Page images
PDF
EPUB

res, ipsá-re, illám-rem, cf. CIL. I, 206 (Lex Iulia Munic., 45 B.C.) 44 eamrim, eare, ib. 161 caires; see further my article in Trans. Am. Phil. Assoc. XXXV, 39, and also G. Ramain, Études sur les groupes de mots, Paris, 1904, pp. 151, 161, 176. I have examined for this purpose all the occurrences in the dramatists of the single case-form rem (acc. sing.), when immediately preceded by meam, tuam, suam, and of the twenty-six cases which actually occur, I find that twenty-two cases, occurring in iambic or trochaic verse, show pretonic syncope; the text or kind of verse is uncertain in two cases (Au. 134; Tri. 139), but appears also to yield the syncopated form, and in two cases only, one in iambic, one in anapaestic, verse (Poe. 659; Per. 781), are the full forms retained. The yévos diλáσlov, which closely represents actual speech, therefore shows pretonic syncope to be more than twenty times as frequent as the original iambic form. The accent tůăm-rem would of course be possible in a few cases, but not in many, e.g. in the dramatists the accent malám-rem is certain in twenty-eight of the thirty occurrences of this locution, while mălăm-rem occurs once in anapaests (Ps. 234) and once in a corrupt passage (Tru. 937). On the other hand, when tuam precedes rcm but is separated by intervening words, synizesis occurs in only half the cases (contrast Poe. 1083; Tru. 965; Tri. 139(?) with Mer. 1011; Ps. 253; Tri. 621), but this low ratio is accidental and is quite unusual in other cases. The examples of meamrem in the dramatists are as follows: Am. 827 t(u)a1m rem cúret; Au. 134 ut(í) t(u)àm rèm (bacch., acc. to ed. mai.); ib. 299 s(u)am rẻm periísse; Cap. 632 m(e)a1m rem nón curés; Cas. 89 non mihi licére m(e)ám rem; Ci. 81 nón potést s(u)a3m rem; Cu. 671 m(e)am rem ágere; Mer. 48 válide s(u)ám rem;

1

3

[blocks in formation]

ib. 454 s(u)am rem esse aéquomst; Mi. 951 t(u)a2m rem cúra; ib. 1117 t(u)ám rem túte agás; Per. 513 ad m(e)a3m

4

[ocr errors]

rem; ib. 781 itaqué měam rém (anap. oct.); Poe. 675 t(u)ám rem tú ages; ib. 659 áge[re] tŭám rem: occásióst; 679 t(u)ám rem, | ádulescéns, loquí; 750 éxplicávi m(e)ám rem;

3

4

1

3

3

3

3

1

4

Ps. 496 m(e)ám rem sápio; Ru. 1399 tun m(e)a3m rem símulas; Tri. 139 s(u)ám rem mélius gésserít (so ed. mai. after Hermann; Mss: s(u)ám meliús rem gésserít); ib. 327 s(u)a3m rem tractavít; 1083 quoi t(u)a3m rem; Phor. 467 álios t(u)a2m rem; Ad. 771 ac t(u)ám rem; Caecil. com. fr. 46 m(e)a3m rem iam omnem; Enn. tr. fr. 220 s(u)a2m rem béne gessére. Similarly we have synizesis in the case-forms m(e)ae-res (nom. pl.), m(e)as-res (acc. pl.) three times: Phor. 820; Tri. 269 (anap.); ib. 446;—in a fourth case, Ci. 719, the scansion is doubtful: egó tibi m(e)ás res, or ego tíbi měás res. Finally, we may take a case-form of the same word where Latin accent-laws would allow us to use shortening, i.e. meis rebus. Three cases occur, always with dimoric pronoun: Cap. 968; Cas. 938 (dactylic); Tri. 446:

6

5

[blocks in formation]

6

Shall we then accept always m(e)ás-res, t(u)ás-res, but decide to scan tuis rébus with shortening? Shall we accept always m(e)órum, (e)ámus, but admit with shortening meo, čunt? Scarcely, I think, unless we are prepared at the same time to give up all belief in the existence of phonetic laws and the regularity of phonetic processes. Still another word-group may be mentioned here. It is well known that the phrase dí vostrám fidem has acquired an accent of its own (Trans. Am. Phil. Assoc. XXXV, 39, and O. Brugman, De iamb. senario, 30) and it is quite evident that the abbreviated phrase t(u)ám-fidem, used in invocation of a single god, is similarly accented, e.g. Au. 692 Iunó Lucína, t(u)ám fidem; Cu. 196 t(u)a1m fidém, Venus. We find also apúd-m(c)as, after the model of the familiar aputme, aputvos, on the monument of Caecilius (CIL. I, 1006, 2): hospés gratúm est, quom apúd meas réstitístei seédes (saturnian), where apud meas, like apud forum, sine mõdõ (A. J.P. XXV, 417) would be inadmissible.

Definite metrical proof of synizesis is also afforded by examples like the following of ea-re, eam-rem (I quote only a few of the numerous cases): Men. prol. 37 Syracúsas de (e)á-re

8

4

4

rédiit; Am. 1087 de (e)a1re sígna; Au. 799 (e)a1-re répudiúm; Phor. 444 quid de (e)á-re; And. 385 ex (e)á-re quíd fiát; Ep. 565 i3lle (e)am-rem ádeo; Mer. 926 (e)a1m-rem núnc exquírit.1 The dative (e)i-rei is inseparable in Pl., and always dissyllabic, the only doubtful case being Men. 234 (Seyffert, Stud. Pl. 25, n. 17). Similarly Am. 1023 quomodo? || Eo2 modo, ut; cf. quomodo, omnimodo, nullomodo, etc., and Trans. Am. Phil. Assoc. XXXV, 48, n. I. Since the demonstrative pronouns are proclitic in general (ib. 36 ff.), I may cite also Per. 194 (e)a fidé; cf. also Mer. 74 atque (e)á pecúnia; Poe. 2 ex (e)á tragoédia; Cu. 551 (i)is tabéllis. Inadmissible would be both the dactylic word éx-ěă and the dactylic foot (__ ~ ~) before the diaeresis in Mi. 246 né titubét, si (ex)quíret e1x (e)a | míles;2 cf. also Tri. 742 ex (e)á; And. 719 vérum ex (e)ó nunc mísera. The full iambic forms of is are rarely used within the verse, and are not even very freely placed in the verse-close (Tri. 405; Ci. 611; Men. 86; Mer. 719, 766, 869, etc.). Finally, the weak vowel is perhaps lost entirely in the compounds eccum (from ecce eúm), eccam, eccos, eccas, and, through analogy, this loss is extended even to the new formation eccă (fem. sing. and nom. pl.), although in the simple we have only ěă, or, at the most, (e)a in elision before a long vowel.3

3

1

Definite proof is also given by the word-groups which quidem forms with possessive pronouns, and which are quite similar to those which it forms with the dissyllabic personal pronouns, i.e. mihi-quidem and mihi-quidem, tibi-quidem and tibi-quidem (but not mihi quidem, tibi quidem). Quidem, when joined to a dissyllabic word, has of course no power of

1 Apart from the very frequent composita ăd-ĕam-rem, ob (propter) ĕam-rem, — cf. hic-ěam-rem (Au. 201),— the full acc, čam-rem occurs in Pl. only in bacch. verse (Mo. 88; Cap. 502), and once in iambic (Ru. prol. 19, cf. Ter. And. 442). 2 Similarly Bömer, .., 42, quotes two verses from Terence in which the forms t(u)ā and t(u)īs are necessary in order to avoid a dactylic foot before the diaeresis: Phor. 1016 nám neque néglegéntia1 t(u)a | néque; ib. 543 nón triúmpho, ex núptii's t(u)is | sí.

3 Stowasser's derivation of eccum from ecc' hum is, however, usually accepted at present, and may well be correct, especially as eum is not really monosyllabic.

shortening it; hence, méumquidem - the scansion which Luchs inadvertently adopts in his discussion (Comment pros. Pl. I, 16 ff.) —is impossible as a regular verse-accentuation in the place of meúm-quidem. Luchs cites four examples in which the possessive has trimoric value (e.g. měá-quidem hércle caúsa, Men. 727; Ru. 139; Poe. 573; Ru. 737), but twenty-two cases of dimoric value: m(e)úm-quidém, Tru. 963; m(e)úm-quidem édepol, As. 190; m(e)ó-quidem ánimo, Au. 478; Ba. 102; Cas. 570; Cu. 499, 514; Men. 200; Mer. 314; Poe. 232; Ru. 1038; Au. 539; cf. Ba. 394;1 m(e)ã-quidém senténtiá, Cas. 563; Men. 81; Poe. 1338; m(e)á-quidem, As. 275; Men. 1029; Per. 537; Tru. 560; t(u)á-quidem, Men. 792; s(u)ám-quidem, Mo. 894. A similar accent is shown by the composita which the possessives form with sibi, mihi, e.g. suōsibi, mea-mihi; cf. the references given in my article A.J.P. XXV, 407, and also Ribbeck, Com. Frgm2, Coroll. xxxiv. Thus we find fully sŭố-sibi, Cap. 81, and even in the nom. sing. tuós tibi sérvos, Ba. 994, although the accents meús, tuós are otherwise almost entirely unknown. The usual scansion, however, is s(u)o-sibi, m(e)á-mihi, e.g. Tru. 698; Cap. prol. 46, 50; Per. 81; Poc. prol. 57, 487; As. 825; Am. 269; cf. Ad. 958, etc.

Further Illustrations from Possessive Pronouns. - It seems necessary at this point, even at the risk of repeating former statements and of making my discussion seem needlessly diffuse, that I should fully explain the inferences which I draw from the facts mentioned in the previous section. The argument which is based upon the occurrence in the γένος διπλάσlov of twenty-two cases of tetramoric tuam-rem to one case of pentamoric tuam-rem may be stated as follows: Whatever occurred in tetramoric tuam-rem, was evidently very acceptable to the dramatists; for they used the results of the process with the greatest freedom. It is true that they have these words in other arrangements, and that tuam . . . rem occurs seven times, rem tuam ten times (including six

...

1 Of doubtful correctness is the scansion of ed. min. Ep. 111 e3s m(e)ŏquidem ánimo.

cases in the verse-close), and rem . . . tuam eleven times (in verse-close only), yet tuam-rem remains the regular and the preferred order, and it is clear that the dramatists have made no special effort to separate the two words by a very free use of tmesis-forms, such as they sometimes employ for metrical convenience, e.g. in the case of manu emittere, animum advortere, magno opere, quapropter.1 We have seen, however, that the process cannot be shortening (p. 187); hence, two conclusions follow: (1) The process employed is synizesis. (2) The frequency of its occurrence shows that synizesis was a favorite use of common speech. It is clear also that the approximate pretonic syncope which is seen in the phrase was first produced through the prose accent t(u)ám-rem, but after this syncope was once definitely produced in the spoken language, it became possible to employ also the verse-accent t(u)am-rém, without affecting in any way the result. On the other hand, when meam stood last in the prose sentence, as in curo rem meam, there could be no slurring, but only the full pronunciation of the form, e.g. měām, in this non-proclitic position, and this remained the case also at the close of a metrical sentence.

Before I enter upon the discussion of other cases of slurring in the sentence, it will be necessary to speak briefly of the position of the possessive in O. Lat., and especially of its position with nouns denoting relationship, since I shall draw my illustrations chiefly from nouns of this class. In classical prose, the possessive is usually placed after its substantive, but, for greater emphasis, may be placed before it (Albrecht, De adiectivi collocat., Marburg, 1890, p. 8 ff.). Nilsson, however, in his careful study, Quomodo pronomina ap. Pl. et Ter. collocentur, Lund, 1901, p. 13 ff., has shown that this rule does not hold good for O. Lat., and that while the possessive has really no preferred or definite position in O. Lat., yet, like the adjective, it tends to retain its original position

1 Or in the case of verumtamen (Seyffert, Bursian's Jahresb. 1894, 317), nemo homo (Asmus, De apposit. collocat. 21), quidem hercle (Kellerhoff, Studem. Stud. II, 64), etc.

« PreviousContinue »