Subject Clitics in Piedmontese: A Diachronie Perspective 1. In recent years, the behaviour of clitics has attracted a considerable amount of attention in the Romance field. Although it is their syntactic function at the various stages of development that arouses the greatest interest, these preliminary observations will concentrate more on morphology and etymology. The term Piedmontese here refers not only to the standardized variety based on the dialect of Turin but also to the numerous other Piedmontese varieties which developed directly from Latin alongside the variety which has gradually emerged as the standard. However, despite the large number of grammatical systems that exist, the pronominal subsystems preserve a degree of similarity that justifies restricting the analysis to a selection of the available data. From the 17 th century onwards, data are drawn from standard Piedmontese and Turinese texts, the dialect of Castellinaldo (ToPPINO 1913), and the dialect of Cairo Montenotte (PARRY 1985), and data recorded in Piedmont during the last ten years 1 . 1.1. Contemporary Piedmontese, like all other Northern Italian dialects, French, and a number of other Romance varieties (cf. VANELLI/ RENZIIBENINCA 1985: 163), possesses a double series of subject pronouns: stressed and unstressed. The latter are clitics that occur with each person of the verb. The written standard, based on the Turinese variety, has the following subject clitic paradigm: sg. pi. I i i II it i III a a 1 I wish to thank the University of Wales: Aberystwyth, the British Academy and the University of Wales Research Fund for supporting various field-work visits to Piedmont. My sincere thanks as usual go to the many informants who gave unstintingly of their time and knowledge. I am also grateful to Professor Glanville Price and to Professor Edward Tuttle for their comments on earlier versions of this article, the substance of which was first presented at the 19 th Romance Linguistics Seminar, Cambridge 1991. Subject Clitics in Piedmontese: A Diachronie Perspective 97 In combination with the optional free stressed subject pronouns a typical verb paradigm runs: mi iparlo 2 ti itparle chiel aparla chila aparla [mi i 'porlu] [ti it 'porle] [kjEl a 'parla] [ c kila a 'parla] noi iparloma voi iparle lor aparlo [nuj i por'luma] [vuj i 'porle] [! ur a 'parlu] My own recordings ofTurinese, as well as BERRUTO (1990: 19), give the variant [e] for i and [t] for it. 1.2. Although the normative grammar of BRERO and BERTODATTI (1988) insists that the subject clitics are obligatory in all contexts 3 , a survey carried out by RENZI and VANELLI (1983) gives only the 2 nd sg. and the 3 rd person clitics (sg. and pl.) as compulsory in Piedmontese; with both 1 st persons and 2 nd pl. as optional. BERRUTO (1990), a native speaker of Piedmontese, confirms their findings: ... forme come veddu, veduma, vedde sono perfettamente grammaticali e alternano con i veddu, i veduma, i vedde 'vedo, vediamo, vedete', mentre non si danno vedde 'vedi', ved 'vede', veddu 'vedono' ... Se si tratti di una variazione libera o governata da regale, e tutto da indagare. (BERRUTO 1990: 19 N33). 1.3. Allowing for this measure of optionality, subject clitics (scl) are required in modern Piedmontese not only when there is no other overt subject but also, as seen above, in conjunction with tonic subject pronouns. In addition, they are obligatory with full noun phrases, regardless of whether these occur in preverbal or postverbal position 4 : Me frei a veul nen 'my brother scl wants not' la seira a finla (Mus. 126,11) 'the evening scl was-ending' ij tre amis a sonpartisne 'the three friends scl are (have) left' a l'e rivaje na fietta (Mus. 123,23) 'scl is (has) arrived a little girl' quandi eh' a arson-o ij bot (Mus. 123,17) 'when that sei resound the strokes' ('when the clock strikes') A mes-dl a-i rivava mare e magna (Mus. 126,18) 'At midday sei there arrived mother and aunt' z ltalics represent written data, in which the phonetic value of the symbols may vary according to time and place of origin, e. g. modern standard Piedmontese orthography has a for [u], whereas u is used in the Val Bormida texts. ltalics have also been maintained for non-IPA phonetic transcription, as in the originals, e.g. BERRUTO (1990) and ToPPINO (1913). 3 «Dinnanzi ai verbi, Ja lingua piemontese usa i Pronomi Personali Verbali, ehe non si possono mai omettere», BREROIBERTODATTI 1988: 72. TELMON (1988: 480) also states that Piedmontese subject clitics are compulsory. 4 The examples come from a collection of short stories (BRERO 1977) and from the Piedmontese language magazine, Musicalbrande (abbrev. Mus.). 98 M. Mair Parry The last example features an impersonal form of the verb used in an unaccusative construction; such structures are characterized by a lack of agreement between verb and following noun phrase, the occurrence of the pleonastic clitic i as well as the expletive pronoun a (formally identical with the 3 rd person [sg. and pl.] subject clitic)S. An expletive clitic pronoun occurs obligatorily in Piedmontese with all impersonal constructions, as in French (il faut, il pleut, etc.): a venta cateje doman 'scl must buy-them tomorrow' a tirava vent 'scl was-blowing wind' The third person subject clitic also occurs after the complementizer ehe of relative constructions: quaidun eh' a veul giuteve 'someone that scl wants to-help-you' dl 'an ch'a ven (Mus. 123,7) 'of the year that scl comes' ( = next year) and after quantifiers: gnun a l'ha pi nen vistlo 'nobody scl has any more seen-him' gnente a saria 'nothing scl would be' chi a parla? 'who scl talks? '; but also: chi sa? 'who knows? ' My own recordings of spoken Turinese confirm the pattern of usage found in the rather conservative, normative register of Piedmontese from which the above examples are taken. 1.4. In the dialect of Cairo Montenotte (Cairese) on the Piedmontese-Ligurian border (today spoken within the administrative region of Liguria but basically a Piedmontese dialect from the syntactic point of view) there is compulsory use of subject clitics in all persons in all the above-mentioned contexts, except after interrogative and relative chi [ki] and after the 1st person sg. tonic subject pronoun, when the subject clitic is optional: mi a parl [mi a 'porl] ti t parli [ti t 'porli] cal u parla [krel u 'porla] chila a parla ['kila a 'porla] u fioca [u 'fioka] 'it is snowing' ni aci a parluma [ni 'ofi a par'luma] vui aci i parli [vuj 'otfi i 'porli] caji i parlu ['kreji i 'porlu] chile i parlu ['kile i 'porlu] chi (u) vug ben [ki (u) vug beJJ] 'chi vede bene' chi u ven cun ti? [ki u ven ku!J ti] 'chi viene con te? ' 1.5. Although RENZI and VANELLI (1983) exclude Piedmontese from the list of dialects showing inversion of verb and subject in interrogative structures, inversion still exists in some dialects, e.g. in those of Mondovl and Biella. lt is being s For a detailed discussion in a generative framework, see BuRzro (1986), especially p. 122-26. Subject Clitics in Piedmontese: A Diachronie Perspective 99 consciously promoted in standard Piedmontese, cf. BREROIBERTODATTI 1988: 117 (note that these interrogative structures have in fact two subject clitics, one on either side of the verb): Cos i farai-ne? 'What sei shall-do sei (I)? ' Cos it faras-to? 'What sei shall-do sei (you [sg])? ' Cos a fara-lo/ la? 'What sei shall-do sei (he/ she/ it)? ' Cos i farom-ne? 'What sei shall-do sei (we)? ' Cos i farev-ne? 'What sei shall-do sei (you [pi.])? ' Cos a faran-ne? or a faran-lo? 'What sei shall-do sei (they)? ' 2. This paper will consider the subject clitics of Piedmontese from a diachronic point of view with only brief references to their syntactic behaviour. The difficulties and drawbacks of basing research on written texts, especially early texts, are well known. In the case of Piedmontese they are compounded by the fact that the few such texts that we have cannot always be precisely located geographically and/ or chronologically and often lack the full range of persons of the verb. Notwithstanding, a close examination of these texts cannot fail to throw light on the development of subject clitics in Piedmontese. 2.1. The various forms of pronouns used as subjects in early Piedmontese texts are listed below (for the 16 th and 17 th centuries, only the clitics are given) 6: a. Sermoni Subalpini, 12 th c. (traditionally regarded as the earliest manifestation of the vernacular in Piedmont, although its Piedmontese origins are by no means universally accepted, cf. WoLF 1991 and DANESI's reply [1991]) 6 The data derive from the following sources: F.A. UGOLINI, Testi antichi italiani, Turin 1942, p.11-69 (Sermoni subalpini); E.BOLLATI/ A.MANNO, «Documenti inediti in antico dialetto piemontese», AS/ 4/ 2 (1878), 375-88 (La Sentenza di Rivalta); G.BERTONI, «Note e correzioni all'antico testo piemontese dei Parlamenti ed epistole», R 39 (1910), 305-14; E.LEVI, «Il detto de! re e della regina, poemetto piemontese de! sec. XIII», in: Melanges de linguistique et literature offerts a M. Alfred Jeanroy, Reprint of Paris 1928 edn, Geneva (Slatkine Reprints) 1972, p.279-90. G.P. Cuvro, «Coj 'd San Giors: a proposit de! prim document ed vej piemonteis ch'a sia databil e localisabil», in: At de'! V Rescontr anternassional de studi an sla lenga e la literatura piemonteisa, Alba 7-8 magg 1988, Alba 1989, p.193-215 (Statuti della Compagnia di San Giorgio); G.GASCA QuEIRAZZA, Documenti di antico volgare, I: Le «Recomendaciones» de! Laudario di Saluzzo; II: Gli Ordinamenti dei Disciplinati e dei Raccomandati di Dronero; III: Frammenti vari da una Miscellanea Grammaticale di Biella (p. 19: Proverbi e Sentenze), Turin 1965-1966; C.CrPOLLA, «Antichissimi aneddoti novaliciensi», in: Memorie dell'Accademia delle Scienze di Torino, 2 nd series 50 (1901), 127-35 (translation exercises from Piedmontese into Latin); F.GA- BOTTO, «Un documento dialettale piemontese de! 1465», in: Bollettino storico-bibliografico subalpino 3 (1898), 278; G.G. ALIONE, Opera Jocunda, 1521, first edition in the Biblioteca Reale, Turin (Comedia de ! 'Homo e de soi sentimenti); C. G1ACOMINO, «La lingua dell'Alione», AG/ 15 (1901), 403-48; G.P. Cuvro, «II dialetto di Torino nel Seicento», ID 37 (1974) 18-120 (Canzoni torinesi). 100 M. Mair Parry sg. pl. I e', eu, ge, eo nos, noi II tu vos, voi III,m. el, il il III,f. ela, ella expl.7 el b. Parlamenti ed epistole, 13 th -14 th c. (Canavese? ) sg. I e, mi II III,m. al expl. al c. Il detto del re e delta regina, 14 th c. sg. IIl,m. al III,f. la expl. el pl. uuy y pl. y d. Statuti delta Compagnia di San Giorgio, 1321 (Chieri) sg. pl. II 0 III,m. al'N 8 , o/ [s] i expl. el, o/ [s] e. Proverbi e Sentenze, end 14 th c. (Vercelli area? ) sg. I yo expl. al 7 The expletive clitic pronoun is usually identical with the 3 rd person m. sg. clitic, but it may differ, as in the case of Castellinaldo i. s / indicates the phonetic context following the subject clitic; V represents a vowel, 'V an accented vowel. Subject Clitics in Piedmontese: A Diachronie Perspective f. Ordinamenti delta Confraternita di Dronero, 14 th -15 th c. sg. pl. I noy III,m. al, ar, (e)l 9 y, i III,f. (e)la expl. al, ar g. La Vita dei Raccomandati delta Vergine, 14 th -15 th c. sg. pl. I noy III,m. i, hy, y III,f. (e)le expl. (e)l h. Le 'Recomendaciones' del Laudario di Saluzzo, 14 th -15 th c. sg. pl. I noy, noe III,m. (e)l y, i III,f. (e)la expl. (e)l i. Translation exercises Piedmontese > Latin, 14 th -15 th c. sg. III tu III,m. a expl. al 101 9 The brackets indicate that groups such as sei, chel, chela may represent either s'el, ch'el, ch'ela or se'l, che'l ehe la. 102 M. Mair Parry j. Sentenza di Rivalta, 1446 (court verdict) sg. pi. I noy III,m. (e)l, lN hi III,f. a le expl. o! [s],ha lN,he lN k. Atto di Poirino, 1465 (legal document) 1. Comedia de ! 'Homo and Parse carnovalesche, G.G. Alione, beg. 16 th c. (Asti) sg. proclitic sg. enclitic I e -i, -y II te, t -tu III,m. el,o,lo,alN,l/V -lo III,f. la, ra, a, a lN -la expl. el, o, a lN, lN m. Canzoni torinesi, 17 th c. sg. I i II t III,m. a, all! 'V III,f. a, all!'V) expl. a, all! 'V, lN -la -lo pi. proclitic pi. enclitic e -y 0 -o, -vo i, y -gle el, lN -gle pi. i 0 a, a lN a, all!'V 2.2. Subject pronominal syntax in the medieval stage of Northern ltalian dialects has been studied in depth by L.Vanelli, L. Renzi, and P.Beninca. A brief summary of their findings will serve as a background to the discussion of the development of the Piedmontese forms. As a rule, affirmative main clauses in medieval texts from Northern ltaly (Venetian to a lesser extent) follow the Verb-second rule: the verb is preceded either by the theme (this may be but is not necessarily the subject) or, in the absence of a theme, by an adverbial phrase. lf the subject is not the theme, it Subject Clitics in Piedmontese: A Diachronie Perspective 103 occurs in immediate postverbal position. An overt subject (either lexical or pronominal, but not both) is obligatory in all clauses, except in those main clauses in which another constituent precedes the verb. As the research of VANELLI, RENZI and BENINCA (1985, also VANELLI 1987) has shown, null subjects are noted only in those contexts in which the subject would otherwise follow the verb, i.e. in main clauses in which some element other than the subject precedes the verb (postverbal subjects are rare in subordinate clauses). 2.3. From the morphological point of view, it is clear that the unstressed subject clitics of the modern dialects derive from the nominative of the Latin personal pronouns 10 , whereas the tonic free forms derive from the oblique stressed forms. In the Medieval period Northern Italian texts show only one series of subject pronouns, those deriving from the Latin nominative, which were free and could be stressed. However, as shown by VANELLI (1984: 285-87), oblique forms of the 1 st and 2 nd sg. and 3 rd sg. and pl. could be used as subjects in certain syntactic contexts, i.e. when not governed by the verb, e.g. with non-finite verb forms; when there is a conjoined subject 11 ; when the subject does not form part of the main sentence; when the verb is not expressed. The 1 st and 2 nd pl. pronouns did not have phonetically distinct alternative oblique forms. When, during the Renaissance period, the originally nominative forms gradually cliticized to the verb, it was the oblique forms that were pressed into general use as tonic, free subject pronouns, first in complementary distribution with the weak pronouns and later in combination with them. Eventually the clitics are no langer perceived as filling the subject slot but 10 Evolution from the Latin personal pronouns to proclitic forms found in Piedmontese at various stages of development can be broadly schematized as follows (V and C represent respectively Vowel and Consonant): EGO> [eo]> [io]> [jo] > [i], [j]N > [e] (Castellinaldo),> [a] (Cairo) TU> [ty]> [t]> ( + prosthetic vowel? ) [it] NOS> [nuj],> ? [n] (PIPINO 1783) vos> [vuj],> [vu],> [u] Third person pronouns developed from the Latin distal demonstrative: ILLE> [el],> [al],> [a]/ C > [ew] before dental and palatal C> [u] (Monferrato) > [eJ] before labial and velar C (Monferrato) (see §4.1.) ILLA> [ela]> [la]> [a] ILLI> [il]> [i], [j]N ILLAE> [ele]> [le] > [el] (Astigiano, Alione). 11 As in the first letter of the Parlamenti ed epistole: «Sapiente per lo certo ehe uuy e-mi ezaschaun nostro successor e areo n-ara semper may gram loso e-honor» (BERTONI 1910: 308). 104 M. Mair Parry function as subject agreement markers (part of the inflection of the verb), with the result that Northern Italian dialects are today classified as Null Subject languages (BRANDI/ CORDIN 1989). 3. Examination of the subject pronouns found in Piedmontese texts reveals that there has been a progressive neutralization of differences between the various persons, especially between the singular and plural. In Turinese the 1 st persons (sg. and pl.) have an identical form, i (e), while 3 rd sg. and pl. are both a; the 2 nd pl. i (e) is also identical with the 1 st persons. For the dialect of Castellinaldo, ToPPINO (1913: 5) gives the following: sg. proclitic sg. enclitic pl. proclitic pl. enclitic I e -ni e -ni II t -ti i -vi/ -vu III,m. i f. ula* 1 a -Zu 1-la i -ni IIl,expl. a, (u), i** * N. B. u r-e 'he is', a r-e 'she is'; ** i is often used as an expletive pronoun with inverted subjects and with impersonal verbs: i ru dis kjal 'expl. scl it says he' ('he says it') i r a diru kila 'expl. sei it has said-it she' ('she has said it') The same phenomenon is found in the more southern dialect of Cortemilia: ii n'e riviiine tre [ij ne ri'vojnE trE] 'expl. sel-there of-them-is arrived-there-of-them three (f.)', 'three of them (f.) have arrived' is sa miii [is so moj] 'expl. sei-SE knows never','one never knows' In Cairese also (see above, §1.4.) the 2 nd pl. is identical with the 3 rd pl. and distinct from the 1 st persons. In all systems it has lost its distinctive form, although ToPPINO (N4) refers to u as being «tutt'ora vivo a Vezza» (near Alba). 3.1. The traditional view held by Romance linguists (e.g. BouRCIEZ 1956: §558, EwERT 1943: §43) was that the use of subject pronouns had become compulsory in order to disambiguate verb endings that had become homophonous as a result of regular phonetic development; the French system was a good illustration of this. Nowadays the obligatory use of subject pronouns is generally ascribed to syntactic causes (PRICE 1971, HARRIS 1978, VANELLI/ RENZIIBENINCA 1985 12), although 12 «Seules les langues du deuxieme groupe [incl. French and N. Ital. dialects,MMP] demandaient [i.e. dans 1a phrase subordonnee] le pronom sujet de fai;on obligatoire. C'est la le noyau du Subject Clitics in Piedmontese: A Diachronie Perspective 105 distinctive verb endings have contributed to the non-emergence or optionality of subject clitics with certain persons of the verb in some dialects. However, the compulsory use of subject pronouns in French meant that the erosion of distinctive verb endings was allowed to proceed unhindered (cf. PRICE 1971). Northern Italian dialects have instead tended to maintain a higher degree of differentiation between the various personal suffixes of the verb, a fact that supports the theory of the syntactic origin of the compulsory use of subject pronouns. RENZI and VANELLI (1983: 133) highlight the fact that the one subject pronoun found in all the varieties examined is the 2 nd sg., which has one of the most stable verb endings (almost as stable as the 1 st and 2 nd pl., those persons in which the subject pronouns are most frequently missing). BENINCA (1986: 466) notes that in Friulian compulsory subject clitics exist alongside «una flessione verbale ehe e la piu ricca di distinzioni in assoluto nel dominio romanzo», and notably more differentiated than the subject clitics. 3.2. One conclusion that may be drawn from these data and from the Piedmontese examples that we have been considering is that, if the verb retains a high degree of inflection, then the subject pronouns themselves may lose their distinctive form. None of the varieties examined by RENZI and VANELLI (1983) has subject clitics with a different phonetic form for each of the six persons of the verb. For those varieties that display subject clitics for each person of the verb, they individuate the following patterns: (a) the 1 st sg., i st pl. and 2 nd pl. are identical, with 3 rd sg. and pl. being different from each other or, as in Turinese, identical; the 2 nd sg. is distinct; (b) the 2 nd sg. and 3 rd sg. may be the only distinct forms, with all the rest identical; (c) 1 s t sg. and pl. are identical as are 2 nd pl. and 3 rd pl., while 2 nd sg. and 3 rd sg. are distinct: this is the pattern found in Castellinaldo, but for the distinctive expletive clitic i, and also in Cairo Montenotte, except for the fact that Cairese 3 rd f. sg. is identical with 1 st sg. and pl. The dialect of Cortemilia, located between these two areas, presents a combination of these two variations on the basic pattern: sg. pl. I a a II t i IIl,m. \ f. u\a i III, expl. i 3.3. Regardless of their increasing failure to identify person, the subject pronouns become ever more necessary components of the verb syntagm. By the end of the pronom personnel obligatoire (et clitique) de ces langues romanes», (VANELLI/ RENZIIBENINCA 1985: 165). 106 M. Mair Parry 18 th century, Piedmontese shows cases of subject clitic doubling: they are being used with tonic pronouns, with full lexical NPs, and with quantifiers: el vost talent a l'era bon '(the) your talent sei was capable' Mi j'eu l'obligassion 'I sei have the obligation' ognun a l'ha scrit 'everyone sei has written' (PIPINO 1783, letters). Where verb endings do become homophonous, as is frequently the case for the 2 nd sg. and pl., the subject clitics for those persons are always distinctive 13 • Although no textual evidence is available for Turinese, neutralization may have occurred first in the 1 st person pronouns 1 4. This is the case in the early-16 th -century text by Giovan GiorgioAlione ofAsti, La Comedia de l'Homo (see § 2.1. (k) above). Here the two proclitic i st person subject pronouns are identical, as are the two enclitic pronouns 15 , while the 3 rd person pronouns are distinct. In Turinese the 3 rd pl. endings of the present tense of regular verbs became homophonous with 1 st sg. as a result of the spread of the -UNT suffix and the fall of the final [n]; in the imperfect tense the 3 rd pl. is homophonous with the l st pl., as it is in both the present and imperfect subjunctive paradigms. A desire to resolve possible ambiguity between 1 st sg. and 3 rd pl. may well have contributed to the replacement, already evident by the 17 th century, of the 3 rd person pl. subject clitic i, by a. a was presumably generalized from the 3 rd sg., possibly following the model of neutralization of the 1 st persons which already had identical forms. No ambiguity results from the generalization of the 3 rd sg. clitic to the plural, since the verb forms always have a different suffix, e. g. Tur. a parla 'he/ she/ it speaks' ~ a parlo 'they speak'. 3.4. A weak form of the 1 st pl. subject pronoun (< Latin NOS) would seem to have developed in Piedmontese in enclitic position, standard -ne, Castellinaldo, -ni, but so far I have no textual evidence of a proclitic form except for the following reference in M. PIPINO's grammar (1783: 31): «Talvota poi anche in vece di ti, noi, voi, sentesi soltanto il suono d'una t', d'una n', e d'una v' con apostrofo, onde in questi casi cosl si scrivera» 1 6. Such a pronoun still exists in the Franco-Provern;:al 13 «Nelle varieta esaminate la neutralizzazione completa di due o piu persone non e ammessa. Dunque le persone della flessione verbale sono differenziate tra di loro o con il morfema desinenziale o con il pronome soggetto o con entrambi» (RENZINANELLI 1983: 133s.). 14 See also VANELLI 1984: 289s. for further observations on the generalization of the 1 st sg. subject clitic to the plural. This is a common feature of Northern French dialects also (cf. ALF Maps 27, 91, 506); the distinctive suffix of the plural form prevents ambiguity. 1s Both sets of forms, e and -i, -y, presumably derive from EGO, the vowel of the former having opened in proclitic position. 16 This statement follows a paragraph explaining the use of i and j in place of mi, ti, noi, voi, with clear examples of subject function: 11 pronome i, ehe da noi usasi in luogo dei pronomi mi, ti, noi, voi, e talvolta si usa per particella riempitiva, si esprimera conj' consonante apostrofato, quando verra preposto a Subject Clitics in Piedmontese: A Diachronie Perspective 107 dialects of the Val D'Aosta (ROBERTS 1990): e.g. Chätillon, N'en minja, St Nicholas, N'en medja 'We have eaten'. The proliferation of proclitic n proforms (complement clitic< Nos, partitive< INDE, and negator< NON, in addition to the subject pronoun) could well have encouraged the early generalization in preverbal position of sg. i to the plural. The phonetic development of the enclitic form -ne is not clear (Castellinaldo has -ni); an original [-nu]< Nos (Fr.-Prov. St Nicholas has enclitic -no) could have been influenced by the corresponding complement clitic, ne, itself modelled on me, te, and possibly also on ne < INDE 17 • Enclitic -ne remained as an interrogative and optative pronoun, then spread presumably in the first instance to the 1 st sg. (following the neutralization of the proclitic forms), with the result that in the 1 st person the number distinction is replaced by a semantico-syntactic distinction between a preverbal (unmarked) clitic on the one hand (i) and a postverbal (interrogative and optative) clitic on the other (-ne). Both clitics were later generalized to the 2 nd pl. 18 and ne also to 3 rd pl. As may be seen in the data from the early texts, vos gave o [u] proclitically, presumably via [vu] (I have no textual evidence for the v' to which Pipino refers) and 18 th -century texts also show the interrogative enclitic subject form -ve (presumably showing the same analogical phonetic development as -ne): Cosa volive d'pi? 'What more do you want? ' (Ignazio Isler). In those Turinese 2 nd pl. verb forms that were reduced to monosyllables, the enclitic -ve according to Salvioni (cited by ToPPINO 1913: 19) has lost its pronominal status, becoming fully integrated into the verb form, which has been generalized to all contexts: i seve 'you are' 1 9 i l'eve 'you have', (also i deve, i steve, i feve in the present paradigms of de 'to give', ste 'to be, remain', and fe 'to do'). verbi, ehe principiano con lettera vocale, o coll'aspirazione h, e percio scriverassi per esempio j'amo, j'hai, j'avomo, j'avi, ecc. perche il suono di questo in simili casi si ode. The following table lists subject clitics referred to or found in Pipino's grammar and letters: sg. pl. I i, j i, j, n' II i, t' i, j, v' III, m., f. a, a l'N, l'N a, a l'N, l'N expl. a, a l'N 11 GrACOMINO 1901: 433 also favours a development from NOS (similarly ve from vos), whilst ALY-BELFADEL (1933: 162) rejects a proposed continuation of Latin interrogative-NE on the grounds that -ne does not appear in all persons of the verb. 1s According to BRERO and BERTODATTI (1988: 117s.) the use of the 2 nd pl. enclitic -nein interrogative structures is optional. 19 Compare the forms given by PrPINO (1783) and ALY-BELFADEL (1933); cf. §4.3. below. 108 M. Mair Parry 4. The forms for the 3 rd person clitics are interesting, particularly in view of the differing interpretations that have been placed on the [l) that occurs before a vowel in the verbs 'to be' and 'to have'. Northern Italian dialects have recently been fertile sources of data for generativists and for testing various hypotheses concerning the status and function of clitics. BuRzro (1986: 172 N47), for example, referring to the following Piedmontese examples: A ! an chersü tüti i presi '(they) scl have increased all the prices' (ib., 123s.) A l'e rumpüse due fnestre 'expl.-scl is broken-themselves two windows' ('Two windows have broken') explains the las an euphonic sound inserted by a phonological rule between subject clitics and the verbs 'be' and 'have'; in some environments it is realized as a glide [j]. Commenting briefly on similar examples in Genoese, u 20 l'e arrivou 'he has arrived', a l'e arriva 'she has arrived', BATTYE (1990) observes that the l cannot be accounted for by the need to avoid a hiatus between the subject clitics u and a and the 3 rd person form of the auxiliary verb, since the absence of these clitics does not remove the need for the [l): le arrivou u Paulu 'Paul has arrived'. ROBERTS (1990) refers to the variation in 3 rd sg. subject pronouns in the Franco-Proven�al dialect of Brusson, Val d'Ayas: [u] occurs before consonants, [l) before vowels: ou mindja 'scl (he) eats' l'a mindja 'scl (he/ she) has eaten' In many Val Bormida dialects on the Piedmontese-Ligurian border today an l clitic follows the usual 3 rd sg. subject clitic before the verb 'to have' and vowel-initial forms of 'to be', in both the full and auxiliary verb functions, e.g. Cairese: [u Ja maJJ'd30] 'scl (he) has eaten' [u IE JJ'da] 'scl (he) is (has) gone' [u 'Iova] 'scl (he) had'. Some 150 years ago, however, as we may see in the Cairese version of the Parable of the Prodigal Son (BroNDELLI 1853: 554), there existed a similar variation to that found in the Franco-Proven�al dialect: [u] before consonants, [l) before vowels: u s'e trovase 'scl (he) SE is found-sE' 'he has been found' l'a die 'scl has said', l'ava 'scl had' but u l'a vist a vni (l here represents the direct object) 'scl (he) him has seen to come' ('he saw him coming'). Some Val Bormida dialects retain this variation to this day, e.g. Osiglia, 20 In these Ligurian examples u represents [u]. Subject Clitics in Piedmontese: A Diachronie Perspective L'om u s ciamova Bepen e l'e pensiuno 'The man scl was called B. and scl (he) is retired ... l'e ndo ... '...scl (he) is gone (went) ...' 109 4.1. In all these varieties the l [l] derives from the same source, namely the Latin masc. sg. demonstrative ILLE. This gave the form [el], but in some dialects the vowel opened in unstressed position to give the open central vowel: [al] (cf. VANELLI 1984: 289). In Turinese, [l] fell before a consonant: [al] > [a], but remained before a vowel, since VCV is a more natural phonological pattern than VV: al'e. As a result of the frequency of the pre-consonantal allomorph a, the syntagm al ewas reinterpreted as a + l'e 21 • In some dialects, e.g. in the Monferrato, the [l] vocalized before dental and palatal consonants: [el] > [ew] > [u], remaining [el] or [er]/ [e.:r] before labials and velars, as in Alione o sa 'he smells' as opposed to el va 'he goes' (cf. GIACOMINO 1901: 433). Over two and a half centuries later, poems in the late eighteenthcentury dialect of Asti reveal a similar phonetically-conditioned allomorphy in the 3 rd m.sg. subject clitic: ar parlava 'he was speaking' ~ au diva 'he was saying' (GASCA QuEIRAZZA 1990: 102). Further investigation is necessary to ascertain whether the a is to be attributed to the increasing influence of the Turinese dialect as the dominant variety in Piedmont; ar and au may represent cases of accretion of the Turinese clitic, while a is apparently the only form used before the negative clitic or complement pronouns other than lo (before which ar is used). In the modern dialect of Agliano, a little to the south of Asti, the original phoneticallyconditioned allomorphy remains vital: [eJ 'porla] 'he talks' [u t 'porla] 'he talks to you' [eJ 'pjuviva] 'it was raining' [u 'suJJa] 'it rings' [u 'seJJt] 'he hears' The verbs 'to have' and 'to be' display: [a 'm] 'he/ she has', [a 'Je] 'he/ she is' Despite the lack of textual evidence for similar phonetically-conditioned allomorphy in the verb paradigms of the Val Bormida, it seems logical to assume, on the basis of precisely this kind of allomorphy in the cognate definite article (as in the dialect of Asti), that it did once exist in this area too. If so, in Cairese and other dialects of the Val Bormida on the southern edge of the Monferrato, the u would then have generalized as the 3 rd m.sg. pronoun, so as to occur also before labial and velar consonants: [u 'porla] 'he talks', [u kodz] 'he falls' 22• Structures such as [u la 21 Examples of a le may be found in Alione where, according to GrACOMINO 1901: 433, they constitute borrowings from Turinese. 22 Ligurian influence may have contributed to the generalization of [u], but data from Agliano show that the alternation of [u/ u] no langer obeys strict phonological rules even here, [ EJ] is 110 M. Mair Parry vift] 'he has seen' would seem to derive from a further generalization of [u] to what was originally a prevocalic context: u was preposed to the subject clitic l, which naturally occurred there: Cairese, l'e > u l'e; l'a > u l'a; a similar generalization of the preconsonantal subject clitic occurred in the case of the feminine as well: l' e > a l'' e. 4.2. The subject clitic u of the Ligurian dialects which, unlike Piedmontese dialects, retain unstressed final vowels, is generally considered as having developed from ILLU by progressive aphaeresis: ILLU > [lu] > [m] > [u] (cf. PARODI 1899-1901: 19). [lu] occurred before all consonants but before vowels the [u] elided leaving just [1] (similarly ILLA > [la] > [1]/ V: l'e, l'a etc). As in the case of Piedmontese, one may assume that the preconsonantal forms u and a were later generalized to this context too, to give u l'e arrivou, a le arriva, etc. 4.3. The finite forms of the standard Piedmontese verb avej 'have' all now begin with [1], as can be seen in the grammars of ALY-BELFADEL (1933) and BRERO/ BERTODATTI (1988). Since 1783, when Maurizio Pipino wrote his grammar, there has been a generalization of the [1] from the 3 rd person: no such generalization has occurred in Ligurian. The present and imperfect indicative paradigms of the verbs 'to be' and 'to have' in the grammars of Pipino and Aly-Belfädel are given below as points of reference for the following discussion: Gramatica piemontese, M. PIPINO 1783, 39-45 (see also note 10 for forms of subject clitics not included in the tables). esse ('to be') aveje ('to have') mison noisomo mij'eu / j'hai noiavomo/ omo tit'ses voise tit'as voiavl col l'e coison col l'ha coil'han mij'era noiero mij'avia noiavfo tit'ere voiere tit'avfe voiavfe col l'era coil'ero col l'avfa coil'avfo replaced by [u] not only before the dental object clitic t 'you (sg.)' but also before bilabial and labiodental object clitics such as m 'me' and v 'you (pl.)', e.g. um beika 'he looks atme', uv beika 'he looks at you'. Tue precise conditioning awaits more detailed investigation, but it is possible that in contexts where consonant-final [EJ] would produce an unacceptable duster, e.g. *[EJ m'bejka] 'he looks atme', speakers have generalized the use of [u], rather than insert a vowel to facilitate pronunciation. Subject Clitics in Piedmontese: A Diachronie Perspective 111 Grammaticapiemontese, A.ALY-BELFADEL 1933: 183s. ese avej mi i sun nuj i suma mi i Z'aj nuj i l'avuma! Z'uma ti i't ses vuj i seve ti't l'iis vuj i l'eve chieZ a Z'e Zur a sun chieZ a Z'ii Zur a Z'iin mi i j'era nuj i j'eru mi i Z'avija nuj i Z'aviju ti i't j'ere/ eres vuj i j'ere ti't l'avijelavijes vuj i Z'avije chieZ a l'era Zur a j'eru chiel a l'avija Zur a Z'aviju The traditional explanation for the generalization of [l] is that this took place for euphonic reasons: to avoid a hiatus or the reduction of [i] to [j]: cf. j'eu in Pipino's grammar. If it is a matter of euphony alone, then it is not clear why [l] was not generalized in the imperfect paradigm of ese. Here we find a generalization of [j] to all persons except the 3 rd sg., which retains [l]. The generalization of [l] in the verb 'to have' is a fairly recent phenomenon: at the end of the 18 th century, in Pipino's letters we find that it has spread only to the 3 rd person pl. along with the a: a l'han. At this time an expression such as i l'eu finilo 'I have finished it' is definitely an instance of the common phenomenon of clitic copying in compound tenses: 'I it have finished-it'. At this stage of Piedmontese, the complement clitic not only appears before the auxiliary, as is normal in the Romance languages, but also appears enclitically attached to the past participle: a mavia dame 'She me had given-me'; a l'han sempre usalo 'They it have always used-it'. (Pipino) 4.1. I propose that these structures, rather than simple euphonic considerations, played a major role in the spread of l, firstly in connection with the auxiliary use of avej in compound tenses. lt is a complex process, since the clitic doubling structures themselves may owe their origin in part to the / 2 3 • The structure may have evolved through the following stages: 23 To date the most convincing explanation for this reduplication of the object pronoun clitics, which in most Piedmontese dialects is later rejected in favour of enclitic positioning on the past participle only (a parallel elimination of redundancy involves Piedmontese negative structures also, cf. PARRY 1989) is that of BENuccr 1989 and 1990: through a process of <destructuring> (destrutturazione, the reverse of Rizzi's restrutturazione (cf. Rrzzr 1976 and 1982) the complex verb of periphrastic constructions (compound tenses as weil as modal/ aspectual structures with the infinitive) is broken down into its two components, with the result that the object clitics attach to the (non-finite) verb of which they are the arguments, e. g. [a vu'ria mus'tremlu] 'he/ she wanted to show it to me' instead of *[a mlu vur'ia mus'tre]. Reduplication is a transitional stage, during which the object clitics still climb to the higher verb but their deep structure link with the embedded infinitive, instead of featuring as a phonologically-null trace, is represented in surface structure by on overt ,copy>. In the compound tenses the past participle to which the pronoun cliticizes never agrees in number or gender with the preverbal object clitic, so that, developing 112 M. Mair Parry up to the 17 th century j'eu fini 'I have finished' i l'eu fini 'I have finished it' a l'ha fini 'he has finished' and 'he has finished it'24. By the 18 th century, for syntactic reasons and also to avoid ambiguity, the complement clitic comes to be copied onto the non-finite verb form, the past participle (cf. PARRY, forthcoming a and b). Thus a I'ha finilo 'he has finished it', a m'ha fame pense 'it made me think', i l'eu fimlo 'I have finished it'. In these examples the preverbal l represents the object pronoun. However, on the basis of a l'hafini 'he has finished' (in which the l derives from the subject pronoun), a l'hafinilo 'he has finished it', instead of being interpreted syntactically as a case of clitic copying, comes to be interpreted as a l'hafini + object lo. The extension of this structural analysis to i l'eufinilo etc., with the loosening of the link between the preverbal l and the object lo/ la, would have opened the way for the widespread generalization of preverbal l in the verb avej. The l then comes to be used, in the 1 st as well as in the other persons, in the absence of an object and serves to keep the ideas put forward in ROBERTS (forthcoming), it may be possible to anaJyse the copies as agreement eJements. In Gallo-Italian dialects other than Piedmontese the process of ,destructuring> has been limited to the infinitival constructions; the characteristic Piedmontese encliticization of object pronouns on the past participles of compound tenses may owe its fortune and grammaticalization to the desire to avoid ambiguity in the third person structures. A further influentiaJ factor couJd have been the topicality of personal pronouns, which often leads to overuse in popular and colloquial registers (see TurrLE, forthcoming), although there a number of reasons for excluding the possibility of its having been the root cause. 24 Tue occurrence of two l's, found in eighteenth-century and Jater texts (e. g. Brofferio: A 'l l'a fin-a mna con chiel, 'He even took her with him' [BRERO 1982: 115)) and noted by TorrrNo (1913: 6 Nl): «11 tor., ehe nella III persona usa a per ogni genere e numero, viene quindi in tali casi ad avere un al ben distinto, ehe si mantiene, anche fuor dei casi indicati, quando segua altro l di pronome, dando cosl. luogo a -ll-: kwand k-al l avrd 'quando J'avra', a pena k-al lu sapia 'appena Jo sappia', al lu ved 'lo vede')» and ALY-BELFA.DEL (1933: 155: a 'l l'a per a lu l'a [J'ha], a 'l l'eper a lu l'e [Jo e]), (cf. also the maintenance of ar before lo in the dialect of Asti, mentioned above) reflects resistance to the neutralization of the difference between, on the one hand [aJ a] 'he has' (later reinterpreted as [a Ja]), and on the other, [a Ja], the expected reduction of [aJ Ja] 'he has it'. lt is not surprising that, as a result of the Jater accretion of l to all finite forms of avej, the two l's come to be reinterpreted, the first now being perceived as the object clitic, the second as part of the verb, as ALY-BELFA.DEL maintains (1933: 167). If, however, a complement clitic occurs with the verb avej, it still causes the l to disappear, e. g. Mare Granda, se i veule e i 'n n'eve da manca, i vad a pijeve j'uciaj, Jit. 'you of-them have need' 'Grandmother, if you wish and you need them, 1'11 go and fetch your gJasses' (BRERO 1988: 78). This written exampJe interestingJy shows a tendency to represent the clitic (here, n < INDE) as double, a fact which calls into question the vaJue of textual examples of a'l l'a etc. as indicators of actuaJ pronunciation. Subject Clitics in Piedmontese: A Diachronie Perspective 113 subject clitics morphologically distinct from the verb form. In the Piedmontese poets Edoardo Ignazio Calvo and Angela Brofferio, writing at the end of the 18 th and in the first half of the 19 th century respectively, we already find a few instances of l generalized beyond the 3 rd person: Calvo: i spero ch'i l'avroma fin1 d'core 'We hope that we shall have finished running' Brofferio: Mi ch'i l'eu pre'esse promoss/ D'grossi titoli 'I who sei have, for being promoted, important qualifications' The proclitic l thus loses both the link with the 3 rd person subject pronoun and with the 3 rd person object clitic, giving the modern forms i l'hai trovate 'I have found you (sg.)', i l'uma trovave 'We have found you (pl.)', in which it replaces the earlier proclitic complement pronouns t' and v' of the clitic copying structures. However, as noted above, the l is not yet completely lexicalized and is still considered a separate particle from the verb form, as is evident from both the graphy, l', and from the fact that when avej is used with its full semantic value, it is replaced by any occurring proclitic complement pronoun (ccl): In'av1a na barba (sei cei Verb; lit. 'I with-it had a beard') 'I was bored stiff'. contrast Il'av1a 'I it had'. 4.3. Euphony has also been proposed as the explanation for the accretion of Li] to forms of the verb ese 'to be'. TüPPINO (1913: 15), referring to the imperfect indicative forms of Castellinaldo, writes, «si osservi per altro ehe a tali forme va sempre premesso un j parassitico quale estirpatore di iato». He suggests that the Li] is a generalization of the proclitic locative adverb [i] 'there', but other possible sources are the 1 st sg. and the 3 rd pl. subject clitics. As in the case of the generalization of / , it seems more satisfactory to consider euphony as a by-product of more fundamental morpho-syntactic pressures. In Pipino's grammar we findj occurring in front of the 1 st sg. of avej: j'eulj'hai, whereas in modern Piedmontese we have i l'hai 'I have'. A comparison of the earlier Turinese form with that of the dialect of Cairo Montenotte, aj'eu [a j!ll] 'I have', suggests the possibility that the two j are the 'same' clitic, deriving from EGO. I propose that in Cairese, as a result of the accretion to the verb form of the 1 st sg. and pl. prevocalic subject clitic j (j'eu, j'uma), prevented in Turinese by the spread of / , the preconsonantal 1 st sg. subject clitic a was generalized to fill a slot which was perceived as vacant (just as we saw above that in the 3 rd sg. u now occurs in front of the original prevocalic clitic / ). Similarly, the 1 st person imperfect indicative forms of the verb esci 'to be', j'era 'I was', j'ermu 'we were', acquired an extra proclitic a: aj'era, aj'ermu. In Turinese, of course, it is the preconsonantal form i that has been generalized in front of j: i j'era, etc., while the j has spread also to the 2 nd sg.: it j'ere 'you were', possibly influenced indeed by structures involving the locative adverb i. lt has been suggested that the initial i of the 2 nd person subject clitic, it, is a separate clitic (see Aly- 114 M. Mair Parry Belfädel's representation as i't 25 ), functioning as a <first position> clitic (BENINCA 1986: 469s.). If so, it could be the result of a generalization of 1st sg. i or it could have emerged originally as a prosthetic vowel (VANELLI 1984: 288), influenced by an incorrect division of the syntagm, ti't (> t'it), with subsequent reconstitution of the tonic pronoun (ti it). 5. The neutralization phenomena witnessed among the subject clitics of Northern ltalian dialects highlight the interesting possibilities of parametric variation in language development: standard French on the one hand and Northern ltalian dialects on the other reacted in quite different ways to the redundancy presented by the fact that the person of the verb came to be marked not only by the verb endings but also by compulsory subject pronouns. Standard French allowed verb endings to atrophy, whilst in Northern ltaly, despite the otherwise widespread loss of unstressed final vowels except [a], the .verb endings retained a greater measure of differentiation and instead several subject pronouns lost their distinctive form. Although often redundant with regard to person marking, the subject clitics of Northern ltalian dialects are clearly crucial to the structure of the sentence, as demonstrated by the appearance of additional forms to make good any loss of morphological distinctiveness. Their configurational role, as well as the reasons for the difference in evolution between these dialects and standard French, will be subjected to more detailed scrutiny in a further study. Aberystwyth Bibliography ALY-BELFA.DEL, A. 1933: Grammatica piemontese, Noale M. Mair Parry BATTYE, A. 1990: Quirky agreement in Genoese. Paper read at the Crucial Languages Seminar, University of Geneva, 6-7 July BENINCA, P. 1986: «Punti di sintassi comparata dei dialetti italiani settentrionali», in: G. HoLTus/ K. RrNGGER (ed.), Raetia antiqua et moderna. W. Th. Elwert zum 80. Geburtstag, Tübingen, p. 457-79 BENUCCI, F. 1989: «'Ristrutturazione', 'destrutturazione' e classificazione delle lingue romanze», in: Medioevo Romanzo 14: 305-37 BENuccr, F. 1990: Destrutturazione. Classi verbali e costruzioni perifrastiche nelle lingue romanze antiche e moderne, Padua BERRUTO, G. 1990: «Note tipologiche di un non tipologo sul dialetto piemontese», in: G. BER- RUTo/ A. A. SOBRERO ( ed.) Studi di sociolinguistica e dialettologia italiana offerti a Corrado Grassi, Galatina, p. 3-24 BIONDELLI, B. 1853: Saggio sui dialetti gallo-italici, Bologna (photographic reprint 1970) 2s PrPINO (1873: 31) also suggests such an analysis when he refers to the substitution of (tonic) ti by i. Subject Clitics in Piedmontese: A Diachronie Perspective 115 BouRcrnz, E. 51967: Elements de linguistique romane, Paris BRANm,L./ CoRDIN,P. 1989: «Two Italian dialects and the Null Subject Parameter», in: 0. JAEGGu/ K. SAFIR (ed.), The Null Subject Parameter, Dordrecht, p.111-42 BRERO,C./ GANDOLFO, R. 1967: La letteratura in piemontese. Dalle origini al risorgimento, Turin BRERO, C. 1977: Conte, faule, e legende piemonteise, Turin BRERO, C. 1982: Storia della letteratura piemontese, Turin BRERO, C./ BERTODATTI, R. 1988: Grammatica della lingua piemontese, Turin BRERO, C. 1989: Sintassi dla lenga piemonteisa, Turin BURZIO, L. 1986: ltalian Syntax: A Government-Binding Approach, Dordrecht DANESI,M. 1991: «The Language of the Sermoni Subalpini Revisited: A Reply to Wolf», in: G.P. Cuvro/ C. PICH (ed. ), At di!l VII Ri!scontr anti!rnassional di! Studi an sla Lenga e la Literatura piemonteisa, Alba, 12-13 magg 1990, Alba, p. 255-62 EwERT, A. 21943: The French Language, London GAscA QuEIRAZZA, G. 1990: «Documenti de! piemontese di Asti nel secondo Settecento: sonetti per il Palio», in: G.P.Cuvro/ C. PrcH (ed.), At di!l VI Ri!scontr anti!rnassional di! Studi an sla Lenga e la Literatura piemonteisa, Alba, 6-7 magg 1989, Alba, p.85-108 GIACOMINO, C. 1901: «La lingua dell'Alione», AG/ 15: 403-48 HARRIS, M. 1978: The Evolution of French Syntax: A Comparative Approach, London PARODI, E.G. 1891-1901: «Studi liguri II», AG/ 15: 1-82 PARRY,M.M. 1985: The Dialect of Cairo Montenotte (SV), unpubl. Ph.D.dissertation, University of Wales PARRY, M.M. 1989: «Strutture negative nei dialetti piemontesi», in: G.P. Cuvro/ C. PrcH (ed.), At di!l V Ri!scontr inti!rnassional di! Studi an sla Lenga e la Literatura piemonteisa, Alba, p.169-77 PARRY,M.M. (forthcoming a): «Some Observations on the Syntax of Clitic Pronouns in Piedmontese», in: M. MAIDEN/ J.C. SMITH (ed.), The Romance Languages and Linguistic Theory, Amsterdam PARRY, M.M. (forthcoming b): «Posizione dei clitici complemento nelle costruzioni verbali perifrastiche de! piemontese», in: G.P. Cuvro/ C. PrcH (ed.), At di!l VIII Ri!scontre anti!rnassional de Studi an sla Lenga e la Literatura piemonteisa, Alba, 4-5 magg 1991, Alba PIPINO, M. 1783: Gramatica piemontese, Turin PRICE, G. 1971: The French Language: Present and Past, London RENZI, L.N ANELLI,L. 1983: «I pronomi soggetto in alcune varieta romanze», in: P.BENINCA et al. (ed.), Scritti linguistici in onore di Giovan Battista Pellegrini, vol.1, Pisa, p.121-45 Rrzzr, L. 1976: «Ristrutturazione», Rivista di Grammatica Generativa 1: 1-54 Rrzzr, L. 1982: Issues in ltalian Syntax, Dordrecht ROBERTS, I. 1990: Inversion and subject clitics in Valdotain. Paper read at the Crucial Languages Seminar, University of Geneva, 6-7 July ROBERTS, I. (forthcoming): Past Participle Agreement and Object Clitics in Franco-Proven<;al Valdotain TELMON, T. 1989: Italienisch: Areallinguistik II. Piemont, in: G. HoLTus/ M. METZELTIN/ C. SCHMITT (ed.) Lexikon der Romanistischen Linguistik, Vol.4: Italienisch, Korsisch, Sardisch, Tübingen, p.469-85 ToPPINO, G. 1913: «II dialetto di Castellinaldo», Studi romanzi 10: 1-48 TuTTLE, E.F. (forthcoming): Dei pronome d'oggetto suffisso al sintagma verbale: nota in calce ad una nota salvioniana de! 1903 VANELLI, L. 1984: «Pronomi e fenomeni di prostesi vocalica nei dialetti italiani settentrionali» RLR 48: 281-95 VANELLI,L. 1987: «I pronomi soggetto nei dialetti settentrionali da! Medio Evo a oggi» Medioevo Romanzo 12: 173-211 VANELLI, L./ RENZI, L./ BENINCA, P. 1985: «Typologie des pronoms sujets dans les langues romanes», in: Linguistique descriptive etc.Actes du xvn e Congres de Linguistique et de Philologie romanes, vol.3, Aix-en-Provence, p. 163-76 116 M. Mair Parry WoLF, H.J. 1991: «La langue des Sermoni subalpini», in: G. P. Cuvro/ C. PrcH (ed.), At di!l VII Ri!scontr anti!rnassional di! Studi an sla Lenga e la Literatura piemonteisa, Alba, 12-13 magg 1990, Alba, p. 237-354