junctive (both present and imperfect)'. The ending et Nay, even the suffix it in the present of the so-called Aéyas legis or legeis [ei = i]3 We find the same quantity again in the third pers. sing. perf. Once, it is even expressed by the spelling ei=i1 Merc. 530, where the ms. A gives redieit, and it is well established by many instances in Plautus and Terence3, to which we have to add about eight different examples from Virgil, Horace and Ovid.—The same remark app. xx plies to the subjunctive ending erit, the fut. perf., īt in 1 I may quote an instance of cum eó simul me mítterēt. ei In this line, Ritschl and Fleck- 2 Most of these originally long 3 Compare scribis Hor. Serm. II 3, 1. 4 For inscriptions see E. Hübner's Index in the C.I.L. 1 p. 601. 5 Corssen II 493 sq. gives a sufficient number. 6 See Nettleship, 1. c., p. 469. Wherever archaic quantities occur in the later poets, they should be considered as the result of imitation of the earlier writers. We may add that the original long quantities are admitted by the later poets quite exclusively in arsi, i. e. when the metrical accent falls upon the ending in question. оз sit velit mavelit, nay even to the simple future erit ('he will be') Capt. 208 and bit in vaenibīt Most. 11601. In the passive, the shortening propensities of the Latin language displayed themselves chiefly in the first person of the singular. In Plautus we find sometimes the original quantities ōr ār3, nay ferur is met with as late as Ovid (Met. vII 61). Analogously, the endings er and rer in the subj. were originally long. It may finally be remarked that es ('thou art') is invariably long in the prosody of the comic poets. B. IRREGULAR SHORTENING OF LONG FINAL VOWELS. All these long vowels are, however, of but occasional occurrence in Plautus and Terence-they are, indeed, nothing more than a few scattered remnants of a period of the language, which was rapidly waning and dying away. The general character of the language in the time of Plautus was quite different. A destructive element had already commenced its powerful influence upon the language, and had already deeply affected and altered the original quantity of many endings and even of many root-vowels of Latin words. The accent in Latin never falls on the last syllable, and its tendency was to destroy the length of this last syllable, especially in case the word was disyllabic and had a short penult. p. xxi We find, therefore, in Plautus a greater number of instances in which the above-mentioned archaic long vowels have been shortened than where they still retain their original quantity-and of this phenomenon we should attribute the main cause to the influence of the accent. But the development, having once commenced, did not stop there; on the contrary, many short quantities are to be found in the comic poets which were either entirely rejected or but exceptionally admitted by later poets. I shall first speak of the final vowels occasionally shortened in the rapid pronunciation of the times between the second and third Punic wars. It will be observed that all the instances which we are about to produce represent disyllabic words which are used as pyrrhichs, instead of their original iambic prosody. This could never have taken place, had they been pronounced with the accent on the last syllable. The long a of the first declension was not only shortened in the nom. and voc. (as it remained indeed ever afterwards), but even in the ablative, e.g. pró mală víta fámam extolles, pró bonă partam glóriam. Ennius ed. Vahlen p. 94. The same happened to the o of the dat. and abl. sing. of the second declension, e. g. the abl. domo stands as a pyrrhich in the following two instances: unde éxit? unde nísi domō :: domŏ? :: mé vide :: etsi vídeo. Mil. gl. 3761. domo quém profugiens dóminum apstulerat, véndidit. Capt. prol. 18. In the abl. ioco the final o is shortened Bacch. 75, where the reading of the mss. is as follows: símulato me amáre :: utrum ego istuc iócon adsimulem an sério? tion destroyed the inflexional cented fast syllables. 1 See Ritschl, praef. Stich. XVII. But see also Brix's note in his recent edition of the Miles gl., p. 138. and so Fleckeisen gives the line, while Ritschl writes utrum ego iocón id simulem an sério. ero (dat. of erus, master) stands as a pyrrhich Aul. 584 and Most. 948. bono is another example of the same kind: haec erít bono génere nata, níl scit nisi verúm loqui. malo falls under the same head: Persa 6451. malo máxumo suo hercle ílico, ubi tántulum peccássit. Cas. IV 4, 6. sét etiam unum hoc éx ingenio málo malum inveniunt suo. Bacch. 546. cavě sís malo. quíd tu málum nam mé [anapæstic]. Rud. Iv 3, 12. In the last passage, Fleckeisen alters the metre by inserting nunc after nam. The abl. modo (which should not be confounded with p. xxii the particle) stands as a pyrrhich Aul. 589: eodem modo servóm ratem esse amánti ero aequom cénseo, and Pseud. 569, where the mss. read as follows 2: nově módě, novom aliquid inventum adferre áddecet. In this case, the words novo modo should be taken as a proceleusmatic, a foot which is very frequent in the first place of a senarius (see Ritschl, Proll. Trin. CCLXXXIX). With the same quantity we have in the Trinummus 602 quó modo tu istuc, Stásime, dixti, nóstrum erilem fílium. Lachmann (on Lucretius p. 116) calls the short quantity of this o'mirabile:' Prof. Key, to avoid recognising a fact like this, proposes the monosyllabic pronunciation 1 In this passage, Ritschl gives bono without the mark of ecthlipsis (Proll. CXLIV), i. e. he considers the final o to be shortened. 2 Ritschl omits inventum and thus restores modo to its usual measure. I am glad to see that Fleckeisen does not follow his example. mo, and to corroborate this conjecture, he appeals to the Roman way of abbreviating the word: mō ('Alphabet' p. 141). But I may observe, that by abbreviating the orthographical representation of a word, nothing is prima facie insinuated as to its pronunciation'. Prof. Key's other argument is drawn from the Romance languages, where quomodo appears in the shape of como come comme: it would, no doubt, prove that quomodo really sounded like quomo (como) in the latest period of the Latin language, but would it explain the real nature of the general law whose slow but steady working at last degraded full words and endings to poor cripples? We recognise in Plautine prosody the beginnings and the first germs of a depravation of the Latin language, which attained its final development in the Romance languages. We need not, therefore, hesitate to explain Romance forms from such shortened endings as are found in Plautus, but great caution should be used in remodelling the pronunciation of Plautine forms upon the analogy of Romance corruptions. The spirit of modern philology requires that the order of time should be observed and forbids us to blend the peculiarities of the different p. xxiii periods of any language. If, however, any further proof 1 If e. g. we were to take the copy-books of German students as the indication of their pronunciation, we should arrive at a great many surprising discoveries in German pronunciation; but unfortunately, they would all be repudiated by the actual pronunciation of those students themselves. 2 The sense of these words is borrowed from Prof. Key himself ('On the so-called A privatiuum' p. 8).-The list of contracted words, given by Prof. Key ('Alphabet' p. 146-148), would require a great many additional observations, if the present writer really intended to examine each separate instance. But he has no intention to criticise all his predecessors, nor does he think it necessary always to state when he deviates from the views of other scholars. He would, however, ask his readers not to think him unacquainted with really excellent labours in the same field, even when he does not expressly quote them; but taking notice of everything would too much increase this Introduction, which the author first thought he could entirely dispense with. He may, however, state that almost the same views as those given here, will |