330 LATIN LANGUAGE But further, forms like those quoted above from the inscriptions, e.g., dedro, oino, cula, &c., led Ritschl and his followers to the recognition of the fact that even at this early time there was a strong tendency to drop the final consonant in Latin ; and this at once furnished a clue to the proper interpretation of many metrical phenomena in Plautus, which had previously been explained on wholly incorrect assumptions. In the case of a line like Trin. , 306, ne tibi aegritiidinem pater parerem, parsi sedulo, it was assumed that pater was pronounced like pere, in order to avoid the apparent neglect of the law of position, which would, according to the practice of the classical poets, have lengthened the syllable -ter. Two considerations suffice to dispose of this hypo thesis : first, there is no evidence whatever that a mute between two vowels was ever dropped in early Latin ; secondly, if jmtcr become by "compression " pere, it would be natural to find mater becoming mere ; but in no case does a form with the first syllable long and the second lengthened by position take the place of one long syllable. On the other hand, there is positive evidence of a varied and unquestionable character to show that a final consonant was frequently dropped in pronunciation, especially in an iambic word. Hence it is clear that pater was pronounced patS, notpere. This shows too that it was not the case, as has been asserted, that a final r was dropped only when it took the place of an earlier s, although this is doubtless the most common instance of its omission. The consonants most commonly dropped are the following : s : e.g., nimi(s) lepide fecit verba ad parsimoniam (Aid., 493). This licence is retained by Ennius (e.g., Ann., 601, turn laterali(s) dolor, certissumu(s) nuntiu(s) mortis), and is common even in Lucretius. Cicero (Orat., 48, 161) speaks as if it had been the usual pronunciation in his own earlier days, and he admits it seven or eight times in his version of Aratus, e.g., magnu(s) Leo, &c. m : e.g., dum qnide(m) ne quid perconteris quod mi haud lubeat proloqui. The practice of eliding a syllable ending in m before a following vowel shows how lightly this consonant was pronounced even by the classical poets. It is very frequently omitted in inscriptions of every period (comp. Corssen, i. 267-74). As Quintilian (ix. 4, 40) says, " m parum exprimitur . . . neque enim eximitur, sed obscu- ratur." In this respect Umbrian quite agreed with popular Latin, but Oscan and Volscian carefully preserved the m (Corssen, i. 276). t : e.g., set arraboni dedi(t) quadraginta minas (Most., 648). So in inscriptions dcde (C. I. L., i. 621fr). d: e.g., hicapu(d) nos magna turba ac magna familiast (Aid., 340). ? : as above ; or, e.g. , pater venit. sed quid pertimui autem, belua (Ter., Phorm., 601). I : e.g., et simu(l) conficiam facilius ego quod volo (Ter., Heaut., 803). n : e.g., aut quid istucestquod vosagitis? non licet, tame(n) siis- picor (Ter., Hcc., 874). It is doubtful whether the last two licences occur in Flautus. Occasionally we find these two tendencies concurring, and pro ducing a short final syllable by the loss of a final consonant and the shortening of a vowel naturally long under the influence of the accent; so that we have forms like ores, boves, vides, rogas, manus, scanned as two short syllables, not only (.) before vowels, but (1} even before consonants, e.g. : (a) asini mordicus me scindunt, boves incursent cornibus (Aid., 232). (b) foras foras lumbrice qui sub terra erepsistr mo Jo (AuL, 620). viros nostros quibus tvi. nos voluisti esse matres familias (Stick., 98). ad papillas mantis ferat, labra a labris nusquam auferat (Bacch., 480). The tendency to drop the final consonant of an iambic word is further extended to groups of words of the same scansion, especially when the second is a preposition, as in quis ad fores est ? &c. (Amphitr., 1014). opta id, quod ut contingat tibi vis (Asin., 713). Accent had also an important effect in inducing the voice to hurry over unaccented syllables, even though long by position, in order to lay full stress upon an accented syllable. But this naturally took place only when the syllable thus shortened was itself preceded by a short syllable so that the formula for this process is " -- = w M - . Under this head we may bring a large number of instances of apparent neglect of quantity. Many of these are cases where the usual spelling is with a double conso nant. Some have argued that as doubled consonants were not used in writing before the time of Ennius (Fest., s. v. " Solitaurilia," p. 293, confirmed from inscriptions by Ritschl, P. L. M. E., p. 123), this is an indication that the pronunciation fluctuated ; but it is doubtful whether this was ever the case except under the influence of the accent ; and this influence was quite as powerful over syllables followed by two different consonants as by a doubled consonant. Thus, per annonam caram dixit me natiim pater (Stick., 179) does not diner in principle from quia omnis bonas bonasque adcurare addecet (Trin., 78); and the unusual quantity of the last two words in nos potius onerenms nosmet vicissatim voluptatibus (Stick., 532) is to be explained in precisely the same way, except that in the latter the voice is hurrying on to dwell upon a long accented syllable, in the former the accent has already fallen on a short accented syllable, a fact which naturally tends to shorten the fol lowing unaccented one. Compare for this configo sagittis fures thensaurarios (Aid., 395) where Goetz after Fleckeisen reads " sagitis." The combinations before which position is most commonly neglected are the following : nt : si id mea voliintate factumst (Trin., 1166). pt : voliiptatem inesse tantam (Rud., 459). st : magistnitus, si quis me hanc habere viderit (End., 477). rn: cassidem in caput,dormiboplaciderf in tabernaculo ( Trin., 7 26). ps : scio apsiirde dictum hoc derisores dicere (Capt., 69). rg : sed sine argento fnistra es . . . (Pseud., 378). It is needless to dwell further upon the details of Plautine scansion. The foregoing instances will have made it clear that, while there are some archaisms still retained, on the whole the language was beginning to suffer from that process of disintegration, which has left such marked traces upon almost every modern language. The introduction of Greek metres for the drama doubt less did much to check this process, and it is probable that, even in the earliest Roman comedies, licences of pronunciation are much less common than they were in the popular language of the time. But the iambic and trochaic measures, especially as employed by the Roman poets, admitted of a free treatment, which left room for much laxity. It was not until the hexameter came to be used for poetry that the laws of prosody were definitely fixed. The rigid canons of dactylic verse required that the pronunciation should be strictly determined ; and hence Ennius, although he does not appear to have introduced any marked changes in generally recognized rules of quantity, was compelled to settle positively much which had previously been fluctuating, and so to lay down the lines to which subsequent poetical works had to conform. From this time forward the literary language of Rome parted company from the popular dialect. It has been said with truth that even to the classical writers Latin was in a certain sense a dead language. Its vocabulary was not identical with that of ordinary life. Literary works, whether in prose or in verge, had to conform to a fixed standard. Now and again a writer of fresh originality would lend new vigour to his style by phrases and con structions drawn from homely speech. But on the whole, and in ever increasing measure, the language of literature was the language of the schools, adapted to foreign models. The genuine current of Italian speech is lost to view with Plautus and Terence, and reappears only in the semi- barbarous products of the early Romance literature. This appears the proper place for a rapid survey of the Pronun- pronunciation of the Latin language, as spoken in its best c j, a * i 2 i 01 -Ljtlw" days. I. CONSONANTS.- 1. Guttural, (a) Sonant G, pronounced as in English, but never softened before about the 6th century after Christ. (b) Surd C, pronounced always as & 1 (except that in some early 1 The evidence for this pronunciation of c will be found lest stated in Corssen, i. 43-67, and Boby, i. xlvii.-liv. It may be summed up as follows: (1) In some words the letter following c varies in a
