Page images

this rarity is the great tendency to fusion of two vowels when only separated by a v. (See preceding paragraph, and comp. Schleicher, Deutsche Sprache, p. 159, ed. 2).

7. Consonantal v is never found before a consonant (Prisc. 1. 23) or final; but always before a vowel. This is quite as it would be if v be equal to w; for w scarcely gains any consonantal power, if indeed it be not absolutely unpronounceable, except before a vowel ; but v is as pronounceable after as before a vowel. Thus sive (older seive), neve when they drop the final e become seu, neu, not siv, nev1. Compare this with Italian, where (the labio-dental) v is frequent before a consonant in the middle of a word; e.g. avro (habebo), courire (cooperire), &c.

8. The English name of the labio-dental voiced fricative is vee. This name is derived from vau, the term applied to the digamma, with which the Latin f, on account of its symbol Ƒ, and the Latin consonantal u, on account of its sound, were identified (cf. Quint. XII. 10. § 29). But in classical times, at any rate, v consonant and v vowel (like i consonant and i vowel) were not distinguished either in symbol or name. Nor were they by Terentianus Maurus. Priscian (1.20) speaks of the name vau being given it from its resemblance to the digamma. But had the sound of English v belonged to it, at the time when the other letters received their name, it would have been called ev. For it is the law of Roman nomenclature2 to denote vowels by their sounds, mute consonants by sounding a vowel after them, be, ce, de, ge, &c.; continuous consonants by a vowel before them (e.g. ef, el, em, en, er, es), probably because in this way each consonant gets its fullest and most characteristic sound (Prisc. 1. 8); the explosives being chiefly distinguishable when they precede a vowel (§ 274), the continuous consonants having when final an opportunity of being prolonged at pleasure. Varro is said to have given va as the name and sound

1 Marius Victorinus (p. 2465), stands alone, I believe, in thinking that obverto, obvius should be ovverto, ovvius.

2 The names of all the letters are given in Pompei. Comm. ad Donat. Vol. v. p. 101 Keil. Cf. also Serg. IV. p. 478. I cannot bring myself to believe that Mr C. B. Cayley, Philol. Soc. Trans. for 1870, pp. 5-16 (the only paper which I have ever seen on the question of the names of the letters), is right in thinking that the Latin names have not been assigned on phonetic principles. Comp. App. A. xxiii.

of the digamma. If the Romans had named their consonantal use of u, they would have denoted it similarly by va or ve (pronounced wa, we), as w like h only obtains its full sound before a vowel.

9. The labio-dental f differs from the labio-dental v only as p from b, t from d, s from z, th (in thin) from th (in then), &c.; i.e. the former is whispered, the latter is voiced. The Saxons and (formerly at least) Welshmen do not notice the difference, or rather they sound the voiced consonants as the voiceless (e. g. pet for bed); we give to each of the symbols, s and th, both the sounds. With so great similarity between f and v is it likely that the Romans, if their v was a labio-dental, would not have confused them or noticed the resemblance? Yet (a) no inscription substitutes F for v (Corssen, Aussprache, 1. p. 136); and (b) the Roman writers (at any rate before the 4th century1) seem not to have noticed this close resemblance, although (as was said before) the symbol F was the ordinary symbol of f, and was borrowed from the digamma to which the Roman v corresponded. Quintilian's description (XII. 10, § 29) of the Roman f indicates strongly its dental and voiceless character. I am inclined to think that no more is meant by his words than 'blown out between the intervals of the teeth with no sound of the voice2.' In the next sentence he speaks of the Æolic letter which we utter in seruum, ceruum,' but seems in no way conscious of any close similarity of it to f. Terentianus Maurus (v. 227) describes f quite correctly as uttered with a gentle breathing while the under lip is pressed against the upper teeth,' and speaks of v consonant at considerable length, but never suggests any resemblance to f.

IO. The ordinary and regular mode of expressing the Latin v in Greek is by ov3, and no distinction is made whether it be a vowel

1 Marius Victorinus (p. 2465), speaks of the 'cognate letters b, f, m, p, u,' which is of course in some sort correct on any supposition.

2 Some think that a still harsher articulation than the ordinary English f is here meant, and no doubt this is possible enough, but, considering that Quintilian regards it as quite peculiar, some emphasis of expression is not unnatural. Even in English f and v are different enough from any other consonants.

3 The sign 8 (originally a T put with its foot in the middle of the o) is not found in inscriptions or coins till the end of the second century p. Chr. (Franz, Elem. Epigraph. Græc. p. 246).

or consonant.

On the other hand, Latin v is never used in the transcription of a Greek word, except as a vowel, usually for o or ov (cf. § 90. ii.).

But Latin v consonant is sometimes expressed in Greek by o, and sometimes by B. Now o was an occasional descendant from a digamma (cf. § 91, and Curt. Gr. Etym. II. 145 = 500, ed. 2), and is certainly, next to ou, the nearest vowel sound to the Latin u. This use of o therefore tends to confirm the inference which may be drawn from the use of ov, viz. that Latin v consonant was the consonantal sound nearest to the vowel u; and that is Engl. w.

The expression of the Latin v consonant by ẞ is one of the main arguments upon which the theory, which makes Latin_v= English v, rests. The argument proceeds, as I understand, thus: 'Greek ẞ either had the sound of Engl. v, or, if not, it had a 'sound, say b, nearer to v than to w. And it is probable that Greek '6 had the sound of Engl. v, for it has this sound in modern 'Greek.'

Now the extent to which ẞ was used to represent Latin ▾ is commonly taken to have been much greater than it really was, Nothing but an undoubting acquiescence in an accredited belief could have caused so vigilant and industrious a philologer as Corssen to treat the question in the superficial way which he has done (Aussprache, I. 311, ed. 2). He gives no authority for the instances in which v in proper names is represented by ẞ, and he quotes, as instances of the same in words which are not proper names, two only from inscriptions (date not specified; they are from Lycia), three from Suidas, and four from Lydus. Lydus was a Byzantine, and not born before A.D. 490; Suidas is later, and indeed is often put as late as the 11th or 12th century p. Chr. Both therefore are witnesses of little weight in such a question; and when we remember that in the 4th century p. Chr, there was a frequent confusion between Latin v and Latin b (which began as early as the 2nd century but not before1), we see that the use by any writers later

1 See § 72. Corssen, Aussprache, I. 131 sq.; Schuchardt, I. 131; Göschen's Pref. to Gaius, p. xxxxii. ed. Lachmann; and Naber's edit. of Fronto, passim. So Priscian (Part. 23=III. 465, Keil) makes the strange statement, that 'all words beginning with vi are written with v, 'except bitumen, bilis and the compounds of bis.'

than the 4th century of a ẞ for v is no evidence whatever of the sound of ▾ in the age of Cicero or of Quintilian.

The Greek writers of most importance for this matter are Polybius (2nd cent. B.C.), Diodorus Siculus (1st cent. B.C.), Dionysius of Halicarnassus and Strabo (Augustan age), Josephus and Plutarch (latter half of 1st cent. p. Chr.), Appian (middle of 2nd cent. p. Chr.), Dio Cassius (end of 2nd or beginning of 3rd cent. p. Chr.). I have examined these attentively, though not exhaustively, and collected a large number of instances of transcription of Latin words, principally proper names. I have since examined Benseler's most painstaking dictionary of Greek proper names, and the result is in both cases the same; viz. that, except in one writer, the instances of v consonant being represented by ẞ are few absolutely, and very few relatively to the instances of its being represented by ov. The one exception is Plutarch, and, so far as I have noticed, most instances commonly quoted have or might have been taken from He has ẞ for v frequently, though not as often as he has ov. The same name appears with ẞ in some of his Lives, in others with ou. Other names are always written one way. But this matter has been so little noticed that some details may be interesting. I have looked particularly through (1) all Plutarch's lives of Romans, and that of Pyrrhus (in Sintenis' edit., Teubner series); (2) the first five books of Polybius (Hultsch's edit.), i.e. all that is preserved in a continuous narrative; and (3) Books IV.—VI. of Dionysius of Halicarnassus (in Kiessling's edit., which in these books rests on a better collation of the most important MS. than in the first three). The result is as follows, the numbers being possibly not strictly accurate, but at any rate accurate enough for the present purpose1.


(1) In Plutarch there are of names of persons (almost all Romans), or places, or peoples, 50 written with and with B;



and the occurrences of these names are, in all, 323 with ou, 180 with B. Of these Valerius, Valeria, Valens, Ventidius, Verginius, Vespasianus, Vibius, Vindicius, Vinius, Vitellius, Volsci occur at least 5 times each (Valerius and Volsci nearly 50 times each), and always

1 I have not included instances where neither ou nor ẞ are used, eg. in Plutarch, Φαώνιος, Νοέμβριος, Σκαιόλας : nor instances of u after q (cf. § 90, 2); though both these speak for a light value being given to v.

with ou; Fulvius, Fulvia, Varro, Verres occur at least 8 times each, and always with B. Others, e. g. Veii, occur both with ov and ß; Volumnius (in Brutus) always with B, Volumnia (in Coriolanus) always with ou; Octavius 16 times (chiefly in Crassus and Pompeius) with ov, 30 times (chiefly in Gracchi and Marcellus) with ẞ; but Octavia (in Antony) 22 times with ou, and only twice (in Marcellus) with ẞ; Servilius 9 times with ou, twice with ß; Servilia once with ov, 14 times with B. Yet other writers have ov in the names which Plutarch writes with ẞ only. For instance, no one else (according to Benseler's Lex.) writes Bappwv (except once Dionys. Hal. I. 14) or Béppηs.

(2) In the first three books of Polybius I find- 10 names, making in all 20 occurrences, all with ov; not a single instance of B. In the 4th and 5th books I find no instance of either. On turning to the extracts from Polybius' lost books I find nothing in those from the 6th and 7th; but in the 8th Ovaλépios once, Λίβιος four times.

(3) In Books IV. to VI. of Dionysius I find 21 names written with ou (besides Avevrîvos), and the occurrences are 184, Valerius, Volsci, and Servilius being exceedingly frequent. There are 5 names only in which v is represented by ẞ; Nævius, Flavus (written in the two best MSS. pλaßios), Servius, Pulvillus, and Elva, the last only occurring twice, the others once.

How much of this comparative frequency of ß in Plutarch is due to the author, how much to his copyists, how much to his editors, I do not know. The text of Polybius and Dionysius may, I suppose, be fairly trusted as far as the editors are concerned. And may be noted that the most trustworthy part of the text of the most trustworthy author (Polybius) gives no instance of ß.


Now in this representation of v by ẞ something doubtless is due to the source of the Greek writer's narrative in each case. Something also to the instinctive desire of assimilating a word to Greek forms; hence the frequent use of ẞ before -tos, e.g. Aíßios (in Plutarch once only Λίονιος), Φλαβιος, Οκτάβιος, Φούλβιος, &c, Something again is due to phonetic reasons. Thus while ou is (in Plutarch) initial in 34 names and medial in 16, ẞ is initial in 17 and medial in 26. In 15 of these 26 ẞ follows λ or p, and in the rest it is between vowels; which are exactly the positions in which

« PreviousContinue »