An Outline of the Phonology and Morphology of Old Provençal

105. Most adverbs of manner were formed by adding -men (-ment, -mens, or -menz) to the feminine singular of the adjective: belamen. These adverbs were originally ablative phrases: serēna mĕnte, etc. In Provençal the specific meaning of the -men was forgotten, but the two parts might still be separated by an intervening word: ẹpsa… ment. When two adverbs in -men were used together, the ending was generally affixed to only one, oftener the first. Bona and mala could be used as adverbs without the suffix.

Comparison.

106. Adjectives and adverbs regularly formed their comparative by prefixing plüs to the positive, and their superlative by prefixing the definite article to the comparative: cara, plüs cara, la plüs cara. This method of comparison goes back to Vulgar Latin times.

107. Some adjectives preserved their old comparative in -ior. These comparatives had an inflection similar to that of sęnher: cf. § 101, (3) and § 101, (3), 1.

108. The following neuter comparatives were used as adverbs: gensẹis gensẹs gensẹtz (＜ génser influenced by longẹis, sordẹis); longẹis longẹitz ＜?✱longĭtius (＜ longĭter + lŏngius); mais ＜ magis; męlhs ＜ mĕlius; mẹns ＜ mĭnus; pęitz ＜ pĕjus; sordẹis ＜ sordĭdius; viatz ＜ vivacius. Mais, męlhs, mẹns, pęitz were used also as neuter pronouns. Viatz lost its comparative sense.

109. A few adjectives, most or all of them learned, preserved the old superlative form with an intensive sense: altisme, carisme, pęsme, prǫsme, santisme.

Numerals.

110. The cardinal numerals are:—

111. The first two numbers were inflected as follows:—

Düi dọi are from Vulgar Latin dŭī = dŭo; dọs is from dŭos, dọas from dŭas. Trẹs has a form trẹi (originally nom. m.), patterned after düi, and a form trẹis, which seems to be a cross between trẹs and trẹi. For the dialect forms of ọnze—sętze, see § 76, (1), Ndc´, and § 80, Dc´. Cen, multiplied by another number, took a plural form when used substantively; when used adjectively, it generally did not, but we find dozentas with a feminine noun. Mil had four plurals, milia miria melia mila; millięrs is a noun.

112. The ordinal numerals had separate forms for the two genders; the masculine forms followed the second declension type, the feminine forms, the first declension. After 5th, they were made by adding to the cardinal numeral the originally distributive ending -ēnus -ēna.

113. Beside primięr we find premięr prümier promięr (§ 44, 1, 3), and also prim and primeiran; for the developments of the ending -ięr, see § 23, 1 and § 73, Ry, 1. Tęrz, tęrza regularly became tęrs, tęrsa (§ 73, Rty). Such forms as secọnda, tęrcia, sęxta, octava, nọna, dęcima are learned.

PRONOUNS.

114. Under this head will be treated not only pronouns and pronominal adjectives, but also articles.

115. In popular Latin the personal, possessive, and demonstrative pronouns and adjectives had two sets of forms, according as they were accented or unaccented (§ 19). Ille, when stressed and used pronominally, became a disjunctive personal pronoun of the third person; when unstressed and used pronominally, it furnished the conjunctive forms of the third person; when unstressed and used adjectively, it developed into a definite article. Ipse had similar uses. These differentiations must have begun in Vulgar Latin times.

116. The declension of ĭlle was considerably altered in Vulgar Latin. The neuter ĭllud disappeared, being replaced by ĭlium. Through the influence of quī cūjus cūī, ✱ĭllī illūjus[89] illūī[89] came to be used beside ĭlle illīus ĭllī. The feminine had, beside illīus ĭllī, a genitive and dative ĭllæ; through the analogy of illūjus illūī, ĭllæ was expanded into illæjus[89] illæi.[89] Illīus then went out of use. In the plural, illōrum (which in some regions, by the analogy of illūjus illūī, had a form ✱illūrum) crowded out illārum; this illōrum came to be used also as a dative.

Ipse and ĭste followed in the main the same course as ĭlle.

Articles.

117. The indefinite article comes from ūnus, which seems to have been occasionally so used even in classic Latin:—

118. (1) The definite article comes from unaccented ĭlle, which, being used as a proclitic, regularly lost its first syllable (§ 19). Ille (✱ĭllī), ĭllum, ĭllī, ĭllos, ĭlla, ĭllas became respectively le (li), lo, li or lhi,[90] los, la, las. Le, lo, li, lhi, la frequently elided their vowel before another vowel (l’an, l’arma), becoming l or lh. Furthermore, le, lo, li, lhi, los, in the intertonic position after a vowel (vé lo páire), regularly lost their vowel (vẹl páire)[91]; and, by analogy, la and las were sometimes reduced to l and ls. We have, then, beside the full forms, the proclitics l, lh, and the enclitics l, lh, ls. Inasmuch as l might be vocalized before a dental,[92] the enclitics l and ls sometimes became u and us (a͡u portęr, e͡uz dias antix).

(2) The particles e and que, with the enclitic l, formed combinations ẹl and quẹl. Quẹl, being understood as qu’el, gave rise to a form ẹl.

(3) In the f. nom. sg. there is a form li or lhi, which is hard to explain. The most likely theory is that when the masculine quī took the place of the feminine quæ (see § 133), the masculine ✱ĭllī came to be used beside ĭlla,[93] for the feminine. The Provençal feminine li (lhi) which resulted was strongly supported by the analogy of a feminine possessive mi, beside ma (see § 127).[94]

(4) The regular forms are, therefore, the following:—

In many texts the objective forms lo, los, ls are used in the nominative.

119. In some southwestern and some southeastern dialects we find forms sọ, sọs, sa, sas, coming from ĭpse.

Personal Pronouns.[95]

120. In Vulgar Latin ĕgo lost its g (§ 55, G). The dative, mĭhi, was preserved only in its contracted form, mī. After the pattern of mī, ✱tī and ✱sī were created for the other persons.

121. Provençal has no nominative forms that are regularly unaccented. In the conjunctive forms of the third person (not reflexive), the direct object is distinguished from the indirect; elsewhere there is no such distinction.

122. Latin mē ＞ mẹ, mī ＞ mi, nōs ＞ nọs; tē ＞ tẹ, ✱tī ＞ ti, vōs ＞ vọs; sē ＞ sẹ, ✱sī ＞ si. Me mi, te ti, se si, used as proclitics before a vowel, or as enclitics after a vowel, were reduced to m, t, s: m’ama, t’apela, s’es; o͡m, be͡t, cosi͡s. Nọs and vọs, used as enclitics after a vowel, became respectively ns and us; que͡ns, no͡us; sī vōs ＞ sius, later sieus (§ 32). The forms (all objective) for the first and second persons and for the third person reflexive are, then:—

123. The conjunctive forms of the third person (not reflexive) come in the main from the proclitic ĭlle: ĭllī, ĭllum, illōrum (✱illūrum), ĭllos, ĭlla, ĭllas became respectively li or lhi, lo, lọr (lür), lọs, la, las. When used proclitically or enclitically, under the conditions described in § 118, (1), li (lhi), lo, los were reduced to l (lh), l, ls; and l was sometimes vocalized. O ＜ hŏc was employed also, meaning ‘it.’ The adverb ĭnde became ẹnt ẹn n (and, through the analogy of me m, te t, se s, also ne), which was often used as a pronoun with the sense ‘of it’, ‘of them’, sometimes ‘of him’, ‘of her’; nọs ẹn ＞ nọn, vọs ẹn ＞ vọn. The adverb hīc became i, meaning ‘here’ or ‘there,’ which served also as a dative pronoun, ‘to it,’ ‘to them’; it was then always an enclitic, forming a diphthong with a preceding vowel; it regularly took the place of li in the constructions lọi = lo li, lai = la li. The forms are:—

Disjunctive Forms.

124. Vulgar Latin ✱ĕo or ✱ĕu ＞ ęu ięu (§ 30), which before an enclitic became ę ię (ę͡l, ię͡n). The other forms explain themselves. The nominative tü, from the beginning of the 13th century, was sometimes used for tẹ after prepositions; this use may have been suggested by the existence of lü = ‘him’; § 125, (1). Nọs + ẹn ＞ nọn, vọs + ẹn ＞ vọn.

125. The disjunctive pronouns of the third person (not reflexive) come from accented ĭlle, with the exception of ǫ from hŏc. Illūi, illōrum, illæjus, illæi lost their first syllable, perhaps through elision after a vowel; illūjus disappeared. Ille, ✱ĭllī gave ẹl ẹlh, il ilh; ẹl sometimes vocalized its l. Illūī became lüi, in some dialects reduced to lü. Illum became ẹl ẹlh. Illōrum (✱illūrum) gave lọr (lür). Illos became ẹls (often ẹus) ẹlhs. Illa, ĭllam both gave ẹla ẹlha. Illæjus became lęis lięis (in some dialects reduced to lięs).[96] Illæi gave lęi (dialectically lę) lięi. Illas became ẹlas ẹlhas.

(2) In the feminine singular nominative there is, beside ẹla ẹlha, a form ilh il. This is probably to be explained, like the feminine article lhi li, as coming from the masculine nominative ✱ĭllī introduced into the feminine, and supported by the feminine possessive mi: see § 118, (3).

(3) Some dialects preserve the final -i of ẹli (m. pl. nom.) and ilhi ili (f. sg. nom.): see § 51, 1.

(4) Occasionally the conjunctive li (f. sg. obj.) and lo (neuter sg. nom.) were used as disjunctive forms. And sometimes the masculine lüi lü was used for the feminine.

(5) The forms are, therefore, the following:—

In many texts the objective forms ẹls ẹlhs, lęis lęi are used in the nominative. We then find occasionally a new objective, ẹlses.

Possessives.

126. Beside mĕus mĕa, tŭus tŭa, sŭas sŭa, there existed in popular Latin the shorter forms ✱mĕs? ✱ma, ✱tŭs ✱ta, sŭs sa. Of the two forms vĕster and vŏster, only the latter was used. To supply the lack of a third person possessive denoting a plural possessor, illōrum came to be employed as a possessive.

127. (1) The primarily atonic possessives come from the shorter Latin forms. The original masculine singular forms of the first person were displaced by mọs mọ, made on the analogy of tọs tọ, sọs sọ, which come regularly from ✱tŭs ✱tŭm, sŭs, sŭm; so in the objective plural we find mọs, corresponding to tọs ＜ ✱tōs, sọs ＜ sōs. Mĕī, tŭī, sŭī gave męi, tọi tüi, sọi süi (§ 34), which, however, were often replaced by the objective forms. ✱Ma ✱mam ✱mas, ✱ta ✱tam ✱tas, sa sam sas became ma mas, ta tas, sa sas; ma, ta, sa often elided their a before a vowel. The formation of midons has been explained in § 91, (2); § 93, (2); § 118, (3): from it came a feminine singular possessive mi, and, by analogy, ti and si.

(2) The forms are:—

They are generally used only adjectively, and without the definite article. In some early texts, however, tọs and sọs, preceded by the article, are used substantively.

128. (1) The primarily tonic possessives come from the longer Latin forms. Mĕus mĕum mĕi mĕos ＞ męus męu męi męus, which regularly became mięus etc. (§ 30); an analogical form mięu is found beside mięi. In the feminine of the first person we have, instead of ✱męa, mięua and mia: the first of these two forms is evidently made up from the masculine; the second may be due partly to the analogy of mi, partly to a proclitic use of the word (§ 44, 4).[97] In the second and third persons the masculine forms are mainly, and the feminine forms partially, replaced by analogical formations based on the possessive of the first person; tŭi, sŭi, tŭa, sŭa, however, give regularly tọi tüi, sọi süi, tọa tua, sọa sua (§ 8).

(2) The forms follow, those of the third person (which correspond exactly to those of the second) being omitted:—

They may be used adjectively or substantively, with or without the definite article.

Plural Possessor.

129. Nŏster, vŏster developed regularly after the pauper model: § 103, (1). The masculine singular nominative often took an -s: cf. § 96; § 101, (1). Some southeastern dialects preserved the -i of nǫstri: cf. § 51, 1. Beside vǫstra we occasionally meet vǫstri, due to the analogy of feminine mi, ti, si.

These forms are used adjectively or substantively, with or without the definite article.

Demonstratives.

130. Latin īdem went out of use. Latin ĭs was preserved only in the phrase ĭd ĭpsum (ad ĭd ĭpsum ＞ adęs), and in the combination ĕccum, in which it ceased to be recognized, so that ĕccu’ was regarded as a synonym of ĕcce.

131. (1) The demonstrative particles ĕcce and ĕccu’ were often prefixed to pronouns in Vulgar Latin. Being thus proclitically used, they frequently lost their first syllable (§ 19); sometimes, however, under the influence of ac (as in ac sīc ＞ aissi), they preserved it, assuming the vowel of ac: ecce ĭlla ＞ aicẹla, eccu’ ĭsta ＞ aquẹsta; cf. § 43, (2).

(2) The suffix -met was used in Vulgar Latin as an intensive prefix. Its change of place was probably due to such phrases as sēmet ĭpsum, understood as sē metĭpsum. The -t, before a vowel, regularly gives -d- (met-ĭpsum ＞ mẹdes); but we find, besides, -z- (＜ Lat. d), introduced perhaps through the analogy of ĭd in ĭd ĭpsum (✱medĭpsum ＞ mezẹis); and also -t- (＜ Lat. tt), which may be the result of a combination of met- and ĭd- (met-ĭd-ĭpsum ＞ ✱metdĭpsu ＞ metẹis).

132. The pronouns preserved, either in their simple form or combined with a prefix, are the following:—

(2) Ipse appears as ẹps ẹpsa, ẹus ẹussa, ẹis ẹissa (with a m. pl. ẹisses and a neuter ẹis); the last forms are the commonest; for the development of the ps, see § 79 and § 80, Ps. Met-ĭpse gives (medips) medẹs, (metẹish) metẹis, and, more commonly, mezẹis (f. mezẹissa, neuter mezẹis); see § 131, (2). ✱Met-ĭpsĭmus becomes medẹsme-s, mesẹsme-s, meẹsme-s (§ 65, D), with feminine forms in -a. Unaccented ĭpsum is probably one source of the neuter sǫ: cf. § 132, (1). For the article (so, sa), see § 119.

(3) Ille, uncombined, developed into an article (§ 118) and a personal pronoun (§§ 123, 125), but went out of use as a demonstrative. Combined with ecce and eccu’ it gave: aicẹl aissẹl, cẹl, sẹl; aquẹl. Echẹl (pronounced ekẹl?) seems to come from eccu’ ĭlle with its original initial vowel preserved. Ipse ĭlle perhaps gave rise also to a sẹl, which ultimately coincided with the form coming from ecce ĭlle. There is a neuter aicelo, perhaps aicẹl + ǫ. Cẹl will illustrate the inflection of all these words; the forms are to be explained like those of the disjunctive personal pronoun (§ 125):—

(4) Iste gave ẹst, ẹstz, ẹsta, ẹstas. Ecce ĭste became aicẹst (not common) and cẹst sẹst; eccu’ ĭste became aquẹst echẹst, and chẹst. Aquẹst will illustrate the inflection; the forms are to be explained like those of cẹl:—

Interrogatives and Relatives.

133. The interrogative and relative pronouns were confused and combined in Vulgar Latin, quī taking the place of quĭs, and quĭd gradually encroaching on quŏd. Furthermore, the masculine forms were used instead of the feminine, which disappeared. We have in Provençal no evidence of the survival of any other cases than the nominative, dative, and accusative singular and the nominative plural:—

The distinction between que ＜ quĕm, que quez ＜ quĭd, and que ＜ quæ could not be maintained; we have, then, simply three forms: a nom. sg. or pl. qui, a nom.-acc. sg. or pl. quẹ (quẹz), a dat. sg. or pl. cüi (sometimes written qui).

134. We have also qualis, which came to be inflected like fezẹls: see § 103, (2); the feminine singular, however, often dropped its -s, and sometimes took the ending -a (cal, cala). Quīnam apparently became quina, which, understood as a feminine form, developed a masculine, quin. There seems to have been also a ✱quīniam (cf. quŏniam?), which gave quinh, quinha. Cf. D. Behrens in the Zeitschrift für französische Sprache, XVII, ii, 67-8, footnote. The phrase de ŭnde became dọnt, dọn, which was often used with the meaning ‘of which’, ‘of whom’.

135. (1) In Provençal the interrogative pronouns are: qui, ‘who’ or ‘whom’; que quez, ‘what’; cüi, ‘to whom’ or ‘whom’, ‘to what’ or ‘what’ (obj.); cals (either alone or preceded by the definite article, inflected as in § 134), ‘which’; quin quinh, quina quinha, ‘which’. Cals is used also as an adjective.

(2) The relative pronouns are: qui, ‘one who’, indefinite (used also, in early texts and in southwestern Languedoc, as the regular relative pronoun for persons); que quez, ‘who’ or ‘whom’, ‘which’; cüi, ‘whom’, ‘which’ (generally used as indirect object of a verb, or after a preposition); lo cals (inflected as in § 134), ‘who’ (‘whom’), ‘which’; don dont, ‘of which’, ‘of whom’.

Indefinite Pronouns and Adjectives.

136. The following words call for special mention:—

(3) Alques alque ＜ alĭquĭd, used as an invariable neuter pronoun or adverb, ‘something’, ‘somewhat’. The -s form, which originally developed before a vowel, was preferred because of the analogy of other neuter pronouns and adverbs. The preservation of the e is due to association with quez que. Alque was sometimes used as an adjective.

(4) Als al au, used as an invariable neuter pronoun, ‘something else’. Al (au) may have been detached from alques, understood as al ques. Meyer-Lübke, however, takes it, as well as Old French el, from ✱alum = aliud: Gram., II, p. 649. Als owes its -s to the analogy of other neuter pronouns, such as alques, ẹis, mais, męlhs, mẹns, pęis, etc.

(5) Altre autre ＜ alter, ‘other’, pronoun and adjective. A dative ✱altrūī, following illūī, goes back to Vulgar Latin. The Provençal forms autrüs, autrü show the influence of alcüs and negüs; autri belongs to the southeastern dialects (cf. aquẹli, ẹli, nǫstri, tüti, etc.). Inflection:—

(6) Altretals autretals ＜ alter talis; by dissimilation, atretals: by substitution of ai- (first syllable of aissi ＜ ac sīc) for atre-, aitals; by fusion of aitals and atretals, aitretals; through analogy of atressi, atrestals. Cf. § 74, 2. Inflection like that of cals (§ 134).

(7) Altretan atretan aitan atrestan etc. ＜ alter + tantum: see altretals.

(8) Cada un ＜ κατά + ūnum,‘every one’. The Greek preposition κατά was introduced into the Latin territory, probably by Greek merchants, in stating prices: καθ᾽ἕνα = cata ūnum, κατά τρεῐς = cata trēs; hence cada ün, cada trẹi. Inflection: cada üs, cada ü(n); cada üna.

(9) Calacom qualacom qualaquom, ‘something’, ‘a little’, seems to be a Provençal compound of cal and acǫ (§ 132), the last syllable of which was perhaps understood as cọm cọ ＜ quōmŏ(do). Cf. quezacom below. There is a diminutive calacomet, which helped to maintain the m of calacom.

(10) Cals que quals que, cal que qual que, ‘whoever’, is a Provençal compound.

(11) Cant quant can quan ＜ quantum quanti, ‘how much’, ‘how many’. Cant, inflected like bęl (§ 103), is used also as an adjective and as a masculine and feminine pronoun.

(12) Cascüs chascüs, ‘everyone’, ‘every’, appears to be a fusion of cada üs and ✱cescüs ＜ ✱cisqu’ ūnus = quĭsque ūnus = ūnus quĭsque. Inflection: cascüs, cascü(n); cascüna.

(13) Ent en n ne ＜ ĭnde,‘some’: cf. § 123.

(14) Maint mant man manh ＜ Celtic ✱mantî, ‘many’, ‘many a’, ‘many a one’. Obj. pl. in -s, f. sg. in -a, f. pl. in -as.

(15) Molt mout mot mul mon ＜ mŭltum, ‘much’. For mọt, mul, mọn, see § 74, (2) and § 74, 1. Mọlt, inflected like bęl (§ 103), is used also as an adjective and as a masculine and feminine pronoun.

(16) Negüs ＜ nĕc ūnus, ‘no one’. Inflection: negüs, negü(n); negüna. Beside negün we find degün, apparently through dissimilation.

(17) Nüls ＜ nūllus, ‘no’, ‘none’. Inflection: nüls, nül, nül, nülh, nüls; nüla, nülas. From nülh ＜ nūllī comes a set of forms with lh: see § 67, (2). Hence, by metathesis suggested by the analogy of negün, lhün. A fusion of nülh and lhün results in lünh, whence a set of forms with nh.

(18) Om ＜ hŏmo, ‘one’.

(19) Pauc ＜ paucum, pauci, ‘little’, ‘few’. There is also a regular adjective, paucs, ‘small’.

(20) Que que, ‘whatever’, is a Provençal compound.

(21) Quecs ＜ quĭsquis (§ 78, 2), ‘everyone’. From quẹcs were formed an objective quẹc and a feminine quẹga (cf. amics amic amiga).

(22) Quesacom (diminutive quesacomet), ‘something’, ‘a little’, is formed like calacọm above, the first element in this case being either quẹs ＜ quĭd or quẹ s = quẹ es.

(23) Qui que, ‘whoever’, is a Provençal compound.

(24) Res re, ‘anything’, ‘something’.

(25) Tals ＜ talis, ‘such’, inflected like cals (§ 134).

(26) Tamanh ＜ tam magnum, ‘so great’; f. tamanha.

(27) Tant tan ta ＜ tantum, tanti, ‘so much’, ‘so many’. Tant, inflected like bęl (§ 103), is used also as an adjective and as a masculine and feminine pronoun.

(28) Totz ＜ tōttus = tōtus (Gram., I, § 547), ‘all’, had a regular inflection: tọtz, tọt, tọt, tọtz; tọta, tọtas. In the masculine nominative plural, however, we find oftener the forms tüch tüich tüit tüt tüti, which point to a Latin ✱tūctī (cf. Italian tutti); for this no satisfactory explanation has been discovered (see Nigra, Rom., XXXI, 525). Hence we occasionally have in the singular tütz, tüt, and in the objective plural tügz tütz; the last form occurs also as a nominative plural. Tọt is frequently used as a neuter pronoun and as an adverb.

(29) Üs, ‘some’; from ūnus, used as an indefinite adjective or pronoun, we have the plural forms; ü(n), ü(n)s; ünas.

2. CONJUGATION.

THE FOUR CONJUGATIONS.

137. (1) In Vulgar Latin there were some shifts, the verbs of the second and third conjugations being particularly unstable: cadĕre, capĕre, sapĕre, for instance, often passed into the second, while mŏvēre, rīdēre frequently followed the third, and mŏri, sĕqui usually went into the fourth. Pŏsse, vĕlle, with the new infinitives ✱potēre, ✱volēre, were made to conform with more or less regularity to the second conjugation type. Beside do, dant, sto, stant, there came into use the forms ✱dao, ✱daunt, ✱stao, ✱staunt. Beside facĕre there doubtless existed a verb ✱fare,[102] strongly influenced by dare and stare; the first suggestion of shortening probably came from the monosyllabic imperative singular fac (or fa[103]), which must have led to a plural ✱fate beside facĭte. Habēre and vadĕre[103] also came under the influence of dare and stare; the former adopted, beside habeo, habes, habet, habent, the forms ✱ho, ✱has, ✱hat, ✱hant or ✱haunt. Vadĕre generally lost its past tenses, which were replaced by īre and, in southern Gaul, by annare.[104]

(2) In Provençal the first conjugation was well preserved, and the fourth lost but little. The second and the third lost many verbs (especially learned words) to the fourth: delir, emplir, envazir, espandir, fugir, iauzir, merir, regir, relinquir, reluzir, vertir;[105] cọzer cozir ＜ consuĕre,[106] devire devezir ＜ divīdĕre, dire dir ＜ dīcĕre, lęire legir (also lire lir) ＜ lĕgĕre, quęrre querir ＜ quærĕre, sęgre seguir ＜ sĕqui, tenẹr tenir[107] ＜ tenēre. Moreover, the second and third conjugations, which in Provençal differed practically only in the infinitive, were much confused: cabẹr, cazẹr, mǫrdre, rire, sabẹr; cọrre accorrẹr, mentavẹr mentaure ＜ mente habēre, movẹr mǫure, quęrre querẹr, redẹbre rezemẹr ＜ redĭmĕre. Uc Faidit, a 13th century grammarian, enumerates about 500 verbs in -ar, about 100 in -er and -re, and a little over 100 in -ir.

138. The inchoative ending -scĕre lost its original sense. The -īsc- type, for verbs of the fourth conjugation, was very widely extended, the -isc- becoming a part of the regular present stem of the fourth conjugation, and disappearing from the infinitive: finīre, ✱finīsco ＞ finir, finisc. An obscure substitute for the Latin -ēsc- type produced an ending -eissir -ezir -zir (dis-pigrēscĕre = despereissir, evanēscĕre = envanezir[108]), which was used in forming some new verbs: enfolezir[109] ＜ fǫl, envelhezir ＜ vęlh, envelzir ＜ vil, esclarzir ＜ clar, escürzir ＜ escür oscür, espaorzir ＜ paọr.[110] The -āsc- and -ōsc- types appear in old verbs: irāscĕre ＞ iráisser, co(g)nōscĕre ＞ conọisser.[111]

139. New verbs were formed, in late Vulgar Latin and in Provençal, only in the first and fourth conjugations. The commonest suffixes were -āre, -iāre, -icāre, -idiāre (＜ -ίζειν: § 57, Z), -īre: ✱oblītāre ＞ oblidar, ✱altiāre ＞ aussar, ✱carricāre ＞ cargar, ✱werridiāre ＞ guerreiar, ✱abbellīre ＞ abelir. Germanic verbs generally went into the first conjugation, except those in -jan, most of which entered the fourth: roubôn ＞ raubar (also raubir), wîtan ＞ guidar; furbjan ＞ forbir.

FUNDAMENTAL CHANGES IN INFLECTION.

140. The Latin perfect passive took the sense of a present; amātus est, for instance, under the influence of such phrases as carus est, came to mean ‘he is loved’. This led to the establishment of an entire passive inflection made up of the perfect participle and the parts of the verb ĕsse; and the old passive forms were gradually abandoned, leaving no trace (save the perfect participle and possibly the gerundive) in the Romance languages. So the passive is constructed in Provençal as in French: ẹs amatz, ęra amatz, fọ amatz, será amatz, etc.; the participle regularly agrees with the subject in gender and number. Latin deponent verbs became active: mŏri ＞ morir, sĕqui ＞ seguir.

141. (1) Such phrases as ĭd habeo factum shifted their meaning from ‘I have it done’, etc., to ‘I have done it’, etc. The Latin perfect came to be restricted to its aorist sense, and the perfect was expressed by compounds of habēre with the perfect participle. In the Romance languages all compound tenses were eventually formed in this way: ai cantat, avia cantat, aurai cantat, etc. In Provençal the auxiliary is sometimes ęsser, instead of avẹr, if the main verb is reflexive, passive, or neuter; ęsser is particularly common with neuter verbs of motion: sọi vengütz.[112] A participle used with avẹr may agree in gender and number with the direct object, if there is one: ai cantat or cantada la cansọ.

(2) The Latin perfect indicative continued to be used as an aorist, and is the source of the preterit in Provençal, as in the other Romance languages: vīdī ＞ vi, ‘I saw’. The pluperfect indicative survived in some regions; in Provençal it is used with the sense of a conditional: fŭĕrat ＞ fọra,‘he would be’. The future perfect indicative and the perfect subjunctive did not remain in Provençal: amavĕro = aurai amat, amavĕrim = aia amat. The pluperfect subjunctive assumed the functions of the imperfect, which disappeared from nearly every part of the Romance territory: audīssem (for audīrem) ＞ auzis. The perfect infinitive left no trace: audīsse = avẹr auzit.

142. The Latin future, which was not uniform in the four conjugations, and, in the third and fourth, was liable to confusion with the present subjunctive, was gradually replaced by various periphrastic constructions: instead of faciam people said factūrus sum, dēbeo facĕre, vŏlo facĕre, habeo (ad) facĕre, etc. The construction that prevailed in the greater part of the Empire was facĕre habeo, a combination of the infinitive with the present indicative of habēre. The verb ĕsse was the only one that ultimately retained the old future beside the new: Pr. ęr, ęrs, ęr, beside serái serás será; in the plural, only serẹm, serẹtz, serán. The new composite future was occasionally used by Tertullian, St. Jerome, and St. Augustine, and became common in Italy by the 6th century.[113]

(2) As an imperfect of the future, there was evolved a combination of the infinitive and the imperfect or perfect indicative. To correspond to dīcit quod venīre habet, was constructed dīxit quŏd venīre habēbat (or habuit); to match sī pŏssum, venīre habeo, was made sī potuĭssem, venīre habēbam (or habuī). In Gaul, as in most of the Empire, only the imperfect of habēre was used for this purpose. Traces of such a construction are found as early as the 3d century. This form is generally called the conditional, and it existed in Provençal side by side with the conditional described in § 141, (2): sería, serías, sería, etc., beside fọra, fọras, fọra, etc. The Romance languages developed also a perfect conditional: auría agüt = ‘I should have had’.

143. (1) The present participle remained in use as an adjective: fīlias placentes ＞ filhas plazẹns; cf. § 101, (3). In its verbal function it was replaced by the ablative of the gerund: vĕnit accŭrrens ＞ vĕnit accŭrrendo ＞ ven acorrẹn. In most Provençal dialects, however, the present participle and the gerund coincided in form (amantem and amando both ＞ amán), the gerund being distinguished from the participle only by its lack of inflection: see § 76, (2).

(2) The gerund retained only the ablative case, the use of which was considerably extended: see above. In its other cases it was replaced by the infinitive: artem dīcendī ＞ artem dīcĕre ＞ art de dire. The supine, too, was replaced by the infinitive: vīsum vĕnit nōs ＞ vĕnit nōs vĭdēre ＞ ven nọs (a) vezẹr.

INFINITIVE, PRESENT PARTICIPLE, AND GERUND.

144. The infinitive endings -āre, -ēre, -īre regularly became -ar, -ẹr, -ir; -ĕre became -re or -er: see § 48, (1) and § 52, (1). Ex.: amāre ＞ amár, vĭdēre ＞ vezẹr, audīre ＞ auzir; tŏllĕre ＞ tǫlre, nascĕre ＞ náisser, dīcere ＞ dire dízer. For shifts of conjugation, see § 137.