(2.) Given the area of distribution of the cognates employed in the reconstruction of Proto-Rom. */molˈton‑e/ (N. It. Friul. Lad. Fr. Frpr. Occit. Gasc. Cat.), the etymon cannot be reconstructed for Common Proto-Romance (or Proto-Romance stricto sensu), nor for Continental Proto-Romance, but only for the more recent stratum of Italo-Western Proto-Romance, subsequent to the separation of Proto-Sardinian (2th century [?], cf. Straka,RLiR 20, 256) and of Proto-Romanian (3th century [?], cf. RosettiIstoria 184; Straka,RLiR 20, 258).

(3.1.) The phonological reconstruction of the etymon brings up interesting questions concerning two issues. On the one hand, on the basis of the initial vowels It. /o/ ~ Friul. /o/ ~ Lad. /u/ ~ Fr. /u/ ~ Frpr. /u/ ~ Occit. /u/ ~ Gasc. /u/ ~ Cat. /o/ (~ /u/), one could hesitate between reconstructing */o/ and */ʊ/ (cf. Gouvert,DÉRom 1, 74). Yet, as we have seen (above 2.), the etymon pertains to Italo-Western Proto-Romance (often called Vulgar Latin in traditional Romanist studies), a language stage marked by the merger of */o/ and */ʊ/ (> */o/, cf. LausbergSprachwissenschaft 1, § 156 [‛das sog. ‛vulgärlateinische‘ System‘]). Thus we propose */‑o‑/, conscious of the fact that our analysis of this matter differs from the one put forward by REW3 (‛*mŭlto‘) and by FEW 6/3, 205b (‛*multo‘).

(3.2.) Furthermore, it should be noted that an exclusively phonological reconstruction, based on the cognate set listed in the data section above, would lead to Proto-Rom. **/molˈton/. The more realistic etymon */molˈton‑e/ that we adopt is the result of a confrontation between phonological and morphological reconstruction. Indeed, the common ancestor of these cognates, all masculine nouns, pertains of necessity to the same part of speech, and thus contains a thematic vowel (cf. HallMorphology 20-27): in general */‑u/ (e.g. */ˈɡɛner‑u/) or */‑e/ (e.g. */ˈɸratr‑e/), rarely also */‑a/ (*/ˈbarb‑a/²). Whereas */‑a/ can be ruled out on purely phonetic grounds (*/‑a/ persists in several of the Romance varieties at issue, cf. LausbergLinguistica 1, § 272-273), it is not possible to decide between the hypotheses */‑u/ and */‑e/ on this basis, for all relevant languages present systematic (or, in the case of Catalan, frequent) deletion both of final */‑u/ and */‑e/ (cf. LausbergLinguistica 1, § 273-274). It was thus necessary to make use of the additional testimony of It. montone, whose /-n-/ is due to an Idio-Romance (i.e. local) innovation (cf. footn. 1), but which is not suspected of having undergone a remorphologisation process, and thus leads to reconstructing the thematic vowel */‑e/, hence */molˈton‑e/.

(4.1.) Comparative reconstruction leads to consider Proto-Rom. */molˈton‑e/ polysemous and to ascribe to it the senses “ram” (above I.) and “wether” (above II.). As these two sememes are reconstructed on the basis of lexemes belonging to Italo-Western Romance languages, they have to be attributed to the same stratum of the proto-language, and it is not possible, on this sole basis, to decide which one is original. Nevertheless, it seems possible, on the basis of areological, semantic-motivational, and documentary considerations, to chronologise the appearance of the two meanings. The areological criterion is perhaps the least powerful: the two relevant areas are almost identical, Friulian, the easternmost tip of the area covered by the reflexes of */molˈton‑e/, being the only language of the “ram”-area not covered by the “wether”-area. Deciding whether “ram” or “wether” is the original meaning is thus not an easy task. The Romanist tradition, which generally puts forward a monosemic etymon, avoids the problem: Meyer-Lübke in REW3 defines “wether” (‛Hammel‘), whereas Müller in FEW 6/3, 209a opts, on the basis of the Gallo-Romance data and Medieval Latin attestations, for the undifferentiated sense “male sheep”, a generic sememe that one can indeed adopt for the more recent stratum of the proto-language where “ram” and “wether” coexist. If we accept Gaulish as the remote origin of the etymon (see below 5.2.), semantic-motivational considerations tend however to give credit to the hypothesis of the anteriority of “ram”. Indeed, a substratum borrowing is particularly plausible for the denomination of the ram, on whom the survival of the flock relies, and whose name, used daily by the indigenous livestock farmers, has thus a particular importance (cf. Pellegrini,AAA 86, 237 and */arˈiet‑e/, section 4.1. of the commentary). Consequently we propose the order “ram” > “wether”. If the more recent sense has not completely ousted the more ancient one in the Romance dialects, the older one is nevertheless the more residual: “ram” has disappeared from contemporary French and Occitan and is retained in only four Catalan locations. It can be assumed that the documentary situation that is observable for Romance represents the last stage of the process of semantic innovation that ensued in Proto-Romance.

(4.2.) As for Fr. mouton m.n. “farm animal with thick wool that eats grass and is kept for its wool and meat (Ovis aries L.), sheep” (since 1188, MatsumuraDictionnaire; WartburgSchaf 4-5, 30; TL; Bloch/Wartburg5; FEW 6/3, 205b-206a; ANDEl; DEAFPré s.v. mouton; DAOSuppl n° 1317; ALF 886), Occit. ˹[molˈto]˺ (since ca. 1100 [.l. moltos], DAOSuppl n° 1317; DAO n° 1317; Raynouard; AppelChrestomathie; Pansier 3; BrunelChartes; BrunelChartesSuppl; Mistral; FEW 6/3, 206a; DOMEl; ALF 886; ALLOr 511* [general])⁷, Gasc. ˹moutoû˺ (since 11th c. [moton], DAG n° 1317; Palay; RohlfsGascon2 158; FEW 6/3, 206a; ALF 886), they do not seem to form a cognate set whose common ancestor could be reconstructed. In our opinion, these lexemes present, on the contrary, a sememe developped at a (preliterary) Idio-Romance stage, in all likelihood in French, from where it would then spread into Occitan and Gascon (by the way, the beginning of this same borrowing process can be observed in contemporary Francoprovençal)⁸. This generic term belongs to the vocabulary of the clients, not of the breeders: male sheep are slaughtered as soon as they reach adulthood, whereas ewes are kept alive as long as they can lamb, so that the consumed product is mostly made up of wether (see Fr. bœuf m.n. “male castrated bovid; edible flesh of oxen, cows, and bulls” and ALLOc 514, 515 “ovins according to age”: masculine and feminine denominations are general for one-year-old ovins, masculine denominations become less frequent for two-years-old ovins, extremely rare for three-years-olds, and exceptional for four-years-olds).

(5.1.) One finds lexemes that are phonologically and semantically close to Proto-Rom. */molˈton‑e/ in Celtic languages: O. Ir. Gael. molt m.n. “ram; wether; sheep” (PedersenKeltisch 1, 137; MacLennanGaelic; LEIA M-62; IEEDCeltic), O. Welsh mollt “ram; sheep” (PedersenKeltisch 1, 137; DelamarreDictionnaire2 227; IEEDCeltic), O. Bret. mout “ram” (DelamarreDictionnaire2 227; IEEDCeltic), Bret. maout “sheep” (HenryBreton; cf. PedersenKeltisch 1, 137), O. Corn. mols “wether” (PedersenKeltisch 1, 137; IEEDCeltic). These items have been identified by Celticists as cognates whose common ancestor is Proto-Celt. */molto-/ m.n. “ram; wether” (Pokorny; IEEDCeltic; cf. Holder s.v. *molto-s).

(5.2.) In the absence of an obvious etymology within Celtic for these items (‛eine überzeugende celt. wurzel fehlt‘), Diez s.v. montone – see also IEEDCeltic s.v. *molto-: ‛a convincing IE etymology is lacking‘ – proposed to attribute the origin of this Romance and Celtic lexical family to Lat. mutilus adj. “mutilated” (cf. REW3 s.v. mŭtĭlus). A variant of this etymology can be found subsequently in Körting in LRW3 (‛*mŭlto, -ōnem m. [aus *mulito für *mutilo v. mutilus?]‘), who in turn passed on to LothBrittoniques 188 the idea that the Celtic lexemes are borrowings from Latin. Yet this etymology identifies “mutilated” as the original seme within the Romance cognate set and thus presupposes the precedence of the sememe “wether” over “ram”, a hypothesis that is ill matched with the documentary evidence, which assigns to “ram” the status of a residual item in relation to its expanding competitor (see above 4.1.). What is more, the area covered by the Romance cognates (N. It. Friul. Lad. Fr. Frpr. Occit. Gasc. Cat.) is almost identical to the area of Gaulish settlement (with an extension into the Catalan region), a fact that can hardly be explained away as a coincidence: why would the oral correlate of Lat. mutilus, whose reconstruction, by the way, is anything but straigthforward⁹, have developped the meaning “wether” precisely on Gaulish territory, but not elsewhere in the Roman Empire? In our opinion, the area covered by the Romance cognates incites, on the contrary, to postulate, with Gröber,ALL 4, 127¹⁰, that it is Proto-Rom. */molˈton‑e/ that is a borrowing from Gaulish¹¹.

(5.3.) There is no written testimony of the presumed Gaulish etymon; it is reconstructed, on the basis of the Celtic cognate series (see above 5.1.), the Romance data and Latin personal names, as [*]multo (BillyThesaurus 110; see also GrzegaKeltizismen 201), *multō (LambertGaulois 200; see also Bolelli,ID 18a, 56), [*]molton- (DelamarreDictionnaire2 227) or, with an inappropriate glottonym, ‛Gallo-Lat. *multo‘ (IEEDCeltic)¹², and its meaning is given as “ram” (DelamarreDictionnaire2 227)¹³ or, less convincingly, as “sheep” (BillyThesaurus 110) or “wether; male animal” (‛Hammel/[männliches Tier]‘, GrzegaKeltizismen 210). Due to the uncertainty surrounding the exact meaning of the Gaulish lexeme, it is difficult to decide which part of the semantics of Proto-Rom. */molˈton‑e/ m.n. “ram; wether” goes back to the etymon and which one might potentially be attributed to Proto-Romance innovation; at most, one can observe that the senses in question (“ram; wether”) are both present in the Celtic cognates of the etymon¹⁴.

(5.4.) During the borrowing process, the Gaulish lexeme, whatever its exact phonetism, was integrated into the Proto-Romance inflectional class in */‑ˈon‑e/ (cf. for instance */tiˈtion‑e/), where the sequence */‑ˈon/ is part of the accusative stem. AlessioLexicon suggests that this morphological adaptation is due, as in */kaˈpron‑e/ m.n. “male of the goat, billy goat” (< */ˈkapr‑a/), to analogy with Lat. petrō, -ōnis “ram” (since Plautus [* ca. 254 – † 184], rare, TLL 10/1-2, 1982), but, as this lexeme does not have a reconstructable correlate in the oral code (Ø REW3; Ø FEW), this hypothesis has to be discarded.

(6.) There is no correlate of */molˈton‑e/ in the written Latin of Antiquity¹⁵. This leads to the conclusion, from a diasystemic point of view, that it pertained to the (spoken) varieties of immediate communication of global Latin and did not have access to the (mostly written) varieties of language of distance, in which Lat. aries m.n. “ram” (cf. */arˈiet‑e/, section 5 of the commentary) and ueruex m.n. “wether; sheep” (since Plautus, IEEDLatin; cf. */βerˈβek‑e/) dominated.

(7.) For further information cf. */arˈiet‑e/