B. Dialects which are detached from the true and proper Italian system, but form no integral part of any foreign Neo-Latin system.
1. Here first of all is the extensive system of the dialects usually called Gallo-Italian, although that designation cannot be considered sufficiently distinctive, since it would be equally applicable to the Franco-Provençal (A. 1) and the Ladin (A. 2). The system is subdivided into four great groups—(a) the Ligurian, (b) the Piedmontese, (c) the Lombard and (d) the Emilian—the name furnishing on the whole sufficient indication of the localization and limits.—These groups, considered more particularly in their more pronounced varieties, differ greatly from each other; and, in regard to the Ligurian, it was even denied that it belongs to this system at all (see Arch. ii. III sqq.).—Characteristic of the Piedmontese, the Lombard and the Emilian is the continual elision of the unaccented final vowels except a (e.g. Turinese öj, oculu; Milanese vǫç, voce; Bolognese vîd, Ital. vite), but the Ligurian does not keep them company (e.g. Genoese öģģu, oculu; vǫže, voce). In the Piedmontese and Emilian there is further a tendency to eliminate the protonic vowels—a tendency much more pronounced in the second of these groups than in the first (e.g. Pied, dné, danaro; vśin, vicino; fnôć, finocchio; Bolognese ćprà, disperato). This phenomenon involves in large measure that of the prothesis of a; as, e.g. in Piedmontese and Emilian armor, rumore; Emilian alvär, levare, &c. U for the long accented Latin u and ö for the short accented Latin o (and even within certain limits the short Latin ó of position) are common to the Piedmontese, the Ligurian, the Lombard and the northernmost section of the Emilian: e.g., Turinese, Milanese and Piacentine dür, and Genoese düu, duro; Turinese and Genoese möve, Parmigiane möver, and Milanese möf, muovere; Piedmontese dörm, dorme; Milanese völta, volta. Ei for the long accented Latin e and for the short accented Latin i is common to the Piedmontese and the Ligurian, and even extends over a large part of Emilia: e.g. Turinese and Genoese avéi, habere, Bolognese avéir; Turinese and Genoese beive, bibere, Bolognese neiv, neve. In Emilia and part of Piedmont ei occurs also in the formulae ĕn, ent, emp; e.g. Bolognese and Modenese beiṅ, solaméint. In connexion with these examples, there is also the Bolognese feiṅ, Ital. fine, representing the series in which e is derived from an í followed by n, a phenomenon which occurs, to a greater or less extent throughout the Emilian dialects; in them also is found, parallel with the ḛi from ḛ, the ou from ǫ: Bolognese udóur, Ital. odore; famóus, Ital. famoso; lóuv, lŭpu. The system shows a repugnance throughout to ie for the short accented Latin e (as it occurs in Italian piede, &c.); in other words, this diphthong has died out, but in various fashions; Piedmontese and Lombard deç, dieci; Genoese dēže (in some corners of Liguria, however, occurs dieže); Bolognese diç, old Bolognese, diese. The greater part of the phenomena indicated above have “Gallic” counterparts too evident to require to be specially pointed out. One of the most important traces of Gallic or Celtic reaction is the reduction of the Latin accented a into e (ä, &c.), of which phenomenon, however, no certain indications have as yet been found in the Ligurian group. On the other hand it remains, in the case of very many of the Piedmontese dialects, in the é of the infinitives of the first conjugation: porté, portare, &c.; and numerous vestiges of it are still found in Lombardy (e.g. in Bassa Brianza: andae, andato; guardae, guardato; sae, sale; see Arch. i. 296-298, 536). Emilia also preserves it in very extensive use: Modenese andér, andare; arivéda, arrivata; peç, pace; Faenzan parlé, parlare and parlato; parléda, parlata; ches, caso; &c. The phenomenon, in company with other Gallo-Italian and more specially Emilian characteristics extends to the valley of the Metauro, and even passes to the opposite side of the Apennines, spreading on both banks of the head stream of the Tiber and through the valley of the Chiane: hence the types artrovér, ritrovare, portéto, portato, &c., of the Perugian and Aretine dialects (see infra C. 3, b). In the phenomenon of á passing into e (as indeed, the Gallo-Italic evolution of other Latin vowels) special distinctions would require to be drawn between bases in which a (not standing in position) precedes a non-nasal consonant (e.g. amáto), and those which have a before a nasal: and in the latter case there would be a non-positional subdivision (e.g. fáme, páne) and a positional one (e.g. quánto, amándo, cámpo); see Arch. i. 293 sqq. This leads us to the nasals, a category of sounds comprising other Gallo-Italic characteristics. There occurs more or less widely, throughout all the sections of the system, and in different gradations, that “velar” nasal in the end of a syllable (paṅ, maṅ; ćáṅta, moṅt)[4] which may be weakened into a simple nasalizing of a vowel (pā, &c.) or even grow completely inaudible (Bergamese pa, pane; padrú, padrone; tep, tempo; met, mente; mut, monte; pût, ponte; púća, punta, i.e. “puncta”), where Celtic and especially Irish analogies and even the frequent use of t for nt, &c., in ancient Umbrian orthography occur to the mind. Then we have the faucal n by which the Ligurian and the Piedmontese (laṅa lüṅa, &c.) are connected with the group which we call Franco-Provençal (A. 1).—We pass on to the “Gallic” resolution of the nexus ct (e.g. facto, fajto, fajtjo. fait, fać; tecto, tejto, tejtjo, teit, teć) which invariably occurs in the Piedmontese, the Ligurian and the Lombard: Pied, fáit, Lig. fajtu, faetu, Lombard fac; Pied. téit, Lig. téitu, Lom. tec; &c. Here it is to be observed that besides the Celtic analogy the Umbrian also helps us (adveitu = ad-vecto; &c.). The Piedmontese and Ligurian come close to each other, more especially by a curious resolution of the secondary hiatus (Gen. réiže, Piedm. réjs = *ra-íce, Ital. radice) by the regular dropping of the d both primary and secondary, a phenomenon common in French (as Piedmontese and Ligurian ríe, ridere; Piedmontese pué, potare; Genoese naeghe = náighe. nátiche, &c.). The Lombard type, or more correctly the type which has become the dominant one in Lombardy (Arch. i. 305-306, 310-311), is more sparing in this respect; and still more so is the Emilian. In the Piedmontese and in the Alpine dialects of Lombardy is also found that other purely Gallic resolution of the guttural between two vowels by which we have the types brája, mánia, over against the Ligurian brága, mánega, braca, manica. Among the phonetic phenomena peculiar to the Ligurian is a continual reduction (as also in Lombardy and part of Piedmont) of l between vowels into r and the subsequent dropping of this r at the end of words in the modern Genoese; just as happens also with the primary r: thus dū = durúr = dolore, &c. Characteristic of the Ligurian, but not without analogies in Upper Italy even (Arch., ii. 157-158, ix. 209, 255), is the resolution of pj, bj, fj into ć, ģ, š: ćü, più, plus; raģģa, rabbia, rabies; šû, fiore. Finally, the sounds š and ž have a very wide range in Ligurian (Arch. ii. 158-159), but are, however, etymologically, of different origin from the sounds š and ž in Lombard. The reduction of s into h occurs in the Bergamo dialects: hira, sera; groh, grosso; cahtél, castello (see also B.2).—A general phenomenon in Gallo-Italic phonetics which also comes to have an inflexional importance is that by which the unaccented final i has an influence on the accented vowel. This enters into a series of phenomena which even extends into southern Italy; but in the Gallo-Italic there are particular resolutions which agree well with the general connexions of this system. [We may briefly recall the following forms in the plural and 2nd person singular: old Piedmontese drayp pl. of drap, Ital. drappo; man, meyn, Ital. mano, -i; long, loyng, Ital. lungo, -ghi; Genoese, káṅ, kḛṅ, Ital. cane, -i; buṅ, buíṅ, Ital. buono, -i; Bolognese, fär, fîr, Ital. ferro, -i; peir, pîr, Ital. pero, -i. zôp, zûp, Ital. zoppo, -i; louv, lûv, Ital. lupo, -i; vedd, vî, Ital. io vedo, tu vedi; vojj, vû, Ital. io voglio, tu vuoi; Milanese quȩst, quist, Ital. questo, -i, and, in the Alps of Lombardy, pal, pȩl, Ital. palo, -i; rȩd, rid, Ital. rete, -i; cor, cör, Ital. cuore, -i; ǫrs, ürs, Ital. orso, -i; law, lȩw, Ital. io lavo, tu lavi; mȩt, mit, Ital. io metto, tu metti; mow möw, Ital. io muovo, tu muovi; cǫr, cür, Ital. io corro, tu corri. [Vicentine pomo, pumi, Ital. pomo, -i; pero, piéri = *píri, Ital. pero, -i; v. Arch. i. 540-541; ix. 235 et seq., xiv. 329-330].—Among morphological peculiarities the first place may be given to the Bolognese sipa (seppa), because, thanks to Dante and others, it has acquired great literary celebrity. It really signifies “sia” (sim, sit), and is an analogical form fashioned on aepa, a legitimate continuation of the corresponding forms of the other auxiliary (habeam, habeat), which is still heard in ch’me aepa purtae, ch’lu aepa purtae, ch’io abbia portato, ch’egli abbia portato. Next may be noted the 3rd person singular in -p of the perfect of esse and of the first conjugation in the Forlì dialect (fop, fu; mandép, mandò; &c.). This also must be analogical, and due to a legitimate ep, ebbe (see Arch. ii. 401; and compare fobbe, fu, in the dialect of Camerino, in the province of Macerata, as well as the Spanish analogy of tuve estuve formed after hube). Characteristic of the Lombard dialect is the ending -i in the 1st person sing. pres. indic. (mi a porti, Ital. io porto); and of Piedmontese, the -éjça, as indicating the subjunctive imperfect (portȩjça, Ital. portassi) the origin of which is to be sought in imperfects of the type staésse, faésse reduced normally to stéjç-, féjç-. Lastly, in the domain of syntax, may be added the tendency to repeat the pronoun (e.g. ti te cántet of the Milanese, which really is tu tu cántas-tu, equivalent merely to “cantas”), a tendency at work in the Emilian and Lombard, but more particularly pronounced in the Piedmontese. With this the corresponding tendency of the Celtic languages has been more than once and with justice compared; here it may be added that the Milanese nün, apparently a single form for “noi,” is really a compound or reduplication in the manner of the ni-ni, its exact counterpart in the Celtic tongues. [From Lombardy, or more precisely, from the Lombardo-Alpine region extending from the western slopes of Monte Rosa to the St Gotthard, are derived the Gallo-Italian dialects, now largely, though not all to the same extent, Sicilianized, from the Sicilian communes of Sanfratello, Piazza-Armerina, Nicosia, Aidone, Novara and Sperlinga (v. Arch. glott. viii. 304-316, 406-422, xiv. 436-452; Romania, xxviii. 409-420; Memorie dell’ Istituto lombardo, xxi. 255 et seq.). The dialects of Gombitelli and Sillano in the Tuscan Apennines are connected with Emilia (Arch. glott. xii. 309-354). And from Liguria come those of Carloforte in Sardinia, as also those of Monaco, and of Mons, Escragnolles and Biot in the French departments of Var and Alpes Maritimes (Revue de linguistique, xiii. 308)]. The literary records for this group go back as far as the 12th century, if we are right in considering as Piedmontese the Gallo-Italian Sermons published and annotated by Foerster (Romanische Studien, iv. 1-92). But the documents published by A. Gaudenzi (Dial. di Bologna, 168-172) are certainly Piedmontese, or more precisely Canavese, and seem to belong to the 13th century. The Chieri texts date from 1321 (Miscellanea di filol. e linguistica, 345-355), and to the 14th century also belongs the Grisostomo (Arch. glott. vii. 1-120), which represents the old Piedmontese dialect of Pavia (Bollett. della Soc. pav. di Storia Patria, ii. 193 et seq.). The oldest Ligurian texts, if we except the “contrasto” in two languages of Rambaud de Vaqueiras (12th century v. Crescini, Manualetto provenzale, 2nd ed., 287-291), belong to the first decades of the 14th century (Arch. glott. xiv. 22 et seq., ii. 161-312, x. 109-140, viii. 1-97). Emilia has manuscripts going back to the first or second half of the 13th century, the Parlamenti of Guido Fava (see Gaudenzi, op. cit. 127-160) and the Regola dei servi published by G. Ferraro (Leghorn, 1875). An important Emilian text, published only in part, is the Mantuan version of the De proprietatibus rerum of Bartol. Anglico, made by Vivaldo Belcalzer in the early years of the 14th century (v. Cian. Giorn. stor. della letteratura italiana, supplement, No. 5, and cf. Rendiconti Istituto Lombardo, series ii. vol. xxxv. p. 957 et seq.). For Modena also there are numerous documents, starting from 1327. For western Lombardy the most ancient texts (13th century, second half) are the poetical compositions of Bonvesin de la Riva and Pietro da Bescapè, which have reached us only in the 14th-century copies. For eastern Lombardy we have, preserved in Venetian or Tuscan versions, and in MSS. of a later date, the works of Gerardo Patecchio, who lived at Cremona in the first half of the 13th century. Bergamasc literature is plentiful, but not before the 14th century (v. Studi medievali, i. 281-292; Giorn. stor. della lett. ital. xlvi. 351 et seq.).
2. Sardinian Dialects.[5]—These are three—the Logudorese or central, the Campidanese or southern and the Gallurese or northern. The third certainly indicates a Sardinian basis, but is strangely disturbed by the intrusion of other elements, among which the Southern Corsican (Sartene) is by far the most copious. The other two are homogeneous, and have great affinity with each other; the Logudorese comes more particularly under consideration here.—The pure Sardinian vocalism has this peculiarity that each accented vowel of the Latin appears to be retained without alteration. Consequently there are no diphthongs representing simple Latin vowels; nor does the rule hold good which is true for so great a proportion of the Romance languages, that the representatives of the ḗ and the í on the one hand and those of the ṓ and the ṹ on the other are normally coincident. Hence plenu (ē); deghe, decem (ĕ); binu, vino (ī); pilu (ĭ); flore (ō); roda, rota (ŏ); duru (ū); nughe, nuce (ŭ). The unaccented vowels keep their ground well, as has already been seen in the case of the finals by the examples adduced.—The s and t of the ancient termination are preserved, though not constantly: tres, onus, passados annos, plantas, faghes, facis, tenemus; mulghet, mulghent.—The formulae ce, ci, ge, gi may be represented by che (ke), &c.; but this appearance of special antiquity is really illusory (see Arch. ii. 143-144). The nexus cl, &c., may be maintained in the beginning of words (claru, plus); but if they are in the body of the word they usually undergo resolutions which, closely related though they be to those of Italian, sometimes bring about very singular results (e.g. ušare, which by the intermediate forms uscare, usjare leads back to usclare = ustlare = ustulare). Nź is the representative of nj (testimónźu, &c.); and lj is reduced to ź alone (e.g. méźus, melius; Campidanese mellus). For ll a frequent substitute is ḍḍ: massīḍḍa, maxilla, &c. Quite characteristic is the continual labialization of the formulae qua, gua, cu, gu, &c.; e.g. ebba, equa; sambene, sanguine (see Arch. ii. 143). The dropping of the primary d (roere, rodere, &c.) but not of the secondary (finidu, sanidade, maduru) is frequent. Characteristic also is the Logudorese prothesis of i before the initial s followed by a consonant (iscamnu, istella, ispada), like the prothesis of e in Spain and in France (see Arch. iii. 447 sqq.).—In the order of the present discussion it is in connexion with this territory that we are for the first time led to consider those phonetic changes in words of which the cause is merely syntactical of transitory, and chiefly those passing accidents which occur to the initial consonant through the historically legitimate or the merely analogical action of the final sound that precedes it. The general explanation of such phenomena reduces itself to this, that, given the intimate syntactic relation of two words, the initial consonant of the second retains or modifies its character as it would retain or modify it if the two words were one. The Celtic languages are especially distinguished by this peculiarity; and among the dialects of Upper Italy the Bergamasc offers a clear example. This dialect is accustomed to drop the v, whether primary or secondary, between vowels in the individual vocables (caá, cavare; fáa, fava, &c.), but to preserve it if it is preceded by a consonant (serva, &c.).—And similarly in syntactic combination we have, for example, de i, di vino; but ol vi, il vino. Insular, southern and central Italy furnish a large number of such phenomena; for Sardinia we shall simply cite a single class, which is at once obvious and easily explained, viz. that represented by su oe, il bove, alongside of sos boes, i. buoi (cf. bíere, bibere; erba).—The article is derived from ipse instead of from ille: su sos, sa sas,—again a geographical anticipation of Spain, which in the Catalan of the Balearic islands still preserves the article from ipse.—A special connexion with Spain exists besides in the nomine type of inflexion, which is constant among the Sardinians (Span. nomne, &c., whence nombre, &c.), nomen, nomene, rámine, aeramine, legumene, &c. (see Arch. ii. 429 sqq.).—Especially noteworthy in the conjugation of the verb is the paradigm cantére, cantéres, &c., timére, timéres, &c., precisely in the sense of the imperfect subjunctive (cf. A. 1; cf. C. 3 b). Next comes the analogical and almost corrupt diffusion of the -si of the ancient strong perfects (such as posi, rosi) by which cantesi, timesi (cantavi, timui), dolfesi, dolui, are reached. Proof of the use and even the abuse of the strong perfects is afforded, however, by the participles and the infinitives of the category to which belong the following examples: ténnidu, tenuto; párfidu, parso; bálfidu, valso; ténnere, bálere, &c. (Arch. ii. 432-433). The future, finally, shows the unagglutinated periphrasis: hapo a mandigare (ho a mangiare = manger-ó); as indeed the unagglutinated forms of the future and the conditional occur in ancient vernacular texts of other Italian districts. [The Campidanese manuscript, in Greek characters, published by Blancard and Wescher (Bibliothèque de l’École des Chartes, xxxv. 256-257), goes back as far as the last years of the 11th century. Next come the Cagliari MSS. published by Solmi (Le Carte volgari dell’ Archivio arcivescovile di Cagliari, Florence, 1905; cf. Guarnerio in Studi romanzi, fascicolo iv. 189 et seq.), the most ancient of which in its original form dates from 1114-1120. For Logoduro, the Condaghe di S. Pietro di Silchi (§§ xii.-xiii.), published by G. Bonazzi (Sassari-Cagliari, 1900; cf. Meyer-Lübke, Zur Kenntnis des Altlogudoresischen, Vienna, 1902), is of the highest importance.]
[3. Vegliote (Veglioto).—Perhaps we may not be considered to be departing from Ascoli’s original plan if we insert here as a third member of the group B the neo-Latin dialect which found its last refuge in the island of Veglia (Gulf of Quarnero), where it came definitively to an end in 1898. The Vegliote dialect is the last remnant of a language which some long time ago extended from thence along the Dalmatian coast, whence it gained the name of Dalmatico, a language which should be carefully distinguished from the Venetian dialect spoken to this day in the towns of the Dalmatian littoral. Its character reminds us in many ways of Rumanian, and of that type of Romano-Balkan dialect which is represented by the Latin elements of Albanian, but to a certain extent also, and especially with regard to the vowel sounds, of the south-eastern dialects of Italy, while it has also affinities with Friuli, Istria and Venetia. These characteristics taken altogether seem to suggest that Dalmatico differs as much as does Sardinian from the purely Italian type. It rejects the -s, it is true, retaining instead the nominative form in the plural; but here these facts are no longer a criterion, since in this point Italian and Rumanian are in agreement. A tendency which we have already noted, and shall have further cause to note hereafter, and which connects in a striking way the Vegliote and Abruzzo-Apulian dialects, consists in reducing the accented vowels to diphthongs: examples of this are: spuota, Ital. spada; buarka, Ital. barca; fiar, Ital. fȩrro; nuat, Ital. notte; kataina, Ital. catḛna; paira, Ital. pḛro; Lat. pĭru; jaura, Ital. ǫra; nauk, Ital. noce; Lat. nŭce; ortaika, Ital. ortica; joiva, Ital. uova. Other vowel phenomena should also be noted, for example those exemplified in prut, Ital. prato; dik, Ital. dieci, Lat. dĕcem; luk, Ital. luogo, Lat. lŏcu; krask, Ital. crḛscere; cenk, Ital. cinque, Lat. quīnque; buka, Ital. bocca, Lat. bčca. With regard to the consonants, we should first notice the invariable persistence of the explosive surds (as in Rumanian and the southern dialects) for which several of the words just cited will serve as examples, with the addition of kuosa, Ital. casa; praiza, Ital. presa; struota, Ital. strada; rosuota, Ital. rugiada; latri, Ital. ladro; raipa, Ital. riva. The c in the formula ce, whether primary or secondary, is represented by k: kaina, Ital. cena; kanaisa, Ital. cinigia; akait, Ital. aceto; plakár, Ital. piacere; dik, Ital. dieci; mukna, Ital. macina; dotko, Ital. dodici; and similarly the g in the formula ge is represented by the corresponding guttural: ghelút, Ital. gelato; jongár, Ital. giungere; plungre, Ital. piangere, &c. On the contrary, the guttural of the primitive formula cū becomes ć (ćol, Ital. culo); this phenomenon is also noteworthy as seeming to justify the inference that the ū was pronounced ü. Pt is preserved, as in Rumanian (sapto, Lat. septem), and often, again as in Rumanian, ct is also reduced to pt (guapto, Lat. octo). As to morphology, a characteristic point is the preservation of the Lat. cantavero, Ital. avrò cantato, in the function of a simple future. Cantaverum also occurs as a conditional. For Vegliote and Dalmatico in general, see M. G. Bartoli’s fundamental work, Das Dalmatische (2 vols., Vienna, 1906), and Zeitschrift für roman. Philologie, xxxii. 1 sqq.; Merlo, Rivista di filologia e d’istruzione class, xxxv. 472 sqq. A short document written about 1280 in the Dalmatic dialect of Ragusa is to be found in Archeografo Triestino, new series, vol. i. pp. 85-86.]
C. Dialects which diverge more or less from the genuine Italian or Tuscan type, but which at the same time can be conjoined with the Tuscan as forming part of a special system of Neo-Latin dialects.
1. Venetian.—Between “Venetian” and “Venetic” several distinctions must be drawn (Arch. i. 391 sqq.). At the present day the population of the Venetian cities is “Venetian” in language, but the country districts are in various ways Venetic.[6] The ancient language of Venice itself and of its estuary was not a little different from that of the present time; and the Ladin vein was particularly evident (see A. 2). A more purely Italian vein—the historical explanation of which presents an attractive problem—has ultimately gained the mastery and determined the “Venetian” type which has since diffused itself so vigorously.—In the Venetian, then, we do not find the most distinctive characteristics of the dialects of Upper Italy comprised under the denomination Gallo-Italic (see B. 1),—neither the ü nor the ö, nor the velar[7] and faucal nasals, nor the Gallic resolution of the ct, nor the frequent elision of unaccented vowels, nor the great redundancy of pronouns. On the contrary, the pure Italian diphthong of ṍ (e.g. cuór) is heard, and the diphthong of ế is in full currency (diéśe, dieci, &c.). Nevertheless the Venetian approaches the type of Northern Italy, or diverges notably from that of Central Italy, by the following phonetic phenomena: the ready elision of primary or secondary d (crúo, crudo; séa, seta, &c.); the regular reduction of the surd into the sonant guttural (e.g. cuogo, Ital. cuoco, coquus); the pure ć in the resolution of cl (e.g. ćave, clave; oréća, auricula); the ś for ģ (śóvene, Ital. giovane); ç for š and ć (péçe, Ital. pesce; çiél, Ital. cielo). Lj preceded by any vowel, primary or secondary, except i, gives ģ: faméga, familia. No Italian dialect is more averse than the Venetian to the doubling of consonants.—In the morphology the use of the 3rd singular for the 3rd plural also, the analogical participle in esto (taśesto, Ital. taciuto, &c.; see Arch. iv. 393, sqq.) and śe, Lat. est, are particularly noteworthy. A curious double relic of Ladin influence is the interrogative type represented by the example crédis-tu, credis tu,—where apart from the interrogation ti credi would be used. For other ancient sources relating to Venice, the estuary of Venice, Verona and Padua, see Arch. i. 448, 465, 421-422; iii. 245-247. [Closely akin to Venetian, though differing from it in about the same degree that the various Gallo-Italian dialects differ among one another, is the indigenous dialect of Istria, now almost entirely ousted by Venetian, and found in a few localities only (Rovigno, Dignano). The most salient characteristics of Istrian can be recognized in the treatment of the accented vowels, and are of a character which recalls, to a certain extent at least, the Vegliote dialect. Thus we have in Istrian i for ệ (bivi, Ital. bevi, Lat. bĭbis; tila, Ital. tḛla; viro, Ital. vero and vetro, Lat. vēru, vĭtru; nito, Ital. netto, Lat. nĭtĭdu, &c.) and analogously u for ǫ (fiur, Ital. fiore, Lat. flōre; bus, Ital. voce, Lat. vōce, &c.); ei and ou from the Lat. ī and ū respectively (ameigo, Lat. amicu, feil, Lat. fīlu, &c.; mour, Lat. mūru; noudu, Lat. nūdu; frouto, Ital. frutto, Lat. frūctu, &c.); ie and uo from ĕ and ŏ respectively in position (piel, Lat. pĕlle, mierlo, Ital. merlo, Lat. mĕrula; kuorno, Lat. cŏrnu; puorta, Lat. pŏrta), a phenomenon in which Istrian resembles not only Vegliote but also Friulian. The resemblance with Verona, in the reduction of final unaccented -e to o should also be noted (nuoto, Ital. notte, &c., bivo, Ital. beve; malamȩntro, Ital. malamente, &c.), and that with Belluno and Treviso in the treatment of -óni, -áni (barbói, -oin, Ital. barboni), though it is peculiar to Istrian that -ain should give -ȩṅ (kaṅ, kȩṅ, Ital. cane -i). With regard to consonants, we should point out the n for gn (líno, Ital. legno); and as to morphology, we should note certain survivals of the inflexional type, amita, -ánis (sing. sía, Ital. zia, pl. siaṅne).] The most ancient Venetian documents take us back to the first half of the 13th century (v. E. Bertanza and V. Lazzarini, Il Dialetto veneziano fino alla morte di Dante Alighieri, Venice, 1891), and to the second half of the same century seems to belong the Saibante MS. For Verona we have also documents of the 13th century (v. Cipolla, in Archivio storico italiano, 1881 and 1882); and to the end of the same century perhaps belongs the MS. which has preserved for us the writings of Giacomino da Verona. See also Archivio glottologico, i. 448, 465, 421-422, iii. 245-247.
2. Corsican[8]—If the “Venetian,” in spite of its peculiar “Italianity,” has naturally special points of contact with the other dialects of Upper Italy (B. 1), the Corsican in like manner, particularly in its southern varieties, has special points of contact with Sardinian proper (B. 2). In general, it is in the southern section of the island, which, geographically even, is farthest removed from Tuscany, that the most characteristic forms of speech are found. The unaccented vowels are undisturbed; but u for the Tuscan o is common to almost all the island,—an insular phenomenon par excellence which connects Corsica with Sardinia and with Sicily, and indeed with Liguria also. So also -i for the Tuscan -e (latti, latte; li cateni, le catene), which prevails chiefly in the southern section, is also found in Northern and Southern Sardinian, and is common to Sicily. It is needless to add that this tendency to u and i manifests itself, more or less decidedly, also within the words. Corsican, too, avoids the diphthongs of ế and ṍ (pe, eri; cori, fora): but, unlike Sardinian, it treats ḯ and ṹ in the Italian fashion: beju, bibo; péveru, piper; pesci; noci, nuces.[9]—It is one of its characteristics to reduce a to e in the formula ar + a consonant (chérne, bérba, &c.), which should be compared particularly with the Piedmontese examples of the same phenomenon (Arch. ii. 133, 144-150). But the gerund in -endu of the first conjugation (turnendu, lagrimendu, &c.) must on the contrary be considered as a phenomenon of analogy, as it is especially recognized in the Sardinian dialects, to all of which it is common (see Arch. ii. 133). And the same is most probably the case with forms of the present participle like merchente, mercante, in spite of enzi and innenzi (anzi, innanzi), in which latter forms there may probably be traced the effect of the Neo-Latin i which availed to reduce the t of the Latin ante; alongside of them we find also anzi and nantu. But cf. also, grȩndi, Ital. grande. In Southern Corsican dr for ll is conspicuous—a phenomenon which also connects Corsica with Sardinia, Sicily and a good part of Southern Italy (see C. 2; and Arch. ii. 135, &c.), also with the northern coast of Tuscany, since examples such as beḍḍu belong also to Carrara and Montignoso. In the Ultramontane variety occur besides, the phenomena of rn changed to r (= rr) and of nd becoming nn (furu, Ital. forno; koru, Ital. corno; kuannu, Ital. quando; vidennu, Ital. vedendo). The former of these would connect Corsican with Sardinian (corru, cornu; carre, carne, &c.); the latter more especially with Sicily, &c. A particular connexion with the central dialects is given by the change of ld into ll (kallu, Ital. caldo).—As to phonetic phenomena connected with syntax, already noticed in B. 2, space admits the following examples only: Cors, na vella, una bella, e bella (ebbélla, et bella); lu jallu, lo gallo, gran ghiallu; cf. Arch. ii. 136 (135, 150), xiv. 185. As Tommaseo has already noted, -one is for the Corsicans not less than for the Sicilians, Calabrians and the French a termination of diminution: e.g. fratedronu, fratellino.—In the first person of the conditional the b is maintained (e.g. farebe, farei), as even at Rome and elsewhere. Lastly, the series of Corsican verbs of the derivative order which run alongside of the Italian series of the original order, and may be represented by the example dissipeghja, dissipa (Falcucci), is to be compared with the Sicilian series represented by cuadiari, riscaldare, curpiári, colpire (Arch. ii. 151).
3. Dialects of Sicily and of the Neapolitan Provinces.—Here the territories on both sides of the Strait of Messina will first be treated together, chiefly with the view of noting their common linguistic peculiarities.—Characteristic then of these parts, as compared with Upper Italy and even with Sardinia, is, generally speaking, the tenacity of the explosive elements of the Latin bases (cf. Arch. ii. 154, &c.). Not that these consonants are constantly preserved uninjured; their degradations, and especially the Neapolitan degradation of the surd into the sonant, are even more frequent than is shown by the dialect as written, but their disappearance is comparatively rather rare; and even the degradations, whether regard be had to the conjunctures in which they occur or to their specific quality, are very different from those of the dialects of Upper Italy. Thus, the t between vowels ordinarily remains intact in Sicilian and Neapolitan (e.g. Sicil. sita, Neap. seta, seta, where in the dialects of Upper Italy we should have seda, sea); and in the Neapolitan dialects it is reduced to d when it is preceded by n or r (e.g. viendę, vento), which is precisely a collocation in which the t would be maintained intact in Upper Italy. The d, on the other hand, is not resolved by elision, but by its reduction to r (e.g. Sicil. víriri, Neap. dialects veré, vedere), a phenomenon which has been frequently compared, perhaps with too little caution, with the d passing into rs (ḍ) in the Umbrian inscriptions. The Neapolitan reduction of nt into nd has its analogies in the reduction of nc (nk) into ng, and of mp into mb, which is also a feature of the Neapolitan dialects, and in that of ns into nź; and here and there we even find a reduction of nf into mb (nf, nv, nb, mb), both in Sicilian and Neapolitan (e.g. at Casteltermini in Sicily ’mbiernu, inferno, and in the Abruzzi cumbonn’, ’mbonn’, confondere, infondere). Here we find ourselves in a series of phenomena to which it may seem that some special contributions were furnished by Oscan and Umbrian (nt, mp, nc into nd, &c.), but for which more secure and general, and so to say “isothermal,” analogies are found in modern Greek and Albanian. The Sicilian does not appear to fit in here as far as the formulae nt and mp are concerned; and it may even be said to go counter to this tendency by reducing nģ and nź to nć, nz (e.g. púnćiri, pungere; menzu, Ital. meźźo; sponza, Ital. spugna, Ven. sponźa).[10] Nay, even in the passing of the sonant into the surd, the Neapolitan dialects would yield special and important contributions (nor is even the Sicilian limited to the case just specified), among which we will only mention the change of d between vowels into t in the last syllable of proparoxytones (e.g. úmmeto, Sicil. úmitu, umido), and in the formula dr (Sicil. and Neap. quatro, Ital. quadro, &c.). From these series of sonants changing into surds comes a peculiar feature of the southern dialects.—A pretty common characteristic is the regular progressive assimilation by which nd is reduced to nn, ṅg to ṅṅ, mb to mm, and even nv also to mm (nv, nb, mb, mm), e.g. Sicil. šínniri, Neap. šénnere, scendere; Sicil. chiummu, Neap. chiummę, piombo; Sicil. and Neap. ’mmidia, invidia; Sicil. sáṅṅu, sangue. As belonging to this class of phenomena the Palaeo-Italic analogy (nd into nn, n), of which the Umbrian furnishes special evidence, readily suggests itself. Another important common characteristic is the reduction of secondary pj fj into kj (chianu -ę, Sicil., Neap., &c., Ital. piano), š (Sicil. šúmi, Neap. šúmmę, fiume), of secondary bj to j (which may be strengthened to ghj) if initial (Sicil. jancu, Neap. janchę, bianco; Sicil. agghianchiari, imbiancare), to l if between vowels (Neap. neglia, nebbia, Sicil. nigliu, nibbio); of primary pj and bj into ć (Sicil. síćća, Neap. séćća, seppia) or ģ respectively (Sicil. raģģa, Neap. arraģģa, rabbia), for which phenomena see also Genoese (B. 1). Further is to be noted the tendency to the sibilation of cj, for which Sicil. jazzu, ghiaccio, may serve as an example (Arch. ii. 149),—a tendency more particularly betrayed in Upper Italy, but Abruzzan departs from it (cf. Abr. jacce, ghiaccio, vracce, braccio, &c.). There is a common inclination also to elide the initial unaccented palatal vowel, and to prefix a, especially before r (this second tendency is found likewise in Southern Sardinian, &c.; see Arch. ii. 138); e.g. Sicil. ’nténniri, Neap. ’ndénnere, intendere; Sicil. arriccamári, Neap. arragamare, ricamare (see Arch. ii. 150). Throughout the whole district, and the adjacent territories in Central Italy, a tendency also prevails towards resolving certain combinations of consonants by the insertion of a vowel; thus combinations in which occur r or l, w or j (Sicil. kiruci, Ital. croce, filágutu, Ital. flauto, salivari, salvare, váriva, Ital. barba; Abr. cálechene, Ital. ganghero, Salevèštre, Silvestro, fęulęmenándę, fulminante, jèreve, Ital. erba, &c.; Avellinese garamegna, gramigna; Neap. ávotro = *áwtro, Ital. áltro, cèvoza = *céwza, Ital. gelso, ajetá side by side with ajtá, Ital. età, ódejo = ódjo, Ital. odio, &c.; Abr. ’nnívęję, indiva, nệbbęję, nebbia, &c.); cattájeve = cattájve, cattivo, goúele = *gowle, gola, &c. &c., are examples from Molfetta, where is also normal the resolution of šk by šek (méšekere, maschera, šekátele, scatola, &c.); cf. seddegno, sdegno, in some dialects of the province of Avellino. In complete contrast to the tendency to get rid of double consonants which has been particularly noted in Venetian (C. 1), we here come to the great division of Italy where the tendency grows strong to gemination (or the doubling of consonants), especially in proparoxytones; and the Neapolitan in this respect goes farther than the Sicilian (e.g. Sicil. sóggiru, suocero, cínniri, cenere, doppu, dopo; ’nsemmula, insieme, in-simul; Neap. dellecato, dilicato; úmmeto, umido; débbole).—As to the phonetic phenomena connected with the syntax (see B. 2), it is sufficient to cite such Sicilian examples as nišuna ronna, nesuna donna, alongside of c’ é donni, c’ è donne; ćincu jorna, cinque giorni, alongside of chiú ghiorna, più giorni; and the Neapolitan la vocca, la bocca, alongside of a bocca, ad buccam, &c.
We now proceed to the special consideration, first, of the Sicilian and, secondly, of the dialects of the mainland.
(a) Sicilian.—The Sicilian vocalism is conspicuously etymological. Though differing in colour from the Tuscan, it is not less noble, and between the two there are remarkable points of contact. The dominant variety, represented in the literary dialect, ignores the diphthongs of ḗ and of ŏ, as it has been seen that they are ignored in Sardinia (B. 2), and here also the ĭ and the ŭ appear intact; but the ḗ and the ṓ are fittingly represented by i and u; and with equal symmetry unaccented e and o are reproduced by i and u. Examples: téni, tiene; nóvu, nuovo; pilu, pelo; miṅnitta, Ital. vendḛtta; jugu, giogo; agustu, Ital. agǫsto; crídiri, credere; vínniri, Ital. vēndere; sira, sera; vina, vena; suli, Ital. sole; ura, ora; furma, Ital. fǫrma. In the evolution of the consonants it is enough to add here the change of lj into ghj (e.g. fígghiu, Ital. figlio) and of ll into ḍḍ (e.g. gaḍḍu, Ital. gallo). As to morphology, we will confine ourselves to pointing out the masculine plurals of neuter form (li pastura, li marinara). For the Sicilian dialect we have a few fragments going back to the 13th century, but the documents are scanty until we come to the 14th century.