eBooks

Greek – Latin – Slavic

Aspects of Linguistics and Grammatography

0424
2023
978-3-8233-9527-0
978-3-8233-8527-1
Gunter Narr Verlag 
Barbora Machajdíková
Ludmila Eliášová Buzássyová
10.24053/9783823395270

The volume is intended for classical philologists and a broad range of scholars working in the fields of theoretical, historical, and comparative linguistics with Ancient Greek, Latin, or Slavic languages as the primary evidence in their research. The contributions address topics ranging from issues of grammatography in a diachronic perspective to historical and comparative linguistics. They encompass both monothematic case studies and comprehensive analyses that capture a linguistic phenomenon in its entirety as well as within a broader context.

<?page no="0"?> ISBN 978-3-8233-8527-1 The volume is intended for classical philologists and a broad range of scholars working in the fields of theoretical, historical, and comparative linguistics with Ancient Greek, Latin, or Slavic languages as the primary evidence in their research. The contributions address topics ranging from issues of grammatography in a diachronic perspective to historical and comparative linguistics. They encompass both monothematic case studies and comprehensive analyses that capture a linguistic phenomenon in its entirety as well as within a broader context. Sprachvergleich Studien zur synchronen und diachronen Sprachwissenschaft Band 3 Machajdíková / Eliášová Buzássyová (Eds.) Greek - Latin - Slavic Barbora Machajdíková / Ľudmila Eliášová Buzássyová (Eds.) Greek - Latin - Slavic Aspects of Linguistics and Grammatography <?page no="1"?> Greek - Latin - Slavic <?page no="2"?> Sprachvergleich Studien zur synchronen und diachronen Sprachwissenschaft Band 3 herausgegeben von Katrin Schmitz (Wuppertal) Joachim Theisen (Athen) Carlotta Viti (Nancy) wissenschaftlicher Beirat Daniel Petit (Paris) Georges-Jean Pinault (Paris) Sabine Ziegler (Berlin) <?page no="3"?> Barbora Machajdíková / Ľudmila Eliášová Buzássyová (eds.) Greek - Latin - Slavic Aspects of Linguistics and Grammatography <?page no="4"?> All studies in this anthology went through an anonymous peer-review process. This anthology was financially supported by the Vega grants of the Ministry of Education, Science, Research and Sport of the Slovak Republic 1/ 0733/ 18 “The Concept of Ancient Grammar in the Grammatic Tradition of the 17th-18th Centuries in the Territory of Today’s Slovakia and in the Wider European Context” and 1/ 0812/ 18 “The Latin Syllable in a Diachronic and Typological Context”. DOI: https: / / doi.org/ 10.24053/ 9783823395270 © 2023 · 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 Ver‐ lages unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, 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 Heraus‐ geber: 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 CPI books GmbH, Leck ISSN 2569-2275 ISBN 978-3-8233-8527-1 (Print) ISBN 978-3-8233-9527-0 (ePDF) ISBN 978-3-8233-0426-5 (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. www.fsc.org MIX Papier aus verantwortungsvollen Quellen FSC ® C083411 ® <?page no="5"?> 7 11 39 73 97 111 141 161 195 213 287 Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anneli Luhtala Milestones in the study of syntax in antiquity and the Middle Ages . . . . Ľudmila Eliášová Buzássyová Word-formation in neo-Latin school grammar . . . . . . . . . . . . . . . . . . . . . . . Gabriela Múcsková Grammatical and grammatographical categories - accord and conflict . Ondřej Šefčík Classicism, Czech language and Jungmann . . . . . . . . . . . . . . . . . . . . . . . . . . Wojciech Sowa Γλῶσσαι κατὰ πόλεις. Greek dialects through the lenses of the ancient lexicography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Václav Blažek Greek γέφῡρα ‘dam; bridge’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Martin Masliš Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Máté Ittzés Speculo claras or speculoclaras? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reiner Lipp The Latin future tense formation of the type faxō and its Italic background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Editors & contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . <?page no="7"?> 1 The terms “grammatography” and “grammaticography” are commonly considered to be synonymous. Along with Kempgen et al. (Die slavischen Sprachen, Ein internatio‐ nales Handbuch zu ihrer Struktur, ihrer Geschichte und ihrer Erforschung, de Gruyter 2009: 1793), we understand the term “grammatography” to be an equivalent to the literacy that sets out the rules of a specific language, whereas “grammaticography” is the scientific (linguistical) representation of grammatography. Preface This anthology was originally conceived as a collection of scholarship from the Greek - Latin - Slavic: Aspects of Linguistics and Grammatography conference. 1 This event, originally scheduled for June 2020, was postponed due to the Covid pandemic to early 2021, and then it was cancelled alto‐ gether. Despite this setback, the abstracts sent to the conference promised an interesting scholarly discussion that would have been a shame to ignore. As a result, the organizers did not give up on the idea of compiling this volume from selected papers from invited participants. From the first stages of its conception, the anthology was intended as a contribution to two fields: the history of the language arts and historical and comparative linguistics. Because this anthology is not a typical conference proceedings, the individual contributions differ in terms of their content and scope. They encompass both monothematic case studies and comprehensive analyses that capture a linguistic phenomenon in its entirety as well as within a broader context. The Greek - Latin - Slavic: Aspects of Linguistics and Grammatography volume is a collection of research papers addressed to classical philologists as well as to a broader range of scholars who work in theoretical, historical, and comparative linguistics with Ancient Greek, Latin, and Slavic languages as the primary material in their research. Its broader approach invites researchers to read into their subjects and to look at the wider context of related subjects. In its first part, the anthology assembles case studies and offers researchers the results of analyses that focus on the theoretical approaches of ancient, medieval, early modern, and modern grammarians to languages and their descriptions. It illustrates how the perspective of <?page no="8"?> contemporary authors of grammars on individual language levels (morpho‐ logical and word-formation, syntactical, and lexical) has evolved over time. In the second part of the anthology, multispectral approaches are en‐ hanced by research into Indo-European languages and classical ancient languages in particular. This research, based on the historical-comparative method as its essential tool, is conducted in the broader background of the philological analysis. As a result, studies of various language phenomena from the domains of phonology, morphology, semantics, and lexicon - as well as etymology and language contact - are all included. In her paper, Anneli Luhtala fills in a gap in the hitherto little-developed research into the history of the description of the syntactic level of Latin, which was a latecomer in ancient language studies. She concentrates on the most central developments that took place in medieval syntactical theory, which was based on late-antique foundations and primarily on Priscianus. The paper by Ľudmila Eliášová Buzássyová analyses the tendencies to systematize knowledge about word-formation in neo-Latin school grammar from the sixteenth to the middle of the nineteenth centuries. Her study fo‐ cuses on the process of the separation of word-formation as an autonomous discipline as well as on the definitions of key word-formation terms and their changes over time. In her paper, Gabriela Múcsková presents several examples from Slovak where grammatical (and terminological) categoriza‐ tion is based on a grammatographical tradition rather than its current properties and abstract grammatical functions resulting from linguistic development. Ondřej Šefčík focuses on an orthographic argument between two prominent philologists - Jungmann and Nejedlý - in the cultural and political context of the national revival movement in Bohemia. The second part of the volume starts with Wojciech Sowa’s paper. Using the example of the anonymous work Glossai kata poleis, he examines the reliability of ancient lexicographical sources when studying Ancient Greek dialectal vocabulary. Václav Blažek discusses the etymology of the Greek word γέφῡρα ‘dam; bridge’ with its dialect variants. He evaluates existing etymological analyses and proposes a highly promising solution. Martin Masliš explores the implications of morphological parsing for the acquisition of semantics from a psycholinguistic perspective; looking at the Ancient Greek nominal lexemes with the suffixes -μοand -ρο-, he tests the proposition that speakers tentatively connected novel words to their possible etymological relatives because of formal similarities. Máté Ittzés 8 Preface <?page no="9"?> discusses how much the construction speculo claras ‘bright like a mirror’ should be regarded as an isolated “positive adjective + ablative” within the system of Latin gradation as a whole and whether there is any support for the assumption of a nominal compound. According to Reiner Lipp, the Old Latin s-future of the type faxō, faxis, faxit represents a durative-telic future tense formation, showing a regular thematic remodelling of a semi-thematic s-future paradigm inherited from Proto-Italic, which is based on an athe‐ matic PIE s-desiderative. He also analyses the associated categories of the subjunctive of the type faxim, faxīs, faxit and the Sabellic s-future. We hope this short anthology can play a part in developing and deepening linguistic and grammatographical research and their interfaces. We are very grateful to all contributors to the volume and to the reviewers for their critical insights. We would like to express our very sincere thanks to Professor Carlotta Viti for the inspiration to compile this publication. Our gratitude is also extended to Tilmann Bub from the Narr Francke Attempto Verlag as well as to Mária Šibalová and Milan Regec for their editorial and technical support and professional advice. Preface 9 <?page no="11"?> Milestones in the study of syntax in antiquity and the Middle Ages Anneli Luhtala Abstract: This essay presents milestones in the study of syntax in the Middle Ages, taking into account not only the development of the mainstream grammar, based on Priscian’s theory, but also alter‐ native approaches to sentence-construction, which drew inspiration from ancient textbooks on rhetoric and dialectic. In these Priscian’s approach, which largely focused on a basic clause, was expanded to cover even phrases as well as compound and complex sentences. In the twelfth century, Priscian’s tools of analysis - transitivity, agreement and government - were developed in association with the study of logic, and the new tools of description included logical word order as well as the subject - predicate division. Keywords: medieval syntactic theory, Priscian’s syntactic theory, an‐ cient rhetoric, ancient dialectic, transitivity, government, word order 1 Introduction The study of syntax is a latecomer in the field of ancient language studies. Textbooks on grammar began to be compiled in the first century BC, but they did not include a section on syntax. Varro’s theoretical treatise On the Latin Language (De lingua Latina) did contain a book on syntax, but it did not represent a distinctly grammatical theory of syntax. The fragments surviving from it are pieces of Stoic logic. It was not until the second century AD that the study of syntax emancipated from Stoic logic, when a theory of syntax began to be developed by the Greek grammarian Apollonius Dyscolus in Alexandria (2 nd cent. AD). The Apollonian theory was adapted into the <?page no="12"?> 1 For the study of syntax in medieval pedagogical grammar, see Luhtala (2013a). 2 In lectione tota sententia periodos dicitur, cuius partes sunt cola et commata (Donatus, Ars maior, Holtz 612.7-8). Institutiones grammaticae of Priscian who taught in Constantinople c. 500 AD. The last two books of the Institutiones, known as the Priscian minor in the Middle Ages, came to represent the mainstream of syntactical study in the twelfth century, when scholarship flourished in the cathedral schools of Northern France. In the thirteenth century it was an obligatory textbook in the Arts Faculties of universities. The study of syntax was not part of secondary grammar education until c. 1200. 1 This is largely due to the importance that the works of the poets had in both ancient and medieval secondary education. In the standard definition of grammar, first recorded in Quintilian’s Institutio oratoria, the grammarian’s task consists in teaching well-formed Latin and the language of the poets, and Quintilian posits a clear division of labour between the teaching of the secondary school and the higher education in the rhetori‐ cian’s school; the study of poetry belongs to the former and the language of prose texts to the latter (1.4.1). In the rhetorical tradition, prose composition had developed tools of analysis distinct from those that were gradually introduced into grammars. In Donatus’s Ars maior (c. 360 AD), a standard textbook on grammar, these rhetorical tools are briefly mentioned: “In classroom reading a complete thought is called period (periodus), its parts being cola and commata.” 2 In this essay, I will present milestones in the study of syntax in the early and high Middle Ages, taking into account not only the development of the mainstream grammar, the Apollonian-Priscianic theory, but also alternative approaches to syntax, which drew inspiration from textbooks on rhetoric and dialectic. Priscian’s theory depended heavily on logic, being mainly restricted to the analysis of the basic clause. Therefore, medieval teachers turned to ancient textbooks on rhetoric and logic when they wanted to expand the scope of syntactic description to include compound and complex clauses, and in fact, both dialectic and rhetoric offered useful tools for analysing Latin prose texts. 12 Anneli Luhtala <?page no="13"?> 3 The account of the Apollonian-Priscianic theory of syntax is largely based on my doctoral thesis (Luhtala 2000a: 146-176) as well Luhtala (2020), see also Blank (1982) and a French translation of Apollonius’s syntactical theory by Lallot (1997). 4 […] tertio, quemadmodum coniungerentur […] (Varro, Ling. 7.10; cf. 8.1). 5 Sed M. Varro in libro de lingua Latina ad Ciceronem quarto uicesimo expeditissime ita finit: ‘Proloquium est sententia, in qua nihil desideratur’ (Aul. Gell. 16.8.6-8). 2 Ancient theory of syntax 3 Varro says that his De lingua Latina consists of three sections, the first dealing with the imposition of names, the second with their inflection and the third part shows how words are joined together (7.10). 4 The third part is lost, but a fragment from it has been preserved in Aulus Gellius’s Noctes Atticae (16.8.6-8). According to Gellius, Varro defined the sentence or rather the proposition in Book 24 of the De lingua Latina as follows: Proloquium est sententia, in qua nihil desideratur ‘A proposition is a meaning, in which nothing is missing’. 5 As Taylor concludes (1987: 6), books fourteen to twenty-five of De lingua Latina did deal extensively with syntax, but probably in manner consistent with its dialectical nature. I agree with Frede (1977: 73; Luhtala 2000a: 23-24) that grammatical syntax belonged to the domain of dialectic before Apollonius Dyscolus. In the later grammatical context, the sentence was termed oratio rather than proloquium. - 2.1 Definition of sentence Priscian defines a sentence as follows: Oratio est congrua ordinatio dic‐ tionum perfectam sententiam demonstrans ‘A sentence is a congruent or‐ dering of words showing a perfect meaning’ (Inst.gram. 2.15, GL II: 53.28-29). The term congruus points at once to the well-formedness of a sentence and, more specifically, to the agreement or concord of the morphological prop‐ erties of various parts of speech when joined together within a sentence. A perfect meaning (perfecta sententia) means that a sentence includes the necessary elements and only them, and the perfection or completeness is both structural and semantic. Structural completeness is reached when a noun in the nominative case is joined to a verb to form a simple sentence, such as Aristoteles ambulat ‘Ar‐ istoteles is walking’. An incomplete expression leaves a question in the mind of the hearer. For instance, a predicate such as ‘walks’ cannot be understood, unless it is joined to a nominative case, e.g. ‘Aristoteles’. Similarly, ‘Cicero Milestones in the study of syntax in antiquity and the Middle Ages 13 <?page no="14"?> 6 These examples are from Priscian’s Inst.gram. (17.14, GL III: 116.25-27). Similar examples, using such names as Socrates, Plato, Aristotle and Cicero, were employed both in grammar and dialectic (e.g. Cicero disputat, Mart. Cap. 4.393) in Antiquity and throughout the Middle Ages. saves’ is not complete, because you must still ask ‘what does he save? ’. The sentence Cicero seruat patriam ‘Cicero saves the fatherland’ is complete and satisfies the mind of the hearer; it is therefore semantically complete. 6 In this verb-centred theory it is the verb that draws both the nominative and the oblique case into the construction, which is a complete statement. This approach reminds us of dependency rather than constituency grammar. - 2.2 Logical word order in grammar Priscian takes a minimal statement as the point of departure for what he calls the ordering of the parts of speech. This ordering is motivated by the question, why were the different parts of speech “invented”, that is, what is their “raison d’être”. The noun and the verb are the principal parts of speech, because they can form complete statements without the other parts of speech. Priscian proceeds to discuss their mutual ordering in philosophical terms, referring to the priority of substance: Ante uerbum quoque necessario ponitur nomen, quia agere et pati substantiae est proprium, in qua est positio nominum, ex quibus proprietas uerbi, id est actio et passio, nascitur (Inst.gram. 18.5, GL III: 211.21-25). “The noun necessarily precedes the verb because acting and being acted upon are properties of substance and name-giving pertains to substances, on which depends the property of the verb, namely action and undergoing of action.” Substance is the first of Aristotle’s ontological categories, and it is the only category of the ten which can subsist on its own, whereas action is an acci‐ dent of a substance, which cannot subsist without a substance. Sentences are said about persons (and things) in the external world. The persons (and things) are substances expressed by means of nouns, whereas their actions and the undergoings of action are signified by verbs. Since substances are logically and ontologically prior to the accidents, the noun is consequently prior to the verb. As regards the ordering of the other three parts of speech, only the posi‐ tioning of the adverb is relevant for the future discussions of word order. 14 Anneli Luhtala <?page no="15"?> According to Priscian, the name of this part suggests a close association with the verb (ad + verbum) which is reflected in its syntactic position; it is placed before the verb just like adjectives are usually placed before nouns (Inst.gram. 15.39, GL III: 89.14-17). This principle, which is alien to the nature of Latin prose, was occasionally adopted by medieval scholars. - 2.3 Transitivity Sentences expressing states of affairs in the external world are depicted in grammar as involving transitive or intransitive action. That is, a sentence expresses a process involving one or two ‘persons’ (persona) or referents, an agent and a patient. This semantic description is related to a morphological aspect, namely the concord or agreement of the inflecting parts of speech. Priscian’s discussion on agreement discloses the fundamental principle, ac‐ cording to which agreement occurs in a construction, when the constituents refer to the same referent (persona). In an intransitive sentence involving only one actant, the verb and the noun/ pronoun show agreement in number and gender, and both parts of speech pertain to the same referent. On the other hand, a transitive sentence exhibits a change of referent, and therefore the two case forms involved in this construction fail to show agreement: “The parts of speech exhibit various inflectional patterns, some being inflected for case and number - e.g. the noun, the pronoun and the participle - others for number and person - e.g. verbs and pronouns - and yet others for tense, like the verb and participle […]. Therefore, in the process of sentence-construction, the inflecting classes of words must be joined to each other in matching agreement, so that the singular is joined to the singular and the plural to the plural, when the constituents pertain to the same referent intransitively, as in ego Priscianus scribo intellegens, and nos oratores scribimus intellegentes. Whenever transitive states of affairs are at issue, it is possible to use different numbers, e.g. docemus discipulum et docemus discipulos […]” (Inst.gram. 17.153, GL III: 182.26- 183.12). 3 Ancient rhetorical theory - 3.1 Periodic and continuous style Several ancient authoritative works on rhetoric established two kinds of prose style, the one being periodic (genus distinctum) and the other contin‐ uous (genus continuum). In each style speech consists of a varying number Milestones in the study of syntax in antiquity and the Middle Ages 15 <?page no="16"?> 7 The terms genus continuum for periodic and genus distinctum for non-periodic style were used in the St. Gall Tractate, edited by Grotans and Porter (1995: 60-61). In Antiquity, the terms oratio perpetua stood for periodic or continuous style and compositio distincta or fracta for non-periodic style (for further details, see Scaglione 1972: 25-30). 8 For cola, commata and periodus, see Quintilianus (9.4.22, 122-130), Martianus Capella (5.527-528), and Grotans (2006: 173-178). 9 “But the period […] consists of two or three or four members, sometimes even of six […]” (transl. by Stahl et al. 1977: 199). 10 For the authenticity of this text, see de Jonge (2008: 40). of segments, which are called cola and commata, and their combination expresses what is called sententia, that is, a thought, an idea or a meaning. 7 Thus, the basic unit of analysis in both grammar and rhetoric is ultimately semantic, the sententia. However, the two arts deal with completely different units of speech. A rhetorical period deals with narrative passages of varying length and complexity, whereas a sentence in grammar (oratio) expresses a meaning, which is minimally perfect (sententia perfecta). The idea of complete meaning is also involved in the definition of the two basic units of a rhetorical period, cola and commata. A comma cannot express a complete meaning on its own, whereas a colon can. Thus, the colon comes close to the idea of a main clause, but even commata can sometimes be short main clauses, for instance commands and exclamations. The comma is used in both styles, but in the periodic style it must always appear together with a colon. 8 In the distinct style, the constituent parts of a passage, cola and commata, are often joined by means of subordination - a concept which was in use neither in Antiquity nor in the Middle Ages. According to Martianus Capella, a period could maximally consist of six cola (5.529). 9 - 3.2 Word order in rhetoric The idea that clarity in prose composition depends on a natural ordering of its components goes back at least to Demetrius in the second century BC. 10 According to the natural ordering, sentences should begin with the nominative case and be followed by the verb; the remaining words come in due succession. Like Apollonius and Priscian, Dionysius of Halicarnassos teaches us to place nouns before verbs, because in the nature of things nouns indicate substances, and substance takes precedence over its accidents. But unlike Apollonius and Priscian, Dionysius places action before its circum‐ 16 Anneli Luhtala <?page no="17"?> 11 […] et quamuis natura illud exigat ut primo nomen et postea uerbum dicatur, ut dictum est, non desinit uerum esse proloquium, etiamsi dicas ‘animal est omnis homo’ (Mart. Cap. 4.392). stances and therefore verbs are placed in front of the adverbs (On Literary Composition, Roberts (ed.) 1910: 98-100; see also Scaglione 1972: 78-79). Yet another guideline offered by Demetrius is to observe the chronological sequence of events, which dictates for example that nouns should precede their appositions or adjectives. Nouns should also come before the pronouns referring to them (On Style, § 199, Roberts (ed.) 1902; see also Grotans 2006: 164-165). - 3.3 Martianus Capella: expanding the logical proposition The idea of a natural ordering of the noun and the verb in a basic clause is also present in Martianus Capella’s (c. 360-428 AD) account of the logical proposition. In Book IV of his popular encyclopaedia, The Marriage of Philology and Mercury, he discusses the proposition Omnis homo est animal ‘Every man is an animal’, which follows the natural order, with the subject preceding the predicate. However, he points out that an inverted word order is equally valid: “Although nature demands that the noun should be uttered first and then the verb, it does not cease to be a true proposition, even if you say ‘An animal is every man’” (§-392, transl. by Stahl et al. 1977: 136-137). 11 In Martianus’s encyclopaedia, which was one of the most popular ancient texts circulating in the early Middle Ages, medieval masters were able to find alternative approaches to syntactical analysis. Martianus starts describing the proposition by comparing it with other kinds of sentences, such as questions and commands. Not any combination of a noun and a verb forming a sentence also constitutes a proposition, he argues, but only a combination that can be affirmed or denied. This can be achieved by the union of the third person of the verb and the nominative case, as in Cicero disputat ‘Cicero is discussing’. Martianus defines the basic components of the proposition Cicero disputat in the standard manner (§ 393, transl. by Stahl et al. 1977: 137): “For what it is we are talking about is ‘laid down’ as subject; and what can be understood about it is ‘declared’ or ‘predicated’ of it.” He then sets out to discover, how the two parts of the proposition can be expanded with various optional components. The constituents in the nominative case (e.g. Romanus) pertain Milestones in the study of syntax in antiquity and the Middle Ages 17 <?page no="18"?> 12 Quicquid accesserit huic sententiae, cui parti accedat, diligenter uidendum est. Nam sunt proloquii partes duae: quae in nomine, una, subiectiua dicitur, quae in uerbo, altera, declaratiua. Subicitur enim quid sit, et declaratur quid de illo possit intellegi. Cum ergo dicimus ‘Cicero disputat’, si accedat huic sententiae ‘in Tusculano’, declara‐ tiuae accessit, si accedat ‘Romanus’, subiectiuae; item si accedat ‘prudenter et copiose’, declaratiuae; item cum dicitur ‘cum Catone’, declaratiuae accessit. Quicquid igitur nominatiuo casu accedit, subiectiuae accedit; quicquid declaratiuae accedit, uariis casibus et modis accedit. Nam subiectiuae non possunt alii casus accedere, declaratiuae nonnisi uarii excepto nominatiuo (Mart. Cap. 4.393). 13 For the study of syntax in the 9 th and 10 th centuries, see Luhtala (1993: 161-179). to the subject part (pars subiectiua) while those in the oblique cases, such as ‘in Tusculanum’, ‘with Cato’ and ‘prudently and copiously’, belong to the predicate part (pars declaratiua). Whatever is added to the predicative part, takes the form of different [oblique] cases, he concludes, whereas only the nominative case can occur in the subjective part. 12 4 Syntactic theory in the early Middle Ages Priscian’s Institutiones grammaticae, poorly known during the first medi‐ eval centuries, began to circulate more widely from the late 8 th century onwards, and its first 16 books, known as the Priscian maior, began to receive commentary. The last two books on syntax, the Priscian minor, only became widely known in the 12 th century. 13 The tools of syntactical analysis offered by Priscian were not sufficient for early medieval learners of Latin who were not native speakers of Latin and had difficulty interpreting and analysing Classical and Christian literary works. A Classical Latin prose text makes use of complex sentences, with inverted word order and heavy participial constructions and subclauses, which did not lend themselves to interpretation without careful analysis. Interaction between the three arts of the trivium, grammar, rhetoric and dialectic, was at the heart of the Carolingian renaissance, being encouraged especially by Alcuin, the primus motor of this educational reform. Dialectic was studied through Boethius’s translations of Aristotle’s so called “old logic” and popular encyclopaedic works, such as Isidore of Seville’s Etymologiae and Martianus Capella’s Marriage. Both encyclopaedias included a short treatise on all the Liberal Arts, in which early medieval teachers were able to find inspiration for analysing Latin sentences. After the Carolingian reform, 18 Anneli Luhtala <?page no="19"?> this new approach was pursued actively in St. Gall in the ninth and tenth centuries. - 4.1 Aristotle’s ontological categories Isidore of Seville introduces the ten Aristotelian categories in a short treatise on dialectic included in the Etymologiae. In his Categories (Categoriae) Aristotle presents a set of ontological categories, which, as he explains, is a classification of things ‘said without combination’: substance (e.g. man, horse), quantity (e.g. four-foot), quality (e.g. white), relation (e.g. double, half), place (e.g. in the Lyceum, in the market-place), time (e.g. yesterday, last year), position (e.g. is lying, is sitting), state (e.g. has shoes on), action (e.g. cutting, burning), passion (e.g. being cut, being burned) (Categoriae 2-4). Propositions, such as ‘Socrates walks’ or ‘Socrates is white’, are combinations of ontological categories. It is obvious that Aristotle used language as a clue to his theory of ontolog‐ ical categories. Aware of the affinities between ontological and grammatical items Isidore combined the ontological categories into a complex sentence, including nominal and adverbial phrases as well as embeddings: ‘Augustine [substance], a great [quantity] orator [quality], the son of that person [rela‐ tion], standing [position] in the temple [place], today [time], adorned with a headband [state], having a dispute [action], gets tired [passion]’ (Plena enim sententia de his ita est: Augustinus, magnus orator, filius illius, stans in templo, hodie, infulatus, disputando, fatigatur, Etym. 2.26.11; see Luhtala 1993: 150-151). Isidore’s identification of the Aristotelian categories with the various constituents of a complex sentence inspired a number of early medieval authors. Alcuin copied this passage into his handbook on dialectic (De dialectica, PL 101: 962C), and it was also used in the Donatus commentary of the ninth-century grammarian Sedulius Scottus (In Donatum maiorem 62.50-61). Furthermore, Martin of Laon exploited it in his teaching of the Liberal Arts (Contreni 1981: 35). - 4.2 The Seven Circumstances The ancient rhetorical handbooks offered a set of items, which invited com‐ parison with grammatical and ontological categories, namely the rhetorical formula of the ‘seven circumstances’ (septem circumstantiae or periochae), including persona ‘person’, res ‘thing’, locus ‘place’, causa ‘cause’, tempus Milestones in the study of syntax in antiquity and the Middle Ages 19 <?page no="20"?> 14 Omnis res argumentando confirmatur […]. Haec quidem omnis res ex illis septem locis argumenta contrahit: quis, quid, cur, ubi, quando, quemadmodum, quibus adminiculis (Rhetores minores 213.43-214.1). Martianus Capella, De Nuptiis 5.52. Narrationis etiam elementa sunt septem: persona, causa, modus, locus, tempus, materia, res (Dick (ed.) 1978: 274.21-275.1). In Rhetores minores, only six circumstances are included in Martianus’s account of rhetoric (Halm (ed.) 1863: 486). 15 Anna Grotans argues in favour of this attribution convincingly (2006: 157). 16 For the reintroduction of the subject - predicate distinction in the twelfth century, see Rosier-Catach (1994). ‘time’, modus ‘manner’, and materia or facultas ‘means’, which often ap‐ pear in the interrogative form quis ‘who’, quid ‘what’, cur ‘why’, quomodo ‘how’, quando ‘when’, ubi ‘where’, quibus facultatibus ‘by what means’. This formula, probably taken over from the rhetorical manual of Marius Victorinus, 14 consists of argumentative loci, which could be used in arguing a case. They served as the point of departure for an analysis of complex sentences in a pedagogical text deriving from the monastery of St. Gall (Grotans/ Porter (eds.) 1995; see Luhtala 1993: 174-176). The author of this treatise, composed in the late ninth or tenth century, was probably Notker Labeo. 15 In accordance with the title, Quomodo VII circumstantie legende ordinande sint (‘How the seven circumstances of things are to be ordered in read‐ ing’), the author sets out to expand a minimal statement, ‘Cicero disputes’, which Martianus Capella had used to illustrate the logical proposition. Notker’s method consists in asking a series of questions: ‘Who did what? When? Where? How? By what means? ’ He answered the question ‘Who did what? ’ using both rhetorical and dialectical terminology. From Martianus he adopted the division of a basic clause into subject (subiectiuum) and predicate (declaratiuum), also identifying the two components ‘Cicero’ and ‘disputes’ as persona and res, the first two items of the rhetorical circum‐ stances (Luhtala 1993: 172-173; Grotans/ Porter (eds.) 1995: 44-46). 16 Having stated that the subject is in the nominative case and the predi‐ cate in the indicative mood, Notker goes on to expand both constituents with various optional elements, accumulating them occasionally into huge complexities. He first elaborates the predicate part with the five optional ‘circumstances’: “Cicero is arguing in Tusculanum for a long time in a remarkable way for the common good with great brilliance” Cicero disputat in Tusculano multo tempore mirum in modum propter communem utilitatem magna excellentia ingenii (Grotans/ Porter (eds.) 1995: 46-48). 17 Then the sub‐ 20 Anneli Luhtala <?page no="21"?> 17 I have introduced capital letters and punctuation marks into the text of the edition in order to make it more readable. 18 Notker adopted the terms used by Martianus: subiectiuum and declaratiuum (Mart. Cap. 4.93). For details, see Luhtala (2000b: 124-125) and for the reintroduction of the subject - predicate division in the twelfth century, see Rosier (1994). ject part is similarly extended with a number of constituents: “Cicero, whose father was a Roman knight, distinguished with consular rank, being of the royal Volscan clan, discusses the nature of the gods, which is awe-inspiring and unknown to mortals” Cicero patre natus equite Romano de regno genere Velscorum rhetor eximius et consulari dignitate praeclarus ipse disputat de natura deorum, quae mirabilis et [i]gnota mortalibus est.  18 Inspired by Martianus, Notker dwells at some length on the ordering of the components in a clause. He maintains that the subject is the foundation on which the predicate is built, and thus, according to the natural order, the subject precedes the predicate as in Deus fecit hominem ‘God made the man’. However, an inverted order, Hominem fecit Deus has the same truth value and meaning. This is, however, not the case in all Biblical contexts. A heretical view results, if one - following the natural or logical order - regards Deus ‘God’ as the subject of the clause Et Deus erat uerbum in the opening passage of John’s Gospel. Reason will never allow ‘Deus’ to be set in first place as the subject and to predicate of God ‘uerbum’, Notker argues. The inverted word order Et uerbum Deus erat ‘And the Word was God’ renders the correct sense (Grotans/ Porter (eds.) 1995: 47). Similarly, the Creed where we confess our faith by saying Ita Deus pater, Deus filius, Deus spiritus sanctus, exhibits inverted word order. ‘Father’, ‘son’ and ‘the Holy Spirit’ are the subjects of these phrases which can be spelt out as follows: Ita pater Deus est, ita filius Deus est, ita spiritus sanctus Deus est (Grotans/ Porter (eds.) 1995: 48). Notker concludes his discussion on word order by stating a pragmatic principle: the organization of the sentence can be regulated by the will (arbitrium/ uoluntas) of the author. To prove his point, he quotes Biblical sentences initiated by each of the circumstantiae in turn (for the examples, see Grotans/ Porter (eds.) 1995: 52). Milestones in the study of syntax in antiquity and the Middle Ages 21 <?page no="22"?> 19 In dealing with continuous style, I have made use of Grotans/ Porter (eds.) (1995) and Grotans (2006) and many of Grotans’s translations of Latin sentences into English. 4.3 The Continuous Style 19 Notker illustrates the continuous style by means of a complex sentence which includes all the seven circumstances, which, as he points out, happens only rarely in a single sentence; the source of this piece of historical narra‐ tive is not known (Grotans/ Porter (eds.) 1995: 62-66; Grotans 2006: 62-66). He first presents this passage as it has been composed in the continuous style and then rewrites it organizing it into a pedagogical order. Its two orders have the same meaning, but the reorganized ordering is the one suitable for teaching, he explains. The author is remarkably aware of the fact that a complex sentence consists of slots which can be filled either with one word or with various complex units of speech. Tempore quo Siluius Aeneas regnauit in Italia, templum domino toto orb[e] famosissimum rex Salomon filius Dauid, cui similis in sapientia nullus ante eum uel post inventus est, quia tabernaculum in Sylo, ubi erat arca, angustum populo uisum est ad orandum et sacrificandum Hierosolimis, loco quem ad hoc elegit dominus ex lapidibus preciosis sectis et quadratis et lignis cedrinis ex libano monte per Iram regem Tyri administratis prospere plusquam credi potest construxit et ad perfectum elimauit. Notker reorders this passage as starting from the subject part: A subiectivo ‘From the subject’. Rex Salomon filius Dauid cui similis in sapientia nullus ante eum uel post inuentus est ‘King Solomon, son of David, whom no one before or after equalled in wisdom’. Ecce persona ‘This is the person’. This entire unit consisting of a noun (Salomon), its apposition (rex), a genitive complement (filius David), and a relative clause makes up a highly complex noun phrase representing the ‘person’ or ‘subject’. It is followed by action or the predicative part. Sequitur actio ‘Action follows’: construxit templum a domino, toto orbe famosissimum et ad perfectum elimauit ‘built a temple for the Lord, the most famous in the whole world, polishing it into perfection’. The predicate part consists of two paratactic clauses. The first includes a verb (construxit) and its object (templum) with an adjectival complement (toto orbe famosis‐ simum), an indirect object (a domino), and the second clause contains a verb (elimauit) and an adverbial phrase (ad perfectum). 22 Anneli Luhtala <?page no="23"?> Sequitur modus, ne aduerbium longe sit a uerbo ‘Then follows the adverb of manner, in order that the adverb should not be placed too far from the verb’. In prospere plusquam credi potes ‘more successfully than one can imagine’, the adverbial part includes a comparative subclause. Ubi? ‘Where? ’ Hierosolimis loco quem ad hoc elegit dominus ‘In a place in Jerusalem chosen for this purpose by the Lord’. In the adverbial phrase two place names, ‘Jerusalem’ and ‘place’, are followed by a relative subclause. Unde? Qua materia? ‘Of which material? ’ Ex lapidibus preciosis sectis et quadratis et lignis cedrinis ex libano monte per Iram regem Tyri administrates ‘From precious stones cut and squared from the cedar trees of Mount Lebanon supplied by Ira, King of Tyre’. The adverbial part is a noun phrase, consisting of a noun and an adjectival complement, ex lapidibus preciosis, into which is incorporated a complex embedding with three past participles, each with complements of their own. Quando? ‘When? ’ Tempore quo Siluius Aeneas regnauit in Italia ‘At the time when Silvius Aeneas reigned in Italy’. This adverbial phrase incorpo‐ rates a relative clause. Quare? ‘Why? ’ Quia tabernaculum in Sylo ubi erat arca angustum populo uisum est ad orandum et sacrificandum ‘Because the temple at Silo, where the arch was, was regarded by people as too small for praying and scarifying’. This unit is expanded with a causal subclause, into which a relative clause is embedded. - 4.4 Distinct style According to Notker, the Bible uses the distinct style (Grotans/ Porter 1995: 60-61). In this style, a colon frequently occurs without any comma, as in Deum nemo uidit umquam ‘No man hath seen God at any time’. This kind of clause is called monocolon by Martianus Capella. The following Biblical passage including six cola is regarded as a specimen of distinct style by Notker, although the six cola are organized paratactically: Omnes manus dissoluentur, et omnia genua fluent aquis, et accingent se ciliciis, et operiet eos formido, et in omni facie est confusio, et in uniuersis capitibus eorum caluitium (Ezechiel 7: 17). “All hands shall be feeble, and all knees shall be weak as water. They shall also gird themselves with sackcloth, and horror shall cover them, and shame shall be upon all faces, and baldness upon all their heads.” Milestones in the study of syntax in antiquity and the Middle Ages 23 <?page no="24"?> 20 Peter’s Summa depends heavily on the earlier version of William of Conches’s commentary on the Priscian Major (1125-1130) (see, for instance, Fredborg 1973 and 1988: 179, 181-182, 184). William taught grammar at Chartres in the second quarter of the 12 th century. The approaches discussed above are remarkable not only in that they expand the scope of syntactic description considerably but also in that they largely draw their examples from Biblical literature. Both of these principles were abandoned when syntactic theory flourished in the twelfth century. Then made-up stock examples tended to be used in syntactic analysis. 5 Syntactic theory in the high Middle Ages The twelfth century was the age of the flourishing of the cathedral schools in Northern France and it is then that dialectic or logic became a major subject. The Priscian minor became widely available and began to dominate the study of syntax. By now, the connection of syntactic description with rhetoric, established in the preceding centuries, was lost, and syntax was studied in association with logic in the mainstream of medieval grammar. Syntactic theory made significant advances in the Middle Ages. I will illustrate these developments with reference to two twelfth-century works. One of them is an independent grammatical compendium composed by Hugh of St. Victor (1096-1141), head of the school of St. Victor in Paris, and entitled De Grammatica (c. 1120). It was designed to be used in the education of clergy. The other work is Peter Helias’s Summa super Priscianum (c. 1140) - a systematic textbook studied at an advanced level which became highly popular throughout Europe up until the early 15 th century. Peter taught in Paris and Poitiers from c. 1135 to c. 1160, specializing in grammar and rhetoric. 20 Both works have incorporated the latest developments in the study of logical semantics into the structure of their works. Taking the simple sentence as the point of departure for their syntactic analysis, they exploit the key notions of Priscian’s theory, namely concord, government, and transitivity. By the mid-twelfth century, a basic clause was regularly described in terms of a logical word order, the nominal constituents being placed before (ante) or after the verb (post verbum). The subject - predicate division absent from Hugh’s treatise (c. 1120) was employed in Peter’s Summa (c. 1140), and 24 Anneli Luhtala <?page no="25"?> 21 For the concept of transitivity in the Middle Ages, see also Rosier (1984) and Kneepkens (1990). from the mid-twelfth century onwards it was a regular part of medieval syntactic theory. While adhering closely to Priscian’s doctrine, Hugh and Peter frequently interpret it slightly differently from Priscian. - 5.1 From clauses (orationes) to phrases (constructiones) Instead of opening the syntactic section of his grammar by defining a sentence (oratio), Hugh focusses on a minor constituent of a sentence, a construction (constructio): Constructio est dictionum congrua in oratione ordination “Construction is a well-formed ordering of words within a sen‐ tence” (De grammatica 106.917). By dropping out the idea of ‘signifying a complete meaning’ (perfectam sententiam significans) from Priscian’s defi‐ nition, the focus is shifted from clauses to phrases. The focus in the study of syntax has indeed shifted from a complete sentence to the components of which it is formed. A similar shift of emphasis from clauses to phrases takes place in Hugh’s interpretation of Priscian’s theory of transitivity 21 : he describes a transitive sentence with two actants, such as Socrates percutit Platonem ‘Socrates is beating Plato’ as consisting of two constructions, intransitio and transitio (De grammatica 106.944-107.962). The first is the intransitio between a noun in the nominative case and a verb (Socrates percutit) and the second is the transitio between the verb and the noun in the oblique case (percutit Platonem). Peter Helias does not fully agree with the kind of analysis proposed by Hugh and other earlier grammarians. The above analysis of the sentence Socrates legit Platonem is mistaken, Peter argues, because constructio can be understood in different ways. Firstly, it is understood “passively”, when we say that Socrates legit is an intransitive construction and that legit Platonem is a transitive construction; and secondly, when we say that Socrates legit Platonem is a transitive construction, we are talking about a construction which is a perfect or “constructed” sentence (oratio constructa) (Summa super Priscianum 900.73-90). Peter restores the concept of sentence (oratio) into this discussion, labelling it oratio constructa. Peter is aware of the confusion prevailing in the use of syntactic metalan‐ guage, and he reports views held by earlier grammarians. Some scholars Milestones in the study of syntax in antiquity and the Middle Ages 25 <?page no="26"?> think, he explains, that oratio is intended only to describe a perfect sentence (oratio perfecta), such as homo currit ‘a man is running’. Some others maintain that the sentence lego ‘(I) am reading’, in which the subject is understood in the verb, is oratio according to meaning but not quantity, that is, it has an implicit rather than explicit subject pronoun. By contrast, homo currit ‘a man is running’ is oratio according to both meaning and quantity, and is called oratio constructa, that is, it includes both an overt subject noun and a predicate. They also maintain that in domo ‘in the house’ cannot be called oratio, whereas homo albus ‘a white man’ can. This is because, they argue, the latter signifies a combination of substance and accidents and is therefore called an imperfect oratio whereas the former does not (Summa 177.38-48). It was customary in the Middle Ages to distinguish two additional sub‐ types of constructions derived from Priscian. In a reflexive construction, such as Socrates diligit se ‘Socrates loves himself ’, the action inhering in Socrates is directed towards himself rather than another person, as Peter explains. Priscian says that this sentence fails to exhibit a transition from one person to another, and therefore Peter prefers to call it a distinct type rather than a subtype of a transitive construction (Summa 899, 49-53). The retransitive construction, such as Socrates rogat Platonem, ut misereatur ‘Socrates asks Plato to have pity on him’ is described as involving two transitions: the action is first directed from an agent to a patient, from Socrates to Plato, and then from the patient towards the agent (Summa 899, 67-900,2). In modern terms, this is an instance of indirect reflexive, that is, a reflexive pronoun in a subclause refers to the subject of the main clause. This is the only type of a subclause that was regularly discussed in medieval treatises on grammar. Peter concludes this discussion by saying that Priscian’s definition of a sentence is generic pertaining to a perfect sentence, such as homo currit ‘a man is running’ but since oratio is a genus, it can be divided into species, which include, for instance, an imperfect oratio, such as homo albus ‘a white man’ (178.66-179.84). These twelfth century discussions show that the distinction between a sentence and a phrase was a topical issue. The closest that these medieval scholars came to the concept of “phrase” was the notion of an imperfect oratio which was a species of the genus oratio. To distinguish between the two phrases homo albus ‘a white man’ and in domo ‘in the house’ in terms of the Aristotelian distinction between substance and accidents 26 Anneli Luhtala <?page no="27"?> seems confusing to us. With hindsight, the source of the confusion is the mixture of criteria - semantic, syntactic, morphological and ontological - used in defining grammatical concepts. - 5.2 Transitivity and noun phrases Hugh extends the theory of transitivity to cover even noun phrases. Thus, intransitio takes place between two or more nominative cases (Scipio Affri‐ canus) or between oblique cases (Scipionem Affricanum), or between a verb and a nominative (Scipio uicit); in each case both constituents point to the same referent. Transitio takes place between a verb and an oblique case (uici Hannibalem), or between a nominative and an oblique case (uincens Hanni‐ balem), or between two oblique cases (parentem Hannibali), whereby the two constituents point to two different referents (De grammatica 107.973- 108.989). Priscian related the criterion of (co)referentiality to the concord or agreement of the morphological features of the constituents. This aspect is absent from Hugh’s presentation. A similar analysis is present in Peter’s Summa but it also includes a novel type of transitive construction, namely one involving prepositional phrases, as in Socrates sedet ad portam ‘Socrates is sitting at the gate’. This sentence does not exhibit a proper transition, Peter argues, because ‘Soc‐ rates is sitting’ is an intransitive construction but is nevertheless ‘somehow’ (quodammodo) transitive because sedet ‘sits’ is associated with an oblique case associated with a preposition (ad portam). It involves a transition from one person to another, because porta is a person according to a wide interpretation of this term. Peter concludes, however, that this sentence is improperly (inproprie) called transitive (Summa 898.39-48). - 5.3 Determinatio When discussing the order of the parts of speech, Hugh draws a distinction between the principal parts of speech which can form a perfect construction (constructio perfecta) on their own, namely nouns, verbs and pronouns, and the rest of the parts of speech that can only modify or determine (determinant) the other parts; they are called adiumenta, that is, ‘assistance’ Milestones in the study of syntax in antiquity and the Middle Ages 27 <?page no="28"?> 22 Cetera omnia adiumenta sunt, quibus perfectus sensus non constat sed expletur. Determinant enim quedam circa oratione, et non constat ex hiis oratio quibus sublatis non destruitur (De grammatica 106.932-934). 23 This term is used in an anonymous commentary on the ‘Priscian minor’ from the 12 th century, entitled Glosa Victorina, as concerning the preposition [Praepositiones] significant enim habitudines quasdam, in illis rebus cum quarum significationibus iungantur et tunc easdem res quoquo modo, et determinatur incerta significatio earum per uoces ei addita (p. 24). It is preserved in one manuscript, Paris, Bibliothèque de l’Arsenal 910, f. 133ra-140ub (12 th cent.), Fredborg (ed.) (2011). 24 The term ‘object’ was not used in the Middle Ages. It is included in the predicate part of the sentence, as we can see from the quotation from John of Genoa’s Catholicon in footnote 26: Verbum uero principale cum toto illo quod sequitur est ‘appositum’. or ‘help’ (De grammatica 106.929-939). 22 The term determinatio was not part of Priscian’s vocabulary; in this context Priscian used the term consignifi‐ care, meaning that these parts of speech can signify only when associated with the principal parts of speech. The term determinatio was adopted by medieval scholars from Boethius. He had employed it in his ‘On Division’ (De diuisione), explaining that it removes the doubt that exists in the mind of the hearer as to the meaning of a word (PL 64 889A-B; Kneepkens 1978: 127). 23 The medieval scholars defined it as follows: dictionem modificare et quodam modo restringere ‘to modify a word or somehow to restrict it’ (Robert Blund, quoted by Kneepkens 1978: 139). Some teachers applied the term determinatio to the relation between a verb and the nominative case noun in a basic clause. - 5.4 Government (regimen) According to Priscian’s verb-centred approach, all verbs ‘demand’ (exigit) or ‘need’ (requirit) one or two nouns to complete their meaning. A systematic theory of government was developed in the 12 th century, whereby Priscian’s terminology began to be replaced by a theory of government (regimen). Peter explains that government has to be understood metaphorically (cf. Kneepkens 1978: 122-123) and uses the terms exigit ‘demands’ and regit ‘gov‐ erns’ synonymously throughout his exposition. It was generally accepted that the verb governs its second actant, ‘the object’ 24 , but the relation between the verb and its first actant was a matter of debate. Some of Peter’s contemporaries defined government, as taking place “when a word causes another word to take on (assumit) a particular case in construction in order to determine its own meaning (ad determinationem sue 28 Anneli Luhtala <?page no="29"?> 25 Dictionem regere dictionem non est aliud quam unam dictionem assumere aliam in constructione ad determinationem sue significationis (Summa 1050.93-95). 26 In a sentence homo albus currit, the meaning of homo remains too vague unless we add an adjective to the noun: […] nimis uaga esset locutio, ideoque ad determinationem significationis necesse est addere ipsi substantiuo adiectiuum nomen (Summa 1053.71- 77). 27 Dictionem regere aliam dictionem nichil aliud est quam trahere secum eam in con‐ structione ad constructionis perfectionem, non autem dico ad significationis determi‐ nationem (Summa 1051.24-26). significationis)”. 25 This definition is at once morphosyntactic and semantic. It is morphosyntactic when we say that the verb governs the nominative case, and it is semantic, when we explain that a verb takes on a nominative case in a construction in order to determine its own meaning, e.g. ‘Socrates is running’ (Summa super Priscianum 1049.86-1050.101). Peter agrees with his contemporaries in that the verb governs the nomi‐ native in a basic clause but not in order to determine its meaning but in order to complete its construction. He draws a terminological distinction between government, which is a morphological relation, and the semantic relation of determination. Peter explains that the verb governs oblique cases in order to perfect its construction. Determination is at issue in a construction such as homo albus ‘a white man’, where homo does not govern albus but rather determines or modifies its meaning. 26 Peter prefers to define government as follows: “a word governs another word by drawing it into a construction with itself in order to complete the construction rather than in order to determine its signification”. 27 Peter’s position differed from the view held by his teacher William of Conches, according to whom the nominative enters into a construction independently in its own right (Kneepkens 1987: 132). Both Hugh and Peter deal with a large number of relations whereby nouns and verbs govern oblique cases, listing them by their semantic or syntactic force. For instance, in filius Herculis ‘the son of Hercules’ the genitive is gov‐ erned by the nominative; the nominative signifies possession and the geni‐ tive denotes the possessor (Summa 1018.77-80, 1017.53-60). Peter frequently makes use of implicit elements in interpreting syntactic constructions. In Priscian’s footsteps, he maintains that when the nominative requires (exigit) the genitive case, it is necessary to understand an implicit substantival verb or its participle in the construction, as in filius Herculis (sum or ens) ‘(I am) the son of Hercules’ (1018.77-80, 1025.44-50). In the noun phrase magne uirtutis uir ‘a very virtuous man’, the genitive is governed by the nominative Milestones in the study of syntax in antiquity and the Middle Ages 29 <?page no="30"?> 28 Nomen enim significat rem suam supponendo eam alicui, et ita non significat rem, ut supponitur rei significate per uerbum, quamuis significet rem de qua dicitur res uerbi, sed non ad hoc significat rem eam (Summa 1052.54-56). uir. The nominative signifying possessor demands a genitive signifying that which inheres in it but expresses praise or blame rather than possession (Summa 1021, 36-47). That all such nominatives signify possession can be shown by resolving (resoluere) these expressions into the verb ‘to have’ [and an object]: uir magne uirtutis amounts to uir habens or possidens magnam uirtutem (1021.52-1022.56). Priscian also frequently resorted to paraphrases in describing such constructions (e.g. Inst.Gram. 18.13, GL III: 214.5-12). - 5.5 Subject (suppositum) and Predicate (appositum) Peter agrees with Priscian’s view that the parts of speech were ‘invented’ with view to their roles in sentence-construction. Peter relates, however, the ‘raison d’être’ (causa inuentionis) of the noun and the verb to their functions as the subject (suppositum) and predicate (appositum) in a perfect sentence. In every perfect sentence, Peter explains, something is said of something. Nouns were invented to express what the sentence is about, whereas the verb was invented to specify that which is said about the noun (Summa 1051.24-32). 28 The division of the basic clause into subject and predicate offered an alternative approach to sentence-construction whereby the verb did not draw the nominative into a construction. It was rather the noun, the subject, that entered into a construction first, and the verb predicated something about it. This was one of the most important syntactic concepts adopted from Aristotelian logic in the Middle Ages. Another novelty of philosophical origin was the role assigned to the substantival verb esse ‘to be’, which had been peripherical in Priscian’s theory. The nature of this verb was a major topic for both philosophers and grammarians. Unlike ordinary verbs, Peter argues, the verb esse does not signify action but mere substance, but it signifies in the manner of the verb, that is, it is said of another sharing tense and verbal endings like ordinary verbs. Substance unites and joins the accidents, and therefore sum is a copula, because it signifies a substance which joins the others into it (Summa 201.43-49). 30 Anneli Luhtala <?page no="31"?> 29 Et scias quod quicquid precedit uerbum principale uel intelligitur precedere est ‘supposi‐ tum’. Verbum uero principale cum toto illo quod sequitur est ‘appositum’ […]. Hic etiam nota quod si unum uerbum ponitur in oratione, illud est principale, si duo uel plura copulata per copulatiuam coniunctionem uel disiunctiuam, illa uerba sunt equaliter principalia, ut ‘Petrus legit et disputat’, ‘Martinus legit uel disputat’. Verbum uero relatiui non potest esse principale, et ponitur hoc relatiuum ‘qui’ in eadem parte orationis cum antecedente cum eo quod in constructione ponitur respectu illius, ut ‘Petrus qui est studens Bononie disputat’, hoc ‘disputat’ solum est appositum, f. 51rb, online p.-108. 30 The term clausula principalis occurs on pages 93-94 of the Compendium totius gram‐ maticae: Nota si due fuerint clausule quarum altera causalem coniunctionem possi‐ deat, prius construenda est clausula principalis, ut ‘Quoniam uirtuti incumbas, te amo’. A uerbo ‘amo’ incipienda est constructio nisi antecedentis et relatiui racio causalem clausulam principali clausule preferendam postulet, ut ‘cum uiderem quam plurimos eloquencie studere eorum diligencie congratulabar’. For Anwykyll’s grammar, see Luhtala (2013b). 6 A note on the concepts of coordination and subordination Syntactic theory made great advances in the Middle Ages, and the achieve‐ ments of medieval scholars are remarkable in the philosophy of language. The ancient and medieval textbooks on grammar, however, failed to make use of a distinction between a main clause and a subclause, and the con‐ cepts of coordination and subordination were absent from their descriptive framework. This does not mean, however, that they ignored complex sen‐ tences wholesale. In the late Middle Ages and in the works of the early Humanists, traces of new terminology can be observed. John of Genoa’s popular encyclopaedia, the Catholicon (1286), contains a short section on grammar in its first book. When defining the concepts of subject and predicate, he explains that anything that comes before the principal verb (verbum principale), is the subject. He goes on to explain that if there is only one verb in a sentence, it is the principal verb. If there are several verbs joined with a copulative or disjunctive conjunction, as in Petrus legit et disputant ‘Peter reads and debates’, Martinus legit uel disputat ‘Martinus reads and debates’, all of them are principal verbs. In the case of the relative pronoun, however, the verb in the relative [clause] is not the principal verb. In the sentence Petrus qui est studens Bononie disputat ‘Peter who is studying in Bologna debates’, disputat is the principal verb. 29 Here we can see that the principal verb is the predicate of the main clause, for which concept the author lacks a term. This concept was termed clausula principalis in the Compendium totius grammaticae (1483) of John Anwykyll, which is regarded as the earliest Humanist grammar in England. 30 The same Milestones in the study of syntax in antiquity and the Middle Ages 31 <?page no="32"?> author uses the term clausula causalis for the causal (sub)clause. Before him, at least the Regule of the Italian Humanist Perotti had used the term uerbum principale. The term clausula, from which the English term ‘clause’ derives, is of rhetorical origin, suggesting that the influence of rhetoric on medieval syn‐ tactic description did not vanish completely after the early Middle Ages. The use of the rhetorical term clausula as standing for ‘clause’ is, in fact, attested earlier in medieval textbooks on versification and prose composition than in grammars. The earliest instance known to me is attested in Geoffrey of Vinsauf ’s Documentum de modo et arte dictandi et uersificandi (c. 1208-1213), in which the term clausula regularly stands for clause, whereas the term sententia represents a unit resulting from a combination of two or more clauses into one (Luhtala 2022). Generally speaking, however, these aspects of language study remained to be explored and codified by early modern scholars. 7 Conclusions The early medieval teachers showed remarkable initiative in exploring ways of analysing Latin prose texts, both Classical and Biblical. All the arts of the trivium had established a basic unit of meaning (sententia) which was complete, calling it a sentence (oratio), a proposition (proloquium) or a period (periodus) respectively. In late Antiquity, only rhetoric and logic (as discussed by Martianus Capella) offered tools for expanding the basic unit with optional elements, including various adverbial phrases, embeddings and subclauses. When Priscian discussed for example noun phrases, he failed to integrate them into the structure of a basic clause. The functional notions of subject and predicate were among the tools of analysis borrowed from logic by scholars active in early medieval St. Gall. In expanding the basic clause, they resorted to Aristotle’s ten ontological categories and the seven rhetorical circumstances. In analysing compound sentences with several optional ‘circumstances’, the author of the St. Gall Tractate showed a remarkably clear understanding that such sentences consisted of slots which could be filled with units of various kinds. He also established a pedagogical word order for interpreting Classical and Biblical texts, which was intended to be used in classroom teaching. Moreover, the author asserted that a complex sentence could be started with any of the 32 Anneli Luhtala <?page no="33"?> circumstances, as depending on the will of the speaker. Thus, the masters in St. Gall discovered that a pragmatic principle governed the ordering of a sentence. After the Carolingian reform there was a decline in intellectual culture, and the approach to the study of syntax as pursued in St. Gall disappeared. When the study of syntax was revived in the cathedral schools of Northern France in the 12 th century, its connection with rhetoric was lost, whereas logical doctrine continued to influence grammatical analysis. However, inspiration for the interaction between grammar and logic was no longer sought in Martianus Capella’s treatise on dialectic. When the division of a sentence into subject and predicate was reintroduced into grammar in the twelfth century, it was adopted from Boethius’s Latin translations of Aristotle’s old logic, and the terms subiectiuum and declaratiuum were replaced with the terms suppositum and appositum. At the same time, the idea of expanding a basic clause with optional elements was lost. On the contrary, the focus tended to shift to units more minor than a clause, that is, to imperfect constructions or phrases. The descriptive framework adopted from Priscian focused on govern‐ ment and transitivity, whereby syntactic relations were largely described as consisting of linkages between two words. In several medieval works, such as Hugh of St. Victor’s De grammatica and the Doctrinale of Alexander of Villadei (c. 1199), Priscian’s definition of a sentence was reworked to account for phrases rather than clauses, to the effect that the idea of a sentence was occasionally obscured. Bibliography - Primary Sources Alcuin: De dialectica. (PL. 101) col. 949-980. Anwykyll, John: Compendium totius grammaticae. Oxford: 1483. <http: / / di‐ glib.hab.de/ inkunabeln/ p-337-4f-helmst-3s/ start.htm> (Last access 22.10.2020) Aristotle: The Categories. On Interpretation (De interpretatione), Harold P. Cook (ed., transl.) (Loeb Classical Library). London: Heinemann/ Cambridge, Massachusetts: Harvard University Press 1938. Aulus Gellius: Noctes Atticae, Peter K. Marshall (ed.). 2 vols. Oxford: Clarendon Press 1990. Boethius, Anicius Manlius Severinus: De divisione, John Magee (ed.). Leiden: Brill 1998. Milestones in the study of syntax in antiquity and the Middle Ages 33 <?page no="34"?> CIMAGL = Cahiers de l’institut du Moyen-Âge grec et latin. Copenhagen: University of Copenhagen 1969- .www.cimagl.dk since 2010. Demetrius of Phaleron: On Style, W. Rhys Roberts (transl.). 1902. (Repr. Hildesheim: Olms 1969.) Dionysius of Halicarnassus: On Literary Composition, W. Rhys Roberts (ed., transl.). London: Macmillan 1910. Donatus: Ars maior: In: GL IV, 367-402; Holtz, Louis (1981): Donat et la tradition de l’enseignement grammatical. Étude sur l’Ars Donati et sa diffusion (IVe-IXe siècle) et édition critique (Documents, Études et Répertoires publiés par l’Institut de Recherche et d’Histoire des Textes 21). Paris: CNRS, 603-674. GL = Grammatici Latini, Heinrich Keil (ed.). 8 vols. Leipzig: B. G. Teubner, 1855-1880. (Repr., Hildesheim: Georg Olms 1981.) Glosa Victorina super partem Prisciani de Constructione (ms Paris Bibliothèque de l’Arsénal 910), Karin Margareta Fredborg (ed.) with the collaboration of Anne Grondeux and Irène Rosier-Catach (Studia artistarum 27). Turnhout: Brepols 2011. Geoffrey of Vinsauf: Documentum de modo et arte dictandi et versificand, Edmond Faral (ed.). Les Arts poétiques du XIIe et du XIIIe siècle. Recherches et documents sur la technique littéraire du Moyen Age. Paris: Librairie ancienne honoré Champion 1924, 263-320. Isidore of Seville: Etymologiae: Isidori Hispalensis episcopi Etymologiarum sive ori‐ ginum libri XX, W. M. Lindsay (ed.), 2 vols. Oxford: 1911. Hugh of St. Victor: De grammatica. Hugonis de Sancto Victore opera propaedeutica, Roger Baron (ed.) (University of Notre Dame Publications in Medieval Studies 20). Notre Dame, Indiana: University of Notre Dame Press 1966, 67-166. John of Genoa: Catholicon. 1286. Quoted from Iohannes <Ianuensis>. Catholicon mit Gedicht ‘hinc tibi sancte pater’. <https: / / www.digitale-sammlungen.de/ en/ view/ bsb00036989? page=1> (Last access 14.1.2022) Marius Victorinus: Rhetores latini minores ex codicibus maximam partem primum adhibitis, Carolus Halm (ed.). Lipsiae: Teubner 1863, 153-304. Martianus Capella: De Nuptiis Philologiae et Mercurii, A. Dick (ed.). Stuttgart: Teubner 1978. Martianus Capella: Martianus Capella and the Seven Liberal Arts. Vol. II. The Marriage of Philology and Mercury, William Harris Stahl and Richard Johnson with E. L. Burge (transl). New York: Columbia University Press 1977. Petrus Helias: Summa super Priscianum, Leo Reilly (ed.). Two vols. Toronto: Pontifical Institute of Mediaeval Studies 1993. PL = Patrologiae (Latinae) cursus completus. 221 vols. Paris: J.-P. Migne 1844-1864. 34 Anneli Luhtala <?page no="35"?> Priscianus: Institutionum Grammaticarum libri XVIII, Martin Hertz (ed.). In: GL II-III. Quintilianus: Institutio oratoria, H. E. Butler (transl.) (Loeb Classical Library, 124). Cambridge, Mass./ London: Harvard University Press 1920. Sedulius Scottus: In Donati Artem maiorem, Bengt Löfstedt (ed.). Grammatici hibernici carolini aevi. Pars III,1-2. CCCM, 40, B-C. Turnhout: Brepols 1977. The St. Gall Tractate: A Medieval Guide to Rhetorical Syntax, Anna Grotans/ David Porter (eds.). Columbia: Camden House 1995. Varro: De lingua latina, Roland G. Kent (transl.) (Loeb Classical Library). 2 vols. Cambridge, Mass./ London: Harvard University Press 1938. - Secondary Sources Ackrill, John L. (1963): Aristotle’s Categories and De interpretatione, transl. with notes. Oxford: Clarendon Press. Blank, David (1982): Ancient Philosophy and Grammar. The Syntax of Apollonius Dyscolus (American Classical Studies 10). California Chico: Scholars. Contreni, John (1981): “John Scottus, Martin Hibernensis, the Liberal Arts, and teaching”. In: Insular Latin Studies: Papers on Latin Texts and Manuscripts of the Brigish Isles: 550-1066, Michael Herren (ed.). Toronto: Pontifical Institute of Mediaeval Studies, 23-44. de Jonge, Casper C. (2008): Between Grammar and Rhetoric. Dionysius of Halicarnassus on Language, Linguistics and Literature. Leiden/ Boston: Brill. Fredborg, Karin Margareta (1973): “The Dependence of Petrus Helias’ Summa super Priscianum on William of Conches’ Glose super Priscianum”. CIMAGL 11: 1-57. Fredborg, Karin Margareta (1988): “Speculative Grammar”. In: A History of Twelfth-Century Western Philosophy, Peter Dronke (ed.). Cambridge: Cambridge University Press, 177-195. Frede, Michael (1977): “The Origin of Traditional Grammar”. In: Historical and Phil‐ osophical Dimensions of Logic. Methodoly and Philosophy of Science, Robert Butts/ Hintikka Jaakko (eds.). Dordrecht: Reidel, 51-79. Grotans, Anna A. (2006): Reading in Medieval St. Gall (Cambridge Studies in Palaeog‐ raphy and Codicology 13). Cambridge: Cambridge University Press. Holtz, Louis (1981): Donat et la tradition de l’enseignement grammatical. Étude sur l’Ars Donati et sa diffusion (IVe-IXe siècle) et édition critique (Documents, Études et Répertoires publiés par l’Institut de Recherche et d’Histoire des Textes, 21). Paris: CNRS. Milestones in the study of syntax in antiquity and the Middle Ages 35 <?page no="36"?> Householder, Fred W. (1981): Syntax of Apollonius Dyscolus, transl. and with comm. (Amsterdam Studies in the Theory and History of Linguistic Science, 23). Am‐ sterdam/ Philadelphia: John Benjamins. Kneepkens, Cornelius Henri (1978): “Master Guido and his View on Government: On Twelfth-Century Linguistic Thought”. Vivarium 16(2): 108-141. Kneepkens, Corneille Henri (1990): “Transitivity, Intransitivity and Related Concepts in 12th Century Grammar. An explorative study”. In: De ortu grammaticae. Studies in Medieval Grammar and Linguistic Theory in memory of Jan Pinborg (Studies in the History of the Language Sciences 43), Geoffrey L. Bursill-Hall/ Sten Ebbesen/ Konrad Koerner (eds.). Amsterdam/ Philadelphia: John Benjamins, 161-189. Lallot, Jean (1997): Apollonius Dyscolus. De la construction. 2 vols. Paris: J. Vrin. Luhtala, Anneli (1993): “Syntax and dialectic in Carolingian commentaries on Pris‐ cian's Institutiones grammaticae”. In: History of Linguistic Thought in the Early Middle Ages (Studies in the History of the Language Sciences 71), Vivien A. Law (ed.). Amsterdam: John Benjamins, 145-191. Luhtala, Anneli (2000a): On the Origin of Syntactical Description in Stoic Logic (The Henry Sweet Society Studies in the History of Linguistics 8). Münster: Nodus Publikationen. Luhtala, Anneli (2000b): “Early medieval commentary on Priscian’s Institutiones grammaticae”. CIMAGL 71: 115-188. Luhtala, Anneli (2005): Grammar and Philosophy in Late Antiquity (Studies in the History of the Language Sciences, 107). Amsterdam/ Philadelphia: John Benjamins. Luhtala, Anneli (2013a): “Pedagogical Grammars before the Eighteenth Century”. In: The Oxford Handbook of the History of Linguistcs, Keith Allan (ed.). Oxford: Oxford University Press, 341-358. Luhtala, Anneli (2013b): “John Anwykyll: Compendium totius Grammatice”. In: Handbuch frühneuzeitlicher Grammatiken, Aino Kärnä/ Anneli Luhtala/ Jenny Ma‐ linen/ Lauri Marjamäki/ Torsten Schaßan (eds.). Wolfenbüttel: 2013. <http: / / diglib.hab.de/ ebooks/ ed000171/ id/ ebooks_ed000171_004/ start.htm> (Last ac‐ cess 22.10.2020) Luhtala, Anneli (2020): “Syntactical Relations in Ancient and Medieval Grammar”. In: The Origin of Dependency Syntax. A Historical Survey from Panini to Tesnière (Studies in Language Companion Series), Andràs Imrényi/ Nicola Mazziotti (eds.). Amsterdam/ Philadelphia: John Benjamins, 24-57. Luhtala, Anneli (2022): “Interpreting and Rephrasing Participial Constructions in Medieval Language Arts”. In: La sintassi sommersa tra Tardo antico ed Età moderna: 36 Anneli Luhtala <?page no="37"?> Fonti, modelli, e strategie interpretative (Lingue, Linguaggi, Metalinguaggio 15), Annamaria Bartolotta (ed.). Roma: Calamo, 63-82. Rosier, Irène (1984): “Transitivité et ordre des mots chez les grammairiens médié‐ vaux”. In: Matériaux pour une histoire des théories linguistiques, Sylvain Auroux et al. (eds.). Lille: Presses Universitaires, 181-190. Rosier, Irène (1994): “L’introduction des notions de sujet et de prédicat dans la grammaire médiévale”. Archives et documents de la SHESL 10: 81-119. Scaglione, Aldo (1972): The Classical Theory of Composition from its Origins to the Present. A Historical Survey. Chapel Hill: The University of North Carolina Press. Taylor, Daniel (1987): “Rethinking the History of Language Science in Classical Antiq‐ uity”. In: The History of Linguistics in the Classical Period (Studies in the History of the Language Sciences 46), Daniel Taylor (ed.). Amsterdam: John Benjamins, 1-16. Milestones in the study of syntax in antiquity and the Middle Ages 37 <?page no="39"?> Word-formation in neo-Latin school grammar Ľudmila Eliášová Buzássyová Abstract: The purpose of this study is to determine the extent of tendencies in systematising knowledge about word-formation in neo-Latin school grammar (the sixteenth century to the mid-nineteenth century) and to investigate how the description of word-formation phenomena changed. Sourcebooks from then German-speaking milieu are examined and compared to late antiquity grammars. The research focuses on two topics: (1) the location of text units with a word-forma‐ tion content in grammars to determine their role in the process of separation of word-formation as an autonomous discipline and (2) the key terms and definitions of these text units. In the first topic, it is argued that the most critical change was the establishment of the opening “word-formation complex” in the Discipline of Words in the sixteenth century. Changes in the second topic identify changes in terminology as well as changes in the hierarchy and typology of definitions. All changes are investigated and explained upon the back‐ ground of word-and-accident model decay. The findings suggest that in terms of grammar structure and the integration of word-formation units, neo-Latin grammar retained its conservative character. On the other hand, more significant changes were detected in formulating definitions of word-formation concepts. Keywords: word-formation, word-and-accident model, Latin and neo-Latin grammar, Donatus, Priscianus, Melanchthon, Ramus, spe‐ cies, figura, primitiva, derivata/ derivativa, simplex/ simplicia, composita, decomposita <?page no="40"?> 1 The independence of word-formation within the hierarchy of linguistic disciplines is a debatable topic. It can be considered a part of lexicology or subsumed together with inflection under the heading of “morphology”. 2 The term was introduced by Barwick (1922: 15). Baratin (2000: 462) interprets this term as “l'enseignement scolaire de base, la vulgate par rapport à laquelle certaines doctrines se démarquent, mais qui sert, dans le dispositif mis en place, à établir une continuité, une permanence doctrinale”. 3 Ramus’s new approach to teaching was motivated by his anti-Aristotelianism and his break with the scholastic tradition; however, in reality Ramus was essentially dependent on both of these paradigms. For more about Ramus’s anti-Aristotelianism, see Padley (1976: 77-96) and Sellberg (2020). 1 Introduction In the history of word-formation as a “self-contained linguistic discipline” (Müller et al. 2015, I: vii), 1 ancient Greek and Latin grammar was the source of the first theoretical approaches and terminology, even though in the ancient era there was no systematic distinction between inflectional and derivational categories. Marcus Terentius Varro’s theory on the difference between declinatio naturalis and declinatio voluntaria was a unique excep‐ tion; however, this theory did not become part of the school curricula in mainstream grammar teaching - “school grammar” 2 - after antiquity. Indeed, school grammar preserved its unchanging character for centuries to such an extent that, as Joachim Latacz (1979: 210) claimed, even during the “observation period” from 1450 to 1850, its main features lasted unchanged compared to late antiquity as far as the grammatical system of Latin was concerned. Although Latacz’s statement can still be broadly agreed with, previous research has stated that some changes occurred in Latin grammar from Renaissance humanism. Findings have shown that a new incentive in the development of ancient concepts in the field of word-formation arose mainly thanks to Julius Scaliger and Petrus Ramus, as Padley (1976: 58- 110) thoroughly described. In particular, Ramus, a French Protestant edu‐ cator (1515-1572), who connected the dichotomous method with his new approach to interpreting word classes, shaped Latin grammatography in various and very diverse parts of Europe, albeit in various modified forms (Padley 1985: 27-83). 3 Ramus’s additional value in the traditional structure of Latin grammar indirectly stimulated a new approach to constructing the grammar of living languages (which with greater self-confidence started to put themselves forward). 4 His initiative was also behind the decision 40 Ľudmila Eliášová Buzássyová <?page no="41"?> 4 The reach of Ramus’s work in the grammars of vernacular languages is documented by Kristol in Auroux (2000: 766); Verrac (ibidem: 772-775); Poppe (ibidem: 801); Häkkinen (ibidem: 807); Sarmiento (ibidem: 866); Fazekas (ibidem: 917). of the German philologist Wolfgang Ratke to dedicate an extra unit to word-formation, remove it from the grammar itself, and after 1630 publish it in an independent textbook entitled Wortbedeutungslehr (Padley 1985: 111; Marjamäki 2018: 47; cf. Ising (ed.) 1959: 271-318). This input was an essential step to further developing word-formation as a discipline in living languages. The reach of this input is also visible in the new distribution of the roles of grammar and dictionaries in the seventeenth century (Kaltz/ Leclercq in Müller et al. 2015: 29). This can be seen, for instance, in Arnauld and Lancelot’s Grammaire générale et raisonnée (1660), where derived and compound words are not dealt with, because, according to the authors, that issue was meant for Dictionnaire général. However, Latin school grammar seemed to change more slowly despite being a springboard and breeding ground of inspiration for vernacular grammars. It was not until the very end of Latacz’s “observation period” that fundamentally new theories were introduced (Burkard 2003). The aim of this paper is to determine how selected authors of neo-Latin school grammar books between the sixteenth century and the middle of the nineteenth century predominantly from German-speaking countries com‐ prehended and explained Latin word-formation phenomena, and whether their descriptions of these phenomena became an autonomous part in the grammars. The main reason to investigate this topic is the fact that researchers have not paid much attention to this issue after Padley’s findings. Monographs and survey studies from the history of Latin word-formation from the neo-Latin period are rare. Even though there are studies by Alfieri (2019) and Lindner/ Rainer (2015), as well as a preliminary review of the development of the discipline by Re (2020), such works are few and far between. Even Burkard (2003), who dealt with the grammar of the eighteenth century and early nineteenth century, did not reflect on word-formation processes in his analysis. On the other hand, a retrospective view on the origins of word-formation research in living languages has provided an excellent opportunity to better understand their Latin foundations (cf., for instance, Marjamäki 2018 and Huterer 2001). Word-formation in neo-Latin school grammar 41 <?page no="42"?> 5 Kaltz/ Leclercq (in Müller et al. 2015: 23) mention a “grammaticographical topic in its own right” in the German tradition. The main reasons for such little interest emerged from two main circum‐ stances. Firstly, there is the communis opinio that suggests that Latin school grammar was very conservative and almost immutable. The other crucial reason was that after vernacular languages started to take over as an object of modern linguistic description and research, Latin school grammar - as a seemingly “exhausted” source - lost a part of its attractiveness as a research topic. This approach was distinctly present in the overview of the history of word-formation research in the Introduction to Word-Formation: An International Handbook of the Languages of Europe (Müller et al. 2015). Here, the primary focus of the authors shifted to the rise of vernacular grammar, which eventually meant the end of attention to Latin grammar (cf. the chapter by Kaltz/ Leclercq in Müller et al. 2015: 22-38) until it started to re-emerge in connection with the historical-comparative method from the middle of the nineteenth century (Lindner in Müller et al. 2015: 38-44). The starting point of this paper suggested two distinct topics of a broad range of subject matters: (1) the formal incorporation of text units with word-formation content in a grammatical text and (2) the use of terms and the formulation of definitions of selected key concepts of word-formation units. The research in the first area was conducted in both grammars written in Latin and German. In this part, the study to some extent had a comparative dimension to grammars of vernacular languages - primarily to grammars of German, Czech, and Slovakized Czech. These grammars from the seventeenth century showed a tendency to describe word-formation phenomena in an individual part. 5 There was also the tendency to show word-formation as an autonomous topic that was separate from grammar. The research aimed to determine the existence of similar tendencies in Latin grammar. The second topic aimed to investigate if and how the definitions of word-formation concepts changed. In contrast to Anneli Luhtala (2002), who analysed grammatical definitions in ancient grammars in terms of semantics and the methods used, this study is based on the classification of definitions according to the purpose of the definition (Zouhar 2014). 42 Ľudmila Eliášová Buzássyová <?page no="43"?> 2 Primary sources The source material of the present research and its comparative basis was provided by the grammarians of late antiquity, Donatus’s Ars maior and Priscianus’s Institutiones grammaticae, which represented the two main branches of the tradition of late antiquity Roman grammar (Ax 1986: 247- 250). The focus area of the present research is neo-Latin grammatography. In terms of its cultural and geographical scope, the research was primarily limited to the adaptation of ancient grammar in then German-speaking milieu that intensively influenced the nature of teaching classical and subsequently living languages in the broader European area and the Slavic part of the continent in particular (Koupil 2015: 38-42; Buzássyová 2019: 417, 431-432). The language of explanation in the analysed texts was initially Latin, and from the end of the seventeenth century it was German. In this study, the choice of neo-Latin grammars chiefly depended on their relevance and distribution as well as their impact on the educational process. While assessing these criteria, the findings of previous research on Latin grammatography (Padley 1976 and 1985; Tavoni 1998; Ising 1970; Latacz 1976; Burkard 2003; Marjamäki 2018; Koupil 2015) were crucial. The analysed texts therefore covered grammars with different approaches. These included works by Philipp Melanchthon (1497-1560) and his various complimented and expanded versions; works by Petrus Ramus; mixed Melanchthonian-Ramean grammars and mixed grammars based primarily on Melanchthon’s texts as well as those featuring his name in the title; and grammars by Johann Rhenius (1574-1639). Latin grammars written in German by Christophorus Cellarius (1638-1707), Immanuel Johann Ger‐ hard Scheller (1735-1803), Georg Friedrich Grotefend (1775-1853), Christian Gottlob Bröder (1745-1819), and Johann Otto Leopold Schultz (1782-1849) were deeply analysed; the so-called donats - the textbooks of the first Latin language basics - were only lightly examined. For comparison purposes, some grammars of vernacular languages - such as German, Czech, and Slovakized Czech - were taken into consideration. Word-formation in neo-Latin school grammar 43 <?page no="44"?> 6 Regarding the concept of the “word-and-accident model” cf. Copeland/ Sluiter (2009: 83); regarding the concept of “accident”, lat. accidens, cf. for instance, Pinborg (1967: 31, 34); and regarding the Greek models of Latin accidentia cf. Vaahtera (1998: 53; 2014). 7 Corrected by the author of this paper, instead of “latter” used by Kaltz/ Leclercq in Müller et al. (2015, II: 23). 8 In adverbs, participles, and conjunctions, only the figura is distinguished; preposi‐ tions and interjections, on the other hand, have neither the accident species nor figura. For more details on joining the accidences species and figura in those gram‐ mars, see Buzássyová (2019: 420, 422, 423). 9 This is due to the application of the principle of universaliter primum formulated by Petrus Ramus (Padley 1976: 85, 86, 1985: 25). 3 Establishing the “word-formation complex” on the background of the word-and-accident model - 3.1 Neo-Latin grammars written in Latin Previous research identified the impact of sixteenth-century Latin gram‐ mars by Scaliger and Ramus as a stimulus for a new approach to explain word-formation phenomena in Latin and vernacular grammatography (Padley 1976 and 1985; Marjamäki 2018; Buzássyová 2019). These changes mainly affected the word-and-accident model 6 as a general model of ex‐ plaining lexical-semantic, derivational, and inflectional properties of word classes - i.e. partes orationis - in Latin school grammar. Within this model, the accident species “allows for a distinction between ‘primary’ [primitiva] and ‘derived’ [derivativa] lexical items, while the accident figura was the base for distinguishing ‘simple’ [simplex], ‘composed’ [composita], and ‘de‐ composed’ [decomposita] words, the [last] 7 term designating nominal deriv‐ atives from compounds” (Katz/ Leclercq in Müller et al. 2015: 23). What was the essence of the changes initiated by Scaliger and Ramus? In grammars, the fixed place of two accidents - species, figura - was either (1) at the beginning of a series of accidents in the case the word classes had both these accidents 8 or (2) at the very beginning of the Discipline of Words - i.e. Etymologia. In this part, these categories were generally described in one joint complex (from now on, the “w-f complex”) for all word classes and did not receive further attention. 9 The analysed primary sources in this study referred to two subtypes of type (2) grammars: (2a) Ramus’s grammars (for example, Grammaticae libri quatuor, 1560 with its numerous further editions) that went without the traditional word-and-accident model, although they continued using 44 Ľudmila Eliášová Buzássyová <?page no="45"?> 10 Ising (1970) provides an exhaustive analysis of the textbooks with the name Donatus and their influence on the grammars of vernacular languages. Some findings in this area of research need to be corrected over time. For more on Ising’s inaccuracies in relation to the donats used in the Kingdom of Hungary, see Eliášová Buzássyová/ Ma‐ chajdíková (2020: 133). 11 For details, see Buzássyová (2019: 428-433). 12 In neo-Latin grammars, similarly to their late antique ancestors, both derivata and derivativa were terms used for naming derivates. For basic information about the use of these terms in late antiquity Roman grammar, see Schad (2007), s.v. derivativa, the terms species and figura, and (2b) Melanchthonian-Ramean grammars (for example, Grammatica Latina Philippo-Ramea 1596) and other mixed grammars based on a modified word-and-accident model within which the “w-f complex” was composed of the accidents species and figura at the very beginning of the Discipline of Words, often called accidentia generalia (for example, Rhenius’s Grammatica Latina cum paralipomenis 1618, 1625: 10; Philippi Melanchthonis Grammatica Latina 1661: 10). This tendency to separate the introductory “w-f complex” confirmed that the word-formation properties of word classes represented an independent area in the language description. Not all grammar books tended to be equally open to the idea of keeping the “w-f complex” separate. This tendency was promoted in different ways in the textbooks of Latin basics known as donats.  10 For example, the well-known Rhenian Donatus Latino-Germanicus entirely removed the issue of deriva‐ tion and composition from the text. Therefore, for instance, in the word class nomen, only four (grammatical in a true sense) accidents - genus, numerus, casus, declinatio - were included (DLG 1663: 2, similarly, for example, DLG 1778: 2). Meanwhile, the bilingual and multilingual Rhenian donats used in the Kingdom of Hungary kept species and figura among the accidents and placed them at the beginning of their sequence (DLH 1697: 2, DLH 1750: 2; DLGHB 1748: 106). The accidents species and figura were also united into one complex in the grammars of vernacular languages (German, Czech, Slovakized Czech, and other Slavic languages) from the seventeenth century onwards. 11 3.1.1 Appendices and other complements in neo-Latin grammars In Latin grammars of type (2b), the separation of the “w-f complex” focusing on a brief description of the classes primitiva - derivata/ derivativa  12 and simplex - composita - decomposita was inconsistent. The main reason for Word-formation in neo-Latin school grammar 45 <?page no="46"?> derivata. A detailed analysis of the use of these terms would be beyond the scope of this study. 13 Authors of vernacular grammars also took the methodological approach of focusing on the description of word-formation phenomena in multiple places from Latin models. Kaltz and Leclercq comment on Schottelius’s case (Ausführliche Arbeit von der Teutschen HaubtSprache 1663) in Müller et al. (2015: 24). this inconsistency was that even though the grammar writers presented the issues of derivation and composition at the beginning of the Discipline of Words according to the Ramean spirit, they kept going back to word-forma‐ tion issues in other parts of their texts. The fundamental cause seemed to be the pressure of the word-and-accident model, which they did not entirely abandon, and, also, they did not fully understand the difference between the lexical and inflectional changes. In Latin grammar books, such repetition was evident, especially in pas‐ sages addressing derivation. In addition to the “w-f complex”, derived lexemes could also be discussed in a particular word class, either under the heading de specie, de speciebus (Rhenius 1625: 102 sq.) or de derivativis (Melanchthon et al. 1661: 100 sq.). In these passages, such types of derivates were described in the word class of nomen as patronymica, possessiva, comparativa, gentilia, deminutiva, denominativa and in word-class verbum derivates such as inchoativa, frequentativa, meditativa (desiderativa); thus, word forms created by derivation were distributed in the text in several places. 13 On the other hand, some authors tried to explain this subject matter in better arranged and more systematic grammar books as far as the struc‐ ture and the proportions of texts were concerned. In these grammars, parts focusing on the description and characteristics of derived lexemes were often referred to as appendices. The term appendix seemed to be appropriate, as the authors and compilers of grammar textbooks perceived its content as superstructural and not as grammatical in a true sense. Such appendices concentrated the traditional material in a reduced space. The mixed Gram‐ matica Philippo-Ramea (1660) could serve as one of the examples in which two well-organized appendices were located: Appendix nominis de derivativis (57-58) and Appendix verbi de derivativis (1660: 101-102). In the first one, the nominal derivatives were methodically classified based on the word classes from which the respective lexeme groups were derived; therefore, special attention was paid to (a) derivatives as lexemes that became essential in terms of form and to (b) the word-formation process. 46 Ľudmila Eliášová Buzássyová <?page no="47"?> 14 Apart from more well-known texts, such as Ratke’s Wortbedeutungslehr (1630) and Claude Irson’s Nouvelle méthode (1662 [1656]) with its lexical section at the end of the grammar, this can be seen in Czech grammars. In the Čechořečnost seu Gram‐ matica linguae Bohemicae (1672) by Schottelius’s successor Rosa, word-formation is an independent part attached to syntax in the chapters De specie, significatione et formatione nominum and De nominibus compositis et contractis (1672: 357-392). Later, following Rosa, Doležal incorporated into his Grammatica Slavico-Bohemica a separate part devoted to the word-formation of all word classes under the name Appendix sistens modum multiplicandi vocabula, per derivationem compositionemque (1746: 226-272). For more details, see Buzássyová (2019: 433-434); for a general (1) Appendix nominis de derivativis, Grammatica Philippo-Ramea (1660: 57-58) I. a nominibus derivata - 1. patronymica, 2. gentilia […]; II. a verbis derivata - verbalia (lectio, venatio […]); III. ab adverbiis adverbialia (hodiernus); IV. item numeralia - cardinalia, ordinalia […] Besides the appendices, there was another form of attachment at the end of some grammar books, called paralipomena (for instance, Frischlin 1609: 259 sqq.; Rhenius 1625: 659 sqq.), which contained topics and examples that were “left aside” from the primary text. In contrast to the appendices men‐ tioned above, paralipomena were more heterogeneous as they represented expanded and detailed knowledge from individual chapters. The content of the original accidents species and figura was just one of many topics discussed in paralipomena; therefore, they seemed irrelevant from the point of view of building thematic units of word-formation. While paralipomena focused mainly on practising grammar and gaining more knowledge within all planes of the language, appendices concentrated on facts of derived word forms in two word classes - nomen and verbum; however, it should be emphasized that in the seventeenth century, these two appendices (on nomen and verbum) did not become the core of an inde‐ pendent part on word-formation in grammar itself, and neither was their content included in the “w-f complex” at the beginning of the Discipline of Words. The link between the introductory “w-f complex” and the appendices on derivatives was not accentuated in the analysed texts expressis verbis through exact references. Unlike in the grammars of living languages, in which the separation of an autonomous topic on composition and derivation could be observed as early as in the seventeenth century, 14 such a process did not seem to exist in Latin grammars. As much as Latin grammars remained Word-formation in neo-Latin school grammar 47 <?page no="48"?> overview on word-formation in grammars of Czech and Slovakized Czech before Dobrovský, see Hauser (1959). 15 The title could be translated as “Latin grammar made easy” as suggested by one reviewer. conservative from the structure and content points of view, yet, at least some efforts to didactic approachability could be detected that helped com‐ prehend the word-formation structure of selected word forms; furthermore, these efforts encouraged more systematic thinking about them. This grammar scenario was also the case during the transition to ver‐ nacular languages as languages of explication. This was a period when grammarians had to come to terms with the legacy of ancient grammar; on the other hand, they wanted to respond to the rational demands of the time and the increasing share of the mother tongue in the school curriculum at the expense of classical languages (Stroh 2007: 244 sq.). - 3.2 Neo-Latin grammars written in German 3.2.1 Christophorus Cellarius (1638-1707) Christophorus Cellarius’s concise grammar of Latin basics, Erleichterte latei‐ nische Grammatica, 15 was written in German and was published for the first time in 1689. In the eighteenth century, his grammar went through several adaptations in different language environments. The brief “w-f complex” focusing on derivation and composition, developed in grammars during the previous periods, also remained in Cellarius’s grammar, which served as an introduction to the Discipline of Words, Etymologia. In his grammar, Cellarius abandoned accidents in all word classes except for nomina. Even in nomina, however, he only included those categories to accidents that he considered genuinely grammatical, i.e. qualitas, genus, motio, comparatio, numerus, casus, declinatio. The species and the figura, on the other hand, were no longer indicated as accidents. This approach only confirmed that in Cellarius’s description of study of words there was a clear dividing line between the accidents of species/ figura and the other accidents, although he never clarified or explained the differences between them. In his brief “w-f complex”, he formulated and kept only two laconic definitions, namely that “words are according to the figura (‘nach der Figur’) simplicia, composita and decomposita” and “according to the species (‘nach der Species’) primitiva and derivativa”. 48 Ľudmila Eliášová Buzássyová <?page no="49"?> 16 In the 1689 edition, the appendix has the shorter name Anhang vom Nomine (p.-24). 17 Cellarius’s original text was substantially revized by J. M. Gesner (1691-1761), as stated in the subtitle: “durchsehen, vermehrt und verbessert von Jo. Matthias Gesner”. The text revized by Gesner reappeared in further modified editions. 18 The names of the editors of Cellarius’s text are not written in the grammar. 19 Understanding the structure of the word correctly was not irrelevant to Cellarius; however, he marked out the territory outside of grammar for this role - namely the dictionary. A part of the full title of his dictionary, Latinitatis probatae et exercitae liber memorialis, is the expression “natural sequence”, naturalis ordo, that reflects the way lexemes are arranged in the dictionary - in alphabetical order but first with the primary word primitivum followed by its derivatives and finally its compounds (Škoviera 1985: 70; Brnadič 2015: 121). This method replaced the word-formation theory upon which the dictionary should be based. Cellarius moved the nominal derivatives of diminutiva, gentilia, patro‐ nymica, possessiva, denominativa as a traditional content of species - to‐ gether with brief details on the declension of composites, such as respublica along with heteroclita, defectiva, and numeralia - to a two-page appendix entitled Anhang von dem Nomine und dessen Anomalien (Erleichterte latei‐ nische Grammatica 1704: 24, 1738: 18-19; 16 Cellarius/ Gesner (ed.) 1740 17 : 21-22, 1786: 22-23). Anhang was a mixture of heterogeneous material from a traditional de specie passage, other appendices, and paralipomena. The title of the Anhang seemed to declass the traditional learning of derived forms to the position of some anomaly. Moreover, in the first editions of Erleichterte lateinische Grammatica, lex‐ emes such as diminutiva, gentilia, and others were not mentioned in Anhang as derivatives. Only in later editions (1720: 21, 1738: 18) 18 are such types of lexemes referred to as derivatives; with precise references to relevant pages, these lexemes were linked with the initial “w-f complex”. Cellarius did not even initially link the introductory “w-f complex” to the verbs frequentativa, meditativa, and diminutiva, inchoativa (known from older grammars as verbal species) nor did he state that they were derived. He named those verbs Abtheilungen, i.e. some “sections” of verbs, and described them as von anders Verbis gemacht, i.e. “made from other verbs” (1720: 28, 1738: 24). In this group of verbs, the German term for derivatives, hergeleitete, could not be found until Gesner’s edition of Cellarius’s text from 1786 (p. 30). A detailed investigation of editions of Cellarius’s brief grammar showed that their terminology was only gradually clarified more precisely, and through these terms the parts of the text regarding word-formation were interconnected. 19 Word-formation in neo-Latin school grammar 49 <?page no="50"?> 3.2.2 Grammars from 1780 to 1840 In the progressively written school grammar books of this period, which, according to Burkard, significantly detached themselves from tradition (2003: 781), there were only slight traces of the accidents. The beginning of this period saw the emergence of Scheller’s grammars: the comprehensive and thorough Ausführliche lateinische Sprachlehre (1779) and its shorter version for teaching the basics, Kurzgefasste lateinische Sprachlehre (1st ed. 1780). The characteristic feature of this new approach is, inter alia, emphasizing the semantics of word classes. In the first part of his grammar, known as the learning of ‘individual words’ - Von einzelnen Wörtern, (Ausführliche lateinische Sprachlehre 1779: 6 sqq.; Kurzgefasste lateinische Sprachlehre 1780: 3 sqq.) - Scheller presented lexemes with a stress on their formal properties (suffixes). As a novelty in Scheller’s grammars, there was a separate “lexicological” chapter entitled Von den Bedeutungen der Wörter (‘About the Meanings of Words’) (1779: 276-288, 1780: 123-126), located at the end of the description of all word classes. Scheller explained how the meaning of a derived word can be revealed from the primary word (1779: 282, 1780: 124) and how the “endings” of the words - such as -quam, -cunque, -o, -osus, -ibilis, -ficus, -eus, -alis - and others determine the meaning of a derivate (1779: 283, 1780: 125). In his grammar, Scheller actually discussed word-formation in four different places. First (1), he briefly demonstrated the “w-f complex” at the beginning of the learning of words (1779: 45, 1780: 12) with typical content (Stammwörter (primitiva) - abgeleitete (derivata), einfache (simplicia) - zusammengesetzte (composita)); then for two word classes, nomen and verbum, he described the traditional learning of (2) patronymica, derivativa, and others - in the case of nomen - (1779: 46 sq., 1780: 13) and (3) frequentativa and others - in the case of verbum - (1779: 133, 1780: 53). Finally (4), in the abovementioned chapter Von den Bedeutungen der Wörter, Scheller determined wahre Bedeutng ‘the correct meaning’ of the word from the form of suffixes. As these text units were already interconnected by terminology, the description was more coherent than in Cellarius’s grammar. Scheller offered a more comprehensive picture of the word-formation process while explaining the content known from traditional accidents. For example, in the chapter about verbs, he described the differences between Stammverba and abgeleitete verba and then defined the four traditional derivation types - inchoativa, frequentativa, desiderativa/ meditativa, and deminutiva - which in older grammars were included in the verbal species. He determined them in terms of semantics (what they meant), morphology 50 Ľudmila Eliášová Buzássyová <?page no="51"?> 20 “Derived verbs are, based on (their) meaning, fourfold - inchoativa, i.e. those that mark the beginning of what is named in the basic word and can be translated using ‘to become’. These words end with -sco and are derived partly from verbs, partly from nouns.” (what suffixes were used to form them), and word-formation (from what word classes and by what word-formation process they were formed): (2) Scheller: Ausführliche lateinische Sprachlehre (1779: 131-132) Die abgeleitete Verba sind ihrer Bedeutung nach vierlerley: Inchoativa, d.-i. welche einen Anfang dessen, was im Stammworte angezeigt ward, bedeuten, und durch ‘werden’ sich übersetzen lassen. Sie endigen sich auf sco […]. Sie werden theils von Verbis, theils von nominibus abgeleitet.  20 Despite their popularity, even far beyond Germany, the grammars by Christian Gottlob Bröder - Practische Grammatik der lateinischen Sprache (1st ed. 1787) and Kleine lateinische Grammatik (1st ed. 1795) - did not produce anything new concerning how word-formation was explained. In his more extensive Practische Grammatik - from the second and revized edition of 1793 - additional paragraphs with a very traditional content on patronymica, deminutiva, epicoena, heterogenea, heteroclita nouns with motio (derivational gender marking), and defectiva could be found after the part about declensions (1793: 19-24, 1821: 19-25). The structure of these paragraphs on nominal derivatives and other forms, their incorporation into the text, and their heterogeneous nature echoed Cellarius’s nominal Anhang; however, they did not have an umbrella title such as Anhang or anything else. On the other hand, the paragraph on verbal derivatives was more adequately compiled by Bröder in the context of didactics; also, it was significantly graphically marked in the text (with the grapheme) as C) Abgeleitete Verba (Practische Grammatik 1787: 58, 1793: 70, 1821: 70, 1832: 74-76). The content and structuring of this paragraph corresponded with the analogous passage in Scheller’s grammar. In his reissued grammars Größere lateinische Grammatik für Schulen, nach Wenck’s Anlage umgearbeitet (1823/ 1824) and Kleine lateinische Grammatik für Schulen (1825), Georg Friedrich Grotefend expanded the introductory complex and presented it in a more general spirit; he abandoned the traditional struc‐ ture of the “w-f complex” that had been limited to the opposites simplicia - Word-formation in neo-Latin school grammar 51 <?page no="52"?> composita (and, in some cases also to decomposita) and primitiva - derivativa. In his introductory complex, Grotefend added examples, such as an illustrative derivation sequence of von rege, richte, kommt rex, regnum, regnare, regnator, regnatrix ‘from rege, richte, comes rex, regnum, regnare, regnator, regnatrix’ (Größere lateinische Grammatik 1823/ 1824: 15; Kleine lateinische Grammatik 1825: 10), which did not occur in traditional school grammars. Over one page, he demonstrated composition and derivation in a general way for all relevant word classes limiting the selection of examples to only the most distinctive ones, namely nominal denominativa, verbalia, diminutiva, and verbal intensiva. Grotefend’s grammar did not cover the traditional passage of nominal forms - patronymica, gentilia, and others - but compared the word-formation proper‐ ties in German and Latin. Grotefend also took a fundamentally new stand on the significance and status of word-formation in his extensive chapter on verbs. In his grammar, word-formation properties namely are one of the criteria for classifying verbs. He divided verbs based on: I.) the meaning into transitive and intransitive; based on II.) the forms for actives and passives; according to III.) inflections to the regular, irregular, defective, and redundant; and according to IV.) their ‘formation’ - Bildung, for einfach (simplicia), zusammengesetzt (com‐ posita), Stammwörter (primitiva), and abgeleitete (derivata) (Größere lateinische Grammatik 1823/ 1824: 108; Kleine lateinische Grammatik 1825: 76). In addition to these more innovative grammars, traditional grammars continued to be written in the nineteenth century. They differed from the Melanchthonian patterns only in their somewhat more precise wording and a more well-arranged curriculum structure. Both textbooks by Otto Schultz - Schulgrammatik der lateinischen Sprache (1815) and Ausführliche lateinische Grammatik für die Oberen Klassen Gelehrter Schulen (1825) - followed the traditional Latin pattern, in which the Latin terms of figura/ species reappeared (Schulgrammatik 1836: 26); however, they did not address the concept of acci‐ dents. His Ausführliche lateinische Grammatik did not cover the “w-f complex” in the introduction to the Discipline of Words. Schultz dealt with this topic quite unsystematically and kept it at the very end of the description of substantives - Von der Species der Wörter überhaupt, und der Substantiva insbesondere (1825: 134). Schultz’s grammar was an evident example which proved Latacz’s statements about the conservative character of Latin school grammar. All things considered, Latin grammars written in German presented some diversity. This can be seen in the Cellarius-Schultz trajectory regarding the incorporation of text units addressing word-formation. A brief “w-f 52 Ľudmila Eliášová Buzássyová <?page no="53"?> 21 Nomen was the main word class from which other word classes were distinguished. Its specific character was to signify substantia and qualitas - substantia corre‐ sponding to the Aristotelian ousia ‘being’; cf. Matthews (2019: 77). For the status of the grammarian’s eight parts of speech as opposed to the dialectician’s two (the noun and the verb), see Luhtala (2005: 129 sq.). 22 Ratke used the term Arth in his grammar as a German equivalent of species and Gestalt as figura (cf. in Ising 1959: Theil II, I. Sprachkunst: 14, 16, 17). Other authors used German complex” devoted to composition and derivation - found at the start of the Discipline of Words in Latin-written grammars and influenced by Ramus - was also a standard element in every analysed grammar written in German apart from Schultz’s one. The grammars differed in whether their introductory complex was limited to an elementary explanation of the opposites simplicia - composita (decomopsita) and primitiva - derivativa, or whether they had expanded and obtained a more general character. Also, other independent textual parts dealing with derivation could still be part of the grammar. Usually, these were the passages in particular word classes, with the traditional content of the former accident species; however, unlike the tradition, these passages manifested a more systematic approach to didactics. Alternatively, some passages with less traditional content were extended to the heterogeneous Anhang. In Scheller’s grammars, the chapter Von den Bedeutungen der Wörter seemed to be a milestone in the explanation of word-formation as subsumed under lexicology. In Grotefend’s grammars, the w-f process was even used as one of the criteria for classifying a particular word class (verb). While describing w-f phenomena, authors of grammars written in German in the eighteenth and early nineteenth centuries dealt with the two main word classes - nouns and verbs - in a more balanced way than their ancestors writing in Latin, who continued focusing on the word class nomen as the legacy of ancient grammar. 21 4 Key terms and definitions of the “w-f complex” - 4.1 Species/ figura terms The terms species and figura were part of the explication in grammars as long as grammar followed the word-and-accident model of traditional Latin grammar. From approximately the sixteenth century, depending on region, the authors of grammars written in vernacular languages started to gradually create terms in these vernacular languages as calques. 22 On the other hand, the terminology Word-formation in neo-Latin school grammar 53 <?page no="54"?> terms together with Latin ones - for instance, Gueintz in his Deutscher Sprachlehre Entwurf (1641: 24). For terminology in German grammars, see for instance Kaltz/ Leclerq in Müller et al. (2015: 24 sq.). 23 See more in section 4.2.1 about definitions. 24 “Those words are also either simplicia, simple, […] or composita, compound […] decomposita, double compound according to the figura.” in Latin grammars that were written in German was much more conservative; this meant that at the end of the eighteenth century, the Latin terms species and figura were still in common use. These terms could even be found in grammars up till the middle of the nineteenth century - albeit quite rarely. German equivalents barely replaced them, and instead, the authors gradually omitted such terms from their explication. This process occurred simultaneously by leaving behind the concept of accidents. While capturing the process of disappearing of the word-and-accident model, the individual editions of Cellarius’s Erleichterte lateinische Gram‐ matica also showed persistence in using traditional Latin terms. The first two editions (Merseburg 1689 and 1697) completely abandoned the terms species and figura, whereas in the editions from 1709 they were reintroduced in the forms mentioned earlier - nach der Figur and nach der Specie. In the last editions, with Gesner’s revision published in Leipzig (1786: 3), these terms were only mentioned in parentheses as supplementary information to traditional terminology, which was already archaic yet which helped main‐ tain the continuity with the heritage of ancient grammar. Consequently, the status of the terms species and figura from a terminological point of view was no longer relevant in the text. They differed fundamentally from the position of the terms simplex/ simplicia, composita, decomposita, and primitiva, derivativa, whose place had remained stable even in grammars with a more forward-looking concept. These terms were therefore also listed with a vernacular equivalent. Here, however, a fundamental fact should be highlighted: the adjectives simplex/ simplicia, composita… refer to the concept of the word and not to accidents. 23 (3) Cellarius: Erleichterte lateinische Grammatica (1709, 1712, 1731, 1738: 2) Solche Wörter sind auch nach der Figur entweder Simplicia, schlechte und einfache, […] oder Composita, zusammen gesetzt, […] Decomposita, doppelt zusammen gesetzte.  24 54 Ľudmila Eliášová Buzássyová <?page no="55"?> 25 “Given their external shape, words are either simple (simplicia) […] per ‘through’, legere ‘to read’, or compound (composita), perlegere ‘read over’[…] The former are either primary (primitiva), such as sta ‘stand’, verus ‘true’[…] or derived (derivativa), such as stabilis ‘solid’, veritas ‘truth’[…] Words composed twice, or more are called decomposita. Thus, for example, the word decompositum itself is composed of de ‘from’ and com ‘together’ and positum ‘laid down’.” Scheller’s and Bröder’s grammars (end of the eighteenth century) made do without the traditional concept of accidents and relevant terms. As far as the approach to previous terms of species/ figura is concerned, the beginning of the nineteenth century brought both innovation and stagnation. The innovation in terminology was represented by Grotefend, who in his “w-f complex” at the beginning of the Discipline of Words made a reference to the ‘external shape of the word form’- äußere Bildung (1823/ 1824: 15), re‐ placing the former term Figur while devotedly retaining the formulation of definitions. His In hinsicht auf Ihre äußere Bildung coincided with the earlier version of nach der Figur. In other places of the text, he used the umbrella term Bildung for derivation and composition instead of previous categories, specifically in the classification of verbs (Größere lateinische Grammatik 1823/ 1824: 108; Kleine lateinische Grammatik 1825: 76). Grotefend had no vernacular equivalent for the term species, and in his grammar he elegantly avoided this old category (cf. the collocation erstere wieder): (4) Grotefend: Größere lateinische Grammatik (1823/ 1824: 15) In hinsicht auf Ihre äußere Bildung sind die Wörter entweder einfache (sim‐ plicia) […] per, durch, legere, lesen […] oder zusammengestezte (composita) perlegere, durchlesen, […] und erstere wieder entweder Stammwörter (prim‐ itiva), wie sta, steh, verus, wahr […] - oder abgeleitete (derivativa), wie sta‐ bilis, stehbar, veritas, Wahrheit […] Doppelt und mehrfach zusammengesetzte Wörter heissen Decomposita; so ist z. B. das Wort decompositum selbst doppelt zusammengesetzt aus, de, ab, con […] zusammen, und positum gesetzt.  25 Schultz was just the opposite - he continued working with the traditional categories of species and figura and added German equivalents to these terms. Like Ratke, he associated figura with the German equivalent äußere Gestalt ‘external form’ and species with Abstammung ‘origin’ which corre‐ Word-formation in neo-Latin school grammar 55 <?page no="56"?> 26 For the term “Stammwörter” and other word-formation terms used in grammars of German, see Kaltz/ Leclerq in Müller et al. (2015: 24-26). 27 X is definiendum (defined entity), Y is definiens (defining entity), “= df ” represents the relation of equality of definition. Cf. Zouhar (2014: 339). 28 Zouhar’s classification of definitions is based on a typology developed by Richard Robinson in his book Definition. Oxford: Oxford University Press 2003 (first edition 1954) with further refinements added. 29 Luhtala refers to this definition as the standard definition of nomen (cf. 2002: 258, 271 and 2005: 38). 30 In Donatus’s Ars maior, the category species does not appear among the accidents. The accident species has become a stable member in the word-and-accident model of school grammar thanks to the authority of Priscianus. Cf. Law (2003: 89). sponded with the noun Stammwörter (primitiva) (Ausführliche Grammatik 1825: 134, 152; Schullgrammatik 1836: 26). 26 - 4.2 Definitions of species/ figura accidents in the word-class definition system 4.2.1 Donatus and Priscianus Within the system of definitions in the Discipline of Words in late antiquity, there were three levels in the description of individual word classes. These levels corresponded with the overall “descending hierarchical structure” in Donatus’s grammar and, depending on that, in Priscianus’s grammar as well (Law 2003: 67, 88). The three definition levels were (1) the word class, (2) the accident of the relevant word class, and (3) the classes of relevant accidents - along with (3a) examples of the relevant class. At the first level (1), word classes were determined by conceptual defini‐ tions in the formula X = df Y, 27 where definiendum and definiens were from the same word class and were functionally interchangeable. As far as the typology of definitions was concerned, it was a stipulative definition (for this term, see Zouhar 2014: 341); 28 for instance, Nomen est pars orationis cum casu corpus aut rem proprie communiterve significans. (Ars Mai. 2.2, GL IV: 373.2-3 = Holtz 614). 29 At the second level (2), all accidents were assigned to the relevant word class: Nomini accidunt sex: qualitas, conparatio, genus, numerus, figura, casus. (GL IV: 373.3-4 = Holtz 614). 30 In terms of the typology of definitions, an enumerative definition could be declared (for this term, see Zouhar 2014: 354). At the third level (3), a specific accident with its related classes (in the case of figura and species with two) was associated with the word class in the form of an enumerative definition. These classes were not 56 Ľudmila Eliášová Buzássyová <?page no="57"?> 31 (3) “Two figura(s) approach a noun, simple and composed.” 32 (3a) “Simple, like doctus, potens, composed, like indoctus, impotens.” 33 (3) “There are two species, both in proper nouns and appellatives that are primary and derived.” 34 (3a) “Primary, like Iulus, mons, derived, like Iulius, montanus.” further defined in a sentence, only characterized (3a) by an example with the adverb ut positioned in front of the given example. (5) Donatus: Ars maior 2.8, GL IV: 377.3 = Holtz 624 (3) Figurae nominibus accidunt duae, simplex et composita. 31 (3a) simplex, ut doctus, potens, composita ut indoctus, inpotens […]  32 (6) Priscianus: Institutiones grammaticae 2.22, GL II: 57.9-11 (3) Species sunt tam propriorum quam apellativorum duae, principalis et derivativa. 33 (3a) Principalis, ut ‘Iulus’, ‘mons’, derivativa, ut ‘Iulius’, ‘montanus’ […]  34 4.2.2 Melanchthon and Ramus The practice for determining the accidents figura/ species with the help of a simple enumeration of their classes (figurae/ species accidunt/ sunt…) did not change over the centuries while it remained in the grammars of the word-and-accident model. Melanchthon held to this model consistently, but, leaning towards Donatus’s grammar, he only used the accident figura from the couple species/ figura in the sequences of accidents in his grammar. Despite the formal similarity and typological identity with Donatus’s definition, Melanch‐ thon’s definition was modified - for instance, in the definition of figura in the word class nomen, levels 3 and 3a (cf. example 5 above) were joined. Also, the definition represented an alternative with the verb esse, not accidere: (7) Melanchthon: Grammatica Philippi Melanchthonis Latina (1542: pages not num‐ bered), Melanchthon: Grammatica Philippi Melanchthonis Latina (1558: 68) (3) + (3a) Figurae item duae sunt, simplex ut aptus, composita ut ineptus. Word-formation in neo-Latin school grammar 57 <?page no="58"?> 35 (3) “Figura is [a category] [used to examine] whether a word is simple or compound.” 36 (3) “Species is [a category] used to examine whether a word is primary or derived.” Since Melanchthon was influenced by both Donatus and Priscianus, there were two typical features from both grammatical traditions that were notable in his grammars. The category species did not have the status of an accident (the influence of Donatus), yet Melanchthon included it in his text (the influence of Priscianus) as passages named de speciebus and de specie for a particular word class without defining this term. There was a different approach in Grammatica Latina by Petrus Ramus, who abandoned the concept of accidents; therefore, species and figura no longer performed in that role. Although the terms themselves remained, they indicated newly defined concepts. (8) Ramus: Grammaticae libri quatuor (1560: 15) (3) Figura est [qua disquiritur], utrum vox simplex aut composita sit. 35 (9) Ramus: Grammaticae libri quatuor (1560: 15) (3) Species est, qua disquiritur, primane sit vox an aliunde flexa. 36 Instead of an enumerative definition, Ramus created an explicative one (for more about this type of definition, see Zouhar 2014: 341). In terms of (compound) sentence structure, Ramus’s definition was represented by a relative clause followed by an indirect question. Although the meaning of the terms figura/ species was defined at the surface structure of the utterance, the focus of the utterance from the point of view of the information structure was the lexeme “word” (vox). Even though the third level was retained in the definition system, changes occurred in the content and typology of the definition. 4.2.3 Rhenius (1574-1639) Ramus’s definitions of the categories species and figura continued in gram‐ mars in different language and cultural territories over centuries. In the German environment, it was mainly in the mixed Melanchthonian-Ramean and Rhenian textbooks where this definition could be found. The only 58 Ľudmila Eliášová Buzássyová <?page no="59"?> 37 Cf. part 3.1 Neo-Latin grammars written in Latin. 38 The English term “head noun” is not an ideal term for designating a noun in the Latin phrase figura simplex and the like. This is because different ways of syntactical analysis are used in English and Latin, so the terminology is not compatible. As an alternative term, “superordinate noun” could be suggested. change noticed regarding the formulation of the definitions was in the term flexa (used by Ramus in the definition of species) that was replaced either with (originally Priscianus’s) derivativa (Grammatica Latina Philippo-Ramea 1596: 16, 1660: ibidem) or the new Rhenian deducta. (10) Rhenius: Grammatica Latina cum paralipomenis (1618, 1625: 10, 11) (3) Figura est, qua disquiritur, utrum simplex aliqua vox sit an composita. (3) Species est, qua disquiritur, primane sit vox aliqua an aliunde deducta. When analysing and comparing the definitions in mixed grammars and exemplary Ramean definitions, an important fact should be highlighted: mixed grammars returned to the concept of accidents; however, this was in a modified form as figura and species were termed accidentia generalia (cf. Rhenius 1618, 1625: 10; Melanchthon et al. 1661: 10). 37 - 4.3 Definitions of classes for species/ figura accidents 4.3.1 Donatus and Priscianus Neither Donatus nor Priscianus defined what the lexemes simplex, compo‐ sita, (decomposita), representing the classes of accident figura, meant on a conceptual and terminological level. They did not explain the lexemes primitiva/ principalis and derivata/ derivativa either, which denominated the classes of accident species. From the morphosyntactic point of view, the lexemes simplex, composita, and others were adjectives connected with their head nouns 38 figura and species with the category of agreement; thus, they were in the feminine gender. At level 3a (see 4.2.1; examples 5 and 6 above), the terms denoting the classes of accidents were determined by default only with an example introduced by the adverb ut positioned in front of the given example. The only exception was Priscianus’s definition of the verbal species, in which the relevant classes were also determined in the form of a relative clause (the remaining word classes did not have such properties). Word-formation in neo-Latin school grammar 59 <?page no="60"?> 39 (3) “There are two species in verbs, primary and derived, which, however, occur in almost all parts of speech. 3a1) The primary [i.e. species] is the one which, by its very nature, acquired the primary position, such as lego, ferveo, domo, facio; 3a2) the derivative, which is/ are derived from primary ones [i.e. species or verba? ], such as […].” 40 According to Luhtala’s classification (2002: 272), this is a semantic definition. 41 Following the principle of universaliter primum, Ramus explained figura and species for all word classes at the beginning of the chapter entitled Etymology; therefore, he also selected examples from several relevant word classes. (11) Priscianus: Institutiones grammaticae GL II, 1: 427.11-14 (3) Species sunt verborum duae, primitiva et derivativa, quae inveniuntur fere in omnibus partibus orationis. 3a1) est igitur [i.e. species] primitiva, quae primam positionem ab ipsa natura accepit, ut lego, ferveo, domo, facio, 3a2) derivativa, quae a positivis derivantur [species or verba? ], ut […]. 39 From the point of view of the typology of definitions, an explicative defini‐ tion was involved in 3a1 and 3a2. 40 If one comprehends the wording of the definition accurately, the subject changes at level 3a2 in the sentence structure of this definition. First, in 3a1, the noun species (feminine, n.sg.) is the subject, and then, in 3a2, the noun verba (neuter, n.pl.) becomes the subject, as it can be deduced from the plural form of the verb derivare. As far as the content is concerned, two questions can be asked: what is the definiendum in 3a2)? Is it a species, or are they verba? 4.3.2 Melanchthon and Ramus Latin grammar from the Renaissance followed late antiquity authors in the practice and form of determining the classes of the related accidents; for example, both Melanchthon and Ramus determined the classes of the figura category by using the same (or almost identical) examples as Donatus or Priscianus; for instance, simplex ut aptus, composita ut ineptus (Melanch‐ thon 1542, pages not numbered, Melanchthon 1558: 68); simplex ut doctus, amo, composita ut indoctus, redamo (Ramus 1560: 15). 41 Despite the formal similarity, a substantial change occurred in the content of the definitions and in the nature of the definiendum. In Ramus’s definition system, the classes of simplex/ composita and primitiva/ flexa became the classes of a new superior category “word”, vox defined at level 3. This implied that they were not classes of figura/ species accidents. 60 Ľudmila Eliášová Buzássyová <?page no="61"?> 42 (3 → 1) “The nouns (‘names’) can be either primary or derived. 3a) Primary are those that do not come from any other, such as equus, aper, etc.; derived are those that derive their origin from elsewhere, such as equuleus, aprugnus, etc.” According to Luhtala’s typology (2002), this is a semantic definition. 43 (3 → 1) “Some of the verbs are primary, called perfect, the others derived. 3a) The primary ones, like lego, dico, the derived ones, such as lectito, dictito.” (12) Ramus: Grammaticae libri quatuor (1560: 15) (3) Figura est [qua disquiritur], utrum vox simplex aut composita sit. 3a) simplex ut doctus amo; composita ut indoctus, redamo. (3) Species est, qua disquiritur, primane sit vox an aliunde flexa. 3a) Prima ut amo, flexa, ut amabilis. Although the lexeme vox was at the centre of the definition, the formulation at level 3a was somewhat ambivalent given that the adjectives simplex/ composita, from a morphosyntactic point of view, could bind to both the head noun vox as well as figura. A similar shift in class definition at level 3a could be seen in Melanchthon’s grammars. In Melanchthon’s definition, terms - the adjectives primitiva/ de‐ rivativa, which in late antique grammars had denominated two classes of the accident species - denoted the classes of a particular word class. In the word class nomen, Melanchthon - through the enumerative definition - specified two classes of nomina, and he classified two classes of verbs in the word class verbum. In his grammar, the key terms, i.e. the adjectives primitiva/ derivativa, therefore referred to the head noun nomina/ verba (neuter, pl.) and not to the noun species (feminine, sg.). (13) Melanchthon: Grammatica Philippi Melanchthonis Latina (1542 pages not numbered), Grammatica Philippi Melanchthonis Latina (1558: 178) (3 → 1) Sunt autem nominum primitiva alia, alia derivata. 3a) Primitiva di‐ cuntur, quae a nullo alio descendunt, ut equus, aper etc. derivata, quae aliunde originem trahunt, ut equuleus, aprugnus etc.  42 (14) Melanchthon: Grammatica Philippi Melanchthonis Latina (1542 pages not numbered), Grammatica Philippi Melanchthonis Latina (1558: 209) (3 → 1) Sunt et verborum alia primitiva, quae perfecta dicuntur, alia derivata. 3a) Primitiva, ut lego, dico. Derivata, ut lectito, dictito.  43 Word-formation in neo-Latin school grammar 61 <?page no="62"?> 44 (3) “Species is a [category] used to examine whether a word is primary or comes from elsewhere. 3a) The primary is the one that is not derived from elsewhere.” This is a semantic definition, according to Luhtala’s classification (2002). Comparing the two definitions, it is evident that while Priscianus’s definition of the verbal species at level 3a2, example 11 (GL II, 1: 427.11-14), was not explicit enough, similar definitions in Melanchthon’s grammar (examples 13 and 14) were more comprehensible. Classes of a particular word class became Melanchthon’s new definienda of the de speciebus passage. The two classes, which the late antiquity system recognized as classes of a word-class accident, developed into classes of a given word class in the new definition system (cf. changes from level 3 to level 1: 3 → 1). 4.3.3 Mixed grammars Mixed grammars were arranged by blending Melanchthonian and Ramean traditions. The components of this “blend” were mixed in a different propor‐ tion. In Rhenius’s grammar, the first explicative definition (3) that originated in Ramus’s grammar was followed by another explicative definition (3a) similar to Melanchthon’s one in terms of content. (15) Rhenius: Grammatica Latina cum paralipomenis (1618, 1625: 10) (3) Species est, qua disquiritur primane sit vox aliqua an aliunde deducta. 3a) Primitiva est, quae non aliunde derivatur, ut schola, ego, clamo, saepe.  44 As early as in Ramus’s definitions of the species/ figura concept at level 3, the lexeme ‘word’ - vox, emerged as the focus of utterance, which corresponded with the definition at level 3a defining the classes of words; however, definition 3a (cf. example 12: simplex ut doctus amo, compo‐ sita ut indoctus redamo) might seem ambivalent. Grammatica Latina Philippo-Ramea, 1660 was one of the examples of how to eliminate these inconsistencies. 62 Ľudmila Eliášová Buzássyová <?page no="63"?> 45 According to Luhtala’s classification, this is a semantic definition combined with an etymological one. Cf. Luhtala (2002: 272 sq.). (16) Grammatica Latina Philippo-Ramea (1660: 16) (3) Species est, qua disquiritur, primane sit vox an aliunde derivata, 3X) unde vox primitiva et derivata dicitur.  45 3a) Primitiva, ferrum, amo, clamo, cras, derivativa, ferreus, amabilis, clamito, crastinus. A new part (3X) was inserted into the system of definitions that arched over parts 3 and 3a. As a result, the hierarchy of the terms species and vox was evident, and it became apparent that the adjectives primitiva and derivata were not related to species but rather to the lexeme vox. The definiendum was the class of the lexical plane. The analysis of key terms and definitions in Latin grammars written in Latin (in section 4.2) suggested changes in the content and structure of definitions regarding key terms of the “w-f complex”. The character of the definienda was changing, both at level 3 (the accidents figura/ species) and at level 3a (the classes of accidents simplex/ composita and primitiva/ derivata), while levels 3 and 3a could merge. In definitions, the word as the unit of the lexical plane received a priority whether it was represented by the lexeme ‘word’- vox, or a specific word class (nomen, verbum). Also, a larger typological variety of definitions emerged than what was known in late antiquity. - 4.4 Definitions of the “w-f complex” in Latin grammars written in German The position of the accidents species/ figura in the role of the definienda was justified only to the time when the word-and-accident model served as the description model of word classes. In neo-Latin grammars, this model gradually lost its value and effectiveness, which exhibited changes in the definienda. The last phase of this process appeared in grammars written in vernacular languages. After this model’s final decline, the word remained at the centre of attention. In the analysed Latin grammars written in German, the accidents species/ figura (apart from some exceptions, such as the ones in Schultz’s texts) were not defined and level 3 was withdrawn from the system of definitions. The key terms from the original ancient level 3a - primitiva, derivativa, simplex, Word-formation in neo-Latin school grammar 63 <?page no="64"?> composita - clearly referred to the term “word/ words”, representing all word classes. The different alternatives of Latin definitions worded as Species est/ Species sunt or Figura est/ Figurae sunt… were replaced by a new definition and a new definiendum - Wörter sind. While in older or more traditionally arranged grammars, there were still some signs of the former definienda - i.e. in expressions (nach der Figur and nach der Specie) - they gradually disappeared. The continuity of the late antique form of the definition was nonetheless retained, and, just like that definition, the German one was also distinctive with its enumerative character listing the two classes of relevant categories. Like those of their ancient ancestors, the definitions in German written grammars were also limited to enumerating examples, where only surface changes arose during the eighteenth and nineteenth centuries, such as the replacement of Latin terms with German equivalents. (17) Cellarius: Erleichterte Lateinische Grammatica (1709, 1712, 1731, 1738: 2) Solche Wörter sind auch nach der Figur entweder Simplicia, schlechte und einfache, […] oder Composita, zusammen gesetzte, […] Decomposita, doppelt zusammen gesetzte […]. (18) Bröder: Practische Grammatik der lateinischen Sprach (1821: 5-6) Diese verschiedenen Wörter sind: […] entweder einfache Wörter. […] oder zusammengesetzte Wörter. Even in the nineteenth century, Latin grammars written in German had the “w-f complex” in the introduction to the Discipline of Words character‐ ized by the same reduced form and definition system as illustrated in the examples above (17, 18). This reduced “w-f complex” was also found in Latin grammars in vernacular languages affected by the German tradition. 5 Conclusion This research aimed to determine whether and how the way of word-forma‐ tion description changed in neo-Latin school grammars in the “observation period” (as referred to by Latacz). The research was limited to sourcebooks 64 Ľudmila Eliášová Buzássyová <?page no="65"?> from then German-speaking milieu between the sixteenth and early nine‐ teenth centuries: firstly, focusing on grammars written in Latin and then in German. The research analysis addressed two subject areas: (1) the location of text units with a word-formation content in grammars to determine their role in the process of separation of word-formation as an autonomous discipline as well as to compare this process in Latin and vernacular grammars (2) selected key terms and definitions of these text units Research findings of the first area suggested that the most critical change in the previously referred process established the “w-f complex” commencing the Discipline of Words, which was initiated by Petrus Ramus. The “w-f complex” developed from the accidents figura and species and was restricted to a description of the elementary classes simplicia - composita - decompo‐ sita and primitiva - derivata/ dirivativa. The other word-formation text units, labelled as appendices (later Anhangs), only covered derivation. After the separation of the “w-f complex” grammar books regarding the descriptions of word-formation phenomena only saw changes in the didactic strategy of relevant topics during the seventeenth and eighteenth centuries. The essential difference between progressive grammars written in German from 1780 to 1850 and their predecessors was that the opening “w-f complex” (with various titles) also covered the material from the original appendices and new illustrative material. On the one hand, the “w-f complex” expanded thematically; on the other hand, it acquired a more general character due to some synthesis and even a comparison with German, which corresponded to its function to briefly impart knowledge to students about the study of words. The successor to this “w-f complex” has been a chapter (and sometimes only a paragraph) about word-formation in grammar books used in current elementary Latin courses. Just like its neo-Latin forerunners, this chapter/ paragraph being placed at the beginning, or directly before the chapter on morphology, has been characterized by its minimal size. Research in the next selected area (terminology and definitions) re‐ vealed that the most considerable change was removing the terms species and figura from grammar, which corresponded with the decay of the word-and-accident model. This structural change was also reflected in the definition system change of individual word classes and their accidents. The Word-formation in neo-Latin school grammar 65 <?page no="66"?> 46 This research was supported by the grant Vega 1/ 0733/ 18, The Concept of Ancient Grammar in the 17 th and 18 th -Century Grammatical Tradition in the Territory of Present-day Slovakia and the Broader European Context. My gratitude goes to two anonymous referees for their comments and additions. definition system defining individual word classes and their accidents in late antique grammars changed to a system that only defined word classes and gradually only words in neo-Latin grammars. The original accidents species and figura lost their position as definienda, which was replaced by the word classes/ words and their classes (primitiva and derivata/ derivativa, simplicia and composita, decomposita). This process culminated in Cellar‐ ius’s grammar books. Also, from a typological point of view, the definitions of word-formation terms changed during the analysed period. The type of definition that seemed to keep its position steadily throughout centuries in Latin grammar was the enumerative one. Although Latin grammar remained conservative in its character throughout centuries, this research has demonstrated that some changes in text structure and didactic approach occurred in the field of word-formation description; however, more significant changes were detected in formu‐ lating definitions of word-formation concepts. All in all, word-formation seems to be a justifiable subject area that would deserve more attention from researchers in the future, primarily regarding the investigation of interconnections between Latin and vernacular grammars. 46 Bibliography - Primary Sources Bröder, Christian Gottlob ( 1 1787): Practische Grammatik der lateinischen Sprache. Leipzig: Crusius. (Further (2 nd ) edition 1793). Bröder, Christian Gottlob ( 15 1821): Practische Grammatik der lateinischen Sprache. Leipzig: Vogel. Bröder, Christian Gottlob ( 4 1803): Kleine lateinische Grammatik mit leichten Lectionen für Anfänger. Leipzig: Crusius. (1 st edition 1795). Cellarius, Christophorus ( 1 1689): Erleichterte Lateinische Grammatica Oder Kurtze doch zulängliche Anweisung zur Lateinischen Sprach. Merseburg: Forberger. (Fur‐ ther editions 1697 (4 th ), 1704 (7 th ), 1709 (9 th ), 1720 (? ), 1738 (21 st ). 66 Ľudmila Eliášová Buzássyová <?page no="67"?> Cellarius, Christophorus ( 1 1689): Latinitatis probatae et exercitae liber memorialis naturali ordine dispositus. Merseburg: Forberger. Donatus: Ars maior. In: GL IV (1864), 367-402; Holtz, Louis (ed.) (1981): Donat et la tradition de l’enseignement grammatical. Étude sur l’Ars Donati et sa diffusion (IVe-IXe siècle) et édition critique (Documents, Études et Répertoires publiés par l’Institut de Recherche et d’Histoire des Textes 21). Paris: CNRS, 603-674. DLG 1663 = Rhenius, Johannes (1663): Donatus Latino-Germanicus seu ratio declinandi et conjugandi cum vocabulis declinationum. Lipsiae: Schürer/ Götz. (Further edition Lipsiae: Lanckisch 1737). DLH 1697 = Rhenius, Johannes (1697): Donatus Latino-Hungaricus, ostendens rationem declinandi, comparandi, conjugandi, formandi, & rudimentalia exercitia compo‐ nendi, ed. Michael Missowicz. Leutschoviae: Brewer. (Further edition 1750). DLGHB 1748 = Doleschalius, Paulus (1748): Donatus Latino-Germanico- Hungarico-Bo‐ hemicus, quem ut desiderata donaret facilitate: summam eorum, quae tironibus linguae Latinae, cognitu in primis necessaria veniunt. Posonii: Royer. Doleschalius, Paulus 1746: Grammatica Slavico-Bohemica… Ratio Accuratae scrip‐ tionis & flexionis […] Praefatus est Matthias Belius. Posonii: Typis Royerianis. Gesner, Johann Matthias (ed.) ( 1 1740): Christoph Cellarii erleichterte Lateinische Grammatic zum Gebrauch der Schulen in den Chur-Braunschweigischen Landen. Goettingen: Vandenhoeck. (Further (2 nd ) edition Leipzig: Böhme 1786). GL = Grammatici Latini (1855-1880), Heinrich Keil (ed.), 8 vols. Leipzig: B. G. Teubner. (Repr., Hildesheim: Georg Olms 1981.) Grotefend, Georg Friedrich ( 4 1823/ 1824): Größere lateinische Grammatik für Schulen, nach Wenck's Anlage umgearbeitet. Frankfurt am Main: Varrentrapp. (1 st edition 1820). Grotefend, Georg Friedrich ( 2 1825): Kleine lateinische Grammatik für Schulen. Frank‐ furt am Main: Varrentrapp. Gueintz, Christian (1641): Deutscher Sprachlehre Entwurf. Cöthen: [s. n.]. Frischlin, Nicodemus (1609): Grammatica Latina compendiose scripta […] Francofurti: Iohannes Spiessius. (1 st edition: Tubingae: Gruppenbachius 1585). Irson, Claude ( 2 1662): Nouvelle méthode pour apprendre facilement les principes et la pureté de la langue française. Paris: Beaudouin. Melanchthon, Philipp (1542): Grammatica Philippi Melanchthonis Latina, iam denuo recognita et plerisque in locis locupletata, [ed. Jacob Micyllus]. Lipsiae: Schumann. Melanchthon, Philipp (1558): Grammatica Philippi Melanchthonis Latina iam denuo recognita […] annotationes recentes de consilio Ioachimi Camerarii. Lipsiae: Bapst. Word-formation in neo-Latin school grammar 67 <?page no="68"?> Melanchthon et al. (1661) = Philippi Melanchthonis Grammatica Latina… ex Iul. Caes. Scaligero, Petro Ramo, Nicod. Frischlino, Josepho Scaligero, Johanne Wanckelio, Jo‐ hanne Rhenio, recognita & locupleta, ed. Erasmus Schmidt. Wittebergae: Selfischius. Melanchthon, Philipp/ Ramus, Petrus (1596): Grammatica Latina Philippo-Ramea, ad faciliorem puerorum captum perspicua methodo breviter conformata, et duobus libris distincta […] Sigenae Nassoviorum: Corvinus. (Further edition Steinfurti: Wellenberg 1660). Priscianus: Institutionum grammaticarum libri XVIII, Martin Hertz (ed.). In: GL II-III. Ramus, Petrus (1560): Grammaticae libri quatuor. Parisiis: Wechelus. Ratke, Wolfgang [after 1630]: “Wortbedeutungslehr der Christlichen Schule […]”. In: Ising, Erika (ed.) (1959): Wolgang Ratkes Schriften zur Deutschen Grammatik (1612-1630). Berlin: Akademie-Verlag. Rhenius, Johannes (1618): Grammatica Latina cum paralipomenis. Lipsiae: Am Ende (Erben). (Further edition Lipsiae: Lanckisch 1625). Scheller, Immanuel Johann Gerhard (1779): Ausführliche lateinische Sprachlehre oder soganannte Grammatik. Leipzig: Fritsch. Scheller, Immanuel Johann Gerhard ( 2 1786): Kurzgefasste lateinische Sprachlehre. Leipzig: Fritsch. (1 st edition 1780). Schottelius, Justus G. (1663): Ausführliche Arbeit von den Teutschen HaubtSprache, Braunschweig, zilligen [rist. 1995, Tübingen, Niemeyer]. Schultz, Otto ( 9 1836): Schulgrammatik der lateinischen Sprache. Halle: Waisenhaus. Schultz, Otto (1825): Ausführliche lateinische Grammatik für die Oberen Klassen Ge‐ lehrter Schulen. Halle: Waisenhaus. (Further edition 1836). - Secondary Sources Alfieri, Luca (2019): “La storia della derivatio, il problema del tempo e le grammatiche “filosofiche” tra il XIII e XVIII secolo”. Linguistica e Filologia 39: 63-105. Ax, Wolfram (1986): Laut, Stimme und Sprache. Studien zu drei Grundbegriffen der antiken Sprachtheorie. Göttingen: Vandenhoeck & Ruprecht. Baratin, Marc (2000): “A l’origine de la tradition artigraphique latine, entre mythe et réalité”. In: History of the Language Sciences. An International Handbook on the Evolution of the Study of Language from the Beginnings to the Present (Handbücher zur Sprach- und Kommunikationswissenschaft; Bd. 18. Vol. 1), Silvain Auroux et al. (eds.). Berlin/ New York: de Gruyter, 459-466. Barwick, Karl (1922): Remmius Palaemon und die römische Ars grammatica (= Philo‐ logus Supplementband 15. 2.). Leipzig: Dietrich 1922. 68 Ľudmila Eliášová Buzássyová <?page no="69"?> Brnadić, Teodora Shek (2015): “The Enlightenment’s Choice of Latin: The Ratio educa‐ tionis of 1777 in the Kingdom of Hungary”. In: Latin at the Crossroads of Identity. The Evolution of Linguistic Nationalism in the Kingdom of Hungary, Central and Ea stern Europe, Volume: 5, Gábor Almási (ed.). Brill, 119-151. Burkard, Thorsten (2003): “Die lateinische Grammatik im 18. und frühen 19. Jahrhun‐ dert. Von einer Wortartenzu einer Satzgliedgrammatik”. In: Germania latina - Latinitas teutonica. Politik, Wissenschaft, humanistische Kultur vom späten Mitte‐ lalter bis in unsere Zeit, Vol. 2, Eckhard Keßler/ Heinrich C. Kuhn (eds.). München: Fink, 781-830. Buzássyová, Ľudmila (2019): “From Ancient Species and Figura Accidents to the Rudiments of the Word-Formation Discipline in Latin and Vernacular Grammars (16th to 18th centuries)”. Listy filologické = Folia philologica 142(3-4): 407-443. Copeland, Rita/ Sluiter, Ineke (eds.) (2009): Medieval Grammar and Rhetoric. Language Arts and Literary Theory AD 300-1475. Oxford: Oxford University Press. Eliášová Buzássyová, Ľudmila/ Machajdíková, Barbora: (2020): “Nie je donát ako donát”. In: Hortus Graeco-Latinus Cassoviensis: zborník príspevkov z klasickej filológie, latinskej medievalistiky a neolatinistiky. Košice: Univerzita Pavla Jozefa Šafárika v Košiciach, 127-148. Fazekas, Tiborc (2000): “Normativ orientierte Sprachforschung in Ungarn”. In: History of the Language Sciences. An International Handbook on the Evolution of the Study of Language from the Beginnings to the Present (Handbücher zur Sprach- und Kommunikationswissenschaft; Bd.-18. Vol. 1), Silvain Auroux et al. (eds.). Berlin: Walter de Gruyter, 916-919. Häkkinen, Kaisa (2000): “Early Grammatical Descriptions of Finno-Ugric”. In: History of the Language Sciences. An International Handbook on the Evolution of the Study of Language from the Beginnings to the Present (Handbücher zur Sprach- und Kommunikationswissenschaft; Bd.-18. Vol. 1), Silvain Auroux et al. (ed.). Berlin: Walter de Gruyter, 806-814. Hauser, Přemysl (1959): “Dobrovského práce o tvoření slov a domácí mluvnická tradice”. In: Sborník Vysoké školy pedagogické: jazyk - literatura, I: Studie o jazyce a literatuře národního obrození. Praha: 25-54. Holtz, Louis (1981): Donat et la tradition de l’enseignement grammatical. Étude sur l’Ars Donati et sa diffusion (IVe-IXe siècle) et édition critique (Documents, Études et Répertoires publiés par l’Institut de Recherche et d’Histoire des Textes 21). Paris: CNRS. Word-formation in neo-Latin school grammar 69 <?page no="70"?> Huterer, Andrea (2001): Die Wortbildungslehre in der Anweisung zur Erlernung der Slavonisch-Rusischen Sprache (1705-1729) von Johann Werner Paus (Slavistische Beiträge 408). München: Otto Sagner. Ising, Erika (1970): Die Herausbildung der Grammatik der Volkssprachen in Mittel- und Osteuropa. Berlin: Akademie-Verlag. Kaltz, Barbara/ Leclercq, Odile (2015): “Word-Formation: From its Beginnings to the 19th Century”. In: Word-Formation. An International Handbook of the Languages of Europe, Vol. 1, Peter O. Müller et al. (eds.). Berlin/ Boston: De Gruyter, 22-37. Koupil, Ondřej ( 2 2015): Grammtykáři: gramatografická a kulturní reflexe češtiny 1533- 1572. Praha: Karolinum. (1st edition 2007). Kristol, Andres Max (2000): “Les premières descriptions grammaticales du français”. In: History of the Language Sciences. An International Handbook on the Evolution of the Study of Language from the Beginnings to the Present (Handbücher zur Sprach- und Kommunikationswissenschaft; Bd. 18. Vol. 1), Silvain Auroux et al. (ed.). Berlin: Walter de Gruyter, 764-770. Latacz, Joachim (1979): “Die Entwicklung der griechischen und lateinischen Schul‐ grammatik”. In: Handbuch der Fachdidaktik. Fachdidaktisches Studium in der Leh‐ rerbildung. Alte Sprachen 1, Joachim Gruber/ Friedrich Maier (eds.). München: Oldenbourg 1979, 193-221. Law, Vivien (2003): History of Linguistics in Europe: From Plato to 1600. Cambridge: Cambridge University Press. Lepschy, Giulio (ed.): History of Linguistics III: Renaissance and Early Modern Linguis‐ tics. London: Longman 1998. Lindauer, Josef/ Pfaffel, Wilhelm ( 3 2019): Roma. Lateinische Grammatik. Bam‐ berg/ München: C. C. Buchner. Lindner, Thomas (2002): Lateinische Komposita: morphologische, historische und lexikalische Studien (Innsbrucker Beiträge zur Sprachwissenschaft, Bd.-105). Inns‐ bruck: Institut für Sprachen und Literaturen der Universität Innsbruck. Lindner, Thomas (ed.) (2011-2015): Indogermanische Grammatik, Band IV: Wortbil‐ dungslehre. Teil 1: Komposition (Indogermanische Bibliothek, Reihe 1: Lehr- und Handbücher, Grammatik). Heidelberg: Winter. Lindner, Thomas (2015): “Word-Formation in Historical-Comparative Grammar”. In: Word-Formation. An International Handbook of the Languages of Europe, Vol. 1, Peter O. Müller et al. (eds.). Berlin/ Boston: De Gruyter Mouton, 38-51. Lindner, Thomas/ Rainer, Franz (2015): “Word-Formation in Neo-Latin”. In: Word-For‐ mation. An International Handbook of the Languages of Europe, Vol. 3., Peter O. Müller et al. (eds.). Berlin/ Boston: De Gruyter Mouton, 1580-1597. 70 Ľudmila Eliášová Buzássyová <?page no="71"?> Luhtala, Anneli (2002): “On Definitions in Ancient Grammar”. In: Grammatical Theory and Philosophy of Language in antiquity (Orbis supplementa 19), Pierre Swiggers/ Alfons Wouters (eds.). Leuven/ Paris: Peeters, 257-285. Luhtala, Anneli (2005): Grammar and Philosophy in Late Antiquity (Studies in the History of Language Sciences, Vol. 107). Amsterdam/ Philadelphia: John Benjamins B.V. Marjamäki, Lauri (2018): Hin zur Wortbildung: Ableitung und Zusammensetzung in vormodernen deutschen Grammatiken und ihren klassischen Vorbildern. Masterar‐ beit. Betreuer: Dr. Aino Kärnä. Universität Helsinki: Institut für moderne Sprachen Germanistik. <https: / / helda.helsinki.fi/ bitstream/ handle/ 10138/ 277503/ Marjamaki _Lauri_Pro_gradu_2018.pdf> (Last access 29.11.2021) Matthews, P. H. (2019): What Graeco-Roman Grammar was about? Oxford: Oxford University Press. Müller, Peter et al. (2015): Word-Formation, An International Handbook of the Lan‐ guages of Europe, Vol. 1. Berlin/ Boston: De Gruyter Mouton. Padley, George Arthur (1976): Grammatical Theory in Western Europe, 1500-1700: The Latin tradition. Cambridge: Cambridge University Press. Padley, George Arthur (1985): Grammatical Theory in Western Europe, 1500-1700: Trends in Vernacular Grammar, I. Cambridge: Cambridge University Press. Pinborg, Jan (1967): Die Entwicklung der Sprachtheorie im Mittelalter. Münster: Aschendorff. Poppe, Erich (2000): “Early Grammatical Descriptions of the Celtic Languages”. In: History of the Language Sciences. An International Handbook on the Evolution of the Study of Language from the Beginnings to the Present (Handbücher zur Sprach- und Kommunikationswissenschaft; Bd.-18. Vol. 1), Silvain Auroux et al. (eds.). Berlin: Walter de Gruyter, 800-805. Re, Alessandro (2020): La compositione nominale latina (Innsbrucker Beiträge zur Sprachwissenschaft; Bd. 165). Innsbruck: Institut für Sprachwissenschaft der Uni‐ versität Innsbruck. Sarmiento, Ramon (2000): “Die Königliche Spanische Akademie und die Pflege der Nationalsprache”. In: History of the Language Sciences. An International Handbook on the Evolution of the Study of Language from the Beginnings to the Present (Handbücher zur Sprach- und Kommunikationswissenschaft; Bd. 18. Vol. 1), Silvain Auroux et al. (eds.). Berlin: Walter de Gruyter, 863-870. Sellberg, Erland (2020): “Petrus Ramus”. In: The Stanford Encyclopedia of Philos‐ ophy (Winter 2020 Edition), Edward N. Zalta (ed.). <https: / / plato.stanford.edu/ ar‐ chives/ win2020/ entries/ ramus/ > (Last access 17.1.2022) Word-formation in neo-Latin school grammar 71 <?page no="72"?> Schad, Samantha (2007): A Lexicon of Latin Grammatical Terminology (Studia Erudita 6). Pisa/ Roma: Fabrizio Serra Editore. Stroh, Wilfried (2007): Latein ist tot, es lebe Latein! : Kleine Geschichte einer großen Sprache. Berlin: List. Škoviera, Daniel (1985): “Cellariov slovník Liber memorialis v úprave Mateja Bela”. Zprávy Jednoty klasických filologů 27(1-3): 70-78. Tavoni, Mirko (1998): “Renaissance Linguistics”. In: History of Linguistics III: Renais‐ sance and Early Modern Linguistics, Giulio Lepschy (ed.). London: Longman, 1-108. Vaahtera, Jaana (1998): Derivation. Greek and Roman Views on Word Formation. Turku: Turun Yliopisto. Vaahtera, Jaana (2014): “Word Formation (paragōgḗ/ súnthesis), Ancient Theories of”. In: Encyclopedia of Ancient Greek Language and Linguistics, Volume 3, Georgios K. Giannakis (ed.). Leiden/ Boston: Brill, 529-532. Verrac, Monique (2000): “Les premières descriptions grammaticales de l'anglais”. In: History of the Language Sciences. An International Handbook on the Evolution of the Study of Language from the Beginnings to the Present (Handbücher zur Sprach- und Kommunikationswissenschaft; Bd.-18. Vol. 1), Silvain Auroux et al. (eds.). Berlin: Walter de Gruyter, 771-777. Zouhar, Marián (2014): “Klasifikácia definícií”. Teorie vědy/ Theory of Sciences 36(3): 337-357. 72 Ľudmila Eliášová Buzássyová <?page no="73"?> Grammatical and grammatographical categories - accord and conflict Gabriela Múcsková Abstract: The article is based on the premise that the grammatical description of a language is a scientific construct determined by the properties of the language and the grammatographical tradition. Some disagreements between grammatical descriptions and properties of the language are usually conditioned by the development of the language, the development of linguistic approaches to relations and categories of the language structure, but also by language planning following the previously approved prescription. That is why some phenomena are often referred to as “problematic” or “controversial” in academic linguistic works, and this is often reflected in their variable interpretation in grammatical, lexicographical, and didactic works. In more detail, the paper presents two phenomena - the classification of the independent grammatical morpheme by in Slovak conditional constructions (e.g. spal by som ‘I would sleep’), and the classification of l-participles as adjectives - in which the grammatical categorization is based on grammatographical tradition rather than the language properties and functions. The aim is to point out that the ways of lin‐ guistic constructing of the picture of grammatical structure as well as the disagreements between the properties of language units and their description reflect the context in which language and its linguistics develop. Keywords: Slovak, grammatical description of a language, grammat‐ ical categories, linguistic levels, word classes, language development, linguistic discussions <?page no="74"?> 1 In the Bibliography listed as Kačala/ Pisárčiková/ Považaj (eds.) (2003). 2 In the Bibliography listed as Buzássyová/ Jarošová (eds.) (2006). 1 Introduction - a reflection of language in grammatography 1.1 Grammatical description of a language, however detailed and extensive, cannot objectively and completely reflect the whole linguistic structure or the specific structural features of all linguistic units. Any grammatical description of language is a kind of scientific linguistic-reflexive construct. On the one hand, it is conditioned by the typological and structural features of the language, but on the other hand, it is determined by the linguistic knowledge and its development, the theoretical or methodological approach to the structural relationships and categories analysis, and - to a certain ex‐ tent - also by the grammatographical tradition as it is preserved or modified in scientific or didactic books. A certain role is also played by the historical development of the language (or by the reflection of the knowledge about this development in the descriptions of the contemporary state) and by its synchronic dynamics and development tendencies. In the Slovak linguistic descriptive tradition, it is indispensable to mention also the role of language policy and prescription as the conscious effort to regulate language use in compliance with the earlier presented and accepted codification or the earlier determined hierarchy of codification criteria. Within the interplay of all the factors, it is evident that despite the best efforts of the scholar describing the language, it is not possible to reach a complete harmony between the natural and real language and its linguistic reflection, or between the language awareness and its linguistic modelling. Even the term grammar does not have only one meaning in linguistic discourse. Sometimes it is delimited as morphology and syntax, at other times as the structure of language as a whole, and in folk perception, it is narrowed down to linguistic correctness in the sense of standard language codification (cf. Dudok 2003: 19; cf. also the entry “grammar” in contempo‐ rary dictionaries of the Slovak language: Krátky slovník slovenského jazyka (‘Short Dictionary of the Slovak Language’) - further herein only as KSSJ 2003 1 ; Slovník súčasného slovenského jazyka A-G (‘Dictionary of Contempo‐ rary Slovak Language A-G’) - further herein only as SSSJ 2006 2 ). As it has been concisely expressed by Dudok (2003: 19): “[…] as if in language and grammar there did not exist one truth. We do not deal here with a hierarchy 74 Gabriela Múcsková <?page no="75"?> 3 In the original: “[…] (v) jazyku a v gramatike akoby nejestvovala jedna pravda. Nemáme tu do činenia s hierarchiou právd, ale s pravdou hierarchií a vďaka tomu, že ide o metajazykovú disciplínu, často aj s hierarchiou autorít”. 4 On various approaches to delimiting grammatical categories cf. Horecký (2003). of truths but with the truth of hierarchies, and, thanks to the fact that what is concerned is a metalinguistic discipline, also with a hierarchy of authorities.” 3 1.2 In comparison with the description of Romance or Germanic lan‐ guages but also in comparison with the description of some closely related Slavic languages, the tradition of the Slovak grammatical description of language is relatively young. Its beginnings go only as far as the time of the beginnings of the standardization of the Slovak language, i.e. the first codification attempts in the 18 th century. Based also on ancient Roman - i.e. Latin and later European - grammatical tradition, Slovak grammars usually distinguish word classes (as naming linguistic categories, namely nominality, qualitativeness, processuality and relationality - cf. Horecký 2003: 39-40) and their grammatical categories as generalized meanings related to a particular word class. Within the general grammatical catego‐ ries and based on elementary grammatical meanings (grammemes), the paradigms of grammatical forms are constructed, in which their members stand in opposition and usually differ by a specific formal element. 4 It is the inflective character of Slovak and variability of morphological forms that - in the grammatical issues - foregrounds the morphology and the description of grammatical endings in morphological structures of inflected word classes and their function in syntactic clauses. Hence, it is the aim to describe the relationship between content, form, and function (often with stress laid on the form), which prevails in the tradition of Slovak grammatographic conceptions. It is mostly focused on the standard variety of language, therefore there is usually present also the prescriptive-codifi‐ cation aspect (cf. Furdík 2003: 42-43). A specific type of grammatical description is represented by lexicographic works that also present a set of selected grammatical information which, however, is determined by the previously defined rules and criteria of selection of the “mapped” grammatical (categorial and paradigmatic) phe‐ nomena. Dictionaries selectively reflect the grammatical structure of the language by selected and limited apparatus of qualifiers and schematized Grammatical and grammatographical categories - accord and conflict 75 <?page no="76"?> 5 In the original: “Je to štandardná deskriptívna morfológia spracovaná v metodolo‐ gickom rámci klasickej systémovej lingvistiky a doposiaľ plní funkciu fundamentál‐ neho kompendia, o ktoré sa opierajú všetky menšie morfologické práce vypracované na špecifické účely.” expressions while, of course, abstracting from more detailed grammatical subcategories or word classes, e.g. light verbs (cf. Ivanová/ Kyseľová 2012), predicative constituents (Ružička 1956: 10-13; Ďurovič 1955: 125-129) etc. On the other hand, Slovak descriptive grammars are often criticized for not paying enough attention to synsemantic lexical units, connectors, prep‐ ositions, or other “peripheral” but communicatively frequent phenomena (cf. Dudok 2003: 21). In this respect, it is paradoxically the dictionaries that provide more complex - above all semantic and functional - grammatical in‐ formation about these “neglected” elements. Moreover, lexicographic works also indicate insufficiently described or categorially ambiguous elements in the metalanguage modelling of grammar and they reveal grammatographic “problems” and “controversial issues”. 1.3 The question of the so-called “problematic phenomena” or “contro‐ versial issues” of the grammatical or morphological description of language recurrently appears in grammatically and lexicographically oriented works or discussions, even in the context of the already existing synthetic gram‐ matical works but especially those, which only are intended to be written. For example, the birth of the Morfológia slovenského jazyka (‘Morphology of the Slovak Language’, Ružička et al. 1966) was preceded by several discussions and a conference on Slovak grammar and lexis (cf. the journal Slovenská reč 1956, No. 1-2, and Oravec 1956). A summary of the basic issues discussed was provided in the paper by J. Ružička Sporné otázky slovenskej morfológie (‘Controversial Questions of Slovak Morphology’, Ružička 1956) and other linguists. Ultimately, the intended project of a complex synchron‐ ical description of Slovak was carried out “only” in the form of the synthetic Slovak morphology description on nearly nine hundred pages. Up to now, this monumental work has not been surpassed and keeps being an inspiring work and a source for later as well as current scholarly and didactic grammatical works. “It is a standard descriptive morphology written within the methodological framework of classical systemic linguistics and up to the present has been fulfilling the role of a fundamental compendium relied upon by all smaller morphological works written for specific purposes” 5 76 Gabriela Múcsková <?page no="77"?> 6 In the original: “[…] sa poznatky zdali pre mnohých ‘posledným slovom’, ktoré sa ťažko prekonáva”. 7 From among the large number of works, at least the following ones can be men‐ tioned: Bosák/ Buzássyová (1985); Sokolová et al. (1999); Furdík (2004); Sokolová/ Iva‐ nová/ Ološtiak (2006); Sokolová et al. (2012); Ološtiak (2015); Ološtiak et al. (2021) etc. 8 For example, Dvonč (1984); Nábělková (1993); Sokolová (1995); Nižníková/ Sokolová (1998); Dolník (2005); Ivanová (2006); Sokolová/ Ivanová (2006); Kačala (2006); Soko‐ lová (2007); Ivanová (2009); Dolník (ed.) (2010); Jarošová (2013a), Jarošová (2013b); Kyseľová/ Ivanová (2013); Kesselová et al. (2013); Janočková (2014); Sokolová/ Žigo (2014); Ivanová (2017); Kyseľová (2017); Ološtiak/ Ološtiaková (2018); Ivanová et al. (2018) and others. 9 Kesselová (2001), Kesselová (2014); Slančová (ed.) (2018) etc. 10 The same also applies for several works listed in footnotes 7 and 8. From among gram‐ matical words based exclusively on corpus data analyses, we could also mention (Dolník 2003: 32). One of its “unsurpassed” qualities is the authority which it has in Slovak linguistics, and thanks to which the “[…] findings for many seemed to be ‘the last word’ that is difficult to overcome” 6 (Furdík 2003: 44). At the beginning of the 21 st century, the Slovak linguistic community returned to the idea of a new academic description of the Slovak morphology and syntax. Again, discussions were carried out, and the conference Tradície a perspektívy gramatického výskumu na Slovensku (‘Traditions and Perspec‐ tives of Grammatical Research in Slovakia’; cf. proceedings by Šimková (ed.) 2003) took place. Although the efforts were directed at the preparation of a new - academic - grammar of Slovak including also syntax, no more up-to-date complex and detailed work arose, and its perspectives are still uncertain. However, it cannot be claimed that Slovak grammatical research has not developed. On the contrary, grammatical linguistic research has markedly increased. However, it concentrated upon various partial areas of grammar and their detailed qualitative and/ or quantitative analysis applying various approaches, e.g. functional, communicational, constructional, corpus-re‐ lated, sociolinguistic, explanatory, or cognitive. Authors mostly deal with morphology and the closely related word-formation 7 topics, or with the par‐ ticular word classes and their grammatical or morphosyntactic categories 8 , including their ontogenesis 9 . The face of grammatical description has been considerably changed in the context of the developing corpus linguistics and natural language processing. 10 Analyses based on material from corpus databases have also led to the re-evaluation of some of the so far generally accepted and by Grammatical and grammatographical categories - accord and conflict 77 <?page no="78"?> Garabík et al. (2016); Majchráková/ Chlpíková/ Bobeková (2017); Ďurčo/ Majchráková (2017); Zumrík/ Majchráková/ Miháliková (2022). 11 At this place, only some examples that are most often stated in conceptual grammat‐ ical and lexicographic works will be mentioned. Two phenomena that we have encountered within our own research aimed at the historical development of lan‐ guage and the processes of grammaticalisation are analysed in more detail. tradition established classifications. On the other hand, unsolved “problem‐ atic questions” in the description of the grammar cause problems also in computer and corpus processing of the Slovak language and creating corpus tools. Nevertheless, the thematically oriented research of some selected gram‐ matical aspects or phenomena - in a more detailed and profound form than which could be brought by the new complex description of the grammar of the language - does not (and cannot) provide clear and unequivocal solutions of those “controversial” and “problematic” phenomena. It rather arrives at the analysis of the nature of the “problematic character” of the particular grammatical phenomenon, or even at revealing new “controver‐ sial” and “problematic” questions. 2 “Problematic” phenomena in the grammatical description of language From the grammatographical point of view, it is interesting to see which morphological phenomena are in grammatical works and case studies most often labelled as “problematic” and “controversial”. Their character is very varied - they range from questions concerning concrete structural morpho‐ logical phenomena and forms to theoretical conceptions of delimiting and understanding the grammatical categories of various types and degrees of generalization. 11 - 2.1 Variant forms Probably the largest group attracting the attention of descriptive as well as prescriptive linguistics is formed by variant grammatical forms, which in lexicographical and normative works are usually labelled as the so-called variant forms (in Slovak ‘dublety’, i.e ‘double forms’). On the synchronic level, these forms can seem to demonstrate a violation of the regularity of the morphological inflectional paradigmatic system, but they are a natural 78 Gabriela Múcsková <?page no="79"?> result of divergent and convergent tendencies within the spontaneous development of language, the impact of analogy and the levelling of the structural formants within the competition, and of intraparadigmatic and interparadigmatic models of forms. From the diachronic point of view, the morphological variants seem to have resulted from the re-evaluation of older structural types of declension and conjugation. In Slovak as a synthetic inflectional language, there is a large number of such examples, many of them also being the subject of discussions or polemics, often with the appli‐ cation of the prescriptive point of view. These cases are not “problematic” in terms of grammatography and, with regard to their number, we will not deal with any particular example. - 2.2 Linguistic levels A different group of questions concerns the specification of the borderlines of grammar itself and within it the borderlines between morphology and syntax, or the place of word-formation within the morphological description of language. In Slovak linguistic tradition, the area of word-formation is understood as part of the lexical level of language. This seems to be “prob‐ lematic” with regard to the affixes that - during the language development - acquired a grammatical function. This area of questions is more generally connected with the issue of the affiliation of some linguistic units to the traditionally delimited language levels, i.e. to the morphonological, morphological, syntactic, or lexical levels. In addition to the above-mentioned inclusion of word-formation into the lexical or grammatical plan, in earlier periods there also appeared discus‐ sions about the place of some grammatical categories in the morphological or the syntactic plan of language (e.g. the grammatical category of person - cf. Horecký 2003: 39-40). Another source of ambiguous classification is the case of transitional units that in the process of development transcend the boundaries of language levels. This concerns above all autosemantic lexemes which in certain con‐ texts (grammatical constructions) undergo a process of grammaticalization, and to various degrees acquire the status of units with a grammatical function, i.e. they pass from the lexical to the morphological level. A typical example of such a process is the lexical and the grammatical status of the present forms of the verb byť ‘to be’, which functions as an autosemantic verb and a copula in predicative constructions, as well as an auxiliary Grammatical and grammatographical categories - accord and conflict 79 <?page no="80"?> 12 In the original: “[…] ako samostatná morféma, pomocné sloveso v zložených sloves‐ ných tvaroch”. 13 Recorded in historical sources are the forms of 1sg. bych, 2/ 3sg. by, 1pl. bychom/ by‐ chme (as well as bychmy), 2pl. byste/ byšte/ by, 3pl. bychu/ byšę/ by, 1du. bychově/ bychova, 2/ 3du. bysta/ byšta, which in their form also reflect the influence, or interference with the forms of Proto-Slavic imperfect (for more details see Kosek 2014: 262; 2017). (e.g. in the forms of passive and resultative). As a grammatical morpheme, it functions in analytical forms of Slovak preterit and conditional what is reflected in academic morphological works; e.g. Ružička et al. (1966: 474, 480); Ološtiak/ Ološtiaková (2018: 39-51); cf. also Múcsková (2016: 138- 147). In spite of the fact that in analytical forms of the Slovak preterit and conditional the forms of the verb byť have only the status of the grammatical morpheme expressing the category of person, in textbooks and handbooks of grammar the authors speak about the “auxiliary” verb byť (e.g. Oravec/ Bajzíková/ Furdík 1984: 148; Považaj (ed.) 2013: 69). In dictionaries, the auxiliary and the grammatical morpheme are presented as synonymous, e.g. “[…] as an autonomous morpheme, an auxiliary verb in complex verbal forms” 12 (SSSJ 2006). Another example is the element sa - originally an ancient Proto-Slavic reflexive pronoun that as such functions also in con‐ temporary Slovak. However, in its petrified accusative form, sa has also acquired the character of a grammatical morpheme (in passive forms or impersonal verbal forms), or of a derivational morpheme in various types of reflexive verbs, or an empty lexical morpheme (cf. Jarošová 2003; Hašanová 2012; Ivanová/ Kyseľová 2018 and others). But in textbooks, the element sa is in all cases uniformly named as a “reflexive pronoun”. ※ In the following, we focus in more detail on the element by that in contem‐ porary Slovak is a grammaticalized component of analytical forms of the conditional (spal by som = ‘I would sleep’). Its origin, its process of gramma‐ ticalization, and the semantic transformation of the whole construction does not correspond to the grammatographic categorization of this element nor its grammatical function in contemporary Slovak. Historically, it is the form of the verb byť, i.e. the form of the original Proto-Slavic aorist bych  13 , which in the construction with the l-form of verbs (originally a participle) was grammaticalized into a grammatical mor‐ pheme 14 within the analytical verbal form of the Proto-Slavic antepreterit. 80 Gabriela Múcsková <?page no="81"?> 14 Analogically, according to the functional development of the form of the present of the verb byť in the forms of the Proto-Slavic perfect (cf. Múcsková 2016) it can be supposed that also the aorist forms of the verb byť at first fulfilled the function of a copula in the predicative syntactic construction with the l-participle (*spalъ bychъ - ‘I was the one who was asleep’). The copula was transformed to an auxiliary verb, and in later grammaticalized analytical forms of antepreterit, the aorist forms of the verb byť can be considered grammatical morphemes representing the category of person. This historical form expressed a past process or state that occurred before another past process, or the connection or parallelism with another past process (on the meaning of the Proto-Slavic antepreterit cf. Gebauer 2007: 551-552; Krajčovič 1988: 141-144; Kosek 2014: 261). The relationship of two past processes could have had various shades in meaning - from only a mere temporal sequentiality through a factual relatedness up to causal con‐ nection. That is why, in addition to its temporal meaning, this relationship also had the meaning of mutual conditionality. The interrelation between the temporal meaning and the conditional meaning became the basis for the later semantic and functional change from the temporal grammatical meaning to the meaning of Actionsart. After this change, the tense form bychъ spalъ (be PSL aorist + sleep l-form = ‘I was sleeping’ or ‘I had slept before…’) acquired the meaning of conditional (= ‘I would sleep’). The semantic trans‐ formation (reanalysis) of the originally temporal - antepreterit - meaning into the category of Actionsart was accompanied by a long and complicated development connected also with the development of the category of mo‐ dality and hypotactic syntactic means (cf. Kosek 2014: 267; 2017). After the aorist forms gradually disappeared from the system of verbal paradigms (since the 13 th and 14 th centuries), also the aorist forms of the verb byť in the function of a grammatical morpheme underwent changes - more exactly, the particular personal forms disappeared and only one petrified form by was preserved. Simultaneously, the analytical conditional form was extended by a new unit - a present form of the verb byť representing the category of person - spal by som (sleep l-form + by + be 1ps.sg = ‘I would sleep’). Within the process of the semantic transformation of the whole construction to the conditional meaning, also the time-related meaning of the element by bleached. The element by became an autonomous grammatical morpheme representing the conditional grammeme (cf. Kopečný et al. 1980: 114), i.e. a Grammatical and grammatographical categories - accord and conflict 81 <?page no="82"?> 15 In fiction or in dialects, by is used also in the function of a subordinate conjunction that is equivalent to the neutral compound conjunction aby ‘in order to’, although in this function the conjunction by is already perceived as obsolescent or poetical. Moreover, the element by is today part of compound conjunctions, conjunctional constructions, or particles: keby ‘if ’, ‘if only’, akoby ‘as if ’ etc. introducing clauses of purpose, conditional, optative, concessive, or comparative clauses which share the meaning of irreality, hypotheticality or the meaning of future validity (for more details cf. KSSJ; Ružička et al.: 715). 16 In the original: “neurčitý čas pomocného slovesa byť”. 17 In the original: “modifikuje modálnosť vety, výpovede na podmieňovací spôsob”. unit with the meaning of irreality or hypotheticality (cf. Sokolová/ Žigo 2014: 112; Múcsková 2021). 15 Hence, from the aspect of grammaticalization and functional grammar, by is an element that in contemporary conditional forms has only a gram‐ matical - morphological - character. It can be classified as an autonomous grammatical morpheme, namely a conditional grammatical formant. The problematic grammatographic categorization of this element in older as well as contemporary grammatical descriptions and lexicographical works stems from the various linguistic approaches in which a different aspect is stressed, namely: 1) The aspect of origin. - Concerning the origin, the element by is classified as the form of the verb byť ‘be’. Such classification can be found only in the oldest works of the first codifiers of standard Slovak. A. Bernolák classified it in this way in his dictionary (1825: 69-70) and Ľ. Štúr characterized the component by in the conditional construction as the “indefinite tense of the auxiliary verb byť” 16 (1846: 73-74). In the attribute “indefinite” we can also see the reflection of the fact that this form already lost its original grammatical meanings, as well as the other forms of the original paradigm. 2) The aspect of morphological status. - From the point of view of the formal morphological status, the element by is delimited as a grammatical morpheme. Such functionally conditioned delimitation is found above all in descriptions of morphology aimed at specialists. For instance, in the above mentioned Morfológia slovenského jazyka (‘Morphology of the Slovak Language’) by is treated as an autonomous grammatical morpheme (Ružička et al. 1966: 463) or as an autonomous conditional morpheme (ibid: 480; cf. also Sokolová et al. 1999: 46; Sokolová/ Žigo 2014: 58 and others). 3) The aspect of the category of modality. - The element by “modifies the modality of the sentence, or the utterance into the conditional mood” 17 (SSSJ 82 Gabriela Múcsková <?page no="83"?> 18 In the original: “pomocné slová, ktorými podávateľ nadväzuje na kontext alebo na situáciu a-pritom vyjadruje svoj vzťah k vecnému obsahu výpovede”. 19 In the original: “vyjadruje rozličné významové odtienky jednotlivých výrazov alebo výpovedí i-kontextových a aktualizačných vzťahov”. 20 In the original: “sa používajú na tvorenie rozličných tvarov a nových slov”. 21 E.g. Pravidlá slovenského pravopisu (‘Rules of Slovak Orthography’) (Považaj (ed.) 2013: 184), Príručka slovenského pravopisu pre školy (‘Manual of Slovak Orthography for Schools’) (Oravec/ Laca 1981: 189) or Príručka slovenského pravopisu pre školy a prax (‘Manual of Slovak Orthography for Schools and for Practical Usage’) (Ripka/ Imrichová/ Skladaná 2005: 130, 200). 2006). With regard to this aspect, the element by is labelled as a particle, i.e. an element of an autonomous word class. This word class started to be used in the Slovak grammatical works in the 1950s, (also) following Russian grammatography, and was named by the term “particle”, often accompanied by its Latin equivalent particulae. Until then, the term “particle” referred to all grammatically uninflected word classes, i.e. adverbs, prepositions, conjunctions, and interjections (cf. Jóna 1951-1952; Ďurovič 1955: 132-136; Ružička 1956: 9-10; Šimková 2003). Today this specific word class includes “auxiliary words by which the author connects the utterance to the context or situation and, at the same time, expresses his/ her relation to the factual content of the statement” 18 (Dudok 2003: 19), or “expresses various shades of meaning of the particular expressions or statements, as well as contex‐ tual and upgraded relationships” 19 (Ružička et al. 1966: 746). It is a quite heterogeneous group of units, and the appurtenance of several of them to this word class reflects the problems connected with the phenomenon of conversion. The roots of categorizing the element by among the particles are connected with the beginnings of establishing this word class at first in school textbooks, and later also in scholarly morphological works. One of the first articles about this new word class by E. Jóna proposes also a further division of particles into subgroups and includes the “particle” by in the subgroup of the so-called non-autonomous particles which “are used in certain forms and new words” 20 (Jóna 1951-1952: 236). Although this subgroup later became the object of critical discussions, the classification of by as a particle has spread and developed into a tradition. It got into textbooks for elementary and secondary schools, but also into university textbooks of morphology (Oravec/ Bajzíková/ Furdík 1984: 145). The element by is also named as a particle in codification and orthographic manuals aimed at the wider public 21 , as well as in user-oriented, codification, or even Grammatical and grammatographical categories - accord and conflict 83 <?page no="84"?> 22 In the original: “zakrytou polemikou proti učeniu o jazykových rovinách”. 23 In the original: “treba mať na zreteli, že jazykové roviny sú metodologickými konštruktmi; nemalo by ísť o deklaratívne popieranie rovinového prístupu, ale o racionálne zisťovanie a využívanie poznatkov získaných z analýzy takto vyčlenených výsekov jazykového systému v prospech celkového poznania jazyka ako zložitého systému verbálneho správania”. academic dictionaries of the Slovak language. Thus, of the three different grammatographic interpretations, the most commonly used is the one that did not take into account the development of language and functional approach to morphological classification. ※ Regarding the division of the language structure into language levels, a dis‐ cussion arose also in Slovak linguistics, whether this division is meaningful. Such division was - in contrast to the then-existing tradition - avoided by E. Pauliny in his work Slovenská gramatika (‘Slovak Grammar’, Pauliny 1981). As he said about this work, it is a “covert polemic against the teaching about linguistic levels” 22 (ibid.: 6; cf. also Furdík 2003: 44). On the margin of the discussion on the usefulness of defining linguistic levels, Juraj Furdík points out that “it is necessary to bear in mind that linguistic levels are meth‐ odological constructs; the level-focused approach should not be refused declaratively, but the approach should involve rational investigation and use of the findings obtained from the analysis of thus defined parts of the linguistic system for the sake of overall knowledge of the language as a complex system of verbal behaviour” 23 (ibid.: 45). - 2.3 Word classes Several “problematic” discussions touch upon defining word classes, and their further internal classification or identification of the boundaries and transitory zones between the particular categories of word classes. In this respect, some “transitory” units or their appurtenance to the traditionally defined word classes are considered problematic. In this case, the word-class origin, the primary form or function, and sometimes also a certain gramma‐ tographic tradition overlap. An often-discussed question is the conversion (word-class homonymy) of uninflected word classes (adverbs, prepositions, conjunctions, interjections, and particles); within the framework of this group of units, particles that 84 Gabriela Múcsková <?page no="85"?> 24 From the history of Slovak grammar, it is interesting to mention here that Ľudovít Štúr, the author of the historical codification of Standard Slovak, on which also the contemporary Standard Slovak is based, delimited participles and other non-finite verbal forms as an autonomous word class which he called “menoslovja” (Štúr 1864: 112-113, 130-131; cf. also Jóna 1951-52: 234). have already been mentioned above are especially problematic. Neither did the basic division of word classes into autosemantic and synsemantic re‐ main without discussion. The principal question was whether synsemantic units constitute an autonomous word class or are a word class of a different type (“superstructural” or “not independent”), or whether they are even only functional or grammatical units (cf. Šimková 2003; Ďurovič 1955; Ondrus 1971). In the group of autosemantic word classes - mainly in the lexicography - those units are denoted as “problematic”, which from the point of view of their origin, paradigmatic characteristics, or function in the utterance, can be classified into various word classes. These units have the character of so-called transitory units that - because of their hybrid char‐ acter - exceed the borders of linguistically delimited word-class categories. ※ In the following analysis, we focus on a group of such problematic units - namely the participles. They are words of verbal origin which in grammat‐ ical descriptions are traditionally perceived as part of verbal paradigms, though having also nominal, or, more exactly, adjectival declension. 24 To a differing degree, these units undergo the process of adjectivization associ‐ ated with the weakening of verbal (dynamic) meaning and with a gradual acquisition of adjectival (static) semantic as well as categorial qualities. There have already been published several works searching for formal structural, as well as semantic criteria for how to delimitate (the originally past) passive -n-/ -tparticiples (as members of the verbal paradigm) and lexicalized deverbal adjectives with the ending -ný, -tý (cf. Ružička et al. 1966: 556; Sejáková 1989, 1995). However, in concrete cases, there still appear uncertainties and in lexicographic works (within a verbal entry or a separate entry as a lexical adjective), these transitory units differ in the way of classification, whether in conceptual rules or in particular details (cf. Ružička 1956: 16). The uncertainty arises from the gradualness of the process of verb-to-adjective transformation, as well as the vagueness of the Grammatical and grammatographical categories - accord and conflict 85 <?page no="86"?> 25 Historically, these are past (or perfect) and active participles with a resultative meaning. They had two types of declension - nominal (short) and composite (long). The nominal forms in collocations with the present forms of the verb byť ‘be’ grammaticalized into the analytical form of the Proto-Slavic perfect that was later semantically transformed into a general preterit (Múcsková 2016). But this analysis is primarily focused on adjectives with the suffix -lý that have developed from the forms of the composite type of declension. 26 The development of the original l-participles in the process of their adjectivization is described in detail in Múcsková (2019); cf. also Horecký (1995) and Kyseľová (2012). boundaries between the degrees of semantic changes and the coexistence of verb and adjectival meanings in polysemic lexemes. In contrast to the passive participles, the original perfect active participles with the formant -l- (so-called l-participles 25 ) are not mentioned in contem‐ porary grammatical works as (etymologically) deverbative adjectives with the suffix -lý. Also, in lexicographical works, all these units are categorized as lexical adjectives in separate entries, although some of them have to a large extent preserved the participial character and the distinct dynamic feature, i.e. they have not fully undergone the process of adjectivization. That is why, this group includes fully adjectivized units as teplý ‘warm’, smelý ‘bold, daring’, svetlý ‘light, bright’, kyslý ‘sour’, which have a disguised verbal participial origin and meaning, as well as units with a preserved resultative meaning, like zrelý ‘ripe, mature’, dospelý ‘adult’, zastaralý ‘outdated, obso‐ lete’, or even units that have a preserved participial character, e.g. došlý ‘who/ what has come’ (e.g. mail, payment, news etc.), ušlý ‘what has been lost’ (e.g. salary, profit, etc.) and others. 26 There are several reasons why the originally active past participles and passive past participles in Slovak grammatical and lexicographic works are presented differently. Evidently, the main reason is that they constitute a smaller group of units, because they are not formed systemically and regularly from all verbs, and most of them have gone through a process of verb-to-adjective transformation. Another reason is that both groups of participles (in terms of their active and passive meaning) were gradually approximated and were understood as synonyms. In later development, the passive participles became more productive while many of the l-participial adjectives gradually ceased to be used and became obsolete. However, the decrease of the number of l-participles was also contributed to by the lan‐ guage policy that mainly in the 20 th century tried to eliminate the influence of Czech upon Slovak. Within this policy, numerous l-participles as well as 86 Gabriela Múcsková <?page no="87"?> lexicalized adjectives were labelled as Czech and therefore non-standard, and it was recommended to replace them with -n-/ -t-participles or by other equivalent forms (Múcsková 2019: 170-171; Nábělková 2014: 80-81). As stated above, in grammatical and lexicographic works, the term “participle” is not used even for those l-forms that have preserved their participial character, although in these cases this would be adequate. Par‐ adoxically, in textbooks, the term “l-participle” is used for labelling those verbal forms, that stem from nominal (short) forms (see footnote No. 26). These forms, however, were fully grammaticalized as part of analytical verbal forms of the preterit, e.g. - spal som (‘I slept’) and of the conditional - spal by som (‘I would sleep’) and the l-ending has here the status of a grammatical element. That is why they should not be referred to as “participles”, but as “l-forms of verbs”. As such, they are called only by some scholarly academic works (e.g. Ružička 1966: 474; Andersen 1987: 26) and university textbook (Oravec/ Bajzíková/ Furdík 1984: 145, 148). ※ - 2.4 Language development and changes in descriptions of its grammar The “problematic character” of some grammatical phenomena sometimes results from the natural development of language and its gradual diver‐ gence from the older grammatographic description of the model of gram‐ matical categories and their elements. For example, in recent prescriptive linguistics, it was (and in the materials for students still is) claimed that analytical forms of antepreterit or conditional preterit should be preserved. However, in current communication, these forms already recede and are spontaneously replaced by the forms of the preterit and present condi‐ tional, or they are perceived as archaic and stylistically marked. This gradual withdrawal is accompanied by variability in linguistic usage (cf. Sokolová/ Žigo 2014: 194-202; Múcsková 2016: 43), as well as numerous polemic discussions. In this way, the prescriptive or didactically oriented language descriptions preserve and petrify the older and at present no more existing state. In such cases, the grammatical description of language lags behind spontaneous linguistic development. However, we also know opposite cases when the description of linguistic structures does not correspond to the real state of language and its dy‐ namics. This concerns a large number of prescriptive decisions that have Grammatical and grammatographical categories - accord and conflict 87 <?page no="88"?> failed in finding their way into actual linguistic usage. Some of them still represent intentional linguistic efforts aimed at regulation within the corpus planning, while others have gradually been re-evaluated. These prescriptive failures are mostly the result of an overestimation of the structural aspect of grammatical forms and their systemic relationships, while the functional aspect and the tendencies of spontaneous development are taken into consideration insufficiently. Another example of this phenomenon is the question of the “presence or absence” of vocative in Slovak. During the course of language development, the special forms of the old Proto-Slavic vocative gradually changed and formally merged with the forms of the nominative. For this reason, this case began to be omitted in grammatical descriptions of declension paradigms or referred to as the “disappeared case” or a case “replaced” by the nominative (cf. Pauliny 1966). The claims about the “disappearance of the vocative” in Slovak are based only on the formal side of the nouns. They do not take into consideration the functional side of the words used in addressing, i.e. the fact that the communication need to address somebody has not disappeared. In addition, in spoken language, there are still present or newly formed special appellative and proprial forms in the function of addressing (Slančová/ Sokolová 1998). The need to address somebody, namely, motivates new forms with different endings in the function of vocative formants to emerge (Nábělková 2019: 628-632). Due to the specific syntactic features of addressing, special communication and pragmatic function, as well as new colloquial forms, several works indicate the need to preserve the vocative in grammatical descriptions of the Slovak declension system (cf. Karčová 2008; Múcsková 2011). 3 Conclusion In this paper, we highlighted some cases of linguistic phenomena and their linguistic categorization where the grammatographic descriptions differed or where there were variant interpretations. The possible reasons for the oc‐ currence of variability (and tensions between the structure of the language and its grammatical image) were also indicated. A distinctive role is played here by a particular linguistic tradition that remains relevant even after the re-evaluation of scholarly knowledge and the more detailed and exact classification of certain linguistic phenomena; however, under the influence 88 Gabriela Múcsková <?page no="89"?> 27 The paper was written within the VEGA 2/ 0133/ 20 project Slovenský pravopis a jeho pravidlá v kontexte súčasnej jazykovedy a jazykovej praxe (Slovak Orthography and its rules in the context of contemporary linguistics and language usage). Translated into English by Ada Böhmerová. of this tradition, as well as the models of foreign descriptive procedures (in Slovak this primarily concerns the influence of Latin, German and Czech grammar), the previously defined categorizations have not considerably changed; among other factors, this is due to their long-term use in education. The variant interpretations of grammatical phenomena reflect the exis‐ tence of “non-categorial” phenomena in language - i.e. phenomena that are difficult to classify within the usual and well-defined linguistic catego‐ ries. Nevertheless, the existence of such “problematic” or “controversial” elements is a natural counterpart of language and they are not realized by its users. In terms of its nature, language represents a categorial con‐ tinuum rather than a system of autonomous categories. This is why elements specified in various scholarly analyses as “problematic” or “controversial” seem to be seen as such merely on the basis of grammatical categories and levels and lexicographical principles constructed by linguists. They function naturally in real language, and they do not present any language or communication problems for language users. In highlighting variable or controversial categorizations and terminolog‐ ical labelling, we do not intend to evaluate the grammatical descriptions of selected language units as “correct” or “incorrect”, nor do we wish to initiate changes in existing or established grammatical descriptions. Our aim is to point out that the ways of linguistically reflecting grammatical phenomena, features of language and disagreements between the properties of language units and their descriptions are an integral part of the context in which language and its linguistics function and develop. 27 Bibliography Andersen, Henning (1987): “From auxiliary to desinence”. In: Historical development of auxiliaries, Martin Haris/ Paolo Ramat (eds.). Berlin: Mouton de Gruyter, 21-51. Bernolák, Anton (1825): Slowár slowenskí, Češko-Laťinsko-Ňemecko-Uherskí. I. Budae: Typogr. Reg. Univers. Hungarie. Grammatical and grammatographical categories - accord and conflict 89 <?page no="90"?> Bosák, Ján/ Buzássyová, Klára (1985): Východiská morfémovej analýzy (Morfematika. Slovotvorba). Bratislava: Veda. Buzássyová, Klára/ Jarošová, Alexandra (eds.) (2006): Slovník súčasného slovenského jazyka. A-G. Bratislava: Veda. Dolník, Juraj (2003): “O novej slovenskej morfológii”. In: Tradícia a perspektívy gra‐ matického výskumu na Slovensku, Mária Šimková (ed.). Bratislava: Veda, 32-37. <https: / / www.juls.savba.sk/ ediela/ tpgvs2003/ tpgvs.pdf> (Last access 31.10.2021) Dolník, Juraj (2005): “Koncepcia novej morfológie spisovnej slovenčiny”. Slovenská reč 70(4): 193-210. Dolník, Juraj (ed.) (2010): Morfologické aspekty súčasnej slovenčiny. Bratislava: Veda. Dudok, Miroslav (2003): “Jazykové domény, gramatika diskurzu a diskurz grama‐ tiky”. In: Tradícia a perspektívy gramatického výskumu na Slovensku, Mária Šim‐ ková (ed.). Bratislava: Veda, 19-23. <https: / / www.juls.savba.sk/ ediela/ tpgvs2003/ tpgvs.pdf> (Last access 31.10.2021) Ďurčo, Peter/ Majchráková, Daniela et al. ( 2 2017): Slovník slovných spojení. Podstatné mená. Bratislava: Veda. Ďurovič, Ľubomír (1955): “K otázke neohybných čiastok reči v slovenčine”. Jazyko‐ vedný sborník SAVU 4: 113-140. Dvonč, Ladislav (1984): Dynamika slovenskej morfológie. Bratislava: Veda. Furdík, Juraj (2003): “O existujúcich a možných morfológiách”. In: Tradícia a per‐ spektívy gramatického výskumu na Slovensku, Mária Šimková (ed.). Bratislava: Veda, 42-48. <https: / / www.juls.savba.sk/ ediela/ tpgvs2003/ tpgvs.pdf> (Last access 31.10.2021) Furdík, Juraj (2004): Slovenská slovotvorba (teória, opis, cvičenia), Martin Ološtiak (ed.). Prešov: Náuka. Garabík, Radovan et al. (2016): Skloňovanie podstatných mien v slovenčine s korpuso‐ vými príkladmi. Bratislava: Vydavateľstvo Mikula. Gebauer, Jan ( 2 2007): Historická mluvnice jazyka českého. Díl IV. Skladba, František Trávníček (ed.). Praha: Academia. Hašanová, Jana (2012): “Spracovanie zvratných slovies v Slovníku súčas‐ ného slovenského jazyka so zameraním na komponent sa”. In: Slovo v slovníku. Aspekty lexikálnej sémantiky - gramatika - štylistika (prag‐ matika), Klára Buzássyová/ Bronislava Chocholová/ Nicol Janočková (eds.). Bratislava: Veda, 184-191. <https: / / data.juls.savba.sk/ julswww/ pub_slovo_v_slov‐ niku? action=AttachFile&do=view&target=slovo_v_slovniku.pdf> (Last access 31.10.2021) 90 Gabriela Múcsková <?page no="91"?> Horecký, Ján (1995): “Prídavné mená na -lý”. Kultúra slova 29(6): 338-342. <https: / / www.juls.savba.sk/ ediela/ ks/ 1995/ 6/ ks1995-6.html#pridavne-mena-na-ly> (Last access 31.10.2021) Horecký, Ján (2003): “Kategoriálna vybavenosť slovenskej morfológie”. In: Tradícia a perspektívy gramatického výskumu na Slovensku, Mária Šimková (ed.). Bratislava: Veda, 38-41. <https: / / www.juls.savba.sk/ ediela/ tpgvs2003/ tpgvs.pdf> (Last access 31.10.2021) Ivanová, Martina (2006): Valencia statických slovies (Acta Facultatis Philosophicae Universitatis Prešoviensis; Monographia 67). Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Ivanová, Martina (ed.) (2009): Aspektuálnosť a modálnosť v slovenčine. Prešov: Filozo‐ fická fakulta Prešovskej univerzity v Prešove. <https: / / www.pulib.sk/ web/ kniznica/ elpub/ dokument/ Ivanova1> (Last access 31.10.2021) Ivanová, Martina (2017): Modálnosť a modálne verbá v slovenčine (Opera Linguistica 15/ 2017). Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. <https: / / www.pulib.sk/ web/ pdf/ web/ viewer.html? file=/ web/ kniznica/ elpub/ doku‐ ment/ Ivanova3/ subor/ 9788055524818.pdf> (Last access 31.10.2021) Ivanová, Martina/ Kyseľová, Miroslava (2012): “Niekoľ ko poznámok o statuse a vymed‐ zení kategoriálnych verb”. In: Slovo v slovníku. Aspekty lexikálnej sémantiky - gra‐ matika - štylistika (pragmatika), Klára Buzássyová/ Bronislava Chocholová/ Nicol Janočková (eds.). Bratislava: Veda, 159-170. <https: / / data.juls.savba.sk/ julswww/ pub_slovo_v_slovniku? action=AttachFile&do=view&target=slovo_v_slovniku.pdf> (Last access 31.10.2021) Ivanová, Martina/ Kyseľová, Miroslava (2019): “Reflexivita ako koncept deiktického odkazovania na subjekt udalosti”. In: Personálna a sociálna deixa v slovenčine, Jana Kesselová (ed.). Prešov: Filozofická fakulta Prešovskej univerzity v Prešove, 113-152. Ivanová, Martina/ Sokolová, Miloslava/ Kyseľová, Miroslava/ Perovská, Veronika (2014): Valenčný slovník slovenských slovies na korpusovom základe. Prešov: Fi‐ lozofická fakulta Prešovskej univerzity v Prešove. <https: / / www.ff.unipo.sk/ sloval/ Slovnik/ slovnik.html> (Last access 31.10.2021) Ivanová, Martina/ Sokolová, Miloslava/ Kyseľová, Miroslava/ Eddy, Eva (2017): Vybrané aspekty valencie verb v slovenčine (Opera Linguistica 16/ 2017). Prešov: Filozofická fakulta Prešovskej univerzity v Pre‐ šove. <https: / / www.pulib.sk/ web/ pdf/ web/ viewer.html? file=/ web/ kniznica/ elpub/ dokument/ Ivanova2/ subor/ 978055519074.pdf> (Last access 31.10.2021) Janočková, Nicol (2014): Sloveso a zmysly. Slovotvorba a vidotvorba. Bratislava: Veda. Grammatical and grammatographical categories - accord and conflict 91 <?page no="92"?> Jarošová, Alexandra (2003): “Umývať sa: predmetová syntagma alebo zvratné slo‐ veso? ”. In: Tradícia a perspektívy gramatického výskumu na Slovensku, Mária Šim‐ ková (ed.). Bratislava: Veda, 204-205. <https: / / www.juls.savba.sk/ ediela/ tpgvs2003/ tpgvs.pdf> (Last access 31.10.2021) Jarošová, Alexandra (2013a): “Problém rozsahu vidovej korelácie: zisťovanie mož‐ ností za hranicami imperfektivizácie a mutácie”. Slovo a slovesnost 74(4): 243-268. Jarošová, Alexandra (2013b): “Vidová opozícia a vidová korelácia v sloven‐ čine”. Jazykovedný časopis 64(1): 5-47. <https: / / www.juls.savba.sk/ ediela/ jc/ 2013/ 1/ JC_2013_1.pdf> (Last access 31.10.2021) Jóna, Eugen (1951-1952): “Zo školskej gramatiky. Častice ako osobitný slovný druh”. Slovenská reč 17(7-8): 233-236. <https: / / www.juls.savba.sk/ ediela/ sr/ 1951-52/ 7-8/ sr1951-52-7-8-lq.pdf> (Last access 31.10.2021) Kačala, Ján (2006): Systém jazykových kategórií. Bratislava: Univerzita Komenského v Bratislave. Kačala, Ján/ Pisárčiková, Mária/ Považaj, Matej (eds.) (2003): Krátky slovník sloven‐ ského jazyka. Bratislava: Veda. Karčová, Agáta (2008): “Príprava a uskutočňovanie projektu morfologického analyzá‐ tora”. In: Varia. 15. Zborník materiálov z XV. kolokvia mladých jazykovedcov, Anna Gálisová/ Alexandra Chomová (eds.). Bratislava: Slovenská jazykovedná spoločnosť pri SAV/ Banská Bystrica: Katedra slovenského jazyka a literatúry FHV UMB v Banskej Bystrici, 286-292. <https: / / www.juls.savba.sk/ ediela/ varia/ 15/ Varia15.pdf> (Last access 31.10.2021) Kesselová, Jana (2001): Lingvistické štúdie o komunikácii detí. Prešov: Náuka. Kesselová et al. (2013): Spojky a spájacie prostriedky v slovenčine (synchrónia - dia‐ chrónia - ontogenéza). Bratislava: Veda. Kesselová, Jana (2014): Funkčno-sémantická kategória relačnosť v ranej ontogenéze reči dieťaťa hovoriaceho po slovensky. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. <https: / / www.pulib.sk/ web/ kniznica/ elpub/ dokument/ Kesse‐ lova5> (Last access 31.10.2021) Kopečný, František et al. (1980): Etymologický slovník slovanských jazyků. Slova gra‐ matická a zájmena. Svazek 2. Spojky, částice, zájmena a zájmenná adverbia. Praha: Academia. Kosek, Pavel (2014): Historická mluvnice češtiny I. Brno: Masarykova univerzita. Kosek, Pavel (2017): “Kondicionál ve starší češtině”. In: CzechEncy - Nový encykloped‐ ický slovník češtiny, Petr Karlík/ Marek Nekula/ Jana Pleskalová (eds.). Brno: Masar‐ ykova univerzita v Brně. <https: / / www.czechency.org/ slovnik/ KONDICIONÁL VE STARŠÍ ČEŠTINĚ> (Last access 31.10.2021). 92 Gabriela Múcsková <?page no="93"?> Krajčovič, Rudolf (1988): Vývin slovenského jazyka a dialektológia. Bratislava: Slov‐ enské pedagogické nakladateľstvo. Kyseľová, Miroslava (2012): “Verbálne particípiá a adjektíva tvorené koncovkou -lý v slovenčine, češtine a poľštine (vývoj a súčasné tendencie)”. In: Prolínání slovanských prostředí, Marek Příhoda/ Kateřina Kedron/ Marcel Černý (eds.). Praha: Filozofická fakulta Univerzity Karlovy v Praze/ Červený Kostelec: Pavel Mervart, 259-267. Kyseľová, Miroslava (2017): Medzi konaním a dianím. Sémantické aspekty verba a valencia. Filozofická fakulta Prešovskej univerzity v Prešove. Kyseľová, Miroslava/ Ivanová, Martina (2013): Sloveso vo svetle kognitívnej gramatiky. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Majchráková, Daniela/ Chlpíková, Katarína/ Bobeková, Kristína (2017): Slovník kolo‐ kácií prídavných mien v slovenčine. Bratislava: Veda. Múcsková, Gabriela (2011): “Zánik ako typ jazykovej zmeny”. In: Jazyk a komunikácia v súvislostiach 3, Oľga Orgoňová (ed.). Bratislava: Univerzita Komenského, 277-286. Múcsková, Gabriela (2016): Jazykové zmeny v historickom vývine gramatických tvarov z aspektu gramatikalizácie (na príklade vývinu slovenského préterita). Bratislava: Univerzita Komenského. Múcsková, Gabriela (2019): “Lexicalization after grammaticalization in the develop‐ ment of Slovak adjectives ending in -lý originating from l-participles”. SKASE Journal of Theoretical Linguistics [online] 16(3): 156-177. <http: / / www.skase.sk/ Volumes/ JTL41/ pdf_doc/ 09.pdf. ISSN 1336-782X> (Last access 31.10.2021) Nábělková, Mira (1993): Vzťahové adjektíva v slovenčine: funkčno-sémantická analýza desubstantívnych derivátov. Bratislava: Veda. Nábělková, Mira (2014): “The case of Czech-Slovak language contact and con‐ tact-induced phenomena”. In: Congruence in contact-induced language change. Language families, typological resemblance, and perceived similarity, Juliane Bes‐ ters-Dilger/ Cynthia Dermarkar/ Stefan Pfänder/ Achim Rabus (eds.). Berlin: Walter de Gruyter, 61-92. Nábělková, Mira (2019): “Prihováram sa ti, Janko môj, ku sviatku…” K osloveniu v slovenčine (a ešte čosi navyše)”. Jazykovedný časopis 70(3): 627-651. <https: / / www.juls.savba.sk/ ediela/ jc/ 2019/ 3/ jc19-03.pdf> (Last access 31.10.2021) Nižníková, Jolana/ Sokolová, Miloslava (1998): Valenčný slovník slovenských slovies. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Ološtiak, Martin (ed.) (2015): Kvalitatívne a kvantitatívne aspekty tvorenia slov v slovenčine. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Grammatical and grammatographical categories - accord and conflict 93 <?page no="94"?> Ološtiak, Martin/ Ološtiaková, Lucia (2018): Morfematika a slovotvorba slovenčiny. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Ološtiak, Martin et al. (2021): Slovník slovotvorných prostriedkov v slovenčine. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Ondrus, Pavel: “O slovách, ktoré netvoria systém slovných druhov”. Slovenská reč 36(1): 28-32. <https: / / www.juls.savba.sk/ ediela/ sr/ 1971/ 1/ sr1971-1-lq.pdf> (Last ac‐ cess 31.10.2021) Oravec, Ján (1956): “Pracovná konferencia o otázkach slovenskej gramatiky a slovníka”. Slovenská reč 21(1-2): 109-113. <https: / / www.juls.savba.sk/ ediela/ sr/ 1956/ 1-2/ sr1956-1-2-lq.pdf> (Last access 31.10.2021) Oravec, Ján/ Laca, Vincent (1981): Príručka slovenského pravopisu pre školy. Bratislava: Slovenské pedagogické nakladateľstvo. Oravec, Ján/ Bajzíková, Eugénia/ Furdík, Juraj (1984): Súčasný slovenský spisovný jazyk. Morfológia. Bratislava: Slovenské pedagogické nakladateľstvo. Pauliny, Eugen (1966): “Zánik vokatívu v slovenčine”. Studia Slavica Academiae Hun‐ garicae 12: 321-323. Pauliny, Eugen (1981): Slovenská gramatika (Opis jazykového systému). Bratislava: Slovenské pedagogické nakladateľstvo. Považaj, Matej (ed.) ( 4 2013): Pravidlá slovenského pravopisu. Bratislava: Veda. Ripka, Ivor/ Imrichová, Mária/ Skladaná, Jana (2005): Príručka slovenského pravopisu pre školy a prax. Bratislava: Agentúra Cesty. Ružička, Jozef (1956): “Sporné otázky slovenskej morfológie”. Slovenská reč 21(1-2): 3-19. <https: / / www.juls.savba.sk/ ediela/ sr/ 1956/ 1-2/ sr1956-1-2-lq.pdf> (Last access 31.10.2021) Ružička, Jozef et al. (1966): Morfológia slovenského jazyka. Bratislava: Vydavateľstvo SAV. Sejáková, Jana (1989): “Adjektivizácia n/ t-príčastí”. In: Horecký, Ján/ Buzássyová, Klára/ Bosák, Ján et al. (1989): Dynamika slovnej zásoba súčasnej slovenčiny. Brati‐ slava: Veda, 200-211. <https: / / www.juls.savba.sk/ ediela/ dynamika/ dszss.pdf> (Last access 31.10.2021) Sejáková, Jana (1995): Adjektivizácia n-/ t-ových príčastí v súčasnej slovenčine. Brati‐ slava: Jazykovedný ústav Ľ. Štúra SAV. (Doctoral dissertation.) Slančová, Daniela (ed.) (2018): Desať štúdií o detskej reči. Lexika - gramatika - prag‐ matika (Opera Linguistica 18/ 2018). Bratislava: Veda/ Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. 94 Gabriela Múcsková <?page no="95"?> Slančová, Daniela/ Sokolová, Miloslava (1998): “Jazykové prostriedky s vokatívnym exponentom v súčasnej slovenčine”. Slovenská reč 63(4): 210-220. <https: / / www.juls.savba.sk/ ediela/ sr/ 1998/ 4/ sr1998_4.pdf> (Last access 31.10.2021) Sokolová, Miloslava (1995): Kapitolky zo slovenskej morfológie. Prešov: Slovacontact. Sokolová, Miloslava (2007): Nový deklinačný systém slovenských substantív. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Sokolová, Miloslava (2009): “Sémantika slovesa a aspektové formy”. In: Aspektuál‐ nosť a modálnosť v slovenčine, Martina Ivanová (ed.). Prešov: Filozofická fakulta Prešovskej univerzity v Prešove, 22-37. <https: / / www.pulib.sk/ web/ kniznica/ elpub/ dokument/ Ivanova1> (Last access 31.10.2021) Sokolová, Miloslava/ Moško, Gustáv/ Šimon, František/ Benko, Vladimír (1999): Morfe‐ matický slovník slovenčiny. Prešov: Náuka. Sokolová, Miloslava/ Ivanová, Martina (eds.) (2006): Sondy do morfosyntaktického vý‐ skumu slovenčiny na korpusovom materiáli. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Sokolová, Miloslava/ Ivanová, Martina/ Ološtiak, Martin (eds.) (2006): Morfematický výskum slovenčiny (možnosti jeho štatistického, elektronického a didaktického spra‐ covania. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Sokolová, Miloslava et al. ( 3 2012): Slovník koreňových morfém slovenčiny. Prešov: Filozofická fakulta Prešovskej univerzity v Prešove. Sokolová, Miloslava/ Žigo, Pavol (2014): Verbálne kategórie aspekt a tempus v sloven‐ čine. Bratislava: Veda. Šimková, Mária (2003): “Tzv. gramatické slová v morfológii a v syntaxi”. In: Tradícia a perspektívy gramatického výskumu na Slovensku, Mária Šimková (ed.). Bratislava: Veda, 233-239. <https: / / www.juls.savba.sk/ ediela/ tpgvs2003/ tpgvs.pdf> (Last access 31.10.2021) Šimková, Mária (ed.) (2003): Tradícia a perspektívy gramatického výskumu na Slov‐ ensku. Bratislava: Veda. Štúr, Ludevít (1846): Nauka reči slovenskej. Prešporok: Tatrín. Zumrík, Miroslav/ Majchráková, Daniela/ Miháliková Lucia (2022): Časovanie slovies v slovenčine. Bratislava: Vydavateľstvo Mikula. Grammatical and grammatographical categories - accord and conflict 95 <?page no="97"?> Classicism, Czech language and Jungmann Ondřej Šefčík Abstract: Classism became a much discussed theme within the Czech National Revival movement. Jungmann’s paper on Classicism was hence not written within the purely academic discourse but as a resolute stand against Nejedlý, who critically rejected the lexical and orthographic innovations brought into the Czech language and litera‐ ture by Jungmann’s circle. We will focus first on the Jungmann’s paper and the orthographic argument between Jungmann and Nejedlý and next the role which the card of Classicism was played within it. Keywords: Czech national revival; Czech orthography; classicism; Czech orthographic reforms; Josef Jungmann 0 Classicism in the Czech national revival of the first half of 19 th century The Czech National Revival serves as one of the best examples of the 19 th century national revivals, especially since its strategies were often copied by later and similar Central European movements. The term Classicism became one of the tools used in the internal debate between Czech Revivalists in the first decades of the 19 th century. We will try to analyse the background, the argument of Classicism and its outcomes on the following lines. Note: The below quoted authors of the 19 th century usually used the contemporary orthography, often (as the theme of the paper is) different from our current use. An everyday praxis of the era was to use the Germanized form of the name and surname if a given publication was in German (Johann Negedly instead Jan Nejedlý). For simplicity, we will use the modern versions both of names and surnames through the whole paper. Jungmann’s personal notes (Zápisky; Jungmann 1973) will be quoted <?page no="98"?> 1 This study has three, slightly different versions: the original manuscript version; the published version of 1827 in Časopis společnosti wlastenského Museum w Čechách, which was partially rewritten by Palacký, the editor of the journal; the version from the year 1841, which is also the final version by Jungmann for his collected writings, which, surprisingly, is based not on the original manuscript, but on the edited version of Palacký. For the use of this paper, the first published version, that edited by Palacký, was used, since this is the version which had the impact on the public. For editorial details, cf. Zelený (1873: 401‒402), Novák (1909: 121‒123), Vodička (1948: 20‒21). Various editions differ even with titles, not speaking about content. from the modern edition; otherwise, the original publications will be quoted. The titles of the original studies are quoted with the actually used orthography of the period both for Czech and German. 1 Jungmann, Nejedlý and Classicism Josef Jungmann (1773‒1874) was the leading figure of the Czech National Revival, who not only became the first of the informal fathers of the nation, being a central figure of the widely spread network of national revivalists (buditelé). Beside this, as a philologist and the pupil of Josef Dobrovský (Joseph Dobrowsky, 1753‒1829) he also re-formulated both literary and linguistic bohemistics. Both his functions were tightly interwoven: the posi‐ tion was given by philology, which also gave Jungmann the authority and esteemed value of the national leader, however purely informal, which gave him the informal power to decide even philological arguments. One of the debates Jungmann took part in was that of Classicism, to which Jungmann devoted the paper O klassičnosti v literatuře wůbec a zwláště české (‘On the Classicism in the Literature in General and Czech in Particular’) published in 1827 1 , which is not only a purely theoretical treatise, but also a hidden polemic on Jan Nejedlý, though Nejedlý is not mentioned in the whole treatise at all (cf. Hattala 1877: 186). For the context, we have to emphasize that Jan Nejedlý (also Johann Negedly, 1776‒1834) was the second professor (since 1801) of the Czech language and literature at Prague University and a pupil of the first professor there, František Martin Pelcl (Franz Pelzel, 1734‒1801). Apart from his wide scientific (especially grammatographic) work, it was Nejedlý who first translated Homer (the first book of Iliad) in Czech (published in 1802). Hence it was Nejedlý who introduced the classical piece of poetry into Czech, 98 Ondřej Šefčík <?page no="99"?> 2 Nejedlý also published another piece with the Classical motive, namely Numa Pom‐ pilius, druhý král Řjmský (1808), written by Jean-Pierre Claris de Florian (original title: Numa Pompilius, second roi de Rome, Paris 1786) and translated by Nejedlý himself. including his personal principles of the Czech hexameter, even printing it using the Antiqua script (blackletter Fraktur was the common script of the Czech print of the period), Antiqua being connected with the classical literature since Humanism as well. 2 Nejedlý was an opponent of language in‐ novations, including neologisms, borrowings (from other Slavic languages) and the orthographic reform. He boasts in his 1829 translation of Salomon Geßner’s Idyllen that the translation is written in Classical Czech, pure and free and unstained by foreign and clumsy words (Sak 2007: 61), which was directly aimed at Jungmann’s translation and adaptation program of the lexical enrichment of the Czech vocabulary (realized fully with his Slownjk česko-německý [‘Czech-German Dictionary’], 1835‒1839). The growing mutual animosity between Jungmann and Nejedlý does not need to be documented indirectly by remarks and hints in their papers; it is also documented in memories of their contemporaries and private letters to friends and colleagues. Jungmann even dedicates one of his private remarks to Nejedlý, written originally in 1846 (Jungmann 1973: 14‒16). However, the contacts between Jungmann and Nejedlý were neither very frequent nor friendly even before the orthographic argument discussed below (Sak 2007: 34). However, in his younger years Jungmann occasionally published his translations in Nejedlý’s journal Hlasatel český (‘Czech herald’) (especially poems by Alexander Pope and Thomas Gray). The treatise on Classicism written by Jungmann is not an academic piece of work, but a fierce polemic on Nejedlý and a part of the broader dispute between two approaches to the state of the art of the Czech language and literature in the first half of the 19 th century. The integral and the most debated part of this dispute in public was the question of orthographic reform, which was also subjected to the question of what the Classicism means and should mean. We have to add that extensive parts of the treatise (roughly the middle section) are just an adaptation of text published originally in Das Gesammt‐ gebiet der teutschen Sprache, nach Prosa, Dichtkunst und Berendsamkeit the‐ oretisch und practisch dargestellt. Erster Band. Philosofie der Sprache by Karl Heinrich Ludwig Pölitz. Jungmann namely quotes from the introduction to Classicism, Czech language and Jungmann 99 <?page no="100"?> 3 Surprisingly, Sak (2007: 126‒126), though quoting Jungmannʼs text directly, does not report the Pölitz's influence on Jungmann at all, though already Jungmann himself expressed it directly within the text. 4 Just within the meaning of the word cultura (animi) as it was used by Cicero in his Tusculanae Disputationes, i.e., as the ‘agricultural’ metaphor derived from Latin colō, colere, coluī, cultus ‘to till, tend, care for, cultivate’. 5 Jungmannʼs original text, as we know it from later published version (cf. Vodičkaʼs edition of 1948: 103) contains a pun on his unnamed opponent (i.e., on Nejedlý), who, according to Jungmann, takes only the first meaning too tightly. Within this first section, Jungmann smears those whose either consider themselves to be living classics or reject any newer enterprise in the field of literature, considering the canon to be closed forever. the first part on the philosophy of language (Pölitz 1825: 1‒34). It is Pölitz who initially shifts the focus from the traditional preservative Classicism towards the cultured living language use. Jungmann usually accepts the general statements by Pölitz but omits the parts of the text focused on the German milieu and other segments useless for his argumentation (cf. Novák 1910: 116‒121). It has to be emphasized that Jungmann quotes Pölitz directly within his text. 3 2 Jungmann and his approach to Classicism Jungmann’s paper has three parts: the first one is an introductory and fully independent work by Jungmann, the middle part heavily depends on Pölitz, the third one is Jungmann’s application of Pölitz’s approaches on the Czech language and literature, in parallel to Pölitz’s application of his thesis on German. It has to be noted that Pölitz’s book was published 1825 and Jungmann’s paper in 1827 (and it had to come through the censor office before), hence Jungmann reflects what he considers the most up-to date state of the debate. Methodically speaking, Jungmann shares with Pölitz (1825: 13‒14) the general semantic shift since the term classical meant cultural  4 both to Pölitz and to Jungmann (cf. 1827: 29‒30). In the first part of the paper, which we can describe using the Latin term introductio, Jungmann adds the etymological observation on the world classical in Latin and its three meanings: 1. the social meaning of the (first social) class as used within the Servian reform; 2. the military meaning of the army, classicum canere; 3. the naval meaning of the navy. 5 100 Ondřej Šefčík <?page no="101"?> In the middle part, which is usually just a direct translation of Pölitz (roughly pages 13‒29 of the original text). Jungmann only omits the sen‐ tences of the original focused too tightly on the German language situations (usually sections on the German situation, on the inability of the French language to express German Philosophy etc.), occasionally adds few lines; especially noteworthy are the characteristics of Greek and Roman writers (cf. Novák 1910: 119‒120). However, the extensive excerpts from Pölitz can be considered a direct transfer (or even better: translatio) of the German Revivalism into Czech sui generis. The following arguments by Pölitz were quoted in the second part: i. that the Classical age is not a result of a work by a single author or a few authors, but by a number of contemporary, educated, original and excellent authors (Jungmann 1827: 177; cf. Pölitz 1825: 13‒14); ii. that only such a classical period sanctifies (Jungmann writes directly: poswěcuge) the national literature and builds foundations for future spiritual and ethical education and richness of language (Jungmann 1827: 177‒178; cf. Pölitz 1825: 14‒15); iii. the claim that the more liberal the government is, the more easily the human spirit and language rise, fuelling the classicism of the given language (Jungmann 1827: 178; cf. Pölitz 1825: 17‒18); iv. emphasis on the vital relationship between the state of religion, science, philosophy, general ethics, poetry and historiography on one side and language on the other side (Jungmann 1827: 178‒179; cf. Pölitz 1825: 18‒19); v. the truly classical language is such language, which can express any idea, to serve for expression of any rational, esthetical and ethical theme (Jungmann 1827: 180; cf. Pölitz 1825: 23‒25); vi. The Golden age is equal to the era in which the given nation has reached the highest physical, intellectual, esthetical and ethical stage of its development. The answer as to which era of the history of the given language is the golden age should be decided by future historians (Jungmann 1827: 181‒182; cf. Pölitz 1825: 27‒29); The third part is, however, fully independent of Pölitz; here Jungmann applies the above-mentioned thesis to the history of the Czech language and literature (here we can use the term adaptatio). Jungmann distinguishes three eras (Jungmann 1827: 35‒39): Classicism, Czech language and Jungmann 101 <?page no="102"?> 6 Surprisingly for modern reader, Jungmann equals the Golden Era with the rule of the Emperor Rudolph II., not with the whole Humanist Era, however, the same equation was the contemporary standard. i. the Oldest Era, corresponding to the modern term of Old Czech language and literature; ii. the Middle Era, corresponding to the traditional Golden Age of his contemporaries, the modern term of Middle Czech. Jungmann states that the golden label of Rudolphine Era 6 does not mean anything else that the number of literary works reached its peak. Jungmann rejects the idea that this peak also means the end for any new production or the ban on any changes and reforms of the language, especially since the opposite point of view would necessarily lead towards the end of the use of the Czech language as the tool of the education and literacy and the Czech nation could then join the momentarily superior neighbour, i.e. Germans (Jungmann 1827: 18); iii. Finally the new era of literacy, which is a σκάνδαλον for many, as Jungmann writes (1827: 38). For Jungmann, it is only the current literary use of the language which causes the full-scale use of the language in all contexts as mentioned earlier and even will bring the new Golden Age in the near future. It is a typical Jungmannian statement of the literature, which can grow not only through originals but also through translated works. The last paragraph states that a nation without the classical works (again in the sense as mentioned earlier) is ingloriously lost and vanishes in time but even a dead language with classical works, though it has perished, survives in such works, which will be forever named after such a language (cf. Jungmann 1827: 39). 3 Czech orthography within Jungmannian Classicism The orthographic disputes of the first half of the 19 th century were already repeatedly examined in details we cannot fully repeat here. From the rele‐ vant literature we have to refer to previous studies: Zelený (1873; especially pp. 176‒199); Hattala (1877; especially pp. 183‒201); Chalupný (1909: 59‒61); 102 Ondřej Šefčík <?page no="103"?> 7 Though not really the “university position”, still the position of the authority in the field. Vondrák (1926; especially pp. 52‒53); Tešnar (2000, 2001, 2003); Sak (2007: 140‒149); Pleskalová/ Šefčík (2007: 516‒519). Czech orthography of the beginning of the 19 th century was based on what is known as the bratrský pravopis [the orthography of the Brethren, i.e., Unitas Fratrorum, Unity of the Brethren, Moravian Church], which, was, contrary to the confessional name, the generally used standard both by Catholic (the prevailing and the religion) and Reformed (a religious minority formed especially by descendants of the Brethren) within the Czech lands and even beyond (but not in German-speaking areas), especially in Prussian Silesia (used by s.-c. Moravci, i.e., by Catholic Moravian-dialect speaking inhabitants of the Upper Silesia) and Kingdom of Hungary (here used by Slovaks, even outside of today’s Slovakia). This was mostly due to the work of the Catholic grammatographer Matěj Václav Štajer (also written Šteyer, 1630‒1692), who transferred the standard of the Brethren orthography into the Catholic milieu, though little changed (Koupil 2015: 155‒156). This standard arose from two theoretical works on the Czech orthography: the anonymous treatise Orthographia bohemica, usually attributed to religious reformer John Hus and to the year 1412 (cf. the critical and commented edition by Voleková/ Dragoun 2019: xxxv‒xxxvii), which introduced diacritic orthography, and the Grammatyka Cžeská w dwogij stránce wywedená by Beneš Optát, Petr Gzel and Václav Philomathes (1533; the part on Czech orthography was written by Optát and Gzel; the latest edition is Optát/ Gzel/ Philomates/ Koupil 2019). The orthographic standard of the Brethren was based on the (lightly modified) standard of Grammatyka Cžeská and this standard was hence roughly three centuries old in twenties of 19 th century. From the point of view of Nejedlý and his followers, Jiří Palkovič (Georg Palkovich; 1769‒1850), who was the professor of the Czech language on the Reformed Lyceum in Bratislava since 1803 7 was especially noteworthy; the Brethren orthography was entirely satisfying, proved through ages and within the best tradition of the Golden Age. Surprisingly, this was not the idea of Jungmann and his followers (here we should name Václav Hanka), who tried to reform the Czech orthography. The idea of the orthographic reform of the writing of y/ i after c/ s/ z was taken from the Josef Dobrovský’s grammar of Czech, printed in 1809 Classicism, Czech language and Jungmann 103 <?page no="104"?> 8 This orthographic manual was subsequently published, later with modified titles also in 1821, 1833, 1835, 1839, 1844, 1845, 1848, 1849. Later versions also included further reforms of the orthography. 9 Lisický (1898: 310) correctly states that, besides the analogy itself, Dobrovský was also motivated by the Slavic etymology. (the second edition, remarkably reworked, was printed in 1819), where he wrote: nach z, s, c schreibt man nur ein y, nie i; doch sollte nach der Analogie in manchen Fällen auch ein i geschrieben werden ‘after z, s, c only y, never i is written, though i should be written in many cases, according to the analogy’ (Dobrowsky 1809: xvii, 8; Dobrowsky 1819: xvi‒xviii, 8). It is notable that this was not a real proposal of orthographic reform, but just the scientific observation of the well-informed Slavist, who personally used the traditional orthography of the Brethren for the rest of his life. The reform itself was introduced by Václav Hanka (1791‒1861) in his treatise Prawopis český podlé základu Gramatyky Dobrowského od geho žáka Wáclawa Hanky (‘The Czech Orthography According to the Grammar by Dobrovský Written by his Pupil Václav Hanka’) (1817 8 : 15, 67‒68). This reform was termed the analogical reform, according to the argumentation of the analogical (and also etymological) writing of i after s, z, c (preserving the same morpholog‐ ical or etymological grounds for writing of y in other contexts). 9 Both Hanka and Jungmann were pupils of Dobrovský, and their ortho‐ graphic cooperation would last for decades. Nejedlý, with his unique position of the professor of Czech language and literature, felt the need to defend the classical orthography and made his first written critique of Hanka’s “indolence” in the third edition of his own Czech grammar (Nejedlý 1821: ii-iv, 22, 24). It should be mentioned that Dobrovský wrote the foreword to Hanka’s grammar of 1822 (Hanka 1822), having been carefully benevolent to the orthographic reform (based on his principles), but he also did not reject the older orthography at all (Hanka 1822: xxvii-xxxii). Nejedlý wrote that the principles of his grammar were taken from the best classical writers of the Golden Age, and even the orthography of the Golden Age was fully preserved and that no reform of the orthography could be accepted in the language, which had already reached the almost superior state of the grammatical education (Nejedlý 1821: ii‒iv). Nejedlý later argued in the same book (on pages 22 and 24) against the analogical orthography 104 Ondřej Šefčík <?page no="105"?> 10 This was a good point: the Brethren orthography has always used y after c, s, z, the analogical orthography used both i and y and pupils have to learn by heart a list of words containing y after s and z (vyjmenovaná slova; lit. ‘listed words’). as a contradiction in the classical orthography, which should be rejected entirely. It should be remarked that Jungmann personally wrote the introduction to Hanka’s edition of the old Czech verses entitled Starobylá skládanie (‘Ancient Poetry’), 1817 using the analogical orthography. More important was the publication of the treatise O počátku a proměnách prawopisu českého (‘On Beginning and Changes of the Czech Orthography’) (1828, cf. 6‒7, 10‒11), which, although published anonymously, was evi‐ dently the work of both Hanka and Jungmann. Jungmann never officially revealed their authorship, but it is well documented by Jungmann himself in his posthumously published autobiographical Zápisky (‘Notes’) (Jungmann 1978: 15). The co-authorship of both Hanka and Jungmann was never doubted (cf. Lisický 1898: 455). Nejedlý’s reaction in the form of the treatise Widerlegung der sogenannten analogisch-orthographischen Neuerungen in der böhmischen Sprache (Nej‐ edlý 1828) was negative. Nejedlý argues that the orthography of the Brethren had already reached a state of perfection in the 16 th century (Nejedlý 1828: 1‒ 2, 9‒12); that the new orthography is more complicated and confuses pupils (Nejedlý 1828: 6-9) 10 ; that the orthography of the Brethren is the orthography of the best Czech written literature of the Golden Age (Sollte es Jemanden von diesen Neurern einfallen zu behaupten dass es nie ein goldenes Zeitalter in der bömischen Literatur gegen habe […] ‘Should it occur to any of these newcomers to claim that there never was a golden Age […]’) (Nejedlý 1828: 13). The orthographic innovation is then dangerous and noxious (Nejedlý 1828: 14). The other critical argument was that the author of O počátku a proměnách was just a scion, supporter, even initiator, of the Pan-Slavic plan to rebuild the Czech language and merge it with other Slavic languages or with the artificially created Slavic universal speech (Nejedlý 1828: 10, 18‒19). This accusation had an enormous political value and its implications could have been disastrous for Jungmann, if taken seriously by the Austrian government and police. It was one of the main reasons why Jungmann himself (and under his own name) wrote a counter-treatise Beleuchtung der Classicism, Czech language and Jungmann 105 <?page no="106"?> Streitfrage über die böhmische Orthographie (Jungmann 1829), where he not only repeated and enhanced the analogical argumentation of O počátku a proměnách (but this time in German, as the argument’s statements were probably focused on the potential readers in the Vienna government). Be‐ sides, Jungmann (and this was probably the more personally dangerous part of the argumentation) rejects any Pan-Slavic connotations. The essential section of this study is §21, devoted directly to the question of the Golden Age. Here Jungmann extensively quotes (in his own German translation) from his O klassičnosti, by fully translating four pages (Jungmann 1829: 60‒63). Hence both treatises almost merge into the single statement (cf. Hattala 1877: 193). These pages are taken from the third part (Jungmann 1827: 36‒ 39), i.e. not the parts taken from Pölitz, but his own description of the Middle and New eras. The following statement on Classicism finished the quotation: Diese Ansicht theilen alle diejenigen, die unsere Literatur etwas näher kenne, und den inhaltsschweren Begriff der Classicität in feiner wahren Bedeutung aufgefasst haben ‘This view is shared by all those who are somewhat more familiar with our literature and who have grasped the content-heavy nature of classicism in a finer, true sense’ (Jungmann 1829: 63). We have to add for a fuller context that this analogous reform was just the first in the series of the orthographic reforms which changed the Czech orthographic system in the first half of the 19 th centuries, all of these reforms were driven by Hanka, with Jungmann as his silent associate. 4 Classicism remodelled Jungmann hence did not view the Classicism as the admiration of the seemingly perfect works of the Golden Age (in the case of the Czech lan‐ guage, positioned roughly between the years 1500‒1620) but as the dynamic process of improvement both of the language and its literature, he even daresays (Jungmann 1829: 83) that under the aegis of the emperor Francis I. a new Golden Age is approaching, better than the previous. His theoretical background was taken from Pölitz, modified according to the Czech situa‐ tion, and used, with success, within the argument between modernists and traditionalists. Surprisingly, even Jungmann’s appraisement of the Golden Age literature, which he shares with Nejedlý, is not an obstacle to change the Classical orthography of the Brethren. 106 Ondřej Šefčík <?page no="107"?> 11 We have to remark that Nejedlý died 1834 and the traditionalists lost their main protagonist, although Palkovič was opposing the new orthography in the following years (cf. Lisický 1898: 487‒493; Tešnar 2001: 104). With publishing houses steadily changing to Antiqua script and the analogical orthography, the Brethren orthog‐ raphy cause was lost soon. 12 Slovaks finally preferred to set up the new written standard ‒ Slovak. Its form given by Štúr was independent of the Czech orthography. Surprisingly, the Slovak professor of Czech language at Prague University, Martin Hattala, introduced the principles into Slovak orthography later in a remarkable manner. The “fall” of the traditionalists was caused by many factors 11 , from which we have to mention the wider social network created by the Jung‐ mannians and supported by those magazines using the reformed orthog‐ raphy: the first Czech language-written scientific journal Krok (printed first in years 1821‒1840); the influential, because widely read within the Catholic priest community, Časopis katolického duchowenstwa (printed 1828‒1851, again 1859‒1949); and Časopis společnosti wlasteneckého Mu‐ seum w Čechách (later many times renamed, printed since 1827 up to today). The last magazine, with the whole directorship of the Museum behind it, became the supreme language authority instead of the univer‐ sity: all the orthography reforms were accepted within its directorship first and later impressed into the public use through Matice česká [lit. “Czech Mother-like (foundation)”], which was both the publishing house and cultural institution, established at the Museum in 1831 under the aegis of František Palacký (1798‒1876). Palacký himself was both Jungmann’s pupil (who later inherited the semi-official status of the Father of the Nation and became the political leader of the Revivalist movement) and the editor of the first edition of Jungmann’s O klassičnosti as well. The analogical orthography was accepted entirely by the Matice-committee in the early 1840s, and it became subsequently the official school standard in 1842 by the decree of the Austrian Ministerium für Cultus und Unterricht (‘Ministry of the Cult and Education’) (cf. Tešnar 2000: 249, 2001: 104, 2003: 30; Pleskalová/ Šefčík 2007: 519). Both private persons and printing houses changed, sooner or later, to the new standard, respecting the need for the orthographic unity. This way the Revivalist orthography became the new standard even outside of Bohemia, where no “official” instructions regarding the orthography were issued, especially in Moravia and Silesia. 12 Classicism, Czech language and Jungmann 107 <?page no="108"?> 13 Here the phototypic edition of both versions is used: Dobrovský, Josef (1940). Po‐ drobná mluvnice jazyka českého v redakcích z roku 1809 a 1819 (‘Concise grammar of Czech language in redactions from years 1809 and 1819’). Praha: Melantrich. 14 Grammatykáři is a consiously archaism by Koupil; we try to reflect it with this translation. Bibliography Chalupný, Emmanuel ( 1 1909): Josef Jungmann. Praha: Spolek výtvarných umělců Mánes. Dobrowsky, Joseph (1809): Ausführliches Lehrgebäude der Böhmischen Sprache, zur gründlichen Erlernung derselben für Deutsche, zur vollkommenern Kenntniss für Böhmen. Prag: bey Johann Herrl. Dobrowsky, Joseph (1819): Lehrgebäude der Böhmischen Sprache, zum Theile ver‐ kürtzt, zum Theile umgearbeitet und vermehrt. Prag: Bey Gottlieb Haase. 13 Hanka, Václav (1817): Prawopis Český podlé Základu Gramatyky Dobrowského od geho žáka Wáclawa Hanky. Praha: U Bohumila Háze. Hanka, Václav (1822): Mluwnice čili Saustawa českého gazyka: podle Dobrowského/ od Waclawa Hanky. Praha: U Bohumila Haze. [Hanka, Václav/ Jungmann, Josef] (1928): O počátku a proměnách prawopisu českého a w čem tak nazwaná analogická od bratrské posawád užjwané orthografie se rozděluge, k ljbeznému a nestrannému uwáženj všem PP. wlastencům w krátkosti podáno. Praha: Tiskem Knjžecj Arcibiskupské knihtiskárny. Hattala, Martin (1877): Brus jazyka českého. Příspěvěk k dějinám osvěty vůbec a slovanské i české zvláště. Praze: Náklad kněhkupectví I. L. Kober. Jungmann, Josef (1827): “O klassičnosti w literatuře wůbec a zwláště české”. Časopis společnosti wlastenského Museum w Čechách 1(1): 29‒39. Jungmann, Joseph (1829): Beleuchtung der Streitfrage über die böhmische Orthog‐ raphie, veranlaßt durch Herrn Joh. Nejedlý´s Widerlegung der so genannten ana‐ logisch-orthographiischen Neuerungen in der böhm. Sprache. Prag: Fürsterzbischö‐ flichen Buchdruckerei. Jungmann, Josef (1973): Zápisky. Praha: Odeon. Koupil, Ondřej ( 2 2015): Grammatykáři: gramatografická a-kulturní reflexe češtiny 1533‒1672 (‘Grammaryans: 14 the Grammatographic and Cultural Reflexion of the Czech Language 1533‒1672’). Praha: Karolinum. Lisický Alois (1898): “Spor o i a y” (‘The i vs y Argument’). Časopis Musea království českého 78: 308‒332, 454‒493. 108 Ondřej Šefčík <?page no="109"?> Negedly, Johann (1821): Praktische Böhmische Grammatik für Deutsche. Prag: Com‐ mission in der Widtmann’schen Buchhandlung. Negedly, Johann (1828): Widerlegung der sogenannten analogisch-orthographischen Neuerungen in der böhmischen Sprache. Prag: Sommerschen Buchdruckerey. Novák, Arne (1910): “Jungmannův článek o klassičnosti v literatuře”. Listy Filologické = Folia philologica 37(2): 110‒123. Optát, Beneš/ Gzel, Petr/ Philomathes, Václav/ Koupil, Ondřej (ed.) (2019): Gramatika česká (1533). Praha: Akropolis. Pleskalová, Jana/ Šefčík, Ondřej (2007): “Pravopis”. In: Kapitoly z dějin české jazyko‐ vědné bohemistiky, Jana Pleskalová et al. (eds.). Praha: Academia, 499‒539. Pölitz, Karl Heinrich Ludwig (1825): Das Gesammtgebiet der teutschen Sprache, nach Prosa, Dichtkunst und Beredsamkeit theoretisch und practisch dargestellt. Erster Band. Philosophie der Sprache. Leipzig: J. C. Hindrissche Buchhandlung. Sak, Robert (2007): Josef Jungmann: život obrozence. Praha: Vyšehrad. Tešnar Hynek (2000): “K pravopisným polemikám v 1. polovině 19. století”. Naše řeč 83: 243‒252. Tešnar Hynek (2001): “K proměnám českého pravopisu v první polovině 19. století”. Sborník prací filosoficko-přírodovědecké fakulty Slezské university v-Opavě, řada jazykovědná D1: 95‒111. Tešnar Hynek (2003): “Ještě k pravopisným polemikám v 1. polovině 19. století”. Naše řeč 86: 27‒37. Vodička, Felix (1948): “Slovo úvodní”. In: Boj o obrození národa. Výbor z díla Josefa Jungmanna, Felix Vodička (ed.). Praha: F. Kosek, 102‒113. Voleková, Kateřina/ Dragoun, Michal (2019): “Úvod”. In: Orthographia Bohemica, Kateřina Voleková (ed.). Praha: Akropolis, xxv-xlviii. Vondrák, Václav (1926): “Vzkříšení českého písemnictví a českého spisovného jazyka koncem stol. XVIII. a počátkem stol. XIX.; Dobrovský; Básnická řeč; Jungman a jeho odpůrci”. In: Vývoj současného spisovného českého jazyka, Václav Vodrák (ed.). Brno: Filosofická fakulta s podporou Ministerstva školství a národní osvěty, 40‒54. Zelený. Václav ( 1 1873): Život Josefa Jungmanna. Praha: Matice česká, v-knihtiskárně Dr. J. B. Pichla a spol. Classicism, Czech language and Jungmann 109 <?page no="111"?> Γλῶσσαι κατὰ πόλεις Greek dialects through the lenses of the ancient lexicography. Wojciech Sowa Abstract: The following paper examines the problem of the reliability of ancient lexicographical sources while studying the Ancient Greek dialectal vocabulary, drawing on the example of an anonymous work Glossai kata poleis, an ancient anonymous list of 100 forms attributed to the various dialects (Urb. Gr. 157). Contrary to the opinions formulated by previous studies, my main hypothesis would be that, from the dialectological point of view, the list is not a reliable source of data but does contain many archaic, mostly poetic (Homeric) formations. The paper seeks also to develop new approaches to linguistic research of Ancient Greek through comparative and interdisciplinary modes of investigation: this approach relies on engaging with the historical-com‐ parative method put into a wider context of philological analysis. The grammatical and lexicographical sources are relatively late in most cases, and also reflect some tradition of understanding the works of Homer as the ultimate source of archaic forms of Greek. It seems, how‐ ever, that by the times of the composition of the lexica of Diogenianus, Hesychius and Kyrillos, the Ancient Greek dialects must have become quite an abstract notion, and a reservoir of strange and obsolete forms. Keywords: Ancient Greek dialects, glosses, etymology, Ancient Greek lexicography <?page no="112"?> 1 Introduction The Ancient Greek language has been attested in a variety of local dialects from its earliest history. Unlike in modern languages, those dialects are not variant forms of a contemporary standard language before the appearance of Koine, but they instead separately reflect the development of features inherited from the Common Greek period. They played an important role as manifestations of cultural diversity and independence of the poleis, despite the fact that the notion of an abstract, superordinate ‘Greek language’ seems to have been present among its speakers. In addition, “an extensive passive knowledge of different dialects” (Morpurgo-Davies 1987: 13) existed (at least in Athens). The two main aims of the dialectology of Greek are both synchronic and diachronic; first, the synchronic description of the attested dialectal features, second, the historical interpretation of dialectal phenomena, in‐ cluding the question of the relations between the individual dialects and the reconstructed Proto-Greek or Indo-European language (Hajnal 2007: 132 ff.). The divergent numbers of documents preserved of the individual dialects, their different character, and the different interactions or orthographical tendencies are the main obstacles to discovering the true dialectal features. A dialectographical approach is also limited since, synchronically, linguistic facts are not always equally present in every place within one dialectal re‐ gion. The sociolinguistic varieties within the dialects are almost completely lost for us, even though efforts have been made to interpret particular facts in a sociolinguistic way. The sources of our knowledge on ancient Greek dialects include: 1. epigraphically attested material, 2. forms occurring in literary works, and 3. ancient lexica and scientific works devoted to the grammar of Greek. Most important among these are, of course, the dialectal inscriptions, which have been found in various parts of the Greek world, at different times and written in different scripts (linear B, alphabet, Cypriot syllabary). In‐ scriptions from the archaic period are extremely rare, and the number of documents varies from dialect to dialect. Thus, depending on the nature of the documents, their value for dialectology also varies. The literary sources include works written in literary forms of particular dialects: in the Greek tradition, certain genres are connected with dialects, e.g. the Ionic dialect is used in historiography or jambs, Doric in choral lyrics, 112 Wojciech Sowa <?page no="113"?> or a particular mixture of Ionic with Aeolic elements in the epic tradition etc. Some authors sporadically refer to dialectal differences, whilst others use them for stylistic reasons. However, the material attested in literary works can be only partly used for dialectological purposes: sometimes it yields true vernacular material, but at other times it is to some extent artificial and reflects other literary influences. Rather than focusing on the study of dominant material in its canonical forms (inscriptions, literary sources) and their legacies, the following study concentrates on the other sort of documentation, namely the ancient testi‐ monies about Ancient Greek dialects, and more precisely, their dialectal lexicon. With the rise of philology in Hellenistic times, ancient grammarians became confronted with divergent forms of Greek. First, they explored lit‐ erary works written in different literary dialects, but from time to time they passed comment on vernacular forms as well. Caution is generally called for regarding the dialectal labels attributed by grammarians. The same problem arises in the case of ancient glosses, which have been collected throughout antiquity and have been transmitted to us in the form of lexica (most notably, that by Hesychius). The material is interesting, although to a great extent it comes from obscure sources. The literary forms are given next to the vernacular ones, and the dialectal provenance is indicated. The glosses are of great importance for the exegesis of classical texts, but the question of how far they can be considered a reliable source of dialectological data, and especially whether the glosses attest to the real state of the vernacular spoken in the various regions of Greece, remains moot. One of the interesting works in this context is the so-called Ποῖαι γλῶσσαι κατὰ πόλεις, an anonymous work that features an ancient list of 100 forms attributed to the various dialects, preserved in an Urb. Gr. 157 and kept in the Vatican Library. The forms were published by Bekker in his Anecdota Graeca III (1095-1096), without commentary, and the list has been used by writers on Greek dialects including Hoffmann, Meister and Bechtel, but always to supplement other sources of information about regional vocabularies. As of now, the document has not been critically edited (cf. Latte 1925: 136 with basic information about the textual tradition), though it seems that alongside the Urb. Gr. 157 fol. 276v there are at least two copies preserved in Madrid (Royal Library, cod. XL and cod. XCV; cf. Iriarte 1769: 146 and 378). Except for short studies by Latte (1925: 136-147) and Bowra (1960: 43-60), the document has not attracted the interests of scholars, and sporadically Γλῶσσαι κατὰ πόλεις 113 <?page no="114"?> occurring references generally consider the list as a source of reliable dialectal data (cf. e.g. Peters 1994: 210). Contrary to other extant ancient sources of scientific reflection on the Greek dialects - mostly preserved in Byzantine sources as Περὶ διαλέκτων by Gregory of Corinth, or Compendia Περὶ Αἰολίδος by Johannes Grammaticus, which treat dialects from the point of view of the phonetic discrepancies or inflection - this document tries to ascribe certain lexical forms to particular regions of Greece. The list, though unimpressive at first sight, brings several questions, which still remain to be answered. For example, what is the list and what was the purpose of its compilation? One may also try to delve deeper into the problem of its origins and sources and ask more precisely what these may have been, whether any of these words can be credited with an ancient lineage or ascribed to a particular dialect with a relative high degree of certainty, and finally, is the list a reliable source of dialectological data? There is a possibility that we are dealing with clearly poetic forms which, according to Bowra, “are familiar from poetry, they are also […] current in various local vernaculars” (Bowra 1960: 45). Contrary to the opinions formulated by Latte and Bowra, my main hy‐ pothesis would be that, from the dialectological point of view, the list is not a reliable source of data, but does contain many archaic, mostly poetic (Homeric) formations. If this could be supported through a dialectological analysis of individual forms (interpreting their provenance based on epi‐ graphic attestations from singular regions) the main problem would be to seek an answer to the question of why the author of the list arranged the forms in a given order, and further, what the relationship of this document is to the preserved lexicon of Hesychius (the most important piece of ancient lexicography), since all forms quoted in the list may be found in Hesychius, but without dialectal labels. For example, as I have suggested elsewhere (Sowa 2011: 172-179), the four forms which in the list have been ascribed to the speakers of Aeolic should not be treated as examples of the lingua vernacular but had instead been taken from the Homeric use and falsely reinterpreted as Aeolic by the Compiler (Αἰολέων· αἰχμή· λόγχη, γοός· κλαυθμός, δῶμα· οἶκος, κεκρυφάλεος· ἀριστερόν). It seems that such presumption can also apply to the following words. 114 Wojciech Sowa <?page no="115"?> 2 Ἀργείων· αἶσα· μοῖρα One of the most striking Homeric-poetic forms quoted under a dialectal label in the list is the form αἶσα, glossed in the meaning of μοῖρα ‘destiny, lot’ (cf. the same entry in Hsch. A 2096 with reference to the Homeric verses, αἶσα· μοῖρα ‘λαχόντα τε ληΐδος αἶσα<ν>’ (Σ 327) καὶ ἡ πεπρωμένη·‘αἶσα γὰρ ἦν ἀπολέσθαι’ (θ 511) καὶ τὸ καθῆκον ‘ἐπεί με κατ’ αἶσαν ἐνείκεσας’ (Γ 59)). The form has been ascribed by the Compiler to the dialect of Peloponnesian Argos, as the term Ἀργείων informs us. The noun seems to be exceptionally old, as it appears in Mycenaean documents, cf. e.g. ]a 3 -sa [ai̯ s(s)a] ‘portion’ used in nom.sg. in PY Un 1426.3 with the indication of the portion of grain GRA 12 or TH. Ug 14 ]a 3 -sa to-so O[ (cf. DM I: 138; Bartoněk 2003: 165, 177, 381, 385). In Homeric Greek, the formulaic character of the form is evident; it has been attested normally with a more abstract meaning, which is then also attested in Lyric and Tragedy (lyric parts), as e.g. as personified goddess Αἶσα (= Μοῖρα), the divinity ‘who dispenses to everyone his lot or destiny’ (φασγανουργός Aesch. Choeph. 648 (lyr.)); as appellative ‘a decree, dispensation of a god’ (ὑπὲρ Διὸς αἶσαν Ρ 321), ἐν αἴσᾳ Aesch. Supp. 545 (lyr.), or one’s lot, destiny, luck (incl. κακῇ αἴσῃ […] ἑλόμην ‘by ill luck’, E 209); it can also refer generally to the idea of sharing in a thing, ληΐδος, ἐλπίδος, αἶσα, cf. e.g. ε 40 (LSJ, cf. also dossier in LfrgrE 1: 375 f.). Examples of αἶσα in post-Homeric poetry haven been generally considered an Epic borrowing and have been used in a figurative way (Ruijgh 1957: 118 f.) The form should be considered the feminine derivate with the motion suffix *-i̯ afrom a t-stem *ai̯ tos, cf. adj. αἴσιος ‘cheap, fitting’ (Frisk GEW I 44), unattested in Greek, therefore one would interpret αἶσα as coming from *(H)ái̯ tih 2 (*h 1 ái̯ tih 2 ) understood as the extension of a *ti-abstract *h 1 ái̯ ti- (cf. Oscan aeteis ‘partis’) with *h 2 . This interpretation can be supported by the existence of an Hom. adjective αἴσιμος ‘destined by fate’ < *h 1 ái̯ ti-mo- (cf. φύξιμος ~ φύξις and Risch 1974: 105; cf. also Peters 1980: 140 94 with further examples of stems in *ti-h 2 derived of ti-stems). Next to *(H)ái̯ tih 2 (*h 1 ái̯ tih 2 ) formation one may probably find traces of casus obliqui in zero grade ἴσσα, cf., also denominative present middle infinitive ἴσσασθαι· κληροῦσθαι. Λέσβιοι in Hsch. Ι 998 ‘to allot, be appointed by lot’, with a clear indication Γλῶσσαι κατὰ πόλεις 115 <?page no="116"?> 1 The word is Hesychean hapax legomenon. According to Hoffmann, the verb is derived from the noun ἴσσα (Aeol. < *ĭt-i̯ a) and may be compared to the Hom. αἶσα (*ă-ĭt-i̯ a). The author does not make any comment on the difference between the two forms, namely, the presence of the / a/ vowel in the Homeric example (Hoffmann 1893: 238). Bechtel postulates the existence of a Lesbian substantive ἴσσα after Fick’s conjecture in the text of Homer’s ι 42 (= 549): δασσάμεθ’, ὡς μή τίς μοι ἀτεμβόμενος κίοι ἴσσης (Bechtel 1921: 120). This seems to have been generally accepted, cf. the entry in Frisk GEW I 738 with ἴσσης instead for ἴσης, ἴσσα ‘Los’ (similarly Chantraine DELG 470). 2 Peters interprets the form as one of the examples “für dessen Herleitung aus idg. HUsowohl positive als auch negative Indizien von etwa gleich großem Gewicht vorliegen” (Peters 1980: 11, 113). 3 The verb occurs in the context of sacrifices, the meaning is “vielleicht ‘aussondern, sortieren’ kurz von Ende der Opferhandlung […] Bd.-Entwicklung geben > zuteilen > verteilen > sortieren” (WOU 72). of dialectal provenance. 1 Peters (1980: 112 f.) points to the fact there are no phonetic obstacles to treating Lesb. ἴσσα as coming from < *iti̯ a-, while *iti̯ a should be interpreted as an “altes Stammallomorph der schwachen Kasus *(H)iti̯ éh 2 des αἶσα < *(H)ái̯ tih 2 zugrundeliegenden proterokinetischen Para‐ digmas”, 2 and thus yielding a preserved zero grade stem of casus obliqui of *h 1 iti̯ éh 2 as against the full grade in *h 1 ai̯ tih 2 , αἶσα. The reconstruction of the verbal root may be found in LIV 2 229 where it has been put to the shape *h 1 ai̯ - ‘give, take’, on the basis of the external evidence, cf. e.g. Toch. B aorist subj. āyu, A em ‘I will give’ < aor. *h 1 ái̯ -/ h 1 ior < *h 1 ái̯ -e-, Hitt. pēhhi, pāi, pianzi < perf. stem *h 1 e-h 1 ói̯ / h 1 iwith an element *pe ‘hin’ (cf. however Peters 1980: 82 f. against the relation of Greek αἰand Hitt. pāi-). It is probable that in its history, the Greek language possessed a nasal present formation from the same root as well, namely αἴνυμαι ‘grab, take’ (< *‘take for oneself ’ in middle) < ? *h 1 i-néu̯ / nu- (LIV 2 , ibidem). The root is also attested in other Greek forms, as ἀπ-αιτέω (denominative < *ai̯ tos), αἴτησις, αἴτημα ‘claim, postulate’ etc. (Frisk GEW I 47). A similar formation occurs in Italic, where in Oscan the forms aeteis ‘partis’ (gen.sg.), a]íttíúm (gen.pl.) meaning ‘part of a property’, have been considered to be a -tiabstract of the same root (WOU 55 f.; cf. the probable primary Umbrian verb aitu 3sg.fut.impv., aituta 3pl.fut.impv. in WOU 71 f. with discussion). 3 The anonymous author of the γλῶσσαι ascribed the form αἶσα to the Doric dialect of Argos. It seems, however, that in the epigraphic context the form is attested in the Linear B tablets, in Arcadian and Cypriot, cf. e.g. αἶσα in a small stone chest from Golgoi (Cyprus), lines 1-2 (ICS 285.2; Golgoi 26) to ti-o-se to wo-i|-no a-i-sa τῶ Διὸς τῶ ϝοί|νω αἶσα as a ‘Zeus’ portion of the 116 Wojciech Sowa <?page no="117"?> 4 Cf. also IG V, 2 40 (4 th cent. BC, Tegea), line 42 ff. [Κ]λέας Φίλωνος, Ἑρμαῖο[ς — — —] [οἵ]δ’ ἔλαχον αἶσαν ἐφ’ ἱερ[εῖ — —]·[Κραριῶται]· Ἀγαθίας; IG V, 2 41 (4th cent. BC, Tegea), col. I-II, line 17 [οἵδ’ ἔλαχον] αἶσαν ἐφ’ ἱερεῖ [— — — ἐπ’] Ἀθαναίαν· πολῖται·; IG V, 2 265 (64-61 BC, Mantineia,), line 33 πέμπειν δὲ αὐτᾶι καὶ αἶσαν ὡσαύτως·; IG V, 2 269 (1 st cent. AD, Mantineia), line 26 δείπνοις ἀποστέλλειν αὐτῆι αἶσαν, and 28 ἐὰν δέ τις μὴ καλέσῃ εἰ μ[ὴ]ἀποστείλῃ τὴν αἶσαν οἷς ἐπιβάλλον ἐστίν; IPArk 11 (61/ 60 BC, Antigoneia Mantineia), line 33 πέμπειν δὲ αὐτᾶι καὶ αἶσαν ὡσαύτως· ε[ἰ δέ τ]ις μὴ καλέσει τῶν ὑποδεχομένων, καλ[είτω ὁ ἐπιγν]ώμα; IPArk 13 (1 st cent. AD, Antigoneia Mantineia), line 26-29 δείπνοις ἀποστέλλειν αὐτῆι αἶσαν, καλεῖν δὲ ἐπὶ τὰ γέρα καὶ Γάϊον Ἰούλιον Στρόβειλον· ἐὰν δέ τις μὴ καλέσῃ εἲ μ[ὴ]ἀποστείλῃ τὴν αἶσαν οἷς ἐπιβάλλον ἐστίν, ὑπόδικος ἔστω δραχμαῖς πεντήκοντα. wine’ (cf. also Egetmeyer 1992: 4; Egetmeyer 2010: 251), and αἶσα 10x in Arcadian official inscriptions in a sense ‘part of a sacrifice, Opferteile’ both in dialectal and later documents (Dubois 1986: 74), e.g. IG V, 2 39, line 13 (4 th cent. BC Tegea) ἔλαχον αἶσαν ἐφ’ ἱερεῖ and juridical document πέμπειν δὲ αὐτᾶι καὶ αἶσαν ὡσαύτως (IPArk 11 Antigoneia Mantineia 61/ 60 BC). 4 None of the attestations yield a poetic context, and so it is therefore quite clear that the form then has to be understood as an authentic one, belonging to the most archaic lexical stock of Greek (by some scholars called ‘Achaean’, Ruijgh 1857: 118 f.). It seems that the probable reason of the qualification of the word as Argivic could be the mention of αἶσα by Athenaeus in his Deipnosphistai (68, 50-53 Kaibel) while quoting Hegesander of Delphi, a Greek grammarian his‐ torian (2 nd cent. BC) and his anecdotical work Commentaries (Ὑπομνήματα): Ἀργεῖοι δ’, ὡς ἐν τοῖς ὑπομνήμασί φησιν ῾Ηγήσανδρος γράφει δ’ οὕτως (FHG IV 419) ‘τὴν συμβολὴν τὴν εἰς τὰ συμπόσια ὑπὸ τῶν πινόντων εἰσφερομένην Ἀργεῖοι χῶν καλοῦσι, τὴν δὲ μερίδα αἶσαν’ - “but the Argives, as Hegesander says in his Commentaries, have other words. He writes as follows: ‘the contribution brought into the symposia by the drinkers is called by the Argives a chôs (heap) while the single share is called an aisa (lot)’” (C. B. Gulick (transl.), Athenaeus Deipnosophistae, Loeb Classical Library, Vol. IV, 1930). The whole picture is, however, complicated, since one may find also a single attestation of the αἶσα in an epigraphic context in Argos, or, better to say with a reference to Argos. In one document of Cretan provenance the two cities conclude a treaty using the mediation of Argos (IC I xxx1, Tylisos, stoich., mid-5 th cent. BC, cf. also Meiggs-Lewis 42, A 10, 17), cf. col A.I line 10 συνβ]άλλεσθαι δὲ τὰν τρίτ‖[αν αἶσ]αν τὸς Ἀργείος τᾶν ψά‖[φōν· ὅκα] τινὰς and col. B.I. lines 5 f. ἐκ Τυλίσō τᾶν ψάφōν τὰν τρί‖ταν αἶσαν. Γλῶσσαι κατὰ πόλεις 117 <?page no="118"?> 5 There is an ongoing discussion as far as the nature of the group and its place next to Doric proper is concerned, cf. e.g. Méndez-Dosuna (1985: 508), who argues that Northwest Doric should not be considered a proper dialectal group (in a genetic sense) but rather merely a case of areal dialectal convergence within the ‘West Greek’ (in the synchronic perspective). It seems however that ‘the term ‘North-West Doric’ has been considered more accurate nowadays emphasizing many features that are common to both groups rather than on their less numerous and largely secondary differences’ (Filos 2018: 227). In both cases the form has been applied in a sense of a ‘third part of the vote’. The problems occur while interpreting the αἶσα as a part of a Doric dialect of Argos since in Doric, one would expect retained geminated *-ssafter i̯ from *h 1 ai̯ tih 2 , cf. Peters 1980: 140 94 ). The inscription yields some of the Argivic features, e.g. aspiration of the intervocalic -sinto -h- (cf. οἱ Κν]ṓℎιοι but retained in Τυλίσō in a context of dialect foreign names), which seems to occur also after glide (cf. Ηsch. K 1119 κατ’αἶαν· κατὰ τὸ πρέπον next to K 1137 κατ’ αἶσαν· κατὰ μοῖραν (Γ 59). In such a case of assuming the inherited character of allegedly Argivic αἶσα one would expect the form *αἰhα (or a sort). We must then suppose that either the form is not inherited, or the orthography does not render the true pronunciation. Since the evidence from the epigraphic sources comes mostly from Arcadian, Cypriot and Mycenaean, one should then probably treat Argivic αἶσα as a borrowing from the Arcadian (Peters 1980: 140 94 ), and not a reliable case of the dialectal vocabulary. 3 Ἀκαρνάνων· στεῖχε· πορεύου An interesting case of the form quoted on the list is an imperative στεῖχε to a verb στείχω (used in place of the forms of a verb ἔρχομαι) in the expected meaning ‘πορεύου’. The form has been ascribed to the inhabitants of the Acarnania region in west-central Greece, speaking a dialect belonging to a Northwest Doric group. 5 The forms of such verb are not attested in epigraphical sources of the region (cf. IG IX,1 2 ), the transmitted imperative does not yield any characteristic features typical for the assumed region. Interestingly, the form can be considered a dialectal element, but rather typical of Lesbian dialect, cf. στείχω in a sense of πορεύομαι, ἐμβαδίζω ‘to go, go inside, to enter’ attested epigraphically, cf. e.g. MYT 04 lines 4-6 (ca. 332 BC; political) […] μ[ηδὲ στειχέτω ἐπὶ μη]δὲν τῶμ παρεχώρησαν αὔτωι 118 Wojciech Sowa <?page no="119"?> οἰ ἐν τᾶι πόλι πρό[σθε ἔοντες, ἀλλὰ σ]τειχόντων ἐπὶ ταῦτα τὰ κτήματα οἰ παρχωρήσα[ντες αὔτωι […] “[…] nor shall he possess anything that those who were living previously in the city surrendered to him; but let those of them living previously in the city who surrendered to him, repossess these properties…” (transl. of Heisserer 1980: 124); ERE 011 lines 1. 10. 11. 18 (beginning of the 2 nd cent. BC; sacral) [---]ς εἰστείχη̣ν̣ εὐσέβεας vac. ἀπὸ μὲν κάδεος ἰδίω [περιμένν]αντας ἀμέραις εἴκοσι […] ‘[---]will enter (as) pious vac. when they wait 20 days from the funeral (first)’; […] δὲ μὴ εἰστείχην μηδὲ προδόταις ‘no traitors will enter’; etc. Of course, the forms of the verb στείχω are attested in the Greek inscrip‐ tions from different regions, the examples occur exclusively in the poetical sort of documents (mostly funeral, the metrum is eleg. distichon). In Lesbian inscriptions, στείχω is attested in political decrees, as well as in sacral law, which expresses the ‘limitations in the access to the temple’, in this case, the meaning is surely ‘to enter’ (the sanctuary), specified by the use of a preverb εισ-, cf. e.g. ERE 01 line 18 μὴ εἰστείχην δὲ μηδὲ γ̣ ̣ ύ[ναικ]α ‘[…] the women will not enter […]’. In the political decree about the ‘return of the exiles’ the forms of ἔρχομαι are also attested (with preverb κατα-), next to the forms of στείχω: οἰ κατεληλύθοντες (lines 1, 3, 9, 10 f., 17), οἰ κατελθόντες (lines 22, 24, 26, 41, 49), οἰ κατελθόντες ἐπὶ Σμιθίνα προτάνιος (line 36), in all the examples they designate different groups of exiles (‘those, who had been banished’; cf. Heisserer/ Hodot 1986: 13 ff.). It seems that the verb στείχω is used in a special context, it should be translated ‘to come back, to return (to one’s possession)’ - ‘to repossess’. In the other dialects and inscriptions, the examples of the verb στείχω are used only in the poetical context, the usual prosaic correspondences are εἰσπορεύομαι, ἔρχομαι (cf. e.g. Megara IG VII 190, 5 line17; Cyrenaic; Ionic, cf. Hernandez-Vazquez 1994: 283; cf. Thess. ἔρχομαι, πορεύεσθαι - ἐπὶ τὸγ χαριστήριον IG IX, 2 1109 line 18). In Homer, the present forms of στείχω are used without any prefixes and have the meaning ‘to march, walk’ (go in an order), cf. also Hsch. Σ 1717: στείχειν· πορεύεσθαι, conversely forms of thematic aorist are accompanied by the prefixes and mean ‘go up, climb’, without prefixes ‘to move, to go away’ (Farmini 1984: 270). Letoublon points to the fact that the subjects are always animated beings as humans or animals; the occurring Sun in λ 17 can Γλῶσσαι κατὰ πόλεις 119 <?page no="120"?> 6 Cf. λ 17: Ἠέλιος φαέθων καταδέρκεται ἀκτίνεσσιν ‖ οὔθ’ ὁπότ’ ἂν στείχῃσι πρὸς οὐρανὸν ἀστερόεντα. be considered a result of a personification; 6 therefore, the suggested original meaning would be ‘to march in an order’ connected to the military usage (Letoublon 1985: 168). The compounds with στείχω may correspond to the other forms of the verbs expressing the meaning ‘to go, come’, e.g. ἀποστείχω = ἀπέρχομαι; προσστείχω = προσέρχομαι. The switching from the original military meaning ‘to march in a line’ (Homer) to ‘march in a direction of; to go to’, and thus the neutralization of the meaning has been accomplished by the use of preverbs. Such an original meaning could explain the use of the form στείχω in the post-Homeric literature in the context of processions or solemn marches (Letoublon 1985: 169). Instead of other forms of the paradigm ἔρχομαι, the use of the verb στείχω can be treated as a specific Lesbian feature (cf. the parallel situation in Arcadian, where the typical poetical verb ἕρπω ‘move slowly’ LSJ is attested in the simple meaning ‘to walk’), therefore the Bechtel’s statement “[d]ie Attische Prosa hat das Verbum στείχω verloren. In der lesbischen, ionischen und dorischen ist es erhalten geblieben” (Bechtel 1921: 162) does not seem to be precise. In the case of Lesbian, the examples are to be found in inscriptions in which the poetic influence is surely to be excluded. In case of Doric, the example of the use of the verb comes from literature, namely from Epicharm (Syracuse; poetry), cf. Fr. 34.1: ἀλλ’ ἄλλος ἔστειχ’ ὧδε τοῦδε κατὰ πόδας, and under the term Ionic Bechtel means the prose of Herodot (cf. 1.9.13: στείχῃ ἐπὶ τὴν εὐὴν κατὰ νώτου τε αυτῆς γένῃ […]). 4 Κρητῶν· φώς· ἀνήρ Under the label Κρητῶν the author of the list quotes ten forms, among which is the form φώς, which, as has been suggested, the inhabitants of Crete used when speaking of ‘man’, as an equivalent of ἀνήρ. Such use would resemble that known from the Homeric Epic, where the form φώς has been noted in a primary sense ‘man, hero’, often coupled with ἀνήρ, cf. e.g. Ρ 17. 3 77f. δύο δ’ οὔπω φῶτε πεπύσθην ‖ ἀνέρε κυδαλίμω… The form has been considered a poetic noun already in the Antiquity, it occurs in Tragedy used mostly with reference to heroes (probably as a poetic/ Epic element), cf. e.g. Aesch. Sept. 499, where the form is attested in the lyric part τοιοῦδε φωτὸς 120 Wojciech Sowa <?page no="121"?> 7 As a typic Epic form, the word has also been used in a stylistically marked context in Comedy, being the parody of the Tragic style, cf. Ar. Pax 528 ΤΡ. Ἀπέπτυσ’ ἐχθροῦ φωτὸς ἔχθιστον πλέκος. 8 Cf. Eustath. Comm ad Hom. Od. I. 376.4f. λέγει ὁ αὐτὸς ῾Ηρακλείδης τοὺς Παμφυλίους ἄλλως χαίρειν τῷ β, προτιθέντας αὐτὸ παντὸς φωνήεντος. τὸ γοῦν φάος φάβος φασί. καὶ τὸ ἀέλιος βαβέλιος. οὕτω δέ φησι καὶ τὸ ὀρούω ὀρούβω. 9 The document is, however, poetic, influenced by the Epic cf. τίς δέ σε ὠκυμόροιο φάους ἀπενόσφισε νοῦσος. 10 The entire word-family forms belong to the IE root *b h eh 2 - ‘to shine’, cf. Ved. bhā́ti (LIV 2 68 f.). πεῖρα εὖ φυλακτέον, although it may have appeared in the general meaning ‘men’ as well, cf. e.g. Aesch. Pers. 242 οὔτινος δοῦλοι κέκληνται φωτὸς οὐδ’ ὑπήκοοι (in trochaic meter, dialogue). Further use also includes the sense ‘man’ as opposite to ‘woman’, or even ‘mortal’, as opposed to ‘god’, cf. P 98 πρὸς δαίμονα φωτὶ μάχεσθαι or φῶτα βρότειον in Eur. Bacch. 542 (lyric part). 7 The form may appear to stand in a quasi-homophonic relationship with the word for ‘light’ φῶς, although it should be emphasized however that the latter does not appear in a contracted form until the 6 th cent. BC (Theognis) and the Classic times, in contrast to the more archaic φάος. In the Homeric Kunstsprache only φάος and φόως, never φῶς was used (cf. dat.sg. φάει and acc.pl. φάεα; the form of the Epic dat.pl. used by Hesiod cf. Fr. 23a8 Λήδη ἐ ̣ [υπλόκαμος ἰκέλη φαέεσσ]ι σελήνης and 252.5 Θηρώ τ’ εὐειδέα ἰκέλην φαέεσσι σελήνης), from there it may have been introduced into lyric by Pindar. On the contrary, in Tragedy, both forms φάος or φῶς may be found in both lyric and dialogue parts, depending on the metric needs and in Comedy φάος may be observed in the lyric parts exclusively (Frisk GEW II 990; Beekes 2010: 1551 f.). Transmitted examples of forms as φάβος ascribed by Heracleides to Pamphylian (cf. Brixhe 1976: 47, 143) 8 or Attic φαους (IG II 2 10073), 9 nom.acc.pl. φάεα may testify to the actual retention of the original */ u̯ / as in φαυο-, 10 which may also be found in numerous derivatives founded on the thematic aorist of the primary verbal formation, as in Homeric φάε ‘lighted up’, cf. compounds δια-, ἐπι-, ὑπο-φαύσκω, -φαῦσαι etc. All such forms are based on the thematic aorist φαϝεῖν (Schwyzer 1939: 747) and continue an IE stem *b h h 2 -u-V, which do not appear to have direct cognates outside Greek. Interestingly the Homeric-Epic φώς in a sense ‘man, hero’ can be derived from the very same root *b h eh 2 - ‘to shine’ as a nominal formation *b h oh 2 tand understood originally as ‘someone (repeatedly) shining’ (Peters 1993: Γλῶσσαι κατὰ πόλεις 121 <?page no="122"?> 11 In E 4-8, at the beginning of Diomedes’ aristeia, Athena causes ‘tireless fire’ to shine forth from his helmet and shield and the flame is then compared to a star. A similar situation is present in Σ 203-214, where at the beginning of Achilles’ aristeia, Athena casts a golden cloud around the hero’s head and causes, again, fire to shine forth; this time, the fire is compared to human-made torches, not to celestial bodies, cf. αὐτὰρ Ἀχιλλεὺς ὦρτο Διῒ φίλος· ἀμφὶ δ’ Ἀθήνη ὤμοις ἰφθίμοισι βάλ’ αἰγίδα θυσσανόεσσαν, ἀμφὶ δέ οἱ κεφαλῇ νέφος ἔστεφε δῖα θεάων χρύσεον, ἐκ δ’ αὐτοῦ δαῖε φλόγα παμφανόωσαν. ὡς δ’ ὅτε καπνὸς ἰὼν ἐξ ἄστεος αἰθέρ’ ἵκηται τηλόθεν ἐκ νήσου, τὴν δήϊοι ἀμφιμάχωνται, οἵ τε πανημέριοι στυγερῷ κρίνονται Ἄρηϊ ἄστεος ἐκ σφετέρου ἅμα δ’ ἠελίῳ καταδύντι πυρσοί τε φλεγέθουσιν ἐπήτριμοι, ὑψόσε δ’ αὐγὴ γίγνεται ἀΐσσουσα περικτιόνεσσιν ἰδέσθαι, αἴ κέν πως σὺν νηυσὶν ἄρεω ἀλκτῆρες ἵκωνται ὣς ἀπ’ Ἀχιλλῆος κεφαλῆς σέλας αἰθέρ’ ἵκανε Cf. Zeitlin (2018: 56): “On the battlefield everything coordinates into a unified and dazzling presence, enhanced by the visual brilliance and auditory resonance of bronze armor. The panoply of arms -of greaves, breastplate, shield, and sword -not only aims at invulnerability. It produces a supernatural radiance, one that sparkles with a thousand bursts of light, whose scintillating effects are magnified by the artful craftsmanship of carved, embossed, and molded designs that glitter, shimmer, sparkle and flash as they catch the shafts of sunlight in refracted pinpoints of light. No wonder this apparition strikes dread in the enemy, and no wonder it evokes similes drawn from the cosmic sphere of sun, star, moon, and meteors, and seems, at its height of intensity, like the force of fire itself, both beautiful and terrible, attesting to a surplus of life for you and death to the adversary.” 12 Cf. also Av. xvarənah, OPers. -farnah, used in the onomastics; for *b h eh 2 in the Greek onomastics, cf. Peters (1993: 104). 101-105), in the Epic language already ‘a hero’ and then ‘man’. The idea of a young ‘powerful’ IE hero manifests in a picture of a glorious wreath of light that surrounds the warrior’s head in battle (Janda 2000: 246 f.), 11 a typical IE idea, “a belief … that a particular radiance, sent by God, issued from the heads of warriors and princes when they were engaged in battle or for the purpose of presaging a glorious destiny form them” (Campanile 1988: 93 f.), which lives on in the Old Irish term known from the epos as lúan láith ‘warrior moon’ or ‘hero light’ and has been used as an attribute of the best heroes as e.g. Cúchulainn. 12 In the Common Greek Indo-European aristocratic culture this idea must have been already lexicalized occurring in the Homeric φώς (homonymous to φῶς in later period). There is no doubt that in the case of φώς ‘man, hero’ we are dealing with an archaic word and meaning which may be traced back to the IE heritage. 122 Wojciech Sowa <?page no="123"?> 13 Cf. IC II xiii 5 (Elyros, 2 nd cent. AD): [Ἀ]γησίφως Ἀδράστοι τῷ πατρὶ [κ]αὶ Γεροίτοι το͂ ι νέννοι, IC II xiii 10 (Elyros, 2 nd cent. AD): Ἀγησίφως Τειμασιθέω ⋮ Εὐθυμίαι Πανκλῆ ⋮ τῇ ἰδίᾳ γυναικὶ ⋮ μνημόσυνον. { 2 palmula} 2 ; further SEG 27, 628 [2] (Itanos, 2 nd / early 1 st cent. BC) Ἀρισστόφω {gen. of Ἀρισστόφοος/ Ἀρισστόφως}; cf. also IC I xxiii 16 (Phaistos, 2 nd cent. BC) Ἀριστόφοος, Κάλλιστος Ἀριστόφω̣ , IC IV 259 (Gortyna ca 200-150 BC) Ἀντιφάτα, Ἀγήσανδρος Μω[ρ]αύτ ̣ α, Ἄμνατος Ἀριστοφόω, Ἐρταῖος; IC I xxv 4 (Pyloros, 1 st cent. AD) [— — —] κἠχετίμας τᾶς γυν[αικός· Π]ανάρης Ἀριστοφόω ἀρ[γύριον φέρει — — — —]κ̣οκ․․․; IC II iii 5 (Aptera 189 BC): δρυς Τυχαμένεος εἶπε· [— —] …π̣οιν Ἀριστοφόω Συβρίτ[ιον] [πρόξενον ἦμεν αὐτὸν καὶ ἔκγονα. The question remains, whether one should consider the attested form to be a Cretan-Homeric isogloss, or in a similar way as the case of the preceding form falsely ascribed to this particular dialect. In his profound study on the etymology of φώς Peters claimed the exclusive character of the comparison and stated that the evidence of the γλῶσσαι κατὰ πόλεις should be taken as a reliable material. The most convincing argument speaking in favour of such suggestion would be the presence of the element ºφως, ºφοος (in gen. ºφόω, ºφω) in the compound names attested in Crete, cf. Ἀγησίφως, Ἀριστόφως (or ºφοος; cf. dossier in Masson, BCH 103, 1979, 73 f.), which may be identified with φώς ‘man’ (Peters 1986: 317 46 ; Peters 1993: 101-105). According to Peters, the unexpected character of this Doric-Homeric isogloss may be explained by ascribing the form to the assumed so-called ‘Old Peloponnese’ lexical stock (Altpeloponnesisch), originally an archaic variant of the Proto-Aeolic dialect once spoken in the region of Peloponnese and the Doric islands, along with the Proto-Ionic (Peters 1986: 317 ff.). Interesting as this theory might be, it does not appear to solve the problem of the allegedly Cretan provenance for the form φώς, as suggested in the list. In the epigraphic documents from Crete, no example of such form may be observed (cf. however 314 examples of a stem ἀνήρ (11), ἀνδρός (303) the latter both as a substantive and as an element of the compound names), one may observe once attested φῶς ‘light’ - here ‘daylight” as a synonym of ‘life’, cf. IC II x 20, the document is however written in poetic style, Ματτία οὔνομα ἐοῦσα, λιποῦσα δὲ φ[ῶς] ὑπὸ [κ]ε ̣ [ύ]θη, the chronology is late (Roman Imperial Period), cf. also late attestation from the 2 nd or 3 rd cent. AD from (H. Nikolaos) Crete (Lato pros Kamara), in a non-clear context Επ[ι ̣ ]|Φ|Ω|Σ (‘loomweight, inscribed on four sides’, SEG 32, 899 (31)). The personal names quoted by Peters are attested relatively late, cf. Ἀγησίφως attested twice in 2 nd cent. AD and Ἀριστόφως, in the 2 nd and 1 st cent. BC, 13 and of course they may be considered examples of the conventional use of Γλῶσσαι κατὰ πόλεις 123 <?page no="124"?> 14 Cf. 88 1-5 [τοὶ] ἐς τῶν νεωτέρων ἐν τὼς ὁπλίτα[ς ἀπε]γράψανθο, Βοιωτοῖς ἄρχοντος Ἰσ[μ]ε ̣ ινίαο, ἐπὶ πόλιος Ἐπιτέλεος· [Ἀ]ριστοφάων Ἀλκιμάχιος [Θ]όρικος Θουέλλιος… . 15 For other examples of PN with the component derived from the same root *b h eh 2 cf. Peters (1993: 104-107); cf. also Janda (2000: 246 ff.). the compound names and not proof of the exclusivity of -φως element in Doric dialects of Crete, especially as the element also seems to be attested in the feminine names in other regions, as e.g. gen. Ἀριστοφώσσης (Kos, IG XII, 4; 1 st cent. AD) or similar form in Attica, cf. IG II 2 10799 (5 th / 4 th cent. BC) Ἀριστοφώσης τῆς μητρὸς τῆς Ἀμφήνορος καὶ Διογνήτο (cf. also female PN Κλεισφύσσα, e.g. Melos IG XII, 3 1190 line8 (2 nd cent. BC) ἐκ δέ γε μητρὸς Κλεισφύσσας, emended into the form Kλειφώσσα by Hiller von Gaertringen 1908: 175 (for the problem of the orthography with geminated / s/ see Peters 1993: 102 76 ). One may further quote evidence from Boeotian Thespiai, where the PN Ἀριστοφάων may be observed (mid. 3 rd cent. BC), 14 a formal and semantic parallel to Ἀριστόφως based on the stem *p h au̯ ont- ‘shining’, a reformed original n-stem *b h ău̯ on-, a remnant of the original heteroclitic neutral abstract noun *b h ah 2 -u̯ r̥ , *b h h 2 -u̯ en-s- ‘the shining’ (cf. casus obliqui in Gr. φαείνω < *p h au̯ en-i̯ e/ o-; cf. Peters 1980: 244f. 198 ). Both names, Ἀριστοφάων and Ἀριστόφως, would yield the different apophonic and morphologic variations of the same root: φώς being an o-grade *b h oh 2 tand φάος from *p h au̯ os actually ‘a contamination from inherited *b h aos < *b h ah 2 os with *b h au̯ ər < *b h ah 2 -u̯ r̥ ’ (Peters 1993: 107 97 ), expressing the idea of the ‘best hero’, as a warrior ‘possessing the best shining’. 15 It seems that the evidence from Greek outside the poetical literary works yields the element φώς in meaning ‘shining, warrior, hero, man’ still in this archaic sense in personal names in various regions of Greece. The regular form for a ‘man’ in Cretan epigraphical sources seems to be a substantive ἀνήρ, cf. the archaic document IC I x 2 (Eltynia, sinistr. (lines1-2); boustr. (lines 3-11), end 6 th / beg. 5 th cent. BC), cf. l 5 αἰ] δέ κ’ ἀν̣ ὴρ | τὸν πηΐσκον παίηι | μὴ [- -]; therefore, the labelling Κρητῶν· φώς· ἀνήρ as proposed by the Compiler of the list seems to be imprecise. There can be no doubt that the word in such form belongs to the Homeric vocabulary, the examples from the onymic area may, in my opinion, point to the conventional use of the element in an archaic semantics, they cannot, however, prove to the alleged dialectal provenance of the form. 124 Wojciech Sowa <?page no="125"?> 5 Κυπρίων· ἄλοχος· γυνή Under the label ‘Cypriots’ the anonymous Compiler of the γλῶσσαι quotes the form ἄλοχος, normally attested in the meaning ‘wife, partner of one’s bed, wife, leman, concubine’ and ascribes the meaning ‘woman’ to the form. ἄλοχος does not pose any formal problems concerning the etymology or word-formation of the noun (*sm̥ -log h -). Being a nominal derivate from the root *leg h - ‘to lie down’ it has been used often in Greek next to the other members of the word-family, as e.g. aor. ἔλεκτο, present λέχεται· κοιμᾶται (S) (Hsch. Λ 761), λέχος ‘lair, bed’, especially of either ‘bridal bed’, or a ‘deathbed’ and e.g. λόχος ‘childbed’, but usually ‘ambush, band (in ambush)’, or a ‘military armed band’ (in this sense in a compound λοχᾱγός, considered a Doric form) etc. (Beekes 2010: 852 f. for the etymology and comparative material). According to Ruijgh, all formations built from the IE root *leg h in Greek belong to the socalled ‘Achaean’ stock of the Epic vocabulary, which should be understood as the presence in Homeric language of an element which is earlier than both the Aeolic and the Ionic stock. In this view, the ‘Achaean’ dialect would mean forms of Greek spoken in the Peloponnese and in adjacent islands of the Aegean during the Mycenaean age. Remnants of this dialect (or group of dialects) may occur in classical times in Arcadia and Cyprus and the record of the earliest Achaean forms may be observed in the Linear B tablets. Ruijgh qualifies the word-family of λεχ- (as e.g. λέκτο, λέκτρον and ἄλοχος) as ‘Achaean’ elements in Homer, being recognized as such chiefly from their use in formulae (Ruijgh 1957: 153 f.), the poetic character seems to have been attached to the forms such as, e.g. λέκτρον and ἄλοχος and kept throughout the history of Greek until late prose (Lucian of Samosata). It is also worth noting that the use of the derivatives from λεχseems to be restricted to Ionic-Attic and limited only to the military terms (λόχος ‘ambush’ and military band’ - probably borrowed from Doric) and the terms expressing the idea of ‘lying’ as in a ‘child-birth’: adj. λόχιος ‘belonging to child-birth’ (cf. also Λοχία, Λοχεία epithets of Artemis and τά λόχια ‘discharge after child-birth’), λόχμη ‘thicket, lair of wild beasts’, (cf. Soph. Ichn. 127 ἐχῖνος ὥς τις ἐν λόχμῃ κεῖσαι πεσών) and λέσχη, ἡ ‘couch Γλῶσσαι κατὰ πόλεις 125 <?page no="126"?> 16 λέσχη seems to already be attested in the Mycenaean documents in a form of a female PN in Pylos, cf. e.g. PY 886.A re-ka te-o-jo-do-e-ra, or Ep 212.1 re-ka te-o-jo-do-e-ra o-na-to e-ke ke-ke me-na ko-to-na pa-ro da-mo to-so pe-mo GRA T 6 (DM II 237). In all cases, we are dealing with a name which should be interpreted as Λέσχα, a female pendant to masc. Λέσχης. 17 Cf. e.g. Attic (Eleusis) IG II 2 4717, 1 st cent. AD (I. Eleusis 328; SEG 30, 172), poetic: Διώνη Προσῆθεν Πασικράτους ἄλοχο[ς εὐχ]ὴν Κου[ροτρόφῳ] ΙΝΛΑ ---vac.; Attic ca. 360 BC - IG II 2 3453 (SEG 30, 140), poetic, [Λυσιμάχη]‖ [ἡ γραῦς ἣν ἐσορᾶτε] Δρακο[ντίδ]ο ἦν [ἄλοχ]ος μέν; Sparta late 3 rd / 2 nd cent. BC (IG V, 1 724), poetic, νῦν δέ νιν Ἀρκαδίας ἀπὸ πατρίδος ὧδε θανόντα ‖ κουριδία Τιμὼ τύμβωι ἔκρυψε ἄλοχος; Thebai, 3 rd cent. AD (IG VII 2539), poetic, τούτου κουριδίη ἄλοχος καὶ [κ]εδνὰ εἰδυῖα; Akarnania (Thyrrheion), 2 nd / 1 st cent. BC; IG IX, 1 2 2: 340, line 9, poetic ἔνθα μοι οὔτ’ ἄλοχος παρακλείνεται οὔτε πατρὸς μου; Kephallenia IG IX, 1 652 Roma Period (prose) τὸ σῆμα τοῦτο Ποπλίου Βειβίου Αἴα<ν>τος ἄλοχος Οὐλπία Ἀπ̣ουσία ἐτῶν λζʹ χαῖρε; Thessaly, Larissa, IG IX, 2 661, line 7 f., 4 th cent. AD, poetic (= SEG 37, 489) Κλαυδίου ἡγεμόνος ἐνθάδε κεῖμ’ ἄλοχος; Macedonia IG X, 2 2 159, line 5, Debrešte: Gradište, 2 nd / 1 st cent. BC (= SEG 26.730), poetic νῦν δ’ ἄλοχός τε φίλη καὶ υἱέες, οὓς ἀπολ[είπει]; Thrace, Philippopolis (Plovdiv), Bolyartsi, IGBulg III, 1 1448, line 4, poetic οἰκοδόμου ἄλοχος ‖ Εὐκλαδίου, ὅς με θανοῦσαν ; Scythia Minor - Kallatis (Mangalia) IGBulg III 131, 3 rd cent. BC, line 5 poetic κείνου ἐς εὐναίους ἦλθε ἄλοχος θαλάμους; Delos, end 3 rd cent. BC, IG XI, 4 1299, line 85, poetic σή τε ἄλοχος· φῶτας γὰρ ἀλιτρο<νό>ους ἐπέδησας; Kalymna, Sanct. of Apollo Tit. Calymnii 130D, b, ca. 260 BC, poetic, line 1 Βίτιόν εἰμι ἄλοχος Διοτίμου, παῖς Στρατονίκου ‖ ἇς τάδε τιμάεντα· ἀγλαὰ τέκνα πέλει; Euboia — Tamynai IG XII Suppl. 540 [--]ά̣λοχο[ς]; Caria, IK Knidos I 131 middle 4 th cent. BC, line 2 poetic Χρυσογόνη[ς] μήτηρ, Ἱπποκράτους δὲ ἄλοχος; Bithynia, Nikomedeia TAM IV, 1 322[1], line 8, poetic μούνη δ’ ἐν [θαλάμοις ἄλοχος ποθέ]ουσα σιωπῇ·; Asia Minor, Galatia, MAMA VII 584, poetic line 9 f. οὗπερ δὴ ἄλοχος πινυτὴ μάλα (metaph. of tombs), lounging place, public building or hall, used as a lounge or meeting-place’ finally ‘talk or gossip, such as went on in the λέσχαι’ (LSJ). 16 According to Peters, ἄλοχος as attested in the list and as suggested by Ruijgh could be a real element of the Cypriot vocabulary, especially due to the presence of one testimony from the Cypriot alphabetic prose-inscription yielding a form of a genitive ἀλόχου (Peters 1994: 205 8 ). The epigraphic evi‐ dence from various regions of Greece generally attests to the use of ἄλοχος in epigraphic context (more than 40 times); except for the above-quoted example, all others come from different regions of Greece, but actually not from Cyprus, cf. e.g. Attica, Laconia, Thebes, Akarnania, Thessaly, Ionic Islands, Macedonia, Thrace and Moesia inferior, Delos, Kalymna, Syros, Thasos, Euboea, various regions of Asia Minor, and Italia; they generally occur either in the funerary or dedicatory documents, as a conventional use until middle of the 5 th cent. AD (Asia Minor). The style of documents is almost exclusively poetic (hexameter, or eleg. distichon). 17 One of the most 126 Wojciech Sowa <?page no="127"?> πολλὰ καμοῦσα; Asia Minor, Galatia, MAMA VII 587, ca. 375-425 AD, line 8, poetic τῷ δ’ ἄλοχος Κλέουσα προσένν̣ [ε]πε μυρομένη περ; Lykaonia, Kana, MAMA VIII 221a, 381-451 AD, line 6 poetic τοῦδ’ ἄ̣[λ]οχο[ς] κε ̣ [δνὴ] στήλην ἐπὶ τύ[μ]βῳ [ἔπ]η̣ξ ̣ εν; Italia, Roma, mid-2 nd cent. AD or shortly after IGUR III 1222 (SEG 49, 1371,1), line 12 f. [ἔ]νθα φίλῃ ἀλόχῳ Κλυ[τ]ίνῃ καὶ παιδὶ φίλοις [τε] / [τ]εῦξα τάφον. 18 IDélos 2 (= Plassart 1950); IG XII 5, 2, 1425b; CEG I 403. The inscription is one of the oldest testimonials of the so called Central Ionic dialect (Inselionisch), the orthography is particularly archaic, and the ductus is still in boustrophedon. The writing points to the so-called “pale blue” Greek alphabets, as known from Paros, Naxos, Thasos (and also Attica), where the use of such signs as “phi” Φ for an aspirated / p h / , “khi” Χ for an aspirated / k h / are present and where a general absence of such signs known from classical Attic as “ksi” and “psi” can be observed (Jeffery 1961: 291 ff.). Such consonant clusters have been rendered in those regional alphabets by a combination of “phi” and “khi” with “sigma” instead, in the archaic period, cf. ΦΣ, ΧΣ, or even by using the letter (/ h/ or / he/ ) with “sigma”, as in this very document, cf. Nαhσιο, Φραhσο (Ναξιō, Φραξō; cf. Schwyzer 1939: 211; Lejeune 1972: 72 f.; 88 f.). There is an ongoing discussion about the reading of the last letter. Alternatively, it also recognizes [ν]. This variant could then be plausibly emended to νῦν, in the plausible sense of ‘[now] wife of Phraxos’. The document yields typical Ionic features as e.g. */ a: / > / e: / , with graphic realization <H> (also after / e/ , / i/ , / r/ ): μητηρ, δημος, ιητρος (cf. ιατρος), πρησσειν (cf. πραττειν); the same phonetical realization as etymological */ e: / (in script E, H); Naxos, Keos different orthographies in archaic material, probably phonetical difference till 5 th cent. BC: H, = [ę] < * / a: / and E = [ē] < */ e: / , cf. κασιγνEτ = κασιγνητη; ανεθΕκεν κ βολοι (= hεκη-) = ἀνεθήκεν ἑκηβόλῳ. 19 Including the typical Epic epithets used while speaking of Artemis, ἑκηβόλος ‘at‐ taining his/ hers aim’ (originally of Apollo), or ἰοχέαιρα ‘arrow-pourer, shooter of arrows’ (of Artemis). The form κασιγνήτη *[kasigne: ta: ] < *km̥ ti-gn̥ h 1 -téh 2 - ‘born with’ seems interesting as it occurs once in Ilias, and outside the Epic it has been attested only in Cypriot (ka-si-ke-ne-to, -ta-, in prosaic texts) and in Aeolic (Lesbian, Thessalian κατιγνειτος); cf. Att. κάσις, -εως ‘brother’ next to ἄδελφος < *sm̥ -g u̯ elb h - ‘from the same womb’. important examples of the archaic use of ἄλοχος is certainly the so-called dedication of Nicandre (of Naxos) found on Delos (ID 2; IG XII, 5, 2 p. XXIV), dated ca. 650 BC, cf. Νικανδρη μ’ανεθεκεν κ βολοι ιοχεαιρ ι | ορ Δεινοδικο το Να ιο ε σοχος αλ(λ)ον | Δεινομενεος δε κασιγνετ Φ ρα ο δ’ αλοχος μ[ην? ]. 18 The document is a dedication of a stele to the goddess Artemis, written in three dactylic hexameter lines, therefore one may assume some degree of Homeric convention applied to the text (similarly as in the case of the Nestor Cup). 19 Clearly observed dialectal features, on the other hand, may actually speak against the assumption made by Ruijgh as far as the distribution of the form ἄλοχος is concerned, testifying to its presence also in Ionic. It seems, however, that the Epic character of the composition is guaranteed Γλῶσσαι κατὰ πόλεις 127 <?page no="128"?> by formulas like ἔξοχος ἄλλων ‘preeminent among others’, which, in a similar manner as in Homeric verse (cf. the acc. ἔξοχον ἄλλων) occupies the verse-final position. The forms ἑκηβόλωι and ἰοχέαιρηι are in their normal formulaic slots. As pointed out by Miller (2013, 102), the switch in traditional language is provided by the transfer of ἑκηβόλος ‘free-shooter’ from Apollo to Artemis, whose epithet is always and exclusively ἰοχέαιρα (mostly verse-final). The situation with Cypriot testimonies is different, though. The sole evidence for the epigraphic use of the noun ἄλοχος comes from an official context, an honorific epitaph of Dionysia, erected by her husband, Tryphon, on a limestone monument in the shape of a kalathos, cf. Palaipaphos (Kouklia): καλάθου μείμημα τρόπαιον ἀνέστησα Τρύφων ἀλόχου Διονυσίας εὐνοίας καὶ σωφροσύνης ἕνεκεν. The document has been written in prose, dates from the 2 nd cent. AD (150-200 AD more precisely, cf. Thompson 1988: 38) and should be analysed as a conventional poetic element used by the Koine Greek in such sorts of inscriptions (cf. the examples quoted above). The analysis of the dialectal documents written in the Cypriot syllabary indicates that the epichoric form used in Cyprus in the meaning ‘wife, spouse’ was ku-na [guna: ], cf. e.g. ICS 88a., funerary stele, 4 th cent. BC. (1) ti-ke-re-to-ti-mo (2) ke-re-te-o-ku-na (Dikreto-(? ) Ti-mokreteo gunā) ‘Dikretô(? ), wife of Timokretes’ (the attested forms of the nom. ku-na, gen. ku-na-i-ko-se, dat. ku-na-i-ki, and maybe in compound ku-na-? -ti-ri-pa-si-? -? , Egetmeyer 1992: 76), paired with po-si-se [posis] ‘husband’ (Egetmeyer 2010: 245 f., 275). It seems that the form ἄλοχος should likewise be considered a typical Homeric element, which of course may be attested in some Greek dialects as well. The distribution of the forms, however, occurring mostly in the poetic, highly conventional context (funerary, or dedicatory, metric verses) points towards the association of the forms with the poetic register. The sole attested example of ἄλοχος in Cypriot documents should fall into the latter category and certainly does not belong to the dialectal vocabulary which, according to the testimony of the dialectal inscription written in the 128 Wojciech Sowa <?page no="129"?> 20 A similar situation may be observed in the case of another member of the word-family λεχ-, namely λέκτρον, which in the γλῶσσαι κατὰ πόλεις has been qualified as Arkadian element, cf. Ἀρκάδων· λέκτρον· κλίνη, but this seems to be attested in poetic inscription everywhere except for Arcadian dialectal region. 21 In spite of the generally negative perception both of Macedonia as a political forma‐ tion and of its social institutions and customs, in Greek historiography and tradition, there was also a feeling of a close relationship between the two nations. Macedo‐ nians were not, it seems, treated as non-Greek foreigners but rather described as original inhabitants of the same general areas as other Greek tribes. For instance, Herodotus calls the Macedonians (Μακεδόνες) a Dorian tribe, living in the area around Mt. Pindos. According to another legend, Perdiccas, the mythical founder of the Macedonian state, came to the northern regions of Greece from the Peloponnesus; elsewhere, as e.g. in Hesiodus’ Catalogue of Women, Macedon and Magnes, brothers from Thessaly, are said to be the cousins of the sons of Hellen - Dorus, Xuthus and Aeolus - to whom the Dorian, Ionic and Aeolic dialects were traditionally attributed, cf. Hes. Catalogue, Fr. 7, cf. Constantinus Porphyrogenn. De them. 2 (p.-86f. Pertusi): ἣ δ’ ὑποκυσαμένη Διὶ γείνατο τερπικεραύνωι υἷε δύω, Μάγνητα Μακηδόνα θ’ ἱππιοχάρμην, οἳ περὶ Πιερίην καὶ λυμπον δώματ’ ἔναιον. 22 For Mimnermos, the name of the land of the Golden Fleece is Αἶα, but by the time of Herodotus, the mythical land was already firmly associated with Colchis (e.g. 7, 193, 2 ἐς Αἶαν τὴν Κολχίδα). As pointed out by West (2007: 198) it may well be that originally syllabary, should be ku-na. The true reasons of ascribing the noun ἄλοχος to Cypriot are not straightforward and are still unclear. 20 6 Μαγνήτων· αἶα· γῆ In the list, the Compiler of the γλῶσσαι κατὰ πόλεις quotes the form αἶα in the meaning ‘earth’ with an indication that the noun should be typical for the inhabitants of Magnesia, in East Thessaly, between Mount Ossa and Pelion. The area plays a role in ancient Greek mythological genealogy, especially providing the link between the Greek tribes and the Ancient Macedonians. 21 The Magnesians are also included in the Greek Army besieging Troy, along‐ side the other tribes living in the regions of what would later become Thrace. The form αἶα is a very interesting case from the philological perspective, as it is a specific Homeric word, used instead of γαῖα. αἶα does not seem to have any clear cognates, it builds no derivatives; and there is no evidence for its existence prior to Homer, whereas γαῖα is quite the opposite. As a personal name in Greek mythology, Αἶα seems to have been the original name of Colchis (‘Land of Dawn’, as in German expression ‘Morgenland’, cf. West 2007: 193-198) 22 and a basis for a derivate Αἰαία, the name of the Γλῶσσαι κατὰ πόλεις 129 <?page no="130"?> Αἶα could also designate other mythical places, as e.g. Okeānos placed in the far east, cf. e.g. Θ 1 ᾿Ηὼς μὲν κροκόπεπλος ἐκίδνατο πᾶσαν ἐπ’ αἶαν which could then mean ‘spread over the whole earth’s eastern rim’. The original meaning must have been forgotten over time and the phrase has been subsequently reinterpreted and used in the simplified meaning ‘over the whole earth’ and applied in other contexts as well, for example as a convenient metrical alternative for γαῖα (cf. also Lesky 1966: 26-62). 23 Nom. sg. 1x in Ilias and 1x in Odyssey, acc. sg. 6x in Ilias and 1x in Odyssey, the opposite situation in case of the gen. sg., which is attested 7x in Ilias and 12x in Odyssey. 24 ῝Ως φάτο, τοὺς δ’ ἤδη κάτεχεν φυσίζοος αἶα / ἐν Λακεδαίμονι αὖθι φίλῃ ἐν πατρίδι γαίῃ. 25 “The idea of the fertile earth being also the repository of the dead has its own ironical or even hopeful paradoxicality”, cf. Kirk (1985: 300). For the interpretation of the epithet in the context, see Floyd (1989: 337-349). 26 Willi (2008: 170) provides a good overview of the existing etymologies. island, house of the sorcerer Circe. Both forms γαῖα and αἶα correspond to each other in the Epic texts, and they remain in some sort of textual relationship in that they might occur instead of one another, but they may also appear in the same verse where they are not purely synonymous. The distribution is striking, as the αἶα in Homeric Epic is attested only 28 times, 23 the occurrences are strictly limited to the verse end. γαῖα, on the other hand, is attested more than 300 times in both Ilias and Odyssey in various positions in verse. As indicated above, there are also cases where the two forms αἶα and γαῖα stay in the same passages, where a clear difference between φυσίζοος αἶα ‘life-giving earth’ and ἐν πατρίδι γαίῃ ‘native land’, cf. Γ 243, 24 may be noted, the latter with a reference of the land of the provenance (of deceased person), hence ‘homeland’, and the former with an emphasis on earth as an ultimate place of dwelling after death, and at the same time, being the source of life. 25 The form is difficult to interpret, plenty hypotheses try to analyse it and combine three nouns αἶα, γαῖα, and μαἶα (‘(foster)mother’) under the concept of the IE ‘Mother Earth’ - which would be then somehow reflected in the assumed connection between αἶα and Lat. avia ‘grandmother’ (cf. Beekes 2010: 130 f.). 26 The other suggestions emphasize the mythical status of Αἶα, as the ‘Land of Dawn’, deriving it from the IE word for ‘dawn’ *h 2 éu̯ sōs, Gr. Hom. Ἠώς (Aeol. Αὔως, Dor. Ἀώς), Lat. Aurōra, Ved. Uṣás-, etc., and seeing in Αἶα reflex of the original female *ih 2 -formation with a zero grade of the suffix, *h 2 éu̯ s-s-ih 2 , which would develop to the Greek *āu̯ hi̯ a, *hāu̯ i̯ a, hāi̯ i̯ a and then, Αἷα, or in a psilotic variant Αἶα (West 2007: 197). 130 Wojciech Sowa <?page no="131"?> 27 Αἶα, πόλις Κόλχων, κτίσμα Αἰήτου, θαλάσσης ἀπέχουσα στάδια τριακόσια, ἣν περιρρέουσι δύο ποταμοί, ῞Ιππος καὶ Κυάνεος, ποιοῦντες αὐτὴν χερρόνησον. ἔοικε δὲ τοὐναντίον ἐξ αὐτῆς ὁ Αἰήτης εἶναι, ὡς ’Ασία ’Ασιάτης, οὕτως Αἶα Αἰάτης καὶ Αἰήτης. ἔστι δὲ καὶ Θετταλίας ἄλλη, ἧς μέμνηται Σοφοκλῆς, τῆς μὲν προτέρας λέγων „εἰς Αἶαν πλέων” τῆς δὲ δευτέρας οὕτως ‘ἔστιν τις Αἶα Θεσσαλῶν παγκληρία’. ἔστιν Αἶα καὶ Μακεδονίας πηγή, ὡς ’Αντίμαχος ἐν Θηβαίδι. τὸ ἐθνικὸν Αἰαῖος ὡς Γαζαῖος, τὸ θηλυκὸν Αἰαία […]. There have been attempts made, trying to consider γαῖα a sort of a con‐ tamination of γῆ and αἶα ‘even though such a derivation means reversing the apparent historical priority of γαῖα over both γῆ and αἶα’ (Haslam 1976: 207), or even to see in αἶα a result of a false segmentation of γαῖα as γ’ αἶα (Brugmann 1903/ 4: 93), which would agree with the ancient tradition, according to which the noun is a result of a hyphaeresis, cf. e.g. Herod. De pros. Cath. 3.1 271.16-21 (= Steph. Byz. Ethn. 36.16-37.4). 27 Τὰ διὰ τοῦ αια καὶ υια δισύλλαβα προπερισπᾶται, μαῖα· ἔστι καὶ πόλις ῾Ελλησποντία, γαῖα ἀπὸ τοῦ γῆ, αἶα, ὅπερ ἀπὸ τοῦ γαῖα γέγονε κατὰ ἀφαίρεσιν τοῦ γ, ἔστι δὲ καὶ Αἶα πόλις Κόλχων· ἔστι δὲ καὶ Θετταλίας ἄλλη, ἧς μέμνηται Σοφοκλῆς, τῆς προτέρας λέγων ‘εἰς Αἶαν πλέων’, τῆς δὲ δευτέρας οὕτως ‘ἔστιν τις Αἶα Θεσσαλῶν παγκληρία’. ἔστιν Αἶα καὶ Μακεδονίας πηγή, ὡς ’Αντίμαχος ἐν Θηβαΐδι, or Hesychius in the same manner, however with reference to Homer, cf. A 1654 αἶα·γῆ (Γ 243) γαῖα κατ’ ἀφαίρεσιν τοῦ γ. καί, ὥς τινες, κρήνη ἐν Παφλαγονίᾳ (Β 850). According to Haslam, the ultimate source for the substantive noun αἶα is to be found in the metrical needs. αἶα is confined to the final foot, unlike γαῖα, which occurs mostly in non-final position (even if the final one would be the most favoured position for γαῖα), and where γαῖα and αἶα are metrically indifferent, γαῖα is invariably the preferred option (Haslam 1976: 209). Furthermore, Haslam points to the expression πατρίδα γαῖαν ‘homeland’ in accusative which would scan , but in nominative expected πατρὶς γαῖα , has been transformed into the more convenient πατρὶς ἄρουρα . Τhe same situation could occur in genitive, where πατρίδος γαίης would be unfitting in a verse (cf. however, one example of the reverse order […] γαίης […] πατρίδος, in κ 50 κλαίοντας, γαίης ἄπο πατρίδος. αὐτὰρ ἐγώ γε) and in such a moment, as Haslam argues, the expression πατρίδος αἴης could come to being as a specific, position-limited variant. This, as a consequence, could lead to the false perception of αἶα as an independent form (Haslam Γλῶσσαι κατὰ πόλεις 131 <?page no="132"?> 28 In Homeric text, all 22 occurrences of πατρίδος are followed by a word with an initial vowel. 29 Cf. Fr. 914 εἰς Αἶαν πλέων and Fr. 915 ἔστιν τις αἶα Θεσσαλῶν παγκληρία (Radt 1999: 581), cf. however Maas 1907: 521 2 , according to whom, the expression αἶα Θεσσαλῶν παγκληρία ‘the land all Thessalians own together’ should actually be understood simply as ‘Thessaly’. 1976: 209 f.), 28 which could “pervade Greek poetry […] especially choral lyric, in the epic tradition” (Silk 2010: 427). It seems that if one accepts the artificial origins of αἶα as only a position variant of γαῖα, which has been borrowed into Greek from Epic poetry, the degree of difficulty in understanding the true reasons behind including it in γλῶσσαι κατὰ πόλεις and ascribing it to a concrete region of Greece increases. It seems that, except for the mythical name of the Colchis, a similar toponym existed in the Mainland Greece, in the form of Αἶα or Αἰαία, a town in Malis in Thessaly, cf. above-mentioned statement of Herodianus, with reference to Sophokles, Αἶα πόλις Κόλχων ἔστι δὲ καὶ Θετταλίας ἄλλη, ἧς μέμνηται Σοφοκλῆς. 29 This could, in principle, point to the existence of the real form. Such a presumption becomes a starting point for Peters, who treats the transmitted entry Μαγνήτων· αἶα· γῆ as a real element of Greek vocabulary. Peters seeks evidence for αἶα in Mycenaean forms as a 3 -wa (probably a name of an oxen, cf. KN C973, as if *ai̯ u̯ āns, cf. Αἴας), a 3 -wa-ja (female name in PY EN 74.22, *ai̯ u̯ ai̯ ā, maybe corresponding to Greek Αἰαία), a 3 -wa-ta (KN Vc7612 without context, maybe *ai̯ u̯ atās, cf. Αἰήτης; cf. DM I 141 for possible interpretations and literature), which could attest the archaic stage in the development of αἶα, namely *ai̯ u̯ a from Common-Greek *hai̯ u̯ ă with ‘already Proto-Aeolic psilosis’ and a short vowel, due to the analogy to the synonymous γαῖα (Peters 1994: 205 f. and 210). In such an interpretation, a proto-form *hai̯ u̯ ā would be understood as an original verbal adjective of the type OInd. pakvá- ‘cooked, ready’, Gr. νεώς/ ναός < *nasu̯ ό- ‘inhabited’ from an IE formation *sh 2 ai̯ u̯ o- ‘wet, moisten’ with a root to be found in the Greek noun αἷμα (*sh 2 ai̯ -), which could be a semantic parallel to γαῖα, γᾶ < *gh 2 n̥ s-i̯ a ‘provided with light liquid’, i.e. ‘inseminated by Ouranos’ (Peters 1994: 209, with parallels for the metaphoric names for ‘sperm’, as λευκὸν μένος , or Ἄγλαυρος). Interestingly Peters connects the noun αἶα with a name of the god of the Netherworld, i.e., Hades, whose name in this scenario would yield a derivate *hai̯ u̯ id, *hai̯ i̯ id or even *(h)ai̯ hid and would stem 132 Wojciech Sowa <?page no="133"?> from the idea of the lower class of the IE society, which, unlike in the case of the warrior-aristocracy had not practised cremation but rather burials (Peters 1994: 206 9 ). This direction has been also followed by Janda, who also links a theonym Hades with the name of earth αἶα, referring to the other IE verbal root, namely *sh 2 ei̯ - ‘tie, bind’ (cf. LIV 2 544). According to Janda the name of Hades should be analysed as *sah 2 i-u̯ o-/ *sh 2 a-i̯ u̯ o-, which can be found in the Lat. adj. saeuus. The author also quotes evidence from the existence of the homophonic verbal root *sah 2 (-i-) in meaning ‘to bind’ (saeua uincula); analysing the rituals attested in the IE languages, Janda states that “die Fesselung des Toten in dern indogermanischen Sprachen und Kulturen noch vielfach greifbar ist, die Fessel ein charakteristisches At‐ tribut des Totengottes” (Janda 2000: 132, passim 114-138). Thus the original semantic of the name of the IE god of death would be ‘the binding one, bond, chain’ and should be interpreted as a development from ‘binding, chained’, *sah 2 i-u̯ ointo the new formation *sah 2 u̯ id- ‘binding one’, yielding the same proportion as νύκτερος adj. ‘nightly’: νυκτερίς *‘belonging to the night’ > ‘bat’ (Janda 2000: 117, 131). The name of ‘earth’ αἶα on the other hand, would reflect, as suggested by Peters, the same IE *sh 2 ai̯ u̯ o-, in fem. *hai̯ u̯ ā and then *hai̯ u̯ ă (analogical shortening of a vowel) in different meaning - the ‘earth’ would be then rather ‘bound, tied’, the image, still present in Vedic poetry, maybe also in the Greek mythological narrative about fixing the still floating Island of Delos before the birth of Apollo and Artemis (Janda 2000: 133). Leaving aside this interesting suggestion, one could still ask about the cri‐ teria allowing us to consider αἶα a real dialectal form. Peters supports such an opinion, quoting defixio from Thebes, dated from the 2 nd cent. BC, where in part B one finds the sequence Μὴ κατ’ αἶαν μηδὲ κατὰ θάλατταν (Audollent 1904: 135). The inscription is, however, quite recent; no real traces of the dialect can be observed, and the documents express mainly traces of the Koine Greek (“ut tabella nostra recentioris aetatis est dialecti vere boeoticae pauca restant vestigia, plera e lingua vulgari sumpta”, Audollent 1904: 136). The quoted sentence seems to have been used in a formulaic way, typical for magic spells (‘the curses of earth and sea’ are widely attested, cf. Strubbe 1997: 293-5), which would be an additional objection against the assumption of the existence of a word αἶα in a vernacular dialect (Boeotian? ) and could probably support the hypothesis of the poetic origins of the gloss. Γλῶσσαι κατὰ πόλεις 133 <?page no="134"?> 7 Concluding remarks As stated above, the Γλῶσσαι κατὰ πόλεις may initially appear to be an unimpressive list of ancient 100 forms, however, it is quite unique and worthy of detailed study. It seems that it may be used as a model in the modern and comprehensive way of the treatment of the dialectal lexicon and has the potential to offer new perspectives in our understanding of the forming scholarly tradition on the auto-reflection of different varieties of language spoken in Greece in ancient times. It could also help to understand the internal relationships between various works (texts) still preserved in manuscript form today. According to Bowra (1960: 49), the anonymous author of the Γλῶσσαι κατὰ πόλεις worked with the material, which could be traced back to Zenodotus, the immediate source being the grammarian Pam‐ phylius (ca. 50 AD). However, until a careful investigation of the preserved text and existing manuscripts is completed, such a statement must remain only a presumption (though it could be the most ancient dialectological tradition attested). According to Latte, at least part of the material had been directly taken from the Hesychius (and Diogenianus), although probably not from the official edition, but rather from a template of the lexicon (Latte 1925: 141), it could also reflect the text of the lexicon still using the ethnic references (Latte 1925: 147), which have been lost in the further version of Hesychius (Marcianus Graecus 622). The folio containing Γλῶσσαι is found in a Codex among various grammatical works of partly anonymous, authors, e.g. Κυρίλλου ἀρχιεπισκόπου ἀλεξανδρείας λεξικὸν κατὰ στοιχείων (Fol. 1), ἐτυμολογία εἰς τά κδ στοιχεῖα (Fol. 145), λεξικόν τῶν ἐνδιαθέτων γραφῶν ἐκτεθὲν παρὰ Στεφάνου καὶ Θεοδορήτου Κασσιανοῦ Λογγίνου φιλοσοφοῦ καὶ ἑτέρων λεξιγράφων (Fol. 225), ῥητορικαί λέξεις (Fol. 257). In folio 147 there is even a special Lexicon of Homeric dialect: λέξεις Ὁμηρικαὶ κατὰ στοιχείον. One could assume the possibility of the intertextual influence from the side of such works. Such a hypothesis seems interesting in light of the Homeric character of the entire list, as demonstrated above, cf. further Ἀκαρνάνων· ἔνεπει· λέγει, attested in Hom., Pi., Bacch., Tragedy, or κῆρ in a sense of ‘ψυχή’, which occurs as such only in Homer, in poetry always κέαρ. The examples are numerous, and the list is actually a collection of Homeric forms, yet it is extremely difficult to say what such a hypothetical influence would look like. The constant mixing of literary forms with real dialectal material (albeit, from our perspective, wrongly interpreted) seems 134 Wojciech Sowa <?page no="135"?> 30 This research was supported by grant BPN/ BEK/ 2021/ 1/ 00104 by NAWA (Polish Agency for Academic Exchange). a very interesting phenomenon, but the question as to what extent such information can be regarded as reliable evidence of the spoken language of different regions of Greece remains unanswered. There is no doubt that Ancient Greek dialectal texts yield heterogenous variants of epichoric idioms: pure dialectal forms often mix with literary (poetic) ones, especially in the funerary context of the metrical inscriptions; obsolete forms are replaced by those recently adopted or more frequent, or forms belonging to a dialect with more prestige at a given chronological stage (e.g. Attic or Koine). Similarly, one may assume that the Greek dialects possessed social varieties connected with various social groups, as in the case of any other language, which could leave traces in the written docu‐ mentation. The epigraphic material does not allow us to uncover all these variants. Therefore, we must deal with secondary evidence in a critical and meaningful way. Our knowledge of the Ancient Greek dialects can come only from written evidence. Here one may observe that, aside from situations where the written message appears to reflect linguistic reality, the written language often has only the most tenuous relationship to the spoken vernacular, or it can even represent the artificial revival - for symbolic reasons - of a dead language or dialect (Brixhe 2007: 489). On the other hand, the grammatical and lexicographical sources are, in most cases, relatively late, and so they also reflect some tradition of understanding the works of Homer as the ultimate source of archaic forms of Greek. They inform us about the epichoric variants of Greek language, mostly in the context of literary works of Epic, Lyric and Tragedy, and we know that, to some extent, the interest in exploring regional forms of Greek also existed. No such works has been transmitted to our times. It seems, however, that by the times of the composition of the lexicon of Diogenianus, Hesychius and Kyrillos, the Ancient Greek dialects must have become quite an abstract notion and a reservoir of strange and obsolete forms. 30 Γλῶσσαι κατὰ πόλεις 135 <?page no="136"?> Bibliography Ancient authors have been quoted according to the electronic editions in the “Thesaurus Linguae Graecae” project prepared by The Packard Human‐ ities Institute. The abbreviations follow the model of LSJ; the Lexicon of Hesychius has been quoted according to the edition of Latte, Kurt (1953- 1966): Hesychii Alexandrini lexicon. 1-2 (A-O). Hauniae: Ejnar Munksgaard editore, and Schmidt, Mauricius (1861-1862): Hesychii Alexandrini Lexicon. 3-4 (Π-Ω). Halle a.d.S.: Max Niemeyer (repr. Amsterdam: Hakkert, 1965). The Lesbian inscriptions follow the standard edition in IG XII,2 = Inscriptiones Graecae, Vol. XII, fasc. 2: Inscriptiones Lesbi, Nesi, Tenedi. Consilio et auctor‐ itate Academiae Litterarum regiae Borussicae, William Paton (ed.). Berlin: Georg Reimer Verlag 1899, their abbreviations and numbers correspond to Hodot 1990. The Cypriot inscriptions follow the standard of the ICS, the other Greek inscriptions have been quoted according to the electronic edition prepared by The Packard Humanities Institute. Audollent, Auguste (1904): Defixionum Tabellae quotquot innotuerunt: tam in Graecis orientis quam in totius occidentis partibus praeter Atticas in Corpore Inscriptionum Atticarum editas. Paris: Albert Fontemoing. Bartoněk, Antonín (2003): Handbuch des mykenischen Griechisch. Heidelberg: Winter. Bechtel, Friedrich (1921): Die griechischen Dialekten. 1. Band. Der lesbische, thessali‐ sche, böotische, arkadische und kyprische Dialekt. Berlin: Weidmann. Beekes, Robert Stephen Paul (2010): Etymological Dictionary of Greek (Leiden Indo-Eu‐ ropean Etymological Dictionary Series 10). Leiden: Brill. Bekker, Immanuel (1821): Anecdota Graeca III. Berlin: Georg Reimer Verlag. Bowra, Cecil Maurice (1960): “Glossai kata poleis”. Glotta 38 (1-2): 43-60. Brixhe, Claude (1976): Le Dialecte Grec de Pamphylie. Paris: Adrien-Maisonneuve. Brixhe, Claude (2007): “A modern approach to the ancient Greek dialects”. In: A History of Ancient Greek: From the Beginnings to Late Antiquity, Anastasios/ Fivos Christidis (ed.). Cambridge: Cambridge University Press, 486-518. Brugmann, Karl (1903/ 1904): “Beiträge zur griechischen, germanischen und slavi‐ schen Wortforschung”. Indogermanische Forschungen 15: 87-104. Campanile, Enrico (1988): “Meaning and pre-history of OIr. lúan láith”. In: Languages and Cultures Studies in Honor of Edgar C. Polomé, Mohammad Ali Jazayery and Werner Winter (eds.). Berlin: De Gruyter Mouton, 89-95. DELG = Chantraine, Pierre: Dictionaire étymologique de la langue grecque. Histoire des mots. Tome I (A-D) 1968, Tome II (E-K) 1970, Tome III (L-P) 1974. Paris: Klincksieck. 136 Wojciech Sowa <?page no="137"?> DM = Diccionario griego-español. Diccionario Micénico (1985-1993): Vol. I-II. Red. por Francisco Aura Jorro. Madrid: Instituto de Filología - Consejo Superior de Investigaciones Científicas. Dubois, Laurent (1986): Recherches sur le dialecte arcadien I. Grammaire, II. Corpus di‐ alectal, III. Notes-index-bibliographie. Louvain-la-Neuve: Bibliothèque des cahiers de l`Institut de Linguistique de Louvain, BCILL 33/ 34/ 35. Egetmeyer, Markus (1992): Wörterbuch zu den Inschriften im kyprischen Syllabar. Unter Berücksichtigung einer Arbeit von Almut Hintze. Berlin/ New York: De Gruyter. (=Kadmos. Zeitschrift für vor- und frühgriechische Epigraphik. Supple‐ mentband 3) Egetmeyer, Markus (2010): Le dialecte grec ancien de Chypre. Berlin/ New York: De Gruyter. Farmini, Luciano (1984): “Osservazioni sulla semantica die verbi apofonetici in Omero”. Annali dell’Università degli Studi di Napoli “L’Orientale”. Rivista del Dipar‐ timento Del mondo classico. Sezione Linguistica 6: 263-283. Filos, Panagiotis (2018): “The dialectal variety of Epirus”. In: Studies in Ancient Greek Dilaects. From Central Greece to the Black Sea, Gregorios K. Giannakis, Emilio Crespo and Panagiotis Filos (eds.). Berlin/ Boston: De Gruyter, 215-247. Floyd, Edwin D. (1989): “Homer and the Life-Producing Earth”. The Classical World 82(5): 337-349. Frisk, GEW = Frisk, Hjalmar (1960): Griechisches Etymologisches Wörterbuch. Heidel‐ berg: Winter. García Ramón, José Luis (1997): “Cuestiones de léxico y onomástica tesalios”. In: Katà diálekton. Atti del III Colloquio Internazionale di Dialettologia Greca. Napoli - Fiaiano d’Ischia, 25-28 settembre 1996, Albio Cassio (ed.). Napoli: A.I.O.N. XIX (1997), 521-552. García Ramón, José Luis. (2004): “Del trabajo en una gramática del tesalio: para una valoración lingüística de las glosas”. In: Dialetti, dialettismi, generi letterari e funzioni sociali. Atti del V Colloquio Internazionale di Linguistica Greca, Milano, 12- 13 settembre 2002, Giovanna Rocca (ed.). Alessandria: Edizioni dell’Orso, 235-264. Hajnal, Ivo (2007): “Die Vorgeschichte der griechischen Dialekte: ein methodischer Rück- und Ausblick”. In: Die altgriechischen Dialekte. Wesen und Werden. Akten des Kolloquiums FU Berlin, 19. -22. September 2001, Ivo Hajnal (ed.). Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft; Bd.-126, 131-156. Haslam, Michael W. (1976): “Homeric Words and Homeric Metre: Two Doublets Examined (λείβω/ εἴβω, γαῖα/ αἶα).” Glotta 54: 201-211. Γλῶσσαι κατὰ πόλεις 137 <?page no="138"?> Heisserer, Andrew J. (1980): Alexander the Great and the Greeks. The Epigraphic Evidence. Norman (Oklahoma): University of Oklahoma Press. Hernández-Vázquez, Asunción (1994): Estudio léxico del jonio minorasiático. Tesis Doctoral. Salamanca. Hiller von Gaertringen, Friedrich (1908): “ΚΛΕΙΦΩΣΣΑ”. Hermes 43(1): 173-176. Hoffmann, Otto (1893): Die Griechischen Dialekte in ihrem historischen Zusammen‐ hange mit den wichtigsten ihrer Quellen. 2. Band. Der nord-achäische Dialekt. Göttingen: Vandenhoeck & Ruprecht. ICS = Masson, Olivier (1983): Les inscriptions chypriotes syllabiques. Recueil critique et commenté [Addenda: 393-406]. Réimpression augmentée [Addenda nova: 407-424]. Paris: de Boccard. Iriarte, Juan de (1769): Regiae bibliothecae matritensis codices graeci mss. Madrid: Antonio Perez de Soto. Janda, Michael (2000): Eleusis. Das indogermanische Erbe der Mysterien. Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft; Bd.-96. Jeffery, Lilian Hamilton (1961): 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. Oxford: Clarendon Press. Kirk, Geoffrey Stephen (1985): The Iliad: A Commentary, Books 1-4. Cambridge: Cambridge University Press. Latte, Kurt (1925): “Glossographika”. Philologus 80: 136-175. Lejeune, Michael (1972): Phonetique historique du mycénien et du grec ancien. Paris: Klincksieck. Lesky, Albin (1966): “Aia”. In: Gesammelte Schriften, Walther Kraus (ed.), Bern/ Mün‐ chen: Francke Verlag, 26-62. Letoublon, Francoise (1985): Il allait, pareil à la nuit. Les verbes de mouvement en grec: supplétisme et aspect verbal. Paris: Klincsieck. LfgrE = Lexikon des frühgriechischen Epos. Im Auftrag der Akademie der Wissen‐ schaften in Göttingen vorbereitet und hrsg. vom Thesaurus Linguae Graecae. Be‐ gründet von Bruno Snell, Bd.-1. Göttingen: Vandenhoeck & Ruprecht 1979. LIV 2 = LIV. Lexikon der indogermanischen Verben: Die Wurzeln und ihre Primärstamm‐ bildungen, 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 Martin Kümmel und Helmut Rix. Wiesbaden: Reichert 2001. LSJ = A Greek-English Lexicon (1996): Henry George Liddell/ Robert Scott (eds.), rev. by Sir Henry Stuart Jones/ Roderick McKenzie. With a revised supplement. Oxford: Clarendon. 138 Wojciech Sowa <?page no="139"?> Maas, Ernst (1907): “Telina. Kaikina”. Zeitschrift für vergleichende Sprachforschung auf dem Gebiete der indogermanischen Sprachen 40: 520-533. Masson, Olivier (1979): “Cretica”. Bulletin de Correspondance Hellénique 103: 57-82. Méndez Dosuna, Julián (1985): Los dialectos dorios del Norosete. Gramática y estudio dialectal. Salamanca: Ediciones de la Universidad de Salamanca. Miller, D. Gary (2013): Ancient Greek Dialects and Early Authors: Introduction to the Dialect Mixture in Homer, with Notes on Lyric and Herodotus. Boston/ Berlin: De Gruyter. Morpurgo-Davies, Anna (1987): “The Greek notion of dialect”. verbum 10: 7-28. (Actes de la première recontre internationale de dialectologie grecque. Nancy/ Pont-à-Mousson 1-3 juillet 1986). Peters, Martin (1986): “Zur Frage einer „achäischen“ Phase des griechischen Epos”. In: o-o-pe-ro-si. Festschrift für Ernst Risch zum 75. Geburtstag, Annemarie Etter (ed.). Berlin/ New York: Walter de Gruyter, 304-319. Peters, Martin (1980): Untersuchungen zur Vertretung der Indogermanischen Lar‐ yngale im Griechischen. Wien: Verlag der Österreichischen Akademie der Wissen‐ schaften. Peters, Martin (1993): “Ein weiterer Fall für Rixsche Gesetz”. In: Indogermanica et Italica. Festschrift für Helmut Rix zum 65. Geburtstag, Gerhard Meiser (ed.). Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft; Bd.-72., 373-405. Peters, Martin (1994): “Griech. γῆ, γαῖα, armen. erkir ‘Erde’”. In: Iranian and Indo-Eu‐ ropean Studies. Memorial Volume of Otakar Klíma, Petr Vavroušek (ed.). Praha: Enigma corporation, 203-213. Radt, Stefan (1999): Tragicorum Graecorum Fragmenta. Vol. 4.: Sophocles. Göttingen: Vandenhoeck & Ruprecht. Risch, Ernst ( 2 1974): Wortbildung der homerischen Sprache. Berlin/ New York: Walter de Gruyter. Ruijgh, Cornelius Jord (1957): L’élément achéen dans la langue epique. Assen: van Gorcum. Schwyzer, Eduard (1939): Griechische Grammatik. Erster Band. München: C. H. Beck’sche Verlagsbuchhandlung. Silk, Michael S. (2010). “The Language of Greek Lyric Poetry”. In: A Companion to the Ancient Greek Language, Egbert J. Bakker (ed.). Malden, MA: Wiley-Blackwell, 424-40. Sowa, Wojciech (2006): “Bemerkungen zum lesbischen dialektalen Wortschatz”. Emerita. Revista de Lingüística y Filología Clásica 74(2): 233-258. Γλῶσσαι κατὰ πόλεις 139 <?page no="140"?> Sowa, Wojciech (2011): “Griechische Dialekte, dialektale Glossen und die antike lexikographische Tradition”. Glotta 87: 159-183. Strubbe, Johan H. M. (1997): Arai Epitymbioi: Imprecations Against Desecrators of the Grave in the Greek Epitaphs of Asia Minor: a Catalogue. Bonn: Dr. Rudolf Habelt Verlag. Thompson, Alexandra Anne (1988): “An Inscription on a kalathos Shaped Monument from Kouklia in Cyprus”. Studies in Greek Linguistics 9: 37-41. West, Martin Litchfield (2007): “Phasis and Aia”. Museum Helveticum 64(4): 193-198. Willi, Andreas (2008): “Demeter, Gê, and the Indo-European word(s) for ‘earth’”. Historische Sprachforschung/ Historical Linguistics 120: 169-194. WOU = Untermann, Jürgen (2000): Wörterbuch des Oskisch-Umbrischen (Handbuch der italischen Dialekte; Bd.-3; Indogermanische Bibliothek: Reihe 1., Lehr- und Handbücher). Heidelberg: Carl Winter Universitätsverlag. Zeitlin, Froma I. (2018): “Constructing the aesthetic body in Homer and beyond”. In: Thinking the Greeks. A Volume in Honor of James M. Redfield, Bruce M. King/ Lillian Doherty (eds.). London/ New York: Routledge 53-69. 140 Wojciech Sowa <?page no="141"?> Greek γέφῡρα ‘dam; bridge’ Václav Blažek Abstract: In the present contribution, the question of the origin of the Greek word γέφῡρα ‘dam; bridge’ with its dialect variants is solved. In the first two parts, a survey of textual evidence and glosses is pre‐ sented. The existing etymological attempts for the internal (Indo-Euro‐ pean) etymology and borrowings from external (non-Indo-European) sources are discussed and evaluated in the third part. In the fourth and final part, the most promising solutions are chosen. Keywords: Indo-European, Greek, substratum, heritage, borrowing 0. The Greek word γέφῡρα remains without any convincing etymology. The present study summarizes its main semantic and phonetic variants. After the basic documentation, the survey of existing etymologies follows, including brief comments. Finally, the most promising solutions are dis‐ cussed. 1. The Greek word γέφῡρα fem. was used in several meanings. 1.1 Homer (8 th cent. BCE? ) used γέφῡρα in the meaning ‘dyke, dam, embankment’ and metaphorically ‘dykes of war’. 1.1.1 The meaning ‘dyke, dam, embankment’ appears in the passage, where Diomedes (son of Tydeus) on the battlefield in front of Troy is compared to the swift flood, destroying dykes: Hom. Il. 5.84-94 ὣς οἳ μὲν πονέοντο κατὰ κρατερὴν ὑσμίνην: Τυδεΐδην δ’ οὐκ ἂν γνοίης ποτέροισι μετείη ἠὲ μετὰ Τρώεσσιν ὁμιλέοι ἦ μετ’ Ἀχαιοῖς. θῦνε γὰρ ἂμ πεδίον ποταμῷ πλήθοντι ἐοικὼς χειμάρρῳ, ὅς τ’ ὦκα ῥέων ἐκέδασσε γεφύρας: τὸν δ’ οὔτ’ ἄρ τε γέφυραι ἐεργμέναι ἰσχανόωσιν, <?page no="142"?> οὔτ’ ἄρα ἕρκεα ἴσχει ἀλωάων ἐριθηλέων ἐλθόντ’ ἐξαπίνης ὅτ’ ἐπιβρίσῃ Διὸς ὄμβρος: πολλὰ δ’ ὑπ’ αὐτοῦ ἔργα κατήριπε κάλ’ αἰζηῶν: ὣς ὑπὸ Τυδεΐδῃ πυκιναὶ κλονέοντο φάλαγγες Τρώων, οὐδ’ ἄρα μιν μίμνον πολέες περ ἐόντες. “Thus toiled they in the mighty conflict; but of Tydeus’ son couldst thou not have told with which host of the twain he was joined, whether it was with the Trojans that he had fellowship or with the Achaeans. For he stormed across the plain like unto a winter torrent at the full, that with its swift flood sweeps away the embankments; this the close-fenced embankments hold not back, neither do the walls of the fruitful vineyards stay its sudden coming when the rain of Zeus driveth it on; and before it in multitudes the fair works of men fall in ruin. Even in such wise before Tydeus’ son were the thick battalions of the Trojans driven in rout, nor might they abide him for all they were so many.” (transl. by Murray 1924) 1.1.2 More frequently the word γέφῡρα appeared in the Iliad in the meta‐ phoric formula ‘dykes of war’, e.g.: Hom. Il. 20.426-427 οὐδ’ ἂν ἔτι δὴν/ ἀλλήλους πτώσσοιμεν ἀνὰ πτολέμοιο γεφύρας. “Not for long shall we any more shrink one from the other along the dykes of war.” (transl. by Murray 1924) 1.2 Pindar (522/ 518-448/ 438 BCE) used the term γέφυρα for the Isthmus of Corinth: Pind. Nem. 6.40-42 πόντου τε γέφυρ’ ἀκάμαντος ἐν ἀμφικτιόνων ταυροφόνῳ τριετηρίδι Κρεοντίδαν τίμασε Ποσειδάνιον ἂν τέμενος· “And the bridge of the untiring sea honored Creontidas in the biennial festival of those who live around, when bulls are slain in the sacred precinct of Poseidon.” (transl. by Svarlien 1990) Pind. Isthm. 4.17-23 νῦν δ’ αὖ μετὰ χειμέριον ποικίλων μηνῶν ζόφον χθὼν ὥτε φοινικέοισιν ἄνθησεν ῥόδοις δαιμόνων βουλαῖς. ὁ κινητὴρ δὲ γᾶς Ὀγχηστὸν οἰκέων 142 Václav Blažek <?page no="143"?> καὶ γέφυραν ποντιάδα πρὸ Κορίνθου τειχέων, τόνδε πορὼν γενεᾷ θαυμαστὸν ὕμνον ἐκ λεχέων ἀνάγει φάμαν παλαιὰν εὐκλέων ἔργων “Now, after the wintry gloom of the changing months, the ground has blossomed as if with crimson roses by the will of the gods. The shaker of the earth who dwells at Onchestus and at the sea-bridge before the walls of Corinth, by offering to that family this marvellous song, wakes from her bed their ancient fame for glorious deeds.” (transl. by Svarlien 1990) 1.3 The meaning ‘bridge’ was introduced into literary texts only in the 5 th cent. BCE: 1.3.1 Aeschylus (525/ 524-456/ 455 BCE) Aesch. Pers. 735-738 Δαρεῖος: πῶς τε δὴ καὶ ποῖ τελευτᾶν; ἔστι τις σωτηρία; Ἄτοσσα: ἄσμενον μολεῖν γέφυραν γαῖν δυοῖν ζευκτηρίαν. Δαρεῖος: καὶ πρὸς ἤπειρον σεσῶσθαι τήνδε, τοῦτ’ ἐτήτυμον; Ἄτοσσα: λόγος κρατεῖ σαφηνὴς τοῦτό γ’, οὐδ’ ἔνι στάσις. Darius: “Met his end, how, tell me, and where? Of his safety is there any hope? ” Atossa: “To his joy he reached the bridge yoking the two continents.” Darius: “And reached our continent in safety? Is this certain? ” Atossa: “Yes, a proven report establishes this. Doubt there is none.” (transl. by Smyth 1926) 1.3.2 Herodotus (484-425 BCE) Hdt. 1.75 ὡς δὲ ἀπίκετο ἐπὶ τὸν Ἅλυν ποταμὸν ὁ Κροῖσος, τὸ ἐνθεῦτεν, ὡς μὲν ἐγὼ λέγω, κατὰ τὰς ἐούσας γεφύρας διεβίβασε τὸν στρατόν, ὡς δὲ ὁ πολλὸς λόγος Ἑλλήνων, Θαλῆς οἱ ὁ Μιλήσιος διεβίβασε. ἀπορέοντος γὰρ Κροίσου ὅκως οἱ διαβήσεται τὸν ποταμὸν ὁ στρατός (οὐ γὰρ δὴ εἶναι κω τοῦτον τὸν χρόνον τὰς γεφύρας ταύτας) λέγεται παρεόντα τὸν Θαλῆν ἐν τῷ στρατοπέδῳ ποιῆσαι αὐτῷ τὸν ποταμὸν ἐξ ἀριστερῆς χειρὸς ῥέοντα τοῦ στρατοῦ καὶ ἐκ δεξιῆς ῥέειν, ποιῆσαι δὲ ὧδε: Greek γέφῡρα ‘dam; bridge’ 143 <?page no="144"?> “When he came to the river Halys, he transported his army across it - by the bridges which were there then, as I maintain; but the general belief of the Greeks is that Thales of Miletus got the army across. The story is that, as Croesus did not know how his army could pass the river (as the aforesaid bridges did not yet exist then), Thales, who was in the encampment, made the river, which flowed on the left of the army, also flow on the right, in the following way.” (transl. by Godley 1920) Hdt. 4.97 Δαρεῖος δὲ ὡς ἀπίκετο καὶ ὁ πεζὸς ἅμ’ αὐτῷ στρατὸς ἐπὶ τὸν Ἴστρον, ἐνθαῦτα διαβάντων πάντων Δαρεῖος ἐκέλευσε τούς τε Ἴωνας τὴν σχεδίην λύσαντας ἕπεσθαι κατ’ ἤπειρον ἑωυτῷ καὶ τὸν ἐκ τῶν νεῶν στρατόν. μελλόντων δὲ τῶν Ἰώνων λύειν καὶ ποιέειν τὰ κελευόμενα, Κώης ὁ Ἐρξάνδρου στρατηγὸς ἐὼν Μυτιληναίων ἔλεξε Δαρείῳ τάδε, πυθόμενος πρότερον εἰ οἱ φίλον εἴη γνώμην ἀποδέκεσθαι παρὰ τοῦ βουλομένου ἀποδείκνυσθαι. ὦ βασιλεῦ, ἐπὶ γῆν γὰρ μέλλεις στρατεύεσθαι τῆς οὔτε ἀρηρομένον φανήσεται οὐδὲν οὔτε πόλις οἰκεομένη: σύ νυν γέφυραν ταύτην ἔα κατὰ χώρην ἑστάναι, φυλάκους αὐτῆς λιπὼν τούτους οἵπερ μιν ἔζευξαν. “When Darius and the land army with him had come to the Ister, and all had crossed, he had the Ionians break the bridge and follow him in his march across the mainland, together with the men of the fleet. So the Ionians were preparing to break the bridge and do Darius’ bidding; but Cöes son of Erxander, the general of the Mytilenaeans, after first asking if Darius were willing to listen to advice from one who wanted to give it, said, ‘Since, O King, you are about to march against a country where you will not find tilled lands or inhabited cities, let this bridge stay where it is, leaving those who made it to guard it.’” (transl. by Godley 1920) 2. Besides these semantic variants of the only form, γέφῡρα, there were recorded several dialect variants differing especially in the initial: 2.1 Boeotian βέφυρα Strattis (end of the 5th - beginning of the 4th cent. BCE), Fragmenta 47.5 = Phoenician Women 49 (apud Athenaeus 621 f): ξυνίετ’ οὐδὲν πᾶσα Θηβαίων πόλις. / οὐδέν ποτ’ ἄλλ’. οἳ πρῶτα μὲν τὴν σηπίαν / ὀπιτθοτίλαν, ὡς λέγουσ’, ὀνομάζετε, / τὸν ἀλεκτρυόνα δ’ ὀρτάλιχον, τὸν ἰατρὸν δὲ / σάκταν, β(λ)έφυραν τὴν γέφυραν, τῦκα δὲ / τὰ σῦκα, κωτιλάδας δὲ τὰς χελιδόνας, / τὴν ἔνθεσιν δ’ ἄκολον, τὸ γελᾶν δὲ κριδδέμεν, / νεασπάτωτον δ’ ἤν τι νεοκάττυτον ᾖ. 144 Václav Blažek <?page no="145"?> 1 Kuhn (1852: 126-28) also added Vedic jámbha- ‘tooth, tusk; (pl.) set of teeth, mouth, jaws’, (VS) jámbhya- ‘inciser, grinder’ and Greek (Hom. Il. 16.489, 19.394) γαμφηλαί pl. ‘jaws of animals’, (Hdt. 9.83) γομφίος ‘grinder-tooth’. Curtius (1866: 159) accepted these ‘tooth/ jaw’-comparanda and further added Old Saxon camb ‘comb’ and Old Church Slavonic zǫbъ ‘tooth’. But with respect to Slavic zthese forms reflect IE *ĝ-, which is incompatible with *g (h) in Vedic ga(m)bhīrá- ‘deep’, not to mention the semantic differences. “You the entire city of Thebes, understand nothing, nothing at all. First, as they tell me, you call the cuttlefish a ‘rear-squirter’, the rooster a ‘chicken’, the doctor a ‘bagman’; for ‘bridge’ you say ‘gridge’, ‘tigs’ for ‘figs’, and for swallows ‘twitterers’. You call a morsel of food ‘a bite’, and for to laugh you say ‘to scratch’. If something is newly stitched, it is ‘new-leathered’.” (transl. by Storey 2011) 2.2 Doric forms 2.2.1 Cretan δέφυρα appears in the following inscription from Gortyn (Collitz/ Bechtel 1905: 5000 iib 6): τὰν δὲ ῥoὰν λεί[π]/ / εν ὄττον κατέχει α’ ἐπ’ ἀγορᾶ/ ι δέφυρα ἢ πλίον, μεῖον δὲ μή 2.2.2 Laconian gloss δίφουρα· γέφυρα. Λάκωνες (Hesychius). 2.3 Gloss of unknown affiliation, probably Northwest Greek (Locrian? ), in acc.pl. βουφάρας· γεφύρας (Hesychius). 3. Survey of etymological attempts (in chronological order) 3.1 The first attempt to explain Greek γέφῡρα ‘bridge; dam’ appeared already in ancient times. Herodian (c. 170-c. 240 CE) apud Eustathius 1235.23: γέφῡρα = γῆ ἐφ’ ὑγρᾶς. Cf. also Etymologicum Magnum 229.9-11: Γέφυρα […] παρὰ τὸ γῆ καὶ τὸ φυρῶ τὸ ἀναμιγνύω, γή-φυρά τις οὖσα ἡ τῇ γῇ φυραθεῖσα καὶ ἐφ’ ὑγρῷ τόπῳ κειμένη. 3.2 Kuhn (1852: 135) saw cognates of Greek γέφῡρα ‘bridge; dam’ in Vedic gab h īrá-/ gamb h īrá- ‘deep’, (VS) gámb h an- ‘depth’ 1 , plus Dardic: Pashai gambā́ ‘deep’< *gamb h aka-, Shina gabū́n ‘foot of mountain, bottom, tree-trunk’ (adopted into Burushaski gʌmūn ‘bottom’) < *gamb h ūna- (Turner 1966, n. 4029). The r-derivatives should be supplemented by the Iranian data: Young Avestan ǰafra- ‘deep’, Pashto žawə́ r, Wanetsi žawár ‘deep’, Parthian jfr / žafr/ ‘deep, low; depth, abyss’, Middle Persian zw(w)pr, Pahlavi zwpl / zufr/ ‘depth’, Persian žarf ‘deep’, Kurdic žiōr, Baluchi ǰahl ‘down, below; south’, ǰuhl ‘deep; south’, Gilaki ǰulf ‘deep’ (ESIJ 3, 91; Durkin-Meisterernst 2004: 198; NEVP 105; EWAI I, 464-67; Nyberg 1974: 232). There is also the Greek γέφῡρα ‘dam; bridge’ 145 <?page no="146"?> Caland-variant in -ipreserved in Young Avestan ǰaiβi.vafra- ‘with deep snow’, and the nasalized and nasal derivative in Young Avestan ǰąfnu- ‘Vertiefung’, corresponding to Sanskrit gámb h an- ‘depth’. It appears that in Indo-Aryan the *g-variant was generalized, while in Iranian it was the *ǰ-variant (Szemerényi 1960: 212), perhaps with exception of Khotanese ggaṃpha- ‘extent, expanse’ (Bailey 1979: 79-80). Avestan gufra- ‘deep’ and maybe Ossetic Iron kwyrf, Digor qulf ‘deep, depth’ (Abaev I 652) can perhaps represent a blend of the expected Iranian *gafraand the hypothetical Iranian continuant of IE *d h ubro- ‘deep’ (Hilmarsson 1996: 194). 3.2.1 Fick (1890: 401) accepted Kuhn’s comparison of Greek γέφῡρα with Vedic gab h īrá- ‘deep’ and added Greek βάπτω ‘to immerse, as to temper or color’ (Hom. Od.), βαπτίζω and the metathesized variant βιπτάζω ‘to dip, plunge; baptize’ [Epich.], βαφή ‘dipping, temper, dye’, besides Hesychius’ gloss βύπτειν· βαπτίζειν (Beekes 2010: 200). This ‘dipping’ addition was accepted by Prellwitz (1905: 93). Szemerényi (1960: 213) demonstrated that Greek βάπτω should be only derived from *g u̯ m̥ b h -i̯ ō. The form without the nasal infix could perhaps be identified in Hesychius’ gloss γυβᾷ· κολυμβᾷ, i.e. ‘dives’ (cf. Pokorny 1959: 466). The differences in the initials and root-vowels have an analogy in Greek γυνή ‘wife, woman’ (Hom. Il.), Mycenaean ku-na-ja vs. Boeotian βανά [Corinna] (cf. Beekes 2010: 291-92). The unaspirated in‐ tervocalic -βinstead of Greek -φindicates some ancient Hellenic language of the Macedonian type; cf. Hesychius’ gloss following Callimachus: κεβαλή· κεφαλή (Schmidt 1867: c. 849). 3.2.2 Twelve years earlier, Fick (1878: 189) had already added the Ger‐ manic counterparts of Sanskrit gab h īráand Greek βάπτω, namely Old Norse kvefja ‘to dip, put under water’, impers. ‘to be swamped, founder, sink’, refl. ‘to dip oneself, duck, dive, be quenched’, kaf n. ‘a plunge into water, dive, diving’, poet. ‘the deep; sea’, Old Swedish kvaf ‘depth’ etc. (cf. de Vries 1962: 296). Another cognate was identified in Tocharian A k u pār ‘deep, deeply; depth’, derivable from Common Tocharian *kwäpāræ and further from *g u̯ b h ə̯ ro- (Hilmarsson 1996: 193-94). 3.2.3 Nikolaev (2019[2021]: § 3) connected the Indo-Iranian terms for ‘deep’ with Greek δυσπέμφελος, applied to turbulent sea or the dangers of seafaring (Hom. Il. 16.748; Hes. Th. 440 & Op. 618). Separating the prefix δυσ-, the stem ºπέμφελος is derivable from *g u̯ h emb h elovia Lex Grassmann and with the Aeolic reflex of the initial labiovelar. This solution, quite legitimate, would exclude any connection of γέφῡρα with the Indo-Iranian 146 Václav Blažek <?page no="147"?> 2 There is an analogical example in Common Slavic *soldъkъ and Lithuanian saldùs ‘sweet’, which is explainable from Late IE *su̯ ādu- ‘sweet’, crossed with Common Slavic *solь ‘salt’, Latvian sâls id. (Machek 1968: 551). word-family ‘deep’. On the other hand, δυσπέμφελος is also derivable from δυσ- + *βέμφελος, where the initial of the second component can reflect *g u̯ and is compatible with the predecessor of γέφῡρα. 3.2.4 Machek (1968: 169-70) connected Sanskrit gab h īrá- ‘deep’ with Common Slavic *glǫbokъ ‘deep’ (ESSJ 6: 141-42), but without any explana‐ tion of the ‘redundant’ -l-. It is tempting to explain it as a result of crossing 2 two different bases, *gъlº, the hypothetical Slavic counterpart of Lithuanian gilùs ‘deep’, Latvian dziļš, Prussian ainan gillin maiggun acc.sg. ‘deep sleep’ (*g (u̯ ) l̥ Hu-), and *gǫbº, the cognate of Sanskrit gamb h īrá- ‘deep’. 3.2.5 It is possible to conclude that the present forms reflect the basic apophonic forms without and with the nasal infix: *g u̯ eb h -/ *g u̯ ob h -/ *g u̯ b h and *g u̯ emb h -/ *g u̯ omb h -/ *g u̯ m̥ b h -. The primary meaning of the root *g u̯ eb h was probably ‘deep’, which was extended to ‘to dip’, similarly as in Old English dēop ‘deep’ and dyppan ‘to dip’ reflecting Germanic *deupaand its derivative *dupjanrespectively (Szemerényi 1960: 212). This conclusion implies the question of the semantic relation between ‘deep’ and ‘bridge’ or ‘dam’, which was never solved. A plausible solution can be found in the idea of Barić (1919: 110 f), who treated the relation of Greek γέφῡρα ‘bridge; dam’ (and Armenian kamowrǰ ‘bridge’) with Albanian urë, pl. ura ‘bridge’. Taking in account the semantics of the cognates in Old Norse vǫr ‘row of stones’, ver ‘dyke’, Sanskrit saṃvaram. ‘dam, mound, bridge’, vártran. ‘dyke, dam’ etc. (Orel 1998: 488; Pokorny 1959: 1162; MW 1116: 922), it is possible to imagine the compound *g u̯ eb h - & *urwith the hypothetical meaning ‘dyke {through the} depth’ vel sim. 3.3 Müller (1877[1878]: 14) was probably the first to connect Greek γέφῡρα ‘bridge; dam’ with Armenian kamowrǰ ‘bridge’. He was followed by Bugge (1889: 22-23) and Meillet (1920: 17-18). Meillet already tried to ex‐ plain the phonological problems, which are discussed in detail by Clackson (1994: 134-35). The Greek-Armenian comparison is acceptable only if we assume the reconstruction of a syllabic sonant in Armenian *g u̯ m̥ b h uri̯ ə̯ 2 , while Greek γέφῡρα should be projected back to pre-Greek *g u̯ eb h uri̯ ə̯ 2 (see Olsen 1999: 66, 812, fn. 56: *-mb h - > Armenian -mas in Armenian camem ‘chew’ vs. Vedic jámbha- ‘tooth, tusk’; cf. also SLA 26). Then, the base would Greek γέφῡρα ‘dam; bridge’ 147 <?page no="148"?> 3 There were also proposed other interpretations, cf. Blažek (1999: 292). be the verbal stem *g u̯ eb h -, while the Armenian form should have arisen from a variant with the nasal infix (cf. § 3.2.4.). The presence of a labial in early Armenian is probably suggested by the Armenian loanword in Georgian ḳiṗorč ̣ i ‘beam serving as bridge’ (Martirosyan apud Beekes 2004: 20). 3.3.1 For completeness, there have been proposed several different ety‐ mological attempts to explain Armenian kamowrǰ ‘bridge’, implying the separation of the Armenian and Greek words: Bartholomae (1893: 269) tried to explain Armenian kamowrǰ ‘bridge’ with help of Common Slavic *kamy, gen.sg. *kamene ‘stone’. Mann (1963: 147) saw in Armenian kamowrǰ a compound consisting of the counterpart of Geeek καμπή ‘curve’ & *b h r̥ i̯ ə̯ 2 ‘bridge’. Ritter (1996: 17; see also Olsen 1999: 66, note 137) derived Armenian kamowrǰ from *gomuri̯ ə̯ 2 , comparing it with Lithuanian gomurỹs ‘roof (of the mouth)’. In this case Greek γέφῡρα should clearly be separated from Armenian kamowrǰ. 3.3.2 Greek γέφῡρα ‘bridge; dam’ and Armenian kamowrǰ ‘bridge’ are probably formed by the same composite suffix *-ur-i̯ ə̯ 2 / -iH 2 as γέργῡρα (Alcm. 132 apud Etymologicum Magnum 228.38) & γόργῡρα (Hdt. 3.145) ‘an underground drain or sewer’. Elwira Kaczyńska (in her yet unpublished article devoted to Ancient Greek collectives) adds the Italic counterparts in Latin decuria fem. coll. ‘a division consisting of ten (esp. persons, soldiers); a decuria, decade’, Umbrian dat.-abl.pl. tekuries, dequrier ‘decuriis’ vs. Latin decem, Umbrian desenº ‘ten’ (Untermann 2000: 167), and Latin centuria fem. coll. ‘a division consisting of hundred (esp. persons, soldiers); a century, company’ vs. Latin centum ‘hundred’, and interprets them as collectives 3 . 3.4 Prellwitz (1892: 59) derived Greek γέφῡρα from *geg h uri̯ a, which would have been formed from the same root as Sanskrit [Up] g h ūrṇáti ‘moves to and from’. But in such a case, the partially reduplicated Greek form should be reconstructed as *keg h uri̯ a thanks to Lex Grassmann. Prellwitz also did not try to explain the semantic development. 3.5 Muss-Arnolt (1893: 75) sought the origin of Greek γέφῡρα in Akkadian gišru(m) ‘bridge, latch’, whence Hebrew gěšūr, Aramaic gišrā, Syriac gešrā, Arabic ǧisr ‘bridge’ were adapted. Apophonically, the following words are connected with ‘bridge’: Akkadian gušūru, New Assyrian gašūru ‘fallen tree 148 Václav Blažek <?page no="149"?> trunk, beam’ > Aramaic g/ kəšūrā id. (Zimmern 1915: 44). The Akkadian word gišru(m) itself represents rather an adaptation of Sumerian GIŠ.ÙR (AHw I, 300: 293) than a continuant of Semitic + giθru expected by Muss-Arnolt. The assumed substitution of the second radical in Semitic words by Greek *p h is highly suspect; in fact, it is only based on a comparison with the Semitic theonym *ʕaθtar-(at-) > Akkadian Ištar etc. and Greek Ἀφροδίτη (from the Iliad onwards; missing in Mycenaen texts) - for a discussion, see West (2000[2002]: 135). The Greek theonym can be explained as a pure Indo-European compound whose components correspond to Greek ἄφαρ ‘suddenly; quick’ & Sanskrit dīti- ‘shining’, as Witczak (1993) demonstrated. 3.6 Prellwitz (1916: 298-99), followed by Pisani (1962: 140), tried to explain Greek γέφῡρα from the IE verb *b h er- ‘to carry’, referring to the Vedic inten‐ sive perfect jab h ā́ra etc. (see MW 764). As a typological parallel concerning the semantic motivation can serve e.g. Lithuanian tíltas ‘bridge’, derivable from IE *telH 2 - ‘to carry’ > Greek τλῆναι, Latin tollō etc. (Schirmer/ Kümmel, LIV 2 622-23; let us mention that e.g. Pokorny 1959: 1060-61 derived tíltas from IE *telH 2 - ‘flacher Boden, Brett’ and not from *telH 2 - ‘to carry’, although their origin can be common - see e.g. de Vaan 2008: 608-609). 3.7 Muller (1926: 209) connected Greek γέφῡρα and its dialect counter‐ parts, plus Armenian kamowrǰ, with Latin vibia fem. ‘plank, cross-piece supported on trestles so as to form a bank’, recorded by Ausonius in Tech‐ nopaegnion 4: in vetere proverbio est: sequitur varam vibia “The old saw runs: the trestle follows the plank”, i.e. “one evil is followed by another to match”. (transl. by Evelyn-White 1919) Latin vibia can be derived from Italic *g u̯ efi̯ ā and further from *g u̯ eb h i̯ ā, which is in fact compatible with Proto-Greek *g u̯ eb h uri̯ a. The specific devel‐ opment of the Latin root vowel has analogy e.g. in Latin familia < *fameli̯ ā, cf. Oscan nom.sg. famelo, Umbrian nom.pl.? fameřias (Untermann 2000: 262), or Latin Sicilia vs. Greek Σικελία (Leumann 1928: 83, § 67). The other Muller’s steps leading to Sanskrit gáb h asti- ‘arm, hand; beam of light’, Gallo-Latin gabalus ‘furca, patibulum, crux’ (Varro apud Nonius Marcellus), Lithuanian g-bana ‘bunch, bundle, handful’ etc. (see Pokorny 1959: 407-09) exclude the initial *g u̯ and should be abandoned. Greek γέφῡρα ‘dam; bridge’ 149 <?page no="150"?> 3.8 Loewenthal (1927: 182) speculated about a connection of Greek γέφῡρα ‘dam, dyke’ with γύψος ‘gypsum, chalk, cement’. 3.9 Schwyzer (1939: 298), deriving Greek γέφῡρα, Boeotian βέφυρα, Cretan δέφυρα from *g u̯ eb h º or *gu̯ eb h º, ascribed the primary meaning ‘dam, dyke’ to this term and speculated about its derivation from the verb δέφω, δέψω ‘to soften, masturbari’ (cf. Beekes 2010: 320), which should lead to the semantic motivation *‘gestampfter Damm’. 3.10 Van Windekens (1961: 546-47) reconstructed the predecessor of Greek γέφῡρα as *βαμῡρα explaing the phonetic changes from the influence of γῆ ‘earth’ and φῡ́ ρω ‘to mix dry and wet’. Similarly, concerning the first component, Haas (1959: 47) and Essabal (1964: 7-9) saw in Armenian kamowrǰ ‘bridge’ the compound of *g u̯ em- ‘to come’ & *b h oro- ‘plank’. 3.11.1 Furnée (1972: 97, 223, 346, 390) proposed the substratal origin of both the Greek forms for ‘bridge; dam’ and Armenian ‘bridge’ seeking sup‐ port in Hattic ḫamurawa- ‘(roof) beam’ (Soysal 2004: 277, 439: glossed by Su‐ merian GIŠ.ÙR). Similarly did Beekes (2004, 2010: 269). Dunaevskaja (1961: 88) identified its relatives in the West Caucasian languages: Abkhaz-Abazin *q w ə(m)bələ-ra ‘beam over the hearth; cross-beam’ (Chirikba 1996: 423; cf. Martirosyan 2010: 352) > Abkhaz a-x̲ w blarə, Abaza of Tapanta q w əmbla, of Ashkar q w əblə. 3.11.2 Martirosyan (2010: 352) finds a more promising parallel to Arme‐ nian kamowrǰ ‘bridge’ in the newly identified Urartian word qaburzani- ‘bridge’ (Çavuşoğlu/ Işik/ Salvini 2010: 45-47; the suffix -niis determined on p. 46), which could be the source of both Armenian kamowrǰ ‘bridge’ and Georgian ḳiṗorč ̣ i ‘beam serving as bridge’ (Martirosyan apud Beekes 2004: 20). A source of Greek γέφῡρα ‘dam; bridge’ should be sought in the expected Hurrian cognate of Urartian qaburzani- ‘bridge’. A promising candidate could be identified in the probable Hurrian borrowing in Akkadian: Old Babylonian kawaru, Middle Assyrian kabaru, New Babylonian kammar(r)u & kamru ‘(garden) wall, ramp or similar earth construction’ (CAD 8, 111; Richter 2012: 182-83). The meaning ‘wall, ramp’ stands close to Homeric ‘dam, dyke’. The difference in the vowel of the first syllable is not surprising; there are also other examples, when the vowel a in Old Anatolian place names written in the cuneiform script corresponds to ɛ in Greek, e.g. Lazpa = Λέσβος or Apaša = Ἔφεσος (cf. García Trabazo 2007: 45). 3.12 Knobloch (1975: 76-77) based his new solution on the assumption that the Boeotian variant βλέφυρα is primary, including its -λ-. The corre‐ 150 Václav Blažek <?page no="151"?> sponding counterparts in Cretan δέφυρα ‘bridge’ and the Laconian gloss δίφουρα· γέφυρα by Hesychius indicate pre-Greek *g u̯ leb h uri̯ ə̯ 2 , character‐ ized by the same suffix as e.g. λέπῡρον ‘Schale, Hülse’ vs. λεπρός ‘schuppig’ or ἁλμυρός ‘salzig’ vs. ἅλμη ‘Meerwasser’ etc. (Knobloch 1975: 77; cf. already Kuhn 1852: 135). The hypothetical root proper, *g u̯ leb h -, was connected by Knobloch with Russian žólob ‘gutter, trough, gully’. The other Slavic cognates, e.g. Church Slavonic žlěbъ ‘groove; crib’, Slovenian žlêb, Old Czech žleb ‘gutter, trough, gully’, indicate Common Slavic *želbъ (Vasmer 2, 61). In addition, one can add another ablaut variant attested in Russian globá fem. ‘beam’. But the traditionally accepted cognate, Old Norse golf m. ‘floor’, derivable from Germanic *gulba- (Falk/ Torp 1909: 137; Kroonen 2013: 194), indicates the initial *g h -, incompatible with *g u̯ -. 3.12.1 Knobloch (l.c.) separated Greek γέφῡρα from Armenian kamowrǰ ‘bridge’ accepting for the latter the comparison with Common Slavic *kamy, gen.sg. *kamene ‘stone’ proposed by Bartholomae (1893: 269). 3.13 Puhvel (1976: 64-66, 1984: 282-83) segmented γέφῡρα as *γ’ἔφῡρα with the particle γε, seeking support in the old designation of Corinth located on the Isthmus, Ἐφύρη (Hom. Od. 1.259), and Hittite epurai- ‘to besiege, dam up’. This innovated form with initial γwould have had to replace the form with initial labiovelar reflected in Cretan δέφυρα and Boeotian βέφυρα. 3.14 Hooker (1979: 387-98) returned to the Semitic origin of γέφῡρα and tried to explain it from the Semitic root √ g-b-ʕ: Akkadian gabʔu ‘peak, height’, Ugaritic gbʕ ‘hill, height, peak’, Hebrew gibʕā ‘hill’ (DRS 97; DUL 288). This propsal can not be rejected, if the primary meaning of γέφῡρα was ‘dam, dyke’. However, Hooker did not explain Greek -ῡρα, which has no parallels in Semitic. 3.15 Van Windekens (1986: 54) derived γέφῡρα from *δέφῡρα, explaining γunder the influence of γέργῡρα (Alcm. 132 apud Etymologicum Magnum 228.38) & γόργῡρα (Hdt. 3.145) ‘an underground drain or sewer’. The expected form *δέφῡρα, confirmed by Cretan δέφυρα, was projected by Van Windekens back to Proto-Greek *γu̯ έφυρɩ̯ α with the initial labiovelar supported by Boeotian βέφυρα. Separating the same suffix as in ἄγκῡρα, he speculated about the metathesis in the root: *γ u̯ -φ < *φ-γ u̯ leading to the root *b h eg u̯ attested e.g. in Greek φέβομαι ‘to flee in fear; scamper, scurry’, Lithuanian bė ́ gti : bė ́ gu ‘to run; escape, flee, take flight; walk fast, hurry; flow, stream (of a river)’, Latvian bêgt : bȩ̂ gu ‘to flee’ etc. (LIV 2 67; Smoczyński Greek γέφῡρα ‘dam; bridge’ 151 <?page no="152"?> 4 One of the referees drew my attention to the Locrian (Northwest Greek) tendency to change ερ to αρ (Mendez Dosuna 1985: 395-396, 408) implying the older form *βουφερα derivable via metathesis from *βεφουρα, which is compatible with Proto-Greek *γ u̯ έφυρɩ̯ α. Despite several speculative aspects of this proposal, the ‘cow’-etymology could be eliminated. 2018: 107). According to Van Windekens the primary semantics of γέφῡρα had to be ‘water-course’. 3.16.1 It is possible to add one more etymology based on Hesychius’ gloss (in acc.pl.) βουφάρας· γεφύρας, without an origin given, which should, considering the order of the Greek alphabet, be corrected to *βουφoρας 4 (Schmidt 1867: 318), while Schulze (1885: 423) proposed the emendation *βοφoύρας following Laconian δίφουρα. Taking in account the first possi‐ bility implying the probable nom.sg. *βουφoρα, it could be easily interpreted as ‘carrying the cattle’ (Knobloch 1975: 76 interpreted it as ‘Ochsenfurt’). A similar paraphrasis can probably be found in Old Irish bóthar ‘path’, origi‐ nally probably ‘livestock footpath or pen’, cp. bó ‘cow, stallion’ (Pedersen II, 51; Buck 1949: 718). O’Rahilly (1950: 160) proposed a beguiling etymology in the form of the compound *bou-itrowhere the second component would correspond to Latin iter. Hamp (2001[2003]: 153) finds support for this interpretation in the toponym Oxgangs near Edinburgh. However, Vendryes pointed to the Welsh dial. (Pembroke) word meidir from fem. meud(y)r ‘crossroad’, with the oldest records from 1531 y veydyr, and from 1600 arian y vidir ‘pecuniae viae’ to which Old Welsh *boudr < *g u̯ ou̯ -trā is to correspond (LEIA B 74-75). In that case, the Old Irish word would not be a compound but an agent noun. This was rejected by Hamp (1994: 172), who showed that the Welsh word reflects Brittonic *bou̯ Vtrā, which could be interpreted as neuter plural, i.e. as a collective to the proto-form *g u̯ ou̯ -itro-. 3.16.2 It remains to be proven whether the other forms represent the same compound. Besides proto-Greek *g u̯ ep h uri̯ a there could exist the var‐ iant *g u̯ op h uri̯ a, similarly as γέργῡρα ‘an underground drain or sewer’ (Alcm. Fragmenta 132; see Etymologicum Magnum 228.38) besides γόργῡρα ‘dun‐ geon’ (Hdt. 3.145) or Κέρκυρα besides Κόρκυρα (Schwyzer 1939: 255; Furnée 1972: 118; Beekes 2010: 283-84). And such forms as *g u̯ op h urº could originate from hypothetical *g u̯ up h orº via metathesis, cf. the forms with υ in the first syllable, which was shifted into the second syllable: Μυτιλήνη vs. Μιτυλήνη, μυστίλη ‘a piece of bread hollowed out like a spoon’ vs. μιστύλη, ὑλακτεῖ ‘bark, howl’ vs. Cretan ἀλυκτεῖ (Schwyzer 1939: 268). It is only in the position 152 Václav Blažek <?page no="153"?> before *ū̆ that the Proto-Greek labiovelar retains its velar realization even in the classical era, although here the power of analogy sometimes overcomes the weight of a sound law, e.g. Greek πρέβυς : Cretan πρεῖγυς ‘a superior’ < *prei̯ s-g u̯ u-, literally ‘going forward’ (Lejeune 1972: 43-45; Frisk II, 592-93). This solution allows us to propose two scenarios, whose results influenced one another: *g u̯ up h orº > *g u̯ op h urº > *g u̯ ep h urº with regular palatalisation of the initial according to rules for post-Mycenaean dialects, besides *g u̯ up h orº > *g u̯ up h urº > *gup h urº > *gep h urº under the influence of the root vowel of *g u̯ ep h urº. The first component *g u̯ ushould represent the zero-grade of Proto-Greek *g u̯ ōu- > Mycenaean qo-u- (qo-u-ka-ra; qo-u-ko-ro; qo-u-qo-ta - see Aura Jorro 1993: 210-11), Classical βοῦς ‘cow, bull, cattle’, similarly as Vedic ºgúin su-gú- ‘with nice cows’ corresponds to nom.sg. gáuṣ, gen.-abl.sg. góṣ, dat.sg. gáve ‘cow, bull, cattle’ (EWAI I, 478-80). 4. Conclusion. From the preceding list of 16 published etymologies, 10 attempts, namely 3.1., 3.4., 3.5., 3.6., 3.8., 3.9., 3.10., 3.12., 3.14. and 3.15., seem to be wrong. The remaining six attempts should be analyzed in mutual connection. 4.1 The etymologies 3.2., 3.3. & 3.7. admit the interpretation of Proto-Greek *g u̯ ep h uri̯ a as the compound *g u̯ eb h -uri̯ ə̯ 2 and similarly the Armenian word kamowrǰ ‘bridge’ as *g u̯ m̥ b h -uri̯ ə̯ 2 , both based on the same semantic motiva‐ tion ‘belonging to the dam through the depth’ → ‘bridge’. They differ only in the use of the simple root *g u̯ eb h in Greek and its zero-grade with the nasal infix *g u̯ m̥ b h in Armenian. On the other hand, among designations of ‘bridge’ or ‘dam’ it has no analogy in either Indo-European or Mediterranean languages. 4.2 The second possibility is based on Hesychius’ gloss βουφάρας· γεφύρας (§ 3.16.). Accepting Schmidt’s emendation, the nom.sg. *βουφoρα allows the projection back to Proto-Greek *g u̯ ōup h ora < Late IE *g u̯ ōu̯ -b h orə̯ 2 ‘carrying cattle’ more or less corresponding to Goidelic *bou̯ -itro-, Brittonic coll. *bou̯ -itrā- *‘livestock path’. It remains questionable whether the other dia‐ lect forms reflecting Proto-Greek *g u̯ ep h uri̯ a are compatible, perhaps from *g u̯ up h ori̯ a via metathesis and vocalic vacillation. 4.3 There was also proposed the third purely Indo-European solution, which was based on the comparison of Latin vibia ‘plank, cross-piece supported on trestles so as to form a bank’ < *g u̯ eb h i̯ eH 2 and Proto-Greek *g u̯ ep h uri̯ a ‘dam; bridge’ < *g u̯ eb h uri̯ ə̯ 2 . The semantics of the Latin word is closer to the meaning ‘dam’ of the Greek counterpart. Greek γέφῡρα ‘dam; bridge’ 153 <?page no="154"?> 5 The study originated with the support of the Specific Research Fund at Masaryk Uni‐ versity. I am grateful to John D. Bengtson for his correction of English, to Alexander Nikolaev, who kindly mediated me his unpublished article, and to two anonymous referees for their valuable comments and additions. 4.4 The fourth alternative assumes an adaptation of Proto-Greek *g u̯ ep h uri̯ a ‘dam; bridge’ from some Hurro-Urartian dialect (§ 3.11.2.) spoken probably in Asia Minor. The hypothetical source-protoform should be close to one of the historically attested forms: Urartian qaburza-ni- ‘bridge’ and Hurrian *kabar- ~ *kam(b)ar- > Old Babylonian kawaru, Middle Assyrian kabaru, New Babylonian kammar(r)u & kamru ‘wall, ramp or similar earth construction’. The same origin, only from another Hurro-Urartian dialect, can be proposed for Armenian kamowrǰ ‘bridge’ and still another for Geor‐ gian ḳiṗorč ̣ i beam serving as bridge’. 4.5 If the alternative name of Corinth, Ἐφύρη, designated originally the Isthmus (cf. § 3.13.), it is possible to speculate about its Hurro-Urartian origin and the Luvian mediation. In Luvian, IE * ( ĝ )(h) before front vowels was deleted: cf. Cuneiform Luvian īš(ša)ra/ i- ‘hand’ vs. Hittite keššar/ kiššer-/ kišrid., or Cuneiform Luvian im(ma)ra/ i- ‘open country’ vs. Hittite gimra- ‘countryside’ etc. (Kloekhorst 2008: 471, 476). 4.6 Summing up, the most probable solution of the origin of the analyzed Greek and Armenian designations of ‘bridge’ (in Greek also ‘dam’) seems to be the adaptation of the Hurro-Urartian substratal term and its interfer‐ ence with Proto-Greek *g u̯ ōup h ora inherited from the Late IE compound *g u̯ ōu̯ -b h orə̯ 2 ‘carrying cattle’. The influence of another inherited appellative, namely Late IE *g u̯ eb h i̯ ā > Latin vibia ‘cross-piece supported on trestles so as to form a bank’, cannot be excluded as well. 5 Bibliography Abaev, Vasilij I. (1958-1995): Istoriko-ėtimologičeskij slovaŕ osetinskogo jazyka, I-V. (Moskva-) Leningrad: Izdateľstvo Akademii nauk. Aeschylus (1926): Persians, with an English translation by Herbert Weir Smyth. Cambridge (MA.): Harvard University Press. AHw = von Soden, Wolfram (1965): Akkadisches Handwörterbuch, I-III. Wiesbaden: Harrassowicz, 1959-1981. 154 Václav Blažek <?page no="155"?> Aura Jorro, Francisco (1985-1993): Diccionario micénico, I-II. Madrid: Instituto de Filología - Consejo superior de investigaciones científicas. Ausonius, Decimus Magnus (1919): “Technopaegnion”. In: Works, Vol. 1, ed. and transl. by Hugh G. Evelyn-White. London: Heinemann/ Cambridge (Mass.): Harvard University Press. Bailey, Harold W. (1979): Dictionary of Khotan Saka. Cambridge: University Press. Barić, Henrik (1919): Albano-rumänische Studien I. (Zur Kunde der Balkanhalbinsel. Quellen und Forschungen VII). Sarajevo: Institut für Balkanforschung. Bartholomae, Christian (1893): “Arica III”. Indogermanische Forschungen 2: 260-284. Beekes, Robert S. P. (2004): “Hom. γέφυρα, and Arm. kamurǰ ‘bridge’”. Glotta 78: 12-21. Blažek, Václav (1999): Numerals. Brno: Masarykova univerzita. Buck, Carl Darling (1949[1992]): A Dictionary of Selected Synonyms in the Principal Indo-European Languages. Chicago/ London: University of Chicago Press. Bugge, Sophus (1889): Beiträge zur etymologischen Erläuterung der armenischen Sprache I. (Christiania Videnskabs-Selskabs Forhandlinger 1889, No. 4: Sitzung der Historisch-Philosophischen Klasse, den 5. April 1889). Christiania: Dybwad. CAD = The Assyrian Dictionary of the Oriental Institute of the University of Chicago Vol. 8: K (1971): A. Leo Oppenheim/ Erica Reiner/ Robert Biggs et alii (eds.). Chicago: Oriental Institute of the University of Chicago. Chirikba, Viacheslav A. (1996): Common West Caucasian. The Reconstruction of its Phonological system and Parts of its Lexicon and Morphology. Leiden: Research School of CNWS - School of Asian, African, and Amerindian Studies. Clackson, James (1994): The Linguistic Relationship between Armenian and Greek. Oxford (UK)/ Cambridge (USA): Blackwell. Collitz, Hermann/ Bechtel, Friedrich (1905): Sammlung der griechischen Dialekt-In‐ schriften; Bd.-3.2. Göttingen: Vandenhoeck & Ruprecht. Curtius, Georg ( 2 1866): Grundzüge der griechischen Etymologie. Leipzig: Teubner. Çavuşoğlu, Rafet/ Işik, Kenan/ Salvini, Mirjo (2010): “New Urartian Inscriptions from East Turkey”. Orientalia. Nova series 79(1): 36-54. DRS = Dictionnaire des racines sémitiques (1970 f): David Cohen (ed.). Paris/ La Haye: Mouton. DUL = A Dictionary of the Ugaritic Language in the Alphabetic Tradition (2015): Gregorio del Olmo Lete/ Joaquín Sanmartín (eds.). Leiden/ Boston: Brill. Dunaevskaja, I.M. (1961): “Principy struktury xattskogo (protoxettskogo) glagola”. In: Peredneaziatskij sbornik I: Voprosy xettologii i xurritologii. Moskva: Izdateľstvo vostočnoj literatury, 57-159. Greek γέφῡρα ‘dam; bridge’ 155 <?page no="156"?> Durkin-Meisterernst, Desmond (2004): Dictionary of Manichaean Texts, Vol. III: Texts from Central Asia and China, Nicholas Sims-Williams (ed.), Part 1: Dictionary of Manichaean Middle Persian and Parthian. Turnhout (Belgium): Brepols. ESIJ = Rastorgueva Vera S./ Ėdeľman, Džoj I. (2007): Ėtimologičeskij slovaŕ iranskix jazykov, Tom 3: f-h. Moskva: Izdateľskaja firma ‘Vostočnaja literatura’ RAN. Essabal, Paul (1964): “IE initial ‘bh’ in Armenian”. Studia Linguistica 18: 1-11. ESSJ = Ėtimologičeskij slovaŕ slavjanskix jazykov, 1-36 (1974 f): O.N. Trubačev et al. (eds.). Moskva: Nauka. Etymologicon Magnum (1816): Friedrich W. Sturz/ Friedrich Sylburg (eds.). Lipsiae: Weigel. EWAI = Mayrhofer, Mannfred (1992-2001): Etymologisches Wörterbuch des Altindoar‐ ischen, I-III. Heidelberg: Winter. Falk, Hjalmar S./ Torp, Alf (1909): Wortschatz der germanischen Spracheinheit. Göt‐ tingen: Vandenhoeck & Ruprecht. Fick, August (1878): “Etymologien”. {Bezzenberger’s} Beiträge zur Kunde der indoger‐ manischen Sprachen 2: 187-189. Fick, August (1890): Vergleichendes Wörterbuch der Indogermanischen Sprachen, I: Wortschatz der Grundsprache, der Arischen und der Westeuropäischen Sprachein‐ heit. Göttingen: Vandenhoeck & Ruprecht. Fraenkel, Ernst (1962-1965): Litauisches etymologisches Wörterbuch, I-II. Göttingen: Vandenhoeck & Ruprecht/ Heidelberg: Winter. Fragments of Old Comedy, Vol. III (2011): Philonicus to Xenophon. Adespota, ed. and transl. by Ian C. Storey (Loeb Classical Library). Cambridge (MA.): Harvard University Press, 260-261, §-49. Frisk, Hjalmar (1973-1991): Griechisches etymologisches Wörterbuch. Heidelberg: Winter. Furnée, Edzard J. (1972): Die wichtigsten Konsonantischen Erscheinungen des Vorgrie‐ chischen. The Hague/ Paris: Mouton. García Trabazo/ José Virgilio (2007): “Ahhiyawafrage y cuestiones conexas. ¿Podemos extraer más datos de las fuentes hititas? ”. In: Las aguas primigenias: el Próximo Oriente Antiguo como fuente de civilización, de Josué Javier Justel Vicente/ Bárbara Eugenia Solans Gracia/ Juan Pablo Vita Barra/ José Ángel Zamora López (eds.). Zaragoza: Instituto de Estudios Islámicos y del Oriente Próximo, 43-67. Haas, Otto (1959): “Die Lehre von den indogermanischen Substraten in Griechen‐ land”. Linguistique Balkanique 1: 29-56. Hamp, Eric P. (1994): “Varia: III. Irish bóthar, Welsh meidir”. Études celtiques 30: 169-173, spec. 172. 156 Václav Blažek <?page no="157"?> Hamp, Eric P. (2001[2003]): “A Rule of the Road”. General Linguistics 41: 147-158. Herodotus (1920): The Histories, A. D. Godley (transl.). Cambridge: Harvard University Press. Hilmarsson, Jörundur (1996): Materials for a Tocharian Historical and Etymological Dictionary (Tocharian and Indo-European Studies, Supplementary Series, Vol. 5). Reykjavík: Málvísindastofnun Háskóla Íslands. Homer (1924): The Iliad, A.T. Murray (transl). Cambridge (MA.): Harvard University Press/ London: Heinemann. Hooker, James T. (1979): “γέφυρα: A Semitic Loan-Word? ”. In: Studies in Dia‐ chronic, Synchronic, and Typological Linguistics, Vol. 1, Bela Brogyanyi (ed.). Am‐ sterdam/ Philadelphia: Benjamins, 387-398. KEWA = Mayrhofer, Mannfred (1956-1980): Kurzgefasstes etymologisches Wörterbuch des Altindischen, I-IV. Heidelberg: Winter. Kloekhorst, Alwin (2008): Etymological Dictionary of the Hittite Inherited Lexicon. Leiden/ Boston: Brill. Knobloch Johann (1975): “Die Brücke - eine griechisch-armenische Isoglosse? ”. Zeits‐ chrift für vergleichende Sprachforschung 89(1): 76-79. Kroonen, Guus (2013): Etymological Dictionary of Proto-Germanic. Leiden/ Boston: Brill. Kuhn, Adalbert (1852): “Die wurzel GAF, GAMF”. Zeitschrift für vergleichende Sprach‐ forschung auf dem Gebiete des Deutschen, Griechischen und Lateinischen 1(2): 123- 141. LEIA = Vendryes Joseph (1959 f.): Lexique étymologique de ľirlandais ancien. Dublin: Institute for Advanced Studies/ Paris: CNRS. Lejeune, Michel (1972): Phonétique historique du mycénien et du grec ancien. Paris: Klincksieck. Leumann, Manu ( 5 1928): Lateinische Grammatik. München: Beck. LIV 2 = LIV. Lexikon der indogermanischen Verben: Die Wurzeln und ihre Primärstamm‐ bildungen, 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 Martin Kümmel und Helmut Rix. Wiesbaden: Reichert 2001. Loewenthal, John (1927): “Wirtschaftsgeschichtliche Parerga, Teil II.”. Wörter und Sachen 10: 180-186. Machek, Václav ( 2 1968): Etymologický slovník jazyka českého. Praha: Academia. MacKenzie, David Neil (1971): A Concise Pahlavi Dictionary. London: Oxford Univer‐ sity Press. Greek γέφῡρα ‘dam; bridge’ 157 <?page no="158"?> Mann, Stuart E. (1963): Armenian and Indo-European (Historical Phonology). London: Luzac. Martirosyan, Hrach (2010): Etymological Dictionary of the Armenian Inherited Lexicon. Leiden/ Boston: Brill. Meillet, Antoine (1920): “Le nom du ‘pont’”. Bulletin de la Société de Linguistique de Paris 22(1): 17-18. Mendez Dosuna, Julian (1985): Los dialectos dorios del Noroeste. Gramatica y estudio dialectal. Salamanca: Ediciones Universidad de Salamanca. Muller, Frederik (1926): Altitalisches Wörterbuch. Göttingen: Vandenhoeck & Ru‐ precht. Müller, Friedrich (1877[1878]): “Armeniaca V”. Sitzungsberichte der Kaiserlichen Aka‐ demie der Wissenschaften (Philosophisch-historischen Classe) 88(1): 9-16. Muss-Arnolt. W. (1893): On Semitic words in Greek and Latin. Ann Arbor: University of Michigan. MW = Monier-Williams, Monier (1899[1993]): A Sanskrit-English Dictionary. Delhi: Motilal Banarsidass. NEVP = A New Etymological Vocabulary of Pashto (2003): By Georg Morgenstierne, J. Elfenbein/ David Neil MacKenzie/ Nicholas Sims-Williams (comp; eds.). Wiesbaden: Reichert. Nikolaev, Alexander (2019[2021]): “Deep Waters: The Etymology of Vedic gabhīrá-”. Historische Sprachforschung 131: 191-207. Nyberg, Henrik S. (1974): A Manual of Pahlavi, Vol. II. Wiesbaden: Harrassowitz. Olsen, Birgit (1999): The Noun in Biblical Armenian. Origin and Word-Formation. Berlin/ New York: Mouton de Gruyter. O’Rahilly/ Thomas F. (1950): “Varia”. Celtica 1(2): 158-160. Orel, Vladimir (1998): Albanian Etymological Dictionary. Leiden/ Boston/ Köln: Brill. Pedersen, Holger (1909-1913): Vergleichende Grammatik der keltischen Sprachen, I-II. Göttingen: Vandenhoeck & Ruprecht. Pindar (1937): The Odes of Pindar including the Principal Fragments, Sir John Sandys (transl.). Cambridge (MA.): Harvard University Press/ London: Heinemann. Pindar (1990): Odes, Diane Arnson Svarlien (ed.). Pisani, Vittore (1961): {Review of KEWA 1960-61}. Paideia 17: 139-142. Pokorny, Julius (1959): Indogermanisches etymologisches Wörterbuch. Bern/ München: Francke. Prellwitz, Walther ( 1 1892). Etymologisches Wörterbuch der Griechischen Sprache. Göttingen: Vandenhoeck & Ruprecht. 158 Václav Blažek <?page no="159"?> Prellwitz, Walther ( 2 1905): Etymologisches Wörterbuch der Griechischen Sprache. Göttingen: Vandenhoeck & Ruprecht. Prellwitz, Walther (1916): “Griechische Etymologien”. Zeitschrift für vergleichende Sprachforschung auf dem Gebiete der Indogermanischen Sprachen 47(3-4): 295-306. Puhvel, Jaan (1976): ΠΟΛΕΜΟΙΟ ΓΕΦΥΡΑΙ. Indogermanische Forschungen 81: 60-66. Puhvel, Jaan (1984): Hittite Etymological Dictionary, Vol. 1: A; Vol. 2: E & I. Berlin/ New York/ Amsterdam: Mouton. Richter, Thomas (2012): Bibliographisches Glossar des Hurritischen. Wiesbaden: Har‐ rassowitz. Ritter, Ralf-Peter, con la colaboración de F. Javier Martinez García (1996): Introducción al armeno antiguo. Madrid: Ediciones Clásicas. Schmidt, Mauricius (ed.) (1867): Hesychii Alexandrini Lexicon. Jena: Duft. Schulze, Wilhelm (1885): “Indogermanische āi- Wurzeln”. Zeitschrift für vergleichende Sprachforschung auf dem Gebiete der Indogermanischen Sprachen 27(4), 420-427. Schwyzer, Eduard (1939): Griechische Grammatik (auf der Grundlage von Karl Brug‐ manns Griechischer Grammatik), I. Band. München: Beck. SLA = Džaukjan, G.B./ Saradževa, L.A./ Arutjunjan, C.R. (1983): Očerki po sravniteľnoj leksikologii armjanskogo jazyka. Erevan: Izdateľstvo AN Armjanskoj SSR. Soysal, Oğuz (2004): Hattischer Wortschatz in hethitischer Textüberlieferung. Leiden/ Boston: Brill. Szemerényi, Oswald (1960): “The Origin of Greek βαϑύς and βόϑυνος”. Glotta 38(3-4), 211-216. Turner, Ralph L. (1966): A Comparative dictionary of the Indo-Aryan Languages. London: Oxford University Press. Untermann, Jürgen (2000): Wörterbuch des Oskisch-Umbrisch. Winter: Heidelberg. de Vaan, Michiel (2008): Etymological Dictionary of Latin and the other Italic Lan‐ guages. Leiden/ Boston: Brill. Van Windekens, Albert J. (1961): “Sur trois mots arméniens”. Handes Amsorya 75: 545-548. Van Windekens, Albert J. (1986): Dictionnaire étymologique complémentaire de la langue grecque. Nouvelles contributions à l’interprétation historique et comparée du vocabulaire. Leuven: Peeters. Vasmer, Max (1986-1987): Ėtimologičeskij slovaŕ russkogo jazyka, Oleg N. Trubačev (transl.). Moskva: Progress. de Vries, Jan ( 2 1962): Altnordisches etymologisches Wörterbuch. Leiden: Brill. West, Martin L. (2000[2002]): “The Name of Aphrodite”. Glotta 46: 134-138. Greek γέφῡρα ‘dam; bridge’ 159 <?page no="160"?> Witczak, Krzysztof T. (1993): “Greek Aphrodite and her Indo-European origin”. In: Miscellanea linguistica Graeco-Latina (Collection d’Études Classiques 7), Lambert Isebaert (ed.). Namur: 115-123. Zimmern, Heinrich (1915): Akkadische Fremdwörter als Beweis für babylonischen Kultureinfluss. Leipzig: Hinrichs. 160 Václav Blažek <?page no="161"?> Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek Martin Masliš Abstract: One strain of research in psycholinguistics focuses on how speakers process and store lexemes based on their morphological structure. Experiments with speakers of modern languages show that words with common root morphemes form families in speakers' mental lexicon. This study applies this framework to Ancient Greek, exploring the impact of morphological parsing on semantic acquisition. We propose that speakers connected novel words to related words based on formal similarities, using related words as cues to intended meaning. This predicts that isolated words were more prone to se‐ mantic reanalysis and change over time. The hypothesis has been tested on two datasets consisting of Ancient Greek nominal lexemes created using the suffixes -μοand -ρο-, respec‐ tively. The experimental groups comprised words for which we have predicted diminished formal transparency and hence a greater chance of being synchronically isolated; the control groups contained words more likely to have been formally transparent. We have observed the prevalence of semantic change in both groups of the two datasets and corroborated this part of the analysis by consulting the Late Antique Greek etymologica. The subsequent statistical analysis has revealed that the words belonging to the groups of the potentially isolated lexemes were more likely to undergo semantic reanalysis than their counterparts from the control groups. Keywords: Ancient Greek, semantic change, mental lexicon, isolated lexemes, word families <?page no="162"?> 1 Bloom (2002: 3-4) and Wagner (2010: 519). 1 Introduction A thought experiment introduced by Quine ( 2 2013: 23-41) presents a scene in which a field researcher investigating an undocumented language finds himself in the following situation. Strolling together with a native informant, they startle a hare, which dashes away. This prompts the informant to utter a sequence that can be captured as gavagai. The researcher is then faced with the problem of how he ought to interpret the linguistic sign. It could be glossed as ‘hare’, ‘food’, or ‘gray’ among many other possibilities depending on which aspect of the scene it refers to. Literature on the acquisition of the lexicon presents modifications of this thought experiment, in which an individual in the process of language acquisition replaces the researcher. 1 Just like the linguist, s/ he must reconstruct the meaning of a sign from all the available hints. Modifying the example still further, let us imagine that, in addition to gavagai, a sign gava exists and that s/ he knows its meaning. Will s/ he then use this piece of data as a cue to the meaning of gavagai if s/ he is also aware that there is a suffix -gai? And what if there once was a pair *gava and gavagai, but the former has since changed to goba, while the latter has remained the same? In this study, we investigate the diachronic implications of such situations for the lexicon of Ancient Greek (henceforth AG). Lexemes that do not participate in synchronic word families (i.e. isolated words) due to formal disparities may have been more susceptible to semantic reanalysis than lex‐ emes transparently connected to the etymologically related words - cf. Hock (1991: 296-297). We propose that when encountering a formally isolated word for the first time, speakers of AG had fewer cues to reconstruct the intended meaning compared to the lexemes that were tentatively embedded in word families due to the matching shape of their derivational bases. To test the hypothesis about the semantic instability of formally isolated words, several preliminary questions need to be addressed. Firstly, the infor‐ mation structure in the mental lexicon of AG speakers must be determined. The hypothesis assumes that lexemes are stored as analysable structures, at least to an extent, and not as unsegmented units. This feature could allow speakers to identify their derivational bases and assign novel words to their appropriate families. Secondly, we will identify root shapes in AG which 162 Martin Masliš <?page no="163"?> 2 See Schreuder/ Baayen (1995: 149-152) for an appraisal of these models together with references. 3 These include exemplar-based models (e.g. Daelemans/ van den Bosch 2005), con‐ nectionist models (cf. McClelland/ Rumelhart 1986), and models based on the Re‐ scorla-Wagner equations describing discriminative learning (e.g. Baayen et al. 2011). were prone to lose their synchronic transparency in the individual deriva‐ tions using language-specific criteria for base transparency. Thirdly, we will elaborate an approach to semantic change that emphasizes reanalysis over the cognitive categorization of possible changes. These preliminaries form the basis of a method developed here to test the hypothesis. Two experimental groups of potentially isolated nominal lexemes will be tested against their control groups consisting of potentially transparent lexemes. These two pairs form two datasets of words built on the synchronically productive suffixes -μοand -ρο-, respectively. We will analyse the presence or absence of a non-trivial semantic change measured against the older meaning, either attested or reconstructed, of the given lexeme or its derivational base depending on the opacity of the given lexeme. 2 Structure of the mental lexicon of Ancient Greek speakers The hypothesis rests on the possibility to segment lexical input into smaller units cueing meanings on the level of cognitive representations. Accord‐ ingly, we need to establish whether AG speakers parsed words as morpheme chains or unanalysable wholes. If they tentatively segmented each novel word into a base and affixes, previously acquired words with the same base could give cues to the semantics of the novel word. Determining the exact way in which speakers process and store lex‐ ical data has been an object of research in psycholinguistics for decades. Whereas the early models oscillated between obligatory parsing and full listing, 2 more recent work applies machine learning and statistics to devise computational models 3 that would replicate experimental results while providing a biologically plausible framework. It is important to note that even if some recent models adopt an a-morphous approach to morphology and discard the layer of morphemic representation altogether, they do not invalidate the approach espoused in this study. One such parsing model envisaged in Baayen et. al. (2011), which is limited to reading, posits the direct mapping of forms onto meanings. Here, the forms equivalent to Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 163 <?page no="164"?> graphic unigrams and bigrams - i.e., individual letters or groups of two letters - cue the meanings of words and affixes. It follows, for example, that the bigram ba has a higher specific cue validity than the unigram a, which appears in many more words. Our approach (see below) applies a similar reasoning. However, we retain the traditional terminology of the Item-and-Arrangement model (Hockett 1954: 214-217), which is a common currency in Greek linguistics, while maintaining that the reasoning pre‐ sented in this study does not depend on it. For our purposes, experimental data provide sufficient insight even without commitment to a specific model - see Diependaele et al. (2012: 316-319) for a literature review. Priming tests measure response latencies when a respondent presented with a word (a prime) evaluates the validity of another word (a target) supplied directly afterwards. In English, pairs like happiness and happy induce shorter latencies than pairs of unrelated words - see Marslen-Wilson et al. (1992: 84-85). This suggests that the exposure to a derived prime word activates the stem representation in the mental lexicon of the subject, which then identifies a related target word faster. Derivational morphology of English, however, allows for bases like {happy} to appear as monomorphemic words. Marslen-Wilson et. al. (1994: 25-26) show that the lack of competition between suffixes in pairs like happiness and happy facilitates shorter response times, while suffixed pairs like attrac‐ tion and attractive do not prime each other. This seems to have negative implications for our hypothesis, as root formations in old Indo-European languages are rare and etymological families cluster around bases which do not appear as monomorphemic lexemes of the type {happy}. The same study, however, ascertains that the suffix competition between related forms has stronger inhibitory effects on the response times compared to pairs of unrelated words like gallon and gallop - see Marslen-Wilson et al. (1994: 26 and 26 n. 31). The stronger form of competition in pairs like attraction and attractive persists because simultaneous activation of the same base with two different suffixes is not possible. The study concludes that this inhibition is probably short-lived, and a delayed testing would likely show facilitatory effects, i.e. shorter latencies, in pairs of suffixed words. The same could hold for AG suffixed pairs like κῦδ-ος ‘glory’ and κυδ-ρός ‘glorious’, which would occupy adjacent positions in the mental lexicon. An investigation of root frequency effects in Italian by Burani/ Thornton (2003) found that low-frequency derived words with high-frequency roots 164 Martin Masliš <?page no="165"?> induced faster and more accurate responses in lexical decision tasks - see also Sandell (2015: 75-76). On the other hand, words with low-frequency roots patterned together with non-derived words. This shows that even if roots do not appear as monomorphemic words, which is the case in Italian, high root frequency provides an advantage in word processing. Our hypothesis assumes that this advantage facilitates semantic mapping during the acquisition, but it is lost if an isolated derivational base cannot be identified by speakers. Another important point is the role of semantics in the facilitatory effects between related words. The hypothesis presumes that speakers could con‐ nect a novel word with no assigned meaning to an already acquired form with the seemingly same base. The decisive factor seems to be the morpho‐ logical type of the given language. Whereas semantically distant words like successor and success do not prime each other in English (Marslen-Wilson et al. 1992: 87-88), the priming effect in languages with more complex morphology does not depend on the semantic component. Derivation in Arabic is based on vocalic alternations in triconsonantal roots of the shape C 1 _C 2 _C 3(_) . Boudelaa/ Marslen-Wilson (2015) show that the priming effect occurs even between semantically dissimilar forms derived from the same Arabic root. This also applies to Hebrew, which combines concatenative and non-concatenative morphology with triconsonantal roots - see Frost/ Forster/ Deutsch (1997). Plaut/ Gonnerman (2000) replicated these results with their connectionist model, in which an artificial language with rich morphology displayed a priming effect between morphemically related words with different meanings. On the other hand, a morphologically impoverished language lacked priming for such pairs. Given the complexity of AG morphology and a common word pattern of C 1 VC 2 roots with a derivational suffix, it is likely that a novel word like κῦδ-ος could activate its relative κυδ-ρός ‘glorious’ in the mental lexicon solely due to the formal resemblance of the base. It is also important to note that for a pair of morphemically related words, the semantic relation no meaning ~ meaning 1 could be less inhibitory than the relation of contrasting meaning 1 ~ meaning 2. The difference between the two pairs is that the first lacks an explicit semantic contrast between its terms, as the first term is underspecified. Experimental data show that lexical processing is also influenced by word family size. For Dutch, Schreuder/ Baayen (1997) found that words belonging to bigger families elicited shorter response times in lexical decision tasks. Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 165 <?page no="166"?> 4 Cf. Sandell (2015: 80-81) for a discussion on the most likely way that family size effects could operate in old Indo-European languages. This effect in Dutch, however, is semantically conditioned, as members with dissimilar meanings did not contribute to the family size. It is important to note that, in morphologically complex Hebrew, semantically distant members produce inhibitory effects within the family. According to del Prado Martín et al. (2005: 506), this interference shows that Hebrew speakers register formal affinities between related words, even if their meanings differ. 4 We assume the same for AG with a slight modification: formal relation could be felt between a novel word with no meaning and a word acquired before. Another hint that AG speakers were aware of morphemic patterns is supplied by AG accentuation. Probert (2006) observed that morphemically unanalysed nominal stems usually received a default recessive accent, while inherently accented suffixes like -ρόretained their accent if speakers identified them as such. Loss of morphemic transparency could lead to default recessive accentuation - see Probert (2006: 289-314) for a summary of the results. To summarize, typological and Greek data allow us to propose that AG speakers could tentatively match novel words with other derivations built on the same base. Comparably to Hebrew or Arabic, complex morphology of AG probably rendered the semantic component non-essential for the initial coupling. An interaction in the mental lexicon could occur due to the perceived similar morphemic make-up of the processed words. 3 Base transparency and its loss The initial coupling depends on the ability of speakers to recognize mor‐ phemically identical bases in different lexemes - see Cutler (1980, 1981) and Baayen (1993: 197-200) on the role of transparency in parsing. For any derived word, the degree of base transparency falls somewhere on a gra‐ dient between fully transparent and opaque. While full transparency can be universally defined as the presence of a formally identical base in different derivations of the same root, e.g. {φοβ(ε)-} in φόβος and φοβερός, its gradual shift towards opacity depends on language-specific and synchronic criteria. Formal differences between variants of the same base can be filtered out 166 Martin Masliš <?page no="167"?> 5 Probert (2006: 295) believes that there was marginal productivity of apophony in AG as late as the 5 th cent. BCE. only if they are perceived as alternations. Alternation, in turn, can be defined as a sufficiently frequent variation of the phonemic material in the base, so that it can usually be identified by speakers - see Meľčuk (2000: 512) for this definition of alternation. For the most part, AG derivational bases overlap at least partially with reflexes of Proto-Indo-European (henceforth PIE) roots. The basic shape of a PIE root was C 1 VC 2 with two obligatory consonants flanking a vowel, which displayed apophonic alternations *-e-, *-o-, and zero according to the inflectional and derivational type. Other shapes built upon this scheme by introducing a resonant on either side of the vowel or an initial group with a fricative *sor *h x ; the latter could also appear as a root final - see LIV 2 5-7. Apophonic alternations in AG roots with three or more phonemes were still common enough 5 to ensure that forms like φέρω ‘bear’ and φορός ‘bear‐ ing’ were recognized as morphemically related (e.g. in the Etymologicum magnum, EM 686.9). Other changes that, assumedly, could be filtered out were allophone variations. A similar mechanism applies to English pairs like vanity / vænɪti/ and vain / veɪn/ , where the terms prime each other despite the different syllable nucleus in the base - Marslen-Wilson et al. (1994: 29-30). We therefore assume that AG speakers could abstract away from the alternations provided that the outcomes were predictable and the overlap between variant shapes of the base was sufficient. This effect could extend beyond semantically similar words to novel lexemes being matched with their tentative relatives due to the potentially identical base. The difference between AG and English, where semantically disparate lexemes do not interact, would be conditioned by AG’s more complex morphology. Relying on the assumption that the phonemic make-up of a base in a specific derivation cues its identity, there appear to be three factors contri‐ buting to lower transparency. These are the small size of the word family, the effects of sound change, and the low number of phonemes in the base. Small word families or words with no base relatives naturally entail lower base frequency. A novel word whose base is less likely to occur in other derivations has a greater chance of staying isolated. If there are no relatives, complete isolation follows per definitionem. Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 167 <?page no="168"?> The effects of sound change differ from alternations in that context-sen‐ sitive diachronic processes can produce outcomes that are hardly predict‐ able by speakers on the synchronic level. Whereas, for example, the root *h 3 meyĝ h - ‘urinate’ appears in the full grade in ὀμείχω ‘id.’, its o-grade reflex {μοιχ-} in μοιχός ‘adulterer’ displays only a partial and rather unpredictable overlap with {ὀμειχ-} due to the Saussure effect (*h x RoC- > *RoC-) - see LIV 2 301-302; an incorrect etymology is provided by EM 589.35. Such changes produce base variants with lower cue validity in comparison with the bases containing just common alternations and allophones. Bases with fewer phonemes provide fewer cues as to their identity. Shorter bases are more likely to display chance resemblances with etymo‐ logically unrelated bases that cannot positively contribute to the semantic mapping. Forms like ὀδμή ‘smell’ from *h 3 ed- ‘start to smell (intr.)’ and ὀδ-ύνη ‘pain’ from *h 1 ed- ‘bite, eat’ contain different roots whose AG reflexes sometimes collapsed together - see EDG, 1047 and 1050. Such base homo‐ phony could facilitate semantic reanalysis due to folk etymologizing rather than help diachronically stabilize the meaning of a word. Lexemes with a shorter base are also more prone to leave the orbit of their etymological families because morphophonological processes affect a greater share of their phonemes. Reflexes of roots like *h 1 ey- ‘go’ provide no mutual overlap in their different apophonic variants in AG. Thus, bases in forms like εἶμι ‘go’ and οἶτος ‘fate’ ( ← ‘going’) do not share a common phoneme despite being derived from the same root - see DELG, 788. The exact threshold where formal intransparency tipped over into sub‐ jective base opacity cannot be precisely defined. For our purposes, however, it is sufficient to assume that the compounded effect of the factors mentioned above sometimes precluded the coupling of a novel word with its etymolog‐ ical relatives. Possible consequences of such lexical isolation are dealt with in the following section. 4 Formal isolation of novel words and the acquisition of semantics The mapping of meanings onto forms relies on several cognitive strategies. For example, children in the process of L1 acquisition first assume that a novel word signifies an object rather than an action and more likely a whole than its part. Other cues are picked up from the pragmatic context, including 168 Martin Masliš <?page no="169"?> 6 See Wagner (2010: 519-520) and Clark (1993: 9) for a discussion on the strategies employed during semantic mapping. the gaze direction of the person who utters the new word. The maxim of the mapping process is to assign a meaning to a word upon the first encounter. 6 Children aged between one and two can analyse the morphemic structure of lexemes and coin new words for referents whose conventional signs they have not yet learned. In English, this strategy utilizes compounding but also derivational suffixes like -er to form agent nouns - see Clark (1993: 39-41). During the acquisition of a language with complex morphology, children gradually register more morphemic items as their level of competence increases - see Clark (1993: 121). For Greek, this would amount to registering affixes and derivational bases while also learning that words sharing a base sometimes share a semantic component. The principles behind the process of coining can be applied to the acquisition of semantics. We have seen that speakers of Hebrew and Arabic recognize formal affinities between words, even if their meanings are not synchronically related. We propose that the same formal relation was felt in AG and that transparently related words provided cues to the meaning of novel words. Thus, besides forming new words from known building blocks, AG speakers could also deduce cues to the meaning of a novel word from its constituent morphemes if the communicative situation and the required contextual meaning did not exclude mapping along these lines. In the realm of languages with concatenative morphology, Sandra (1994: 260-264) discusses several studies that tested the import of morphemic segmentation in the process of L2 vocabulary learning, sometimes with non-sense words substituting real lexemes. According to the findings, ado‐ lescent and adult subjects retained new foreign vocabulary better if they were aware of the semantic motivation behind derived words with bases that also appeared elsewhere in the system. In addition, learning was faster for derived words with familiar bases in comparison to novel, also non-sense, monomorphemic words, regardless of whether the subjects received a hint to the semantic motivation for the former or not. According to Sandra, this suggests that, at least in L2 acquisition, the learning of derived lexemes is facilitated by awareness of the semantic motivation and that speakers resort to segmenting the input and deducing the motivation if the morphemic make-up allows this. We propose that the same mechanism Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 169 <?page no="170"?> 7 We adopt the notion of motivation from the structuralist works associated with the Prague linguistic circle and its later offshoots - e.g. Mathesius (1966 [ 1 1942]: 14-16) and later Dokulil (1962: 103-117). helped facilitate semantic mapping in L1 acquisition of AG. The difference between L2 and L1 acquisition is that in L2 the segmentation and the perceived motivation of a sign aid retention and speed up the learning process, while identifying the base in L1 provides cues to the meaning of a novel word because it highlights a feature of extralinguistic reality. If we assume that a new derived word partially takes its meaning from a derivational base whose semantics are already present in the system, it follows that the referent, or one of its salient features, motivates the presence of the base in the new word. 7 It is important to note that the link between the word and its referent must be established in the mental lexicon of every individual speaker anew. If a subjectively novel word carries a transparent cue (i.e. its base) to the correct referent, the space for semantic reanalysis is narrowed and the lexeme is not likely to radically change its original meaning. The semantic stability in this context does not mean that the meaning cannot change at all. It suggests that the mapped meaning tends to remain within the range of meanings reconcilable with the semantics of the base whose identity is recognized by speakers. A simple test for this kind of semantic stability would have speakers connect a word with other derivations with the same root or base. Lexemes with either too dissimilar meanings or too different shapes of their deriva‐ tional base would not occur to the speakers as being related. Such a test could delineate the range of possible meanings for a given base across a synchronic word family. Consequently, unspotted base relatives would supply examples of the forms not exposed to any semantic pressure from the word family. The meanings of such synchronically isolated but historically related forms undergoing reanalysis are more likely to end up beyond the boundaries of the semantic range compatible with the base. If the new reanalysed meaning spreads in the language community, the diachronic perspective registers this as an instance of semantic change. To return to the original proposal, the core hypothesis of this study presupposes that such non-trivial differences in the meaning of etymolog‐ ically related words in AG could be associated with diminished formal transparency. A synchronically opaque base, however, did not necessarily 170 Martin Masliš <?page no="171"?> 8 This can be described as de-etymologization, a notion which has some currency in Slavic linguistics, see Janyšková (2017). bring about a radical change in meaning. It is quite possible that an isolated word could retain its original semantics or remain within the boundaries of the semantic stability, the disparate shape of its base being the only disqualifying factor precluding membership in the word family. 8 Likewise, an identical shape of the base in words with radically different meanings is not excluded either. The hypothesis merely aims to explain the data in a probabilistic manner rather than by applying a strictly causal framework. Our approach to semantic change agrees with Fortson (2003), who stresses the importance of reanalysis over the cognitive categorization of changes into metaphors, metonymies, extensions, specializations, and other types - see Blank (1999: 70-82), Campbell (2013: 222-232), Fortson (2003: 648-650), and Traugott (2017) for common typologies and general accounts of semantic change. According to Fortson, different cognitive strategies speakers use when employing words with defined meanings in a particular context to express concepts are important only inasmuch as they provide input for possible reanalysis. Organic acquisition of lexical items centres around abstracting the meaning of a sign based on the cues present in the communicative situation. It follows that the intended referent is somehow present in the context, but a speaker not familiar with the word can reconstruct anything reconcilable with the communicative situation. For example, Eng. bead is related to Ger. Gebet ‘prayer’, the original context of the reanalysis in English being the use of rosary prayer beads. Thus, syn‐ chronic isolates do not invite an analysis which would lay bare a particular cognitive or onomasiological strategy that led to the word being used in the given context. Such a linguistic sign is then felt as an arbitrary token whose meaning needs to be reconstructed from scratch. Even if the newly acquired meaning encompasses the same referent as was initially intended, any possible motivation is lost. The newly defined meaning can then be extended further without any stabilizing effects that synchronic relatives could provide. For example, AG δῆμος, structurally signifying ‘division, dividing’ (cf. δαίομαι ‘divide’, see EDG, 325), came to mean ‘district’ ( ← ‘a division of land’) before acquiring its common classical sense ‘the people’. From the diachronic perspective, such a scenario represents a series of metonymies that were petrified thanks to reanalysis and loss of motivation. Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 171 <?page no="172"?> 5 Methods The hypothesis about the semantic instability of the formally isolated words in AG was experimentally tested on two datasets. The first dataset comprises 82 simplex words formed with the suffix -μο-; the second includes 102 simplex words with the suffix -ρο-. Each dataset was split into two groups of potentially transparent and potentially opaque lexemes, respectively. The words were culled from Buck/ Petersen (1945). The datasets are appended at the end of the study. The suffix -μοwas the reflex of the PIE *-mothat derived action nouns, or instantials, and less numerous primary adjectives in the protolanguage - see Chantraine (1933: 132-147), Buck/ Petersen (1945: 184-185), Nussbaum (2014: 249), and Probert (2006: 238-239). The Greek -μοremained produc‐ tive in the historical period, forming deverbal action nouns and secondary adjectives. Besides the simple suffix, our dataset also includes words with the extended variant -αμο-. Lexemes containing -αγμοand -ισμοderived from verbs in -άζω and -ίζω respectively have been excluded because of their exceptional productivity, which could skew the data. Lexemes with the synchronic -σμοwere retained only if the -σbelonged to the root morpheme or replaced a root-final alveolar in the sequence *-D-smo-. The extended -θμοhas been excluded as well. The dataset contains adjectives and masculine nouns in -μος. The suffix -ροwas the reflex of the PIE *-rothat formed adjectives and, to a lesser extent, deverbal nouns in the protolanguage - see again Chantraine (1933: 221-236), Buck/ Petersen (1945: 311-315), Nussbaum (1976: 108-113), and Probert (2006: 155-158) for the properties of the PIE *-ro-, its position in the ‘Caland System’, and how this carried over to its AG reflex -ρο-. Formation of adjectives in -ροwas still productive in AG, with new words usually being built on derived bases like the -ηand -(ι)αstems (ἀνιαρός ‘troublesome’ from ἀνία ‘trouble’), thematic ο-stems (φοβερός ‘fearful’ from φόβος ‘fear’), -ε(ϝ)-/ -υstems (λιγυρός ‘shrill’ from λιγύς ‘shrill’), and the -ος/ -εσstems (στυγερός ‘hateful’ from στύγος ‘hatred’). In the protolan‐ guage, adjectives in *-rocould be built directly on a root and usually appeared alongside other derivations with the PIE suffixes *-i-, *-u-, *-ont-, *-es-, and *-eh 1 -. These formed the so-called ‘Caland System’, whose reflexes were present in AG, but the system as a whole was no longer productive in the daughter language. 172 Martin Masliš <?page no="173"?> 9 With the exception of the classical ζ-, whose biphonemic value / zd/ reflects *y-. 10 In this framework, trivial changes do not obfuscate the motivation of a linguistic sign and speakers are aware, at least intuitively, of this motivation, even if the semantics do not completely agree with the structural meaning of the morphemic chain. Even though it is difficult to assess how native speakers of AG perceived the motivation behind individual words, the Late Antique etymologica and the employment of figura etymologica in texts provide hints as to which words were synchronically transparent. The lexemes included in the -ροdataset also cover words with the extended suffixes -αρο-, -ερο-, and -υρο-. Derivations from the thematic ο-stems in -εροand the type in -ηροhave been excluded, as these were productive and closely connected with their source. Thematized forms of the heteroclitic *-r/ nstems have been included because the derived forms could appear as synchronic -ροstems, especially if the source heteroclitic noun did not survive. The dataset contains adjectives, masculine and feminine nouns in -ρος, and neuters in -ρον. The groups of potentially transparent lexemes comprise words whose base contained at least three AG phonemes, without the initial aspiration, after stripping off the derivational suffix. Two groups of potentially opaque lexemes include words whose base had no more than two AG phonemes 9 or words without any surviving relatives in AG. The meaning of each lexeme has been compared with the semantics of the base in other etymologically related derivations. If the meaning of a word did not seem closely connected with the meaning of its relatives and it could be established that the disparate meaning was a Greek innovation, the word has been marked as having undergone a semantic change. To avoid an artificial skew in favour of our hypothesis, only words displaying non-trivial development 10 have been marked as changed in the opaque groups. For the same reason, the criteria were less strict for the potentially transparent groups. The data on word families and prior meanings stem from the three standard etymological dictionaries of Greek, GEW, DELG, and EDG. The final authority, however, were LIV 2 and NIL. The lexemes that acquired disparate semantics already in PIE have not been labelled as changed. The words without a plausible etymology have been excluded from the datasets. When available, subjective assessment of the semantic differences has been corroborated using the Etymologicum magnum (EM). Although this lexicographical work was compiled in the mid-12 th cent., the data stem from Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 173 <?page no="174"?> 11 The same approach is adopted in Probert (2006: 253-256) to assess synchronic analyzability of nominals. earlier etymologica like the Etymologicum Gudianum (the 10 th cent.), the Etymologicum genuinum (the 9 th cent.), and Orion’s Etymologicum (the 5 th cent.), all of which utilize still earlier scholia and lexica, sometimes building upon Hellenistic scholarship - see Alpers (1990: 27-32), Reitzenstein (1905: 15-16), and Dickey (2015: 472) for the details of their textual transmission; Sluiter (2015) and Zucker/ Le Feuvre (2021) discuss the cultural and intellec‐ tual background of ancient etymologizing. Even if some of the observations are too late to reflect the intuitions of AG speakers in the classical and pre-classical period, lemmata in EM come very close to the test we have envisaged in Section 4. Thus, if EM does not link a word with its base or root relatives, excluding the words that are derived from the observed word itself, we consider the lexeme to be positively isolated, i.e. either its form or meaning were disparate enough to preclude the connection with the appropriate word family. 11 If a seemingly disparate meaning arisen through an innovation coincided with a wrong analysis in EM, the change of meaning has been considered non-trivial. A correct analysis of a word in EM usually meant that a non-trivial change has been ruled out. To return to the example of δῆμος, the lemma in EM illustrates that speakers failed to identify the motivation for the sense ‘the people’; the same was very probably the case for the sense ‘division of land’ even with the earliest occurrences in the epics. The change is easier to be reckoned with if we suppose that the label δῆμος with the word sense ‘division (of land)’ lost the cue structure that would otherwise point speakers to the feature that initially motivated this derivate to be used as a sign for that particular concept, namely a division of a whole. (1) <δῆμος>: τὸ πλῆθος· παρὰ τὸ δέω, τὸ δεσμεύω. […] ὁ συνδεδεμένος ὄχλος […]. Ἢ παρὰ τὸ δέμας, δῆμος, οἷον ἓν σῶμά τι ὑπάρχων. Ἢ δίδημι, τὸ δεσμῶ. - “<δῆμος>: a multitude; it comes from δέω ‘bind’ and refers to the united people. Alternatively, it could come from δέμας ‘a human body’, as if it signified something like a body. It may also come from δίδημι ‘fetter’.” (EM 264.41) 174 Martin Masliš <?page no="175"?> In the case of αἰσχρός ‘shameful’ and the other primary -ροadjectives listed under the same lemma, the presented etymologies are correct except for σαθρός. This indicates that speakers were aware of the lexical neighbour‐ hood of the adjectives and that we can rule out a non-trivial semantic change with a reasonable degree of certainty. (2) <αἰσχρός>: Ὁ αἰσχύνης μετέχων. αἶσχος, αἰσχρὸς, ὡς κῦδος κυδρὸς, ψύχος ψυχρός· καὶ ἀπὸ τοῦ θηλυκοῦ σάθη, σαθρός· καὶ λύπη, λυπρός. Οὐ γὰρ ἀπὸ τοῦ αἰσχηρὸς συγκέκοπται. - “<αἰσχρός>: someone who shares in shame. αἶσχος ‘shame’, αἰσχρὸς ‘shame‐ ful’, like κῦδος ‘glory’, κυδρός ‘glorious’, ψύχος ‘cold’ (subst.), ψυχρός ‘cold’ (adj.); also, from the feminine σάθη ‘penis’, σαθρός ‘unsound’; also λύπη ‘grief ’, λυπρός ‘wretched’. It does not come from αἰσχηρός through a syn‐ cope.” (EM 39.53) The data from EM, however, were not accepted without exceptions. Occa‐ sionally, a word analysed correctly by EM has been labelled as having undergone a change if the meaning had been deemed sufficiently disparate. The mismatch applied mostly to the groups of potentially transparent words where the semantic criteria were applied more loosely to avoid accidentally lowering the number of words exhibiting a change. An explanation for correct analyses in EM could also be the nature of the concentrated effort of a lexicographer that can outperform natural associative processes in the mental lexicon. We used the one-sided Fisher’s exact test to see if the presence of a semantic change was independent of the base transparency in each data set. The results are reported in the following section. 6 Results Out of 184 lexemes in the data, 135 have a corresponding lemma or note in EM. The performance of the lexicographers agrees to a great extent with the prediction made in Section 4. A word preliminarily labelled by us as semantically disparate was usually not connected to its etymological relatives in EM. On the other hand, lexemes labelled as not innovated were mostly analysed correctly. Out of 135 EM lemmata, 14 must be set Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 175 <?page no="176"?> 12 κόσμος (12), δοχμός (30), ἄνεμος (13), κάλαμος (31); ἀγρός (16), ὑγρός (18), ἱερός (20), μηρός (23), παῦρος (25), ταῦρος (27), πυρός (28), τυρός (29), ζωρός (10), πικρός (65). 13 οἶμος (19), κρυμός (68), ἰωχμός (80); σφοδρός (52), βόθρος (101). 14 ἀτμός (9), ἑσμός (8), ὅρμος (7), ἰταμός (33), κορμός (36), ψωμός (40), φορμός (38), πότμος (39); ἱλαρός (2), στιβαρός (34), θαλερός (37), κάπρος (38), ζέφυρος (40). aside 12 because the words had no comparable relatives in AG; thus, a correct analysis could not be expected. Of the remaining 121 lemmata (see Table 1), 103, or 85.12%, agree with the expected data structure. The remaining 18 lemmata break down into 5 that were not analysed correctly despite their stable meaning, 13 and 13 that were successfully connected with their relatives despite disparate semantics. 14 Data structure EM lemmata - no change + correct EM 80 Expected change + incorrect EM 23 Expected no change + incorrect EM 5 EM underachieved change + correct EM 13 EM overachieved Table 1: Lemmata in the Etymologicum magnum Out of 13 lexemes that were unexpectedly analysed correctly, 9 belong to the groups of potentially transparent words. This means that even if our supposed instances of EM overachieving could be explained by too loose criteria for semantic change, re-labelling all 13 words as not changed would only strengthen our alternative hypothesis, as it would disproportionally increase the rate of the semantically stable words for the groups of the potentially transparent lexemes. On the other hand, the instances of EM underachieving can be explained as mere mistakes or as the lexemes not participating in a family solely due to formal discrepancies. Table 2 contains two contingency tables placed adjacently. Its left half comprises the data of the -μοdataset: the first column shows the distribu‐ tion of semantic (in)stability in the group of potentially opaque lexemes, the second column shows the same for the group of potentially transparent words. The right half of Table 2 presents the -ροdataset in the same manner. Each contingency table has been evaluated by applying the one-sided Fisher’s exact test with the null hypothesis that the potentially opaque words 176 Martin Masliš <?page no="177"?> changed their meaning as much as or less than the potentially transparent words; our alternative hypothesis is that the opaque words changed more often than the transparent ones. - -μοpoten‐ tially opaque -μοpotentially transparent -ροpoten‐ tially opaque -ροpotentially transparent change 13 9 12 13 no change 19 41 21 56 Table 2: Contingency tables of the datasets For the suffix -μο-, the p-value of the test is 0.023, meaning that observing our data if the null hypothesis were true would occur only with a 2.3% chance. The odds ratio is 3.07 with a 95 % confidence interval of 1.19 to ∞ . Thus, the observed data for -μοindicate that the words that have undergone a non-trivial semantic change are about three times as likely to belong to the group of potentially isolated rather than potentially transparent words. The p-value under 0.05 and the confidence interval not including 1 suggest that the outcome is statistically significant and that our hypothesis explains the data better than the null hypothesis. A similar picture emerges for the suffix -ρο-. Here, the p-value is 0.048, the odds ratio equals 2.44, while its 95 % confidence interval ranges from 1.08 to ∞ . These results are likewise statistically significant and support the alternative. 7 Discussion and concluding remarks The quantitative analysis has shown that the proposed hypothesis about low base transparency facilitating semantic change adequately describes the structure of the datasets. The argumentation developed in Sections 2 to 4 employs the typological and psycholinguistic perspective to provide a plausible explanation for this effect. Speakers of Ancient Greek (AG) could register relations between words due to formal and semantic similar‐ ities. The complex concatenative morphology of AG probably ensured that speakers felt such relations, even if the semantic component did not directly contribute to the linking. A stronger form of this effect has been observed in Hebrew and Arabic, where matching overrides even dissimilar semantics. In the case of novel AG words, the initial lack of semantic properties helped Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 177 <?page no="178"?> 15 This research is financially supported by the Charles University Research Centre program No. 204053 and the Charles University Grant Agency (GA UK), grant number 424420 “Semantic (in)stability of formally isolated lexemes exemplified on Ancient Greek” [“Sémantická (ne)stabilita formálne izolovaných slov na príklade starej gréč‐ tiny”] awarded at the Faculty of Arts, Charles University. facilitate their connection with tentative etymological relatives because the formal criteria of matching could apply without any inhibition stemming from disparate meanings. Whether such word matching could operate in AG even against inhibitory semantic effects like in the template systems of Hebrew and Arabic remains an open question. For our purposes, even the weaker form of this effect in AG would suffice to support our argumentation. The novel AG words tentatively connected with their probable relatives carried more cues as to their intended referents than the formally isolated lexemes. The transparent derived words pointed speakers to the salient ex‐ tralinguistic feature that initially motivated the presence of the derivational base in the word. From the speakers’ perspective, this relation reduced the arbitrariness of the linguistic sign. The semantics of the recognized base then limited the space for any possible reanalysis. This mechanism, however, could function only if speakers recognized the base and found a plausible motivation for its presence. The motivation could be lost because of unspecified cultural changes or if some other salient feature of the referent ousted the one that initially motivated the derivational base. The latter can be illustrated by φορμός ‘basket’ ( ← ‘carrying’) coming to signify ‘mat’ because the perceived salient feature shifted from the act of carrying to the material aspects of plaited wickerwork. This suggests that speakers could sometimes disregard the cue provided by the word itself if the apparent referent forced a more acceptable defining feature. Similarly, a formally opaque base provided no cues at all and the reanalysis could run its course unchecked. The described mechanism was probabilistic rather than strictly causal. Using the one-sided Fisher’s exact test, we have confirmed our hypothesis that the derived AG nominal lexemes with no synchronic relatives or no rec‐ ognized derivational neighbourhood were prone to change their meaning more often than the words built on transparent bases. 15 178 Martin Masliš <?page no="179"?> 16 The note on the author in the Lexeme column indicates the first occurrence, a ‘+’ sign indicates subsequent occurrences; a ‘+’ sign in the Sem. change column stands for a non-trivial semantic change; if available, the EM column supplies the lemma in the Etymologicum magnum, and a ‘+’ sign indicates correct etymology. 8 Appendix 16 Dataset of the words in -μο-: • potentially opaque lexemes # Lexeme Sem. change Etymology EM 1 κύαμος, ὁ ‘bean’; Hom. Il. + + κυέω ‘be pregnant’; GEW, Ι, 37 N/ A 2 δῆμος, ὁ ‘district, land, the people’; Hom. Il. + + δαίομαι ‘divide’; GEW, I, 381 264.41 − 3 λιμός, ὁ ‘hunger’; Hom. Il. + + λιάζομαι ‘incline, sink, fall’; DELG, 641; for the root *leyh 2 - ‘cease, vanish’ see LIV 2 406 568.20 − 4 λοιμός, ὁ ‘plague’; Hom. Il. + + λιάζομαι ‘incline, sink, fall’; DELG, 641; for the root *leyh 2 - ‘cease, vanish’ see LIV 2 406 568.20 − 5 σιμός ‘snub-nosed’; Xenoph. + + *tih 2 -mó- ‘struck’; also σῆμα ‘mark’, σῶμα ‘corpse’, σῖτος ‘grain’; Janda (2005) reconstructs a root *tyeh 2 - ‘strike’ 713.15. − 6 ὅλμος, ὁ ‘bowl-shaped body, trunk’; Hom. Il. + + εἰλέω ‘turn’; EDG, 1070 622.53 − 7 ὅρμος, ὁ ‘cord, necklace, anchorage, haven’; Hom. Il. + + εἵρω ‘string’; EDG, 1105- 1106 631.30 + / ( − ) 8 ἑσμός, ὁ ‘swarm’; Α. + + either ἕζομαι ‘sit down’ or ἵημι ‘set forth’; DELG, 378 383.9 + (ἵεμαι) / ( − ) 9 ἀτμός, ὁ ‘vapour’; Α. + + ἄημι ‘blow (of winds)’; EDG, 164 164.40 + 10 θυμός, ὁ ‘soul, spirit’; Hom. Il. + + *d h uh x -mó- ‘smoke’; also θυμιάω ‘fumigate’; EDG, 564 458.7 − Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 179 <?page no="180"?> # Lexeme Sem. change Etymology EM 11 ζωμός, ὁ ‘soup, sauce’; Ar. + + perhaps to ζέω ‘boil’ from *yes- ‘boil’, cf. EDG, 504 and Hackstein (2002: 191) for the phonology; probably not from PIE *y(e)uh x s- ‘soup, broth’, NIL, 405 n. 1 N/ A 12 κόσμος, ὁ ‘order, pro‐ priety, good behaviour, ornament’; Hom. Il. + + no relatives in AG; *k̂ eNs- ‘proclaim’; LIV 2 326, EDG, 759 532.10 − 13 ἄνεμος, ὁ ‘wind’; Hom. Il. + + no relatives in AG; *h 2 enh 1 - ‘breathe’; LIV 2 267 103.37 − 14 οὐλαμός, ὁ ‘throng’; Hom. Il. + − εἱλέω ‘press together, col‐ lect, huddle’; EDG, 1125 640.29 + 15 ἀγμός, ὁ ‘fracture, cleft, cliff’; Hp. + − ἄγνυμι ‘smash’; EDG, 14 N/ A 16 ὄγμος, ὁ ‘furrow, swa‐ the’; Hom. Il. + − ἄγω ‘lead, drive’; EDG, 1045 613.34 + 17 ὠγμός, ὁ ‘a cry of oh! ’; A. + − ὤζω ‘cry oh! ’ N/ A 18 δειμός, ὁ ‘fear’; Hom. Il. + − δείδω ‘be afraid’; EDG, 310 261.40 + 19 οἶμος, ὁ ‘way, path, tract’ Hes. + − either to ἵεμαι ‘hasten’ from *weyh 1 - ‘strive for’ or εἶμι ‘go’ from *h 1 ey- ‘go’; cf. DELG, 784, EDG, 1057-1058, LIV 2 668, 232- 233 617.55 − 20 ἁρμός, ὁ ‘joint, bolt, peg’ Ε. + − ἀραρίσκω ‘fit together’; EDG, 123. 144.42 + 21 εἱρμός, ὁ ‘train, series’ Arist. + − εἵρω ‘string’ 103.45 + 22 ᾀσμός, ὁ ‘song’; Pl. Com. − ἀείδω ‘sing’ Ν/ Α 23 ὠσμός, ὁ ‘pushing’; D. S. − ὠθέω ‘push’; EDG, 1676 N/ A 24 καυμός, ὁ ‘fever’; Call. − καίω ‘burn’ N/ A 25 ῥυμός, ὁ ‘pole of a cha‐ riot, trail’; Hom. Il. + − ἐρύω ‘pull, draw’; EDG, 468 706.35 + 26 χυμός, ὁ ‘humour’; Hp. + − χέω ‘pour’; EDG, 1653 265.8 + 27 ὄχμος, ὁ ‘fortress’; Lyc. − ἔχω ‘hold, have’ 645.41 + 180 Martin Masliš <?page no="181"?> # Lexeme Sem. change Etymology EM 28 βωμός, ὁ ‘altar, plat‐ form’; Hom. Il. + − ἔβην ‘stepped’, *g w eh 2 - ‘step’; LIV 2 205 217.48 + 29 θωμός, ὁ ‘heap’; A. + − τίθημι ‘put, stack’; EDG, 569 459.52 + 30 δοχμός ‘oblique’; Hom. Il. + − no relatives in AG; cf. Skt. jihmá- ‘slanted’; EDG, 350 285.15 − 31 κάλαμος, ὁ ‘reed’; Hdt. + − no relatives in AG; cf. Lat. culmus ‘stalk’, Old High German halm ‘stalk’, Ru. solóma ‘reed, straw’; EDG, 622 485.27 − 32 ἠθμός, ὁ ‘strainer’; E. + − ἤθω ‘strain’ 422.33 + • potentially transparent lexemes 33 ἰταμός ‘headlong, hasty, eager’; A + + base ἰτα-, semantically contrastive to εἶμι ‘go’, de‐ rived from ἴτης ‘headlong, hasty’; EDG, 604 479.37, 43.37 + 34 ποταμός, ὁ ‘river’; Hom. Il. + + πέτομαι ‘fly’; EDG, 1225- 1226 685.4 − 35 τριμμός, ὁ ‘worn road’; X. + + τρίβω ‘rub, grind, wear down’; EDG, 1508-1508 N/ A 36 κορμός, ὁ ‘branchless trunk’; Hom. Od. + + κείρω ‘cut short, clip’ 141.28 + 37 τορμός ὁ ‘socket, nave, peg’; Hdt. + + Hom. ἔτορε ‘pierced’; EDG, 1495 N/ A 38 φορμός, ὁ ‘basket, mat, plaited cloak’; Hes. + + φέρω ‘bear’; EDG, 1587 798.55 + 39 πότμος. ὁ ‘lot, destiny’; Hom. Il. + + πίπτω ‘fall’; EDG, 1195 685.30 + 40 ψωμός, ὁ ‘morsel’; Hom. Od. + + inf. ψῆν ‘rub, grate, scratch’; EDG, 1665 819.45 + 41 σκαλμός, ὁ ‘thole’; h.Hom. 7 + + σκάλλω ‘stir up, hoe’ 715.22 − 42 πλόκαμος, ὁ ‘lock of hair, braid’; A. + − πλέκω ‘braid, knit’; EDG, 1206-1207 677.7 + 43 φραγμός, ὁ ‘enclosing’; S. + − φράσσω ‘fence in, en‐ close’; EDG, 1588-1589 613.36 + Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 181 <?page no="182"?> 44 δηγμός, ὁ ‘bite, sting, gnawing pain’; Hp. + − δάκνω ‘bite, sting’; EDG, 299 N/ A 45 κρεγμός, ὁ ‘sound of a stringed instrument’; Epich. + − κρέκω ‘strike a stringed instrument’ N/ A 46 ἑλιγμός, ὁ ‘winding’; Hdt. + − ἑλίσσω ‘turn around’ N/ A 47 φλογμός, ὁ ‘flame, glow, inflammation’; E. + − φλέγω ‘ignite, burn’ 795.48 + 48 λυγμός, ὁ ‘sobbing, loud cry’; Hp. + − λύζω ‘sob, have a hiccup’ N/ A 49 μυγμός, ὁ ‘moaning’; A. + − μύζω ‘mutter, moan’ 572.48 + 50 βρυγμός, ὁ ‘biting, gobbling, chattering, gnashing of teeth’; Hp. + − βρύκω ‘bite, eat greedily’ 215.46 + 51 διωγμός, ὁ ‘pursuit’; X. + − διώκω ‘pursue, drive away’ 481.34 + 52 κλωγμός, ὁ ‘clucking sound’; X. + − κλώζω ‘cackle’ 322.52 + 53 ἀρδμός, ὁ ‘means of wa‐ tering’; Hom. Il. + − ἄρδω ‘irrigate’ 137.43 + 54 παλμός, ὁ ‘quivering mo‐ tion’; Alciphr. + − πάλλω ‘sway, rock, poise’ 648.17 + 55 ψαλμός, ὁ ‘twitching, twanging’; E. + − ψάλλω ‘pluck, twitch, play a stringed instrument’ 817.20 + 56 τιλμός, ὁ ‘plucking of hair’; A. + − τίλλω ‘pluck hair’ 817.23 + 57 στολμός, ὁ ‘equipment’; E. + − στέλλω ‘make ready, fit out, dispatch’ N/ A 58 κομμός, ὁ ‘striking, dirge’; A. + − κόπτω ‘strike, smite’ 526.39 + 59 πταρμός, ὁ ‘sneezing’; Hp. + − πτάρνυμαι ‘sneeze’ N/ A 60 ἀγερμός, ὁ ‘collection of money, call-to-arms’; Arist. + − ἀγείρω ‘collect, gather’ N/ A 61 θερμός ‘warm’; Hom. Il. + − inherited, cf. θέρομαι ‘be‐ come warm’; EDG, 542 447.15 + 62 ὀδυρμός, ὁ ‘lamentation’; A. + − ὀδύρομαι ‘lament, bewail’ N/ A 182 Martin Masliš <?page no="183"?> 63 συρμός, ὁ ‘sweeping mo‐ tion’; Pl. + − σύρω ‘draw, drag along’ N/ A 64 φυρμός, ὁ ‘mixture’; D. S. + − φύρω ‘mingle, confuse’ N/ A 65 δασμός, ὁ ‘division of spoil’; Hom. Il. + − δατέομαι ‘divide’; EDG, 305-306 249.3 + 66 σεισμός, ὁ ‘shaking’; E. + − σείω ‘shake’ N/ A 67 δρυμός, ὁ ‘bush’; S. + − δρῦς ‘tree, oak’; EDG, 355- 356 228.49 + 68 κρυμός, ὁ ‘frost’; Hdt. + − κρύος ‘cold, frost’ 542.1 − 69 ἔτυμος ‘true’; S. + − ἐτε(ϝ)ός ‘real, true’; EDG, 474-475 388.1 + 70 πλοχμός, ὁ ‘lock of hair’; Hom. Il. + − πλέκω ‘braid, knit’ 645.41 + 71 μυχμός, ὁ ‘moaning, groaning’; Hom. Od. + − μύζω ‘mutter, moan’ N/ A 72 ὀδαγμός, ὁ ‘irritation’; S. − ὀδάξομαι ‘feel irritation’ N/ A 73 ἄργεμος, ὁ ‘white speck on the eye’; S. + − ἀργός ‘shining, white’ 135.37 + 74 ἀλοιμός, ὁ ‘polishing, plastering’; S. + − ἀλείφω ‘anoint, polish, plaster’ 69.42 + 75 ἰνδαλμός, ὁ ‘form’; Hp. + − ἰνδάλλομαι ‘appear, seem’ 471.40 + 76 σκυλμός, ὁ ‘rending’; LXX. + − σκύλλω ‘tear, maltreat’ N/ A 77 θλιμμός, ὁ ‘pressure’; LXX. − θλίβω ‘press’ N/ A 78 ἐρετμός, ὁ ‘rowing’; Hdn. + − ἐρέσσω ‘row’ 370.57 + 79 ἀμυχμός, ὁ ‘sword-wound’; Theoc. + − ἀμύσσω ‘scratch, tear, lac‐ erate’ N/ A 80 ἰωχμός, ὁ ‘rout’; Hom. Il. + − ἰωκή ‘rout’ 481.33 − 81 ῥωχμός, ὁ ‘cleft’; Hom. Il. + − ρήγνυμι ‘break’ 281.12 + 82 βρῶμος, ὁ ‘food’; Arat. − βιβρώσκω ‘eat up’ 201.17 + Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 183 <?page no="184"?> Dataset of the words in -ρο-: • potentially opaque lexemes # Lexeme Sem. change Etymology EM 1 ἰθαρός ‘cheerful, nim‐ ble’; Alc. + + *h 2 eyd h - ‘kindle’, cf. αἴθω ‘kindle’; LIV 2 259, EDG, 582 N/ A 2 ἱλαρός ‘cheerful’; Ar. + + ἵλαος ‘gracious, propi‐ tious’, ἱλάσκομαι ‘ap‐ pease’, *selh 2 - ’become merciful’; EDG, 586-587, LIV 2 530 470.154 + 3 ὕδερος, ὁ ‘dropsy’; Arist. + + no relatives in AG, cf. Lat. uterus ‘belly, womb’, Skt. udára- ‘(swollen) belly’; EDG, 1526 N/ A 4 ἁδρός ‘thick, stout, strong, violent’; Hdt. + + ἄδην ‘to one's fill’, *seh 2 (y)- ‘become sati‐ ated’; EDG, 21-22, LIV 2 520 18.45 − 5 ἵμερος, ὁ ‘longing’; Hom. Il. + + *sih 2 meró- ‘spellbinding’ to *sh 2 ey- ‘bind’, cf. ἱμάς ‘thong, rope’; EDG, 591, Weiss (1998: 47-56), LIV 2 544 467.8 − 6 λῆρος, ὁ ‘nonsense’; Ar. + + *leh 2 - ‘bark’, cf. λαίειν·φθέγγεσθαι Hsch.; EDG, 858, LIV 2 400 564.15 − 7 χῆρος ‘bereaved, empty’; E. + + *ĝ h eh 1 - ‘leave, abandon’, cf. also χώρα ‘land’; EDG, 1630-1631, LIV 2 173 811.29 − 8 ἄκρος ‘topmost’; Hom. Il. + + *h 2 ek̂ - ‘be/ make/ become sharp’, cf. ὄκρις ‘jagged point’; EDG, 57, NIL, 294 n. 21, LIV 2 261 53.6 − 9 οὖρος, ὁ ‘fair wind’; Hom. Od. + + *h 1 osu-ro- ‘bringing good’; Nussbaum (1998: 152) 642.27 − 10 ζωρός ‘pure (of wine)’; Α. R. + + no relatives in AG, *yeh 3 -roor *yoh x -ro-, cf. OSC jarŭ ‘strong, hard, se‐ rious’; EDG, 505 414.34 − 184 Martin Masliš <?page no="185"?> # Lexeme Sem. change Etymology EM 11 χῶρος, ὁ ‘place’; Hom. Il. + + *ĝ h oh 1 -rofrom *ĝ h eh 1 - ‘leave, abandon’; EDG, 1654, LIV 2 173 813.53 − 12 λαρός ‘pleasant to the taste, sweet’; Hom. Il. + + ἀπο-λαύω ‘enjoy, have the benefit’; EDG, 835 557.6 − 13 πιαρός ‘fat’; Hp. + − πῖαρ ‘fat’ N/ A 14 μιαρός ‘polluted’; Hom. Il. + − μιαίνω ‘stain’ 274.6 + 15 ναρός ‘flowing’; A. + − νάω ‘flow’; EDG, 1000 597.43 + 16 ἀγρός, ὁ ‘field’; Hom. Il. + − inherited, cf. Lat. ager ‘field’ and Skt. ájra- ‘id.’ from *h 2 eĝ-ro-; EDG, 16 13.30 − 17 ὕδρος, ὁ ‘water-snake’; Hom. Il. + − inherited, *ud-ró- ‘water animal, otter’; EDG, 1526 465.41 + (indirect) 18 ὑγρός ‘wet, fluid’; Hom. Il. + − no relatives in AG; from *ug w -róto the root *weg w - ‘make wet’; EDG, 1525- 1526, LIV 2 663 n. 1 774.48 − 19 δαερός ‘hot, burnt’; Emp. − δαίω ‘burn’; EDG, 298-299 246.36 + 20 ἱερός ‘supernatural’; Hom. Il. + − *h 1 eysh 2 - ‘strengthen, urge forward’, the seman‐ tics of the adj. seem to be inherited, cf. Skt. iṣirá- ‘strong (of deities); EDG, 580-581, LIV 2 234 468.8 − 21 διερός ‘alive, wet’; Hom. Od. + − διαίνω ‘moisten’ 274.1 + 22 δηρός ‘lasting long’; Hom. Il. + − δήν ‘for a long while’; EDG, 326-327 N/ A 23 μηρός, ὁ ‘fleshy thigh’; Hom. Il. + − inherited, *mems-ro- ‘meaty’ from *mé(m)s- ‘flesh’, cf. Lat. membra ‘limbs’; EDG, 947, NIL, 486 and 488 585.23 − 24 γαῦρος ‘exulting in, haughty’; Archil. + − γάνυται ‘be happy’ from *geh 2 u- ‘be happy’; LIV 2 184 222.34 + 25 παῦρος ‘small’; Hes. + − inherited, *peh 2 u-ro-, cf. Lat. parvus ‘small’; EDG, 1158 653.32 − Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 185 <?page no="186"?> # Lexeme Sem. change Etymology EM 26 αἶθρος, ὁ ‘chill morning air’; Hom. Od. + − αἴθρη ‘clear sky’; EDG, 36 107.31 +, 339.28 + 27 ταῦρος, ὁ ‘bull’; Hom. Il. + − inherited, *teh 2 u-ro-, cf. Lat. taurus ‘bull’; EDG, 1455-1456 747.38 − 28 πυρός, ὁ ‘wheat’; Hom. Il. + − inherited, *puh x -ro-, cf. Lith. pūraī ‘winter corn’, Ru. pyréj ‘dog-grass’; EDG, 1263-1264 697.47 − 29 τυρός, ὁ ‘cheese’; Hom. Il. + − inherited, *tuh 2 -rofrom *teuh 2 - ‘swell, become strong’, cf. Av. tūiri-, n. ‘cheeselike milk’; EDG, 1520, LIV 2 639-640 772.5 − 30 ὀχυρός ‘firm, lasting’; Hes. + − ἔχω ‘have, hold’ 405.15 + 31 ἐχυρός ‘strong, secure’; Th. + − ἔχω ‘have, hold’ 405.15 + 32 δῶρον, τό ‘gift’; Hom. Il. + − inherited, *déh 3 -ro-, Old Church Slavonic darŭ ‘gift’; EDG, 363 293.23 + 33 πτερόν, τό ‘feather, wing’; Hom. Od. + − πέτομαι ‘fly’ 694.11 + • potentially transparent lexemes 34 στιβαρός ‘strong, sturdy’; Hom. Il. + + στείβω ‘tread, trample’; EDG, 1393-1394 727.42 + 35 λαγαρός ‘hollow, sunken, slack, emaciated’; X. + + λαγαίω ‘release, let go’ (Cret.), λαγωός ‘hare’ (i.e. ‘slack-ear’), *sleĝ- ‘slack’; EDG, 819, LIV 2 565 192.36 − 36 τάλαρος, ὁ ‘basket, wicker cage’; Hom. Od. + + cf. τελαμών ‘bearing strap’, *telh 2 - ‘bear’; EDG, 1445-1446 744.56 − 37 θαλερός ‘blooming, fresh, big, stout, sturdy’; Hom. Il. + + θάλλω ‘sprout, grow’ 441.32 + 38 κάπρος, ὁ ‘boar’; Hom. Il. + + from *keh 2 p- ‘take’; cf. κώπη ‘handle’, κάπτω ‘gulp down’, Lat. caper 234.31 + 186 Martin Masliš <?page no="187"?> ‘he-goat’ (i.e. ‘snatcher’); EDG, 639-640 39 καπυρός ‘brittle, crackly’; Epich. + + κάπνος ‘smoke’; EDG, 640 N/ A 40 ζέφυρος, ὁ ‘westerly wind’; Hom. Il. + + ζόφος ‘darkness, west’ 408.50 + ( − ) 41 ἐλαφρός ‘nimble (Hom.), light in weight (Hom.), easy (Pi.), gentle (Isoc.)’; Hom. Il. + + 403: ἐλαχύς ‘small, short’, *h 1 leng wh - ‘move easily (intr.)’; EDG, 403, NIL, 243 326.15 − 42 τάφρος, ἡ ‘trench’; Hom. Il. + + θάπτω ‘bury’, τάφος ‘bur‐ ying’; EDG, 543 748.43 − 43 βλῆχρος ‘faint, gentle, small’; Thphr. + + βλάξ ‘indolent, stolid, stu‐ pid’; EDG, 221 200.13 − 44 ψαφαρός ‘friable’; A. + + ψῆφος ‘pebble’; EDG, 1666 N/ A 45 μύδρος, ὁ ‘red-hot mass, anvil’; A. + + μυδάω ‘be damp’ 593.32 − 46 λάφυρα, τά ‘spoils’; A. + + *lemb h- ‘grab’, cf. ἀμφιλαφής ‘wide-spreading, abun‐ dant’; LIV 2 411-412 556.51 − 47 σοβαρός ‘rushing, haughty’; Ar. + − σοβέω ‘scare away, move rapidly, walk in a pom‐ pous way’; EDG, 1371 721.35 + 48 χλιαρός ‘warm’; Epich. + − χλιαίνω ‘warm’ 566.44 + 49 γαμβρός ‘son-in-law, brother-in-law’; Hom. Il. + − γαμέω ‘marry’ 221.8 + 50 λυγρός ‘miserable’; Hom. Il. + − λευγαλέος ‘wretched’; EDG, 850 N/ A 51 φαιδρός ‘bright, cheer‐ ful’; Pi. + − φαίδιμος ‘shining’; EDG, 1544 N/ A 52 σφοδρός ‘vehement’; Hom. Od. + − σφεδανός ‘vehement, vio‐ lent’; EDG, 1429 739.4 − 53 κυδρός ‘glorious’; Hom. Il. + − κῦδος ‘glory’ 39.54 + 54 τακερός ‘melting’; Ar. + − τήκω ‘melt, dissolve’; EDG, 1477 468.10 + 55 γλυκερός ‘sweet’; Hom. Od. + − γλυκύς ‘sweet’ N/ A 56 φανερός ‘visible, shi‐ ning’; Hdt. + − φαίνω ‘bring to light’ 246.37 + Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 187 <?page no="188"?> 57 κρατερός ‘strong’; Hom. Il. + − κράτος ‘strength’ 535.40 + 58 αἰψηρός ‘quick, sudden’; Hom. Il. + − αἶψα ‘quickly’ N/ A 59 σκυθρός ‘angry, grumpy’; Men. + − σκυδμαίνω ‘be angry’; EDG, 1360 N/ A 60 ἐρυθρός ‘red’; Hom. Il. + − ἔρευθος ‘redness’ 379.6 + 61 νωθρός ‘sluggish’; Hp. + − νωθής ‘slow, indolent, stubborn’ 608.2 + 62 ὄνειρος ‘dream’; Hom. Il. + − ὄναρ ‘dream, vision’ N/ A 63 μακρός ‘long’; Hom. Il. + − μῆκος ‘length’ 818.1 + 64 νεκρός, ὁ ‘dead’; Hom. Il. + − νέκυς ‘corpse’ 599.55 + 65 πικρός ‘sharp, shrill, bit‐ ter’; Hom. Il. + − *peik̂ - ‘carve, scratch’ ( → ‘paint’ already in PIE), cf. ποικίλος ‘varicoloured’; EDG, 1190, LIV 2 465-466 671.50 − 66 σαπρός ‘rotten’; Hippon. + − σήπομαι ‘rot’ 707.26 + 67 λεπρός ‘scaly, uneven, scabby’; Hp. + − λέπω ‘peel’, λέπος ‘rind, scale’ 561.3 + 68 λαμπρός ‘bright’; Hom. Il. + − λάμπω ‘glow’ 555.20 + 69 λυπρός ‘distressful’; Hom. Od. + − λύπη ‘pain’ 39.55 + 70 κέντρον ‘sting, goad’; Hom. Il. + − κεντέω ‘sting, goad’ 325.35 + 71 σταυρός ‘pole, cross’; Hom. Od. + − inherited, cf. Old Norse staurr ‘pole’ from *steh 2 -u-ro-, ἵστημι ‘make stand’; EDG, 1391 725.29 + 72 ἄργυρος ‘silver’; Hom. Il. + − *h 2 erĝ- ‘brilliant, white’, cf. ἀργός ‘shining white, brilliant’, Lat. argentum ‘silver’; NIL, 317 137.24 + 73 γλαμυρός ‘blear-eyed’; Hp. + − γλάμων ‘blear eyed’, γλάμος·μύξα Hsch.; EDG, 274 N/ A 74 ἁλμυρός ‘salty’; Hom. Od. + − ἅλμη ‘seawater’ 69.18 + 188 Martin Masliš <?page no="189"?> 75 ξυρόν, τό ‘razor’; E. + − inherited, cf. Skt. kṣurá- ‘razor’, ξύω ‘shave, smoothen’; EDG, 1138- 1139 611.25 + 76 στιφρός ‘firm, compact’; Ar. + − στῖφος ‘body of man in close array’ N/ A 77 αἰσχρός ‘dishonoring’; Hom. Il. + − αἶσχος ‘shame’ 39.53 + 78 γλίσχρος ‘sticky, penur‐ ious’; Hp. + − γλίχομαι ‘cling to, strive for’; EDG, 276 234.29 + 79 ψυχρός ‘cold’; Hom. Il. + − ψῦχος ‘cold’ 39.54 + 80 χλωρός ‘greenish-yel‐ low’; Hom. Od. + − *ĝ h lh 3 -ró- ‘green, yellow’, cf. χλόη ‘young verdure’, χόλος ‘gal’; EDG, 1638- 1639 813.2 + 81 κλαδαρός ‘fragile, quiv‐ ering, wavy’; Plb. + − κλάω ‘break’ influenced by κραδαίνω ‘shake’; DELG, 537 N/ A 82 μαδαρός ‘wet’; Hp. + − μαδάω ‘be moist’ 564.30 + 83 φλυδαρός ‘soft’; Hp. + − φλυδάω ‘have an excess of moisture, become soft’ N/ A 84 χαλαρός ‘slack’; Hp. + − χαλάω ‘slacken’ N/ A 85 σθεναρός ‘strong’; Hom. Il. + − σθένος ‘strength’ N/ A 86 σιναρός ‘damaged’; Hp. + − σίνομαι ‘plunder, harm’ N/ A 87 λιπαρός ‘oily, shining, rich’; Hom. Il. + − λίπα ‘richly (with oil)’ 566.47 + 88 γεραρός ‘honourable, old’; Hom. Il. + − γέρας ‘gift of honour’, γῆρας ‘old age’; EDG, 267- 268 and 271 227.20 + 89 μυσαρός ‘foul’; E. + − μύσος ‘defilement’ 87.27 + 90 ψεφαρός ‘gloomy’; Hp. + − ψέφας ‘gloom’ N/ A 91 ψυδρός ‘lying’; Thgn. + − ψεῦδος ‘falsehood’ N/ A 92 φλογερός ‘blazing’; E. + − φλόξ ‘flame’ N/ A 93 στυγερός ‘loathed’; Hom. Il. + − στύγος ‘hatred’ 73.29 + 94 πλακερός ‘broad’; Theoc. − πλάξ ‘flat land’ N/ A 95 σφαλερός ‘slippery, tot‐ tering’; Hdt. + − σφάλλω ‘cause to fall’ 637.18 + Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 189 <?page no="190"?> 96 νυκτερός ‘nocturnal’; A. + − νύξ ‘night’ N/ A 97 κρυερός ‘icy’; Hom. Il. + − κρύος ‘frost’ 541.51 + 98 τραφερός ‘well-fed’; Theoc. + − τρέφω ‘feed, thicken’ 764.32 + 99 ἀλιτρός ‘sinful’; Hom. Il. + − ἀλιταίνω ‘sin’ 65.16 + 100 κυλινδρός, ὁ ‘rolling stone’; Chrysipp.Stoic. + − κυλίνδω ‘roll’ N/ A 101 βόθρος, ὁ ‘hole’; Hom. Od. + − βόθυνος ‘hole’ 204.11 − 102 ἄλευρον, τό ‘flour’; Hdt. + − ἀλέω ‘grind’ 65.34 + Bibliography Alpers, Klaus (1990): “Griechische Lexikographie in Antike und Mittelalter”. In: Welt der Information, Hans Koch/ Andreas Krup-Ebert (eds.). Stuttgart: Metzler, 14-38. Baayen, Harald R. (1993): “On Frequency, Transparency, and Productivity”. In: Year‐ book of Morphology 1992, Geert E. Booij/ Jaap van Marle (eds.). Dordrecht: Kluwer, 181‒208. Baayen, Harald R. et. al. (2011): “An amorphous model for morphological processing in visual comprehension based on naive discriminative learning”. Psychological Review 118(3): 438-481. Blank, Andreas (1999): “Why do new meanings occur? A cognitive typology of the motivations for lexical semantic change”. In: Historical Semantics and Cognition, Andreas Blank/ Peter Koch (eds.). Berlin/ Boston: De Gruyter Mouton, 61-89. Bloom, Paul (2002): How children learn the meanings of words. Cambridge, MA.: MIT Press. Boudelaa, Sami/ Marslen-Wilson, William D. (2015): “Structure, form, and meaning in the mental lexicon: Evidence from Arabic”. Language, Cognition and Neuroscience 30(8): 955-992. Buck, Carl D./ Petersen, Walter (1945): A Reverse Index of Greek Nouns and Adjectives. Chicago: University of Chicago Press. (Repr. Hildesheim and New York: Olms, 1970.) Burani, Cristina/ Thornton, Anna M. (2003): “The Interplay of Root, Suffix and Whole-Word Frequency in Processing Derived Words”. In: Morphological Structure in Language Processing, Harald R. Baayen/ Robert Schreuder (eds.). Berlin: de Gruyter, 157-208. 190 Martin Masliš <?page no="191"?> Campbell, Lyle ( 3 2013): Historical linguistics. Edinburgh: Edinburgh University Press. Chantraine, Pierre (1933): La formation des noms en grec ancien. Paris: Kliencksieck. Clark, Eve V. (2003): The Lexicon in Acquisition. Cambridge: Cambridge University Press. Cutler, Anne (1980): “Productivity in Word Formation”. In: Papers from the Sixteenth Regional Meeting of the Chicago Linguistic Society, Jody Kreiman/ Almerindo E. Ojeda (eds.). Chicago: Chicago Linguistic Society, 45-51. Cutler, Anne (1981): “Degrees of transparency in word formation”. Canadian Journal of Linguistics/ Revue canadienne de linguistique 26(1): 73-77. Daelemans, Walter/ Van den Bosch, Antal (2005): Memory-Based Language Processing. Cambridge: Cambridge University Press. Dickey, Eleanor (2015): “The sources of our knowledge of ancient scholarship”. In: Brill’s Companion to Ancient Greek Scholarship, Franco Montanari/ Stephanos Matthaios/ Antonios Rengakos (eds.). Leiden/ Boston: Brill, 459- 514. Diependaele, Kevin et al. (2012): “Derivational Morphology and Skilled Reading: An Empirical Overview”. In: The Cambridge Handbook of Psycholinguistics, Michael J. Spivey/ Ken McRae/ Marc F. Joanisse (eds.). Cambridge: Cambridge University Press, 311-332. del Prado Martín, Fermín M. et al. (2005): “Changing places: A cross-language perspec‐ tive on frequency and family size in Dutch and Hebrew”. Journal of Memory and Language 53(4): 496-512. DELG = Chantraine, Pierre (1968-80): Dictionnaire étymologique de la langue grecque: Histoire des mots. Paris: Klincksieck. Dokulil, Miloš (1962): Tvoření slov v češtině I. Teorie odvozování slov. Praha: Naklada‐ telství Československé akademie věd. EDG = Beekes, Robert (2010): Etymological dictionary of Greek, 2 vols. Leiden/ Boston: Brill. Etymologicon Magnum (1816): Friedrich W. Sturz/ Friedrich Sylburg (eds.). Lipsiae: Weigel. Fortson, Benjamin W. (2003): “An approach to semantic change”. In: The Handbook of Historical Linguistics, Richard D. Janda/ Brian D. Joseph (eds.). Malden/ Oxford/ Mel‐ bourne/ Berlin: Blackwell, 648-666. Frost, Ram/ Forster, Kenneth I./ Deutsch, Avital (1997): “What can we learn from the morphology of Hebrew? A masked-priming investigation of morphological repre‐ sentation”. Journal of Experimental Psychology: Learning, Memory, and Cognition 23(4): 829-856. Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 191 <?page no="192"?> GEW = Frisk, Hjalmar (1960-72): Griechisches etymologisches Wörterbuch, 3 vols. Heidelberg: Winter. Hackstein, Olav (2002): Die Sprachform der homerischen Epen: Faktoren morpho‐ logischer Variabilität in literarischen Frühformen: Tradition, Sprachwandel, spra‐ chliche Anachronismen. Wiesbaden: Reichert. Hock, Hans (1991): Principles of Historical Linguistics. Berlin/ New York: De Gruyter Mouton. Hockett, Charles F. (1954): “Two models of grammatical description”. Word 10(2-3): 210-234. Janda, Michael (2005): “Σῶµα-σῆµα and Socrats’ snub nose: The new IE root *tiah 2 - ‘to strike’”. Paper presented at the Colloquium ‘Indo-European and its neighbours’ in combination with the 2. Indogermanistischer Arbeitstagung Münster/ Leiden, Leiden, 6 June 2005. Janyšková, Ilona (2017): “Deetymologizace”. In: CzechEncy - Nový encyklopedický slovník češtiny, Petr Karlík/ Marek Nekula/ Jana Pleskalová (eds.). < https: / / www.czechency.org/ slovnik/ DEETYMOLOGIZACE > (Last access 19.1.2021) LIV 2 = LIV. Lexikon der indogermanischen Verben: Die Wurzeln und ihre Primärstamm‐ bildungen, 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 Martin Kümmel und Helmut Rix. Wiesbaden: Reichert 2001. Marslen-Wilson, William D. et. al. (1992): “Abstractness and transparency in the mental lexicon”. In: Proceedings of the Fourteenth Annual Conference of the Cogni‐ tive Science Society: July 29 to August 1, 1992, Cognitive Science Program, Indiana University Bloomington. New York: Psychology Press, 84-88. Marslen-Wilson, William D. et. al. (1994): “Morphology and meaning in the English mental lexicon”. Psychological Review 101(1): 3-33. Mathesius, Vilém (1966): Řeč a sloh. Praha: Československý spisovatel. McClelland, James L./ Rumelhart, David E./ PDP Research Group (1986): Parallel Dis‐ tributed Processing. Explorations in the Microstructure of Cognition. Cambridge, MA.: MIT Press. Meľčuk, Igor (2000): “Suppletion”. In: Morphology, An International Handbook on Inflection and Word-Formation 1, Geert Booij/ Christian Lehmann/ Joachim Mugdan (eds.). Berlin/ New York: Mouton de Gruyter, 510-524. NIL = Wodtko, Dagmar S./ Irslinger, Britta S./ Schneider, Carolin (2008): Nomina im indogermanischen Lexikon. Heidelberg: Winter. Nussbaum, Alan J. (1976): Caland’s “Law” and the Caland System. Ph.D. dissertation (Harvard University). 192 Martin Masliš <?page no="193"?> Nussbaum, Alan J. (1998): Two Studies in Greek and Homeric Linguistics. Göttingen: Vandenhoeck & Ruprecht. Nussbaum, Alan J. (2014): “Greek τέκμαρ ‘sign’ and τέκμωρ ‘sign’: Why both? ” In: Das Nomen im Indogermanischen, Norbert Oettinger/ Thomas Steer (eds.). Wiesbaden: Reichert, 215-260. Plaut, David C./ Gonnerman, Laura M. (2000): “Are non-semantic morphological effects incompatible with a distributed connectionist approach to lexical processing? ” Language and Cognitive Processes 15(4-5): 445-485. Probert, Philomen (2006): Ancient Greek Accentuation: Synchronic Patterns, Frequency Effects, and Prehistory. Oxford: Oxford University Press. Quine, Willard V. O. ( 2 2013): Word and object. Cambridge, MA.: MIT Press. Reitzenstein, Richard (1905): “Etymologika”. In: Paulys Realencyclopädie der classis‐ chen Altertumswissenschaft (RE). Band V,2. Stuttgart: Metzler, 807-817. Sandell, Ryan P. (2015): Productivity in Historical Linguistics: Computational Perspec‐ tives on Word-Formation in Ancient Greek and Sanskrit. Ph.D. dissertation (Univer‐ sity of California). Sandra, Dominiek (1994): “The morphology of the mental lexicon: Internal word structure viewed from a psycholinguistic perspective”. Language and Cognitive Processes 9(3): 227-269. Schreuder, Robert/ Baayen, Harald R. (1995): “Modeling morphological processing”. In: Morphological Aspects of Language Processing, Laurie B. Feldman (ed.). Mahwah, NJ.: Lawrence Erlbaum Associates, 131-154. Schreuder, Robert/ Baayen, Harald R. (1997): “How complex simplex words can be”. Journal of Memory and Language 37(1): 118-139. Sluiter, Ineke (2015): “Ancient Etymology: A Tool for Thinking”. In: Brill’s Companion to Ancient Greek Scholarship, Franco Montanari/ Stephanos Matthaios/ Antonios Rengakos (eds.). Leiden/ Boston: Brill, 896- 922. Traugott, Elizabeth C. (2017): “Semantic Change”. In: Oxford Research Encyclopedia of Linguistics. <https: / / doi.org/ 10.1093/ acrefore/ 9780199384655.013.323> (Last access 11.4.2021) Wagner, Laura (2010): “The acquisition of semantics”. Wiley Interdisciplinary Reviews: Cognitive Science 1(4): 519-526. Weiss, Michael L. (1998): “Erotica: on the prehistory of Greek desire”. Harvard Studies in Classical Philology 98: 31-61. Zucker, Arnaud/ Le Feuvre, Claire (2021): “Introduction”. In: Ancient and Medieval Greek Etymology: Theory and Practice, Arnaud Zucker/ Claire Le Feuvre (eds.). Berlin/ Boston: De Gruyter, 1-18. Formal opacity and semantic (in)stability of derived nominal lexemes in Ancient Greek 193 <?page no="195"?> 1 The abbreviations for authors, texts and corpora follow the practice of the ThLL. The following abbreviations are used for marking the constituents of the grading constructions: cpree = comparee, par = parameter, pm = parameter marker, stand = standard, stm = standard marker (on this terminology see most recently Key‐ dana/ Hock/ Widmer 2021: 11‒13). I would like to express my thanks to my anonymous reviewers for their comments and suggestions. Any remaining shortcomings are of course my own. Speculo claras or speculoclaras? Máté Ittzés Abstract: The phrase speculo claras ‘bright like mirror’, attested once in Plautus Most. 642, in which the positive adjective claras is con‐ strued with the ablative speculo, has been found suspicious by various scholars (R. Ellis, E. Hoffmann, W. M. Lindsay, etc.) due to its unusual syntax and emended as a compound adjective in the form speculiclaras or speculoclaras (cf. OLD s.v. speculoclarus). Others (E. Löfstedt, A. H. Salonius, J. B. Hofmann/ A. Szantyr, etc.), however, have preferred reading speculo claras in two words, referring to various alleged or real, closer or more distant syntactic parallels. In my paper, I will treat the adduced linguistic arguments in detail and try to answer the question, how much the construction “positive adjective + ablative” is to be regarded as isolated within the system of Latin gradation as a whole, and whether there is any support or need for the assumption of a nominal compound as hypothesized by Ellis and others. Keywords: Plautus, syntax, gradation, ablative, nominal compound 1 Introduction The construction speculo claras in Plaut. Most. 642, 1 as transmitted in the manuscripts, has generated much discussion in Plautine philology as well as <?page no="196"?> 2 The interpretation of speculo as an ablative is widely accepted, although it must be added that theoretically it could also be taken as a dative case form. The latter was the opinion, for instance, of Fay, who interpreted it as a dativus iudicantis, the dative of the person judging (Fay 1902: 119: “bright to a mirror, i.e. even a mirror would call the house splendid; speculo is dativus iudicantis, as in line 251, though perhaps we may regard the dative here as a variation for ad, ‘in comparison with’”; note that in connection with verse 251, Fay 1902: 92 also refers to the fairy tale of Snowwhite, “the mirror being the official judge of beauty”, which does not seem to me to be a particularly strong scientific argument). A weighty, but admittedly not completely cogent counterargument against Fay’s interpretation of Most. 642 is that the role of the dativus iudicantis seems to be taken only by human entities (cf., e.g., Luraghi 2010: 78; Panhuis 2015: 102). To be sure, we have also examples such as Iam lauta es? ‒ Iam pol mihi quidem atque oculis meis Plaut. Truc. 378 “Are you washed now? ‒ Yes, at least to myself and my own eyes.” (transl. Pinkster 2015: 1203), but here too, oculis is connected by meis to the speaker, i.e. a human being. As regards Most. 251, I think that speculo is not a dativus iudicantis in this passage either but is better taken as a dativus finalis: “for (the purpose of) a mirror, you’re the most eminent mirror yourself”. 3 This type is usually, but in my opinion erroneously, interpreted as an equative construction (see Ittzés 2021b in detail and cf. also below). Latin linguistics. The reason for this is the fact that an ablative of compar‐ ison, which the noun form speculo in all probability is, 2 can be regarded as anything but regular in connection with a bare positive adjective. It is generally taken for granted that speculo claras comprises a generic comparison, semantically related to the special comparative type 3 melle dulcior, which involves a comparative adjective and an ablative of compar‐ ison, or the similative type niger tamquam coruus, which contains a positive adjective and a nominative standard marked by a particle (in this case tamquam). The passage including speculo claras runs in the edition of Leo (1896; this is the edition that is included, for instance, in the PHI Latin Texts corpus) as follows: (1) Plaut. Most. 637‒642 - - TR. Aedis filius - tuos emit. TH. Aedis? TR. Aedis. TH. Euge, Philolaches - patrissat: iám homo in mercatura uortitur. - ain tu, aedis? TR. Aedis inquam. sed scin quoius modi? - TH. Qui scire possum? TR. Vah. TH. Quid est? TR. Ne me roga. - TH. Nam quid ita? TR. Speculo claras, candorem merum. - - - TR. Your son has bought 196 Máté Ittzés <?page no="197"?> 4 In the apparatus Ritschl even ventures to suggest a possible supplement of the text: “TH. Nam quid ita? [TR. cor tibi exsultaturum arbitror: / Magis hercle] speculo claras, clarorem merum.” (Note Ritschl’s clarorem instead of Leo’s candorem for clamorem of the manuscripts.) 5 Note, incidentally, that this compound adjective is translated by Erich Segal as “splendiddifferiffic” (Segal 1996: 161). 6 The uniqueness of this hypothesized compound was apparently the reason for Leo (1896: 106) to reject the emendation of Spengel: “speculo quasi ut speculum claras singularis constructio; compositum speculoclaras contra analogiam”. See also a house. TH. A house? TR. A house. TH. Well done! Philolaches - takes after his dad; he’s a man with a turn for business. - A house, you say? TR. A house, I say. But, do you know what kind? - TH. How can I? TR. Wow! TH. What does ‘wow’ mean? TR. Don’t ask. - TH. Why’s that? TR. A mirror-bright one, pure radiance. - - (transl. by Slater 2016: 54) Some scholars have thought that the transmitted text must be corrupt and decided to emend it in one way or another. A few editors and commentators, like Ritschl (1852: 91) and Lorenz (1883: 119), assumed a lacuna before speculo, which must have originally contained the parameter marker magis of the comparative degree or aeque of the equative (for more on the latter possibility cf. below). Ritschl, followed by Lorenz, prints the text accordingly as follows: (2) TH. Nam quid ita? TR. * * * * * * * - * * * speculo claras: clarorem merum. 4 Although this emendation obviously solves the syntactic problem of the positive adjective in connection with the ablative of comparison, it also means that the ablative speculo necessarily separates the two parts of the analytic grading construction (comparative magis claras or equative aeque claras), which is not a very frequent phenomenon in Latin comparative or equative constructions either (on this point cf. below). Others (e.g. Spengel 1865: 48; Ellis 1882: 167; Hoffmann 1885: 232; Lindsay 1905) have chosen to solve the problem in a different way, namely by re‐ moving the word division between the two words and taking speculoclaras (or speculiclaras) as a compound adjective, an opinion which seems to be the prevalent one nowadays (cf. OLD: 1802 s.v. speculoclarus; Bader 1962: 326; Lindner 1996: 178; Slater 2016: 54‒55; 5 not accepted by ThLL III 1272, 59-60). Lindner adds that when creating this hapax legomenon, 6 Plautus might have Speculo claras or speculoclaras? 197 <?page no="198"?> Löfstedt (1911: 190): “Allein dies Wort ist ja nicht nur sonst unbekannt, sondern überhaupt morphologisch unbefriedigend.” had in mind a Greek compound adjective such as ὑαλοειδής ‘looking like glass, glassy’ (cf. LSJ: 1840). However, there have been some noted scholars (Löfstedt 1911: 190‒191; Salonius 1920: 110‒111; Hofmann/ Szantyr 1972: 110) who have sticked to the transmitted reading speculo claras divided into two words. They have argued that in spite of being unique, the construction should not at all be regarded as grammatically incorrect and have referred to various syntactic parallels, which might explain the seemingly irregular combination of the positive adjective with the ablative of comparison. In this paper, my purpose will be to weigh the arguments that have been adduced in favour of the reading speculo claras (i.e. the combination of positive adjective + ablative of comparison) by Löfstedt and his followers, and see whether we can really dispense with changing the transmitted text in one of the ways mentioned above. At the end of the paper I will also deal briefly with some questions related to the other two solutions (i.e. assuming a lacuna before speculo or removing the word division). 2 The arguments for retaining the transmitted text and their evaluation According to Löfstedt (1911: 190), who seems to have been the first to formulate this view, “es liegt eigentlich eine Kontamination von speculo clariores und ut speculum claras vor, um so natürlicher, als diese beiden Wendungen hier sachlich gleichwertig waren.” Löfstedt refers as a parallel (“eine ähnliche Mischkonstruktion”) to the use of the ablative together with the equative parameter marker aeque (for the attestations cf. below) as well as (“ein direktes Gegenstück”) to the well-known wall inscription from Pom‐ peii (CIL IV 1895; ex. 9), which reads tan durum saxso, “eine volkstümliche Umgestaltung” instead of the original phrase magis […] saxo durum of Ovid (ars 1.475; ex. 10), which is “die korrekte literarische Ausdrucksweise”. He finishes his argumentation by claiming that, similarly to the case of speculo claras, saying magis durum saxo was factually the same as using tam durum quam saxum (“sachlich war es ja gleichgültig”), and this must have been the ultimate reason for the grammatical contamination of the two patterns. 198 Máté Ittzés <?page no="199"?> 7 A notable exception is the short remark by Risch (1954: 215): “Jedenfalls scheinen mir die griech. Beispiele für eine angeblich adäquative Bedeutung gegenüber den zahlreichen eindeutig komparativen noch nicht genügend gesichert zu sein.” The alleged equivalence (“gleichwertig”, “gleichgültig”) of the two types of gradation (speculo clariores ~ ut speculum claras; magis durum saxo ~ tam durum quam saxum) is based on the assumption that morphological comparatives in connection with quality-typifying substantives (i.e. the so-called melle dulcior type) are instances of the equative degree rather than true comparatives, since nothing “can be endowed with more of the quality than the object symbolizing it” (Rosén 1999: 193; for this view see, among many others, Löfstedt 1956: 238‒242; Puhvel 1973; Stefanelli 1984: 205; Bertocchi/ Orlandini 1996: 210; Rosén 1999: 189‒193; Orlandini/ Poccetti 2010: 194; Cuzzolin 2011: 604 7 ). However, as I have tried to clarify in a recent article (Ittzés 2021b), this view cannot be upheld. To be sure, in terms of pragmatic inference, melle dulcior constructions can certainly be conceived as elatives in just the same way as corresponding constructions of the morphologically equative type, since both can indeed be understood as expressing a very high degree. For instance, melle dulcior may in fact be interpreted as a means of expressing the notion ‘very sweet’ (for such an elative interpretation see, first of all, Orlandini/ Poccetti 2010; cf. also Bertocchi/ Orlandini 1996: 210). This does not mean, however, that on a “lower” level, i.e. in semantic terms, the two types should necessarily be exponents of the same category, the equative degree. As for the impossibility of surpassing the prototype with respect to the given parameter (cf. the formulation of Rosén above), I think this opinion is again based on a misunderstanding. It should go namely without saying that the degree relations implied by melle dulcior constructions hold true in the universe of discourse invoked by the speaker (in other words, on a rhetorical level) and not, or at least not necessarily, in objective reality. For instance, even if it is technically indeed probably impossible for anything to be whiter than snow, that does not mean that niue candidior (e.g. Ov. met. 8.373; Mart. 1.115.2-3) could not be used by the speaker as a comparative of superiority. Remember also that, for a human or any other living being, to be “only” as white as snow (i.e. niue candidior in its alleged equative meaning) is impossible in the same way, which means that similar considerations would apply to the interpretation of such constructions anyway. Speculo claras or speculoclaras? 199 <?page no="200"?> 8 Rosén (1999: 193) maintains that the rare constructions involving a standard with the same meaning but marked by the particle quam instead of the ablative of comparison (e.g. homo leuior quam pluma Plaut. Men. 487-488 “you fellow lighter than a feather”) contain an “unnatural comparison”, while the formally fixed ablative pattern has an equative force. Having in mind the fundamental characteristics of quam standards (cf. Tarriño 2011: 385: “the quam standard rather has an adversative and implicit negative value: whereas the ablative case presupposes the property in the standard, quam does not presuppose it”), the difference of the two variants seems to be that the quam pattern comprises an even more radical exaggeration, since it implies that in comparison with the comparee, the prototypical standard seems not to have the respective property at all. Thus, in my view there is no difference between the two patterns in terms of gradation type (i.e. equative vs. comparative). It is obvious that all proverbial phrases of the melle dulcior type essentially and inherently comprise a hyperbole or exaggerated comparison. In my view, what they are intended to express is precisely that, according to the speaker and in the context of the discourse, the comparee is indeed endowed with more of the particular quality than the entity which is conventionally regarded as having its maximum degree. Accordingly, I think that the melle dulcior type has to be considered as a special instance of the comparative of superiority. 8 3 Supposedly parallel constructions Let us turn now to the alleged parallelism of speculo claras and the two other aforementioned equative constructions involving the ablative of com‐ parison. The construction with the equative parameter marker aeque, which occurs “öfters” according to Löfstedt, is actually attested only six times in the extant corpus of Latin literature, mainly in Old Latin. It appears four times in Plautus (3rd-2nd c. B.C.), once in Pliny the Elder (1st c. A.D.) and once in a letter of Sidonius Apollinaris (5th c. A.D.). (3) nullust hoc metuculosus aeque (Plaut. Amph. 293) -- cpree stand. stm par pm - “there is not in existence another such cowardly fellow as this” (4) neque est neque fuit me senex quisquam amator / adaeque miser (Plaut. Cas. 684‒685) -- - stand. stm cpree p m par - “there neither is nor ever was any old man in love so wretched as I” 200 Máté Ittzés <?page no="201"?> (5) qui me in terra aeque fortunatus erit, si illa ad me bitet […]? (Plaut. Curc. 141) -- cpree stand.stm - pm par - “who on earth will be so blest as myself, if she comes to me? ” (6) quo nemo adaeque iuuentute ex omni Attica / antehac est habitus parcus -- stand.stm cpree pm par - nec magis continens (Plaut. Most. 30‒31) - “a person, with whom not one of all the young men of Attica was before deemed equally frugal or more steady” (7) picturae […], quibus equidem nullas aeque miror (Plin. nat. 35.17) -- - stand.stm - cpree pm par - “paintings […], with which I do not admire any in equal measure” (8) in homine […], quo nihil aeque miserum destitutumque nascendi condicio -- - stand.stm cpree pm par par - produxit (Sidon. epist. 7.14.3) - “in the human race […], with which, by the conditions of its birth, nothing has been produced equally miserable or helpless” If we have a closer look at these passages, we can immediately observe that the aeque (pm) + positive adjective (par) + ablative (stand) pattern has a very limited distribution as far as the syntactic configuration is concerned. Namely, in all the six known cases, 1. the stand precedes the par and the pm, and 2. the stand is always a pronominal ablative and never a lexical noun phrase and thus never a “quality-typifying substantive”. On the other hand, the Pompeii inscription (see ex. 9), which is referred to by Löfstedt (cf. above), displays none of these two properties: firstly, the stand is not a pronoun, but a lexical noun phrase (saxso), and secondly, it is not followed, but preceded by the pm+par constituents (tan durum). (9) quid pote tan durum saxso aut quid mollius unda (CIL IV 1895) -- cpree - pm par stand. stm - cpree par.pm stand. stm - “what can be as hard as stone, what softer than water? ” Thus it turns out that it is not possible to speak of a parallelism of the two equative patterns (tam + positive + ablative standard in ex. 9 and aeque + positive + ablative standard in exx. 3‒8). To be more precise, the tam + ablative pattern attested in the Pompeii inscription is completely isolated and idiosyncratic, it has no exact or even close parallels, and was, in my opinion, never “sprachwirklich” in Latin. The simplest and most obvious explanation is that it owes its “existence”, or rather its incidental attestation, Speculo claras or speculoclaras? 201 <?page no="202"?> 9 Thus, contrary to Rosén (1999: 190), I think that the epigraphic version is in fact “in peius mutatum” (Bücheler 1895: 433 ad No. 936). to an imperfect memory of Ovid’s hexameter (ex. 10) by a not particularly erudite Pompeian person (cf. e.g. Rebling 1883: 9). 9 (10) quid magis est saxso durum, quid mollius unda? (Ovid. ars. 1.475) -- cpree pm stand.stm par - “what can be as hard as stone, what softer than water? ” Furthermore, even if one considered the inscriptional construction genuine, there would be some points which have to be taken into account. Namely, it can be observed that the second of the abovementioned restrictions (the presence of a pronominal standard) does not apply to the phrase speculo claras either, to which, on the other hand, the presence of a “quality-typi‐ fying substantive” is common with the Pompeian construction tam durum saxo. However, this does not mean that speculo claras and alleged tam durum saxo could be regarded as close relatives. There are namely two significant differences between the two constructions. Firstly, the inscrip‐ tional construction involves an explicit parameter marker (tam), which, remarkably enough, speculo claras is lacking. Secondly, and more importantly, the inscriptional construction, identi‐ cally to all instances of the aeque + ablative pattern (exx. 3-8), occurs in a negative context (i.e. in clauses containing some sort of negation or in rhetorical questions, which are equivalent to negative statements), which is again not the case for speculo claras, at least in the transmitted text. This fact is crucial inasmuch as it is only in negative contexts that a neutralization of the comparative and equative degrees takes place and the inference of both constructions is necessarily a superlative one. For instance, all patterns listed in (11) are equivalent to the superlative construction X est maxime par. (11) a. nihil / quid? tam par - est quam X n o m ~ X n o m est maxime par - b. nihil / quid? aeque par - est ac X n o m - c. nihil / quid? - par -ius est quam X n o m - d. nihil / quid? - par -ius est - X a b l Note furthermore that in exx. 3‒8 the parameter and the parameter marker are always preceded by the neutralizing factor, the item indicating the neg‐ ative polarity (which is nearly always the same constituent as the comparee). 202 Máté Ittzés <?page no="203"?> 10 See, however, above on the apparent restrictions of this pattern (cf. exx. 3-8). 11 Not actually attested, but implied by the existence of quid mollius unda in the second half of the verse (cf. ex. 10) and other melle dulcior constructions. 12 Löfstedt (1911: 190) refers rather to Plaut. Merc. 335 (homo me miserior nullust aeque “there is not any person more wretched than myself”, involving a comparative adjective in addition to aeque; cf. also Capt. 828; in Mil. 551 the standard is not mentioned explicitly) as an argument proving that a confusion with comparative constructions took place in the case of the aeque-pattern in exx. 3‒8, but in my view the simultaneous presence of the comparative morpheme -iorbeside aeque sets these examples somewhat more apart. All this means that even if the pattern tam + ablative stand (cf. ex. 9) had in fact existed in Latin, which I think it probably did not, and thus quid tam durum saxo? and quid magis durum saxo? had been pragmatically equivalent (both implying a superlative inference), this would not mean that they necessarily contained the same grading construction type. It is entirely possible that one of them is an equative pattern, the other one a comparative, which are neutralized in, or rather due to, the negative context (cf. ex. 12). (12) a. quid tam (? ? ) / aeque 10 durum - saxo? ~ saxum durissimum est - b. (quid - dur- -ius saxo? ) 11 - c. quid magis durum - saxo? I agree with those scholars who interpret the aeque + ablative pattern as re‐ sulting from some sort of contamination (e.g. Kühner/ Stegmann 1992: 2, 467: “eine an sich unlogische, nur durch Kontamination zu erklärende Konstruk‐ tion”; Hofmann/ Szantyr 1972: 110: “durch Kontamination entstanden”; Löf‐ stedt 1956: 241: “Gedankenkontamination”; cf. also Núñez Romero-Balmas 2002: 162-163). This hypothesis is made particularly plausible by example (6), in which adaeque […] parcus and magis continens are coordinated and share the same stand, expressed in the form of an ablative of comparison. 12 It has to be added, however, as a crucial point that such a contamination may occur only in a negative context (cf. above), which is responsible for the neutralization of the two types of degree relation. 4 Grading constructions without an explicit parameter marker A potentially valid argument in favour of Löfstedt’s interpretation could be if, similarly to the hypothesized interpretation of speculo claras, we could find other grading constructions in Latin which lack an overt parameter marker under similar circumstances. Let us examine now if there exist Speculo claras or speculoclaras? 203 <?page no="204"?> 13 The use of the positive form of the adjective obviously cannot be attributed to a general ban on the superlative inclitissimus, since the latter is widely attested (e.g. Cato orig. frg. 83; Colum. 1.4.2; Gell. 3.7.19; 13.7.6 [in the latter case it is in fact construed with a partitive genitive: inclutissimi poetarum]). any such patterns. On the basis of a comprehensive survey of the grading constructions in Latin (cf. Ittzés 2021a; see also Cuzzolin 2011), we can say that, apart from the potential equative or similative reading of the very construction speculo claras, there seem to be altogether five patterns in which the absence of an explicit parameter marker has been assumed, but as we will see immediately, not all of them are equally well attested. One of these patterns is represented by a single, isolated superlative construction (cf. ex. 13), in which the partitive genitive standard omnium philosophorum, including a universal quantifier, accompanies the positive adjective inclito (instead of superlative inclitissimo 13 ). The citation in (13) reproduces the text of the 1972 Teubner-edition of Iustinus by O. Seel. Note, however, that the genitival phrase omnium philosophorum has been deleted by some older editors, which solution, if accepted, would mean that (13) is not a genuine example of this pattern and thus no instance of the absence of a parameter marker. (13) Iust. 12.16.8 -- per quinquennium sub Aristotele doctore inclito omnium philosophorum creuit -- - cpree par stand.stm - “he (scil. Alexander the Great) improved himself for five years under his instructor Aristotle, the most famous of all philosophers” Apart from this somewhat dubious (and also rather late, i.e. 3rd century A.D.) example, it seems that there are four other gradation patterns which may lack an explicit parameter marker. 1. Similative constructions involving a standard marked by various particles such as tamquam, quasi, sicut, ceu or quomodo (cf. Ittzés 2021a: 485-486): This type (ex. 14) is frequently used for expressing generic comparisons of idiomatic nature, but it is apparently less frequent in “ordinary” compari‐ sons than the corresponding type with an explicit parameter marker. On the other hand, this type often occurs in constructions that have some specific function on what Functional Grammar calls the “interpersonal level”, i.e. if they are concerned with the speech event rather than the narrated event itself (see Tarriño 2011: 408-419). 204 Máté Ittzés <?page no="205"?> 14 Nevertheless, note also the arguments of Pinkster (2015: 928) (cf. Torrego 2002: 271-277; Tarriño 2011: 379 n. 8) against the interpretation of such constructions in terms of gradation. (14) Petr. 43.7 -- […] fuit […] niger tamquam coruus -- - par stm stand - “was black as a crow” 2. Equative constructions with a standard marked by the particle quam (cf. Ittzés 2021a: 499): Such constructions (ex. 15) usually involve a negation; examples without a negation occur first in Late Latin. Many instances of this type are actually “pseudo-comparative” constructions (cf. also below). Note that since the pm is omitted and since this pattern shares the stm quam with the comparative type mentioned below (4.), it is only the lexical content which allows non… quam constructions to be taken either as (negated) comparisons of equality (in the sense of non tam… quam) or superiority (in the sense of non magis… quam). However, the meaning and the presence of the stand suffice for the construction itself to be recognized (Tarriño 2011: 387). Note also that the presence of the negation, due to its neutralizing effect (on which cf. above), makes the difference between equative and comparative in this case actually irrelevant. (15) Plaut. Rud. 943 -- non edepol piscis expeto quam tui sermonis sum indigens -- - | ― cpree ― | stm | ――――stand―――― | -- “by Pollux, I’m not looking for fish as much as I am in need of a conversation with you” 3. Comparative constructions with a standard marked by various adposi‐ tions such as ante ‘before’, prae ‘before, in front of ’, praeter ‘beyond’, super ‘above’ (also supra quam), occasionally ultra ‘over’ (cf. Ittzés 2021a: 515): The existence of this type (the phenomenon is called positivus pro com‐ parativo) has been controversial (see e.g. Cuzzolin 2011: 556, who claims, following Timpanaro, that the occurrence of the pm in the comparative strategy was obligatory in Latin and could not be omitted under any circum‐ stances). Although most of the apparent examples are indeed uncertain, there seem to be some occurrences (e.g. ex. 16) that are, in my view, difficult to be explained otherwise (cf. Kühner/ Stegmann 1992: 2, 468). 14 Speculo claras or speculoclaras? 205 <?page no="206"?> 15 There is yet another pattern with an implicit parameter marker, but in that case, there is no standard involved (cf. Ittzés 2021a: 558). I am referring to the use of various adjectives with an appropriate semantics to express the excessive degree (16) Cic. fam. 4.4.2 -- tu […] prae nobis beatus -- cpree stm stand par -- “you [seem] happier than ourselves” 4. Comparative constructions with a standard marked by the particle quam (cf. Ittzés 2021a: 527-528): Similarly to the previous type, the existence of this pattern has also been controversial (cf. again Cuzzolin 2011: 556). While it is true that a large part of the alleged examples do not stand scrutiny and have to be textually emended (see already Kühner/ Stegmann 1992: 2, 463; cf. Cuzzolin 2011: 556 on Enn. ann. 134; Bertocchi/ Orlandini 1996: 205 and Conte 2013: 103-104 on Plaut. Rud. 1114), the approach that completely denies the existence of this pattern is again, in my view, unsupported (many potential or secure exam‐ ples are cited by Hofmann/ Szantyr 1972: 593-594). Most of the examples are found in the works of Tacitus (Tarriño 2011: 387). However, it has to be added that the largest part of the attestations are in fact “pseudo-comparatives”. (17) Tac. ann. 4.61 -- Asinius Agrippa, claris maioribus quam uetustis -- - par c p r e e cpree+stand stm par s t a n d -- “Asinius Agrippa, with ancestors more illustrious than old” 5 Conclusions As it is clear from this survey, none of the abovementioned four types involves a standard marked by a specific case ending, but only such that is marked by a separate and therefore more salient morphological means, i.e. a particle or an adposition, which may in some sense be regarded as a device compensating for the lack of the parameter marker in order to make the grading construction recognizable. It appears that Latin system‐ atically avoids such grading constructions which lack an explicit param‐ eter marker and a morphologically “separate” (i.e. particle or adposition) standard marker at the same time, 15 a configuration which would precisely 206 Máté Ittzés <?page no="207"?> without any overt pm (Burkard/ Schauer 2012: 67). It is clear that the excessive reading of such sentences is determined only by the context. See e.g. Cic. orat. 33: nihil difficile amanti puto, which may be translated as follows: “I think nothing is too difficult for the person who loves”. be characteristic of speculo claras in its Löfstedt-style interpretation. These circumstances obviously speak against taking a two-word speculo claras at face value and regarding it as semantically equivalent either to equative tam claras quam speculum or similative claras ut speculum (or, for that matter, comparative speculo clariores). Having all this in mind, we have to conclude that the argumentation of Löfstedt and his followers suffers from considerable shortcomings, which means that the transmitted text speculo claras with its alleged equative (‘as bright as mirror’) or similative (‘bright like mirror’) interpretation cannot be confirmed by means of any syntactic parallels. Therefore, the emendation of the transmitted text seems to be indeed inevitable. - Outlook: Remarks on the suggested emendations Although my purpose in this short article has been mainly to evaluate the arguments that have been adduced in favour of the retention of the transmitted text, after having found that these arguments are invalid, let us now shortly deal with some issues concerning the two aforementioned possibilities that may solve the remaining textual and linguistic problem: i.e. either assuming a lacuna before speculo, which originally contained a parameter marker (as per Ritschl and his followers) or taking speculoclaras as a compound adjective (as per Spengel and others). Let us first have a look at the question already hinted to above whether there are other instances that could support the assumption of the word order *pm (*magis or *aeque) stand (speculo) par (claras) (i.e. the separa‐ tion of the pm and the par by the ablative standard), which is implied by Ritschl’s approach. In Plautus, I have found just a single example with magis that seems to display the same order of the constituents (ex. 18), which suggests that in Old Latin it was obviously a marked choice. Note, however, that in this example there are in fact two functionally identical parameter markers (the particle magis as well as the comparative suffix -ius), which means that the Speculo claras or speculoclaras? 207 <?page no="208"?> 16 In accordance with exx. 3‒6, there are no such examples involving (ad)aeque in Plautus. I have found no instances of the two patterns in Terence, Cato and Sallust either. construction may also be interpreted as an instance of the basic word order stand par+pm, amplified by the redundant addition of the particle magis. 16 (18) Plaut. Capt. 644 -- nihil […] inuenies magis hoc certo certius -- cpree pm stand par+pm -- “you’ll find nothing more certain than this certainty” Nonetheless, there are a number of later examples of the pattern under investigation, the most famous one probably being Ovid’s magis […] saxo durum (cf. ex. 10). This indicates that, as far as the order of the constituents of the grading construction is concerned, Ritschl’s hypothesis cannot be rejected out of hand, although the assumed pattern is not very well repre‐ sented in Old Latin. As far as the “compositional approach” is concerned, if we have a look at the secondary literature, it becomes clear that the assessment of the hypothesized compound speculoclarus has been rather ambiguous, mainly because, if accepted, it would be a hapax legomenon in the extant Latin litera‐ ture. For two sharply opposing views see e.g. the old verdict of Coulter (1916: 158 n. 19) (“A form so unlike any other compound as to raise considerable doubt about the correctness of the emendation”) vs. the recent evaluation of Lindner (1996: 178) (“Die Konjektur […] erscheint sehr plausibel”). As I have mentioned above, the view taken by Lindner seems to be the most popular nowadays (it has been accepted most recently by De Melo in his Loeb edition of Plautus’ comedies published in 2011). There have been attempts in the secondary literature to remove the com‐ pound speculoclarus from its isolation. For instance, Stein (1971: 599) tried to connect it to other compound adjectives consisting of an adjective and a noun: “An adjective and a noun are combined in albicapillus, semnisomnus, siccoculus, and unoculus. In speculoclarus the arrangement is reversed.” However, Stein fails to mention the important fact that while albicapillus etc. are possessive (i.e. bahuvrīhi) compounds, speculoclarus not only has a reversed order, but also belongs to a fundamentally distinct category, that of the determinative compounds. 208 Máté Ittzés <?page no="209"?> 17 The parallels mentioned by Bader in a footnote (1962: 326 n. 66) are indeed seman‐ tically appropriate and could of course be multiplied, but, for our purposes, suffer from the obvious weakness that they are not Latin compounds, but come from other languages (e.g. Germ. steinhart). Note, incidentally, that the compound “skr. ū́rṇā-mṛdu- ‘doux comme la laine’” is attested first in a Middle Vedic prose text (Taittirīya-Brāhmaṇa) and it obviously replaced the older possessive compound ū́rṇa-mradas- ‘having the softness of wool’ with the noun mrádasas second member, which appears as early as the Rigveda (RV 5.5.4a; 10.18.10c) and which is a completely different category (cf. Wackernagel 1905: 232). The data in Wackernagel (1905: 235) clearly show that this type of compound involving a comparison is not very old in Old Indo-Aryan (the earliest example Wackernagel mentions is Old Vedic śúka-babhru- ‘reddish as a parrot’). Bader is therefore right in claiming that “le procédé a dû se développer indépendamment dans plusieurs langues”. The problem is that in absence of any unequivocal parallels it cannot be proven beyond doubt that Latin belonged to these languages (contrary, for instance, to Greek, which obviously did: cf. e.g. μελιηδής ‘honey-sweet’). The material collected by Bader (1962: 326), which Lindner explicitly refers to as supporting the plausibility of the emendation speculoclarus, is in fact very limited and also rather controversial. The closest parallels that may be mentioned from Bader’s list for the noun+adjective structure in Old and Classical Latin are compound adjectives with the final member -cupidus (e.g. turpilucricupidus ‘greedy for dishonest gain’ Plaut. Trin. 100), but even these have a quite different internal syntax. The other examples mentioned by Bader in the same section are semantically or syntactically very remote and/ or late. 17 These observations show that the assumption of a compound speculo‐ clarus is, contrary to some recent opinions, actually far from trivial and readily acceptable. This is true even if it cannot be denied that, at least from the point of view of textual criticism, removing the word division between speculo and claras would be more economical than assuming a larger lacuna before speculo. Nevertheless, what has hopefully become clear from this paper is that the text as it is transmitted has indeed to be emended in one way or another. Bibliography Bader, Françoise (1962): La formation des composés nominaux en latin (Annales Littéraires de l’Université de Besançon 46). Paris: Les Belles Lettres. Speculo claras or speculoclaras? 209 <?page no="210"?> Bertocchi, Alessandra/ Orlandini, Anna (1996): “Quelques aspects de la comparaison en latin”. Indogermanische Forschungen 101: 195-232. Burkard, Thorsten/ Schauer, Markus ( 5 2012): Lehrbuch der lateinischen Syntax und Semantik. Begründet von Hermann Menge. Darmstadt: Wissenschaftliche Buchge‐ sellschaft. Bücheler, Franciscus (ed.) (1895): Carmina Latina epigraphica. Vol. 1. Leipzig: Teubner. Conte, Gian Biagio (2013): Ope ingenii. Experiences of Textual Criticism. Berlin/ Boston: De Gruyter. Coulter, Cornelia C. (1916): “Compound Adjectives in Early Latin Poetry”. Transactions and Proceedings of the American Philological Association 47: 153‒172. Cuzzolin, Pierlugi (2011): “Comparative and Superlative”. In: New Perspectives on Historical Latin Syntax. Vol. 4. Complex Sentences, Grammaticalization, Typology, Philip Baldi/ Pierluigi Cuzzolin (eds.). Berlin/ Boston: De Gruyter Mouton, 549-659. De Melo, Wolfgang (ed., transl.) (2011): Plautus: The Merchant. The Braggart Soldier. The Ghost. The Persian (Loeb Classical Library 163). Cambridge, MA: Harvard University Press. Ellis, Robinson (1882): “On the Mostellaria of Plautus”. The Journal of Philology 11: 161‒174. Fay, Edwin W. (ed.) (1902): T. Macci Plauti Mostellaria. Boston: Allyn and Bacon. Hoffmann, Emanuel (1885): “Speculoclarus”. Archiv für lateinische Lexikographie und Grammatik 2: 232. Hofmann, Johann B./ Szantyr, Anton (1972): Lateinische Syntax und Stilistik. Verbess. Nachdr. München: Beck. Ittzés, Máté (2021a): “Latin”. In: Comparison and Gradation in Indo-European (The Mouton Handbooks of Indo-European Typology 1), Götz Keydana/ Wolfgang Hock/ Paul Widmer (eds.). Berlin/ Boston: De Gruyter Mouton, 477‒561. Ittzés, Máté (2021b): “Melle dulcior: Equative or Comparative? ” In: Linguisticae Disser‐ tationes. Current Perspectives on Latin Grammar, Lexicon and Pragmatics. Selected Papers from the 20th International Colloquium on Latin Linguistics (Las Palmas de Gran Canaria, Spain, June 17-21, 2019), Antonio María Martín Rodríguez (ed.). Madrid: Ediciones Clásicas, 235‒248. Keydana, Götz/ Hock, Wolfgang/ Widmer, Paul (2021): “Comparison and Gradation in Indo-European: Introduction and Overview”. In: Comparison and Gradation in Indo-European (The Mouton Handbooks of Indo-European Typology 1), Götz Keydana/ Wolfgang Hock/ Paul Widmer (eds.). Berlin/ Boston: De Gruyter Mouton, 1‒33. 210 Máté Ittzés <?page no="211"?> Kühner, Raphael/ Stegmann, Carl ( 5 1992): Ausführliche Grammatik der lateinischen Sprache. 2. Teil. Satzlehre. 1.-2. Band. Hannover: Hahnsche Buchhandlung. Leo, Friedrich (ed.) (1896): Plauti Comoediae. Vol. 2. Berlin: Weidmann. Lindner, Thomas (1996): Lateinische Komposita. Ein Glossar vornehmlich zum Wort‐ schatz der Dichtersprache. Innsbruck: Institut für Sprachwissenschaft der Univer‐ sität Innsbruck. Lindsay, Wallace M. (ed.) (1905): T. Macci Plauti Comoediae. Vol. 2. Oxford: Clarendon. Lorenz, August O. F. ( 2 1883): Ausgewählte Komödien des T. Maccius Plautus. Bd.-2. Mostellaria. Erklärt von August O. F. Lorenz. Berlin: Weidmann. Löfstedt, Einar (1911): “Plautinischer Sprachgebrauch und Verwandtes”. Glotta 3: 171-191. Löfstedt, Einar ( 2 1956): Syntactica. Studien und Beiträge zur historischen Syntax des La‐ teins. 2. Teil. Syntaktisch-stilistische Gesichtspunkte und Probleme. Lund: Gleerup. Luraghi, Silvia (2010): “Adverbial Phrases”. In: New Perspectives on Historical Latin Syntax. Vol. 2. Constituent Syntax: Adverbial Phrases, Adverbs, Mood, Tense, Philip Baldi/ Pierluigi Cuzzolin (eds.). Berlin/ New York: De Gruyter Mouton, 19‒107. LSJ = A Greek-English Lexicon (1996): Henry George Liddell/ Robert Scott (eds.), rev. by Sir Henry Stuart Jones/ Roderick McKenzie. With a revised supplement. Oxford: Clarendon. Núñez Romero-Balmas, Salvador (2002): “El ecuativo latino con atque. ¿Extensión o reanálisis? ”. In: La comparación en latín, Empar Espinilla/ Pere J. Quetglas/ Maria Es‐ peranza Torrego (eds.). Madrid/ Barcelona: Universidad Autónoma de Madrid/ Uni‐ versitat de Barcelona, 149-170. OLD = Oxford Latin Dictionary (1968): Peter G. W. Glare (ed.). Oxford: Clarendon. Orlandini, Anna/ Poccetti, Paolo (2010): “À propos des tournures exprimant une com‐ paraison élative (« melle dulcior ») et de leurs évolutions romanes”. In: Le syntagme nominal en latin. Nouvelles contributions. Actes de l’atelier du Centre Alfred Ernout Université de Paris-Sorbonne (Paris IV), 11 octobre 2008, Olga Spevak (ed.). Paris: L’Harmattan, 183-198. Panhuis, Dirk (2015): Lateinische Grammatik, übers. von Roland Hoffmann. Berlin/ München/ Boston: De Gruyter. Pinkster, Harm (2015): The Oxford Latin Syntax. Vol. 1. The Simple Clause. Oxford: Oxford University Press. Puhvel, Jaan (1973): “Nature and Means of Comparison in Proto-Indo-European Grammar”. Journal of Indo-European Studies 1: 145-154. Rebling, Oskar (1883): Versuch einer Charakteristik der römischen Umgangssprache. Kiel: Lipsius/ Tischer. Speculo claras or speculoclaras? 211 <?page no="212"?> Risch, Ernst (1954): “Zusammenfassender Literaturbericht für die Jahre 1939-1950”. Glotta 33: 179-227. Ritschl, Friedrich (ed.) (1852): T. Macci Plauti Mostellaria. Bonn: H. B. Koenig. Rosén, Hannah (1999): Latine loqui. Trends and Directions in the Crystallization of Classical Latin. München: Wilhelm Fink. Salonius, Aarne H. (1920): Vitae patrum. Kritische Untersuchungen über Text, Syntax und Wortschatz der spätlateinischen Vitae patrum (B. III, V, VI, VII). Lund: Gleerup. Segal, Erich (transl.) (1996): Plautus Four Comedies. The Braggart Soldier, The Brothers Menaechmus, The Haunted House, The Pot of Gold. Oxford: Oxford University Press. Slater, Niall W. (2016): “Speculating in Unreal Estate: Locution, Locution, Locution”. In: Roman Drama and its Contexts, Stavros Frangoulidis/ Stephen J. Harrison/ Gesine Manuwald (eds.). Berlin/ Boston: De Gruyter, 43-65. Spengel, Andreas (1865): T. Maccius Plautus. Kritik, Prosodie, Metrik. Göttingen: Vandenhoeck & Ruprecht. Stefanelli, Rossana (1984): “Problemi di sintassi indoeuropea: alcune riflessioni sulla comparazione di disuguaglianza”. Studi e Saggi Linguistici 24: 187-225. Stein, J. Peter (1971): “Compound Word Coinages in the Plays of Plautus”. Latomus 30: 598‒606. Tarriño, Eusebia (2011): “Comparative clauses”. In: New Perspectives on Historical Latin Syntax. Vol. 4. Complex Sentences, Grammaticalization, Typology, Philip Baldi/ Pierluigi Cuzzolin (eds.). Berlin/ Boston: De Gruyter Mouton, 373-426. ThLL = Thesaurus linguae Latinae (1900 ff.): Michael Hillen et al. (eds.). Vol. 1 ff. Leipzig/ München/ Berlin: Teubner/ Saur/ De Gruyter. Torrego, Maria Esperanza (2002): “Los SN comparativos: el segundo terminó de la comparación”. In: La comparación en latín, Empar Espinilla/ Pere J. Quet‐ glas/ Maria Esperanza Torrego (eds.). Madrid/ Barcelona: Universidad Autónoma de Madrid/ Universitat de Barcelona, 251-279. Wackernagel, Jakob (1905): Altindische Grammatik. Bd. II/ 1. Einleitung zur Wortlehre. Nominalkomposition. Göttingen: Vandenhoeck & Ruprecht. 212 Máté Ittzés <?page no="213"?> The Latin future tense formation of the type faxō and its Italic background Reiner Lipp Abstract: According to the approach of the present study, the Old Latin s-future of the type faxō, faxis, faxit represents a durative-telic, i.e. ter‐ minative, future tense formation, showing a regular thematic remod‐ elling of a semi-thematic s-future paradigm inherited from Proto-Italic, i.e. *fak-s-ō, *fak-s-si, *fak-s-ti etc. (with thematic vowel of o-colour in 1.sg., 1.pl., 3.pl. and athematic forms in the rest of the paradigm), which is based on an athematic PIE s-desiderative with inchoative-in‐ tentional value. The associated subjunctive of the type faxim, faxīs, faxit structurally represents the optative in *-īof the original athematic paradigm. Accordingly, the Sabellic s-future category represented by Oscan fust, Umbrian fust, fust = / fu-s-t/ ‘will be’ goes back to the same Proto-Italic semi-thematic paradigm type with athematic 3.sg. *fū-s-ti (see the conclusions in §§-18.1.-4.). Keywords: faxo type, future tense, desiderative, terminative, Old Latin, Sabellic, Proto-Italic 1. The Italic branch possesses a future tense category formed with the suffix -sthat structurally continues the inherited class of the so-called PIE “desiderative”, i.e. an athematic formation in *-sin inchoative-inten‐ tional function with the meaning ‘to be going to’. In Old Latin and in occasional archaic use within Classical Latin, this category is represented by the synchronically thematic future formation of the type faxō, -is, -it ‘will make/ arrange/ achieve’ (< *fak-s-ō, *fak-s-e-s, *fak-s-e-t) with subjunctive faxim, -īs, -it ‘may make/ arrange/ achieve’ (< optative *fak-s-ī-m, *fak-s-ī-s, *fak-s-ī-d), which is paradigmatically associated with the present faciō, -ere ‘make’. In Sabellic, this category is represented by athematic future <?page no="214"?> forms of the type of Oscan fust, Umbrian fust, fust = / fu-s-t/ ‘will be’, which are paradigmatically associated with the present Oscan est, íst, Umbrian est, est ‘is’. The underlying PIE formation has left its traces in different branches of Indo-European. Partly the respective formations have become lexicalized, partly they have kept and further developed a distinct functional grammatical value. 2.1 Among the lexicalized forms of the Proto-Indo-European desiderative based on the same morphological stem formation, we find, for instance, Old Avestan cōišt ‘determines, assigns’ ( ← *čai̯ š < *čai̯ š-d̥ < PIE *k u̯ éi̯ s-t, with unreleased final dental *-t = *[-d̥ ] being lost in Avestan after sibilant and for morphological reasons partially restored as fortis -t, in contrast to its pres‐ ervation in OAv. cōrǝt̰ ‘makes’ < *čar-d̥ < PIE *k u̯ ér-t; cf. Hoffmann/ Forssman 2004: 98 f., Lipp 2016: 258 f.) as well as Old Irish ad: cí ‘sees’ (< *ad-k u̯ is-e-t(i)), Gaulish ni … appisetu ‘you shall not see’ (< *nē ad-k u̯ is-e-tōd) as reflexes of the allomorphic stem variance of underlying Proto-Indo-European *k u̯ éi̯ -s-/ *k u̯ i-s- ‘to be going to perceive’ = ‘to pay attention to, to see to it’, which represents an athematic manner of action formation of the amphikinetic class (i.e. displaying mobile accent) that is based on the PIE root *k u̯ ei- ‘to perceive’, as underlying Vedic cā́yati ‘perceives, observes’, Old Church Slavonic čajǫ, čajati ‘expect, hope’ < *k u̯ ḗi̯ -e/ o- ← PIE *k u̯ ḗi̯ -/ *k u̯ éi̯ - (LIV 2 381 f., 377 f.; on the athematic amphikinetic desiderative of the PIE type *u̯ éi̯ d-s-/ *u̯ id-s- ‘to be going to see’ beside the reduplicated thematic desiderative of the PIE type *u̯ i-u̯ n̥ -sé/ ó- ‘to be going to win/ gain’, see LIV 2 23 f., 39 with refs.; for an overview of the morphological types of Indo-European sigmatic desiderative/ future formations with different analyses, cf. Jasanoff 1988: 232-234, 2003: 133; Szemerényi 1989: 307-312, 1996: 285-288, and the comprehensive treatment by Hill 2004; for a classification in the frame of the IE verbal s-stem formations in general, cf. Søborg 2020a). 2.2 A functional telic value of the stem formations with s-extension as represented by lexicalized roots of the type *k u̯ ei̯ -shas been advocated by Cohen (2017), Friis (2018: 77-84), Søborg (2020b: 640). Yet, there are quite a number of unextended roots underlying s-extensions like *mei̯ ḱ- ‘to mix’ (as basis of *mi̯ eḱ-s- ‘to adhere’, represented by root aorist Vedic ámyak ‘has adhered’ < *é-mi̯ eḱs-t; LIV 2 445 with n. 1; on the root shape with Schwebeablaut, see below § 4.) that either show a root aorist indicating a telic value of the unextended root (like *mei̯ ḱ-/ *miḱas continued by Homeric Greek ἔμικτο ‘mixed with, joined together (itr.)’; LIV 2 428f. with n. 1) or presuppose 214 Reiner Lipp <?page no="215"?> it for their paradigm structure, for instance due to the existence of a perfect formation designating the state of the subject resulting from a previous verbal event with endpoint, thus implying a telic meaning of the underlying root as in the case of *k u̯ ei̯ - ‘to perceive’ with perfect *k u̯ e-k u̯ ói̯ -/ *k u̯ e-k u̯ i- ‘to have perceived, to look over’ > Ved. cikā́ya ‘inspects, recognizes’ beside a paradigmatically independent, but etymologically perhaps related root aorist *k u̯ éi̯ -/ *k u̯ i- > Ved. ví acet ‘has separated’, OAv. vī-š ́ iiātā = / vi-ši̯ -ata/ 3.pl. injunctive middle ‘they separate’ (LIV 2 377-379 s.v. 1.*k u̯ ei̯ - ‘to perceive’ with n. 1; 2.*k u̯ ei̯ - ‘to collect, layer’ with nn. 1, 2 also on the possible para‐ digmatic disjunction of etymologically related stem forms, in this case by functional split via the semantically transitive nasal present *k u̯ i-néu̯ -/ -nu- ‘to perceive with effect on the object’ > Ved. cinóti ‘layers, collects’, OAv. vī-cinaot̰ ‘separates’). Therefore, the function of the s-extensions cannot have been confined to the telic value alone. Consistent with this, the functional class of Latin s-futures of the type faxō ‘will make/ arrange/ achieve’ turns out to be durative-telic, i.e. terminative (see below §§-10., 13.4.3.-4.). This reflects the use of the PIE desiderative as an inchoative-intentional category, which was not restricted to the expression of the desire of the subject, but implied the intention for the verbal act to be fulfilled, comparable to the function of the reduplicated desiderative formation in *-sas continued by the Indo-Aryan desiderative (e.g. Ved. vívāsati ‘seeks to gain’ ← PIE *u̯ i-u̯ n̥ -sé/ ó-; see above § 2.1.), which by the ancient Vedic grammarians is described by paraphrases with iṣ ‘to seek, strive for’ (cf. Kümmel 2010: 180-182, based on Heenen 2006: 36-73, who for the desiderative, however, o.c. 50 f. designates the not yet finished course of action towards the endpoint as non-terminative, which should rather be labelled as not terminated or unfinished; yet, for such an action in progress with an extrapolated endpoint beyond the frame of the reference period, the term ‘durative-telic’ = ‘terminative’ is used in the present paper; on the terminology, see below §-10.). 3.1 A related type of formation is to be seen in the Indo-Iranian imperative in *-si, which continues the 2.sg. of the indicative of an athematic stem in *-s-, i.e. structurally a form in *-s-si. It is attested above all in Indo-Aryan, cf. Vedic sátsi ‘sit down! ’ (< PIE */ sétsi/ = *séd-s-si, appearing in a paradigm with a root aorist *séd-/ *sd-, as represented by Ved. subj. 2.dual sádathas ‘you will sit down’ and thematic aorist ásadat ‘has sat down’; LIV 2 513), and, as the only example from Iranian, Old Avestan dōiši (< Proto-Indo-Iranian *dái̯ č-ši = */ dai̯ ć-ši/ < PIE */ déi̯ ḱ-si/ = *déi̯ ḱ-s-si, appearing in a paradigm showing as a The Latin future tense formation of the type faxō and its Italic background 215 <?page no="216"?> post-PIE new formation the s-aorist *dḗi̯ ḱ-s-/ *déi̯ ḱ-s-, as represented within Indo-Iranian by OAv. injunctive dāiš ‘show! ’, subjunctive 1.sg. dōišā ‘I will show’, beside an inherited root aorist *déi̯ ḱ-/ *diḱ- > Ved. middle ádiṣṭa ‘has shown’ with zero-grade of the root; Narten 1964: 22 n. 22; 27, 42 n. 85; 80, 140 f.; LIV 2 108 f. with n. 2). Often such forms appear in combination with an imperative, e.g. RV 1, 14, 1 ā́ devébhir yāhi yákṣi ca ‘with the gods come up (= yāhi imperative) and sacrifice (= yákṣi si-imperative of yaj-)! ’ (Delbrück 1888: 365). As the paradigmatic distribution shows, the si-imperative is not conditioned by the occurrence of an s-aorist in the same paradigm, despite the frequent paradigmatic co-ocurrence of both formations. Instead, the si-imperative represents an original formation independent from the system of grammatical tense stems (Narten 1964: 39, 45-49). As such, in Vedic it partially appears beside a root aorist (or a thematized root aorist), like in the aforementioned case of si-imperative sátsi ‘sit down! ’ beside root aorist subj. sádathas (and thematic aor. ásadat); e.g. si-imperative śróṣi ‘listen! ’: root aorist áśrot ‘has heard’; si-imperative nákṣi ‘attain! ’ : root aorist ā́naṭ ‘has attained’ (Narten 1964: 47); and partially the si-imperative can be found beside a root aorist with a secondary subjunctive in -sa- (probably triggered by the presence of the imperative in -si); e.g. darṣi ‘break open! ’ : root aorist adar ‘you have split apart’, injunctive dárt ‘splits apart’ : sa-subjunctive 2.sg. darṣasi, 3.sg. darṣati, middle darṣate ‘will split apart’ (Narten 1964: 39, 145); and frequently it occurs beside an inherited or newly formed s-aorist (as replacement of an original root aorist) with a paradigmatically regular subjunctive in -s-a- (the -sformant being part of the stem of the s-aorist); e.g. si-imperative yakṣi ‘sacrifice, worship! ’ : s-aorist ayāṭ, middle ayaṣṭa ‘has sacrificed, has worshipped’ : subjunctive yakṣat ‘will sacrifice, will worship’ (Narten 1964: 200-202); si-imperative vakṣi ‘drive! ’ : s-aorist ávāṭ ‘has driven’ : subjunctive vakṣati, vakṣat ‘will drive’ (= YAv. uz-uuažat̰ ‘will drive out’; Narten 1964: 240 f.; LIV 2 661 f.); si-imperative yaṁsi ‘stretch out! ’ : s-aorist 1.sg. ayāṁsam, 3.sg. ayān ‘have/ has stretched out’ beside root aorist with 3.pl. ayamur ‘have stretched out’, injunctive 1.sg. yamam, 3.pl. yaman (with a corresponding inherited root aorist in OAv. injunctive middle apa-iiaṇtā ‘takes away’; Narten 1964: 204 f. with n. 609; LIV 2 312). 3.2 Since forms of the si-imperative occur in a number of paradigms containing a root aorist not associated with a newly formed s-aorist or a renewed subjunctive in -s-a-, there is no basis for the assumption that they resulted within the paradigm of an s-aorist from haplology of a 2.sg. sub‐ 216 Reiner Lipp <?page no="217"?> junctive in *-s-e-si > *-sasi to *-si (on this assumption going back to Cardona 1965, 1968: 55-60; Szemerényi 1966; cf. Gotō 2013: 114-116 with nn. 255-257 for further refs., the alleged haplology *-s-e-si > *-si being regarded as an already PIE process by Jasanoff 2003: 181 n. 384, 182 f., 2012: 116 f., 2019: 21). Accordingly, the fact that forms of the si-imperative are also used in relative clauses and subordinate clauses with the relative conjunction yád ‘that, when’ in the modal function of a subjunctive (Narten 1964: 202, 205 with n. 609; Gotō 2013: 115 with n. 257) does not at all mean that they continue an original subjunctive formation. Instead, this usage reflects the original status of the so-called “desiderative” with its inchoative-intentional value, which underlies not only its use as an imperative, e.g. Ved. hoṣi ‘pour out, sacrifice! ’ (< PIE */ ǵ h éu̯ si/ ‘you are going to pour out’ = *ǵ h éu̯ -s-si, belonging to the desiderative stem *ǵ h éu̯ -s-, e.g. in Greek future χέω, χεύω ‘will pour out’), but also the subjunctive-like prospective function evidenced in subordinate clauses like RV 3, 30, 18 yán mahī ́ r íṣa āsátsi pūrvī ́ ḥ rāyó vantā́ro br̥ hatáḥ syāma ‘when you sit down (= ā-sátsi) to many great refreshments we would like to be the winners of abundant wealth’, with ā-sátsi ‘you will sit down’ (< PIE */ sétsi/ = *séd-s-si) continuing the desiderative meaning *‘you are going to sit down’. 4. Anatolian offers lexicalized forms like Hittite tamāšzi, damašzi = Cunei‐ form Luwian damasti ‘(op)presses’, Old Hittite pret. tamāšta = / damā́s-/ < Anatolian *dḿ̥ h 2 s-t(i). The Anatolian paradigm either results from levelling of the PIE desiderative *démh 2 -s-/ *dm̥ h 2 -s- ‘to be going to subdue’ from *demh 2 - ‘to subdue, tame’ and corresponds morphologically to Ηomeric δαμᾷ ← *δεμάει < *dema-s-e/ o- ← *démh 2 -s-, which in Greek represents the functional future tense formation of δάμνημι ‘tame’ (LIV 2 116 f. s.v. 2.*demh 2 with nn. 7-11). Or Anatolian shows the levelled outcome of PIE *dméh 2 -s-/ *dm̥ h 2 -s-, being remodelled from the stem *démh 2 -s-/ *dm̥ h 2 -swith Schwebeablaut, i.e. showing metathesis of the resonant in order to prevent a triconsonantal cluster within the obstruent-containing stem final, as in the case of some desiderative formations that became lexicalized as PIE roots, such as *h 2 leksfrom *h 2 elk- ‘to ward off’, *h 2 u̯ eksfrom *h 2 eu̯ g- ‘to become strong, grow’, *mi̯ eḱs- ‘to adhere’ from *mei̯ ḱ- ‘to mix’ (Kloekhorst 2008: 823 f., followed by Søborg 2020b: 636-639, o.c. with comprehensive discussion of the s-extended verbs in Hittite, on which see below § 12.4.4.; on original desiderative stems in *-sshowing Schwebeablaut, see above § 2.2. and below § 5.3.). Furthermore, the Hittite imperative paḫši ‘protect! ’ The Latin future tense formation of the type faxō and its Italic background 217 <?page no="218"?> has been regarded by Jasanoff (2003: 182 f.; 2012, 2019: 20-22) as a case of an inherited si-imperative comparable to that of Indo-Iranian (see above §§ 3.1.-2.), but, according to Oettinger (2007), Søborg (2020a: 193-195), the form is rather to be analysed as / pahs-i/ with the synchronic imperative ending -i (used with actives and transitive middles and alternating with -Ø), built on the verbal stem / pahs-/ ‘to protect’ as represented by active 1.sg. paḫḫašḫi, paḫḫašmi and in old texts prevailingly by middle 1.sg. paḫḫašḫa, 3.sg. paḫša(ri) ‘protect; (in the middle also: ) to seek protection’ (cf. Kloekhorst 2008: 611 f.), being a continuation of PIE *peh 2 -s- ‘to seek to protect’ = Old Church Slavonic pasǫ, pasti ‘to graze, herd’ as desiderative of *peh 2 - ‘to protect’ = Ved. pā́ti ‘protects’ (cf. LIV 2 460). 5.1 Ηomeric δαμᾷ < *dema-s-e/ orepresents the functional future tense of δάμνημι ‘tame’ and belongs morphologically to the Greek future tense category of the type ὄψομαι ‘I will see’ < *ok u̯ -s-e/ oshowing the suffix form *-s-e/ o-, which seems to go back to a modal formation of a verbal s-stem with thematic vowel, i.e. a subjunctive. But for the lack of an s-aorist formation of the root *h 3 ék u̯ in PIE and Greek, the form cannot continue a subjunctive of the s-aorist (LIV 2 297f. s.v. *h 3 ék u̯ -). Likewise, the future tense forms ἐλεύσομαι ‘I will come’ < *eleu̯ t h -s-e/ oand πείσομαι ‘I will suffer’ < *k u̯ ent h -s-e/ oappear within the paradigm beside the thematic aorist forms Hom. ἤλυθον ‘came’ < *elut h -e/ oand ἔπαθον < *k u̯ ɔt h -e/ o- ‘suffered’, which are based on PIE root aorists (*h 1 léu̯ d h -/ *h 1 lud h -, LIV 2 248 f. with n. 14; *k u̯ énd h -/ *k u̯ n̥ d h - < *k u̯ ém-d h h 1 -/ *k u̯ m̥ -d h h 1 or *k u̯ énth 2 -/ *k u̯ n̥ th 2 -, LIV 2 390 with n. 6; on the root shape LIV 2 390 n.1; Lipp 2009a: 58 n. 156 and Neri 2005: 220 f. n. 72, who pleads for *k u̯ enth 2 - > Proto-Greek *k u̯ ent h with *Th 2 > *t h after a continuous sound = resonant, laryngeal, s), without there being an s-aorist inherited from PIE or as a new formation in Greek (for ἐλεύσομαι and πείσομαι as continuations of a PIE desiderative, cf. Rix 1976: 223, 225; Weiss 2009: 419). Due to the missing evidence in the respective paradigms for an s-aorist in PIE or Greek and of a verbal stem with e-grade of the root in Greek and because of the paradigmatically isolated status of ὄψομαι as suppletive future of ὁράω ‘see’, an alternative interpretation of the Greek s-future as a formation based on the subjunctive of the s-aorist or as a productive new formation is excluded (pace Rasmussen 1985: 396 f. n. 36; Willi 2018: 441-452). Moreover, the frequent inflection of the Greek s-future as medium or medium tantum has to be taken into account as it is consistent with its originally desiderative function (in view of the use 218 Reiner Lipp <?page no="219"?> of the medium as a functional expression of an action in the interest of the subject or implying special involvement of the subject as in verbs of affect and perception; on the connection of the use of the medium with the desiderative meaning, cf. Chantraine 1984: 247; Weiss 2009: 419). This corresponds to the frequent (but not prevailing) use of the medium and occurrence of media tantum in the category of the (reduplicated thematic) Vedic desiderative (on the use of diathesis in this category, cf. Heenen 2006: 17-22), cf. with middle inflection Gk. ὄψομαι ‘I will see’ and etymologically related Ved. ī ́ kṣe ‘I look at’ (Heenen 2006: 56 f., 82-89), Gk. γνώσομαι ‘I will recognize’ and etymologically related Ved. jijñāse ‘I seek to recognize’ (Heenen 2006: 136-138). Likewise, in the case of Hittite inchoatives, there is a tendency to use middle inflection (Sergio Neri p. c.), cf., as denominative of / hassuezzi-/ ‘royal status’, the verbal stem / hassuezzii̯ e/ a-/ ‘to become king’ with 3.sg. preterite middle lugal-ez-zi-i̯ a-at-ta-[at], lugal-u-ez-zi-et-ta-at beside active lugal-ez-zi-at. 5.2 For these reasons, the future tense forms of the type ὄψομαι rather result from a remodelling of the inherited athematic formation of the basic PIE desiderative *h 3 ék u̯ -s-/ *h 3 k u̯ -s- ‘to be going to see’, which is due to the wide-spread, but conditioned Greek replacement of the endings *-énti and *-éntoi̯ in the 3.pl. active and middle by *-onti and *-ontoi̯ , thereby triggering the thematization of the whole paradigm. Except for the verbs ἐστί(ν) ‘is’ = / es-tí/ with 3.pl. εἰσί(ν) = / ẹ̄ sí/ < *ẹ̄nsi < *ehenti < *es-enti (< PIE *h 1 s-énti) and εἶσι(ν) ‘(s)he will go’ = / ẹ̄ ̃ -si/ (< *h 1 éi̯ -ti) with 3.pl. Hom. εἴσ’ = / ẹ̄ ́ s(i)/ < *ensi < *enti (without aspiration after the rest of the paradigm) ← *henti (< PIE *h 1 i̯ -énti; acc. to Neri 2018: 260 f. n. 24, Schindler’s Law *Hi̯ - > Gk. his to specify as *h 1/ 2 i̯ - > Gk. / h-/ vs. *h 3 i̯ -, *i̯ - > Gk. / dz-/ ), as configurations implying early loss of regular morphological segmentation, such a mechanism of replacement was effective in all paradigms where the vowel colour of the ending allomorphs *-énti and *-éntoi̯ was not backed by the rest of the paradigm. Such a remodelling is therefore in a distributional contrast to the preservation of these ending allomorphs in athematic paradigms with stems ending in the vowel -elike 3.pl. present active Hom. τιθεῖσι < *tí-t h e-nsi < *t h i-t h -enti ( ← PIE *d h é-d h h 1 -n̥ ti = Ved. dádhati) beside 1.pl. τίθεμεν = / tí-t h e-men/ , 2.pl. τίθετε = / tí-t h e-te/ and in 3.pl. present middle τίθενται = / tí-t h e-ntai/ < *t h i-t h -entoi̯ ( ← *d h é-d h h 1 -n̥ toi̯ = Ved. dádhate) beside 3.sg. τίθεται = / tí-t h e-tai/ , 1.pl. τιθέμεθα = / ti-t h é-met h a/ , 2.pl. τίθεσθε = / tí-t h e-st h e/ etc. On The Latin future tense formation of the type faxō and its Italic background 219 <?page no="220"?> the other hand, in the case of the present stem PIE *ǵi-ǵénh 1 -/ *ǵi-ǵn̥ h 1 - ‘to beget, bear’ (from *ǵénh 1 -), there is 3.pl. middle PIE *ǵi-ǵn̥ h 1 -éntoi̯ > *ǵi-ǵn-éntoi̯ (reduplicated formation with PIE laryngeal loss after resonant as in the second member of a compound like Gk. νεογνός ‘newborn’ < PIE *neu̯ o-ǵnó- < *neu̯ o-ǵn̥ h 1 -ó-; cf. Lipp, 2009b: 367 f. with n. 39 for further refs.; for an alternative proposed by Neri, see immediately below), which in Greek undergoes a regular remodelling to *gign-ontoi̯ > γίγνονται, triggering the thematic inflection of the present tense paradigm with 1.sg. γίγνομαι, 3.sg. γίγνεται etc. (instead of continuing the representation of *ǵi-ǵn̥ h 1 -mói̯ , *ǵi-ǵn̥ h 1 -tói̯ > *ǵi-ǵn̥ -mói̯ , *ǵi-ǵn̥ -tói̯ with expected post-resonant laryngeal reduction in a reduplicated formation as in *ker-kr̥ (H)-ti- > Ved. carkr̥ tí- ‘praising, glory’ beside *kr̥ H-ti- > Ved. kīrtí- ‘fame, glory’, in accordance with compounds as *h 1 sú-su(H)ti- > Ved. sú-ṣuti- ‘easy birth’ beside *suH-tí- > Ved. sūtí- ‘birth’; for examples Mayrhofer 1986: 129, 140, 149, 2005: 98 ff., 103 f.; however, following Neri 2017: 275-341, esp. 317, with conclusion 343, in amphior hysterokinetic nominal paradigms including compounds as well as in mobile verbal paradigms this phenomenon can probably be traced back to the pre-tonic PIE laryngeal reduction according to the “Wetter-Regel” as in *h 2 u̯ eh 1 -tró- ‘the blowing’ = ‘air in motion’ > *h 2 u̯ etró- > Proto-Germanic *weðra n ‘weather’, i.e. *VHTRV́ , *R̥ HTRV́ > *VTRV́ , *R̥ TRV́ , whereas, on a post-PIE stage after laryngeal loss, a reduplicated form as *ǵi-ǵn̥ h 1 -éntoi̯ > *ǵi-ǵn̥ -éntoi̯ > *ǵi-ǵn-éntoi̯ or a compound as *neu̯ o-ǵn̥ h 1 -ó- > *neu̯ o-ǵn̥ -ó- > *neu̯ o-ǵnómight show an optional post-Proto-PIE reversal of Lindeman’s Law in a plurisyllabic sequence, as in the cases described by Neri 2017: 156 f., 165 with n. 290, 222 n. 71, 224 n. 80, 262, 271, 292 n. 73, 2019: 50). This process corresponds to the wide-spread thematization of root aorists in Greek, as in the case of PIE *g u̯ élh 1 -/ *g u̯ l̥ h 1 - ‘to hit by throwing’ with 3.pl. *g u̯ l̥ h 1 -ént (*-t = *[-d̥ ] = */ -d/ ), which in Greek undergoes reshaping of *g u̯ ɔl-end to 3.pl. *g u̯ ɔl-on(d) > Hom. βάλον, ἔβαλον, thereby resulting in the emergence of the thematic aorist represented by 1.sg. βάλον, ἔβαλον, 3.sg. βάλε(ν), ἔβαλε(ν) etc. Likewise, in the associated nasal infix present *g u̯ l̥ -né-h 1 -/ *g u̯ l̥ -n-h 1 with 3.pl. *g u̯ l̥ -n-h 1 -énti, Greek shows a reshaping of 3.pl. *g u̯ ɔln-enti to *g u̯ ɔln-onti > *g u̯ ɔln-onsi > βάλλουσι with subsequent thematic inflection of the present tense paradigm as 1.sg. βάλλω, 3.sg. βάλλει etc. (Harðarson 1993: 160-162 with n. 69, 165 f.). 5.3 In Greek, moreover, comparable forms based on a PIE desiderative in *-sare found as lexicalized items like ἀλέξω ‘I ward off’ = Ved. rákṣati 220 Reiner Lipp <?page no="221"?> ‘protects’, representing the secondary root form *h 2 leks-, which goes back to the original desiderative stem *h 2 élk-s-/ *h 2 l̥ k-s- (with the strong stem *h 2 élk-s- > *h 2 léksshowing Schwebeablaut by resonant metathesis to avoid consonant clustering in stem-final position), as derived from the root *h 2 elk-/ *h 2 l̥ k- ‘to ward off’ in Gk. redupl. aor. ἄλαλκε ‘warded off’, ἀλκή ‘defence’, Goth. alhs ‘temple, shrine’, Lith. al ̃ kas ‘holy grove’ (LIV 2 264, 278); likewise, Gk. Hom. ἀέξομαι ‘grow’, ἀέξω ‘increase’ < *au̯ eks-e/ o- ~ OAv. aor. vaxšt ‘grows (in strength [as inner object]), makes grow (the strength)’ represents the secondary root form *h 2 u̯ eks- ‘to grow’, which is based on the original de‐ siderative stem *h 2 éu̯ g-s-/ *h 2 ug-s- (with the strong stem *h 2 éu̯ g-s- > *h 2 u̯ éksshowing Schwebeablaut), as derived from *h 2 eu̯ g-/ *h 2 ug- ‘to become strong, grow’ in Ved. ójas-, OAv. aogah-, YAv. aojah- ‘strength’ < *h 2 éu̯ g-os/ -és-, Goth. aukan ‘increase (itr.)’, Lat. augeō, -ēre ‘inrease (tr.)’ (LIV 2 274 f., 288 f.). (For the phonological motivation of Schwebeablaut and an overview of the respective cases, cf. Ozoliņš 2015: esp. 91-98, 99 on the forms cited; on Schwebeablaut in desiderative stems in *-s-, see above §§-2.2., 4.). 6. Furthermore, in the paradigm belonging to the present Old Lithuanian duosti ‘gives’ (< *dōt s̱ ti < *dōd- ← *ded- < *de-dh 3 -, reduplicated present of *dō- < *deh 3 -), there is the regular type of the Lithuanian future tense formation with 3.sg./ 3.pl. duõs ‘will give’ (metatony) < *dṓ-s-t(i), dial. 1.pl. dúosme < *dṓ-s-me going back to the PIE desiderative *déh 3 -s-/ *dh 3 -sfrom *deh 3 - ‘to give’ (cf. the Greek future δώσω as equivalent with secondary thematization as described above § 5.2.; on the Baltic paradigm cf. Stang 1966: 397-399; see the detailed historical treatments by W.P. Schmid 1963; Hill 2004: 73- 100, 152-155; 2014). Such forms appear in paradigmatic connection with Lithuanian future tense formations showing the formant -si-, as 1.sg. dúosiu, 2.sg. dúosi, 3.sg. Old Lith. duosi, and 1.pl. dúosime, 2.pl. dúosite. These future formations in -siare either supposed to be generalized from original 3.pl. *dṓ-si-nt(i) < *dō-s-n̥ ti, i.e. from a paradigmatic item of the same athematic s-stem formation (as per Jasanoff 1978: 105-107, 1988: 233, 2003: 133 n. 298; Weiss 2009: 419; Villanueva-Svensson 2010: 218-221), or they are considered as the regular Baltic semi-thematic representation of the inherited future suffix *-si̯ e-, with Lith. 1.sg. dúosiu, Old Lith. 3.sg. duosi delivering the direct match of Ved. dāsyāmi*, dāsyáti < PIE *deh 3 -si̯ é/ óand corresponding morphologically to Celtic formations like Gaulish pissíiumí ‘I will see’ < *k u̯ is-si̯ é/ ó- (as prevailingly assumed, cf. Szemerényi 1989: 307-312, 1996: 285-288 with further refs.), which, however, leaves unexplained a form like The Latin future tense formation of the type faxō and its Italic background 221 <?page no="222"?> duõs ‘will give’. Therefore, Hill (2004, 2014) pleads for a Proto-East-Baltic conflation of the athematic s-formation and the si̯ e/ o-formation in one single future tense paradigm, rejecting for good reasons the aforementioned derivation from an athematic paradigm in *-svia 3.pl. *-si-nt(i) < *-s-n̥ ti (cf. Hill 2004: 94-100, 2014: 54-57). 7. Moreover, there is a possible, but disputable, correspondence in the case of the Old Irish s-subjunctive, e.g. 3.sg. téis, conj. : téi, : té ‘shall go’ < *tēss-ti, *tess-t# ← Insular Celtic *tēss-e-ti, *tēss-e-t# < *tēχs-e-ti (initial position), *tēχs-e-t# (final position, with *-t# < *-ti as secondarily generalized primary ending from the indicative of the present tense) ← Proto-Celtic *tēχse < *stēksed < *stei̯ g h -s-e-t (with *-t = *[-d̥ ] = / -d/ ), with the original thematic inflection of Insular Celtic *tēss-e-t(i) being presupposed by the cluster *-ss-V < *-χs-V, in contrast to *-χt- < *-χstas in the paradigm of OIr. aigid, : aig ‘drives, leads’ with preterite 3.sg. acht, conj. : acht < Insular Celtic *aχti, *aχt# < *aχs-ti, *aχs-t# ← Proto-Celtic *aχs < *aks-d < *ag-s-t (with *-t = *[-d̥ ] = / -d/ ), representing an original athematic form in an otherwise later thematized preterite paradigm. This original athematic s-preterite, in turn, influenced the originally thematic paradigm of the s-subjunctive to become semi-thematic in Goidelic with a renewed athematic 3.sg. like *tēss-t(i) ← *tēss-e-t(i) (< *tēχs-e-t(i)). For its original thematic inflection, the category of the semi-thematic Insular Celtic s-subjunctive can thus either be interpreted as the subjunctive of the inherited desiderative or as the subjunctive of the productive s-aorist (for an original thematic inflection of the OIr. s-subjunctive Schumacher 2004: 52 f.; discussion in connection with the unreduplicated OIr. s-future by Hill 2004: 148-152 with refs.; assignment to the desiderative advocated by Rix 1977: 148, 151-154, assignment to the s-aorist by McCone 1991: 71-80). 8.1 In Latino-Faliscan, there are, on the one hand, early lexicalized items continuing a desiderative formation, as Lat. vīsō, -ere ‘visit, go and see some‐ one’ < *u̯ ei̯ d-s-e/ oor *u̯ id-s-e/ o-, representing a regular thematic remodelling of the athematic desiderative PIE *u̯ éi̯ d-s-/ *u̯ id-s- ‘to be going to see’, as related to Lat. videō, -ēre ‘see’ < *u̯ id-ē-i̯ e/ ofrom *u̯ ei̯ d- (LIV 2 665-667 s.v. *u̯ ei̯ dwith n. 16); Lat. quaesō, -ere ‘request, ask’ < *ko-ai̯ s-s-e/ o-, representing a regular thematic remodelling of the athematic desiderative PIE *h 2 éi̯ s-s-/ *h 2 is-s- ‘to be going to search’, as related to Lat. quaerō, -ere ‘seek, search’ < *ko-ai̯ s-e/ ofrom *h 2 ei̯ s- (cf. LIV 2 260 s.v. *h 2 ei̯ swith nn. 8, 10). The involved mechanism of thematic remodelling in the Italic languages was pervasive, 222 Reiner Lipp <?page no="223"?> concerning the vast majority of the inherited athematic verbal stems (see below §§-9.4.-5.). 8.2 On the other hand, within Latino-Faliscan, the inherited category of the desiderative in *-sis continued in the functional value of a future tense by the Old Latin category of indicative faxō ‘will make/ arrange/ achieve’ and the corresponding subjunctive faxim ‘may make/ arrange/ achieve’, which are associated with the paradigm of the present tense faciō, -ere ‘to make’. This morphological future tense formation in -sis attested in the Law of the XII Tables, going back with transmission-related adaptions to a text redacted in 451-450 BC according to Livy (on the question of the chronology of the redaction, cf. Warmington 1967b: xxvi-xxxiii, Crawford 1996: 556 f.), and subsequently in the early literary transmission of the 3 rd and 2 nd centuries BC with most evidence coming from Plautus. But in comparison to the later standard future tense formation of Latin (on which see below §§ 10., 17.1.- 2.) the future tense of the type faxō is paradigmatically and functionally restricted (see below §§ 10., 15.3.2.), being on the decline already from the beginning of the literary transmission (Benveniste 1922: 60: “l’histoire des formes sigmatiques de futur et de subjonctif en latin n’est que l’histoire de leur disparition.”). Despite the synchronically thematic character of the indicative future faxō, faxis, faxit < *fak-s-ō, *fak-s-e-s, *fak-s-e-t, the future subjunctive faxim, faxīs, faxit < *fak-s-ī-m, *fak-s-ī-s, *fak-s-ī-d displays the suffix *-īfrom a PIE athematic optative formation in *-ih 1 -. For the historical interpretation of the peculiarity of this Latin future tense paradigm, one has to take into account the athematic character of the morphologically related future tense formation in Sabellic, e.g. Oscan fust, Umbrian fust, fust = / fu-s-t/ ‘will be’ ( ← *fīst < *fū-s-t; inherited formation of the s-future based on the root), Oscan didest = / dide-s-t/ ‘will give’ (innovatory formation of the s-future derived from present stem *did-e/ oas in Vest. didet ‘gives’) with the corresponding subjunctive formation in -īfrom the old athematic optative, as only once evidenced for Sabellic by South Picene aisis = [ajsīs] ‘you may do/ achieve’ < *aɟezīs (with velar palatalization followed by medial syllable syncope) < *age-s-ī-s (innovatory formation of the s-future derived from present stem *ag-e/ oas represented by Osc. imptv. II actud ‘shall treat with’ < *ag-e-tōd = Lat. agitō; on this modal form of the Sabellic s-future, see below § 14.6.), matching OLat. āxim < *ag-s-ī-m (as inherited formation of the s-future based on the root; on the difference in derivation of the stem of the s-future in Sabellic and Latin, see below §§-14.1., 14.2., 14.5., 15.1.). The Latin future tense formation of the type faxō and its Italic background 223 <?page no="224"?> 9.1 As a continuation of the athematic PIE desiderative in *-s-, the Proto-Italic forerunner of the Old Latin s-future faxō ‘will make/ arrange’ had acquired a semi-thematic pattern (a state that is thus also to be presupposed for Sabellic; see below §§ 15.2., 15.3.1.), since in Proto-Italic all athematic verbs (except for *es-/ *s- ‘to be’ < PIE *h 1 és-/ *h 1 s-; see below § 16.5.) replaced the athematic endings of the 1.sg. *-mi and 3.pl. *-enti by the thematic end‐ ings *-ō and *-o-nti, and mostly that of the 1.pl. *-mos by thematic *-o-mos, as can still be seen from the present tense paradigms of the Latin irregular verbs īre ‘to go’, ēsse ‘to eat’, velle ‘to want, wish’, e.g. 1.sg. volō < *u̯ el-ō, 1.pl. volumus < *u̯ el-o-mos, 3.pl. volunt < *u̯ el-o-nti versus athematic forms such as 3.sg. vult < *u̯ el-ti in the remaining paradigm (on the semi-thematic inflectional pattern, cf. Rix 2003a: 154 f.). 9.2 These irregular verbs represent a class of root presents with conso‐ nant-final basis that in Proto-Italic got a semi-thematic inflection, which is preserved as such in the following Latin paradigm patterns (on the respec‐ tive semi-thematic paradigms, cf. Sommer 1914: 536-538, 540-542, 533 f.; Leumann 1977: 521, 525-527, 528-530; Meiser 1998: 222-224): From īre ‘to go’, 1.sg. eō, 2.sg. īs, 3.sg. it = OLat. īt, 1.pl. īmus, 2.pl. ītis, 3.pl. eunt < Proto-Italic 1.sg. *e-ō (< *ei̯ -ō), 2.sg. *ei̯ -si, 3.sg. *ei̯ -ti, 1.pl. *ei̯ -mos (athematic ending kept), 2.pl. *ei̯ -tes, 3.pl. *e-onti (< *ei̯ -o-nti) (from the amphikinetic root present PIE *h 1 éi̯ -/ *h 1 i-, with strong stem *h 1 éi̯ - > *ei̯ generalized under the influence of imptv. *h 1 éi̯ -te > *ei̯ -te > Lat. īte, Pael. eite). From ēsse ‘to eat’, 1.sg. edō, 2.sg. ēs, 3.sg. ēst, 1.pl. edimus, 2.pl. ēstis, 3.pl. edunt < Proto-Italic 1.sg. *ed-ō, 2.sg. *ēs-si (< *ēt-si < *ed-si with phonetic vowel lengthening before restored media */ d/ at morpheme juncture acc. to Lex Lachmann, with due possibility, however, that this effect is phonematized only in Latin), 3.sg. *ēs-ti ( ← *ēssi < *ēt s̱ ti < *ed-ti acc. to Lex Lachmann), 1.pl. *ed-o-mos, 2.pl. *ēs-tes ( ← *ēsses < *ēt s̱ tes < *ed-tes acc. to Lex Lachmann), 3.pl. *ed-o-nti (from the acrostatic root present PIE *h 1 ḗd-/ *h 1 éd-, with Proto-Italic generalization of the weak stem *h 1 éd- > *ed-). From velle ‘to want, wish’, 1. sg. volō, 2.sg. vīs (< OLat. veis < *u̯ els < *u̯ ell-s, with preserved e before palatal *l from the geminate *ll, due to restoration on the basis of the assimilated pre-stage of vel ‘or’, as represented by dactylic OLat. vel-uti Ennius Ann. 340V with long scansion as *u̯ ell ‘(as) you want’ < *u̯ els < *u̯ el-si; Leumann 1977: 142; Cowgill 1978; Weiss 2009: 430 with n. 23), 3.sg. volt, vult, 1.pl. volumus, 2.pl. voltis, vultis, 3.pl. volunt < Proto-Italic 1.sg. *u̯ el-ō, 2.sg. *u̯ el-si, 3.sg. *u̯ el-ti, 1.pl. *u̯ el-o-mos, 224 Reiner Lipp <?page no="225"?> 2.pl. *u̯ el-tes, 3.pl. *u̯ el-o-nti (from the acrostatic root present PIE *u̯ ḗlh 1 -/ *u̯ élh 1 -, with Proto-Italic generalization of the ante-vocalic variant of the weak stem as in 3.pl. *u̯ élh 1 -n̥ ti > *u̯ el-ǝnti → Proto-Italic *u̯ el-onti > Lat. volunt and optative *u̯ élh 1 -ih 1 -m > Proto-Italic *u̯ el-ī-m > Lat. subjunctive velim; Proto-Italic 1.pl. *u̯ el-o-mos > Lat. *u̯ elǝmos > volumus by vowel weakening with <u> as rendering of the rounded realization of the weakened vowel *ǝ before labial after rounded vowel in the preceding syllable, in contrast to the result of vowel weakening in *ag-o-mos > Lat. *agǝmos > agimus; cf. Dunkel 1998: 97; Nishimura 2010: 220). 9.3 From this it is evident that the semi-thematic inflection in Italic paradigms of original athematic verbs shows a thematic vowel in the 1.sg., 1.pl. and 3.pl., i.e. paradigmatic slots with o-vowel corresponding to the thematic inflection. According to Dunkel (1998: esp. 96 f.), the occurrence of the thematic vowel in these paradigms was triggered by the integration of original subjunctive forms of the 1.sg. and 1.pl. with voluntative and hortative function, e.g. *ed-ō ‘I want to eat’, *ed-o-mos ‘let’s eat’ > Lat. edō ‘I eat’, edimus ‘we eat’, with subsequent alignment to the pattern of thematic forms in *-oas in 3.pl. *ed-o-nt > Lat. edunt. The case of Lat. 1.pl. īmus ‘we go’ < *ei̯ -mos without thematic vowel might reflect an occasional use of the indicative in a pragmatically corresponding function, i.e. *‘we go! ’ = *‘we’ll go’ ~ *‘let’s go’. However, in view of the Latin subjunctive forms sīmus, velīmus, edīmus going back to optatives in *-ī-, Dunkel (1998: 97 f. with nn. 60, 61) assumes that “due to the formal and functional merger of the IE subjunctive and optative in all Italic dialects īmus can be derived from a correctly formed optative *h 1 i-ih 1 -mós”, with the same merger implying “that this form could once have been used voluntatively”. 9.4 As postulated above for consonant-final athematic stems, except for *es-/ *s- ‘to be’, there was a Proto-Italic stem levelling with introduction of the thematic endings of 1.sg. *-ō, 1.pl. *-o-mos, 3.pl. *-o-nti. Thus, for instance, in the paradigm of the nasal infix present of PIE *b h ei̯ d- ‘to split’ with the stem apophony *b h i-né-d- : *b h i-n-d- (> Ved. bhinát-ti : bhind-ánti), Proto-Italic generalized the weak stem *find- < *b h i-n-dand displayed a semi-thematic inflection: 1.sg. *find-ō, 2.sg. *find-si > *fint-si > *fins-si, 3.sg. *find-ti > *fint s̱ ti > *finssi → *fins-ti, 1.pl. *find-mos > *fin-mos (cf. *ei̯ -mos > Lat. īmus ‘we go’) → *find-o-mos, 2.pl. *find-tes > *fint s̱ tes > *finsses → *fins-tes, 3.pl. *find-o-nti. For the further development of these originally athematic paradigms, which had become semi-thematic in Proto-Italic, it The Latin future tense formation of the type faxō and its Italic background 225 <?page no="226"?> can be observed that in the post-Proto-Italic period, i.e. during the individual histories of Latino-Faliscan and Sabellic, most of these semi-thematic para‐ digms inherited from Proto-Italic underwent full thematization (except for single semi-thematic Proto-Italic paradigms preserved as such, like Lat. eō, it, eunt ‘go’, edō, ēst, edunt ‘eat’, volō, vult, volunt ‘want’; see above § 9.2.). Therefore, Latino-Faliscan shows thematic *find-ō, *find-e-s, *find-e-t, *find-o-mos, *find-e-tes, *find-o-nt > Lat. findō, findis, findit, findimus, finditis, findunt (cf. Meiser 1998: 185). 9.4.1 A differently motivated case of a thematic nasal infix present in Latin is provided by the paradigmatically isolated form of present 3.sg. ninguit ‘it’s snowing’ (without 3.pl. as a source of thematization) < Proto-Italic *sning u̯ -e-ti < *sni-n-g u̯ h -e-ti (*g u̯ h > */ χ u̯ / = *[-γ u̯ -] with occlusive realization as *[-g u̯ -] after nasal; on *ND h > Proto-Italic *ND, cf. Kümmel 2014), which represents a present tense form based on an analogically levelled present subjunctive PIE *sni-né-g u̯ h -e-t(i) ‘(snow) will make s. o. stick’ belonging to the factitive nasal infix present PIE *sni-né-g u̯ h -/ *sni-n-g u̯ h - ‘to make s. o. stick’ (on the semantically transitive, i.e. factitive, function of nasal presents, cf. Meiser 1993b), as associated with the intransitive root aorist PIE *snéi̯ g u̯ h -/ *snig u̯ h - ‘to stick’ as basis of the thematized aor. Ved. asnihat ‘(warrior) has remained prostrate’ = *‘has stuck (to the ground)’ and the derived present formation (Epic, Classical) Old Indic snihyati ‘sticks to’; comparable is the etymologically related present tense Gk. νείφει, OLat. nīvit (Pacuvius), OHG snīwit etc. ‘it’s snowing’ < *snéi̯ g u̯ h -e-ti as a presumably original aorist subjunctive meaning ‘(snow) will stick’ from the aforementioned root aorist, which is akin to the root noun Gk. acc. νίφ-α, Lat. nix, nivis etc. < *snig u̯ h - ‘snow’ = ‘what sticks’ (LIV 2 573 with nn. 1, 2, 3 and refs.). 9.5 At the Proto-Italic stage, the s-future, based on the athematic PIE desiderative with stem formant *-s-, still displayed a semi-thematic pattern of the structure 1.sg. *fak-s-ō, 2.sg. *fak-s-si, 3.sg. *fak-s-ti, 1.pl. *fak-s-o-mos, 2.pl. *fak-s-tes, 3.pl. *fak-s-o-nti. Therefore, like most athematic stems, which had become semi-thematic in Proto-Italic, this paradigm underwent full thematization in the post-Proto-Italic period during the individual history of Latino-Faliscan (in accordance with the reshaping of originally athe‐ matic structures underlying forms such as aforementioned semi-thematic Proto-Italic *find-ō, *fins-si, *fins-ti etc. → Latino-Faliscan *find-ō, *find-e-s, *find-e-t etc.; see above § 9.4.). Without the usual recourse to original sub‐ junctive forms in *-ō, *-e-si, *-e-ti etc. for the indicative forms of the faxō-type, 226 Reiner Lipp <?page no="227"?> this formally regular process of thematization explains the emergence of the fully thematic paradigm of *fak-s-ō, *fak-s-e-s, *fak-s-e-t, *fak-s-o-mos, *fak-s-e-tes, *fak-s-o-nt > OLat. (and Class.Lat. as archaisms) faxō, faxis, faxit, faximus* (in this lexeme, there is no attestation of 1.pl. indicative), faxitis, *faxunt → faxint* (like future perfect *-is-ont > *-erunt → *-erint; but, apparently in all attestations of this lexeme, the form faxint is subjunc‐ tive, i.e. *fak-s-ī-nd as continuation of the old optative in *-ī-). Due to the original athematic character of the paradigm of the s-future, the associated subjunctive forms in *-īcontinue athematic optative forms in *-ī- < PIE *-ih 1 -: 1.sg. *fak-s-ī-m, 2.sg. *fak-s-ī-s, 3.sg. *fak-s-ī-d, 1.pl. *fak-s-ī-mos, 2.pl. *fak-s-ī-tes, 3.pl. *fak-s-ī-nd > OLat. (and Class.Lat. as archaisms) faxim, faxīs, faxīt or faxit (on the scansion of the final syllable of the subjunctive 3.sg., cf. de Melo 2007b: 66 f. with n. 317), + faxīmus (Plautus Tru. 60; facimus cod.), faxītis, faxint (for the attested forms of this paradigm, cf. de Melo 2007b: 112 f.). The optative suffix in its zero-grade form is either taken over from root-accented acrostatic paradigms, cf. Lat. velim ‘I would like’ (beside indicative volō < *u̯ el-ō) < *u̯ el-ī-m < *u̯ élh 1 -ih 1 -m, Lat. edim ‘I would like to eat’ (beside indicative ed-ō) < *ed-ī-m < *h 1 éd-ih 1 -m, or it represents the generalized weak form of the mobile optative suffix *-i̯ éh 1 -/ *-ih 1 in amphikinetic paradigms, with the strong stem in *-i̯ ē- < *-i̯ éh 1 being only kept in the originally monosyllabic forms of the subjunctive of the copula as Palaeo-Lat. sied (Duenos inscr. CIL I 2 4; / s-i̯ ē-d/ < *h 1 s-i̯ éh 1 -d), and OLat. siem, siēs, siet ‘shall/ may be’, which, however, already in OLat. sim, sīs, sit show the effect of secondary levelling in favour of *-ī- < *-ih 1 after OLat. sīmus, sītis, sint ← sient ( ← *s-ii̯ -end < *h 1 s-ih 1 -énd) (on the mechanism of the Latin levelling of the original optative suffix to *-ī-, cf. Harðarson 1993: 85 f.). 10. Counter to the upcoming productive future tense formations of the type faciam ‘I’ll make, will be doing’, which is based on the present subjunc‐ tive denoting unmarked imperfective verbal events (actions and processes), and, in other paradigm classes, of the type of specific future forms in -bō as monēbō ‘I’ll remind, will be reminding’, which are based on a periphrastic construction with a basic durative value (see below §§ 17.1.-2.), the recessive future tense formation of the type faxō had a marked terminative value in the sense of ‘I’ll get it done, I’ll arrange it’, which in subordinate clauses sometimes seems to have led to its usage as a future perfect, e.g. peribo si non fecero, si faxo vapulabo ‘I’ll die, if I don’t do it. I’ll be beaten, if I do’ (Plautus Fr. 62W, apud Gellius 3,3,7). However, as is evident from The Latin future tense formation of the type faxō and its Italic background 227 <?page no="228"?> main clauses, the faxō-type represented an imperfective future tense with specific terminative value. This means, as a verbal category it was not perfective, but durative-telic (on the definition of the term ‘terminative’ as ‘durative-telic’, see Malzahn 2019: 227-229, 234). With other words, the future tense formation of the type faxō denoted a durative manner of action implying an extrapolated endpoint of the respective verbal event beyond the frame of the reference period (on the terminative value of the s-future, see below §§ 13.4.3.-4.). Therefore, the Latin future tense of the type faxō represents terminative expressions as ‘I am getting it done, I’ll get it done’, which functionally result from the inchoative-intentional value of the s-desiderative such as ‘I am going to do/ to make it’, implying a verbal configuration with starting-point and intended purpose to be fulfilled, comparable to the usage of the reduplicated desiderative in Vedic as expression of the intention of the subject (see above § 2.2.). According to Kieckers (1931: 272), Meillet/ Vendryes (1979: 296), the functional value of the original desiderative formation is still transparent in cases like siquidem te quicquam quod faxis pudet ‘if you are really feeling shame about whatever you are going to do’ (Plautus Mil. 624) and the gloss faxit ‘libenter facit’ (CGL IV 75, 2, V 456, 13 and 500, 34). For the terminative reading of the basic type ‘will get it done, will arrange/ achieve’ in the indicative and ‘may get it done, may arrange/ achieve’ in the subjunctive (not perfective as ‘will have done it’ and ‘may have done it’), cf. the usage of the indicative in faxo haud quicquam sit morae ‘I’ll arrange that there be no delay’ (Plautus Amph. 972), nam faxo - mea eris sponsa ‘for I’ll arrange - you will be my wife’ (Plautus Poen. 1228), and of the subjunctive in ita di faxint ‘so may the gods arrange it’ (Plautus Aul. 149) and quod ille faxit Iuppiter ‘may Jupiter up there so arrange it’ (Plautus Amph. 461) (following the argument by Weiss 2009: 419 f. with examples in nn. 15-17 against a categorial perfective value of the faxō/ faxim type and for its derivation from the PIE desiderative, as opposed to de Melo 2007a: 306-327, who considers it as analogically derived from certain forms of the perfective subjunctive and optative of the s-aorist, which had become paradigmatically independent). The frequently used form of the indicative 1.sg. faxō ‘I will make/ arrange/ achieve’, and possibly the 3.sg. faxsit ‘he will make’ (Pacuvius 424R), capsit ‘he will take’ (Ennius Ann. 319V), if not to be interpreted as s-future subjunctives in the meaning ‘he may make’, ‘he may take’ (cf. de Melo 2007a: 191 n. 2), are the only indicative forms of the s-future category that appear in main clauses, whereas all the other indicative forms 228 Reiner Lipp <?page no="229"?> of the s-future are attested exclusively in subordinate clauses (cf. Rix 1998a: 621 f.; de Melo 2007a: 180; Weiss 2009: 419 n. 14; for the attested forms of the s-future indicative in Plautus, Terence and others as well as for the clause types of their occurrence, cf. de Melo 2007b: 65-74). On the other hand, the forms of the s-subjunctive of the type faxim appear mostly in main clauses to express a wish or, in the negative, a prohibition (with attestations of the 1.sg. being rather voluntative than potential), but sometimes forms of the s-subjunctive occur in subordinate optative and final clauses as well as, serving the function of an eventualis, in conditional clauses, however without denoting anteriority or possibly associated syntactic dependency as the subjunctive perfect does (cf. Rix 1998a: 623; Weiss 2009: 420 n. 17; for the attested forms of the s-subjunctives in Plautus, Terence and others as well as for the specific clause types of their occurrence, cf. de Melo 2007b: 112-118). (On the non-anterior and non-perfective value of the indicative and subjunctive forms of the Old Latin s-future showing -r- < *-s-, see below §§-13.4.1., 13.4.3., 13.4.4.). 11.1 The s-future of the type of indicative faxō ‘I’ll make/ arrange/ achieve’ = / fak-s-ō/ with subjunctive faxim ‘(that) I may make/ arrange/ achieve’ = / fak-s-i-m/ (< *-ī-m based on the inherited optative) was mostly, but not exclusively formed in correspondence to the shape of the perfect passive participle (ppp) from a verbal base devoid of present stem markers; cf. faxō = / fak-s-ō/ with / fak-/ as in ppp factum, in contrast to present tense faciō, facit ‘make(s)’ < *fak-i/ i̯ o- ← *fak-i̯ e/ o-; but cave […] deixis = / dẹ̄ k-s-ī-s/ ‘you shall not say’ (in the reading of codex Ambrosianus of Plautus Mercator 484) is congruent with that of deico = / dẹ̄ k-ō/ (in Mercator 300, 465 etc.), i.e. with pres. dīcō, as against ppp dictum (Rix 1998a: 620, 630). This means, it was formed from the root or a verbal base derived from the present stem by subtracting the suffixes characterizing the present stem as such. Therefore, the respective verbal base did neither contain the thematic vowel *-e/ oor the suffixes *-i̯ e/ o-, *-ske/ onor a nasal infix (Rix 1998a: 620, 628, 633; Meiser 1998: 183, 2003: 40; see below §§ 13.5., 14.1.). The pattern for this rule of formation came about already in the pre-Proto-Italic period. It was conditioned by the paradigmatic setting of primary root-based s-future for‐ mations, as represented by *fak-s- (semi-thematic with allomorph *fak-s-oand full thematization in Lat. faxō, faxit; see above §§ 9.1., 9.5.), which is built on the root morpheme *fak- (< *d h h 1 kfrom PIE *d h éh 1 -k-/ *d h h 1 -k-, The Latin future tense formation of the type faxō and its Italic background 229 <?page no="230"?> a fossilized manner of action formation of *d h eh 1 - ‘to set, put’; LIV 2 139) and appears beside the present stem *fak-i̯ e/ o- ( → Proto-Italic *fak-i-/ -i̯ o- > Lat. faciō, facit, Osc. αfα̣κειτ = / am-fakẹt/ < *fak-i-ti; see below § 17.1.). This configuration triggered analogical new formations, as is still evident from cases of the s-future like attested OLat. subjunctive noxit ‘should someone harm; may it (not) hurt’ < *noke-s-ī-d (undergoing medial syllable syncope) from the stem *noke-sbased on *noke-, as abstracted from the present stem *noke-i̯ e/ o- (< PIE causative *noḱ-éi̯ e/ o-) = Lat. noceō, -ēre ‘harm, hurt’ (see below § 13.2.1.; on the analogical pattern going back to the pre-Proto-Italic stage until the period of the loss of intervocalic *-i̯ -, see Hill 2004: 139 f., based on Rix 1998a: 625-629; for secondary s-future formations built on a verbal base with truncation of the present stem marker -i̯ e/ o-, there are further examples revealing the underlying structure, on which see below §§ 13.1.4., 13.2.2., 13.5., 18.3.). In the case of inherited s-futures as OLat. subjunctive rupsit beside a nasal infix present like rumpō, -ere ‘break’, the root-based formation of the s-future was kept, thereby displaying a form without the specific present tense marker as represented by the fossilized nasal infix (see below §-13.3.1.). 11.2 After a consonant, the suffix of the Latin s-future appears in its basic form -slike in faxō = / fak-s-ō/ , after an original long vowel, however, as -sslike in cantassō* < *kantā-s-ō with the attested subjunctive excantassit < *kantā-s-ī-d (Rix 1998a: 625-627; Meiser 1998: 183, 2003: 40; this future form is based on *kantāas abstracted from the stem of the denominal present cantō, -āre ‘sing’ < *kantā-i̯ e/ o-). The representation of suffixal *-sin fut. cantassō* < *kantā-s-ō is in contrast to s-aorist formations of secondary verbs as 3.pl. *kantā-s-ond > perf. cantā-r-unt (beside common cantā-ver-unt < *kantā-u̯ is-ont from a periphrastic construction with fossilized perfect participle and copula, i.e. *kantā-u̯ os-is + *sont (? ); for the aorist formation cf. Meiser 1998: 207, 204 f., based on Rix 1992: 230-236). This treatment of *-safter a long vowel is specific to the category of the Old Latin s-future, which, for the purpose of formal distinction from preterites in *-s- > -rlike aforementioned cantārunt, came up as a morphonological regularization of the effect of the so-called littera-rule (cf. Rix 1998a: 625-627, 633; Meiser 1998: 77), describing an occasional quantity metathesis between long vowel and consonant of the type VCC < V̅ C in synchronically accented position, i.e. showing stress on the third last or second last syllable according to the penultima-rule, as in Lat. littera ‘letter’ < / lītera/ = inscr. OLat. literai CIL I 2 230 Reiner Lipp <?page no="231"?> 595, 6, leiteras CIL I 2 583 § 35 (inverse orthography like vīta = ueitam CIL I 2 364, 7 after cases with / ī/ < *ei̯ ) < *lītes-ā ← *līt-os/ -esntr. ← *lī-to- (< PIE *h 2 liH-tófrom *h 2 lei̯ H- ‘to smear’ in Lat. linō, -ere < *h 2 li-n-H-, cf. LIV 2 277 f.; for etymology and formation, cf. Walde/ Hofmann 1938: 814 f., discussion of different attempts by Blažek 2018), and with application to a long vowel resulting from monophthongization of an original diphthong Lat. Iuppiter < Iūpiter < *di̯ ou-pater (< PIE *di̯ éu̯ + *ph 2 ter) (on more specific context regulations of the littera-rule, cf. Weiss 2009: 144; 2010a: 149-151; 2010b; Sen 2015: 42-78, which however do not account for the s-futures of the type cantassō* < *ºā-s-ō beside faxō = / fak-s-ō/ ; on the post-Proto-Italic effect of the littera-rule as a recurrent irregular process not only in Latin with stress under penultima-rule, but also in the individual Sabellic languages without stress regulation after the penultima-rule, cf. Lipp 2021: 311 f. with refs.; for an approximate dating of the inner-Latin preliterary operation of this process in the category of the s-future to the time around 400 BC, see below §§-13.1.2.-3.). 11.3 For the pattern of formation of the Old Latin future tense category represented by faxō, faxim, an in-depth analysis was undertaken by Rix (1998a), followed by Meiser (1998: 182-184, 2003: 40). A morphological description of the formations of the Latin s-future and the single attestations of the forms was given by Neue/ Wagener (1897: 506-517), and an overview on the single formation types was offered by Sommer (1914: 584-587), Leumann (1977: 621-624). For an up-to-date analysis of use, function and forms of this class, see de Melo (2007a). 12.1 According to the approach of the present study, the Old Latin s-future of the type faxō, faxis, faxit shows a regular thematic remodelling of a semi-thematic s-future paradigm inherited from Proto-Italic, i.e. *fak-s-ō, *fak-s-si, *fak-s-ti etc. (with thematic vowel of o-colour in 1.sg., 1.pl., 3.pl. and athematic forms in the rest of the paradigm), which is based on an athematic PIE s-desiderative with inchoative-intentional value (see above §§ 9.1., 9.5.). The associated subjunctive of the type faxim, faxīs, faxit struc‐ turally represents the optative in *-īof the original athematic paradigm. Accordingly, the Sabellic s-future category represented by Osc. fust, Umbr. fust, fust = / fu-s-t/ ‘will be’ goes back to the same Proto-Italic semi-thematic paradigm type with athematic 3.sg. *fū-s-ti, but, due to the limited evidence from the preserved corpus of Sabellic, instances from distinctive thematic paradigm slots are not attested (see below §§-15.2., 15.3.1.). The Latin future tense formation of the type faxō and its Italic background 231 <?page no="232"?> 12.2 In the research to date, on the one hand, the Old Latin s-future of the type faxō, faxis, faxit has been considered to be based on the short vowel subjunctive in *-e/ oor a thematic enlargement of an athematic s-desider‐ ative formation (underlying the attested athematic forms of the Sabellic s-future), with the associated synchronic subjunctive of the type faxim, faxīs, faxit as a continuation of the corresponding optative in *-ī-; see Benveniste (1922), Thomas (1938: 27-66, regarding the future tense of the type faxō as a modified form of an inherited desiderative, with the morpheme *-sbeing genetically identical to that of the s-aorist), W.P. Schmid (1963: 41, 46), Rix (1976: 224 f., 1977: 148 with n. 45, 2003a: 153 f., 2003b: 4 f.), Meillet/ Vendryes (1979: 192 f., 281, 296 f.), Monteil (1986: 322, 329-331), Jasanoff (1988: 233 with n. 15, 1991: 85 with n. 3, 2003: 133 f.), Harðarson (1993: 86 n. 116, 114 n. 75, 1998: 338 with nn. 46, 47), Meiser (1993a: 177 f. with n. 33 for further refs.; 1998: 182 f., 2003: 33 f., 39 f.), Stempel (1998: 277 with n. 11), García Castillero (2000: 74-80), Weiss (2009: 418-421), and the comprehensive dossier for the derivation from a PIE athematic s-future formation by Hill (2004: 115-146, 152-154, offering a critical assessment of previous approaches with detailed references and o.c. 133 ff. with nn. 42, 43, 44 a comprehensive discussion of the formations underlying the distinct s-future types of Latin and Sabellic). Rix, Meiser and Hill consider the Latin s-future of the type faxō = / fak-s-ō/ , which is built on a verbal base without present stem suffix, i.e. fakas against the present stem fak-i̯ o-/ -iin Lat. faciō, facit, and the Sabellic s-future of the type Oscan didest = / dide-s-t/ ‘will give’, which is built on the present stem *did-e/ oas in Vest. didet ‘gives’, as two distinct formations already on the Proto-Italic level. Although likewise starting from an athematic PIE s-desiderative formation, in the present investigation, however, the latter view is contested, since the Sabellic formation of the s-future on the basis of the present tense stem is to be explained as a post-Proto-Italic remodelling (see below §§ 14.2., 14.5., 15.1.), starting from a Proto-Italic formation as structurally reflected by the Latin faxō-type, which is built on a verbal base devoid of present stem markers (see above §§-11.1., 14.1.). 12.3 On the other hand, the view has been advocated that the Old Latin future tense formation represented by indicative faxō goes back to the prospective subjunctive of the s-aorist and that of the associated subjunctive faxim to the original optative of the s-aorist, often under the assumption of an early functional split of this future tense type from the paradigm of the earlier s-aorist or a preterite originating from it; see Sommer (1914: 232 Reiner Lipp <?page no="233"?> 584-587), Kieckers (1931: 272-274, 226 f.), Narten (1973: 142-144 with nn. 25, 26), Pisani (1974: 287), Leumann (1977: 573, 576, 621), McCone (1991: 70 f.), Prosdocimi/ Marinetti (1993: 315-23), Sihler (1995: 558 f.), with Szemerényi (1989: 307-312, 1996: 285-288), Euler (1992: 31-34 with n. 32) also pointing to the existence of the aorist stem *fak-sin the preterite form vha.g.s.to ‘he made, offered’ of closely related Venetic; this view is basically shared by de Melo (2007a: 306-327, esp. 313 f.; 2007c, although o.c. 12-18 taking Venetic vha.g.s.to quite rightly as merely typological evidence for an s-preterite stem *fak-s-, not as a shared Italo-Venetic innovation underlying Lat. faxō; likewise Meiser 2003: 35, 89 n.40), Bertocci (2016), Willi (2018: 453), Søborg (2020a: 87-89), Clackson (2020: 67 f.). 12.4.1 In his interpretation of the Latin s-future formation, Pedersen (1921: 12 ff., 22 f.) starts from s-aorist forms, with faxō being the original aorist subjunctive and faxim the original aorist optative, whereas the Sa‐ bellic future formation of the type pertemest ‘will prevent’ is derived from *em-es-ti as a clearly athematic form containing the future tense suffix *-es- (reconstruction shared by G. Schmidt 1986: 35, 38-47, 58 with n. 97). In the same vein, de Vaan (2011: 34, following Schrijver 2006: 61, 62) derives most of the Latin forms of the so-called perfective future in -sfrom the stem of the s-aorist, with the indicative type of faxō ‘will make’ continuing the Proto-Italic aorist subjunctive with primary endings and the subjunctive type of faxim ‘may make’ continuing the Proto-Italic aorist optative. The underlying aorist stem *fak-sis considered to be confirmed for Proto-Italic by Venetic vha.g.s.to ‘made, offered’. For the prehistorical background of the sigmatic future formations in Italic, de Vaan (2011: 34) posits a system with “s-present *fak-es-ti, *fak-es-t, aor. *fak-s-t, aor.sb. *fak-s-e-t, aor.opt. *fak-s-ī-t.” But as just mentioned, de Vaan himself advocates a derivation of the Latin s-future faxit from an s-aorist subjunctive with primary ending, thus *fak-s-e-ti, not *fak-s-e-t. 12.4.2 Regarding this analysis, there is however the problem that, except for the 1.sg. in *-ō, in old subjunctive forms that are structurally inherited from the Proto-Italic stage including those in prospective function to express the future tense, the Italic languages do not reflect the so-called primary endings (with hic-et-nunc marker *-i), but only the so-called secondary or “primitive” endings (without hic-et-nunc marker) like in Palaeo-Latin esed ‘shall/ will be’ (Forum inscr. CIL I 2 1) < *es-e-d (< PIE *h 1 és-e-t = *[h 1 ésed̥ ] = The Latin future tense formation of the type faxō and its Italic background 233 <?page no="234"?> */ h 1 és-e-d/ ; on the realization of final stops in PIE as unreleased lenes, cf. Lipp 2016 with refs.) as precursor of the later future tense form erit with secon‐ darily introduced primary ending -t < *-t-i after the synchronically basic present indicative. Correspondingly, Sabellic retained the original endings as e.g. in the Oscan forms of the present subjunctive deiuaid ‘shall swear’ < *dei̯ u̯ ā-i̯ ē-d, preterite subjunctive hipid < *χēp-ē-d ‘shall keep’, Oscan fuid ‘shall be’ < *fuu̯ -ē-d and imperfect subjunctive fusíd ‘would be’ < *fu-s-ē-d, in contrast to corresponding Lat. foret with only secondary introduction of the so-called primary ending -t < *-t-i from the present indicative. 12.4.3 According to Kortlandt’s model adopted by de Vaan (2011: 32, 34), Sabellic future formations like Osc. didest are regarded to represent an inherited structure / did-es-t/ based on an athematic s-present with apo‐ phonic suffix *-és-/ *-s- (pursuing the aforementioned proposal by Pedersen 1921: 12 ff.), which was reanalysed as / did-e-s-t/ after the thematic present / did-e-t/ , as represented by Vestinian didet (Kortlandt 2007: 152). Therefore, for the athematic form of the Sabellic s-future fust = / fu-s-t/ , which tradi‐ tionally is regarded as the continuation of a Proto-Italic s-future *fū-s-ti, Kortlandt (ibid.) reconstructs, instead, the alternating stem *fu-es-/ *fu-s-, for which one would expect, however, a representation by Sabellic †fuu̯ -es-t, since in Sabellic there was no vowel weakening in closed syllable. 12.4.4 Furthermore, the assumption of an apophonic verbal suffix of the form *-és-/ *-s-, as postulated by Pedersen not only for the Sabellic future formations, but also for Hittite verbs with s-extensions (Pedersen 1938: 94-97), is not backed by any comparative evidence (for the rejection of the reconstruction of an apophonic future suffix *-és-/ *-s-, cf. Hill 2004: 153 f. with n. 52). A proto-form of this type, reconstructed in the shape *dm̥ h 2 -és-/ *dm̥ h 2 -sby van den Hout (1988, following Pedersen o.c.) as underlying Hitt. 3.sg. damašzi/ 3.pl. dameššanzi ‘(op)press’, could only be proven by a reflex of assumed *dm̥ h 2 -ésas Hitt. *damḫas-. But, according to Søborg (2020b: 636 ff.), verbal forms with s-extensions showing an apophonic degree *-escannot be verified in Hittite (for the derivation of the Hittite form from *démh 2 -s-/ *dm̥ h 2 -sor *dméh 2 -s-/ *dm̥ h 2 -swith Schwebeablaut, see above §-4.). 13. According to Rix (1998a), in the case of the Latin s-future, the following morphological patterns can be found within the Old Latin formations, which are still occasionally attested in quotations and archaizing passages of the later period, with single forms like indicative faxō in the meaning ‘I’ll 234 Reiner Lipp <?page no="235"?> arrange it, I’ll see to it’ (e.g. Vergilius Aen. 9, 154. 12, 316) and subjunctive ausim ‘I may dare’ (e.g. Livius Praef. 1, Tacitus Ann. 1, 81) still appearing as isolated lexemes of the elevated style in Classical Latin. 13.1.1 In the I. conjugation consisting of stems in -ā-, on the one hand, there are the denominatives of the type Lat. cantō, -āre ‘sing’ (with perf. cantā-vī, ppp cantā-tum), which as a derivative of *kan-to- (as the orig‐ inal ppp of canō, -ere ‘sing’ < *kan-e/ o-) shows the present tense stem / kant-ā-/ < *kant-ā-i̯ e/ ocontaining the denominative suffix -ā- < *-ā-e/ o- < *-ā-i̯ e/ o- (< PIE */ -ah 2 -i̯ e/ o-/ ), and the associated s-future of the type cantassō*, cantassim* (excantassit, occentassit) < *kantā-s-ō, *kantā-s-ī-m (quantity metathesis VCC < V̅ C according to the littera-rule; see above § 11.2.), with the derivational stem base *kant-ābeing devoid of the present tense marker *-i̯ e/ o- (according to the formation rule of this category that the derivational basis contains no specific marker of the present tense; see above § 11.1.). On the other hand, there are dever‐ batives as Lat. amō, -āre ‘like, love’ (with perf. amā-vī, ppp amā-tum on the analogy of the aforementioned denominative type, instead of †amuī < *ama-u̯ ai̯ , †amitum < *ama-to-), showing the present tense stem / amā-/ < *ama-e/ o- < *ama-i̯ e/ o- (vowel-final athematic present stems being regularly extended by *-i̯ e/ o-) ← *ama-/ *ama- ‘to grab, hold on’ (< PIE *h 2 émh 3 -/ *h 2 m̥ h 3 - = Ved. ámī-ti ‘holds fast, swears’; Rix 1999: 523 f.; LIV 2 265 f.; thus, the original paradigm pattern of Lat. amō, -āre < *ama-i̯ e/ omust structurally have corresponded to the type of Lat. domō, -āre ‘tame’ < *doma-i̯ e/ o- ← *domai̯ e/ o- < PIE *domh 2 -éi̯ e/ owith Lat. perf. domuī < *doma-u̯ ai̯ , ppp domitum < *doma-to-; LIV 2 116 f. with n. 5). The associated s-future amassō, amassim* (amassīs, amassit) is remodelled (like perfect and ppp) on the analogy of the more frequent denominative type with present tense stem / kantā-/ < *kantā-i̯ e/ o- : s-future *kantā-s-ō > cantassō*. The levelled future formation of amassō, amassim* appears instead of the unattested pattern of †amerō, †amerim < *amezō, *amezīm < *ama-s-ō = *[amazō], *ama-s-ī-m = *[amazīm], based on the original derivational basis *ama-, as descriptively abstracted from the present tense stem *ama-i̯ e/ o- > / amā-/ . But the unattested structure of an s-future †amerō < *ama-s-ō built on the bare root stem *amacorresponds to the original formation of the s-future of deverbatives as represented by attested ad-iuverō ‘will render help’ < *ad-ii̯ ou̯ a-s-ō, with *ii̯ ou̯ aas unexended form of the secondarily The Latin future tense formation of the type faxō and its Italic background 235 <?page no="236"?> extended present tense stem *ii̯ ou̯ a-i̯ e/ o- > / i̯ ou̯ ā-/ in ad-iuvō, -iuvā-re (see below §-13.1.4.). 13.1.2 Nevertheless, a stem shape as assumed for the unlevelled future formation *ama-s-ō (from the original basis *ama-) > *amezō > †amerō might be presupposed for the formation of the conditional derived from the stem of the Latin s-future in its original form *ama-s-ō = *[amazō] ‘I’ll love’. This is the formation underlying the so-called imperfect subjunc‐ tive (see below §§ 15.3.1.-2.): *ama-s-ē-m, *ama-s-ē-s, *ama-s-ē-d ‘would love’ = *[amazēm], *[amazēs], *[amazēd] > *amezēm, *amezēs, *amezēd > *amerem, *amerēs, *amerēd → amārem, amārēs, amāret, with late remodel‐ ling after the associated present tense stem / amā-/ (< *ama-i̯ e/ o-) in amō, amā-re. An analogical process of this kind surfaces in the relation of the aforementioned s-future ad-iuverō < *ad-ii̯ ou̯ a-s-ō (built on the verbal base *ii̯ ou̯ aas unextended form of the secondarily extended present tense stem *ii̯ ou̯ a-i̯ e/ o- > / i̯ ou̯ ā-/ in ad-iuvō, -iuvā-re) and the conditional derived from it, i.e. *ad-iuverem < *ad-ii̯ ou̯ a-s-ē-m, which, counter to the s-future ad-iuverō, shows reshaping into the attested imperfect subjunctive form ad-iuvārem = / ad-i̯ uu̯ ā-r-e-m/ after the synchronic present stem / i̯ ou̯ ā-/ ~ / ad-i̯ uu̯ ā-/ in ad-iuvō, -iuvā-re. It is however worth considering that, in the case of existing analogical s-future forms like amassō : pres. amō, -āre after s-future cantassō* : pres. cantō, -āre (see above § 13.1.1.), the levelling of the derived conditional might have taken place somewhat earlier, that is at the onset of rhotacism, i.e. *ama-s-ē-m = *[amazēm] > *amezēm (> *amerem) → *amā-s-ē-m = *[amāzēm]. According to the littera-rule, this form could have produced a doublet form *amassēm, which, on the other hand, would probably have been avoided as being identical with the fore-runner of the so-called pluperfect subjunctive *ama-s-s-ē-m → *amāssēm > amāssem ‘would have loved’, which as an original conditional represented the sub‐ junctive of an s-future based on the secondary sigmatic preterite stem *ama-s- (abstracted from present stem *ama-i̯ e/ o- > *amā-) → *amā-sas preserved in perfect 2.sg. amāstī, 2.pl. amāstis, 3.pl. amārunt beside the normal forms of the perfect paradigm of amāvī, i.e. amāvistī, amāvistis, amāvērunt with pluperfect subjunctive amāvissem (on the traces of sigmatic preterite formations of secondary verbs in the Latin perfect paradigm, cf. Meiser 1998: 207). Therefore, in the case of the imperfect subjunctive, the recent analogical basic form *amā-s-ē-m = *[amāzēm] would undergo the effect of incipient rhotacism, thereby leading to a form of the shape of later 236 Reiner Lipp <?page no="237"?> attested amārem, which for its congruence with the stem of the functionally associated present tense amō, amā-re became the prevailing form (as modal expression of the irrealis belonging to the present tense; on the related formal alignment of the imperfect subjunctive to the present tense stem, see below § 15.3.2.). Thereby, a congruent formation of the conditional derived from the future tense type of cantassō* (excantassit, occentassit) < *kantā-s-ō = *[kantāzō] (with original long-vowel basis *kant-āas derived from the denominative present stem *kant-ā-i̯ e/ o-) would have been stabilized, i.e. the imperfect subjunctive cantārem < *kantā-s-ē-m = *[kantāzēm] would preserve the basic long vowel -ādue to the functional alignment to the present tense stem / kantā-/ (< *kant-ā-i̯ e/ o-) in cantō, cantā-re, presumably prevailing in competition over a doublet form *kantassēm < *kantā-s-ē-m = *[kantāzēm] with effect of the littera-rule, which however might rather have been avoided due to the existence of a nearly identical pluperfect subjunctive form *kantā-s-s-ē-m > cantāssem, representing an original s-fu‐ ture subjunctive = conditional based on the secondary s-preterite *kantā-sas attested in perf. 3.pl. cantārunt (on this form, see above § 11.2.); for the possible occurrence of such doublets, cf. later attested forms as Iūpiter ~ Iuppiter and pāricīda ~ parricīda. Such a late reshuffling of imperfect subjunctive forms after the corresponding present tense stem would also be compatible with the process to be supposed for the conditional based on the s-future faxō = / fak-s-ō/ ‘I’ll make’, i.e. *fak-s-ē-m ‘I would make’ → *fak-i-s-ē-m = *[fakizēm] by remodelling after the present stem *fak-iin Lat. faciō, facit, followed by the further phonological change to the attested shape of imperfect subjunctive facerem, with secondary intervocalic *-s- = *[-z-] > -rby still operative rhotacism during the period of continuing or potential effectiveness. 13.1.3 Taking into account cases of secondary analogical reshaping of future tense forms as amassō (via intermediate *amā-s-ō = *[amāzō] ← *amezō < *ama-s-ō after denominative cantassō* < *kantā-s-ō), one can state that, as a relatively late and recurrent phonetic process in the history of the single branches of Italic, in Latin the littera-rule worked only in a stressed sequence of the structure V̅ C > VCC (see above § 11.2.), that is, within the s-future formations, only in forms such as *kantā-s-ō > cantassō* with *kant-āas a derivational basis in a long vowel. In these formations, therefore, the littera-rule probably functioned as a morphonological rule to prevent the effect of the incipient rhotacism (*s = *[z] > r in the first half The Latin future tense formation of the type faxō and its Italic background 237 <?page no="238"?> of the 4 th c. BC; for the date of rhotacism cf. Meiser 1998: 95 f.; Weiss 2009: 151 f.) and the concomitant loss of morphological transparency, i.e. *-āsō = *[-āzō] → -assō (instead of †-ārō), showing {-s-} as preserved future marker after original stem final long vowel in analogy to faxō = / fak-s-ō/ . Counter to the early dating by Rix (1998: 626 f.), who reckons for these formations with an already Proto-Italic application of the littera-rule as a posttonic quantity metathesis, as if corresponding to the stress regulation of Classical Latin with accented length under the penultima-rule, despite generally assumed initial stress in Common Italic (cf. the critical remarks by Hill 2004: 136-139 with n. 45 and de Vaan 2011: 33 f.), the inferable late date of the operation of the littera-rule in the period around 400 BC explains the otherwise irregular lack of the effect of vowel weakening in closed syllable (i.e. amassō, not †amessō), which had already taken place between the end of the 6 th c. BC and the middle of the 5 th c. BC. Thus, in contrast to secondary analogical amassō < *amā-s-ō = *[amāzō] ← *amezō < *ama-s-ō, one can observe the effect of the phonologically regular vowel weakening in the s-future form ad-iuverō ‘I’ll render help’ < *ad-ii̯ ou̯ ezō < *ad-ii̯ ou̯ a-s-ō, representing an inherited structure with preserved short-vowel basis *ii̯ ou̯ a- (see below § 13.1.4.). The morphonological process with the littera-rule preserving the future marker {-s-} after original long vowel as in cantassō* < *kantā-s-ō (in contrast to *-s- > -rafter preserved short vowel as in ad-iuverō < *ad-ii̯ ou̯ a-s-ō) is functionally comparable to the restitution of the future tense morpheme *-safter a stem-final long vowel in Greek, cf. future λῡ́ σω ← *lūhō < *lū-s-ō (pres. λύω ‘loosen’ < *luu̯ ō < PIE *luH-ō from *leu̯ H- ‘to loosen’) in analogy to cases with preserved -safter a stop like δείξω = / dei̯ k-s-ō/ , whereas after a short vowel the *-sis mostly not restored, cf. future πετῶ < πετάω < *petahō < *peta-s-ō (pres. πετάννῡμι, Hom. πίτνημι < PIE *p e t-né-h 2 from *peth 2 - ‘to spread out’) (cf. Rix 1976: 223 f.). 13.1.4 A special case of s-future formation with intervocalic *-s- > -r-, i.e. with effect of rhotacism, is provided by a minoritarian type of deverbatives that has not undergone remodelling on the analogy of the denominative type. Such a formation is represented by (ad-)iuvō, -āre ‘support, help’, as well as in Old Latin inscriptions by imperative II adiouanto (CIL I 2 2676, post n. 586 inserendum), subjunctive iouent (CIL I 2 364, 11 from the Faliscan area; Bakkum 2009: 498-500 nr. 217) < *ii̯ ou̯ a-i̯ ē-nd (with perf. (ad-)iūvī < *ii̯ ou̯ a-u̯ ai̯ , ppp (ad-)iūtum < *ii̯ ū-to-), showing the present tense stem / i̯ ou̯ ā-/ ~ / ad-i̯ uu̯ ā-/ < *(ad-)ii̯ ou̯ a-i̯ e/ o- (with stem-final long vowel °ā- < 238 Reiner Lipp <?page no="239"?> *°a-i̯ e/ odue to regular postvocalic present stem extension by *-i̯ e/ o-) beside OLat. iuve ‘help! ’ (Accius tr. 489R) from the thematic stem */ i̯ uu̯ -e/ o-/ < *ii̯ uu̯ -e/ o- (based on 3.pl. *ii̯ uu̯ -onti), as continuation of an apophonic present tense configuration *ii̯ ou̯ a-/ *ii̯ ū-, 3.pl. *ii̯ uu̯ -onti ← *ii̯ uu̯ -enti (< PIE *h 1 i-h 1 éu̯ H-/ *h 1 i-h 1 uH- ‘to help, foster’ ≈ Hitt. ii̯ a(u)u̯ atta med. ‘recovers (itr.)’; for the paradigmatic analysis Rix 1999: 520; LIV 2 243 f. with n. 4, for the forms cf. Walde/ Hofmann 1938: 736 f.; Ernout/ Meillet 1967: 331; Leumann 1977: 596; de Vaan 2008: 318, and the critical stance of Hill 2004: 135 f.), displaying the associated s-future paradigm adiuverō (Ennius Ann. 335V = Warmington 1967a: 120 f. nr. 327; Skutsch 1985: 99 nr. 337, with commentary 512; see below § 13.4.4.), (ad)iuverim* (indicative adiuverit Terence Ph. 537, subjunctive adiuverit Plautus Ru. 305, iuverint Catullus, Propertius) < *(ad-)ii̯ ou̯ a-s-ō, *(ad-)ii̯ ou̯ a-s-ī-m (for the analysis as a formation belonging to the category of the s-future, see Rix 1998: 624 f., 630 f., 632). This is an s-future formation with intervocalic *-s- > -r-, which is built on a verbal base consisting of the unextended present stem *ii̯ ou̯ a-, being devoid of its later regular postvocalic present stem extension *-i̯ e/ oas in *ii̯ ou̯ a-i̯ e/ o- > *ii̯ ou̯ ā- > (ad-)iuvā-re. Thus, the form of the s-future adiuverō (showing short -uas against perf. (ad-)iūvī < *ii̯ ou̯ a-u̯ ai̯ and future perf. iūverō) displays intervocalic *-s- = [-z-] > -rafter short vowel, in contradistinction to s-future forms showing a quantity metathesis according to the littera-rule, which could only work with stem bases in a long vowel as in the case of *kantā-s-ō > cantassō* etc. 13.2.1 In the II. conjugation consisting of stems in -ē-, on the one hand, there are the essive formations expressing a state like Lat. habeō, -ēre ‘hold, have’, with the associated perf. hab-uī, ppp habi-tum < *hab-e-u̯ ai, *hab-e-tobeing built in analogy to the causative-iterative type in *-e-i̯ e/ o- ← *-éi̯ e/ omentioned below, instead of former *hab-ē-u̯ ai̯ , *hab-ē-to-, in structural correspondence to the noun acētum ‘vinegar’ < *ak-ē-toas a lexicalized form of the verbal adjective of aceō, -ēre ‘be sour’, which delivered the verbal base *ak-ēas abstracted from the present stem in *-ē-i̯ e/ o- (on the morphological relic acētum, see Harðarson 1995: 223 f. n. 51; Rix 1998a: 629; Meiser 1998: 190). The respective present stem / hab-ē-/ contains the essive suffix -ē- < *-ē-e/ o- < *-ē-i̯ e/ o- (< *-eh 1 -i̯ e/ o- ← PIE *-h 1 i̯ e/ oin a present stem formation derived from the stem of the fientive in *-éh 1 -/ *-h 1 -, which denotes an accomplished and therefore telic process as represented by the intransitive Greek aorist in -η- = / -ē-/ ; reconstruction of form and function The Latin future tense formation of the type faxō and its Italic background 239 <?page no="240"?> of the essive category, its Italic remodelling and PIE derivation from the fientive formation according to Harðarson 1998, followed by Lipp 2009a: 41 with n. 100 for further refs.; cf. LIV 2 25). For the formation of the para‐ digmatically associated verbal stems outside the forms of the present stem, in pre-Proto-Italic a verbal base *χab-ēwas abstracted from *χab-ē-i̯ e/ oby dropping the specific present marker *-i̯ e/ o- (see above § 11.1.). This verbal base *hab-ē- < *χab-ēis therefore underlying the formation of the associated s-future OLat. habessō*, habessim* (e.g. habessit, prohibessīs, prohibessit, prohibessint) < *hab-ē-s-ō, *hab-ē-s-ī-m (quantity metathesis VCC < V̅ C according to the littera-rule; see above § 11.2.). On the other hand, there are original causative and iterative formations like Lat. noceō, -ēre ‘harm, hurt’, showing the present stem / nok-ē-/ < *nok-e-e/ o- < *nok-e-i̯ e/ o- ← *nokei̯ e/ o- (< PIE *noḱ-éi̯ e/ o- = Ved. nāś-áya-ti ‘makes disappear, destroys’ as causative of PIE *neḱ- ‘to disappear’; LIV 2 451 f.), with a verbal base *nok-eas appearing in the secondary inner-Latin formations of perf. nocuī, ppp nocitum < *noke-u̯ ai̯ , *noke-to-. For the formation of the tense stems outside the present stem category, the aforementioned verbal base *nok-ehad been abstracted from the reanalysed present stem formation *nok-e-i̯ e/ oby drop‐ ping the specific present marker *-i̯ e/ o-. Correspondingly, the verbal base *nok-eis presupposed by the stem of the associated Italic s-future *nok-e-s-, underlying the Old Latin s-future subjunctive noxit ‘should someone harm; may it (not) hurt’ (XII Tabulae 12, 2; Lucilius 1195M) < *nok-e-s-ī-d, which was liable to medial syllable syncope with loss of the short base vowel *-e-. Moreover, there is a structurally rare type of denominative formation within the II. conjugation (beside the more frequent denominative essive type of albeō, -ēre ‘to be white’ with stem / alb-ē-/ < *alβ-ē-e/ o- < *-ē-i̯ e/ o- < *-e-h 1 -i̯ e/ ofrom *h 2 alb h -o- > *alβo- > Lat. albus ‘white’) that occurs in the case of Latin audeō, -ēre ‘like to do, dare’, which represents a derivative of the Italic adjective continued by Latin avidus ‘eager, greedy, avaricious’ < *au̯ iδo- (< *h 2 au̯ -id h ofrom the root PIE */ h 2 au̯ -/ = *h 2 eu̯ - ‘to enjoy, desire’ in the Latin essive verb aveō, -ēre ‘to be eager, desire’ < *h 2 au̯ -eh 1 -i̯ e/ o-; cf. LIV 2 -274; de Vaan 2008: 60 f., 65). This means, it is a derivative consisting of the denominal verb stem *au̯ iδe-i̯ e/ o- ‘to act as an eager one, to act eagerly’ > Proto-Italic *au̯ iδe-e/ o- > Latino-Faliscan *au̯ iδ-ē- > Latin / au̯ d-ē-/ ‘like to do, dare’ (with medial syllable syncope; analysis of the formation according to Rix 1998a: 627 f.). By subtracting the present stem suffix *-i̯ e/ ofor the tense stems outside the present stem category, this stem provided the verbal 240 Reiner Lipp <?page no="241"?> base *au̯ iδethat appears, with effect of medial syllable syncope, in the secondary inner-Latin perfect formations OLat. ausī < *au̯ ssẹ̄ < *au̯ d(e)-s-ai̯ < *au̯ iδe-s-ai̯ and Lat. ausus sum ‘I have dared’ from the verbal adjective *au̯ sso- < *au̯ d(e)-to- < *au̯ iδe-to-, but served already in Proto-Italic for the formation of the associated s-future *au̯ iδe-sthat is underlying the Old Latin s-future subjunctive ausim ‘I may dare’, which as a lexeme still used in elevated style of Classical Latin goes back to *au̯ ssim < *au̯ d-s-ī-m < *au̯ de-s-ī-m (medial syllable syncope with eventual loss of the short base vowel *-e-) < *au̯ iδe-s-ī-m. On the historically different stem formations of the II. conjugation as partially reflected by the respective forms of the Old Latin s-future, see Rix (1998a: 627-629, 632 f.). 13.2.2 Moreover, in the II. conjugation, there is a special case of s-future formation with intervocalic *-s- > -r-, which is provided by the original causative formation Lat. moneō, -ēre ‘remind, admonish’, showing the stem / mon-ē-/ < *mon-e-e/ o- < *mon-e-i̯ e/ o- ← *monei̯ e/ o- (< PIE *mon-éi̯ e/ o- ‘to make think (of)’ = OAv. mānaiieiti ‘admonishes’, Ved. mānáyati ‘honors, esteems’ from *men- ‘to get an idea, think’; LIV 2 435 f.), with the s-future indicative moneris (Pacuvius 30R), subjunctive monerint (Pacuvius 112R) < *mon-e-s-es, *mon-e-s-ī-nd (with avoidance of vowel syncope in *-nVs- = *[-nVz-] > †-nz-; Rix 1998a: 628), representing the Proto-Italic stem *mon-e-sbased on *mon-efrom the reanalysed present stem formation *mon-e-i̯ e/ o-; the verbal base *mon-eis likewise underlying the secondary inner-Latin formations of perf. monuī, ppp monitum < *mone-u̯ ai̯ , *mone-to- (on the formal and functional analysis of moneris, monerint as forms of the s-future and not of the s-aorist, see Rix 1998a: 628, 632; for the details, see below §§-13.4.3.-4.). 13.3.1 The III. conjugation comprises formations with the original the‐ matic present suffixes *-e/ o-, *-i̯ e/ oin postconsonantal position, and *-ske/ o-. It is represented by the present Lat. faciō, -ere ‘make’ < *fak-i/ i̯ o- (perf. fēcī, ppp factum) and the well-attested associated s-future faxō, faxim < *fak-s-ō, *fak-s-ī-m (indicative faxō, faxis, faxit, faxitis, passive faxitur, subjunctive faxim, faxīs, faxit, faxīmus* [Plautus Tru. 60; facimus cod.], faxītis, faxint; 3.pl. faxint appears nearly always as subjunctive form in the combination dī faxint and is apparently not attested as indicative); other examples are provided by forms like the present capiō, -ere ‘take, catch’ (perf. cēpī, ppp captum), showing the stem */ kap-i/ i̯ o-/ , and the associated s-future capsō, capsim* (capsīs, capsit, capsimus) < *kap-s-ō, *kap-s-ī-m; present agō, -ere The Latin future tense formation of the type faxō and its Italic background 241 <?page no="242"?> (perf. ēgī, ppp āctum), showing the stem */ ag-e/ o-/ , and the associated s-fu‐ ture subjunctive āxim (adāxint) < *ag-s-ī-m (with vowel lengthening before preconsonantal media according to Lachmann’s Law); present caedō, -ere ‘strike’ (perf. cecīdī, ppp caesum), showing the stem */ kai̯ d-e/ o-/ , and the associated s-future subjunctive occisit ‘may kill, should someone kill’ (Leges regiae - II Numa Pompilius 3; XII Tabulae 8, 12; for the transmitted form, cf. Bruns 1893: 8, 31; Warmington 1967b: 482 f.; Crawford 1996: 562, 571, 578, 610, 611, 612), which represents the historical shape / ok-kẹ̄ s-s-ī-t/ of the form of redaction + occessit (as presumably reflected by the gloss CGL V 630, 21 occelsit ‘occiderit’) < *(ob-)kai̯ d-s-ī-d (see below § 14.6.); present rumpō, -ere ‘break (tr.)’ (perf. rūpī, ppp ruptum), showing the nasal infix stem */ rump-e/ o-/ ← 3.pl. *rump-onti ← *ru-m-p-enti (< PIE *ru-né-p-/ *ru-n-p- = Ved. lumpáti ‘breaks (tr.)’ from the root *reu̯ pwith root aorist continued by Lat. perf. rūp-ī; LIV 2 510 f.), and the associated s-future subjunctive rupsit ‘should someone break’ (XII Tabulae 8, 2) < *rup-s-ī-d (from PIE *réu̯ p-s-/ *rup-s-), which, as an inherited formation based on the root, regularly lacks the nasal infix of the respective present stem. 13.3.2 Like the aforementioned s-future ad-iuverō from ad-iuvō, -āre ‘help’ of the I. conjugation and the aforementioned s-future moneris from moneō, -ēre ‘remind, admonish’ of the II. conjugation, also in the III. con‐ jugation, there is a special case of s-future formation with intervocalic *-s- > -r-. It is connected with the original nasal infix present Lat. sinō, -ere ‘let (happen), allow’ (perf. sīvī, ppp situm), showing the thematic stem */ sin-e/ o-/ ← Proto-Italic 3.pl. *sin-onti ← *sin-enti, resulting from a semi-the‐ matic Proto-Italic remodelling of an athematic paradigm (see above §§ 9.3.- 4.), based on the stem *si-na- ← *si-nē-/ *si-na- (< PIE *si-né-h 1 -/ *si-n-h 1 from *sh 1 ei̯ -/ *sih 1 - < *sh 1 i- ‘to let go, loosen, relax’; cf. LIV 2 518 s.v. *seh 1 (i̯ )with n. 5; probably not with Rix 1993: 86 f., 1998a: 631 as *‘let settle’ from *tḱei̯ - ‘to settle’, as also advocated by LIV 2 643 f. and nn. 5, 6, but referring to the critical remarks by Lipp 2009b: 314, 325 f.; on underlying *sh 1 eisee below). The associated s-future subjunctive sīrīs (Plautus, Pacuvius, Cato) (sīrit, sīrītis, sīrint Plautus+) < *sei̯ rīs < *sei̯ -s-ī-s with the stem *sei̯ -sdid not undergo the effect of the littera-rule (VCC < V̅ C; see above § 11.2.), since, at the time of the rhotacism of intervocalic *-s- = *[-z-] > *-r-, the preceding diphthong *ei̯ had not yet contracted to the long vowel / ī/ as represented by the attested form sīrīs. The stem of the s-future *sei̯ -sregularly lacks the nasal infix of the respective present stem and continues the PIE desiderative 242 Reiner Lipp <?page no="243"?> formation *sh 1 éi̯ -s-/ *sih 1 -s- < *sh 1 i-s-. Together with the nasal infix present Lat. sinō, -ere ← *si-né-h 1 -/ *si-n-h 1 -, the s-future form sīrīs is based on the root *sh 1 ei̯ -/ *sih 1 - < *sh 1 i- ‘to let go, loosen, relax’ (laryngeal metathesis CUHC < CHUC), as represented by Ved. ava-sāyáiati ‘he lets unharness (the horses)’ < *sh 1 oi̯ -éi̯ e/ oand perhaps Faliscan salues seite ‘stay safely at rest! ’ (EF 4) < *sh 1 éi̯ -te (philological interpretation of the Faliscan forms according to Rix 1993: 86 f.; 1998a: 631, however with etymological derivation from the root *tḱei̯ - ‘to settle’; on the other hand, seite is regarded as an error for s{e}ite or s<ie>te reflecting the old PIE optative *h 1 s-ih 1 -té from *h 1 es- ‘to be’ and equivalent to Lat. subjunctive sītis ‘you may be’ by Bakkum 2009: 60, 157, 158, 171, 172, 197, 299, 413). The root goes back to the stem *sh 1 -éi̯ -/ *sh 1 -i-, which constitutes a pre-PIE manner of action formation based on the root *seh 1 - > aor. Ved. (áva, ví) sāt ‘loosens, unharnesses’ (reconstruction of the verbal root *sh 1 ei̯ according to Lipp 2009b: 205-209; for the structural derivation, cf. PIE *tḱei̯ - ‘to produce crops, settle’ from *teḱ- ‘to produce offspring, beget’, LIV 2 643 with n. 1; PIE *d h g u̯ h ei̯ - ‘to fade/ perish by heat’ from *d h eg u̯ h - ‘to burn (tr.)’, LIV 2 150 f. with n. 2). 13.4.1 In ancient grammar, the forms of the s-future were often identified with the corresponding forms of the future perfect and the subjunctive perfect, e.g. Paulus ex Festo 26L: amasso ‘amavero’; 25L: adaxint ‘adegerint’. Consequently, also in modern philology, the aforementioned forms of the type adiuverō, iuverint (with short -u-), moneris, monerint, sīrīs, sīrint with -rreflecting a morpheme *-shave been classified as extra-paradigmatic relics of subjunctive and optative formations of a former s-aorist, i.e., on the one hand, as forms of the future perfect based on a former prospective aorist subjunctive and, on the other hand, as forms of the subjunctive perfect based on a former optative of an aorist stem. Such a derivation from the s-aorist, however, would imply a functional expression of anteriority or perfectivity, as is usual with forms of the perfect subjunctive (except for extra-temporal prohibitions of the 2.sg./ pl.), for example in the case of perieris ‘may you have perished’ (Plautus Men. 295), nec quod una esca me iuverit magis ‘neither could any meal have tasted better to me at all’ (Plautus Most. 691 with perfect subjunctive / i̯ ūverit/ ‘could have helped = pleased/ tasted’ showing long vowel / ū/ ), utinam […] vere auguraverim ‘may I have divined correctly’ (Cicero Rep. 4,8). In contrast, s-future subjunctives have no relation to anteriority, cf. opsecro, ne quid in te mali faxit ira percita The Latin future tense formation of the type faxō and its Italic background 243 <?page no="244"?> ‘I beseech thee, lest blind rage do in thee some evil’ (Plautus Cas. 627 f.), ne boa noxit ‘may a varicose vein not hurt’ (Lucilius 1195M). 13.4.2 On the alleged origin of the aforementioned forms from the s-aorist, cf. Sommer (1914: 580-582), Meiser (1998: 184) (for further refs. on the dis‐ putable derivation of the entire category of the Latin s-future of the type faxō from the s-aorist, see above § 12.3.); for the attestations of the single forms of the type of adiuverō (with short u as against future perfect adiūverō), moneris (as against future perfect monueris) with -r- < *-s-, cf. Harðarson (1993: 113 f.). The philological and formal interpretation of these forms, as presented here, follows Rix (1998a: 624 f., 628 f., 630 f., 632), who conveys a congruent classification of the forms in question as instances of the s-future and not of the s-aorist, and whose view is shared by Meiser (2003: 40). A detailed philological interpretation with similar results is offered by de Melo (2007a: 215-223, 353), classifying these forms as belonging to the s-future, which in that frame (o.c. 306-327) is nevertheless considered as a category derived early from the paradigm of the s-aorist or a preterite based on it. 13.4.3 As stated above, however, the forms of the s-future are dura‐ tive-telic, i.e. terminative, not anterior in function (see above § 10.), and this is also the case with the subjunctive forms in question, i.e. those displaying -r- < *-s-, e.g. ita me divi […] iuverint ‘so may the gods help me’ (Catullus 66, 18 with s-future / i̯ uu̯ erint/ showing short / u/ in contrast to perfect iūvī with subjunctive / i̯ ūverit/ ; see above §§ 13.4.1.-2.), di monerint meliora atque amentiam averuncassint tuam ‘may the gods advise better things and eradicate your madness’ (Pacuvius 112R ≈ Lucilius 653M with s-future subjunctive monerint being in line with the following s-future subjunctive averuncassint), te precor […] utique fruges […] bene evenire siris, pastores […] servassis ‘I pray that you will let the fruit flourish and protect the shepherds’ (Cato Agr. 141, 2-3 with s-future subjunctive sīrīs being in line with the following s-future subjunctive servassīs). Thus, the forms in question correspond in function to normal s-future subjunctives like ita di faxint ‘so may the gods arrange it’ (Plautus Aul. 149 etc.; see above § 10.). In accordance with the analysis offered by Rix (1998a), this means that these forms are not relics of the s-aorist, but morphologically as well as functionally forms of the s-future category of the type faxō, faxim. 13.4.4 Despite the fact that in subordinate clauses the indicative forms of the future perfect as adiūverit ‘will have helped’, monueris ‘you will have admonished/ advised’, which express anteriority or perfectivity, are 244 Reiner Lipp <?page no="245"?> pragmatically indeed similar to the corresponding indicative forms of the Old Latin s-future, also in the case of the indicative it has to be stated that the s-future is functionally distinct, since, as pointed out above, it is dura‐ tive-telic, i.e. terminative (see above §§ 10., 13.4.3.). In correspondence to the subjunctive forms, this is not only the value of the indicative forms as faxō ‘I’ll make/ arrange’ appearing in main clauses, but also to be presupposed for the indicative forms in subordinate clauses; cf. Ennius Ann. 335/ 7V (= War‐ mington 1967a: 120 f. nr. 327/ 9; Skutsch 1985: 99 nr. 337/ 9 with commentary 512): o Tite, si quid ego adiuvero curamve levasso/ ecquid erit praemi ‘O Titus, if I really render any help or ease the sorrow, will there be any reward? ’ (Warmington’s reading adiuvero correctly represents adiuuero with <uu> = [uu̯ ] as transmitted in the manuscripts, which in the other philological text editions is unjustly rendered as adiuero in contradistinction to forms like future perfect adiuuero = / ad-i̯ ūu̯ erō/ ); the indicative form adiuverō showing short / u/ as s-future with the terminative value ‘will achieve to help, will really render help’ (< *ad-ii̯ ou̯ a-s-ō) is in line with levassō ‘will achieve to ease, will really ease’ (on the short scansion of / u/ , cf. Sommer 1914: 581; Leumann 1977: 596; Skutsch 1985: 512; Rix 1998a: 624 with n. 15; on the passage Benveniste 1922: 55: “C’est seulement le fait accompli qui peut valoir à son auteur une récompense”). Pacuvius 30R: quod me moneris, effectum dabo ‘what you will give me as advice, to that I’ll give effect’, with moneris ‘(what) you will give as advice’ being likewise a terminative s-future (< *mone-s-e-s), but undergoing re-interpretation in later tradition as future perfect monueris ‘(what) you will have advised’, as glossed by Nonius 507, 25. 13.5 As shown above, the Latin s-future shows a morphological set-up built on a verbal base abstracted from the present stem, which also gives an indication of the time-horizon of its genesis. Helmut Rix (1998a: 627-629, 633) has pointed out that this formation pattern already belonged to an early stage of Proto-Italic, since the differentiated shape of some formational types of the s-future presupposes a state with an as yet preserved present stem marker *-i̯ e/ oin the paradigmatically associated present stem (prior to the Proto-Italic loss of intervocalic *-i̯ -), as can be seen in the II. conjugation: On the one hand, there are forms like present Lat. habeō, -ēre ‘have’ with the stem / hab-ē-/ resulting by vowel contraction (after the loss of intervo‐ calic *i̯ ) from the essive formation *χab-ē-i̯ e/ oand the associated s-future subjunctive OLat. habessit < Proto-Italic optative *χabē-s-ī-d with the stem The Latin future tense formation of the type faxō and its Italic background 245 <?page no="246"?> OLat. / habess-/ < Proto-Italic *χab-ē-s-, built on the verbal base *χab-ēwith final long vowel, as obtained in the pre-Proto-Italic period by subtracting the specific present marker *-i̯ e/ ofrom the present stem *χab-ē-i̯ e/ o- (see above § 13.2.1.). On the other hand, this conjugation displays forms of the type of Lat. noceō, -ēre ‘harm’ with the stem / nok-ē-/ resulting by vowel contraction (after the loss of intervocalic *i̯ ) from the causative formation *nok-e-i̯ e/ o- ← *nok-ei̯ e/ oand the associated s-future subjunctive OLat. noxit < Proto-Italic optative *noke-s-ī-d (undergoing syncope of the medial short vowel in the prehistory of Latin) with the stem / noks-/ < Proto-Italic *nok-e-s-, built on the verbal base *nok-ewith final short vowel, as obtained in the pre-Proto-Italic period by subtracting the specific present marker *-i̯ e/ ofrom the present stem *nok-e-i̯ e/ o- (see above § 13.2.1.). Likewise, in the I. conjugation, there is Lat. (ad-)iuvō, -āre ‘support, help’, inscr. OLat. adiouanto (perf. (ad-)iūvī < *ii̯ ou̯ a-u̯ ai̯ , ppp (ad-)iūtum < *ii̯ ū-to-) with the deverbative present stem / i̯ ou̯ ā-/ ~ / ad-i̯ uu̯ ā-/ < *(ad-)ii̯ ou̯ a-i̯ e/ o- (with *°a-i̯ e/ o- > °ācontaining the de‐ rivative suffix *-i̯ e/ o-, the attested stem-final long vowel °ābeing otherwise inexplicable) from athematic *ii̯ ou̯ a-/ *ii̯ ū- (< PIE *h 1 i-h 1 éu̯ H-/ *h 1 i-h 1 uH-) and the associated s-future ind. OLat. adiuverō, adiuverit, subj. adiuverit (with short u) < Proto-Italic *(ad-)ii̯ ou̯ a-s-ō, *(ad-)ii̯ ou̯ a-s-ti, *(ad-)ii̯ ou̯ a-s-ī-d (in the prehistory of Latin, s-future indicative *ii̯ ou̯ a-s-ti → *ii̯ ou̯ a-s-e-ti > ad-iuverit by regular thematization of the athematic forms of the Proto-Italic semi-the‐ matic paradigm), with *ii̯ ou̯ a-sbased on the unextended present stem *ii̯ ou̯ ain final short vowel, stemming from the period before its regular extension in the present tense paradigm to *ii̯ ou̯ a-i̯ e/ o- > *ii̯ ou̯ ā- > (ad-)iuvāre (see above § 13.1.4.). This formation is in contrast to denominatives as Lat. cantō, -āre ‘sing’ (perf. cantā-vī, ppp cantā-tum) with the stem / kant-ā-/ , containing the denominative suffix -ā- < *-ā-i̯ e/ o- (< PIE */ -ah 2 -i̯ e/ o-/ with basic suffix *-i̯ e/ o-), and the associated s-future cantassō*, subjunctive excantassit < Proto-Italic *kantā-s-ō, optative *kantā-s-ī-d from the long-vowel basis *kant-ā-, as abstracted in pre-Proto-Italic from the present stem *kant-ā-i̯ e/ oby dropping the specific present marker *-i̯ e/ o- (see above § 13.1.1.). Only during the prehistory of Latin, this denominative type triggered the remod‐ elling of the paradigms of deverbative present formations as in the case of Lat. amō, -āre ‘like, love’, showing the present stem / amā-/ < *ama-i̯ e/ o- ← *ama-/ *ama- ‘to grab, hold on’ (< PIE */ h 2 ámh 3 -/ = *h 2 émh 3 -/ *h 2 m̥ h 3 -) and the remodelled s-future amassō instead of †amerō < Proto-Italic *ama-s-ō = *[amazō]; in view of the synchronic present stem / amā-/ , the form of the 246 Reiner Lipp <?page no="247"?> future amassō results from a reshaping by analogy with the more frequent denominative type, displaying a configuration with present stem / kantā-/ : s-future *kantā-s-ō > cantassō* (triggering also the remodelling in the rest of the deverbative paradigm type of / amā-/ < *ama-i̯ e/ owith perfect amā-vī, ppp amā-tum, for †amuī < *ama-u̯ ai̯ , †amitum < *ama-to-; see above §§ 13.1.1., 13.1.2., 13.1.3.). 14.1 In view of these facts, one must consider that, in the prehistory of Proto-Italic, there existed the inherited type of the s-future, continuing the PIE desiderative as originally amphikinetic s-formation built on the root (see above § 2.). With levelling of the root morpheme, this future tense type is represented, on the one hand, by Proto-Italic formations of the s-future like indicative 3.sg. *fū-s-ti ‘will be’, *ei̯ -s-ti ‘will go’ (> Sabellic *fī-s-t → *fu-s-t > Umbr. fust, fust, Oscan fust; Sabellic *ei̯ -s-t > Umbr. eest, est = / ēst/ ) beside athematic presents as *es-ti ‘is’, *ei̯ -ti ‘goes’ (> Sabellic *est > Umbr. est, est, Osc. est, íst = Lat. est ‘is’; Sabellic *ei̯ -t ‘goes’ not attested, but imptv. II *ei̯ -tọ̄d > Umbr. etu, etu, eetu ‘shall go’; cf. *ei̯ -t > OLat. īt, Lat. it). On the other hand, this future tense type is represented by Proto-Italic formations like indicative 3.sg. *ag-s-ti ‘will drive, act, do’ (optative 1.sg. *ag-s-ī-m > subjunc‐ tive OLat. āxim), *fak-s-ti ‘will make’ (with 3.pl. *fak-s-onti ← *fak-s-enti in the semi-thematic indicative paradigm of the Proto-Italic s-future, followed by a full thematization in the prehistory of Latin with 1.sg. *fak-s-ō, 2.sg. *fak-s-si → *fak-s-e-si, 3.sg. *fak-s-ti → *fak-s-e-ti > OLat. faxō, faxis, faxit etc.), 3.sg. *sei̯ -s-ti ‘will let (happen), allow’ (optative *sei̯ -s-ī-d > subjunctive OLat. sīrit) beside thematic presents as *ag-e-ti ‘drives, acts, does’ (> Lat. agit), *fak-i̯ e-ti ‘makes’ → Proto-Italic *fak-i-ti (> Lat. facit = Osc. αfα̣κειτ / am-fakẹt/ ; on the secondarily semi-thematic stem formation, see below § 17.1.) and secondarily thematized athematic forms like 3.pl. nasal present *si-n-enti → Proto-Italic *sin-onti (triggering a semi-thematic indicative paradigm in Proto-Italic with subsequent inner-Latin thematization: 1.sg. *sin-ō, 2.sg. *si-na-si → *sin-e-si, 3.sg. *si-na-ti → *sin-e-ti > Lat. sinō, sinis, sinit etc.). This was a configuration that, according to Rix (1998a: 620, 628 f., 633), already in the prehistory of Proto-Italic (when intervocalic *-i̯ was still preserved) conditioned a rule to form s-futures from a de-characterized verbal base, resulting from the subtraction of specific markers for the present stem (like *-e/ o-, *-i̯ e/ o-, *-ske/ o-, nasal infix). As shown above (§ 13.5.), this applied, for instance, to paradigms belonging to the II. conjugation with synchronic present stems in -ēeither from pre-Proto-Italic essives in *-ēi̯ e/ oor dever‐ The Latin future tense formation of the type faxō and its Italic background 247 <?page no="248"?> batives in *-ei̯ e/ o-, which show a corresponding difference in their s-future formation. Thus, there were cases like the Proto-Italic s-future *χabē-s-ti ‘will hold’ (optative *χabē-s-ī-d > OLat. habessit), displaying the stem *χab-ē-sbased on *χab-ē-, as abstracted from the present stem *χab-ē-i̯ e/ o- ‘to hold’ (> *habē- > Lat. habeō, -ēre ‘hold, have’, Umbr. habetu, habitu ‘shall hold’), and the Proto-Italic s-future *noke-s-ti ‘will harm’ (optative *noke-s-ī-d > OLat. subjunctive noxit), displaying the stem *nok-e-sbased on *nok-e-, as abstracted from the present stem *nok-ei̯ e/ o- (original causative formation with o-grade of the root) by re-segmentation as *nok-e-i̯ e/ o- ‘to harm’ (> *nokēin Latin noceō, -ēre). 14.2 A post-Proto-Italic innovation of the Sabellic branch is the introduc‐ tion of the present tense stem as basis of the s-future formation. This means, the future stem in -sis synchronically built on a basis that includes the specific markers of the present stem (such as thematic vowel *-e- ← *-e/ o-, suffix *-i̯ e- ← *-i̯ o-/ -i-, essive suffix *-ē- < *-ē(i̯ )e- ← *-eh 1 -i̯ e/ o-). Thus, there is the thematic present stem / fer-e/ o-/ , represented by Marruc. feret = / fer-ent/ ‘they carry away/ receive’, Osc. amfret = / amfrent/ ‘they carry around’ < *amb(i) + *fer-ent ( ← *fer-ont = Lat. ferunt < Proto-Italic *fer-o-nti with regular Sabellic replacement of 3.pl. ending *-ont by *-ent), Umbr. fertu = / fertu/ ‘shall carry, bring’ (with medial syllable syncope) < *fer-e-tọ̄d, and, derived from the thematic present stem, the s-future Umbr. ferest ‘will carry/ bring’ = / fer-e-s-t/ for Proto-Italic *fer-s-ti; likewise, there is the thematic present stem */ did-e/ o-/ , represented by Vest. didet = / dẹdet/ ‘gives’ < *did-e-t, Umbr. imptv. II titu, tetu, ditu = / dẹtu/ ‘shall give’ (with medial syllable syncope) < *did-e-tọ̄d (like Lat. reddō, -ere ‘give back, render’ < *re-did-e/ othematized via 3.pl. *did-ont < Proto-Italic *did-onti < *di-dh 3 -enti ← PIE *dé-dh 3 -n̥ ti = Ved. dádati ‘they give’, based on the PIE present stem *dé-doh 3 -/ *dé-dh 3 -, cf. 3.sg. Ved. dádāti, Gk. δίδωσι < *di-dō-ti), and, derived from the thematic present stem, there is the s-future Osc. didest ‘will give’ = / dẹdest/ < */ did-e-s-t/ , replacing Proto-Italic *dō-s-ti or *da-s-ti < PIE *déh 3 -s- = */ dóh 3 -s-/ (= Lith. duõs and thematized Gk. δώσω) ~ *dh 3 -s- (see above § 6.). 14.3 The semi-thematic present stem / her-i̯ o-/ -i-/ ‘to wish, desire’ re‐ sults from the remodelling of pre-Proto-Italic *χer-i̯ o-/ -i̯ einto Proto-Italic *χer-i̯ o-/ -i- (on the mechanism of the remodelling, see below § 17.1.), with the basic suffix form *-i̯ ebeing still reflected in conditional phrases by the archaic subjunctive Umbr. heriiei, heri = / heri̯ ẹ/ ‘(si) velit/ (if) anyone wishes’ < *her-i̯ ẹ̄-d < Proto-Italic *χer-i̯ ē-d ‘should anyone wish’ (eventualis), 248 Reiner Lipp <?page no="249"?> showing *-i̯ ē- = *-i̯ e-ein the subjunctive as lengthened form of the suffix *-i̯ e- (as against the productive Sabellic subjunctive form Osc. heriiad); this semi-thematic present is represented by Umbr. passive herter, herte, herti, hertei = / hertẹr/ ‘is desired, necessary’ < *her-i-ti-r (syncope of short vowel in open medial syllable), infinitive herifi ‘to wish’ < *her-i-fi̯ ẹ̄ -r, showing the associated s-future Umbr. heriest, Osc. herest = [heri̯ est] = / her-i-es-t/ instead of */ her-i-s-t/ (based on the semi-thematic present stem */ her-i-/ with the synchronic future stem suffix / -es-/ ; see below) for Proto-Italic *χer-s-ti (without present stem suffix) (on the paradigm, cf. Meiser 1986: 130; LIV 2 176 f. s.v. 1. *ǵ h erwith nn. 4a, 5; Lipp 2022: 519 f.). Yet, Meiser (2003: 61) interprets the related Umbrian particle / heri̯ ẹz/ < *her-i̯ ẹ̄-s in heri(s) […] heri(s), heriei […] heriei etc. ‘vel […] vel/ either […] or’ as a fossilized optative aorist with secondary full grade of the root for *χor-i̯ ē- < *ǵ h r̥ -i̯ éh 1 -, which is however unparalleled in view of a relic form such as OLat. duim ‘I would give’ < *dū-ī-m ← *dū-i̯ ē-m < *duh 3 -i̯ éh 1 -m (Harðarson 1993: 86 with n. 115; LIV 2 107 with nn. 1, 2; verbal u-stem PIE *deh 3 -u-/ *dh 3 -u- > *duh 3 - ‘to hand over, give’ from *deh 3 - ‘to give’ as a telic or terminative manner of action formation like PIE *terh 2 -u- ‘to overcome’ from *terh 2 - ‘to pass through’; cf. LIV 2 633 f.) with generalized *-ī- < *-ih 1 outside monosyllabic forms (see above, § 9.5.). Moreover, this analysis can functionally not apply to the inflectional 3.sg. / heri̯ ẹ/ < *her-i̯ ẹ̄-d in conditional phrases, reflecting the older pattern of the present subjunc‐ tive (on the Sabellic forms, cf. Untermann 2000: 321-324, 325). There is also the semi-thematic present stem */ hab-i̯ o-/ -i-/ ‘to seize’ in Umbr. imptv. II hahtu, hatu, hatu ‘shall seize’ < *hab-i-tọ̄d (medial syllable syncope) and future Umbr. habiest ‘will seize’ = [habi̯ est] = / hab-i-es-t/ instead of */ hab-i-s-t/ (based on the semi-thematic present stem / hab-i-/ with the syn‐ chronic future stem suffix / -es-/ ) for Proto-Italic *χab-s-ti (without present stem suffix) (on the paradigm, cf. Meiser 1986: 126, 2003: 101 f.; LIV 2 195 s.v. *g h eHbwith n. 2). On the basis of an analysis proposed by Harðarson (1998: 336 with n. 40), this formal remodelling with a synchronic future tense suffix -escan be explained as an effect of analogy with morphological oppositions such as s-future */ em-e-s-t/ ‘will take away’ > Osc. pertemest ‘will prevent’ (with analogical retention of the root vowel after present *em-e-t = Lat. emit ‘buys’, as represented by Umbr. imptv. II sumtu ‘shall take up’ < *sub-em-e-tọ̄d with medial syllable syncope beside vowel weakening in inf. pertumum < *pert-em-om) : future perfect */ emm-us-t/ ‘will have taken The Latin future tense formation of the type faxō and its Italic background 249 <?page no="250"?> away’ > Osc. pertemust ‘will have prevented’ (like Lat. ēmī ‘have bought’ with stem of lengthened grade perfect *ēm- > *emmas in Osc. emmens ‘they have bought’ by effect of the littera-rule, as above § 11.2.; on the forms, cf. Rix 1998a: 632 n. 37; Meiser 2003: 73 with n. 137, 199 with n. 18, considering for Oscan also a thematic preterite *em-e/ owith future perfect *em-us-t from root aor. *em-, as allegedly preserved in OLat. em ‘take! ’, likewise Untermann 2000: 219-221); together with the imperative present form *em < *em-e = OLat. em ‘take! ’ (not root aor. as against Meiser o.c.) from emō, -ere ‘buy’ ( ← ‘take’), the structure of future perfect */ emm-us-t/ (with suffixal -Vs-t following the root stem) could trigger subsequent reanalysis of s-future */ em-e-s-t/ as */ em-es-t/ (for the difference in the shape of the root morpheme in the future perfect stem */ emm-us-/ and the future tense stem */ em-es-/ , the assumed reanalysis of the future tense stem as a synchronic formation in / -es-/ , based on proportional analogy with the future perfect in / -us-/ , was doubted by Martzloff 2006a: 442 f., but, despite the slight variance in the root shape, the reanalysis obviously started from the fact that it is a matter of allomorphs of one and the same root morpheme). A comparable configuration can be found in s-future */ did-e-s-t/ ‘will give’ > Osc. didest (after present *did-e-t > Vest. didet) : future perfect */ ded-us-t/ ‘will have given’ (with reduplicated perfect stem *de-das in */ ded-e-d/ > Osc. deded, Umbr. dede ‘has given’) → */ did-us-t/ (with / i/ after present */ did-e-t/ ) > Umbr. teřust, dirsust = / dẹřust/ , with possible reanalysis of s-future */ did-e-s-t/ as */ did-es-t/ (like future perfect */ did-us-t/ ). A further pattern for the emergence of a synchronic future suffix -esmight be provided by the s-future Umbr. ferest with re-segmentation as / fer-es-t/ ( ← / fer-e-s-t/ ), showing a descriptive present stem base / fer-/ (combined with the deriva‐ tional future suffix / -es-/ ) by secondary athematic interpretation of 3.pl. / fer-ent/ = Marruc. feret ( ← *fer-ont = Lat. ferunt) after pres. 2.sg. (*fer(r) <) *fer-s (on final *rs in Sabellic Buck 1928: 77) < *fer-e-s (Proto-Sabellic final syllable syncope) = Lat. fers ← *fer < *ferr < *fers < *fer-e-s (restitution of the ending after inner-Latin phonological change *RVs# > *Rs# > *RR# > *R, cf. *u̯ iros > *u̯ irr > Lat. vir ‘man’) and imptv. 2.sg. *fer < *fer-e, 2.pl. *fer-te < *fer-e-te, imptv. II *fer-tọ̄d = Umbr. fertu < *fer-e-tōd (allegro forms with early apocope and presumably premature syncope) = Lat. fer, ferte, fertō (as against the normal thematic type Lat. age, agite, agitō < *ag-e, *ag-e-te, *ag-e-tōd; on subjunctive Lat. ferrent < *fer-s-ē-nd based on *ferfrom the 250 Reiner Lipp <?page no="251"?> synchronically semi-thematic present tense paradigm ferō, fers, fert etc., see below §-15.4.2.). 14.4 Beside Umbr. 1.sg. stahu = / sta'u/ ‘I stand’ (with hiatus marker h) < *sta'ọ̄ = Lat. stō, stat (stāre) < *sta-ō, *sta-e-t from the present stem *sta-e/ o- ‘to stand’ < *sta-i̯ e/ o- < *sth 2 -h 1 i̯ é/ ó- (formation with the PIE essive suffix *-h 1 i̯ é/ ó-; see Harðarson 1998: 336 f. with n. 38, followed by Lipp 2021: 277, 300; cf. LIV 2 590-592 with nn. 24, 25, 26), the Sabellic branch shows the morphologically reshaped present stem *sta-ẹ̄ -e/ o- ‘to stand’, with the essive marker *-ẹ̄applied for functional recharacterization on the analogy of the semantically related present stem *sed-ẹ̄ -e/ o- ‘to sit, to be sitting’, as represented by Umbrian sersitu ‘sedeto/ shall sit’ < *sedẹ̄ tọ̄ d = Latin sedeō, -ēre < *sed-ē-(i̯ )e/ o- < *sed-eh 1 -i̯ é/ ó- (with renewed essive suffix *-eh 1 -i̯ é/ óas remodelling of PIE *-h 1 i̯ é/ ó-; on the paradigm of the verb ‘to stand’ in the Italic languages, but with different historical analyses of the stem formation, cf. Untermann 2000: 697-700; García Castillero 2000: 357-361; Martzloff 2006a: 238-270 with refs.). This characterized essive stem is well represented in present tense paradigms of the individual Sabellic languages, appearing in Umbr. imptv. II stahitu = / sta'ẹtu/ ‘shall stand’ (with hiatus marker h) < *sta'ẹ̄tọ̄d, South Picene praistaít = / prai̯ -sta'ẹ̄ t/ ‘stands out, towers’, Osc. staít = / sta'ẹt/ ‘stands, belongs to’ < *sta'ẹ̄ t, South Picene praistaínt = / prai̯ -sta'ẹ̄ nt/ ‘they stand out, tower’, Osc. stahínt, stahint = / sta'ẹnt/ ‘they stand’ (with hiatus marker h) < *sta'ẹ̄ nt, Osc. eestínt = / ē-stẹnt/ ‘they stand out’ < *eχ-sta'ẹ̄ nt (with syncope of short vowel *-ain open medial syllable), Osc. staíet = / sta'ẹ'ent/ ‘they stand’ < *sta'ẹ̄ 'ent ← *sta-ẹ̄ -ont (Sabellic replace‐ ment of 3.pl. ending *-ont by *-ent as Osc. fiíet, fiiet = / fīi̯ -ent/ ‘happen, are performed’ ← *fīi̯ -ont = Lat. fīunt ‘become’ < Proto-Italic *fī i̯ -o-nti < *fū-i̯ o-nti). Following the regular formational pattern, the associated Sabellic s-future is built on this recharacterized present stem *sta-ẹ̄ - (< *sta-ẹ̄ -e-), as exemplified by Umbr. fut. staheren = / sta'ẹrent/ ‘they will stand’ (with hiatus marker h; form attested in TI Ib 19 staheren: termnesku ‘stabunt apud termina’ with haplography for *staherent: tº) < *sta'ẹ̄ -z-ent ← *sta-ẹ̄-s-ont (Sabellic 3.pl. in *-ent for thematic *-ont within a semi-thematic future tense paradigm inherited from Proto-Italic). 14.5 In conclusion, one can state that Sabellic introduced the present tense stem as derivational basis for the Sabellic future formation with the characterizing suffix *-s-, as exemplified by */ did-e-s-t/ ‘will give’ > Osc. didest after present */ did-e-t/ ‘gives’ > Vest. didet (for Proto-Italic s-future The Latin future tense formation of the type faxō and its Italic background 251 <?page no="252"?> *dō-s-ti or *da-s-ti from the PIE desiderative *déh 3 -s-/ *dh 3 -s-), but, counter to Latino-Faliscan, Sabellic retained the inflection of the athematic forms in the presumably semi-thematic paradigm of the s-future, as inherited from Proto-Italic. 14.6 As evidence for a modal form of the s-future in Sabellic, one can adduce South Picene aisis TE 5. The form appears in the sequence povaisis, which, taking into account the context, is explained by Martzloff as ‘quo axis’ = ‘whereby you may do’ ~ ‘so that you may achieve’, with aisis as a mor‐ phological match of the attested future subjunctive OLat. āxim < *ag-s-ī-m (Martzloff 2006a: 221, 325, 359, 444, 447, 470, 471, 472, 485, 750; 2006b: 75-77, 2009: 363 f., 369-371), which continues the optative of the desiderative stem *ag-sand is paradigmatically linked to the present stem *ag-e/ o- > Lat. agō, -ere ‘to drive, do, achieve’. However, the derivation of SPc. aisis from *aksīs < *ag-s-ī-s is at variance with the normal Sabellic development of inherited *-ks- > -ss-, which occurs also in the closely related Pre-Samnite dialect of Tortora in the case of εσ[ε]ι Ps 20 < *ekso- ‘this’ (Lazzarini/ Poccetti 2001: 51, 153 f.). Nevertheless, according to Lipp (2021: 284), Martzloff’s functional identification of the form can be defended if one takes into account that in Sabellic the sigmatic future formation based on an athematic desiderative in -swas reshaped on the basis of the respective present stem. Thus, the inherited desiderative *fer-swas reshaped to *fer-e-s- > Umbr. ferest ‘will carry, bring’ as paradigmatic future belonging to the present stem *fer-e- > Marruc. feret = / fer-ent/ ‘they carry away, receive’, Osc. amfret = / amfrent/ ‘they carry around’ < *amb(i) + *fer-ent = Lat. ferunt ‘they carry, bring’ < *fer-ont. Therefore, in Sabellic the desiderative stem *ag-shad to be reshaped to *ag-e-sas regular paradigmatic future formation belonging to the present stem *ag-ein imptv. II Osc. actud, Umbr. aitu, aitu ‘shall drive/ move’ (with regular vowel syncope) = Lat. agitō < *ag-e-tōd. Accordingly, for the original desiderative one can posit an optative form *ag-s-ī-s, with subsequent reshaping to Sabellic *ag-e-s-ī-s, whence by dialectal South Picene velar palatalization *aɟezīs and, by recent vowel syncope in open medial syllable, finally the attested form aisis = [ajzīs] in the inferable meaning ‘you may do/ achieve’ emerges (velar palatalization attested in the same inscription TE 5 with <ś> = [š] ~ [č] < *k in śidom ‘this’ < *kid-om, eśelsít ‘they erect, elevate’ < *ek-kelsẹ̄ nt; cf. Eichner 1993: 55 with n. 28, 61 n. 39; Martzloff 2006b: 66; Lipp 2021: 289, 291). South Picene aisis < *aɟezīs < *ag-e-s-ī-s therefore displays a phonetic development like 252 Reiner Lipp <?page no="253"?> the respective Umbrian imperative form aitu, aitu (syncope) < *aɟ-e-tọ̄ d̥ (velar palatalization) < *ag-e-tōd (= Lat. agitō), matching Oscan actud (with syncope, but regularly without previous palatalization). A Sabellic parallel for the reshaping of the s-future SPc. aisis < *ag-e-s-ī-s ← *ag-s-ī-s after the thematic present stem *ag-e/ oappears in Palaeo-Hernican ni kait[sis He 2 ‘don’t break (the vessel)’ (Rix 1998b: 253-255) as a subjunctive form of the s-future related to Lat. caedō, -ere ‘strike, smash’, representing the Proto-Italic present stem *kai̯ d-e/ othat triggered the Sabellic remodelling of *kai̯ d-s-ī-s > *kai̯ ssīs to *kai̯ d-e-s-ī-s > / kai̯ tsīs/ (with medial syllable syncope), which in fact has a morphologically unaltered cognate in OLat. occisit ‘should someone kill’ = / ok-kẹ̄ s-s-ī-t/ according to the form of redaction + oc‐ cessit (presumably reflected by the gloss CGL V 630, 21 occelsit ‘occiderit’) < *(ob-)kai̯ d-s-ī-d, as attested e.g. XII Tabulae 8, 12 Si nox furtum factum sit (codices; faxsit em. Cuiacius), si im occisit, iure caesus esto ‘Should theft have been done by night (should someone commit theft by night), should someone (= owner) kill him (= thief), he (= thief) shall be held killed according to law’ (for the text of the statute, cf. Bruns 1893: 31, Warmington 1967b: 482 f. with nn. 12 and c, and Crawford 1996: 578 f., 609-613 proposing the text reconstruction si nox furtum fa<x>it, <ast> im occisit, iure caesus esto; on the formation, see above §-13.3.1.). 15.1 In contrast to the Latin future tense category of the type represented by the synchronically thematic forms Lat. faxō, faxis, faxit, Sabellic shows athematic future forms with stem suffix *-s-, as represented in its proto-type by Oscan fust, Umbrian fust, fust ‘will be’ = / fu-s-t/ ← *fīst (Sabellic *ū > ī in monosyllabic forms and final syllables; Buck 1928: 41) < Proto-Italic *fū-s-ti (with secondarily introduced *fufrom reduplicated or antevocalic forms, see below § 15.2.). Supplying the future of the present tense est, est ‘is’ < *es-ti, it still continues the Proto-Italic athematic form *fū-s-ti, levelled from the inherited PIE desiderative *b h u̯ éh 2 -s-/ *b h uh 2 -sas derivate of the root PIE *b h u̯ eh 2 -/ *b h uh 2 - ‘to become’ > Italic *f(u)u̯ ā-/ *fū-, represented by OLat. subjunctive fuat from prohibitive nē fuat ← *nē fuu̯ ā-d (with effect of Lex Lindeman after heavy syllable) ← injunctive *méh 1 b h u̯ ah 2 -d̥ (< *b h u̯ éh 2 -t) and OLat. perf. fūit after fūimus ← perfect *fu-fū-mos < *fe-fū-mos (cf. LIV 2 98 f. with nn. 1, 2, 5, 6, Lipp 2021: 307 f.; see below §§ 15.2., 16.5., 17.2.). A modified form of this inherited future tense formation appears in the pre‐ vailing Sabellic type represented by Oscan didest ‘will give’, Umbr. ferest ‘will carry/ bring’ (replacing the inherited Proto-Italic formations *dō-s-ti, The Latin future tense formation of the type faxō and its Italic background 253 <?page no="254"?> *fer-s-ti based on the root), which shows reshaping on the analogy of the thematic present stem, as attested by Vestinian didet, Marrucinian 3.pl. feret = / fer-ent/ . As is obvious from the contrast between 3.sg. present didet and future didest, the lack of the thematic vowel in the ending *-s-t of the future tense form cannot be due to a syncope of †-s-e-t to *-s-t (in Proto-Sabellic there was only a syncope in final syllables that concerned short vowels in sequences with the pattern *-CVs# and *-Ci̯ / u̯ Vm#; Benediktsson 1960: 215-234, esp. 230-233; Meiser 1986: 59-61; Hill 2004: 120 f.). Therefore, it is excluded that this future tense formation goes back to the prospective subjunctive of the s-aorist with 3.sg. in *-s-e-d, displaying furthermore the so-called secondary ending *-d (< PIE *-t = *[-d̥ ] = */ -d/ ), which would have been preserved in Sabellic, as it is otherwise represented in the forms of the 3.sg. subjunctive, e.g. Osc. deiuaid ‘shall swear’ = [dei̯ u̯ a'ẹd] < *dei̯ u̯ ā'ẹ̄ d < *dei̯ u̯ ā-i̯ ē-d from the denominative present stem *dei̯ u̯ ā-i̯ e/ o- ‘to invoke the god(s)’ (see above § 12.4.2.). This means that we are clearly dealing with a genuinely athematic formation that structurally continues the athematic formation of the desiderative, as inherited from Proto-Indo-European. 15.2 As shown above (§§ 9.1., 12.1.), like the Proto-Italic formation under‐ lying the Latin future tense of the type faxō, the paradigm of the Sabellic s-future was presumably semi-thematic with the thematic vowel *-oin the paradigmatic slots where the thematic inflexion had the vowel *-o-, but athematic in the paradigmatic slots where the thematic inflection had the thematic vowel *-e- (see below § 15.3.1.). The patterning is thus comparable to the preserved semi-thematic inflection of the type of Latin volō, vīs, vult, volumus, vultis, volunt ‘want’ < *u̯ el-ō, *u̯ el-s, *u̯ el-t, *u̯ el-o-mos, *u̯ el-tes, *u̯ el-o-nt (see above § 9.2.). Therefore, for the pre-stage of the Sabellic s-future one can reconstruct the pattern inherited from Proto-Italic: *fū-s-ō, *fū-s-si, *fū-s-ti, *fū-s-o-mos, *fū-s-tes, *fū-s-o-nti. Oscan fust and Umbrian fust, fust ‘will be’ = / fu-s-t/ ← *fīst < *fū-s-t show secondary replacement of *fū- > *fīby *fuwith short vowel from reduplicated or prevocalic formations like *fu-fuu̯ - (with assimilation of the reduplication syllable) < *fe-fuu̯ - (prevocalic form by generalization of the plural stem *fe-fū-), as in Pre-Samnite fufuhud ‘has been’ < *fu-fuu̯ -o-d ← *fu-fuu̯ -e-d < *fe-fuu̯ -e-d, Oscan fufens ‘they have been’ < *fufu̯ ens < *fu-fuu̯ -ens ← *fu-fuu̯ -o-nd = Pre-Samnite fυfυϝοδ (with preserved thematic ending in 3.pl. preterite) < *fe-fuu̯ -ond. In Oscan and Umbrian there was a regular replacement of the thematic primary ending *-o-nt by athematic *-ent, as usual, cf. 3.pl. Oscan 254 Reiner Lipp <?page no="255"?> fusent (on a newly discovered inscription of the so-called iuvilas group; Poccetti 2018), Umbrian furent ‘they will be’ < *fus-ent ← *fūs-o-nt like 3.pl. present Osc. fiíet, fiiet / fīi̯ -ent/ ‘happen, are performed’ ← *fīi̯ -ont = Lat. fīunt ‘become’ < Proto-Italic *fī i̯ -o-nti < *fū-i̯ o-nti. For the fragmentary character of the Sabellic evidence, except for the 3.pl. with regular replace‐ ment of thematic *-o-nt by *-ent, up to now there is no attestation of the paradigmatic future tense forms of the 1.sg. in *-ō and of the 1.pl. with the thematic vowel *-o-. 15.3.1 On the other hand, like for the Proto-Italic forerunner of the Latin s-future, there is at least indirect evidence for the semi-thematic character of the Sabellic s-future as inherited from Proto-Italic: It is provided by the so-called imperfect subjunctive for the expression of the conditional, which is derived from the stem of the s-future. Structurally this category represents a long-vowel subjunctive in *-ē- (< *-e-e-), which presupposes a thematic pri‐ mary stem, from which it is derived (on the original status of the imperfect subjunctive as long-vowel subjunctive derived from the Proto-Italic s-future as a category with semi-thematic inflection, cf. Harðarson 1998: 338 n. 46). Thus, there is the type of the imperfect subjunctive formation represented by Oscan fusíd = / fu-s-ẹ-d/ = [fuzẹd] ‘would be’ from the Cippus Abellanus (Cm 1A19. 23; late 2 nd c. BC), which represents the original long-vowel subjunctive *fu-s-ẹ̄ -d corresponding to the form of the Latin imperfect sub‐ junctive foret ← *fu-s-ē-d. As such, this Sabellic imperfect subjunctive form is derived from the semi-thematic s-future type *fu-s-t/ *fu-s-o-nt ‘will be’ (with regular replacement of the thematic 3.pl. ending *-o-nt by athematic *-ent in Sabellic), underlying the attested future forms Osc. fust, fusent, Umbr. fust, fust, furent (see above §-15.2.). Therefore, the imperfect subjunctive formation corresponds structurally to the long-vowel subjunctive continued by the future forms belonging to the semi-thematic presents Lat. volō, vīs, volt ~ vult, volumus, voltis ~ vultis, volunt ‘want’ and Lat. edō, ēs, ēst, edimus, ēstis, edunt ‘eat’ (see above § 9.2.). In these paradigms, the derived long-vowel subjunctive appears in the functional future tense forms volam ← *u̯ elem < *u̯ el-ē-m; volēs < *u̯ el-ē-s; volet < *u̯ el-ē-d; volent < *u̯ el-ē-nd etc. ‘will want’ and edam ← *edem < *ed-ē-m; edēs < *ed-ē-s; edet < *ed-ē-d; edent < *ed-ē-nd etc. ‘will eat’ (with the endings of 3.sg. -t and 3.pl. -nt < *-ti, *-nti generalized from the present tense). If the primary paradigm had been athematic, the respective subjunctive would only display the short vowel of the proper subjunctive marker, i.e. *-e/ o-, but not the fusional The Latin future tense formation of the type faxō and its Italic background 255 <?page no="256"?> long vowel from the combination of thematic vowel + subjunctive marker, leading to *-ē- (< *-e-e-)/ *-ō- (< *-o-o-) with secondary generalization of *-ē-; cf. the athematic indicative present Lat. est = / es-t/ ‘is’ (with -t < PIE *-ti) as derivational base for the future Lat. erit ‘will be’ ← subjunctive *es-e-d ‘shall be’ (short vowel as subjunctive marker) = Palaeo-Lat. esed (Forum inscr. CIL I 2 1; with the so-called secondary ending -d < PIE *-t = [-d̥ ] = / -d/ ), as against the thematic indicative present Lat. agit < *ag-e-t ‘acts, does’ as derivational basis for the future Lat. aget ‘will act/ do’ ← *ag-ē-d ‘shall act/ do’, showing the fusional subjunctive marker *-ēas preserved or reflected also by the other future tense forms as 2.sg. agēs, 1.pl. agēmus, 2.pl. agētis, 3.pl. agent ← *ag-ē-nd vs. the take-over from the synchronic ā-subjunctive in 1.sg. agam, replacing *agem < *ag-ē-m; cf. the inherited form in South Picene CH 2 putih knúksem = / putih gnọ̄ skem/ ‘uti gnoscam’ < *gnōskēm with subjunctive stem *gnōsk-ē-, derived from present stem *gnōsk-ein Lat. gnōscit (see below §-17.1.). 15.3.2 Therefore, also the s-future of the type faxō, faxis, faxit, faxint* ( ← *faxunt) ‘will make/ arrange’ as a remodelling of the semi-thematic paradigm *fak-s-ō, *fak-s-si, *fak-s-ti, fak-s-o-nti formed a long-vowel subjunctive with generalized *-ē- (< *-e-e-), i.e. *fak-s-ē-m, *fak-s-ē-s, *fak-s-ē-d, *fak-s-ē-nd etc. Thus, it represented a prospective formation based on the stem of the future tense, which was originally a manner of action formation for the expression of the so-called desiderative with functionally inchoative-inten‐ tional meaning as ‘to be going to do’. Therefore, a form like *fak-s-ē-d as a derived prospective formation (to express expectation) had the functional meaning ‘(s)he is expected to be going to make’ and became the expression of the conditional ‘(under such and such conditions) (s)he would make’. Whereas the underlying future tense formation of the type faxō became restricted in its use by the emergence of new formations to express the future tense (in this case by the future tense paradigm faciam, faciēs, faciet, facient etc. ‘will make’ ← *-ē-m, *-ē-s, *-ē-d, *-ē-nd as formal continuation of the former prospective subjunctive of the present tense; see below § 17.1.), the prospective formation of the type *fak-s-ē-m, as derived from faxō = / fak-s-ō/ , became functionally productive as a conditional in the sense of ‘would make’, i.e. as a verbal category denoting non-reality or imagina‐ tion, in functional contradistinction to the respective indicative present faciō ‘I make’, denoting reality in the view of the speaker. For its un‐ folding productivity as expression of the non-real, however, the condi‐ 256 Reiner Lipp <?page no="257"?> tional aligned not only paradigmatically, but also morphologically with the functionally associated imperfective present stem *fak-i̯ o-/ *fak-i- (syn‐ chronically semi-thematic), which triggered the reshaping from *fak-s-ē-m, *fak-s-ē-s, *fak-s-ē-d, *fak-s-ē-nd etc. into *fak-i-s-ē-m, *fak-i-s-ē-s, *fak-i-s-ē-d, *fak-i-s-ē-nd > *fakizēm, *fakizēs, *fakizēd, *fakizēnd > Lat. facerem, facerēs, faceret, facerent etc. ‘would make’ (for the late date of this inner-Latin remodelling, see above §§ 13.1.2.-3.), thereby following the proportion in stem pattern of other semi-thematic verbs, which select the stem form without thematic vowel as derivational basis of the imperfect subjunctive, like Lat. eō, it = OLat. īt < *e(i̯ )-ō, *ei̯ -ti with imperfect subjunctive īrem, īrēs, īret, īrent < *ei̯ -s-ē-m, *ei̯ -s-ē-s, *ei̯ -s-ē-d, *ei̯ -s-ē-nd, and Lat. volō, vult < *u̯ el-ō, *u̯ el-ti with imperfect subjunctive vellem, vellēs, vellet, vellent < *u̯ el-s-ē-m, *u̯ el-s-ē-s, *u̯ el-s-ē-d, *u̯ el-s-ē-nd. Consequently, one can state that the productive Latin imperfect subjunc‐ tive of the type facerem = */ fak-i-s-ē-m/ shows formal alignment with the imperfective present tense faciō = / fak-i-ō/ , facit = / fak-i-t/ since functionally it represents a conditional belonging to the imperfective present stem. Therefore, as so-called irrealis in the sense ‘would make / do’, it serves to express the non-real (i.e. verbal events happening under certain conditions in the sense of a thought experiment), with reference to the present as well as indistinctively, as still in Old Latin, to the past (in its imperfective aspect; also still in Classical Latin in deliberative function as quid facerem ‘what should I do then? / what should I have done? ’), and furthermore, in subordinate clauses with syntactic dependency on a superordinate verb in a past tense form, it serves to express simultaneity or finality vis-à-vis the superordinate verb (cf. Meiser 1993a: 172-174, 181 f.; de Melo 2007a: 72). In contrast to this alignment of the productive imperfect subjunctive to the imperfective present, the formation of the recessive basic type of the future faxō, with its marked durative-telic, i.e. terminative character (see above § 10.) and its restricted formal applicability (base final l or r being avoided for the assimilation l-s, r-s > ll, rr; Meiser 1998: 184), remained unaffected by such a remodelling. Early enough, in fact since the beginning of the literary transmission, as mentioned above (§§ 8.2., 10.), there were competing formations that became productive as expressions of the future tense with unmarked durative value and were finally prevailing (for the competing future formations, see below §§-17.1.-2.). The Latin future tense formation of the type faxō and its Italic background 257 <?page no="258"?> 15.4.1 This inner-Latin process of alignment of the conditional of the Proto-Italic s-future (continuing the PIE desiderative category) to the pattern of the present stem, which resulted in the formation of the imperfect subjunctive, has a formal parallel in the aforementioned case of the Sabellic s-future and the so-called imperfect subjunctive as the conditional derived from it (see above §§ 15.3.1.-2.). However, in contrast to the Latin imperfect subjunctive, in Sabellic the motivation for this alignment was not based on aspect, since Sabellic did not possess an imperfect category of the type in †-βā- = Lat. -bāthat would be in aspectual distinction to the syncretistic preterite category conventionally labelled “perfect”, which, besides some new formations, consists of original aorist, perfect and imperfect forms (in the latter case displaying a fossilized present stem with preterite endings; for that type of Sabellic preterite forms, see Lipp 2021: 300 f., 305-308). Regarding that the so-called imperfect subjunctive originally represents the subjunctive of the s-future, the trigger for the Sabellic remodelling of this category after the present tense stem is provided by the general Sabellic alignment of the s-future to the paradigmatic reference form of the present tense; cf. pres. / did-e-t/ = Vestinian didet ‘gives’ with future / did-e-s-t/ = Osc. didest ‘will give’ for inherited Proto-Italic *dō-s-ti, and pres. / fer-ent/ ( ← *fer-o-nt = Lat. ferunt) = Marruc. feret ‘they carry away/ receive’ with future / fer-e-s-t/ = Umbr. ferest ‘will carry’ for inherited Proto-Italic *fer-s-ti and, based on the semi-thematic future stem fer-e-s-(o-) ( ← Proto-Italic *fer-s-(o-)), the inner-Sabellic new formation of the so-called imperfect subjunctive *fer-e-s-ẹ̄-nd → Osc. f]erríns ‘they would carry away/ receive’ corresponding to Lat. ferrent < *fer-s-ē-nd (on the formation see below § 15.4.2.; on the Sabellic alignment of the s-future to the present tense, see above §§ 14.2.-3., 14.5.; on the status of the Italic imperfect subjunctive in *-s-ēas a conditional represented by the subjunctive form of the s-future formation as preserved in Sabellic, cf. Meiser 1993a, 1998: 201 f., 2003: 32-34, 39 f.; Rix 2003a: 153 f., 2003b: 5-7; Hill 2004: 121-133, with critique of the analysis as an optative/ subjunctive in *-(i̯ )ēof a preterite stem in *-s-, as per Jasanoff 1991, or as a prospective subjunctive built on a preterite stem in *-se-, as per Stempel 1998: 277-279). 15.4.2 Likewise, in the case of the PIE s-desiderative *peth 2 -sbelonging to the root *peth 2 - ‘to spread out’, which is represented by the Greek future πετῶ < πετάω < *petahō < *peta-s-ō (paradigmatically connected with pres. 258 Reiner Lipp <?page no="259"?> πετάννῡμι, Hom. πίτνημι < PIE *p e t-né-h 2 -; see above § 13.1.3.; on the secondary thematization of the Greek s-future, see above §-5.2.), for Proto-Italic one ex‐ pects a corresponding semi-thematic s-future *peta-s-ō, *peta-s-ti, *peta-s-onti ‘will spread out’ (on the innovative Proto-Italic semi-thematic pattern, see above §§ 9.1.-5., 15.2., 15.3.1.) with a long-vowel subjunctive (showing sec‐ ondarily generalized *-ē-; see above §§ 15.3.1.-2.) in the function of the pertaining conditional, i.e. *peta-s-ē-m, *peta-s-ē-d, *peta-s-ē-nd ‘would spread out’. However, due to the post-Proto-Italic alignment of the conditional with the respective present tense stem that happened in the wake of the extinction of many opaque future forms of the faxō-type as in this case *pessō < *pet(a)sō, Latin shows the attested forms panderem, panderet, panderent ‘would spread out’, which represent subjunctive imperfect formations of the present tense pandō, -ere ‘spread out, extend, open’ with stem pande/ o- < *padne/ o- < *patne/ o-, resulting by medial syllable syncope from Proto-Latino-Faliscan *patan-e/ o- ← Proto-Italic 3.pl. *patan-onti ( ← PIE *p e t-n̥ -h 2 -énti = *[pǝtn̥ h 2 ánti] from the nasal infix present *p e t-né-h 2 -/ *p e t-n̥ -h 2 -, as represented by Homeric πίτνημι ‘spread out’; LIV 2 478f.; Lipp 2009a: 42 n. 102 with refs.); thus, these forms continue the inner-Latin imperfect subjunctive stem *patne-sē- (after medial syllable syncope) > *padne-sē- = *[padnezē-] > *pande-sē- = *[pandezē-] (with secondary *-tnbecoming *-dn- > -nd-; cf. Meiser 1993c: 263 f., 1998: 121 f.), underlying Latin / pande-rē-/ (with intervocalic *-s- = *[z] > r). Correspondingly, in Sabellic there is the imperfect subjunctive Oscan patensíns ‘they would open’, attested on the Cippus Abellanus (= Cm 1B24. 25; late 2 nd c. BC), which, as the morphological equivalent of independently remodelled Latin panderent, represents an original subjunctive formation of a Sabellic s-future *patane-s- > *patne-s-, which is built on the unattested Sabellic present stem *patn-e/ o- (medial syllable syncope) < Proto-Sabellic *patan-e/ o- ← Proto-Italic 3.pl. *patan-onti. Starting for the Oscan imperfect subjunctive formation from the Proto-Sabellic new formation *patane-s-ẹ̄ - = *[patanezẹ̄ -], reflecting the thematic present stem *patan-e/ oas derivational basis of the Proto-Sabellic s-future *patane-s- (instead of representing an original s-future formation built on the root like aforementioned Proto-Italic *peta-sor, in general terms, a verbal base devoid of characteristic present stem markers, as per § 11.1.), one arrives at the inner-Sabellic imperfect subjunctive stem *patne-sẹ̄ - = *[patnezẹ̄ -] (by medial syllable syncope) > *patn̥ -sẹ̄ - = *[patn̥ zẹ̄ -] (successive Sabellic medial syllable syncope with sec‐ ondarily resulting syllabic *n̥ ) > / paten-zẹ-/ (with anaptyxis of secondary *n̥ The Latin future tense formation of the type faxō and its Italic background 259 <?page no="260"?> > en) in Oscan patensíns ‘they would open’ = / paten-zẹ-ns/ (Rix 1995: 405; cf. Meiser 1986: 96, 133, 1993c: 262-264; Lipp 2009a: 42 n. 102), which thus represents the form of the Proto-Sabellic s-future subjunctive *patane-sẹ̄ -nd (= Lat. panderent), introducing regularly the secondary ending -ns ( ← *-n < *-nn < *-nd < PIE *-nt = *[-nd̥ ] = */ -nd/ ). This Sabellic formation in the function of the imperfect subjunctive belongs to the unattested present indicative form *patenent ‘they open’ (with Oscan vowel anaptyxis beside resonant) < *patn-ent (medial syllable syncope) < *patan-ent ← Proto-Italic *patan-onti ( ← PIE *p e t-n̥ -h 2 -énti = *[pǝtn̥ h 2 ánti]). Likewise, also on the Cippus Abellanus (= Cm 1B29), there appears the aforementioned imperfect subjunctive Osc. f]erríns ‘they would carry away/ receive’ (see above § 15.4.1.), which represents a derivationally independent formation from its Latin equivalent ferrent ‘they would carry/ bring’ < *fer-s-ē-nd based on *fer-, corresponding to the synchron‐ ically semi-thematic present stem Lat. *fer-o-/ fer- ( ← *fer-e-) in ferō, fers, fert, ferimus, fertis, ferunt, imptv. fer, ferte, imptv. II fertō (see above § 14.3.), as against subjunctive sererent < *sere-s-ē-nd and *sese-s-ē-nd after the thematic present stems *ser-o-/ -e- ‘to link together’ and *ses-o-/ -e- ‘to sow’ in the homon‐ ymous paradigm forms serō, seris, serit, serimus, seritis, serunt, imptv. sere, serite, imptv. II seritō. For Sabellic, Osc. f]erríns = / ferrẹns/ (medial syllable syncope) < *fere-sẹ̄ -ns = *[ferezẹ̄ ns] presupposes a specific post-Proto-Italic, i.e. Proto-Sabellic, new formation *fere-s-ẹ̄ -nd = *[ferezẹ̄ nd], morphologically representing a subjunctive of the Sabellic s-future *fere-s- = Umbr. ferest, which, as a form remodelled on the present stem *fer-e/ o- = Marruc. feret 3.pl. (Meiser 1998: 224, 2003: 40, 50), substituted the Proto-Italic semi-thematic s-future *fer-s-ti/ *fer-s-o-nti ‘will carry’ (from the root-based desiderative stem *fer-s-) showing the subjunctive *fer-s-ē-d/ *fer-s-ē-nd ‘would carry’ as expres‐ sion of the conditional (with the long-vowel subjunctive being conditioned by the thematic indicative stem *fer-s-o-; on semi-thematic paradigms with long vowel subjunctives, see above §§ 15.3.1.-2.). For morphonological reasons, Osc. f]erríns = / ferrẹns/ clearly represents the Proto-Sabellic new formation *fere-s-ẹ̄ -nd based on the thematic present stem *fer-e-, since with / -rr-/ < *-rs- < *-rVsit shows the effect of medial syllable syncope in its immediate pre-stage *fere-sẹ̄ -ns = *[ferezẹ̄ ns] as against the representation of original *-Vrsby Oscan / -V̅ r-/ in terúm, teer[úm ‘territorium’ = / tēro-/ < *terro- (cf. Lat. terra) < *ters-o- (cf. Buck 1928: 48 f., 76). 15.4.3 The Sabellic evidence of the imperfect subjunctive does not show the forms in independent use, but only embedded in subordinate clauses, where, 260 Reiner Lipp <?page no="261"?> depending on a superordinate verb in the preterite, from the standpoint of the past they express simultaneity or finality according to the consecutio temporum, as is also the case in Latin. So far, there is no evidence in the limited corpus of Sabellic for the use as an irrealis corresponding to one of the main usages in Latin (Meiser 1993a: 172 f. with refs.; Rix 2003b: 5-7). Accordingly, the aforementioned Oscan forms of the imperfect subjunctive attested on the Cippus Abellanus (late 2 nd c. BC), i.e. fusíd ‘foret, esset/ would be’, patensíns ‘aperirent/ they would open’, f]erríns ‘acciperent/ they would receive’, appear in subordinate clause constructions belonging to a superordinate verb of de‐ cree in the preterite, which appears in the prescript of the inscription, i.e. ekss kúmbened […] ‘ita convenit: […]/ thus it was decreed that […]’, with the rest of the text being subordinate to this formula (for details, cf. Meiser 1993a: 172 with n. 19, Clackson 2020: 57 f., 61; for the text of the Cippus Abellanus and its interpretation, cf. Crawford 2011: 887-892). For this reason, one occasionally observes a formal alignment of an imperfect subjunctive form with that of the paradigmatically correlated preterite. This is the case with the Paelignian form upsaseter (Pg 2) = / ūpsasẹtẹr/ ‘ut (templum) exstrueretur/ that the temple would be built’ < *ọ̄ pesā-sẹ̄ -tįr ( ← *opesā-s-), which is syntactically dependent on the preterite coisatens ‘curaverunt/ they took care, saw to it’ in the main clause (for text and interpretation of the inscription from late 2 nd c. BC, cf. Vetter 1953: 151 f. = inscr. 216; Crawford 2011: 247 = inscr. Superaequum 4; Clackson 2020: 58 f.; for the morphonological analysis, cf. Meiser 1996: 206 with n. 65; Zair 2016: 307 n. 25; Lipp 2021: 310). Clackson (2020: 72-74) considers this Paelignian and the aforementioned Oscan examples of imperfect subjunctives with the suffix *-sēas late calques of the corresponding Latin formations foret, ferrent, operarētur and, as to its structural pattern, panderent; this seems however excluded in view of the specific morphonological shape of the Sabellic forms in question, which display the effect of the medial syllable syncope of the late 6 th century BC and whose formation must go back to the Proto- Sabellic period, as outlined above (in § 15.4.2.). Besides the effect of syncope, the Paelignian imperfect subjunctive form upsaseter = / ūpsasẹtẹr/ shows the specific influence of the paradigmatically assigned preterite, as attested by Oscan 3.sg. Osc. upsed Sa 18.33, upse[d Hi 8, ups(e)d Sa 34, ups(ed) Sa 32 = / ūps-ed/ ‘fecit / (has) made’, 3.pl. uupsens Po 1 f., upsens Po 8, ουπσενσ Me 1.2.3 = / ūps-ens/ ‘fecerunt/ (have) made’ (Oscan ū < Sabellic *ọ̄ in initial syl‐ lables). The associated preterite forms represent an analogical lengthened grade perfect that appears as a hyper-characterized variant alongside the The Latin future tense formation of the type faxō and its Italic background 261 <?page no="262"?> unmarked basic form of the preterite with short stem initial vowel as in 3.sg. Osc. úpsed = / ops-ed/ and 3.pl. Vestinian ośens = / oss-ens/ (< *ops-ens), Umbr. opset = / ops-ent/ (for / -ens/ ). These preterite forms betray the effect of the Sabellic medial syllable syncope (with structurally older -ps- < *-pVsas against assimilated -ss- < *-ps-) and as such represent a remodelling with regularized thematic preterite endings on the basis of the original imper‐ fect forms 3.sg. *opesād, 3.pl. *opesā̆ nd, which display the denominative present stem *opesā- < *opes-ā-i̯ e/ o- (derived from *op-os/ -es- > Lat. opus, -eris ntr. ‘work’), as attested by Osc. úpsannam = / opsa-nnam/ ‘(domum) faciendam/ (house) to be built’ < *opesā-dn-ā-, Umbr. osatu = / ossa-tu/ ‘facito/ shall make’ < *opesā-tōd (Lipp 2021: 301, 308-310, 312-314). This Sabellic formation is probably older than semantically different and late attested Lat. operor, -ārī ‘to be busy, to be at work (itr.)’ (Plinius+), which is assumed to be back-formed from earlier attested operātus ‘busy, occupied’ (Lucilius, Vergilius+) as direct derivative of the abstract opera ‘effort, activity’ < *opes-ā (Postgate 1999, followed by de Vaan 2008: 432, Weiss 2009: 269). 16.1 In view of the fact that the Latin imperfect subjunctive of the type represented by facerem ‘would make’ < *fakizēm < *fak-i-s-ē-m is built on the stem of the present tense, i.e. on *fak-ias in the present forms Lat. faciō, facit = / fak-i-ō/ , / fak-i-t/ , an apparent alternative would be offered by the assumption that, just like the form of the Latin imperfect subjunctive, the Old Latin future tense formation of the type of faxō ‘will make’, faxim ‘may make’ is a formation built on the present stem and represents a syncopated form of *fakizō, *fakizīm < *fak-i-s-ō, *fak-i-s-ī-m, in accordance to the syncope in the phonotactically resembling OLat. s-future formation noxit < *nokezīd < *nok-e-s-ī-d, which is built on the verbal base *nok-eresulting from the subtraction of the present stem suffix *-i̯ e/ ofrom the stem form *nok-e-i̯ e/ o- (< PIE causative *noḱ-éi̯ e/ o-) as underlying Latin noceō, -ēre ‘to harm, hinder’. Correspondingly, just like the Latin imperfect subjunctive agerem < *agezēm < *ag-e-s-ē-m, belonging to the present tense agit ‘drives, acts, does’ < *ag-e-t with stem *ag-e-, the s-future subjunctive formation āxim could then be derived from *agezīm < *ag-e-s-ī-m, comparable to South Picene aisis ‘you might do’ < *aɟezīs < *ag-e-s-ī-s (see above §§ 8.2., 14.6.). Under this premise, it looks as if the formation of the Old Latin future type faxō would correspond to the Sabellic s-future formation, which is morphologically built on the present stem, and could therefore even continue a formation structurally 262 Reiner Lipp <?page no="263"?> inherited from Proto-Italic. On closer inspection, however, this impression proves to be inaccurate (see below §-16.3.). 16.2 In this context, it is indeed possible to point out that the Sabellic remodelling of the inherited root-based s-future formation represented by Osc. fust, Umbr. fust, fust ‘will be’ (from the root fu- ← *fū-) can be clearly recognized in the Sabellic relation between pres. 3.sg. *did-e-t ‘gives’ > Vest. didet (synchronic present stem / did-e-/ ) and fut. 3.sg. *did-e-s-t > Osc. didest = / did-es-t/ , with the present tense stem as the renewed basis for the s-future (instead of Proto-Italic *dō-s-ti); likewise, pres. 3.sg. *fer-e-t ‘brings’, 3.pl. (*fer-o-nt = Lat. ferunt → ) *fer-ent ‘carry, bring’ > 3.sg. *fer-e-t or remodelled *fer-t (= Lat. fert analogous to the allegro forms of the imptv. Lat. fer < *fer-e with apocope, Lat. ferte < *fer-e-te with premature syncope), 3.pl. Marruc. feret = / fer-ent/ (synchronic present stem / fer-e-/ or / fer-/ ) and fut. *fer-e-s-t > Umbr. ferest = / fer-es-t/ (instead of Proto-Italic *fer-s-ti); pres. 3.sg. *her-i-t ‘wishes, desires’ (synchronic present stem *her-ias in Umbr. herter ‘is desired, necessary’ < *her-i-tir) and fut. *her-i-s-t → *her-i-es-t > Umbr. heriest, Osc. herest (instead of Proto-Italic *χer-s-ti; for the morphological resegmentation *-e-s- → *-esas future tense marker in Sabellic, see above §-14.3.). 16.3 For Latin, however, one has to consider that the stem *nok-e-sun‐ derlying the Old Lat. s-future formation noxit < *nok-e-s-ī-d is not based on the present stem / nok-ē-/ -eo-/ < *nok-e-i̯ e/ oas represented in Lat. noceō, -ēre ‘harm’, but, as explained above (§§ 13.2.1., 13.5.), it represents a secondary formation built on the verbal base *nok-e-, which came about by subtraction of the suffix *-i̯ e/ ocharacterizing the present stem. Correspondingly, the Old Latin s-future form rupsit is not based on the stem of the corresponding present tense form rumpō, -ere ‘break’, characterized by the nasal infix, but, as an inherited desiderative formation, on the root (see above §-13.3.1.). Thus, since the formation of the Latin s-future is generally not based on a stem containing a suffix characteristic of a present stem, but on the root or a de-characterized verbal base that does not contain a suffix specific to a present tense stem (see above §-11.1.), the Old Latin future forms faxō, faxim and āxim cannot go back to †fak-i-s-ō, †fak-i-s-ī-m and †ag-e-s-ī-m and do not have as derivational bases the present tense stems *fak-iand *ag-e-, as appearing in pres. facit < *fak-i-t and agit < *ag-e-t. Instead, forms of the type OLat. faxō, faxim and OLat. āxim originate from Proto-Italic formations of the structure *fak-s-ō, *fak-s-ī-m and *ag-s-ī-m. The Latin future tense formation of the type faxō and its Italic background 263 <?page no="264"?> 16.4 Therefore, these s-future forms go back to Proto-Italic semi-the‐ matic indicative forms that were remodelled from an athematic paradigm type, e.g. 1.sg. *fak-s-ō (thematic), 2.sg. *fak-s-si, 3.sg. *fak-s-ti (athematic), 3.pl. *fak-s-o-nti (thematic) ‘will make’, and to inherited optative forms as *fak-s-i̯ ē-m, *fak-s-i̯ ē-s, *fak-s-i̯ ē-d, *fak-s-ī-nd ‘may make’, formed with the apophonic suffix *-i̯ éh 1 -/ *-ih 1 - > *-i̯ ē-/ *-īon the basis of the athematic s-stem of the desiderative (being a manner of action formation of the mobile accent class; see above § 2.) and later levelled to *fak-s-ī-m, *fak-s-ī-s, *fak-s-ī-d, *fak-s-ī-nd (on the general levelling of the original optative suffix *-i̯ ē-/ *-īto *-īexcept in the copula, with monosyllabic forms such as *s-i̯ ē-d > Palaeo-Lat. sied, OLat. siet being occasionally kept until the Old Latin period, see above § 9.5.). Paradigmatically associated was an innovative subjunctive formation derived from the thematic basis of the Proto-Italic semi-thematic paradigm, which therefore surfaced as a long-vowel subjunctive (with generalized *-ē-), i.e. *fak-s-ē-m ( ← *fak-s-ō), *fak-s-ē-s, *fak-s-ē-d, *fak-s-ē-nd ( ← *fak-s-ō-nd) ‘would make’. This means that the forms of the imperfect subjunctive of the type Lat. facerem ‘would make’ < *fak-i-s-ē-m and agerem ‘would drive/ act/ do’ < *ag-e-s-ē-m, whose precursors on the Proto-Italic stage were derivationally connected with the faxō-type (‘will make’) as pertaining subjunctives in the function of a conditional (‘would make’), are not inherited in their attested structure but, on the contrary, represent post-Proto-Italic transformations which came about morphologically under the analogical influence of the paradigmatically associated imperfective present stems such as *fak-iand *ag-e- (see above §-15.3.2.). 16.5 The innovative character of this productive pattern also becomes structurally evident from the fact that in Latin, on the model of the athematic present of the copula est ‘is’ = / es-t/ with sum ‘I am’ = Palaeo-Lat. esom (Garigliano bowl from around 500 BC; on the inscription Vine 1998; Rix 2003a: 161 f.; Hartmann 2005: 147-153 with further refs.), OLat. esum (acc. to Varro, Ling. 9, 100) and es ‘you are’ = OLat. es = / es-s/ ~ / es/ (< 1.sg. *es-m̥ with secondary *-m̥ > -om due to i-apocope, 2.sg. *es-s, 3.sg. *es-t < *es-mi, *es-si, *es-ti; Joseph/ Wallace 1987: 687-689, Meiser 1998: 73, 221, Rix 2003a: 154 f.), there arose the new formation of the imperfect subjunctive Lat. essem, essēs, esset ‘would be’ < *es-s-ē-m, *es-s-ē-s, *es-s-ē-d. Due to its productive pattern of alignment to the present tense stem, this formation pushed back the use of the ancient form of the imperfect subjunctive Lat. forem, forēs, foret ‘would be’ < *fu-s-ē-m = *[fuzēm], *fu-s-ēs = *[fuzēs], *fu-s-ē-d 264 Reiner Lipp <?page no="265"?> = *[fuzēd], which structurally continues a subjunctive formation of the Proto-Italic s-future. This root-based formation corresponds exactly to the Sabellic form *fu-s-ẹ̄-d = [fuzẹ̄ d] < *fu-s-ē-d, as represented by the imperfect subjunctive Osc. fusíd. Both the Latin and the Oscan forms show a probably post-Proto-Italic remodelling of the Proto-Italic form *fū-s-ē-d by introducing the root allomorph *fu- (with short *u) after the pre-vocalic root variant *fuu̯ - < *fū- (OLat. = Lat. fuit ← thematic aorist *fuu̯ -e-d vs. OLat. fūit after fūimus ← perfect *fu-fū-mos < *fe-fū-mos) and the reduplicated perfect stem *fu-fuu̯ - < *fe-fū- (Osc. fufens < *fu-fuu̯ -ens ← *fu-fuu̯ -en(d) ← *fu-fuu̯ -ond = Pre-Samnite fυfυϝoδ < *fe-fuu̯ -ond). Proto-Italic *fū-s-ē-d is genetically a subjunctive formation in the function of a conditional that belongs to the semi-thematic Proto-Italic s-future with 3.sg. *fū-s-ti (athematic), 3.pl. *fū-s-o-nti (thematic) ‘will become, will be’, which is continued by Osc. fust, fusent, Umbr. fust, fust, furent < Proto-Sabellic *fu-s-t, *fu-s-ent ~ *fu-s-ont (with *fufor *fūand regularized inner-Sabellic 3.pl. in *-ent for *-ont, as against preserved *-ond in the preterite Pre-Samnite fυfυϝoδ = / fu-fuu̯ -ond/ ; on this s-future formation, see above §§-15.2., 15.3.1.). 16.6 Beside the archaic root-based subjunctive formation of Lat. foret < *fu-s-ē-d = *[fuzēd] (= Osc. fusíd ← Proto-Italic *fū-s-ē-d, belonging to future indicative *fū-s-ti), the original morphological pattern of the later imperfect subjunctive is continued by the semi-thematic paradigms with preserved athematic present indicative forms, as is the case with the paradigms of Lat. eō (thematic), īs, it, OLat. īt (athematic) ‘go’, Lat. edō (thematic), ēs, ēst (athematic) ‘eat’, Lat. volō (thematic), vīs, vult (athematic) ‘want’, showing the root-based imperfect subjunctive forms Lat. īrem, īrēs, īret < *ei̯ -s-ē-m, *ei̯ -s-ē-s, *ei̯ -s-ē-d ‘would go’ (as original subjunctive form based on the Proto-Italic s-future *ei̯ -s-ti ‘will go’ > Umbr. eest, est), Lat. ēssēm, com-ēssēs, -ēsset < *ed-s-ē-m, *ed-s-ē-s, *ed-s-ē-d ‘would eat’, Lat. vellem, vellēs, vellet < *u̯ el-s-ē-m, *u̯ el-s-ē-s, *u̯ el-s-ē-d ‘would want’ (on the formation of the imperfect subjunctive in Latin and Sabellic, cf. Meiser 1993a: 174 ff., 1998: 201 f., who, however, regards its formation on the basis of the characterized stem of the present tense as a trait inherited from Proto-Italic; on the respective semi-thematic paradigms, see above §§ 9.1.-3.). However, in the Latino-Faliscan semi-thematic paradigms with present stem suffix *-i̯ o-/ *-i- (III. conjugation) and *-ii̯ o-/ *-ī- (IV. conjugation), the stem form lacking the thematic vowel *-e-, i.e. the stem containing the suffix form *-ior *-ī- (going back to pre-Proto-Italic *-i̯ eor *-ii̯ eby vowel assimilation or reduction in The Latin future tense formation of the type faxō and its Italic background 265 <?page no="266"?> unstressed open syllable; for the suffix formation, see below § 17.1.), was secondarily introduced as derivational basis of the imperfect subjunctive, which therefore is also formed from a synchronically athematic stem. Thus, there is the conditional Proto-Italic *fak-s-ē-d ‘would make’ as subjunctive of the Proto-Italic semi-thematic s-future stem *fak-s-o-/ *fak-s- ‘will make’, which is paradigmatically associated with the present stem *fak-i̯ o-/ -i- ← *fak-i̯ o-/ -i̯ e- (on the formation of the Proto-Italic s-future from a verbal base without specific present stem markers such as *-i̯ o-/ -i̯ e-, see above §§ 11.1., 13.5., 14.1.), triggering the Latino-Faliscan remodelling to the imperfect subjunctive form *fak-i-s-ē-d > Lat. faceret as a form congruent to the synchronically athematic present stem *fak-iin Lat. facit ‘makes’. Likewise, there is the Proto-Italic conditional *seru̯ -i-s-ē-d ‘would guard’ as subjunctive of the Proto-Italic semi-thematic s-future stem *seru̯ -i-so-/ *seru̯ -i-s- ‘will guard’ from a stem base *seru̯ -i-, as abstracted from the paradigmatically associated present stem *seru̯ -i-i̯ o-/ -i̯ eby re-segmentation of *seru̯ -ii̯ o-/ -ii̯ e- > *seru̯ -ii̯ o-/ -ī- (with preserved glide *i̯ after i before *o as non-homorganic vowel and with vowel assimilation in the unstressed palatal sequence *-ii̯ e- > *-ii̯ i- > *-ī-; form resulting by Sievers effect from *seru̯ -i̯ o-/ -i̯ e- ‘to act as guard, to guard’ as denominative of *ser-u̯ o- ‘guard’ > Lat. servus ‘servant, slave’; see below § 17.1.), which triggered the Latino-Faliscan remodelling to the imperfect subjunctive form *seru̯ -ī-s-ē-d > Lat. servīret ‘would serve’ as a form congruent to the synchronically athematic present stem *seru̯ -īin OLat. servīt, Lat. servit ‘serves’ ( ← *‘guards, is observant’ with semantic shift under the influence of servus ‘servant, slave’). 17. Counter to Sabellic, which continues the s-future (originally based on the root as Osc. Umbr. fust = / fu-s-t/ ‘will be’ ← *fīst < Proto-Italic *fū-s-ti) with regular secondary remodelling after the present stem (Oscan didest ‘will give’ ← Proto-Italic *dō-s-ti after present stem dide/ o-, as attested by Vesti‐ nian didet ‘gives’; see above § 15.1.), in Latin the corresponding future tense category of the type faxō, -is, -it (< *fak-s-ō, *fak-s-e-s, *fak-s-e-t ← Proto-Italic *fak-s-ō, fak-s-si, fak-s-ti) was pushed back in its use for functional and formal reasons by two innovatory future tense formations (as hinted at above §§-8.2., 10., 15.3.2. and described in the following). 17.1 On the one hand, there is the original prospective subjunctive in -ē- (with 1.sg. in *-ē-m being replaced by secondary forms in -a-m < *-ā-m from the later productive subjunctive in -ā-), which in Latin took over the future function in the III. conjugation showing present stems in -i-/ -u- < 266 Reiner Lipp <?page no="267"?> *-e-/ -o- (thematic), cf. Lat. 3.sg. agit/ pl. agunt with future tense aget/ agent ← *-ē-d/ *-ē-nd, and present stems in -i-/ -iu- < *-i-/ -i̯ o- (semi-thematic, remodelled from *-i̯ e-/ -i̯ o-), cf. Lat. 3.sg. facit/ pl. faciunt with future tense faciet/ facient ← *-ē-d/ *-ē-nd, and the IV. conjugation showing present stems in -ī-/ -iu- < *-ī-/ -ii̯ o- (semi-thematic, remodelled from *-ii̯ e-/ -ii̯ o-), cf. Lat. 3.sg. servit, OLat. servīt/ pl. serviunt with future tense serviet/ servient ← *-ē-d/ *-ē-nd (on the stem formations of the III. and IV. conjugations, cf. Meiser 1998: 191-196; Weiss 2009: 399 f., 404-408; on the associated future formation Meiser 1998: 199; Weiss 2009: 415). For the future tense in Old Latin, there are also attestations of the morphologically inherited 1.sg. in -e(m) < *-ē-m as transmitted by grammarians and manuscripts, e.g. dīce and facie for dīcam and faciam belonging to present dīcō, -ere and faciō, -ere (evidence provided by Churchill 2000), which formally correspond to the South Picene subjunctive form in CH 2 putih knúksem = / putih gnọ̄ skem/ ‘uti gnoscam’ < *gnōskēm (see above § 15.3.1.). As mentioned above, as precursor of the productive future tense formation, the original long vowel subjunctive in *-ē- (= *-e-e-) also appears in the paradigms of the III. and IV. conjugations that are characterized by the synchronically semi-thematic present stem suffixes *-i-/ -i̯ oand *-ī-/ -ii̯ o-, which don’t show an overt reflex of the thematic vowel in the shape *-e-. This phenomenon is due to the fact that in Proto-Italic these suffix forms resulted from a reduction of the original suffix forms *-i̯ e-/ -i̯ oand *-ii̯ e-/ -ii̯ o-, which contained the regular thematic vowel *-ebe‐ side *-o- (cf. the well-structured consideration of different morphonological approaches by Buzássyová 2012: 8 with nn. 21-23 for further refs.). In this respect, one has to keep in mind that the Sabellic equivalent of Lat. pres. 3.sg. facit ‘makes’ shows the regular reflex of the inherited Proto-Italic short vowel */ i/ by Sabellic open i = *[į] > / ẹ/ (close e), as attested by the Oscan forms in Lu 13 (Tricarico/ Lucania; 300-250 BC) αfα̣κειτ = / am-fakẹt/ ‘makes/ offers a sacrifice’ (<ει> = / ẹ/ ) and Lu 18 (helmet/ Lucania; 375-350 BC) ανα: fακετ ‘id.’ = / an a -fakẹt/ (with secondary Oscan vowel anaptyxis; <ε> = / e/ , / ẹ/ in old orthography) < Sabellic *an(o)-fakįt < Proto-Italic *fak-i-ti (LIV 2 139 s.v. *d h eh 1 kwith n. 1b for further refs.). This fact excludes a derivation of the form by generalization of the result of the post-Proto-Italic post-resonant vowel syncope before final *-s (as in the 2.sg.) within the individual histories of Sabellic and Latin (pace Meiser 1998: 195), which would have given Sabellic 2.sg. †fakis < †fak-i̯ e-s with close i = [ị] > Osc.Umbr. / i/ resulting from the secondarily syllabified glide *i̯ , as in Osc. lúvkis, λωϝκισ = / lou̯ kis/ , Umbr. The Latin future tense formation of the type faxō and its Italic background 267 <?page no="268"?> vuvçis = / u̯ ou̯ šis/ < Proto-Sabellic *lou̯ kis = *[lou̯ kịs] < Proto-Italic *lou̯ k-i̯ o-s = Lat. Lūcius (cf. Meiser 1986: 59 f.). Thus, the respective semi-thematic suffix shape of the III. conjugation might have come about via Proto-Italic assimilation in the sequence *-C-i- < *-C-i̯ i- < *-C-i̯ edisplaying an unaccented open syllable within the initial stress system, as e.g. in 3.sg. *-C-i̯ e-ti > *-C-i̯ i-ti > *-C-i-ti (a development not contradicted by any inferable evidence from the Proto-Italic level, taking into account the earlier loss of intervocalic *i̯ in Proto-Italic *-V-e- < *-V-i̯ eas in Lat. amat, monet < *amāt, *monēt < Proto-Italic *amaeti, *mon‐ eeti < *ama-i̯ e-ti, *mon-ei̯ e-ti as well as the typologically comparable later forms as Lat. facimus < *fak-i̯ i-mos < *fak-i̯ o-mos with vowel weakening as against analogical societās, -tātis ‘fellowship’ resulting by dissimilation from *soki̯ i-tāti- < *sok u̯ (i)i̯ o-tātiwith suffix retention after the base word socius ‘fellow’ < *sok-i̯ o- < *sok u̯ -(i)i̯ o-). According to Martzloff (2006a: 295 f., 297 f., 300-304, 305-307), this morphonological change is due to the analogical influence from the much more frequent formations of the IV. conjugation, showing *-C-ī- < *-C-ii̯ i- < *-C-ii̯ eby assimilation of *e in open syllable to preceding *i within unaccented position of the sequence, e.g. 3.sg. *-C-ii̯ e-ti > *-C-ii̯ i-ti > *-C-ī-ti. As an alternative, Weiss (2009: 122, 468; 2012: 152 f.) postulates an early pre-Proto-Italic phonological reduction in the III. and IV. conjugations alike, showing *-C-i- (III. conj.) and *-C-ī- (IV. conj.) ← *-C-i̯ eand *-C-ii̯ eby loss of *e in open syllable after *i̯ , comparable to voc. Lat. fīlī ‘o son! ’ < *fēl-ii̯ e. Cf. III. conjugation: Lat. facit ‘makes’, imptv. II facitō = Osc. αfα̣κειτ = / am-fakẹt/ ‘makes/ offers a sacrifice’ (<ει> = / ẹ/ < Proto-Italic *i), Osc. imptv. II factud (with Sabellic medial syllable syncope) < Proto-Italic *fak-i-ti, *fak-i-tōd ← *fak-i̯ e-ti, *fak-i̯ e-tōd with original stem paradigm *fak-i̯ e-/ -i̯ o- (LIV 2 139 f. s.v. *d h eh 1 kwith nn. 1b, 2; see above); IV. conjugation: OLat. servīt, Lat. servit ‘serves’, imptv. II servītō = Umbr. imptv. II seritu, seritu ‘guard, look after! ’ < Proto-Italic *seru̯ -ī-ti, *seru̯ -ī-tōd ← *seru̯ -ii̯ e-ti, *seru̯ -ii̯ e-tōd with original stem paradigm *seru̯ -ii̯ e-/ -ii̯ o- < *seru̯ -i̯ e-/ -i̯ o-, showing the effect of Sievers’ Law after heavy syllable (*seru̯ -i̯ o-/ -i̯ e- ‘to act as guard, to guard’ as denominative of *ser-u̯ o- ‘guard’ > Lat. servus ‘servant, slave’; cf. Rix 1994: 71-81, with objections by Martzloff 2006a: 633-635; see above § 16.6.) (on the representation of the original presents in *-i̯ e/ owithin Sabellic and its prehistory, cf. García Castillero 2000: 187-208; differently Schrijver 1991: 411 with elaboration in 2003, starting for the Italic precursor of the Latin III. conjugation in -iō from athematic i-presents 268 Reiner Lipp <?page no="269"?> with regular thematization in 1.sg. and 3.pl. and for the IV. conjugation in -iō from thematic presents in *-i̯ e/ o-). This means that in Latin there is the present agō, agis, agit ‘drive, act, do’ < *-ō, *-e-s, *-e-t with the productive future agam, agēs, aget from the original thematic subjunctive in *-ē-m, *-ē-s, *-ē-d, likewise the present faciō, facis, facit ‘make’ < *-i̯ ō, *-i-s, *-i-t with the productive future faciam, faciēs, faciet ← *-i̯ ē-m, *-i̯ ē-s, *-i̯ ē-d (based on the original suffix shape *-i̯ e-), the present serviō, -īs, -it, OLat. -īt ‘serve’ < *-ii̯ ō, *-ī-s, *-ī-t with the productive future serviam, serviēs, serviet ← *-ii̯ ē-m, *-ii̯ ē-s, *-ii̯ ē-d (based on the original suffix shape *-(i)i̯ e-). 17.2 On the other hand, Latin shows the originally periphrastic future tense formation of the type carēbō, -bis, -bit ‘will lack’ < *-bō, *-bes, *-bet, continuing *-bu̯ ō, *-bu̯ es, *-bu̯ ed < *-βou̯ -ō, *-βou̯ -e-s, *-βou̯ -e-d with medial syllable syncope (pres. Lat. careō, -ēre ‘lack, be without’ < *kaz-ēe/ o- = Osc. kasit ‘is necessary’ < *kaz-ẹ̄ -t < essive *kas-ē(i̯ )e/ o- ‘to be lacking’ < *-eh 1 -i̯ e/ o-, related to Lat. castus ‘free from, pure, chaste’; LIV 2 329; de Vaan 2008: 92 f.). This formation is also presupposed for Faliscan, where, in direct correspond‐ ence to Latin carēbō, there appear Middle Faliscan future tense forms from the 4 th century BC such as carefo MF 59, MF 60 ‘I will lack, be without (wine)’ = / karēfō/ = [karēβō], showing the termination -βō < *-βu̯ ō < *-βou̯ -ō with preserved voiced fricative [-β-] as allophone of / f/ between sonorous segments in the interior of a word, and with effect of vowel syncope in medial syllable (for which, according to Bakkum 2009: 101, 102 f., due to the scarcity of the material, there are only rare attestations in Middle Faliscan, cf. maximo MF 89 [with weakening -i- < -o-], max]om[o] MF 90, Ṃ axo Ṃ [o] MF 98 = / māksomo-/ < *magsomo- = Lat. māximus, māxumus < *magsǝmo- [vowel weakening] < *magsomo-, resulting by medial syllable syncope from the Latino-Faliscan superlative *mag-isomo- < Proto-Italic *mag-isǝmo- < PIE *-is-m̥ h 2 o-; on the phonological representation of *m̥ (H)V and the superlative suffix in the Italic languages, cf. Lipp 2009b: 130-132 with n. 250, 2021: 294 with further refs.). For the future formation in question, Middle Faliscan offers two further examples from the same inscriptions: pipafo MF 59 ‘I will drink’ = / pibāfō/ = [pibāβō] or / bibāfō/ = [bibāβō] (perhaps from a verbal intensive formation *pib-ābeside thematic *pib-e/ o- > *bibe/ o- > Lat. bibō, -ere ‘drink’, cf. Lat. bibāx ‘fond of drinking’, cognomen Bibāculus < *bib-ā-k- ← *-āaccording to Leumann 1977: 376; Weiss 2009: 305; further the Latin doublets ē-ducō, -āre ‘bring up, educate’ [with short u]: ē-dūcō, -ere ‘pull out, lead out’ with the simplex dūcō, -ere ‘pull, lead’; on this derivational The Latin future tense formation of the type faxō and its Italic background 269 <?page no="270"?> type Meiser 1998: 189 with refs.), MF 60 pafo ‘I will drink’ = / bafō/ = [baβō] or / pafō/ = [paβō] (perhaps with short / a/ like fut. Lat. dabō belonging to dare ‘to give’ with pres. sg. dā-/ pl. dain dō < *dāō, dās, dat < *dāt/ damus, date, dant; imptv. sg. dā/ pl. date ← aor. imptv. *dō/ *da-te in OLat. cedo/ cette ‘give it here! ’ < PIE *déh 3 / *dh 3 -té; perhaps likewise imptv. sg. *pā/ pl. *pate ← aor. imptv. *pō/ *pa-te < PIE *péh 3 / *ph 3 -té; cf. Leumann 1977: 577 f.; Meiser 2003: 105 f.) (on the Faliscan future forms Bakkum 2009: 101, 156, 157, 161, 169 f., 434 f. with further refs.; especially on the question of the stem base in -āor -aunderlying pipafo and pafo, cf. the discussion by Lejeune 1990: 29 f.; Bakkum 2009: 101, 169 f. and the different views by Giacomelli 1963: 152, 238 f.; Leumann 1977: 577 f.; Steinbauer 1989: 236 n. 8; Schrijver 1991: 412 f.; Seldeslachts 2002; Dupraz 2006; Hadas-Lebel 2011: 164). This future tense formation is based on a periphrastic construction of the present participle and the subjunctive form of the copula. The morpho‐ logical shape of the copula *-βou̯ -e-d corresponds to present Ved. bhávati ‘becomes, is becoming’, resulting from an early grammaticalization of the aorist subjunctive *b h éu̯ h 2 -e-ti, *b h éu̯ h 2 -e-t [-d̥ ] ‘will become’, which shows Schwebeablaut of the root morpheme *b h u̯ eh 2 - = */ b h u̯ ah 2 -/ in the position before its vowel-initial subjunctive suffix *-e/ o-. The original shape of the root is still reflected by Latin nē fuās, nē fuat ‘shall not be’ as continuation of the aorist injunctive *nē fuu̯ ās, *nē fuu̯ ād (with effect of Lex Lindeman after heavy syllable) from the PIE prohibitive construction *méh 1 b h u̯ ah 2 -s, *méh 1 b h u̯ ah 2 -d̥ < *b h u̯ éh 2 -s, *b h u̯ éh 2 -t with the unmarked set of the so-called secondary endings *-s, *-t = *[-d̥ ] = / -d/ (on the respective verbal root forma‐ tions, see above §§ 15.1., 15.2., 16.5.; on the root shape *b h u̯ eh 2 and its Schwebeablaut variant *b h eu̯ h 2 -, see LIV 2 98-101 nn. 1, 2, 5, 11, 24, 29, 31, 33, in detail Rix 2003c; Neri s.v. Budike in Neri/ Ziegler 2012: 34, 36-38). In Italic the aorist became part of a syncretistic preterite category without distinctive aspect function, as presupposed by the Latin imperfect formation of the type agēbat ‘(s)he drove/ was driving’ < *agenz-βu̯ ā-d < *ag-en(t)-s = *[agenz̥ ] + *fu̯ ād = *[βu̯ ād̥ ] ‘(s)he was driving (forward)’ (with / f/ = [β] in close connection like in the interior of a word), representing an original progressive formation against the background of an aspectually unspeci‐ fied preterite category (Lipp 2021: 306-308 with refs.). The type of the corresponding future formation in *-bō, *-bes, *-bet < *-bu̯ ō, *-bu̯ es, *-bu̯ ed < *-βou̯ -ō, *-βou̯ -e-s, *-βou̯ -e-d was introduced into the I. conjugation with present stems in -ā- < *-ai̯ e-/ -ai̯ oor *-āi̯ e-/ -āi̯ olike Lat. domō, domās, domat 270 Reiner Lipp <?page no="271"?> ‘tame’ < *-aō, *-ā-s, *-ā-t < *-ai̯ ō, *-ai̯ e-si, *-ai̯ e-ti (with *domai̯ e/ o- < PIE iterative */ domh 2 -ái̯ e/ o-/ = *domh 2 -éi̯ e/ o-) and Lat. dōnō, donās, dōnat ‘give as a gift’ < *-āō, *-ā-s, *-ā-t < *-āi̯ ō, *-āi̯ e-si, *-āi̯ e-ti (with *dōn-ā-i̯ e/ o- < PIE */ -ah 2 -i̯ e/ o-/ = *-eh 2 -i̯ e/ oas denominative of the thematic noun *dōno- > Lat. dōnum ‘gift’), displaying the regular future tense formations domābō, -bis, -bit and dōnābō, -bis, -bit. In this conjugation the inherited thematic subjunctive formation in *-ē- (= *-e-e-) was kept as functional subjunctive of the present tense without shift to the use as expression of the future tense as in the III. and IV. conjugations, since the new subjunctive in -āof the type agat ← *-ā-d belonging to present agit < *-e-t would not have been distinctive against the indicative forms of the I. conjugation as domat, dōnat, leading to a consistent functional preservation of the inherited present subjunctive forms domet ‘shall tame’, dōnet ‘shall give’ ← *domēd, *dōnēd < *domai̯ ēd, *dōnāi̯ ēd. The same morphological innovation in the future tense formation as in the I. conjugation was introduced into the II. conjugation with present stems in -ē- < *-ei̯ e-/ -ei̯ oor *-ēi̯ e-/ -ēi̯ oas Lat. moneō, monēs, monet ‘remind, admonish’ < *-eō, *-ē-s, *-ē-t < *-ei̯ ō, *-ei̯ e-si, *-ei̯ e-ti (with *mon-ei̯ e/ o- < PIE causative *mon-éi̯ e/ o- ‘to make think (of), remind’) and Lat. careō, carēs, caret < *-ēō, *-ē-s, *-ē-t < *-ēi̯ ō, *-ēi̯ e-si, *-ēi̯ e-ti (with *kas-ē-i̯ e/ o- < *-eh 1 -i̯ e/ o- ← PIE essive *ḱ e s-h 1 i̯ é/ óafter fientive aor. in *-éh 1 -/ *-h 1 -, as represented by the Greek intr. aor. in -η-), displaying the regular future tense formations monēbō, -bis, -bit and carēbō, -bis, bit (on the stem formations of the I. and II. conjugations, cf. Meiser 1998: 186-190; Weiss 2009: 399- 404; on the associated future formation Meiser 1998: 199 f.; Weiss 2009: 415). In these paradigms the original thematic subjunctive in *-ē- (= *-e-e-) as *mon-ei̯ ē-d > *monēd → †monet and *kas-ēi̯ ē-d > *kazēd → †karet had become indistinguishable from the present indicative monet < *monēt < *mon-ei̯ e-ti and caret < *kazēt < *kas-ēi̯ e-ti and was therefore not available as a distinct prospective formation to express the future. Because of this formal coincidence, the original present subjunctive of the II. conjugation was substituted in the inherited function of the voluntative and the eventualis by the innovative ā-subjunctive of the form moneam, moneās, moneat ‘shall remind’, caream, careās, careat ‘shall lack’. 18.1 In conclusion, it can be stated that the Old Latin s-future of the type faxō, faxis, faxit, faximus*, faxitis, faxint* ( ← *faxunt) ‘will make/ arrange/ achieve’, which is functionally marked as durative-telic, i.e. terminative (but not perfective, as usually assumed), represents a regular thematic remodel‐ The Latin future tense formation of the type faxō and its Italic background 271 <?page no="272"?> ling of a semi-thematic s-future paradigm inherited from Proto-Italic, con‐ sisting of thematic forms as 1.sg. in *-ō and 1.pl., 3.pl. with the thematic vowel *-oand of athematic forms in the rest of the paradigm, i.e. 1.sg. *fak-s-ō (thematic), 2.sg. *fak-s-si, 3.sg. *fak-s-ti (athematic), 1.pl. *fak-s-o-mos (the‐ matic), 2.pl. *fak-s-tes (athematic), 3.pl. *fak-s-o-nti (thematic) ‘will make’. In its structure, it corresponds to the semi-thematic inflectional type as preserved by some irregular verbs going back to PIE athematic verbs, such as Lat. velle < *u̯ el-s-i ‘to want, wish’ with 1.sg. volō < *u̯ el-ō, 2.sg. vīs ← *u̯ el-si, 3.sg. vult, volt < *u̯ el-ti, 1.pl. volumus < *u̯ el-o-mos, 2.pl. vultis, voltis < *u̯ el-tes, 3.pl. volunt < *u̯ el-o-nti. The semi-thematic Proto-Italic forerunner of the paradigm of OLat. faxō goes back to an athematic PIE s-desiderative with an inchoative-intentional value of ‘to be going to make’. The associated subjunctive of the type faxim, faxīs, faxit, + faxīmus, faxītis*, faxint structur‐ ally represents the optative in *-īof the original athematic paradigm. In contrast to the few preserved irregular semi-thematic verbs of the type velle, the future indicative faxō shows full thematization as most verbs of the Proto-Italic semi-thematic type based on inherited athematic paradigms, such as Lat. findō, -is, -it, -unt ‘split’ < *-ō, *-e-s, *-e-t, *-o-nt ← Proto-Italic *find-ō, *-si, *-ti, *-o-nti ← PIE *b h i-né-d-mi, *-si, *-ti, *b h i-n-d-énti etc. 18.2 The Sabellic s-future inflection as exemplified by 3.sg. Osc. fust, Umbr. fust, fust = / fu-s-t/ , 3.pl. Osc. fusent = / fu-s-ent/ , Umbr. furent < *fu-s-ent ‘will be’ goes back to the same Proto-Italic semi-thematic paradigm type, but, due to the limited evidence from the preserved corpus, only instances of athematic forms of the respective paradigm are attested, taking into account however that the original 3.pl. *fu-s-ont ← Proto-Italic *fū-s-o-nti underwent the paradigmatically regular Sabellic replacement of the thematic ending *-ont by athematic *-ent as also in the etymologically related present tense paradigm with 3.pl. Osc. fiíet, fiiet ‘happen, are performed’ = / fīi̯ -ent/ ← *fīi̯ -ont = Lat. fīunt ‘become’ < Proto-Italic *fī i̯ -o-nti < *fū-i̯ o-nti. The presence of thematic forms within the Proto-Italic paradigm triggered for the associated subjunctive a formation with long thematic vowel *-ē-/ -ōas in the case of the subjunctives of thematic paradigms in *-e/ o-. Therefore, the subjunctive of the s-future *fak-s-ō (thematic), *fak-s-si, *fak-s-ti (athematic), *fak-s-o-nti (thematic) ‘will make’ had the form *fak-s-ō, *fak-s-ē-s, *fak-s-ē-d, *fak-s-ō-nd, whence presumably early *fak-s-ē-m, *fak-s-ē-s, *fak-s-ē-d, *fak-s-ē-nd by generalization of *-ē-. Being a paradigmatic derivative of an inherited de‐ siderative category with the functional meaning ‘is going to make’ → ‘will 272 Reiner Lipp <?page no="273"?> make’ as underlying the Proto-Italic s-future, this subjunctive formation had the original value of a related prospective in the meaning ‘is expected to be going to make’. In Proto-Italic this formation was used as the associated conditional of the s-future in the sense of ‘would make (under certain conditions)’. As the expression of the non-real, it was however the functional counterpart of the imperfective indicative present faciō, facit ‘make’ with the stem *fak-i̯ o-/ -i-, to which in Latino-Faliscan it aligned morphologically as the corresponding imperfect subjunctive, showing forms such as facerem, facerēs, faceret, facerent ‘would make’ < *fak-i-s-ē-m, *fak-i-s-ē-s, *fak-i-s-ē-d, *fak-i-s-ē-nd, following the proportion in stem pattern of other semi-the‐ matic verbs like present Lat. volō, vult < *u̯ el-ō, *u̯ el-ti with imperfect subjunctive vellem, vellēs, vellet, vellent < *u̯ el-s-ē-m, *u̯ el-s-ē-s, *u̯ el-s-ē-d, *u̯ el-s-ē-nd (i.e. a long vowel subjunctive built on the semi-thematic base of the underlying s-future). A comparable new formation can be seen in the case of the imperfect subjunctive Lat. essem, essēs, esset, essent ‘would be, were’ < *es-s-ē-m, *es-s-ē-s, *es-s-ē-d, *es-s-ē-nd as shaped after present indicative Palaeo-Lat. esom (Garigliano bowl) = Lat. sum, est ‘I am, (s)he is’ < *es-m̥ , *es-t < *es-mi, es-ti, replacing archaic Lat. forem, forēs, foret, forent ‘would be, were’ < *fu-s-ē-m, *fu-s-ē-s, *fu-s-ē-d, *fu-s-ē-nd = Osc. fusíd < *fu-s-ē-d as original conditional of the s-future preserved in Osc. fust, Umbr. fust, fust = / fu-s-t/ ‘will be’. In contradistinction to the indicative as the expression of the real from the view-point of the speaker, the imperfect subjunctive functionally represents a conditional belonging to the imper‐ fective present stem and as so-called irrealis expresses the non-real with reference to the present and in Old Latin indistinctively also to the past (in its imperfective aspect) or, as a dependent form in a subordinate clause, the simultaneity or finality with regard to a superordinate verb in a past tense form. 18.3 An innovation dating back to the Proto-Italic period is the formation of the s-future from a verbal base that was created from the present stem by subtraction of the distinctive present stem suffix *-(i̯ )e/ o- (still without effect of contraction with the preceding vowel). In Old Latin, this structure is still recognizable in forms as ad-iuverō ‘will render help’, monerint ‘may admonish’, noxit ‘should someone harm; may it (not) hurt’ < *(ad-)ii̯ ou̯ a-s-ō, *mone-s-ī-nd, *noke-s-ī-d with the verbal bases *ii̯ ou̯ a-, *mone-, *noke-, as descriptively abstracted from the present stems in Lat. ad-iuvō, -āre ‘help’ The Latin future tense formation of the type faxō and its Italic background 273 <?page no="274"?> 1 For valuable advice and important hints on the topic and digital research tools, I express my gratitude to my friends Bela Brogyanyi (Freiburg), Alberto Calderini (Perugia), Martin Capitanio (Offenbach), Martin Gális (Prague), Barbora Machajdíková (Bratislava), Riccardo Massarelli (Perugia), Sergio Neri (Munich), Alexandre Trébu‐ chon (Freiburg), Petr Vlach (Prague), Jan “Jen” Žídek (Prague). During my lectures on the Sabellic languages, held at Charles University Prague in the winter and summer semesters of the academic year 2020/ 21 via Zoom transmission, I fortunately had the opportunity to discuss with them research questions on the linguistic history of the Italic languages. The present study is deeply indebted to the fundamental research of my former teacher Prof. Helmut Rix (1926-2004), whose inspiring instruction I had the privilege to pursue for several years at the University of Freiburg. < *ii̯ ou̯ a-(i̯ )e/ o- ← PIE *h 1 i-h 1 éu̯ H- (redupl. pres.), Lat. moneō, -ēre ‘admonish’ < *mone-(i̯ )e/ o- ← PIE *mon-éi̯ e/ o- (causative), Lat. noceō, -ēre ‘harm, hurt’ < *noke-(i̯ )e/ o- ← PIE *noḱ-éi̯ e/ o- (causative). Beside these formations built on a basis in a short vowel, there is the prevailing type built on a basis in a long vowel, which shows the inner-Latin effect of the littera-rule, e.g. OLat. excantassit ‘may charm out’ < *kantā-s-ī-d from cantō, -āre ‘sing’ < *kantā-(i̯ )e/ oas denominative of *kan-to- (original verbal adjective of the basic verb canō, -ere ‘sing’ < *kan-e/ o-), OLat. habessit ‘may hold’ < *χabē-s-ī-d from habeō, -ēre ‘hold, have’ < *χabē-(i̯ )e/ oas essive in *-eh 1 -i̯ e/ o- ← PIE *-h 1 i̯ e/ o- (after fientive aor. in PIE *-éh 1 -/ *-h 1 -). 18.4 In contrast to Latin, Sabellic has renewed the formation of the s-fu‐ ture on the basis of the characterized present stem, i.e. including the suffix of the present stem into the derivational basis of the s-future; cf. present Marruc. feret = / fer-ent/ , Umbr. imptv. II fertu (medial syllable syncope) < *fer-e-tọ̄d with the thematic stem *fer-e/ o- ‘to carry, bring’ as basis of the future Umbr. ferest = / fer-e-s-t/ ← Proto-Italic *fer-s-ti, as well as of the imperfect subjunctive Osc. f]erríns (medial syllable syncope) < *fer-e-s-ẹ̄-nd ← *fer-s-ē-nd, representing the Proto-Italic s-future subjunctive *fer-s-ō-nd, which corresponds to Lat. ferrent < *fer-s-ē-nd, built on synchronically semi-thematic ferō, fers, fert, ferunt < *fer-o-/ *fer- (as against Lat. sererent < *ser-e-s-ē-nd after thematic present serō, -ere ‘link together’ < *ser-e/ o-); present imptv. II Umbr. aitu (medial syllable syncope) < *aɟ-e-tọ̄d (Umbrian velar palatalization), Osc. actud (medial syllable syncope) < *ag-e-tọ̄d (= Lat. agitō) with thematic stem *ag-e/ o- ‘to drive, act, do’ as basis of the future subjunctive South Picene aisis (medial syllable syncope) < *aɟ-e-s-ī-s (dialectal South Picene velar palatalization) < *ag-e-s-ī-s ← Proto-Italic future optative *ag-s-ī-s, corresponding to OLat. āxim, āxīs*. 1 274 Reiner Lipp <?page no="275"?> Bibliography Latin literary texts including grammarians, lexicographers and fragmentary tradition are largely cited after the editions listed in OLD, except in the case of specified references to the text passages concerned. Latin inscriptions are cited according to CIL I 2 , Faliscan inscriptions according to Bakkum (2009), and Sabellic inscriptions according to Rix (2002 [South Picene inscriptions being quoted however without the prefix Sp and the Umbrian Tabulae Iguvinae with TI]), under due consideration of Crawford (2011). In this article small caps are used to represent forms from Latin inscriptions, while in the case of the other languages of Ancient Italy, small caps in italics are used for forms from inscriptions in the respective native non-Latin alphabets, whereas lower-case italics are used for forms from Sabellic epichoric inscriptions in the Latin alphabet and, as usual, for forms from Latin literary texts. Bakkum, Gabriël C. L. M. (2009): The Latin dialect of the Ager Faliscus. 150 years of scholarship. 2 parts. Amsterdam: Amsterdam University Press. Benediktsson, Hreinn (1960): “The vowel syncope in Oscan-Umbrian”. Norsk Tids‐ skrift for Sprogvidenskap 19: 157-295. Benveniste, Émile (1922): “Les futurs et subjonctifs sigmatiques du latin archaïque”. Bulletin de la Société de Linguistique de Paris 23 (1921-1922): 32-63. Bertocci, Davide (2016): “Remarks on the type faxō/ faxim”. In: Latinitatis rationes: Descriptive and Historical Accounts for the Latin Language, Paolo Poccetti (ed.). Berlin/ Boston: De Gruyter, 22-37. Blažek, Václav (2018): “Toward the etymology of Latin littera”. Graeco-Latina Bru‐ nensia 23(2): 5-11. Bruns, Carolus Georgius (1893): Fontes iuris Romani antiqui, Carolus Georgius Bruns (ed.), editio sexta cura Theodori Mommseni et Ottonis Gradenwitz. Pars Prior: Leges et negotia. Friburgi in Brisgavia et Lipsiae: in libraria academica I.C.B. Mohrii. Buck, Carl D. (1928): A Grammar of Oscan and Umbrian - with a Collection of Inscrip‐ tions and a Glossary, reprinted, with additions and corrections. Boston: Athenaeum Press. 2. Nachdruck der Ausgabe Boston 1904 mit den Zusätzen der erweiterten Auflage 1928. (Repr., Hildesheim: Georg Olms, 1995.) Buzássyová, Ľudmila (2012): “Some remarks on the genitive of the Latin third de‐ clension and its Faliscan pendant”. In: Greek and Latin from an Indo-European Perspective 3 (GLIEP 3), Proceedings of the Conference held at the Comenius Univer‐ sity Bratislava, July 8th - 10th 2010, Wojciech Sowa/ Stefan Schaffner (eds.) (IJDL Supplements 2). München: peniope, 3-16. The Latin future tense formation of the type faxō and its Italic background 275 <?page no="276"?> Cardona, George (1965): “The Vedic imperatives in -si”. Language 41: 1-18. Cardona, George (1968): On Haplology in Indo-European. Philadelphia: University of Pennsylvania Press. Chantraine, Pierre (1984): Morphologie historique du grec, deuxième édition, revue et augmentée, nouveau tirage. Paris: Klincksieck. Churchill, J. Bradford (2000): “Dice and Facie: Quintilian, Institutio Oratoria 1.7.23 and 9.4.39”. The American Journal of Philology 121: 279-289. CGL = Corpus glossariorum Latinorum, a Gustavo Loewe incohatum … composuit, recensuit, edidit Georgius Goetz [VII volumina]. Lipsiae et Berolini: in aedibus B. G. Teubneri, 1888-1923. CIL I 2 = Corpus Inscriptionum Latinarum, Vol. I 2 Pars II. 1918-1986. Vol. I. Inscriptiones Latinae antiquissimae ad C. Caesaris mortem, Th. Mommsen (ed.). Editio altera. Pars posterior. Cura E. Lommatzsch. Fasc. I 1918; II 1931: Addenda; III 1943: Addenda altera. - Fasc. IV 1986: Addenda tertia. Cura A. Degrassi. Auxit et edenda curavit H. Krummrey. Berlin: de Gruyter. Clackson, James (2020): “The Latin and Oscan imperfect subjunctive in *-sē-”. In: Pro‐ ceedings of the 31st Annual UCLA Indo-European Conference, Los Angeles, November 8th and 9th 2019, David M. Goldstein/ Stephanie W. Jamison/ Brent Vine (eds.) with the assistance of Angelo Mercado. Hamburg: Buske, 55-77. Cohen, Paul S. (2017): “PIE telic s-extensions and their diachronic implications”. In: Usque ad Radices: Indo-European Studies in Honour of Brigit Anette Olsen, Bjarne Simmelkjær Sandgaard Hansen/ Adam Hyllested/ Anders Richardt Jørgensen/ Guus Kroonen/ Jenny Helena Larsson/ Benedicte Nielsen Whitehead/ Thomas Olander/ To‐ bias Mosbæk Søborg (eds.). Copenhagen: Museum Tusculanum Press, 61-77. Cowgill, Warren (1978): “The source of Latin vīs ‘thou wilt’”. Die Sprache 24: 25-44. Crawford, M.H. (1996): Roman Statutes, M.H. Crawford (ed.). II Volumes. London: Institute of Classical Studies, School of Advanced Study, University of London. Crawford, M.H. (2011): Imagines Italicae: A Corpus of Italic Inscriptions, M.H. Crawford with W.M. Broadhead/ J.P.T. Clackson/ F. Santangelo/ S. Thompson/ M. Watmough (eds.), E. Bissa/ G. Bodard (comp.). III Volumes. London: Institute of Classical Studies, School of Advanced Study, University of London. Delbrück, Bertold (1888): Syntaktische Forschungen V: Altindische Syntax. Halle a. S.: Verlag der Buchhandlung des Waisenhauses. Dunkel, George E. (1998): “On the ‘thematisation’ of Latin sum, volo, eo, and edo and the system of endings in the IE subjunctive active”. In: Mír curad: Studies in Honor of Calvert Watkins, Jay Jasanoff/ H. Craig Melchert/ Lisi Oliver (eds.). Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck, 83-100. 276 Reiner Lipp <?page no="277"?> Dupraz, Emmanuel (2006): “La forme falisque pipafo/ pafo et le futur latino-falisque”. In: LALIES, Actes des sessions de linguistique et de littérature 27. Paris: Éditions Rue d’Ulm/ Presses de l’École normale supérieure, 325-344. Eichner, Heiner (1993): “1919 oder 1991? Zur Entwicklung der oskisch-umbrischen Studien nebst einer neuen Interpretation des Textes von Fonte Romito (Vetter Nr.-147)”. In: Oskisch-Umbrisch: Texte und Grammatik, Arbeitstagung der Indoger‐ manischen Gesellschaft und der Società Italiana di Glottologia vom 25. bis 28. September 1991 in Freiburg, Helmut Rix (ed.). Wiesbaden: Reichert, 46-95. Ernout, Alfred/ Meillet, Antoine (1967): Dictionnaire étymologique de la langue latine: Histoire des mots, quatrième édition, deuxième tirage augmenté de corrections nouvelles. Paris: Klincksieck. Euler, Wolfram (1992): Modale Aoristbildungen und ihre Relikte in den alteuropäischen Sprachen. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Friis, Louise Skydsbjerg (2018): The Sigmatic Aorist and Hittite: Twists and Turns in Search of a Core-Indo-European Innovation. Prize paper/ Master’s thesis, University of Copenhagen. García Castillero, Carlos (2000): La formación del tema de presente primario oscoumbro. Vitoria-Gasteiz: Universidad del País Vasco/ Euskal Herriko Unibertsitatea. Giacomelli, Gabriella (1963): La lingua falisca. Firenze: Olschki. Gotō, Toshifumi (2013): Old Indo-Aryan Morphology and its Indo-Iranian Background, in co-operation with Jared S. Klein and Velizar Sadovski (Sitzungsberichte der Österreichischen Akademie der Wissenschaften, philosophisch-historische Klasse, 849. Band). Wien: Österreichische Akademie der Wissenschaften. Hadas-Lebel, Jean (2011): “La variante falisque”. In: La variation linguistique dans les langues de l’Italie préromaine, Gilles van Heems (ed.). Lyon: Maison de l’Orient et de la Méditerranée Jean Pouilloux, 155-168. Harðarson, Jón Axel (1993): Studien zum urindogermanischen Wurzelaorist und dessen Vertretung im Indoiranischen und Griechischen. Innsbruck: Institut für Sprachwis‐ senschaft der Universität Innsbruck. Harðarson, Jón Axel (1995): “Nord. metta, altengl. mettian, altind. mādáyate und der Ansatz einer uridg. Wurzel *med- ‘satt/ voll werden’”. Historische Sprachforschung 108: 207-235. Harðarson, Jón Axel (1998): “Mit dem Suffix *-eh 1 bzw. *-(e)h 1 -i̯ e/ ogebildete Verbal‐ stämme im Indogermanischen”. In: Sprache und Kultur der Indogermanen, Akten der X. Fachtagung der Indogermanischen Gesellschaft, Innsbruck, 22.-28. September 1996, Wolfgang Meid (ed.). Innsbruck: Institut für Sprachwissenschaft der Univer‐ sität Innsbruck, 323-339. The Latin future tense formation of the type faxō and its Italic background 277 <?page no="278"?> Hartmann, Markus (2005): Die frühlateinischen Inschriften und ihre Datierung: Eine linguistisch-archäologisch-paläographische Untersuchung. Bremen: Hempen. Heenen, François (2006): Le désidératif en védique (Leiden Studies in Indo-European 13). Amsterdam/ New York: Rodopi. Hill, Eugen (2004): “Die sigmatischen Modus-Bildungen der indogermanischen Spra‐ chen. Erste Abhandlung: Das baltische Futur und seine Verwandten”. International Journal of Diachronic Linguistics and Linguistic Reconstruction 1: 69-171. Hill, Eugen (2014): “Using stem suppletion for semantic reconstruction: The case of Indo-European modals and East Baltic future tense formations”. Indo-European Linguistics 2: 42-72. Hoffmann, Karl/ Forssman, Bernhard (2004): Avestische Laut- und Flexionslehre, 2., durchgesehene und erweiterte Auflage. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. van den Hout, Theo P.J. (1988): “Hethitisch damašš-/ dameššmi ‘(be)drücken’ und der indogermanische sigmatische Aorist”. In: A Linguistic Happening in Memory of Ben Schwartz: Studies in Anatolian, Italic, and other Indo-European languages, Yoël L. Arbeitman (ed.). Louvain-la-Neuve: Peeters, 305-319. Jasanoff, Jay H. (1978): Stative and Middle in Indo-European. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Jasanoff, Jay H. (1988): “PIE *ĝnē- ‘recognize, know’”. In: Die Laryngaltheorie und die Rekonstruktion des indogermanischen Laut- und Formensystems, Alfred Bammes‐ berger (ed.). Heidelberg: Universitätsverlag C. Winter, 227-239. Jasanoff, Jay H. (1991): “The origin of the Italic imperfect subjunctive”. Historische Sprachforschung 104: 84-105. Jasanoff, Jay H. (2003): Hittite and the Indo-European Verb. Oxford/ New York: Oxford University Press. Jasanoff, Jay H. (2012): “Did Hittite have si-imperatives? ”. In: The Sound of Indo- European 2: Papers on Indo-European Phonetics, Phonemics and Morphophonemics, Roman Sukač/ Ondřej Šefčík (eds.) (LINCOM Studies in Indo-European Linguistics 41). Muenchen: LINCOM, 116-132. Jasanoff, Jay H. (2019): “The sigmatic forms of the Hittite verb”. Indo-European Lin‐ guistics 7: 13-71. Joseph, Brian/ Wallace, Rex (1987): “Latin sum/ Oscan súm, sim, esum”. The American Journal of Philology 108(4): 675-693. Kieckers, Ernst (1931): Historische lateinische Grammatik - mit Berücksichtigung des Vulgärlateins und der romanischen Sprachen. Zweiter Teil: Formenlehre. München: Hueber. 278 Reiner Lipp <?page no="279"?> Kloekhorst, Alwin (2008): Etymological Dictionary of the Hittite Inherited Lexicon (Leiden Indo-European Etymological Dictionary Series, Volume 5). Leiden/ Boston: Brill. Kortlandt, Frederik H.H. (2007): Italo-Celtic Origins and Prehistoric Development of the Irish Language. Amsterdam/ New York: Rodopi. Kümmel, Martin J. (2010): “Rezension von Heenen 2006”. Indo-Iranian Journal 53: 179-184. Kümmel, Martin (2014): “Etymologie und Phonologie: Umbrisch amboltu”. Die Sprache 50 (2012/ 13): 31-43. Lazzarini, Maria Letizia/ Poccetti, Paolo (2001): Il mondo enotrio tra VI e V secolo a.C.: Atti dei seminari napoletani (1996-1998), Lʼiscrizione paleoitalica da Tortora. Napoli: Loffredo Editore. Lejeune, Michel (1990): “Notes de linguistique italique, XL : «Bois! » disait ce sicule; «Je boirai» répond ce falisque”. Revue des études latines (Paris) 68: 28-30. Leumann, Manu (1977): Lateinische Laut- und Formenlehre, Neuausgabe 1977 der 1926-1928 in 5. Auflage erschienenen ‘Lateinischen Laut- und Formenlehre’. Mün‐ chen: C.H. Beck. Lipp, Reiner (2009a): Die indogermanischen und einzelsprachlichen Palatale im Indoiranischen, Band I: Neurekonstruktion, Nuristan-Sprachen, Genese der indoarischen Retroflexe, Indoarisch von Mitanni. Heidelberg: Universitätsverlag Winter. Lipp, Reiner (2009b): Die indogermanischen und einzelsprachlichen Palatale im Indoiranischen, Band II: Thorn-Problem, indoiranische Laryngalvokalisation. Heidel‐ berg: Universitätsverlag Winter. Lipp, Reiner (2016): “Final stops in Indo-European: Their phonological classification as a key to the Proto-Indo-European root structure constraint”. In: Indoevropská srovnávací jazykověda: nedávný vývoj a aktuální stav. Indo-European Comparative Linguistics: Recent Development and State-of-the-Art, Václav Blažek/ Ondřej Šefčík (eds.). Slovo a slovesnost 77(4): 251-299. Lipp, Reiner (2021): “The medial syllable syncope in the South Picene inscriptions”. In: Studies in General and Historical Linguistics Offered to Jón Axel Harđarson On the Occasion of his 65th Birthday, Matteo Tarsi (ed.). Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck, 269-328. Lipp, Reiner (2022): “Umbrian fefure as a relic form of the Proto-Indo-European perfect”. In: eqo : duenosio. Studi offerti a Luciano Agostiniani, Alberto Calderini/ Riccardo Massarelli (eds.). Ariodante - Linguistica e epigrafia dell’Italia antica, Università degli Studi di Perugia I: 499-534. The Latin future tense formation of the type faxō and its Italic background 279 <?page no="280"?> LIV 2 = LIV. Lexikon der indogermanischen Verben: Die Wurzeln und ihre Primärstamm‐ bildungen, 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 Martin Kümmel und Helmut Rix. Wiesbaden: Reichert 2001. Malzahn, Melanie (2019): “How the Indo-Europeans managed TO OVERCOME and TO GET OLD: The behavior of telic roots in PIE”. In: QAZZU warrai: Anatolian and Indo-European Studies in Honor of Kazuhiko Yoshida, Adam Alvah Catt/ Ronald I. Kim/ Brent Vine (eds.). Ann Arbor/ New York: Beech Stave Press, 225-238. Martzloff, Vincent (2006a): Le thèmes de présent dans l’épigraphie italique et en latin archaïque. Ph.D. dissertation, Université Lumière - Lyon II. Martzloff, Vincent (2006b): “Les syntagmes picéniens povaisis pidaitúpas, me{nt}fistrúí nemúne í , trebegies titúí”. Revue de philologie, de littérature et dʼhistoire anciennes 80: 63-104. Martzloff, Vincent (2009): “Questions d’exégèse picénienne”. In: Autour de Michel Lejeune: Actes des Journées d’etude organisées à l’Université Lumière-Lyon 2 - Maison de l’Orient et de la Méditerranée, 2-3 février 2006, Frédérique Biville/ Isabelle Boehm (eds.). Lyon: Maison de l’Orient et de la Méditerranée, 359-378. Mayrhofer, Manfred (1986): Indogermanische Grammatik, Manfred Mayrhofer (ed.); Bd I: 2. Halbband: Lautlehre (Segmentale Phonologie des Indogermanischen). Hei‐ delberg: Universitätsverlag C. Winter. Mayrhofer, Manfred (2005): Die Fortsetzung der indogermanischen Laryngale im Indo-Iranischen (Sitzungsberichte der Österreichischen Akademie der Wissen‐ schaften, Philosophisch-historische Klasse, 730. Band). Wien: Österreichische Aka‐ demie der Wissenschaften. McCone, Kim (1991): The Indo-European origins of the Old Irish nasal presents, sub‐ junctives and futures. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Meillet, Antoine/ Vendryes, Joseph (1979): Traité de grammaire comparée des langues classiques, 5 e édition, tirage revue par J. Vendryes. Paris: Champion. Meiser, Gerhard (1986): Lautgeschichte der umbrischen Sprache. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Meiser, Gerhard (1993a): “Uritalische Modussyntax: Zur Genese des Konjunktiv Im‐ perfekt”. In: Oskisch-Umbrisch: Texte und Grammatik, Arbeitstagung der Indoger‐ manischen Gesellschaft und der Società Italiana di Glottologia vom 25. bis 28. September 1991 in Freiburg, Helmut Rix (ed.). Wiesbaden: Reichert, 167-195. Meiser, Gerhard (1993b): “Zur Funktion des Nasalpräsens im Urindogermanischen”. In: Indogermanica et Italica: Festschrift für Helmut Rix zum 65. Geburtstag, Gerhard 280 Reiner Lipp <?page no="281"?> Meiser (ed.). Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck, 280-313. Meiser, Gerhard (1993c): “Das Gerundivum im Spiegel der italischen Onomastik”. In: Sprachen und Schriften des antiken Mittelmeerraums: Festschrift für Jürgen Untermann zum 65. Geburtstag, Frank Heidermanns/ Helmut Rix/ Elmar Seebold (eds.). Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck, 255- 268. Meiser, Gerhard (1998): Historische Laut- und Formenlehre der lateinischen Sprache. Darmstadt: Wissenschaftliche Buchgesellschaft. Meiser, Gerhard (2003): Veni Vidi Vici: Die Vorgeschichte des lateinischen Perfektsys‐ tems (Zetemata 113). München: C.H. Beck. de Melo, Wolfgang David Cirilo (2007a): The Early Latin Verb System: Archaic Forms in Plautus, Terence, and Beyond (Oxford Classical Monographs). Oxford/ New York: Oxford University Press. de Melo, Wolfgang David Cirilo (2007b): The Early Latin Verb System: Archaic Forms in Plautus, Terence, and Beyond. Appendices (Oxford Classical Monographs) [pdf file]. de Melo, Wolfgang David Cirilo (2007c): “The sigmatic future and the genetic affiliation of Venetic: Latin faxō ‘I shall make’ and Venetic vhag.s.to ‘he made’”. Transactions of the Philological Society 105(1): 1-21. Monteil, Pierre (1986): Éléments de phonétique et de morphologie du latin. Paris: Nathan-Université. Narten, Johanna (1964): Die sigmatischen Aoriste im Veda. Wiesbaden: Harrassowitz. Narten, Johanna (1973): “Zur Flexion des lateinischen Perfekts”. Münchener Studien zur Sprachwissenschaft 31: 133-50. Neri, Sergio (2005): “Riflessioni sull’apofonia radicale di proto-germanico *namō n ‘nome’”. Historische Sprachforschung 118: 201-250. Neri, Sergio (2013): “Zum urindogermanischen Wort für ‘Hand’”. In: Multi nominis grammaticus: Studies in Classical and Indo-European Linguistics in Honor of Alan J. Nussbaum on the Occasion of his Sixty-Fifth Birthday, Adam I. Cooper/ Jeremy Rau/ Michael Weiss (eds.). Ann Arbor/ New York: Beech Stave Press, 185-205. Neri, Sergio (2017): Wetter: Etymologie und Lautgesetz. Università degli Studi di Perugia: Centro Stampa Morlacchi. Neri, Sergio (2018): “Genitiv und Lokativ: Zur Herkunft der urindogermanischen Genitivendung *-si̯ o”. In: Vina diem celebrent: Studies in Linguistics and Philology in Honor of Brent Vine, Dieter Gunkel/ Stephanie W. Jamison/ Angelo O. Mercado/ Ka‐ zuhiko Yoshida (eds). Ann Arbor/ New York: Beech Stave Press, 256-266. Neri, Sergio (2019): “Riflessi delle laringali indoeuropee in germanico”. In: Dall’ The Latin future tense formation of the type faxō and its Italic background 281 <?page no="282"?> indoeuropeo al germanico: Problemi di linguistica storica, Roberto Rosselli del Turco (ed.) (XVIII Seminario avanzato in Filologia germanica). Alessandria: Edizioni dell’Orso, 41-66. Neri, Sergio/ Ziegler, Sabine (2012): „Horde Nöss”: Etymologische Studien zu den Thüringer Dialekten. Bremen: Hempen. Neue, Friedrich/ C. Wagener (1897): Formenlehre der lateinischen Sprache, Dritter Band: Das Verbum, dritte, sehr vermehrte Auflage. Leipzig: Reisland. Nishimura, Kanehiro (2010): “Patterns of Vowel Reduction in Latin: Phonetics and Phonology”. Historische Sprachforschung 123: 217-257. Oettinger, Norbert (2007): “Der hethitische Imperativ auf -i vom Typ paḫši ‘schütze! ’”. In: Tabularia Hethaeorum: Hethitische Beiträge Silvin Košak zum 65. Geburtstag, Detlev Groddek/ Marina Zorman (eds.) (Dresdner Beiträge zur Hethitologie 25). Wiesbaden: Harrassowitz, 561-568. OLD = Oxford Latin Dictionary. Oxford: Clarendon Press, 1968. Ozoliņš, Kaspars (2015): Revisiting Proto-Indo-European Schwebeablaut. PhD disserta‐ tion, University of California, Los Angeles. Pedersen, Holger (1921): Les formes sigmatiques du verbe latin et le problème du futur indo-européen (Det Kongelige Danske Videnskabernes Selskab. Historisk-filologiske Meddelelser. III, 5.). København: Høst & Søn. Pedersen, Holger (1938): Hittitisch und die anderen indoeuropäischen Sprachen (Det Kgl. Danske Videnskabernes Selskab. Historisk-filologiske Meddelelser. XXV, 2.). København: Levin & Munksgaard. Pisani, Vittore (1974): Grammatica latina storica e comparativa, quarta edizione riveduta (Manuale storico della lingua latina, vol. II.). Torino: Rosenberg & Sellier. Poccetti, Paolo (2018): “Ein neuer Beleg des oskischen Futurums des Verbums ‚sein‘ zwischen Textinterpretation und Modussyntax”. In: Priscis libentius et liberius novis: Indogermanische und sprachwissenschaftliche Studien, Festschrift für Gerhard Meiser zum 65. Geburtstag, Olav Hackstein/ Andreas Opfermann (eds.). Hamburg: baar, 131-144. Postgate, J. 1899. “Operatus and operari”. The Journal of Philology 26: 314-320. Prosdocimi, Aldo/ Marinetti, Anna (1993): “Appunti sul verbo latino italico III: Sulla morfologia del tema-base del perfetto latino. I. I perfetti in -ue in -s-”. In: Sprachen und Schriften des antiken Mittelmeerraums: Festschrift für Jürgen Untermann zum 65. Geburtstag, Frank Heidermanns/ Helmut Rix/ Elmar Seebold (eds.). Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck, 297-328. Rasmussen, Jens E. (1985): “Der Prospektiv - eine verkannte indogermanische Verbal‐ kategorie? ” In: Grammatische Kategorien: Funktion und Geschichte, Akten der VII. 282 Reiner Lipp <?page no="283"?> Fachtagung der Indogermanischen Gesellschaft, Berlin, 20.-25. Februar 1983, Bern‐ fried Schlerath (ed.) unter Mitarbeit von Veronica Rittner. Wiesbaden: Reichert, 384-399. Rix, Helmut (1976): Historische Grammatik des Griechischen: Laut- und Formenlehre. Darmstadt: Wissenschaftliche Buchgesellschaft. 2., korrigierte Auflage 1992. Rix, Helmut (1977): “Das keltische Verbalsystem auf dem Hintergrund des indoiranisch-griechischen Rekonstruktionsmodells”. In: Indogermanisch und Keltisch: Kolloquium der Indogermanischen Gesellschaft am 16. und 17. Februar 1976 in Bonn, Karl Horst Schmidt (ed.). Wiesbaden: Reichert, 132-158. Rix, Helmut (1992): “Zur Entstehung des lateinischen Perfektparadigmas”. In: Latein und Indogermanisch: Akten des Kolloquiums der Indogermanischen Gesellschaft, Salzburg, 23.-26. September 1986, Oswald Panagl/ Thomas Krisch (eds.). Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck, 221-240. Rix, Helmut (1993): “Rezension von: La civiltà dei Falisci. Atti del XV Convegno di Studi Etruschi ed Italici. Cività Castellana - Forte Sangallo, 28-31 maggio 1987. Firenze: Olschki, 1990”. Kratylos 38: 83-87. Rix, Helmut (1994): Die Termini der Unfreiheit in den Sprachen Alt-Italiens. Stuttgart: Steiner. Rix, Helmut (1995): “Einige lateinische Präsensstammbildungen zu Seṭ-Wurzeln”. In: Analecta Indoeuropaea Cracoviensia, vol. II: Kuryłowicz memorial volume, Part one, W. Smoczýnski (ed.). Cracow: Universitas, 399-408. Rix, Helmut (1998a): “Bemerkungen zu den lateinischen Verbformen des Typs faxo faxim”. In: Mír curad: Studies in Honor of Calvert Watkins, Jay Jasanoff/ H. Craig Melchert/ Lisi Oliver (eds.). Innsbruck: Institut für Sprachwissenschaft der Univer‐ sität Innsbruck, 619-634. Rix, Helmut (1998b): “Eine neue frühsabellische Inschrift und der altitalische Prä‐ ventiv”. Historische Sprachforschung 111: 247-269. Rix, Helmut (1999): “Schwach charakterisierte lateinische Präsensstämme zu Seṭ-Wurzeln mit Vollstufe I”. In: Compositiones Indogermanicae in memoriam Jochem Schindler, Heiner Eichner/ Hans Christian Luschützky (eds.) unter redaktio‐ neller Mitwirkung von Velizar Sadovski. Praha: enigma corporation, 515-535. Rix, Helmut (2002): Sabellische Texte: Die Texte des Oskischen, Umbrischen und Südpikenischen. Heidelberg: Universitätsverlag C. Winter. Rix, Helmut (2003a): “Ausgliederung und Aufgliederung der italischen Sprachen”. In: Languages in Prehistoric Europe, Alfred Bammesberger/ Theo Vennemann (eds.) in collaboration with Markus Bieswanger/ Joachim Grzega. Heidelberg: Universitäts‐ verlag C. Winter, 147-172. The Latin future tense formation of the type faxō and its Italic background 283 <?page no="284"?> Rix, Helmut (2003b): “Towards a reconstruction of Proto-Italic: The verbal system”. In: Proceedings of the Fourteenth Annual UCLA Indo-European Conference, Los Angeles, November 8-9, 2002 (Journal of Indo-European Studies monograph series), Karlene Jones-Bley/ Martin E. Huld/ Angela Della Volpe/ Miriam Robbins Dexter (eds.) (Journal of Indo-European Studies monograph series, no. 47). Washington, DC: Institute for the Study of Man, 1-24. Rix, Helmut (2003c): “The Latin imperfect in -bā-, the Proto-Indo-European root *b h u̯ eh 2 and full grade I forms from seṭ-roots with full grade II”. In: Language in Time and Space: A Festschrift for Werner Winter on the Occasion of his 80th Birthday, Brigitte L.M. Bauer/ Georges-Jean Pinault (eds.). Berlin/ New York: Mouton de Gruyter, 363-384. Schmid, Wolfgang P. (1963): Studien zum baltischen und indogermanischen Verbum. Wiesbaden: Harrassowitz. Schmidt, Gernot (1986): “Zum indogermanischen s-Futur”. In: o-o-pe-ro-si: Festschrift für Ernst Risch zum 75. Geburtstag, Annemarie Etter (ed.). Berlin/ New York: de Gruyter, 33-59. Schrijver, Peter (1991): The reflexes of the Proto-Indo-European laryngeals in Latin. Amsterdam/ Atlanta: Rodopi. Schrijver, Peter (2003): “Athematic i-presents: the Italic and Celtic evidence”. Incontri Linguistici 26: 59-86. Schrijver, Peter (2006): “Review of Meiser 2003”. Kratylos 51: 46-64. Schumacher, Stefan (2004): Die keltischen Primärverben. Ein vergleichendes, etymolo‐ gisches und morphologisches Lexikon, unter Mitarbeit von Britta Schulze-Thulin und Caroline aan de Wiel. Innsbruck: Institut für Sprachen und Literaturen der Universität Innsbruck. Seldeslachts, Herman (2002): “Le présent de ‘boire’ en proto-indo-européen et les futurs falisques pipafo/ pafo”. Études classiques 70: 53-63. Sen, Ranjan (2015): Syllable and Segment in Latin. Oxford: Oxford University Press. Sihler, Andrew L. (1995): New Comparative Grammar of Greek and Latin. New York/ Oxford: Oxford University Press. Skutsch, Otto (1985): The Annals of Q. Ennius, ed. with introduction and commentary by Otto Skutsch. Oxford: Clarendon Press. Søborg, Tobias Mosbæk (2020a): Sigmatic verbal formations in Anatolian and Indo- European: A cladistic study. PhD. Dissertation, University of Copenhagen. Søborg, Tobias Mosbæk (2020b): “Sieving the sources of s-extended verbs in Hittite”. In: vácām̐ si miśrā́ kr̥ ṇavāmahai: Proceedings of the International Conference of the Society for Indo-European Studies and IWoBA XII, Ljubljana 4-7 June 2019, 284 Reiner Lipp <?page no="285"?> Celebrating One Hundred Years of Indo-European Comparative Linguistics at the University of Ljubljana, Luka Repanšek/ Harald Bichlmeier/ Velizar Sadovski (eds.). Hamburg: baar, 621-643. Sommer, Ferdinand (1914): Handbuch der lateinischen Laut- und Formenlehre, zweite und dritte Auflage. Heidelberg: Winter. Stang, Christian S. (1966): Vergleichende Grammatik der Baltischen Sprachen. Oslo/ Bergen/ Tromsø: Universitetsforlaget. Steinbauer, Dieter Hubertus (1989): Etymologische Untersuchungen zu den bei Plautus belegten Verben der lateinischen ersten Konjugation. Unter besonderer Berücksich‐ tigung der Denominative. Inaugural-Dissertation, Universität Regensburg. Alten‐ dorf b. Bamberg: Druckerei Gräbner. Stempel, Reinhard (1998): “Zur Vorgeschichte und Entwicklung des lateinischen Tempus- und Modussystems”. Historische Sprachforschung 111: 270-285. Szemerényi, Oswald (1966): “The origin of the Vedic ‘imperatives’ in -si”. Language 42: 1-6. Szemerényi, Oswald (1989): Einführung in die vergleichende Sprachwissenschaft, 3., vollständig neu bearbeitete Auflage. Darmstadt: Wissenschaftliche Buchgesell‐ schaft. 4., durchgesehene Auflage 1990. Szemerényi, Oswald (1996): Introduction to Indo-European Linguistics, translated from Einführung in die vergleichende Sprachwissenschaft, 4th edition, 1990, with additional notes and references. Oxford: Clarendon Press. Thomas, François (1938): Recherches sur le subjonctif latin: Histoire et valeur des formes. Paris: Klincksieck. Untermann, Jürgen (2000): Wörterbuch des Oskisch-Umbrischen. Heidelberg: Univer‐ sitätsverlag C. Winter. de Vaan, Michiel (2008): Etymological Dictionary of Latin and the Other Italic Lan‐ guages (Leiden Indo-European Etymological Dictionary Series, Volume 7). Leiden/ Boston: Brill. de Vaan, Michiel (2011): “PIE i-presents, s-presents, and their reflexes in Latin”. Glotta 87: 23-36. Vetter, Emil (1953): Handbuch der italischen Dialekte, I. Band: Texte mit Erklärung, Glossen, Wörterverzeichnis. Heidelberg: Universitätsverlag C. Winter. Villanueva Svensson, Miguel (2010): “Baltic sta-presents and the Indo-European desiderative”. Indogermanische Forschungen 115: 204-233. Vine, Brent (1998): “Remarks on the Archaic Latin ‘Garigliano bowl’ inscription”. Zeitschrift für Papyrologie und Epigraphik 121: 257-262. The Latin future tense formation of the type faxō and its Italic background 285 <?page no="286"?> Walde, Alois/ Hofmann, Johan B. (1938): Lateinisches Etymologisches Wörterbuch, 3., neubearbeitete Auflage, Erster Band. Heidelberg: Winter. Warmington, E.H. (1967a): Remains of Old Latin I: Ennius and Caecilius, E.H. War‐ mington (ed., transl.), revised and reprinted (Loeb Classical Library 294). Cam‐ bridge, Mass./ London, England: Harvard University Press. Warmington, E.H. (1967b): Remains of Old Latin III: Lucilius; The Twelve Tables, E.H. Warmington (ed., transl.), revised and reprinted (Loeb Classical Library 329). Cambridge, Mass./ London, England: Harvard University Press. Weiss, Michael (2009): Outline of the Historical and Comparative Grammar of Latin, second, corrected printing 2011. Ann Arbor/ New York: Beech Stave Press. Weiss, Michael (2010a): Language and ritual in Sabellic Italy: The ritual complex of the third and fourth Tabulae Iguvinae. Leiden/ Boston: Brill. Weiss, Michael (2010b): “Observations on the Littera Rule”. ECIEC XXIX, Cornell University, June 19, 2010 [Handout]. Weiss, Michael (2012): “Italo-Celtica: Linguistic and cultural points of contact be‐ tween Italic and Celtic”. In: Proceedings of the 23rd Annual UCLA Indo-European Conference, Los Angeles, October 28th and 29th, 2011, Stephanie W. Jamison/ H. Craig Melchert/ Brent Vine (eds.), with the assistance of Angelo Mercado. Bremen: Hempen, 151-173. Willi, Andreas (2018): Origins of the Greek Verb. Cambridge: Cambridge University Press. Zair, Nicholas (2016): “Vowel weakening in the Sabellic languages as language con‐ tact”. Indogermanische Forschungen 121: 295-315. 286 Reiner Lipp <?page no="287"?> Editors & contributors Editors Mgr. Barbora Machajdíková, PhD. Assistant Professor, Department of Classical and Semitic Philology Faculty of Arts, Comenius University, Bratislava, Slovakia barbora.machajdikova@uniba.sk doc. Mgr. Ľudmila Eliášová Buzássyová, PhD. Associate Professor, Department of Classical and Semitic Philology Faculty of Arts, Comenius University, Bratislava, Slovakia ludmila.buzassyova@uniba.sk Contributors prof. RNDr. Václav Blažek, CSc. Professor, Department of Linguistics and Baltic Languages Faculty of Arts, Masaryk University, Brno, Czech Republic blazek@phil.muni.cz doc. Mgr. Ľudmila Eliášová Buzássyová, PhD. Associate Professor, Department of Classical and Semitic Philology Faculty of Arts, Comenius University in Bratislava, Slovak Republic ludmila.buzassyova@uniba.sk Dr. hab. Máté Ittzés Associate Professor, Indo-European Linguistics Research Division Department of Indian Studies Eötvös Loránd University, Budapest, Hungary ittzes.mate@btk.elte.hu Dr. phil. Reiner Lipp M.A. Institute of Comparative Linguistics Faculty of Arts, Charles University, Prague, Czech Republic reiner.lipp@capitanio.org <?page no="288"?> Dr. hab. Anneli Luhtala Associate Professor, Department of Languages Faculty of Arts, University of Helsinki, Finnland anneli.luhtala@helsinki.fi Mgr. Martin Masliš Institute of Greek and Latin Studies Faculty of Arts, Charles University, Prague, Czech Republic martin.maslis@ff.cuni.cz doc. Mgr. Gabriela Múcsková, PhD. Associate Professor, Department of Slovak Language and Theory of Commu‐ nication Faculty of Arts, Comenius University in Bratislava, Slovak Republic gabriela.mucskova@uniba.sk Dr. hab. Wojciech Sowa Prof. UJ University Professor, Institute of Classical-Philology Jagiellonian University, Cracow, Poland wojciech.sowa@uj.edu.pl doc. PhDr. Ondřej Šefčík, Ph.D. Associate Professor, Department of Linguistics and Baltic Languages Faculty of Arts, Masaryk University, Brno, Czech Republic sefcik@phil.muni.cz Editorial & technical support Mgr. Mária Šibalová, PhD. and Mgr. Milan Regec, PhD. 288 Editors & contributors <?page no="289"?> Sprachvergleich Studien zur synchronen und diachronen Sprachwissenschaft herausgegeben von Katrin Schmitz, Joachim Theisen und Carlotta Viti Die Reihe widmet sich in Monographien und Sammelbänden dem Sprachvergleich in seiner ganzen Bandbreite und will dabei bewusst über Grenzen gehen: Grenzen der Sprachfamilien, Grenzen der Theorien und Modelle, Grenzen in Zeit und Raum. Die Studien verbinden eine genaue Datenanalyse der jeweiligen Texte und Sprachen in ihren synchronen oder diachronen Aspekten mit den aktuellsten Erkenntnissen der linguistischen Theorie. Sie erkunden dabei die Möglichkeiten des methodischen Spektrums moderner Sprachwissenschaft und leisten einen Beitrag zu seiner Erweiterung. Die Reihe wird herausgegeben von Katrin Schmitz (Wuppertal), Joachim Theisen (Athen) und Carlotta Viti (Nancy). Publikationssprachen sind Deutsch und Englisch. Über die Aufnahme neuer Titel entscheidet das Herausgeberkollegium mit Unterstützung eines wissenschaftlichen Beirats nach eingehender Prüfung. 1Pauline Weiß Die innere Struktur der DP in den altindogermanischen Artikelsprachen Eine Analyse der Funktion und Verwendung der Artikeltypen 2018, 502 Seiten €[D] 88,- ISBN 978-3-8233-8184-6 2Chiara Zanchi Multiple Preverbs in Ancient Indo- European Languages A comparative study on Vedic, Homeric Greek, Old Church Slavic, and Old Irish 2019, 436 Seiten €[D] 88,- ISBN 978-3-8233-8274-4 3Barbora Machajdíková, Ľudmila Eliášová Buzássyová (eds.) Greek - Latin - Slavic Aspects of Linguistics and Grammatography 2023, 288 Seiten €[D] 78,- ISBN 978-3-8233-8527-1 Bisher sind erschienen: <?page no="290"?> ISBN 978-3-8233-8527-1 The volume is intended for classical philologists and a broad range of scholars working in the fields of theoretical, historical, and comparative linguistics with Ancient Greek, Latin, or Slavic languages as the primary evidence in their research. The contributions address topics ranging from issues of grammatography in a diachronic perspective to historical and comparative linguistics. They encompass both monothematic case studies and comprehensive analyses that capture a linguistic phenomenon in its entirety as well as within a broader context. Sprachvergleich Studien zur synchronen und diachronen Sprachwissenschaft Band 3 Machajdíková / Eliášová Buzássyová (Eds.) Greek - Latin - Slavic Barbora Machajdíková / Ľudmila Eliášová Buzássyová (Eds.) Greek - Latin - Slavic Aspects of Linguistics and Grammatography