1. In poetry ía is sometimes counted as one syllable: avi͡an, devi͡an.

2. For some subsequent developments of western dialects, see Meyer-Lübke, Gram., II, p. 326.

3. For the personal endings, see §§ 164-169.

4. Esser has: ęra, ęras, ęra, erám, erátz erás, ęran ęron ęro.

PRETERIT, OLD CONDITIONAL, AND IMPERFECT SUBJUNCTIVE.

171. These parts are all formed from the same stem, that of the Latin perfect: cf. § 141, (2). Ex.: cantęi, cantęra, cantęs; vendęi, vendęra, vendęs; partí, partíra, partís; vi, vira, vis; dẹc, dẹgra, deguẹs.

Preterit.

172. Preterits which stress the ending throughout are called weak; those which do not stress the ending throughout are called strong: partí, partíst, partí, partím, partítz, partíron is weak; saup, saubíst, saup, saubém, saubétz, sáubron is strong. Verbs of the first and fourth conjugations regularly have weak preterits (amęi, finí). Verbs of the second and third, with very few exceptions, originally had strong preterits (placuī > plac, fēcī > fis): many of them, however, developed weak preterits either in Vulgar Latin or in Provençal (irasquęi, nasquęi, tessęi tesquęi, visquęi); some assumed a weak form in -í in the 1st pers. sg. (dis dissí, pris prenguí, remas remanguí, trais traguí: cf. §§ 173, 177); quęrre, on the other hand, substituted a strong preterit (quis, etc.) for a weak one.

173. (1) Final , in the first pers. sg., doubtless remained through the earlier stages of Provençal (habuī > águi, dīxī > díssi): cf. § 51, (2). Before it fell, it changed an accented ẹ in the preceding syllable to i (vēnīvēnuī > ✱vẹngui vinc): cf. § 27; occasionally, however, the ẹ was kept, through the analogy of the other persons (pris prẹs). Sometimes, instead of falling, the -i took the accent (following the analogy of the fourth conjugation) and remained: águi > ac or aguí, díssi > dis or dissí (cf. § 177).

When the -i was immediately preceded by an accented vowel, it regularly formed a diphthong with that vowel, and did not fall (fuī > füi): cf. § 51, (3); but -íi was simplified to -i (partīvī partīī > partí).

Before enclitic l, -ei -iei were often reduced to -e -ie: cantiel.

(2) In the 2d pers. sg., -stī became -st, a preceding ẹ being changed to i (§ 27): partīstī > partíst, debuĭstī > deguíst; sometimes, through the analogy of the 2d pers. pl., ẹ remains (venguẹst: cf. § 27, 2). Occasionally the final -t disappears: anięst anięs, fezíst fezís.

(3) The -t of the 3d pers. sg. was lost in strong preterits: placuit > plac, vīdit > vi. In weak preterits, it was retained by most dialects after é, and by many after í: donęt donę, vendęt vendę; partí partít. Cf. § 82, T.

(4) In the 1st pers. pl., -mus -mu’ (see § 167) was reduced to -m: vīdĭmu’ > vim.

(5) The -stis of the 2d pers. pl. regularly became -tz (§ 78, 2), later in many dialects -s (§ 64): debuĭstis > deguẹtz deguẹs.

(6) The -runt of the 3d pers. pl. regularly gave -ron or -ro (§ 83, Nt): partīrunt > partíron partíro, vīdĕrunt > viron viro. In some dialects -en is substituted for -on: fŭĕrunt > fọron fọren (cf. § 169).

The e before -runt, which in classic Latin was usually long, was always short in Vulgar Latin when it was preserved at all: amavĕruntamārunt, fēcĕrunt.

Weak Preterits.

174. (1) In the first and fourth conjugations we find in Latin the following endings:—

-āvī -āī-āvĭmus-īvī -īī-īvĭmus
-āvĭstī -āstī-āvĭstis -āstis-īvĭstī -īstī-īvĭstis -īstis
-āvit -aut-āvēre -āvē̆runt -ārunt-īvit -īit -īt-īvēre -īvē̆runt -īrunt

The popular speech preferred in every case the shortened form, and generally reduced -āvĭmus, -īvĭmus to -āmus, -īmus (in southern Gaul -āmu’, -īmu’: § 167), on the analogy of the 2d pers. sg. and pl.

(2) In the second conjugation a few verbs (delēre, flēre, nēre, -olēre, -plēre, viēre) had similar endings (delēvī, etc.), which were doubtless contracted in like fashion in so far as these words were in common use. Most verbs of this conjugation, however, had strong preterits (tacēre, tacuī; vidēre, vīdī; etc.).

(3) The third conjugation had in classic Latin no weak endings corresponding to those of the first, second, and fourth; but the vulgar speech developed a set in the following manner. Compounds of dare formed their perfect in -dĭdī (perdĭdī); this -dĭdī, in accordance with the principle set forth in § 16, 3, came to be pronounced -dę́dị (condédi); and -dędị, probably through dissimilation[124], was shortened to -dęi (✱credéi). With this form as a starting-point, a weak preterit was created on the analogy of those of the other conjugations, the endings being something like -ęi, -ęstị, -ęt, -ęmus -ęmu’, -ęstis, -ęrunt. This inflection was probably extended to some verbs outside the -dĕre class (✱battéi, etc.?).

175. (1) In Provençal the weak inflection disappeared from the second conjugation, delēre and -plēre passing into the fourth, and the other weak verbs going out of use.

(2) Verbs of the fourth conjugation (except venir) all took the weak endings -í, -íst, -í, -ím, -ítz, -íron: partí, partíst, partí, partím, partítz, partíron. Irregular verbs either disappeared or became regular (sensī = sentí), with the exception of venīre > venir (vinc).[125]

(3) The new weak endings of the third conjugation developed into -ęi, -ęst, -ęt, -ém, -ętz, -ęron: vendęi, vendęst, vendęt, vendém, vendętz, vendęron. In the 1st pers. sg. the ę often broke (vendięi), and the diphthong was sometimes carried into the 2d pers. sg. (vendięst). These endings were considerably extended in Provençal (cazęt, etc.), and were occasionally attached to a strong preterit stem (nasquęt, tesquęt, venquęt, visquęt). Most verbs, however, kept their strong preterit (mis, conọc). The -īvī perfect disappeared from the third conjugation: quæsīvit > ✱quæsit > quẹs.

(4) The first conjugation discarded its own weak endings, and substituted those of the third: cantęi cantięi, cantęst cantięst, cantęt, cantém, cantętz, cantęron. This strange phenomenon seems to have originated as follows: dare, dĕdī > dar, dęi; from dar the ending -ęi was readily extended to estar (estęi); and from these two very common verbs it spread to the whole first conjugation.

Irregular verbs (except dare, stare) either disappeared or became regular.

1. According to Meyer-Lübke, Gram., II, p. 304, Latin -ai became by phonetic process -ęi in Vulgar Latin, and -ęi or -ięi in Provençal. There seems to be no evidence to support this theory. Cf. § 23, 2.

2. In the dialects of Béarn and Catalonia the original a remains in some parts of the preterit.

176. A final -c, which developed in the strong -ui preterits (§ 184), often became attached to the 3d pers. sg. of weak preterits of the fourth conjugation: floríc, fugíc, iauzíc, partíc.[126] It was sometimes extended to other weak preterits: chantęc, entendęc, nasquęc,[127] paręc.[128] We find also a 3d pers. pl. cazęgron, etc., and even a 1st pers. sg. ameguí, etc. In some western dialects the final -c was adopted by the whole first conjugation: donęc, portęc, etc.

177. Some strong preterits occasionally assumed weak endings:—

(1) In the 1st pers. sg. several verbs in -s sometimes either added an -í or shifted the stress to an originally unaccented final -i (cf. §§ 172, 173): dis dissí, pris presí, quis quesí, respos respozí. A few verbs in -c did the same: aic aiguí, bẹc beguí, conọc conoguí, saup saubí, vinc venguí, vǫlc volguí. An ending -guí being thus established, this syllable was sometimes added to preterits not of the -c class: costrenguí, destrenguí, prenguí, remanguí, restrenguí, traguí.

(2) In the 3d pers. sg. weak endings are rare: ac aguęt, vẹnc venguęt.

(3) In the 3d pers. pl. the weak ending is not uncommon in -s preterits: diron dissęron, düistrent düissęron, mẹsdren mezęron, prẹson presęron, remastrent remazęron, traissęron. We probably have to deal here, as in (1), with a shift of accent—dīxĕrunt > ✱dísseron > dissęron, etc.: see § 49, (2). The same thing may be true of such a form as aguęron, beside ágron, from ✱áβwerunt = habuĕrunt; such a form as visquęron, on the other hand, is doubtless imitative.

Strong Preterits.

178. (1) The reduplicative perfects were discarded in Vulgar Latin, with the exception of dĕdi (and its compounds) and stĕti, whose reduplicative character was no longer apparent. Cecĭdī became ✱cadui or ✱cadéi; the rest either disappeared or passed into the -sī class: cucŭrrī > ✱cŭrsī, momŏrdī > ✱mŏrsī, pepĕndī > ✱pē(n)sī, pupŭgī > ✱punxī, tetĕndī > ✱tē(n)sī, tetĕgī > ✱taxītanxī.

(2) The -i perfects were greatly reduced in number in Vulgar Latin. Some disappeared (ēgī), some became weak (fūgī > ✱fugīī > fügí); others passed into the -sī or the -uī class: prehĕndī > ✱prē(n)sī > pris; bĭbit > ✱bĭbuit > bẹc, vēnitvēnuit > vẹnc. In Provençal only three verbs remained: fēcī > fis, fuī > füi, vīdī > vi.

(3) Of the -sī class (including -ssī and -xī) over twenty verbs were preserved in Vulgar Latin (dīxī, excŭssī, mīsī, traxī, etc.), and about the same number passed into this class from others (absco(n)sī, ✱fraxīsŭrsī, etc.): cf. (1) and (2) above. In Provençal nearly half the verbs of the second and third conjugations have -sī preterits: rema(n)sī > remas, ✱respō(n)sī > respọs[129].

(4) The -uī class held its own very well in Vulgar Latin (placuī, etc.) and received some additions (natus sum > ✱nacuī, sustŭlī > ✱tŏluī, vēnī > ✱vēnuī, vīcī > ✱vĭncuī, vīxī > ✱vīscuī, etc.)[130]. To this class belonged, in Vulgar Latin (and, according to Meyer-Lübke[131], in classic Latin also), all perfects in -vī, this ending being pronounced -wŭī, later -wwị or -βwị: cognōvī > ✱conōvuī > conọc, crēvit > ✱crēvuit > crẹc, mōvī > ✱mŏvuī > mǫc. Cf. § 148. In Provençal not far from half the verbs of the second and third conjugations have -uī preterits. For a combination of a -c < -uī stem with a weak ending, see § 175, (3). For the extension of -c < -uī to other conjugations, see § 176.

179. In the 1st pers. pl. the accent was shifted to the ending, to make this form correspond to the 2d pers. sg. and pl.: fēcĭmus > ✱fēcĭ́mu’ > fezẹm (cf. fecĭstī > fezist, fecĭstis > fezẹtz), ✱prē(n)sĭmus > ✱presĭ́mu’ > prezẹm, debŭĭmusdeβwĭ́mu’ > deguẹm. Exceptions are fŭĭmus > fọm, vīdĭmus > vim; in these verbs the 2d pers. forms also are monosyllabic (füst, fọtz; vist, vitz).

180. We find in some verbs an irregular 3d pers. pl. without -r-, made by adding -on or -en to the 3d pers. sg., the final consonant of which is voiced in all verbs in which it is voiced in the other persons of the plural: (aucire) aucis, aucíson; (plánher) plais, pláisson; (prenre) prẹs, prẹson; (remanre) remas, remáson; (venir) vẹnc, vẹnguen; (volẹr) vǫlc, vǫlgon.

1. Prenre has preiron (beside preson preseron), probably through the analogy of feironfēcĕrunt. Mairon, from maner, is perhaps to be explained in the same way.

181. (1) Through the change of -e- to -i- by the influence of a final -ī, as described in § 173, (1), a distinction was established between the first and the third person singular of some preterits: crēvī > cric, crēvit > crẹc; fēcī > fis, fēcit > fẹs; ✱prē(n)sī > pris, ✱prē(n)sit > prẹs; tĕnuītēnuī[132] > tinc, tĕnuittēnuit > tẹnc; vēnīvēnuī[132] > vinc, vēnitvēnuit > vẹnc. Mẹtre, also, has mis, mẹs, which may come from ✱mĭssīmĭssit (cf. mĭssum) = mīsī, mīsit; or perhaps mis comes from mīsī and mẹs is analogical. Through the analogy of such forms, quęrre has quis, quẹs. In the preterit of podẹr, both pŏtuī and pŏtuit would regularly have given pǫc puǫc puęc (§ 37), but pǫc was kept for the 3d person, and puǫc puęc was used for the 1st. The preterit of volẹr differentiates the two persons similarly—vuęlc, vǫlc; here the diphthong (perhaps under the influence of puęc) is borrowed from the present, where we have ✱vŏleo > vuęlh, ✱vŏlet > vǫl (§ 37). Avẹr, likewise, borrows a distinction from the present: aic, ac reproduce the vowels of ai, a; aic + aguí > aiguí.

(2) For -í as a characteristic of the first person, see § 177, (1).

(3) For -c as a distinctive mark of the third person, see § 176.

182. The three -ī perfects developed in Provençal as follows:—

(1) Facĕre > faire (✱fare > far) has:

fēcī>fis, fezífēĭmusfēcĭ́mu’>fezẹm
fēcĭ́stī>fezíst fezísfēcĭstis>fezẹtz fezẹs
fēcit>fẹtz fẹsfēcĕrunt>fẹiron fẹiro

1. We do not find, in the 1st pers. sg., as we should expect (§ 65, C´), fitz beside fis; doubtless the form came early under the influence of mis, pris, quis, etc. For fezí, see § 177, (1). There is also a form fi, due, perhaps, to the analogy of vivīdī; corresponding to fi are 3d pers. sg. fe, and pl. fem, fes, feron. A rare figuí is evidently made on the model of aiguí, etc. In the 3d pers. sg. we find also fei, which seems to be patterned after feiron or after the present fai.

(2) Esse (> ✱ĕssĕre > ęsser ęstre) had originally a long u in the perfect. In literary Latin the u was shortened, but the popular speech seems to have kept ū beside ŭ. The Provençal 1st and 2d pers. sg. apparently come from fūī, ✱fūstī = fuĭstī (although Pr. füi might be taken from fŭī), while the other forms presuppose ŭ:

fūī>füifŭĭmusfŭmu’>fọm
fūĭstīfūstī>füst füsfŭĭstisfŭstis>fọtz fọs
fŭitfŭt>fọ, fọn, fọncfŭĕruntfŭrunt>fọron fọro, fọren

1. A rare fo in the 1st pers. sg. seems to be simply borrowed from the 3d. In the 3d pers. sg., fon beside fo is due to the analogy of -on -o in the 3d pers. pl., and, in general, of such double forms as bon bo, mon mo, son so, ton to: cf. § 63, (5). Fonc shows the influence of tenc, venc.

(3) Vidēre > vezẹr has:

vīdī>✱viði ✱við vi, vicvīdĭmusvīdĭmu’>✱viðmu ✱viim vim
vīdĭstī>vist visvīdĭstis>vitz vis
vīdit>✱við vi, vit, vicvīdĕrunt>✱viðrun viron viro

1. The 1st pers. sg. vic is patterned upon aichabuī, criccrēvi, etc. The 2d pers. forms are irregular, as we should expect ✱vezist, ✱vezetz: evidently the 2d pers. followed the analogy of the 1st and 3d. In the 3d pers. sg., vit and vic follow the model of partit, partic, etc.: see § 173, (3), and § 176.

183. In the -sī perfect the 3d pers. pl. presented difficulties. If the -e- of the penult fell, an s or z and an r were brought together. Most dialects apparently preserved the -e-, and shifted the accent to it (aucizęron, condüissęron, dissęron, prezęron, remazęron, traissęron), or else borrowed outright the weak ending (respondęron): cf. § 49, (2), and § 177, (3). Dialects which lost the -e- too early to follow this method, generally suppressed the sibilant (aucíron, diron, mẹron from mẹtre, remáron), or omitted the -r- and formed the 3d pers. pl. directly from the 3d pers. sg. (aucízon, pláisson, prẹzon, remázon: § 180), or else imitated a preterit of another class (mairon from manẹr, prẹiron from prenre, doubtless patterned after fẹiron < fēcĕrunt); some borderland dialects kept the sibilant and the r, and developed a dental between them (düystrent < dūxĕrunt, mẹsdren < mīsĕrunt + ✱mĭssĕrunt: § 70, Sr, Zr).

As examples of the -sī perfect we may take the preterit of dire < dīcĕre and penre prenre < pr(eh)ĕndĕre:—

(1)

dīxī>dis, dissídīxĭmusdīxĭ́mu’>dissẹm
dīxĭstī>dissístdīxĭstis>dissẹtz dissẹs
dīxit>disdīxĕrunt>dissęron, diron diro

(2)

prē(n)sī>pris, prẹs, presíprē(n)sĭmusprēsĭ́mu’>presẹm
prē(n)sĭstī>presístprē(n)sĭstis>presẹtz presẹs
prē(n)sit>prẹsprē(n)sĕrunt>presęron, prẹson, prẹiron

(3) Escriure < scrībĕre has, beside escris < scrīpsī, a preterit escrius (cf. p. p. escriut escrit escrich), in which the u is probably due to the influence of the infinitive.

(4) For dissí, presí, quesí, respozí, see § 177, (1). For pris prẹs, etc., see § 173, (1).

184. In the -uī perfect the development depends somewhat upon the consonant preceding the u. The treatment of the various cons. + w groups, which was discussed in § 72, may be illustrated by habuit > ac[133], crēvitcrēvuit > crẹc[134]; cuit > nǫc[135]; sēditduit > sęc, tuit > pǫc; valuit > valc[136], tĕnuitnuit > tẹnc[137], ruit > męrc; sapuit > saup[138]: the noteworthy features are the change of u to -c (through w, gw, g), the absorption of the preceding consonant unless it be a liquid, a nasal, or a p, the preservation of the liquid or nasal, and the metathesis of the p.

Avẹr < habēre, podẹr < ✱pŏtēre pŏsse, volẹr < ✱vŏlēre vĕlle, sabẹr < ✱sapēre sapĕre will serve as examples (for the accentuation of the 3d pers. pl., see § 16, 2):—

(1)

habuī>ac, aguí, aic, aiguíhabuĭmusaβwĭ́mu’>aguẹm
habuĭstī>aguísthabuĭstis>aguẹtz aguẹs
habuit>achabuĕrunt>ágron ágro, aguęron

1. For aguí (beguí, conoguí), see § 177, (1). For aic, aiguí, (cric), see § 181, (1). For aguęron (visquęron), see § 177, (3).

(2)

pŏtuī>pǫc puǫc puęcpotuĭmuspotwĭ́mu’>poguẹm
potuĭstī>poguístpotuĭstis>poguẹtz poguẹs
pŏtuit>pǫc, pǫtpotuĕrunt>pǫgron pǫgro

1. For puoc, see § 181, (1). Pot is apparently due to the combined influence of weak preterits and the parts of poder in which the dental is preserved.

(3)

vŏluī>vǫlc, vuęlc, volguívoluĭmusvolwĭ́mu’>volguẹm
voluĭstī>volguístvoluĭstis>volguẹtz volguẹs
vŏluit>vǫlcvoluĕrunt>vǫlgron vǫlgro

1. For vuelc (tinc, vinc), see § 181, (1); for volguí (venguí), § 177, (1).

(4)

sapuī>saup, saubísapuĭmussapwĭ́mu’>saubẹm
sapuĭstī>saubístsapuĭstis>saubẹtz saubẹs
sapuit>saupsapuĕrunt>sáubron sáubro, sáupron

1. For saubí, see § 177, (1). For sáupron (sáupra, saupés, saupút), see § 65, P, 3; cf. § 148, (2).

Old Conditional.

185. The old conditional came from the Latin pluperfect indicative, which had been supplanted in its pluperfect sense by a compound form, and was gradually restricted in its use to the functions of a preterit, a perfect conditional, and a simple conditional: see § 141, (2). In Provençal it had only the conditional meaning; and as the new conditional rendered it superfluous, it fell into disuse (with the exception of ágra and fọra) in the 13th and 14th centuries: see § 142, (2).

186. In the fourth conjugation the old conditional comes from the contracted form of the pluperfect (audīramaudīvĕram). Weak verbs of the third conjugation constructed a similar form (✱vendęram). First conjugation verbs started with the contracted pluperfect (amāramamāvĕram), but in Provençal substituted ę for á, as in the preterit: § 175, (4). The Provençal types of the old conditional of weak verbs are, therefore, represented by: amęra, vendęra, auzíra. The inflection is as follows:—

amęraamerám
amęrasamerátz
amęraamęran
auzíraauzirám
auzírasauzirátz
auzíraauzíran

187. Strong verbs of the and the -uī classes regularly took their old conditional directly from the Latin pluperfect: fēcĕram > fẹira, fŭĕram > fọra, vīdĕram > vira; habŭĕramáβwĕram16, 2) > ágra, pŏtŭĕram > pǫgra, vŏlŭĕram > vǫlgra, sapŭĕram > sáubra sáupra (§ 65, P, 3). Of course the Latin pluperfect, and therefore the Provençal conditional, followed the shift of the perfect if it changed from one class to another: vēnī > ✱vēnuī, hence ✱vēnŭĕram > vẹngra. The inflection is as follows:—

fọraforámágraagrám
fọrasforátz foráságrasagrátz agrás
fọrafọranágraágran

1. Faire has féra (cf. feron) beside féira.

2. For sáupra, cf. § 148, (2), and § 184, (4), 1, and § 192.

3. Devér has beside dégra a form déura, evidently influenced by the new conditional, deuría.

4. Páisser, plazér have beside págra, plágra the forms paisséra, plazéra.

188. Strong verbs of the -sī class regularly form their old conditional on the same plan as the 3d pers. pl. of the preterit (§ 183): (dīxĕram) díra, cf. díron; (✱prēsĕram) prẹira, cf. prẹiron; (arsĕram) arsęra, cf. arsęron.

189. It will be noted that in all verbs, weak and strong, the old conditional may be constructed from the 3d pers. pl. of the preterit by changing -on to -a.

Imperfect Subjunctive.

190. The Provençal imperfect subjunctive came from the Latin pluperfect subjunctive, which in Vulgar Latin assumed the functions of the imperfect and generally displaced it, its own place having been taken by a compound form: see § 141, (2).

191. For weak verbs the basis was the contracted form of the first and fourth conjugations (amāssemamāvĭssem, audīssemaudīvĭssem); weak verbs of third conjugation had a similar analogical form (✱vēndęssem). First conjugation verbs substituted ę for á, as in the perfect and the old conditional: § 175, (4); § 186. The Provençal types are: amęs, vendęs, auzís. The inflection is:

amęsamessẹmauzísauzissẹm
amęssesamessẹtz -ẹsauzíssesauzissẹtz -ẹs
amęsamęssen -on -oauzísauzíssen -on -o

192. Strong verbs regularly made their imperfect directly from the Vulgar Latin form of the pluperfect: fecĭssem > fezẹs, fŭĭssemfŭssem > fọs, vidĭssem > vezẹs, venĭssemvenuĭssem > venguẹs; dixĭssem > dissẹs, ✱pre(n)sĭssem > prezẹs; habuĭssem > aguẹs, potuĭssem > poguẹs, voluĭssem > volguẹs, sapuĭssem > saubẹs saupẹs (§ 65, P, 3). The inflection is:—

fọsfossẹmaguẹsaguessẹm, acsẹm
fọssesfossẹtz -ẹsaguẹssesaguessẹtz -ẹs, acsẹtz -ẹs
fọsfọssen -on -oaguẹsaguẹssen -on -o

1. The syncopated forms in the 1st and 2d pers. pl. are common to the -uī class: decsém, iacsém, pocsém, saupsém.

2. In the 3d pers. pl. -an sometimes takes the place of -en or -on: mezéssan, saubéssan. This ending is doubtless borrowed from the present subjunctive and the old conditional.

3. Vezér has vis beside vezés. From faire we find in the 3d pers. pl. fésson.

4. Metre has mezés, due, no doubt, to the analogy of mes and of prezés.

193. Some dialects have an ending -a, -as, -a, -ám, -átz, an, borrowed from the present subjunctive and the old conditional, but added to the stem of the imperfect subjunctive: chantęssa, vendęssa, floríssa; fọssa.


FOOTNOTES

[1] There should now be added: J. B. Beck, Die Melodien der Troubadours, 1908.

[2] See B. Schädel in Rom., XXXVII, 140.

[3] See E. Bourciez, Les Mots espagnols comparés aux mots gascons.

[4] See C. Chabaneau, la Langue et la littérature du Limousin, in the Revue des langues romanes, XXXV, 379.

[5] See G. Paris, Origines de la poésie lyrique en France au moyen âge.

[6] This h (coming from f) is peculiar to Gascon; the other dialects have no h.

[7] G, b, d are sounded k, p, t only at the end of a word or before a final s.

[8] G, b, d are sounded k, p, t only at the end of a word or before a final s.

[9] Rr is generally distinguished from r, but there are a few examples of their confusion in rhyme.

[10] Ts is usually written c at the beginning of a word, z or tz at the end.

[11] G has the sound of tš only at the end of a word or combined with final z.

[12] Cf. French.

[13] Cf. Spanish.

[14] For some exceptions see Rom., XXXII, 591; P. Marchot, Phon., p. 9.

[15] Cf. R. Karch, Die nordfranzösischen Elemente im Altprovenzalischen, 1901.

[16] See § 175, (4).

[17] There is no diphthong in the preterit ending -ęc: cazęc, etc.

[18] This view is a modification of the theory developed by C. Voretzsch in his admirable treatise, Zur Geschichte der Diphthongierung im Altprovenzalischen, Halle, 1900. That ę is not affected by an i in the following syllable is shown by such words as empęri, evangęli, saltęri, which must have been adopted fairly early. The same thing is true of ǫ: apostǫli, ǫli, etc.

[19] The diphthong of ǫ occurs, however, in this text, v. 203, in uelŏculi.

[20] Derrier (derer, dereer), beside dereire, is manifestly due to the influence of primier. To the influence of the same ending -ier, as in carr(i)eira, is to be ascribed the diphthong in cad(i)eiracathĕdra.

[21] The things just said of ę are true of ǫ: there is no breaking before u < l (tǫut = tǫlt) nor before ts, dz, s, z (nŏcet > nǫtz, ✱nŏptias > nǫssas).

[22] The conditions are not quite the same as for e: an ę does not break before a labial (nęps) nor before n´ (vęnha). Breaking before g and k seems more general for ǫ than for ę.

[23] So the second person forms cuebres, uebres, uefres, and the third person forms cuebre, uebre, uefre; cf. cǫbron, ǫbri, etc.

[24] For a discussion of the date, see K. Nyrop, Grammaire historique de la langue française (Copenhagen, 1899-1903), I, § 187.

[25] For the accent, see § 16, 1.

[26] Spelled drictus: see Schuchardt, Vokalismus des Vulgärlateins, II, 422.

[27] The period of the fall of the intertonic vowel covers, in part, the period of the voicing of intervocalic surds (§ 65); sometimes the vowel fell too soon for the surd to be voiced, sometimes it did not. The relation of the fall of unstressed vowels to the development of intervocalic consonants, in French, has been examined by L. Clédat in the Revue de philologie française, in a series of articles beginning XVII, 122. Cf. P. Marchot, Phon., pp. 84-90.

[28] Cf. H. Wendel, Die Entwicklung der Nachtonvokale aus dem Lateinischen ins Provenzalische, 1906.

[29] Domnus may be the older form.

[30] The change of accent, in this verb and others, was due to the analogy of the first and fourth conjugations (cantęron, sentíron) and to the influence of the second person plural (dissętz).

[31] The feminine forms cobéza, tebéza, etc., show a change of accent.

[32] In most of the modern dialects (but not in Gascony and lower Languedoc) this a has become o: rosaroso. But in the Limousin dialects and some others -as > -a: rosasrosa.

[33] The tg in this word is probably due to the influence of iutiarjūdĭcāre.

[34] The forms with r may be due to dissimilation or to the influence of clergue.

[35] Alvernia is attested: cf. Zs., XXVI, 123. The usual form is Arvernicum.

[36] Compare, in English, the c of coo and the k of key.

[37] Compare the old-fashioned pronunciation of words like card, kind.

[38] For final -ci, -gi in plurals, see § 92, (2).

[39] Before this, frīgĭdus had become frįgdus in Italy and Gaul.

[40] It is natural to suppose that the n, in falling, nasalized the vowel; but no trace of this nasality remains.

[41] Fes, nut, which quite supplanted the regular forms, perhaps show the influence of res, mut.

[42] By analogy of such double forms, n is sometimes added to a few words ending in a vowel: fŭitfo fon, prōpro pron.

[43] Cf. § 63, (4).

[44] Also auvir, probably a northern local development of auir; and aurir, doubtless from auzir in a dialect that confuses r and z. See R, 2 and S, 2.

[45] Cf. § 63, (3).

[46] The i from ð fuses with the preceding i.

[47] Trachor has been influenced by trach, past participle of traire.

[48] Intervocalic c and g have been studied by H. Sabersky, Zur provenzalischen Lautlehre, 1888, pp. 8-19.

[49] Mica micha are from ✱micca = mīca + cīccum.

[50] Original Latin g seems more prone to fall than g < c.

[51] For the reduction of au to a see § 41.

[52] Clerc is from ✱clĕrcum, which must have existed contemporaneously with clĕrĭcum.

[53] Possibly it comes from N. Greek σράδη: cf. Romanische Forschungen, XV, 880.

[54] See Zs., XXXII, 434.

[55] Magis was probably reduced to mais in Vulgar Latin.

[56] Legir may have been reconstructed on the basis of leglĕgit.

[57] For the groups ending in y, cf. L. J. Juroszek, Ein Beitrag zur Geschichte der jotazierten Konsonanten in Frankreich, in Zs., XXVII, 550 ff. The groups ending in y and those containing c or g have been studied by H. Sabersky, Zur provenzalischen Lautlehre, 1888.

[58] S is generally written ss between vowels, to distinguish it from s = z.

[59] Most of the words in this category are semi-learned: cf. fabla and faula. See § 55, B.

[60] See § 47, (2).

[61] See § 47, (2).

[62] Also faur: cf. § 52, (1), 1.

[63] Sozersŏcĕrum: cf. § 49, (1).

[64] For the vowel of nęr nięr, see § 25, 1, (e).

[65] Enteir, neir seem to have lost final e under the influence of numerous adjectives in -er -ier -ieir-arium.

[66] In the modern dialects the d is probably commoner than it was in the old literary language; it occurs in Bordeaux, Languedoc, and Provence.

[67] Ratie is perhaps French.

[68] After o, the u disappears.

[69] Glai is due perhaps to the analogy of ney65, β, 3), perhaps to such double forms as fatz fai = facit.

[70] Seti (pronounced with two syllables) seems to be an improperly constructed post-verbal noun from assetiar. For metgue, etc., see § 65, G, (3).

[71] Apparently maniar, escomeniar developed in the region where g became y before a: cf. § 65, G, (1).

[72] The r remained palatal long enough to cause breaking: cf. §§ 30, 37.

[73] These sounds lost their palatal quality too early to cause breaking: cf. §§ 30, 37. Cf. Einf., § 133.

[74] Palaitz, however, is used by Marcabru, A. Daniel, and P. Vidal. Poizon occurs in Flamenca and in modern Limousin (beside pozon), raizo is found in the Boeci and other texts.

[75] Cf. § 49, (3).

[76] Daun, etc., are Gascon.

[77] Such forms as amdaamĭta, comde condecompŭtum, show a late fall of the medial vowel.

[78] The i in ueich seems to be merely graphic.

[79] The i in voig seems to be merely graphic.

[80] Fret, freda occur also.

[81] The phenomena of dissimilation have been well classified by M. Grammont in La dissimilation consonantique dans les langues indo-européennes et dans les langues romanes, 1895. For metathesis, see Zs., XXVIII, 1.

[82] Cato uses fructi.

[83] The process began in classic Latin: materies materia, etc.

[84] See § 89, 1.

[85] See § 47, (3).

[86] See § 16, 1.

[87] See § 52, (1), 1.

[88] Perhaps from a fusion of plūs and pluriōres = plūres. Cf. Fr.

[89] These forms existed as early as the 1st century of our era. See Zs., XXVI, 600, 619. Ejus, ei may have had some influence.

[90] See § 67, (2).

[91] § 45.

[92] § 74, (2).

[93] Cf. Meyer-Lübke, Gram., II, p. 104.

[94] See Suchier in Grundriss, I, p. 627.

[95] Cf. A. von Elsner, Ueber Form und Verwendung des Personalpronomens im Altprovenzalischen, 1886.

[96] Cf. Thomas in Rom., XII, 334; Meyer-Lübke in Gram., II, page 104. For a different explanation, see Ascoli in Archivio glottologico italiano, XV, 314, 396.

[97] For a different explanation of mia, see Gram., I, pp. 246-248, also Horning in Zs., XXV, 341. Cf. Fr. moie.

[98] Cels shows the influence of masculine nouns and adjectives.

[99] Aquel has also aquilli. Cilha is evidently a combination of cilh and celha.

[100] Aquel has also aqueli.

[101] Aquel has also aquelz and aquelses.

[102] See G. Rydberg, Le développement de facere dans les langues romanes, 1893.

[103] See A. Zimmermann in Zs., XXV, 735.

[104] See C. C. Rice in Publications of the Modern Language Association of America, XIX, 217.

[105] Cf. § 138.

[106] Cf. § 72, Sw.

[107] According to Raimon Vidal, a 13th century grammarian, tenir is French.

[108] Esvanuir seems to come from the perfect, evanuī.

[109] Enfolhetir shows the influence of follet.

[110] For two different explanations of this ending, see A. Thomas, Essais de philologie française, 25, 281; E. Herzog, in Bausteine zur romanischen Philologie, 481.

[111] See K. Sittl in Archiv für lateinische Lexikographie und Grammatik, I, 465.

[112] For ai estat we sometimes find soi agutz. The confusion arises perhaps from the use of both es and a in the sense of ‘there is’: hence es estat = a agut; and by a mixture of the two, es agut. Cf. L. Gauchat, Sone avuto, in Scritti vari di filologia (dedicated to E. Monaci), 1901, p. 61.

[113] Cf. P. Thielmann, Archiv für lateinische Lexikographie und Grammatik, II, 48 and 157.

[114] For the accented vowels in these forms, see §§ 167, 168.

[115] See § 168.

[116] See § 37.

[117] See § 30.

[118] Also, by analogy, iassér.

[119] Raimon Vidal says that crei, vei are the proper forms for the 1st pers. sg. of the pres. indicative.

[120] According to Raimon Vidal, trac is the only correct form.

[121] Volemus occurs repeatedly in 7th century Latin.

[122] Cf. O. Schmidt, Ueber die Endungen des Præsens im Altprovenzalischen, 1887.

[123] The loss of -s is not confined to the Provençal territory: it occurs also in western France, Catalonia, and the Engadine.

[124] Cf. the reduction of habēbam to aβea: § 153.

[125] Tenér tenír really belongs to the second conjugation.

[126] According to Raimon Vidal, this is the regular ending of the 3d pers. sg. of the fourth conjugation.

[127] In nasquec the ui ending occurs twice.

[128] Beside parẹc, coming perhaps from a V. L. ✱parēvitparēvuit.

[129] All verbs in -ndĕre took the perfect in -sī: ascos, defes, pris, respos, etc. Lĕgĕre took ✱lĕxīleis through the analogy of the p. p. lĕctum. So fĭngĕre took ✱fĭxīfeis through fĭctum; frangĕre, pĭngĕre, tangĕre did likewise (frais, peis, tais); and in Provençal cénhercĭngĕre, esténherexstĭnguĕre, plánherplangĕre followed the example of these (ceis, esteis, plais): hence all verbs in -nher have the preterit in -s.

[130] See Zs., XXVIII, 97.

[131] Gram., II, p. 357.

[132] Tĕnuī and vēnī influenced each other.

[133] So bĭbuitbec, debuitdec.

[134] So cognōvitconoc, mōvitmoc.

[135] So ✱cŏcuitcoc, jacuitiac, ✱nascuitnasc, ✱pa(s)cuitpac, placuitplac, tacuittac, ✱tescuittesc, ✱vĭncuitvenc, ✱vīscuitvisc.

[136] So caluitcalc, ✱tŏluittolc, vŏluitvolc.

[137] So ✱vēnuitvenc.

[138] So erĭpuitereup, recĭpuitreceup.