STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. THE USE OF THE OE-DIPHTHONG IN PLAUTUS By Andrew R. Anderson In connection with my Doctor's dissertation, in which the status of the ez-readings in the Plautine MSS was examined, the late Professor Minton Warren had originally suggested that I undertake a similar study of the oe-diphthong. The purpose of this paper is to carry out that suggestion. First let me define the three categories in which the oe-diph- thong may be considered: I. Words in which the oe-diphthong was never reduced to u. If a reduction ever took place, it was not to u but to e, and the change was generally confined to Low Latin or Romance; cf. Ital. pena^poena; Span, hedo( foedus, "filthy." In this category are native Latin words containing oe for original I.-E. ot-diphthong, amoenus, coenum (but cf. Walde hat. et. Worterbuch, s. v. "caenum"), foetor, foetidus, foedus "filthy," foedus "treaty," moenia, proelium, oiei; words contain- ing an oe-diphthong produced by contraction, Cloelius, coepi, coetus, oboedio; Greek loan-words in which the oe is the tran- scription of 01 or a>, poena, comoedia. II. Words in which it stands for I.-E. oi, and in which in Clas- sical Latin it was reduced to u, as munus, munera, munio, *mune, milniceps, municipium, communis, inmunis, murus, muto, mutuus, spuma, utor, etc., cunae, cunio, euro, cura, etc., ludus, ludo, etc., plures, plurumi; punio, punicus, puniceus, and probably huius, cuius, 1 etc. For convenience and clearness I give uniformly u for the earlier oe. Obviously the discussion of this category will be the most important part of my paper. III. Loan-words from the Greek, in which oe is explained as the transcription of v, as Antamoenides, lagoena. For further examples see Schuchardt Vokalismus, II. 278 ff. The first category requires no further treatment in this paper, and the third, in my opinion, can best be understood after a full 1 1, e., if the development be hoiius > huius, quoiius (pronounced eoiius) > cuius. [Classical Philology IV, July, 1909] 291 292 Andrew R. Anderson discussion of the second; consequently I begin by citing those instances in which the MSS of Plautus give or point to the archaic oe for the classical u. Bacch. 926: moenitum. Capt. 254: circummoeniti. Cist. 540: admoenivi. Pers. 554: moenitum (munitum A, monitum P). 559: moenita muro A, P n. 1. Poen. 990, 991 : Punice .... Poenus Poenior, Poenior Bothe, punior libri. Pseud. 229: Phoenicium poeniceo. 384: admoenire. 5856 : admoenire (a quotation of 384). Stick. 695: moenia (=munera). Trin. 24: inmoenest, immunest A, In menest P. True. 102: oenus B, unus CD. Morphologically noenum, Aul. 67, Bacch. xxiii ? (Goetz and Schoell) belongs here, but it was used only as an archaic form, and is therefore in a class by itself; vid. Walde s. v. "ndn." For some conjectural oe-readings see the Triumvirate edition Mil. 223, 228, Rud. 934, True. 2. Rud. 692 moenia BCD, which Schoell changes to moeniam, is hardly worth considering. The reduction of oe to u is generally regarded as having taken place in the course of the second century B. c. ; so Sommer Handbuch, p. 88. Lindsay LL., p. 240, puts it as far back as the beginning of the century, and I feel confident that even he did not date the change too early. Probably the earliest epi- graphic evidence for the change is found in utier of CIL I. 33: Quei apice insigne Diah's y?aminis gesistei, Mors perfect tua ut essent omnia brevia, Honos fama virtusque gloria atque ingenium, Quibus sei in longa liewset tibe utier vita, Facile facteis superases gloriam maiorum; 5 Qua re lubens te in gremiu, Scipio, recipit Terra, Publi, prognatum Publio, Corneli. This P. Cornelius Scipio has not been certainly identified. Mommsen puts the close of his life anywhere from 204 to 154, and on account of such linguistic peculiarities as the lack of -d in the ablative and the reduction of oe to u in utier 4, scholars have The Use of the OE-Diphthong in Plauttjs 293 been inclined to favor a later rather than an earlier date. How- ever, after the reading of in agro Teurano, postscript to the SO de Bacchanalibus CIL I. 196, of 186 b. a, and the discovery of the decree of Aemilius Paulus, CIL II. 5041, of 189 b. a, the abla- tival -d no longer returns to plague us in Plautus, and inasmuch as in every other respect also the orthography of this Scipio inscription agrees with that of Plautus, e. g., the differentiation of I in insigne 1, vita 4, Publi .... Cornell 7, from the ei- diphthong in quei 1, gesistei 1, sei 4, tibe 4, facteis 5, the failure to use double consonants in gesistei 1, licuiset 4, superases 5 (cf. essent 2, terra 7), it will probably be correct to regard the spell- ing utier as part of the evidence for the reduction of oe to u at an earlier date than commonly supposed. 1 Even if a later date is insisted on for the inscription, it should be borne in mind that as orthography always lags behind, generally far behind, changes in pronunciation — this being especially true of the Scipio elogia — there is still nothing in this inscription to disprove the view that the oe-diphthong had already been monophthongized to u in the times of Plautus. Indeed, I hope to present some evidence in favor of the view that it had been reduced at least as early as the earliest of his datable plays — the Miles Gloriosus, which was brought out shortly after 206 B. o. But before presenting this additional evidence, let me consider two passages that seem to point to a contrary conclusion — Pseud. 229: Cras Phoenicium poeniceo corio in vises pergulam. The date of the Pseudolus is 191. The classical form olpoeni- ceus is puniceus, and the latter orthography might be expected for Plautus, if the statement of Lindsay above referred to be cor- rect. But in the present passage the diphthong has probably not been reduced, and that poeniceo should be read here is indispu- table. At first sight this passage would seem to prove that the pronunciation as well as the orthography would be poeniceus, a On the assumption that both oi and ou in the SC. de Bacch. are archaisms stand- ing for an actual pronunciation u, it would be possible to regard PLOVS (which occurs three times) as a mistaken restoration for *plois *ploes (cf. eouro for coero, CIL 1. 1419) and so bring it into line with PLOIRVME, CIL I. 32, and ploeres, Cic. De leg. iii. 3. 6. 294 Andbew K. Anderson poenicus, poenio, moenio, commoenis, oenus, etc. It has been brought out by L. Buchhold 1 that paromoeosis was at its height in Plautus and Ennius and that this practice has given us much evi- dence regarding pronunciation and orthography in our author. Yet in his fondness for punning Plautus sometimes even went so far as to use exceptional or archaic forms; cf. Capt. 67: iudices iustissumi domi duellique duellatores optumi. Here duelli and likewise duellatores were archaisms even for the time of Plautus; duelli isn't even found in a single MS for this passage. Yet Plautus' fondness for like sounds leaves no doubt as to the reading. So Amph. 1058 : corrumpta sum atque apsumpta sum, where I believe Fleckeisen is perfectly correct in reading cor- rumpta against corrupta of the MSS. Mil. 1407: dispennite hominem divorsum et distennite. Here the reading dispennite reported by Nonius for dispendite carries with it the change of distendite of the MSS to distennite. So in Ennius Epig. ii (CPL, Ennius Sat. 1. 66) we should probably read with Bergk: Nemo me dacrumis decoret nee funera fletu faxit. We must not then close our eyes to the possibility that poenieeo in Pseud. 229 may have been given an archaic pronunciation by Plautus for the sake of closer similarity to Phoenicium. The other passage is Poen. 990, 991 : Null us me est hodie P oenus Poenior. Poenior is the correction of Bothe; the MSS read punior. Poenior would seem to point to poenio, poenicus, etc., but here, as in the passage from Pseud., the form is shown to be exceptional, being in fact a coinage for the nonce, and standing therefore alto- gether by itself. 1 Deparomoeoseos (allitterationis) apud veteres Romanorum poetas usu, Lips. 1883 ; ef . E. B. T. Spencer Adnominatio in the Plays of Plautus with Special Reference to Questions of Pronunciation and Orthography, Rome, 1906. Unfortunately Mr. Spen- cer's investigation leaves much to be desired. The Use of the OE-Diphthong in Plautus 295 Plautus has left us no statement regarding his orthography of oe-u as he has about his use of ei-%. Cf . Bud. 1305, 1306 ; True. 262-64. The evidence for his pronunciation of oe must be gath- ered principally from instances of assonance, the most important being Mil. 324: Sc. abi, ludis me, Palaestrio. Pa. turn mihi sunt manus inquinatae. Sc. quidum? Pa. quia ludo luto. This is one of a great many instances where Plautus puns a long with a short vowel and I cite the following: Amph. 1: Ut vos in vostris voltis; cf. 1006. 318: exossatum os; cf. 342. 498: uxore usuraria. As. 142: pane in pannis. 568: fideli infidus. Bacch. 362: crucisalum me ex Chrysalo (=crusalo); cf. 687. 490: malis malim modis. Merc. 83: amens amansque. 161: a portu apporto; cf. Ep. 595; Mil. 316; Rud. 1225. It may be noted that generally (but not always) when Plautus puns a long and a short vowel in arsi, there is a double consonant or its equivalent after the short vowel. In thesi he is freer in his treatment both of sounds and of quantities. This passage from Mil. 324 is not conclusive for the identity of the quality of the sounds compared until confirmed by other evidence, and this is found in Bacch. 129: Non omnis aetas, Lyde, ludo convenit a passage that, as far as the quality of the sounds compared is concerned, is to be grouped with Bacch. 362: Crucisalum me ex Chrysalo, 687 : in cruciatum Chrysalum, 1183: Chrysalus .... excruciem. Together they prove that at this period Greek v was represented in Latin by u, so that in 129 the sounds compared are both in arsi and identical in both quality and quantity, and would be repre- 296 Andrew K. Anderson sented phonetically by Lude, liido, and etymologically by Lude loedo. In other words, the sound of the oe (<^ oi) in loedo (y ludo) had been reduced to u. I cannot here agree with Lindsay LL, p. 248, who separates Bacch. 129 from Bacch. 362, 687, 1183, and would represent the phonetics of Lyde, ludo by Lude, ludo. I hold therefore that in Mil. 325 the pronunciation is faithfully represented by ludo, luto and in Bacch. 129 by Lude, ludo. Hoffmann, in BB XXVI. 137ff., denies the possibility of ludus, liido ever having come from the I.-E. stem loid- (cf. v. Grien- berger Wiener Sitzungsber. CXLII. VIII. 151), which he says could have resulted only in leid- y lid. To make good his conten- tion he finds it necessary to treat the epigraphic evidence for loed [loid) as of no value, and the instances of its occurrence as pseudo- archaisms, interdependent in error. Accordingly he takes ludo from an I.-E. stem *ghleu-d-, cf. %Xeu-a£<». If the conclusions of Hoffmann and v. Grienberger are correct, then the puns ludo luto, Lyde ludo have no bearing on the question of the oe-diphthong in Plautus. Had they known of these puns they would probably have treated them as confirmatory of their own conclusions (and to complete their case they would only have been obliged to dis- prove the original oi-diph thong in utor, euro, etc., as well). But the conclusions of Hoffmann are to be rejected for the following reasons: a) His objection to the phonetics loid-yiud- is not valid; cf. Walde, s. v. "ludo;" Sommer, pp. 88, 91; and I am not aware that Lindsay has changed his view of the etymology of ludo given in Led. Lang. 248. Furthermore, if the stem *leub- (e\ev0e/>o?) gives *loub- > loib- y leib- > lib-, why should not according to Hoff- mann (since in his argument he ignores the labial as a factor in this development) *ghleu-d give loud- yloid-yieid-yiid-? b) To deny the development of liido from loido, is to exclude all possible connection of these words with Osc. luisarifs: cf. Buck Oram, of Osc. and Umbr., art. 138 and p. 248, No. 21. If the reduction of oi to ii was already an accomplished fact in the time of Plautus in ludo, it is reasonable to suppose that the reduction had become uniform throughout this whole category ; and Amph. 498 uxore usuraria seems to confirm this view, as The Use of the OE-Diphthong in Plautus 297 this phrase was probably pronounced ussor(e) ussuraria. 1 Trin. 181, 182: Neque adeo hasce emi mini nee ussurae meae: Illi redemi russum is of doubtful value on account of the distance of ussurae and russum from each other. Here may be quoted Most. 209: Cur obsecro non curem t and Poen. 354: Qur ego id cureml nam qui istaec magis meast cwratio? as reasonably sure proof of the reduction of oe to u in euro, etc. The value of the two quotations is not impaired by the fact that cur does not contain I.-E. u (yet cf . J. Schmidt KZ XXXII. 405, and Persson IF II. 248), or by the fact that the orthography in the time of Plautus may have been quor. The latter form, how- ever, rests only on the express testimony of Velius Longus GL VII. 77. 9K, and lacks epigraphic confirmation under the Repub- lic, whereas in CIL I. 1454. 1 (time of the Gracchi) we have que = cur. Etymologically the quality of the u in cur is the same as in fur j cf. Gk. (fxbp, and Lorenz' note on furinum .... forum, Pseud. 790. Quintilian i. 7. 26 (cf. i. 4. 10, 11) has told us that by the teachers, of his boyhood seruus, ceruus, etc., were still spelled seruos, ceruos, in order that the gemination of the u might not cause them to be pronounced serus cerus. Cur was probably pronounced in the same way in the time of Plautus as in the time of Augustus, and nothing was to be feared from the simpli- fied spelling of quor y quur y qur or cur. The spelling quor cited by Velius Longus and used by Varro was etymological, and conservative, and theoretical, and not representative of its pronun- ciation, which was cur even for the time of PI. Here might be quoted Velius Longus GL VII. 58. 4K, to the effect that words like primitiuus, though spelled -uos until in the eighth century of the city, were pronounced -uus and not -uos. Combining this fact with Bersu Die Gutturalen, p. 53, where, on the basis of such spellings as oquoltod ( = occultod) CIL I. 196, of 186 b. o. and 'For the weakened pronunciation of x to ss cf. the MSS in Ep. 248; Bud. 729; Stick. 163, 175, 272; True. 913, 940; Pseud. 1107 luxantur, lustrantur. Nonius 40, 24 M, and the development of x in Romance. 298 Andrew R. Andebson quom ( = prep, cum) OIL I. 34. 3, end of the sixth century of the city, he proves that in Republican Latin labialized gutturals before u (<(o) were unknown, we have still further proof that guor was pronounced ciir even by Plautus. Cf. Lindsay LL iv. 137. 3, p. 300. The instances in Plautus in which an original M-sound is punned with an original oi-diphthong are unfortunately not more numerous than those already cited, unless we include here the very attractive interpretation of Aul. 562 f . by Prescott in C. P. II, 335 f., which is possible only on the hypothesis that coera had already been reduced to cura. Above has been presented the evidence for believing that the oe (en) which we find monophthongized to u in Classical Latin had already under Plautus been thus reduced in pronunciation. Admittedly the evidence is inconclusive, but the hypothesis has the virtue of explaining the phenomena met with in a more satis- factory way than the opposite view put forth, e. g., by Brugmann KVG I, p. 84, where he says, "oe gait zur Zeit des Plautus." Here he means pronunciation as well as spelling. Whether the Romans were swift enough innovators in spelling reform to have made the change in orthography also is doubtful, but I should be inclined to use the orthography punio, munio, communis, plures, euro, utor except where there seem to be special reasons for doing the contrary. The instances cited on pp. 2 ff . may be considered individually. Whenever any of them are rejected, it is only their diphthongal pronunciation that is denied. Noenum, Aul. 67, Bacch. xxiii; Poenus Poenior, Poen. 991; Phoenicium poeniceo, Pseud. 229, may be accepted for reasons previously stated. In addition to these there is much to be said for Pleckeisen's moenit (before moenia) Mil. 228, both because of the assonance and because of the dignified nature of the passage. There is less to be said for his moeni 223, as there is no assonance there. In the remaining passages, where the MSS give oe for ' u, they should not be fol- lowed except for special or additional reasons, and it must be admitted that these special or additional reasons are likely to be more or less subjective: e. g., moenitum, Bacch. 926, stands in a The Use of the OE-Diphthong in Plautus 299 passage that has a heroic tone and might gain in impressiveness by the diphthongal pronunciation. The same reasons could be adduced for Cist. 540; Pers. 554, 559; Pseud. 384, 585 a (=384). However, it seems unreasonable that we should read moenitum in Pers. 554, and moenita 559, unless we extend the oe also to muro 559, and emend munitum muro to moenitum moero 553, and murus to moerus 560. The Triumvirate editors have admitted commoe- nibo against the MSS in Bud. 934. Their reading of moeris for vostris, True. 2, is not felicitous. No special reason seems to exist for circummoeniti, Capt. 254, moenia ( = munera), Stick. 695, inmoene, Trin. 24, oenus, True. 102. I now come to the consideration of the readings of category III, namely, those in which oe is said to be the transcription of Greek v. lagoenam, Cure. 78 ;* Antamoenides, Poen. cast of characters. The Greek originals of these words are respectively \dyvvos and 'AvTafiwlSi}?; cf. K. Schmidt in Hermes XXXVII. 356. The precarious position of oe as a transcription for v may be gathered from Schuchardt's discussion of it, Vokalismus II. 278-87. The case for such transcription is not supported by inscriptional evidence, but rests on a few MS readings, and on the very doubtful readings in a single passage in a single gram- marian — Maximus Victorinus OL VI. 196. 3K: "Quae sunt litterae peregrinae? y et z. Quare peregrinae? Siquidem a nobis propter Graeca nomina adsumptae sunt, ut puta Hylas, zephyrus. Quae si adsumpta non essent, Hulas et sdepherus (sdephurus?) diceremus." Here hoelas (sic) is given only by one MS, Goth, and by c, while sdephoerus is the conjecture of Kibbeck BM XII. 431. Schuchardt loc. cit., argues in favor of these two readings. Kitschl Bonner Sommerkat. 1856 had denied that oe could be used as transcription for v, but had admitted that it was the regular transcription for v. The ques- tion of chronology, however, must be considered, and I have no hesitation in rejecting the grammarian's statement (reading 1 It may be that Fleckeisen is right in reading lagaenam (from \dyi)vos ; of. scaena < o-Kr)vj) making thus a better pun on leaenam of the preceding line. If so, its dis- cussion is outside the scope of this paper. But on the whole, I cannot help regarding the reading lagaenam as improbable on the ground of the questionable standing of \iyrivos, and even with lagunam the pun is by no means a bad one. 300 Andrew R. Andeeson Hoelas and sdephoerus) as applied to Plautus. For in our poet's time v of Greek loan-words in Latin was pronounced u, as turan- nus, sucopanta. This is so well established as not to require proof here. Consequently if these words — lagoenam Antamoeni- des — are to stand thus in our text of Plautus, they are to be pro- nounced as though spelled with u, or, putting the matter in another way, with the oe that in the time of Plautus had been monoph- thongized to ii. So it is that K. Schmidt loc. cit., reads Anta- munides, and Vendryes Mecherches sur I'histoire et les effets de VintensitS initiate 284 regards the form lagoena as a reversed form of writing u. When then did the transcription of v by oe come into vogue? Ribbeck JJ LXXV. 317 holds that it came in with Accius, and in frg. 122 we find Froegiae. This view of Ribbeck's is dis- proved by Schuchardt II. 287 : " Hatte Accius wirklich oe fur v eingeftlhrt, und ware diese Schreibung schon im 7. Jahrhundert d. St. gang und gabe gewesen, so wtirde uns gewiss inschriftliche Belege derselben nicht fehlen" — a process of reasoning which, by the way, is equally fatal when turned against the conclusions of Schuchardt himself. We are safe then in concluding that the transcription of v by oe, if ever practiced, did not take place until imperial times, and did not affect the text of Plautus except so far as this was corrupted by having such forms as lagoenam and Antamoenides introduced into it. The fact that for Plautus' time oe was not used for v is still further evidence against Lindsay's interpretation of the sounds Lyde, ludo, Bacch. 129, where he holds, LL, p. 248, that ludo still has a diphthongal pronunciation of oe approaching the sound of v in Lyde. Pkinceton University