Page images

The pronunciation of a final vowel before an initial vowel is somewhat uncertain. But that it was not omitted, but either lightly pronounced separately, or formed into a diphthong with the initial vowel, seems both in oratory and poetry to be the right conclusion, both from the language of Cicero (Orat. 13. § 77; 14. § 150 sqq.), and Quintilian (IX. 4, § 33, XI. 3. 34), and from the fact that the vowel was written, not omitted. (See Corssen, Ausspr. II. pp. 770 -793). The chief points of usage in this matter in Latin verse are given (after Luc. Müller) in §§ 288-291. (I have there used the terms elided and elision in conformity with general usage and for brevity).

The modern analogies are thus stated by Mr Ellis (p. 324). 'In common French discourse the final e and many medial e's 6 may be said to be entirely elided..... When singing, the French 'not merely pronounce these e's, but dwell upon them, and give 'them long and accented notes in the music. This recognition is absolutely necessary to the measure of the verse, which, depending ⚫ solely upon the number of the syllables in a line, and having no ' relation to the position of accent, is entirely broken up and destroyed when these syllables are omitted. And yet when they 'declaim, the French omit these final e's without mercy, producing 'to English ears a hideous, rough, shapeless, unmusical result, 'which nothing but a consciousness of the omitted syllables can 'mass into rhythm.' Again (p. 329 n.), 'In German and French 'poetry the omission of the vowel is complete and absolute. It is 'not in any way slurred over, or rapidly pronounced in connection 'with the following vowel, as is the case in Italian and Spanish 'poetry, and even in Italian singing. The Germans, like the Greeks, 'do not even write the elided vowel. The Latins wrote the elided ' vowel, as the Italians do, and may therefore have touched it briefly, 'as in the English custom of reading Latin verse; whereas it is 'the German custom to omit such vowels altogether, even in reading 'Latin verse. Except in a few instances as l', t', &c. the French do 'not make the elision of a final e before a following vowel, and in 'old English the vowel was written even when elided.' thinks Chaucer sounded, at least usually, his final e's.

Mr Ellis

Final m before an initial vowel was, according to Quintilian (IX. 4. 40), sounded, though slightly: 'Etiamsi scribitur, tamen parum 'exprimitur, ut multum ille et quantum erat, adeo ut pæne cujus

[ocr errors]

'dam novæ litteræ sonum reddat. Neque enim eximitur, sed 'obscuratur et tantum in hoc aliqua inter duas vocalis velut nota 'est, ne ipsæ coeant.' Mr Ellis (Phil. Soc. Trans. 1867, Suppl. p. 20) suggests that the m may have simply nasalized the preceding vowel, as is the case with m frequently in Portuguese and French, and with n always in the latter language.

The omission of the initial vowel in est is mentioned in § 721. Perhaps also the same may have taken place in istuc, &c. (§ 375).

The chief rules of accentuation are given in Book I. Chap. XIII. I confess to entertaining some doubts as to a short syllable, when followed by an enclitic, receiving the accent, e. g. primáque. As the Romans would not have accented primaque on the penult, if it had been one word, I do not see why the i should have lost the accent by the addition of the enclitic. But the grammarians no doubt are against me, and I cannot pretend to any great confidence in my own judgment in matters of accentuation and quantity.

Division of words into Syllables.

The general doctrine given (in §§ 14-16, 272-274) is, I think, in fair accordance with the teaching of Mr Bell and Mr Ellis. To a pamphlet of the latter I owe the first hint of what I believe to be the truth. Their views will be found in Bell's Visible Speech, p. 69 sq., Principles of Speech, p. 87 sq.; Ellis's Early English Pronunciation, p. 51 sqq.

The application of this doctrine to Latin brings me at once into collision with the doctrine faithfully transmitted from Priscian (Lib. 2), and even with the same doctrine as modified by Krüger (Lat. Gr. §§ 32, 33) and Madvig (Bemerkungen, p. 17). Madvig's

1 But the mode of representing the pronunciation is often different from what they appear to recommend. For instance, Mr Bell (Visible Speech, p. 119; and comp. Ellis, p. 55 note) says critical is pronounced cri-ti-cal not crit-ic-al. I am quite aware that his ear is far better than mine, but I cannot think, if we are to assign the t to one syllable more than the other, that it would be generally thought to belong to the second syllable. However, write the division how we may, I do not mean more in what I say of Latin pronunciation of mutes than that the consonant was pronounced as much with the vowel before it as t is (invariably I believe) in this word critical. And this is not the notion which I get from the ordinary statement.

account of both is as follows (Lat. Gr. § 13): 'A consonant be'tween two vowels belongs to the last vowel, and with this it is ' combined in pronunciation. Of two or more consonants the last, 'or, if they can begin a Latin word, the two last, belong to the 'following vowel, the remaining consonant or consonants to the preceding vowel (pa-tris, fa-scia, ef-fluo, perfec-tus, emp-tus). 'The double x is best united with the preceding vowel. In words 'compounded with prepositions the final consonant of the preposi'tion is not separated from it (ab-eo, ad-eo, præter-eo, also prod-eo, 'red-eo).



[ocr errors]

According to a generally spread custom' [this is Priscian's doctrine] words are in many books so divided, that all consonants 'which in Greek can begin a word, and all mutes with liquids '(even though they could not begin a Greek word, e.g. gm), and 'similar combinations of two mutes (e.g. gd as ct) are drawn to 'the syllable following (i-gnis, o-mnis, a-ctus, ra-ptus, Ca-dmus, 'i-pse, scri-psi, Le-sbos, a-gmen, Da-phne, rhy-thmus, smara-gdus).'

I assert, on the contrary, that the Roman pronunciation tended to unite a consonant with the preceding, not with the following, vowel; and I have briefly mentioned in § 273, and need not here repeat, the indisputable facts of Latin etymology and prosody, which seem to me to justify this inference. I have in § 274, p. 89, briefly noted (in some words of Mr Bell's) the probable basis of the ordinary doctrine, and will now remark on some objections which may possibly be urged to three of my four arguments.

I. It may be said that the retention of o after v (instead of allowing the change to u § 93) shows a connexion with the following, not with the preceding, vowel. Unquestionably it does, and the reason is that the vowel u only becomes consonantal at all by its rapid pronunciation before a following vowel. V (=w) is not a consonant standing independently between two vowels (as it ought to have been to invalidate my principle), but a vowel, which, if it be distinctly pronounced as such, does not necessarily affect either the preceding or subsequent sounds, but, when coming before a different vowel, naturally gains a semiconsonantal character. W is hardly pronounceable at the end of a syllable. See above, p. xxxyi,

2. It may be said that a change of the final sound of a word is

sometimes caused by the initial sound of a word following; e.g. ἀμ-πέδον for ανα πέδον ; imprimis for in primis; and that therefore such a change does not imply the union in one syllable of the consonants so affecting each other. I do not deny that sounds in different syllables may affect one another; the law of assimilation or dissimilation does undoubtedly extend over several syllables, and in some languages, I believe, prevails much more largely than it does in Latin; but when we find, as we do in Latin, such changes frequent and regular, almost invariable indeed, in the case of contiguous consonants, and very rare, in the case of separated consonants, it seems to me difficult to suppose that these contiguous consonants were separated in speech. And such instances of the influence of initial sounds of a word on the terminal sound of a preceding word rather show that the two words run into one another in pronunciation. This is confirmed (a) by the express statement of the Latin grammarians, that prepositions with a case had no separate accent (§ 299): (b) by their being constantly written as one word in inscriptions (Corssen, Aussprache, 11. 863-872); (c) by the change of vowel in (for instance) illico for in loco (Ib. p. 869).

3. It may be said that the prosodiacal law, of a syllable being long if its vowel has two consonants after it, applies just as much when these two consonants are in different words, as when they are in the same word as the vowel; and therefore, if the lengthening of the syllable proves that the consonants are in the same syllable, it proves also that the initial consonant of a word must be regarded as in the same syllable as the end of a preceding word. This is so, no doubt, but how else is the fact to be accounted for? The Romans did not arbitrarily invent the laws of prosody: these laws must in substance rest on sounds actually heard. Part of the solution of the apparently strange confusion of word with word is, I think, to be found in the fact that words were not divided in writing, and that consequently a law strictly applicable to consonants in the same word was applied also to consonants in different words, partly from a real confusion in rapid speech, partly from a want of distinction in writing. When both consonants are in the second word, the Romans were much more reluctant (cf. § 293) to

admit in theory, because they were less liable to produce in practice, the same prosodiacal effect. The confusion of two words here supposed finds an analogy in French, when the final consonant otherwise mute is revived in order to be pronounced, not with its own word, but with the following word. (Comp. also Ellis, Early

Eng. Pron. p. 56.)

To the 4th argument I do not see what answer can be made.


Madvig (Bemerk., pp. 17, 26. n.) points to the vowel found in perfectus, nomen compared with i in perficio, nominis, and considers it to be due to the syllable being closed in the first two words, open in the last two. And it may be urged that on my theory, though perfect-us compared with perfic-io may admit of explanation, there are not two consonants to account for the e in nomen. True, but neither is there a closed syllable to account for mare compared with maris; and still more clearly in words like iste, ante, compared with istic, antistes, &c. (see § 234, 3.), the open syllable exhibits the e, but becoming closed takes i instead. The true explanation of the e in nomen, I am not at all sure of: it might perhaps be held to be the result of the suffix having once been, as some philologers (e.g. Leo Meyer, II. 263) suppose, ment (for mant), in which case the e has remained as in eques for equets, &c. But it is enough to observe that on examining carefully the laws of change as set forth (more systematically than I have elsewhere seen) in § 224, 3, it will be found that nomen, nominis is quite consistent with other words, and that these laws, be their basis what it may, do not depend on the syllable being open or shut.

The following is a summary statement of the probable pronunciation of educated Romans in the period from Cicero to Quintilian, say 70 A.C. to 90 P.C. (The references in brackets are to pages of the preface or sections of the book where arguments are given.)

« PreviousContinue »