S remained: amīcus ＞ amics, cŏrpus ＞ cǫrs, facias ＞ fassas, fŏrtes ＞ fǫrtz, ŏpĕras ＞ ǫbras, sŭbtus ＞ sọtz. Between a palatal, or an n that did not fall, and an s, a t developed in some dialects: annos ＞ anz, fīlios ＞ filz; cf. § 63, (1), (8).

T, in independent words, fell very early, except in the preterit of verbs; there it was retained in most dialects in weak preterits of the first and third conjugations, in many dialects in weak preterits of the fourth, but disappeared in strong preterits: amat ＞ ama, cantābat ＞ cantava, dar’ ✱hat ＞ darạ, dōnet ＞ dọn, stat ＞ estạ, partībat ＞ partia, placet ＞ platz, tenēr’ + -ē(b)at ＞ tenria, vĕnit ＞ ven; donāvit ＞ donęt donę, vēndĭdit ✱vendĕ́dit ＞ vendęt vendę, partīvit partīt ＞ partị partịt, placuit ＞ plac, vīdit ＞ vi. In the proclitics et and ✱ot (= aut), the t fell before consonants; before vowels it became d, which, under the influence of ad and quĭd, developed like an original d: et ＞ e, eð ez; ✱ot ＞ o, oð oz; later, e and o came to be used often before vowels also.

Final Groups.

83. Ks remained in Vulgar Latin at the end of monosyllables only (§ 55, X); there it became, in Provençal, is: rēx ＞ rẹis, sĕx ＞ sęis. Grecx, nicx are Latinisms.

Nt was generally reduced to n; but in the extreme north and some parts of the south the t was retained in -ant: amant ＞ aman, habē(b)ant ＞ avian aviant; cantent ＞ canten; vēndunt ＞ vẹndon. In some dialects the n fell after o, u (vẹndo, au); -on and -o were used concurrently by the poets.

St ＞ s in ĕst ＞ ęs. Cf. § 28, 5.

SPORADIC CHANGE.

84. For certain consonant changes no laws have been established.[81] Some of them doubtless originate in the language of children, which is governed by principles different from those which regulate the speech of adults. Others are due to vague associations of sound or sense. Borrowed and learned words are especially exposed to such whimsical alteration.

Insertion.

85. The insertion (or addition) of a consonant, in such cases as those mentioned below, is probably always due to some false association or wrong etymology, but the specific cause often cannot be ascertained; the added consonant seems to be generally a liquid or a nasal:—

Metathesis.

86. Metathesis is not very common in Provençal, although a few texts offer many examples; it is apparently restricted to liquids and nasals:—

Dissimilation.

87. Dissimilation, like the other irregular phenomena, affects mainly liquids and nasals, particularly r; it is not, however, entirely confined to these classes. The two nasals, m and n, are similar enough to undergo dissimilation. Some of the cases go back to Vulgar Latin, while others are peculiar to Provençal or to Provençal and French. In the table below, a dash indicates the total disappearance of the consonant in question.

m + n ＞ m + r: ✱comĭnicāre ＞ comenegar comergar, ✱indomĭnicātum ＞ endomeniat endomergat, mancĭpium ＞ mansip massip marsip, manĭcum ＞ margue, mŏnăchum ＞ mongue morgue.

n + m ＞ r + m: anĭma ＞ anma arma, ✱mĭnimāre ＞ mermar.

n + n ＞ n + r, r + n; ? d + n: canŏnĭcum ＞ canónegue canorgue, venēnum ＞ verin; nec ūnum ＞ negün degün? (cf. Andalusian and Asturian dengun, Catalan dingu, apparently from nec ūnum + nĭngŭlum).

r + r ＞ r + —, — + r, l + r: ✱Bernhardum ＞ Bernart Bernat, marmor ＞ marbre marme, prŏprium ＞ prǫpri prǫpi; dīe Mércūrī (influenced by dīe Vénĕris) ＞ dimęrcres dimęcres, grandem rem ＞ granrẹ ganrẹ, pr(eh)ĕndĕre ＞ prenre penre, prĕsby̆ter ＞ pręstre pęstre; arbĭtrium ＞ albire, ✱Arvĕrnium ＞ Alvęrnhe, peregrīnus ＞ pele(g)rīnus ＞ peleris, pŭrpŭra ＞ pọlpra.

s + s ＞ — + s: ✱spasmāre ＞ (espasme) pasmar (cf. French pâmer), perhaps through confusion of the initial es- with the prefix ex-.

t + t? ＞ — + t: statiōnem? ＞ sazọ (cf. French saison, Spanish sazon).

y + y ＞ y + —: ✱disjejūnāre ＞ ✱disjeunāre ＞ ✱disy’nāre ＞ dis´nar disnar.

III. MORPHOLOGY.

88. The most important morphological developments are common to all, or nearly all, the Romance languages. They may therefore be ascribed, in their early stages, to Vulgar Latin, although direct evidence of their beginnings is scanty.

89. (1) During the late Vulgar Latin and early Romance period neuter nouns gradually became masculine; this change was doubtless due in part to phonetic developments which obliterated distinctive endings: dōnum ＞ dọn, m.; nōmen ＞ nọm, m. Mare, however, became almost always feminine in Gaul: la mar. Some neuter plurals in -a, used mainly in a collective sense, were preserved and eventually became feminine singulars: fŏlium fŏlia ＞ fǫlha, f. sg.; lĭgnum lĭgna ＞ lẹnha, f. sg.; so luǫgua, pọma, prada, beside lǫc, pọm, prat (and, by analogy, grasa, beside gras ＜ gradus); similarly labia ＞ lavias, f. pl.

(2) Masculine and feminine nouns usually kept their original gender. Abstract nouns in -or, however, regularly became feminine in Gaul, other abstract nouns being mostly feminine in Latin: honōrem ＞ onọr, f.; sapōrem ＞ sabọr, f. With the exception of manus, which generally retained its gender, feminine nouns of the second and fourth declensions, unless they passed into the first declension (pĭrus ＞ pẹra), became masculine, to conform to the usual -us type: fraxĭnus ＞ fraisnes, m.; pīnus ＞ pins, m. Attracted by such words as these, arbor became masculine. There were some other less important shifts.

90. Some nouns passed from the fourth to the second declension in the classic Latin period (dŏmus, fīcus); the rest doubtless followed in Vulgar Latin (frūctus,[82] gradus, manus). Fifth declension nouns in -ies went over, for the most part, to the first declension:[83] dīes ＞ dia, facies ＞ fassa, glacies ＞ glassa, rabies ＞ rabia; but we find also di, fatz, glatz (ratge is probably French), following the third declension type. Fifth declension nouns which did not shift to the first came to be declined after the model of the third (fides, res, spes). The five declensions were therefore reduced to three, presumably in Vulgar Latin times. Among these there were some exchanges: polvera, vergena; cf. § 89, (1), (2), 1.

91. The use of cases became more and more restricted in Vulgar Latin, prepositional constructions taking the place of pure case distinction. At the beginning of the Romance period, nouns probably had, in unstudied speech, only two cases in constant use: a nominative and an accusative or accusative-ablative. These two cases were generally retained in Provençal, for the second and third declensions, until the literary period: we may call them nominative and objective.

(1) The locative, which had almost vanished in classic Latin, lingered in Vulgar Latin only in names of places. It has left no sure traces in Provençal.

(2) The vocative, in classic Latin, was like the nominative for most words; in Vulgar Latin it probably disappeared, except in Church phrases, such as mī dŏmĭne. In Provençal we find the nominative regularly used in address (chanzọs, companh, emperaire, ioglars, Papiols), although the objective occasionally occurs in its stead (barọns pl., ioglar malastrüc, trachọr).

(3) The genitive, in the popular language, was little by little replaced by other constructions—commonly by the ablative with de or by the dative; the beginnings of this substitution may be observed as early as Plautus. Among Provençal nouns—aside from such learned forms as ancianọr, christianọr, companhọr, paianọr, parentọr—we find remnants of the genitive only in a few compound words, as diiǫus ＜ dīe Jŏvis, and in the standing phrase ẹs mestięr ＜ est ministĕrii.

(4) The dative, which in most words had the same ending as the ablative, came to be replaced, in the greater part of the Empire, by the accusative with ad; this construction, too, goes back as far as Plautus. Provençal nouns retain no traces of the dative.

(5) The ablative, after the fall of final m (§ 55, M) and the loss of quantitive distinctions in unstressed syllables (§ 21), differed little or not at all from the accusative in the singular of nearly all nouns: causăm causā, dōnŭm dōnō, patrĕm patrĕ, frūctŭm frūctū, dīĕm dīē. Furthermore, some prepositions (especially in) were used both with the accusative and with the ablative. It was inevitable, then, that the two cases should be confounded in the singular, and we have evidence of such confusion as early as the first century of our era; this led gradually to a substitution of the accusative for the ablative in the plural, the accusative plural being somewhat commoner and frequently simpler than the ablative. We may, therefore, take the accusative as the basis of the Provençal objective, remembering, however, that this accusative has been more or less blended with the ablative.

(6) The two-case declension remained theoretically in use in Provençal literature through the 14th century; but in texts later than the 12th, cases are often confused. From the spoken language the declension disappeared, in the west (as in Catalan), before the literary period; in the centre and east, probably in the 12th century; in the north, in the 13th. The case preserved was usually the objective, but sometimes the nominative. Some nouns in -aire -adọr kept both forms, with a differentiation of meaning.

92. In the discussion of declensions some phonetic peculiarities must not be overlooked:—

(1) In the nom. pl. of the 2d declension, a stressed ẹ, followed in the next syllable by final -ị, would regularly give ị (cf. § 27, 1); but the ẹ is preserved by the analogy of the nom. and obj. sg. and the obj. pl.: capĭllī ＞ cabẹl, mĭssī ＞ mẹs, quētī ＞ quẹt, sērī ＞ sẹr. We do, however, find cabil, and (perhaps by analogy) auzil ＜ aucĕllī.

(2) In the nom. pl. of the 2d declension, a c or g before the final -ị would regularly be palatalized (cf. § 55, C, G); but it is preserved from palatalization by the analogy of the other three forms: amīcī ＞ amic, lŏngī ＞ lonc.

(3) For the development of a t between a palatal or an n and a final s, see § 82, S: annos ＞ anz, fīlios ＞ filz.

(4) For the simplification of final scs, sts to cs, ts, see § 78, 2: ✱bŭscus? ＞ bǫcs, trīstes ＞ tritz.

(5) For the history of -arius and -tōrius, see § 23, 1 and § 73, Ry, 1.

93. (1) Nouns whose objective singular ended in s were invariable in the earlier part of the literary period; bracchium ＞ bratz, cŏrpus ＞ cǫrs, imperatrīcem ＞ emperairitz, fascem ＞ fais, latus ＞ latz, lūcem ＞ lütz, mĭssum ＞ mẹs, nasum ＞ nas, ŏpus ＞ ǫps, ŭrsum ＞ ọrs, pĭscem ＞ pẹis, pĕctus ＞ pęitz, prĕtium ＞ prętz, tĕmpus ＞ tems, vĕrsum ＞ vęrs, vīsum ＞ vis, vōcem ＞ vọtz. Later, however, a plural (originally obj. pl.) was made for such words by adding -es, generally at a time when final ts had been reduced to s (§ 64): brasses, cǫrses, mẹsses, pẹisses, vęrses; examples occur as early as the end of the 12th century.

(2) Other invariable nouns are midons, sidons, and often laus and rẹs; the last two sometimes have an objective lau, rẹ. Midons comes from the Church Latin mī dŏmĭne, which was popularized by the substitution of the Provençal don for dŏmĭne and the addition of the nom. -s; the term was transferred from religious to feudal, and thence to amatory use, and came to mean ‘my lady.’ Mi domina was common in Church Latin. Sidons is formed on the model of midons.

(3) For nouns in tš, see § 63, (1): ✱disdūctum ＞ desdüg, frūctum ＞ früch, gaudium ＞ gaug, nŏctem ＞ nuech. Such words were very often written in the plural with -gz, which was pronounced either ts or tš. The pronunciation ts is attested by such rhymes as malfagz: alumenatz.

94. Infinitives used substantively conformed to the 2d declension type: lauzars lauzar (like fǫcs fǫc), rire-s rire (like fabre-s fabre): see § 96. The same thing is true of masculine post-verbal nouns: (getar) gętz gęt, (guidar) guitz guit, (lansar) lans (invariable).

First Declension.

95. This declension came to include a part of the fifth and also some neuter plurals of the second and third. With the exception of dia (nearly always masculine) and of a few learned words, it contained only feminine nouns. As the nominative, accusative, and ablative singular early became identical, leaving only one form in the singular, the plural forms were reduced to one, the accusative crowding out the nominative; this substitution, which must have been begun before the Provençal period, was doubtless helped by the identity of nominative and accusative plural in feminine nouns of the third declension. Causa will serve as a model:—

96. This declension came to include the fourth. With the exception of mas, ‘hand’ (generally feminine), it contained only masculine nouns. The different types may be illustrated by fŏcus, dōnum, faber:—

For the c of fŏcī, see § 92, (2). For ✱dŏnus, etc., see § 89 (1). Nom. fabre is due to the analogy of the other three cases; the s of fabres is borrowed from the prevailing fǫcs type.

Third Declension.

97. This declension absorbed a part of the fifth: cf. § 90.

98. Nouns whose stem was different in the nominative and the accusative singular, reconstructed the nominative to correspond to the accusative, the new form being similar to the original genitive: papĭlio papiliōnem ＞ papiliōnis papiliōnem, pēs pĕdem ＞ pĕdis pĕdem. The change began in the Vulgar Latin period. Exceptions to the rule are names of persons, unless they ended in -ans or -ens: nĕpos nepōtem ＞ nęps nebọt; but amans amantem ＞ ✱amantis amantem ＞ amáns amán.

99. Masculine nouns of the third declension, early in the Provençal period, made their nominative plural conform to the second declension type, thus distinguishing it from the objective plural: pater patrem patres patres ＞ paire paire paire paires (cf. Old French and Italian). Feminines, on the other hand, kept the nominative plural in -s: mater matrem matres matres ＞ maire maire maires maires.

100. A few neuter nouns, becoming masculine in Vulgar Latin, developed distinctively masculine forms in the singular: gĕnus gĕnus ＞ ✱gĕneris ✱gĕnĕrem ＞ genres genre; so fŭlgur (＞ fŭlger) ＞ ✱fŭlgĕrem ＞ fọuzer. Most neuters, however, kept in the singular their original stem: sēmen ＞ sẹm, tĕmpus ＞ tems. But those in -men regularly, and those in -r sometimes, took an -s in the nominative singular: flūmen flūmen ＞ flüms flüm, marmor marmor ＞ marme-s marme; cǫr, in the literary language, usually has no nominative -s. In the plural most neuters brought their forms into harmony with the masculine type, but those in -us kept the -s throughout: (caput >) capus ✱capum capĭta capĭta ＞ caps cap cap caps, cŏr cŏr cŏrda cŏrda ＞ cǫr cǫr cǫr cǫrs, nōmen nōmen nōmĭna nōmĭna ＞ nọms nọm nọm nọms; but cŏrpus cŏrpus cŏrpŏra cŏrpŏra ＞ cǫrs cǫrs cǫrs cǫrs. Mare, becoming feminine, was declined thus: mars mar mars mars.

101. The third declension comprises three principal types: (1) nouns which in Latin had no difference of stem or of accent between the nominative and the accusative singular; (2) those which had a difference of stem but not of accent; (3) those which had a difference of accent.

(3) Nouns with a difference of accent:—

ADJECTIVES.

102. What has been said concerning the inflection of nouns applies also to adjectives: see §§ 91-101. For pronominal adjectives see §§ 114 ff.

103. We must recognize two classes of adjectives: (1) those which in Latin distinguish the feminine from the masculine; (2) those which do not.

104. In impersonal constructions we frequently find a nominative singular without -s, which is apparently a survival of the Latin neuter: m’es bel (greu, parven, semblan, etc.) que.… But the form with -s sometimes occurs in the same constructions: m’es greus que.…