¹ REW³ §212.

² Grandgent, From Latin to Italian 31 (Cambridge, 1927; hereinafter referred to simply as ‘Grandgent’; other books by the same author will always be referred to by their own titles), with question mark. Meyer-Lübke (REW³ §6038a) fails to list it in this connection, but puts it under ūrere (§9081), again with no explanation of the curious phonetics. (Ever since Castelvetro it has been recognized that -r- > -r- is ‘lautgesetzlich’ in Italian; the -gg- in forms like chieggo = chiedo = quaerō, etc., is not phonological but analogical. Cf. Grandgent 122, Meyer-Lübke, Grammatica Storica della Lingua Italiana 177 [tr. Bartoli and Braun, Torino, 1931; hereinafter referred to as ML].)

³ E.g., as taken over from the first syllable of uggioso ‘disgusting’ < odiōsus, with dialectal development of o > u in pre-tonic syllable: Grandgent 31, 44; ML 40.

⁴ REW³ §8873.

⁵ Etymon suggested by De Gregorio, Studî Glottologici 8.303, but without explaining the -r-, whose presence it is our purpose to explain.

The opinion is ascribed to Coromines, Homenatje a Rubió y Lluch 3.283 (I have been unable to check this reference) that It. attraccare is from the Catalan. However, it is difficult to explain Italian [k:] from Catalan [k], whereas the Iberian forms on the other hand point to an original double consonant as the source for intervocalic [k]. In any case, the problem of the intrusive -r- remains the same.

⁶ Zimmermann, ALLG 5.567.

⁷ See REW³ §2795. For the various etymologies proposed, references, and discussion, cf. Wartburg, FEW 3.179; Walde-Hofmann, LEW 1.371, 381, s. v. dōnec, dunc. Dēnique as an etymon was suggested by Förster, RF 1.322–325, but without sufficient explanation of Rom. -- in place of --; cf. Paris' criticism of Förster, Romania 12.133.

⁸ Meyer-Lübke, Grammaire des Langues Romanes 3. §222, §485, §558 (New York, 1923, reprint; hereinafter referred to as Rom. Gramm.).

⁹ The examples cited by Meyer-Lübke (Rom. Gramm. 3.§558) from the Divina Commedia and elsewhere as showing temporal use of (a)dunque are all capable of interpretation as ‘deductive’ adverbial or conjunctive use. The passage in Inf. 22.64 (Lo Duca: 'Dunque or di '…') is punctuated differently by different editors; in some editions dunque, in others or, begins the words spoken by Vergil. For this type of use in the earliest OProv. also, cf. Boecis 110: e cum es velz, donc estai bona ment 'and when he is old, then (at last) he is well off'.

¹⁰ Rom. Gramm. 3.§483. According to the explanation offered above, therefore, Lat. dunque must have been prior in formation to dunc, not vice versa as assumed by Meyer-Lübke (Rom. Gramm. 2.§626, Ital. dunque and Fr. donc < dunc + que).

¹¹ Grandgent 25.

¹² ML 18, 44.

¹³ The other examples cited in support of this law (ML 44; followed by Grandgent 23) are also weak; mischia ‘mixes’ < ∗misculat, being a verb-form, is subject to analogical influences from forms with original -ī- such as the form mīxtus (> It. misto); vischio ‘bird-lime’ < visculum had original -i- (vīscus ‘bird-lime’, vīscāre 'to besmear'); biscia 'snake' directly < bēstia ‘beast’ is phonologically, as well as semantically, difficult, as recognized by Meyer-Lübke, REW³ §1061. Among other considerations, moreover, the parallel back vowel -- is not checked in its normal development by following -sk- or -st-, cf. angstia > angcia.

For biscia, cf. Tuttle, RR 6.344 (× vīpera, probably the best explanation). Rohlfs, ZRPh 41.354 f.; 52.74–5, contributes nothing new to the problem.

¹⁴ Grandgent 25.

¹⁵ REW³ §4263; cf. also Menéndez Pidal, Romania 29.357.

¹⁶ Meyer-Lübke and d'Ovidio, Gröbers Grundriss² 1.445, 464 (Strassburg, 1907); followed by Ernout, Les éléments dialectaux du vocabulaire latin 156 (Paris, 1909; for attestations of a Late Latin elex); REW³ §4259; Ernout-Meillet, Diet. étym. de la langue latine 452 (Paris, 1932); Rohlfs, ZRPh 46.158. The last-mentioned cites other examples of an assumed Osco-Umbrian ē for Latin ī in the South Italian dialects—which can all be explained, however, on other grounds. ∗Vēpera for vīpera (seen in Basilicata v) appears to be another case of reciprocal exchange of tonic vowel, cf. biscia with -ī- from vīpera (above, fn. 13).

¹⁷ Buck, A Grammar of Oscan and Umbrian² 44 f. (Boston, 1928). I am indebted to Professor Buck for his assistance in this connection.

¹⁸ LEW 1.678 (Heidelberg, 1936).

¹⁹ Studî di etimologia italiana e romanza §359 (Firenze, 1878).

²⁰ Revue de linguistique romane 1.122.

²¹ Grandgent 110.

²² Language 11.240.

²³ ZRPh 32.498; REW³ §5709.

²⁴ Moccio may have gotten its -čč- from the plural, as Meyer-Lübke suggests, or else may come from ∗mcceus (REW³ §5707). Related to the Italian words from this latter source (It. and It. dial, mozzo, etc. ‘servant-boy’, see Puşcariu, Jahresbericht des Instituts für rumänische Sprache zu Leipzig 11.108) are also Span, mozo, Port, moço ‘young; boy’ and Span, muchacho ‘boy’, rather than to musteus ‘musty’ and mutilus ‘mutilated’ respectively. Musteus in all other cases (REW³ §5779) gives words like It. moscio ‘flabby, soft’, etc., showing a different line of semantic development. Span, mozo, Port, moço are better understood as normal developments from ∗mcceus, with the same semantic change as that pointed out by Puşcariu for the It. forms, namely ‘Rotzbub > Junge > Knecht’ (It. moccichino, -one, etc.).

In deriving muchacho from mutilus, the difficulties are both semantic and phonological (as recognized by Meyer-Lübke, REW³ §5791). OSp. mochacho furnishes a key to the problem; we have here a formation on mozo with the suffix -acho (cf. Rom. Gramm. 2.§420), first ∗mozacho, then mochacho with assimilation of -z- to the following -č- (conversely, therefore, to the explanation offered in Rom. Gramm. l.§513).

²⁶ Grandgent 28.

²⁷ Studî §459.

²⁸ REW³ §6877.

²⁹ REW³ §209.

³⁰ Arch. Gl. It. 16.430.

³¹ Grandgent 105.

³² A dialectal development; cf. Grandgent, Introduction to Vulgar Latin 119 (Boston, 1907).

³³ Studî §292.

³⁴ REW³ §3795; cf. also Menéndez Pidal, Romania 29.344, on Span, en cuclillas. Direct connection with Lat. glōc- is particularly unlikely in view of the disparity between the initial consonants.

³⁵ Nòvo dizionario universale della lingua italiana 1.494 (Milano, 1894).

³⁶ REW³ §2059.

³⁷ Studî §362.

³⁸ ZRPh 30.300.

³⁹ REW³ §4414.

⁴⁰ Ascoli, Miscellanea Caix-Canello 428 (Firenze, 1886); cf. also Ernout, Éléments dialectaux 131; Meyer-Lübke, REW³ §1355, both of whom consider the second element of bubulcus as unexplained.

⁴¹ Grandgent 42.

⁴² REW³ §116.

⁴³ Vocabolario della lingua italiana 835 (Milano, 1917).

⁴⁴ Lea, Italian Popular Comedy 230 f. (Oxford, 1934).

⁴⁵ For example, in the list of members of the company of the ‘Gelosi’ (Lea 261 f.), of the eleven members of the company, seven came from northern Italy, three from Bologna, and one from Florence.

⁴⁶ Quoted by Lea (66) from Riccoboni, Histoire du Théâtre Italien.

I am indebted to Mr. Harold M. Barnes, Jr., for his assistance in checking references and for valuable suggestions.