» UNIVERSITY OF CALIFORNIA PUBLICATIONS CLASSICAL PHILOLOGY 'ol. 1, No. 7, pp. 205-262 Junes 190? •g t JTH o ^^^=i o ^==— z_ I 1 5 3 ^^^=| , ^p^ ji ^Sc 5 ^S3 ", 6 <" SOME PHASES OF THE RELATION OF THOUGHT TO VERSE IN PLAUTUS HENRY W. PRESCOTT BERKELEY THE UNIVERSITY PRESS UNIVERSITY OF CALIFORNIA PUBLICATIONS Note.— The University of California Publications are offered in exchange for the publications of learned societies and institutions, universities and libraries. Complete lists of all the publications of the University will be sent upon request. For sample copies, lists of publications or other information, address the Manager of the University Press, Berkeley, California, U. S. A. All matter sent in exchange should be addressed to The Exchange Department, University Library, Berkeley, California, U. S. A. CLASSICAL PHILOLOGY.— Edward B. Clapp, William A. Merrill, Herbert C. Nutting, Editors. Price per volume $2.00. Volume I (in progress). No. 1. Hiatus in Greek Melic Poetry, by Edward B. Clapp. Price, $0.50 No. 2. Studies in the Si-clause, by Herbert C. Nutting. . . " .60 No. 3. The Whence and Whither of the Modern Science of Lan- guage, by Benj. Ide Wheeler " .25 No. 4. On the Influence of Lucretius on Horace, by William A. Merrill .......... " .25 No. 5. The Priests of Asklepios (A New Method of Dating Athenian Archons), by William Scott Ferguson .... " .50 No. 6. Horace's Alcaic Strophe, by Leon Josiah Richardson . " .25 No. 7. Some Phases of the Relation of Thought to Verse in Plautus, by Henry W. Prescott . . " .50 The following series in Graeco-Roman Archaeology, Egyptian Archaeology, Ameri- can Archaeology and Ethnology and Anthropological Memoirs are publications from the Department of Anthropology: GRAECO-ROMAN ARCHAEOLOGY. (Large Octavo). Vol. 1. The Tebtunis Papyri, Part 1. Edited by Bernard P. Grenfell, Arthur S. Hunt, and J. Gilbart Smyly. Pages 690, Plates 9, 1902 Price, $16.00 Vol. 2. The Tebtunis Papyri, Part 2 (in press). Vol. 3. The Tebtunis Papyri, Part 3 (in preparation). EGYPTIAN ARCHAEOLOGY. (Quarto). Vol. 1. The Hearst Medical Papyrus. Edited by G. A. Reisner. Hieratic text in 17 facsimile plates in collotype, with introduction and vocabulary, pages 48, 1905. (J. C. Hinrichs, Leipzig, 25 Marks) . . . Price, $8.00 AMERICAN ARCHAEOLOGY AND ETHNOLOGY. (Octavo). Vol.1. 1903-1904. 378 pp., with 30 plates Price, $4.25 Vol. 2. 1904-1907. 392 pp., with 21 plates and map. . . Price, 3.50 Vol. 3. The Morphology of the Hupa Language, by Pliny Earle Goddard. Pages 344, June, 1905 Price, 3.50 Vol. 4. No. 1. The Earliest Historical Relations between Mexico and Japan, by Zelia Nuttall. Pages 47, April, 1906. . Price, .50 No. 2. Contributions to the Physical Anthropology of California, by A. Hrdlicka. Pages 16, Tables 5, Plates 10, June, 1906. Price, .75 No. 3. Shoshonean Dialects of California, by A. L. Kroeber. Pages 100, February, 1907 Price, 1.50 No. 4. Indian Myths of South Central California, by A. L. Kroeber. Pages 84, May 1907. Price, .75 Vol. 5. No. 1. The Phonology of the Hupa Language: Part I, The Indi- vidual Sounds, by Pliny Earle Goddard. Pages 20, Plates 8, March, 1907 Price, .35 Vol. 6. The Ethno-Geography of the Porno Indians, by S. A. Barrett (in press). Vol. 7. No. 1. The Emeryville Shellmound, by Max Uhle. Pages 106, Plates 12, June, 1907 Price, 1.25 UNIVERSITY OF CALIFORNIA PUBLICATIONS IN CLASSICAL PHILOLOGY Vol. 1, No. 7, pp. 205-262. June 17, 1907 SOME PHASES OF THE RELATION OF THOUGHT TO VERSE IN PLAUTUS BY HENRY W. PRESCOTT. In his study of the Saturnian verse Leo has recently stated his conception of the relation of thought to verse in early Latin poetry: " in early Latin verse," Leo says, with reference espec- ially to the Saturnian, " verse and sentence are identical; art- poetry in its beginnings" (and he refers to Plautus as illustra- tive of the principle)," when sentence-structure was developing, resisted this inherent recpiirement and limited itself to the norm by which words in the sentence intimately connected in thought were not separated by the verse unless the separation was justi- fied by special considerations: externally, by reason of length, or by the colligation of words through alliteration or other means of connection; internally, by reason of emphasis or some stylistic effect of the word thus separated." 1 xtl Vers und Satz fallen urspriinglich zusaminen; . . . Die Kunstporsir hat in ihren Anfiingen, wie sich die Satzbildung lniiehtig entwickelte, mit dieser der Pocsie innewohnenden Forderuug gekiinipft und sie auf die Norm besclnankt, dass im Satze eng zusaminengekorige Worter nicht durch den Vers getrennt werden diirfen, werni sick nickt die Trennun^ dureh einen besonderen Umstand als berechtigt erweist; aiisserlick durch Liinge, durck allitterirende oder andere einander suchende und anziehende Wortverbindungen. innerlick durck Nackdruck oder sonst stilistiscke Absickt des jrcsoiulprten Worts. So ersckeint der Gebrauck bei Plautus ausgebildet. " Der saturniscke Vers 14 = Abkandl. Gutting. Gesell. (1905). In 1S81 Buecheler reminded Schoell tkat only prouoininal adjectives were separated from tkeir nouns by tke verse-end, that almost no other adjectives were so treated, in the text of Plautus (Truculent us, ed. Sckoell, praefatio XLV, n. 1). Buecheler repeated this admonition in 206 University of California Publications. [Class Phil. Leo has left to others the task of testing the validity of his law. I have attempted to gather and study the evidence offered by one group of examples in Plautus, the cases in which adjectival words, whether ordinary attributives, pronominal adjectives, or numerals, are separated from their substantives by the verse. In many respects the study must be descriptive: the lack of similar studies in Greek poetry, and the fragmentary remains of earlier Latin poetry, usually of uncertain metrical constitution, retard a convincing account of Plautus 's position in the histor- ical development of verse-technique. Nor will it be just to con- firm or refute Leo's theory until other phases of the problem in Plautus, and the corresponding phenomena in Greek poetry are investigated. For the present, the study may suggest points of view and methods of approach, which will doubtless need read- justment as the problem is studied in its larger aspects. Among the features that Leo enumerates as justifying separa- tion is length: this element may be a matter of syllables, or in addition to syllables may include an extension of thought. That is, a given word may be long, or a thought-unit involving several words may be long. In either case, it is not at once clear that Length occasions the separation. If, however, as appears to be Rh. Mus. 41 (1887) 312. In 1893 Appulm published his dissertation: Quaestiones Plautinae. Quae rationes inter versus singulos sententiasque intercedant Plauti exemplo comprobatur (Marburg). Interpretative analysis was impossible in this attempt to cover a Large field within the Compass of a doctor's dissertation. Nbrden summarizes the usage of Vergil in Aeneis Buch VI, 390-391. For references to studies of the general question of the collocation of words, as well as of the special question under consideration, cf. the same work 382 n. 1, and the same author's Die antike Kunstprosa I 68 n. 1. In the present paper the song-measures are excluded; I have not know- ingly included examples from such passages except for comparative pur- poses, ami then their provenance is stated. I may be open to criticism in nut dividing the material with reference to the metre of the verses con- cerned; bul the results show no important differences between the tech- nique of the iambic and trochaic verses, or of the shorter and longer verses, except such as may more convenient ly be described parenthetically, and a metrical classification interferes with clearness of presentation. Vol. l] Prescott.— Thought and Verst in Plautus. 207 the case, 2 words of five or more syllables that are metrically suit- able regularly tend to the end of the verse, or less frequently to the beginning, it follows that, if such a word is a substantive or adjective, the difficulties in combining the two members of the pair in one verse are much greater than they otherwise would be. And similarly, a thought-unit consisting of a substantive and several adjectives, wherever they may be disposed in the verse, will by reason of the number of syllables, easily overflow into the next verse. In a thorough treatment of Leo's theory predicative expres- sions should be included. The consciousness of verse-unity could not be better illustrated than in these two couplets : isque hie compressit virginem adulescentulus (vi), vinulentus, multa nocte, in via. (Cist. 158) quom hasce herbas huius modi in suom alvom eongerunt formidulosas dictu, non essu modo. (Ps. 823) But such cases of predicative expressions, involving long words, are apart from our immediate purpose. There are, however, a few cases of adjectives following their substantives (either adjec- tive or substantive is of great length) and not so clearly pre- dicative. Their position makes it possible that they amplify the meaning, in which case this amplifying force as well as length justify the separation. Most of these adjectives are derived from proper nouns; and since in almost all cases the adjectives stand at the beginning of the second verse it is significant to note that in Oscan and Umbrian proper adjectives usually follow their nouns : 3 Philopolemum vivom, salvom'et sospitem vidi in publica celoce, ibidemque ilium aduleacentulum Aleum una et tuom Stalagmum servom (Capt. 873) - In the Mostellaria, for example, out of 90 cases of words of five or more syllables, two-thirds stand at the end of a verse; of the remain- ing third all but two are metrically impossible at the end. On the other hand words of four syllables are freely disposed in the interior of the verse. Five syllables is, therefore, assumed to be the minimum of length which may be regarded as offering difficulty. 3 Nilsson, de collocatione pron. adj. apud Plant nni et Terentium 10 = bunds Pniversitets Ars-skrift 37 (1901). 208 University of California Publications. [Class Phil. non ego te : i ■ 1 ilium duco dentatum virum Macedoniensem, qui te nunc flentem facit: (Ps. 104n ) quern propter, o mea vita? — propter militem Babyloniensem, qui quasi uxorem sibi (True. 391) sed illi patruo huiua qui vivit senex Carthaginiensi duae fuere filiae: (Poen. 83) These examples are of somewhat different value. In the first case, the length of adulescentulum and its consequent position (of fifteen occurrences of the word thirteen are at the end of the verse) are the controlling factors : Aleum is no more amplifying than in vs. 169 of the same play (nam cecum hie captivom adulescentem (intus) Aleum, \ prognatum genere sunnno et sum m is (I if iis) where the adjective is kept in the same verse with its shorter noun. The next two examples are alike in having the separated adjective followed by the caesural pause and an ex- planatory gwi-clause. 4 In the last example, too, we have the caesura! pause. Plautine usage of these adjectives points to length as the influential factor. Carthaginiensis occurs only at the beginning of the verse (Poen. 59, 84, 963, 997, 1377) with one exception (1124). Babyloniensis is less constant: at the beginning in True. 81, penultimate word in True. 203 (here, however, iambic septenarius; in the other cases, senarii) ; in all three cases the same phrase occurs. So we get militem j Baby- loniensem (391), | Babyloniensem militem (81), Babyloniensis mil' s | (203). It is clear that length and metrical conditions are potent. Macedonicnsis does not occur again: Mace don i us takes its place (Ps. 51, 316, 616, 1090, 1152, 1162), and in all the cases except one (346) it stands at the end, different metrical constitution making it convenient in that position: in all the cases of Macedonius, however, separation is avoided except in the following couplet: 'ili minis constitutes an almost inseparable The alliteration in Ep. 707 is also to be noted. 1 < If. mutuos, below, p. 23 \. Vol. l] Prcscott. — Thought and Verse in Plautus. 211 unit (usually at the end or beginning of a verse) adds to the difficulty. This brings us to examples of long thought-units Such thought-units may be of two sorts : a substantive attended by a succession of adjectives of equal value, e. g. "a long, lean, rascally, devil of a fellow " ; or a substantive accompanied by attributive modifiers of unequal value, e. g. "my own twin sister." Our author is fond of billingsgate, and offers a richer store of the first variety of compounds than we may quote. In general it may be said that such a succession of adjectives is usually so disposed as to accentuate the unity of the verses: the substantive usually precedes or is embraced between groups of attributives ; the thought is in a measure complete, and the vir- tues or vices or indifferent qualities either run over into several verses or occasionally are bound within a single verse, in either case without serious disturbance of verse-unity. A few examples will illustrate these characteristics: nisi mihi supplicium virgeum (MSS. virgarum) de te datur longum, diutinumque, a mane ad vesperum. (M. G. 502) stat propter virum fortem atque fortunatum et forma regia. (M. G. 9, cf. 56-57) ecquem recalvom ad Silanum senem, statutum, ventriosuin, tortis superciliis, contraeta fronte, fraudulentum, deorum odium atque hominum, malum, mali viti probrique plenum, qui dueeret mulierculas duas secum satis venustas? (Rud. 316) For other examples, Bacch. 280 (if Leo's strigosum is accepted), Cas. 767, Men. 402, 487, M. G. 88, Ps. 724, 974, Rud. 125, 313, True. 287. In the examples quoted other obvious features will be noticed : in the first, intensification of one idea in one verse ; in the second, initial rhyme. There are a few cases of a succes- sion of two or three adjectives in which the unity is not so obvious : ut aliquem hominem strenuom benevolentem adducerem ad te. (Ps. 697) post altrinsei'ust securicula ancipes, itidem aurea litterata: ibi matris nomen in securiculast. 10 (Bud. 1158) Cf. Rud. 478, 1156-1157. 212 University of California Publications. [Class Phil. il>i nunc statuam volt dare auream solidam" faciundam ex auro PMlippo, (Cure. 439) Tii all of these the noun and one adjective (or two) stand in the lirst verse so that the thought is practically complete; bcncvolcn- iem, and aurea (as we shall see presently), are metrically con- venient in the places which they occupy; the separated adjectives all stand at the beginnings of their respective verses and are not without emphasis; it is also to he noticed that Jittcrata is explained in the rest of the verse. Of the second variety of thought-units, two occur with sufficient frequency to he of significance. These are the expressions for " own twin sister, brother, son," often accompanied by a pleo- nastic numeral when the expression is in the plural; and the phrase for a sum of money in which nummi aurei Philippi in various arrangements, with an accompany in g numeral or further attribute, makes an elaborate complex. This latter phrase is usually from eight to thirteen syllables in extent, and on five occasions the longer varieties run over into a second verse: 12 sunt tibi intus aurei trecenti minimi Plrilippi? — sescenti quoque. (Poen. 165) qui mini mille minimum crederet PMlippum, (Trim 954) atque etiam Philippum, numeratum illius in mensa maim, mille numnmni. (Trin. 965) hie sunt numerati aurei trecenti minimi qui vocantur Philippei. (Poen. 713) nam dueentis aureis Philippis redemi vitam ex flagitio tuam. (Baech. 1010) On the contrary, in a large majority of cases similar varieties of the same phrase, not always with aureus, are included in a single verse: As. 153, Bacch. 230, 590, 882, 934, 1026, Poen. 670, 732. Trin. 152, 959, 1158. 13 11 The proximity of faciundam gives solidam predicative force in our passage: cf. < icero, de div. I, 24, 48. ; -'lii i'ers. 438 probi numerati are probably amplifying, as Leo brings out iii liis punctuation: cf. Pers. 52<>. J3 It is not likely that in any of these phrases there was any violent separation (cf. for the usage of the various forms Langen, Beitriige zur Kritik u. Erklarung des Plautus 85 ff., Brix on Trin. 844). At least in Vol. l] Prescott— Thought and Verse in Plautus. 213 There are a dozen instances of the first phrase, including more than six syllables, and of these only two escape into a second verse; these two are of eleven and ten syllables: geminam germanam meant hie sororem esse indaudivi: earn veni quaesitum. (M. G. 441) spes mihi est ves inventurum fratres germanos dims geminos, una matre natos et patre vino uno die. (Men. 1102) The second of these (and possibly the first 14 ) is only apparent separation: geminos is followed by a sense-pause which empha- sizes the idea as amplifying, and the elaboration of the same idea in the rest of the same verse gives a distinct unity to that verse. Indeed, geminus is elsewhere in the same play a sub- stantive : Men. 26, 40, 68, 69, and if the prologue is of dubious authorship in parts, at least once in the play itself, vs. 1120. In nine cases long forms of this complex are confined to a single verse: Amph. 480, cf. 1070, Men. 18, 232, 1082, 1125, M. G. 238, 383, 391, 717. To be sure, our impression that this situation points to a sensitiveness to verse-unity is momentarily disturbed when we find a much shorter form of the same phrase running over : sicut soror eius hue gemina venit Ephesum et mater accersuntque earn. (M. G. 974) Only momentarily, for again gemina may be substantival; Palaestrio may be working upon the soldier very tactfully. the 'separation of nummus Philippus, the use of Philip pus alone, and the examples above (Trin. 954, 965, with qui vocantur Philippei in Poen. 714), suggest that the words are separable, either one amplifying the other. When aureus (convenient at the verse-end, cf. above and As. 153, Bacch. 230, 590, 934, Trin. 1139) is a part of the phrase, the separation seems more violent ; if, however, Bentley 's emendation of Bacch. 230 is right, there would be some evidence of a substantival aureus, similar to the usage of later times; and one should compare the usage of xpvaovc as a substantive without nrari/p in the fragments of Greek comedy: Jacobi, comicae dictionis index s. v. xP vao ^ r >- The separation of aureus is no more than that of a material genitive as in Hipponax, 22, 4: teal aa/ii3a?JoKa HaonepiaKa ml xf/vooi) GTarr/fxir eHqKOvra rovripov voi v" 1 . But Plautus does not separate the genitive nuri in this phrase. 14 The resumptive earn in the same verse with sororem may help to strengthen the unity of the verse. 214 University of California Publications. [Class Phil. slowly unloading his ammunition, "a sister, her twin." (So, perhaps, also in vss. 47.'} -174.) And Leo might add that the alliteration in sicut soror reasserts the unity of the first verse. 15 II. In so far as he overcomes the obstacle offered by length in a large majority of cases, Plautus may be said to show respect for the integrity of the verse. But the poet's aversion to separation or his indifference to verse-unity is best tested by conditions in which there are no obstacles in the length of words or thoughts. Some general considerations will help us to appreciate the ex- amples. In the later Republican prose the substantive is often sepa- rated from its attributive by intervening words, and much more frequently in poetry ; so far as I know, no effort has been made to discover whether such separation is regulated by any laws or not 10 — whether, for example, certain attributives are more separ- able than others, whether the intervening words are of some special character, etc. Xorden 17 has already pointed out that such separation in early Latin prose is, as regards the number and the nature of the intervening words, subject to limitations. Altenburg 18 has collected the material : usually only one word intervenes, or if more, they constitute a unit of thought. From our present point of view we should like to know whether the attributives themselves show degrees of separability: whether, 15 Under the head of long thought-units should come Ep. 559, in which the genitive and the adjectives constitute an inseparable compound and perhaps account for the escape of mulierem: accipe, aerumnosam et miseriarum compotem mulierem retines. The same would apply to Nonius's reading aerumnarum. 10 Even the interpretation of the material under discussion in this paper would be facilitated lis a study of the collocation of adjective and sub- stantive within the verse, quite apart from the question of separation by the verse. 17 Die antike Kunstprosa 1 179-180, and 180 n. 2. 18 De sermone pedestri Italorum vetustissimo = JHB. Supplbd. 2-1 (1898) 523 ff. Vol. l] Prescott— Thought and Verse in Plautus. 215 for example, the separation of certain pronominal adjectives does not appear earlier than that of ordinary attributives. Perhaps the material is too scanty to lead to convincing gener- alization; tne fact that in Oscan the relative adjective is very regularly separated from its noun and stands at the opposite extreme of the clause lends significance to a similar situation in Plautus. 19 Such observations as Kaibel makes in his study of Aristotle's Athenian Constitution 20 would affect our interpreta- tion of many examples if early Latin prose showed similar char- acteristics: he notes that certain pronominal adjectives are separated from their substantives with greater frequency and by more intervening words than ordinary attributives; he men- tions in the order of such frequency o 7 >to<>, ttSs, oAo?, dAAot, the relative, totos, tis ,• but the last seven are naturally represented only by one or two examples; he also re- fers to numerals, but without mentioning the frequency of separation in such cases. Altenburg's examples show that some of the corresponding words in Latin are separated in early prose. 21 When we add thereto that in Plautus, quite apart from the question of separation within the verse, the cases of separation by the verse and, often, by intervening words as well, show a relatively large number of pronominal adjectives and numerals, we may suspect that some influence made the disturb- ance of verse-unity either less violent or more imperative than it appears to us and than it perhaps was in the case of ordinary attributives: in Plautus 20 per cent, of the cases of separation by the verse-end are pronominal adjectives, 25 per cent, posses- sive adjectives, 15 per cent, numerals. That is, more than half arc pronominal words and numerals. A step towards the explanation of some of these phenomena has been taken by Wackernagel, 22 though without reference to the matter of verse-unity. His investigations in Indogermanic languages, especially Greek and Latin, bring to Light survivals 1U Altenburg, 1. c. 530; Norden, 1. c. I 181 n. 1. -° Stil u. Text der TlnMTEia 'Adr/vaiuv des Aristoteles 99 ff. 21 For example, ceteri, omnes, numerals including nullus, alter, tantus, qui (rel.), quis (indef.). -Indog. Forseh. I 4 IT. ( 'f. Delbriiek, Synt.-ikt. Forseli. Ill 47. 216 University of California Publications. [Class Phil. of a law by which short enclitic words lend to the beginning of ;i sentence, usually to the second place. Pronominal words are often enclitics, and some pronominal adjectives are directly affected by this law. Others arc indirectly affected; for the law of pronominal attraction, combined with Wackernagel's law, will sometimes bring pronominal words that may or may not be enclitics to at least the third place in the sentence. Snch laws have precedence of the natural attraction of the adjective to its substanl ive. A few other laws affect the collocation of words so fundamen- tally that verse-unity must waive its claims, whenever it con- flicts. Words of the same category are attracted to one another. Certain formulas exist for the expression of certain ideas, e. g.. of oaths. Groups of words in Plantns have been studied and peculiarities of collocation discovered. Most of these conditions reflect the usage of ordinary speech. But there are other arti- ficial combinations — whether due to the influence of rhetoric or nof we may not always say — resulting often in the interlocking of words and the consequent separation of words that are syn- tactically connected. All snch factors must he appreciated. Apparent violation of verse-unity may he only conservation of these natural or artificial collocations. 23 Some nf these general considerations account for the separate treatment of ordinary attributives, possessive adjectives, other pronominal adjectives, and numerals. All of them will make more intelligible the discussion of individual passages. In this discussion I do not wish to he understood as represent- ing flic attendant features to he the cause of separation or atone- ment for separation; that would he begging an important ques- tion. In viewing the problem of verse-unity with reference to Leo's theory, it is apparent that the cases of separation are often all ended hy snch features as Leo regards to be justifications for - 3 On the various matters here briefly referred to cf. Langen, Eh. Mus. 12 (1857) 426 ff.; Kellerhof, de collocatione verborum Plautina = Stude- mund-Stud. II 49 ff.; Kampf, de pronominum personaliuro usu et collo- catione ap. poet, scaen. Rom. 16 ff. = Berliner Studien III (1886); Leo, Bemerkungen iiber plautinische Wortstellungen u. Wortgruppen = Nach- richt. Gotting. Gesell. (1895) 416, 432-433; Norden, A.eneis Buch VI, 386. Vol. l] Preseott. — Thought and Verse in Plautus. 217 separation : a descriptive paper notes the appearance of such features. Quite apart from this descriptive treatment is the important question which Leo's theory involves, namely: is Plautus, under the influence of earlier Latin poetry, conscious of verse-unity in the sense that all cases of separation must be justified by special considerations? Granting that these fea- tures attend separation, there is the further question : may any or all of these be proved to be necessarily involved in the relation of thought to verse ? For example, alliteration is inherent in Plautus 's style: may not its appearance have nothing to do with verse-unity?- 4 Furthermore, granting that Plautus is conscious of the individuality of each verse, which may hardly be denied, such consciousness may arise in one of several ways : a poet may be under the influence of a primitive form of verse in which verse and sentence are identical — so Plautus in Leo's theory; or he may be far removed from any such influence and yet pre- serve the unity of the verse — which is not necessarily lost sight of entirely even in advanced stages of verse-development — eithei for the purpose of bringing into relief units of thought, or as a concession to an artificial tendency of his time. 20 On a priori grounds Plautus 's attitude towards verse-unity may well be sus- pected of being affected by the Saturnian verse: he is, however, adapting Greek comedies, and the verse-technique of his Greek sources had reached a much higher point than contemporary Latin verse. This counter-influence must be reckoned with in any a priori reasoning. Leo would be the first to recognize the validity of this contention. None of these important questions is begged-in the following descriptive treatment. Some of them may be considered by way of conclusion, but many of them cannot he settled in a study <>i' a few phases of verse-unity. The division of adjectives is but a small part of word-division, and word-division is but a part of -* Of course the fad thai alliterative groups are usually limited to a single verse in itself shows a consciousness of verse unity. The question at issue is whether a noun or adjective is separated lor the purpose of bringing it into an alliterative group. - : 'Such an artificial preservation of unity appears in Bion: ef. Wilamo- witz, Adonis ;;s_:i»>. 218 University of California Publications. [Class Phil. a larger topic which includes the division of the larger units of thought, phrases and clauses. III. When an attributive follows its substantive it is often possible that the adjective is amplifying; each case must be interpreted with reference to the context, but the mere possibility justifies us in distinguishing between (a) adjectives that follow, and (b) those that precede their substantives. Further classification might be desirable, for example, with reference to whether or not words intervene between the adjective and noun ; but this would confuse the discussion. I have persuaded myself from an inspection of the Mostellaria that the number and the nature of the words that intervene between adjective and noun within the verse are the same in the corresponding situation when a verse-end also intervenes. In some cases it may well be argued thai verse-unity was sacrificed to the normal collocation of words. The equally important question whether within the verse the collocation of adjective and noun and intervening words is ever abnormal for the sake of preserving verse-unity is not within the limits of this paper. (a) It is not easy to draw a line between purely predicative and amplifying adjectives. The former, as we saw in examples of long adjectives, are often set off in a separate verse; many are participial : is ex se hunc reliquit qui hie nunc habitat filium pariter moratum ut pater avosque huius fuit. (Aul. 21) cur inclementer dicis lepidis litteris lepidia tabellis lepida conscriptis manuf (Ps. 27) vilirus is cum corona, candide vestitus, lautus, exornatusque ambulat. (Cas. 7G7) Somewhat different in effect, but equally separable are these participial adjectives: miles lenoni Ballioni epistulam conscriptam mittit Polymaehaeroplagides, (Ps. 998) Vol. i] Prcscott. — Thought and Verse in Plautus. 219 hominem cum ornamentis omnibus exornatum addueite ad me iam ad trapezitam Aeschinum. (Ps. 756) et tu gnatarn tuam ornatam adduce lepide iu peregrinum modum. (Pers. 157) "Writes and sends," "dress up and bring" may suggest the effect of such separation. Such examples, in which the verbal element is prominent, are hardly within the scope of this paper. 20 I take it that the following group of cases will not be regarded as illustrating real separation ; predicative or amplifying as you please, the suggestion of physical or emotional distress is an afterthought, which separation by the verse-end and intervening words, and position in close connection with caesura or diaeresis accentuate : item parasiti rebus prolatis latent in occulto miseri, victitant suco suo. (Capt. 82) ecastor lege 'dura vivont mulieres multoque iniquiore miserae quam viri. (Merc. 817) itaque nos ventisque rluctibusque iai-tatae exemplis plurumis miserae 27 perpetuam noctem; (Eud. 369) ilia autem virgo atque altera itidem ancillula de navi timidae desuluerunt in scapham. (Eud. 74) ibi me nescio quis arripit timidam atque pavidam, nee vivam nee mortuam. (Cure. 6-48) A similar pathetic effect is evident in mulierculas video sedentis in scapha solas duas. (Eud. 162) 26 Nor present participles as in nam istaec quae tibi renuntiantur, filiuin te velle amantem argento circumducere, (Ps. 430) 27 So, preceding a pronoun, in a lyrical context: sed muliebri animo sum tamen: miserae (quom venit) in mentem mi hi mortis, metu membra occupat. (Eud. 685) Note the alliteration carried through the couplet with pathetic effect. Another example, of miseni following a pronoun: P<>1 me quidem miseram odio enicavit. (As. 920) 220 University of California Publications. [Class Phil. Nor will there be any doubt that these adjectives are inde- pendent : nunc equos iunetos inl es capere me indomitos, ferocis, (Men. 862) Conviva commodus in M. G. 612 does not prevent the same adjective from becoming an amplifying expression with the same noun in eonvivaa volo reperire nobis commodos, qui una sient. (Poen. 615) Here the noun and adjective appear at the extremes of the sen- tence after and before pauses. 28 In the following case the con- text shows that frigidam is predicative; calefieri finds its anti- thesis in adponi frigidam: calefieri iussi reliquias — pernam quidem ins est adponi frigidam postridie. (Pers. 105) "Served up cold" is clearly the idea. 29 Nor may I admit as indubitable cases of real separation such substantival adjectives as virgo and posticum: eius cupio filiam virginem mini desponderi. (Aul. 172) est etiam hie ostium aliud posticum nostrarum harunc aedium: (St. 419) Filiola virgo (Rud. 39) and virginem gnatam suam (Trin. 113) may support the adjectival force of the first adjective, but in any case the separation in our passage defines filia and contrasts the daughter of Euclio with the middle-aged woman of Mega- dorus's previous remarks (162). 30 As for posticum, it is clearly. 28 The adjective moJcstum in the following verses is more closely con- nected with the infinitive: et impudicum et impudentem hominem addecet molestum ultro advenire ad alienam domum, (Bud. 115) Ai.| une will not take luculentum (luculente P) as anything but predica- tive (Ep. 158) after comparing vs. 311 of the same play. - ■• If. nieniini: ut muraena et conger ne calefierent : nam nimio melius oppectuntur frigida. '(Pers. 110) 80 So. luit with dearly expressed contrast in the second verse, the com- pound virgo civis is divided in an paulum hoc esse tilii videtur, virginem vitiare civen:.' cnnservain esse credidi. (Ter. Eun. 857) Vol. ij Prescott. — Thought and Verse in Plant us. 221 a substantive in Most. 931, and so its diminutive in Trin. 194, 1085; in the Stichus, if not an appositive, it defines ostium." 1 The separation of aliud does not here concern us. In connection with substantival adjectives another passage in the Aulularia is to be considered: namque hoc qui dicat : quo illae nubent divites dotatae, si istud ius pauperibus ponitur? (Aul. 489 ) M The contrast between divites and pauperes suggests that the for- mer is substantival ; but it does not at once follow that dotatae is purely adjectival. For vss. 534-5 of the same play show how easily the participial adjective becomes substantival : nam quae indotata est, ea in potestate est viri; dotatae mactant et malo et damno viros. Similarly Ter. Phor. 938, 940. If, however, it is adjectival in our passage, it adds to and explains divites very much as facti osum in venit hoc mihi, Megadore, in mentem ted esse hominem divitem factiosum, me autem esse hominem pauperum pauperrimum. (Aul. 226) In both passages we have the contrast between rich and poor, and in factiosum as in dotatae the happy isolation at the begin- ning of the verses of a more specific attribute of the rich class: in each case the emphasis is accentuated by the sense-pause which follows the separated adjective. From a different point of view hominem divitem \ factiosum should be compared with hominem strenuom \ benevolentem (Ps. 697, above, p. 211). Most of such amplifying ideas are similarly brought into prominence by their position at the beginning of the second verse; often they are followed by a decided sense-pause; some- times this separation brings them into the vicinity of contrasted 81 The verse immediatly following in the Stichus (450a) contains pos- ticam partem, but this verse is not in A, and the division of 450a and 451 in B is suspicious: cf. Leo ad loc. If vs. 450a is genuine, as Lindsay seems to regard it in his Oxford text, a purely adjectival force gains some support. Cf . Pauli Pestus, 220 M = 276 de Ponor. 3; In a similar context Menander (585 K.) has a similar separation: iirsr/c )vi>a'iK' ewiiikfipov imdvfiel Xaf3elv irTiovrnvaav 222 University of California Publications. [Class Phil. iilr;is." ;; All of these features, with attendant alliteration, are illustrated in ego te, Philocrates false, faciam at verus hodie reperiare Tyndarus. (Capt. 609) The separation of an adjective from a vocative is similarly ar- ranged, but here in a succession of epithets (referred to on p. 211), in Quid ais, homo levior quam pluma, pessume et nequissume, flagitium hominis, subdole ac minimi preti? (Men. 487) The surprise of the opprobrious epithet is made more effective by separation and prominent position. The element of surprise, which false and levior, like subvolturium and volaticorum among the long adjectives, illustrate, recurs in another example of the vocative ; the parasite greets his patron as a veritable god on earth: o mi Iuppiter terrestris, te coepnlonus eompellat tuos. (Pers. 99) Without the element of surprise and without so distinct a sense- pause, but, I think, with emphasis paterni is separated in nonne arbitraris eum adulescentem anuli paterni signum novisse. (Trin. 789) So in Poen. 1080 the same adjective stands with emphasis in the same position, though not separated. Contrast is heightened by alliteration 34 in quodque concubinam erilem insimulare ausus es probri pudicam meque summi flagiti, (M. G. 508) and here prominent position is given to the crime rather than the adjective, that the two crimes may occupy the extremes of 33 For contrasted ideas brought into the same verse by the separation of an adjective cf. Caecilius 221 E 3 : egon vitam meam Atticam contendam cum istae rustic-ana (tua), Syra? unless it is an octonarius, as C. F. W. Midler supposes. Bergk's asticam brings out the contrast more plainly: cf. rusticatim . . . vrbanatim in Pom- ponius 7 R 3 (Leo, Analecta Plautina: de figuris sermonis II 32). 34 Cf. probrum, propinqua partitudo (Aul. 75), probrum . . . parfitudo prope . . . palam (Aul. 276). Vol. ij Prescott.— Thought and Verst in Plant us. 223 the verse and the two abused innocents be juxtaposed in pudicam meque. Contrast and comprehensiveness are obtained in this separation of dexteram : age rusum ostende hue manum dexteram. — em. — nunc laevam ostende. — quin equidem ambas profero. (Aul. 649) Somewhat different is the collocation in nixus ]aevo in femine habet laevam manum, dextera digitis rationem eomputat, ferit femur dexterum. (M. G. 203) Here the contrasted parts occupy different verses; dexterum echoes dextera of the preceding verse, 35 and the actor's gestures doubtless contributed to the effect; the alliterative features are plain, whether or not part of the poet's intention in separating the adjective. An adjective expressive of size is naturally liable to separation and prominence; 36 in this example maxumi is practically predic- ative ; number and size are postponed with dramatic effect : postquam in eunas conditust devolant angues iubati deorsum in impluvium duo maxumi: continuo extollunt ambo capita. (Amph. 1107) Essentially attributive, but in effective juxtaposition, the same adjective is postponed with more injury to verse-unity in sumne probus, sum lepidus civis, qui Atticam hodie civitatem maxumam maiorem feci atque auxi civi femina? (Pers. 474) The postponement of the verb makes the thought less complete, but the alliterative juxtaposition 37 of the superlative and com- parative more than compensates for the separation. When the verb comes in the first verse, the adjective escapes into the second verse with less violence to unity, and in this example is brought 35 Cf. usque . . . \ usque . . . | faciebatis . . . | fxigiebatis . . . (As. 210- 213) ; iussin (As. 424-426) ; deam . . . | deum . . . (As. 781-782) ; omnes (Aul. 114-115); itaque (Cist. 513-515); peril (Merc. 124-125); egomet (Merc. 852-854) ; ferreas, ferream, fcrreas (Pers. 571-573) ; p< rqvn (Poen. 418-420), pater . . . | 'pater . . . (Poen. 1260-1261). 80 Cf . Norden, Aeneis Buch VI, 390. 37 Cf. Cas. 1006, Amph. 704, Capt. 1034, M. G. 1218, Rud. 71, St. 739. 224 University of California Publications. [Class Phil. into associations of thought and sound that give the second verse a unity of its own : nulla igitur dicat : equidem dotem ad te adtuli maiorem multo quam tibi erat pecunia. (Aul. 498) So with elaborated emphasis on size : verum nunc si qua mi obtigerit hereditas magna atque luculenta, 38 (True. 344) A necessary specification is added to the noun in ut opinor, quam ex me ut unam faciam litteram lon(gam, me)um laqueo collum quando obstrinxero. (Aul. 77) 30 Alliterative possibilities may have helped attract the adjective into the neighborhood of laqueo; the alliteration in litteram | long am is merely an unavoidable accident. This prominent position, combined with a sense-pause, some- times introduces an elaboration of the idea 40 expressed in the separated adjective ; so in the elaboration of a joke : si hercle illic illas bodie digito tetigerit invitas, ni istunc istis invitassitis (Rud. 810) or with further explanation of the idea as in the examples quoted above (p. 211) in Rud. 1158, and (p. 209) 421, 372. In two examples in which the long adjective inhonestus is set at the beginning 41 of the verse the amplifying idea occupies the entire second verse with predicative effect : nunc hie occepit quaestum hunc fili gratia inhonestum et maxime alienum ingenio suo. (Capt. 98) 38 Note the balance between magna atque luculenta (345) and dulce atque amarum (346). 3J According to the reading of the MSS. Bacch. 279 belongs here: ego lembum conspieor longum strigorem maleficum exornarier. But strigorem is dubious. 10 Cf. Norden, Aeneis Buch VI, 391. 41 The same adjective stands in the same position in Ter. Eun. 357. For the occupation of the entire second verse cf. Trin. 750: sed ut ego nunc adulescenti thensaurum indicem indomito, pleno amoris ac lasciviae? vol. l] Prcscott. — Thought and Verse in Plautus. 225 verum quom multos multa admisse aceeperim inhonesta propter amorem atque aliena a bonis: (M. G. 1287) 42 A few cases remain in which the added ideas, set off at or near the beginning of the second verse, are rather conspicuously linked by alliteration to neighboring words in the same verse; some such cases have been already mentioned, but in the follow- ing the alliteration is even more conspicuous: turn quae hie sunt scriptae litterae, hoc in equo insunt milites armati atque animati probe." (Bacch. 941) quid istic? verba facimus. huic homini opust quadraginta minis celeriter calidis, danistae quas resolvat, et cito. (Ep. 141) quibus hie pretiis porci veneunt sacres sinceri? (Men. 289) Diaeresis or caesura contribute to the emphasis and independent unity of the amplifying ideas; in the second example the entire second verse has a unity of its own, of which the alliteration is a superficial indication. 44 In the following example, referred to among the cases of successive epithets, the alliteration in both verses brings into relief the distinct unity of each, and the sepa- rated adjective, being only the last in an accumulation of epi- thets, escapes into the second verse without violence : iam hercle ego istos fictos compositos crispos eoneinnos tuos unguentatos usque ex cerebro exvellam. (True. 287) In M. (4. 508 we noted a certain artificiality in probri pudicam meque summi flagiti (above, p. 222). The employment of the ends of a verse to set in relief a pair of balanced ideas appears in " erne, mi vir. lanam, unde tibi pallium malacum et calidum conficiatur, tunicaeque hibernae bonae, " (M. G. 687) The adjectives here are less evidently amplifying, though con- ceivably separable; the striking feature is the position of each 42 Omitted in A. 43 Cf. Aecius 308 K 3 : ut nunc, cum animatua iero, satis armatua sum. 44 For alliterative groups including calidus ef. Cas. 255, 309, Ep. i!-">i»; and especially, in connection with our passage: reperi, comminiscere, cedo calidum consilium cito, (M. (!. 226) 226 University of California Publications. [Class Phil. pair of adjectives at the opposite extreme of the verse, the first pair varied by the connecting particle et. The two substantives are divided between the verses; the verb common to both stands before the diaeresis of the second verse; the alliteration is com- paratively unimportant. Cf. Norden, Aeneis Buch VI, 383 on similar phenomena in Vergil. The regularity with which adjectives, following their substan- tives and separated, stand at the beginning of the second verse, is not appreciably disturbed by a few examples of different dis- positions of the separated ideas. So the adjective sacerrumus, which regularly appears at the end of a verse in Plautus (Rud. 158, Most. 983), is effectively placed in a verse which constitutes a unity by itself and with alliteration that hisses out the oppro- brious epithet: 45 praesenti argento homini, si leno est homo, quantum hominum terra sustinet sacerrumo. (Poen. 89) Similarly Plautus sets off the accomplishments of the parasite's sun-clial ; again superlatives, and to be sure in one case metrically convenient (cf. Merc. 206) ; and again in a verse that is an in- dependent unit; both this and the former example are essen- tially predicative: nam(unum) me puero venter eral solarium, multo omnium istorum optimum et verissumum. (Boeotia, 1, 4) The separated adjective stands after a diaeresis, w 7 ith reiteration of the same idea at the end of the same verse and in the next verse, in quia enim filio nos oportet opitulari unico. — at quamquam unicust, nihilo magis ille unicust mihi fllius quam ego illi pater: (('as. 202) (Cf. Capt. 150: tibi ill' unicust, mi (Ham unico magis aniens.) A somewhat similar but less explicable separation occurs in si itast, tesseram conferre si vis hospitalem, eecam attuli. (Poen. 1047) Here the adjective is not demonstrably amplifying (cf. 958, a Cf. Ter. Hec. 85: minime equidem me oblectavi, quae cum milite Corinthum nine .sum profecta inliumanissumo : Vol,. i] Prescott. — Thought and Verse in Plautus. 007 1052, where it precedes the noun), though it may be felt as an afterthought; the association of thought in eccam attuli may have drawn it from its noun ; but the interruption, by the verse- end, of the artificial interlocking of tesseram conferre si vis hos- pitalem—a, thought-unit embraced between. noun and adjective is striking. The examples above (Poen. 615, Pers. 105, p. 220) are similar, but the adjectives in those cases are more clearly amplifying or predicative. We have reviewed the cases in which the separated adjectives follow their substantives : 40 such adjectives have very regularly stood at the beginning of the second verse and usually with a caesura or sense-pause immediately following; with few excep- tions they have been added ideas, the separation of which was accomplished without violence to verse-unity; many of them, indeed, were almost if not quite predicative; most of them gained by separation, through acquiring emphasis, or producing antithesis or sound-effects. There is perhaps only one doubtful case : quin potius per gratiam bouam abeat abs te. (M. G. 1125) It may hardly be said that bonam adds to the thought, for per gratiam is sufficient in itself (M. G. 979, 1200, and St. 71 accord- ing to Leo, Bemerkungen fiber pi. Wortstellungen etc. 118 and Lindsay, Class. Rev. 8 [1891] 159). Bona gratia is, of course, Plautine (Bacch. 1022, Rud. 516). The same idea, expressed in the same play, vs. 979, vin tu illam actutum amovere, a te ut abeat per gratiam .' makes us suspect that in 1125 the poet availed himself of the pleonastic adjective and of separation for the sake of the reitera- 40 Most. 501 should be added: hospes me hie uecavit, isque mi' defodit insepultum clam (ibidem) in hisce aedibus, scelestus, auri causa, nunc tu hinc emigra: BCelestae hae sunt aedes, impiast habitatio. The afterthought scelestus is echoed in scelestae. Insepultum needs no comment: cf. defodit in terrain dimidiatos in Cato's .Speeches, XXXVII 3. 228 University of California Publications. [Class Phil. tion of a- and b-sounds, just as a consideration for a- and t- sounds affected the structure of vs. 979. 47 It is obvious that the cases of separation in which the adjec- tive appears in the first verse, and the substantive in the second, necessarily involve the incompleteness of the first verse. In most of the cases enumerated in the previous paragraphs the adjectives ranged from purely predicative to loosely amplify- ing; the thought was in a measure complete in the first verse, especially if the verb came in that verse; the separation was apparent rather than real. The examples about to be discussed may seem, per se, to impair the validity of Leo's theory; it is important, therefore, to note that they are few in number. Nor h it impossible that in spite of the separation the noun or adjec- tive may be so related to the context as to reinforce to some extent the unity of the verses. It may be well to cpiote at once a striking example of the reali- zation of this possibility. In one passage already quoted we have seen some evidence of a rather studied disposition of adjec- tives and substantives (M. G. 687, above, p. 225). The case before us shows evidence of even more care in the collocation of words : aequo mendicus atque ille opulentissimus censetur censu ad Acheruntem mortuos. (Trin. -493) It is perhaps annoying to enumerate the features of this couplet, which are sufficiently plain to any sympathetic reader or hearer. In the first place, the thought is incomplete until the caesura of the second verse is reached. Yet the separation of aequo from censu is attended by an effective juxtaposition of ideas in the first verse, which gives to that verse a partial unity. 48 The sep- 47 Appuhn, 1. c. 67-68, distinguishes sharply between dissyllabic and trisyllabic adjectives, and maintains that the former may not be separated. There does not seem to me to be any evidence to warrant such a dis- tinction, and it lacks inherent probability. His contention that bonam is unemphatic and absorbed in the first foot, may ease the separation, but does not explain it. "Cf. <'ist. 532: postremo quando aequa lege pauperi cum divite non licet, Vol. l] Prcscott. — Thought and Verse in Plautus. 229 aration of censu results in a figura etymologica and consequent unity of sound- and sense-effect. And mortuos at the end car- ries us back to the nouns of the first verse in such a way as to establish the unity of the couplet by the close interlocking of ideas. 49 A phase of a-n-6 kolvov is illustrated in the following case : decet innocentem qui sit atque innoxium servom superbum esse, apud erum potissumum. (Ps. 460) The thought is again incomplete until we reach the caesura of the second verse; yet there is a fitness in the transference of servom to the side of superbum, with which it belongs as much as with the adjectives of the preceding verse, and to which allit- erative opportunities (cf. As. 470) attract it. The significance of this example is clearer on comparing it with the recurrence of the same thought without separation of the adjective in decet innocentem servom atque innoxium confidentem esse, suom apud erum potissumum. (Capt. 665) In both passages the verse preceding the couplet contains the adverb confidenter, and this adverb prompts the commonplace in each case : in the Capt. the poet repeats the idea of the adverb iii the corresponding adjective ; in the Ps. he chooses a synonym. It is not, of course, possible to discover whether in the latter case his choice was determined by a desire to avoid the recur- rence of the same stem or whether the alliterative unit servom superbum came to his mind independently of any consciousness of monotony in the repetition confidenter — confidt it I cm. But in any case the comparative artificiality of the couplet from the Ps. is evident: the development in freedom of technique is clear."" Without discounting the value of other factors may we 49 Nor is the emphasis on aequo to be overlooked; cf. the Greek equiva- lent in Menander 538 K: koivov rbv ' Auh/r iaxov ol iravreg vxiroi. The tragic seriousness of the speaker in the Trinummus perhaps explains the artificial style, which adds dignity to the expression (Leo, Plaut. Forsch. 122 and note 5). 50 The hiatus in ('apt. 665 is perhaps a part of tlie crudity of composi- tion. 230 University of California Publications. [ r 'LAss Phil. not say that when once the alliterative unit occurred to him the unity of sound proved superior to the affinity of the attributive for its noun, and that this conservation of unity of sound was made easier or perhaps suggested by the fact that there was a strong unity of thought as well which linked servom to si iuju,i, \ need nut detain us: cf. Men. 1089 quoted above. 232 University of California Publications. [Class Phil. from the different sound-effects of each verse — a-sounds predom- inating in the first verse, 1-, m-, u-, and n-sounds in the second ; certainly there does seem to be something conscious in the change from attrcam of the first verse to aureolum of the second; we may properly maintain that the unity of form and of sound- effects in the second verse could have arisen only from a con- sciousness that the second verse was a distinct entity. At the same time the fact that the consciousness expresses itself only in a superficial or external preservation of verse-unity, and that unity of thought is interrupted, suggests that "art-poetry" in Plautus's hands was on occasion further advanced than the chronological proximity of Saturnian verse would lead us to suspect. In contrast with merely superficial observance of unity stand a few cases of separation in which the thought serves to reassert the unity of the verse: hosticum hoc mihi domicilium est, Athenis domus est Atticis; ego istam domum neque moror neqiie vos qui homines sitis novi neque scio. (M. G. 450) Alliteration, to be sure, may have attracted hosticum to hoc, but the dominant factors are emphasis and contrast. Hosticum is first in the sentence because emphasis brings it to that position. Domicilium is first in the verse 54 because emphasis again de- mands for it a prominent position : it must stand in the same verse with domus to bring out the contrast between "house" and "home." The effect may be suggested in English by "Stop! a stranger's | house you point me to; my home's in Athens; for your home | I care not, nor know I who you gentle- men may be." Another passage in which at first sight unity seems to be dis- regarded, when studied in the light of the context, shows con- siderable consciousness of the intimate association of verse-unit and thought-unit : habui uumerum sedulo: hoc est sextum a porta proxunmm angiportum, in id angiportum me devorti iusserat; quotumas aedis dixerit, id ego admodum incerto scio. (Ps. 960) 14 But est domicilium in CD. Xote also host in m {ost — ) BCD. Vol. l] Prescott. — Thought and Verse in Plautus. 233 Here, again, it may be said that porta has attracted the allitera- tive proxumum, but the verse-division represents a correspond- ing division of thought. The beginning of the first verse leads up to the number and precise location ; angiportum, separated from its two adjectives, stands out at the beginning of the sec- ond verse, again with emphasis, and is repeated 55 with the re- sumptive pronoun — all of which heightens the contrast with aedis of the third verse. The effect is: "I've got the number right: the sixth, (in going from the gate), | alley-way, that's the alley-way I was told to take ; | but the number of the house, that I've clean forgotten." Perhaps the existence of any unity in the following example will be less readily granted : coepi observare eequi maiorem filius mihi honorem haberet quam eius habuisset pater. (Aul. 16) There seem to be two prominent factors in the separation : the comparative degree is attracted to the ablative of degree of dif- ference; 56 alliteration brings together honorem and haberet." Yet is it too fanciful to say that in spite of the separation the position of filius and pater at the ends of their verses 58 suggests a unity of thought quite apart from and above the syntactical and alliterative unity of each verse? The two verses are com- parable to the two pans of the scale, the son balancing the father, and maiorem alongside of filius marking the turn of the balance which the expectant Lar anticipates. 59 55 Examples . of such repetition may be found in Bach, de usu pron. demonstrat. = Studemund-Stud. II 353-354. 66 See the examples in Fraesdorff , de comparativi gradus usu Plautino 31 ff. Other factors, external or internal, may have precedence over the natural juxtaposition of the ablative of degree and the comparative, but the generalization above is not thereby endangered. 67 Cf. honos homini Trin. 697, meque honorem Mi habere True. 591, mihi honor es suae domi habuit maxumos Pers. 512, habuit, me habere honorem As. 81. 58 To be sure, they owe their position in some measure to metrical con- venience: cf. vss. 12, 21, 30 of the same prologue. 59 It is not likely that the following example involves separation (but note vinum Chium in Cure. 78) : ubi tu Leucadio, Lesbio, Thasio, Chio, vetustate vino edentulo aetatem inriges. (Poen. 699) 234 University of California Publications. [Class Phil. Nor can I be sure that my understanding of the next ease will prove convincing. The adjective mutuos is occasionally sepa- rated in expressions of the ideas of borrowing and lending; in two of the cases the adjective follows the noun, in one the adjec- tive precedes. For purposes of comparison I include them all here, although the former belong in the previous section : tecumque oravi ut nummos sescentos mihi dares utendos mutuos. (Pers. 117) sed quinque inventis opus est argenti minis nmtuis, quas hodie reddam: (Ps. 732) sed potes nunc mutuam drachumam dare unam mihi, quam eras reddam tibi? (Ps. 85) The frecpient collocation of this adjective with dare and rogare in commercial phrases may have given it a substantival force corresponding to the English "loan": so, for example, exorare mutuom in Pers. 43 (with argentum far distant in 39) suggests that the adjectival force is approximately substantival, 60 and eventually this substantival usage becomes established; even in Plautus we have tide si pudoris egeas, sumas mutuom (Amph. 819). If this is granted, the separation becomes innocuous, even if the adjective precedes; the alliteration in the last example perhaps adds to the unity of the verse, but no such additional feature is necessary if mutuam is in effect appositional. The cases hitherto discussed have shown, in varying degrees, consciousness of verse-unity and conservation of it to some ex- tent in spite of the separation of the attributives. The exam- ples we have now to consider do not so plainly point to a sensi- tiveness to the identity of verse- and sense-unit. There are often extenuating circumstances, but in most cases we must admit that the separation involves a distinct interruption of a thought-unit with less effectual employment of the features that in other examples reinforced the unity of the verse. Prominent 60 Cf. Ps. 294: nullus est tibi quem roges mutuom argentum? — quin nomen quoque iam interiit " mutuom." As. 248 and Trin. 1051 also show mutuos in a sense approximately sub- stantival. The various forms of facere mutuom are hardly parallel. Vol. l] Prescott— Thought and Versi in Plautus. 235 among these is a group of superlatives of cretic measurement which may owe their separation in part to metrical convenience ; occasionally there result sound-effects that may have conduced to separation, but in general the violation of unity is unmistak- able, and the palliating or counteracting features are superficial. It is, however, always to be remembered that the cases of sepa- ration are extremely few in proportion to the number of occur- rences of a given adjective at the end of a verse. The most important member of this group is maxumus, which we have already found separated, but following its noun and standing at the beginning of the second verse with emphasis. This adjec- tive appears 86 times in Plautus: 39 times at the end of the verse, 38 times in the interior, nine times at the beginning. It is not likely that, under normal conditions, the position at the verse-end is prompted by a desire to emphasize; 01 generally un- emphatic words occupy this position. A collection of all the examples of the phrase opere maxumo, with and without sepa- ration, will illustrate the feature of metrical convenience: 02 nam rex Seleucus me opere oravit maxumo (M. G. 75) nunc te hoc orare iussit opere maxumo (Most. 752) pater Calidori opere eclixit maxumo (Ps. 897) rogare iussit ted ut opere maxumo (St. 248) iussit maxumo opere orare, ut patrem aliquo absterreres modo, (Most. 420) non hercle vero taceo. nam tu maxumo me opsecravisti opere, Casinam ut poscerem uxorem mihi (Cas. 992) Cf. Terence, Thais maxumo te orabat opere, ut eras redires. (Eun. 532) 61 Such a position for emphasis is occupied at least once by the very words with which we are now concerned: ego miserrumis periclis sum per maria maxuma vectus, capitali periclo per praedones plurumos me servavi, (Trin. 1087) 62 The significance of the cases of separation is somewhat more appar- ent when we note that magno opere, maiorc opere, nimio opere, tanto "/<< re are never separated in Plautus by the verse-end. 236 University of California Publications. [Class Phil. It is evident that opere is attracted to orare and opsecrare, but so far as the thought is concerned, there is nothing to diminish the violence in the division of maxumo opere in Most. 420, or the division of the larger word-groups in Cas. 992 and Eun. 532. And in the first of the two following cases of maxumus there are no sound-effects to relieve the separation ; in the sec- ond, separation brings together m- and a-sounds ; these are, how- ever, from lyrical passages: ubi quisque institerat, concidit crepitu. ibi nescio quia maxuma voce exclamat: (Amph. 1063) quam malum? quid machiner? quid commiuiscar? maxumas nugas iiieptus incipisso. 63 haereo. (Capt. 531) Cf. Terence, Geta, hominem maxumi preti 84 te esse hodie iudicavi animo meo ; (Ad. 891) Consideration for sound and the artificial arrangement of words may have played some part in the structure of these verses: Alexandrum magnum atque Agathoclem aiunt maxumas duo res gessisse: quid mihi fiet tertio, qui solus facio facinora inmortalia ? (Most. 775) The a-sounds are prominent in the first verse; magnum and maxumas are perhaps not unintentionally put in the same verse; duo, interlocked between maxumas and res, is in contrast with tertio at the other extreme of the same verse. 65 Another superlative optumus occurs at the end of the verse in one third of the total number of its occurrences ; in only one case does its position result in separation: Ineptias incipissc is the reading of the MSS. 01 Contrast with this the stereotyped position at the end of the verse, without separation, of minimi preti, parvi preti, magni preti, quantiiis preti in Plautus (cf. Kassow, de Plauti substantivis s. v. pretium GS.= JHB. Supplbd. 12 (1881) 710). 85 Cf. altera . . . altera, Aul. 195; superi . . . inferi, Aul. 368; miserius . . . dignius, Bacch. 41; malefactor em . . . beneficum, Bacch. 395; meam . . . tuam, Capt. 632. It is interesting to note in this connection a couplet in baccbiac verse: sed vero duae, sat scio, maxumo uni populo cuilubet plus satis dare potis sunt, (Poen. 226). Vol. l] Prescott — Thought and Versi in Plautus. 237 sed, ere, optuma vos video bpportunitate ambo advenire. (Ep. 202) With this should be compared optuma opportunitate ambo advenistia. (Merc. 964) Next in significance to the rarity of the separation is the fact, attested by the verse from the Merc, that the initial sounds op — v — v — op are the external manifestation of unity which is cer- tainly interrupted by the end of the verse. Such a case is far from disturbing Leo's theory. Such interlocked complexes of thought and sound, which are characteristic of the language, must burst the bonds that confine units of thought within the verse: that they do it so rarely is significant. A third superlative that, like optumus, stands at the end of the verse in one third of the total number of its occurrences is plurumus. The singular and the plural of this word are perhaps on a different footing : the plural is conceivably analogous to the separation of omnes; 66 so, for example, in this case of plurumi in the interior of a verse, the separation seems less violent than in cases of the singular : GT plurumi ad ilium modum periere pueri liberi Carthagine. (Poen. 988) Whether this is true in the case of the following feminine plural is not at once patent to an English auditor: O Gripe, Gripe, in aetate hominum plurumae mint transennae, ubi decipiuntur dolis. (End. 1235) In any case, the singular seems at first to be rather rudely sepa- rated in miles Lyconi in Epidauro hospiti suo Tlierapontigonus Platagidorus plurumam salutem dicit. (Cure. 429) Here the conventional phrases of epistolary address run along naturally and result in two separations, with the first of which 86 Cf. below, j,.. 258. 07 Jn Eph, 393 plummet (plurwmum MSS.) is predicative. 238 University of California Publications. [Class Phil. we are not now concerned, but' verse-unity is suggested in the alliterative colligation of Platagidorus plurumam; the effect is as if plurumam were an adverb and salutem (licit no more than sah't ii iubet, as the following example suggests: eruni atque servom plurumum Philto iubet salvere, Lesbonicum et Stasimum. (Trin. 435) in which, again, we have similar alliteration — plurumum Philto. pronounced J'ilfo. So, too. our explanation is confirmed by multam me tilii salutem iussit Therapontigonus dicere (Cure. 420) in which, as in the other cases, mid tarn me are attracted to each other, while salutem iussit like salutem elicit and solvere stands at the beginning of the second verse. 68 The adjective parvolus occurs thirteen times: nine times at the end of a verse, three times with separation. Of these three, one belongs in our examples of adjectives following their nouns, and is a mere afterthought : nam mihi item gnatae duae cum nutrice una sunt surruptae parvolae. (Poen. 1104) The other two, both from the same play and of the same situa- tion, are cases of violent and absolute separation : 6a nam ego illanc olim quae hinc flens abiit parvolam puellam proiectam ex angiportu sustuli. (Cist. 12o) nam mihi ab hippodromo memini adferri parvolam puellam eamque me mihi supponere. (Cist. 552) A comparison with three cases in Terence justifies us in attrib- uting the separation in large measure to metrical convenience: ibi turn matri parvolam puellam dono quidam mercator dedit (Eun. 108) 68 On the other hand, without separation, but again in alliterative col- ligation in Veneri dicito multam meis verbis salutem. (Poen. 406) 60 The alliteration, interrupted by the verse-end, in parvolam | puellam has no significance, for it is accidental: the range of expressions tor the idea is too limited to admit our regarding it as genuine alliteration. Vol. lj Prescott— Thought and Verst in Plautus. 239 nisi si ilia forte quae olim periit parvola soror, hanc se intendit esse, ut est audacia. (Eun. 524) ah, stultitiast istaec, non pudor. tarn ol> parvolam rem paene e patria! (Ad. 274) In the second example sense as well as sound may connect p< Hit parvola — "died in infancy," and in the last there are sound- effects that reassert the unity of the verses. 70 So much for this group of cretic adjectives ; 71 the following participial adjectives may be more easily separable because of 70 Something might be said for a substantival force in parvola, though it could hardly apply to the last example from Terence: such a force is possible in Ter. Eun. 155: parvola hinc est abrepta ; the substantival force is evident in Terence's a parvolo (And. 35, Ad. 48)= a puero. The nearest approach to it in Plautus is in Poen. 896, 1346, but it is not certain in either place; nor is Ps. 783 a clear case. Cf. Lorenz, Pseudolus, Einleiturig p. 59. 71 Before leaving these examples in which metrical convenience seems to be a factor in the separation, I may call attention to a closely re- lated phenomenon which, it seems to me, is not always recognized. Is not the stereotyped position of certain words in the verse often nothing more than the working of the poet's mind along the path of well-worn " grooves," as a psychologist might express it"? For example, in the cases above in which salutem iussit or dicii, or solvere hibet, appear, the position of salutem and salver e (rather regularly at the beginning of the verse, though not uniformly) can hardly be attributed to metrical conve- nience alone : it is to some extent a matter of habit. A better example is furnished by these examples from Euripides 's Iphigeneia in Tauris: TuTifUj-EOV 701 !-EGTOV £K VCIOV ?M i 3eIV ayaKfia naaaq TrpoGtykpovTE fiifxavaq, (Ill) <\'/v mdapfiara. (1315) ro r' o'vpavov nicy/ia, rf/r Siur xuprjq aya?/ia. vaug d'kic /tfntjc i^Bky^aro po/fTn;- (1384) Those of us who are reluctant to admit metrical convenience as a factor may find some comfort in emphasizing the part that mental habit plays in the regular appearance of certain words in the same part of the yerse. 'Aya'Afia in the verses above seems to me to owe its position to this rather than anything else. 240 University of California Publications. [Class Phil. the peculiar nature of the adjective, and the balanced isolation of pah /•; salve, insperate nobis pater, to complecti nos .sine. — cupite atque exspectate loiter, salve. (Poen. 12.")!)) The greetings are from two sisters with artificial variation of the conventional terms: the imperatives and vocatives are ar- ranged in chiastic order; pater stands at the beginning of each vnsc, 7 - leaving the adjectives at the end in each ease. The collocation is the same as in o salve, insperate multis amiis post quern conspicor f rater, (Men. 1132) according to MS. B, but the other members of the Palatine fam- ily (and A apparently agrees) introduce a change of speakers before frater. Even if we agree with the editors in following A and the majority of the Palatine family, the isolation of the participial vocative, and the relative clause that modifies it, may point to a certain degree of separability in the participial adjec- tives insperate, cupite, and exspectate in our passage. 73 There remains a small group of cases in which verse-unity seems to be lost sight of, and which are alike in that the adjec- tives are of four syllables metrically convenient at the end of the verse : pol istic me baud eentesumam partem laudat quam ipse meritust ut laudetur laudibus. (Capt. 421) baud eentesumam partem 'lixi atque, otium rei si sit, possum expromere. (M. G. 763) si quisquam banc libera li causa maim assereret, (Cure. 490) ne epistula quidem ulla sit in aedibus nee cerata adeo tabula; et si qua inutilis pietura sit, earn vendat: (As. 763) Centesumus occurs only in these two places in Plautus; libcrali causa occurs in the interior of the verse in Poen. 906, 964, 1102, '■- < !f. above, p. 223, n. 35. "Ferger, de voeativi usu Plautino Terentianoque 32, defends the read- ing of B in Men. 1132 on the ground that insjii rate is not found in Plau- tus without an accompanying noun. Vol. l] Prescotl. — Thought and Versi in Plautus. 241 and so liberali manu in Cure. 668, 700: inutilis occurs again in Ps. 794 and at the end of the verse. But the separation in these cases is not entirely a matter of length and metrical conveni- ence : the collocation of the other words in the sentence is so fixed by almost inviolable laws that it is not surprising that the adjective should escape into the second verse. For to any- body familiar with Plautus and with Wackernagel's study of the position in the sentence of enclitic words it will be clear that the collocations pol istic me, si quisquam hanc, and < i si qua are to a considerable extent fixed in the usage of the language ; the increased difficulty of conserving verse-unity is obvious. 74 The very fact that in some 15,000 verses so few cases of sepa- ration occur — and this in spite of the fondness of the Koman for interlocked complexes which would seem to make the preser- vation of verse-unity difficult — clearly attests the sanity of Leo's contention. The further fact that in so many of the few cases of separation the unity of the verse reasserts itself through asso- ciation of thought or sound confirms in large measure his re- cpiirement of special justification when separation does occur. The existence of a few cases in which unity is not apparent need not affect the validity of the principle ; the essential unity of the verse so far as attributive adjectives are concerned is clear at once from comparison with a tragedy of Euripides or of Seneca — clearer than any statistics could make it. IV. The large proportion of possessive adjectives among the cases of separation deserves an explanation. They represent one- fourth of the total; indeed if we eliminate cases of merely ap- parent separation the proportion would be even larger. No small part of the explanation is found, of course, in the relative frequency of the possessive adjectives in the conversa- 74 In As. 763 ff. there is perhaps some effect in the position of the nouns tpisl ilia, ccrala tahitln, pictura at or near the beginning of suc- cessive verses. The resumptive earn may also reinforce the unity of the last YCl'St'. 242 University of California Publications. [Class Phil. tional Latin of the plays. That among 3000 75 eases of posses- sive adjectives only about 60 should be separated from their substantives by the verse-end may seem in itself some slight tribute to wrsc-unity rather than a contravention of it. Yet the obvious violence to the unity of thought, at least from an English standpoint, in dividing "thy son"' between two verses makes even a small percentage seem inexplicably large. We must not. however, allow our English standpoint to influence us. The separation of "thy son" by the verse-end in English i- not altogether analogous to the separation of filius from tuos. For in the Latin sentence the phrase corresponding to "thy son" is much less of an independent unit of thought than in the English sentence: in the Latin sentence, largely because the possessives mens, tuos, suos are generally unemphatic and often without accent in the phrase- or sentence-unit, the division by the verse-end does not separate "thy" from "son," but rather divides a larger unit of thought. It is clear, for example, that tims emit aedis filius (Most. 670) constitutes a unit of thought: and so, too, does aedis filius tuos emit (Most. 637, cf. 997). The separation in this latter case, if any is felt, is rather that of aedis filius from tuos emit than merely of filius from tuos. Furthermore, since the possessive adjectives meus, tuos, suos are generally unemphatic in our examples, it is pos- sible and likely that in this example tuos was absorbed in the rhythmical unit tuos emit without much consciousness of any violence in separating tuos from filius by the verse-end; the fre- quency and ease with which words intervene between these pos- sessives and their substantives (quite apart from separation by the verse-end) may support this contention. Even if the pos- sessive had some slight stress upon it, as in the beginning of trochaic verses and rarely in an iambic verse (filiam \ sudm despondit, Cist. 600), certainly such stress was subordinate: suam, despite some quantitative prominence, must have been merged in the surrounding words. 76 Of course it may he ob- "•'- Xilsson, 1. c. 1-. "Some such idea is expressed by Appuhn. 1. c. 63, but in a way that fails to account for trochaic verses and Cist. 600. I hope it is clear that Vol. l] Prescott. — Thought and Verse in Plautus. 243 jected that the thought would lead us to merge it in the preced- ing, rather than in the following word, in the example quoted, and that the possessive is enclitic," not proclitic. For our pres- ent purpose it is enough that the possessive is absorbed in a larger unit, and that the separation by the verse-end is by no means the same as that involved in the division between verses of the English possessive and its substantive. 78 In the second place it is to be noted that the possessives are subject to at least one influence from which ordinary attribu- tives are free: Kampf, 79 and others before him, observed the attraction of pronominal words to one another. Such attrac- tion appears in a relatively small number of our examples: 80 earn meae | uxori (Men. 480), illam quae meam | gnatam (Cist. 547), tu mini tua | oratione (As. 112), ad illam quae tuom | . . . filium (Baceh. 406), fores conservas | meas a te (As. 386), the paragraph above is not intended to offer any complete explanation of the separation, but only to suggest that the separation, such as it is, is probably by no means so harsh as it appears to us. The point that I wish to make' is that the unemphatie possessive has very little independent force and is not merely " swallowed up " (Appuhn) metrically, but ab- sorbed in larger thought-units even of ordinary speech. "Lindsay, Latin Language 167; but cf. E. Wallstedt, Fran Filologiska Foreningen: Sprakliga Uppsatser III (Lund 1906) 189 ff; also Kadford, Trans. Amer. Phil. Assoc. 36 (1905) 190 ff. Neither of these last two articles was accessible to me in time to use them for the discussion above. "The fact that the genitive case is used in appositional relation to the possessives (e.g. mea unius opera) might lead to the suggestion that the separation is not more serious than that of a possessive genitive. This would be a helpful suggestion if the possessive genitive in Plautus were regularly or even frequently separated from its noun by the verse; cases do occur (e. g. Bacch. 901, Bud. 1079, Cist. 544), but rarely; and the possessive genitive with pater, uxor, filius, mater, which are the nouns most frequently appearing in our cases of the separated possessive adjective, is in Plautus almost inseparable from its noun even by intervening words. "Kampf, 1. c. 16 ff. 80 A few cases, though too few to he significant, of a verse-end intrr- v.Miing between pronominal words thus combined are worth noting: tua \ me ('as. 279-280, meam | me Cist. 98-99, me \ meam Ep. 480-481, mea meae M. G. 738-739, sc \ suamque Trin. 109-110, tibi \ tua Ps. 112 L13. L'44 University of California Publications. [Class Phil. filio I moo te esse amicum et ilium intellexi tilii (('apt. 140), sine dispendio | tun tu.-uu libertam (Poen. 163), servos... | suos mihi (Most. 1087). If alliteration appears in such eases, it is, of course, incidental and results from the attraction: it is not a primary factor. Wackernagel I Endog. Forsch. I. 406 ff.) does not include mens, tuos, suos among his examples of enclitic words that drift to the beginning of the sentence. There are cases of separation that mighl have been affected by his law, but they are too few to suggest the direct influence of his law; these few show the enclitic possessive* immediately following the introductory word; they seem more significant when other words intervene between the possessive and the noun: e. g. True. 855, Aul. 733, St. 416. Since Wackernagel 's law affects particularly certain monosyllabic and dissyllabic pronouns, it follows that in com- bination with tlie law of pronominal attraction there results in many cases the necessity of placing the possessive in the third or fourth place: take, for example, these two cases, one of sepa- ration, one without separation i eonteris tu tua me oratione, nmlier, quisquia es. (Cist. (309) protect < i nemo est quern iam dehinc metuam mihi ne quid nocere possit, cum tu mihi tua oratione oiiiiiem animum ostemlisti tuom. (As. Ill) To say nothing of other features, the rule of collocation that makes I u second in the sentence, in combination with the attrac- tion that joins tu I iia me and tu mihi tua. undoubtedly regulates to a considerable degree the disposition of the words: and it is clear that the existence of such laws of collocation must appear seriously to interfere with the poet's consideration of verse- unity, at least in many cases. Such laws affect the spoken language; if Plautus is more ob- servant of them than of verse-unity, it is no more than we should expect of a dramatic poet who is reproducing the conversational Latin of his daw The same general truth applies to ordinary attributives, but they are not as a class subject to these particu- vol. ij Prescott. — Thought and Verst in Plauhis. 245 lar regulations. In addition to the observance of laws control- ling the arrangement of words in speech the poet is governed by the conditions of his verse. It is easy to overestimate the force of metrical convenience. It is seldom more than one of many factors. But it may hardly be denied that the iambic or pyr- rhic possessives found a comfortable habitat at the end 81 and at the beginning of certain iambic and trochaic verses. Indeed, quite apart from the metrical convenience of the possessives that do not involve separation, the cases of separated possessives of iambic or pyrrhic measurement lead to two conclusions : 1) in all cases of separation in which mens, tuos, or suos fol- lows a substantive, whether with or without intervening words, the possessive stands at the beginning of the second verse; 82 2) in all cases of separation in which mens, tuos, or suos pre- cedes a substantive, whether with or without intervening words, the possessive stands at the end of the first verse. 83 The exceptions to these principles 84 only test their validity. It is of course evident that in the cases covered by the first rule there is no reason why the possessive should not stand at the end of the second verse ; such a position is unusual, probably because the separation by intervening words is thereby abnormally great; an example from Terence is qui turn illam amabant, forte ita ut fit, filium perduxere illuc, secum ut una esset, meum. (And. 80) Similarly under the second rule there is no reason why the pos- sessive should not stand at the beginning of the first verse; but here, again, such position is unusual probably because of the extent of the intervening words; an isolated example is 81 For statistics cf. Nilsson, 1. c. 37. 82 Amph. 134, 135, As. 387, 434, Aul. 289, Bacch. 880, Capt. 141, 873, Cist. 586, 601, Cure. 347, 430, Ep. 391, 401, 482, 583, M. G. 543, Mfost. 638, 998, 1088, Poen. 164, 192, 1375, Ps. 483, 650, 850, Bud. 743, Trin. 1101, 1144, True. 293. "As. 16, 112, 785, Aul. 733, Bacch. 406, 777, Cist. 184, 547, 772, Ep. 279, Men. 420, 480, 518, 740, M. (i. 563, . 791>, Rud. 1392, St. 416. Trin. 1147, True. 355. 84 The hiatus, therefore, after the first word of Ps. 650 is not to be cured by changing suam hue to hue suam (Bothe), and Trin. 141 becomes suspicious. 246 University of California Publications. [Class Phil. meamne hie Mnesilochus, Nicobuli filius, per vim ut retineat mulieremf (Bacch. 842) Tn both cases the rare position is attended by other features: in the first,, the postponement of meum perhaps suggests the pathos of the situation: in the second, emphasis, alliteration, and collo- cation with hie are contributory factors. Finally, such an ex- ception to these rules as appears in the following example is due to the peculiar nature of the formula and the greater conveni- ence of obs( cro at the end : adsum, Callicles: per tua ohsecro genua, ut tu istuc insipienter factum sapienter feras (True. 826) Cf. Cure. 630, where per tua genua te <>hs< <■>■<> concludes the verse, and Poen. [1387], where, again at the end of the verse, we find per ego tua te genua obsecro. s:> We have thus noted several features that make the compara- tively large number of separated possessives more easily under- stood. As in the case of ordinary attributives, there are occa- sionally special conditions which emphasize the unity of the verse in spite of the separation. The accidental alliteration arising from pronominal attraction w r e have already noticed , there are a few cases of genuine alliteration : nl>i erit empta, ut aliquo ex urbe amoveas; nisi quid est tua secus sententia. (Ep. 279) nam hominem servom suos domitos habere oportet oculos et manus (M. G. 563) oeulos volo meoa delectare munditiis meretriciis. (Poen. 191) There are a few cases, allied to those of pronominal attraction, in which pronominal words are not immediately juxtaposed but are grouped together in the same verse: all, salus mea, servavisti me. (Bacch. 879) vel cyn. ijui duduni fili causa coeperam cyu nii'il I'xcruciare aninii, quasi quid films 98 Cf. Langen, Beitrage zur Kiitik u. Erklarung d. PL 335; Kampf, I. c. 21. Vol. l] Prescott — Thought and Verse in Plautus. '-'47 meus deliquisset me erga (Ep. 389)" O fili:i mea, quom banc video, rnearum me absens miseriarum commones; (Rud. 742) Tn the following example )iuar belongs to both nouns: inscitiae meae et stultitiae ignoscas. (M. G. 542) The possessive adjectives of the plural pronouns of the first and second persons occur naturally with much less frequency than mens, tuos, suos, and cases of separation are proportion- ately fewer. They are subject to fewer special regulations and conditions : they are not enclitics ; metrical convenience does not affect their position so significantly ; they are to be sure subject to the principle of pronominal attraction : 87 saluto te, vieine Apollo, qui aedibus propinquos nostris aceolis, venerorque te, (Baccb. 172) tonstricem Suram novisti nostram? (True. 405) qua re filiam credidisti nostram? (Ep. 597) meritissumo eius quae volet faciemus, qui bosce amores nostros dispulsos compulit. (As. 737) nam meus formidat animus, nostrum tarn din ibi desidere neque redire filium. (Baccb. 237) In these cases there is little to suggest the entity of individual verses. The possessive and its noun in every example but one bracket other words, and the word-group thus formed shows no respect for verse-unity. Such word-groups appear in very sim- ple form in Altenberg's examples from early prose: in Plautus's verse — we may not here enquire into the causes — they are often 86 Note also ego, ego med, metis at or near the beginning of successive verses. "This does not happen to appear in our examples, but note Terence Haut. 711: ut quom narrel senex voster nostro esse istam amicam gnati, non credat tameu. 248 University of California Publications. [Class Phil. elaborate, as the Las1 example above illustrates/ 8 The signifi- e.int fad is that in spite of the employment of such interlocked phrases the poet so seldom allows them to escape into the second verse. It is true that when the ordinary attributive escapes, verse-unity seems more often to reassert itself than when a pos- sessive is separated, but such difference as there is, is accounted for by the relative frequency of the possessives, the unemphatic nature of most of them, and their metrical character, which draws some of them to the extremities of the verse. Inasmuch as nostcr, voster are subject only to the second of these influ- ences, lack of emphasis may properly be regarded as the most important factor in the separation. sr ' ss In the cases of mens, tuos, suos, visually the possessive is separated from its noun only by a verb (Aul. 733-734, Ps. 849-850). There are a few cases of more elaborate interlocking: ad ilium quae tuom perdidit, pessum dedit tibi filium unice unicum. (Bacch. 406) Special effects are usually produced by Buch arrangements; an interesting case is sicut tuom vis unicum gnatum tuae superesse vitae sospitem et superstitem, (As. 16) Here the couplet is securely linked together by the connection between the noun of the first verse and the adjectives of the second; but as the connection is predicative, the unity of the second verse, reinforced by the sound- and sense-effect, is paramount; tuae is separated from vitae, and the separation also divides the group tuae superesse vitae, but if our con- clusions above are correct, the weak force of tuae made the separation inoffensive to the Roman. Another interesting case is quid ais? ecquam scis filium tibicinam meura amare .' (Ps. 482) The eriss cross ecquam . . . filium tibicinam | meum brings together the con- trasted objects and suggests the father's indignation, while wrum is too weak to interrupt seriously the unity id' the verses except so far as it is already interrupted. 9 The evidence does not suffice to include Greek influence as an addi- tional factor. The ways of expressing the possessive idea in Greek are more varied, and the conditions inherent in the words are different from those of their Latin equivalents. The fragments of the New Comedy offer almost no parallels to the separation in Plautus. In Menander's (307 K.) j u aavrbv inm\ hv to Kp&yfiara j eldyg ra aav~ov ) the article with the |m>s Vol. i] Prescott. — Thought and Verse in Plautus. 249 V. These special conditions also affect many other pronominal adjectives, so that it is not surprising that, for example, the demonstrative pronouns in their adjectival usage are second, in frequency of separation, to the possessive adjectives. Again, however, the cases of separation, viewed with reference to the total number of occurrences of such adjectives, are extremely few. The fact that these words are pronominal as well as ad- jectival may in many cases have mitigated the separation; and the effect of Wackernagel's law and of the law of pronominal attraction, working either separately or in common, is very pro- nounced in many of our examples. The studies of Langen, Bach, Kampf, Kellerhof, taken in connection with Wacker- nagel's different and broader point of view, explain the position not only of the demonstratives, but of the determinative, and of the indefinite quis and its derivatives. If these words find their natural habitat immediately after the introductory word of the sentence, and if the closeness of the adjectival relation is something much less binding than the operation of Wacker- nagel's law — as is quite evident — it is remarkable that cases of separation are so infrequent. The examples that follow will show the pronominal word in close connection with the introductory word of the sentence ; so nunc is immediately followed by hoc: nunc hoe deferam argentum ad hanc, quam mage amo quam matrem meam. (True. 661) sessive genitive may suggest an amplifying idea. I have not found any cases of efioc, aoq thus separated. In Euripides, however, parallels occur, but they are less frequent than in Plautus; e.g. yr//ia^ rvpawov nai Kaaiyv^rovg renvois | f/in'ir Qvthvuv; (Med. 877, possibly with emphasis on kfiolq), ripf? Ffif/v KOfiii^ofiai | Zaf-iuv a6r.fyr/v (Iph. T. 1362), ou^t ryv ifii)v \ (povia vouifav x el P a (Iph. T. o85), i-uaav ftc <"/'' ffwi fki>ir ayn'/ftan. j yalav (Iph. T. 1480). So, too, o6v . . . | TrpdouTrav (Ion 925), narpoQ \ roi'fiov (Ion 725, Med. 746), i .saepe causam dixeris pendens advorsus octo artutos, audacis viros, valentis virgatores. (As. 564) ubi saepe ad languorem tua duritia dederis octo va lidos lictores, ulmeis adfeetos lentis virgis. (As. 574) (atque) auditavi saepe hoc volgo dicier, solere elephantuni gravidam perpetuus decern esse annos; (St. 167) noil quinquaginta modo, (jiunlringentos filios hain't atque equidem lectos sine probro: (Bacch. 973) 112 Verbum nullum without separation by the verse in Bacch. 785 (by emendation), Ter. Eun. 88. I'llns, with neque preceding, is separated in Ter. .M. 85. 258 University of California Publications. [Class Phil. The last passage is from a canticum, and is ascribed by Leo to an ampliticator. In the other examples some attendant features are worth noting. Respect for unity is shown in duae — duobus (St. 539), duo—tertio (Most. 775), 113 and in the isolation of adjectives and nouns in the second verse in the two examples from the Asinaria. In most of the cases the numeral follows the noun, or if it precedes the separation brings into prominence important elements (As. 564, 574, St. 168). A few cases of omnes are in place here: 114 hariolos, haruspices mitte omnes; (Amph. 1132) quin edepol servos, ancillas domo certum est omnia mittere ad te. (^as. 521) deartuasti dilaceravisti atque opes confecisti omnes, res ac rationes meas: (Capt. 672 ap. Nonium) ita res divina mini fuit: res serias omnis extollo ex hoc die in alium diem. (Poen. 499) Ebodum venimus, ubi quas merces vexeram omnis ut volui vendidi ex sententia: (Merc. 93) servos pollicitust dare suos mihi omnis quaestioni. (Most. 10S7) ubi ego omnibus parvis maguisque miseriis praefulcior : (Ps. 771) atque me minoris facio prae illo, qui omnium legum atque iurum fictor, conditor cluet; (Ep. 522) fateor me omnium hominum esse Athenis Atticis minimi preti. (Ep. 501) The first six examples, in which the adjective follows in the second verse, involve no violation of verse-unity ; the last three, however, are certainly, from an English standpoint, more de- structive of unity. (Cf. also the separation of tot in Poen. 582.) It is likely that the adjective is more separable than the corresponding word in English : the evidence for this is found in the apparent separability of numerals in general, and 113 Cf. Poen. 898. 114 For omnes in Ter. cf. And. 77, 667, Eun. 1032. Similarly complures, Ter. Ad. 229 (cf. plurumi in Plautus, above, p. 237); pauci, Ter. Hec. 58; aliquod, Ter. Phor. 312. Cf. Norden, Aeneis Buch VI, 390. Vol. l] Prescott— Thought and Verse in Plautus. 259 in the usage of the corresponding words in Greek verse. 115 Cer- tainly the explanation of the separation of numerals is more likely to be found in inherent qualities of the numerals as such than in such attendant features as the metrical convenience of the cretic omnium at the end of a verse. VII. Proper and improper numerals, pronominal adjectives, and in particular possessive adjectives were separated without essen- tial disturbance of verse-unity. This inherent separability seems to be proved not only by the treatment of these words in Plautus, but by the evidence furnished by early Latin prose, and by Greek prose and verse: the nature of the evidence sug- gests that this separability was an inherited trait. The opera- tion of Wackernagel's law and of the law of pronominal attrac- tion is a further manifestation of the looseness of the bond that binds pronominal adjectives to their nouns. The separation of possessive adjectives was probably promoted by the unemphatic nature of the words, which suffered a loss of their individuality. These conclusions do not differ essentially from those of Appuhn. In the treatment of attributive adjectives, however, I hope that something has been gained by an atteiript to interpret, within the limits set by the paper, the passages illustrating sep- aration. We found that attributives following the noun and separated were regularly expressions of ideas ranging from pre- dicative to amplifying, and the separation was usually attended by features that reinforced the unity of the verse. We found, too, that when the separated attributives preceded their nouns, although from an English standpoint the unity of the verse was 115 For the ordinary numerals in Ter. cf. Eun. 332, Phor. 638, Ad. 46. For Greek examples cf . elg, ovdeig, fiqdeig, Menander 535, 3 ; 282 ; 382 ; 397 ; 128, 3; Philemon 4, 13; 28, 9; other numerals, Menander 7, 1; 357; 547- 548; Philemon 12; 89, 7; nag, Menander 13, 2; 173; b/ioc, Menander 67, 2; KdvTeg, arravTcg, Diphilus 17, 2; Philemon 91, 7; Menander 292, 4; 363, 7; 404, 7; 532, 1; tto'aM, Menander 593. And for numerals in early Latin prose, cf. Altenburg, 1. c. 524 ff. 260 University of California Publications. [Class Phil. impaired, there were almost always associations of sound or sense that reasserted the unity of the verse ; more often the unity was apparent in the organization of the thought than in the superficial colligation resulting- from sound-effects. We may not always he confident that the resultant effects rep- resent efficient causes: in the matter of alliteration this is espe- cially true. The confinement, in most cases, of alliterative, groups to a single verse attests the entity of the verse, but allit- eration is seldom more than an incidental factor in separation : usually other and stronger factors appear along with alliteration. Metrical convenience is evident in the position of some words, especially those of considerable length, cretic words, and the possessive adjectives of pyrrhic and iambic measurement : the position convenient for such words may have conduced to sepa- ration. Again, however, other factors are usually discernible. Indeed, the total effect of a verse or couplet is a product of many factors: it is not easy to say that one is more important than another. But it seems to me noteworthy that in so large a number of separated attributives, the unity of the verse, if my interpretation is correct, is effected by internal organization rather than by superficial colligation. So much so that in cases like maxumo | me opsecravisti opere, optuma | vos video oppor- tunilate, tesseram | conferre si vis hospitalem I prefer to recog- nize the beginnings of a freer technique rather than admit metrical convenience and alliteration as really dominant factors in the separation. Such cases are rare; nor may anybody deny the essential unity of verse, the practical identity of verse and thought, in the ex- amples under discussion. The effect is often crudely simple, but in many cases the poet is far from being a clumsy crafts- man; he shows no little competency in making verse-unity a means of bringing into effective relief associated thoughts and sounds; and occasionally he uses the beginning and the end of the same verse, the beginnings of successive verses, in ways that indi- cate a consciousness of the opportunities, not merely of the limi- tations, presented by verse-unity. It is also significant that we can find so little positive proof of Vol. l] Prescott— Thought and Verse in Plautus. 261 the influence of his Greek sources: 116 he seems rather to be work- ing out his own problems in the spirit of his own language, fash- ioning his verse with nice adjustment of sound-effects peculiar to Latin, often producing a neat balance or antithesis which has yet to be proved to result from a study of Greek rhetoric, and happily conserving, even within the limits set by verse-unity, the simpler forms of interlocked word-groups, which are as characteristic of the organizing power of the Roman mind as any phase of their political administration. These same word- groups, however, must sometimes break down the barriers, and maxumo | vie opsecravisti opere, optuma \ vos video opportuni- tate, Usseram \ conferre si vis hospitalem perhaps point the way which leads to greater freedom. Only after further investigation is it safe to take the historical point of view and ask ourselves what is Plautus 's precise posi- tion in the development of verse-technique. In the answer to that question we must not be too hasty in placing him near the beginning of art-poetry in Latin : the comic verse under discus- sion is the most capacious of the commoner forms of metre ; and this verse conveyed the conversational Latin of the day to an audience that must catch at once the effects of sound and thought. Epic verse and tragedy were created under different conditions. Some of the simple directness of Plautus 's verse is perhaps to be attributed to these conditions rather than to the chronological proximity of the Saturnian verse. But in the present paper we have been interested only in suggesting some 16 Without further investigation of Greek technique the statement must remain in this vague form. It would be easy to find parallels from Eurip- ides, and some cases from the New Comedy, of Plautus 's postponement of adjectives and nouns to the beginning of the second verse, and of post- ponement for antithetical effects, but the running over of the thought to the caesura of the second verse, familiar to readers of Greek tragic poetry, is the exception rather than the rule in Plautus; nor are the features common to Greek and Plautine verse too hastily to be regarded as merely imitative in Latin verse, especially in the case of antithetical effects. Investigation, particularly of the technique of Aristophanes, Euripides, and the New Comedy, based upon sympathetic interpretation, must precede any more precise statement of Plautus 's relation to his models in these respects. 262 University of California Publications. [Class Phil. ways of interpreting a small part of the evidence that bears upon the question which Leo has answered, forestalling the investiga- tion of the subject in his admirable statement of the historical position of Plantns in this phase of verse-technique. °°0 015 3, UNIVERSITY OF CALIFORNIA PU BLICATION S- CONTI N u EDi ANTHROPOLOGICAL MEMOIRS. (Quarto). Vol. I. Explorations in Peru, by Max Uhle (in preparation). No. 1. The Ruins of Moche. No. 2. Huamachuco, Chincha, lea. No. 3. The Inca Buildings of the Valley of Pisco. ASTRONOMY.-W. W. Campbell, Editor. (Lick Observatory, Mt. Hamilton, Cal.) Publications of the Lick Observatory.— Volumes I-V completed. Volume VI (in progress). No. 1. A Short Method of Determining Orbits from Three Observations, by A. O. Leuschner. No. 2. Elements of Asteroid 1900 GA, by A. O. Leuschner and Adelaide M. Hobe. No. 3. Preliminary Elements of Comet 1900 III, by R. H. Curtiss and C. G. Dall. Contributions from the Lick Observatory.— Nos. I-V. Lick Observatory Bulletins.— Volumes I— II I completed. Volume IV (in progress). BOTANY.— W. A. Setchell, Editor. Price per volume $3.50. Volume I (pp. 418) completed. Volume II (in progress). No. 8. The Resistance of Certain Marine Algae to Changes in Osmotic Pressure and Temperature, by W. J. V. Osterhout. No. 9. The Role of Osmotic Pressure in Marine Plants, by W. J. V. Osterhout. No. 10. On the Importance of Physiologically Balanced Solutions for Plants, by W. J. V. Osterhout. No. 11. The Antitoxic Action of Potassium on Magnesium, by W. J. V. Osterhout. No. 12. Cytological Studies in Cyanophyceae, by Nathaniel Lyon Gardner. ... . ... . . . Price, No. 13. On a small Collection of Mosses from Alaska, by J. Cardot and T. Theriot Price, EDUCATION.— Edited by the Department of Education. Price per volume $2.50 Volume I (pp. 424). Notes on the Development of a Child, by Milicent W. Shinn Price, Vol. II (in progress). — No. 1. Notes on Children's Drawings, by Elmer E. Brown . . . ' Price, Vol. Ill (in progress). —No. 1. Origin of American State Universities, by Elmer E. Brown Price, No. 2. State Aid to Secondary Schools, by David Rhys Jones Price, ENGINEERING. — Edited under the direction of the Engineering Departments. This series will contain contributions from the Colleges of Mechanics, Mining, and Civil Engineering. Volume I in progress. No. 1. Bulletin of the Department of Civil Engineering. I.— Moment Diagrams and Typical Live Loads, by Charles Derleth, Jr., Pages 1-9, Plate 1. January, 1907. . Price, GEOLOGY. — Bulletin of the Department of Geology. Andrew C. Lawson, Editor. Price per volume $3.50. Volumes I (pp. 428), II (pp. 450), III (475) and IV (462), completed. Volume V (in progress). 4. New Cestraciont Teeth from the West American Triassic, by Edna M. Wemple. Price, .10 5. Preliminary Note on a New Marine Reptile from the Middle Triassic of Nevada, by John C. Merriam Price, .10 6. Lawsonite, Columbite, Beryl, Barite, and Calcite, by Arthur S. Eakle. Pp. 81-94, pi. 10. January, 1907. . Price, .10 7. The Fossil Fishes of California, with Supplementary Notes on other Species of Extinct Fishes, by David Starr Jordan. Pp. 95-144, pis. 11-12. April, 1907 Price, .50 8. Fish Remains from the Marine Lower Triassic of Aspen Ridge, Idaho, by Malcolm Goddard. Pp. 145-148, June, 1907. Price, .05 2.25 .50 .50 .75 10 UNIVERSITY OF CALIFORNIA PUBUCATIONS-(CONTINUED) PATHOLOGY.— Alonzo Englebert Taylor, Editor. Price per volume, $2.50. Volume I (pp. 347) completed. No. 5. On the Autolysis of Protein, by Alonzo Englebert Taylor. I one No. 6. On the Reversion of Tryptic Digestion, by Alonzo Englebert Taylor, f cover No. 7. Studies on an Ash-Free Diet, by Alonzo Englebert Taylor. Price, .20 No. 8. On Fermentation, by Alonzo Englebert Taylor. . . Price, 2.00 No. 9. On the Synthesis of Protein through the Action of Trypsin, by Alonzo Englebert Taylor. Price, .05 PHILOSOPHY.— Volume I, completed. Price, $2.00. PHYSIOLOGY.— Jacques Loeb, Editor. Price per volume $2.00. Volume I (pp. 217) completed. Volume II (pp. 215) completed. Volume III (in progress). No. 7. On the Counteraction of the Toxic Effect of Hypertonic Solutions upon the Fertilized and Unfertilized Egg of the Sea-Urchin by Lack of Oxygen, by Jacques Loeb. No. 8. On the Production of a Fertilization Membrane in the Egg of the Sea-Urchin with the Blood of Certain Gephyrean worms (a preliminary note), by Jacques Loeb. No. 9. Note on the Synthesis of Protein through the Action of Pepsin, by T. Brailsford Robertson. John Bruce MacCallum : On the Mechanism of the Physiological Action of the Cathartics. ....... Price, 75c. ZOOLOGY.— W. E. Ritter, Editor. Price per volume $3.50. Volumes I (pp. 317) and II (pp. 382) completed. Volume III (in progress). Commencing with Volume II, this series contains Contributions from the Laboratory of the Marine Biological Association of San Diego. No. 6. Dinoflagellata of the San Diego Region, II. On Triposolenia,] a New Genus of the Dinophysidae, by Charles Atwood Kofoid. Pages 24, Plates 3. No. 7. A Discussion of the Species Characters in Triposolenia, Charles Atwood Kofoid. Pages 10. No. 8. On the Significance of the Asymmetry in Triposolenia, Charles Atwood Kofoid; Pages 7. No. 9. Ostracoda of the San Diego Region. II.— Littoral Forms, by] * n Chancey Juday. Pages 22, Pis. 3. ' ' one No. 10. Cladocera of the San Diego Region, by Chancey Juday. Pages | Price 2, text-figure 1. *- J .25 No. 11. The Marine Fishes of Southern California, by Edwin Chapin Starks and Earl Leonard Morris. Pages 93, Plate 1. . Price, .75 No. 12. Biological Studies on Corymorpha. II. — The Development of C. palma from the Egg, by Harry Beal Torrey. Pages 46, 33 text figures. . Price, .50 No. 13. Dinoflagellata of the San Diego Region, III. Description of New Species, by Charles Atwood Kofoid. Pages 42, Plates 12. Price, .50 UNIVERSITY CHRONICLE.— An official record of University life, issued quarterly, edited by a committee of the faculty. Price, $1.00 per year. Current volume No. IX. Address all orders, or requests for information concerning the above publications to The University Press, Berkeley, California. European agent for the series in American Archaeology and Ethnology, Classical Philology, Education, Philosophy, and Semitic Philology, Otto Harrassowitz, Leipzig. For the series in Botany, Geology, Pathology, Physiology, and also American Archae- ology and Ethnology, R. Friedlaender & Sobn, Berlin.