II. PHONOLOGY.

14. Inasmuch as Provençal, like the other Romance languages, grew out of the Latin commonly spoken under the Roman Empire, we must take this latter language as our starting-point. The transformation was so gradual and continuous that we cannot assign any date at which speech ceases to be Latin and begins to be Provençal; since, however, the various Latin dialects—destined to become later the various Romance languages—began to diverge widely in the 6th and 7th centuries, we may, for the sake of convenience, say that the Latin period ends at about this time. Before this, certain changes (which affected all the Romance tongues) had occurred in the popular language, differentiating it considerably from the classic Latin of the Augustan writers. Although the most important of these alterations have to do with inflections rather than with pronunciation, the sound-changes in Vulgar Latin are by no means insignificant.

15. It is essential at the outset to distinguish “popular” from “learned” words. The former, having always been a part of the spoken vocabulary, have been subject to the operation of all the phonetic laws that have governed the development of the language. The latter class, consisting of words borrowed by clerks, at various periods, from Latin books and from the Latin of the Church, is naturally exempt from sound-changes that occurred in the vulgar tongue before the time of their adoption. The form of learned words depends, in the first place, on the clerical pronunciation of Latin at the date of their borrowing; then, if they came into general use, their form was subject to the influence of any phonetic laws that were subsequently in force. The fate of borrowed terms differs, therefore, according to the time of their introduction and the degree of popularity which they afterwards attained.

1. ACCENT.

16. The place of the primary accent, which in Classic Latin was determined by quantity, remained unchanged in Vulgar Latin even after quantitative distinctions were lost. A short vowel before a mute followed by a liquid may, in Classic Latin, be stressed or unstressed; in Vulgar Latin it is usually stressed: cathédra, tenébræ.[14]

There are some exceptions to the rule of the persistence of the accent in Vulgar Latin:—

1. An accented e or i immediately followed by the vowel of the penult transfers the stress to this latter vowel, and is itself changed to y: filíŏlus ＞ filyólus, mulíĕrem ＞ mulyére. This shift is perhaps due to a tendency to stress the more sonorous of two contiguous vowels.

2. An accented u immediately followed by the vowel of the penult transfers the stress to the preceding syllable, and is itself changed to w: habúĕrunt ＞ ábwerunt, tenúĕram ＞ ténwera. This shift cannot be explained on the same principle as the foregoing one; it is perhaps due in every case to analogy—hábuit, ténui, for instance, being responsible for the change in habúerunt, tenúeram.

3. Verbs compounded with a prefix, if their constituent parts were fully recognized, were usually replaced in Vulgar Latin by a formation in which the vowel and the accent of the simple verb were preserved: défĭcit ＞ disfácit, réddĭdi ＞ reddédi, rénĕgo ＞ renégo, réquĭrit ＞ requærit. In récĭpit ＞ recípit the accent but not the vowel was restored, speakers having ceased to associate this verb with capio. In cóllĭgo, érĭgo, éxĕo, ínflo the composite nature of the word was apparently not recognized.

4. The adverbs ĭllāc, ĭllīc accented their last syllable, by the analogy of hāc, hīc.

17. In Provençal the primary accent falls on the same syllable as in Vulgar Latin: bonitātem ＞ V. L. bonitáte ＞ Pr. bontát, compŭtum ＞ V. L. cómputu ＞ Pr. cónte; cathĕdra ＞ V. L. catédra ＞ Pr. cadéira; filiŏlus ＞ V. L. filyólus ＞ Pr. filhóls, tenuĕram ＞ V. L. ténwera ＞ Pr. téngra, requĭrit ＞ V. L. requærit ＞ Pr. requér, illac ＞ V. L. illác ＞ Pr. lai.

1. Some learned words have an irregular accentuation, apparently due to a mispronunciation of the Latin: cándĭdum ＞ quandí, grammátĭca ＞ gramatíca, láchry̆mo ＞ lagrím, spírĭtum ＞ esprít (perhaps from the formula spirítui sancto). Others were adopted with the correct stress, but shifted it later: fábrĭca ＞ fábrega ＞ fabréga (and fárga), fémĭna ＞ fémena ＞ feména (and fémna), láchry̆ma ＞ lágrema ＞ lagréma, sémĭnat ＞ sémena ＞ seména (and sémna), vírgĭnem ＞ vérgena ＞ vergína (and vérge).

2. Dimércres ＜ dīe Mercūrī (perhaps through ✱dīe Mércŏris) has evidently been influenced by divénres ＜ dīe Vĕnĕris.

3. Some irregularities due to inflection will be discussed under Morphology.

18. The secondary accent, in Vulgar Latin, seems not to have followed the Classic Latin quantitative rule, but to have fallen regularly on the second syllable from the primary stress: cṓgĭtó, cupĭ́dĭtā́tem. If this secondary accent followed the tonic, its vowel probably developed as an unstressed post-tonic vowel; if it preceded, its vowel was apparently treated as a stressed vowel. This treatment was doubtless continued in Provençal until the intertonic vowel dropped out: cógĭtó cógĭtánt ＞ cug cúian (cf. cánto cántant ＞ can cántan), ✱comĭnĭtĭāre ＞ comén’tiáre ＞ coménzár ＞ comensár. As may be seen from this last example, after the fall of the intertonic vowel, the secondary stress, being brought next to the primary, disappeared, and its vowel was henceforth unaccented. Cf. § 45, 1.

19. Short, unemphatic words had no accent in Vulgar Latin, and were attached as particles to the beginning or the end of another word: te vídet, áma me. Such words, if they were not monosyllabic, tended to become so; a disyllabic proclitic beginning with a vowel regularly, in Vulgar Latin, lost its first syllable: illum vídeo ＞ V. L. lu véyo ＞ Pr. lo vei. A word which was used sometimes independently, sometimes as a particle, naturally developed double forms.

2. VOWELS.

QUANTITY.

20. Latin had the following vowels, which might be long or short: a, e, i, o, u. The diphthongs, æ, œ, au, eu, ui, were always long: æ and œ, however, were simplified into monophthongs, mainly in the Republican epoch, æ being sounded ę̄, œ probably ẹ̄; au retained (save in some popular dialects) its old pronunciation; eu did not occur in any word that survived; ui, in cui, illui, in Vulgar Latin, was accented úi (as in fui). The simple vowels, except a, were, doubtless from early times, slightly different in quality according to their quantity, the long vowels being sounded close, the short open: ẹ̄, ị̄, ọ̄, ụ̄; ę̆, į̆, ǫ̆, ų̆.

21. Between the 1st and the 7th century of our era, the Classic Latin quantity died out: it had apparently disappeared from unstressed vowels as early as the 4th century, from stressed by the 6th. It left its traces, however, as we have seen, upon accentuation (§ 16), and also upon vowel quality, the originally long and short remaining differentiated in sound, if they were accented. Of the unaccented vowels, only i shows sure signs of such a differentiation, and even for i the distinction is evident only in a final syllable: vēnī vēnĭt ＞ vẹnị vẹnįt.

ACCENTED VOWELS.

22. The vowels of Vulgar Latin are a, ẹ, ę, ị, į, ọ, ǫ, ụ, ų, with the diphthongs áu and úi; the old æ and œ had become identical in sound with ę and ẹ. As early as the 3d century of our era, į was changed, in nearly all the Empire, to ẹ, and thus became identical with the vowel coming from original ē. A little later, perhaps, ų, in the greater part of the Empire, became ọ, thus coinciding with the vowel that was originally ō. Ypsilon, in words taken from the Greek, was identified, in early borrowings, with Latin u; in later ones, with Latin i: βύρσα ＞ Pr. borsa, γῠρος ＞ Pr. girs. Omicron, which apparently had the close sound in Greek, generally (but not always) retained it in recently borrowed words in Vulgar Latin: τόρνος ＞ tọrnus (cf. Pr. tọrn), but κόλαφος ＞ cŏlăphus = cọlapus or cǫlapus (cf. Pr. cǫlp).

The development of the Vulgar Latin vowels in Provençal will now be examined in detail:—

a

23. Cl. L. ā, ă ＞ V. L. a ＞ Pr. ą: ărbŏrem ＞ ąrbre, grātum ＞ grąt, măre > mąr.

1. The ending -arius shows an irregular development in French and Provençal, the Provençal forms being mainly such as would come from -ĕrius; as in parlier, parleira. In the earliest stage we find apparently -ęr´ and -ęr´a; then -ęr´ and -ęir´a; next -ęr, -ięr and -ęira, -ięira; finally, with a reciprocal influence of the two genders, -ęr, -ięr, -ęir, and -ęra, -ięra, -ęira, -ięira: caballarium ＞ c(h)avaler -ier, -eir, ✱man(u)aria ＞ manera -iera -eira -ieira. The peculiar treatment of this suffix has not been satisfactorily explained. See E. R. Zimmermann, Die Geschichte des lateinischen Suffixes -arius in den romanischen Sprachen, 1895; E. Staaff, Le suffixe -arius dans les langues romanes, Upsala, 1896, reviewed by Marchot in Zs., XXI, 296, by Körting in Zeitschrift für französische Sprache, XXII, 55; Meyer-Lübke, Gram., I, 222, § 237; Zimmermann in Zs., XXVI, 591; Thomas in Rom., XXXI, 481 and in Bausteine zur romanischen Philologie, 641. The likeliest theory is that of Thomas: that -arius was associated with the Germanic ending -ari and participated in the umlaut which affected the latter; cf. Phon., pp. 34-36.

2. In Gascony and Languedoc ei is used for ai ＜ habeo. The ei perhaps developed first as a future ending (amar -ei) by analogy of the preterit ending -ei (amei): see Morphology, §§ 152, 1, 162, (4), 175, (4), where this latter ending is discussed also. For a different explanation, see Meyer-Lübke, Gram., I, 222, § 237.

3. A few apparent irregularities are to be traced to the vocabulary of Vulgar Latin. For instance, Pr. sereisa represents, not Cl. L. cĕrăsus, but V. L. cĕrĕsĕa: see Meyer-Lübke, Einf., § 103. Uebre is from ✱ŏpĕrit, or apĕrit modified by ✱cŏpĕrit = cōperit. Voig is from ✱vŏcĭtum = vacuum: Einf., § 114.

4. Such forms as fontaina = fontana ＜ fontāna, etc., and tres = tras ＜ trans, etc., are French or belong to the borderland between French and Provençal.

24. In some dialects, particularly in Rouergue, Limousin, Auvergne, and Dauphiné, a became ạ before a nasal, and at the end of a monosyllable or an oxytone: canem ＞ cạn, grandem ＞ grạnt, cadit ＞ cạ, stat ＞ estạ.

1. The conditions differ somewhat in the various dialects, according as the nasal consonant falls or remains, and is followed by another consonant or not. In Limousin the sound is ą before an n that cannot fall: see § 11, (5). In Rouergue and in Dauphiné, ạ appears before all nasals. The poets generally follow the Limousin usage. See F. Pfützner, Ueber die Aussprache des provenzalischen A, Halle, 1884.

ẹ

25. Cl. L. ē, ĭ, œ ＞ V. L. ẹ ＞ Pr. ẹ: habēre ＞ avẹr, mē ＞ mẹ, mensem ＞ mẹs, plēnum ＞ plẹn, rēgem ＞ rẹi, vēndĕre ＞ vẹndre; ĭnter ＞ ẹntre, fĭdem ＞ fẹ, malĭtia ＞ malẹza, mĭnus ＞ mẹns, mĭttĕre ＞ mẹtre, sĭccum ＞ sẹc, vĭrĭdem ＞ vẹrt, pœna ＞ pẹna.

1. Some words have ę instead of ẹ:—

(a) The ending -ētis in the present indicative becomes -ętz through the analogy of ętz ＜ ĕstis.

(b) Camęl (also ẹ), candęla (also ẹ), cruzęl, fizęl (also ẹ), maissęla have ę through the analogy of the suffix -ęl ＜ -ĕllus. In camel the substitution probably goes back to Vulgar Latin.

(c) Many learned words, including proper names, have ę for ẹ: decręt, Elizabęt, Moysęs, pantęra, requięs, secręt (ẹ), sencęr.

(d) Espęr for espẹr ＜ spēro, quęt for quẹt ＜ qu(i)ētum are perhaps bad rhymes. Bartolomeo Zorzi, a Venetian, rhymes -ẹs with -ęs; in Catalan these two endings were not distinguished.

(e) Individual cases: adęs, ‘at once,’ probably from ad id ĭpsum, seems to have been affected by pręs and apręs ＜ ad prĕssum; mostięr ＜ monastērium shows the influence of ministĕrium; nęr nięr (also nẹr nẹgre) ＜ nĭgrum perhaps shows the influence of entęr entięr and the numerous adjectives in -ęr -ięr; nęu nięu nęy ＜ nĭvem has been attracted by bręu gręu, lęu; senęstre (cf. late Lat. sinexter) is evidently influenced by dęstre.

2. Many words have i instead of ẹ:—

(a) Berbitz = vervēcem, camis = camĭsia, come from alternative V. L. forms, berbīcem, ✱camīsia. Planissa (also -eza), sebissa, etc., probably show -īcia for -ĭtia. For dit = dĭgĭtum see § 65, Y, 1.

(b) In many learned words Latin ĭ is represented by i in Provençal: albir, martire, edifici, iuzizi, servizi, vici, etc.; iusticia, leticia, tristicia, etc. Aurilha (also ẹ) ＜ aurĭcula, cilh, (also cieilh, sobreselhs) ＜ cĭlium, issilh ＜ exĭlium, familha ＜ famĭlia, maístre (also maẹstre maiẹstre) ＜ magĭstrum, meravilha (also ẹ) ＜ mirabĭlia, perilh ＜ perĭculum, etc., are probably learned forms. Máistre and mestre are French.

(c) Ciri (cere) = cēreum, iure (cf. ebriac) = ēbrium (or ✱ĕbrium), marquis (ẹ), merci (ẹ), país (ẹ) = ✱pagēnsem, plazir (ẹ), pris (ẹ), etc., are French.[15] For a discussion of iure and a different explanation of ciri, see P. Savj-Lopez, Dell’ “Umlaut” provenzale, 1902, p. 4.

(d) Ins (also entz) ＜ ĭntus, dins (also dens) ＜ de ĭntus, dintre (cf. en, entre) ＜ de ĭnter have not been satisfactorily explained. Regular forms with ẹ are found in Béarn, Gascony, Dauphiné, and the Alps.

(e) Individual cases: tapit ＜ ταπήτιον shows the modern pronunciation of Greek η; verin = venēnum is an example of substitution of suffix.

3. Arnei, fei, mei = me, palafrei, perquei, sei = se are French or Poitevin forms; they are common in William of Poitiers. Mercey, rey = re, used by Marcabru, seem to be due either to an imitation of such forms as the preceding or to the analogy of crei cre ＜ crēdo. Cf. § 65, N, 3.

4. Contránher seems to be a fusion of constrĭngere and contrahere; vendanha ＜ vindēmia shows French influence.

26. An ẹ in hiatus became i: lĭgat ＞ lia, vĭa ＞ via.

27. When there was in the next syllable a final ī, V. L. ẹ was changed in Provençal to i: ecc’ĭllī ＞ cilh, ecc’ĭstī ＞ cist, fēcī ＞ fis, ✱prēsī ＞ pris, ✱vēnuī ＞ vinc, vigĭntī ✱vĭntī ＞ vint.

1. In the nominative plural of masculine nouns and adjectives this change was regularly prevented by the analogy of the singular and the accusative plural: mĭssī ＞ mes, plēnī ＞ plen. We find, however, cabil ＜ capĭllī.

2. Dec for ✱dic ＜ dēbuī seems to have been attracted by the dec ＜ dēbuit of the third person. Venguest for venguist ＜ ✱venuĭstī is due both to the influence of the plural forms venguem, venguetz and to the analogy of the weak preterits, such as cantest, vendest.

ę

28. Cl. L. ĕ, æ ＞ V. L. ę ＞ Pr. ę: infĕrnum ＞ enfęrn, fĕrrum ＞ fęr, pĕdem ＞ pę; cælum ＞ cęl, quærit ＞ quęr.

1. Such forms as glisia, lire, pire, pis, profit are French. Profich may be a cross between profieg and profit, or it may be due to the analogy of dich.

2. Cossint, mint, sint, used by Arnaut Daniel, are perhaps faulty rhymes.

3. Auzil ＜ avicĕllī, in the Boeci, may be due to the analogy of such plural forms as cabil ＜ capĭllī, il ＜ ĭllī, etc. Briu, sometimes used for breu ＜ brĕvem, is evidently connected with abrivar, ‘hasten,’ the origin of which is uncertain. Elig shows the influence either of eligir (beside elegir) or of dig. Ginh = genh ＜ ingĕnium evidently follows ginhos ＜ ingeniōsus and its derivatives. Isme (esme) is a post-verbal noun from ✱ismar (cf. azismamen), a dialect form of esmar ＜ æstimare. Quis ＜ ✱quæsi, tinc ＜ tĕnui are due to the analogy of pris ＜ ✱prēsī, vinc ＜ ✱vēnuī.

4. Beside nęula ＜ nĕbula, we find nebla, neble, presumably from the same source, and also nible, niól, nióla, niúl, niúla, nivól. According to Nigra, Archivio glottologico italiano, XV, 494, nūbes ＞ nūbĭlus ＞ ✱nĭbŭlus (and ✱nĭbūlus?), whence might be derived ✱níŭlus ✱niúlus, which would account for niól-a, niúl-a, and perhaps for a ✱nívol ＞ nivól. Nible might be regarded as a cross between neble and niul. Cf. § 38, 3.

5. In ẹs ＜ ĕst the ẹ probably comes from such combinations as mẹ’s, quẹ’s, understood as m’ẹs, qu’ẹs. Espẹlh ＜ spĕculum shows the influence of cossẹlh, solẹlh. Estẹla presupposes a Latin ✱stēla or ✱stēlla for stĕlla: cf. the Fr. and It.

6. Plais, ‘hedge’ seems to be a cross between plĕxus and paxillus, ‘fence.’ Vianda (＜ vivenda?) is probably French.

7. Volon ＜ volentem shows the influence of the ending -ŭndus.

8. Greuga ＜ con-gregar has been influenced by greu ＜ ✱grĕvem = gravem influenced by lĕvem. Cf. grey ＜ grĕgem.

29. Before a nasal, in most of the dialects of Limousin, Languedoc, and Gascony, ę became ẹ: bĕne ＞ bẹn, dicĕntem ＞ dizẹn, tĕmpus ＞ tẹms, tĕnet ＞ tẹn, vĕniam ＞ vẹnha, vĕntum ＞ vẹnt.

30. Early in the history of Provençal, before u, i, or one of the palatal consonants l´, r´, s´, z´, y, tš, dž, an ę broke into ię, except in a few dialects of the west and north: dĕus ＞ dięus, mĕum ＞ mięu; amāvi ＞ ✱amai ＞ amęi amięi,[16] ✱fĕria ＞ fięira, ✱ec(c)lĕsia? (Cf. Zs., XXV, 344) ＞ glięiza, lĕctum ＞ lięit, pĕjus ＞ pięis; vĕtŭlum vĕclum ＞ vięlh, ministĕrium ＞ mestięr, ✱ec(c)lĕsia? ＞ glięza, mĕdia ＞ mięia, lĕctum ＞ lięg. There seems to be also, at least in some dialects, a tendency to break the ę before a g or a k: lĕgunt ＞ lięgon; ✱sĕquit ＞ sęc sięc, subjunctive sięgas (sęga), but infinitive sęgre ＜ ✱sĕquere.[17]

The breaking was probably due to a premature lifting of the tongue under the influence of a following high vowel or a palatal (or velar) consonant.[18] Before u it occurred everywhere except in the extreme west; before palatals the ę apparently remained intact both in the extreme west and in Quercy, Rouergue, Auvergne, and Dauphiné. At first, no doubt, the diphthong was less marked than it became in the 12th and 13th centuries. It is not indicated in our oldest text, the Boeci (breu, deu, eu, mei, meler, vel)[19], and it frequently remains unexpressed even in the writings of the literary period.

It is to be noted that ę does not break before u ＜ l nor before i ＜ ð: bĕllus ＞ bęls ＞ bęus, pĕtra ＞ ✱pęðra ＞ pęira, Pĕtrum ＞ ✱Pęðre ＞ Pęire, rĕtro ＞ ✱ręðre ＞ ręire[20]. The breaking must, therefore, have occurred before these developments of l and ð, both of which apparently antedate the Boeci: cf. euz = els, v. 139; eu = el, v. 155; Teiric ＜ ✱Teðric ＜ Theodorīcum, v. 44, etc. On the other hand, there is no diphthong before ts, dz, s, z coming from Latin c´, cy, pty, tty, ty: dĕcem ＞ dętz, pĕttia (or pĕcia) ＞ pęssa, nĕptia ＞ nęssa, ✱prĕtiat ＞ pręza, prĕtium ＞ prętz[21]. The breaking, therefore, took place after these consonants had ceased to be palatal. We may ascribe it with some confidence to the period between the seventh and tenth centuries.

1. A number of cases of ię before r are doubtless to be explained by analogy. Hĕri ＞ ęr; autre + er ＞ autręr, which, through the influence of adjectives in -ęr -ięr, became autrięr: hence the form ięr. Fĕrio, mĕreo ＞ fięr, mięr; hence, by analogy, the first person forms profięr, quięr, then the third person forms fięr, mięr, profięr, quięr, sięrf (but sęrvon, sęrva), and the subjunctives ofięira, sofię(i)ra.

2. Ięsc (= ĕxeo), ięscon, ięsca receive their diphthong either from earlier forms with s´ or from ięis ＜ ĕxit.

ị

31. Cl. L. ī ＞ V. L. ị: amīcum ＞ amịc, fīnem ＞ fịn, trīstem ＞ trịst.

1. Frẹg, frẹit are from V. L. ✱frĭgdum = frīgĭdum, the ĭ being perhaps due to the analogy of rĭgĭdum.

32. In the 13th century or earlier the group iu, in most dialects, became ieu: captīvum ＞ caitiu caitieu, æstīvum ＞ estiu estieu, revīvĕre ＞ reviure revieure, sī vās ＞ sius sieus.

ọ

33. Cl. L. ō, ŭ ＞ V. L. ọ ＞ Pr. ọ, which developed into ụ probably during the literary period: dolōrem ＞ dolọr, spōnsa ＞ espọsa, flōrem ＞ flọr; bŭcca ＞ bọca, gŭla ＞ gọla.

1. An irregular ǫ, which is found in some words, goes back to Vulgar Latin: cǫbra = re-cŭperat, cǫsta (also ọ) = cōnstat, nǫra = nŭra, ǫu = ōvum, plǫia = plŭvia, redǫbla = ✱redŭplat, sǫbra = sŭperat, suefre = sŭffero. V. L. ✱cŏperat may be regarded as a fusion of cŭperat and ✱cŏperit (§ 40, 1; cf. Rom. XXXI, 9); ✱cŏstat is unexplained; ✱nŏra shows the influence of sŏror and sŏcĕra; the ŏ of ✱ŏvum has been explained as due to differentiation from the following v; ✱plŏia is to be connected with the popular plŏvĕre (cf. Meyer-Lübke, Einf., § 142); ✱sŏperat follows the analogy of ✱cŏperat; ✱sŏffero evidently follows ŏffero. Redǫbla (also ọ) is not accounted for. If trǫba has anything to do with tŭrbat, it was perhaps influenced by prŏbat (cf. Zs., XXVIII, 50). Engǫissa ＜ V. L. ✱angŏstia = angŭstia. See A. Thomas, Nouveaux essais de philologie française, 1904, 339.

2. Some words have ü: iüs (also iọs) ＜ deōrsum shows the influence of süs ＜ sūrsum; lür (usually lọr) ＜ illōrum (cf. lur in the dialects of Navarre and Aragon) comes through an ✱illūrum due to the analogy of illūi = illi; melhüra (ọ), peiüra (ọ) perhaps follow aüra ＜ ✱a(u)gūrat; rancüra is a mixture of rancōrem and cūra; üis is from V. L. ūstium = ōstium (cf. Zs., XXV, 355); üpa ＜ ŭpŭpa is due to onomatopœia.

3. The adverbs ar, ara, er, era, eras, meaning ‘now,’ are hardly to be connected with hōra. Meyer-Lübke takes era, etc., from a Latin ✱era corresponding to Greek άρα; ara, ar may come directly from άρα, άρ: cf. Gr., III, 552, note.

4. Tonleu, ‘tariff,’ from τελώνιον, shows double metathesis. For adoutz, ‘fount,’ see A. Thomas, Essais de philologie française, 1897, 205.

34. Before tš, dž (and it, id), before n´, and before final i, an ọ becomes ü in various dialects: cōgĭtat ＞ cüia cüida, ✱stŭdiat ＞ estüia, fŭgit ＞ füg, refŭgium ＞ refüg; jŭngĕre ＞ iünher, ŭngĕre ＞ ünher, pŭgnum ＞ pünh; dŭī ＞ düi, sŭm ＞ sọ + i ＞ süi. The ü before tš, dž apparently occurs everywhere except in Dauphiné; before n´ it is to be found in nearly all the dialects of the north and west; before final i it seems to be limited to Bordeaux, Auvergne, and a part of Languedoc.

ǫ

35. Cl. L. ŏ ＞ V. L. ǫ ＞ Pr. ǫ: cŏr ＞ cǫr, cŏrpus ＞ cǫrs, mŏrtem ＞ mǫrt, ŏpĕra ＞ ǫbra, rŏta ＞ rǫda.

1. For demọra (also ǫ) ＜ ✱demŏrat, see Meyer-Lübke, Gram., I, 204, § 220. For prọa (also prǫa, prueva) ＜ prŏbat, see Rom., XXXI, 10, footnote 3.

36. Before a nasal, in most of the dialects of Limousin, Languedoc, and Gascony, ǫ became ọ: bŏnum ＞ bọn, fŏntem ＞ fọnt, pŏntem ＞ pọnt. Cf. E. Levy in Mélanges de philologie romane dédiés à Carl Wahlund, 1896, p. 207.

1. If the nasal was n´, the vowel remained open in most or all of these dialects: cŏgnĭta ＞ coinda cuenda cuenhda, lŏnge ＞ lonh luenh, sŏmnium ＞ sonh suenh.

37. Early in the history of Provençal, before u, a labial consonant, a g or a k, an i, or one of the palatal consonants l´, n´, r´, s´, z´, y, tš, dž, an ǫ broke, in most dialects, into a diphthong which developed into ue, üo, üe, or ü[22]: bŏvem ＞ bǫu büọu büeu, ✱ŏvum ＞ ǫu üou üeu, nŏvus ＞ nǫus nüous nüeus; ✱cŏpero ＞ cǫbri cüebre[23], nŏva ＞ nǫva nüeva, ŏpus ＞ ǫps üops, prŏbat ＞ prǫa prüeva, ✱trŏpo? ＞ trǫp trüeb; cŏquus ＞ cǫcs cüocs cüex, fŏcum ＞ fǫc füoc füec füc, crŏcus ＞ grǫcs grüocs grüecs, jŏcum ＞ iǫc iüoc iüec iüc, lŏcus ＞ lǫcs lüocs lüecs, lŏcat ＞ lüoga, pŏtui ＞ püec, sŏc(ĕ)rum ＞ (sǫzer) sǫgre süegre (fem. süegra); ✱ingrŏssiat ＞ engrǫissa engrüeissa, ✱angŏstia ＞ engǫissa engüeissa, nŏctem ＞ nǫit nüoit nüeit, ŏcto ＞ ǫit üeit, pŏstea ＞ pǫissas püeissas, prŏximus ＞ prǫymes prüeymes; fŏlia ＞ fǫlha füolha füelha fülha, ŏcŭlus ŏclus ＞ ǫlhs üolhs üelhs ülhs, lŏnge ＞ lǫnh lüenh, sŏmnium ＞ sǫnh süenh, cŏrium ＞ cǫr cüer, pŏstea ＞ pües, prŏximum ＞ prǫsme prüesme, ✱plŏia ＞ plǫia plüeia plüia, ✱inŏdiat ＞ enǫia enüeia enüia, ✱pŏdiat ＞ pǫia püeia püia, nŏctem ＞ nüoch nüech nüh, ŏcto ＞ üeg.

The breaking was probably due to a premature lifting of the tongue under the influence of a following high vowel or a palatal or velar consonant, or to a premature partial closure of the lips in anticipation of a following labial. Before i or a palatal the diphthong was at the start presumably üo; before u or a labial or velar consonant, uo: from these two types, the first of which influenced the second, came the later developments. Ü is a reduction of üo or üe; it apparently does not occur before u.

The dialect conditions are mixed, the development in each region depending somewhat on the following sound. In the southwest, ǫ and ue seem to prevail; in the northwest, ü; in the west, in Limousin, and in Auvergne, üe; in Languedoc, üo; in the east and south, üe, üo, ǫ.

The date of breaking is discussed in § 30.

1. In some words where a diphthong would be expected, none is found, although it may have existed: mǫu ＜ mŏvet, nǫu ＜ nŏvem, plǫu ＜ ✱plŏvit; trǫp ＜ þrop; brǫcs ＜ ✱brŏccus, iǫgon ＜ jŏcunt, lǫgui ＜ lŏco. The form püoc or püec ＜ pŏtui is regularly reserved for the first person, pŏtuit being represented by pǫc.

2. A few cases of irregular breaking are easily explained: püosc püesc (= pŏssum) and püosca püesca (= pŏssim) owe their diphthong either to earlier forms with s´ or to the analogy of püec; sǫfre süefre süfre (= sŭffert) are from ✱sŏfferit, formed upon ✱ŏfferit = ŏffert (cf. § 33, 1); vüelc (= vŏlui) follows the analogy of vüelh (＜ ✱vŏleo = vŏlo) and of püec.

ụ

38. Cl. L. ū ＞ V. L. ụ ＞ Pr. ü: ✱habūtus ＞ avütz, jūstum ＞ iüst, mūrum ＞ mür, mūtus ＞ mütz, nūdus ＞ nütz, plūs ＞ plüs.

The date of the change of ụ into ü is not known; there is no ü in Catalan, and there may have been none in early Gascon. It seems likely that the Celts, when they adopted Latin, pronounced ū a little further forward in the mouth than did the Romans; that their ụ continued to advance gradually toward the front of the mouth until it became ü; and that this ü spread to the parts of France that were not originally Celtic.[24] In the literary period the sound was probably ü in most or all of the Provençal dialects.

1. Pr. ọnze represents a V. L. ✱ŭndĕcim, which in Gaul and Spain replaced ūndĕcim. Lọita lücha, trọcha trücha probably go back to Latin double forms, ✱lŭcta lūcta, ✱trŭcta trūcta.

2. Nǫssas ＜ ✱nŏptias = nūptias, by analogy of ✱nŏvius, ‘bridegroom,’ from nŏvus.

3. Before u, Pr. ü apparently became i: nūbem ＞ ✱nüu ＞ niu, pūlĭcem ＞ ✱püuze ＞ piuze. See §§ 63, (4); 74, (2).

au

39. Cl. L. au ＞ V. L. au ＞ Pr. au: aurum ＞ aur, gaudium ＞ gaug, paucum ＞ pauc, thesaurus ＞ tesaurs.

1. Bloi ＜ blauþr, ioi, ioia, ioios, lotia ＜ ✱laubja, noiza, onta ＜ hauniþa, or, sor, tesor, etc., are French or Poitevin; ioi is a good Poitevin form. Iai, ‘joy,’ seems to be a fusion of ioi and Pr. iai = gai.

2. Anta ＜ hauniþa is unexplained.

UNACCENTED VOWELS.

40. (1) The fate of an unaccented vowel depended largely upon the syllable in which it stood: in general, unstressed vowels in the initial syllable remained intact, while all vowels, except a, fell (at different dates) in the other syllables. The fall of unaccented vowels resulted in many new consonant groups: collocáre ＞ colcár, hóminem ＞ ómne, sábbatum ＞ sápte.

(2) The vowels e̯ and i̯, instead of falling or remaining unchanged, became y in Vulgar Latin, early in our era: alea ＞ alya, diŭrnus ＞ dyųrnus, mĕdium ＞ mędyu. Similarly u̯ became w: placui ＞ placwi, tĕnuis ＞ tęnwis.

1. Apparently, however, e̯é, i̯é ＞ e; o̯ó, u̯ó ＞ o: prĕhĕndĕre ＞ prĕndĕre; abĭĕtem ＞ ✱abētem, facĭēbat ＞ ✱facēbat, parĭĕtem ＞ parētem, quĭētus ＞ quētus; cŏhŏrtem ＞ cōrtem, cŏŏpĕrit ＞ cōpĕrit ✱cŏpĕrit; dŭŏdĕcim ＞ ✱dōdĕcim. The short e and o in prĕndĕre and ✱cŏpĕrit are not accounted for. In mulĭĕrem[25] ＞ Pr. molhęr the i̯ remained long enough (perhaps under the influence of the nominative mŭlier) to palatalize the l.

Initial Syllable.

41. Usually, in the literary language, Latin a ＞ Pr. a; Latin æ, œ, and e, i (without regard to quantity) ＞ Pr. e; Latin o, u (long or short) ＞ Pr. o; Lat. au ＞ Pr. au, unless the next syllable contained an ú, in which case the au was reduced (in the Vulgar Latin time) to a. Ex.: amīcum ＞ amic, caballus ＞ cavals; æquālem ＞ egal, ✱pœnĭtĕre ＞ penẹdre, dēbēre ＞ devẹr, mĕliōrem ＞ melhọr, dīlĕctum ＞ delęit, dīvīnum ＞ devin, dīvīdĕre ＞ devire, fīnīre ＞ fenir, mĭnōrem ＞ menọr; plōrāre ＞ plorar, sōlātium ＞ solatz, cŏlōrem ＞ colọr, ✱vŏlēre ＞ volẹr, mūstēla ＞ mostẹla, sŭbĭnde ＞ sovẹn; aucĕllum ＞ auzęl, audīre ＞ auzir, augŭstum ＞ aọst, ✱augūrium ＞ aür.

1. An initial vowel is occasionally lost, either through elision with the article (✱eclĕsia ＞ ✱eglęisa, la eglęisa ＞ la glęisa) or through the dropping of a prefix (ingĕnium ＞ engẹnh gẹnh): epĭscŏpus ＞ bisbes, alauda ＞ lauzeta, occasiōnem ＞ ocaiso caiso.

2. In a few words the vowel of the initial syllable disappeared, for some unknown reason, before r in Vulgar Latin: ✱cŏrrŏtŭlāre ＞ ✱crŏtŭlāre ＞ crollar, dīrēctus ＞ drēctus[26] ＞ dreitz, quĭrītāre ＞ ✱crītāre ＞ cridar.

3. Domne, used familiarly as a proclitic (§ 19), lost its first syllable, and, before a vowel, was reduced to n. The combinations de n, que n (followed by a proper name) were understood as d’en, qu’en; hence the title en, ‘Sir.’ See Schultz-Gora in Zs., XXVI, 588; Elise Richter in Zs., XXVII, 193; V. Cescini, Manualetto provenzale, 2d ed., 1905, 168 ff.

4. The proclitic ọ probably comes from a V. L. ot, not from aut.

42. The vowel of the initial syllable, especially in verbs, was extremely subject to the influence of analogy: cülhir (ǫ) through cülh (ǫ) ＜ cŏllĭgit, dizẹn ＜ dīcĕntem through dire ＜ dīcĕre, dürar through dür ＜ dūrum and düra ＜ dūrat, finir through fin ＜ fīnem, fivęla through fibla ＜ fībula, puęiar (ǫ) through puęia (ǫ) ＜ ✱pŏdiat.

1. Avangęli (e) is perhaps influenced by avan; blisọ (e) ＜ blas may possibly have been influenced by tiso; gazardọ ＜ wiðarlôn shows the influence of gazanhar; in piucęla (pülcęla) ＜ ✱pūellicĕlla (Zs., XXV, 343) the püu of the first syllable was changed to piu just as pūlĭcem became piuze (see § 38, 3); in vas = ves ＜ ve(r)sus the a is due to the analogy of az ＜ ad; vais is unexplained, vaus follows daus (§ 44, 6). If dessẹ is from de exín, the first syllable is irregular. Beside maniar ＜ manducare are unexplained forms meniar miniar. In duptar (o), suritz (o) the u doubtless represents ụ or ọ, not ü. Girofle ＜ Καρυόφυλλον and olifan orifan ＜ elephantem are French.

43. Sometimes the initial syllable was altered by a change of prefix or a false idea of etymology: aucire ＜ occīdĕre (cf. the Italian and Rumanian forms), diman (e) ＜ de máne (cf. di ＜ dīem), dementre ＜ dum ĭntĕrim (cf. de ＜ de), engǫissa ＜ ✱angŭstia (cf. en ＜ in), envanezir ＜ evanēscĕre, escür ＜ obscūrum (cf. es- ＜ ex-), preọn prefọn (o) ＜ profŭndum, redọn ＜ rotŭndum (re- in V. L.: Schuchardt, Vokalismus des Vulgärlateins, II, 213), trabalh ＜ trepalium (cf. tra- ＜ tra- = trans-). Dimenge (also ditmenge) is from dīe domĭnĭco.

1. On the same principle are doubtless to be explained such double forms as evori (a), saboros (e), socors (e), somondre (e), soror (e). Serori occurs in a Latin inscription.

2. The prefix eccu-, under the influence of ac and atque, became ✱accu- in southern Gaul and elsewhere: aco ＜ ✱accu’hoc, aquel ＜ ✱accu’ĭllum, aquest ＜ ✱accu’ĭstum, aqui ＜ ✱accu’hīc. Eissi ＜ ecce hīc sometimes becomes aissi through the analogy of aissi ＜ ac sīc.

3. In such forms as tresanar, the prefix tres- is French.

44. Local or partial phonetic changes affected the initial syllable of many words: demandar (do-) ＜ demandāre, emplir (üm-) ＜ implēre; ciutat cieutat ＜ cīvĭtātem; eissir issir ＜ exīre, getar gitar ＜ ✱jĕctāre; crear criar ＜ creāre; mercẹ (mar-) ＜ mercēdem; delgat (dal-) ＜ delicātum.

1. Nearly everywhere there is a tendency to change e to o, u, or ü before a labial, especially before m: premier promier prumier, remas romas, semblar somblar, trebalh trubalh. So de ves ＞ ✱do ves ＞ dous.

2. In the 13th century, nearly everywhere, iu ＞ ieu: piucela pieucela.

3. Many dialects of the north and west change ei and e to i: deissendre dissendre, eissam issam, eissi issi, eissilh issilh, leisso lisso, meitat mitat; degerir (i), denhar (i), disnar, en in, enfern (i), entrar (i), envers (i), escien icient, proclitic est ist, estar (i), estiers (i), Felip (i), gelos (i), genhos (i), genolh (i), gequir (i), guereiar (i), guerensa (i), i(n)vern, isnel irnel, peior pigor, proclitic per pir, premier (i), semblar (i), serven (i), serventes (i), sevals (i), trebalhar (i), tremblar (i). In disnar, ivern, isnel only i is found. In some dialects there is an alternation of e and i, e being used when there is an í in the next syllable, i when there is none, fenít, sirvén. In vezí ＜ vīcīnum the e probably goes back to V. L.: cf. Fr.

4. In a few dialects e in hiatus with a following vowel becomes i: crear criar, leal lial, prear priar, preon prion, real rial.

5. In many dialects of the north and west e has a tendency to become a before r: guerentia garensa, merce marce, pergamen pargamen.

6. In some dialects there is a tendency to assimilate e to an á in the next syllable: delgat dalgat, gigant iaian, deman (a), semblar (a), serrar (a), tremblar (a). So de vás (§ 42, 1) ＞ da vás ＞ dávas; hence daus, under the influence of deus ＞ de ves.

Intertonic Syllable.

45. The term intertonic is applied to the syllable that follows the secondary (§ 18) and precedes the primary accent. In this position all vowels, except a, regularly disappeared in popular words, probably between the 5th and the 8th century[27]; a apparently remained: ✱bŭllĭcāre ＞ boiar (bollegar), bŏnĭtātem ＞ bontat, ✱carrĭcāre ＞ carcar cargar, caballĭcāre ＞ cavalcar cavalgar, cĕrĕbĕllum ＞ cervęl, cīvĭtātem ＞ ciutat, cŏllŏcāre ＞ colcar colgar, dēlĭcātum ＞ delcat delgat, excommūnĭcāre ✱excommĭnĭcāre ＞ escomeniar, vĕrēcŭndia ＞ vergọnha; calamĕllum ＞ calamęl, invadĕre ✱invadīre ＞ envazir, margarīta ＞ margarida, mīrabĭlia ＞ miravilha, parav(e)rēdus ＞ palafrẹs.

1. The vowel is preserved in a number of words in which it originally bore the secondary accent (§ 18): abbréviáre ＞ abreuiar, calúmniáre ＞ caloniar, ✱eríciónem ＞ erisso; on the other hand, ✱cominítiáre (through ✱comín’tiáre) ＞ comensar, partítiónem (through ✱pártiónem) ＞ parso. Cf. Zs., XXVII, 576, 684, 693, 698, 701, 704. When kept, the vowel is sometimes altered: ✱carōnea ✱caróneáta ＞ caraunhada, ✱cupídietósus ＞ cobeitos cobitos, papíliónem ＞ pabalho.

2. The prefix mĭnus- was reduced to mis- (or mes-) in Gaul, perhaps at the close of the Vulgar Latin period: ✱mínus-prétiat ＞ mespreza. Menes- was used also. Cf. P. Marchot, Phon., pp. 43, 44.

3. Mostier is from ✱monistĕrium, altered, by the influence of ministĕrium, from monastērium. Comprar is from V. L. comperare. Calmelh calmelha (cf. calamel above) are Provençal formations from calm. Caresma or caresme seems to be from V. L. ✱quarrēsĭma = quadragēsĭma. Anedier ＜ anatarium shows the influence of anét ánet ＜ anătem (§ 48, 1).

4. In learned words the vowel is generally preserved: irregulár, irritár, pelicán, philozophía. The vowel is, however, often altered, the exchange of e and i being particularly frequent: esperít, femeníl, orifán, peligrí (e), soteirán (sotrán) ＜ subterraneum influenced by dereirán and primeirán.

46. Very often the intertonic vowel was preserved by the analogy of some cognate word or form in which that vowel was stressed: devinár through devín, finimén through finír, guerreiár through guerréia, noiridúra through noirír, oblidár through oblít, pertusár through pertúsa, reusar through reúsa, servidór through servíre.

1. In such cases the preserved vowel is sometimes altered, the exchange of e and i being especially common: avinén, covinén, sovinénsa, cf. venír, ven; enginhár, enginhós, cf. genh; envelzír, cf. vil; gememén, cf. gemír; issarnít (eissernít), from excĕrnĕre; randóla, from hirŭndŭla, perhaps influenced by randón; temerós (o), from ✱timorōsus, influenced by temér; traazó (i), from traditiōnem, with a substitution of suffix; volentiérs, from voluntarius, under the influence of volén ＜ volentem.

Penult.[28]

47. (1) The vowel of the penult of proparoxytones fell in many words in Vulgar Latin, especially between a labial and another consonant, and between two consonants one of which was a liquid: ✱avĭca ＞ ✱auca, cŏm(i)tem, cŏmp(u)tum, dēb(i)tum, dŏm(i)nus[29]; alt(e)ra, vĭg(i)lat, cal(i)dus, vĭr(i)dem; frig(i)dus, nĭtĭdus ＞ ✱nĭttus, pŏs(i)tus, pūtĭdus ＞ ✱pūttus.

(2) The classic Latin -culus comprises an original -clus (sæclum) and an original -culus (aurĭcŭla). In popular Latin both were -clus (✱macla, ŏclus, etc.), to which was assimilated -tŭlus in current words (vĕtŭlus ＞ vĕclus, etc.).

(3) Many popular words which in Vulgar Latin had very generally lost the vowel were for some reason introduced into southern Gaul in their classical forms, and not a few were adopted both in the uncontracted and in the syncopated state: fragĭlem ＞ frágel (cf. Fr. fraile, It. frale), jŭvĕnem ＞ iọve (cf. Fr. iuevne); clĕrĭcum ＞ clęrgue clĕr’cum ＞ clęrc, dēbĭtum ＞ dẹute dēb’tum ＞ dẹpte, flēbĭlem ＞ frẹvol flēb’lem ＞ frẹble, mal’habĭtum ＞ malaute mal’hab’tum ＞ malapte, nĭtĭdum ＞ nẹde ✱nĭttum ＞ nẹt, hŏmĭnem ＞ ome hŏm’nem ＞ omne, pŏpŭlum ＞ pǫbol pŏp’lum ＞ pǫble.

1. Cŏgnĭtum seems to have become ✱cónhede, whence coinde cuende conge. Cf. § 79, Gnd, Gnt.

48. The unaccented penult vowels that had not already fallen disappeared, in most cases, in the transition from Latin to Provençal: ✱carrĭcat ＞ carca, cŏllŏcat ＞ cǫlca, cŭrrĕre ＞ cọrre, spathŭla ＞ espatla, ✱ĕssĕre (= ĕsse) ＞ ęstre, ī(n)sŭla isla, pĕssĭmum ＞ pęsme, pōnĕre ＞ pọnre, ✱rīdĕre ＞ rire, tabŭla ＞ taula, tŏllĕre ＞ tǫlre.

1. A apparently was more tenacious than other vowels, and frequently remained as an indistinct e: anătem ＞ ánet, which, being associated with the diminutive ending -ét, became anét (cf. modern Pr. anèdo); cannăbim ＞ cánebe (learned?); cŏlăphum ＞ ✱cólebe ＞ cǫlbe, but cŏl’phum ＞ cǫlp; Stĕphănum ＞ Estęve; lampăda ＞ lámpeza; ŏrgănum ＞ órguene (later orguéne) órgue; ŏrphănum ＞ ǫrfe; raphănum ＞ ráfe; Rhŏdănum ＞ Rǫzer; ✱sēcăle (= sĕcāle) ＞ séguel (but cf. modern segle selho). Cf. A. Thomas in the Journal des savants, June, 1901, p. 370. See also P. Marchot, Phon., pp. 90-94. Cf. § 45, footnote. It is noteworthy that ✱cólebe ultimately lost its penult, while the other words lost the final syllable or none.

49. Under certain conditions, however, a vowel which had not fallen in the Latin of southern Gaul was often kept in Provençal. It was then probably indistinct in sound, and was written usually e, but occasionally o.

(1) After c´, g´, or y the vowel was apparently retained in some dialects and lost in others. When the c´, g´, or y was intervocalic, forms with and without the vowel are about equally common; when the c´, g´, or y was preceded by a consonant, forms with the vowel predominate, and after cons. + c´ the vowel was apparently never lost. After intervocalic c´: cŏcĕre (= cŏquĕre) ＞ cǫire cǫzer, dīcĕre ＞ dire dízer, dūcĕre ＞ düire ✱düzer (condücir dedüzir), facĕre ＞ faire ✱fázer (fazedọr, etc.), gracĭlem ＞ graile, ✱nŏcĕre (= nŏcēre) ＞ nǫire nǫzer, placĭtum ＞ plach, sŏcĕrum ＞ sǫzer (sǫgre is from sŏcrum), ✱vŏcĭtum (= vacuum) ＞ vuech. After intervocalic g´ or y: bajŭlus ＞ bailes, fragĭlem ＞ frágel, imagĭnem ＞ imáge, lĕgĕre ＞ lęire legír (through ✱lęger?), rĭgĭdum ＞ rẹide rẹge, rĭgĭda ＞ rẹgeza, ✱tragĕre (= trahĕre) ＞ traire tragír (through ✱tráger?). After cons. + c´: carcer ＞ cárcer, crēscĕre ＞ crẹisser, nascĕre ＞ náisser, pascĕre ＞ páisser, parcĕre ＞ párcer, ✱tŏrcĕre (= tŏrquēre) ＞ tǫrzer. After cons. + g´ or y: angĕlum ＞ ángel (learned?), ✱cŏll’gĕre (= cŏllĭgĕre, through cŏllĭgo etc.) ＞ cǫlre cuelher colhír, ✱dē-ēr’gĕre (= ērĭgĕre) ＞ dẹrdre dẹrzer, ✱fŭlgĕrem (from fŭlger = fŭlgur) ＞ fọuzer, jŭngĕre ＞ iọnher, margĭnem ＞ marge, plangĕre> planher, vĭrgĭnem ＞ vẹrgena vẹrge.

(2) After ks, s, ss, and sy the vowel was apparently retained in some dialects and lost in others: dīxĕrunt ＞ diron dissęron (through ✱dísseron)[30], dūxĕrunt ＞ düystrent düissęron (✱dúisseron), fraxĭnum ＞ fraisne fraisse, traxĕrunt ＞ traissęron (✱tráisseron), tŏxĭcum ＞ tuęissec; asĭnum ＞ asne ase, mĭsĕrum ＞ miser (learned), ✱prē(n)sĕrunt ＞ prẹson prezęron (✱prẹzeron), rema(n)sĕrunt ＞ remastrent remasęron (✱remáseron); ✱ĕssĕre (= ĕsse) ＞ ęstre ęsser (used in Rouergue, Limousin, Marche, and Dauphiné), ✱mĭssĕrunt (= mīsĕrunt) ＞ mestrunt (mẹsdren) mesęron (✱mẹsseron), passĕrem ＞ pásser; ✱cō(n)sĕre (= consuĕre) ＞ coser (cozír is from V. L. ✱cosīre).

(3) Between a labial and a dental the vowel was apparently kept: cŭpĭdum ＞ cọbe,[31] fēmĭna ＞ fẹmena fẹme (but fēm’na ＞ fẹmna), jŭvĕnem ＞ iọve, ✱lūmĭnem ＞ lüme (lūmen ＞ lüm), hŏmĭnem ＞ ómen óme (but hŏm’nĕm ＞ omne), tĕpĭdum ＞ tębe,[31] tĕrmĭnum ＞ tęrme. Cf. § 48, 1.

(4) Between a dental and a guttural the vowel remained long enough for the guttural to become y (§ 52; § 65, G): mĕdĭcum ＞ ✱mędegu ＞ ✱mędeye ＞ mędže (= mege). If the first consonant was a liquid or a nasal, the vowel apparently allowed the guttural to become y in some dialects, but not in others: ✱carrĭcat ＞ caria carga, clĕrĭcum ＞ clęrie clęrgue, mŏnăchum ＞ monie mongue. Caballĭcat ＞ cavalga, cŏllŏcat ＞ cǫlca cǫlga show an earlier fall. In clĕr´cum ＞ clęrc the fall goes back to Latin times.

(5) Between lv and r the vowel was kept in some dialects and lost in others: sŏlvĕre ＞ sǫlver sǫlvre, vŏlvĕre ＞ vǫlver vǫlvre, pŭlvĕrem ＞ polvęra.

50. Some learned proparoxytones kept for a while both post-tonic vowels (usually written e), but most of them ultimately either shifted their accent to the penult (§ 17, 1) or dropped their final syllable: domĕstĭcum ＞ domęstegue, lacrĭma ＞ lágrema, mĕrĭtum ＞ męrite, hŏrrĭda ＞ ǫreza, rēgĭmen ＞ régeme; fĭstŭla ＞ festóla, fragĭlem ＞ fragíl, mĕrĭtum ＞ merít, tĕrmĭnum ＞ termíni; diacŏnum ＞ diágue, flēbĭlem ＞ frẹvol (cf. flēb´lem ＞ frẹble), nĭtĭdum ＞ nẹde (cf. ✱nĭttum ＞ nẹt), ōrdĭnem ＞ órde, pŏpŭlum ＞ pǫbol (cf. pŏp´lum ＞ pǫble), prīncĭpem ＞ príncep prínce. Cf. § 47, (3).

Final Syllable.

51. As early as the 8th century, in popular words, the vowels of final syllables fell, the fall occurring first, perhaps, after liquids: hĕrĭ ＞ ęr, malĕ ＞ mal; bŏnŭs ＞ bos, cŏlăphŭm ＞ cǫlp, cōgĭto ＞ cüg, panĕm ＞ pan, prĕtiŭm ＞ prętz.

(1) Latin a, however, remained, being generally pronounced ạ: audiăm ＞ auiạ, bŏnă ＞ bonạ, fīliās ＞ filhạs.[32]

(2) Latin final ī probably remained in all dialects later than the 8th century, and in some until the beginning of the literary period: hábuī ＞ águi ＞ aguí. Before it fell, it changed an accented ẹ in the preceding syllable to ị: see § 27.

(3) Latin i and u remained if they were immediately preceded by an accented vowel: fuī ＞ füi, mĕī ＞ męi, sŭī ＞ sọi; cavum ✱caum ＞ chau, dĕus ＞ dęus, ĕgo ✱ĕo ✱ĕu ＞ ęu, rīvum rīum ＞ riu. In such cases the two vowels formed a diphthong.

(4) Before final nt Latin e, u remained as e, o: cantent ＞ canten, vēndunt ＞ vẹndon.

1. In Aude, Tarn, Aveyron, Corrèze, and a part of Haute-Garonne, final ī was preserved as late as the 12th century: pagadi, salvi, soli. See Rom., XIV, 291-2 and XXXIV, 362. Such forms occur also in Vaud and Dauphiné. Cf. Gram., II, p. 82.—In the dialect of some texts, -ī, before falling, palatalized a preceding l (or ll), n (or nn), nd, nt, or t: annī ＞ anh, bellī ＞ beill, ✱infantī ＞ efanh, ✱spiritī ＞ esperih, mundī ＞ monh. See Rom., XXXIV, 353.

2. In the extreme east there are traces of final -ōs: aquestos, ellos, tantos.

3. Grau for gra ＜ gradum, niu for ni ＜ nīdum are Catalan. Amiu for amic ＜ amīcum, chastiu for chastic ＜ castīgo belong to the dialect of Forez, and point to a very early fall of the guttural in that dialect. Cf. § 65, D, G.

4. Aire, vaire, beside air ＜ aĕrem, vair ＜ varium, probably show the influence of the numerous nouns in -aire (amaire, etc.); cf. § 52, (1). Fores beside fors ＜ fŏris, nemes beside nems ＜ nĭmis, senes beside sens ＜ sĭne probably developed the e when the next word began with a consonant: see § 62, (3). For colbe, see § 48, 1. Reide rede perhaps owes its -e to rege: § 49, (1). Beside volp ＜ vŭlpem there is a volpe.

5. Coma, beside com, con, co ＜ quōmŏ(do), apparently owes its -a to the analogy of the adverbs bona and mala and other adverbs of manner. For a different explanation, see J. Vising in the Tobler Festschrift (Abhandlungen Herrn Prof. Dr. Tobler… dargebracht, 1895), p. 113.

6. E seems to have been inserted in the second person singular of some verbs, to distinguish it from the third person: co(g)nōscis ＞ conọisses, co(g)nōscit ＞ conọis.

52. When the fall of the vowel would have resulted in an undesirable consonant group at the end of a word, the vowel was retained as an indistinct e: dŭbĭto ＞ dọpte, lŭcrum ＞ lọgre.

The principal groups that call for a supporting vowel are: (1) a consonant and a liquid; (2) a labial and a dental; (3) in proparoxytones, a consonant and a c´ or c originally separated by the vowel of the penult; (4) in proparoxytones, a consonant and an m or n originally separated by the vowel of the penult. Ex.: ĭnter ＞ ẹntre; aptum ＞ apte; ✱dōdĕcim ＞ dọtze, jūdĭco ＞ iütge; ✱Jacŏmus ＞ Iacmes, asĭnum ＞ asne.

If the word was a paroxytone, and the first consonant was a palatal and the second an r, the supporting vowel stood between the two: major ＞ maier, mĕlior ＞ męlher, mŭlier ＞ mọlher, pĕjor ＞ pęier, sĕnior ＞ sęnher. Otherwise the supporting vowel followed the consonant group.

The four classes of groups (aside from the palatal + r just mentioned) will now be examined in detail:—

(1) Examples: alter ＞ autre, Carŏlus ＞ Carles, dŭplus ＞ dọbles, ✱ĕssĕre ＞ ęstre, fabrum ＞ fabre, ✱mĕr(ŭ)lum ＞ męrle, nŏster ＞ nǫstre, pauper ＞ paubre, pŏp(ŭ)lum ＞ pǫble, pōnĕre ＞ pọnre, recĭpĕre ＞ recẹbre, rŭmpĕre ＞ rọmpre, tŏllĕre ＞ tǫlre, vŏlvĕre ＞ vǫlvre; mascŭlum ＞ mascle, etc.; flēb(ĭ)lem ＞ frẹble, etc. Under this head is included r-r (cŭrrere ＞ cọrre, quærĕre ＞ quęrre), but not ll and rr (bĕllum ＞ bęl, fĕrrum ＞ fęr). In Provençal the first element was often changed, later than the 8th century, into a vowel, original b and v becoming u, and d, t, c, g, and y being turned to i: bĭbĕre ＞ bẹure, scrībĕre ＞ escriure, ✱mŏvĕre ＞ mǫure, plŏvĕre ＞ plǫure, vīvĕre ＞ viure; latro ＞ laire, matrem ＞ maire, radĕre ＞ raire, ✱rīdĕre ＞ rire, vĭtrum ＞ vẹire; desīdĕro ＞ desire, etc.; amātor ＞ amaire, servītor ＞ servire, etc.; dīcĕre ＞ dire, dūcĕre ＞ düire, facĕre ＞ faire, gracĭlem ＞ graile, ✱tacĕre ＞ taire; frīgĕre ＞ frire, weigăro gaire, lĕgĕre ＞ lęire; bajŭlum ＞ baile. Apparent exceptions to the rule are intervocalic cl, gl, which were probably reduced to single consonants before the 8th century: ŏc(ŭ)lum ＞ ǫlh, vĭg(ĭ)lo ＞ vẹlh.

1. The rare forms frair, mair, pair, Peir (still used in Gascony), beside regular fraire, maire, paire, Peire, are probably due to proclitic use; so sor beside sorre ＜ sŏror, and possibly faur beside faure ＜ faber. The learned albir = albire ＜ arbĭtrium may be due to the analogy of other double forms. Dimerc for dimercre (§ 17, 2) perhaps follows dimenc.

2. Rr requires a vowel in a few dialects: corre = cor ＜ cŭrrit, ferre = fer ＜ fĕrrum, torre = tor ＜ tŭrrem.

(2) Examples: cŭbĭtum ＞ cọde; cŏmĭtem ＞ comte; dēbĭtum ＞ dẹpte dẹute, § 47, (3); dŏmnum ＞ domne; dŭbĭto ＞ dọpte; hŏspĭtem ＞ ǫste; sabbătum ＞ sapte.

1. Azaut seems to be post-verbal from azautar ＜ adaptāre. Escrit ＜ scrīptum shows the influence of dit ＜ dīctum. Malaut, beside malaute malapte ＜ mal´habĭtum, is reconstructed from the feminine malauta on the model of aut, auta. Set ＜ sĕptem must have developed as a proclitic.

(3) Examples: jūdicem ＞ iütge[33]; pŏllĭcem ＞ pǫuze; quīndĕcim ＞ quinze; salĭcem ＞ sauze; sēdĕcim ＞ sẹdze;—canŏnĭcum ＞ canonge canọrgue,[34] § 49, (4); clĕrĭcum ＞ clęrge clęrgue (§ 48, 2); mĕdĭcum ＞ mętge; mŏnăchum ＞ monge mongue mǫrgue,[34] § 49, (4); vĭndĭco ＞ vẹnie; viatĭcum ＞ viatge, etc.

1. The forms poutz, sautz, beside pouse, sauze, would seem to indicate that lc´ did not require a supporting vowel in all dialects.

2. ✱Ficotum (= jēcur), a fusion of συκωτόν (‘fig-fattened’) and fīcus, combined with ✱hēpăte (= hēpar), became ✱fẹ́catu ✱fẹcitu ✱fẹgidu, and then, through the influence of the familiar ending -igu (= ĭcum), ✱fẹdigu ＞ fẹtge. See G. Paris in Miscellanea linguistica in onore di G. Ascoli, 1901, p. 41; H. Schuchardt in Zs., XXV, 615, and XXVIII, 435; L. Clédat in Revue de philologie française et de littérature, XV, 235. Pege, for peich ＜ pĕctus, seems to be due to the analogy of fetge.

(4) Examples: æstĭmo ＞ esme; dĕcĭmum ＞ dęsme; fraxĭnum ＞ fraisne; incūdĭnem ＞ enclütge (cf. § 80, Dn); ✱metĭpsĭmum ＞ medẹsme; pĕssĭmum ＞ pęsme; prŏxĭmus ＞ prǫsmes.

1. Faim ＜ facĭmu(s) doubtless lost its -e through the analogy of the alternative form fazẹm ＜ ✱facímu(s) and of the usual endings -ám, -ẹm.

(5) In some dialects, at least, by, mby, mny, py, rny required a supporting vowel: rŭbeum ＞ rọtge, cambio ＞ camie, sŏmnium ＞ songe suenh, apium ＞ ache api, ✱Arvĕrnium[35] ＞ Alvęrnhe; ratge (= rabiem) is probably French. Original lm, rm, sm required a supporting vowel in some dialects but not in others: hëlm ＞ ęlme ęlm, ŭlmum ＞ ọlme ọlm, palmum ＞ palm; ✱ĕrmum ἔρημον ＞ erm, fĭrmum ＞ fẹrm, gĕrmen ＞ gęrme; spasmum ＞ espasme.

(6) Many verbs regularly have an -e in the first person singular of the present indicative: desire, dọpte, iütge, etc. By the analogy of these, -e often appears in the first person singular of verbs which need no supporting vowel: remīro ＞ remir remire. By the analogy of the preterit (águi, füi, etc.), -i is very often substituted for this -e: azọr azọri, cant canti, prętz pręzi, etc.

53. Many late words preserve the final vowel as -e: benigne, bisbe ＜ epĭscŏpum, digne (cf. denhar), mixte (cf. mẹst), regne (cf. reing), signe (cf. sẹnh). Cf. § 50; (for cǫlbe) § 48, 1; and (for cǫinde, etc.) § 47, 1. Learned formations from nouns in -ium usually end in -i, simply dropping the -um: capitǫli, edifici, empęri, iüzízi, martíri (martíre), negǫci, ofíci, periüri, remęzi, servízi, vici. Similar forms in -i were sometimes taken from the accusative of nouns and adjectives in -ius: Boęci ＜ Boëthium, prǫpri (prǫpre) ＜ prŏprium, savi ＜ sabium.

1. It should be remembered that the Latin words, at the time of their adoption, had undergone various phonetic changes in the clerical pronunciation: cf. § 15. A form remezi, for instance, presupposes a pronunciation of remĕdium as remęðiu(m).

3. CONSONANTS.

54. The Latin consonants which we have to consider are: b, c (= k), d, f, g, h, j (= y), l, m, n, p, qu (= kw), r, s, t, v (= w), x (= ks). To these we must add the Vulgar Latin w coming from u̯, and y coming from e̯, i̯: see § 40, (2). Furthermore, in words borrowed from Germanic dialects we find b, ð, h, k, þ, w, which call for special notice; and, in words borrowed from Greek, ch, k, ph, th, z.

The Latin d, f, j, l, p, t call for no remark at present. Latin h, in popular speech, became silent very early (hŏc ＞ ŏc, hŏmo ＞ ŏmo), and, although an attempt was made to restore it in polite speech, it left no trace in the Romance languages: cf. Rom., XI, 399. Double consonants were pronounced distinctly longer than single ones: annus, ĭlle, ŏssum, tĕrra.

55. Latin b, c, g, m, n, qu, r, s, v, w, x, y show the following developments in popular Latin speech:—

B between vowels became, through failure to close the lips tightly, β (bilabial v), from the 1st to the 3d century of our era: habēre ＞ aβẹre. The same change took place, to a certain extent, when the b was not intervocalic, but we have few, if any, traces of it in Provençal. Between vowels, even in learned words, the clerical pronunciation was probably β or v until the 7th century. Cf. V.

C before a front vowel (e, i), as early as the 3d century, doubtless had, in nearly all the Empire, a front or palatal articulation; that is, it was formed as close as possible to the following vowel[36]: cĕntum ＞ c´ĕntu, dūcĕre ＞ dūc´ĕre. The next step was the introduction of an audible glide, a brief y, between the c´ and the vowel[37]: c´yęntu, dục´yere. By the 5th century this c´y had developed into a kind of ty, the c´ having been drawn still further forward: t´yęntu dụt´yere. Through a modification of the y-glide, the group then became, in the 6th or 7th century, tš or ts: tšęntu tsęntu. See H. Schuchardt, Voc., I, 151, and Ltblt., XIV, 360; P. E. Guarnerio, in Supplementi all’ Archivio glottologico italiano, IV (1897), pp. 21-51 (cf. Rom., XXX, 617); G. Paris, in the Journal des savants, 1900, 359, in the Annuaire de l’École pratique des Hautes-Études, 1893, 7, in the Comptes rendus des séances de l’Académie des Inscriptions, 1893, 81, and in Rom., XXXIII, 322; W. Meyer-Lübke, Einf., pp. 123-126; F. G. Mohl, Zs., XXVI, 595; P. Marchot, Phon., pp. 51-53; W. Meyer-Lübke, in Bausteine zur romanischen Philologie, 313. Cf. G and X.[38]

G between vowels, before the accent, disappeared in some words in at least a part of the Empire: le(g)ālis, li(g)āmen, re(g)ālis, (realis is attested for the 8th century); ĕgo, generally used as a proclitic, everywhere lost its g; on the other hand, g was kept in castigāre, fatigāre, ligāre, negāre, pagānus. G before a front vowel (e, i), by the 1st or 2d century, was pronounced g´ (cf. C): gĕntem ＞ g´ĕnte, fragĭlis ＞ frag´ĭlis. As early as the 4th century this g´, through failure to form a close articulation, opened into y[39]: yęnte, fráyilis. Before an accented e or i an intervocalic y disappeared, in the greater part of the Empire, being fused with the vowel: magĭster ＞ mayįster ＞ maẹster, ✱pagēnsis ＞ payẹsis ＞ paẹsis, regīna ＞ reyịna ＞ reịna.[38]

M and n, when final, were weak and indistinct from the earliest times, except in monosyllables; by the 3d or 4th century they had probably disappeared altogether from the end of polysyllables: damnu, nọme; but jam, non.

N before spirants (f, j, s, v), except in the prefixes con- and in-, became silent during the Republican period, the preceding vowel, if it was short, being lengthened by compensation[40]: mē(n)sis, pē(n)sare. If the syllable con- or in- was not recognized as a prefix, the n fell: co(n)sul, co(n)ventum, i(n)fas. In learned and newly constructed words the n was pronounced. Cf. M.

Qu, gu before o or u were reduced to c, g in the 1st or 2d century: see W.

R before s, in a number of words, became s in the Republican period: deōrsum ＞ deōssum, dŏrsum ＞ dŏssum, sūrsum ＞ sūssum; so, in a part of the Empire, pĕrsĭca ＞ pĕssĭca, vĕrsus (preposition) ＞ vĕssus. Early in our era ss after a long vowel was reduced to s: deōsu, sūsu.

S was probably always voiceless, or surd, in classic Latin, but became voiced between vowels, in Gaul, at the end of the Vulgar Latin period: casa. To initial s + consonant an i or e was prefixed, at first, no doubt, after a word ending in a consonant: in schŏla ＞ in iscŏla; this process began in the 2d century and had become general by the 4th.

V, originally pronounced w, became β probably in the 1st century: vīvĕre ＞ βīβĕre. Before u, v regularly disappeared, but it was restored by analogy in many words: flavus ＞ flaus, ōvum ＞ ŏum, rīvus ＞ rīus; but also ŏvum, rīvus, by the analogy of ova, rivi. In the greater part of the Empire v apparently fell also before an accented o: pavōnem ＞ paōne, pavōrem ＞ paōre. Cf. W. When a β, representing either b or v, became contiguous to a following consonant, it changed to u: ✱avĭca ＞ aβĭca ＞ auca, gabăta ＞ gaβata ＞ gauta, ✱flavĭtat ＞ flaβĭtat ＞ flautat. In several words rv became rb in Latin: vervēcem ＞ berbēce berbīce, cŏrvus ＞ cŏrbus, cŭrvus ＞ cŭrbus.

W coming, in the 2d or 3d century, from u̯ (§ 40) differed from Latin v, then pronounced β, but was probably identical with Germanic w: dēbuī ＞ dẹbwị, placuī ＞ placwị, sapuit ＞ sapwit, tĕnuis ＞ tęnwis. W fell between a consonant and o or u: antīquus ＞ antịcus, battuo ＞ batto, carduus ＞ cardus, cŏquus ＞ cǫcus, distĭnguo ＞ distįngo, mŏrtuus ＞ mǫrtus; so eccu’hŏc ＞ Pr. acǫ. Cf. Qu.

X (= ks) was reduced to s, in the 2d or 3d century, before a consonant or at the end of a word of more than one syllable: sĕstus, sĕnes; but sĕx. So the prefix ex- ＞ es- before any consonant but s: ✱exgaudēre ＞ Pr. esiauzir, ✱exlucēre ＜ Pr. esluzir, ✱exmĭttĕre ＞ Pr. esmẹtre; excernĕre ＞ ✱escernīre ＞ Pr. eissernir. Ex- + s apparently became either ex- or ess-: ✱exsanguinātum ＞ Pr. eissancnat, ✱exsaritāre ＞ Pr. eissartar, ✱exsĕquĕre ＞ Pr. essęgre, ✱exsŭrgĕre ＞ Pr. essọrger, ✱exsūcāre ＞ Pr. eissügar essügar.

Y coming, in the 2d or 3d century, from e̯ or i̯, (§ 40) coincided with Latin j: habeam ＞ abya, eāmus ＞ yamus, tĕneat ＞ tęnyat; audio ＞ audyo, fīlia ＞ fịlya, vĕniat ＞ vęnyat. As early as the 4th century the groups dy, gy were reduced to y; and ly, ny probably became l´, n´: mĕdius ＞ mędyus ＞ męyus, corrĭgia ＞ corrįgya ＞ corrẹya; mĕlior ＞ męlyor ＞ męl´or, tĕneo ＞ tęnyo ＞ tęn´o.

56. Germanic b, ð, h, k, þ, w call for special mention:—

B did not participate in the change of Latin intervocalic b to β: roubôn ＞ Pr. raubar. The words containing it were evidently adopted after this phonetic law had ceased to operate.

ð, þ were pronounced by the Latins as d, t: ✱waiðanjan ＞ ✱wadanyāre ＞ Pr. gazanhar (It. guadagnare), þrëscan ＞ ✱trescāre ＞ Pr. trescar.

H, at the beginning of a word, was lost in the greater part of the Empire, including southern Gaul: hapja ＞ ✱apya ＞ Pr. apcha. H between vowels was lost in some words and replaced by kk in others: spëhôn ＞ Pr. espiar, fëhu ＞ Pr. fęu, jëhan ＞ ✱yekkīre ＞ Pr. gequir. Ht was regularly replaced by tt: slahta ＞ ✱sclatta ＞ Pr. esclata; but wahta, perhaps borrowed at a different period, became Pr. gaita.

K, in southern Gaul, did not take the palatal pronunciation before front vowels: skërnon ＞ Pr. esquernir, skina ＞ Pr. esquina, skiuhan ＞ Pr. esquivar, ✱rîk-ĭtia ＞ Pr. riquẹza; only the derivatives of Franko (doubtless Latinized early) show palatalization, as ✱Francia ＞ Pr. Fransa. G, however, seems to have been palatalized: gîga ＞ Pr. giga, geisla ＞ Pr. giscle. Before a, in words introduced early, k and g were treated like Latin c and g: kausjan ＞ Pr. cauzir chauzir, gâhi ＞ Pr. gai iai; see § 11, (1).

W was vigorously pronounced, and, through reinforcement of its velar element, came to be sounded gw: warjan ＞ ✱warīre gwarīre ＞ Pr. garir, wërra ＞ ✱węrra gwęrra ＞ Pr. guęrra.

57. Greek ζ, θ, κ, φ, χ did not exactly correspond to any Latin consonants:—

Z, whatever may have been its original pronunciation, received in Vulgar Latin the value dy, which then, like any other dy, became y: ✱zelōsus (from ζῆλος) = dyelọsus yelọsus ＞ Pr. gelọs. The infinitive ending -ίζειν, introduced in such words as βαπτίζειν, ＞ baptizāre = bapti(d)yāre, became very common in the form -įdyāre -įyāre, and was used to make new verbs: wërra + ίζειν ＞ ✱werrįdyāre gwerrįyāre ＞ Pr. guerrẹiar.

θ, in the popular speech of Rome, was replaced by t: similarly χ was replaced by c: σπαθή ＞ spatha = spata; χορδή ＞ chŏrda = cǫrda.

κ was apparently intermediate in sound between Latin c and g; it was generally replaced by the former, but sometimes by the latter: κατά ＞ cata, κυβερνᾰν ＞ gubernare.

φ, in Greek, was in early times (perhaps until the 4th century of our era) a strongly explosive p; it then developed into f. In words borrowed by the Romans in the early period it was replaced by p; in later words it was sounded f: κόλαφος ＞ cŏlăphus = cǫlapus, φασίολος ＞ phaseŏlus faseŏlus.

58. The fate of all these consonants in Provençal depended largely on their position in the word: we must therefore distinguish initial, medial, and final consonants. In a general way, the first tended to remain unchanged, the second to weaken, the third to disappear. Furthermore we must separate single consonants from consonant groups: the latter resisted change better than the former; but a group consisting of dissimilar elements tended to assimilate them.

INITIAL CONSONANTS.

59. A consonant preceded by a prefix was treated as an initial consonant as long as the character of this preceding syllable was recognized: de-cadĕre ＞ decazẹr, de-pĭngĕre ＞ depẹnher, præ-parāre ＞ preparar, re-cordāre ＞ recordar, re-patriāre ＞ repairar, re-pausāre ＞ repausar, se-dūcĕre ＞ sedüire. If, however, the initial syllable ceased to be recognized as a prefix, the following consonant was treated as a medial consonant: præpŏsĭtum ＞ prebǫst, retŏrta ＞ redǫrta; so, perhaps, profŭndum ＞ preọn. The rare rebọnre (beside repọnre) ＜ re-pōnĕre has the special sense ‘to bury’.

Single Initial Consonants.

60. B, d, l, m, n, p, r, s, t underwent no change: ben, dọn, lǫc, mẹ, nau, pauc, rius, si, tü.

1. For cremetar ＜ ✱tremitāre, see Meyer-Lübke, Einf., § 194. For granolha ＜ ✱ranŭcŭla, see Körting, ranuculus.

61. C, c´, f, g, g´, β, y suffered some change. C, g must be distinguished from c´, g´: § 55, C, G.

C, g before o, u remained unchanged: colōrem ＞ colọr, cūra ＞ cüra; gŭla ＞ gọla, gŭtta ＞ gọta. Before a they changed only in the north and northeast, where they became (perhaps from the 7th to the 9th century) respectively tš and dž: campus ＞ camps champs; gaudēre ＞ gauzir iauzir.

C´ ＞ ts, which just before and during the literary period was reduced to s: cælum ＞ cęl sęl, cīvitātem ＞ ciutat ciptat siptat. For g´, see Y.

Y, comprising Latin dy, g´, gy, j, and z, became dž (except in Béarn, where it remained y): diurnālem ＞ iornal (yornal), deō(r)sum ＞ iọs; gĕlus ＞ gęls, gentīlem ＞ gentil (yentil), gȳrāre ＞ girar; jam ＞ ia, jŏcum ＞ iǫc (yǫc), jŭvĕnem ＞ iọve; ✱zelōsus ＞ gelọs.

F remained unchanged, except in Béarn and a part of Gascony, where it became h: famem ＞ fam ham, fĭdem ＞ fẹ hẹ, fŏcum ＞ fǫc hüc, fŏlia ＞ fuelha huelha.

β ＞ v (the dentilabial spirant), except in Béarn, Gascony, and parts of Languedoc, where it became b: vĕnit ＞ ven be, vĕntum ＞ vent bent, vĕrsus (§ 55, R) ＞ vęs bęs, vōs ＞ vọs bọs.

1. In a few words β, owing to Germanic influence, was replaced by w ＞ gw: vadum + watan ＞ gua, vastare + wôst ＞ guastar. So vagīna ＞ guaína, Vascŏnia ＞ Gasconha. Cf. gw below.

Initial Groups.

62. There are three classes of groups: those ending in l or r, those ending in w, and those beginning with s:—

(1) Bl, br, cl, cr, dr, gl, gr, pl, pr, tr underwent no change: blasphemāre ＞ blasmar, brĕvem ＞ bręu, clarus ＞ clars, crŭcem ＞ crọtz, drappus ＞ draps, glaciem ＞ glatz, gradum ＞ gra, plēnum ＞ plẹn, precāre ＞ pregar, trans ＞ tras. Gras is from grassus, a fusion of crassus and grossus. For grǫcs ＜ κρόκος see § 57, κ.

(2) Gw (Germanic w) and kw (Latin qu) were reduced, perhaps in the 10th century, to g and k, except in the west, where the w was retained: wahta ＞ gaita guaita, warjan ＞ garir guarir, wërra ＞ gęrra guęrra, wīsa ＞ gisa guisa; quando ＞ can quan, quare ＞ car quar. It should be noted that the u was commonly kept in the spelling (especially before e and i) after it had ceased to be pronounced, gu and qu being regarded merely as symbols for “hard” g and c. For cinc ＜ quīnque, see § 87, kw. Sw remained in suavem ＞ suau.

(3) To groups beginning with s a vowel had been prefixed in Vulgar Latin (§ 55, S); this vowel appears in Provençal as e. Sc´ apparently did not occur in any popular word; sl early became scl; the other groups (sc, scl, scr, sp, st, str) remained unchanged, except that in the north and northeast sc ＞ stš before a: scala ＞ escala eschala, schŏa ＞ escǫla, slahta ＞ ✱sclatta ＞ esclata, scrībĕre ＞ escriure, spīna ＞ espina, stare ＞ estar, strĭngĕre ＞ estrẹnher.

MEDIAL CONSONANTS.

63. (1) It is well to note at the outset that when, through the fall of an unaccented vowel (§ 51), an early Provençal b, d, dz, dž, g, z, or ž was made final or contiguous to a final s, it became voiceless: ŏpus ＞ ǫbus ＞ ǫbs ǫps, ŏrbum ＞ ǫrbu ＞ ǫrb ǫrp: datum ＞ dadu ＞ dad dat, vĭrĭdem vĭrdem ＞ vẹrde ＞ vẹrd vẹrt; prĕtium ＞ prędzu ＞ prędz pręts (written pretz), vōcem ＞ vọdze ＞ vọdz vọts (written votz); mĕdium ＞ męyu mędžu ＞ mędž mętš (written meg or mech); amīcus ＞ amigus ＞ amigs amics, largum ＞ largu ＞ larg larc; rīsum ＞ rizu ＞ riz ris; basium ＞ bažu ＞ baiž baiš (bais). The combination tšs, however, loses either its second or its third element: ✱gaudios ＞ gautšs ＞ gautš or gauts (both of them often written gaugz); so nŏctes ＞ nuetšs ＞ nuetš or nuets (nuegz). For local variations of g, see (6). For apud ＞ ab ap am an, see § 65, P. 2.

(2) Under the same conditions, y became i: vĭdeo ＞ vẹyo ＞ vẹy vẹi, pĕjus ＞ pęyus ＞ pęys pięis.

(3) Under the same conditions, ð, coming from intervocalic d, fell when final, but became t before s: audit ＞ auði ＞ auð au; crūdus ＞ cruðus ＞ cruds crüts. So crūdum ＞ crü, fĭdem ＞ fẹ, fraudem ＞ frau, gradum ＞ gra, nīdum ＞ ni, nōdum ＞ nọ, pĕdem ＞ pę, sapidum ＞ sabe, tĕpidum ＞ tębe; grados ＞ grats, nōdus ＞ nọts, nūdus ＞ nüts, pĕdes ＞ pęts. The two sets of forms influenced each other: hence degras, fes,[41] nis, pes, etc.; crut, grat, not, nut,[41] etc.

(4) Under the same conditions, β, coming from v or from intervocalic b, became u if preceded by a vowel, but fell if preceded by a consonant: bĭbit ＞ bẹβi ＞ bẹβ bẹu, vīvit ＞ βịβi ＞ βịβ vịu, claves ＞ claβes ＞ claβs claus, vīvus ＞ βīβus ＞ βịβs vịus; salvet ＞ sal, salvum ＞ sal, sĕrvit ＞ siẹr, nĕrvos ＞ nęrs, salvus ＞ sals, sĕrvus ＞ sęrs. Sometimes, however, final β preceded by a consonant, instead of falling, became f: salvet ＞ salf, salvum ＞ salf, sĕrvit ＞ sięrf, vŏlvit ＞ vǫlf; it may be that these are the only regular forms for cons. + β when final, and that sal, sier are due to the analogy of sals, siers.

(5) Under the same conditions, final n, if preceded by a vowel, was kept in the extreme west, parts of the north, and all the southeast and east, but fell everywhere else; n before s was generally kept only in the southeast and east: bĕne ＞ be ben, canem ＞ ca can, sŏnum ＞ so son[42]; bŏnus ＞ bos bons, mansiōnes ＞ maisọs maisọns. In mĭnus ＞ mẹns the n was kept, perhaps through the analogy of menọr. If the n was preceded by a consonant (r), the fall seems to have been even commoner: cŏrnu ＞ cǫr cǫrn, tŏrno ＞ tọr tọrn; diŭrnus ＞ iọrs iọrns. For iọrnh ＜ diŭrnī, see § 51, 1. Provençal n coming from nn never falls: annus ＞ ans.

(6) Under the same conditions, g, representing original c or g, became c after o or u, and after other vowels either became c or was changed to i (which fused with a preceding i): fŏcum ＞ fǫc, lŏcus ＞ lǫcs, paucum ＞ pauc, Hūgo ＞ Uc; Aureliācum ＞ Aurelhac, dīco ＞ dic di, Henrīcum ＞ Enric Enri, ✱trago ＞ trac trai. The forms with c are the commoner; they have been most persistent in the west.

(7) The vocalization of l before s (malus ＞ maus) is a different phenomenon from the foregoing. See § 65, L.

(8) An m or an n that becomes contiguous to final s often develops into mp or nt, but oftener (judging from the spellings) does not: nĭmis ＞ nems nemps; annos ＞ ans anz.

(9) Between a liquid or a nasal and a final s, a b or a p generally fell, unless supported by the analogy of a form in which the b or p was final: ambos ＞ ams ambs, cŏrpus ＞ cors, tĕmpus ＞ tems temps; cf. balbs (balb), orbs (orb).

1. The d, n, t of the proclitics ad, quid, in, aut, et will be treated under Final Consonants.

64. Final ts from any source, in Provence, Limousin, and a part of Languedoc and Gascony, was reduced, during the literary period, to s: amātis ＞ amatz amas, habētis ＞ avętz avęs, dīcit ＞ ditz dis, grandes ＞ granz grans, latus ＞ latz las, prĕtium ＞ prętz pręs. On the other hand, in a part of Limousin (especially in Limoges), and also in Dauphiné, -ts, in the second person plural of verbs, became t: habētis ＞ avęt.

Single Medial Consonants.

65. The single medial consonants will now be considered separately, in alphabetical order:—

β, coming from b or v, became v; except in the west and a part of the centre, where, if it remained intervocalic, it changed to b: habētis ＞ avętz abętz, debēre ＞ devẹr debẹr, faba ＞ fava faba; avārum ＞ avar, æstīva ＞ estiva, brĕvem ＞ bręu,[43] clavem ＞ clau, dīe Jŏvis ＞ diiǫus, lĕvat ＞ lęva, novĕlla ＞ novęla nabęra (Gascon), vīvus ＞ vius. When the preceding or following vowel was o or u, a β before the accent fell in most dialects, being fused with the vowel: abŭndare ＞ aondar abondar, gŭbĕrnare ＞ goernar governar, proclitic ŭbĭ ＞ ọ, prŏbare ＞ proar, sŭbĭnde ＞ soẹn sovẹn sobẹn, trĭbūtum ＞ treüt; Lŭdovīcus ＞ Lozoics, Provĭncia ＞ Proẹnsa Provẹnsa, novĕllum ＞ noęl novęl, novĕmbrem ＞ noembre novembre, pavōnem ＞ paọn, pavōrem ＞ paọr (cf. § 55, V).

1. The perfect endings -avi etc., -ivi etc. had lost their v in Latin. For avia, etc., see § 87, β.

2. Abans, beside avanz, avan ＜ ab ante, apparently shows the influence of Pr. ab = apud. Abet ＜ abiĕtem (§ 40, 1) is unexplained: cf. Italian abete. Abora is a Provençal compound of ab and ora. Trap, beside trau ＜ trabem, is doubtless from the nominative traps ＜ trabs, which seems to have been differentiated in meaning from the V. L. nominative trabis.

3. Brey, grey, ney, beside breu ＜ brĕvem, greu ＜ ✱grĕvem, neu ＜ nĭvem (cf. § 25, 1, e), have been subjected to the attraction of grey ＜ grĕgem, ley ＜ lēgem. Greug is a post-verbal noun from greuiar ＜ ✱grĕviare.

4. Natiz = natius ＜ natīvus seems to have been influenced by mestis ＜ mixtīcius. Massis is from ✱massīcius.

5. Paziment = pavamen owes its z perhaps to the analogy of aizimen.

6. In purely learned words, b and v were written as in Latin: diabol, diluvi.

C, from the 4th to the 6th century, was voiced to g, and then developed like any other g. See G.

1. After au, apparently, c did not change: ✱auca (＜ ✱avĭca ＜ avis) ＞ auca, pauca ＞ pauca, rauca ＞ rauca, ✱traucare (? ＜ ✱trabucare) ＞ traucar. Cf. § 65, P, 3.

2. In purely learned words, c remained unchanged: vocal. Alucar aluchar, aluc seem to be learned formations patterned after antelucānus and Low Latin lucānus.

C´, when it became contiguous to a consonant, through the fall of the unaccented vowel of the penult, was reduced to i: cŏcĕre ＞ cǫire, dīcĕre ＞ diire dire, dīcĭtis ＞ ditz, facĕre ＞ faire, facĭmu(s) ＞ faim, facĭtis ＞ faitz, fēcĕram ＞ fẹira, fēcĕrunt ＞ fẹiron, gracĭlem ＞ graile. When it remained intervocalic, it was assibilated during the transition period (§ 55, C); in most of the Provençal territory it became dz, which during the literary period was simplified to z; but in some dialects of the south and the northwest it resulted in idz (later iz), an i-glide having developed before the consonant while it was still palatal: aucĕllum ＞ auzęl, jacēre ＞ iazẹr, lĭcēre ＞ lezẹr, lūcēre ＞ lüzẹr lüzir lüisir, placēre ＞ plazẹr plaizẹr; crŭcem ＞ crọz crọiz crọis (see §§ 63, 64), dīcere ＞ dízer, dīcit ＞ ditz dis, dūcit ＞ dütz düs, facit ＞ fatz fas, jacet ＞ iatz ias iays, pacem ＞ patz pas pais, placet ＞ platz plas plais, verācem? ＞ verais, vōcem ＞ vọtz vọiz.

1. Aucel, beside auzel, perhaps belongs to a dialect in which c´ was not voiced after au: cf. C, 1. See § 80, Bc´.

2. Iasser, beside iazer, seems to be due to ias ＜ jacet and iassa ＜ jaceat.

3. For desma deima, see S, 1.

4. In purely learned words, c´ ＞ ts: acĭdum ＞ aci.

D, in a part of the west, remained unchanged; elsewhere, during the Vulgar Latin period, it opened into ð, which fell in the 11th century and earlier in parts of the north and east, and in the rest of the Provençal territory became z as early as the first part of the 12th century: audīre ＞ auzir auir audir,[44] audit ＞ au,[45] cadit ＞ ca, crudēlem ＞ cruzęl cruęl crudęl, fīdat ＞ fia, ✱gaudo ＞ gau, hŏrrĭda ＞ ǫreza, hŏrrĭdum ＞ ǫre, laudo ＞ lau, alauda ＞ lauzẹta laudẹta, rīdat ＞ ria, tradĕre ＞ trazir trair tradir, vidēre ＞ vezẹr vẹr vedẹr, vĭdet ＞ vẹ. When ð became contiguous to a following consonant (except final s), it changed to i: divīdĕre ＞ divire,[46] traditōrem ＞ traidọr (which, influenced by traïr, was pronounced traïdọr).[47]

1. Crey, beside cre ＜ crēdo, follows dei ＜ dēbeo, vei ＜ vĭdeo. Mercey, beside merce ＜ mercēdem, shows the influence of grey ＜ grĕgem, lei ＜ lēgem, and perhaps French fei ＜ fĭdem. Cf. β, (3).

2. Grau = gra ＜ gradum, niu = ni ＜ nīdum belong to the Catalan dialect, in which ð fell before the 8th century: gradum ＞ graðu ＞ gra-u ＞ grau, the u being preserved through combining into a diphthong with the a.

3. In purely learned words, d remained: odi ＜ ŏdium.

F is very rare (cf. § 59). The few examples appear to show that f (presumably in the 6th century or earlier) became β, and then developed like any other β (see β): Stĕphănum ＞ Estęve, co(n)fortāre? ＞ ✱coβortar ＞ ✱coortar ＞ conortar (through the common use of the double forms, con-, co-), gryphum ＞ griu, raphănum ＞ rave rafe, ✱refusāre ＞ rehusar refusar, ✱prefŭndum (= pro-) ＞ preọn. Nevertheless, cofịn ＜ cŏphĭnum, defǫrs ＜ de fŏris, grifọ, profięg ＜ profĕctum, rafe, would seem to indicate that in some words, possibly less popular at the outset, f was retained.

1. In purely learned words, f was kept: antifona, Caifas, philozophia.

G, representing original c and g, had a varied development.[48] For the fall of g in some words in Vulgar Latin, see § 55, G.

(1) Before a, g remained in the greater part of the territory, but in the north and east it early became y; and this y was generally retained in the eastern dialects (often fusing with a preceding i), while in most of the northern it developed into dž (cf. Y): amīca ＞ amiga amiia amia, dīcam ＞ diga dia, mīca ＞ miga miia mia,[49] pacāre ＞ pagar paiar, precāre ＞ pregar preiar; castigāre ＞ castigar castiar, legālem ＞ leial leyal lial, ligāmen ＞ liam, lĭgātum ＞ legat liat, plaga ＞ plaga plaia, regālem ＞ reial, rūga ＞ rüa.[50]

(2) Before o and u (ü), g was preserved, except in a few words which (doubtless in Vulgar Latin times) lost it either in all or in many dialects: acūtum ＞ agüt, secŭndum ＞ segọn, secūrus ＞ segürs; ✱a(u)gūrium ＞ agür aür, a(u)gŭstum ＞ agọst[51] ahọst, proclitic ĕgo ＞ ęu, ✱fagŏttum ＞ fagǫt, figūra ＞ figüra, Hugōnem ＞ Ugọ. For a g that becomes final or contiguous to final s, see § 63, (6): amīcus ＞ amics amis, Auriācum ＞ Auriac, cŏcum (= cŏquum) ＞ cǫc, jŏcus ＞ iǫcs, Ludovīcum ＞ Lozoic Lozoi, prĕco ＞ pręc; castīgo ＞ chastic chasti.

(3) Between the last two vowels of a proparoxytone, g, early in the Provençal period, became y, which developed into dž before the literary epoch; cf. § 49, (4): clĕrĭcum ＞ clęrge, ✱coratĭcum ＞ coratge, domĭnĭcum ＞ dimẹnge, manĭca ＞ mania, mĕdĭcum ＞ męge, mŏnăchum ＞ monge, ✱paratĭcum ＞ paratge, viatĭcum ＞ viatie. In some dialects, however, the vowel of the penult, after liquids and nasals, fell too early for the g to become y: clęrgue,[52] esta(t)ga, mętgue, mongue.

1. Amiu, chastiu belong to the dialect of Forez; so perhaps fau ＜ fagum, preu ＜ prĕco. These forms indicate a very early fall of the g in the dialect to which they belong. Cf. § 51, 3.

2. In purely learned words, Latin g remains unchanged: paganōrum ＞ paganor.

G´ became y during the Vulgar Latin period (§ 55, G). See Y.

1. In purely learned words the letter g was retained, but it was doubtless pronounced dž: astrologia.

L remained: colōrem ＞ colọr, male ＞ mal, ✱volēre (= velle) ＞ volẹr. Before final s, l became u in most dialects, in some as early as the 10th century: malos ＞ maus, talis ＞ taus; l was written, however, long after l had been vocalized. Under the influence of forms in which -ls ＞ -us, final l became u in the southwest and in some other regions: Aprīlem ＞ abriu. Cf. § 74, (2). Au ＜ -al is common in William of Poitiers.

1. For Gascon l ＞ r, see § 10.

2. Orifan, beside olifan ＜ elephantem, is probably French.

L´ will be considered, as ly, under Groups, § 73, Ly.

M remained: amāre ＞ amar, hŏmo ＞ om, timōrem ＞ temọr.

1. Occasionally -am rhymes with -an (afan: fam, portam: avan); this seems to show an indistinct pronunciation of the final nasal in some dialects. Cf. aven = avem ＜ habēmu(s) in the Nobla Leyczon. Cf. § 167, 2.

N remained: bŏnas ＞ bonas, donāre ＞ donar, lūna ＞ lüna. For n final or contiguous to final s, see § 63, (5): fīnis ＞ fis fins, panem ＞ pa pan.

1. In canorgue, dimergue, morgue, etc., beside canonge, dimenge, monge, etc., the r may be explained partly by dissimilation, partly by the analogy of clergue and of words with double forms (§ 87).

2. Menhs meins, beside regular mens ＜ mĭnus, show the influence of the alternative forms genhs geins and gens from ingĕnium (see § 73, Ny).

3. Iassey (= iasse, the latter part of which may be from exín = exĭnde), tey (= te ＜ tĕnet), used by Marcabru, are doubtless due either to a mistaken imitation of conventional borderland forms (see § 25, 3) or to the analogy of crei = cre ＜ crēdo (crei itself being due to the analogy of dei ＜ dēbeo, vei ＜ vĭdeo).

N´ will be considered, as ny, under Groups, § 73, Ny.

P, from the 4th to the 6th century, was voiced to b: capĭllum ＞ cabẹl, rīpa ＞ riba, ✱sapēre (= sapĕre) ＞ sabẹr, trepalium ＞ trebalh; capit ＞ cap (§ 63), sapis ＞ saps.

1. In some borderland dialects p ＞ v, as in French: saver. Evescat, evesque, beside bisbat, bisbe, are French.

2. Apud, used as a proclitic, became for some reason in Vulgar Latin ✱apu, which developed regularly into ✱abu and, after the fall of intertonic vowels, ab. This ab assimilated its b more or less to a following consonant, becoming ap before voiceless consonants, am before nasals; am, used before dentals, became an: hence we have four forms, ab, ap, am, an. Amb perhaps developed first from am before l, as in am l’autre; when used before a consonant with which mb did not readily combine, it expanded into ambe. See Elise Richter, Zs., XXVI, 532; J. Huber, Zs., XXX, 583.

3. In some dialects, apparently, p was not voiced after au: sapuĕrunt ✱sapwĕrunt ✱saupĕrunt ＞ saubron saupron. Cf. § 65, C, 1.

4. In purely learned words, p remains: epifania.

R remained: amāra ＞ amara, durāre ＞ dürar, ĕrat ＞ ęra. Final rs was reduced to s, in most dialects, during and after the literary period: priōres ＞ priọrs priọs (Girart); the reduction apparently began in Limousin as early as the 12th century (Bertran de Born rhymes iọs and flọrs).

1. Final r began to fall in many dialects in the 14th century. At present it has disappeared all through the south and west: amōrem ＞ amou, flōrem ＞ flou.

2. In some dialects (especially those of Gard and Hérault) intervocalic r and z were confused, probably during the literary period: gyrāre ＞ girar gisar; conversely audīre ＞ auzir aurir. Cf. Revue des langues romanes, XL, 49, 121.

S was voiced to z, probably from the 4th to the 6th century: pausa ＞ pausa, presĕntem ＞ presen; rīsum ＞ ris (§ 63).

1. An s that became contiguous to n was changed, in a few dialects, to r: almosna almorna, disnar dirnar. In modern Limousin and some of the dialects of Dauphiné, Languedoc, and Gascony, s has disappeared before nasals: asne ane, caresma carema, disnar dinar (so blasmar blamar, desma dema); the fall began during the literary period. S before a consonant in many of the modern dialects, and final s in some, has become i: asne aine, caresma careima (so perhaps desma deima, pruesme prueime); some traces of this change occur in texts of the literary period. Cf. Zs., XXIII, 413. Isla, in Limousin, became ilha (perhaps through iyla): cf. Zs., XXIII, 414. Cf. § 78.

2. In some southeastern dialects intervocalic z after au has changed to v: causa cauva (so auzir auvir); possibly the auvent of the Boeci, v. 23, is to be connected with this.

T, from the 4th to the 6th century, was voiced to d: amāta ＞ amada, natālis ＞ nadals, servitōrem ＞ servidọr; habētis ＞ avętz avęs avęt (§§ 63, 64), latus ＞ latz las, natum ＞ nat. For a t which became contiguous to r (amātor ＞ amaire), see § 52, (1), and § 70, Tr.

1. In some dialects of the south and southeast, final t fell shortly after the literary period: amātum ＞ amat ama.—Appoestat is French.

2. Tōtus, in Gaul, became tōttus as early as the 4th century: hence Pr. tota totas. For meteis ＜ met-ĭpse see § 131, (2).

3. Espaza (beside espada) ＜ spatha, was perhaps influenced in its pronunciation by the spelling of the Latin word.[53] Ez, coming from et before a vowel, shows the influence of az (＜ ad + vowel) and quez (＜ quĭd + vowel). Grazal, ‘grail’, is perhaps a cross between ✱cratella ＜ crater and gradale, ‘service-book’; so grazalet. Grazir grazire (cf. agradar) is perhaps altered from an earlier ✱grazar ＜ ✱gratiare. Mezeis ＜ met-ĭpse, mezesmes, meesmes (beside medesmes) ＜ ✱met-ĭpsĭmus have been subjected to the analogy of ez ＜ et and quez ＜ quĭd or of ĭd ĭpsum: § 131, (2).

4. In some dialects tī became a sound written h: peccatī ＞ peccah. See § 51, 1.

5. In purely learned words, t remains: eternal.

W will be considered, as gw, under Groups, § 72, βw.

X is a symbol for ks: see Groups, § 79, Ks.

Y, representing Latin dy, g´, gy, j, and z (cf. § 55, G, Y; § 57, Z) had a varied development.

(1) When it became contiguous to a following consonant (§§ 45, 49) it changed to i: adjutāre ＞ ayudāre ＞ ay’dar ＞ aidar, medietātem ＞ meitat; cōgitāre ＞ cüidar, frīgĕre ＞ frire, lĕgĕre ＞ lęyre, propagĭnem ＞ probaina, rĭgĭda ＞ rẹida, ✱tragĕre ＞ traire; bajŭlus ＞ bailes.

(2) When it remained intervocalic, it became dž in most of the territory, but in the northeast and parts of the north it was not changed:[54] audiam ＞ auia, in-ŏdiare ＞ enoiar, invĭdia ＞ envẹia, invĭdiōsus ＞ enveiọs enveyọs, ✱gladia ＞ glaya, mediānum ＞ meian, ✱pŏdiāre ＞ poiar, radiāre ＞ raiar, sordĭdior ＞ sordẹier, vĭdeat ＞ vẹia; fragĭlem ＞ fragel; ✱exagiāre ＞ assatiar essaiar essayar, corrĭgia ＞ corrẹia corrẹya, fagea ＞ faia faya, regiōnem ＞ reiọ; dīe Jŏvís ＞ diiǫus, major ＞ maier, pĕjor ＞ pięier, pejōrem ＞ peiọr, trŏja ＞ trǫia; baptizāre ＞ bateiar. For a dž or a y that became final or contiguous to final s, see § 63, (1), (2): audio ＞ auch, in ŏdio ＞ enuęg (plural enuętz enuęg) enǫi, gaudium ＞ gauch, gladium ＞ glai, mĕdium ＞ męg męi, hŏdie ＞ ǫi, pŏdium ＞ puęg pǫi, radium ＞ rai; fŭgit ＞ füg füi, grĕgem ＞ gręy, lēgem ＞ lẹg (pl. lẹitz) lẹi, lĕgit ＞ lięg, magis mais,[55] rēgem ＞ rẹi, exagium ＞ essai; pĕjus ＞ pięis.

(3) Before accented e or i, y disappeared (doubtless in Vulgar Latin: § 55, G), except in some western dialects, where it became dž: vagīna ＞ guaīna, ✱legīre (= lĕgĕre) ＞ legir,[56] magĭster ＞ maẹstre maiẹstre magẹstre, ✱pagē(n)sis ＞ paẹs pagẹs, regīna ＞ reïna, sagĭtta ＞ saẹta saiẹta sagẹta.

1. Detz ditz ＞ dĭgĭtus are irregular and unexplained. The word is irregular in some other Romance languages, notably in Italian. Cf. Gröber’s Grundriss, I, p. 507.

2. Glavi, beside glai (and learned glazi) ＜ gladium, is supposed by some to show the influence of Celtic cládibo. Cf. Körting; also H. Schuchardt, Zs., XXV, 345.

3. Messér seems to be a contraction (due to proclitic use) of ✱messeyer = mes, ‘my’, + ✱seyer? ＜ ✱sĕyor = sĕnior (cf. A. Lindström, L’analogie dans la déclinaison des substantifs latins en Gaule, 1897-8, pp. 292-3).

4. In purely learned words, di, g, gi, z are retained, the g being pronounced presumably as dž, the z as z: odi, fragil, regio, canonizar.

Medial Groups.

66. Medial groups may be conveniently classified as follows:—A. Double Consonants (1); B. Groups of Dissimilar Consonants: groups ending in l (2), groups ending in r (3), groups ending in w (4), groups ending in y (5), groups beginning with l, m, n, r, or s and not ending in l, r, w, or y (6), all other groups (7). This order will be followed.[57]

1. It should be noted that the prefixes ad-, sub- regularly assimilate their d or b to the following consonant: ✱ad-rīpāre ＞ arribar, sub-venīre ＞ sovenir. Sosrire, sosterrar, sostraire show a substitution of prefix, due, no doubt, to the analogy of sospirar, sostener.

1. Double Consonants.

67. In general, the double consonants became single, in the 9th or 10th century (perhaps earlier before the accent), but underwent no other change save those described in §§ 63, 64: abbātem ＞ abat, sĭccum ＞ sẹc, rĕddo ＞ ręt, affībulāre ＞ afiblar, aggregāre ＞ agregar, flamma ＞ flama, pĭnna ＞ pẹna, cappa ＞ capa, passum ＞ pas,[58] mŭttum ＞ mọt, advenīre ✱avvenīre ＞ avenir.

(1) Cc before a, in the east and northeast, became tš; elsewhere, c; bŭcca ＞ bọca bọcha, vacca ＞ vaca vacha.

(2) Ll, in some southern dialects, became l´; elsewhere, l: capĭllum ＞ cabẹl cabẹlh, grȳllum ＞ gril grilh, mantĕllum ＞ mantęl mantęlh, villānus ＞ vilas vilhas. It is possible, however, that -llī regularly became l´ in Limousin, while ll before other vowels was not palatalized: caballum ＞ caval, caballī ＞ cavalh; ĭllī ＞ ilh, ĭllōs ＞ ẹls; this would account in part for the frequent occurrence of lh in the poems. So in some dialects -nnī ＞ n´: annī ＞ anh. Cf. § 51, 1. For final ls and l, see § 65, L: ĭllos ＞ ẹls ẹus, vallem ＞ val vau. For Gascon l ＞ r, see § 10: appĕllat ＞ apęla apęra.

(3) Rr, when intervocalic, seems generally to have been distinguished from r during the literary period and later: cŭrrĕre ＞ cọrre, ✱corrŭptiāre ＞ corrossar, errāre ＞ errar, tĕrra ＞ tęrra. Occasionally, however, rr is found in rhyme with r.

2. Groups Ending in L.

68. The groups of two consonants will be treated in alphabetical order. It will be seen that bl, rl, sl remained unchanged; ml developed a glide consonant between its two members; pl, tl, βl and yl respectively voiced, assimilated, and vocalized their first element; while cl, gl were fused into l´. For an explanation of this last phenomenon, see § 79.

Bl ＞ bl: nĕbŭla ＞ nębla, ✱oblītāre ＞ oblidar, sabulōnem ＞ sablọn.[59]

βl ＞ ul: fabŭla ✱faβla ＞ faula, sibilāre ✱siβlāre ＞ siular, tabŭla ✱taβla ＞ taula.

Cl ＞ l´: genŭcŭlum ＞ genọlh, ŏcŭlum ＞ ǫlh, sĭtŭla ✱sĭcla[60] ＞ sẹlha, vĕtŭla vĕcla[60] ＞ vęlha. In learned words we find gl, cl: ✱e(c)clĕsia ＞ glęiza, joculārem ＞ ioglar, sæcŭlum ＞ sęgle sęcle.

C´l ＞ il: gracĭlem ＞ graile.

Dl ＞ dl, which during the literary period became ll and then l: mŏdŭlum ＞ ✱mǫdle mǫlle.

Gl ＞ l´: rēgŭla ＞ rẹlha, ✱strĭgŭla ＞ estrẹlha, vĭg(ĭ)lat ＞ vẹlha. Lẹula ＜ lĕgŭla, tẹula ＜ tēgŭla are irregular: cf. Archivio glottologico italiano, XIII, 439, 459.

Ml ＞ mbl: sĭmĭlāre ＞ semblar, trĕmŭlat ＞ trẹmbla. In sembrar we find an r ＜ l due perhaps to the analogy of membrar ＜ memorāre. In semel gives ensẹmble essẹms.

Pl ＞ bl: cōpŭla ＞ cọbla, dŭplum ＞ dọble. Learned words have pl: duplicar.

Rl ＞ rl: Carŏlus ＞ Carles, hōrolŏgium ＞ orlǫi, ✱paraulare ＞ parlar.

Sl ＞ sl: ī(n)sŭla ＞ isla. For ilha, see § 65, S, 1.

Tl ＞ tl, which during the literary period became ll and then l: rŏtŭlum ＞ rǫtle rǫlle, spathūla ＞ espatla espalla espala. In really popular words tl had become cl in Vulgar Latin.[61]

Yl ＞ il: bajŭlus ＞ bailes. Cf. § 65, Y, (1).

69. A group of three consonants remained unchanged, except that double consonants became single: ambulāre ＞ amblar, implēre ＞ emplir, avŭncŭlus ＞ avọncles, cĭrcŭlus ＞ cẹrcles, mascŭlus ＞ mascles, ✱afflammāre ＞ aflamar, inflāre ＞ enflar, ŭngŭla ＞ ọngla, emplastrum ＞ emplastre.

1. Selcle, beside cercle, seems to show an assimilation of the r to the l of the next syllable. Empastre, beside emplastre, has been influenced by pasta. Emblar is probably from V. L. ✱imbolare = involare.

3. Groups Ending in R.

70. The groups of two consonants will be treated in alphabetical order. It will be seen that br, gr, lr, nr generally remained unchanged; mr, sr, zr (and sometimes lr, nr) developed a glide consonant; cr, pr voiced, and βr, c´r, dr, tr, yr vocalized their first element.

Br ＞ br, also ur, occasionally ir (through ür): fabrum ＞ fabre faure[62], fĕbrem ＞ fębre fęure, lībra ＞ liura; Octōbrem ＞ ochọyre. For brg, see § 80.

βr ＞ ur, occasionally ir: bĭbĕre ＞ bęure, débēr’ hábḙo ＞ deβr’áyo ＞ deurái, ✱rōbŏrem ＞ rọure rọire; ✱mŏvĕre (= mŏvēre) ＞ mǫure, plŏvĕre ＞ plǫure.

Cr ＞ gr: acrem ＞ agre, lacrĭma ＞ lagrema, lŭcrum ＞ lọgre, macrum ＞ magre, sacrāre ＞ sagrar, sŏcrum ＞ sǫgre.[63] In late learned words we find cr: secret.

C´r ＞ ir: cŏcĕre ＞ cǫire, dīcĕre ＞ dire, dūcere ＞ düire, facĕre ＞ faire. Cf. § 49, (1).

Dr ＞ ðr ＞ ir: divīdĕre ＞ devire, quadrum ＞ caire, vídēr’ hábe̯o ＞ veð’r-áyo ＞ veirái. After au, apparently, ð simply disappeared: claudĕre ＞ claure. Late learned words have dr: quadrupedi.

Gr, in popular words, was reduced to r in Vulgar Latin in parts of the Empire fra(g)rāre ＞ ✱frarar ✱flarar flazar, intĕ(g)rum ＞ entęr entięr, nĭ(g)rum ＞ nęr nięr,[64] pere(g)rīnum ＞ peleri, pĭ(g)rĭtia ＞ perẹza; these forms occur in Gascony, Rouergue, and Limousin, but forms with gr are found in the same region. Elsewhere, in these same words, and everywhere, in more bookish words, gr remained in Vulgar Latin; this gr was kept in most of the Provençal territory, but was changed to ir in Dauphiné, Auvergne, and Languedoc: fragrāre ＞ flairar, integrāre ＞ enteirar, intĕ́grum ＞ entęgre entęir,[65] nĭgrēscĕre ＞ negrezir, nĭgrum ＞ nẹgre nẹir,[65] peregrīnum ＞ pelegri, pĭgrĭtia ＞ pigręza. Purely learned words have gr everywhere: agricultura.

G´r: see Yr.

Lr usually remained unaltered, but in some dialects became ldr[66]: vălēr’ hábe̯o ＞ valrai valdrai, ✱vólēr’ hábe̯o ＞ volrai voldrai.

Mr ＞ mbr: camĕra ＞ cambra, memorāre ＞ membrar, nŭmĕrum ＞ nọmbre.

Nr usually remained unaltered, but in some dialects became ndr[66] cĭnĕrem ＞ cẹnre cẹndre, dīe Vĕnĕris ＞ divenres divendres, in-gĕnerāre ＞ engenrar, Henrīcum ＞ Enric, expōnĕre ＞ espọnre espọndre, ✱gĕnĕrem (= gĕnus) ＞ genre, gĕnĕrum ＞ genre gendre, honorāre ＞ onrar ondrar, tenēr’ habeo ＞ tenrai tendrai, venīr’ habeo ＞ venrai vendrai.

Pr ＞ br: capra ＞ cabra, cŭperāre ＞ cobrar, erĭpĕre ＞ erẹbre, ŏpĕra ＞ ǫbra, pauper ＞ paubre, ✱pĭperāta ＞ pebrada, recĭpĕre ＞ recẹbre, sŭperāre ＞ sobrar, sŭper ＞ sọbre. Purely learned words have pr: caprin. It is uncertain whether paupre (beside the usual paubre) is a Latinism or represents some dialect in which au prevented voicing.

Sr ＞ str: ✱ĕssĕre (= ĕsse) ＞ ęstre. For ęsser, see § 49, (2).

Tr ＞ dr ＞ ðr ＞ ir: amātor ＞ amaire, fratrem ＞ fraire, latro ＞ laire, mater ＞ maire, ŭtĕrem ＞ ọire, pĕtra ＞ pęira, Pĕtrus ＞ Pęires, petrōnem ＞ peirọ, petrōsus ＞ peirọs, pre(s)by̆tĕrum (Einf., § 140) ＞ prevẹire, servītor ＞ servire, vĭtrum ＞ vẹire. Learned words have dr and tr: ✱poenĭtĕre ＞ penedre (penedir), impetrāre ＞ impetrar.

Yr ＞ ir: frīgĕre ＞ frire, lĕgĕre ＞ lęyre.

Zr ＞ zdr: mīsĕrunt + ✱mĭssĕrunt ＞ ✱mẹzron mẹsdron.

1. Redebre (beside rezemér) ＜ redĭmere has apparently been influenced by recebre. The Burgundian sor for sobre comes from the prefix sŭr- (sŭr-rīdēre, etc.). Perri ＜ ✱pētrīnum is probably French.