¹ From W. Meyer-Lübke, Grammatik der romanischen Sprachen 1.§§404–31 (Leipzig, 1890) to H. Lausberg, Romanische Sprachwissenschaft 2.5–29, §§299–539 (Berlin, 1956), and in virtually every historical grammar of the individual Romance languages.

² I have used this expression here and elsewhere (e.g. in ‘The Papal States in Italian linguistic history’, Lg. 19.125–40 [1943]) as a useful cover term, although it by no means implies that all WRomance dialects have undergone this shift, nor that it characterized the entire WRomance territory from earliest times.

³ E.g. E. Cross, ‘The Latin initial syllable in the Romance languages’, PMLA 49.994–1004 (1934), especially ‘It is a matter of common observation and knowledge that initial vowels, consonants, and syllables, even when in unaccented position, suffer far less loss and change than do final vowels, consonants, and syllables’ (994).

⁴ E.g. R. G. Kent, ‘The alleged strength of the initial syllable’, Lg. 7.179–87 (1931); ‘Again the initial syllable’, Lg. 11.151–2 (1935).

⁵ E.g. F. Jungemann, La teoría del substrato y los dialectos hispano-romances y gascones 196–7 (Madrid, 1956).

⁶ Cf. the remarks of H. Weinrich, Phonologische Studien zur romanischen Sprachgeschichte 63–4, §74 (Münster/Westfalen, 1958), especially ‘Es ist weder die labile Wortgrenze die Folge eines “acento de intensidad no demarcativo débil” noch die stabile Wortgrenze die Folge eines ich weiβ nicht wie gearteteten Intensitätsakzents, wie F. Jungemann mit Martinet annimmt …‘ (64).

⁷ E.g. Weinrich, Phonologische Studien 64–5, §75; 73–4, §85.

⁸ As suggested by F. Diez, Etymologisches Wörterbuch der romanischen Sprachen ⁴ 64–5 (Bonn, 1878), and accepted by R. Menéndez Pidal, Orígenes del español ³ 332 (Madrid, 1950), but rejected by others, including J. Corominas, Diccionario crítico-etimológico de la lengua castellana 1.512–3 (Madrid, 1954), preferring /bárbaru/ as etymon.

⁹ Cf. M. L. Wagner, Historische Lautlehre des Sardischen 206, §375 (Halle/Saale, 1941; ZRPh. Beiheft no. 93).

¹⁰ Apparently the only mention of syntactic doubling that has been made hitherto in this connection is by A. Martinet (Économie des changements phonétiques 281–5, §§11.30–6 [Bern, 1956]), but only with reference to the development of initial /r- n- 1-/ in Ibero-Romance.

¹¹ The literature on ‘raddoppiamento sintattico’ is extensive; cf. the references given in my Bibliografia della linguistica italiana ² §§3134–8 (Firenze, 1958), and also A. Leone, ‘A proposito del raddoppiamento sintattico’, Bollettino del Centro di studî filologici e linguistici siciliani 7.163–70 (1962).

¹² As suggested by A. Camilli, Pronuncia e grafia dell'italiano ² 122–3 (Firenze, 1947; Biblioteca di Lingua nostra no. 2).

¹³ As in my article ‘Italian inflection’, Lg. 20.11–21 (1944), and in my Descriptive Italian grammar (Ithaca, N.Y., 1948), especially 15–6, §1.622.1.

¹⁴ As in my textbook Italian for modern living (Ithaca, N.Y., 1959; reedition Philadelphia, 1961).

¹⁵ M. L. Wagner, Historische Lautlehre des Sardischen, devotes no separate section to syntactic doubling, but instances of it are quoted passim, especially 196–205, §§349–73, under ‘Die Auslautkonsonanten’.

¹⁶ Beginning with sao cco kelle terre … in the ‘Formula testimoniale di Sessa’ of March 963, reprinted in virtually all anthologies of Old Italian.

¹⁷ E.g. e ccando, e ffecila in the ‘Privilegio logudorese’ (ca. 1080–5), given in A. Monteverdi, Manuale d'avviamento agli studî romanzi 145–6 (Milano, 1952).

¹⁸ Cf. R. G. Kent, The sounds of Latin ³ 152, §192 (Baltimore, 1945), or any complete manual of Latin grammar, e.g. Allen and Greenough, Bennett, Gildersleeve, Hale and Buck.

¹⁹ As pointed out in my article ‘The reconstruction of Proto-Romance’, Lg. 26.6–27 (1950), Sardinian belongs in a totally different branch of Romance from the other subdivisions of the family, and the presence of a nonborrowed phenomenon in Sardinian and one or more other branches of Romance is justification for our ascribing it to Proto-Romance; on the other hand, of course, a later and probably borrowed development in Sardinian such as the West Romance sound-shift is not evidence for the state of affairs in Proto-Romance.

²⁰ When we speak of intermediate proto-stages, we are of course saying nothing about where they may have been spoken or what their geographical extent may have been. That such intermediate stages existed, however, is assumed in practice by all scholars, even when they deny it in theory.

²¹ For the justification of ascribing the contrast C ~ CC to at least the Proto-Gallo-Romance and Proto-Ibero-Romance stages, cf. the discussion in Lg. 26.13–4 (1950). When A. Martinet states (Économie des changements phonétiques 279) ‘nous n'y [in Gallo-Romance] trouvons aucune trace directe des oppositions -n-/-nn- et -1-/-ll-‘, he is obviously wrong.

²² As printed in R. T. Hill and T. G. Bergin, Anthology of the Provençal troubadours 111–2 (New Haven, 1941). This passage was chosen for its easy narrative style and probable closeness to everyday speech.

²³ My assumptions concerning pausai junctures were inevitably somewhat arbitrary; I set up pauses wherever the editors had supplied commas, and in other places where it seemed reasonable to assume that a speaker would have ended a moderately long breath-group, neither breaking the utterance into very small fragments nor continuing for a very long time.

²⁴ Weinrich, Phonologische Stadien 63, §75.

²⁵ Weinrich, ibid. 73–4, §86.

²⁶ As does, for instance, G. Rohlfs, IF 68.306–7 (1963).

²⁷ Romanische Sprachwissenschaft 2.93, §579.

²⁸ As suggested by R. L. Politzer for central Italian lenition, ‘Latin lenition and Italian consonants’, General linguistics 1.70–8 (1955), especially 71–3, with a very vague concept of ‘early Latin lenition’.

²⁹ Cf. M. L. Wagner, Historische Lautlehre des Sardischen 68, §§101–2.