Page images

Dziatzko's papers has followed Donatus on the tibiae, but so long as his variations remained unexplained, grounds for the preference accorded to the manuscript readings were not entirely convincing.

A third, and perhaps the most important, result achieved in this paper is the discovery of new evidence about the history of the text of Terence. From this study of the paraphrases and Didascaliae, it appears that the new or alphabetical order of plays was based upon a y order, and so presumably upon a y codex; that the & order is not the original alphabetical order, but has suffered an interchange of position between the last two plays; that the & family has derived its Didascaliae not from a manuscript similar to that used by Donatus, but from a manuscript of the y family; and that, in part at least, the y recension was the basis of the recension, not vice versa, and is, therefore, the older of the two. All these points are more or less directly opposed to the theories set forth by Leo and Dziatzko, who agreed in substance, though differing in details. I have elsewhere expressed the conviction that these theories are opposed to the facts, and are, therefore, utterly untenable. This view

is confirmed by the results reached in this paper. But since limitations of space prevent any adequate presentation of this subject here, for the present I content myself with pointing out above the most general application of my results to the question of the manuscript tradition. At some time in the future I hope to return to this subject, and, in connection with this and other sources of evidence, to indicate what I conceive to be the true history of the text of Terence.

1 Harvard Studies in Classical Philology, XIV, p. 171 f.

VIII. - Plautine Synizesis.

A Study of the Phenomena

of Brevis Coalescens.1



THE purpose of the present paper is, first, assuming the real existence of Plautine synizesis, to indicate more clearly, through a discussion of syllable values, its probable nature; secondly, to establish the fact of its actual existence, if possible, to the exclusion of any other hypothesis; thirdly, to investigate the laws and limitations to which it is subject. In view of the doubts which have prevailed in some quarters upon the subject of early Latin synizesis, this may seem to some too large an undertaking, but, fortunately, it will not be necessary for me to discuss these questions in reliance upon my own unaided resources. Many Plautine scholars have made just and careful observations upon this general subject, and even those who have advocated erroneous views, notably Müller and Skutsch, have often contributed something of real value to the discussion. Hence it will be my task chiefly to collect and to harmonize the important single facts which my predecessors in this field have already pointed out.2

1 D. M. et perpetuae memoriae GUILELMI ELIshae Peters, praeceptoris sui carissimi in Universitate Virginiensi, viri in fabulis Plautinis maxime versati sacrum esse voluit pientissimus auctor.

2 Bibliography: Ritschl, Proleg. cap. xii; Spengel, Einleitung zu Andria, § 8; C. F. W. Müller, Plaut. Pros. 456 ff.; L. Müller, Res Metr.2 lib. iv (279 ff.); Leppermann, De correptione iamb. ap. Pl., Münster, 1890 (unfor tunately, L.'s statistics upon synizesis and non-synizesis forms are in part extremely inaccurate); Bömer, De correptione iamb. Terent., Münster, 1891 (a careful dissertation); Abraham, Stud. Plaut., Neue Jahrb. f. Phil. XIV, 204 f.; Nilsson, Quomodo pronomina, quae cum subst. coniunguntur, ap. Pl. collocentur, Lund, 1901; Leo, Plaut. Forsch. 323; Havet, De Saturnio, 32, 79 f.; Gleditsch, Metr3 258, 295; Skutsch, Sat. Viadr. 135 ff.; Audouin, De Plaut. anapaestis, Paris, 1898, p. 69 f., 121, 228; Ahlberg, De proceleusmaticis antiquae poesis Lat. I, 85 ff.; Lindsay, Introduction to the Captivi, 26 ff.; Neue, II3, 366 ff., 377 ff., etc.; Thurneysen, K.Z. XXX, 499 f.; Bronisch, Die oskischen i- und e- Vocale,


That the prosody of Plautus is different in many respects from the prosody of the Augustan poets, is universally admitted to be true, yet, in spite of our earnest efforts to draw a clear line of demarcation between the usage of the earlier and that of the later period, we are in constant danger of bringing to the study of Plautine verse a set of conceptions based upon the familiar Augustan prosody, and a vocabulary of metrical terms unconsciously colored by their adaptation to the Augustan use. This traditional vocabulary, which we are under the necessity of using, is often also quite inadequate and quite inexact, and, in view of the unfortunate associations with which some of these terms are almost indissolubly connected, their use often increases our perplexity and embarrassment. No one of the ancient metrical phenomena seems to me more inadequately named than that which I propose to discuss in the present paper, viz. early Latin synizesis, since the term 'synizesis,' as is well known, is employed in several quite different senses and is often applied to quite disparate phenomena; before I enter fully upon the discussion of this question, however, it will be best to give another and an even more striking illustration of the inadequacy of our common metrical terms as applied to Plautine prosody.

The Problem of Ille. — It is important to recognize that the system of quantitative measurements employed by Plautus is different in several respects from the quantitative system of Augustan poetry. I do not mean to imply by this that either one of the two systems is less good or less genuine than the other; on the contrary, both systems seem to me to be based as a whole upon the most genuine and substantial foundation possible, viz. a definite and original sense of quantity on the

Leipzig, 1892; Sturtevant, Contraction in the Case Forms of deus, is, and idem, Chicago, 1902; Engelbrecht, Wien. Stud. VI, 236 ff.; Brock, Quaest. gramm., Dorpat, 1897; Hodgman, Harvard Studies, IX, 151 f.; Christ, Metr.2 30 ff.; Corssen, Ausspr. II2, 744 ff.; Skutsch, Tépas, Göttingen, 1903, p. 108 ff.; etc.

part of those who used the Roman language.1 But I mean that the two systems differ somewhat in the solution of certain practical problems of exact measurement, such as must confront every form of quantitative speech. An apt illustration of this fact is afforded, as it seems to me, by the treatment in Plautus of those dissyllabic pronouns and conjunctions, which, according to the grammarians, are without the "high tone" (fastigium, cf. Keil, VII, 360, 10), viz. ille, inde, immo, nempe, etc. As is well known, ille not only has the value of three morae (— ~) in Plautus, but also at times the value of two morae (~~), but Skutsch has shown in his Plaut. Forsch. that the latter value belongs to the pronoun, only when it is dissyllabic and does not suffer elision of the final syllable; monosyllabic ill[e], ill[a], etc., have invariably the value of two morae (-). In view of these facts, Skutsch maintains that the conclusion is certain that the ultima of ille, immo, nempe, etc., was often totally suppressed by syncope in the rapid pronunciation of colloquial speech, and he offers us the following "mathematical demonstration" of the correctness of this view (Plaut. Forsch. 40):

Nempe before consonants has the value of
The first syllable has the value of

The second syllable has the value of

[ocr errors]
[ocr errors]

2 morae
2 morae

o mora

Such a solution of the problem appears at first sight most attractive, and, in the case of nempe, it has been accepted as a correct solution by a very large number of Plautine scholars; yet, in my judgment, the proposed solution is very far from being really convincing. Since, however, I propose to discuss the question of these pronouns more fully elsewhere, only the general character of my criticism need be indicated here.2 Professor Skutsch's argument appears to

1 Cf. Trans. Am. Phil. Assoc. XXXV, 50 f.; A.J.P. XXV, 425 ff.

2 For this reason do not dwell here upon the fact that many other pronouns, such as nēquis, nèqua, quisquis, hocin, omnis, etc., should be added to the list given by Skutsch, nor upon the fact that the pronominal word-orders cannot be easily changed for the colloquial language, and hence would quickly develop

me to involve two very doubtful assumptions: (1) The assumption that the syllable nemp has precisely the same value and the same pronunciation, whether it belongs to the dissyllable nempe, or to the monosyllable nemp[e]. Such a supposition would be obviously incorrect, for example, if applied to the English sentence; for in our own language the dissyllable manly and the monosyllable man have quite different relations to accent and quantity (cf. Dabney, Musical Basis of Verse, 32). (2) The assumption that, in the verse of Plautus (which closely conforms to actual speech), all of the syllables can be properly divided into those of one, and those of two morae respectively.

This last consideration is a fundamental one, and has to do with the essential difference which exists between Plautine and Augustan prosody. For, however regular the usage of the Augustan poets may be, it is not certain that, in the earlier period, either one of the syllables of nempe had normal value. Thus final short syllables, especially final syllables in -e, may well be shorter in many cases than the normal short, as is shown by the fact that they suffer actual syncope in a limited number of cases, and in many cases they are treated, by preference, with elision (Langen, Philol. XLVI (1887), 419; Lindsay, L.L. 203). It is even more certain that the long penult of toneless pronouns and conjunctions is shorter than the normal long of a penultimate syllable. Let us suppose then that the penultimate syllable of nempe has the value of 1 morae, instead of 2 morae, and that the ultima has the value of of a mora, instead of I mora. Would not these assumed values account equally well for the metrical usage of Plautus? The dissyllable nempe, which would really have the value of 2 morae, could still be always treated as two shorts, and the monosyllable nemp[e], which would really be worth only 1 morae, would still constitute, in actual usage, a long. And even if this explanation should be rejected for the single word nempe, it is evident

the use of any metrical license for which a genuine ground once existed. Further, in any complete discussion, the question would arise why the supposed syncope does not apply to all well-worn words alike, such as esse, saepe, curre, etc.

« PreviousContinue »