zesis forms such as myus, myă, yă, or such as tuos, meus, mea, ea;1 such strange and abortive forms are rejected at the present day by all careful critics (e.g. Lindsay, L.L. 439, 426; Ahlberg, De procel. I, 88 ff., 154), and, in accordance with Spengel's well-considered views (Einl. zu Andria, xxxi, n. 3), synizesis is confidently restricted to those cases in which a short vowel is 'subordinated' to a following long one. Thus, with reference to the supposed form meus, eu is not a regular Latin diphthong, but is of secondary origin in seu, neu, and the result of reduction from sive, neve; even in the case of genuine compounds, neutiquam is pronounced nyutiquam, něũter is regularly a trisyllable (Consentius, Keil, V, 389, 28), etc. A similar disproof might be given of the other supposed combinations also, which in no case involve the contraction of qualitatively similar vowels such as is seen in nēmo, dēsse, debeo, nīl, cōnesto, etc. Further, pūri, Lucr. iv, 1026, does not represent puĕri, as L. Müller supposes (R.M.2 298), but must be referred to the adj. pūrus, as Munro has shown; puĕritiem, Auson. Prof. 10, 17, is not to be explained as a case of synizesis, but is a proceleusmaticus used in place of a dactyl, since resolution of the arsis is sometimes allowed in both the early and the late hexameter, e.g. Enn. A. 267 M. cápitibu(s); CIL. I, 542 fácilia fáxseis; cf. Priscian, Keil, II, 14; Christ, Metr.2 145; Exon, Hermathena, XIII, 157 f. Finally, it may be added that examples of iambic měús, tuos, with final s making position, are excessively rare in the dramatists, e.g. Hau. 219 non út měūs, qui (Fleck.: non út měúst, qui).

1

5

2

1 These last are retained in Neue, II3, 371 ff., no doubt through lack of careful revision.

2 Similarly the Augustan poets do not admit synizesis in the case of two short vowels unless the second is long by position or by pause, eg. Verg. Ecl. vi, 30, Orph(e)a, where the short a is lengthened by pause; cf. also L. Müller, Einl. zu Sat. des Horaz, xxvii, § 3. Hence dein and proin, which always show synizesis, occur only before consonants (Skutsch, Forsch. 87 ff.; Birt, Rhein. Mus. LI, 267 f.).

6

8 Similarly it is almost sheer wantonness for L. Müller to interpret pūrē as puere in Lucil. xxvi, 83 M., where Marx now reads (xxvi, 662): laútum e ménsa púrē cápturús cibúm.

(2) Synizesis in Cases of Elision. Synizesis occurs both when the iambic sequence is contained in a single word, as in tuō, and when, as the result of elision, it is contained in two words, as in (u)o arbitrátu, Cap. 867 (for numerous other examples, see C. Müller, Pl. Pr. 457 ff.). In the first class of cases, the use of synizesis is relatively more frequent, and occurs with the possessives, as we have seen, about nine times as frequently as the dissyllabic measurement. In cases of the second class, however, where a small fragment of the 'bruised' vowel is doubtless retained in pronunciation and intervenes between the two syllables of the iambic sequence, the employment of synizesis is only twice as frequent as the dissyllabic measurement; thus m(e)o, t(u)o, s(u)o arbitratu occur in Pl. twelve times, meo arbitratu occurs six times; m(e)am uxorem occurs seven times, měam uxorem four times. The examples of supposed méa uxor, which are cited by Ahlberg, De corrept. Pl. 70, should be read m(e)a úxor, as is necessary in Am. 522 m(e)a uxor. For qu(i)a in elision, cf. p. 180, n. 3.

(3) Principle of C. F. W. Müller: Total Elision of Synizesis Forms not Allowable before a Short Syllable. Müller correctly observed (Pl. Pr. 457 f.) that a combination like meo ănimo is necessarily always read as tetrasyllabic in early Latin verse; since then a combination like meo arbitratu may always, or almost always, be conceivably pentasyllabic, i.e. read as mão arbitratu, he concluded that the supposed monosyllabic forms never suffer 'total elision,' and are therefore never really treated like monosyllabic forms in any particular. Müller's conclusion, however, is manifestly lacking in logical cogency, and is not warranted by his premises. For he was only warranted in concluding with certainty that these forms do not suffer 'total elision' before a short vowel.1 This latter is undoubtedly the case, for the vocal organs experienced no special difficulty in pronouncing the sequence meo

1 Compare the careful observations of Bömer, .., 43, in refutation of Müller's view: “Maximam offensionem huiusmodi vocabuli synalœphe cum brevi vocali habebat Synaloephe cum longa vocali minus erat insueta. Loci, quales sunt mea Antiphila, meo arbitratu, etc., saepius inveniuntur."

6

7

ănimo, and hence had no occasion to seek a special relief. Nevertheless a few apparently well-attested cases (cf. Hauler, Einl. zu Phor. 56, n. 6) before a short vowel are found in the whole drama, viz. St. 39 pól m(eo) animo ómnis (anap. dim.); 275 núnc m(eae) erǎe núntiábo; Tri. 724; Mi. 262; Poe. 1070; Cap. 666; Cas. 542 (?); Titin. tog. 40; Hec. 238 (Umpf. and Dz., following A). These examples may either be rejected entirely in view of their extreme rarity (Skutsch, Sat. Viadr. 143; Ahlberg, De procel. I, 91),1 or they may be considered as evidences of the close approach of meō to a genuine monosyllabic pronunciation, or- the most probable solution in my judgment - they may be interpreted, like Lucretius's or(i)undi, simply as cases of the total suppression of a weak semivowel in hiatus; cf. p. 169 above.

(4) Doubtful Existence of Monosyllabic Stem *SO-. - The statement is commonly made on the authority of Festus that, in addition to the usual forms of the possessive, there existed in O. Lat. the forms sām, sõs, sās, sīs, etc., and that these latter were freely used by Ennius. Comparative grammarians have been dubious about identifying these O. Lat. forms with the synizesis-results s(u)ām, s(u)ōs, and have preferred to connect them with the I.-Eur. monosyllabic stem *suo-, Skt. svá-s, Gk. ofó-s, etc.; cf. Stolz, Müller's Handb. II3, 2, 137; Lindsay, L.L. 426; Sommer, Lat. Lautlehre, 445, § 279.2 They have thus been led tacitly to assume the existence in historical Latin of *sos and *să in the nom. sing. as well as of sām, sās, sōs in the oblique cases. Unfortunately for this

1 Cf. also Skutsch, Tépas, III.

2 Very many scholars, however, as L. Müller, R.M.2 322, 297, and Neue, II3, 366, 369, 371, are content to accept the explanation sis (Enn., Lucr.) = suis; sõ CIL. V, 2007) = suo; mieis (CIL. I, 38), i.e. mīs = meis; tīs (Inser. Or. 4847) = tuis; cf. also Lindsay, l./., 268, and Sommer, l.l., 446. The particular identifications just named are, in my judgment, undoubtedly correct. The Plautus Mss also sometimes indicate the monosyllabic pronunciation of the dat.-abl. pl. meis by the orthography mieis, miis, or mis; for examples, cf. Neue, II3, 366, and Sturtevant, .., 35. In addition, we find in the vulgar language not only the barbarous form suobus (dat.-abl. pl.), but also sybus (CIL. VI, 26896). The latter is probably not formed after sibi, as Stolz (Müller's Handb. II3, 2, 134) suggests, but after sis; cf. dibus (CIL. VI, 214), formed after dis.

hypothesis, we find only the latter forms ascribed to Ennius, and certainly there is no trace whatever of the forms *sos, să in the dramatists; for had they been in colloquial use, the Plautine plays would have been filled with examples like the following: út sos-ămicus dicerét, ut sos erus aúferát, út marepóscam múnerá, út sa-děpéndít, cf. út (e)ă dixit, etc. In other words, at the very point where the monosyllabic possessive stem might demonstrate its independent existence, no traces of it are found either in Ennius or in Pl. I am led, however, to the conclusion that if the alleged literary (not vulgar) forms sām, sās were used by Ennius, they cannot be certainly identified with the synizesis-forms suam, suas, which are not absolute monosyllables, nor can they be certainly referred to Ennius's well-known attempt to introduce phonetic or quasi-phonetic principles in the writing of Latin words, as Édon, Écriture du lat. populaire, 82, would refer them. Yet, with the exception of the abl. pl. sīs2—an undoubted synizesis form—which is itself probably only quasiphonetic, but is warranted by the analogy of dīs, īs, mieis, miis, mis, tis (vulgar), these alleged forms are very poorly attested and seem likely to have arisen from a palpable misinterpretation by Festus (repeated later by Sch. on Per. i, 108) of Enn. A. 102 M: Virgines nam sibi quisque domi Romanus habet sas. In commenting on this passage, Verrius Flaccus had expressly explained O. Lat. sās (from the demonstrative stem *so- *sā) as used in place of eas,3 as indeed the

1 The above statement should be put more strongly; for the only monosyllabic forms which would come directly from the I.-Eur. stem *suo-, viz. the forms sos and som, are the missing ones. Cf. Sommer, l.l., 445: “*Suo- musste nach § 94, 2, lateinisch in allen Formen, wo o auf folgte, zu să- werden; diese Gestalt wurde dann durchs ganze Paradigma durchgeführt."

2 Enn. A. 141 M.: Postquam lumina sis oculis bonus Ancus reliquit ; cf. Lucr. iii, 1023.

8 Fest. 476, 17 Th.: sas Verrius putat significare eas teste Ennio qui dicit, etc., . . ., cum suas magis significare videatur. With the exception of this misinterpreted passage, Festus quotes no actual verses of Ennius for the alleged monosyllabic possessives sam, sas, sos, but we find the quasi-monosyllabic form spelled suos, in the Mss of Festus, 324, 17 Th., Paulus, 325, 6, and Nonius, 158, 20: (Dis) Poeni soliti suos sacrificare puellos (A. 233 M.); cf. also Lindsay, L.L. 429. – Verrius's interpretation of sas as eas in the passage quoted above is accepted also

sense requires, but Festus evidently misunderstood the meaning of the verse, and was thus led to ascribe to Ennius false possessive forms sas, sos, sam, etc.; for a similar view, see L. Müller's note ad loc.

(5) Retention of a Pair of Shorts through Logical Analogy. - I have already spoken of the absence of slurring in anapaestic groups like sed-či, ab-eis, in-eisdem (p. 183). It should be further noted that the monosyllabic forms and is had apparently not established themselves in all uses in the time of Pl., e.g. they were not in use after a short monosyllable, and there are no well-attested cases in the drama of sed i (dicunt), ab is (véniunt), with iambic shortening. At a later period, however-probably first in the Ciceronian agethe simple forms and the group-forms of is and idem became more fully assimilated; thus compare Pl. Mi. 758, se1d (c)idém, with Hor. C. iii, 2, 27, sub ísdem, and Manil., Astron. iii, 73, sémper ut (e)ídem (dat. sing.). Owing to a tendency, which is not consistently carried out, to interpret Ms is as is, the ed. min. gives here at times un-Plautine forms, but quid-iis (Mss), or quid-cis is necessary Poe. 167, and ut-iís (Mss), or ut-eis, should be read Am. 68; Men. 972; Ru. 647, etc., as well as in his (CD), Ps. 1109 (ed. min.: in is); of very doubtful scansion are Mo. 862, Ps. 1111, and the scansions of the ed. min., sed i, neque is, could only be defended as a license of anap. verse.1 In short, at this period all the group-forms of is were trisyllabic like sed-čo, and the intrusion of a dissyl

by Lachmann (on Lucr. vi, 1067) and by Vahlen in his second edition of Ennius (Leipzig, 1903), who explains: “virgines, nam sibi quisque eas domi Romanus habet, reddi non possunt." Skutsch (T'épas, 144; Pauly-Wissowa, Realencyclopäd. VI, 2625) attempts to explain sas as the possessive pronoun in this passage, but characterizes this use of the word as an arbitrary and tasteless invention of Ennius in imitation of Homeric ös (eine Erfindung des Ennius... willkürlich ... krass'). This supposition, which confessedly does little honor to Ennius, is also quite unnecessary; it had been put forward before by L. Müller (K.M2 322), but was afterwards definitely abandoned by him. On the other hand, L. Müller explains sis as a genuine synizesis form.

1 On the other hand, of the seven necessary cases (inclusive of Men. 972) in the drama of dissyllabic nom. pl. m. and abl. pl. of is, which are cited by Sturtevant, l., 24, one is the fourth foot of the senarius, two are verse-closes, and four are trisyllabic groups.