 1876. CIL XI = Bormann 1888. Imag. Ital. = Crawford 2011. ET = Meiser 2014. ST = Rix 2002. Bibliographical references Allen, R.M & R. Wallace (2021) “Two Etruscan scarab rings engraved with Old Latin inscriptions”, Zeitschrift für Papyrologie und Epigraphik 118: 307-314. Bakkum, G.C. (2009) The Latin dialect of the Ager Faliscus. 150 years of scholarship. Parts I and II, Amsterdam, Amsterdam University Press. Bormann, E. (1888) (ed.) Corpus Inscriptionum Latinarum, Inscriptiones Aemiliae, Etruriae, Umbriae Latinae, Vol. XI, pars I, Berlin, Reimer. 248 Rex Wallace <?page no="249"?> Borman, E. & W. Henzen (1876) (eds.) Corpus Inscriptionum Latinarum, Inscriptiones Urbis Romae Latinae, Vol. VI, pars I, Berlin, Reimer. Chastagnol, A.; M. Leglay; P. Le Roux (1984) (eds.) “Année 1981. Italie, no. 374”, L’Année Épigraphique-1984: 93-94. Cencetti, G. (1957) “Ricerche sulla scrittura Latina nell’eta arcaica”, Bullettino dell’Archivio Paleografico Italiano 2/ 3: 175-205. Crawford, M. (2011) Imagines Italicae. A corpus of Italic inscriptions, Volumes 1-3, London, The Institute of Classical Studies. Degrassi, A. & H. Krummery (1986) (eds.) Corpus Inscriptionum Latinarum, Inscriptiones Latinae antiquissimae ad C. Caesaris mortem, Vol. I², fasc. 4, Berlin, de Gruyter. Gagé, J. & M. Leglay (1966) (eds.) “Année 1965. Périodiques, no. 261”,-L’Année Épigra‐ phique-1966: 85. Gasperini, L. (2009) “Epitafio latino inedito dal Blerano”, Epigraphica 71: 373-375. Giacomelli, G. (1963) La lingua falisca, Firenze, Olschki. Gordon, A.E. (1973) The letter names of the Latin alphabet, Berkeley, University of California Press. Haack, M. L. (2007) “Apollon médecin en Etrurie”, Ancient Society 37: 167-190. Maras, D.F. (2009a) Il dono votivo. Gli dei e il sacro nelle iscrizione etrusche di culto, Pisa & Roma, Fabrizio Serra. Maras, D.F. (2009b) “Note in margine al CIE II, 1, 5”, Studi Etruschi 73: 237-247. Meiser, G. (1998) Historische Laut- und Formenlehre der lateinischen Sprache, Darmstadt, Wissenschaftliche Buchgesellschaft. Meiser, G et al. (2014) Etruskische Texte. 2 Editio minor. Bd. I. Einleitung, Konkordanz, Indices; II. Texte, Hamburg, Baar. Rallo, A. (2005) “Appunti preliminary sulle aree sacre di San Giuliano”, in Comella, A. & S. Mele (eds.) Depositi votive e culti dell’Italia antica dal period arcaico a quello tardo-republicano. Atti del Convegno di Studi (Perugia 1-4 giugno 2000), Bari, S. Spirito, 229-236. Rix, H. (2002) Sabellische Texte. Die Texte des Oskischen, Umbrischen, und Südpikenische, Heidelberg, Winter. Schulze, W. (1904) Zur Geschichte lateinishcer Eigennamen, Berlin, Weidmann. SofrediA. & G. Susini. (1966) “No. 9. Barbarano Romano”, Epigraphica 28: 188. Steingräber, S. (2018) “Zum Phänomen de “follia terapeutica” der etruskisch-italischen anatomichen Votiv - soziale, ökonomische und religiose Aspekte”, in Aigner-Foresti, L. & P. Amann (eds.) Beiträge zur sozialgeschichte der Etrusker. Akten der internatio‐ nalen Tagung (Wien, 8-10.6.2016), Wien, Holzhausen, 399-408. Villa D’Amelio, P. (1963) “San Giuliano. Scavi e scoperte nella necropolis dal 1957 ai 1959”, Notizie degli scavi di antichità 17: 1-76. Comments on an Old Latin Inscription from San Giuliano (CIL I 2 , 2780) 249 <?page no="250"?> Vine, B. (1993) Studies in Archaic Latin inscriptions, Innsbruck, Instrucher Beiträge zur Sprachwissenschaft. Wallace, R. (2011) “The Latin alphabet and orthography”, in Clackson, J. (ed.) A companion to the Latin language, London & New York, Cambridge University Press, 9-28. Weiss, M. (2020) 2 Outline of the historical and comparative grammar of Latin, Ann Arbor & New York, Beech Stave. Acknowledgements I thank my colleague, Stephan Steingräber, for sharing his photographs of the inscribed statue base. The inscription was drawn by Valerie Woelfel with the support of a grant from the University of Massachusetts Amherst Retired Faculty Association. Images Fig. 1 Inscribed tufa statue base. Early 3 rd c. BCE (? ). Photograph courtesy of Stefan Steingräber. 250 Rex Wallace <?page no="251"?> Fig. 2 Inscription on tufa statue base. Early 3 rd c. BCE (? ). Image courtesy of Stefan Steingräber. Fig. 3 Illustration of inscription on tufa statue base. Early 3 rd c. BCE (? ). Image courtesy of the University of Massachusetts Amherst. Drawing by Valerie Woelfel. Comments on an Old Latin Inscription from San Giuliano (CIL I 2 , 2780) 251 <?page no="253"?> 1 On the periodisation of Latin, see Adamik (2015). Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? Luca Rigobianco (Ca’ Foscari University of Venice) Abstract: This paper deals with the vowel reduction and deletion in non-initial syllables which occurred in Latin after the earliest documents between the 5 th and the 3 rd century BC. Specifically, it examines the well-known hypothesis according to which these phenomena would be contact-induced. In this regard, it aims to demonstrate, on the one hand, that the borrowed feature is not (only) the stress position on the first syllable but (also) a rhythmic tendency and, on the other, that there are linguistic and extra-linguistic reasons for assuming that it spread from the Sabellic languages to Etruscan and, probably later, to Latin. Keywords: contact-induced change; Archaic Latin; Sabellic languages; Etruscan; vowel reduction; vowel deletion. 1 Introduction This paper deals with the vowel reduction and deletion in non-initial syllables which occurred in Archaic Latin 1 after the earliest documents between the 5 th and the 3 rd century BC. Specifically, it aims to examine the well-known hypothesis according to which these phenomena would be contact-induced. First, I will recall the relevant Latin data (§2) and then discuss the role of stress position and rhythmic pattern in determining vowel reduction and deletion (§2.1). Subsequently, I will focus on the vowel reduction and deletion attested in coeval neighbouring languages, that is to say Faliscan (§3.1), the Sabellic languages (§3.2), and Etruscan (§3.3), in order to precisely assess their chronology and outcomes for comparison with the Latin data. In light of the foregoing, I will assume that the vowel reduction and deletion in Latin <?page no="254"?> 2 Unless vhevaked should be emended in vhaked, as suggested by Maras (2015) and Mancini (2021). 3 Some scholars hold that the change of -ou̯Vinto -ūshould be considered apart from the phenomena in question (see, for example, Meiser 1998: 88). In any case, see also iouoand ioueon the Corcolle Altar fragments, dated to approximately 500 BC (Prosdocimi 1979a: 197-216; Vine 1993: 65-83). 4 According to Weiss (2020: 149), the change of -ed to -it would not be a case of vowel reduction. 5 Gnade & Colonna (2003). See also mamarc[ on a urn from Osteria dell’Osa dated to 630-620 BC (Colonna 1980a). 6 On a possible explanation of the relationship between the two phenomena in Archaic Latin, see Rix (1966). are probably contact-induced phenomena and I will attempt to reconstruct their origin and development (§4). Specifically, I intend, on the one hand, to demonstrate that the borrowed feature is not (only) the stress position on the first syllable but (also) a rhythmic tendency and, on the other, to confirm the linguistic and extra-linguistic reasons for assuming that it spread from the Sabellic languages to Etruscan and, probably later, to Latin. 2 Vowel reduction and deletion in Archaic Latin In the earliest Latin texts, there is no evidence of vowel reduction or deletion (cf. Nishimura 2010a): see, for example, vhevhaked  2 and numasioi in the Praenestine fibula (second quarter of the 7 th century BC; Colonna 1999: 439), iouesat  3 and feced  4 in the Duenos inscription (first half of the 6 th century BC; Colonna 1979: 167), sacros, esed, and iouestod in the Forum Cippus (mid-6 th century BC; Coarelli 1983: 130), mamarcom in a fragmentary inscription on a dolium from Satricum 5 (mid-6 th century BC; Gnade & Colonna 2003: 19), and mamartei in the Lapis Satricanus (end of the 6 th / beginning of the 5 th century BC; Colonna 1980b: 48). Later, but prior to the beginning of the literary tradition in the 3 rd century BC, vowel reduction and deletion occurred in non-initial syllables 6 . Specifically, on the one hand, vowel reduction is recognisable in prefixed, compounded, and univerbated forms - mostly verbs - (see, for example, conficio beside facio), but also in consonant-stem nouns (see, for example, genitive capitis beside nominative caput), third conjugation verbs (see, for example, legimus beside legere), and early Greek loanwords (see, for example, Tarentum beside Greek Τάρας, Τάραντος). The trigger and outcome of vowel reduction depend on various factors, such as syllable position, syllable structure, and phonetic context (Leumann 1977: 79-91; Meiser 1998: 67-73; Weiss 2020: 126-131; cf. also Nishimura 2010b). Furthermore, analogical levelling may obscure the effects of 254 Luca Rigobianco <?page no="255"?> 7 Cf. Cic. Fam. 9.21.2 „L. Papirium Crassum, qui primum Papi<s>ius est vocari de‐ situs“ (Lucius Papirius Crassus was consul in 336 and 330 BC). 8 As to the preliterary stage, Nishimura (2011: 14-17) has argued that vowel deletion took place under a metric constraint which would have prevented the sequence of two light syllables: see, for example, ūsurpā- < *oi̯sŭrŭpāand repperī < *rĕpĕpărī. 9 Such a hypothesis has been rejected on different grounds by Ballester (1990; 1996), Oniga (1990; 2006), Pultrová (2006; 2011), and Ohannesian (2020), who have provided different alternative explanations. However, the traditional hypothesis seems to more convincingly account for the whole phenomenology in question - including vowel reduction in early Greek loanwords - than those put forward by Oniga, Pultrová, and Ohannesian. On Ballester’s hypothesis, see below (n. 65). 10 Some phonetic changes which affected Italic languages may be explained assuming the persistence of the Proto-Indo-European mobile accent in early Proto-Italic (see Vine 2012 and references therein as well as Höfler 2017). 11 See, by way of example, Fortson (2011: 102; “all the linguistic evidence strongly indicates that Latin had a strong expiratory stress-accent throughout its history”). vowel reduction by restoring the original vowel. On the other hand, it is very difficult to determine the rules of vowel deletion, also because the phenomenon seems to have occurred several times throughout the history of Latin (Leumann 1977: 95-99; Meiser 1998: 66-67, 73-74; Weiss 2020: 132-135). Apparently, the earliest occurrences of vowel deletion predate rhotacism, which took place in the 4 th century BC 7 : see, for example, pōnō < *posnō < *posinō (Ernout & Meillet 1959: 520-521; Walde & Hofmann 1938-1954: II, 335-336; de Vaan 2008: 479) instead of **pornō < **porinō < *posinō  8 . 2.1 Vowel reduction and deletion as stressand rhythm-related phenomena Since in Latin vowel reduction and deletion affect only non-initial syllables, such phenomena are commonly ascribed to the effect of a first-syllable stress, which would have predated the Penultimate Law of stress placement (Leumann 1977: 246-248; Meiser 1998: 53; Weiss 2020: 119) 9 . Specifically, such a stress has been often considered as a late Proto-Italic innovation as compared to the Proto-Indo-European mobile accent 10 . In any case, it should be noted that stress position is not sufficient by itself to determine the phenomena of vowel reduction and deletion being considered. This, as well as the opposition to the French School’s hypothesis of a pitch accent in Latin (Leumann 1977: 248-254), could be the reason why the accent responsible for such phenomena is commonly described as a dynamic, expiratory, or intensive accent which would have been particularly strong 11 . Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? 255 <?page no="256"?> 12 In this regard, Flemming (2005: 3) claims that “[i]t is typical correlates of lack of stress that condition neutralization, not stress per se”. However, other factors should also be considered 12 and, especially, rhythmic factors seem to have played an important role. Several theoretical models have been developed to describe how rhythm works from both a phonetic and a phonological point of view. In this regard, the traditional distinction between stress-timed languages and syllable-timed languages, which has been recently revised by Bertinetto as a distinction between compensating languages and controlling languages (Bertinetto 1989; Bertinetto & Bertini 2008; Bertinetto & Bertini 2010), may be particularly convenient in explaining vowel reduction and deletion phenomena. Specifically, Bertinetto starts from Pike’s proposal to distinguish a syl‐ lable-timed rhythm, in which all the syllables tend to have the same duration, and a stress-timed rhythm, in which, on the contrary, it is the intervals between stresses which tend to have the same duration (Pike 1945: 34-36). The stress-timed rhythm, which would characterise languages such as English, would produce a “rhythmic crushing of syllables into short time limits” and, therefore, would be “partly responsible for many abbreviations - in which syllables may be omitted entirely - and the obscuring of vowels” (Pike 1945: 34). Although such a distinction has been contradicted by experimental evi‐ dence, according to Bertinetto, the core hypothesis may be maintained by replacing the notions of syllable-timed and stressed-time rhythm with those of compensation and control. Control and compensation refer to “how vocalic and consonantal gestures are coupled in the articulatory flow”. In particular, in an ideal controlling language “all segments receive the same amount of expenditure, i.e. articulatory effort, and (ideally) tend to have the same duration”, whereas an ideal compensating language is characterised by an “increased gestural overlap in unstressed syllables, where the segment most liable to compression / coarticulation is of course the vocalic nucleus” (Bertinetto 2009: 427). Within such a framework, vowel reduction and deletion in non-initial syllables would be accounted for by the interaction between a first-syllable stress and the tendency to compensate, that is to say, on the prosodic level, the tendency to the tolerance of unstressed vowels towards very high levels of coarticulation. Evidently, such a tendency may have different outcomes depending on its relative strength as well as language-specific features, such as the acceptability of complex consonants clusters. 256 Luca Rigobianco <?page no="257"?> 13 Monte di Badi is a village in the Apennines about sixty kilometres south west of Bologna. 14 For an up-to-date review, see Bakkum (2009: 100-103). 15 Bakkum 98, 162, and 220. Faliscan inscriptions are cited according to Bakkum (2009). 3 Vowel reduction and deletion in Faliscan, the Sabellic languages, and Etruscan 3.1 Faliscan The identification of vowel reduction and deletion phenomena in Faliscan is quite controversial. Such phenomena are usually said to be absent in Faliscan (cf. Giacomelli 2006: 104), though Bakkum (2009: 101) considers this unlikely due to the close kinship between Faliscan and Latin. In this regard, it should be noted, however, that close dialects may instead show a very different phenomenology in relation to vowel reduction and deletion. This is the case, for example, of the contemporary Bolognese dialects, in which the tendency towards vowel reduction and deletion increases as one goes from the Apennines to the city of Bologna (cf., for example, Monte di Badi 13 dialect [sel'vadigo] and Bologna dialect [sal'va: dg] ‘wild’ < Latin siluāticum; Filipponio 2012: 71-79). Anyway, the available data in Faliscan are scarce and their analysis with regard to the phenomena in question is often uncertain 14 . Here, I limit myself to recalling the form maxomo, which occurs as a cognomen in at least three Faliscan inscriptions 15 . Such a form, which corresponds to Latin maximus/ maxumus and may be likely traced back to an earlier form *magisomos (Wallace 2020: 90 n. 8; cf. Ernout & Meillet 1959: 377-379; Walde & Hofmann 1938-1954: II, 14; de Vaan 2008: 358-359), is particularly interesting as it seems to show at the same time the deletion of -i- (*magisomos > *magsomos) and the preservation of the medial-syllable vowel -o-. In the light of this, it may be speculated that also in Faliscan there would be a tendency towards rhythmic compensation (§2.1) - which would allow to explain the deletion of -iin *magisomos - with partially different results with respect to Latin. In any case, such a tendency would post-date the earliest documents, if the form pe⁝parai, attested in the so-called Ceres-inscription (7 th century BC), is a reduplicative perfect with preservation of the medial syllable vowel -acorresponding to Latin peperi ‘I gave birth’ (Bakkum 2009: 169, 192), where the medial syllable vowel -eis the expected outcome of vowel reduction. 3.2 The Sabellic languages The reconstruction of vowel reduction and deletion processes in the Sabellic languages is made difficult by the fragmentary nature of their documentation. Since the beginning of their respective writing traditions, the Sabellic languages Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? 257 <?page no="258"?> 16 ST Um 4 = ImIt Caere 1. See Rix (1992). Sabellic inscriptions are cited according to Rix (2002) = ST and Crawford (2011) = ImIt. 17 ST Um 1 (II a 18, 28, IV 4 struhçla, III 34 struçla, II a 41, IV 1 struhçlas, VI a 59 strusla, VI b 5, 23, VII a 8, 42, 54 struśla). 18 Nishimura (2012: 387). See also Nishimura (2016) on syncope of u-vocalism in Sabellic. 19 ST Ps 20 = ImIt Blanda 1. For the dating, see Lazzarini & Poccetti (2001: 16). 20 For an overview, see Nishimura (2012: 381-386) and Zair (2016: 300-312). 21 ST Lu 1 = ImIt Bantia 1 (line 7). 22 ST Lu 1 = ImIt Bantia 1 (line 4). 23 ST Lu 1 = ImIt Bantia 1 (line 7). 24 ST Um 1 (Ib 27). 25 ST Ps 1 = ImIt Nerulum 1. 26 Rix (1997: 146-147). See, however, the remarks on the correction διποτερε<ς> for διποτερεμ by Crawford (2011: 1341). show the deletion of short vowels before s in final syllables, which may therefore be assumed as a Protoor at least Common-Sabellic phenomenon (Clackson 2015: 10): see, for example, setums < *septumos ‘Setums (male anthroponym)’ in a Palaeo-Sabellic inscription of the 7 th century BC 16 . On the contrary, vowel deletion in medial syllables probably occurred independently in the different Sabellic languages, as it applies after language-specific sound changes: see, for example, Umbrian struhçla 17 ‘sacrificial pastry’ < *struu̯ikelā, where the deletion of *-eoccurred after the palatalization of the preceding velar stop, indicated by the use of the letter <ç> 18 . However, vowel deletion in medial syllables seems also to have occurred in archaic varieties (Nishimura 2012: 388-389; cf. also Lipp 2021): see, for example, Pre-Samnite (σ)τα[.]ιοσϙτοδ in the Tortora inscription (end of the 6 th century BC 19 ), which may be analysed as a so-called future imperative (-τοδ < *-tōd) of a *-sk-e/ overbal base with deletion of -e- (-σϙ- < *-ske-; Lazzarini & Poccetti 2001: 143, 173). In the different Sabellic languages vowel reduction seems to affect different vowels, to be triggered by different phonetic contexts, and to have different outcomes 20 . For example, in Oscan -a-, -e-, and -ochanged into -uin medial open syllables before or after a labial sound (see, for instance, pertumum  21 ‘to prevent’ < *pertemom; cf. pertemust  22 ‘(he/ she/ it) will have prevented’, per‐ temest  23 ‘(he/ she/ it) will prevent’; Zair 2016: 302-203), whereas in Umbrian -ain medial syllables changed into -o- (see, for instance, the female theonym prestota < *praistata-; cf. prestate 24 ; Zair 2016: 303-306). Such phenomena seem to be relatively late both in Oscan and Umbrian (Zair 2016: 299, 311-312). As for the archaic varieties, the presence of vowel reduction is not ascertained (Nishimura 2012: 383-386): a possible example could be Pre-Samnite διποτερε<ς> 25 ‘Jupiter’ (beginning of the 5 th century BC), if -ποτερis the outcome of *-pateras supposed by Rix 26 . 258 Luca Rigobianco <?page no="259"?> 27 ST Sa 1 (A 16, B 19) = ImIT Terventum 34 (A 16, B 19). 28 Thurneysen (1909). See Nishimura (2014: 186) for an analysis of the distribution of <Vh>, <VhV>, and <VV> as an orthographic device for [V̄] in Umbrian. The use of alliteration in archaic Italic poetic texts is also commonly considered an indication of a first-syllable stress: for an overview, see Nishimura (2014: 168 n. 23). 29 ET 2 Ve 3.11, 3.29 (avi]le), Cr 2.99 (avilesca), 3.28 (ạvịḷe), Ta 7.18, 7.35 (avi[le), AT 2.9, Vs 1.26, 1.84, 1.100 (ạviles), 1.165, 1.166, Vc 1.78 (av(i)les), 2.71, 3.4, 3.5 (av[ile), 0.23, Ru 2.1 ([a]vileś), Vt 1.154 (aviḷeś), AS 2.1 (avil[eś), 2.14, Pe 1.115 (ạviles), Fs 1.4, 1.5, 0.3 (aṿịḷẹṣ́ ), Pa 1.1, OA 2.6. Etruscan inscriptions are cited according to Meiser (2014) = ET 2 . The notation of [s] as either <s> (sigma) or <ś> (san) depends on the orthographic habits of the different areas of Etruria. 30 ET 2 Cr 3.23. 31 ET 2 Cm 2.127, Ve 2.1, Vs 1.5, 1.13, 1.33 (avele{-}s), 1.38, 1.39, 1.45, 1.56, 1.121 (aṿ[e]les), 1.137 (aveḷes), 1.328, 1.344, AV 1.23 ([a]veles), Po 2.5 (ave(le)), Cl 2.15 (avel[e), OA 3.1 (aṿẹ[lesi), 6.1. 32 ET 2 Cr 1.63, AV 2.17. 33 For the many occurrences of avle/ aule, see the index in Meiser (2014). 34 For the many occurrences of muluvanice/ muluvanike, see the index in Meiser (2014). The notation of [k] as either <c> or <k> depends on the orthographic habits of the different areas of Etruria. Summing up, the Sabellic languages probably shared a first-syllable stress and the tendency towards rhythmic compensation (cf. §2.1). Such a tendency led to vowel reduction and deletion phenomena in non-initial syllables, which mostly occurred separately in the single languages. Besides, the hypothesis of a first-syllable stress at least in Oscan is supported by the occurrence almost exclusively in the first syllable of a strategy for signalling vowel length (see, for example, aasaí 27 ‘at the altar’, with aafor [a: ] 28 ). 3.3 Etruscan The studies concerning the Etruscan stress system are extremely limited and, to date, there is no attempt to describe its functioning on the basis of an examination of the entire corpus of inscriptions. Nevertheless, a first-syllable stress is commonly reconstructed for Etruscan (Rix 2004: 949). Such a stress as well as the tendency towards rhythmic compensation (§2.1) would be responsible for the vowel reduction and then deletion attested in non-initial syllables: see, for example, avile  29 ‘Avile (male praenomen)’ > av[ǝ]le (which could be written as avile, avale  30 , avele  31 , and avule  32 ) > avle/ aule  33 . Vowel deletion is particularly extensive and has been regularly written down since the beginning of the 5 th century BC (Rix 2004: 950). However, evidence of this phenomenon may be traced back to the 7 th century BC. Thus, for example, alongside the form muluvanice/ muluvanike  34 ‘gave’, there are also attested the Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? 259 <?page no="260"?> 35 ET 2 Vt 3.5. 36 ET 2 Cl 3.2 (muluṿẹneke). 37 ET 2 Cr 3.27 (mu]luvnice). 38 ET 2 Ve 3.18 (mulṿanice), Cr 3.11, Cr 3.15. 39 ET 2 Ru 3.1 (mulvaniḳe). 40 ET 2 Cr 3.14. The gemination of -n-, which also occurs in the form θanursiannaŝ in the same inscription, remains to be clarified. 41 ET 2 Cl 2.3 (ṃụḷvenike). 42 ET 2 Vc 3.3. 43 See also Rigobianco (2017). In any case, the innovative character of the tendency towards rhythmic compensation in Etruscan is confirmed by the absence of vowel reduction and deletion in Rhaetic, a language phylogenetically related to Etruscan (see Schumacher 2004: 316-317). 44 For the many occurrences of seχ, see the index in Meiser (2014). 45 ET 2 Cr 3.33, Vs 3.9, Pe 1.21 (śeχis), 1.1101 (seχiś). 46 The reconstruction of *-si as the original genitive ending is based on the palatalisation of the stem-vowel in the ablative case (see, for example, nominative-accusative rasna ‘army (? ), people (? )’, genitive rasnas, ablative raśneś), whose ending would arise from the addition to the genitive ending *-si - with -i responsible for the palatalisation - of a further morpheme *-sV (raśneś < *rasna-si-sV); see Rix (2004, 952-953). forms muluvunike  35 (end of the 7 th century BC) with reduction of -ato [ǝ] written as <u>, muluveneke  36 (last quarter of the 7 th century BC) with reduction of -aand -ito [ǝ] written as <e>, muluvnice  37 (last quarter of the 7 th century BC) with deletion of -a-, mulvanice  38 (end of the 7 th / beginning of the 6 th century BC)/ mulvanike  39 (second half of the 7 th century BC) and mulvannice  40 (end of the 7 th / beginning of the 6 th century BC) with deletion of -u-, mulvenike  41 (third quarter of the 7 th century BC) with deletion of -uand reduction of -ato [ǝ] written as <e>, as well as mulvenece  42 (end of the 7 th / beginning of the 6 th century BC) with deletion of -uand reduction of -aand -ito [ǝ] written as <e>. It seems therefore conceivable that the spread of the notation of vowel deletion in the beginning of the 5 th century BC is the result of an orthographic reform by writing schools aimed at fixing in the spelling a phonetic change dating back, at least in some varieties of Etruscan, to over a century earlier. Furthermore, it should be noted that the first-syllable stress might be an innovation in Etruscan. Specifically, on the basis of a prehistoric phenomenon of apocope reconstructed by Rix (1984: 204; 1989: 173-182), Prosdocimi (1986: 612-613) has hypothesised that the Etruscan stress system was originally characterized by a penultimate-syllable stress 43 . Accordingly, for example, the nominative-accusative form seχ  44 ‘daughter’ would be the outcome of a previous *séχi, and the corresponding genitive form seχis  45 of a previous *seχísi  46 . 260 Luca Rigobianco <?page no="261"?> 47 See, for example, Skutsch (1913: 196). For an overview, see Leumann (1977: 247-248). 48 On such a tendency, see Prosdocimi (1995: 41-42). 4 Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? Phenomena of vowel reduction and deletion occurred in Archaic Latin between the earliest epigraphic records (7 th -6 th century BC) and the beginning of the literary tradition (3 rd century BC) are similar to vowel reduction and deletion occurred in Etruscan and in the Sabellic languages since the earliest epigraphic records (7 th century BC). In particular, all these phenomena reflect a tendency towards rhythmic compensation and affect non-initial syllables, thus implying a first-syllable stress (cf. §2.1). Since the beginning of the 20 th century, the sharing of a first-syllable stress between Latin, Sabellic languages, and Etruscan has commonly been considered as a feature which would have spread by contact 47 . Specifically, Etruscan has been identified as the source language. Such a proposal seems to be based, on the one hand, on the past tendency to exaggerate the influence of Etruscan on Latin and the Sabellic languages 48 , and, on the other, on the pervasiveness of vowel deletion in Etruscan after the beginning of the 5 th century BC (cf. §3.3). However, in case first-syllable stress and tendency towards rhythmic compensation are contact-induced, it cannot be excluded a priori that the strength of such a tendency depends on a language-specific development and may in principle be greater in the target-language. Recently, the hypothesis of a contact-induced feature has been proposed, though in general terms, by Zair (2016). Conversely, this hypothesis has been questioned several times by Nishimura (2012: 392-394; 2014: 168-169) in his works on the stress system and related vowel reduction and deletion phenomena in Latin and the Sabellic languages. Nishimura holds that an Etruscan influence should be ruled out because the earliest phenomena of vowel reduction and de‐ letion in the Sabellic languages would precede similar phenomena in Etruscan. However, as shown above (§3.3), traces of vowel deletion in Etruscan may be seen as early as the second half of the 7 th century BC. Changes in prosodic features as a result of contact are not unexpected. As noticed by Salmons (1992: 25-27), there are several direct and indirect pieces of evidence of such changes (cf. van der Hulst & Goedemans & Rice 2017). This is the case, for example, of the Zaonežje dialect of Russian, in which stress, when on the final syllable, was shifted onto the initial syllable following contact with Karelian, a Uralic language which has a fixed initial-syllable stress: see, for example, Zaonežje dialect žýv’æt (he/ she/ it) lives’ beside standard Russian živ’ót Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? 261 <?page no="262"?> 49 See, for example, Thomason & Kaufman (1988: 246), who consider it “an erroneous assumption that an exact correspondence between source-language structures and target-language structures is to be expected”. 50 A comprehensive review of such loanwords is still missing. For example, one may cite the male praenomina mamarce, puplie-, and kavie (cf. Latin Māmercus, Publius, Gāius; Rix 1995: 723), the theonyms uni, menerva, and neθuns (cf. Latin Iūnō, Minerua, and Neptūnus; Rix 1981), the vase name putlum- (cf. Latin pōculum; Prosdocimi 1979b: 159-161), the institutional term macstre- (cf. Latin magister; Maggiani 1996: 114), and the kinship term nefts ‘nephew’ (cf. Latin nepōs; Maggiani 2019). (Pronk 2018: 555). Furthermore, in several geographic areas, phylogenetically unrelated languages show similar stress systems (Salmons 1992: 43-49). In any case, it should be noted that, as in the case of the just cited Zaonežje dialect, changes in stress system as a result of contact do not entail that the receiving language acquires the whole stress system of the source language (Pronk 2018: 564-567). More generally, such a lack of exact correspondence applies to any contact-induced change 49 . In light of this, taking into-account the similarity of the phenomena - namely, the vowel reduction and deletion in non-initial syllables -, their chronology, the geographic proximity, the lack of phylogenetic relatedness between Etruscan on the one hand and Latin and Sabellic on the other, as well as the existence of interference phenomena between them (see below), it seems reasonable to assume these phenomena to be contact-induced. Hence, the problem arises to reconstruct their origin and development. Theoretically, two broad scenarios are possible: a contact-induced phenomenon «in language maintenance» or a contact-induced phenomenon «in language shift». In the first scenario, we should assume an «intensive contact, including much bilingualism among borrowing-language speakers over a long period of time», in the second, a «large shifting group and imperfect learning» (Thomason & Kaufman 1988: 50). On the basis of linguistic and archaeological evidence, a scenario similar to the former could be reconstructed for the Etruscan and Sabellic ethne in the centuries immediately before the beginning of their respective writing traditions, approximately between the 10 th and the 8 th century BC. Several language-contact phenomena attributable to either the Sabellic languages or Latin are attested in Etruscan since the earliest times. First of all, numerous loanwords may be identified - such as anthroponyms, theonyms, technical terms, institutional terms, and kinship terms - 50 . For some of these, Rix (1981: 123-125) has shown that there are formal reasons for considering more probable a Sabellic rather than a Latin origin. Furthemore, derivational morphemes of probable or at least possible Sabellic and/ or Latin origin may be identified as well. Specifically, such morphemes, mostly attested since the beginning of the 262 Luca Rigobianco <?page no="263"?> 51 ET 2 AV 1.29 (6 th century BC). 52 ET 2 Pa 1.2 (second quarter of the 6 th century BC). 53 ET 2 Fa 2.5 (archaic), Ve 3.47 (last quarter of the 7 th century BC), Cr 2.36 (third quarter of the 7 th century BC), 2.80 (end of the 6 th / beginning of the 5 th century BC), 2.158 (end of the 7 th century BC), 7.1 (third quarter of the 7 th century BC), AT 2.1 (second half of the 7 th century BC), 2.2 (veleliia[s; 7 th century BC), Vs 1.66 (end of the 6 th / beginning of the 5 th century BC), 1.85 (end of the 6 th / beginning of the 5 th century BC), 1.112 (veleliạ[s; end of the 6 th / beginning of the 5 th century BC). 54 ET 2 Cm 3.2 (third quarter of the 7 th century BC). 55 van Heems 2008, pp. 86-87 (such a morpheme is attested only in late Etruscan). 56 ET 2 Cl 1.356, 1.1146, 1.1148, 1.1674 (venzileṣ́ ), 1.2366 (venzi[le). 57 ET 2 Co 3.4, 4.11 (culśanś[). 58 ET 2 Vs 1.81 (6 th -5 th century BC), Vn 3.2 (l[a]uχu[s]hie; end of the 7 th / beginning of the 6 th century BC), Vt 1.71 (end of the 6 th century BC). Etruscan writing tradition, are: -θe/ -te (cf. Proto-Italic *-ti-), for deriving adjec‐ tives from nouns, in particular ethnics from toponyms (Steinbauer 1999: 126-127; see, for example, kaiseriθe-  51 ‘Kaiseriθe (male gentilicium); lit. Caeretan’); -i (cf. Proto-Italic *-ī), for deriving female gentilicia (Rigobianco 2013: 139-146; see, for example, puleisnai  52 ‘Puleisnai (female gentilicium)’); -ia (cf. Proto-Italic *-jā), for deriving female anthroponyms (Rigobianco 2013: 171-178; see, for example, velelia(-)  53 ‘Velelia (female praenomen)’); -ie (cf. Proto-Italic *-jo-), for deriving adjectives, in particular gentilicia, from nouns (De Simone 1989: 271-275; see, for example, rasunie-  54 ‘Rasunie (male gentilicium); lit. military (? ), popular (? )’); -le (cf. Proto-Italic *-lo-), for deriving hypocoristics 55 (see, for example, venzile-  56 ‘Venzile (male gentilicium)’); -ns (cf. Proto-Italic *-no-s), for deriving theonyms (Rix 1998: 213-216, 222; see, for example, culśanś-  57 ‘Culsans (theonym)’); -sie (cf. Proto-Italic *-sjo-? ), for deriving praenomina (De Simone 2006: 126-131; see, for example, lauχusie-  58 ‘Lauχusie (male praenomen)’). Accordingly, the contact between Etruscan and the Sabellic languages and/ or Latin seems to be, in Thomason’s terminology, a «[m]ore intense contact» in a four-degree scale which goes from a «[c]asual contact» and arrives, through a «[s]lightly more intense contact» and then a «[m]ore intense contact», to an «[i]ntense contact» (Thomason 2001: 70-71). In particular, such a «[m]ore intense contact» entails the possible borrowing of structural features such as prosodic features. From a historical point of view, such a linguistic situation may be explained in the light of the processes of formation of proto-urban centres in the 9 th century BC by Etruscan populations with the participation of Sabellic groups (see, for example, Maggiani 2012: 402). However, as already mentioned, vowel reduction and deletion in Latin are evidently later than in the Sabellic languages and Etruscan (cf. §§2, 3.2, 3.3). Furthermore, apart from some Etruscan and Sabellic loanwords, there is no Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? 263 <?page no="264"?> 59 On the contact between Latin, on the one hand, and Oscan, Umbrian, and Etruscan, on the other, see, for example, Adams (2004: 112-184). 60 On language shift as a trigger for phonological reshaping of the secondary language on the model of the primary language, see also Ross (2004: 191-193). 61 Furthermore, some Etruscan inscriptions from the archaic period have been found in Rome and Latium. 62 Liv. 2.16.3-4. 63 I thank Professor Cuzzolin for this remark. 64 I thank an anonymous reviewer for pointing this out. decisive evidence in Latin of contact-induced phenomena involving much Latin-Sabellic or Latin-Etruscan bilingualism over a long period of time 59 . Nevertheless, the rhythmic tendency towards compensation which led to vowel reduction and deletion in Latin could be a contact-induced phenomenon due to large Etruscan and/ or Sabellic speaking groups shifting to Latin with consequent imperfect learning. In this regard, as noticed by Thomason and Kaufman (1988: 39), «unlike borrowing, interference through imperfect learning does not begin with vocabulary: it begins instead with sounds and syntax» 60 . From an historical point of view, such a scenario seems entirely plausible, taking into account, for example, the period of Etruscan dominance in the 6 th century BC and the presence in Rome of a vicus Tuscus (cf. Colonna 1987) 61 , as well as the migration to Rome of Sabellic groups, such as on the occasion of the arrival of Attius Clausus from the Sabine territory at the end of the 6 th century BC (“Attius Clausus, cui postea Appio Claudio fuit Romae nomen, […] ab Inregillo, magna clientium comitatus manu, Romam transfugit” 62 ). In such processes, the evident prestige of the imperfect learners in the Roman society may have played an important role in promoting an imitation by the original speakers of Latin. One might object that vowel reduction and deletion are natural phenomena which are widespread in the world’s languages, so it cannot be excluded a priori that these phenomena occurred independently in Latin 63 . However, beyond the general difficulty of establishing whether a linguistic change is contact-induced (cf. Thomason 2001: 91-95), such an objection dissolves, at least in part, once we assume that internal and external factors often jointly contribute to determining a linguistic change (cf. Thomason 2001: 61-63, Joseph 2002: 51-55) and a natural tendency may be reinforced by external pressure, as in the case of contact-induced grammaticalization (Heine & Kuteva 2003) 64 . In the specific case, linguistic as well as historical considerations (see above) firmly point to contact being the reason or at least one of the reasons which would have induced the change in the rhythmic pattern. 264 Luca Rigobianco <?page no="265"?> 65 In this regard, it should be noted that, assuming a tendency towards rhythmic compen‐ sation borrowed from Etruscan and/ or the Sabellic languages, phenomena of vowel reduction and deletion in Latin may also be explained if the hypothesis of a Proto-Italic first-syllable accent is rejected and an archaic Latin stress system as reconstructed by Ballester (1990) is accepted. Specifically, according to Ballester, the Penultimate Law would have been preceded by a stress placement on the leftmost long vowel or, if the word does not contain any long vowel, on the leftmost syllable. 5 Conclusions On the basis of what has been seen above, it seems possible to reconstruct the following scenario. In late Proto-Italic, the Proto-Indo-European mobile accent would have been replaced by a first-syllable stress (cf. §2.1). Subsequently, Proto-Sabellic would have developed a tendency towards rhythmic compensa‐ tion, which would have been responsible for the deletion of short vowels before s in final syllables and, then, for vowel reduction and deletion in non-initial syllables, which occurred independently in the different Sabellic languages (cf. §3.2). As for Etruscan, the original penultimate-syllable stress, responsible of a prehistoric phenomenon of apocope, would have been replaced by a first-syllable stress. Together with a tendency towards rhythmic compensation, this new stress would be responsible for vowel reduction and then deletion in non-initial syllables attested since the second half of the 7 th century BC and regularly written down since the beginning of the 5 th century BC (cf. §3.3). Both stress position and rhythmic pattern would be the outcome of a contact-induced change following a widespread Etruscan-Sabellic bilingualism among Etruscan speakers, which is also reflected by the numerous Etruscan words and derivational morphemes borrowed from the Italic languages (§4). Latin would have inherited a first-syllable stress from late Proto-Italic (§2.1). The vowel reduction and deletion phenomena in non-initial syllable, attested only after the earliest texts (§2), would be due to the interaction between the first-syllable stress, inherited from Proto-Italic, and an innovative tendency towards rhythmic compensation. Such a tendency would have spread in Latin following the imperfect learning of large Etruscan and/ or Sabellic speaking groups shifting to Latin 65 . 6 Bibliographical references Adamik, B. (2015) “The Periodization of Latin. An Old Question Revisited”, in Haverling, G. V. M. (ed.) Latin Linguistics in the Early 21 st Century, Uppsala, 638-650. Adams, J.N. (2004) Bilingualism and the Latin Language, Cambridge. Bakkum, G.C. L. M. (2009) The Latin Dialect of the Ager Faliscus, Amsterdam. Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? 265 <?page no="266"?> Ballester, X. (1990) “La posición del acento prehistórico latino”, Emerita 58: 33-50. Ballester, X. (1996) “La tipología y el acento prehistórico latino”, Emerita 64: 59-63. Bertinetto, P.M. (1989) “Reflections on the dichotomy «stress» vs. «syllable-timing»”, Revue de Phonétique Appliquée 91/ 93: 99-130. Bertinetto, P.M. & C. Bertini (2009) “On modeling the rhythm of natural languages”, in Barbosa, P. et al. (eds.) Speech Prosody 2008. Fourth Conference on Speech Prosody. May 6-9, Campinas, Brazil. Proceedings, Campinas, 427-430. Bertinetto, P.M. & C. Bertini (2010) “Towards a unified predictive model of Natural Language Rhythm”, in Russo, M. (ed.) Prosodic Universals. Comparative Studies in Rhythmic Modeling and Rhythm Typology, Roma, 43-78. Clackson, J. (2015) “Subgrouping in the Sabellian branch of Indo-European”, Transactions of the Philological Society 113: 4-37. Coarelli, F. (1983) Il Foro Romano. Periodo arcaico, Roma. Colonna, G. (1978) “Etruria e Lazio nell’età dei Tarquini”, in Cristofani, M. (ed.) Etruria e Lazio arcaico, Roma, 55-66. Colonna, G. (1979) “Duenos”, Studi Etruschi 47: 163-172. Colonna, G. (1980a) “Graeco more bibere: l’iscrizione della tomba 115 dell’Osteria dell’Osa”, Archeologia laziale. III, Roma, 51-55. Colonna, G. (1980b) “L’aspetto epigrafico”, in Stibbe, C.M. et al. (eds.) Lapis Satricanus. Archaeological, epigraphical, linguistic and historical aspects of the new inscription from Satricum, The Hague, 41-69. Colonna, G. (1999) “Epigrafi etrusche e latine a confronto”, in XI Congresso Internazionale di Epigrafia Greca e Latina. Roma, 18-24 settembre 1997. Atti. I. Roma, 435-450. Crawford, M. H. (2011) Imagines Italicae, London. De Simone, C. (1989) “Etrusco Acvilna ~ latino Aquilius. Un problema di intercambio onomastico”, La Parola del Passato 44: 263-280. De Simone, C. (2006) “I ‘Rossi’ in Etruria: il nome dei Rutuli”, Incidenza dell’Antico 4-: 111-139. de Vaan, M. (2008) Etymological Dictionary of Latin and the other Italic Languages, Leiden & Boston. Ernout, A. & A. Meillet (1959) Dictionnaire étymologique de la langue latine, Paris. Filipponio, L. (2012) “Reflexionen über die Diachronie des galloitalienischen Rhythmus”, in Schafroth, E. & M. Selig (eds.) Testo e ritmi. Zum Rhythmus in der italienischen Sprache, Frankfurt am Main, 61-87. Flemming, E. (2005) A Phonetically-Based Model of Phonological Vowel Reduction, manu‐ script, https: / / web.mit.edu/ flemming/ www/ paper/ vowelred.pdf. Fortson, B.W. (2011) “Latin Prosody and Metrics”, in Clackson, J. (ed.) A companion to the Latin language, Oxford, 92-104. Giacomelli, R. (2006) Nuove ricerche falische, Roma. 266 Luca Rigobianco <?page no="267"?> Gnade, M. & G. Colonna (2003) “Dolio con iscrizioni latine arcaiche da Satricum”, Archeologia classica 54: 1-21. Joseph, B. D. (2002) “Language Contact and Historical Linguistics”, in Mufwene, S. S. & A.M. Escobar (eds.) The Cambridge Handbook of Language Contact. Volume 1, Cambridge, 43-63. Heine, B. & T. Kuteva (2003) “On contact-induced grammaticalization”, Studies in Language 27: 529-572. Höfler, S. (2017) “Observations on the palma rule”, Pallas 103: 15-23. Lazzarini, M.L. & P. Poccetti (2001) Il mondo enotrio tra VI e V secolo a.C. Atti dei seminari napoletani (1996-1998), Napoli. Leumann, M. (1977) Lateinische Laut- und Formen-lehre, München. Lipp, R. (2021) “The medial syllable syncope in South Picene inscriptions”, in Tarsi, M. (ed.) Studies in General and Historical Linguistics Offered to Jón Axel Harðarson on the Occasion of his 65 th Birthday , Innsbruck, 269-328. Maggiani, A. (1996) “Appunti sulle magistrature etrusche”, Studi Etruschi 62: 95-138. Maggiani, A. (2012) “La religione”, in Bartoloni, G. (ed.) Introduzione all’Etruscologia, Milano, 395-418. Maggiani, A. (2019) “Papals, nefts, prumts. Termini di parentela in etrusco. Due nuove proposte di lettura”, Studi Etruschi 82: 145-167. Mancini, M. (2021) “Does Prenestinian fe⋮faked actually exist? ”, Journal of Latin Linguis‐ tics 20: 75-108. Maras, D.F. (2015) “La prima stesura dell’iscrizione di Manios e l’uso epigrafico dell’in‐ terpunzione espuntiva”, Bullettino di Paletnologia italiana 99: 113-122. Meiser, G. (1998) Historische Laut- und Formenlehre der lateinischen Sprache, Darmstadt. Meiser, G. (2014) Etruskische Texte. Editio Minor, Hamburg. Nishimura, K. (2010a) “Chronology between word-internal and word-final vowel reduc‐ tion in Latin”, Incontri linguistici 33: 167-177. Nishimura, K. (2010b) “Patterns of Vowel Reduction in Latin: Phonetics and Phonology”, Historische Sprachforschung 123: 217-257. Nishimura, K. (2011) “Syncope in Latin: a Historical Restatement of its Conditions”, Internation Journal of Diachronic Linguistics and Linguistic Reconstruction 8: 1-41. Nishimura, K. (2012) “Vowel reduction and deletion in Sabellic: a synchronic and diachronic interface”, in Whitehead, B.N. et al. (eds.), 381-398. Nishimura, K. (2014) “On accent in the Italic languages: nature, position, and history”, Studia Linguistica Universitatis Iagellonicae Cracoviensis 131: 161-192. Nishimura, K. (2016) “On syncope of u-vocalism in Sabellic”, Indogermanische For‐ schungen 121: 199-211. Ohanessian, M. (2020) “Prefijación y apofonía en latín”, Lingue antiche e moderne 9: 185-204. Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? 267 <?page no="268"?> Oniga, R. (1990) “L’apofonia nei composti e l’ipotesi dell’‘intensità’ in latino”, in Danese, R. et al. (eds.) Metrica classica e linguistica, Urbino, 195-236. Oniga, R. (2006) “Un’eccezione all’apofonia latina”, in Bombi, R. et al. (eds.) Studi linguistici in onore di Roberto Gusmani, Alessandria, 1337-1339. Pike, K. L. (1945) The intonation of American English, Ann Arbor. Pronk, T. (2018) “Language contact and prosodic change in Slavic and Baltic”, Diachronica 35: 552-580. Prosdocimi, A.L. (1979a) “Studi sul latino arcaico”, Studi Etruschi 47: 173-221. Prosdocimi, A.L. (1979b) “Le iscrizioni italiche. Acquisizioni, temi, problemi”, Le iscrizioni prelatine in Italia, Roma, 119-204. Prosdocimi, A.L. (1986) “Sull’accento latino e italico”, in Etter, A. (ed.) o-o-pe-ro-si. Festschrift für Ernst Risch zum 75. Geburtstag, Berlin & New York, 601-618. Prosdocimi, A.L. (1995) “Filoni indoeuropei in Italia. Riflessioni e appunti”, in Landi, A. (ed.) L’Italia e il Mediterraneo antico antico, Pisa, volume II, 7-163. Pultrová, L. (2006) The vocalism of Latin medial syllables, Praha. Pultrová, L. (2011) “From the Proto-Indo-European to the Classical Latin accent”, Listy filologické 134: 219-243. Rigobianco, L. (2013) Su numerus, genus e sexus., Elementi per una grammatica dell’etrusco, Roma. Rigobianco, L. (2017) “Per una grammatica dell’etrusco. Considerazioni morfonologiche sulla derivazione di nomi e aggettivi in etrusco arcaico”, Mediterranea 14: 185-203. Rix, H. (1966) “Die lateinische Synkope als historisches und phonologisches Problem”, Kratylos 11: 156-165. Rix, H. (1981) “Rapporti onomastici fra il panteon etrusco e quello romano”, Gli Etruschi e Roma, Roma, 104-126. Rix, H. (1984) “La scrittura e la lingua”, in Cristofani, M. (ed.) Gli Etruschi. Una nuova immagine, Firenze, 199-227. Rix, H. (1989) “Zur Morphostruktur des etruskischen s-Genetivs”, Studi Etruschi 55: 169-193. Rix, H. (1992) “Una firma paleo-umbra”, Archivio Glottologico Italiano 67: 243-252. Rix, H. (1995) “Etruskische Personennamen”, in Eichler, Ernst et al. (eds.) Namenfor‐ schung / Name Studies / Les noms propres. 1. Halbband, Berlin & New York, 719-724. Rix, H. (1997) “The Pre-Lucanian Inscriptions of Southern Italy”, in Adams, D. (ed.) Festschrift for Eric P. Hamp. Volume II, Washington D.C., 144-154. Rix, H. (1998) “Teonimi etruschi e teonimi italici”, Annali della Fondazione per il Museo «Claudio Faina» 5: 207-229. Rix, H. (2002) Sabellische Texte, Heidelberg. Rix, H. (2004) “Etruscan”, in Woodard, R.D. (ed.) The Cambridge Encyclopedia of the World’s Ancient Languages, Cambridge, 943-966. 268 Luca Rigobianco <?page no="269"?> Ross, M. (2004) “Diagnosing prehistoric language contact”, in Hickey, R. (ed.) Motives for language change, Cambridge, 174-198. Salmons, J. (1992) Accentual Change and Language Contact, London. Schumacher, S. (2004) Die rätischen Inschriften, Innsbruck Skutsch, F. (1913) “Der lateinische Accent”, Glotta 4: 187-200. Steinbauer, D.H. (1999) Neues Handbuch des Etruskischen, St. Katharinen. Thomason, S. G. & T. Kaufman (1989) Language Contact, Creolization, and Genetic Linguistics, Berkeley Thomason, S. G. (2001) Language contact, Edinburgh. Thurneysen, R. (1909) “Italisches. 1. Die Betonung des Oskischen”, Glotta 1: 240-242. Untermann, J. (2000) Wörterbuch des Oskisch-Umbrischen, Heidelberg. van der Hulst, H. & R. Goedemans & K. Rice (2017) “Word Prominence and Areal Lin‐ guistics”, in Hickey, R. (ed.) The Cambridge Handbook of Areal Linguistics, Cambridge, 161-203. van Heems, G. (2008) “Diminutifs, sobriquets et hypochoristiques étrusques”, in Poccetti, P. (ed.) Les prénoms de l’Italie antique, Pisa & Roma, 69-110. Vine, B. (1993) Studies in Archaic Latin Inscriptions, Innsbruck. Vine, B. (2012) “PIE mobile accent in Italic: Further evidence”, Whitehead, B.N. et al. (eds.), 545-575. Walde, A. & J.B. Hofmann (1938-1954) Lateinisches etymologisches Wörterbuch, Heidel‐ berg. Weiss, M. (2020) Outline of the Historical and Comparative Grammar of Latin. Second Edition, Ann Arbor & New York. Whitehead, B.N. et al. (eds.) (2012) The sound of Indo-European. Phonetics, phonemics, and morphophonemics, København. Zair, N. (2016) “Vowel weakening in the Sabellic languages as language contact”, Indogermanische Forschungen 121: 295-316. Vowel reduction and deletion in Archaic Latin: contact-induced phenomena? 269 <?page no="271"?> 1 Mein Schüler und Freund Herr Dr. Marcus Heckenkamp hat meine deutsche Fassung korrigiert und verbessert, ihm sei mein herzlichster Dank ausgesprochen. Nausistrata (Ter. Phorm. 784-1055) 1 Gualtiero Calboli (Universität Bologna, Alma Mater Studiorum) Meinem Schüler und Freund dem Oberstudiendirektor Hansbert Bertsch in memoriam Abstract: In this article I took into account a character, namely Nau‐ sistrata, the wife of Chremes and the mother of the young Phaedria. Nausistrata is the owner of a farm in the island of Lemnos which she inherited from her father. She says (789-791) that her father brought from Lemnos two talents pro year, while her husband Chremes reduced a lot this annuity (vir viro quid praestat! What a difference between the two men! ). Thus, we know from Nausistrata two important notices, namely that her father was one of the rich landowners who went from Athens with the other κληροῦκοι in the IV cent. BC and enlarged his own farm by buying the farms of some other poor farmers. Lemnos produced oil, vine and wheat and acted as middleman in the importation of wheat from Skythia to Athens (as today the Ucraine). In the III cent. BC, this activity was strong reduced, and this corresponds to the fact deplored by Nausistrata that Chremes, her husband, brought from Lemnos much less than her father. But in the comedy Nausistrata is infomed by the parasit Phormio that Chremes in Lemnos had another wife and another family, and Nausistrata ascribes the reduction of the rents from Lemnos to this situation of Chremes. The model of Terence’s Phormio was the Ἐπιδικαζόμενος of Apollodorus, a pupil of Menander, a comedy we have not. Therefore, Eckard Lefèvre, an authority on Terence, supposed that the whole character of Nausistrata was invented by Terence. In <?page no="272"?> this article I tried to demonstrate that, on the contrary, Nausistrata was already in Apollodoros’ comedy, not only because of some specific notices (Lemnos’ crisis in III century could be known also in Rome), but because some peculiarities as the extension of the latifundia are more specific and connected with the development of the plot: the reduction of the rents from Lemnos as caused by the double family of Chremes and the consequent Nausistrata’s anger. On his side, Terence added some other features from Plautus to Nausistrata starting from this point and made her a kind of uxor saeua as we find at the conclusion of the comedy. Anyhow, we find a strong connection of the parts where a fundamental position is played by Nausistrata who cannot disappear in the Greek model. This seems to be a condition of the whole Roman culture: Greek culture cannot disappear. Keywords: Lemnos; Athenian colonization; Latin comedy and Greek models; uxor saeua and Plautus. Ich habe schon den Einfluss des Griechischen auf die lateinische Syntax un‐ tersucht (Calboli 2009) und jetzt möchte ich meine Aufmerksamkeit auf einen anderen Einfluss richten, der eine literarische Figur insgesamt betrifft, näm‐ lich die Ehefrau des Phormio des Terenz, Nausistrata. Man weiß, dass in den dem Terenz gewidmeten Untersuchungen ein wichtiges Untersuchungsfeld, vielleicht das wichtigste, darin liegt, zu unterscheiden, was sich an das griechi‐ sche Vorbild, sei es von Menander oder Apollodor, anlehnt und was auf Terenz selbst zurückgeht. Auch die Komödie des Phormio hat sich dieser Gewohnheit nicht entzogen. Das entsprechende griechische Modell ist die Didaskalie nach Apollodoros’ Komödie Epidicazomenos gewesen, wobei es umstritten ist, ob das schon gegebene Urteil, nach dem der junge Antiphon gezwungen wurde, seine geliebte Phanium zu heiraten, wiederhergestellt werden darf oder nicht. Phanium auf der anderen Seite ist die Tochter von Antiphons Onkel Chremes: Phanium wurde auf Lemnos von Chremes gezeugt, der eine Art Bigamie auf Lemnos führte. Dann ist Phanium mit ihrer Mutter nach Athen umgezogen; die Mutter, die ältere Frau, die Chremes auf Lemnos als Art Gattin hatte, ist gestorben. Die Waise Phanium ist in Athen von Antiphon in Abwesenheit des Vaters Antiphons Demipho und des Onkels Chremes verheiratet worden. Chremes seinerseits hat in Athen eine andere Frau, die Nausistrata heißt, und nicht zögert, über ihren Mann Chremes zu schimpfen, sobald sie unterrichtet wird, dass Chremes in Lemnos eine andere Frau hatte. Ja, am Ende wird ein echter Skandal nach den Sitten einer römischen Gesellschaft dargestellt: die Entscheidung wird nicht dem pater familias gegeben, sondern der Frau, 272 Gualtiero Calboli <?page no="273"?> 2 Aber Nausistrata (1040-41) beschimpft ihren Mann Chremes und sagt: NA. Adeo hoc indignum tibi uedetur,/ homo adulescens si habet unam amicam, tu uxores duas? So wird im Stück vorausgesetzt, und ohne Ζögern, dass Chremes zwei Ehefrauen, eine in Lemnos und die andere in Athen, hatte. namentlich Nausistrata, und Nausistrata überträgt die Entscheidung dem Sohn (1044-1046 NA. neque promitto quicquam neque respondeo / priu’ quam gnatum uidero: ei(u)s iudicio permitto omnia: / quod is iubebit faciam). Dies ist eine Umwälzung der römischen Gewohnheit des alltäglichen Lebens und konnte normalerweise nur in Griechenland geschehen. Jedoch, dass die Teile, wo Nausistrata eintritt (784-819; 990-1054), auf Terenz zurückgehen und nicht auf Apollodor, hat schon Eckard Lefèvre (1978) be‐ wiesen, und ich war zuerst überzeugt, dass er überwiegend richtig lag, obgleich nicht in allen Einzelheiten. Denn nicht alles in seiner Beweisführung besitzt freilich dieselbe Beweiskraft. Die Entdeckung des ursprünglichen Texts von Apollodor bleibt in einigen Punkten und Besonderheiten eine schwierige Arbeit, wie, z. B. wo Demipho seinen Bruder Chremes versucht zu verteidigen (vs. 106-1020) und von Lefèvre im allgemeinen für Ernst genommen und Apollodor zugeschrieben wird, während mir seine Verteidigung, wenn sie auch von Apollodor abhängt, eine vollständig unglaubliche wegen der Verteidigung verursachte Herabsetzung zu sein scheint: E.Lefèvre (1978: 42) hat darüber geschrieben: „Aufgrund dieser Voraussetzungen dürfte die Version, nach der Chremes mit Phaniums Mutter lediglich ein Verhältnis gehabt hat, auf Apol‐ lodor zurückgehen und die Doppelehe eine Erfindung des freier schaltenden römischen Dichters sein“. Dies sei lediglich auf Demiphos Wörtern begründet (1016-1020), der seinen Bruder um jeden Preis zu entschuldigen versucht. Nach dieser Erzählung dagegen und Lefèvres Interpretation hätte Chremes nur ein einziges Mal mit Phaniums Mutter auf Lemnos Verkehr gehabt 2 , und später sei er zufälling auf Lemnos gewesen. Aber so weit möchte auch Lefèvre nicht gehen, der schreibt (S. 43): „Freilich könnte Chremes auch schon bei Apollodor des Öfteren nach Lemnos gefahren sein, um heimlich der Tochter und ihrer Mutter zu helfen“. Unten werden wir sehen, dass dies nicht die einzige Ursache für Chremes war, nach Lemnos zu fahren, denn er sollte den Grundbesitz der Nausistrata, den von Nausistrata geerbten Grundbesitz, pflegen. Auf diese Besonderheit hat auch Karl Büchner (1974: 357) seine Aufmerksamkeit gerichtet und einige wichtige Beziehungen hervorgehoben: Aber weder Büchner noch Lefèvre haben die besondere Situation der Insel Lemnos ausführlich in Betracht gezogen. Dies hat in mir Zweifel aufkommen lassen, so dass ich entschied, die Situation war, von Lemnos eingehender zu untersuchen. Denn auf Lemnos war eine Athenische Kleruchie und es waren Nausistrata (Ter. Phorm. 784-1055) 273 <?page no="274"?> Athener Kleruchen tätig. Weizen, Wein und Öl wurden produziert und auf dieser Insel wurde der Weg überwacht, auf dem der Weizen von der Chersones nach Athen gebracht wurde (vgl. Laura Ficuciello, 2012). Aber dass Chremes gewohnt nach Lemnos zu fahren, nur um der Tochter und ihrer Mutter zu helfen, ist unsinnig. Dazu passt auch die Situation, die Ficuciello dargestellt hat, nämlich der Organisation und des Arbeitsablaufs auf Lemnos. Denn man hat eine Reihe von Pyrgoi (Türme) gefunden, die zur Verteidigung der Insel und zur Kontrolle der Sklavenarbeit benutzt wurden. Diese Türme wurden auch als Lager benutzt. Man hat auch Reste von Ölpressen und Weinkeltern gefunden, die eine starke, halbindustrielle Produktion beweisen, eine Produktion jedoch, die sich mangels einer modernen Düngung in Gefahr befand, abzunehmen. Eine starke Reduktion der Erträge war fast unvermeidlich. So wird einerseits eine zufällige, kurze Anwesenheit eines Großgrundbesitzers, wie Nausistratas Vater es war (der zwei Talente pro Jahr bekam) und Nausistratas Mannes, Chremes (der zwei Talente bekommen sollte) ganz unmöglich, andererseits konnten die Renten von Chremes (und Nausistrata) nicht so groß wie die von Nausistratas Vaters sein. Auf jeden Fall sollte die Anwesenheit und Tätigkeit von Chremes auf Lemnos die ganze Zeitdauer vom Mähen bis zur Weinlese und weiter umfassen, d. h. den ganzen Sommer, und noch mehr Zeit, um Weizen, Wein und Öl zum Export vorzubereiten, und die Erträge waren trotzdem reduziert. Deshalb scheint mir der Schluss unhaltbar zu sein, zu dem Lefévre (1978: 46) gekommen ist: „Eine weitere Annahme, Nausistratas Antritt V 3 stamme von Terenz, mag darin zu sehen sein, dass sein materialistisches Thema mit Nausistratas diesbezüglicher Argumentation in der terenzischen Version von V 9 zusammenstimmt, nach der Chremes permanent Beziehungen zu Phaniums Mutter unterhalten habe: ‘Also das war der Grund für seine häufigen Reisen nach Lemnos und die längere Aufenthalte? Das die schlechten Preise, die unsere Erträge minderten? ’ (1012-1013) […] Bestand bei Apollodor kein Anlaß, daß Nausistrata in der Schluß-Szene V 9 erschien, und geht auch ihr Dialog mit Demipho in V 3 auf Terenz zurück, ist die Vermutung naheliegend, daß sie bei Apollodor überhaupt nicht aufgetreten war“ (1978: 46) und später (S. 110): „Auf Chremes ‘Verlogenheit’ wird bei Apollodor kein Gewicht gelegen haben, da Nausistrata bei ihm wahrscheinlich nicht aufgetreten war und erst die gemeinsame Konfrontation der Ehegatten mit Chremes’ Bindung an Lemnos die Voraussetzung ist, diese zu einem moralischen Problem werden zu lassen“. Dass Lefèvre die Geschichte Athens und Lemnos‘ etwas vernachlässigt hat, hat auch Bochard (1980: 58 Anm.2) richtig bemerkt. Wenn Chremes bei Apollodor auf Lemnos zufällig ein einziges Mal mit Phaniums Mutter vinolentus Geschlechtsverkehr gehabt hatte, gab es keinen 274 Gualtiero Calboli <?page no="275"?> Grund, dass er zu diesem Zweck oft nach Lemnos zurückkam, und wenn Chremes oft nach Lemnos zurückkam, bedeutet es, dass er als Ehemann von Nausistrata eine Art Verwalter des von Nausistrata ererbten Großgrundbesitzes geworden war und deshalb Nausistratas Meinung nach gewohnt war, dorthin häufig und regelmäꞵig zu fahren. Das ist selbsverständlich. Was aber Nausistrata angeht, scheint es, wie Lefèvre anzunehmen, dass sie eine Erfindung des Terenz sei, m. E., unmöglich, denn sie bringt in der Unterhaltung mit ihrem Schwager Demipho Besonderheiten hervor, die Terenz nicht kennen konnte, oder nur ganz zufällig kannte, Apollodor aber naturgemäß kannte. In der Tat, haben wir es mit zwei Erzählungen zu tun, der Erzählung von Nausistrata (vs. 784-795, wir möchten diesen ersten Auftritt von Nausistrata „Nausistrata1“ benennen, und dies, um deutlich den ersten Auftritt der Nausistrata von dem zweiten zu unterscheiden, Nausistrata2). Hierbei erklärt Nausistrata1 dem Demipho, dass sie in einer großartigeren Weise hätte ihm helfen wollen, aber die Erträge aus Lemnos, im Vergleich zur vorherigen Situation, als ihr Vater lebte und auf Lemnos arbeitete, reduziert wurden. Deshalb war sie gezwungen, ihre finanzielle Hilfe einzuschränken, wofür sie sich entschuldigt. Die zweite Erzäh‐ lung (Nausistrata2, sozusagen, vs. 885-1055) wird von Demipho benutzt, um in der Unterhaltung mit Nausistrata2 nachzuweisen, dass Chremes in Lemnos keine wirkliche, zweite Ehe geschlossen hat. Darauf stützt sich Lefèvre, um zu beweisen, dass „Glücklicherweise ist noch zu erkennen, daß es sich bei Demiphos Charakteristik des Verhälnisses seines Bruders um die Konzeption des Originals handeln muß“ (Lefèvre, 1978: 43). Dann hat Professor Lefèvre auch die Stelle hinzugezogen, wo Demipho und Chremes sich über den Umzug von Chremes‘ Frau und Tochter vom Lemnos nach Athen erkundigen. Das war auch der Plan seines Bruders Demipho. Es lohnt sich die Mühe, beide Terenz-Stellen genauer anzusehen: Ter. Phorm.567-576 DE. Quid? qua profectu’ causa hinc es Lemnum, Chreme, Adduxti tecum filiam? CH. Non. DE. quid ita non ? CH. postquam uidet me ei(u)s mater esse hic diutius, simul autem non manebat aetas uirginis meam neglegentiam: ipsam cum omni familia ad me profectam esse aibant. DE. quid illi tam diu quaeso igitur commorabare, ubi id audiueras ? CH. pol me detinuit morbus. DE. unde ? aut qui? CH. rogas ? senectus ipsast morbu’. sed uenisse eas saluas audiui ex nauta qui illas uexerat. ----570 ----575 Lefèvre (1978: 43) hat daraus einen Schluss gezogen, der mich nicht überzeugt hat: „An dieser Stelle gibt Chremes selbst eine zumindest mehrjärige Abwe‐ Nausistrata (Ter. Phorm. 784-1055) 275 <?page no="276"?> 3 Aber Chremes musste in Lemnos als Verwalter des Groꞵgrondbezichtes der Nausis‐ trata arbeiten und daher scheint eine mehrjärige Abwesenheit des Chremes von Lemnos gar unmöglich zu sein. 4 Dass Stilpo richtig war und aus Apollodorus kam, hat schon Kuiper (1938: 51), hervorgehoben. 5 A.Michel (1987: 122) muß nicht falsch gelesen werden, wenn er schreibt: „Demiphon est avare: il doit donner beaucoup d’argent“. Michel hätte besser schreiben müssen: „il devrait donner beaucoup d’argent“, denn das Geld wird von seinem Bruder Chremes gezahlt, d.-h., handelt es sich immer um Geld aus Lemnos. senheit und vor allem eine Vernachlässigung der Tochter und ihrer Mutter zu. Daß dieses Motiv von Apollodor stammt, geht daraus hervor, daß die ganze Anlage des Stücks von ihm getragen wird: Nur darin, daß Chremes sich nicht hinreichend um Phaniums Mutter gekümmert hat, ist es motiviert, daß sie von sich aus, d. h. ohne daß es Chremes weiß, nach Athen übergesiedelt ist“ 3 . Auch hier hat Lefèvre nicht die ganze Situation in Bezug auf Chremes und die Bedingungen auf Lemnos berücksichtigt. Zunächst erzählt Chremes, dass er auf Lemnos erkrankt und daran gehindert war, sofort von Lemnos zurückzufahren. Aber dies betrifft nur die Zeit, die Chremes auf Lemnos verbracht hat, nachdem die zweite Frau oder (wie Lefèvre, S. 41, vorzieht) Mätresse nach Athen schon, gerade um Chremes (besser gesagt, um Stilpo-- unter diesem Namen wurde Chremes von Phaniums Mutter gekannt) 4 zu treffen, übergesiedelt war (ipsam cum omni familia / ad me profectam esse). Chremes nennt seine Krankenheit senectus ipsast morbu’, d. h. er entschuldigt seine Langsamkeit. Chremes war also wegen seines Alters langsam geworden, aber mehr als ein Jahr / ein paar Jahre durfte Chremes nicht von seinem, besser gesagt, Nausistratas, Großgrundbesitz abwesend sein, denn er musste die Arbeit und die Produktion kontrollieren; dann brauchte er das Geld aus Lemnos (er hatte vielleicht auch andere Renten, aber nur das Geld von Lemnos wurde in der Komödie erwähnt 5 , 664-681 CH. Opportune adeo argentum nunc mecum attuli / fructum quem Lemni uxori’ reddunt praedia : / ind’ sumam). Es scheint also, dass Chremes, um Sonderausgaben (und vielleicht auch Haushaltskosten) zu bestreiten, nur auf das Geld aus Lemnos rechnen konnte, auch Nausistrata hat nur Geld aus Lemnos zur Verfügung. Wenn Chremes ungefähr ein paar Jahre oder ein Jahr von Lemnos abwesend geblieben ist, passt dies gut dazu, daß er viel Geld (nämlich die Rente von zwei Jahren oder eines ganzen Jahres) besaß. Daß er länger abwesend war, ist wegen der Kontrolle der Sklavenarbeit auf Lemnos und der Notwendigkeit von Geld in Athen, m. E. aus‐ geschlossen. Aus Demiphos Worten erfahren wir, dass Chremes nach Lemnos gefahren war, um seine Tochter (Phanium) zu holen, weil er mit seinem Bruder Demipho geplant hatte, die Ehe zwischen Phanium und seinem Neffen Antipho 276 Gualtiero Calboli <?page no="277"?> 6 Schon Kuiper 1938: 74, hatte alles dies für eine Lüge gehalten, und schrieb: “It is impossible, I think, that Demipho should be telling a lie here”. Die Ansicht Kuipers wird auch von allem bestätigt, was vor der Lüge Demiphons geschieht (vs.957-989), wo Demipho mit der List beginnt (958-961 uides peccatum tuum esse elatum foras / neque iam id celare posse te uxorem tuam : / nunc quod ipsa ex aliis auditura sit, Chreme, / id nosmet indicare placabilius est) und am Ende sogar auf der Gewalt nicht verzichtet (988 DE. nisi sequitur [sc. Phormio], pugnos in uentrem ingere), so dass Phormio gezwungen wird, zweimals Nausistrata anzurufen. Dies geht mit der Folge zusammen, vorzubereiten (DE. Quid? qua profectu’ causa hinc es Lemnum, Chreme, / adduxti tecum filiam? ). Aber das war nicht der einzige Zweck, zu dem er nach Lemnos fuhr, weil er die Einkünfte des Großgrundbesitzes, den er als Nausistratas Ehemann besaß, aus Lemnos mitbringen mußte und wirklich mitgebracht hat (679-681 CH. Opportune adeo argentum nunc mecum attuli, / fructum quem Lemni uxori’ reddunt praedia : / ind’ sumam). So mußte Chremes zwei Ziele auf Lemnos erreichen, seine Tochter nach Athen zu holen und die Einkünfte von Nausistratas praedia einzutreiben. Wenn Chremes in Demiphos Worten sofort aus Lemnos abfahren konnte, als er erkannte, seine Lemnische Frau (oder Mätresse) sei nach Athen umgezogen, bedeutet dies, dass Chremes in Lemnos seine Arbeit schon abgeschlossen und alle Einkünfte gesammelt hatte. Oder es ist nur eine falsche Entschuldigung, wie es aus der Antwort auf Demiphos Frage erscheint (DE. unde aut qui? / CH. rogas? senectus ipsast morbu’). Es scheint, als ob Chremes seine Tätigkeit in Lemnos nach Terenz/ Apollodor(? ) verbergen wollte und er sich geschämt hätte, als Verwalter der Nausistrata zu handeln. Sehen wir jetzt die Stelle, wo Demipho versucht, die zweite Ehe seines Bruders Chremes herunterzuspielen und als eine zufällige Episode darzustellen: Ter. Phorm.1014-1025 DE. ego, Nausistrata, esse in hac re culpam meritum non nego; sed ea qui sit ignoscenda. PH. uerba fiunt mortuo. DE. nam neque neglegentia tua neque odio id fecit tuo. uinolentu’ fere abhinc annos quindecim mulierculam eam compressit unde haec natast; neque postilla umquam attigit. ea mortem obit, e medio abiit qui fuit in re hac scrupulus. quam ob rem te oro, ut alia facta tua sunt, aequo animo hoc feras. NA. quid ego aequo animo? cupio misera in hac re iam defungier; Sed quid sperem? aetate porro minu’ peccaturum putem? Iam tum erat senex, senectus si uerecundos facit. Hierbei erzwingt Demipho die Wahrheit und versucht, die Ehe seines Bruders als eine vorübergehende Kleinigkeit schönzureden, aber er überzeugte Nausis‐ trata nicht und uns auch nicht 6 . Jedoch Lefèvre (1978: 42) hat daraus den Nausistrata (Ter. Phorm. 784-1055) 277 <?page no="278"?> nur wenn es sich um eine Lüge handelt, und passt zu Apollodor, nicht zu Terenz, und es ist undenkbar, dass Terenz all dies hinzugefügt hat, das zu der folgenden Einschränkung von Demipho (vs.1014-1020) nicht passt: die lemnische Ehe war ein großes Hindernis für Chremes und liefert die Rechtfertigung, die Nausistrata am Ende benützt: vs.1040-1042 CH. Hem quid ais? NA. adeo hoc indignum tibi uidetur, filius/ homo adulescens si habet unam amicam, tu uxores duas? 7 Diese Voraussetzungen basieren auf einer großen Vermischung vom juristischen Geschichtspunkte aus und gründen sich nur auf die hypothetische Basis, namentlich dass Phanium eine νόθη war. Aber es kommt darauf an, dass eine attische Bürgerin, eine ἀστή, die Mutter von Phanium, in Lemnos vergewaltigt wurde, ohne dass eine Ehe stattfinden sollte. Dies ist gerade, was Demipho sagt, und Lefèvre annimmt, es sei von Apollodor gemacht. Dass andererseits Phanniums Mutter die attische Staatsangehörigkeit besass, war eine ‘condicio sine qua non’ für Phanium, um eine attische Bürgerin zu sein. Wie konnte also Chremes eine regelmässige Ehe vermeiden? Tatsächlich hat er sie nicht vermieden, nur einen falschen Namen gegeben, Stilpo. 8 Aber Demipho spricht nicht von einem Verhältnis, sondern von einer Vergewaltigung: uinolentus’ […] mulierculam eam compressit unde haec natast. Wie kann man Demiphos Erzählung verwenden, um Apollodors Stück wiederherzustellen? Schluẞ gezogen, auf diese Weise hätte Apollodor sein Stück gebaut, obgleich er mich, wie ich sagte, gar nicht überzeugt hat. Auf jeden Fall hat er, wie folgt, geschrieben: „Aufgrund dieser Voraussetzungen 7 dürfte die Version, nach der Chremes mit Phaniums Mutter lediglich ein Verhältnis 8 gehabt hat, auf Apol‐ lodor zurückgehen und die Doppelehe eine Erfindung des freier schaltenden römischen Dichters sein. Chremes’ Verhalten war bei Apollodor menschlich verständlich motiviert, bei Terenz ist es in ein bedenkliches Licht gerückt“. Natürlich lösst Lefèvre alle Schwierigkeit mit dem Ansatz auf, dass Nausistrata eine Erfindung von Terenz war. Aber dies passt mit der Situation der Insel Lemnos nicht gut zusammen, nämlich einer Situation, die wir aus den auf der Insel Lemnos durchgeführten archäologischen Ausgrabungen kennengelernt haben, die dem Apollodor gut bekannt sein sollten, aber kaum dem Terenz. In der Tat handelt es sich um die Abnahme der Einkünfte aus Lemnos, die im III. Jahrhundert stattfand, und die Notwendigkeit, dass Chremes lange Zeit auf Lemnos verbringen musste, um die Arbeit der Sklaven bei Produktion und Export des Großgutbesitzes seiner Frau Nausistrata zu organisieren. Aber jetzt müssen wir alles ausfühlicher hervorheben und nachprüfen. Außer den von Nausistrata dargestellten Besonderheiten, ist eine Tatsache, dass Chremes’ Tochter Phanium, die auf Lemnos geboren war, die attische Staatsangehörigkeit besaß, und dies nach Perikles’ Gesetz des Jahres 451, ohne dass beide Eltern attisch waren, unmöglich war. Das bedeutet, daß Phaniums Mutter die attische Staatsangehörigkeit besaß, und normalerweise, dass eine regelmässige Ehe zwischen Chremes und der vergewaltigten Mutter von Phanium stattfand. Welche Ursache gab es ohne diese Annahme, dass 278 Gualtiero Calboli <?page no="279"?> 9 Auch S.Ferrucci (2006: 192-194) hat erwähnt, dass die athenische Ehe nicht so einfach und unzweideutig war und deren Entscheidung normalerweise den Richtern überlassen wurde, weil präzise Regeln in der Ehe zu Verfügung standen. Klar war nur, dass beide Eheleute athenisch sein mussten. Chremes einen falschen Namen angab? Es spielt sicher keine Rolle, ob es sich um eine vollgültige Ehe handelte oder nicht 9 . „Besonders wichtig (schreibt Lefèvre, 1978: 41) ist die Beurteilung des Falles der Phile (Isae. 3,45 bis 52), die beweise, „daß die T o c h t e r eines Bürgers und seiner athenischen Mätresse in eine vollgülltige Ehe gegeben werden konnte, daß sie also das Verbot der Mischehe mit Fremden nicht berührte“. Daraus folgt für den Ἐπιδικαζόμενος, daß Phaniums Mutter nicht mit Chremes in v o l l g ü l t i g e r Ehe gelebt zu haben brauchte, damit die Tochter ἀστή war. Sie mußte freilich selbst Bürgerin gewesen seinˮ. Es geht also nicht um die Ehe, sondern um die Staatsangehörigkeit beider Eltern, die notwendig war, um den Kindern die Staatsangehörigkeit zu übertragen. Wenn Phanium athenisch war, bedeutet dies, dass beide Eltern athenisch waren, wenn auch in keiner regelmäßigen Ehe wohnten, abgesehen dass man in der Komödie immer von Ehe gesprochen hat. Warum hat dann Chremes dem vergewaltigten Mädchen den falschen Namen Stilpo gegeben, eines Atheners, der nicht existierte? Wie konnte Pha‐ nium ein athenisches Mädchen sein, wenn sie Tochter einer nicht athenischen Frau war? Dann besaßen die κληροῦχοι der Insel Lemnos, deren Gruppe aller Wahrscheinlichkeit nach auch Phaniums Mutter angehörte, die athenische Staatsangehörigkeit. Kein Wunder, dass sie mit der ganzen familia nach Athen kam, um Stilpo zu finden. Dass sie mit Stilpo, nachdem sie vergewaltigt wurde, in keinem Kontakt mehr war (numquam attigit), um Demiphos Wörter zu benutzen, ist unglaublich (credat Iudaeus Apella! ). Denn Chreme brauchte dem vergewaltigten Mädchen seinen namen nicht zu geben, wenn er keine Bezie‐ hung mehr haben wollte. Es scheint also, dass Chremes einen anderen Zweck zu verfolgen hatte, nämlich den angefangenen Geschlechtsverker nicht zu unterbrechen, sei es durch die Ehe, sei es durch eine andere Beziehungsweise. Außerdem war Phaniums Mutter keine so verächtliche muliercula, dass man sie straflos vergewaltigen dürfte. Denn Chremes sollte den falschen Name Stilponem nehmen und ihn der neuen zweiten Frau mitteilen. Warum diesen Namen mitteilen in einem Fall von Vergewaltigung, wenn in der Hecyra (829 dicitque sese illi anulum, dum luctat, detraxisse) nur ein entgerissener Ring gelassen wurde, um den Vergewaltiger zu erkennen? Die Lüge des Demipho ist evident. Dies scheint kaum einer Terenz folgenden Bearbeitung zuzuschreiben zu sein, oder, besser gesagt, könnte aller Wahrscheinlichkeit nach eher ein Erzeugnis des Originals von Apollodorus sein als der terenzischen Wiederher‐ Nausistrata (Ter. Phorm. 784-1055) 279 <?page no="280"?> 10 Nach athenischem Eherecht behielt die Ehefrau ihren Grundbesitz, um die Bezie‐ hungen zur vorigen Familie nicht zu trennen (vgl. M.Apolito, 2009: 108). Das passt gut zu der Erwähnung ihres Vaters von Nausistrata. 11 Dass der Status der Pächter des staatlichen Landes in Attika (Lemnos) besonders gut war, hat R. Osborne (1988, 291 f.) nachgewiesen. Ein ähnliches Phänomen wurde auch für die in den Tafeln von Heraclea überlieferten Maßnamen vermutet (IG XIV, 645); vgl. F. Coarelli (1998). Nach dem epigraphischen Befund scheinen dieselben Familiengruppen lange Zeit auf der Insel Lemnos gewohnt zu haben, die auch in klassischer Zeit anerkannt wurden, vgl. dazu E. Culasso Gastaldi (2007). stellung zuzuschreiben sein. Aber Chremes hatte auch andere Gründen, lange Zeit in Lemnos zu verbringen, und zwar als Verwalter des Großgrundbesitzes, den Nausistrata in jeder Insel besaß 10 . Das IV Jahrhundert v. Chr., als die öffentliche Entstehung der lemnischen Kleruchie stattfand, wurde durch gesellschaftliche Zerrüttungen gestört; jedoch die Insel Lemnos wurde gerade in jener Zeit vollständig von den Athenern mit mittleren bis kleinen Gutshöfen belegt, und zwar in den östlichen Teilen der Insel, insbesondere im Gebiet Kaminia. Nach den archäologischen Ausgra‐ bungen und epigraphischen Zeugnissen scheint es, dass der Teil der Insel, der dem Gebiet Hephestia entsprach, geteilt und in viele κλῆροι gegliedert wurde, die den κληροῦχοι gewährt wurden, wie es in den athenischen Inscrhriften des IV Jahrhunderts v.-Chr. bezeugt ist. Die Besitzer dieser Gutshöfe jedoch konnten, wie es scheint, die Ausgaben der Eigenbewirtschaftung nicht bestreiten, und hielten besonders der Konkur‐ renz der Großgrundbesitzer nicht stand, die sich zwar in den westlichen Teilen der Insel Lesbos, im Gebiet Myrrina, entwickelt hatten und mit einer billigeren Sklavenarbeit, wie in Attika, rechnen konnten. Manchmal zogen die kleineren Ackerbesitzer vor, ihe Gutshöfe bald zu verkaufen. So sind die mittleren und kleinen Gutshöfe allmählich, nämlich innerhalb eines Jahrhun‐ derts verschwunden, denn sie wurden aller Wahrscheinlichkeit nach von den Großgrundbesizern aufgekauft, die vielleicht auch schon ab dem V. Jahrhundert v. Chr. auf den verpachteten Boden des Staats rechnen konnten (Ficuciello 2012: 153), wichtige Teile des Landes besaßen und seit langem gute Geschäfte auf der Insel Lemnos machten 11 . Diese Besitzer hatten die Verschuldung der mittleren und kleinen Besitzer ausgenutzt, denen besonders die von Athenern im IV. Jahrhundert v. Chr. ge‐ sendeten κληροῦχοι angehörten. Die neuen κληροῦχοι wurden, wie es scheint, gezwungen ihre Besitzungen billiger zu verkaufen und auf diese Weise wurden die Güter der Großgrundbesitzer vergrößert. So wurden die Groꞵgutbesitzer mächtiger und hatten sogar ein politisches Übergewicht gewonnen. Sie waren auf jeden Fall in III. Jahrhundert und in hellenistischer Zeit eine kleine Gruppe 280 Gualtiero Calboli <?page no="281"?> 12 Vgl. Suda α 3404 Adler. Apollodor wurde auch der Athener genannt, weil er mit der athenischen Staatsangehörigkeit ausgezeichnet wurde. von Landbesizern. Diese Situation entspricht der Tätigkeit des Nausistratas Vater. Aber diese Landbesitzer bedurften einer besonderen Fähigkeit und Ag‐ gressivität, mit denen Chremes nicht gut angerüstet zu sein scheint (vs.790-793 NA. Vir uiro quid praestat! […] uirum me natam uellem: ego ostenderem, DE. Certo scio. NA. Quo pacto …). Laura Ficuciello (2012: 154) hat bewiesen, dass sich im III. Jahrhundert die wirtschaftliche Situation auf Lemnos in diese Richtung entwickelt hatte, schwieriger wurde und in der Folge eine Herabsetzung der Einkünfte verursachte, und besonders - können wir hinzufügen - für jemanden, der den anderen Großgutbesitzern ebenbürtig, wie Chremes (vielleicht auch wegen der häufigen Abwesenheiten) nicht war. Wenn wir also alles zusammenfassen, was aus der neuen archäologischen Untersuchung und aus einer ausführlicheren Prüfung aller diesbezüglichen Besonderheiten hervorgeht, ist das neue Bild von Chremes’ Lage wie folgt: 1. Auf Lemnos hatten die ersten κληροῦχοι während des IV. Jahrhunderts eine blühende Wirtschaft entwickelt, die trotzdem die Entstehung von Groẞgrundbesitzen nicht verhindern konnte; diese Situation entspricht der Zeit, als Nausistratas Vater in Lemnos arbeitete und viel Geld verdient hatte. Die ihm gehörenden Groẞgrundbesitze wurden gerade in jener Zeit entwickelt, und die Einkünfte wuchsen. 2. Im dritten Jahrhundert (III saec.) schrumpften die Einkünfte stark, denn der Ackerbau in Lemnos machte eine Krise durch, wie durch die Archäologie gezeigt wurde, abgesehen von den Ursachen, die solch eine Krise gebracht hatten. Lemnos blieb fast immer im Machtbereich der Athener, und den Athenern und einem Einwohner von Athen, wie Apollodor 12 , musste Lemnos‘ Wirtschaft vollständig bekannt sein, während verwunderlich wäre, dass solche Kentnisse auch einem Römer vorlagen. 3. Chremes hatte auf Lemnos eine zweite Ehe (oder eine standesgemässe Ehe) unter dem Namen Stilpo mit einer Frau geschlossen, die attische Staatangehörigkeit besaß, sonst wäre seine Tochter nicht ein attisches Mädchen, wie sie es tatsächlich war. 4. In der Komödie Phormio wird kein anderes Geld erwähnt als das Geld aus Lemnos, das in jedem Fall für die großen Sonderausgaben und auch für die gewönlichen Haushaltskosten benützt wurde. Um dieses Geld zu bekommen, musste Chremes jedes Jahr (ode jede zwei Jahre) nach Lemnos fahren und so lange auf Lemnos bleiben, bis er das Geld eingefordert hatte. Nausistrata (Ter. Phorm. 784-1055) 281 <?page no="282"?> 13 Am Ende des Eunuchus erwähnt Terenz die Attica elegantia, die ihm deshalb, allerdings mit einem ironischen Unterton, nicht unbekannt war (vgl. G.Calboli, 1991: 629). 5. Alles dies stand schon in Apollodors Original, Terenz hat nur eine Idee am Ende hinzugefügt und Plautus’ uxorem saeuam zum Teil ganz vorsichtig und in einer genialen Weise nachgeahmt, die wir später hervorheben werden. Jetzt müssen zwei Besonderheiten behandelt werden, nämlich (1) was konnte genau für Terenz Plautus’ uxor saeua meinen, (2) wie könnte Terenz als Plautus’ Nachahmer und Fortsetzer gelten, ohne zur Anklage des furtum zu kommen, um dagegen die Anklage furtum zu vermeiden und ein in jeder Hinsicht ‚At‐ tisches‘ Stück 13 zu verfassen? So müssen wir einige Belege von einer uxor saeua in Betracht ziehen, die bei Plautus vorkommen, zumal dass auch Nausistrata als uxor saeua (v.744) betrachtet wird. Um einen weiteren Eindruck zu bekommen, was bestimmte Komiker mit ‚Ehefrau‘ und ‚Hochzeit‘ verbinden, muss man die Ehefrau in der plautinischen Komödie „Menaechmi“ in den Blick nehmen. Der Plot läuft auf einen Austausch zwischen Menaechmus1 und Menarchmus2 (ursprünglich Sosicles) hinaus, einen Austausch, der den Höhepunkt erreicht, als nicht einmal mehr der Sklave von Menaechmus2, Messenius, seinen Herrn erkennt und mit Fausthieben einschreitet, um den Zwilling seines Herren Menaechmus1 zu retten, den vier Männer zum Medicus bringen, da er verrückt sei. Das Vorbild des Stücks ist unsicher, vielleicht ist es Posidippus. Die uxor saeua findet sich: Plaut. Asin. 851-947, wo Artemona, eine ‘dotata uxor’ ihren Mann Demaenetum accubantem cum corona amplexum amicam ertappt, und zwar nachdem er, angestiftet von seinem Sohn, Argyrippus (der als Gast spielt), seine Frau harsch angefahren hatte hat: 899-901 ARG. quid ais, pater? ecquid matrem amas? DE. egone illam? nunc amo, quia non adest. / ARG. quid quom adest? DE. periisse cupio. Plaut. Ep.173-180 P E . Reuereor filium. A P . At pol ego te credidi uxorem, quam tu extulist, pudore exsequi, quoiius quotiens sepulcrum uides, sacruficas ilico Orco hostiis, neque adeo iniuria, quia licitumst eam tibi uiuendo uincere. P E . Oh! Hercules ego fui, dum illa mecum fuit; neque sexta aerumna acerbior Herculi quam illa mihi obiectast. A P . Pulcra edepol dos pecuniast. P E . Quae quidem pol non maritast. 282 Gualtiero Calboli <?page no="283"?> 14 Zur Bedeutung und Verwendung des Worts iudicium vgl. G.Focardi (1972: 81). Merc.586 metuo ego uxorem cras si rure redierit / ne illa hic offendat (der alte Lysimachus, aber schon vorher der alte Kaufmann Demipho, verliebt in die Freundin seines Sohns Carinus (vs.544-561), versucht, eine Orgie im Haus seines Freunds in verborgenen von seiner Frau zu machen, uxor: uxor me expectat iam dudum essuriens domi; / iam iurgio enicabit, si intro rediero). Dann Lysimachus’ Frau fällt auf, die zehn Talente als Mitgift mitgebracht hat, und zu Hause eine Dirne (meretrix) antrifft (vs.700ff). Aber wir haben noch mehr zu berücksichtigen: denn es scheint, dass dem Terenz, als er den Phormio beendete, der Schluß des Mercator von Plautus in Betracht und das besondere Gesetz vorschwebte, das der adulescens Eutycus vorzuschlagen plant, nämlich dass den sechzigjährigen Alten verboten werde, mit den jüngen Söhnen an einem Wettkampf in scortari teilzunehmen, und ihnen nicht erlaubt werde, dies den Jungen zu verbieten: Plaut. Merc.1015-1026 DE. eamus intro. EV. immo dicamus senibus legem censeo priu’ quam abeamus, qua lege teneant contentique sint. annos natus sexaginta qui erit, si quem scibimus maritum siue hercle adeo caelibem scortarier, cum eo nos hic lege agemus: inscitum arbitrabimur et per nos quidem hercle egebit qui suom prodegerit. neu quisquam posthac prohibeto adulescentem filium quin amet et scortum ducat, quod bono fiat modo; siquis prohibuerit, plus perdet clam <qua>si praehibuerit palam. haec adeo | uti ex hac nocte primum lex teneat senes. bene ualete; atque, adulescentes, haec si uobis lex placet, ob senum hercle industriam uos aequom est clare plaudere. Dies bedeutet, dass Plautus die ganze Situation umgekert hat: der Junge sollte als Hurer wegen seiner Unmässigkeit verurteilt und bestraft werden, und statt des Jungen ist sein Vater, meint Eutycus, der verurteilt und bestraft werden sollte, weil er es gegen das Naturgesetz wagte eine Dirne zu verführen. Aber das ist kein Staatsgesetz, sondern nur ein Privatgesetz (dicamus senibus legem). Fast dasselbe oder eine sehr ähnliche Regel hat Nausistrata festgelegt, nämlich dass ihr Sohn Phaedria entscheiden muß, ob Chremes bestraft oder freigesprochen werden muss (ei(u)s iudicio  14 permitto omnia : / quod is iubebit faciam), obgleich Phaedria als Freundin eine Dirne gekriegt hat, die mit dem Geld der lemnischen Landwirtschaft von Nausistrata gekauft und freigelassen worden ist. Es scheint, Nausistrata (Ter. Phorm. 784-1055) 283 <?page no="284"?> ein Hurer hätte die Entschuldigung dank einem umgekehrten von Terenz als Plautus’ Nachahmer bestätigten Gesetz erreicht, wofür auch Nausistrata, als saeua uxor, nicht wenig mitgebracht hatte. Genau eine schiefe Nachahmung, die Terenz gerne, aber immer mit einem wenigem Angst versuchte (und es handelte sich immer um ein Privatgesetz). Plaut.Mil.673-74 nam in mala uxore atque inimico si quid sumas, sumpstus est, / in bono hospite atque amico quaestus est quod sumitur: 679-681 nam mihi, deum uirtute dicam, propter diuitias meas / licuit uxorem dotatam genere summo ducere; / sed nolo mi oblatatricem in aedes intro mittere. Dann verteidigt Periplectomenus die persönliche Freiheit gegen Ehre und schließt die ganze Stelle mit einer Art Gnome im Meinungsaustausch mit dem Sklave Palaestrio aus, wie folgt (l.684) P ALA E S T R I O : qur non uis? Nam procreare liberos lepidumst opus. / P E R I P L . Hercle uero liberum esse tete, id multo lepidius. So wurde die Ehefrau wegen der Mitgift geschätzt und auch Nausistrata ist eine reiche Frau, die nicht nur eine respektwürdige Mitgift besitzt, sondern auch Acker in Lemnos als Erbschaft erhalten hat, wovon ihr Vater zwei Talenta pro Jahr zu bekommen gewohnt war, während Chremes, Nausistratas Mann, weit weniger bekommt und bringt zu Hause. Ursache dieser Senkung liegt eben darin, dass Chremes im Lemnus eine zweite Familie nähren muss. Aber das geschah voran, jetzt, nämlich in der Zeit der Action ist die zweite Familie von Chremes nach Athen umgezogen, die Frau ist gestorben und die Tochter hat Antiphonem, den Neffen von Chremes, verheiratet. Aber die wirkliche Ursache, warum die Einkünfte im Lemnos reduziert wurden, war, weil der Ackerbau in Lemnos im III Jahhundert in eine Krise geraten war, und dies hatte mit der Nahrung einer zweiten Familie nichts oder nur ganz wenig zu tun. Ursache dieser Krise scheint die Unfruchtbarkeit des Landes wegen Mangel an Düngemitteln gewesen zu sein oder wegen einer unvernünftigen Ausmergelung, was, wie gewöhnlich und auch in Nausistratas Zeit geschah, der Unfähigkeit des Anbaus zugeschrieben wurde. Es ist auch nicht vollschständig ausgeschlossen, dass solche eine Unfähigkeit eine Rolle in dieser Krisis ausgespielt hat, aber sie war weder die einzige noch die wichtigste Ursache. Diese könnte wahrscheilicher eine Folge der Entwicklung des Großgrundbesitzes und der Sklavenarbeit sein. Eine Situation, die in Terenzens Zeit den Römern noch nicht gut bekannt sein konnte. Es scheint also zu einer ausführlicheren Kenntnis der Insel Lemnos nicht zu passen, Apollodor vollständig auszuschliessen (Szene V III) und wir können m. E. dem Terenz lediglich eine Umarbeitung des Endes (V 9), und diese auch in einer sehr beschränkten Weise, zuschreiben. Aber was genau konnte dem Terenz nicht und nur dem Apollodor bekannt sein? Es sind zwei Besonderheiten, 284 Gualtiero Calboli <?page no="285"?> 15 Diese Beziehung Eunuchus - Phormio leitet man aus den Didaskalien ab, und die Glaubwürdigkeit der Didaskalien wurde von Gestri und Mattingly in Zweifel gebracht, aber von vielen Anderen Gelehrten besonders von M.R.Posani (1942) und F.Arnaldi (1947: 103-110) verteidigt (vgl. H.Marti, 2968: 20-23; D.Nardo, 1969: 123-125; G.Calboli, 1974: 145-147 [ich habe auch die Didaskalien verteidigt, denn kein Argument der Kritiker stützt sich auf festen antiken Zeugnissen]. Jedoch der Lesart des Bembinus nach würde der Eunuchus während der Ludi Romani und Phormio während der Ludi Megalenses angeführt. Mattingly (1959) hat dem Bembinus gefolgt, und auch Donatus bezeugt (S.-346), dass der Phormio während der Ludi Megalenses angeführt wurde. die wir nur durch archäologische Belege erkannt haben, nämlich zunächt (1) dass am IV Jahrhundert, in der Zeit, als Nausistratas Vater in Lemnos tätig war, sich in Lemnos die Wirtschaft zu Großgrundbesitzungen entwickelt hatte, so dass einem von diesen Groβgrundbesitzen Nausistratas Vater gehören konnte und er zu einem Einkommen von zwei Talenten pro Jahr gekommen war, und zweitens (2), dass im III. Jahrhundert der Ertrag der Großgrundbesitze sank, und dies weder von Chremes Unfähigkeit, noch von der Notwendigkeit eine zweite Famile auszuhalten abhing, sondern von einer Beschränkung der ganzen Ackerproduktion und Landwirtsaft, die auf Lemnos zu Chremes‘ und Nausistratas Zeit stattfand. Was den ersten Punkt angeht, den könnte Terenz erfunden haben, den zweiten nicht oder wenigstens unwahscheinlich, und er muβ dem Apollodor zugeschrieben werden. Aber Terenz hat m. E. das Ende in V 9 ein wenig verändert. Wie hat aber Terenz diese Veränderung erreicht? Ich vermute, daß er den Charakter der Nausistrata aus Apollodor genommen und umgearbeitet hat und als Muster für eine solche Umarbeitung Plautus genommen hat. Denn Plautus hat die uxor saeua in seine Komödien oft eingeführt. Aber es scheint, dass es sich um ein eingeschränktes Modell handelte, dem Terenz sich nicht zu sehr annähern durfte, denn er versuchte die Anklage eines furtum zu vermeiden, wie er sie für den Eunuchus gekommen hatte. Tatsache ist, dass Terenz in den Adelphoe das furtum aus Plautus auf jedem Preis zu vermeiden versuchte (vgl. Calboli, 2012). Dann wurde der Phormio während der Ludi Romani (September) in demselben Jahr geführt, wie der Eunuchus während der Ludi Megalenses (April) 15 . Also fand nur vier Monaten vorher der Angriff statt, in dem Terenz von Lucius Lanuvinus als fur angeklagt wurde, der aus Naevius und Plautus die personae des miles und des parasitus gestohlen habe. Terenz konnte keine andere Verteidigung finden als zu antworten, dass er beide Personen aus dem Colax des Menanders direkt genommen hatte. Die Anklage und die Verteidigung fanden daher kurz vor Phormios Vorführung statt und Terenz sollte auf eine solche Aufklage besonders aufmerksam sein. Es scheint mir die Mühe zu lohnen, diese Besonderheit eingehender zu untersuchen. Nausistrata (Ter. Phorm. 784-1055) 285 <?page no="286"?> Wie Nausistrata das Wort res (eine Art Schwamm- oder Passe-partout Wort) verwendet, d. h. als Geld, hat R. Müller, 1997: 288, behandelt: „Von Plautus abweichend, der für ‟Besitz, Wohlstand, Vermögenˮ und ‟Geld(vermögen) res singularisch und pluralisch gebraucht, beschränkt sich Terenz in den genannten Funktionen ganz auf den Singular, als ob sich eine morphologische Eingrenzung durchgesetzt hätte“. Die Stelle, wo Nausistrata als eine Großgrundbesitzerin erscheint, sind die Verse Phor. 784-793 (siehe unten). Seinerseits hat Karl Büchner (1974: 434-438) versucht, das Theater von Apollodoros von Karystos zu definieren und die Wahl von Apollodoros durch Terenz zu rechtfertigen, mit anderen Worten das Problem zu lösen, warum Terenz zwei verschiedene Vorbilder, nämlich Menander und Apollodor, gewählt hat, und dann die Originale so umgearbeitet hat, dass die Verschiedenheiten zwischen Menander und Apollodor nicht abgemildert, sondern bewahrt und sogar hervorgehoben wurden. In der Tat bin ich aus dem Kommentar der letzten Szenen des Phormio ausgegangen, in denen zwei Figuren, Antiphon und Nausistrata auftreten. Was Antiphon angeht, kann man nichts erfahren, denn es handelt sich um eine Phantasie-Person ohne Beziehung zu einer wirklichen Persönlichkeit. Was dagegen Nausistrata von seinem Mann Chremes und den großen Besitzungen erzählt, die sie von ihrem Vater in Lemnos geerbt hatte und die von ihrem Man Chremes verwaltet wurden, diese können und müssen ausführlicher untersucht werden, denn es handelt sich um präzise Tatsachen, die der wirklichen Situazion der Insel Lemnos entsprachen. Was ich zu meinem großen Stauenen entdeckte, war, dass die Situation der Insel Lemnos in der entsprechenden Zeit der Erzählung von Nasistrata vollständig passte, und es sich um eine Situation handelte, die nur dem Apollodor, und kaum dem Terenz bekannt sein konnte. Der Schluß war natürlich, dass man vermuten kann, dass nur eine Umarbeitung, nichts mehr, von Apollodors Ἐπιδικαζόμενος von Terenzens Seite stattfinden konnte, obgleich ich meine, Nausistrata sei zum Teil von Terenz umgearbeitet worden, wie später deutlich werden wird. Dass sie vollständig dem Terenz (Lefèvre) und wenigstens zum großen Teil (Büchner) zugeschrieben werden sollte, scheint mir wirklich unmöglich. So versuchen wir jetzt wieder, die ganze Sache eingehend zu untersuchen. Zunächst muß man die Stelle in Betracht ziehen, wo Nausistrata sich mit Demipho, dem Bruder ihres Mannes, unterhält und deutlich erzählt, woher sie und ihr Mann Chremes das eingesetzte Geld bekommen haben, nämlich aus den Ackerbesitzungen auf Lemnos, die freilich nicht mehr so reich sind, wie zu der Zeit, als Nausistratas Vater auf Lemnos tätig war, und zwar, wie es scheint, als Großgrundsbesitzer: 286 Gualtiero Calboli <?page no="287"?> Ter. Phorm.784-793 DE. agedum, ut soles, Nausistrata, fac illa ut placetur nobis, ut sua uoluntate id quod est faciumdum faciat. NA. faciam. DE. pariter nunc opera me adiuues ac re dudum opitulata es. NA. factum uolo. ac pol minu’ queo uiri culpa quam me dignumst. DE. quid autem? NA. quia pol mei patris bene parta indiligenter tutatur; nam ex is praediis talenta argenti bina statim capiebat; vir viro quid praestat! DE. bina quaeso? NA. ac rebu’ uilioribus multo tamen duo talenta. DE. hui. NA. quid haec videntur? DE. scilicet. NA. virum me natam vellem: ego ostenderem, DE. certo scio. 785 ----790 Am Ende des ganzen Stücks handelt Nausistrata wie ein vir. Bemerkenswert ist der Unterschied zwischen talenta argenti bina und tamen duo talenta, wodurch wir erfahren, dass talenta argenti bina ein Jahresertrag war und dass die Acker‐ besitzungen größer geworden waren (rebus vilioribus multo). Wie es scheint, hatte Nausistratas Vater vor seinem Tod den Ertrag sogar vergrößert. Auf jeden Fall war Nausistrata eine reiche Frau Gemahlin. Die Wiederherstellung des Ἐπιδικαζόμενος von Lefèvre wurde von A.Blanchard (1980) in Zweifel gezogen mit Argumenten, die man nicht leicht vernachlässigen kann. Aber auch die Wiederherstellung von Blanchard selbst kann sich nicht dem Gegegenargument entziehen, dass es auf persönlicher Einbildungskraft basiert, beide Wiederher‐ stellungen, die von Lefèvre and die von Blanchard, sind bemerkenswert, wenn auch die von Lefèvre sich m. E. zu weit von Apollodors Text entfernt. Blanchart hat dagegen versucht, Terenz mehr von Plautus zu distanzieren. Ich versuchte mich dagegen zu einer zu ausführlichen Behandlung der ganzen Komödien nicht hinreißen zu lassen und mich nur auf die Nausistrata zu beschränken. Lefèvre hat Nausistrata, nämlich Terenzens Nausistrata, vollständig aus dem Ἐπιδικάζομνος übernommen, Blanchard ist in dieser Hinsicht zurückhaltender: „De même enfin le portrait de Nausistrata me paraît avoir été quelque peu durci par Tèrence, mais je ne suis pas certaine que le poète latin a créé le rôle“ (Blanchard, 1980: 64). Dass Terenz sich auf jeden Fall dem Plautus annährt (wenn man sich Lefèvres Ansicht anschließt) gibt Blachard zweifelsohne zu: „s’imposerait même sans conteste“ (Blanchard, 1980: 52). Ein anderer Fall von uxor saeua findet sich, oder besser, wird erwähnt in Plaut. Rud.893-905 (es spricht der alte Daemones) DA. … iam clientas repperi [nämlich die zwei jungen Mädchen Palaestra und Ampelisca] / atque ambas forma scitula atque aetatula. / sed uxor scelesta me omnibus seruat modis, / ne quid significem quippiam mulierculis. […] (vs. 1045-1047) sed ad prandium uxor me vocat. redeo domum. / iam meas opplebit auris vaniloquentia. Nausistrata (Ter. Phorm. 784-1055) 287 <?page no="288"?> Auf jeden Fall hat Nausistrata das erreicht, was der adulescens Eutycus in Mercator geplant hatte, nämlich ein Gericht, bei dem die Entschscheidung nicht dem pater familias, d.-h., ihrem Mann, Chremes, überlassen wird, sondern dem Sohn, dem schuldigem Phaedria, der das Geld von Lemnos (vs.679-681) dazu benutzt hat, um seine Geliebte freizukaufen (v.830). Alles dies wurde in der Tat nicht direkt von Phaedria, sondern vom Diener Geta und dem ‘parasitus’ Phormio durchgeführt: und Nausistrata entscheidet, ihr Sohn Phaedria sei eine Art Berufungsrichter (iudex). Der Parasitus Phormio erzählt, was er gemacht hat, um dies Geld zu finden, Chremes, der Vater, protestiert, aber Nausistrata vertraut jede Entscheidung ihrem Sohn an, er wird entscheiden, und der Onkel Demipho verpflichtet, dass er seinem Bruder Chremes gehorsam sein wird: Ter. Phorm.1038-1046 PH. Ego minas triginta per fallaciam ab illoc abstuli: dedi tuo gnato: is pro sua amica lenoni dedit. CH. Hem quid ais? NA. adeo hoc indignum tibi videtur, filius homo adulescens si habet unam amicam, tu uxores duas? Nil pudere! quo ore illum obiurgabi’? responde mihi. DE. faciet ut voles. NA. immo ut meam iam scias sententiam, neque ego ignosco neque promitto quicquam neque respondeo prius quam gnatum videro: ei(u)s iudicio permitto omnia: quod is iubebit faciam. PH. Mulier sapiens es, Nausistrata. NA. sati’ tibi<n> est? PH. Immo vero pulchre discedo et probe et praeter spem. Auf diese Weise hat Terenz eine plautinische Idee verwirklicht, ohne gleichzeitig Plautus wortlich und unvorsichtig nachzuahmen, was er besonders vermeiden musste und wollte. Mein Schluß ist also, dass es für die Abnahme der Einkünfte aus Lemnos wahrscheinlicher ist auf einem athenischen Dichter wie Apollodor als auf einem römischen Dichter wie Terenz zu rechnen, und es besser ist, für einen Römer, wie Terenz, einen Einfluß der plautinischen uxor saeua als eine einfache, glatte Beibehaltung des Originals zu vermuten. Außerdem, auch Terenz nennt Nausistrata eine uxor saeua (744 CH. Conclusam hic habeo uxorem saeuam). Jedoch haben wir noch eine andere Frage zu beantworten: wie konnte ein römischer komischer Dichter es wagen so gründlich gegen die römischen Sitten zu gehen und alle Macht der Frau Gemahlin dem Sohn des paterfamilias zu übertragen? Ja er brauchte zwei Freiheiten, die des ‚genus comicum‘ und die der Griechen. Das bedeutet Apollodorus. 288 Gualtiero Calboli <?page no="289"?> 16 Dazu eine große Literatur, zu deren ich jetzt mich zu wenden verzichte, nicht desto weniger vgl. H.Marti (1964: 23). Die iuristische Tätigkeit ist zweifellos das Leitmotiv von beiden Stücken, dem Phormio und dem griechischen Ἐπιδικαζόμενος des Apollodorus. Freilich könnte man auch im Zweifel sein, denn, wenn wir den Didaskalien und der Calliopianischen Überlieferung und dem Donatus glauben, wurde das Stück Phormio während der Ludi Romani (4-19 Sept.) aufgeführt, nachdem während der Megalenses (4-10 April) wurde der Eunuchus aufgeführt, und Terenz musste sich gegen den Angrif von Luscius Lanuvinus im Eunuchus (30-34) verteidigen, der in der Aufführung (Experimentieraufführung) gerufen hatte, dass Terenz parasiti personam […] et militis ‘a Naeui et Plauti ueteribus fabulis’ gestohlen hatte (Eun.25sq.). Terenz antwortete, er hatte beide Personen aus Menander genommen. Das ist wahr und Terenz hatte zwischen die Ludi Megalenses und die Romani Zeit genug, um den Phormio zu korrigieren und die Anklage, Plautus nach‐ geahmt zu haben, zu vermeiden. Es scheint, wie ich schon oben sagte, dass Eunu‐ chus und Phormio in demselben Jahr aufgeführt wurden, Eunuchus während der Ludi Megalenses (April) und Phormio während der Ludi Romani (September) 16 . In jener Zeit dürfte Terenz auf die Anklage besonders aufmerksam gewesen sein, die Luscius Lanuvinus gegen ihn erhoben hatte und die im Eunuchus wiedergegeben wird (Eun.15-34). In jenem Fall wurde Terenz gezwungen zuzu‐ geben, dass er vielleicht einen Fehler gemacht hatte (si id est peccatum, peccatum imprudentiast / poetae, non quo furtum facere studuerit). Es ist selbsverständlich, dass Terenz sich im Phormio in derselben Verlegenheit wiederzufinden, vermied, in der er sich im Eunuchus gefunden hatte. Gerade in den Adelphoe habe ich bewiesen, dass Terenz alles versuchte, um sich nicht in solch einer Situation sich zu befinden (vgl. G.Calboli 2011). Es wäre nämlich kein Wunder, hätte er versucht zu vermeiden, sich dem Plautus zu viel zu nähern. Es ist natürlich, dass Terenz auf den Schutz des griechischen Modells nicht verzichten wollte. Denn er konnte damit sagen, er habe nicht aus einem römischen Dichter genommen (ein furtum), sondern aus dem griechischen Original, das als Erzeugnis der Griechen, nämlich von deditici, den Römern zur Verfügung stand. Ich habe schon bewiesen, wie Terenz in seinem letzten Stück ‘Adelphoe’ sorgfältig sogar auf Kleinigkeiten aufpasste. So war das „Familiengericht“ am Ende des Phormio, nach der die Entscheidung dem Sohn Phaedria beauftragt wird, m. E. dem Apol‐ lodor zuzuschreiben, obgleich auch in Plautus’ Mercator ein Vorschlag gemacht wird, der nicht zuweit davon geht. Dass Terenz dieselbe Gewissenhaftigkeit, die er in den Adelphoe bewiesen hat, schon in Phormio so entwickelt hatte, Nausistrata (Ter. Phorm. 784-1055) 289 <?page no="290"?> 17 Vgl. S.Ferrucci (2006). dass er nicht wagte, Apollodorus’ Ende zu viel zu verändern, hätte man hierbei, als eine glaubwürdige Hypothese, vielleicht als die glaubwürdigste Hypothese darstellen können. Falls man Mattinglys (1959) und Bembinus’ Chronologie annehmen möchte, würde der Phormio vor den Eunuchus gestellt, und diese Argumentation wäre nicht mehr gültig, und man könnte denken, Luscius Lanuvinus hätte im Phormio auf den Angriff, der im Eunuchus stattfindet, wegen einer unbekannten Ursache verzichtet. Auf jeden Fall scheint es nicht leicht gegen Donatus’ Stellungsname zu gehen und diese Hypothese scheint kaum annehmbar. Doch wichtiger scheint eine andere Besonderheit: Nausistratas abschließende Stellungnahme, Phaedria sollte entscheiden (ei(u)s iudicio permitto omnia: quod is iubebit faciam), stellt uns vor eine interessante Frage: Nach dem vorigen Fall des Staatsrechts (Phor.447-458) wäre es fast lächerlich, mit dem Staatsrecht zu rechnen und das in Plautus’ Mercator (Plaut. Merc.1015-1026) vorgeschlagene Gesetz dürfte kein öffentliches und Staatsgesetz sein, sondern nur ein Familien- und Privatgesetz. Um alles zusammenzufassen, können wir sagen, dass man in dieser Komödie mit zwei Motiven zu tun hat, nämlich (1) dem Motiv des οἴκος, 17 des Gesetzes, das in der ἐπιδικασία, wenn sie vom klugen Phormio verwaltet wird, zu einem guten Erfolg bringt, als restitutio in integrum. Dagegen war die Beratung der drei Juristen (Phor.446-459) katastrophal und widersprüchlich, so dass sich Nausistrata als dritten Schritt auf ein privates Familiengesetz verlegt. In dieser Weise stellt sich Apollodors Stück als ein Iter legum dar, wodurch die Beziehung des οἴκος und der Stadt mit den verschiedenen Fratrien behandelt wird. Und dies kann freilich nur aus Apollodor kommen. Das zweite Motiv (2) ist gerade die uxor saeua, die besonders dem Plautus zuzuschreiben zu sein scheint. Was jedoch dieses Motiv angeht, müssen wir zwei Fragen beantworten: (a) wie konnte Terenz die Anklage vermeiden, Plautus nachgeahmt zu haben (d. h., die Anklage von furtum), diese hätte gegen ihn entweder von Luscius Lanuvinus, oder von jemanden anderem erhoben werden können; (b) warum sollte Terenz plötzlich die Aufmerksamkeit vernachlässigen, die wir (Calboli, 2011) in den Adelphoe bemerkt haben? Was die Frage (a) betrifft, wie wir oben gesehen haben, muss man zwei Möglichkeiten in Betracht ziehen, (a’) der Phormio wurde nach den Eunuchus aufgeführt, und in diesem Fall war Luscius schon gestorben oder Terenz dachte, die athenische Farbe der apollodorischen Nausistrata sollte ein ausreichender Schutz gegen jeden Angriff, sogar von Lucius, sein, denn er könnte immer antworten, dass Nausistrata schon athenisch 290 Gualtiero Calboli <?page no="291"?> 18 Eine Besonderheit des Phormio zu sein scheint, dass der Phormio als einzige römische Wiederherstellung des Έπιδικαζόμενος war, und daher ein anderer parasitus und miles nicht existierte, womit die entsprechende Persona des Phormio verglichen werden konnte. Aber kein miles im Phormio besteht, es bleibt also nur der parasitus, der mit Naevius und Plautus, wie im Eunuchus, verglichen werden konnte. Denn im Eunuchus 23-26 liest man: exclamat [sc. poeta vetus, ex Phormione 1] furem, non poetam fabulam/ dedisse et nil dedisse uerborum tamen ; / Colacem esse Naeui, et Plauti ueterem fabulam / parasiti personam inde ablatam et militis. Terenz scheint sich darum nicht viel zukümmern. Denn er schrieb (Phorm.25-28) Epidicazomenon quam uocant comoediam / Graeci, Latini Phormionem nominant / quia primas partis qui aget is erit Phormio / parasitu’. Aber kümmertete sich Donatus, p. 345 Hanc comoediam manifestum est prius ab Apollodoro sub alio nomine, hoc est Έπιδικαζομένου, Graece scriptam esse quam Latine a Terentio Phormionem. Quamobrem nulla dubitatio est hanc solam esse, cui nomen poeta mutauerit, et errare eos, qui in hac Phormionem parasitum putant a formula litis, quam intenderit, nominatum, cum Graeca fiscus sparteus et stramen nauticum sic dicatur: a cuius rei uel capacitate uel utilitate etiam ab Apollodoro parasitus Phormionis nomine nuncupatur. Wie kann man diese Verschiedenheit zwischen Terenz und Donatus erklären? M.E., weil Terenz vor dem Eunuchus schrieb, Donatus nach dem Eunuchus. Aber dies is nur eine Vermutung. war und aus Apollodors Ἐπιδικατζόμενος kam 18 . Wenn dagegen der Phormio vor dem Eunuchus angeführt wurde, galt Luscius’ Angriff nicht diesem Stück, nämlich dem Phormio, sondern später dem Eunuchus, also einige Monaten später als dass Terenz sich darum in Phormio kümmern konnte. Was dann (b) die Aufmerksamkeit angeht, die Terenz in den Adelphoe walten lassen hat, wie ich bewiesen habe, hat Terenz zwei Zwecke verfolgt, einerseits den Angriff, den jemand durchführen könnte (abgesehen davon, dass Luscius Lanuvinus schon verstorben zu sein scheint, Eun.15) zu verhindern, und anderseits die Wandelbarkeit des Aischinus zu zeigen (vgl. Calboli 2011: 9-10). Damit konnte Menander/ Terenz den Fehler der zu strengen und der zu gemäßigten Erziehung zeigen, und schließlich die Notwendigkeit, die Extreme zu vermeiden - hierbei auch mit einem doppelten Gesicht. So konnte der Stoff der Nausistrata aus Apollodor kommen, die Farbe der uxor saeua (sie wurde Phor.744 genau so genannt), aber nur die Farbe, eher aus Plautus. Was die Beziehung der römischen Komiker, und insbesondere des Terenz, zu den entsprechenden griechischen Authoren angeht, haben wir gesehen, wie kompliziert es war und wie vieler Vorsicht auch ein moderner Kritiker bedarf. Bibliographie Apolito, M. (2009) L’adulterio nel diritto Greco. Riflessioni sul diritto dell’antico oriente mediterraneo, Napoli, Museopolis Press. Arnaldi, F. (1947) Da Plauto a Terenzio II: Terenzio, Napoli, Loffredo. Nausistrata (Ter. Phorm. 784-1055) 291 <?page no="292"?> Blanchard, A. (1981) „La composition du « Phormion » et l’originalité de Térence“, Revue des Études Latines 58: 49-66. Calboli, G. (1974) „Un frammento di C.Laelius Sapiens? “, in AA.Vv., Poesia latina in frammenti, miscellanea filologica, Genova, Istituto di Filologia Classica, 141-172. Calboli, G. (1991) „Il miles gloriosus di Terenzio e l’infinito storico latino“, AAVv, Studi di filologia classica in onore di G.Monaco, II, Palermo, Istituto di Filologia Classica, 599-632. Calboli, G. (2009) ”Latin syntax and Greek“, in Baldi, Ph. & P. Cuzzolin (eds.) New perspectives on historical Latin syntax, Volume 1: Syntax of the sentence, Berlin, de Gruyter, s65-193. Calboli, G. (2011) „Donat à la recherche de Ménandre chez Térence“, Communication présentée à la première rencontre du Groupe de Recherche International (GDRI) ‘Grammaire et littérature’, Paris, Nov. 2011, Interferences 6. Calboli, G. (2020) Cornifici seu Incerti Auctoris Rhetorica ad C. Herennium, Vol.3, Lexicon, Berlin, de Gruyter. Coarelli, F. (1998) „Problemi e ipotesi sulle tavole greche di eraclea“, in Siritide e Metapontino. Storie di due territori coloniali (Atti dell’incontro di studio, Policoro 1991), Naples-Paestum, Cahiers du Centre Jean Berard 20: 281-290. Culasso Gastaldi, E. (2010) „L’epigrafia ateniese dell’isola di Lemnos“, in Culasso, E. & D. Marchiandi (ed.) Gli Ateniesi fuori dall’Attica. Modi d’intervento e di controllo del territorio, Paestum & Atene, Pandemos. Ferrucci, S. (2006) „L’OIKOS nel diritto attico, pubblico, privato e individuale nella democrazia ateniese classica“, DIKE, Rivista di storia del diritto greco ed ellenistico 6: 183-210. Ficuciello, L. (2012) „Il territorio di Myrina (Lemno): indizi sull’occupazione e sullo sfruttamento delle risorse“, Annuario della Scuola Archeologica di Atene e delle Missioni Italiane in Oriente 88, ser. III, 10: 237-269. Focardi, G. (1972) „Linguaggio forense nei prologhi terenziani“, Studi Italiani di Filologia Class. Ν.S. 44: 53-88. Goldberg, S. M. (1986) Understanding Terence, Princeton, NJ, Princeton University Press. Kuiper, W.E.J. (1938) Two commedies by Apollodorus of Carystus, Terence’s Hecyra and Phormio, Leiden, Brill. Lefèvre, E. (1978) Der Phormio des Terenz und der Epidikazomenos des Apollodor von Karystos, München, Beck. Marti, H. (1964) „Terenz 1909-1959“, Lustrum 8: 5-101. Mattingly, H.B. (1959) „The terenzian didascaliae“, Athenaeum n.s. 37: 148-173. Michel, A. (1987) „Esthétique et moralité dans le Phormion; le decorum de Térence“, Filologia e Forme Letterarie, Studi offerti a Francesco Della Corte, Vol. II, Urbino, Universita degli Studi di Urbino, 111-126. 292 Gualtiero Calboli <?page no="293"?> Müller, R. (1997) Sprechen und Sprache, Dialoglinguistische Studien zu Terenz, Heidelberg, Winter. Osborne, R. (1988) „Social and economic implications of the leasing of land and property in classical and hellenistic Greece“, Chiron 18: 279-323. Posani, M.R. (1942) „Le didascalie delle commedie di Terenzio e la cronologia“, Rendiconti dell’Accademia d’Italia, 1. Suppl. vol. 3, 247-280. Nausistrata (Ter. Phorm. 784-1055) 293 <?page no="295"?> Syntactic Theory and Textual Criticism in Plautus: Camerarius’ Emendation of Amphitruo 260 Michael Weiss & Ryan Windhearn (Cornell University, Ithaca, NY) Abstract: In this paper, we combine the methods of traditional textual criticism with the insights of syntactic theory. We bring new evidence against an emendation of Plautus, Amphitruo 260 proposed by the Renais‐ sance scholar Camerarius and found in most modern editions by arguing that the word order that this emendation produces was likely dispreferred by Old Latin speakers. We conclude that Pareus’ (1641: 12) emendation of the same line shows a word order more in keeping with what we would expect from Old Latin-syntax.- Keywords: Plautus; Old Latin word order; syntactic theory; textual criti‐ cism. 1. Introduction In this paper, we bring new evidence against an emendation of Plautus, Amphitruo 260 proposed by Camerarius (1545: 13) and found in most modern editions by arguing that the word order that this emendation produces was likely dispreferred by Old Latin speakers. We conclude that Pareus’ (1641: 12) emendation of the same line shows a word order more in keeping with what we would expect from Old Latin syntax. Our argument stems from the recent syntactic analysis of the auxiliary constructions of Amphitruo and Cato’s De Agri Cultura in Windhearn 2020, part of a project comparing and reconstructing clausal syntax across the oldest attested Indo-European languages. Our analysis here is based on an expanded Old Latin corpus that includes all of the auxiliary <?page no="296"?> 1 The complete data set is posted at https: / / cornell.academia.edu/ michaelweiss. The authors would like to thank Michael Fontaine, Benjamin W. Fortson IV, and John Whitman for valuable comments on a draft of this paper. Michael Weiss is responsible for the discussion of metrics and textual issues and for the expanded corpus of auxiliary constructions in Old Latin. Ryan Windhearn is responsible for the syntactic analysis and the Amphitruo and De Agri Cultura auxiliary corpus. 2 The manuscript can be seen here: https: / / digi.ub.uni-heidelberg.de/ diglit/ bav_pal_lat_ 1615/ 0028/ image,info. D, Vaticanus Latinus 3870, 10 th -11 th century, which can be seen here https: / / digi.vatlib.it/ view/ MSS_Vat.lat.3870, agrees in all essentials. The Amphitruo does not survive in C (Codex Palatinus Heidelbergensis (Pal. lat. 1613)) nor in the Ambrosian palimpsest nor is any line of it quoted by Turnebus from the lost Codex Turnebi of the Palatine family. The lesser lights of the Palatine family, J and E, are not reported to differ in any essential way in this passage, though we have not been able to examine digital images of them. 3 The reading poare in the main text of B is corrected to potare by a second hand and is also the reading of a corrector in D and is the only reading in E. The reading before correction in D is positare which has been emended to the metrically required potitare in the Italici. 4 All translations from Plautus are from the Loeb edition of Wolfgang de Melo. constructions in the works of Livius Andronicus, Naevius, Ennius, and the entire corpus of Plautus. 1 2. Amphitruo 260: History and Textual Criticism In act I, scene 1 of Plautus’ Amphitruo Sosia recounts Amphitruo’s victory over the Teloboae and the Teloboae’s surrender (256-261) Here the text is given as in B (Palatinus Latinus Vaticanus 1615, 10 th -11 th cent.= Codex uetus Camerarii): 2 (1) postridie in castra ex urbe ad nos ueniunt flentes principes uelatis manibus orant ignoscamus peccatum suum deduntque se, diuina humanaque omnia, urbem et liberos in ditionem atque in arbitratum cuncti Thebano poplo post ob uirtutem ero Amphitrioni patera donata est aurea. qui Pterela potare 3 rex solitus est. haec sic dicam erae - “The next day their leaders came from the city to our camp, crying, and with covered hands-they asked us to forgive them their transgression. They all surrendered themselves, all their sacred and profane possessions, their city and their children, into the power and sway of the Theban people. Afterwards-my master Amphitruo was presented with a golden bowl for his valor, the one from which King Pterela used to drink. That’s how I’ll tell my mistress the story.” 4 Though the meaning of the passage is clear, the meter, iambic octonarii with a caesura after the first element of the third metron, has been corrupted in lines 296 Michael Weiss & Ryan Windhearn <?page no="297"?> 5 So Bothe 1809: 15; Naudet 1830: 31; Fleckeisen 1850: 13; Leo 1885: 13; Goetz and Schoell 1898: 13; Lindsay 1904 l. 260 (no page number); Ernout 1932: 23; Christenson 2000: 91; de Melo 2011: 32, 34. 6 The 1545 edition can be viewed here https: / / www.digitale-sammlungen.de/ en/ view/ bs b10171214? page=80,81. In the first edition of 1535 Camerarius still printed donata’st aurea (1535: 13). The 1535 edition can be viewed here http: / / digital.onb.ac.at/ OnbViewer/ viewer.faces? doc=ABO _%2BZ17594810X at p. 17. 260-261 and various emendations have been proposed since the Renaissance to fix the problems. We are here only interested in the solutions that have been offered to heal line 260 which is presented below with metrical notation in the style of Gratwick 1999. (2) A B C D a B c c D A/ b b c D A B C-? ? - post ob uirtut(em) er(o) Amphitrioni patera donat(a) est aurea The transmitted text is clearly corrupt. The final C position must be occupied by a short syllable, which is not the case in the scansion above, and the final two syllables of aurea need to occupy one metrical position, but resolution is not permitted in the final foot of an iambic line. Obviously a change is required and if we compare the major editions of last 213 years, we find that they almost all agree in giving lines 260 as follows: 5 (3) A B C D a B c c DA/ b b c D A B c D - post ob uirtut(em) er(o) Amphitruoni patera donat(a) aure(a) est, The favored solution, which goes back to the Renaissance scholar Joachim Camerarius ( Joachim Liebhard, known as Kammermeister 1500-1574) 1545: 13, 6 and which was quickly adopted by e.g. Lambinus 1576: 15, Dousa 1594, Gruterius 1621: 20 etc., restores the meter by switching the order of the last two words. This allows the final a of aurea to be elided producing the needed final c D sequence. Metrically and paleographically, this solution is perfect. We have demonstrable cases of the linear order of words being flipped. Lindsay 1896: 31 comments that “[t]ransposition of words is perhaps the commonest error of MSS., so that a change in the order of the words is usually the least violent remedy that an editor can apply to an unmetrical line. As an instance from Plautus we may take Stich. 293 a trochaic septenarius, which in all the minuscule MSS. has the impossible ending censeo aequum, but in A ends correctly with aequom censeo.” Syntactic Theory and Textual Criticism in Plautus 297 <?page no="298"?> Nevertheless Camerarius’ solution is not the only one to have been proposed. Ritschl, Loewe, Goetz, and Schoell 1882: 30 report that “est Pylades patera, Pareus Amphitruoni uocabulis adiungunt” “Pylades attaches est to the word patera and Pareus to the word Amphitruoni”. Ussing 1875: 260 notes “est metri causa transposuit Pylades” “Pylades transposed est for metrical reasons” and Ussing himself follows this emendation, printing paterast in his own text (Ussing 1875: 15). Pylades refers to the Lombard humanist Giovan Francesco Boccardo Pilade and Pareus refers to Johannes Philippus Pareus (= Johann Philipp Wängler, 1576-1648) who published three editions of Plautus in 1610, 1619, and 1641. In his 1506 edition of all the plays of Plautus Pylades does not print what both Ritschl and Ussing report but instead offers the manuscript reading. The transposition presumably was suggested in his 1522 edition of just the Amphitruo, though we have not be able to confirm this by autopsy. Pareus more or less kept the manuscript reading in the second edition (Pareus 1619: 13 donata’st) but offered his own correction in the third edition (Pareus 1641: 12 Amphitruonist). Pylades’ reading and Pareus’ reading are both metrically fine: Pylades: (4) A B C D a B c c D A/ b b c D A B c D - post ob uirtut(em) er(o) Amphitruoni patera’st donat(a) aurea, Pareus: (5) A B C D a B c c D A/ b b c D A B c D - post ob uirtut(em) er(o) Amphitruoni’st patera donat(a) aurea Paleographically, Camerarius’ suggestion is a bit stronger because one could see how a line-final est in a long line such as the iambic octonarius could have been written above the line and then inserted in the wrong place and this is presumably the reason why Camerarius’s transposition has been generally adopted. The only alternative proposal that we are aware of that does not involved changing the position of est is the more radical solution offered by Havet 1895: 18 who suggested: (6) A B C D a B c c DA b b c d d A B c D - post ob uirtut(em) er(o) Amphitruoni patera dat(a) ibist aurea In support of this emendation Havet writes: 298 Michael Weiss & Ryan Windhearn <?page no="299"?> data ibist L. Havet cf. illi 761 766 780 : donata est ortum fortasse ex dccτaιuιsτ (in ceteris locis ut 418 763 771 aut datur alicui pateră aut donatur aliquis paterā) “L. Havet (proposes) data ibist. Cf. illi ‘there’ at lines 761, 766, and 780. donata est has arisen from dccτaιuιsτ (in other loci like 418, 763, 772, (Plautus writes) either a bowl (pateră nom. sg.) is given (datur) to somebody (alicui) or someone (aliquis) is gifted (donatur) with a bowl (paterā abl. sg.)” But Havet’s approach is too bold. The three other passages cited by Havet do have a local particle illi ‘there’ but the presence of illi in those passages does not support the presence of ibi in this passage. In all of the cases cited by Havet there is implicit or explicit contrast between the giving of the patera to Amphitruo by the defeated Teloboae in the past (illic et tunc) with Amphitruo’s supposedly having given it to Alcumena now in Thebes today (hic et nunc) (7) Al. opsecro, etiamne hoc negabis, te auream pateram mihi dedisse dono hodie, qua te illi donatum esse dixeras? “Please, will you also deny that you gave me the golden bowl as a gift today, the one you’d said you were presented with there? ” Amph. 760-761 (8) Amph. mane, mane, opsecro te. nimis demiror, Sosia, qui illaec illi me donatum esse aurea patera sciat, “Wait, wait, please. (to Sosia) I’m highly surprised, Sosia, how she knows I was presented with a golden bowl there…? ” Amph. 766-767 (9) Al. estne haec patera qua donatu’s illi? “Isn’t this the bowl you were presented with-there? ” Amph. 780 In the context of Sosia’s monologue there is no contrast at all because there is no second recent act of giving for the former giving to be contrasted with. The second point about the syntactic constructions of the two verbs is also not airtight. Although it is true that donare is often construed with an accusative of the recipient, which becomes the nominative when the sentence is made passive, and an ablative of the gift, this is not an exceptionless rule either in Latin or in Plautus. For example in the Persa 775 a we read hoc mea manus tuae (dat.) poclum (acc.) donat, “my hand gives this cup to your (hand)”. 3. The Clausal Syntax of Old Latin Syntactic analysis of Old and Classical Latin has historically been difficult due to a combination of the lack of negative data from native speakers and the relatively Syntactic Theory and Textual Criticism in Plautus 299 <?page no="300"?> free word order of the language, but over the past few decades there have been increasing numbers of analyses within modern generative frameworks (e.g. Ledgeway 2012, Danckaert 2012 and 2017, etc.). Windhearn 2020 is one such analysis that specifically addresses the syntax of Old Latin in Plautus’ Amphitruo and Cato the Elder’s De Agri Cultura, featuring the innovative approach of using auxiliary constructions to more precisely determine the hierarchical relationships between verbal elements in the clause. Auxiliary and modal constructions are uniquely suited to showing the syntactic relationship between tense and the lower verbal domain, as demonstrated by the following New High German examples: (10) a. Jill weiß dass ich den Apfel - esse. - - Jill knows that I the apple - eat - - “Jill knows that I am eating the apple.” - b. Jill weiß dass ich den Apfel gegessen habe. - - Jill knows that I the apple eaten have - - “Jill knows that I have eaten the apple.” - c. *Jill weiß dass ich den Apfel habe gegessen - - Jill knows that I the apple have eaten The first sentence with just a single inflected verb in the subordinate clause doesn’t tell us as much about the clause structure of German as the second sentence does: the fact that b) is grammatical and c) isn’t shows us that German requires a very specific configurational relationship between the inflectional features in T and the participle in the lower verbal domain, and greatly restricts the number of possible syntactic analyses of these data as a result. Auxiliary data like this is a large part of why many syntactic analyses of modern German V2 syntax (Haider 1993, etc.) posit right-headedness in T and V to derive these relationships: 300 Michael Weiss & Ryan Windhearn <?page no="301"?> 7 We define “auxiliary construction” as a monoclausal construction consisting of an inflected auxiliary (or embedded non-finite auxiliary) with an accompanying participle. Multiclausal constructions consisting of inflected verbs and embedded non-finite verbs, e.g. a verb plus an infinitive, were not considered, though they also overwhelmingly show the word order we would expect in a right-headed analysis of Old Latin syntax. Furthermore, auxiliary constructions are doubly useful in languages like Latin that show a large variety of syntactic movement processes. When we look at the Old Latin corpus specifically, as free as the word order is, we find that there are only a relatively small number of attested configurations of participles and their inflected auxiliaries, all of which must be derivable from whatever basic structural relationship exists in the language between V (instantiated by the participle) and T (instantiated by the auxiliary). Windhearn 2020 looked at 191 examples of auxiliary constructions 7 in Cato’s De Agri Cultura and Plautus’ Amphitruo to determine the structural relationship between the inflected auxiliary and the participle lower in the clause, and found that 70% of the time participles and their inflected auxiliaries appear at the end of their clauses, in that order, i.e. Participle-Auxiliary. This behavior is consistent across both pragmatically unmarked clauses and marked constructions like wh-questions, as seen in the examples below. (11) certe aduenientem hic me hospitio pugneo accepturus est. “He’s definitely going to give me a fisty welcome-on my arrival.” Amph. 296 (12) id in suggestu inter dolia positum habeto “Keep it on the elevation among the jars.” Agr. 154.1 (trans. Hooper and Ash 1934) Syntactic Theory and Textual Criticism in Plautus 301 <?page no="302"?> 8 Unless there are other factors. It’s common in traditional English poetry to see ungrammatical word orders for metrical or stylistic reasons. Compare Milton’s proem of Paradise Lost (OF Mans First Disobedience, and the Fruit / Of that Forbidden Tree, whose mortal tast/ Brought Death into the World, and all our woe,/ With loss of Eden, till one greater Man/ Restore us, and regain the blissful Seat, Sing Heav’nly Muse…) or a line from Milton’s Sonnet IX (The better part with Mary and with Ruth Chosen thou hast) or Pope’s An Essay on Criticism (Our sons their fathers’ failing language see/ And such as Chaucer is shall Dryden be). See Ferber 2019: 104-141 for a discussion of English poetic syntax. But the poetry of Plautus is rather different than the high style of Milton or Pope, who were themselves, of course, highly influenced by the possibilities of Latin syntax. The Plautine plays were written mostly in flexible iambo-trochaic meters which must have permitted a skilled poet like Plautus to stay close to natural language. Further the plays of Plautus were performed as popular entertainment which suggests that they must have been accessible to the average Latin speaker without too much post-syntactic processing. The syntax of poetry does not always conform to the rules of spoken syntax, but we should, especially in a genre like Roman comedy, favor the analysis that fits with syntactic norms. (13) sed quid tu foras egressa es? “But why did you come out? ” Amph. 1078 Furthermore, when embedded within another clause, we commonly see em‐ bedded clauses showing up immediately prior to the clause-final participle and auxiliary of the matrix clause: (14) nisi [etiam hoc falso dici] insimulaturus es “[…] unless you’re going to allege that this is also a lie.” Amph. 901 This is exactly the behavior we would expect if embedded clauses were com‐ plements of right-headed verbal domains. From all of this data, Windhearn hypothesized a clausal syntax of right-headedness in T and V for Old Latin to account for this behavior, similar to the analysis of modern German subordinate clauses above. If this analysis is correct, it makes predictions about what word orders we expect to see (and not see) in Old Latin texts: 8 namely, all of them should be derivable from a right-T, right-V structure. Windhearn 2020 demonstrated that this is the case for all attested auxiliary constructions in Amphitruo and De Agri Cultura except one: Amphitruo 260. In preparing the present paper, we expanded this corpus to include all examples of auxiliary constructions in the works of Livius Andronicus, Naevius, Ennius, the complete works of Plautus as well as Cato’s De Agri Cultura. This gave us a corpus of 1,324 auxiliary constructions, 944 of which 302 Michael Weiss & Ryan Windhearn <?page no="303"?> (71 %) show the expected participle-auxiliary word order, with the remaining 380 tokens falling into the following three categories: 1. Participle-X-Auxiliary (“PXA”, 95x) instances, where the participle and auxiliary occur in the expected order, but with unexpected intervening elements 2. Auxiliary-Participle (“AP”, 226x) instances, where the participle and auxil‐ iary occur in the opposite order we would expect 3. Auxiliary-X-Participle (“AXP”, 59x) instances, where the participle and auxiliary occur in the opposite order and have intervening elements From the data, it’s clear that metrical factors do play a role in Plautine word order. For example, when the auxiliary is sumus, the most common order is participle-sumus which is normally localized at line end (17x: Amph. 731, Asin. 361, Asin. 367, Bacch. 720, Capt. 253, Cist. 121, Men. 147, Men. 654, Men. 743, Mil. 83, Mil. 257, Mil. 612, Most. 365, Poen. 285, Stich. 740, Tri. 942, Tri. 1104), since sumus provides a convenient iamb for the line end of iambic senarii and trochaic septenarii. (vs. 7x in the order sumus-participle in other loci: Epid. 406, Epid. 719, Men. 235, Mil. 117, Poen. 1185, Poen. 1201, Pseud. 543). The figures for the 2 nd plural form estis in periphrastic constructions are more evenly divided. The order estis-participle is attested 6x (Epid. 376, Merc. 750, Pers. 465, Pseud. 148, Rud. 268, Stich. 529), vs. 3x for the order participle-estis (Amph. 89, Most. 62, Truc. 4). It seems unlikely that this distribution has a purely syntactic explanation. Participle-sumus sumus-Participle Ratio Participle-estis estis-Participle Ratio 17 7 2.4: 1 3 6 .5: 1 Even so, we find that the vast majority of instances are still derivable from right-headed TP and VP syntax, as demonstrated below. Syntactic Theory and Textual Criticism in Plautus 303 <?page no="304"?> 9 Note that this postposition analysis follows from the right-headed structural assump‐ tions of this paper. These examples are usually analogous to the weak focus construc‐ tions discussed in Devine and Stephens 2006: 176-179, in which a left-headed VP and TP are generally assumed for Latin. Devine and Stephens analyze these as involving focus of the tensed verb around the other constituents in their own lower topic and focus positions, following the project’s information structure-based account of Latin word order. In this paper we instead primarily rely on syntactic explanations for the word orders seen in the data, and though we agree that these examples involve weak focus, we rely on an analysis of right-dislocation instead of verb focus. This is advantageous for two reasons. First, in cases like these where the participle and auxiliary are pragmatically unmarked and show up before the constituents bearing weak focus appearing to their right, their unmarkedness is reflected in their base-generated positions, the right-headed VP and TP. Thus, the constituents being weakly focused are themselves the phrases being displaced from the expected linearization. Second, auxiliaries are phrasal heads, and as such can only undergo head movement, and not phrasal movement to a focus position. As such, raising of the verb to a focus position is an insufficient explanation when the verb in question is an auxiliary, and our in-situ explanation of these examples is better motivated syntactically. 3.1. Right-Dislocation The first thing to note about the Old Latin data is the frequency of right-dislo‐ cation, in which one or more constituents may be postposed after the rest of the clause 9 : (15) nam noctu hac soluta est [nauis nostra] 1 [e portu Persico] 2 “This night we set sail from Port Persicus.” Amph. 412 (16) ut cum exercitu / hinc profectus sum [ad Teloboas hostis] “ever since I and the army went away from here to our enemy, the Teloboians,” Amph. 733-734 Following Windhearn 2021, we treat these as right dislocation of the sort described in, e.g., Kuno 1978, Kayne 1994, and Whitman 2000. After Kayne 1994, we assign them a biclausal structure in which the first clause has been raised to the specifier position of a functional category that takes the clause containing the right-dislocated elements as its complement. Only the higher copy of the raised clause is pronounced. 304 Michael Weiss & Ryan Windhearn <?page no="305"?> 10 Note that “focus of the participle” here refers to a syntactic operation, not a pragmatic one, and that movement to this position can be for a number of different pragmatic reasons in addition to prototypical informational or contrastive focus. Ben Fortson (per litteras electronicas) notes that in (18) it’s really in gratiam that matters, and that reuentum is gotten out of the way because it’s semantically extraneous. We agree, but propose that spec-FocP is the syntactic position this “getting out of the way” operation is likely targeting as reuentum still constitutes discourse-new information. Structure of Right Dislocation (from Whitman 2000) Languages with null pronominals, such as the old Indo-European languages, often do not show the mandatory resumption we see in the English examples above, and often allow for multiple right dislocation (Endo 1996), which is also common in the Old Latin data. One pattern we commonly see in clauses with participle-auxiliary-X (PAX) word order across the old Indo-European languages, including Latin, is the right dislocation of a part of a constituent, usually a modifier, resulting in discontinuous constituency. This is shown in the example below where uostra ‘your’ is the right dislocated part of the phrase [opera uostra] ‘your work’. (17) intro abi: [opera huc conducta est uostra], non oratio. “Go inside. Your were hired for your work here, not for your talk.” Aul. 455 In these examples we assume speakers are pronouncing the higher copy of the head, and the lower copy of the modifier. 3.2. Accounting for participle-X-auxiliary (PXA) Word Order The class of instances, which show the intervention of some element between the participle and auxiliary, usually involves focus of the participle 10 . (18) rursum si reuentum in gratiam est “and if they’re reconciled again” Amph. 942 Syntactic Theory and Textual Criticism in Plautus 305 <?page no="306"?> 11 Both these examples and the AXP examples below belong to the set of phenomena Devine and Stephens 2006: 181-83 analyze as auxiliary raising to focus. 12 For a thorough overview of this incipient V2 behavior seen in early Latin, cf. Ledgeway 2017. 13 Here a subscript “F” indicates focus, and “R” indicates right-dislocation. (19) sed tamen, tu nempe eos asinos praedicas / uetulos, claudos, [quibus suptritae ad femina iam erant ungulae]? “But are you talking about those old, lame donkeys [whose hooves had been worn away up to the thighs]? ” Asin. 339-340 In the second example we see a combination of participle-focus (suptritae) and right dislocation (ungulae). 3.3. Accounting for AP Word Order The second and most common class of instances in Old Latin, Aux-Participle (AP), show the opposite order of auxiliary and participle that we would expect clause-finally 11 : (20) quia pudicitiae huius uitium me hinc apsente est additum “because her chastity’s been violated during my absence from here.” Amph. 811 (21) ita tanta mira in aedibus sunt facta “Such strange things have happened in the house.” Amph. 1057 (22) Fateor: nam sum optusus pugnis pessume “I admit it: I was beaten up horribly with fists.” Amph. 606 These examples we analyze as showing either T-to-C movement 12 (22) or participle right dislocation (20 and 21), a phenomenon that, according to Samek-Lodovici (2015), still exists in Latin’s daughter Italian over two thousand years later (along with the multiple right-dislocation also widespread in Latin): (23) Lo abbiamo a MARCO F , 13 il tavolo R , riportato R . - it have-1PL to Mark the table bring.back-PART - “We brought the table back to MARK.” Samek-Lodovici (2015: 186) 3.4. Accounting for AXP Word Order The final class of instances show reversed order of the participle and auxiliary, with some other element also intervening between the auxiliary and participle. 306 Michael Weiss & Ryan Windhearn <?page no="307"?> 14 Note also that many of these examples are in embedded clauses, where Devine and Stephens 2006: 184-5 note that fronting of the auxiliary is common. It’s possible that one early Latin trigger of T-to-C constructions was clausal subordination. This is the smallest, but most diverse, class of subregularities, usually showing right-dislocation combined with various constructions featuring T-to-C move‐ ment of the auxiliary, such as the est-ne question in (26) 14 . (24) quo pacto sit donis donatus plurumis “how he was presented with a great many gifts.” Amph. 137 (25) Ne quoquam exsurgatis, [donec a me erit signum datum] “don’t get up to go anywhere, [until you get a sign from me].” Bacch. 758 (26) estne empta mi istis legibus? “Is she mine on those terms? ” Epid. 471 In (24) we have T-to-C movement of sit and right-dislocation of plurimis. (25) shows the common topicalization of the prepositional phrase combined with T-to-C movement of erit. (26) shows T-to-C movement of est and right dislocation of mi and istis legibus. In this manner the vast majority of the auxiliary constructions seen across the Old Latin corpus can be derived from a right-headed TP/ VP structure. In the next section we will refocus our discussion on the one case that cannot be accounted for in this fashion. 4. A theoretical-syntactic perspective on Amphitruo 260 The only example in our corpus not straightforwardly derivable from a right-T, right-V structure in the manner detailed above is Amphitruo 260 as emended by Camerarius, reproduced below. (27) post ob uirtut(em) er(o) Amphitruoni patera donat(a) aure(a) est Here, at the end of the clause we have the head of a discontinuous subject, patera, immediately followed by the participle of the auxiliary construction, donata. After that comes the rest of the discontinuous subject, the adjective aurea, finally followed by the inflected auxiliary. None of the movement operations discussed above will easily account for this word order in any combination, which is what spurred us to check the Syntactic Theory and Textual Criticism in Plautus 307 <?page no="308"?> 15 It’s worth mentioning that some researchers may resort to scrambling here, but we think this unlikely both because of the lack of parallels for the word order of this verse elsewhere in Plautus, as well as the lack of other examples in our Old Latin auxiliary corpus requiring a scrambling analysis in general. transmission in the first place and to discover the emendation of Camerarius. If the headedness analysis in Windhearn (2020) is correct, the expected base word order for the non-topicalized portion of the clause before any movement processes would be patera aurea donata est: From here, the main movement processes available to us are topicalization, focus, and right dislocation. Multiple topicalization is common in Latin, but topicalization stereotypically involves the fronting of familiar information to recenter that constituent in the discourse (Chomsky 1965, Gundel 2004, etc.). This is why ero Amphitruoni is topicalized: to reconnect the specific sentence with the topic of the discourse, namely Amphitruo. The patera, however, which is mentioned here in the play for the first time, and the fact that it was given, constitute new discourse information, so it is not the case that patera or donata can be topicalized here in Camerarius’ text. Another option is focus, but while discourse-new information is a prime candidate for informational focus, the most commonly accepted structure for the left periphery, that of Rizzi (1997), only allows for one element to be focused, and both patera and donata would have to be focused to give us the word order seen in Camerarius. 15 This leaves us with right dislocation of aurea and est as the only remaining solution. The problem here is that tensed verbs are rarely considered to show canonical right dislocation cross-linguistically, as seen in Lambrecht’s 2001 308 Michael Weiss & Ryan Windhearn <?page no="309"?> 16 Though see, e.g., Durham 2011 and other analyses of English auxiliary right-dislocation, e.g. She’s got a very good degree has Julie, as an extension of subject right-dislocation, which Durham notes is often viewed as non-canonical. In any case the Old Latin data doesn’t show any examples of unambiguous tensed verb right-dislocation, potentially because they would be in all cases undifferentiable from clause-final matrix verbs, creating a learnability issue given Latin’s extensive use of the left periphery. This makes sense pragmatically as well: as Lambrecht 2001 states, right dislocations “serve to secure the continued attention of an addressee”, and there’s no need to secure the continued attention of an addressee to a constituent that’s already clause-final. This is mirrored in the fact that the right dislocated elements in English are clause-initial subjects or, in some varieties, medial verbs, not clause-final objects. 17 See Ram-Prasad 2022: 138 for discussion of the landing sites of wh-words in relatives clauses in Proto-Indo-European and Old Latin vs. Classical Latin. survey of dislocation constructions across the world’s languages. 16 Thus, if this right-headed analysis of Old Latin TP structure is correct, there is no straightforward way through any combination of the mechanisms above to generate this word order. Furthermore, there are no other examples in our Old Latin corpus that show the word order seen in Camerarius’ emendation. The closest we get are examples like the following, where the participle occurs near the beginning of the clause in clear examples of topicalization or focus, in contrast to the unexpectedly low landing site of the participle required by Camerarius’ emendation of Amphitruo 260. (28) sed tamen, tu nempe eos asinos praedicas / uetulos, claudos, [quibus suptritae 17 ad femina iam erant ungulae]? “But are you talking about those old, lame donkeys [whose hooves had been worn away up to the thighs]? ” Asin. 339-340 We should note that it may still be possible that the word order seen in Camerarius is original, deviating slightly from grammaticality in order to fit the meter, but since the word order in Camerarius is already an emendation with existing alternatives, it seems to us preferable to favor an explanation in line with a syntactic analysis that can derive the other word orders seen in the work, especially when no other examples in the corpus show the word order of the emendation. It should also be noted that the corruptly transmitted text itself shows exactly the sort of syntax we would expect, with patera donata est aurea showing a discontinuous subject DP followed by the participle, inflected auxiliary, and the right-dislocated remnant of the subject constituent. This is a typical example Syntactic Theory and Textual Criticism in Plautus 309 <?page no="310"?> of the PAX-type above, one of the most frequently attested sub-patterns for auxiliary constructions in Old Latin and across the old Indo-European languages in general. In sum, at this point we have an emended text that fits the meter but shows unlikely syntax according to Windhearn 2020, along with a transmitted text that shows expected syntax, but doesn’t fit the meter. From these facts, a few possible explanations present themselves. If the trans‐ mitted text shows exactly the syntax we would expect, but doesn’t fit the meter, then it’s possible that the transmitted text itself has been changed from its original form to make better syntactic sense to whoever was responsible for the change. So, we turn next to consider the other proposed emendations to see if we have a likely candidate for the original word order. In Havet we find exactly the same syntactic issue we see in Camerarius in addition to the other shortcomings noted above, which leaves the emendations of Pylades and Pareus. In Pareus (repeated below as (29)) we find an unproblematic ordering of constituents. (29) post ob uirtut(em) er(o) Amphitruoni’st patera donat(a) aurea Immediately after the topicalized ero Amphitruoni we find T-to-C movement of the inflected copula, followed by the head of the discontinuous subject and participle in the expected order, and concluded by the right dislocated modifier. This word order is mirrored by the few dozen other APX examples in the corpus which show raised auxiliaries and right dislocations, of which a large number also similarly feature a topicalized dative or ablative adjunct immediately before the focused auxiliary in C (Fin 0 ). We have also included an AXP example to illustrate the similarly common pattern of topicalized adjunct followed by the raised auxiliary, continuous subject, and participle. (30) tum quae hic-sunt-scriptae litterae “The letters which are written here” Bacch. 941 (31) Seruos eius qui hinc a nobis-est-mercatus mulierem, “the slave of that [Macedonian soldier] who bought a woman from us here.” Pseud. 617 (32) Ne quoquam exsurgatis, donec a me-erit-signum datum. “don’t get up to go anywhere, [until you get a sign from me].” Bacch. 758 310 Michael Weiss & Ryan Windhearn <?page no="311"?> 18 The word order in this line is unanimously transmitted but it has not gone unchallenged. As transmitted, the caesura falls after the sixth element (quo pacto sit donis / donatus plurimis A B C D A B / C D A B c D), a relatively rare type in Plautine iambic senarii. See Questa 2007: 333. In addition this line might violate Meyer’s Law which is formulated by Fortson 2008: 33 as follows: “If an unaccented word-end falls in the fourth or eighth position in the senarius, or the seventh or eleventh position in the septenarius, it cannot be preceded by a heavy syllable (or its resolution of two shorts). Salmasius (as reported in Ritschl, Loewe Goetz, and Fridericus Schoell 1882: 17) suggested a transposition quo pacto donis sit donatus plurimis which solves both problems and is adopted by some editors, e.g. Leo 1885: 9. Nevertheless the line is not unmetrical though it may violate some lower level constraints. Marx 1922: 171 defended the transmitted reading and pointed out that on the basis of inscriptional evidence donis donatus is almost a fixed and inseparable unit. If we accept the transmitted reading, this example also shows that the auxiliary can be separated from its participle by a caesura. Another example of separation of the auxiliary and participle by the caesura is Poen. 802 Non sum nequiquam miles-factus: paululum A B C D A/ B C D A B c D. Contraction of est before the caesura can be paralleled by Ter. And. 44 quas(i) exprobratiost / immemoris benefici a B c D a B / C dd A bb c D. We even have one example in the corpus, Amphitruo 136, that nearly perfectly matches the syntax of Pareus’ emendation, with a topicalized adjunct followed by the focused auxiliary, then a discontinuous, partially-right-dislocated con‐ stituent split by the participle: (33) quo pacto sit donis donatus plurumis “how he was presented with a great many gifts.” Amph. 137 18 Another example, Stichus 540, shows a similar overall word order in which a discourse-old subject has been topicalized instead, followed by the raised auxiliary, a discontinuous adjunct, and then the participle: (34) eae-erant-duobus nuptae fratribus “they were married to two brothers” Stichus 540 All of this makes Pareus’ emendation a strong contender for the original word order of the line. In Pylades, however, we find similar problems to the issues we saw in Camera‐ rius, though to a lesser degree: (35) post ob uirtut(em) er(o) Amphitruoni patera’st donat(a) aurea Syntactic Theory and Textual Criticism in Plautus 311 <?page no="312"?> 19 Another solution suggested to us by Ben Fortson is to transpose donata est and patera yielding post ob uirtut(em) er(o) Amphitrioni donat(a) est pater(a) aurea A B C D a B c c D A/ A B C a a B c D. This line is also synactically unobjectionable since the patera aurea can be a case of right dislocation. Here we see the copula occurring after the subject, but before the participle and clause-final adjective, in an APX configuration. The participle is occurring before the second half of its NP constituent subject, and right-dislocated constituents must maintain base word order, so the participle and the adjective can’t have both been right dislocated. The left periphery only allows for once instance of focus, so we can’t focus both patera and the auxiliary. The only possibility would be to assume that patera was also topicalized for some reason, and then followed by the focused auxiliary. And even then, we only ever find topicalized discontinuous subjects appearing before topicalized adjuncts in similar examples elsewhere in Plautus. Either way, even allowing for an unexpected order of multiple topicalization in this example, Pylades’ emendation doesn’t fit the pragmatic context of the clause as mentioned earlier, as there is no reason for patera to have been topicalized. As such, this emendation remains less likely than that of Pareus, even though this emendation would put the auxiliary closer to its transmitted position after donata. If Pareus’ emendation is indeed the correct one, it’s possible that the trans‐ mitted text had been altered due either to pure accident or to confusion by a less-than-native reader at seeing the copula in so high a position in the clause. If this error were due to grammatical confusion, then it makes sense that the metrically irregular transmitted text would show exactly the syntax we would expect. If, on the other hand, the transmitted text is merely a copying accident, then the error could have been due to the copyist’s familiarity with the more frequent word orders seen in Old Latin syntax, and in Plautus specifically. This could have resulted in accidental transposition, placing the inflected auxiliary where we find it in the transmitted text, right before the right dislocated adjective, which is the most likely position in the clause to expect it given the order of the other words. 19 5. Conclusion In the preceding sections we have shown that, given a right-headed analysis of Old Latin’s TP and VP syntax, the word order seen in Camerarius’s emendation of Amphitruo 260 is unexpected, while that of the transmitted text is quite 312 Michael Weiss & Ryan Windhearn <?page no="313"?> common and expected. We further demonstrated that the most likely original word order is that of Pareus’ emendation, placing the inflected auxiliary in C above the subject. This paper illustrates how modern theoretical-syntactic analyses may be productively applied to textual criticism, acting as an additional tool to decide between potential restorations of a text. Bibliographical references Bothe, F.H. (1809) (ed.) M. Atti Plauti comoediarum tomus primus, Berlin, Kuhn. Camerarius, J. (1535) (ed.) M. Accii Plauti Sarsinatis comici festiuissimi comediæ xx, Basel, Hervagius. Camerarius, J. (1545) (ed.) M. Accii Plauti comoediae V. magna cum cura emendatae a Ioachimo Camerario Pab(ergensi) modo editae cum annotationibus eiusdem, Leipzig, Bapst d. Ä. Chomsky, N. (1965) Aspects of the theory of syntax, Cambridge (MA), MIT Press. Christenson, D. (2000) (ed.) Plautus Amphitryo, New York, Cambridge University Press. Danckaert, L. (2017) The development of Latin clause structure: a study of the extended verb phrase, New York, Oxford University Press. Devine, A.M. & L.D. Stephens (2006) Latin word order: structured meaning and informa‐ tion, New York, Oxford University Press. Dousa, J. (1594) (ed.) M. Accii Plauti comici fabulae superstites XX, Leiden, apud Fran‐ ciscum Raphelengium. Endo, Y. (1996) “Right dislocation”, MIT Working Papers in Linguistics 29: Proceedings of Formal Approaches to Japanese Linguistics 2: 1-20. Ernout, A. (1932) (ed.) Plautus, Paris, Les belles lettres. Ferber, M. (2019) “Unusual word order and other syntactic quirks in poetry”, in-Poetry and language: the linguistics of verse, Cambridge, Cambridge University Press, 104-141. doi: 10.1017/ 9781108554152.005. Fleckeisen, A. (1850)-(ed.) T. Macci Plauti comoediae, Leipzig, Teubner. Fortson IV, B.W. (2008) Language and rhythm in Plautus. Synchronic and diachronic studies, Berlin, de Gruyter. Goetz, G. & F. Schoell (1898) (eds.) T. Macci Plauti comoediae, Fasciculus I, Leipzig, Teubner. Gratwick, A. (1999) (ed.) Terence: The Brothers, 2 nd ed, Warminster, Aris & Phillips. Gruterius, J. (1621) (ed.) Comoediae, ex recognitione Jani Gruteri, qui bonâ fide contulit cum MSS. Palatinis, Wittenberg, apud Zachariam Schurerum. Gundel, J.K. & Th. Fretheim (2004) “Topic and focus”, in Horn, L.R. and G. Ward (eds.) The handbook of pragmatics, Malden (MA), Blackwell, 175-196. Havet, L. (1895) (ed.) Plauti Amphitryo, Paris, Bouillon. Syntactic Theory and Textual Criticism in Plautus 313 <?page no="314"?> Hooper, W.D. & H.B. Ash (1934) (eds. and trans.) Marcus Porcius Cato, On Agriculture; Marcus Terentius Varro, On Agriculture, Cambridge (MA), Harvard University Press. Kayne, R.S. (1994) The antisymmetry of syntax, Cambridge (MA), MIT Press. Kuno, S. (1978) Danwa no Bunpo, Tokyo, Taishukan. Lambinus, D. (1576) (ed.) Plautus ex fide atque auctoritate complurium librorum manu‐ scriptorum opera Dionys. Lambini Monstroliensis emendatus, Paris, Apud Ioannem Macæum. Lambrecht, K. (2001) “Dislocation”, in Haspelmath, M.; E. König; W. Oesterreicher; W. Raible (eds.) Language typology and language universals: an international handbook, vol. 2, Berlin, de Gruyter, 1050-1078. Ledgeway, A. (2012) From Latin to Romance: morphosyntactic typology and change, New York, Oxford University Press. Ledgeway, A. (2017) “Late Latin verb second: the sentential word order of the Itinerarium Egeriae”, Catalan Journal of Linguistics 16: 163-216. Leo, F. (1885) (ed.) T. Macci Plauti comoediae, vol. 1, Berlin, Weidmann. Lindsay, W.M. (1896)-Introduction to Latin textual emendation---An introduction to Latin textual emendation, based on the text of Plautus, Oxford, Macmillan. Lindsay, W.M. (1904) (ed.) T. Macci Plauti comoediae, tomus I, Oxford, Clarendon. Marx, F. (1922) Molossische und Bakcheische Wortformen in der Verskunst der Griechen und Römer, Leipzig, Teubner. de Melo, W.D.C. (2011) (ed. and trans.) Plautus, Cambridge (MA), Harvard University Press. Naudet, J. (1830) (ed.) M. Accii. Plauti comœdiæ, cum selectis uariorum notis et nouis commentariis, vol. 1, Paris, Nicolaus Eligius Lemaire. Pareus, J. Ph. (1610) (ed.) M. Accii Plauti Comoediae XX superstites, Frankfurt, Impensis Jenæ Rhodii. Pareus, J. Ph. (1641) (ed.)-[Plauti] Comoediae XX superstites, et deperditarum frag‐ menta,-Frankfurt, in officina Fischeri. Pylades, G.F. (1506) (ed.) Lector-optime-scito-has-comoedias-uiginti plautinas, Brescia, Britannicus. Pylades, G.F. (1522) (ed.) Amphitrio Plauti a Pylade castigata, Zwolle, Corver. Questa, C. (2007) La metrica di Plauto e di Terenzio, 2 nd ed., Urbino, QuattroVenti. Ram-Prasad, K. (2022) The syntax of relative clauses and related phenomena in Proto-Indo-European, Ph.D. diss., University of Cambridge. Ritschl, F.; G.L.G. Goetz; F. Schoell (1882) (eds.) T. Macci Plauti Comoediae recensuit instrumento critico et prolegomenis auxit Fridericus Ritschelius sociis operae adsumptis Gustavo Loewe, Georgio Goetz Friderico Schoell. Tomi II Fasiculus II Amphitruo, Leipzig, Teubner. 314 Michael Weiss & Ryan Windhearn <?page no="315"?> Rizzi, L. (1997) “The fine structure of the left periphery”, in Haegeman, L. (ed.) Elements of grammar, Dordrecht, Springer 281-337. Samek-Lodovici, V. (2015)-“The interaction of focus,-givenness, and prosody.-A study of Italian clause structure”,-Oxford, Oxford University-Press. Ussing, I.L. (1875) (ed.) T. Maccii Plauti Comoediae, vol. I, Copenhagen, Sumptibus Librariae Gyldendalianæ (F. Hegel). Whitman, J. (2000) “Right dislocation in English and Japanese”, in Takami, K.; A. Kamio; J. Whitman (eds.) Syntactic and functional explorations in honour of Susumu Kuno, Tokyo, Kurosio Pub, 445-470. Windhearn, R. (2020) Evidence from innovation: reconstructing disharmonic headedness for Proto-Indo-European, Ph.D. diss., Cornell University. Windhearn, R. (2021) “Disharmonic headedness in Homeric Greek and Tocharian”, Journal of Historical Syntax 5: 1-29. Syntactic Theory and Textual Criticism in Plautus 315 <?page no="317"?> 1 The paper evolved from a project of national relevance newly funded by the Italian Ministry of Education, University and Research with the title ‘Parts of speech meet rhetorics: searching for syntax in the continuity between the Middle Ages and the Modern Age’ (Prot. 20172F2FEZ). The creation of linguistic metalanguage in Antiquity and Middle Ages as result of translational processes Paola Cotticelli-Kurras (University of Verona) Abstract: The dimension of multilingualism and multiculturalism was a characteristic in the Greco-Roman Antiquity which let many traces in different cultural domains and their specific language. Latin and Greek played the main role in this history even after Theodosius’ division of the Roman Empire in the Latin-speaking Western part and the Greek-speaking Eastern part. The interplay between these two languages was only one aspect of the complex picture of relations in the different regions or provinces of the Empire where other languages were spoken or could arise. The Greco-Roman cultural exchange brought to the development of translation techniques, the dimension of translation was itself a peculiarity of a cultural heritage which continued through time. The direction of the language influence in particular is visible especially in the creation of metalanguages and specialized languages. One thinks of the direct role of Greek in Egypt, of Latin on Gaul, of Greek in Syriac dialects, but also on the direct and indirect influence of both Latin and Greek on the Germanic and Romance languages through the grammatical tradition. This paper aims at providing some insights into the metalinguistic croft of grammars and grammarians as special case of language contact in the Late Antiquity and the early Middle Ages. 1 Keywords: multilingualism in Late Antiquity, Greek-Latin cultural con‐ tact, creation of Latin linguistic metalanguage, grammatical tradition <?page no="318"?> in Late Antiquity and Middle Ages, Latin grammaticography, origin of syntactic metalanguage 1. Focusing on a multilingual society in the late Antiquity The Roman Empire, at its height, encompassed a vast territory stretching from modern-day Britain to Egypt and the Middle East. Latin was the administrative and literary language of the empire. However, many other languages coexisted, especially in the provinces, and generated a multilingual society. Since the Classical Latin period, the Roman society was at least a bilingual one, because Greek was very widespread in large parts of the Empire. Local languages continued to thrive alongside Latin. For example, Greek was widely spoken in the Eastern part of the Empire, including regions like Egypt and Anatolia. In some areas, native languages persisted, and Latin remained a secondary language for everyday use. Bilingualism was common in urban centers, where people could speak both their local language and Latin. Code-switching, the practice of switching between languages within a conversation, was also prevalent in these multilingual communities. Latin was the dominant language of learning, religion, and administration in medieval Europe. It served as the lingua franca among scholars, clerics, and educated elites, facilitating communication and the exchange of knowledge across different regions. Latin was also the language of the Church and religious texts, contributing to its widespread use and influence. Language and linguistics in medieval Europe were shaped by a combination of historical, social, and cultural factors. During this period, which roughly spans from the 5 th to the 15 th century, several important developments took place that influenced the languages spoken, written, and studied across the regions. The use of Latin was not uniform throughout the continent. In addition to Classical Latin, medieval Europe saw the development of various Latin varieties, including Chancery Latin used for administrative purposes and L2 Latin, which was Latin influenced by the native languages of non-native speakers of Latin. In the historical period from the 3 rd to the 7 th century AD, marked by the transition from the classical world to the early Middle Ages, the Roman Empire was undergoing significant changes, and various regions experienced diverse linguistic and cultural interactions due to conquests, migrations, and trade. Furthermore, the Roman late Antiquity saw significant cultural exchanges between different regions. Intellectual and philosophical ideas from Greece, Egypt, and the Near East influenced the Roman world, leading to a rich synthesis of diverse cultures. While Latin was the language of the educated elite, the 318 Paola Cotticelli-Kurras <?page no="319"?> common people spoke various vernacular languages specific to their regions (see Haßler 2018). These vernaculars evolved from the Latin spoken by the local populations, mixed with influences from invading tribes and neighboring languages. The most significant vernaculars during this time were Old English, Old French, Old High German, Old Spanish, and Old Italian. The rise of Christianity further contributed to the linguistic diversity in late Antiquity. The Bible and other Christian texts were translated into various local languages, allowing for broader dissemination of religious ideas. Later on, the movement of different tribes and peoples, such as the German ones, Visigoths, Vandals, and Ostrogoths, also played a role in shaping multilingual societies as they interacted with the local populations and often adopted or imposed new languages. As the Western Roman Empire faced political and economic challenges, Latin’s importance as a unifying language started to decline. Medieval Europe was characterized by extensive language contact due to trade, conquests, and migrations. As a result, languages influenced one another, leading to the adoption of loanwords and linguistic features from neighboring languages. Overall, late Antiquity was a period of dynamic cultural and linguistic interactions, characterized by the coexistence and mutual influence of various languages and traditions, paving the way for the linguistic diversity that defines many societies to this day. Language played a crucial role in defining regional identities and fostering a sense of community. Vernacular languages were instrumental in the develop‐ ment of national identities in later centuries. Overall, medieval Europe witnessed a dynamic linguistic landscape, with Latin as the dominant language for scholarly and religious activities and a rich variety of vernacular languages shaping everyday communication and cultural expressions. The interaction between Latin and vernaculars, as well as the study of linguistics and grammar, laid the groundwork for the linguistic diversity and complexity that characterizes Europe to this day. In this frame, we consider the interplay between Greek and Latin in the different Latin grammars and their later tradition. 1.1 The grammars in the late Antiquity Late Antiquity saw the development of Latin grammars that systematized the rules of the Latin language. Prominent grammarians, such as Donatus and Priscian, authored Latin grammatical treatises that covered topics like phonetics, morphology, syntax, and rhetoric. Ancient and Medieval linguistic metalanguage as translational process 319 <?page no="320"?> 2 We refer for an edition and detailed discussion of this text to Holtz, 1981, and Amsler 1989, among others. The period of Late Antiquity covers roughly the 3 rd to 7 th centuries AD and is characterized by the transition from the classical world to the early Middle Ages. In this period, grammars played a significant role in shaping the study and understanding of languages, particularly Latin and Greek. During this time, various grammarians and scholars contributed to the development of grammatical theories and the preservation of linguistic knowledge. Late antique grammarians focused on preserving and codifying the knowledge of classical languages, primarily Latin and Greek. Their efforts were crucial in maintaining and transmitting the literary and scholarly heritage of the ancient world to later generations. The study of Greek grammar continued to be important in this period. Scholars like Apollonius Dyscolus and Herodian worked on Greek grammar and syntax, refining the understanding of the language’s structure and usage. In fact, late antique grammarians built upon the linguistic traditions of earlier periods, including those traditions established by ancient Greek grammarians like Dionysius Thrax and Latin grammarians like Varro. As Hayden 2017 remarks, “among the most influential grammatical texts in the medieval European world were the works of Aelius Donatus (ca. 350 AD). His Ars minor, a short treatise on the parts of speech written in question-and-answer form, remained the standard Latin primer throughout the Middle Ages. Donatus’ more advanced Ars maior was also a popular and widely known text, which was divided into three parts: the first book dealt with sound (uox), letter/ speech-sound (littera), syllables, metrical feet, accents, and punctuation; the second, with the parts of speech; and the third, with stylistic matters, such as barbarisms, solecisms, metrical faults, metaplasms, schemes, and tropes. Many grammarians in Late Antiquity wrote commentaries on classical texts, particularly on the works of famous authors like Virgil, Cicero, and Homer. These commentaries provided linguistic and grammatical analyses to aid in the interpretation and understanding of these influential works. The Ars maior provided the standard structure for any treatment of grammar until it was replaced by the fourfold division into orthographia, prosodia, etymologia, and diasintastica (‘syntax’) current in the later Middle Ages (Copeland - Sluiter, 2009: 82).” 2 In addition to this, rhetoric, the art of persuasive speaking and writing, was a vital component of late antique grammars. The study of rhetoric was closely intertwined with grammatical analysis and language usage, as it aimed to cultivate eloquence and effective communication. 320 Paola Cotticelli-Kurras <?page no="321"?> 3 See Hayden (2017: 2f.). 4 See Dionisotti (1988: 204). Late antique grammarians were concerned with language instruction and the teaching of Latin and Greek. Their grammatical treatises served as essential resources for educators and students alike. For this tradition, it is noteworthy that different types of grammars circulated. Vivien Law (1994: 88-92; 2003: 65 and 83) in her work on Medieval Latin grammars classified these sources into two principal but overlapping categories. 3 The so-called Schulgrammatik genre encompasses works that systematically introduce fundamental concepts such as parts of speech and their attributes. These texts are characterized by a strict hierarchical and systematic structure, often emphasizing semantic categories, assuming familiarity with Latin forms among native speakers. On the other hand, the Regulae genre comprises grammars that primarily consist of extensive lists of examples showcasing morphological phenomena. These grammars were likely intended as references rather than comprehensive study materials, targeting advanced learners who already possessed a grasp of basic language forms. Examples of this genre within a multilingual context include the detailed accounts of nominal declension and verbal conjugation by Byzantine scholars Phocas and Eutyches. Some grammarians merged elements of both the Schulgrammatik and Regulae approaches, notable among them being Consentius (5 th century), Diomedes (4 th century? ), and Priscian (4 th -5 th century). The assumption was that Donatus presupposed knowledge of the Latin language in which he focused on analyzing the language. If we look at the Western and Eastern traditions in Latin grammar, the real difference in their structure is the grammarian’s point of view: Donatus wrote a grammar for people who knew Latin, while Priscian taught Latin as a second language. 4 Around the 5 th century some changes are noted: it seems that some grammars written by grammarians such as Charisius and Diomedes, Cledonius, Priscian, and Eutyches living in the Eastern part of the Empire - therefore written for Greek speakers who learned Latin as L2 - were imported into the West where an increase of non-Latin speakers was noted, belonging to the Germanic populations, in addition to the diversified competence of those who only spoke Latin and also read it. From here we understand the need to import these “Eastern” grammars into the West, bringing elements of Greek with them. Donatus’ commentators and the different manuscripts and area of his tradi‐ tion help understanding exactly such a situation, testifying the relationships between the Eastern and Western part of the Empire. According to a study by Dionisotti (1984, 204f.), one of the manuscripts of the Donatian tradition is a Ancient and Medieval linguistic metalanguage as translational process 321 <?page no="322"?> 5 See Russell (2012: 196). copy written in Northern Italy in the 5 th century though the manuscript came from East, according to Dionisotti who quotes the analysis of the manuscript by Bischoff. Besides the difference between Donatus’ and Priscian’s grammars and their later traditions, the later copies of these grammars testify the changes not only of the language but also of the society and its speakers. In addition, there are also two ninth-century treatises on word order that show that Carolingian interest in sentence interpretation went beyond the relatively elementary analysis of syntax in Priscian’s Institutiones to include issues such as compound sentences and embedded constructions (Covington 1984, Luhtala, 2013: 347- 348), Eriugena, (Law 1987, 1993, 2003, 1995, Luhtala, 1996), Sedulius Scottus, (Löfstedt, 1977), and Alcuin (Rizza 2022). Donatus’ Artes remained the basic introduction to grammar in the late Middle Ages, and both these textbooks and Priscian’s Institutiones continued to be the subject of study and commentary. 1.2 Specialization of grammars in the target language Considering the evidence of multilingualism in northwestern Europe on the fringes of the former Roman Empire and beyond in the three or four centuries after the fall of Roman rule, what makes this area particularly interesting is that it includes both areas with a long tradition of spoken Latin, such as Spain and Gaul, and areas where there was no continuous (or very little continuous) tradition of spoken Latin. 5 Latin has been studied by non-native speakers as a high-ranking language. Adding to the mix of speakers of Germanic and Celtic languages from the borders of the former Roman Empire and beyond, many of whom had learned Latin as a second language saw no compelling reason why Latin should be considered a superior language. Since many different versions of Latin would gradually become problematic, the need for a standardized form of Latin became more pressing. The establishment of a classically spoken Latin and the resulting and growing disconnection between a more formally spoken Latin and its Roman-spoken descendants appear to have had practical consequences. The example of the Carolingian Renaissance is a good one for such a process. The example of the Irish traditions is similar: the Auraicept na n-èces is a text concerned with the status of the Irish language, is mainly concerned with 322 Paola Cotticelli-Kurras <?page no="323"?> 6 Given that languages such as Irish, Welsh and Old Norse had vernacular literature from a much earlier phase of the Middle Ages, it is therefore not surprising to find contemporary textual sources examining the structure and characteristics of these languages. Such sources often involve a contrastive analysis between Latin and vernacular through vernacular glosses on Latin grammars and other texts or through collections of glossaries. In some cases, however, we find grammars of both Latin and vernacular written using the vernacular. While mediaeval grammars of the Middle Ages often derive to a significant extent from the format and content of Latin textbooks, their authors often also show considerable insight into the specifics of the particular language with which they deal. The following section examines some of the most important surviving sources for the study of European vernaculars in the Middle Ages, before the advent of printing in the first half of the 15 th century. To this tradition see Law (1987; 1994; 1997). 7 One of the best overviews and well-documented work on the heritage of the Greco-Latin grammaticography in the Middle Ages is the research by Swiggers 1995. poetry, although it purports to be in some respects a grammar thereby offering Irish what Priscian and Donatus intended for Latin. 6 2. The role of grammaticography in a multilingual society Grammaticography can be considered as a point of contact between translation, teaching of a second language, creation of terminology and grammatical re‐ flection in the comparison of different approaches and cultural perspectives. Translation techniques were established in Antiquity, often, but not exclusively, from Greek to Latin, which accompanied the history of grammars at least from Donatus, and later from Priscian to the modern world. In Late Antiquity, scholars from the Eastern and Western Roman Empire engaged in scholarly exchange, which facilitated the cross-pollination of grammatical ideas and techniques between Greek and Latin grammarians. 7 The role of Greek grammar in Latin grammar was significant, particularly during Late Antiquity and the early medieval period. Greek and Latin were the two primary classical languages of the ancient Mediterranean world, and their close relationship influenced the study and development of both grammars. Latin grammarians were heavily influenced by Greek grammatical concepts and terminology. Many grammatical terms used in Latin grammars were borrowed directly from Greek, as Greek grammatical analysis was well-established and sophisticated. The research has focused particularly on the role of Greek terms and their Latin equivalents in the medieval grammars establishing the history of gram‐ matical thoughts and ideas. Among them, particularly the field of the syntactic terminology in the time after Priscian built the bridge between the Greek Ancient and Medieval linguistic metalanguage as translational process 323 <?page no="324"?> 8 Here I refer to the works by Householder (1981); Lallot (1997); Brandenburg (2005); Lambert - Bonnet (eds., 2021). terminology by Apollonius Dyscolus 8 and the tradition of the Ars grammatica during the Middle Ages. Late antique grammars laid the foundation for later developments in lin‐ guistic studies, particularly during the medieval period. The grammatical theories and linguistic analyses of this era provided essential insights into the structure and usage of languages, contributing to the continued development of Latin and Greek as well as the emerging vernacular languages of the medieval world (see the early medieval Irish and the German tradition). With the spread of Christianity in Late Antiquity, some grammarians inte‐ grated Christian themes into their works, reflecting the growing influence of the Church on intellectual pursuits. Medieval Europe was marked by the rise of Scholasticism, a philosophical and theological movement that heavily relied on language and logic. In this perspective, grammar assumed an important role as the tool to clarify and understand the Holy Scriptures, not only the classical texts. Scholars engaged in grammatical and linguistic studies to ana‐ lyze and interpret texts, leading to the development of medieval grammatical theories. Finally, the preservation and dissemination of grammatical knowledge heavily relied on manuscript culture. Many grammatical works were copied and transmitted through handwritten manuscripts, ensuring their survival and transmission to later centuries. Our problem is that many of those manuscripts are not edited so that only a small part of the linguistic and grammatical handbooks is directly accessible. Besides some well-known examples, such as different translations and equivalents of Greek λόγος as oratio, sententia since the beginning of the Latin grammaticography, among others, the creation of terms connected with the growing theories on syntax and definitions of syntactic concepts will be highlighted. Below, we will discuss some terms, e.g. congruitas used in grammatical contexts as a calque of Greek σύμβαμα and meaning ‘a complete predicate, according to Priscian, Gramm. 3. 211. 20 ff.), that is referred to an ‘intransitive construction’ (i.e. Plato philosophatur, see also Baratin 1989: 391), or the pair Greek ἀξίωμα and Latin constructio (among other terms). In the modern history of grammaticography and lexicography, the termi‐ nological interest culminated in the lexicon by Schad 2007 who collects the linguistic terms until Priscian, so that the terms in Late Latin, from the grammars of the Carolingian period, the Early Medieval treatises, commentaries and (Irish, English) glosses have not been collected yet and only partially analyzed. In 324 Paola Cotticelli-Kurras <?page no="325"?> 9 We can draw on other research, e.g. on the origin of French grammar (Merrilees 1988), which deals both with the beginnings of the French metalanguage, attested in manuscripts from at least the twelfth century, as a consequence of the influence of Donate’s Ars minor on the one hand, and the transposition of the paradigms of Latin grammar in the 14 th century, on the other side. It is important for us to emphasize that (not only) French grammarians of the Renaissance partly owe their terminology and their concepts to their medieval predecessors, since they do not obtain all their material from classical grammarians (cf. Chevalier, pp. 133-170). Merrilees 1988: 399 demands that the history of (French) grammar denying the medieval contribution mentioned above must be corrected (refers to Julien, 1988, Städler 1988). The author added (1988: 406 with fn. 15) an interesting list of French technical terms from a medieval glossary which footed in the medieval grammarians. I list the examples of the syntactic meta‐ language since the choice is partially coincident with the terms I will comment in the following: “composición, congruité, conjoncïon, conjonctif, conjugación, consignifier, consonnant, construcción, constructible (adj. et subst.), construire, copulatif, declinable, dérivatif, diction, gouverner, acenter, reciproque, régularité, relación, relatif (adj. et subst.), transitif, transition”. Here we refer also for this tradition to the Traité de syntaxe de Metz 647 (éd. Mok 1975), and also to Heinimann (1963: 26 f.). 10 Here I refer to Baratin (1989), Swiggers (1995); Swiggers-Wouters (1999, 2015a, 2015b); Luhtala (2021), and the contributions in Lambert-Bonnet (eds.) (2021). fact, one of the first works on the creation of a grammatical terminology is that by Heinimann (1963), who focused on the origin of this special field of terminology in the Middle Ages. No other works treating the medieval grammatical terminology have been published until now. 9 3. Focusing on the syntactic field: new concepts or translated labels? Greek grammar served as a model for linguistic analysis in Latin. Latin gram‐ marians turned to Greek grammatical treatises and theories as examples of how to systematically study the structure and syntax of a language. 10 Besides that, the study of Greek literature and rhetoric inspired Latin writers and grammarians. Latin authors often translated Greek works, and the imitation of Greek rhetorical styles and techniques influenced the development of Latin literature and rhetorical theories. As already mentioned, in some regions of the Roman Empire, bilingualism in both Greek and Latin was common. This led to code-switching, where speakers would switch between Greek and Latin within the same discourse. The study of Greek grammar helped Latin grammarians understand and analyze these language interactions. As a result of cultural interactions, Latin borrowed numerous words from Greek, especially in the creation of some technical languages. The grammatical one is a good example. Understanding Greek grammar assisted Latin grammarians in integrating Greek Ancient and Medieval linguistic metalanguage as translational process 325 <?page no="326"?> 11 In order to give only one example of the interchange and interplay of the different fields of knowledge and languages in the Middle Ages in the formation of the technical languages in the Trivium, here some illustrative terms: fundamental musical concepts as modus, tropus, harmonia, systema, diastema, and phthongus are recurring in Latin texts though they reflect the inconsistency found in Greek treatises. Furthermore, the terminological confusion in Latin reflects the interdisciplinary borrowing from different fields. In fact, terms as comma, colon, periodus, arsis, thesis, levatio, and positio are among those borrowed from the fields of grammar, rhetoric, and metrics. The grammatical term comma sometimes refers in music to a self-standing portion of a melodic line; it is the musical equivalent of a section or division of a periodus and corresponds loosely to a clause (Latin incisum) within a sentence. But comma can also identify a very small (spatial) interval, a second musical meaning derived from mathematics and therefore unconnected with the grammatical or rhetorical comma. Since the chant as the focus of the musical treatises is closely related to the sacred text it represents, and because its structure is derived from, and subordinate to the textual structure, the language of grammar and rhetoric is used to define musical phrase structures. In addition to such classical terms as colon and comma, one finds particula, membrum, incisum, distinctio, and periodus; arsis and thesis (which aid in the definition of phrase length as well as metrical forms) often appear in their Latin forms, (e)levatio and positio, levandum and deponendum. From the language of rhetoric there are borrowings such as protrahere, contrahere, productus, correptus, and protensio to describe the manner in which a musical line is projected, just as these terms had once been applied to public speaking. See Philips in Mantello---Rigg (1999: 297 ff.). loanwords into Latin sentences while maintaining proper grammatical struc‐ tures. Latin grammarians studied and commented on Greek literature, analyzing its style, rhetoric, and grammar. These studies helped Latin writers adopt various rhetorical devices and linguistic features into their own compositions. Overall, Greek grammar played a crucial role in shaping the development of Latin grammar during late antiquity and beyond. The interaction between the two classical languages contributed to the sophistication and refinement of grammatical analysis in both traditions. This linguistic cross-fertilization paved the way for the continued evolution of Latin as a language of literature, scholarship, and culture in the medieval and early modern periods. 11 The grammatical structure based on the eight parts of speech is taken from the Greek tradition and spread through the Western grammaticography over time lasting until now (see Vineis 1994). In the following I present some cases of the linguistic interaction of the continuation between the Greek and the Latin grammaticography, but also of the sometimes necessary metalinguistic elaboration for the creation of proper terms and denomination of categories or concepts in Latin. I refer to the material collected by Schad (2007) in her lexicon as starting point of the analysis and give then, when relevant, the continuation of the grammatical path until the Medieval treatises. 326 Paola Cotticelli-Kurras <?page no="327"?> 12 We refer to Milani (2004: 9-32). 13 See the other denominations such as comparativus, Latinus, septimus, sextus. 14 For the attestations of the denominations see Schad, s.v.: “Varro seems not to have known this term since he uses sextus and Latinus for the ablative case: Diom. 1.302.4 -um Graeci non habent, hunc tamen Varro sextum (e.g. Var. 10.62), interdum Latinum appellat, quia Latinae linguae proprius est, cuius vis apud Graecos per genetivum explicabitur (Don. 624.13; Pomp. 5.171.12; Cons. 5.351.9). Char. 170.13 Plinius ait inter cetera etiam istud G. Caesarem dedisse praeceptum, quod neutra nomina ‘ar’ nominativo clausa per ‘i’ dativum -umque singulares ostendant (Char. 156.13). Lersch (1838-41: ii, 231) attributes these words to Caesar, Job (1893: 68), Nettleship (1889: 6) and Steinthal (1890-91: ii, 266) rather to Pliny. The earliest direct attestation is in Quint. 1.4.26. Fest. 282.4; Paul. Fest. 140.4; Vel. 7.77.7 -o casu; Scaur. 7.25.2 -o plurali; Gell. 1.16.13 casum … -um; Sacerd. 6. 427.5; Prob. 4.75.1 -us casus, Serv. 4.408.28 -us singularis. Prisc. 2.186.2 -us etiam comparativus, ut ‘aufero ab Hectore’ et ‘fortior Hectore’ (Dosith. 7.401.5 comparativum gradum casum -um trahere). “Char. 15.8 casus sunt sex, nominativus genetivus dativus incusativus, qui et accusa‐ tivus, vocativus -us (Diom. 1.301.35; Dosith. 7.392.9; Aud. 7.341.19; Vict. 6.189.20; Prob. 4.74.18; Don. 586.16, 624.12; ps. Asp. 5.550.12).” 15 Its ancient etymology is given in Schad, s.v.: “expl. in Don. 4.535.1 -us (dictus), quod per eum auferre nos ab aliquo aliquid signifi‐ cemus, ut ‘ab hoc magistro’. 544.14 -us ab auferendo dictus, aufer ab eo. Pomp. 5.183.9 -us dictus est, quod per ipsum aliquid auferamus, ‘aufero a Cicerone’. 3.1 ablativus; ablativus absolutus 3.1.1. ablativus The term ablativus (ablative) to denote the grammatical case in Latin has a long history that spans from ancient Roman grammarians to the medieval period. The denomination of the case ablativus and its recognition in grammatical analysis can be traced through the following key stages. 12 Ancient Roman grammarians recognized and classified the various cases in Latin, including the ablative case, which had distinct functions in expressing different relationships such as means, instrument, and manner. However, the early Latin grammarians did not always use the term ablativus  13 to refer to this case as also Schad s.v. remarks. 14 Latin grammarians did not have a model for the denomination of such case in Greek grammars since the case was lacking in Greek. Aelius Donatus (4 th century AD) in his work Ars Minor - one of the earliest comprehensive treatments of Latin grammar - used first the term ablativus to refer to the ablative case explicitly. 15 His work, which became widely used as a standard textbook for teaching Latin grammar, contributed to the widespread recognition and adoption of the term ablative in grammatical discussions. After Donatus, Ancient and Medieval linguistic metalanguage as translational process 327 <?page no="328"?> 16 See Thurot, Notices, (1868: 165 f.), see further Jeep, (1893: 152 ff.); the occurrences in Mittellateinisches Wörterbuch, s.v. adiectivus, and finally, Baebler, (1885: 62 f.). later Latin grammarians, including Servius (4 th -5 th century AD) and Priscian (5 th - 6 th century AD), continued to use the term ablativus to refer to the ablative case in their grammatical treatises. Priscian’s influential work Institutiones Grammaticae provided a comprehensive analysis of Latin grammar, including detailed explanations and examples of the ablative case. In Medieval Latin Grammars, which were heavily influenced by the works of earlier grammarians like Donatus and Priscian, the term ablativus continued to be used. Grammarians during the medieval period, such as Isidore of Seville (6 th -7 th century AD), continued to refer to the case as ablativus. Schad (2007, s.v.) quotes the Greek correspondences with ἀφαιρετικός, ἀπενεκτικός, ἀποκομιστικός which is the product of a calque from Latin: “Gk. has no ablative, and no label for this case is found in the Gk. grammarians. The three equivalents to ablativus (casus) in the Gk. text of Dosith. are probably therefore later coinages created to match the Latin term. Dositheus 7.392.11 ἀφαιρετικὴ, ἢ καὶ ἀπενεκτική. 401.5 τὴν πτῶσιν τὴν ἀποκομιστικήν. Ἀφαιρετική, from ἀφαιρέω, and ἀπενεκτική, from ἀποφέρω, represent loan translations of ablativus, from auferre (Bécares Botas 1985: s.v.)”. Overall, the term ablativus to denote the ablative case in Latin became firmly established and widely recognized through the works of ancient and medieval grammarians. The term ablativus has persisted since the Middle Ages in the study of Latin grammar to this day and continues to be used in modern Latin grammatical discussions. 3.1.2 absolutus For the special syntactic construction of the ablativus absolutus the denomi‐ nation is first in the Middle Ages established, since Priscian defined the construction according to its meaning in consequentiae significatione (“which mean a consequence of facts”, see Ars, GL 3, 221.25-222.3), while the adjective absolutus has been used by Priscian to denote some intransitive verbs (s. also Schad 2007: 362). Especially for the use of the adjective absolutus, together with the case ablative, and corresponding to Gr. ἀπόλυτος to designate a non-governed, independent case, we refer to Petrus Helias in the early Middle Ages, who as first defines the term and name of the construction of the ablativus absolutus. 16 328 Paola Cotticelli-Kurras <?page no="329"?> 17 See Rosier-Catach (1983), Marmo (1994), and the recent analyses by Cotticelli-Kurras (i.p. and 2023) with all the references to the topic. 18 Further references are: Gasti (ed.) (2003); Swiggers - Wouters (1999: 3-13). 3.2. congruitas and its related field 17 The first attestations of congruitas in Priscian show the mechanical transposition or creation of the word as calque from Greek σύμβαμα or ἀξιώματα which designated so-called intransitive constructions, and this according to the Stoic teaching and the Apollonian use. Another term for congruitas in this meaning is Latin dignitas. congruitas: σύμβαμα ‘intransitive construction’ (from Schad, 2007, s.v.) “Prisc. 3.211.20 ff. constructiones, quae per nominativum absolvuntur, Stoici ἀξιώματα vel συμβάματα, id est dignitates vel -es, vocabant - ut……, illas vero, …in quibus necesse est cum nominativo etiam obliquum aliquem casum proferri, παρασυμβάματα dicebant, hoc est minus quam -es, ut ‘Cicero servat patriam’ (opp. ἀσύμβαμα, id est incongruitatem … ut ‘placet mihi venire ad te’). Gk.: σύμβαμα, AD synt. 43.16, 430.3. Prisc. gives a rather garbled version of the Stoic doctrine reported by Apollonius but uses the same examples (AD synt. 43.16 περιπατεῖ, γράφει). Opp. incongruitas.” 3.2.1 incongruitas We consider in this context its opposite incongruitas. “Compare with Gr. ἀσύμβαμα: ‘construction with two oblique forms’ (from Schad, 2007, s.v, with reference to Baratin 1989: 391) Prisc. 3.211.26 quando … ex duobus obliquis constructio fit, ἀσύμβαμα, id est -em, dicebant, ut ‘placet mihi venire ad te’. Gk.: found only in Prisc. Opp. congruitas.” 3.2.2 dignitas The further Latin term involved is dignitas. Here I quote first from Schad, 2007, s.v. (who referes to Baratin 1989: 391): 18 “dignitas ‘intransitive construction’ “Prisc. 3.211.19 has quidem constructiones, quae per nominativum absol‐ vuntur, stoici ἀξιώματα vel συμβάματα, id est -es vel congruitates, vocabant - ut ‘ego Priscianus scribo, Apollonius ambulat, Plato philosophatur’. Gk.: ἀξιώμα (sic) Arist. metaph. 997 a 7, 1005 b 33, etc. The Stoics use ἀξιώμα (sic) for propositions of all types, both simple and complex (see DL 7.68ff. for a list of categories). Prisc.’s examples are of simple propositions of the affirmative type (κατηγορικόν): DL (7.70) gives the example Δίων περιπατεῖ. Ancient and Medieval linguistic metalanguage as translational process 329 <?page no="330"?> 19 To this explanatory gloss and its rendering see Swiggers 2009: “I would propose ‘subject-predicate construction’ (corresponding to Greek ἀξίωμα)”. 20 The definition and the related terms transitivus and intransitivus are very intriguing and they deserve a particular and detailed treatment on which we cannot go into detail for reasons of space. We will therefore limit ourselves to giving some essential references: The Greek term μετάβασις refers to a phenomenon in Latin grammar where a verb’s meaning changes when it takes on a different case. Specifically, this concept arises when a verb that usually governs one case takes on another case, and this change affects the verb’s meaning. This notion of μετάβασις in Latin grammar is based on a similar concept in ancient Greek grammar, where certain verbs exhibited changes in meaning when paired with different cases. In Latin the concept of transitio has been use for its rendering. The Latin verb fruor, which typically governs the ablative case means in its usual sense “to enjoy” or “to make use of.” However, when it takes on the accusative case, the verb’s meaning changes to “to consume” or “to use up.” Greek μετάβασις in Latin grammar is concerned with the morpho-syntactic and semantic shifts that occur when certain Varro translates ἀξιώμα (sic) by proloquium (Gell. 16.8.2, 6, 8), Cicero by pronun‐ tiatum (Gell. 16.8.8). Prisc. may have chosen dignitas to translate ἀξιώμα (sic) in a grammatical context because the two terms share the nongrammatical senses ‘worth; rank’. See congruitas. Non-gram.: ‘worth’ (Pl.+); ‘rank’ (Cic.+).” The explanation given by Schad is that dignitas as Latin correspondence to Greek ἀξίωμα is driven by the non-technical meaning of the word dignitas as ‘worth’ or ‘rank’. 19 Nevertheless, Greek ἀξίωμα refers to an axiom or a self-evident truth. In the context of late Latin grammars, as mentioned before, it was often used in rhetorical and dialectical studies as a fundamental and universally accepted principle on which further reasoning and argumentation could be based. The further developments of the terms can give us better information of their uses: dignitas means in Late Latin grammars “worth” or “dignity”. Deriving from the Latin word dignus, which means “worthy” or “deserving”, in the context of grammar, dignitas refers to the appropriateness of a word, or a language or a linguistic structure within a sentence or discourse to perform a certain meaning. Different words or grammatical elements were assigned varying levels of dignitas based on their role and function in a sentence. Understanding the dignitas of words and their proper usage was crucial for eloquent and effective communication in both spoken and written Latin. Considering the few occurrences in Priscian, one can claim that the term congruitas designates certain constructions or verb types, transitive and intran‐ sitive one. 20 Therefore, the term is not yet one of the parameters for the definition 330 Paola Cotticelli-Kurras <?page no="331"?> verbs interact with different cases, highlighting the nuances of verb usage and their syntactic patterns. On these concepts in Priscian see Meneghel (2020) and Graffi (2021). 21 Compared to previous grammars, the Modistae introduced new definitions and new distinctions, the syntax is better developed and the structure is more systematic and coherent than what had been achieved up to that point. of a complete sentence (congrua et perfecta oratio), as the use of the adjective congruus in the late Latin and especially in the later grammars of the early Middle Ages suggested. The substantive congruitas in grammatical sense had therefore a very restricted use before the medieval grammarians. Instead of the substantive, the root related adjective found large employment also in the Late Latin grammars. Congruitas translates in later (medieval) grammatical texts “congruity” or “suitability.” In the context of Late Latin and medieval grammars, congruitas was used to describe the principle of linguistic harmony and appropriateness in composition. It referred to the proper arrangement and combination of words, phrases, and clauses to create a well-structured and rhetorically pleasing sentence. Following the principle of congruitas ensured that the elements of a sentence were in harmony with each other, resulting in clear and effective communication. The topics covered in the diasynthetica are constructio, congruitas and per‐ fectio. The constructio taken in its formulation given by Thomas of Erfurt (13 th -14 th century): 21 “A construction is accomplished by the dependence of one component on another; but a dependency relationship exists simply between two elements, viz. between the dependent and the determinant. Consequently, exactly two principal components belong to a construction, namely the dependent and the determinant”. constructio describes the principle of dependence in terms of valence, argu‐ ment saturation or relationality. In the following, I give only few examples for the use of the adjective congruus. 3.2.3 Constructio congrua et perfecta; constructibile It is worth quoting the famous definition of Priscian for sentence (oratio): congrua dictionum ordinatio, sententiam perfectam demonstrans (Priscianus, Ars II. 15), the parameters of which are the ordering of the word and the fulfillment of the meaning. In order to keep the relationship to the Greek antecedens of such a definition, Kneepkens (1985) claimed that the terms congruitas and perfectio in Priscian are to be traced back to the corresponding terms of Ancient and Medieval linguistic metalanguage as translational process 331 <?page no="332"?> 22 The Modistae were a small group of scholars active during the thirteenth and fourteenth century, most of them based at the University of Paris, who attempted to trace a systematic relationship among reality, thought, and words via the modi significandi (“modes of signifying”), grammatical properties that directly reflected the modi intelli‐ gendi (“modes of understanding”) of the mind, which were in their turn a reflection of the modi essendi (“modes of being”) or properties of real-world entities. In general, see Marmo (1994, 2011); Kelly (2002), Law (1993, 1999); Copeland - Sluiter (2009); Luhtala (2018). To take an example, the noun signified by means of the modus entis, the “mode of being,” and the modus determinatae apprehensionis, the “mode of fixed reference,” whereas the pronoun, which also signified through the modus entis, was distinguished from the noun by signifying by means of the modus indeterminatae apprehensionis, because a pronoun such as “that” may signify anything from a stone to a shoelace, unlike a noun. Here an interesting definition of the concepts of the modi related to congruitas by John Aurifaber. The author refers to Priscian: “modi significandi sunt inventi propter congruitatem et perfectionem. Sed ista possunt haberi sine hiis, quad patet per experientiam, tum quia antiqui, scilicet Priscianus et Donatus insuper et gramatici sapientissimi, nichil tradiderunt de hiis salvantes nichilominus congruitatem et perfectionem, cum etiam hodie salvantur congruitas et perfectio in multis mundi partibus, videlicet in Francia, Lumbardia et in Anglia et in maiori parte Alcmanie, scilicet in altera parte Reni, id est in Brabantia et in Piccardia tota”. 23 Latin Grammarians used the denomination suppositum and appositum for the concepts which even in Petrus Helias’ definition are oriented to Aristotle’s terms of hypokei‐ menon and kategoroumenon. The quote from Petrus Helias is the following: In omni perfecta oratione dicitur aliquid et de aliquo. Fuit igitur repertum nomen ad discernendum de que est sermo, verbum vero ad discernendum quid dicitur de eo (Thurot, 1868: 217). Apollonius Dyscolus. Perfectio is said to correspond to the phrase αὐτοτελὴς λόγος (‘Proposition, perfect, complete statement’ as a Stoic concept). The bridge to the Modistae 22 is given by some later grammarians, Alcuin (735 - 19 May 804), who quotes Priscian word-by-word, as well as Sedulius Scotus (9 th century; oratio est ordinatio congrua). In Hugh of S. Victor (1096 - 1141) we find a different expression, namely the use of constructio for congrua oratio; according to this step, also the word construction acquires a new and more technical and specific meaning: littera est congrua ordinatio dictionum, quod etiam constructionem uocamus (Hugh of Saint Victor, Didascalicon de studio legendi lib. 3, 58, 16). Another important author is Petrus Helias, who combined the former meaning of congruitas in Priscian as Latin translation of σύμβαμα as “intransitive construction”, adding that these constructions are rightly combined (congrue iunguntur dictiones), and in this way he introduced the new nuance of congruitas through the adverb in order to express the right way of constructing syntactic structures. He adds in his definition that a ‘perfect expression’ is made up by a subject and a predicate, 23 which undergo the agreement, opening the pathway 332 Paola Cotticelli-Kurras <?page no="333"?> 24 As example I give only one complex definition of the perfect sentence, that by Martin of Dacia, Modi significandi 113: Dicunt quidam quod principia perfectionis sunt ista, scilicet suppositum et appositum, et quod suppositum sit per se stans et fixum, et quod appositum sit modi finiti. (. . .) Et illi qui sic dicunt, supponunt quod perfectio non praesupponat congruitatem, quia omnia illa principia reperiuntur in figurativa constructione (. . .). Et isti etiam plus dicunt. Dicunt enim quod figurativa constructio plus est congrua et perfecta quam non figurativa, ut ista ‘turba ruunt’ quam ista ‘turba ruit’, quia magis movet intellectum quam non figurativa. “Some say that the principles of perfection of an utterance are these, that is (that there is) a subject and a predicate, that the subject is independent and fixed (it has two ways of meaning independence and permanence), and that the predicate is finite (…) And those who say this believe that perfection does not require good training, since these principles are (also) found in figurative construction (…). You say more. In fact, they say that the figurative construction is more congruent and perfect than the non-figurative, like this “people noise” versus this “people noises”, since (the former) moves the intellect more than the non-figurative”. 25 See Baratin (1979). to the syntactic theory and definition by the Modistae of a sentence according to its three passiones: constructio, congruitas and perfectio. 24 Petrus Helias develops a syntactic theory, absent in Priscian, 25 which addresses the principles of the construction that brings together noun and verb, which introduces the concepts of agreement, congruitas, regimen and rectio at the center of the diasyntactica or diasynthetica elaborated in the 12 th century commentaries. The transitio is deepened from a theoretical point of view, for example from the possessor to the possessed. The perspective of the Modistae then introduces novelties with respect to Priscian (see also Swiggers 1999: 181ff.). One of them is among others, the new concept of constructibile introduced by the Late Medieval grammarians who were devoted to syntactic theories and who used it in the sense of “word” or “element” as a syntactic unit. A constructibile had no model in the Greek grammars. It would represent a constituent or a part of a larger syntactic unit. Each constructio could have (no more then) two constructibilia, which stand in a hierarchical relationship to each other. This hierarchical relationship between the two constructibilia implies that they are connected in a specific manner within the constructio. This concept reflects the medieval grammarians’ efforts to analyze and describe the structure of language, and to understand how words and elements come together to form meaningful expressions. It’s worth noting that this approach to grammar and syntax is specific to the historical period and the grammatical theories of that time. However, in contemporary linguistic theory and practice, the concept and terminology have evolved, and Ancient and Medieval linguistic metalanguage as translational process 333 <?page no="334"?> modern linguists use different frameworks and approaches to study and analyze language structures. 3.3 The realm of diasynthetica Syntax, i.e. constructio according to Priscian, diasintastica and the like in the later Middle Ages was given very little attention, apart from the famous discussion in books 17 and 18 of Priscian’s Ars grammatica, until the eleventh century. In the early Middle Ages, there was, however, a well-developed repertoire of terms to denote a verb or a preposition “taking” or “governing” a case, i.e. servire + dative, adiungi + dative, trahere + accusative. By the twelfth century sociari, iungi, and construi cum were in common use to describe syntactic relations among the different parts of speech, but a different metaphor was gaining currency: exigere (used by Petrus Helias) and regere (found almost universally in late medieval texts and also in former grammars). The novelties in the terminology of the late medieval syntax are the theory on the term regimen, the study of constructions in which one word “governs” another, obliging it to be in a particular case, and congruitas (“agreement, grammaticality”), as mentioned above. 3.3.1 The place for constructio, its constructibilia and related features Regimen and its theory will replace all previous terms in use in the tradition of Priscian (sociari, iungi, and construi cum), realized as relation by virtue of a feature such as transitivity (ex vi transitionis) or by the copulative function of the verb “to be” (ex vi copule). A constructio, made up of syntactic units, constructibilia, was described in terms of the relation between dependens (“gov‐ erning element”) and terminans (“governed element”); for example, in Socrates legit, legit is dependens, Socrates is terminans, whereas in Socrates albus, albus is dependens and Socrates is terminans (see Covington 1984: 37-61). 3.3.2 Sententia, oratio, clausula Terminology for larger units was less precise, oratio, for example, covering anything from “word” to “sentence” to “text.” Although the terms sententia and clausula were in use in the rhetorical tradition, they did have a technical sense also in grammar. Especially sententia and oratio have many nuances, some of them coming from loan translations from Greek terms: Virgilius Maro Grammaticus (7 th cent.), alone among medieval grammarians, created precise terminology to cover two aspects of the sentence: sententia for “sentence” as a semantic unit, testimonium and quassum for “sentence” as a 334 Paola Cotticelli-Kurras <?page no="335"?> 26 We refer to Cope’s edition of Aristotle’s Rhetoric. Here the translation of Aristotle: ‘The cola and the periods should be neither stubby nor long. A short one often causes the hearer a bump; for when [his mind] is rushing toward what is to come and its measure, of which he has his own definition, he is pulled up short by the speaker’s pausing and trips, as it were, at the abrupt close. Long ones cause him to be left behind, as do those racers who go wide at the turning point; for they, too, lose contact with their fellows. Similarly, long periods turn into a logos and are like a prelude.’ (Transl. Burkett 2011: 213). See also Rapp (2002). 27 Here only the translation of the passage: ‘The kolon is the composition of two or more words, as in ‘I’m missing the reason why I should help Nikia’…. A komma is a part of the kolon, as in ‘know yourself ’ and ‘Nothing in excess’…’ (my translation). For the text see Patillon 2002. For the discussion of the comparison of the two terms in Greek and Latin see also Cotticelli-Kurras (2020). formal unit. During the twelfth century subiectum and praedicatum were taken over from logic to denote “subject” and “predicate” with associated elements; suppositum and appositum replaced them by the end of the century but were later restricted to the subject and (verb) predicate alone, without such additional elements as modifying adjectives, direct objects, and the like. For clausula I refer to the recent research by Cotticelli-Kurras (2022), where I could confirm and enlarge the description of the grammatical meaning given by Lewis & Short, s.v. (the term is not listed in Schad 2007). 3.3.3 Membrum, incisum Major parts of a clause or sentence, phrases and similar, but also single parts, are named incisa or membra. Though the denominations are semantic large, they are used in a linguistic sense, and their Greek origin is undeniable. The pathway of their use and transmission runs through the rhetoric, maybe starting from Aristotle (Arist., Rhet. 3.9.6 for κῶλα which structure the period, 26 and Ael. Arist., Techn. rhet. § 1-4, 27 ) also in this case, but there are different correspondences. Traditionally, in the technical sense, membrum means an element of speech, a member or clause of a sentence. According to Cic.: quae Graeci κόμματα et κῶλα nominant, nos recte incisa et membra dicimus. Cf. Auct. Her. 4, 19, 26. (L&S, s.v. membrum). The definition is still general, since membrum and incisa in the history of grammars do not mean the same, but their technical use is recorded in this way. The two terms are not listed in Schad 2007. 3.4 Articulus The theory of the article, which the post-antique grammarians took over from the Romans and, since the Renaissance, also from the Greeks, is therefore quite confusing. Ancient and Medieval linguistic metalanguage as translational process 335 <?page no="336"?> 28 Cf. also Donatus IV 356, 6: haec Musa, huius Musae etc. This peculiar part of speech extant in Greek but not in Latin is commented by Diomedes underlining the absence of the first in Latin and of the second in Greek: Latini articulum, Graeci interiectionem non adnumerant, (Diomedes 1301, 1). Among the parts of speech, pronouns and articulus are now distinguished according to the model of the Alexandrians, but the relative, together with the form-similar interrogative pronoun, are merged in one and the same category. Apart from the fact that the concept of the article also varies in modern languages, late Roman grammarians do not agree on the distinction between pronoun and articulus. For some, the adjectival demonstrative is an articulus. Priscian always regard hic, haec, hoc in their pragmatic employment within the speech (in oratione) as pronouns and allow them to be valid as articuli or pronomina articularia only in the paradigms of grammarians: Hic vero et haec et hoc indubitanter pronomina sunt quae nisi in declinatione nominum a grammaticis loco articulorum non recipiuntur. (Priscian Ars, III 120, 4) Furthermore, he adds as an explanation the following statement: Nec mirum, cum apud Graecos quoque articuli inveniantur loco pronominum positi. (Priscian Ars, III 120, 7; similar II 581, 24). The Latin grammarians, following the Greek model, are in the habit of placing the demonstrative hic in front of the nouns as gender and case signs: Unde commune [sc. genus] articulum sive articulare pronomen tam masculini quam feminini generis assumit, ut hic sacerdos et haec sacerdos (Priscian II 141, 10). 28 The emphasis with which Priscian repeatedly underlines that Latin has no actual article (II 54, 16; 55, 2; III 11, 25; 119, 28) suggests a stubborn school tradition which does not want to recognize this essential difference between Latin and Greek. See also Schad 2007 s.v. for the quotations by Quintilian, and Priscian, as articuli adiecti from the Stoic attestations, while Charisius 247.1 quotes: “-o, id est τῷ ἄρθρῳ, deficiente (apud Romanos), and the following explanation: ‘By the time of AD, the article and the pronoun have been classified as two separate parts of speech and designated ἄρθρον and ἀντωνυμία respectively (also in DT 23.1-2). AD is at pains to emphasise this point (synt. 118.12ff.). Correspondingly the later Latin grammarians use articulus for article and pronomen for pronoun, although for Prisc., following AD, the scope of the term pronomen is restricted to personal and demonstrative pronouns.’” Later on, Isidore bears witness of a tradition already attested earlier: Inter articulum autem et pronomen hoc interest, quod articulus tunc est, quum nomini coniungitur, ut hic sapiens. Cum vero non coniungitur, demonstrativum pronomen 336 Paola Cotticelli-Kurras <?page no="337"?> 29 Since arthron as PoS is matter of very large references, I omit any reference here. For articulus I limit the reference to Colombat - Lahaussois (2019, 159-201). For a short grammaticographical discussion on articulus from Latin to the romance languages see Heinimann 1965. See also Denecker 2017 for the medieval period. est, ut hic, haec, hoc. (Etymol. I 8, 4). The term articulus is a semantic calque from Greek arthron, though the meaning of the Greek part of speech is not unique. 29 4. Conclusions Through the excursus across the late Roman grammatical tradition and its interplay with the Greek one it was possible to show how the bilingual society and the continuous intermingling of the two cultures have created a mutual influence that gave an exchange on different levels in different periods. Especially in the grammatical tradition the bilingual high formation of the cultivated people could provide an interplay of the two languages. In a first phase, some of the Latin grammatical terms have been created as translations or calque from the Greek ones. In some cases, they were a mechanical process which did not lead to a continuation of the original meaning of the term, as in the case of congruitas. It has been used in the meaning of the corresponding Greek term only in Priscian, but the later use of the term showed an own development in the direction of a new syntactic traditions. Nevertheless, Priscian’s grammar was the first step of the convergence of some categories of the Greek grammatical tradition led by Apollonius Dyscolus with the Latin one. Furthermore, the different needs mirrored in the Latin grammars written in the Eastern regions of the Roman Empire for Greek speaking people who learned Latin as second language have been combined with the interest to enrich the Latin tradition with other categories. The attempt was also to explain the linguistic differences between the two languages, for example in the case of the “article” (articulus and arthron) and the complicated metalinguistic consequences which derived from that. In the Early Middle Ages, the Latin grammatical tradition developed fusing the two principal streams, that by Donatus and that by Priscian, through the commentaries written since the 9 th century. After that, a new impulse came again from the Greek culture thanks to the diffusion of Aristotle’s works, first in the Latin and Arabic translations, later in the original versions, so that a renewed interplay occurred between the two worlds. The logic works by Aristotle gave a new frame to the grammars by the Modistae in the 13 th to the early 15 th centuries, a fact that led to the development of new syntactic theories on the basis of Priscian’s syntactic chapters. In this frame, both new concepts such as constructibile have been created as well as Ancient and Medieval linguistic metalanguage as translational process 337 <?page no="338"?> traditional ones have been connected with new meanings, or filled by new nuances, such as congruitas. Antiquity did not provide the Middle Ages and the Renaissance with a synchronic method that would allow for a sharp categorical delimitation of certain parts of speech with the same form and similar functions. However, the syntax of Apollonius Dyscolus, the first and only comprehensive syntax of antiquity, did a good service to later generations: for the Middle Ages in relation to breaking the tradition of Priscian’s Ars grammatica, and for the Renaissance, moreover, by restoring its original text. Cultural contact from the Late Antiquity onwards did not create solution of continuity in the history of grammar, on the contrary, it provided growth and deepening of ideas towards the formations of the modern syntactic theories. Bibliographical references Ps.-Aelius Aristide (2002) Arts rhétoriques, Livre I (Le discours politique). Livre II (Le discours simple). Texte établi et traduit par M. Patillon, révisé par A.-Ph. Segonds, Paris, Belles Lettres. Alexander de Villa Dei, Doctrinale = Dietrich Reichling (ed.) Das Doctrinale des Alexander de Villa-Dei. Kritisch-exegetische Ausgabe mit Einleitung, Verzeichnis der Handschriften und Drucke, nebst Registern. Berlin 1893; reprint Leipzig without year (= Monumenta Germaniae Paedagogica. Band-12); reprint (edition 1893) Franklin Books, New York 1974. Amsler, M. (1989) Etymology and grammatical discourse in late antiquity and the early Middle Ages, Amsterdam & Philadelphia, Benjamins. Baebler, J.J. (1885) Beiträge zu einer Geschichte der lateinischen Grammatik im Mittelalter, Halle. Baratin, M. (1979) “Priscien et la constitution d’une syntaxe latine”, L’Information Grammaticale 3: 33-37. Baratin, M. (1989) La naissance de la syntaxe à Rome, Paris, Editions de Minuit. Biondi, B.; F. Dedè; A. Scala (2022) (eds.) Ubi homo, ibi lingua. Studi in onore di Maria Patrizia Bologna, Alessandria, Edizioni dell’Orso. Brandenburg, Ph. (2005) Apollonius Dyskolos, Über das Pronomen, Einführung. Text, Übersetzung und Erläuterungen, München & Leipzig, Sauer. Burkett, J.W. (2011) Aristotle, Rhetoric III: a commentary, Los Angeles, Texas Christian University. Chevalier, C. (1968) Histoire de la syntaxe: naissance de la notion de complément dans la grammaire française (1530-1579), Genève. 338 Paola Cotticelli-Kurras <?page no="339"?> Colombat, B.; I. Rosier-Catach; M. Baratin & J. Lallot (1999) “Dictionnaire de la termi‐ nologie linguistique : concordantia, congruitas, consequentia, katallèlotès”, Histoire Épistémologie Langage, 21/ 2, Constitution de la syntaxe: 149-156; doi: https: / / doi.org/ 10.3406/ hel.1999.27 Cope, E.M. (1877) The rhetoric of Aristotle with a commentary of the late Edward Meredith Cope, Cambridge, Cambridge University Press. Cope, E.M. & J.E. Sandys (2011) Aristotle: Rhetoric, volume 3, Cambridge, Cambridge University Press. Copeland, R. & I. Sluiter (2009) Medieval grammar and rhetoric. Language arts and literary theory, AD 300-1475, Oxford, Oxford University Press. Cotticelli-Kurras, P. (2021a) “Constructio and related terms in Medieval grammars: towards a theory of syntax”, in M.-L. Aliffi, A. Bartolotta, C. Nigrelli (eds.), Perspectives on Language and Linguistics. Essays in honour of Lucio Melazzo, Palermo University Press, Springer, Palermo, pp.-93-115. Cotticelli-Kurras, P. (2021b) “The legacy of Priscian and the doctrine of syntax in the medieval grammars”, in Lambert, F. & G. Bonnet (eds.), 155-164. Cotticelli-Kurras, P. (2021c) “Regimen vs. ordo: how to build phrases with words in ME grammars”, in Bartolotta, A. (ed.) Submerged syntax between Late Antiquity and the Modern Age.- Sources, models, and interpretive strategies, Roma, Il Calamo, 7-40. Cotticelli-Kurras, P. (2022) “From clausula to clause, 2000 years of metalinguistic history”, in Biondi, B. et al. (eds.), 1307-327. Cotticelli-Kurras, P. (2023) “The metalanguage of clause structure in Medieval grammars: what about dependent clauses? ”, in Cotticelli-Kurras P. (ed.) Metalanguage, glossing and conceptualization in the grammars of the Middle Ages, Münster, Nodus. Cotticelli-Kurras, P. (i.p.) “Congruitas als Kriterium einer bewussten Sprachform. Ein excursus durch die Latinität”, in Neid, C. (ed.) Proceedings of ‚Sprache und Sprachbe‐ wusstsein in der Geschichte der Sprachwissenschaft: XXXI. Internationales Kolloquium des ‘SGdS’ vom 8.-12. Juni 2022 in Flensburg‘, Münster, Nodus, 35-65. Cotticelli Kurras, P. (2020) (ed.) Word, phrase and sentence in relation: ancient grammars and contexts, Berlin, De Gruyter. Cotticelli-Kurras, P. (2023) (ed.) Metalanguage, glossing and conceptualization in the grammars of the Middle Ages, Münster, Nodus. Covington, M. A. (1984) Syntactic theory in the High Middle Ages, Cambridge, Cambridge University Press. Denecker, T. (2017) Ideas on language in Early Latin Christianity. From Tertullian to Isidore of Seville, Leiden & Boston, Brill. Dionisotti A. C. (1984) “Latin grammar for Greeks and Goths, rev. of La Grammatica dell’Anonymus Bobiensis (GL I 533-565 Keil) by M. De Nonno; Tre Testi Grammaticali Ancient and Medieval linguistic metalanguage as translational process 339 <?page no="340"?> Bobbiesi (GL V 555-566; 634-654; IV 207-216 Keil) by M. Passalacqua, The Journal of Roman Studies, 74: 202-208. Gasti, F. (2003) (ed.) Grammatica e grammatici latini: teoria ed esegesi, Pavia Collegio Ghislieri, Ibis. Graffi, G. (2021) “Transitivity’ — the (partially) classical roots of a problematic concept”, in Cotticelli-Kurras, P. (ed.) From Aristotle to Chomsky. Essays in the history of linguistics, Münster, Nodus, 97-108. Haßler, G. (2018) History of European vernacular grammar writing, https: / / doi.org/ 10.10 93/ acrefore/ 9780199384655.013.386 Householder, F.W. (1981) The syntax of Apollonius Dyscolus, translation and comment, Amsterdam, Benjamins. Hayden, D. (2017, April 26). Language and Linguistics in Medieval Europe. Oxford Research Encyclopedia of Linguistics, from https: / / oxfordre.com/ linguistics/ view/ 10.1 093/ acrefore/ 9780199384655.001.0001/ acrefore-9780199384655-e-380. Heinimann, S. (1963) “Zur Geschichte der grammatischen Terminologie im Mittelalter”, Zeitschrift für romanische Philologie, 79: 23-37. Heinimann, S. (1965) “Die Lehre vom Artikel in den romanischen Sprachen von der mittelalterlichen Grammatik zur modernen Sprachwissenschaft”, Vox Romanica 26: 180-192. Holtz, L. (1981) Donat et la tradition de l’enseignement grammatical: étude et édition critique, Paris, CNRS. Holtz, L. (2011) “Una nuova fonte manoscritta dell’Arte Bernese”, in Munzi, L. (ed.) Custos Latini sermonis. Testi grammaticali latini dell’alto medioevo, Saggi e note testuali di L. Munzi, Pisa & Roma, Fabrizio Serra, 5-29. Isidor of Sevilla, Etymologiae, ed. by Lindsay, M.W., 2 vll., Oxford 1911. Jeep, L. (1893) Zur Geschichte der Lehre von den Redetheilen bei den lateinischen Gramma‐ tikern, Leipzig, Teubner. Julien, M.J. (1988) “La terminologie française des parties du discours et de leurs sous-classes au XVF siècle”, Langages 92: 65-78. Keil, H. (1855-1880) (ed.) = KGL, Grammatici Latini, ex recensione Henrici Keilii, I-VII, Lipsiae (= Hildesheim 1961). Kelly, L.G. (2002) The mirror of grammar. Theology, philosophy and the Modistae, Am‐ sterdam & Philadelphia, Benjamins. Kneepkens, C.H. (1985) “Roger Bacon on the double intellectus: a note on the development of the theory of Congruitas and Perfectio in the first half of the thirteenth century”, in Lewry, P. O. (ed.) The rise of British logic, Toronto, Pontifical Institute of Medieval Studies, 115-143. Lallot, J. (1997) Apollonius Dyscole. De la construction. Texte grec accompagnée de notes critiques, introduction, traduction, notes exégétiques, index, vol. 1, 2, Paris, Vrin. 340 Paola Cotticelli-Kurras <?page no="341"?> Lambert, F. & G. Bonnet (eds.) Apollonius and Priscian: transmission, translation, tradition. A reception history, Turnhout, Brepols. Law, V. (1987) The insular Latin grammarians, Woodbridge, Boydell. Law, V. (1993) History of linguistic thought in the Early Middle Ages, Amsterdam, Benjamins. Law, V. (1994) “The study of grammar”, in McKitterick, R. (ed.) Carolingian culture: emulation and innovation, Cambridge, Cambridge University Press, 88-110. Law, V. (1995) Wisdom, authority and grammar in the seventh century: decoding Virgilius Maro grammaticus, Cambridge, Cambridge University Press. Law, V. (1997) Grammar and grammarians in the early Middle Ages, London, Longman. Law, V. (1999) “Grammar”, in Mantello, F.A.C. & A.G. Rigg (eds.) Medieval Latin, an introduction and bibliographical guide, Washington D.C., The Catholic University of America, 288-295. Law, V. (2003) The history of linguistics in Europe from Plato to 1600, Oxford, Oxford University Press. Löfstedt, B. (1977) Sedulius Scottus: commentum in maiorem Donatum grammaticum, Turnhout, Brepols. Luhtala, A. (1996) “Grammar and dialectic: a topical issue in the ninth century”, in van Riel, G.; C. Steel; J. McEvoy (eds.) Johannes Scottus Eriugena. The Bible and hermeneutics. Proceedings of the ninth international colloquium of the Society for the Promotion of Eriugenian Studies held at Leuven and Louvain-la-Neuve, June 7-10, 1995, Leuven, Leuven University Press, 279-301. Luhtala, A. (2013) “Pedagogical grammars before the eighteenth century”, in Allan, K. (ed.) The Oxford handbook of the history of linguistics, Oxford, Oxford University Press, 341-358. Luhtala, A. (2018) The Art of grammar in the Middle Ages, Turnhout, Brepols. Luhtala, A. (2020) “Syntactic relations in ancient and medieval grammatical theory”, in Imrényi, A. & N. Mazziotta (eds.) Aspects of the theory and history of dependency grammar, Amsterdam, Benjamins, 23-58. Mantello, F.A.C. & A.G. Rigg (1999) (eds.) Medieval Latin: an introduction and bibliograph‐ ical guide, Washington D.C., Catholic University of America Press. Marmo, C. (1994) Semiotica e linguaggio nella Scolastica: Parigi, Bologna, Erfurt 1270-1330. La semiotica dei modisti, Roma, Istituto Storico Italiano per il Medio Evo. Marmo, C. (2011) La semiotica del XIII secolo. Tra arti liberali e teologia, Milano, Bompiani. Meneghel, R. (2020) “On the metalinguistic passage from διάβασις and μετάβασις by Apollonius Dyscolus to transitio by Priscian”, |in Cotticelli-Kurras, P. (ed.), 117-150. Merrilees, B.S. (1988)-“Les débuts de la terminologie grammaticale en français: à propos de quelques travaux récents”, Romania 109-: 397-411. Ancient and Medieval linguistic metalanguage as translational process 341 <?page no="342"?> Milani, C. (2004) “Il septimus e l’octavus casus nel pensiero dei grammatici latini”, in Graffi, G. (ed.) Fortuna e vicissitudini di concetti grammaticali, Padova, Unipress, 9-43. Mittellateinisches Wörterbuch, (1939-1996). München, Bayerische Akademie der Wissen‐ schaften, https: / / mlw.badw.de/ mlw-digital.html Mok, Q.I. (1975) “Un traité médiéval de syntaxe latine en français”, Mélanges de linguis‐ tique et de littérature offerts à Lein Geschiere, Amsterdam, Rodopi, 37-53. Phillips, N. (1999) “Music”, in Mantello, F. & A.G. Rigg (eds.), 296-306. Priscien, Grammaire Livre XVII — Syntaxe, 1. (2010)-Texte latin, traduction introduite et annotée par le Groupe Ars Grammatica animé par M. Baratin et composé de F. Biville, G. Bonnet, B. Colombat, A. Garcea, L. Holtz, S. Issaeva, M. Keller, D. Marchand, Paris, Vrin. Priscien, Groupe Ars Grammatica (trad.) (2017) Grammaire Livre XVIII - Syntaxe, 2, Paris, Vrin. Rapp, Ch. (2002) Aristoteles: Rhetorik. Übersetzt und erläutert, Berlin, Akademie-Verlag. Rizza, A. (2022) “Vox, constructio, sententia. Particolarità in Alcuino di York”, in Biondi, L. et al. (eds.), vol. 2, 1119-1129. Rosier, I. (1983) La Grammaire spéculative des Modistes, Lille, Presses Universitaires. Rosier, I. (1992)-La terminologie linguistique latine médiévale, in Auroux, S. (ed.) Histoire des idées linguistiques, Tome 2: Le développement de la grammaire occidentale, Liège, Mardaga, 590-597. Russell, P. (2012) “An habeas linguam latinam? Non tam bene sapio: views of multilin‐ gualism from the early medieval West”, in Mullen, A. & P. James (eds.) Multilingualism in the Graeco-Roman worlds, Cambridge, Cambridge University Press, 193-125. Schad, S. (2007) A lexicon of Latin grammatical terminology, Pisa & Roma, Fabrizio Serra. Städtler, Th. (1988) Zu den Anfängen der französischen Grammatiksprache, Berlin & Boston, De Gruyter. Swiggers, P. (1995) “L’heritage grammatical greco-latin et la grammaire au Moyen Âge”, in Welkenhuysen, A.; H. Braet; W. Verbeke (eds.) Mediaeval antiquity, Leuven, Leuven University Press, 159-195. Swiggers, P. (2009) “Review to Samantha Schad, A Lexicon of Latin Grammatical Terminology”, Bryn Mawr Classical Review, https: / / bmcr.brynmawr.edu/ 2009/ 2009.02. 03/ Swiggers, P. & A. Wouters (1999) “Translation ancient grammatical texts”, in Cram, D.; A.R. Linn; E. Nowak (eds.) History of linguistics 1996, Volume 2, From classical to contemporary linguistics, Amsterdam & Philadelphia, Benjamins, 3-13. Swiggers, P. & A. Wouters (2015a) “Priscian on the distinction between adverbs and conjunctions”, in Xenis, G.A. (ed.) Literature, scholarship, philosophy, and history. Classical studies in memory of Ioannis Taifacos, Stuttgart, Steiner, 265-275. 342 Paola Cotticelli-Kurras <?page no="343"?> Swiggers, P. & A. Wouters (2015b) “Description of the constituent elements of the (Greek) language”, in Montanari, F.; S. Matthaios; A. Rengakos (eds.)-Brill’s companion to Ancient Greek scholarship, 2 vols., Leiden, Brill, 759-797. Swiggers, P. & A. Wouters (2003) (eds.) Syntax in Antiquity, Leuven, Peeters. Thomas of Erfurt, Grammatica Speculativa, An edition with translation and commentary by G.L. Bursill-Hall, London, Longman 1972. Thurot, Ch. (1964) Extraits de divers manuscrits latins pour servir à l’histoire des doctrines grammaticales au Moyen Âge. (Reprint of the edition Paris 1868, Notices et extraits des manuscrits de la Bibliothèque Impériale et autres bibliothèques 32, 2), Frankfurt/ M., Minerva. Vineis, E. (1994) “La linguistica medievale”, in Lepschy, G. (ed.) Storia della linguistica, vol. 2, Bologna, Il Mulino, 13-102. Ancient and Medieval linguistic metalanguage as translational process 343 <?page no="345"?> La négation comme moteur de l’évolution linguistique Anna Orlandini & Paolo Poccetti (Centre A. Ernout, Paris-Sorbonne & Université de Rome 2) Abstract: Negation and related phenomena are very rarely considered in the perspective of language change. Both main negative structures, i.e. D(ouble) N(egation), and N(egative) C(oncordance), which typologi‐ cally split the languages of the world, are observed as synchronic and diachronic variations in the history of ancient Greek and Latin, as well as in their evolutionary paths towards Modern Greek and the Romance languages, respectively. Particularly, some basic facts, generally involved in the change of negative structures, are focused, namely, the system of the indefinite pronouns and the interplay between different types of comparatives and superlatives under negation. A special attention is devoted to expressions with ‘separative’ or ‘exclusive’ functions, such as disjunction, and to terms indicating otherness or contrast between two items, mostly signalled by the inherited morpheme -tero-, evidenced by pronouns, particles and comparatives of almost all I.E. languages. The evolution from DN to NC shown by the Romance languages with respect to literary Latin is consistent with the tendency to shift from logical onto natural structures, which is universally observed in the world’s languages as well as in sociolinguistically low registers of a language. Keywords: Negative structures, language change, diachronic vs. syn‐ chronic variations, Latin, Greek, Romance Languages Introduction Dans toutes les langues, le système de la négation est assujetti à des variations et synchroniques et diachroniques, qui se répercutent dans l’évolution langa‐ gière. Néanmoins, parmi les phénomènes généraux de l’évolution des langues, les changements sémantico-lexicaux et sémantico-syntaxiques mis en œuvre <?page no="346"?> 1 Pour d’autres explications étymologiques de non, à notre avis, moins plausibles, voir Dunkel (1987, 23 ; 2014, 533), mais voir aussi les remarques là-dessus dans Orlandini-Poccetti (2022, 5-6). 2 Orlandini-Poccetti (2022). par la négation ont été très rarement pris en compte. De ce point de vue, les variétés du latin et des langues romanes constituent un domaine privilégié pour mettre en valeur le rôle de la négation en tant que véritable moteur de l’évolution linguistique. En effet, l’analyse contrastive de plusieurs contextes du latin et des langues romanes fait ressortir le rôle joué par la négation dans le changement. L’évolution de la négation met en valeur le noyau essentiel qui distingue la langue de Rome de ses langues-filles, notamment le parcours qui amène de l’emploi de la négation « logique » vers la négation « non logique », qui caractérise la plupart des langues naturelles. Ce parcours cohérent et linéaire, qui se manifeste dans le système de la négation, touche plusieurs domaines sémantiques, qui n’ont jamais été mis en corrélation sous cette perspective. 1. Négation standard et négation volitive La manifestation la plus saisissante de l’évolution de la négation du latin aux langues romanes, qui est parallèle à celle d’autres langues, concerne l’effacement de la distinction entre la négation standard, employée dans les propositions assertives, et la négation des propositions non-assertives, incluant les questions, les ordres, les souhaits et les propositions subordonnées de force volitive. La majorité des langues indo-européennes a développé une négation assertive par le biais d’un élargissement morphologique de la particule *n qui est la seule proto-forme reconstruite, qui se retrouve dans la quasi-totalité des langues appartenant à cette famille. Le plus souvent, l’élargissement morphologique est issu de l’univerbation avec un terme à polarité négative (NPT) ou une particule coordonnante. Ainsi, à partir de *n , le latin a créé la particule nōn < n +oinom  1 et nec < n -k w e pour les propositions assertives, en la distinguant de n , réservée aux propositions non-assertives (interrogatives et volitives). De leur côté, les langues romanes ont effacé cette distinction, en remplaçant l’expression de la négation dans les volitives par celle des assertives, comme, par ex. en français «-tu ne parles pas-» vs. «-ne parle pas-», «-afin que tu ne puisse pas parler-». On n’insistera pas davantage sur ce phénomène, auquel on a consacré un autre travail 2 , si non pour souligner que chaque langue, indépendamment l’une de l’autre, a parcouru le même chemin qui a amené à la création d’une particule réservée à la négation standard, qui, par la suite, s’est substituée à celle de la 346 Anna Orlandini & Paolo Poccetti <?page no="347"?> 3 Haspelmath (2005). 4 Viti (2011). négation non-assertive (interrogative et volitive). Cet aboutissement distingue, dans la quasi-totalité, les langues modernes des langues anciennes du domaine indo-européen, s’inscrivant dans le parcours général, selon lequel l’élément de l’expression non marquée remplace celui de l’expression marquée. 2. Le rôle de la négation dans la positivisation des indéfinis (N-mots) dans le passage d’une structure à « double négation » (DN) à une à «-concordance négative-» (NC) Le latin classique est une langue ‘compositionnelle’ à double négation (DN), c’est-à-dire que la présence de deux items négatifs fait ainsi qu’ils effacent réciproquement leur charge négative, engendrant une interprétation positive pour les éléments quantifiés. Cette langue présente la particularité d’être une langue de type V-NI («-verbe non nié-»---«-pronom indéfini négatif à négation incorporée », ex. uenit nemo), alors que la plupart des langues romanes sont de type NV-NI (« négation verbale » et « pronom indéfini négatif »). En latin classique, nemo, nullus, nihil, numquam, nusquam, etc. sont des quantifieurs négatifs à part entière, ils ne se bornent pas à nier le mot auquel ils sont incorporés, mais ils engendrent à eux seuls une négation propositionnelle. En revanche, la NC, qui cumule plusieurs éléments morphologiquement négatifs, mais où un seul exprime la négation, est en contraste avec le « principe de compositionalité-», qui pose que l’interprétation sémantique d’une proposi‐ tion est le résultat de la somme des sens de ses composants. Par ailleurs, la non-compositionalité est plus naturelle et fréquente dans les langues du monde 3 , si bien qu’elle est connue même en ancien indien 4 . La double négation (DN), qui est normalement un phénomène non marqué en latin classique, devient plus rare, marquée, contrastive et subordonnée à des contraintes dans les langues romanes, alors qu’elle est la structure privilégiée de l’allemand, de l’anglais, du néerlandais. La DN est toujours binaire, logique, elle opère sur couples d’items pleinement négatifs qui interagissent entre eux, comme c’est le cas du passage suivant : (1) emo Arpinas non Plancio studuit (Cic. Planc. 22) «-Pas un homme d’Arpinum qui n’ait été le partisan de Plancius » (=Tous les Arpinates ont été les partisans de Plancius). La négation comme moteur de l’évolution linguistique 347 <?page no="348"?> 5 Sur la question de la variété langagière du Bellum Africum, voir Cioffi (2019). 6 Neque est la leçon des mss. SLN (Ashburnhamensis, Louaniensis et Neapolitanus), alors que les mss. MUR (Mediceus, Ursinianus, Riccardianus) et TV (Thuaneus et Vindobonensis) ont necne. 7 Elle doit être insérée le plus tôt possible, selon le Negative First Principle de Jespersen, réformulé par Horn (1989,293) pose : « There is a tendency to put the negative element as early as possible in an utterance because the contribution of the negation to the meaning of a larger constituent is particularly dramatic and the hearer needs to get this information as soon as possible». 8 de Swart (2010,22) écrit : « In languages such as Italian, we can see the ‘self-licensing’ nature of n-words at work …, where no licensor is available for the n-word in subject position (l’italique est nôtre). The ‘self-licensing’ nature of n-words is harder to illustrate in languages in which a marker of sentential negation is obligatory present in all sentences containing an n-word. Such languages are labeled ‘strict’ negative concord languages … and are opposed to the ‘non-strict’ variety of negative concord found in Italian-». Selon Ö. Dahl (2010,14) la DN est une structure symétrique parce que deux termes également négatifs y sont impliqués, alors que la concordance négative (NC) relève d’une construction asymétrique, parce que seulement un terme est un marqueur négatif et l’autre est quelque chose d’autre-: « traiting one of them (= negative markers) as a negative marker and the other as something else-». Mais, à notre avis, c’est précisément la présence d’une négation cumulative dans une structure non-compositionnelle qui engendre la positivisation des « N-mots », en effaçant leur charge négative en relation avec le rôle sémantique qu’ils jouent dans la proposition. Une telle tendance, qui se rapproche de l’état des langues romanes, est manifestée par certaines variétés du latin, comme dans le cas du passage suivant-du Bellum Africum  5 : (2) neque 6 locum excusatio nullum haberet (Bell. Afr. 8) « Et il n’y avait pas à dis‐ cuter les possibilités d’exécution » «-impossible n’importe quelle excuse / toute excuse-». Au niveau de la représentation sémantique, les énoncés de ce type ne présentent qu’un seul opérateur NEG, qui peut être engendrée par la négation préverbale 7 ou par le quantifieur négatif, selon le rôle sémantique de ce dernier. Par ex., en italien le « N-word » nessuno dans le rôle du sujet et en position préverbale ne nécessite pas la négation du prédicat : il est « self-licensing » 8 . S’il y a dans phrase d’autres éléments négatifs, notamment d’autres « N-words », ils n’ont pas un sens pleinement négatif, comme dans la proposition (3) Nessuno mi ha detto niente «-personne ne m’a rien dit-». 348 Anna Orlandini & Paolo Poccetti <?page no="349"?> 9 Meyer-Lübke (1890-1900, III, 778 sq.). 10 Cela correspond à l’idée de Giannakidou (2000,90) : « Of course, the extent to which NC constitutes a problem for compositionality depends on whether we taken n-words to be semantically negative or not». 11 Cf. Viti (2011), Giannollo (2018). 12 Comme le disent les auteurs : (1977, 96) « other work indicates that Accessibility Hierarchy may play a more general role in determining the accessibility of noun phrases as candidates and targets for syntactic process, although the precise delimitation of the area of relevance of Accessibility Hierarchy remains a task for future research.” L’exemple de l’italien, parallèle à celui du français, excepté l’emploi de la négation préverbale exclue de l’italien, requise en français, met en relief le fonctionnement différent du quantifieur négation par rapport à son rôle séman‐ tique. Une telle variation subordonnée au contexte a joué un rôle essentiel dans l’évolution diachronique de l’emploi comme quantifieurs pleinement négatifs réalisant la DN en latin classique, à un emploi qualitatif comme quantifieurs indéfinis génériques (« free choice ») ou quantitatif comme Minimiseurs ou comme Termes à polarité Négative (TPN), aboutissant à la NC en latin tardif et dans les langues romanes. Déjà Meyer-Lübke avait signalé que la négation cumulative serait à l’origine des emplois “positifs” des « sémi-négations » (les « N-Mots ») des langues romanes 9 . Cela est la vraie clé interprétative du phénomène de la NC : les emplois positifs des indéfinis qui, dans la négation cumulative, gardent encore une morphologie négative mais qui ont affaibli leurs sémantisme négatif 10 . Dans une perspective plus fonctionnelle que sémantique, plusieurs savants qui se sont occupés de NC ont fait l’hypothèse que la fonction focalisante soit à l’origine de telle structure 11 . C’est une hypothèse fonctionnellement tout à fait plausible, mais elle ne nous donne pas l’explication sémantique du procès évolutif. Celui-ci prend en compte principalement la positivisation des indéfinis, comme le suggère Meyer-Lübke, ainsi que, ajoutons-nous, le jeu réciproque et alterné de l’affaiblissement de la négation propositionnelle non et du n-mot, qui s’affaiblit plus ou moins en relation avec son rôle argumental, comme nous le verrons dans la suite. D’une manière peut-être surprenante, mais nous expliquerons pourquoi, la possibilité pour un indéfini négatif de garder intactes les propriétés quantifica‐ tionnelles et le sémantisme négatif ou, en revanche, de s’affaiblir, semble pouvoir être indiquée par l’échelle hiérarchique d’accessibilité à la relativisation pour un SN, formulée par Keenan et Comrie (1977, 66) 12 -: La négation comme moteur de l’évolution linguistique 349 <?page no="350"?> 13 Ainsi déjà Horn (2020, 11) : « quantifier negation in subject position functions as ’more negative’ than that in nonsubject position-». Accessibility Hierarchy (AH) SU>DO>IO>OBL>GEN> OCOMP Figure 1. Cette échelle, où > signifie “plus accessible que”, nous prédit qu’à l’extrême pôle à droite il est possible d’avoir seulement un TPN ou un « free choice » (lat. quisquam, angl. anybody, osc. pis), alors qu’à l’ extrême pôle à gauche le seul admis est le quantifieur négatif (lat. nemo, nullus, nihil, numquam, etc., osc. nepis, it. nessuno, fr. personne, esp. nadie, angl. nobody). Au fur et à mesure qu’on se déplace vers la droite, le quantifieur négatif s’affaiblit et la négation propositionnelle se renforce. Ainsi dans la proposition : Non è venuto nessuno, l’indéfini sujet nessuno garde sa valeur quantificationnelle pleine et il négativise à lui seul toute la proposition 13 , non est, dans ce cas, pléonastique. En revanche dans : Non l’ho detto a nessuno, l’indéfini a le rôle d’objet indirect, il est un « N-mot » (il a la valeur d’un « free choice » a chicchessia) et la négation non est une vraie négation phrasale. Mais l’échelle d’accessibilité de Keenan & Comrie (1977) peut se rapprocher de la hierarchy of negative items de T. Espinal (2007,55) dans le schéma reproduit ci-dessous-: Hierachy of negative items NQ / Minimizers> N-words> PI Figure 2. Cette échelle représente une variation décroissante de force négative et quantifi‐ cationnelle selon un cycle négatif. Dans ce schéma les vrais quantifieurs négatifs, représentés par tous les pronoms à négation incorporée du latin classique, ont une force quantificationnelle majeure que les « N-words » (les pronoms négatifs non sujet du latin tardif et des langues romanes) qui ont une force quantificationnelle variable et moindre et enfin les PI (items à polarité) qui sont des simples variables sans force quantificationnelle. Il faut donc postuler un basculement réciproque entre la force négative de la négation de phrase et le pronom négatif. La force négative de la négation de phrase, en cas de concordance négative, est moindre en présence d’un pronom négatif sujet, mais elle est la plus grande en présence d’un TPN. C’est pourquoi l’on peut esquisser deux parcours sur l’échelle de Keenan-Comrie (1977) : une force amoindrissant pour le pronom négatif en relation avec son rôle sémantique, une force croissante pour la négation de phrase. 350 Anna Orlandini & Paolo Poccetti <?page no="351"?> 14 C’était bien connu de Dante, Inf. XXVII, 118-123 : «-…ch’assolver non si può chi non si pente / né pentere e volere insieme puossi / per la contradizion che nol consente / Oh me dolente ! come mi riscossi / quando mi prese dicendomi : “Forse / tu non pensavi ch’io loico fossi”. 3. Le rôle de la négation dans l’évolution de la disjonction. 3.1. Dans une disjonction exclusive. 3.1.1. Latin Aut Du point de vue sémantique, la disjonction exclusive engendre un sens négatif, parce qu’elle implique l’exclusion de l’un des deux éléments de l’ensemble considéré, c’est-à-dire X aut Y « ou X ou Y » implique « X, mais non Y » ou, inversement, « Y, mais non X ». La disjonction exclusive, tout comme la DN, appartient au langage logique 14 . En latin classique, aut exprime la disjonction exclusive, hors de la portée de la négation, en rapport avec le monde réel, souvent en présence de la modalité objective. Cette particule exprime une disjonction forte, marquée, qui ajoute, au sens de base inclusif, l’implicature conversationnelle, de nature pragmatique : « l’un, mais pas les deux disjoints à la fois », donc une implicature d’exclusion de l’un de deux éléments de l’ensemble considéré. La même implicature est partagée par le pronom indéfini latin alteruter, qui véhicule le même sens d’exclusion entre deux éléments ou ensembles, notamment « l’un ou l’autre (entre deux)-». La négation a des répercussions sur l’évolution de la disjonction exclusive, exprimée en latin par aut. En effet, la négation permet le passage d’une structure binaire, symétrique, compositionnelle et marquée, qui se réalise dans des contextes de re, grâce à la disjonction exclusive latine aut, à une structure cumulative, asymétrique, non marquée, qui se réalise normalement dans des contextes de dicto, non assertifs et qui correspond à la disjonction inclusive latine uel. La négation ayant une portée phrasale a l’effet d’effacer l’implicature conversationnelle d’exclusion liée à la disjonction exclusive aut, et au sémantisme de l’indéfini alteruter (non alteruter =uteruis). Par le biais de la négation externe (non aut), on passe, en effet, d’une disjonction exclusive (aut), dans une série fermée de deux éléments (ou p ou q), à une la “disjonction inclusive”, (les deux éléments à la fois, où plusieurs éléments s’il s’agit d’une série ouverte). L’ouverture vers d’autres éléments, qui dépassent une relation limitée à deux termes, est commune et à l’évolution de la disjonction latine et à l’évolution de la structure à double négation (DN). En effet, la structure à (DN) repose sur un principe binaire, dans la mesure où elle est symétrique et compositionnelle, alors que la structure à concordance négative (NC), en tant La négation comme moteur de l’évolution linguistique 351 <?page no="352"?> qu’asymétrique et non-compositionnelle, est non-binaire et cumulative. Ainsi, la structure à (DN) est la plus proche de la logique, dont le noyau essentiel est la compositionalité de deux éléments, alors que la structure à (NC) s’éloigne du langage de la logique, et caractérise plutôt les langues naturelles, parce qu’elle prend en compte les variations de sens engendrées par la possibilité de-«-positivisation-» des «-N-words-», selon leur rôle syntactico-sémantique. En latin, la disjonction inclusive peut être introduite par uel dans des contextes positifs, et par nec, neque ou par non … aut dans des contextes négatifs, alors que la disjonction exclusive est toujours introduite par aut qui échappe à la portée de la négation. À notre avis, l’emploi inclusif de aut sous négation (qui se réalise dans des contextes non-assertifs ouvrant sur des mondes possibles) peut être considéré comme l’équivalent d’une conjonction de coordination dans la portée de la négation de phrase, autrement dit, d’une coordination de deux propositions négatives : (4a) non famem aut sitim, neque frigus neque lassitudinem opperiri (Sall. Catil. 13,3) «-on n’attendait pas la faim ou la soif, ni le froid, ni la fatigue-». (4b) Nec aut Persae aut Macedones dubitauere quin ipse rex esset occisus (Curt. 4,15,28) «-Ni les Perses ni les Macédoniens ne doutèrent que c’était le roi qui avait été tué-». Dans les deux contextes ((4a) et (4b)) aut ne fonctionne pas comme disjonction exclusive, mais inclusive des deux éléments disjoints. En revanche, aut exclusif, qui est contextuellement déterminé, ne peut jamais se trouver dans la portée de la négation, alors qu’il implique, à lui seul, une négation d’exclusion (« l’un des deux disjoints, pas les deux à fois”)-: (5) “Aut hoc, aut illud. Hoc autem; non igitur illud”. Itemque: “Aut hoc, aut illud. Non autem hoc; illud igitur”. […] Deinde addunt coniunctionum negantiam, sic: “NON [ET hoc ET illud]” . Hoc autem; non igitur illud. […]”NON [ET hoc ET illud]”. Non autem hoc; illud igitur.” (Cic. top. 56-57). « ‘C’est l’un ou l’autre. Or c’est l’un. Donc ce n’est pas l’autre’. De même : “ ‘C’est l’un ou l’autre. Or ce n’est pas l’un ; donc c’est l’autre’ […] ‘Ce ne peut être à la fois ceci et cela. Or c’est ceci. Donc ce n’est pas cela’ […]’Ce ne peut être à la fois ceci et cela. Or ce n’est pas ceci. Donc c’est cela’-». La disjonction exclusive, qui concerne le monde réel et la modalité objective, n’admet pas l’hypothèse que les propositions p et q peuvent être vraies ou 352 Anna Orlandini & Paolo Poccetti <?page no="353"?> 15 Cf. Prisc. III 97,17-27 GLK Disiunctiuae sunt, quae, quamuis dictiones coniungunt, sensum tamen disiunctum et alteram quidem rem esse, alteram uero non esse significant […] et est quando indicatiue dicuntur, est quando dubitative : indicatiue, ‘aut lux est aut tenebrae’ ; dubitatiue, ut ‘aut prodest aut non prodest diuitias quaerere’. fausses en même temps. La disjonction exclusive est ainsi conforme à la “loi du tiers exclu” 15 . Le choix exclusif ne concerne que l’un des deux volets d’une alternative ; dans ce sens, la disjonction exclusive implique une négation d’exclusion (“p, et non pas q” ou “q, et non pas p”). La nature de cette opposition est objective, l’exclusivité est donnée par la situation contextuelle en relation avec l’état des affaires du monde réel. Aut de la disjonction exclusive équivaut à une assertion hypothétique alternative : “si non p, alors q “ou : “si p, alors non q”. La disjonction exclusive (aut…aut), toujours hors de la portée de la négation, exprime l’impossibilité de coexistence, dans le monde réel, de p et q, où l’une est vraie et l’autre est fausse, mais non les deux à la fois. Ainsi non [et … et] = non [et hoc et illud]. Par rapport à la négation, l’emploi inclusif (dans la portée de la négation) et l’emploi exclusif (hors de la portée de cette négation) sont en distribution complémentaire. Cela nous confirme que l’emploi inclusif, le seul admis dans la portée de la négation, est l’emploi non marqué de l’opérateur de disjonction. En effet, l’emploi exclusif doit être toujours mis en relation avec un contexte particulier, et il est porteur d’un trait distinctif, notamment l’implicature con‐ versationnelle d’exclusion de l’un des deux disjoints. En diachronie, aut s’impose sur uel qui cesse son existence, en passant aux dérivés de aut son sens. Dans les langues modernes, un seul marqueur (o, ou, or) réalise à la fois, selon les contextes, et la disjonction exclusive et celle inclusive, quand l’implicature conversationnelle d’exclusion est effacée. 2.1.2. Alteruter vs. uteruis Un fonctionnement analogue, dans le domaine de la quantification binaire en latin, se retrouve avec alteruter, qui lexicalise la disjonction exclusive (« ou » bloqué vers le haut : « l’un, mais pas les deux à la fois ») et uteruis qui lexicalise la disjonction inclusive, dont la valeur de base est celle de « ou » non bloquée vers le haut (« l’un, et peut-être les deux à la fois »). La différence essentielle entre les deux pronoms est manifestée par leur formation : alors que uteruis signifie-: «-ou a ou b, indifféremment-», alteruter signifie : «-ou a ou b, mais un seul des deux, pas les deux à la fois » avec l’idée négative d’exclusion. Par conséquent le pronom / adjectif alteruter est le quantifieur binaire le plus proche de aut. En revanche, uteruis signale que le choix entre les deux options est indifférent La négation comme moteur de l’évolution linguistique 353 <?page no="354"?> 16 Selon l’équivalence logique entre non (A ou B) et (non A) et (non B), posée par la loi A. De Morgan. ou non significatif pour le locuteur. Il peut concerner un seul ou bien les deux membres de l’ensemble. La valeur de disjonction inclusive exprimée par uteruis est à mettre en relation avec la valeur de uel, ce qui se manifeste aussi du point de vue formel, parce que les deux se composent d’éléments issus de la racine de uelle. En revanche, hors de la portée de la négation, l’emploi de alteruter est fonctionnellement proche de celui de aut-: (6) ‘Si enim’, inquit, ‘alterutrum concessero necessarium esse, necesse erit cras Hermarchum aut uiuere aut non uiuere’ (Cic. ac. 2,97) «-Si en effet, des deux termes de l’alternative l’un s’impose nécessairement, il sera nécessaire que demain Hérmarchus vive ou qu’il ne vive pas-». En (6) alterutrum a comme domaine de référence deux prédications contradic‐ toires uiuere et non uiuere et dans ce domaine il quantifie nécessairement sur l’une des deux contradictoires. La lecture qu’on obtient grâce à alteruter est ainsi-: «-L’un ou l’autre, mais pas les deux à la fois-». La négation, effaçant l’implicature conversationnelle d’exclusion, favorise pour alteruter une interprétation inclusive, cumulative, tout comme pour aut dans des contextes négatifs 16 -: (7) neque (se) rem publicam alterutro exercitu priuare uoluisse (Caes. civ. 3,90,2) «-Il n’avait pas voulu priver l’État de l’une des deux armées-». Lorsque l’indéfini est dans la portée de la négation, le renvoi à un seul élément d’un ensemble de deux est neutralisé, puisque l’on nie à la fois les deux éléments : le passage (7), se référant aux deux armées, réalise le sens « ni (le priver) de l’une ni de l’autre »). Neuter peut alors apparaître en corrélation avec alteruter dans un contexte où alteruter n’a plus de force exclusive-: (8) Si duo et alteri duo numeri ponuntur sic, ut neuter ex prioribus cum alterutro ex insequentibus componi possit, ne is quidem, qui ex duobus prioribus factus est, componi cum alterutro insequentium potest. (Mart. Cap. nupt. 7,782) «-Si l’on pose deux nombres, puis deux autres, de sorte que ni l’un ni l’autre des deux premiers ne puisse se comparer à l’un ou à l’autre des deux suivants, la somme des deux premiers non plus ne peut se comparer ni à l’un ni à l’autre des deux suivants-». 354 Anna Orlandini & Paolo Poccetti <?page no="355"?> 17 Lehmann (1986, 199, s.v. ƕaƥar) ; Kluge (1989, 780, s.v. weder) ; Onions (1966, 304, s.v. either-; 630, s.v. or). En osque, un indéfini qui correspond, du point de vue et formel et fonctionnel, à alteruter du latin apparaît dans la forme du locatif alttreí putereípíd : (9) alttreí putereípíd akeneí (T.A. B 21=ImIt 1200-1205) «-tous les deux ans-». Dans ce syntagme nominal, se référant à un rite religieux qui doit se dérouler tous les deux ans, l’indéfini véhicule le même sens d’exclusion : le décalage temporel implique l’exclusion d’un troisième an. Les indéfinis du latin alter, uter(que), neuter et leurs composés, tout comme l’osque alttreí putereí, se relient, du point de vue sémantique, par la référence à un ensemble de deux éléments, qui peut comporter l’alternative entre l’un ou l’autre (alter, uter), l’exclusion des deux (neuter) ou l’inclusion de deux (uterque) et, du point de vue morphologique, par la dérivation par le suffixe -tero-. Le rôle de marquer une alternative est montrée par les langues germaniques, qui ont développé certains dérivés en -teroen fonction de l’exclusion, ce qui est encore partiellement visible dans les langues modernes. Ainsi les deux particules corrélatives, qui marquent la disjonction exclusive de l’anglais either… or remontent respectivement l’une au vieux angl. āwther, issu de la contraction de āhwaether, et l’autre à oƥƥe, pourvu de la terminaison en -(e)r de son corrélatif en -ther. Encore plus clairement, le parallélisme morphologique de dérivés en -terose manifeste dans les deux particules corrélatives de la disjonction en allemand entweder…oder « ou…ou », ayant la fonction de marquer l’exclusion d’un élément dans un ensemble de deux. Plus précisément, l’une (weder), tout comme l’anglais whether et le got. ƕaƥar, doit être reconduite à une formation de marqueur d’alternative commune à plusieurs langues indo-européennes, aboutissant à uter en latin, à πότερος en grec, à kataráḥ en a.ind., alors que l’autre (oder) a suivi un parcours de réfection analogique parallèle à celui de or en anglais 17 . 3.2. Le suffixe -teroet son évolution Le suffixe -teroa attiré à plusieurs reprises l’attention des comparatistes, parce qu’il se retrouve dans la plupart des langues indoeuropéennes, où il assure trois fonctions essentielles, notamment-: La négation comme moteur de l’évolution linguistique 355 <?page no="356"?> 18 Baldi & Cuzzolin (2010, 48-50). a. la formation d’indéfinis, se référant à l’un de deux éléments d’un ensemble, tels qu’en latin alter, uter, ou à la totalité négative d’un ensemble de deux, lat. neuter, en grec οὐδέτερος, πότερος, ἕτερος, ἑκάτερος, en osque alttreí putereí (loc. sg.), en indien ancien kataráḥ, itara-, en lituanien katràs, en gotique ƕaƥar ; b. l’expression du comparatif, spécialement des adjectifs à base thématique -o, d’inégalité, comme, par ex., gr. σοφώτερος, ind.anc. juṣṭa-taraḥ, ou d’égalité, donnant lieu à l’équatif du vieux irlandais (ex. demnithir, lérithir) ; c. la dérivation lexicale à partir de bases nominales et pronominales (lat. mater «-mère-» vs. matertera «-tante maternelle-»-; a.ind. aśvaḥ «-cheval-» vs. aśvataraḥ « mulet » ; lat. nos, uos vs. noster, uester ; gr. ἡμεῖς, ὑμεῖς ~ ἡμέτερος, ὑμέτερος; ) ou adverbiales (lat. magis-ter, minis-ter ; exter-ior, inter-ior, ulter-ior, citer-ior). Les trois fonctions de -terose concentrent en indien ancien, alors qu’elles se distribuent différemment dans d’autres langues. L’origine de ce morphème ainsi que son rôle sémantique ont été l’objet de discussion 18 , mais, depuis longtemps, à la base des emplois différents, on reconnaît généralement une fonction contrastive ou oppositive ou séparative entre deux éléments. La référence à un couple est impliquée par les trois classes de mots mentionnées ci-dessus, notamment, par la quasi-totalité des indéfinis pourvus de ce suffixe, par l’expression du comparatif (d’égalité ou d’inégalité), par les dérivés lexicaux, impliquant la référence à un élément considéré comme modèle exemplaire. Ainsi en latin matertera « tante maternelle » par rapport à la « mère », en indien ancien aśvataráḥ « mulet » par rapport à aśvaḥ « cheval », vatsataráḥ « jeune taureau » par rapport à vatsáḥ « veau ». En effet, ces types de dérivés se rapprochent de la comparaison, parce que le lexème de base est une sorte de modèle exemplaire des qualités partagées, parfois seulement partiellement, par le référent. En conclusion, les trois types de formation en -tero-, qui sont, en partie, grammaticalisés, et, en partie, lexicalisés, pivotent sur la mise en relation de deux éléments pourvus de qualités susceptibles d’être mises en comparaison. A bon droit, donc, on a reconnu au suffixe -terola propriété fondamentale de mettre en contraste deux éléments couplés, qui partagent une qualité à des degrés tantôt égaux tantôt différents. La fonction ‘différentielle’ intrinsèque à -teroa été mise en valeur par Benveniste-: « Tous ces faits, qui sont du vocabulaire et de la ‘parole’, montrent que -terocomporte une valeur différentielle. Cette valeur se manifeste à l’occasion d’emplois, où un 356 Anna Orlandini & Paolo Poccetti <?page no="357"?> 19 Benveniste (1948, 119). 20 Benveniste (1948, 143). 21 Rix 1976, 179-181. 22 Daniel (2013). terme est caractérisé : il est par là posé comme distinct d’un autre terme lexicalement différent avec lequel il n’a aucune relation syntaxique. Les deux termes entrent dans deux phrases distinctes, du type «-A est X-teros, B est Z-» 19 . Ainsi -terose distingue de l’autre suffixe employé dans les langues indo-europé‐ ennes pour l’expression du comparatif, notamment *i̭ōs/ -i̭es-, auquel Benveniste attribue une « valeur dimensionnelle », selon laquelle « le terme comparé est l’objet d’une évaluation dimensionnelle à partir d’une norme typique […] c’est bien d’une ‘qualité dimensionnelle’ que le comparé est pourvu par rapport à un comparant fixe-» 20 . Une telle distinction servait à Benveniste pour expliquer l’emploi de *i̭ōs/ -i̭esdans les comparaisons qui n’admettent pas d’échelle graduée, comme les compa‐ ratifs où le comparant représente un modèle exemplaire, conventionnellement identifié comme le degré le plus élevé dans l’échelle du paramètre de la compa‐ raison. Il s’agit des types de comparatifs appelés « à parangon », exemplifiés par la tournure « plus/ aussi doux que le miel » (ex. lat. melle dulcior, gr. μέλιτος γλυκίων, a.ind. ghṛtāt svādīyaḥ). Ces expressions ne formulent pas une véritable comparaison, mais fonctionnent comme des élatifs. Le même critère s’applique à certaines formes de comparatifs latins en -ior, bâtis sur des adverbes en -tero- (extrā vs. intrā ; ultrā vs. citrā) indiquant la position par rapport à un point de repère. En fait, les dérivés exter-ior vs. inter-ior, indiquant l’intérieur ou l’extérieur, et ulter-ior vs. citer-ior, indiquant l’au-delà ou l’en deçà d’une limite ou d’un bornage, expriment une qualité qui ne peut pas être nuancée, et qui ne permet que son inclusion ou son exclusion totale. En ce cas, -ior marque la qualité possédée au degré maximum et, par conséquent, n’implique aucune comparaison d’inégalité. La fonction ‘différentielle’ de -terose retrouve dans la morphologie des pro‐ noms possessifs du pluriel en latin (noster, uester) et en grec (ἡμέτερος, ὑμέτερος, d’après lesquels le grec a bâti σφέτερος « leur » et σφωΐτερος « leur entre deux ») 21 . Dans ce cas, le suffixe -teroest lié à la valeur de base des pronoms de 1 ère et 2 nde personne du pluriel, qui se prévoient l’inclusion ou l’exclusion d’autres individus au-dehors de la sphère respectivement du locuteur et de l’interlocuteur. Dans certaines langues, une telle distinction à abouti à une différence formelle entre les pronoms pour exprimer l’un ou l’autre de deux valeurs 22 . La négation comme moteur de l’évolution linguistique 357 <?page no="358"?> 23 Wackernagel (2009, 102-105). La fonction de -terode focaliser un ensemble de deux éléments est manifestée en grec par l’ordinal δεύτερος « deuxième » et le pronom ἀμφότερος « les deux-», où -terorenchérit sur l’expression de la dualité, liée aux bases lexicales elles-mêmes, notamment δύω et ἄμφω. En outre, la forme renforcée ἀμφότερος a remplacé ἄμφω, apparenté à ambo du latin, en conséquence de l’affaiblissement sémantique de l’ancien morphème du duel. En conclusion, le rôle sémantique du suffixe -terodans certains domaines de la grammaire et du lexique doit être reconduit au principe cognitif sous-jacent la catégorie du duel, en tant que nombre grammatical, distinct du pluriel. Au niveau perceptif, deux éléments qui sont mis en relation, forment un ensemble fermé, qui se situe hors du calcule et de la numérotation, alors que le pluriel est cumulatif et se projette vers une série ouverte. Au niveau de l’expression, le nombre duel, en tant que catégorie grammaticale autonome, est très répandu dans les langues du monde, mais il se manifeste d’une manière différente dans les langues, se distribuant dans la flexion nominale et/ ou pronominale, la conjugaison verbale, le lexique et la syntaxe dans la mesure où l’accord grammatical est impliqué. Toutefois, plusieurs langues montrent, au fil du temps, une tendance à la disparition du nombre duel au profit du pluriel. Comme l’avait déjà souligné Wackernagel, presque toutes les langues indo-européennes, qui possédaient le duel à l’état le plus ancien, l’ont perdu dans leurs parcours évolutifs 23 . À une telle tendance n’a pas échappé le suffixe -tero-, qui, d’une part, a été totalement remplacé en tant qu’outil grammatical, par ex., dans le comparatif, comme nous le montrent, parmi les langues modernes, le grec, l’irlandais, le hindi par rapport à leurs ancêtres qui le possédaient, et, de l’autre part, il a perdu le renvoi aux éléments d’un couple pour s’ouvrir vers la pluralité. Les mêmes parcours ont été suivis par les indéfinis en -terodu grec et du latin. Ainsi certaines formes telles que uter, uterque du latin ont totalement disparu des langues romanes, tout comme le grec ancien πότερος « quel entre deux » inconnu en grec moderne. D’autres, comme neuter du latin et son correspondant grec οὐδέτερος « ni l’un…ni l’autre », ont acquis le sens lexical de l’adjectif «-neutre-», qui dans les langues romanes, n’exclut plus deux éléments à la fois, mais renvoie, d’une manière floue, à quelque chose génériquement dépourvue de qualités. De la même manière, alter finit par renvoyer génériquement à «-autrui-», «-l’autre-» (fr. «-comme disait l’autre-»). Les traces de ces tendances évolutives se retrouvent déjà dans langues anciennes. Ainsi, par ex., en grec, l’on constate des flottements, signalés par 358 Anna Orlandini & Paolo Poccetti <?page no="359"?> 24 Wackernagel (2009, 523-526; 737). 25 Blass-Debrunner (1961, 35) ; Mayser (1933, II,2, ). 26 Bader (1985). Une raison phonétique, notamment la dissimilation, est, vraisemblable‐ ment, le responsable de l’absence de dérivés en -teroà partir de *ar-. Ainsi, on laissera de coté la question de savoir si *alet *arne sont que des variantes d’une même racine, différemment distribuées. Wackernagel, entre πότερος «-lequel-? (entre deux)-» et πoίος «-lequel-? (entre plusieurs) », le remplacement de ἑκάτερος « chacun (entre deux) » par ἕκαστος « chacun (parmi plusieurs) », de οὐδέτερος et μηδέτερος par οὐδάλλος et οὐδαμός 24 . Tous les trois, notamment πότερος, ἑκάτερος, οὐδέτερος/ μηδέτερος, sont presque inconnu du grec tardif, spécialement de la langue du Nouveau Testament et des papyri  25 . 3.3 Les avatars du pronom «-l’autre-» et les enjeux de la négation Dans plusieurs langues indo-européennes, le suffixe -tero-, bâti sur des racines différentes, sert à marquer les pronoms indiquant « l’autre » : ainsi alter du latin appartient à la racine *al-, d’où sont issus lat. alius, gr. ἄλλoς, gaul. allo- (< *al-i̭o-), alors que de la racine *ansont issus l’a.ind. ántara- « ce qui se trouve à l’intérieur » (lat. inter-ior) et anyá- « autre », mais aussi les formes du germanique, telles que le got. anϸar, le vieux all. andara, l’angl. other, le lit. añ(a)tras. En revanche, en grec ἕτερος / ἅτερος est composé de *sṃ- « un », en parallèle à l’a.ind. ekatará- « un des deux », composé de eka- « un », se distinguant de ekatamá- « un parmi plusieurs ». La racine *ar-, qui, toutefois, n’a pas engendré des dérivés en -tero- 26 , a aussi contribué à la formation d’expressions indiquant « autre ». Dans ce panorama, l’on peut retracer des parcours communs aux dérivés en -tero-, qui, dans l’évolution de chaque langue, ont perdu le renvoi marqué et contrastif à l’un de deux éléments d’un ensemble fermé, au profit d’un renvoi non marquée à une multiplicité indistincte. Ce développement se retrouve dans alter du latin, mais aussi dans son correspondant en osque, dans ἕτερος / ἅτερος du grec, ainsi que dans les langues germaniques. L’histoire de alter en latin et celle ἕτερος / ἅτερος en grec, dans une certaine mesure, se déroulent en parallèle. En principe, les deux pronoms signalent « l’autre », « l’un de deux », par le renvoi à l’un de deux éléments d’un ensemble fermé, avec une implicature conventionnelle, véhiculée par le suffixe -tero-, d’exclusion d’un troisième élément. Cette valeur appartient à ἕτερος, en grec homérique, pour indiquer des parties binaires du corps : La négation comme moteur de l’évolution linguistique 359 <?page no="360"?> (10a) φολκὸς-ἔην,-χωλὸς-δ’-ἕτερον-πόδα (Il. 2, 217)-«-il était camus et boiteux d’un pied-» (10b) ὡς δ’ ὅτε ποιμὴν ῥεῖα φέρει πόκον ἄρσενος οἰὸς χειρὶ λαβὼν ἑτέρῃ (Il. 12,451) « Comme un berger qui porte sans effort d’une seule main a toison d’un bélier et il la tient d’une seule main-» ainsi qu’alter, en latin classique, par ex.,-: (10c) dextra tenet calamum, strictum tenet altera ferrum (Ovid. epist. 11,3) « À la main droite tient une plume ; l’autre tient un fer nu-». (10d) humili et corpore exiguo, claudus altero pede (Nep. Ages. 8,1) «-il était de petite taille, chétive apparence et boiteux d’un pied-». De même, par rapport à l’un des membres d’un couple contextuellement déterminé : (11a) καὶ Bάττος μὲν ὁ ἕτερος τῶν στρατηγῶν δύω γὰρ ἦσαν ἐν τῇ μάχῃ oἱ παρόντες (Tuc. 4,43,1) «-et Battus, l’un des chefs de l’armée (en effet, ils étaient deux présents sur le champ de bataille) » (11b) Mil. ei duae puellae sunt meretrices servulae/ sorores : earum hic alteram efflictim perit (Plaut. Poen. 1094 sq.) « Il a chez lui deux jeunes esclaves, deux courtisanes : il est amoureux fou de l’une d’entre elles-». Toutefois, en diachronie, ἕτερος et alter développent un parcours d’affaiblisse‐ ment de la valeur exprimée par le suffixe -tero-, réalisant progressivement une ouverture vers le pluriel, ce qui comporte la disparition de l’implicature conventionnelle d’exclusion d’un troisième élément. Ce parcours s’insère dans la tendance plus générale, déjà mentionnée, à effacer le duel dans la flexion des noms et dans la conjugaison des verbes dans toutes les langues indo-europé‐ ennes. Autrement dit, la disparition du duel au niveau lexical suit, d’une manière différenciée d’une langue à l’autre, la disparition du nombre duel en tant que catégorie grammaticale. 3.3.1 L’affaiblissement fonctionnel de -tero- L’affaiblissement de la fonction de -teroest centrale dans l’évolution de ἕτερος et de alter, qui dépassent la relation entre deux éléments, permettant un renvoi à la multiplicité. On peut retracer le moteur de ce développement dans deux types de contextes. Le premier est l’effet de sens produit par la numérotation, c’est-à-dire lorsque « l’autre » entre deux est conçu comme « ce qui suit le premier-», le «-second-». 360 Anna Orlandini & Paolo Poccetti <?page no="361"?> 27 LEIA, A-31, s.v. aile-; Bader (1986,61). 28 PY Ma 365: Bartoněk (2003, 443). L’emploi du terme utilisé pour « l’autre » comme numéral ordinal avec le sens de « second » est très répandu : parmi les langues germaniques, il est connu du gotique anϸar ainsi que du vieux anglais ōϸer (=other). Les langues celtiques ont grammaticalisé *al-i̭ocomme ordinal courant pour « second » : vieux irl. aile, gall. ail-; gaul. allo- 27 . Quant à ἕτερος / ἅτερος, les traces du glissement vers « le suivant, le second » remontent à l’époque du grec mycénien, comme le prouve l’expression a 2 -te-ro we-to «-l’année suivante-» 28 . Chez Homère, l’on trouve ἕτερος, employé dans une liste ouverte, avec la valeur de «-deuxième-», dans une série où il précède les ordinaux «-troisième-» et «-quatrième-»-: (12) ἀμφίπολοι-δ᾽-ἄρα-τέως-μὲν-ἐνὶ-μεγάροισι-πένοντο τέσσαρες,-[….]. τάων-ἡ-μὲν-ἔβαλλε-θρόνοις-ἔνι-ῥήγεα-καλὰ πορφύρεα-καθύπερθ᾽,-ὑπένερθε-δὲ-λῖθ᾽-ὑπέβαλλεν: ἡ-δ᾽-ἑτέρη-προπάροιθε-θρόνων-ἐτίταινε-τραπέζας ἀργυρέας,-ἐπὶ-δέ-σφι-τίθει-χρύσεια-κάνεια: ἡ-δὲ-τρίτη-κρητῆρι-μελίφρονα-οἶνον-ἐκίρνα ἡδὺν-ἐν-ἀργυρέῳ,-νέμε-δὲ-χρύσεια-κύπελλα. ἡ-δὲ-τετάρτη-ὕδωρ-ἐφόρει-καὶ-πῦρ-ἀνέκαιε πολλὸν-ὑπὸ-τρίποδι-μεγάλῳ: -ἰαίνετο-δ᾽-ὕδωρ (Od.10, 349-359) - « Quatre servantes travaillaient dans ce palais avec zèle […] L’une d’elles étend sur des sièges de superbes tapis de pourpre et les recouvre encore d’un riche tissu de lin ; une autre dresse devant les sièges des tables d’argent sur lesquelles elle place des corbeilles d’or ; la troisième mêle dans un cratère d’argent un vin suave aussi doux que le miel et distribue des coupes d’or-; la quatrième enfin allume le bois desséché sous le large trépied et fait tiédir de l’eau-». La numérotation est impliquée par l’emploi de ἕτερος en corrélation avec ἄλλ oς répété plusieurs fois, comme dans le contexte suivant, où ἕτερος occupe la troisième position dans la liste-: (13) ἄλλῳ-μὲν-γὰρ-ἔδωκε-θεὸς-πολεμήϊα-ἔργα, ἄλλῳ-δ᾽-ὀρχηστύν,-ἑτέρῳ-κίθαριν-καὶ-ἀοιδήν, ἄλλῳ-δ᾽-ἐν-στήθεσσι-τιθεῖ-νόον-εὐρύοπα-Ζεὺς ἐσθλόν (Il 13, 730-736) - «-A l’un la divinité octroie l’œuvre de guerre, à tel autre la danse, à tel autre encore la cithare et le chant-; à tel enfin Zeus à la grande voix met dans la poitrine un bon esprit-». La négation comme moteur de l’évolution linguistique 361 <?page no="362"?> 29 Cf. Bortolussi (2015, 173-182)-; Bertocchi & Orlandini (à paraître). De même pour alter, pour lequel la valeur numérative de « second » était présente déjà chez les Comiques-: (14a) SY. Lesbonicum hic adulescentem quaero, in his regionibus ubi habitet, et item alterum ad istanc capitis albitudinem (Plaut. Trin. 874) « Je cherche ici un jeune homme appelé Lesbonicus, je veux savoir où il habite, dans ce quartier, et aussi un autre homme à la tête aussi blanche que toi-». (14b) CHR. ducentos nummos Philippos militi quos dare se promisit dabit. nunc alteris etiam ducentis usus est (Plaut. Bacch. 969-971) «-Il va maintenant donner au soldat mes deux cents philippes qu’il a promis. Et nous avons encore besoin de cents autres-». Tout comme ἕτερος en (12), alter reçoit la valeur de « deuxième » quand il précède les numéraux ordinaux «-troisième-» et «-quatrième-» : (15a) quidam […] Gallus […] ab latere dextro traiectus exanimatusque concidit [….]. Exanimato altero successit tertius et tertio quartus (Caes. Gall. 7,25,3) « il y avait un Gaulois […] : un trait de scorpion lui perce le flanc droit ; il tombe mort. Un deuxième de ses plus proches voisins est tué de la même manière ; un troisième lui succède-; à celui un quatrième-». Par là, alter est employé comme véritable numéral ordinal dans le calcul-: (15b) anno trecentesimo altero quam condita Roma erat (Liv. 3,33,1) « 301 ans après la fondation de Rome (= pendant la trois-cents deuxième année après ….)-» (15c) post quadragesimum et alterum mensem (Colum.12,41) «-41 mois après = pendant le quarante-deuxième mois après ….-». La valeur de « ce qui suit, le second » favorise le glissement de la valeur de «-l’autre-», qui est propre de alter, à «-un autre-», qui est propre de alius  29 : (16a) ut iterum periremus et alterum excidium patriae uideremus (Liv. 28,39,7) « pour périr à nouveau et voir détruire une seconde fois (une autre fois) notre patrie » (16b) rursum dormiuit et uidit alterum somnium (Vulg. gen. 41,5) « Il s’endormit de nouveau et il fit un autre rêve-». En latin tardif, cet emploi devient plus fréquent. Tout comme pour ἕτερος qui remplace ἄλλoς, l’emploi de alter dans une liste ouverte se superpose à celui de alius renvoyant à plusieurs éléments. Ainsi dans 362 Anna Orlandini & Paolo Poccetti <?page no="363"?> une liste de huit unités prises en compte, alter…alter alterne dans la corrélation avec pars…pars, haec…illa-: (17a) Nam timor unus erat, facies non una timoris / pars laniat crines, pars sine mente sedet / altera maesta silet, frustra vocat / haec queritur, stupet haec; haec manet, illa fugit (Ovid. ars 121-124) «-L’épouvante est partout la même, mais les symptômes en sont différentes. Les unes s’arrachent les cheveux, les autres tombent sans connaissance, celle-ci pleure et se tait, celle-là appelle en vain sa mère, d’autres restent plongées dans la stupeur. L’une demeure immobile, l’autre fuit-». De son coté, alius aussi peut être répété dans une liste ouverte-: (17b) Aliud genus est narrandi, aliud persuadendi, aliud docendi (Cic. orat. 180) «-une chose c’est le genre de la narration, une autre c’est celui de la persuasion, et encore une autre celui de l’enseignement-» (17c) milites Romani, perculsi tumultu insolito, arma capere alii, alii se abdere, pars territos confirmare (Sall. Jug. 38,5) « les soldats romains sont décontenancés par cette attaque soudaine, les uns prennent les armes, les autres se cachent, une partie rassure leurs camarades effrayés-». Inversement, en latin tardif, alius peut remplacer alter dans la corrélation de deux éléments, ce qui révèle la tendance à effacer la distinction entre les deux formes pronominales : (17d) gnatos duos creauit. Horum alterum in terra linquit, alium sub terra locat (CIL I 2 1211 = CLE 52,6) « Il a eu deux fils. Dont l’un est resté sur la terre et l’autre est sous la terre-». Par un échange analogue et inversé, le sens de « différent » propre à alius, peut être partagé par alter dans les tournures exprimant une idée de séparation : alter quam, alius quam, alius ac : (18a) quamquam aliud dicit ac sentit, non aliud tamen simulat (Quint. inst. 9,2,45) « bien qu’il présente un sens et en renferme un autre, ce dernier sens est moins déguisé-». (18b) SI. nunc mihi certum est alio pacto Pseudolo insidias dare quam in aliis comoediis fit (Plaut. Pseud. 1239) «-Et maintenant, j’ai l’intention de guetter Pseudolus, mais autrement que dans les comédies-». La négation comme moteur de l’évolution linguistique 363 <?page no="364"?> 30 D’après la définition ex pluribus dans ThLL 1736, 4 ss. (18c) idem schiston et alterius generis quam haematiten tradit, quem vocat anthraciten (Plin. nat. hist. 36,148,4) « Lui-même rapporte encore qu’il existe un schiste d’un autre genre que l’hématite qu’il appelle anthracite-». La présence de la négation est l’autre phénomène qui a fait glisser ἕτερος et alter vers la multiplicité d’éléments, voire la totalité. Sous négation, ἕτερος et alter ne renvoient plus à l’autre, mais à nul autre, à la totalité négative 30 , par ex. : (19a) neque enim spes altera restat (Ovid. fast. 3, 625) «-il ne lui reste aucun autre espoir-». (19b) quodque vident sternunt nec metus alter inest (Ovid. fast. 2, 224) « « ils détruisent tout ce qu’ils voient, et il n’y a plus aucune autre crainte-». Dans ces contextes négatifs, alter prend le sens de « nul autre » « aucun autre », indiquant la totalité négative, l’exclusion de toute autre possibilité, qui est le sens de nemo (~∃x). Ce sens est confirmé par l’occurrence des indéfinis ullus, quis, quisquam-en présence d’une négation : (20b) nec ulla uirtus altera melior (Sen. clem. 1,5,3) «-nulle autre vertu n’est meilleure-» (20c) ne cuius alterius opere egeret (Quint. inst. 12,11,21) « pour ne pas avoir besoin de l’aide d’aucun autre-» ou des indéfinis à négation incorporée nemo et nullus, qui à eux seuls engendrent une négation de phrase: (21a) Aristoteles tradit et simul plures cerni, nemini compertum alteri, quod equiden sciam (Plin. nat. hist. 2,91) « Aristote rapporte qu’on en voit souvent plusieurs à la fois, observation que personne autre n’a faite, à ma connaissance-». (21b) nullius alterius fiduciam ducis habebant (Iust. 6,4,10) «-ils ne faisaient confiance à aucun autre commandant-». De cette manière, alter se superpose à alius qui sous négation prend la même valeur de «-nul autre-»-: (22a) alius bonus nullus erit (Cat. agr. 6, 2) «-nul autre ne sera bon-». 364 Anna Orlandini & Paolo Poccetti <?page no="365"?> 31 Selon Serbat (1984,525) alius exprime une pure opération intellectuelle de mise en relation, le fait de se présenter comme «-différent de-», «-autre-». (22b) quando aliud nihil auxilii habeat (Liv. 2,55,3) « comme il n’a nul autre aide ».- De même, en grec, ἕτερος, s’accompagnant de la particule négative, signifie « aucun autre », et il se retrouve souvent dans des expressions comparatives de sens élatif, comme, par ex., en français, « personne n’est aussi courageux que toi-» ou «-personne sauf toi-n’a autant de courage »-: (23a) τοῦτ’ ἔστιν οὐχ ἕτερον ἢ ἡ τῆς ψυχῆς ὄντως καὶ πάντως ἀτιμία (Plat. leg. 727d « Et encore, préférer la beauté à la vertu, ce n’est là rien d’autre que déshonneur réel et total de l’âme ». (23b) ἔλπομαι οὐχ ἕτερον τόδε τλήμεναι Αἰγιαλήων ἠὲ σέ, (Theocr. 25, 174) «-Je ne pense pas qu’un autre que toi, eût accompli un exploit pareil-». Ainsi dans la tournure, assez fréquente en grec, οἵoς οὐχ ἕτερος « comme aucun autre-» indiquant ce qui n’admet pas de comparaison-: (24a) παθήματά τε ξυνηνέχθη γενέσθαι ἐν αὐτῷ τῇ Ἑλλάδι οἷα οὐχ ἕτερα ἐν ἴσῳ χρόνῳ (Tuc. 1,23) « au cours de cette guerre des malheurs fondirent sur la Grèce en une proportion jusque-là inconnue-». (24b) καὶ ἦν καρτερὰ καὶ οἵα οὐχ ἑτέρα τῶν προτέρων (Tuc. 7,70) « (Le combat) fut acharné et sans analogie avec ceux qui l’avaient précédé-». Dans ces cas, la négation, ne concernant plus le couple fermé de deux éléments, entraîne ἕτερος et alter respectivement vers le domaine sémantique de ἄλλ oς et de alius, qui est un domaine relationnel 31 et non-compositionnel. Il est bien connu que les concepts d’identité et de différence échappent à l’analyse par les opérateurs logiques. Selon le principe de « compositionalité » de la logique classique, les propositions doivent toutes être vraies pour satisfaire le vrai de l’énoncé. En revanche, pour expliquer le fonctionnement sémantique des relations d’identité et de différence, il faut envisager une relation symétrique à l’intérieur d’une seule proposition. Autrement dit, l’énoncé : Pierre est différent de Paul ne ressort pas du vrai de deux propositions : Pierre est différent et Paul est différent, mais de la relation symétrique : Pierre est différent de Paul = Paul est différent de Pierre= Pierre et Paul sont différents entre eux. Selon Platon (Sophist. 257B), la négation ne relève pas des concepts de contradictoire, ni de celui de contraire (τὸ ἐνάντιον), mais du concept d’alterité (τὸ ἕτερον), La négation comme moteur de l’évolution linguistique 365 <?page no="366"?> 32 Schol. Hor. carm. 4,10,6 immutationem formae ..dolebit. 33 Cf. Bader (1985). de différence. Par ex., l’expression pas grand (μὴ μέγα) ne se réfère pas à quelque chose qui est opposé à grand (= petit), mais à quelque chose qui est différent de grand. Cette interprétation concerne surtout la négation non véridictoire des énoncés non assertifs, où le concept de discordance est crucial pour analyser la négation pragmatique, et plus en général, la négation des structures non-compositionnelles. Alter et ἕτερος peuvent ainsi exprimer le sens de « différent », qui est la valeur relationnelle propre de alius-: (25a) quotiens te speculo uideris alterum 32 (Hor. carm. 4,10,6) «-que de fois tu diras en te voyant tout autre dans le miroir ». Le sens de « différent » est attribuable ἕτερος quand il s’oppose à « le même »-: (25b) ὡς ἔσθ’ οὗτος ὁ σκοπὸς οὐχ ἕτερος, ἀλλ’ ὁ αὐτός (Plat. leg. 693 c) « ce ne sont pas là des buts différents, c’est le même but ». Pour exprimer la différence, ἕτερος se construit de la même manière que ἄλλoς avec le génitif à valeur séparative-: (25c) Ἀγαθὸν δὲ κοινῶς μὲν τὸ τὶ ὄφελος, ἰδίως δ’ ἤτοι ταὐτὸν ἢ οὐχ ἕτερον ὠφελείας (Diog. Laert. 7,94) «-le bien, pris d’une manière générale, est l’utile, avec cette distinction plus particulière : d’une part l’utile, de l’autre ce qui n’est pas contraire à l’utilité » (25d) Πότερον οὖν, ὦ Ἱππία, τοὺς θεοὺς ἡγῇ τὰ δίκαια νομοθετεῖν ἢ ἄλλα τῶν δικαίων; (Xen. Mem. 4,4,25) « Crois-tu donc, Hippias, que les dieux établissent des lois justes, ou qu’ils puissent en établir de contraires à la justice ? -».- Ainsi alter et ἕτερος, à travers les parcours d’affaiblissent de la fonction originaire de -teroqu’on vient d’esquisser, ont fini par envahir le domaine respectivement de alius et ἄλλoς, les deux issus de *al-i̭o-, dont le sémantisme véhicule le sens de séparation, de différence, de privation, d’éloignement 33 . À partir de la valeur séparative, ἕτερος et alter peuvent rejoindre la valeur exceptive, commune à ἄλλoς et à alius, aboutissant au sens de « excepté, sauf », qui se réalise sous négation. Ainsi les tournures grecques : οὐδὲν ἕτερον ἤ / εἰ μή et οὐδέν’ ἄλλο ἤ / εἰ μή résultent parallèles à celles latines : nihil aliud quam/ nisi, 366 Anna Orlandini & Paolo Poccetti <?page no="367"?> 34 Bader (1985, 59-61). 35 Untermann (2000, 81), s.v. allo. 36 LEIA s.v. uile U 18. nihil alterum quam/ nisi avec la valeur exceptive et focalisante de «-rien d’autre sauf, excepté-»-: (26a) ἔχοντες οὐδὲν ἕτερον τῶν ὅπλων ἢ μάχαιραν (Plut. Mor. 696c) « n’ayant aucune autre arme, sauf l’épée-» (26b) οὐδέν’ ἄλλον ἢ τὸν θύλακον (Aristoph. Eccl. 382) «-rien d’autre qu’un sac-» (26c) οὐδὲν’εἰπεῖν - ἕτερον εἰ μὴ νὴ Δία νεανίσκου τινὸς ἦν ἐκφορὰ (Plut. Sol. 6,2,4) « Rien d’autre - s’exclama-t-il - sauf les funérailles d’un jeune homme-» (26d) nullius alterius consilio quam Scipionis (Liv. 30, 23, 4) « selon l’avis de nul autre, sauf Scipion-». 3.3.2 « Autre » marqueur de différence, d’exclusion et de totalité négative Les langues modernes ont développé les valeurs d’exclusion véhiculées par alter et ἕτερος, sous négation. Ainsi, dans le standard du comparatif, «-autre-», s’accompagnant d’un élément négatif, sert à marquer l’exclusion de tout terme de comparaison, ce qui réalise le sens de superlatif ou d’élatif comme dans les expressions du français « il est intelligent comme nul autre » ou de l’italien «-è intelligente come nessun altro » ou « è intelligente quant’altri mai ». Un tel emploi de alter ainsi que de alius est exclu du latin classique qui exprime l’élatif par d’autres tournures, telles que, par exemple, quam qui maxime. Dans ces contextes, « autre » fonctionne comme renforcement de l’élément négatif pour nier la totalité, à l’exclusion d’un seul élément. En fait, les expressions de la totalité et celles de l’altérité et de la différence se reconduisent à la même racine *al-, se distribuant avec des morphèmes différents parmi les langues indo-européennes : les dérivés en -teroet -i̭oont fourni les expressions de l’altérité et de la différence (ex. lat. alter, alius, osque alttreí-; vieux irl. aile), alors que le dérivé *al-no  34 est à la base de l’indéfini « tout » des langues germaniques (ex. got. alls, angl. all, all. alle) ainsi que de l’osque (allo) 35 . Des morphèmes dérivationnels différents sur la même base distinguent les indéfinis pour « tout » dans les langues celtiques, tels que uile < *al-i̭ode l’irlandais ancien et oll < *al-nodu brittonique 36 . En grec ancien, ἄλλoς, s’accompagnant de l’article, véhicule l’idée de totalité prenant le sens de « tout ce qui reste » (par ex. τὸν-ἄλλον-χρόνον «-durant tout le temps qui reste-») et au pluriel οἱ-ἄλλοι signifie normalement «-tous les autres-», lat. ceteri. La négation comme moteur de l’évolution linguistique 367 <?page no="368"?> 37 A remarquer la définition dans le ThLL. s.v. alius I 1625, 75 : liberiore ratione interdum iis substantiuis additur, quae neque ut synonyma neque ut notiones generales aliis opponuntur; 38 Löfstedt (1933, 189). Le glissement sémantique de l’idée de différence et d’altérité à celle de la totalité se retrouve synchroniquement dans certains emplois de ἄλλoς en grec et de alius en latin. Parfois ces indéfinis n’expriment pas une véritable altérité, mais servent à rajouter d’éléments qui s’intègrent réciproquement ou complètent un ensemble, atteignant la totalité 37 -: (27a) αὐτὰρ-Ὀδυσσεὺς τέρπετ’-ἐνὶ-φρεσὶν-ᾗσιν-ἀκούων-ἠδὲ-καὶ-ἄλλοι Φαίηκες δολιχήρετμοι, ναυσικλυτοὶ ἄνδρες (Od. 8,367-369) « Ulysse les écoute avec ravissement et aux accents du chanteur se réjouissent les Phéaciens, navigateurs habiles, dont les longues rames sillonnent la mer-» (27b) παρεκάλεσαν-τοὺς-ἐν-τῷ-καταλόγῳ-ὁπλίτας-καὶ τοὺς-ἄλλους-ἱππέας (Xen. Hell 2,4,9) «-ils convoquent les hoplites dont les noms sont sur la liste, et tous les cavaliers-» (27c) hinc nubila, tonitrua et alia fulmina (Plin. nat. hist. 2,102) « de là les nuages, les tonnerres et les éclairs-» (27d) eo missa plaustra iumentaque alia (Liv. 4,41,8) «-on y envoya de la ville des chariots et des chevaux-». Dans ces contextes, les pronoms alius et ἄλλoς ont été interprétés comme ayant un emploi pléonastique 38 , parce que leur suppression ne modifie pas le sens. Comme l’on peut le vérifier dans les exemples donnés ci-dessus, les traductions modernes de ces passages ne comportent pas le mot pour « autre », remplacé tout simplement par l’article défini. Toutefois, il nous semble que ces pronoms véhiculent d’une manière emphatique une autre nuance : le sens résomptif de «-tous-». En revanche, dans les langues romanes c’est l’indéfini «-tout-», s’accompag‐ nant de « autre », qui peut être omis sans nuire au sens, comme dans la phrase : « parmi eux, une partie a accepté, (tous) les autres ont refusé ». Mais, en union à un indéfini négatif, «-autre-» peut être supprimé sans aucune variation sémantique. Cela se réalise dans la comparaison, où « nul autre » figure dans le rôle du comparé ou du comparant (standard) : dans ce cas « nul autre » ne se distingue de « personne » que par une nuance plus emphatique. En fait, ces occurrences de « nul autre » / « personne » dans le comparatif expriment 368 Anna Orlandini & Paolo Poccetti <?page no="369"?> 39 Comme l’on le verra en détail au point 4. 40 Cf. au point 1, le classement du sujet (SU) selon le rôle syntactico-sémantique en relation à l’échelle hiérarchique de Keenan et Comrie (1977), fig.1. des superlatifs ou élatifs 39 . Le même sens de superlatif ou élatif est obtenu si « nul autre » / « personne » constituent le comparé ou le comparant (standard), indifféremment dans une comparaison d’égalité ou d’inégalité, comme dans les expressions du français-: (28a) « nul autre n’est plus intelligent que lui » / « il est intelligent comme nul autre » ou de l’italien : (28b) «-nessun altro è intelligente come lui-» / «-è intelligente come nessun altro-». Dans ces contextes, qui expriment d’élatif, le mot pour « autre » ne figure pas en latin classique. Cette langue distingue, selon la place d’occurrence, les pronoms/ adjectifs négatifs (nemo, nihil, nullus), possibles seulement dans le comparé, où le rôle de sujet leur assure une force quantificationnelle négative pleine 40 , et les pronoms/ adjectifs à polarité négative (quisquam, ullus), qui sont les seuls possibles dans le comparant, qui n’est qu’orienté négativement et jamais négatif. En revanche, en grec ancien ἄλλoς présente un emploi flottant sous négation dans les comparatifs du type (28a,b) exprimant un élatif. Ce pronom peut s’accompagner d’un indéfini négatif-: (28c) Ζεῦ πάτερ οὔ τις σεῖο θεῶν ὀλοώτερος ἄλλος (Il. 3,365) « Ah, Zeus père, il n’y a pas de dieu plus exécrable que toi-» ou bien remplacer l’indéfini-: (28d) ἐπεὶ οὐ σέο κύντερον ἄλλο (Il. 8,483). « parce que personne n’est plus chien que toi-» ou être omis-sans modifier le sens : (28e) οὐ μὲν γάρ-τί-πού-ἐστιν-ὀϊζυρώτερον-ἀνδρὸς (Il. 17,446) «-rien n’est plus misérable que l’homme-». La négation comme moteur de l’évolution linguistique 369 <?page no="370"?> 41 Cf. aussi au point 1, le classement du comparant (OCOMP) selon le rôle syntactico-sé‐ mantique en relation à l’échelle hiérarchique de Keenan et Comrie (1977), fig.1. 42 Denniston (1970, 24-26). Les langues romanes ne distinguent pas formellement le pronom négatif selon la place du comparé et celle du comparant (mais du point de vue sémantique, au comparé il s’agit d’un pronom négatif plein, alors qu’au comparant il s’agit d’un « N-word » 41 ). Dans ces langues, le terme pour « autre » sous négation, peut être omis sans aucun changement de sens. Toutefois, le mot pour « autre » demeure dans l’expression de l’italien ayant le même sens élatif : « è intelligente quant’altri mai » : dans ce cas, le terme pour « autre » ne peut pas être supprimé, parce qu’il a un rôle pronominal et non adjectival. Ainsi, en parallèle au grec (28d), dans certains contextes de l’oralité, l’élément négatif est supprimé, mais il transmet à « autre » la fonction négative avec force focalisante et emphatique. C’est le cas, par exemple, des occurrences de « altro » dans les réponses, témoignées par quelques variétés de l’italien : ce mot devient un adverbe d’intensité, il remplace « rien d’autre » et exprime la réponse négative-: «-non, pas du tout-», par ex.-: (29a) Q. «-desidera qualcos’altro-? -» R. «-Altro-» (=« Non, rien d’autre.-») Ainsi, en grec moderne, l’expression καθε αλλo réalise une réponse négative polémique avec le sens de-: « Non,-au contraire ». En revanche, l’omission de l’élément négatif dans la formule restrictive «-rien d’autre que cela » aboutit à un sens affirmatif emphatique et focalisant. C’est le cas de l’expression italienne «-altro che-», qui dans les réponses véhicule le sens de «-tout à fait, bien sûr-»-: (29b) Q. « sei sicuro di averlo visto ieri ? » R. « Altro che ! » (= Oui, bien sur ! Et comment ! ). En grec ancien, ἄλλoς et ἕτερος, employés pour marquer la différence, présen‐ tent la même construction syntaxique que le comparatif d’inégalité, notamment le génitif et la particule ἤ qui sert aussi pour la coordination disjonctive 42 : les deux constructions partagent la valeur de séparation ou d’éloignement des deux éléments comparés. De même, dans la comparaison, la fonction séparative est exprimée souvent par la construction syntaxique du comparant de plusieurs langues, comme, par ex., par certains emplois de l’ablatif en latin ou par l’emploi de la négation en indien ancien et dans les langues slaves. 370 Anna Orlandini & Paolo Poccetti <?page no="371"?> 43 Biraud (1983, 179). 44 Viti (2007, 33). En revanche, en latin pour la même fonction, alius (ainsi que alter dans son évolution) se joint aux particules de la coordination connective, ac, atque, et (ex. alius ac, atque) ou à la particule quam commune au comparatif (alius quam). Quam en latin fonctionne à la fois pour le comparatif d’inégalité et d’égalité, alors qu’en grec ἤ s’emploie seulement pour le comparatif d’inégalité, mais non pour celui d’égalité. Cette distinction se retrouve aussi, en parallèle, pour l’expression de la différence et de l’identité ou de la similarité. En latin, on constate une convergence dans les expressions de la différence et de l’identité ou de la similarité, qui font recours aux particules de la coordination connective (alius / dissimilis / dispar ac, atque vs idem / similis / par ac, atque), alors que le grec présente une divergence entre les deux expressions du point de vue syntaxique. L’identité ou la similarité s’expriment ou par la corrélation (ex. τοσοῦτον…ὅσον ; οἶον) ou par le pronom relatif ὅς ou des éléments apparentés, tels que ὡς, ὡσπερ, ou par les particules connectives καί, ἠύτε 43 , tout comme en latin tam…quam, idem qui, par, similis ac/ et. En revanche, la différence en grec est réalisée par la particule disjonctive ἤ, alors qu’en latin ce sont les particules connectives ac, atque, et qui la signalent. Bref, le grec exprime la différence par la stratégie syntaxique de séparation et d’exclusion, commune à la comparaison d’inégalité et à la coordination disjonctive, alors que le latin l’exprime à travers les particules employées pour la comparaison d’égalité et pour la coordination connective. En revanche, dans l’expression de l’égalité ou de la similarité, le grec et le latin convergent dans la même stratégie commune à la coordination connective. En conclusion, les expressions de la différence et la similarité, tout comme la comparaison d’égalité et d’inégalité, font recours aux stratégies qui sont communes tantôt à coordination disjonctive tantôt à la coordination connective. Le védique nous montre que les particules distribuées entre les deux fonctions sémantico-syntaxiques remontent à deux types de déictiques : ceux d’éloigne‐ ment (“distal deictics”) sont à l’origine de la particule disjonctive vā, alors que les déictiques de la proximité (“proximal deictics”) sont en rapport avec la particule coordonnante copulative ca  44 . La relation entre l’expression de la différence et de l’identité ou similarité, d’une part, et celle, de la coordination connective et disjonctive de l’autre, permet de mieux retracer le chemin qui a amené à la disparition de alius dans les langues romanes, remplacé, du point de vue formel et fonctionnel, par les continuateurs de alter, qui a ajouté à son sens primaire, le sens relationnel La négation comme moteur de l’évolution linguistique 371 <?page no="372"?> de alius. Une telle évolution est à mettre en parallèle, dans la coordination disjonctive, à la disparition de uel au profit de aut, et, dans la coordination connective, à la disparition de -que et atque au profit de et. Tout comme les opérateurs de la coordination, les aboutissements de alter dans les langues romanes, tels que it. altro, sp. otro, fr. autre, cumulent deux sens, notamment la référence à un ensemble de deux éléments, propre en origine à alter, et celle à plusieurs éléments, propre en origine à alius, qui véhicule la valeur de « différent ». Ce sont les contextes, syntaxiques ou pragmatiques, qui permettent de distinguer l’un de l’autre. Ainsi dans les langues romanes, les principaux opérateurs de la distinction syntaxique sont les deux types d’articles, défini et indéfini (l’autre et un autre), les indéfinis (it. qualcun altro / nessun altro), et les marqueurs de séparation (fr. autre que, it. altro da). De même, en anglais, les fonctions de other se différencient au niveau syntaxique (par ex., the other (one), another, other than), alors que, après un indéfini, l’anglais fait recours à un autre élément lexical (else), probablement apparenté à alius (par ex., anybody else). Dans les langues germaniques anciennes, les valeurs de « second » et de « autre » se retrouvent dans les emplois de anϸar du gotique ainsi que dans ceux de ōϸer du vieux anglais. 4. La négation et l’évolution de la comparaison Comme nous le disions, les constructions syntaxiques du latin et du grec montrent un parallélisme saisissant entre les expressions de la différence et de la similarité et celles du comparatif. Par ailleurs, la fonction essentielle du comparatif est d’exprimer une différence ou une similarité, respectivement par le comparatif d’inégalité (les notions de « plus » ou « moins ») ou d’égalité (la notion de « autant que »). Néanmoins, les deux types de comparatif sont susceptibles de se croiser, aboutissant à des résultats différents : d’une part, le renversement réciproque, le passage de l’inégalité à l’égalité et vice-versa et, de l’autre, l’effacement de la distinction entre inégalité et égalité, qui porte la comparaison au degré le plus haut possible, expression du superlatif ou d’élatif. Ces effets différents sont engendrés par deux types de contextes : 1) lorsque la négation porte sur les opérateurs d’inégalité ou de différence (notamment « plus », « moins », « différent ») ou d’égalité ou de similarité (notamment «-aussi-», « autant-», «-égal-»), le résultat est le renversement réciproque-: on passe de l’inégalité à l’égalité et vice-versa ; 2) lorsque le comparée ou le standard de comparaison sont niés, en présence d’un pronom indéfini négatif ou d’un « N-word », la distinction entre égalité et inégalité est effacée et la comparaison est projetée au degré le 372 Anna Orlandini & Paolo Poccetti <?page no="373"?> 45 Bertocchi & Orlandini (1996). plus élevé que possible, aboutissant à un superlatif absolu. Ce dernier cas se réalise également lorsque le comparant (le standard) est représenté par un terme exemplaire, une sorte de ‘parangon’ de la qualité paramètre de la comparaison. On peut résumer ces transformations sémantiques, en les exemplifiant dans le schéma suivant-: 1a) X n’est pas plus / moins doux que Y = X est aussi doux que Y 1b) X n’est pas aussi doux que Y = X est plus / moins doux que Y 2a) Rien n’est pas aussi doux que X / Rien n’est pas plus doux que X = X est doux au degré le plus élevé. 2b) X est aussi doux que le miel / X est plus doux que le miel = X est doux au degré le plus élevé, très doux. Dans les contextes de comparaison, la négation joue un rôle essentiel, se distribuant en manière différente. Dans 1a) et 1b) la négation renverse l’inégalité et l’égalité, permettant de passer de l’une à l’autre et vice-versa, par ex-: (30a) (animalia) nec minus atque homines inter se nota cluere (Lucr. 2,349) «-les animaux entre eux se connaissent, non moins bien que les hommes-». (30b) οὐχ ἧττον ἦν δεινὸς ἐν ταῖς τοῦ πολέμου παρασκευαῖς (Plut. Vit. Demetr. 19,10,6) « il n’était ni moins soigneux, ni moins diligent à faire tous les préparatifs nécessaires à la guerre ». (30c) PA. tum argenti montis, non massas, habet. Ethna mons non aeque-altust (Plaut. Mil. 1065) « il a une quantité d’argent, non par monceux, mais par montagnes : l’Etna n’est pas aussi haut-». (30d) καὶ οὐκ ἐν τῷ ὁμοίῳ στρατευσόμενοι καὶ ὅτε ἐν τοῖς τῇδε ὑπηκόοις ξύμμαχοι ἤλθετε ἐπί τινα (Tuc. 6,21) « les conditions de la campagne ne seront pas du tout ce qu’elles étaient, quand, ici, entourés de vos sujets, vous marchiez en alliés contre un adversaire-». En revanche, dans 2a) la négation du comparé permet au comparant d’atteindre le degré le plus élevé, exprimant ainsi un superlatif ou un élatif. Une telle structure consiste d’un « parallélisme négatif », qui fonctionne d’une manière inversée à la négation du comparant (le standard), qu’on vient de mentionner ci-dessus (ex. (28a), (28b)). Lorsqu’un pronom négatif à part entière tel que nemo, nihil figure dans le comparé, on obtient un degré superlatif de la qualité pour le comparant 45 : la négation du comparé réalise ce que E. Sapir a appelé La négation comme moteur de l’évolution linguistique 373 <?page no="374"?> 46 Sapir (1951, 132). 47 Ainsi Hofmann & Szantyr (1965, 108; 162; 874), qui parlent de « Komparativ statt Superlativ-». 48 Fraenkel (1960,7). un « unconditionned superlative » 46 . Suivant les grammairiens anciens on a parlé de comparatiuus pro superlatiuo  47 , car il correspond aux structures morpho-syntaxiques du comparatif, mais au niveau sémantique il réalise un véritable superlatif. Bref, la négation change le statut fonctionnel de la compa‐ raison : une comparaison d’inégalité ou d’égalité niée dans le comparé efface sa nature, devenant une expression qui n’admet pas de comparaison, c’est-à-dire un superlatif/ élatif. Dans ce cas, la distinction entre égalité et inégalité se neutralise, parce que la qualité, poussée au niveau le plus élevé, sort de toute comparaison. Les exemples suivants montrent le même effet de sens en présence d’une marque de comparatif d’inégalité (miserior, miserius dans (31a) et (31b)) ou d’une marque de comparatif d’égalité (aeque feracem dans 31c) : (31a) AL. Miseriorem ego ex amore quam te uidi neminem (Plaut. Cas. 520) «-Jamais je n’ai vu d’homme plus follement amouraché que toi-». (31b) BA. Miserius nihil est quam mulier (Plaut. Bacch. 41) «-Y a-t-il rien de plus malheureux qu’une femme ? -» (31c) EP. Nullum esse opinor ego agrum in agro Attico / aeque feracem quam hic est noster Periphanes (Plaut. Epid. 306 sq.) « Non, je ne crois pas qu’il y ait dans toute l’Attique un fond de terre d’aussi bon rapport que notre Périphane-». En outre, un indice formel qui plaide en faveur de la neutralisation de la distinc‐ tion entre inégalité et égalité, est fourni par l’exemple suivant, où l’expression d’inégalité (l’adjectif pourvu du morphème -ior signalant la supériorité) et celle d’égalité (l’adverbe aeque) co-existent dans le même énoncé-: (32) CHA. homo me miserior nullust aeque (Plaut. Merc. 335) «-Il n’est personne, je pense, qui soit plus malheureux que moi-». Comme l’avait signalé E. Fraenkel 48 , la structure comparative avec le pronom indéfini négatif (le type : nemo miserior me) se rencontre chez Plaute surtout au début des monologues pour exalter la figure du locuteur, en la présentant comme sans pareil. 374 Anna Orlandini & Paolo Poccetti <?page no="375"?> 49 La présence à la fois de mollior et de l’adverbe magis dans un même énoncé montre que la forme en -ior ne devait pas être perçue ici comme une gradation quantitative de l’adjectif, cf. aussi - AR. Quin nihil, inquam, inuenies magis hoc certo certius (Plaut. Capt. 644) « Oui, te dis-je, tu ne trouveras rien de plus certain que cette certitude ») où il s’agit aussi d’une comparaison qualitative. Comme nous le disions, le latin distingue, pour le comparé, les quantifieurs universels pleinement négatifs (nemo, nihil, nullus) et, pour le comparant, les quantifieurs à orientation négative (quisquam-/ ullus) : (33a) Ac uidete quanto taetrior hic tyrannus Syracusanis fuerit quam quisquam superiorum (Cic.Verr. II 4,123) «-Ce tyran fut plus cruel envers les Syracusains qu’aucun de ses prédécesseurs ». (33b) mollior 49 magis quam ullus cinaedus (Plaut. Aul. 422) « Le corps plus souple que celui d’un mignon-». L’emploi des indéfinis à polarité négative quisquam / ullus révèle qu’un sens négatif implicite est sous-jacent à ce type de comparaison. Même si aucune négation syntaxique n’y figure, le sens négatif est impliqué par l’exclusion de la totalité à l’exception du comparé. La négation du comparant n’est, donc, qu’un renversement de la négation du comparé, et produit le même effet de sens d’un superlatif. Ainsi la traduction des exemples (33a) et (33b) peut être reformulée par la négation du comparé de la manière suivante : 33a) «-aucun des tyrans antérieurs n’a été plus / aussi cruel-» 33b) «-aucun mignon-n’a eu un corps plus / aussi souple-». Par là, l’on parvient à la neutralisation de la distinction entre égalité et inégalité, aboutissant à la valeur d’un superlatif de la même manière que dans les exemples (31a), (31b), (31c). En parallèle, l’anglais, langue à double négation comme le latin classique, exprime la négation du comparant par les indéfinis à polarité négative à base «-any-» qui se rapprochent des quantifieurs de «-free choice-», par ex.-: (34a) he is sweeter than anybody else- se distinguant d’un quantifieur pleinement négatif employé dans le rôle de comparé, tout comme en latin nemo, nihil, nullus, par ex.-: La négation comme moteur de l’évolution linguistique 375 <?page no="376"?> 50 Viti (2002, 65-72). (34b) nobody is sweeter than him. - En revanche, les langues romanes, caractérisées par la structure à concordance négative, emploient les mêmes pronoms indéfinis morphologiquement négatifs dans le rôle du comparé et dans celui du comparant. Mais ces pronoms fonctionnent sémantiquement de manière différente : en fonction du comparé, où ils jouent le rôle sémantique du sujet, ils gardent une force pleinement négative, équivalant à celle de nemo, nihil en latin, alors qu’en fonction du comparant (le standard), ils sont des «-N-words-», avec force négative faible et seulement implicite, et peuvent fonctionner comme « free choice » (n’importe qui en français, chiunque en italien), par ex.-: (35a) fr. «-il est plus doux que nul autre / personne / n’importe qui-» (35b) it. «-è più dolce di nessuno / chiunque (altro)-». De son coté, le latin opère une distinction non seulement entre les pronoms indéfinis dans le rôle du comparé et dans celui du comparant, mais aussi dans la construction syntaxique du comparant. Comme le montrent les exemples (33a), (33b), le comparant exprimé par les indéfinis quisquam, ullus est introduit par quam, alors que l’emploi de l’ablatif de comparaison est exclu de ces contextes. Une telle contrainte s’explique à la lumière d’une distinction dans l’expression du comparant, que C. Viti a observé dans d’autres langues 50 -: c’est le critère d’identification, selon lequel en védique se distribuent la marque iva, introduisant un comparant (standard) plus ou moins identifié, et la négation ná, s’appliquant aux comparants non identifiés ou moins identifiables. Or, les indéfinis négatifs dans le comparant se situent au niveau zéro de l’identification du référent, ils ont une valeur générique, en cohérence avec leur nature de « free choice-». Pour cette raison, ils n’admettent que le connecteur quam. L’effet élatif, hyperbolique réalisé par un pronom indéfini (pleinement ou partiellement) négatif dans l’un ou l’autre terme de comparaison est atteint de la même mesure, lorsque le comparant (ou standard) est représenté par un terme exemplaire, une sorte de parangon de la qualité qui est le paramètre de la comparaison : c’est le cas exemplifié dans le schéma ci-dessus sous 2b) par l’expression «-X est aussi doux que le miel / X est plus doux que le miel-». Ce type de comparaison, représenté par la formule latine melle dulcior «-plus doux que le miel-», est très répandu dans plusieurs langues pour exprimer une 376 Anna Orlandini & Paolo Poccetti <?page no="377"?> 51 Benveniste (1948, 135). 52 Rosén (1999, 190). 53 Sans aucun fondement, donc, Ittzés (2021, 246) parle d’un “special instance of the comparative of superiority”. 54 Voir, par ex., Löfstedt (1942, 327 ss.). valeur élative, par ex. en grec μέλιτος γλυκίων, identique à celle du latin, en indien ancien ghṛtāt svādīyaḥ « plus doux que le beurre », en irlandais ancien maissiu máenib « plus splendide que les trésors », en vieux anglais sunnan beorhtra « plus lumineux que le soleil », en serbe vjetra brže « plus rapide que le vent ». Parfois, le terme employé comme parangon est issu de l’adjectif lui-même, exprimant la qualité qui est le paramètre de comparaison, comme, par ex., en védique svādhoḥ svādīyaḥ (RV I 114,6) « plus doux que le doux », en vieux slave-: zlo zla zlěje « un malheur plus malheureux que le malheur », en latin O salute mea salus salubrior (Plaut. Cist. 644) « Ô, toi qui es plus salutaire que le dieu Salut ». Cette stratégie donne lieu à un polyptote, une figure rhétorique, qui renchérit sur le sens élatif de l’expression, dont le style sort du langage ordinaire. Déjà Benveniste avait souligné la convergence sémantique entre la négation du comparé ou du comparant et la comparaison ‘à parangon’, qu’il avait nommé « adéquatif » « par référence à un objet pris comme norme immuable [….] qui s’identifie avec la qualité qu’elle symbolise » 51 . Autrement dit, l’effet intensif, hyperbolique, réalisé par la négation du pronom indéfini coïncide avec la prédication superlative exprimée par comparant, présenté comme un ‘parangon’, atteignant le degré le plus élevé de la qualité, tout comme dans le type melle dulcior. Ainsi les deux expressions réalisent des superlatifs ou élatifs, ce qui exclut leur interprétation comme comparatifs d’égalité 52 ou d’inégalité 53 . En latin, le parallélisme fonctionnel du comparatif avec la négation du comparé (le type nemo me miserior) et la comparaison ‘à parangon’ (le type melle dulcior) est signalé aussi par la construction du comparant à l’ablatif (avec certaines contraintes) au lieu de l’emploi de quam. On a beaucoup discuté sur l’origine de l’ablatif de comparaison par rapport aux valeurs, apparemment opposées entre elles, de l’ablatif latin, notamment, celle d’éloignement ou séparation, qui lui est propre, et celle sociative, qui relève du syncrétisme avec le cas instrumental 54 . En fait, les deux fonctions de l’ablatif coexistent, se distribuant différemment dans les deux expressions de comparatif ayant la valeur de superlatif. On reconnaîtra la fonction sociative de l’ablatif au type melle dulcior, où le comparé est soudé à l’élément modèle exemplaire de la qualité paramètre de la comparaison. À ce niveau, il fonctionne comme un équatif au niveau le plus élevé, n’admettant aucune gradation. En revanche, dans la La négation comme moteur de l’évolution linguistique 377 <?page no="378"?> 55 Ascoli (1891a-; b); de même Meid (1967). 56 Meid (1967, 184). 57 LEIA s.v. co (2) C-133-; com- C-161. 58 Meid (1967, 225). comparaison, où le comparé est nié, l’ablatif exprime une séparation, parce qu’il sert à marquer l’exclusion du comparant par rapport à une totalité qui ne possède pas la qualité-paramètre. De même, la fonction sociative et celle séparative se distribuent dans l’ex‐ pression de la comparaison à l’intérieur de plusieurs langues. Par exemple, en indien ancien, le comparant figure tantôt au cas instrumental tantôt à l’ablatif. En irlandais ancien, le comparant se construit à l’accusatif dans l’équatif, alors qu’il se présente au datif dans le comparatif d’inégalité. Une particularité de cette langue c’est l’équatif à suffixe -ithir, généralement reconduit à -tero-, qui figure le plus souvent dans des tournures du type latin melle dulcior. Comme l’avait déjà remarqué Ascoli 55 , les occurrences de l’équatif irlandais en -ithir se présentent constamment à union avec un comparant conçu comme parangon, un modèle exemplaire de la qualité, tel que, par exemple, gilithir snechta «-aussi / plus blanc que la neige-», dans des formules tout à fait parallèles à celles du type latin melle dulcior. De cette manière, dans l’équatif de l’irlandais ancien et moyen, toute gradation, impliquée par la comparaison d’égalité ou d’inégalité, est effacée au profit du sens superlatif. Ces expressions d’élatif caractérisent le style élevé ou recherché de certains genres de textes. C’est pourquoi l’irlandais ancien distingue l’équatif à suffixe -ithir, ayant valeur de superlatif, du comparatif d’égalité proprement dit, où le paramètre de la comparaison est marqué par le préfixe com, par ex. maith « bon » : com-maith « aussi bon » 56 . Significativement ce préfixe est issu de la préposition co(n) signifiant « avec » 57 , ce qui révèle la fonction sociative à l’origine de ce procédé 58 . L’ordre syntaxique est un aspect commun aux expressions de sens élatif, le type latin melle dulcior et l’équatif irlandais. Dans la formule latine melle dulcior ainsi que dans celles semblables des autres langues, par ex., μέλιτος γλυκίων du grec, ghṛtāt svādīyaḥ de l’indien ancien, zlo zla zlěje du vieux slave, snjó hvitari du vieux norrois, le terme ‘à parangon’ souvent précède immédiatement la qualité paramètre de la comparaison, à laquelle il se soude strictement (tout comme dans l’ordre syntaxique du déterminant suivi du déterminé). Cet ordre syntaxique, issu de la comparaison ‘à parangon’, est à l’origine de quelques composés des langues germaniques, tels que les adjectifs de l’allemand honigsüss et de l’anglais honey-sweet signifiant « si-doux que le miel-». 378 Anna Orlandini & Paolo Poccetti <?page no="379"?> 59 Vendryes (1980, 128). 60 Pinault (1985, 140). Cet ordre syntaxique se distribue différemment dans la construction de l’équatif de l’irlandais ancien, qui distingue le cas où le comparant est représenté par un nom ou un syntagme nominal, du cas où il est représenté par une proposition. L’un est postposé au paramètre de comparaison (par ex., gilithir snechta « aussi / plus blanc que la neige », sonartaithir slebe « aussi/ plus fort que les montagnes »), alors que l’autre est antéposée au paramètre de comparaison, par ex.,-: (36a) amal as suthain riuth gréne síc bith suthainidir sin ainm Solmon « le nom de Salomon sera aussi éternel que la course du soleil » (littéralement « qu’est éternelle la course du soleil, aussi éternel sera le nom de Salomon ») 59 . L’antéposition du comparant au comparé caractérise les tournures de la langue védique exprimant un équatif à valeur élative au moyen de la particule négative ná,-comme, par ex.,-: (36b) ghrtáṃ ná suci matáyaḥ pavante (RV VI 10.2 d) «-(devenant) pures comme le beurre-fondu, (mes) pensées se clarifient-» (traduction par L. Renou). Comme l’a observé J.-G. Pinault 60 , le principe à la base d’une telle tournure c’est l’exclusion du comparant, ce qui a comme effet de mettre en relief que la qualité possédée par le comparé atteint le degré le plus élevé. L’équatif/ élatif de l’exemple (36b) est construit de la manière suivante : l’énoncé se divise en deux parties : d’abord on introduit le comparant à ‘parangon’ (ghrtáṃ « beurre ») tout au début, suivi par la particule négative ná et, dans la seconde partie, le comparé (matáyaḥ « pensées »), suivi, en clôture, du paramètre de comparaison (pavante « devenir pur, clair »). Le sens, qui en résulte, est le suivant : « ce n’est pas le beurre-fondu qui se clarifie (exclusion du comparant), mais ce sont mes pensées qui se clarifient au même degré (sens équatif pour la qualité du comparé) », aboutissant à « mes pensées deviennent aussi clairs que le beurre-fondu (sens élatif, la qualité est possédée par le comparé au plus haut degré)». Une telle structure, bien connue du védique, a été reconduite par J-G. Pinault au schéma S ná P, signifiant que P, même s’il n’est pas S (à savoir, le standard ou comparant), possède les qualités de S, qui est le modèle exemplaire. En fait, une telle structure syntaxique se rapproche de certaines tournures d’équatif du latin, qui ont la même valeur élative, telles que-: La négation comme moteur de l’évolution linguistique 379 <?page no="380"?> 61 Comme, en revanche, malheureusement, le fait Ittzés (2021, 245). (37a) PYTH. Pumex non aeque est ardus atque hic est senex (Plaut. Aul. 297) «-la pierre ponce n’est pas si sèche que ce vieux-là-» qui sont également possibles dans les langues romanes, par ex. en français-: (37b) la pierre n’est pas si dure que ta tête. Au fond, du point de vue sémantique, ces expressions ne se différencient pas de celles, où le comparé ou le comparant sont niés, du type-: (37c) ME. Nullust hoc metuculosus aeque (Plaut. Amph. 293) « Il n’y a pas un poltron comme lui-». Les types de comparaison représentées par (37a) et (37c) font ressortir le rôle de la négation, qui permet de renverser l’ordre du comparant et du comparé sans modifier le sens. On peut même supprimer la négation, en réalisant une proposition positive, comme, par ex.-: (38a) la pierre n’est pas si dure que ta tête vs (38b) ta tête est si dure que la pierre Ou bien, on peut déplacer la négation du comparé au-comparant : (38c) Personne n’a la tête si dure que toi vs (38d) Tu as la tête si dure que personne. Les deux couples de tournures, notamment (38a), (38b) et (38c), (38d), montrent une convergence dans la valeur élative, réalisée grâce aux transformations syntaxiques opérées par la négation. Cette propriété sémantico-syntaxique distingue le fonctionnement des expressions ayant valeur élative, du comparatif proprement dit. C’est pourquoi le type (38b), qui se rattache au type melle dulcior quant à sa force de superlatif, doit être séparé du comparatif de supériorité, proprement dit, auquel à tort on a prétendu de l’intégrer 61 . Le rôle important de la négation dans la comparaison, que l’on peut saisir en synchronie, se manifeste aussi en diachronie. C’est surtout la comparaison à valeur élative qui est affectée par des changements, qui relèvent, d’une part, des 380 Anna Orlandini & Paolo Poccetti <?page no="381"?> 62 Haspelmath & Buchholz (1998, 306-309). 63 Viti (2015, 36, 125). variations du système de la négation et de l’ordre syntaxique dans l’évolution langagière et, de l’autre, des variations stylistiques entre genres différents de textes. En effet, comme nous le disions, la comparaison ‘à parangon’ s’applique souvent aux tournures caractérisant un registre stylistique élevé, comme le montrent les occurrences de l’équatif irlandais et des formules védiques avec négation postposée au comparant. Dans les deux cas, il s’agit d’expressions emphatiques adéquates aux hymnes poétiques, aux formules religieuses et, généralement, aux registres stylistiquement marqués. Le changement diachronique concerne la marque du paramètre, comme dans le cas de l’équatif de l’irlandais ancien qui a disparu dans la langue moderne, tout comme le morphème du comparatif latin dans les langues romanes, ou la marque du comparant comme les tournures védiques, qui sont inconnues de la langue sanscrite. Les deux langues, situées aux extrémités du domaine indo-européen, mon‐ trent une tendance à la « case transparency » 62 , c’est-à-dire l’emploi du même cas dans le comparé et dans le comparant. Ce phénomène est très répandu dans l’évolution de la comparaison de plusieurs langues ainsi que dans celle d’autres relations syntaxiques 63 . En général, il s’agit d’un procédé morpho-syntaxique qui amène au déclin d’un cas régime exclusif du comparant (sélectionné par chaque langue entre l’ablatif, l’instrumental, le génitif, le datif, etc.) au profit d’une particule qui lui permet d’adopter le même cas du comparé. Au niveau synchronique, une telle tendance est témoignée déjà en grec homérique, significativement dans les comparaisons ‘à parangon’, où l’emploi du génitif dans μέλιτος γλυκίων « plus doux que le miel » (Il. 1, 249) alterne avec μελάντερον ἠΰτε πίσσα « noir comme la poix » (Il. 4, 276). Ce flottement entre l’emploi du génitif, propre du comparatif d’inégalité et ἠΰτε, qui sert au comparatif d’égalité, démontre, une fois de plus, que, dans les comparaisons ‘à parangon’, l’opposition entre égalité et inégalité est neutralisée. En diachronie, la particule qui s’impose est, le plus souvent, celle qui fonctionne comme connecteur servant à mettre en corrélation deux éléments dans une échelle graduable, non bornée. Par là, les connecteurs de ce type se prêtent à mettre en relation sur des échelles différentes, non seulement des noms, mais aussi des propositions et, par conséquent, ils se prêtent à devenir des outils de la coordination et de la subordination. C’est le cas du quam en latin et de ἠΰτε en grec, ainsi que de iva et yáthā en sanscrit, qui fonctionnent à la fois dans la comparaison, comme La négation comme moteur de l’évolution linguistique 381 <?page no="382"?> 64 Viti (2007, 44-47). 65 Viti (2002, 73-80). 66 Viti (2002, 81). marques du comparant, et dans la syntaxe des propositions, pour introduire la coordination et la subordination 64 . Aux sources de telles particules se trouvent des pronoms déictiques, marquant la distance ou la proximité (comme pour iva et ἠΰτε) ou pronoms relatifs (comme pour yáthā et quam). La propriété de ces connecteurs d’atteindre le degré le plus élevé de l’échelle d’un paramètre fixé favorise leur emploi dans les comparaisons ‘à parangon’ (le type melle dulcior). Cet emploi se retrouve déjà dans les langues anciennes, comme le témoigne la disparition de ná dans la langue sanscrite classique au profit de iva  65 pour les expressions d’équatif à valeur élative, en parallèle aux occurrences de quam au lieu de l’ablatif en latin-: (39) PEN. homo levior quam pluma (Plaut. Men. 488) « personnage plus léger que plume-» Les deux phénomènes présentent des convergences quant à leur origine et à leur aboutissement. Comme l’a montré C. Viti 66 , les particules ná et iva, en tant que marqueurs du comparant, se distribuent différemment dans la langue védique, selon un critère d’identification : iva s’accompagne d’éléments qui dans l’échelle d’identification se situent à un degré plus élevé que ceux marqués par ná. Le même critère s’applique en latin à la distribution de l’ablatif et de quam comme marqueurs du comparant : dans ce cas, c’est l’ablatif qui marque les éléments qui sont référentiellement identifiables. En effet, le comparant à l’ablatif est toujours un individu, indiqué par pronom personnel (me, te) ou un terme concret en fonction de parangon (melle, luce, sidere, etc.). Cette contrainte dans l’emploi de l’ablatif est signalée par les contextes où le comparé est nié. Ainsi, par exemple, l’ablatif marque le comparant exprimé par un pronom personnel, qui assure le degré le plus élevé de l’identification, alors qu’il est exclu en présence d’un terme à valeur générique, qui est marqué par quam. Cette distinction dans la marque du comparant est montrée d’une manière contrastive par les exemples suivants-: (40a) = (31b) Miserius nihil est quam mulier (Plaut. Bacch. 41) « Y a-t-il rien de plus malheureux qu’une femme ? -» (40b) = (32) CHA. homo me miserior nullust aeque (Plaut. Merc. 335) «-Il n’est personne, je pense, qui soit plus malheureux que moi-». 382 Anna Orlandini & Paolo Poccetti <?page no="383"?> Ainsi, un quantifieur universel à orientation négative, correspondant au degré zéro de l’identification, dans le rôle du comparant, n’admet que quam, tels que quam quisquam, quam ullus (33a), (33b), alors que l’ablatif ne saurait être réalisé par *quoquam ou *ullo. Dans les comparaisons ‘à parangon’ (le type melle dul‐ cior), le terme à l’ablatif est universellement identifié par la référence à un modèle exemplaire. Néanmoins, on trouve parfois, même si assez rarement, l’emploi de quam à la place de l’ablatif, comme dans l’exemple (39). Le même effet de sens élatif pour le comparant, exprimé par un pronom indéfini à orientation négative et par un terme à parangon, a mis en œuvre l’alignement des deux constructions, aboutissant à l’emploi d’une marque commune pour le comparant, qui permet la « case transparency », réalisée par quam. Le parallélisme morpho-syntaxique qui se produit entre quam quisquam et quam pluma n’obéit qu’à un principe cognitif visant à créer une symétrie entre rôle fonctionnel et sémantique et isomorphisme formel et iconique. En latin, la marque quam du comparant, étant obligatoire lorsqu’il s’agit d’un pronom indéfini à orientation négative (quam quisquam), a entraîné ensuite, occasionnellement, celle du terme ‘à parangon’ (quam pluma). La diffusion de quam en latin, parallèle à celle de iva et yáthā en sanscrit, aboutit au même résultat : uniformiser la marque du comparant dans la comparaison ‘à parangon’ à celle du comparatif d’inégalité ou d’égalité. Cela s’impose dans les langues romanes, où la comparaison ‘à parangon’ (le type melle dulcior) ne se distingue pas formellement du comparatif d’égalité ou d’inégalité : par ex., en français l’expression « X est plus/ aussi doux que le miel » s’exprime de la même manière que « X est plus/ aussi doux que Y ». De même, en italien «-X è più dolce del miele-» et «-X è più dolce di Y ». Conclusions À partir de l’analyse de phénomènes langagières apparemment séparés et isolés nous avons essayé de retracer un parcours commun, où la négation joue un rôle primaire dans leur évolution. Ainsi nous avons montré que l’évolution des dérivés à suffixe -tero-, réalisée par les indéfinis marquant l’altérité, s’inscrit dans un chemin cohérent qui, d’une part, au niveau cognitif, est parallèle à la disparition du duel, en tant que catégorie de la grammaire, et, de l’autre, au niveau sémantique, ouvre vers la multiplicité et la variété, caractérisant les langues naturelles. Les contextes sous négation ont favorisé un tel changement, qui se signale d’une manière spéciale dans l’évolution de alter du latin aux langues romanes. La négation comme moteur de l’évolution linguistique 383 <?page no="384"?> La même démarche, le passage d’une structure fermée, logique, composition‐ nelle à une structure plus libre et non-compositionnelle, est aussi à la base de l’évolution du système de la négation des structures à double négation (DN) vers celles à concordance négative (NC), ainsi que de l’évolution de aut de la valeur exclusive du latin à la valeur inclusive des langues romanes. D’une manière plus saisissante, la comparaison fait ressortir universellement au niveau synchronique la fonction de la négation, qui permet le passage du comparatif de l’égalité à celui l’inégalité et inversement. En outre, la négation du comparé et du comparant par l’emploi d’un indéfini fait acquérir à un comparatif le sens d’un superlatif ou élatif. Ces types d’expressions pivotent autour de l’exclusion de la totalité, excepté le comparé, qui, de cette manière, devient le seul caractérisé par la propriété paramètre de la comparaison. Le même sens de superlatif est obtenu par la comparaison dite ‘à parangon’, qui signale que le terme ‘à parangon’ joue le même rôle que la négation du comparé ou du comparant. Ce n’est pas par hasard si, dans certaines langues, ces expressions sont formulées par la négation du terme servant comme ‘parangon’. Significativement, dans tous ces phénomènes, les enjeux de l’évolution se retrouvent déjà dans les phases les plus anciennes de chaque langue, suivant le principe général que les procédés de l’évolution diachronique plongent leurs racines dans les tendances décelées par l’état synchronique d’une langue. Dans tout cela, la négation a joué un rôle essentiel, qui, auparavant, n’avait jamais été pris en compte dans la perspective unitaire et cohérente d’une évolution ouvrant les structures d’une langue dite « logique » vers la liberté innovante d’une langue plus naturelle. Bibliographie Ascoli, G.I. (1891a) “Sulla storia generale delle funzioni del suffisso -terocon speciale considerazione del riflesso irlandese”, Archivio Glottologico Italiano Supplemento I: 53-72. Ascoli, G.I. (1891b) «-Celtica. Ancora dei pareggiativi irlandesi in -ithir-», Archivio Glottologico Italiano Supplemento I: 98-99. Bader, F. (1986) « De skr. anyáà skr. ā́rya-. Noms i.e. de l’ ‘autre’ », Bulletin de la Société de Linguistique de Paris 80: 57-90. Baldi, Ph. & P. Cuzzolin (2010) “Osservazioni sulla morfologia del comparativo in greco antico”, dans Putzu, I.; G. Paulis; G.F. Nieddu; P. Cuzzolin (éds.) La morfologia del greco tra tipologia e diacronia, Milano, Angeli, 46-60. Bartoněk, A. (2003) Handbuch des mykenischen Griechisch, Heidelberg, Winter. 384 Anna Orlandini & Paolo Poccetti <?page no="385"?> Benveniste, É. (1948) Noms d’agents et noms d’action en indo-européen, Paris, Maison‐ neuve. Bertocchi, A. & A. Orlandini (à paraître) “Latin ALTER and ALIUS”, dans Amaral, P. (éd.) Other: Ambiguity, constraints and change. Bertocchi, A. & A. Orlandini (1996) “Quelques aspects de la comparaison en latin”, Indogermanische Forschungen 101: 195-232. Biraud, M. (1983) “Les expressions de l’idée comparative en Grec classique : coréférence et disjonction”, Glotta 61: 167-182. Blass, F. & A. Debrunner (1961) Greek grammar of the New Testament and other early Christian literature, Chicago, University of Chicago Press. Bortolussi, B. (2015) Syntaxe des indéfinis latins, Quis, quisque, alius, Paris, Presses de l’Université Paris-Sorbonne. Cioffi, C. (2019) “Bellum Africum tra vecchi pregiudizi e nuove prospettive”, Bollettino di studi latini 49: 220-235. Dahl, Ö. (2010) “Typology of negation”, dans Horn, L.R. (éd.) The expression of negation, Berlin & New York, De Gruyter, 9-38. Daniel, M. (2013) “Plurality in independent personal pronouns”, dans Dryer, M. & M. Haspelmath (éds.) The world atlas of language structures online, http: / / wals.info/ chapter/ 35 de Swart, H. (2010) Expression and interpretation of negation. An OT typology. Dordrecht, Springer. Denniston, J.D. (1970) 2 The Greek particles, Oxford, Clarendon. Dunkel, G. (1987) “A typology of metanalysis in Indo-European”, dans Watkins, C. (éd.) Studies in memory of Warren Cowgill, Berlin & New York, De Gruyter, 7-37. Dunkel, G. (2014) Lexikon der indogermanischen Partikeln und Nominalstämme, Heidel‐ berg, Winter. Ernout, A. & F. Thomas (1953) 2 Syntaxe latine, Paris, Klincksieck. Espinal, M.T. (2007) “Licensing espletive negation and negative concord in Catalan and Spanish”, dans Floricic, F. (éd.) La négation dans les langues romanes, Amsterdam, Benjamins, 49-74. Fraenkel, E. (1960) Elementi plautini in Plauto, Firenze, La Nuova Italia. Giannakidou, A. (2000) “Negative concord and the scope of universals”, Transactions of the Philological Society 98: 87-120. Giannollo, C. (2018) “Indefinites and negation in Ancient Greek”, DiGS 20, University of York. Haspelmath, M. & O. Buchholz (1998) “Equative and similative constructions in the languages of Europe”, dans van der Auwera, J. (éd.) Adverbial constructions in the languages of Europe, Berlin & New York, de Gruyter. La négation comme moteur de l’évolution linguistique 385 <?page no="386"?> Haspelmath, M. (2005) “Negative indefinite pronouns and predicate negation”, dans Haspelmath, M.; M. Dryer; D. Gil; B. Comrie (éds.) The world atlas of language structures, Oxford, Oxford, University Press, 466-469. Hofmann, J.B. & A. Szantyr (1965) Lateinische Syntax und Stilistik, München, Beck. Horn, L. (2020) -“Negation and opposition: contradiction and contrariety in logic and language”, dans Déprez, V. & M.T. Espinal (éds.) The Oxford handbook of negation, Oxford, Oxford University Press. Horn, L. (1989) A natural history of negation, Chicago, The Chicago University Press. Ittzés, M. (2021) “Melle dulcior. Equative or comparative? ”, dans Martín Rodriguez, A.M. (éd.) Linguisticae dissertationes. Current perspectives on Latin grammar, lexicon, and pragmatics, Madrid, Ediciones Clásicas, 235-248. Keenan E.L. & B. Comrie (1977) “Noun Prase Accessibility and Universal Grammar”, Linguistic Inquiry 8: 63-99. Kluge, F. (1989) Etymologisches Wörterbuch der deutschen Sprache, Berlin & New York. Kühner, R. & C. Stegmann (1955) Ausführliche Grammatik der lateinischen Sprache, Satzlehre. 3. Auflage durchgesehen von A. Thierfelder, Leverkusen, Gottschalk. Lehmann, W. (1986) A Gothic etymological dictionary, Leiden, Brill. LEIA = Lexique étymologique de l’irlandais ancien, dirigé par J. Vendryes (1959-), Dublin & Paris, Institute for Advanced Studies & Centre National de la Recherche Scientifique. Löfstedt, E. (1942) 2 Syntactica, I Teil: Studien und Beiträge zur historischen Syntax des Lateins, Űber einige Grundfragen der lateinischen Nominalsyntax, Lund, Gleerup. Mayser, E. (1933) Grammatik der griechischen Papyri aus der Ptolomäerzeit, Berlin, De Gruter. Meid, W. (1967) “Zum Aequativ der keltischen Sprachen, besonders des Irischen”, dans Beiträge zur Indogermanistik und Keltologie J. Pokorny zum 80. Geburtstag gewidmet, Innsbruck, Institut für Sprachwissenschaft, 223-242 (réimpr. dans Ausge‐ wählte Schriften zum Indogermanischen, Keltischen und Germanischen, Innsbruck 2012, 183-210). Meyer-Lübke, W. (1890-1900) Grammaire des langues romanes (trad. A. et G. Doutrepont), Paris, Welter. Muller, C. (1991) La négation en français, Genève, Droz. Napoli D.J. & M. Nespor (1976) “Negatives in Comparatives”, Language 52: 811-838. Onions, C.T. (1966) The Oxford dictionary of English etymology, Oxford, Clarendon. Orlandini, A. & P. Poccetti (2022) “In principio erat nē/ nĕ : une analyse de la négation latine non-standard”, Revue de Linguistique Latine du Centre Alfred Ernout (De Lingua Latina), 22. L’expression de la volonté en latin, Université Paris Sorbonne, https: / / hal.s orbonne-universite.fr/ hal-03777976. Orlandini, A. (2001) Négation et argumentation en latin. Grammaire fondamentale du latin. Tome VIII, Louvain & Paris, Peeters. 386 Anna Orlandini & Paolo Poccetti <?page no="387"?> Pinault, G. (1985) “Négation et comparaison védique”, Bulletin de la Société de Linguistique de Paris 80: 103-144. Rosén, H. (1999) Latine loqui, München, Fink. Sapir, E. (1951) “Grading: a study in semantics”, Selected writings of E. Sapir in language, culture and personality, Berkeley & Los Angeles, 122-149. Serbat, G. (1984) “Is : un super-nom”, Latomus 43: 554-559. Untermann, J. (2000) Wörterbuch des Oskisch-Umbrischen, Heidelberg, Winter. Vendryes, J. (1908) Grammaire du vieil-irlandais, Paris, Guilmoto. Viti, C. (2002) “Comparazione e individuazione : uno studio sugli equativi rgvedici ina e ná”, Archivio Glottologico Italiano 87: 46-87. Viti, C. (2007) Strategies of subordination in Vedic, Milano, Angeli. Viti, C. (2011) “Osservazioni sulla negazione cumulativa”, dans Lasagna, M.; A. Orlandini; P. Poccetti (éds.) Attorno alla negazione. Analisi di contesti negativi dalle lingue antiche al romanzo, Linguarum Varietas 1: 13-25. Viti, C. (2015) Variation und Wandel in der Syntax der alten indogermanischen Sprachen, Tübingen, Narr. Wackernagel, J. (2009) Lectures on Syntax, éd. par D. Langslow, Oxford, Oxford University Press. La négation comme moteur de l’évolution linguistique 387 <?page no="389"?> Language Contact in Antiquity. Participial Constructions in Hellenistic Greek, Hebrew/ Aramaic and Old Church Slavonic Vit Bubenik (Memorial University of Newfoundland) Abstract: Three salient morpho-syntactic features of Hellenistic Greek will be examined in terms of language contact with Hebrew and Aramaic: (i) increased use of periphrastic constructions, (ii) constructions of the genitive and the dative absolute, and (iii) increased use of the nominalized articular infinitive. In the second part of this paper we will examine their replications in Old Church Slavonic as translated from the books of the New Testament. Keywords: Participles, translations, Greek, Hebrew, Aramaic, Old Church Slavonic 1. Translation in Antiquity The large field of language contact has recently been covered in terms of theory and methodology (and 40 case studies) in the Wiley-Blackwell volume (cf. Hickey 2010). Historical scenarios of language contact in antiquity deserves a special attention in our attempts to provide an ultimate ‘explanation’ of language changes in Hellenistic Greek, Mishnaic Hebrew and Middle Aramaic. In current historical sociolinguistics the social and linguistic factors are distin‐ guished. The linguistic ‘predictors’ (Thomason 2010: 39) include the degree of ty‐ pological distance, differing degrees of contact intensity, universal markedness and degree of integration within a linguistic system). For antiquity our success in this endeavor depends crucially on the availability of bilingual corpora such as that of biblical books of the Old Testament (OT) (Septuagint) and the New Testament (NT) (translated from Hebrew and Aramaic, and composed in Hellenistic Greek by men whose command of Greek varied to various degrees). <?page no="390"?> Given the prestige of these texts, their translation is literal (calquing) verbum e verbo as opposed to free sensus de sensu (in Saussurean terminology emphasis on signifiant versus emphasis on signifié (Brock 2001: 879)). The prestige of the Bible in Late Antiquity ( Jerome’s Vulgata in the West) was instrumental in the spread of the literal style in almost all translations (Syriac, Armenian, Old Church Slavonic). For the purposes of historical soci‐ olinguistics it is important to realize that their literal renderings reproduce grammatical categories of the original, especially syntactic categories (word order) and morpho-syntactic categories (the ‘redundant’ use of the correlative demonstrative pronoun in the relative clause, phrasal word order, positioning of pronominal clitics, expressing the notion of repetitivity/ iterativity, rendering the Hebrew gerund (‘quotative particle’) and periphrastic formations in the imperfective aspect). In this paper we will examine renderings of the participial constructions, absolute constructions and the use of the nominalized articular infinitive in Hellenistic Greek and Old Church Slavonic which might be indica‐ tive of ‘more intense’ language contact. In the syntax of OT and NT texts there are structural deviations from the usus of contemporary Hellenistic authors. Taken at their face value, they should be evaluated with a caveat that we are dealing with the literary version of a certain register of the Hellenistic Koine with a complicated relationship to its spoken variety due to the process of translation. In earlier scholarship, the intensity of contact between Greek and Hebrew was measured in terms of “Semitisms” (Maloney 1981). It is generally acknowledged that the most striking among them is the ‘overuse’ of the conjunction καὶ ‘and’ in imitation of its Hebrew equivalent wə= ‘and’ at the beginning of a new paragraph and in instances of conditional subordination (see under 2.5). In the first part of this paper (Sections 2.1 - 2.5), we will examine the realizations of participial and absolute constructions in Hellenistic Greek and Hebrew/ Aramaic which could be judged to reflect more intense language con‐ tact than that of the borrowing of basic vocabulary (Section 3). In the second part (Sections 4 - 5), along the same lines, we will examine the renderings of Greek participial constructions, the exploitation of case in absolute constructions, and the use of the articular infinitive in Old Church Slavonic translations (of the 10/ 11 th c.). Our main aim will be to assess the degree of the intensity of contact between two IE languages which are typologically closer. 390 Vit Bubenik <?page no="391"?> 2. Verb system of Late Classical/ Hellenistic Greek The system of finite forms of Hellenistic Greek is practically identical with that of Classical Greek with its three aspectual categories (imperfective, perfective and perfect) with some noticeable innovations towards analyticity. The most salient general differences are found in the formation of the future, which has retreated in several ways. Most notably, the active reduplicative future perfect (γε-γράψω) and its mediopassive counterpart (λε-λύ-σ-εται) have dropped out, and the simple future can now be formed analytically by means of the auxiliary μέλλω γράφειν ‘I intend to write’ > ‘I will write’. Its participial system is sketched in Table 1. It displays participial counterparts to all the finite aspectual categories (Imperfective - Perfective - Perfect-- Future). - Imperfective Perfective Perfect Future (phased out) Active -ont- -όnt-, -s-ant- -(w)όt- -s-ont- Med/ Pass ó-meno- -ó-meno-, -sá-meno- -méno- -só-meno- Pass - - - -tó -té-o Table 1. Greek participial system Greek could express the contrast between the perfective and perfect (retro‐ spective) aspect in main clauses by finite means (ἔγραψα vs. γέγραφa) and in subordinate clauses by participles: γράψας ‘having written’ vs. γεγραφώς ‘having written (with a present result)’. However, during the post-Classical centuries this contrast was weakened to judge by the indiscriminate use of the aorist and the perfect in describing the same situation by the Hellenistic authors. At the end of the development in Medieval Greek, the rich participial system with four aspectual categories (IPFV, PFV, PERF and FUT), declinable for gender, number and case, was reduced to two genderless, numberless and caseless absolutives whose main syntactic function was the implementation of the relative time of simultaneity and anteriority: γράφοντας ‘(while) writing’ and ἔχοντας γράψει ‘(after) having written’. 2.1 Increased use of periphrastic constructions In addition, a number of periphrastic constructions combining participles with the copula are documented. The Classical language had already made use of compounding in the formation of the mediopassive modal forms (subjunctive Language Contact in Antiquity. Participial Constructions 391 <?page no="392"?> and optative). These formations are based on the imperfective mediopassive participle in -μένο in combination with the modal forms of the copula: (1) Mediopassive perfect indicative: λέ-λυ-ται; subjunctive: λε-λυ-μένον ᾖ - Mediopassive pluperfect indicative: ἐ-λέ-λυ-τo; optative: λε-λυ-μένον ἐίη Already in the Classical language this periphrasis had been extended to the active modal forms combining the active perfect participle with the modal forms of the copula: (2) Active perfect indicative: λέ-λυ-κ-ε(ν); subjunctive: λε-λυ-κ-ὼς ᾖ - Active pluperfect indicative: ἐ-λελύ-κ-ει(ν); optative: λε-λυ-κ-ὼς ἐίη In terms of linguistic diffusion in our principal document of Hellenistic Greek, the New Testament, Blass and Debrunner (1961: 179) observe that the ‘periphra‐ sis’ with the copula is rare in the present, while in the imperfect, future and the infinitive it is widely employed: (3) Present: ἡ διακονία οὐ μόνον ἐστὶν προσαναπληροῦσα… ‘for the rendering of this service not only supplies the wants of the saints’ [2 Corinthians 9: 12] (4) Imperfect: ἤμην φυλακίζων καὶ δέρων τοὺς πιστεύοντας ἐπὶ σέ ‘I was impris‐ oning and beating those who believed in you’ [Acts 22: 19] - (It should be observed that this is a translation of Paul’s speech given in Ἑβράιδι διαλέκτῳ, i.e. in Aramaic). (5) Future (active): ἀνθρώπους ἔσῃ ζωγρῶν ‘You will be catching men’ [Luke 5.10] - Future (passive): ἔσεσθε μισούμενοι ὑπὸ πάντων ‘you will be hated by all’ [Mt 10.22] As Blass-Debrunner (1961: 180) observe, here “the reason for periphrasis is the emphasis on duration’, which the monolectal passive future μιση-θή-σε-σθε does not mark morphologically. An isolated example of the periphrastic imperative is in (6): (6) Imperative: ἴσθι εὐνοῶν τῷ ἀντιδίκῳ σου [Mt 5.25] ‘be well-minded to your accuser’ (freely ‘make friends with your accuser’). The verb εὐνοῶ ‘to be well minded’ does not appear elsewhere in the NT. 392 Vit Bubenik <?page no="393"?> The analytic formations in the perfective aspect exploiting the copula with the aorist participle are rare in Classical authors, and their aspectual value is dubious (e.g. Herodot’s οὐδέ τι νεώτερόν εἰμι ποιήσας ‘I have not been acting any differently now …’[4.127.1]). Another combination with the aorist participle involves the verb of possesion ἔχω ποιήσας. This construction (and also the combination with the perfect participle ἔχω πεποιηκώς) may be linked with the development of the ‘have’ Perfect with the mediopassive participle indicative of its aspectual uncertainty (Drinka 2017). As far as the periphrastic constructions for the monolectal perfect are concerned, in addition to the analytic modal constructions in Classical Greek, exemplified in (1) and (2), Hellenistic Greek has instances where monolectal forms of the perfect and pluperfect in the indicative alternate with their analytic counterparts: (7) γέγραπται ‘it has been written’ (very often) ~ γεγραμμένον ἐστίν [ Jn 6.31, 20.30] - ἐπεγέγραπτo ‘it had been written’ [A 17.23] ~ ἦν γεγραμμένον [ Jn 19.19] It has been also observed (Blass-Debrunner 1961: 179) that in some instances periphrasis provides “a rhetorically more forceful expression”, as in (8): (8) ἑστὼς ἐπὶ τοῦ βήματος Καίσαρός εἰμι [A 25.10] ‘I am standing before Caesar’s tribunal’ which is “better” than monolectal ἕστηκα ἐπὶ τοῦ βήματος Καίσαρoς ‘I have stood on the tribune’. Here the analytic formation emphasizes the present result of Paul’s stepping on the tribune. During the Roman period, there appeared sporadic instances of the possessive construction habeo + direct object + perfect participle in agreement with the object (examples in Brixhe 2007: 906), which could be calqued in Greek translations of official Roman texts such as Res Gestae in (9): (9) Italia autem XXVIII colonias … mea auctoritate deductas habet [Res Gestae 28.2] Ἰταλία δὲ εἴκοσι ὀκτὼ ἀποικίας ἔχει ὑπ’ ἐμοῦ καταχθείσας We may wonder whether these Latin constructions contributed to the expansion of parallel constructions in Hellenistic Greek. According to Horrocks (2010: 131) this ‘have’-Perfect is a “wholly unclassical” construction documented in the Language Contact in Antiquity. Participial Constructions 393 <?page no="394"?> writings of the historian Diodorus Siculus (during Augustus) and the essayist Plutarch (50 - 120 C.E.). This ‘latinizing’ construction disappeared in the 2 nd c. CE but it reappeared in popular varieties of Greek after the Latin conquest of Constantinople (in 1204) under Romance influence. Before the Frankish domi‐ nation there are only isolated examples which inaugurate later ‘have’-Perfects with the adjacent mediopassive participle agreeing with the object as in Digenés Akrítes ( Jeffreys 1998): (10) τὸν γρόθον του εἰς τὸ μάγουλον εἶχεν ἀκουμπισμένον [Digenés Akrítes E, 418] ‘He held his fist at rest against his cheek’ > ‘He has/ had rested his fist against his cheek’ 2.2 Periphrastic formations in Hebrew and Aramaic On the whole, the language-contact explanation of periphrastic constructions as ‘Semitisms/ Septuagintisms / Hebraisms’ in the imperfective aspect in Hellenistic Greek of biblical books remains contentious. As was shown in (1 and 2), the classical language could provide an analogical model for its extension from the perfect into the imperfective aspect, but all this would not explain its frequency in the NT. Given the fact that only the periphrasis in the Imperfect of the type ἦν διδάσκων is relatively common, we have to admit that we are not dealing with a grammatical category of the progressive aspect as in Aramaic, whose aspectual system did not possess the form corresponding to the Greek Imperfect ἐδίδασκε. The beginnings of this analytic realization of former monolectal categories lie in pre-Hellenistic centuries (cf. Aerts 1965: 5-26) but it became mu