¹ Not that the phenomena are found only in compound verbs, but because the demonstration of them to scholars in other linguistic fields is simpler if such examples be used.

² Cf. especially F. Sommer, Hdb. d. lateinischen Laut- und Formenlehre² 83–115 (1914); K. Brugmann, Grundriss d. vergl. Sprachwiss. d. indog. Sprachen ¹l. 548–51 (1886), ²1. 2. 971–2 (1897), also ²1. 1. 213–33 (1897).

³ Literature in Stolz-Schmalz Lat. Gram. (Müller-Otto Handbuch der Alter-tumswiss. 2. 2.), fifth edition, revised by M. Leumann and J. B. Hofmann, 186–7 (1926); F. Sommer, op. cit. 87.

⁴ Cf. especially A. Meillet et J. Vendryes, Traité de Grammaire comparée des Langues classiques (1924) : §164, Les voyelles brèves sont souvent instables. En syllabe initiale, elles sont relativement résistantes et ne subissent d'autres altérations que celles qui proviennent de l'action des phonèmes voisins. En syllabe intérieure, elles tendaient à s'abréger; il est résulté de cette tendance que la plupart se sont fermées et que quelques-unes méme ont disparu (phénomène dite de syncope)…. §176. La difference de traitement entre les voyelles brèves suivant qu'elles sont initiales ou intérieures, suppose aux syllabes initiales un caractère spécial, qui se reconnaît dans l'emploi que fait la langue de l'alliteration. (This suggests that the ‘special character’ is the possession of an accent of energy; for the Germanic alliterative verse employs alliteration of accented syllables only!) Cf. also §189, §190; and A. Meillet Esquisse d'une Histoire de la Langue latine 133–5 (1928), without any interpretative theory on the difference in the development of short vowels and diphthongs, according as they were in initial syllables or in non-initial syllables.

⁵ This, in connection with the changes in Latin, is taken as indicating that ‘weakening’ is in the direction of shortness, and therefore in the direction of e to i, and of o to u. But such is not always the case, cf. OHG bilîban, MHG belîben.

⁶ His conclusions are to be found in Rivista Indo-greco-italica 7. 65–9 (1923), from which the quotations are made. For a complete bibliography of his articles on the subject, cf. J. Marouzeau, Dix Années de Bibliographie classique 1914–24, 2. 555–6.

⁷ The words quoted are documented by references to G. Körting, Lateinischromanisches Wörterbuch³, Paderborn 1907, and W. Meyer-Lübke, Romanisches etymologisches Wörterbuch², Heidelberg 1924. It seems needless in most instances to gloss the words, since the semantics are not in question here; nor is it worth while to mark the vowel quantities.

⁸ In most instances a familiar word which retains the initial syllable is given for comparison.

⁹ To be found in Tommaseo e Bellini, Dizionario della Lingua Italiana (1865); Petrocchi, Novo Dizionario Universale della Lingua Italiana, 1887; Pianigiani, Vocabulario Etimologico della Lingua Italiana, 1907.

¹⁰ It is true that most of the Romanic examples may be explained as losing the initial syllable by wrong division after the article or after the preposition di or de; but this will not account for the loss in Fr. briller, dresser, droit, and their cognates. Other unquestionable examples might be added.

¹¹ In these words, and in others to be cited later, it has no importance for the purpose at hand at what date the loss took place, nor in what language or languages; the only feature of importance is the difference between the start-form and the final form, which lacks the initial syllable of the original.

¹² A. Leskien, Litauisches Lesebuch 250.

¹³ Except those from minor Romanic dialects.

¹⁴ Quite similar is the formation of personal nicknames by the omission of syllables before the main accent: thus (in various languages) Zita for Carmencíta, Etta for Margarétta, Gretchen for Margarétchen, Lizzie and Lizbet for Elízabeth, Sepp for Giuséppe, Hans for Johánnes, Milly for Amélia, Retta for Lorétta, Nurnie for Minérva, Viah for Zevíah, (Greek) Nasos for Athanásios. (The last few are perhaps not standard, but are personally known to the writer.)

¹⁵ App. Probi 4. 198. 5–6, 4. 199. 8–9 Keil.

¹⁶ Such an accent in no way conflicts with an accent of stress or energy, which may be present in the same word on the same syllable as the pitch-accent, or separately on another syllable. If such a stress-accent was present in ancient Greek, it was only as the accent which gave the rhythm of verse—and I believe that Greek verse did have such an accent—without interfering with the pitch-accent of the words.

¹⁷ It should be remembered that in Modern Greek the voiced stops β δ γ have become voiced spirants, and that the voiceless stops π τ κ have become voiced stops when preceded by a nasal; further, that the voiced stops [b d g] can be graphically represented only μπ ντ γκ.

¹⁸ In continuous discourse this whole word would be proclitic, i.e., unaccented, and the initial syllable therefore subject to loss.

¹⁹ Aperçu d'une Histoire de la Langue grecque ³268 (1930) = ¹296 (1913): ‘A cet état du grec ancien s'oppose celui du grec moderne; le grec moderne n'a pas de voyelles longues ou brèves par elles-mêmes. Sont longues les voyelles intérieures frappées de l'accent; sont brèves les voyelles intérieures inaccentuées. Les voyelles toniques sont aujourd'hui encore prononcées avec une élévation de la voix: l'accent de hauteur n'a donc pas disparu. Il n'y a pas d'intensité fortement marquée, ni d'effets marqués de l'intensité; car les alterations et les chutes de voyelles qu'on observe dans les parlers modernes, surtout dans la région septentrionale, s'expliquent par des différences de quantité plus que par des différences d'intensité.‘ Cf. also ³317 = ¹358.

²⁰ Cf. Meillet-Vendryes, op. cit. §189, §190.