Page images

ent and originally independent parts of a compound, as in coerce, coēmisse, coēgi, deôsculer, dehortari, presēndo, etc. (cf. Klotz, Grundz. 140; C. F. W. Müller, Pl. Pros. 451 ff.'). These phenomena are usually grouped under the name of synizesis, but as I have already pointed out (p. 159), the use of this name cannot be considered fortunate. To classical scholars the term 'synizesis’ is chiefly associated with certain poetical artifices, certain metrical licenses, which dactylic poets like Homer and Vergil have freely used in order to adapt words of a difficult kind - especially cretic words like ναυτέων vautéwv or aurcis — to the exigencies of the hexameter verse, and, although the ancients sometimes defined the term naturally and scientifically -- as, for example, Scholia A upon Hephaestion, p. 119 W., where it is implied that, in Aoi for Deol, the e is naturally so weak as to be scarcely audible yet the artificial sense has always predominated in the use of the term ; cf. Christ, Metr.2 28, § 37.2 Hence there can be little ground for wonder that so conscientious a student of Plautus as C. F. W. Müller has put it on record (Pl. Pros. 456, n. 1) that he considered all Roman synizesis artificial and of Greek origin, and there can be little doubt that he was largely led by this view of the subject to substitute iambic shortening as an explanation for all supposed cases of synizesis in Plautus. Before we undertake, then, to prove the reality of Plautine synizesis or to explain its laws, we must clearly differentiate this synizesis of the living speech and of the preferred rhythm from certain other phenomena which possess a wholly different character, and owe their origin to a widely different line of development.

A. The phenomena which we propose to discuss are quite different from the synizesis of Greek origin (Synisesis Grae

1 Here, as in every other part of the present subject, authorities differ about the proper name for the phenomena. C. F. W. Müller, for example, rejects the term 'synizesis,' and compares the method of procedure to elision between independent words.

2 In the passage of Scholia A cited above, the natural and artificial elements are strangely mixed, e.g. tas ka kouet plas åpalpeî tûv otixwv (artificial) ... δισύλλαβον υποκλέψασα (natural) διά τήν χρείαν (artificial) και γέγονε θοι TPÓTOY Tivá (natural).

canica, L. Müller, R.M.2 283, 325 ff.), which was first introduced by Catullus and his contemporaries into Roman epic poetry, and which may best be illustrated by the familiar Vergilian examples aureis (Aen. i, 726), alveõ (ib., vi, 412), Eurysthéā (ib., viii, 292). I may point out just here, by way of anticipation, one important and striking difference which exists between Vergilian and O. Lat. synizesis. The latter is chiefly connected with the weakening of an initial syllable in words which begin with an iambus, e.g. (e)os, (e)āmus, t(u)ām-rem (but cf. èxčámus, aúreò), while the former assumes the weakening of a medial syllable in polysyllabic words, e.g. aureò. This treatment of the initial syllable of iambic words in O. Lat. is probably to be explained on the wellknown principle of the preferred rhythm of a language; in other words, the ear of the Roman people originally preferred the fierce energy (yopyótns) of the trochaic rhythm to the comparative tameness of the iambic movement, and this national preference for the trochaic rhythm made itself felt in word-forms, wherever the conditions were otherwise favorable, as in (e)ámus, èxěámus, ádfüìt, aúrěd, etc.

Greek synizesis, on the other hand, although doing no actual violence to this principle, does nothing to promote it, and hence stands in no close relation to distinctive Roman tendencies or the cadences of Roman speech. Yet widely as the Greek and the Roman forms of synizesis differ in these respects, they are both subject at certain points to the same limitations (see p. 202).

B. The phenomena which we discuss are wholly different in character from the treatment of vowel i and vowel u as consonant i and consonant u respectively, in such a way as to make the preceding short syllable long by position, e.g. āvyum (Enn. Ann. 91 M.), insīdyantes (ib., 443 M.), ābyete (Verg. Aen. ii, 16), tēnvia (id., Geor. ii, 121), consīly[um] (Hor. C. iii, 4, 41); for other examples, see L. Müller, R.M.2 299 ff.; Christ, Metr.2 32. According to L. Müller (1.1., 283, 301), this hardening of i and u is confined to epic and lyric poetry, where it is due almost wholly to metrical necessity (extrema necessitate), and its use was never admitted in Latin iambic and trochaic metres. In fact, as many metricians recognize, this phenomenon is not, properly speaking, synizesis at all, and it is not subject to the strict laws which we shall find always observed in Plautine synizesis (p. 178).

The consonantization of vowel i and vowel u, as we have seen, is wholly unknown to Latin iambic verse. An entirely different treatment of these weak vowels is occasionally found both in O. Lat. iambic and dactylic verse, and consists in their total suppression (so far as concerns the metre), in cases where they stand in inner hiatus. This usage is rare, but is thoroughly well attested, and, unlike synizesis, it is subject to no special restrictions. Examples are: evenat, advenat, pervenat, etc. (Pl.), monerim (Pacuv.), augura (Accius), progen(i)e[m] (CIL. I, 38), or(i)undi (Lucr.), oper(i )untur (Laevius); see L. Müller, R.M.2 289 f.; Christ, Metr.2 32; Klotz, Grundz. 140; Lindsay, L.L. 465, 506.1 For a similar use in Greek poetry, cf. Christ, Metr.2 30, 29; L. Müller, l.l., 289. Under this head belong also the very rare cases in Plautus of trisyllabic me[o] animo, su[0] ăliquem cited below (p. 203). Cf. also Pl. Per. 100:

Terrestris té c(oěpulónus compellát tuós.?



We have seen that both the synizesis which is imitated from Greek usage and the hardening of i and u are artistic devices which are employed to introduce difficult words into special kinds of verse. The question remains whether there was, in such word-forms as were named above (meo, meām, eo, etc.), a natural synizesis characterizing the living speech, and somewhat similar to the pronunciation which is almost

1 Most of these forms are well attested, but Plautine evenat, pervenat, have little Ms authority, and are sharply called in question by Exon, Hermathena, XIII, 138 f. The suppression of vowel i in these forms also seems, however, the most prol able solution. A different explanation is possible also for some of the o her examples cited above; a case of suppression that should be added to the list is periero, which, according to Professor Warren's probable derivation in Trans. Am. Phil. Assoc. XXXII, 112 f., represents periuero (reduced from periovero); cf. Walde's Lexicon (appendix), which accepts this explanation in part.

2 A numeral placed below the line denotes a foot of iambic verse; in the line, a foot of trochaic verse.

universally admitted as the normal and regular one for dein, deinde, proinde, proût, quoàd, dehinc, etc. In spite of the objections raised by some Plautine scholars, such as C. F. W. Müller, Havet, and Skutsch, I believe it certain that there was in many such word-forms? a weakly uttered syllable in hiatus, which, in comparison with the much longer adjacent syllable, appeared to diminish greatly and to fall sensibly below the value of a mora. It was therefore nearly always neglected in the middle of a verse, and was freely allowed the value of a mora only through the conventions of the verse-close. This so-called synizesis occurs in connection with the short vowels e, i, and 11, all three of which (especially i and u) readily tend to assume a semi-vocalic character 3 Yet in the O. Lat. phenomena these vowels have not become consonants, but remain slight and weakly uttered vowel sounds. This appears from the fact that elision con

1C. F. W. Müller (1.1., 451), as has already been noted, rejects deinde, proinde, etc., and supposes a species of elision between the two parts of the compound word, eg. d[e]inde, etc. From the list of words given above, I have omitted the dissyllabic eidem (idem), eisdem (īsdem), also the somewhat less frequent eadem, eosdem, etc., of Augustan poetry (L. Müller, l.l., 297, 322), because these latter may seem to some capable of a different explanation.

2 Of course, the synizesis here described does not apply to all the case-forms of meus, etc., nor to elided forms under all conditions ; cf. pp. 178, 201 f.

8 Ritschl seems to be very nearly correct in saying (Opusc. II, 600 f.) that the whole conception of O. Lat. synizesis rests upon the semi-vocalic character of the two sounds, i and u— a statement which I wish, however, to interpret in the sense that i and u are half-vowels, and not full vowels, in these words. Of course, in the words which show synizesis, e sometimes represents an original i, and we find the spelling iamus (for eamus) in Inscrr., iam for eam (acc. sing. fem.) in Mss of Varro (Neue, 113, 381) and in the grammarians, initial i- in all the Oscan forms of the demonstrative pronoun is (Bronisch, 1.1., 97), etc. Yet, in reality, all the forms just named and especially, for example, the O. Lat. form mius (which is attested by the grammarians and, according to Lindsay, L.L. 21, by some of the derivative Romance forms, but which is needlessly explained away by Sommer, Lat. Lautlehre, 446), seem to show rather the general tendency of ě in hiatus to become close e and so approach the sound of i; for numerous examples of this tendency, see Lindsay, L.L. 19 ff.; Seelmann, Ausspr. d. Lat. 187 f.; note further Vergilian miis, attested by Quint. viii, 3, 25, according to a probable conjecture (Mss mus), and ium on a Luceria inscr., CIL. IX, 782. On the part plaved by the half-vowels i and u in synizesis, cf. also Corssen, Ausspr. 11°, 767; Ahlberg: De procel. I, 86.

stantly occurs before monosyllabic (e)ām, (e), etc., e.g. Tri. 197 iuxtáque (e)am cýro. Further, the weak vowel is wholly suppressed in the quasi-phonetic spelling of Ennius, i.e. sis, (abl. pl.; cf. the double forms suāvium — which at times was doubtless very nearly sữavium and sāvium 1), and in the vulgar forms do, dae, quattor, des, quescas, etc.; note also its loss in the compound forms ecc-um, ecc-am, etc., while the consonantal i of iam gives rise to the trisyllabic compounds nunctam, etiam, etc. Although a total suppression of the vowels e, i, and u does not ordinarily occur in 0. Lat. synizesis, it would yet be more nearly correct to say that we have their suppression under certain conditions, than it would be to say that we have their consonantization. For, as we have already seen, the tendency in O. Lat. was rather to suppress vowel i and vowel u than to fully consonantize them; cf. O. Lat. ăbicio 2 with Augustan ābyicio (Vendryes, L'intens. init. 266 f.), and similarly tenia and abete would possibly be nearer the 0. Lat. colloquial pronunciation in some respects than tēnvia and ābyete.3

To sum up: The vowels in question have not lost the musical quality by virtue of which they are vowels, and have not degenerated into the consonantal lack of tone. They are vowels in the true sense of a sound that has tone, but they are slurred or faded tones. We may still further illustrate

1 For numerous cases of the loss of post-consonantal u before the accent, in vulgar Latin, see Lindsay, L.L. 268.

2 One should rather say occasional O. Lat. ăbicio; for Exon's convincing study (Hermathena, XIII (1904), 129 ff.) shows clearly that ăbicio is an exceptional scansion in O. Lat., and its occurrence not wholly free from doubt in any case.

8 It must be remembered, however, that the consonantal character of j and v does not appear to have been fully developed in 0. Lat. (cf. Lindsay, L.L. 45), and it is probable, on the whole, that 0. Lat. j and v were half-vowels (i, y) rather than spirant consonants (y, w). Thus, according to one view (cf. Lorenz on Mo. 642), synizesis takes place not only through h, as in nihil, prehendo, but also through j, e.g. in hūius, quoius, eius; but, in any case, it is clear that ; assumes a vocalic character in these words, cf. conctus from coionctus.

4 A number of our editions of Plautus and Terence fall into the inaccuracy of describing vowel i and vowel u in synizesis as consonants. Thus Hallidie, in his edition of the Captivi (London, 1891, xliii), says expressly: " Another form of contraction is caused by the vowel i being pronounced as the consonant i (y). This method of pronunciation is adopted by the Augustan poets in such words as

« PreviousContinue »