THE CLASSICAL QUARTERLY APRIL, 1927. ‘MOX.’ THAT mox means, or can mean, ‘soon,’ is an assertion which is made often and positively, in works of all sorts, from the ordinary dictionaries, such as Facciolati-Forcellini and Lewis and Short, and valuable writings on lexico- graphy such as Merguet’s lexicon to Cicero and Krebs-Schmalz’ A ntibarbarus, down to the latest little school book at which I have looked. As I had never been able to find a passage in which it clearly and unambiguously had that meaning, in any classical author, I have thought it worth while to examine all the instances in a few good writers, with a view to finding out how they really do use it. I have examined Plautus, Terence, and the oratorical and philosophical works of Cicero, for republican usage; Horace, as a transitional author in point of style; and finally Suetonius, as representing the practice of the Silver Age at a time when it had become fixed and settled, as understood by a man of little or no originality but a good deal of learning, who did not archaize. In all, I count one hundred and fifty-two instances, most of them textually certain. The first point to notice is that the three republican authors between them give only thirty-six, omitting the instances of quam mox, the meaning of which remains certain however we explain it. Mox is therefore, in the usage of the Golden Age and of the earlier period, a comparatively rare word, occurring on the average about four times in a medium-sized Teubner volume. This is worth noting, at least by teachers, for it shows at once that, whatever mox means, it is not the ordinary word for ‘soon,’ or it would certainly be commoner. Yet many prose-books, in their vocabularies, give it as the only equivalent. Further, an examination of all the passages (and I think I have omitted none, save, as already explained, those which contain guam mox) indi- cates that it does not mean ‘soon’ in any of them, but ‘ not now,’ ‘after a while,’ ‘by and by.’ I. ‘Mox’ CONTRASTED WITH ‘ Now.’ If a word means ‘soon,’ we should naturally expect to find it now and then contrasted and paired with some word meaning ‘late.’ This is particu- larly true of Latin, which is so fond of contrasting pairs, such as noctes ac dies, | femina uirque, aperte tecte (Cic. ad Ait., I, 14, 4 where Matthiae and Boot are NO, II. VOL. XXI. E 58 H. J. ROSE clearly right in rendering siue aperte siue tecte), Patulcius Clustus. We should expect, that is, to find some such phrase as aut mox aut sero or the like. What we actually do find is the contrast mox) (nunc, as follows : 1. Ter, Eun. 219, nam aut iam reuortere aut mox noctu te adiget horsum insomnia. 2. Cicero, in Pisonem 38, nondum commemoro rapinas . . uidebitur, ut cum fure . . . disputabo; nunc meam spoliatam fortunam con- feram, etc. 3. Cicero, Phil. IV. 12, urgetur nunc eis copiis quas iam habemus, mox eis quas paucis diebus noui consules comparabunt. 4. Cicero, De fin. I. 35, quae fuerit causa, mox uidero; interea hoc tenebo, etc. 5. De fin. IV. 5, qui enim sit finis bonorum, mox; hoc loco tantum dico, etc. 6. De fin. V. 60, quorum omnium quae sint notitiae quaeque significentur rerum uocabulis quaeque cuiusque uis et natura sit, mox uidebimus ; hoc autem loco, etc. 7. Tusc. disp. II. 26, sed hoc mox uidebo, interea unde isti uersus ? 8. De nat. deor. III. 37, hoc totum quale sit, mox ; nunc autem concludatur illud, etc. g. De re pub. I. 20, sed ista mox; nunc audiamus Philum. . Mox, Si It is to be noted that several of these examples not only contrast the future with the present, not a nearer with a more distant future, but show that the event looked forward to is not very near. In (1) Phaedria will come back either that very day (iam) or in the night, perhaps that same night (mox). In (3) the forces with which Antony mox urgebitur are as yet not even recruited, nor will the recruiting begin for some few days. ‘Soon’ is a relative term; anyone who said that our present industrial troubles would be soon over, and then explained that he meant next century, would be regarded as giving us very cold comfort; but to a geologist, a thing which was to happen in so short a time as a century would seem to be coming very soon indeed. Now, surely, when the whole action of the play is but twelve hours or so, as is the case in the Eunuchus, ‘soon’ and ‘not till night’ are very different things; and with events moving as rapidly as they were in 43 B.c., Cicero’s audience are meant to be cheered not so much by the thought that the new army will soon take the field, as by the assurance that it will do so in due course. In the other passages the reference is to a later passage in the same work; in the case of (8) no more seems to be said of Kleanthes’ theories (hoc totum) in the work as we have it, and much the same is true of (g); but the promise given in (4) and in (5) is indeed soon fulfilled (see sections 42 and 14 respectively of the same books); nor have we to wait long for the discussion of particular jAonesta promised in (6), for it begins seven sections further on. The same is true of some passages in which nunc or its equivalent is not expressed (see the next sectior not m« paralle passag like (4) ‘soon ’ ‘soon,’ the cor ‘after long. I 3 Cf. after th i has hea sj comes.’ him inte 4. J So The until nig 5. C The 6. F She short. 7: C. ‘Then aft 8. 7? ue! qui > should he like. horsum mox, Si um con- mox eis rea hoc tantum ificentur 1c autem ? cludatur rast the how that me back ox). In ecruited, e term; ver, and ‘iving us so short 7, surely, . case in ind with e meant oon take he other ise of (8) ‘k as we ) and in he same - honesta ; true of the next ‘MOX’ 59 section). But allthis does not prove that mox means ‘soon,’ merely that it does not mean ‘after a long while,’ which no one supposes it does. It is exactly parallel to the common uses of supra and infra, which may refer equally to a passage half a dozen lines away or one many pages off. We must put passages like (4) and (5) side by side with passages like II. 14 and 16 below, where ‘soon’ would not be an appropriate word at all. Granted that mox means ‘soon, I do not see how it could be used also of an event which is shown by the context, or explicitly said, to be somewhat remote ; but if it merely means ‘after a while,’ it can very naturally be used whether that ‘while’ is short or long. This appears further from the next section. II. ‘Mox’ USED OF AN EVENT NOT IMMEDIATELY FOLLOWING. 1. Plautus, Men. 47, ne mox erretis iam nunc praedico prius. Cf. M.G. 150 (deferred prologue) et mox ne erretis. ‘To avoid confusion after the play (the next act) begins.’ 2. M.G. 1168, ne ille mox uereatur introire, ‘lest he fear to enter when he has heard our story.’ 3. Poen. 596, ut sciamus quid dicamus mox pro testimonio, ‘ when the time comes.’ The pretended soldier has to complete his deceit of Lycus and bring him into court before any evidence can be given. 4. Rud. 111, quon furatum mox uenias uestigas loca ? So Trin. 864, credo edepol, quo mox furatum ueniat speculatur loca. The supposed thief, presumably, is not expected to begin his burglary until night. Cf. Rud. 417, accipiam hospitio, si mox uenies uesperi. 5. Casin. 84. mox hercle uero, post transactam fabulam, argentum si quis dederit, ut ego suspicor, ultro ibit nuptum. The play has hardly begun yet. 6. Poen. 404. AG. da ergo, ut credam, sauium. ADE. mox dabo, quom ab re diuina rediero. She is just setting out for the ceremony, which will not be particularly short. 7. Capt. 192. ibo intro atque intus subducam ratiunculam, quantillum argenti mi apud tarpezitam siet ; ad fratrem quo ire dixeram, mox iuero. ‘Then after that I will go to my brother’s.’ 8. Truc. 546. uerum abibo. quid ais? nunc tu num neuis me, uoluptas mea, quo uocatus sum ire ad cenam ? mox huc cubitum uenero. ‘I shall come after dinner and spend the night here.’ 60 H. J. ROSE g. Men. 430, ergo mox auferto tecum quando abibis. Menaechmus is just entering Erotium’s house, where he means to spend some time. 10. Terence, Andria 483. nunc primum fac istaec lauet, poste deinde quod iussi dari bibere et quantum imperaui, date; mox ego huc reuertor. Lesbia clearly does not expect to visit her patient again for some hours at least ; she has assured Archilis that all is going very well so far. In fact, we see no more of her. 11. Adelph. 204, de argento somnium ‘mox; cras redi.’ ‘ Not now, call again to-morrow.’ 12. Cicero, Acad. prior. II. 66, ut praepostere tecum agam; mox referam me ad ordinem. Mox here is precisely equivalent to av@s. 13. De diuin. I. 47, discedo parumper a somniis, ad quae mox reuertar. He does so in the next section, but compare 14. Der. p. II. 65, puto nobis mox de instituendis et conseruandis ciuita- tibus aptiorem, Tubero, fore disserendi locum. So far as can be judged from the fragments and Augustine’s synopsis, de ciuit. Det II. 21, the aptior locus was in the next book. 15. De fin. IV. 8, nam de summo bono mox, ut dixi, uidebimus. This refers back to 5 (see above, I. 5) and forward to 14. Contrast 16. Acad. prior. II. 87, sed de physicis mox. This refers to 118. What meaning more exact than ‘ by and bye’ can be given to a word which is used indifferently to refer one section, six sections, some thirty sections, and a whole book ahead? Add de legg, 1. 54, where the promise to consider mox the views of Antiochus seems nowhere to be fulfilled in the surviving parts of the work, and de diwin. II. 38, where mox de sortibus refers forward to 85. III. SomE DovusBTFUL USEs. First, for the sake of completeness, let us get rid of a few passages in which it is highly doubtful if any such word as mox occurs. Plaut. Amph. 550, mox is a conjecture of Leo’s, and to my thinking a poor one. Pseud. 573, again mox is Leo’s conjecture only. Terence, Andr.507, some MSS. have ‘sed nilo setius mox puerum huc deferent ante ostium ;’ but the greater part omit the word, which is necessary neither to scansion nor to sense; the line has then a hiatus in diaerest, after puerum, which is at least as likely as an elision across the diaeresis. 1. Plaut. Merc. 781, haec uassa aut mox aut cras iubebo aps te peti. If perhap licebit uiro nl use, an TI several transla tion to other t also ust used it exampl mox in Th the half can ma: 2. | Th In does nc certainly summer the seas repeat tl The 5 ¢ Eur rupts he: ) spend ours at 1ct, we ww, call -eferam tar. cluita- nopsis, a word thirty mise to in the s refers ages in A mph. td. 5735 ie ‘sed rt omit ine has -elision ‘MOX’ 61 If mox is right, it means ‘ later in the day,’ and is contrasted with cras, or perhaps ‘to-night.’ The latter is apparently its meaning, ibid., 802, eadem licebit mox cenare rectius, and certainly Cist. 43, haec quidem ecastor cottidie uiro nubit, nupsitque hodie, nubet mox noctu. But in any case it is a strange use, and Bothe was perhaps right in conjecturing nox. Thus we have so far found not one instance where mox must mean ‘ soon,’ several where it cannot, and a number where ‘ soon’ is at least not the natural translation. We now proceed to examine Horace, whose style shows a transi- tion to Silver usage in some ways. In particular, we find him using mox with other than present and future tenses, as for instance Carm. III. 27, 69. He also uses it more freely than the earlier authors, nineteen times in all. If they used it proportionately as often, they would furnish some two hundred examples. In another way he leads up to the Silver Age—namely, the use of mox in enumerations, aS appears from the following examples : IV. I. Carm. IV. 4, 5 sqq. olim iuuentus et patrius uigor nido laborum protulit inscium, - mox in ouilia demisit hostem uiuidus impetus, nunc in reluctantes dracones egit amor dapis atque praedae. Three stages are enumerated : first the eaglet which can hardly fly, next the half-grown bird which feeds on lambs, finally the full-grown eagle which can master the hereditary foe of its race. 2. Carm. IV. 7, 11, et mox bruma recurrit iners. This ends a catalogue of the seasons. In neither of the above is the translation ‘soon’ appropriate; an eagle does not pass from half to full growth particularly soon, and midwinter certainly does not follow upon autumn any faster than autumn on summer, or summer on spring. Horace’s point in the whole passage is, not so much that the seasons pass rapidly, as that they pass, and we, unlike the moon, cannot repeat them indefinitely. V. ‘ Mox’ witH PAstT TENSEs. The instances are not many. 1. Carm. III. 27, 66. aderat querenti ... Venus. mox ubi risit satis, ‘ abstineto’ dixit. Europa has been lamenting for some thirty-five lines when Venus inter- rupts her. ‘Soon’ would be sheer nonsense here. 2. IV. 14,14, maior Neronum mox graue proelium commisit. 62 H. J. ROSE This mox corresponds exactly to Cassius Dio’s ére:ra (liv. 22, 3) when he, like Horace, passes from the description of the earlier campaign of I5 B.C. to deal with the later one. 3. A.P. 220. carmine qui tragico uilem certauit ob hircum mox etiam agrestis satyros nudauit. Non tta multo post tinuentam tragoediam tllud quoque genus inuentum esse, et quidem a Pratina, Aeschyli aequali, is Orelli’s note, and, like most of his notes, it is sensible. But the sense of his non ita multo post inuentam tragoediam is contained in carmine ... hircum, which in itself denotes an early stage. Mox is ‘ next,’ ‘afterwards.’ Of the categories I. and II. given above, there are plenty ot examples in Horace. Of the former we have instances in Carm. I. 4, 20, ‘Lycidan... quo calet iuuentus nunc omnis et mox uirgines tepebunt ’; here mox denotes the passage of enough time to make a pretty boy into a handsome young man. See also Carm. II. 1, 11, where mox is the length of time Pollio will take to write his history; III. 6, 25, where mox is used in contrasting the actions of a depraved woman before and after marriage (tam nunc, 23); I. I, 17, the merchant’s behaviour during and after the storm; Sat. II. 4,6; Epp. I. 1, 12, a particularly good instance of what the word really means, ‘ condo et compono quae mox depromere possim,’ ‘for future use.’ No one stores away what he will soon need. There is also Epp. II. 1, 126, where tam nunc and mox mark two successive stages of early education; 2, 95, where mox is contrasted with primum (93). In category (II.) we may reckon Carm. III. 6, 47, the famous phrase ‘ mox daturos progeniem uitiosiorem.’ Perhaps this had better been included under (IV.), since it gives a list of four successive generations. But in any case Horace’s wicked contemporaries will need some thirty years before they are replaced by their yet more vicious offspring. Again in Efp. I. 18, 76, we have the passage of a not inconsiderable time : qualem commendes etiam atque etiam aspice, ne mox incutiant aliena tibi peccata pudorem. The person incautiously recommended will surely not show himself impossible all at once. A shorter time, but still not particularly short, is indicated by mox in Sat. II. 3, 192, ‘ergo consulere et mox respondere licebit ?? Agamem- non must be questioned and give his answer before the inquisitive plebetus can criticize him. In Ep. II. 1, 191, the procession of captive kings which mox trahitur comes on after the long battle-scene is over; 2, 161, the wzlicus is still harrowing the land, mox frumenta daturus only when he has got in the harvest. In A.P. 168 and 183 mox simply means ‘ later.’ In one or two passages indeed, mox almost equals denique or tandem. Of this kind is Epp. I. 17, 55: Till in It | ‘N(ach fiir poste have be to this : Also th paucts » the best If little or here th will suf a tura prc Th is 2b1d. | iam tu! princep mox stl fautores mox ’); II. most st his law errorem in Thra after th T1b. 67, Same s mox eo Caligul: which a be addu infreque in Hisp life), Ca quite in its etyn maksubl hen he, B.C. to 1m esse, of his roediam y stage. ples in ae tes the g man. take to ‘ions of 17, the - X, 12, mpono yhat he ¢ mark d with e ‘ mox under y case ley are re have ossible ted by amem- us can h mox is still arvest. n. Of ‘MOX’ 63 meretricis acumina, saepe catellam, saepe periscelidem raptam sibi flentis, uti mox nulla fides damnis uerisque doloribus adsit. Till in the end she is not believed when her losses and griefs are real.’ It therefore appears that Krebs-Schmalz are too hasty in their statement, ‘N(ach)kl(assisch) ist es auch in der Bedeutung hernach, in der Folge, spaterhin, fiir postea.’ In most of the above passages, postea or posthac could perfectly well have been used, as will be seen by looking through them. I would attribute to this fact the comparative rarity of the word in the best republican authors. Also the Plinian idiom to which Antibarbarus objects (paulo mox, mox paulo, paucis mox horis), while it is not Ciceronian nor in keeping with the usage of the best authors generally, has in itself nothing irrational. If we look now at the Silver usage as illustrated by Suetonius, we find little or nothing which is not already in Horace. I do not propose to give here the whole of the ninety-four examples I have found in his works; a few will suffice, which I arrange under the categories already given. 1 Iul. 71, Masintham . . . abstrahentibus statim eripuit et mox ex prae- tura proficiscens . . . auexit. This is the familiar nunc)(mox transferred to past time. Of the same kind is 1bid. 94 statim)(mox (immediate contrasted with later result), Cal. 12, ‘Seiano iam tunc suspecto mox et oppresso’; Claud. 41, ‘ priuatus adhuc . . . mox princeps’; Ner. 35, ‘Octauiae consuetudinem cito aspernatus . . . eandem mox strangulare meditatus’; ibid., ‘olim adoptionis mox dominationis suae fautores atque rectores, 44 (primam curam . .. mox)’; Vesp. 3 (‘olim... mox’); De rhet. 1,‘ olim . .. synthesis uocabant, mox controuersias.’ II. Examples here are particularly abundant. I give half a dozen of the most striking: Jul. 28, Pompey forgot to exempt Caesar from the working of his law, ‘ac mox, lege iam in aes incisa et in aerarium oblata, corrigeret errorem’; Aug.g4, ‘tantam mox fiduciam fati Augustus habuit’; his confidence in Thrasyllus’ prediction, as shown by the publication of his nativity, was long after the delivery of the prophecy, which was before the death of Julius; Tib. 67, ‘ne mox... impar tantis honoribus inueniretur,’ compare in the same section, ‘multo ante . .. prospexisse’; Calig. 8, ‘ uersiculi imperante mox eo diuulgati’ (the interval is that between the birth and the accession of Caligula). Cf. Claud. 45, Ner. z, 6, Oth. 10, Dom. 1, De gramm. 23, in all of which a considerable lapse of time is indicated. Several passages could also be adduced in which a short time is indicated, as Iul. 22, 23. Indeed, it not infrequently could be rendered ‘ next,’ as Aug. 8, ‘ profectum mox auunculum in Hispanias . . . subsecutus’ (it was the next important event in Octavian’s life), Cal. 35, ‘ Thraeci et mox hoplomacho comparauit.’ This signification is quite in keeping with the use of the word generally, and also, I would add, with its etymology. It is connected (see Walde, Lat. etym. Wort., s.u.) with Skt. maksubhis, festinantibus (instrumental), and hence may well contain the idea of 64 H. J. ROSE passing rapidly on to the next point in the story, or the next item in an enumeration. III. In Suetonius, again, there are some usages which are not easy to classify. Aug. 6, ‘huc introire . . . religio est, concepta opinione ueteri.. . sed et mox confirmata,’ seems almost to mean ‘recently corroborated,’ which suggests the late Latin’ use of the word to mean ‘just,’ as Aug. Solil. II. 36, in puero mox nato, ‘in a new-born child’; however, it may mean simply ‘later.’ Aug. 47, ‘nec mox occasio aut causa traiciendi fuit,’ ‘ he never at any later time had any opportunity or reason for crossing’; Tib. 53, ‘nec ullo mox sermone dignatus est,’ ‘ never deigned to speak to her again’; zbid. 61, ‘ mox in omne genus crudelitatis erupit,’ following on an enumeration of particular acts of cruelty ; this suggests the occasional use of mox = denique, which we have already noted. IV. Mox in enumerations is so common in Suetonius that I quote but one or two instances. It is used in giving a list of the successive wives of an emperor, as Aug. 62, ‘Claudiam . . . duxit uxorem . . . mox Scriboniam in matrimonium accepit ...cum hac quoque diuortium fecit ... ac statim Liuiam ... abduxit.’ Cf. Claud. 26, Vit. 6. It also is found in lists of successive offices held by a person, as Galb. 4, ‘ praetor . . . exin prouinciae . .- praefuit . .. mox consulatum ... gessit’; Vesp. 2, ‘aedilitatis ac mox praeturae candidatus’ (cf. De gramm. g), of a series of conspiracies, Aug. 19 ‘(coniurationes) Lepidi iuuenis, deinde Varronis Murenae et Fanni Capitonis, mox M. Egnati, exin Plauti Rufi Lucique Pauli progeneri sui, ac praeter has L. Andasi.’ Cf. in various contexts of this sort, ibtd. 38 (series of enactments), 50 (series of seals), Claud. 6 (series of senatus consulta). This, of course, is much the same as the meaning ‘ next’ already mentioned. Thus, after examining a series of authors spread over some three hundred years, we have found no certain example of mox meaning ‘soon.’ It remains to ask why it was ever supposed to mean that. The mistake is not due to modern lexicographers, small or great, but goes back at any rate to Festus, p. 314, Lindsay: ‘quam mox significat quam cito, sed si per se ponas mox, significabit paullo post, uel postea’ (cf. pp. 123, 315). This, if we could suppose it due to Verrius Flaccus, would be a testimony of great authority ; but the very next article shows that in questions of lexicography Festus is by no means dependent upon Verrius alone: ‘ quatere ... non, ut Verrius putat, ferire.’ Therefore it is, or may be, no more than the opinion of one who lived in the fourth century A.D.,—i.e., whose native speech was late, not classical Latin. And that in Latin of about that time ‘mox’ could mean ‘soon’ is plain from the Vulgate of Philipp. 2, 23, ‘ hunc igitur spero me mittere mox, ut uidero quae circa me sunt,’ where the original is todrov pév otv érATrivw Téuat, ws av adidw ta mepi éeué, eEavTis. As might be expected, the glossators have similar definitions, whether their ultimate source is Festus or 1 Also in Columella, see Krebs-Schmalz, s.u, another. (CO 61, has ‘ mo Glossari This wa ing how in our oO through original] classical sense ‘Nn of the B come sd Wk and for | mox, OD republic among sequenc is thus, if we re (a) have me will sor mentior (5) consequ So also may Co! Poen. 5 evidenc (c) ‘eadem sapias, time an (4) quaerit immora respond Agamer daturus of the p (¢) ‘ea gen 1 iN an easy to Oe ie sie cn ’ which IT. 36, simply at any lo mox » ‘mox ‘ticular e have ut one of an iam in statim ists of linciae > mox ug. 19g itonis, er has ents), rse, is ndred mains lue to estus, mox, could ority ; is by errius e who ssical yn’ is mox, Tr if@ . the us or ‘MOX’ 5 another. ‘Confestim: statim, continuo: mox,’ says the Abstrusa glossary (CO 61, Lindsay-Thompson), and again (MO 28), ‘ Mox: statim.’ The Abolita has ‘ mox : statim, sine mora’ (MO 1), derived from a Vergilian gloss; the Glossarium Ansileubi copies this, MO 556, Lindsay-Mountford-Whatmough. This was quite enough to set modern dictionaries on the wrong track, consider- ing how prone most of their compilers are (we have seen honourable exceptions in our own age and country) to follow each other through a blunder like sheep through a gap in the hedge. But it does not fully explain how the mistake originally came about, nor how, in the development of the language from the classical to the later form, the change of meaning arose. For although the sense ‘next’ is allied to ‘soon,’ it is far from identical with it; every adherent of the Blank party in politics hopes that the next Dash government will not come soon, What seems to me the true reason alike for the post-classical development and for the lexicographers’ blunder is a very subtle shade of meaning in the word mox, one which explains at once why the word is rather rare in the best republican authors and why it is used at all, instead of dividing up its work among posthac, postea, and post. That is, that whereas these words express sequence in time only, mox very often indicates a causal connection as well, and is thus, as it were, intermediate between postea and propterea. This will appear if we re-examine some of our examples. (a) Cicero der. p. I. 20, ‘sed ista mox; nunc audiamus Philum.’ They have mentioned a subject only to drop it; it is not only a probability that they will sometime discuss it again, but a natural consequence of their having mentioned it. And so with other cases of that kind. (6) Plautus, M.G. 1168, ‘ne ille mox uereatur introire.’ He may in consequence of what he is told (as well as after hearing it) be afraid to enter. So also Rud. 111, ‘quon furatum mox uenias,’ cf. Trin. 864 ; the supposed thief may come to steal in consequence of his reconnoitring of the place. Again, Poen. 596, ‘quid dicamus mox pro testimonio,’ the trial and therefore the evidence will be a consequence of the trick played on Lycus. (c) This helps to explain one or two rather odd passages, as Merc. 802, ‘eadem licebit mox cenare rectius.. The preceding words are ‘iubeas, si sapias, haec intro auferrier.’ The good dinner will follow both in time and in consequence on the taking into the house of the supplies. (d) A very good instance is Horace, Carm. III. 6, 25, ‘mox iuniores quaerit adulteros,’ The wife’s disreputable conduct follows naturally upon her immoral girlhood, just described. Also Sat. II. 3, 192, ‘consulere et mox respondere ’; the answer will follow naturally upon the information given by Agamemnon when consulted. Ep. II. 2, 160, ‘segetes occat tibi mox frumenta daturus,’ the corn-dues could never be paid but for the harrowing which is part of the process of making the corn grow. (e) This is still to be found here and there in Silver usage: Suet. Aug. 2, ‘ea gens a Tarquinio Prisco rege inter minores gentes adlecta in senatum, mox 66 ‘MOX’ a Seruio Tullio in patricios traducta,’ etc. It was a natural consequence of Tarquin’s action that Servius should further honour a family already honoured. Now apart from the fact that nearness in logical sequence is easily con- founded with nearness in time, so subtle a shade of meaning as this is exactly the thing one might expect to be lost in the wear and tear of daily speech. When once mox was so worn down as to mean no more than postea, the next stage, in a healthy living language, such as Latin continued to be long after the last classical author was dead, would be to give it a new meaning—namely ‘soon,’ as distinguished from words like continuo, statim, cito, which all mene ‘at once.’ Finally, as the language degenerated and the Dark Ages drew near this distinction in turn was more or less completely lost. Mox, then, is the counterpart of olim, meaning ‘in the future’ as that means ‘in the past’; its radical significance of hurrying is applied, not to the rapid sequence of events, but to the action of the speaker in getting quickly on to the next point of his remarks, or in later usage, the next event in his story; and his reason for wishing to get quickly to it is, that the word frequently implies a causal connection, and one naturally wants to mention the effect soon after speaking of the cause. As to quam mox, it clearly means ‘how much in the future,’ hence ‘ how soon.’ H. J. Rose. NoTE.—On page 60, line 6 from the bottom, the new text of Terence by Kauer-Lindsay, which I have just received, reads sed nilo setius referetur mox huc puer ante ostium. If either reading with mox is what Terence wrote, we have here a normal contrast of the word with zam in the next line.—H. J. R. eee A P A s obtainec once at still exis at Bam Paris, t that (at regardec they eac Ra text} of yet exa Book 4c than th II. of the sixteent Book 3 of Book III and thi The ear was not century here be any er! importa A traditio of the « about 4 1 Befo1 and stops 2 For # 4 Krey: rest of de uence of onoured. sily con- S exactly y speech. the next ng after “namely, il meant ‘ew near, as that t to the ickly on t in his 1e word tion the y means ROSE. nce by uy Mmox ‘ote, we sé a erm maga LIVY’S FOURTH DECADE. A PRELIMINARY ENQUIRY INTO THE EVIDENCE OF MSS, A SUMMARY view of the main evidence at qur disposal may be soon obtained. Three traditions appear at the outset. The first depends on a MS. once at Mainz, and now no longer extant, but of which part, at any rate, still existed in the sixteenth century ; the second on an eleventh century MS. at Bamberg; and the third on a number of later MSS. in Rome, Florence, Paris, the British Museum, Oxford, Holkham, and other places. The fact that (at any rate for preliminary investigation) these three traditions must be regarded as separate may be seen first from the parts of the decade which they each omit. I. The later MSS., with whose tradition it is clear that the early printed text! of Livy (before 1518, ed. Mog.) was closely associated, all (so far as yet examined) lack Book 33 entirely, and none of them go beyond Book 40. 37. 3 (‘conciliabulisque edixerunt’), some of them stopping earlier than this.” II. The Moguntinus (henceforth called here Mg), as is clear from all three of the sources which throw light upon it, lacked (when examined in the sixteenth century) Books 31 and 32 and the first seventeen chapters of Book 33,2 but contained the rest of the decade from 33. 17. 6 to the end of Book 4o. Ill. The Bambergensis (henceforth called here B) contains Books 31, 32, and the whole of Book 33, and continues as far as 38. 46. 4 (‘ igitur inci’). The early chapters of Book 33 were published from it in Rome in 1616, but it was not properly examined before the work of Kreyssig in the early nineteenth century.* [A few fragments of an uncial MS. at Bamberg (fifth century) may here be mentioned.® Clearly among these scanty portions of the decade, where any error of B can be traced to the Bamberg uncial also, it is of great importance in the study of the origin of error. | A superficial glance shows us dependent upon each of these three traditions. On the word of B, and B only, we must rely for our knowledge of the early chapters of Book 33. On the second-hand evidence supplied us about Mg we must rely for our knowledge of the last chapters of Book 40. At 5 The fragments consist of the following por- tions: 33. 34. 9—33- 35- 5. 33- 36. 5—33- 37- 6; 34. 29. II—34. 29. 14, 34. 31. 19-34. 32. 2; 35. 5. 1O—35. 6. I, 35. 8. 4—35. 8. 9 5 39. 30. 4— 39. 36. 16, 39. 37. I—39. 37. 16. 1 Before ed. Mg, 1518, printed text omits 33 and stops at 40. 37. 3. 2 For Aganensis see below, p. g. 3 Mg began 33. 17. 6 ‘eis partibus,’ 4 Kreyssig, 1839 ed. of Book 33 (collated B for rest of decade). 68 S. K. JOHNSON first sight there is no such absolute dependence upon the late MS. tradition. But for tackling honestly the problem of the fourth decade, the evidence of these MSS. is indispensable for the following reasons : Firstly, B has many omissions, many of which are supplied by the early printed text and by these ‘ deteriores.’ That this (abundant) supply is due to interpolation is inconceivable. The character and the quantity of readings thus supplied do not allow such a supposition. Such omissions in B occur most frequently where no reading of Mg is recorded (i.e. both in the early part of the decade where, even in the sixteenth century, evidence from Mg was unavailable, and in numerous passages in the central part of the decade where no reading from Mg is given us). For this reason alone a correct estimation of the value of the ‘ deteriores’ and of their relation to each other is clearly requisite. Secondly, for the portion of the work from Book 38. 46. 4 to 40. 37. 3 our only existing MS. evidence is that of the ‘dett.’ Incomplete and often unreliable extracts from Mg are of course frequent within this portion of the work ; but in any case these extracts were only given where a discrepancy between Mg and the printed text was noticed by those who had access to Mg. It is therefore important to discover how close the ‘dett.’ are to the early printed text, and to find out how many varying traditions there may be among them. Thirdly, if all or some of the ‘ dett.’ are derived from an independent tradition or from more than one such tradition, it is obvious that we cannot dispense with their testimony. They are the more indispensable because the material for evidence about Mg is most uncertain. A short description of that material seems necessary here. In 1518! an edition was printed at Mainz. The text included for the first time the last part of Book 33 (from 17. 6 onwards) and the last part of Book 40 (from 37. 3 onwards). These portions of the work were apparently taken straight from the ‘ancient codex’ (Mg). For the rest of the decade the edition followed the text of the previous printed editions, but at the end of the work appear extracts from Mg made by Car- bachius (and Wolfgang) purporting to give the places in which that MS. differed from the traditional text. The copyists admit that their MS. was written in old characters and was hard to decipher. That they passed over many differences and misread many their successor Gelenius does not let us forget. But there is no reason to doubt that these extracts of Carbachius are genuine attempts at reproduction of his MS. It is highly probable, however, that he was not particular on questions of order, especially as in all passages in which he gives an order of words different from that of the received text he is always giving also a variation of reading. Moreover, after a careful study of his readings, I feel convinced that he sometimes omitted words in quoting a passage when these words occurred also in the text that he was comparing, 1 As regards Franciscus Modius (the third p. xlv) has shown the worthlessness of his inspector of Mg), Drakenborch (Vol. VII., work. and that to compl Som He corr number decade h in both s citing nc been prc not the « carried ' weighing Gelenius separatel’ Mor we canr accessib! to them absolute point is second |! and 32, after 40. notes OI 32. 10. I on Book singular a better the plu he neve Spirensi examine 1 E.g., i tion of J frumenti i eaeque pr Carbachiu ‘ frumeniti the fact t ‘ binae,’ a tion by v theory is either with gloss (made anno’ in | ‘eo’ for ' ‘anno.’ and one ir ‘adition. >vidence ne early Ss due to eadings B occur tly part Mg was e where imation clearly }O- 37> 3 d often 1 of the repancy }to Mg. le early may be sendent cannot use the of that Mainz. n 17. 6 yortions * (Mg). printed oy Car- differed written * many forget. renuine r, that assages text he tudy of juoting paring, ; of his LIVY’S FOURTH DECADE 69 and that he often gives the words that are new or changed, adding only enough to complete the sense or grammar of his quotation.! Some fifteen years later the scholar Gelenius had access to the same MS. He corrected some misreadings of Carbachius and added further notes, a number of which were clearly based on what he saw in Mg. Twice in the decade he mentions Mg by name; he often refers to his ‘archetype’ or ‘codex’ in both singular and plural, and often, too, he says merely ‘lege’ or ‘ legendum,’ citing no authority. That many of these latter passages are MS. readings has been proved almost conclusively by their correspondence with B. But this is not the case with all; and though Weissenborn’s suspicion had probably been carried too far, no study of the fourth decade can be conducted without weighing each individual statement of Gelenius and placing it in its context in Gelenius’ writing [i.e. it is most misleading to take each note of Gelenius separately (e.g. as quoted by Drakenborch) ]. Moreover, when we are convinced that a given reading is based on a MS., we cannot be sure that it is based on Mg. For Gelenius had two MSS. accessible for his notes on the fourth decade. In his note on 36. 22. 8 he refers to them by name as ‘ Moguntinus et Spirensis.’ There is no means of being absolutely sure in Books 34-39 which of these two he is quoting. But one point is clear and (before making a study of the ‘dett.’) most relevant. This second MS. (the ‘ Spirensian’) possessed at least the greater part of Books 31 and 32, it was without Book 33 entirely, and it ceased altogether somewhere after 40. 36. 5 and before 40. 39. 8. It cannot be only a coincidence that in his notes on Books 31 and 32 he refers to a ‘uetus codex,’ etc. [mentioning in 32. 10. 11 that his other codex (Mg) lacked these books], and that in his notes on Book 33 (of which the first part was unknown to him) he again uses the singular ‘ codex,’ etc., only, and is entirely concerned in correcting ed. Mg from a better reading of Mg itself, whereas from 34 to 40. 36. 5 he frequently uses the plural ‘ archetypa’ and ‘ codices,’ etc. Before Book 34 and after 40. 36. 5 he never uses this plural. It may be therefore provisionally stated that the Spirensian lacked the same portions of the decade as are lacked by the examined ‘deteriores’ (i.e. it lacked Book 33 and the end of Book 40).? 1 E.g., in 37. 2. 12 the text (before investiga- if we admit that Carbachius gave an incomplete tion of Mg) gave ‘binae eo anno decumae frumenti imperatae ’; Carbachius gives us ‘ binae eaeque proximae decumae imperatae.’ Surely Carbachius found ‘eo anno’ in Mg and also ‘frumenti’'? He was wishing to comment on the fact that ‘eaeque proximae’ occurred after ‘ binae,’ and hecompleted the sense of his quota- tion by writing ‘decumae imperatae.’ If this theory is not accepted, then we must suppose either with Drak. that ‘aeque proximae’ as a gloss (made by an arithmetical reader) replaced ‘ eo anno’ in Mg, or with Madvig that B+dett. had ‘eo’ for ‘aeque ac proximo,’ while Mg omitted ‘anno.’ This would imply one omission in Mg and one in B+ dett., which seems unlikely. But quotation, Madvig’s emendation is comparatively easy : (Madv.) BINEAEQU’AC PROXIMO ANNO was Cor- rupted into (say) BINEAEQUEPROXIMOANNO, which became (say) BINEAEQUEPROXIMEOANNO, which became (say) BINEEAEQUEPROXIMEEOANNO, which, I think, Carbachius thought Mg read ; cf. 35. 14. 7. 2 Since writing this, I have had access to Aduchs ‘de Gel. codice Spirensi.’ I find that the same point is made here and the relation between ‘S' and the ‘ dett’ demonstrated. 70 S. K. JOHNSON Madvig (Em. L., pp. 458 sqq.) says that in Books 31 and 32 the readings of this MS. (‘S’) prove to be close to those of B; but he has not pointed out— what seems a likely supposition—that from ‘S’ the majority at any rate of the ‘deteriores’ are ultimately derived. Such a supposition is purely hypothetical, But it is supported by several points. First, it has been noticed that ‘S$’ must have been closely related to B, but did not suffer from B’s faults of omission. Now, as I have hinted, it is partly in supplying B’s omissions that the late MSS. are valuable. Secondly, if we suppose that this lost MS. was the parent of the ‘ deteriores,’ it accounts for Gelenius’ apparent preference for Mg’ where the two were available. For ‘S’ may well have been close enough to the established text of printed editions for him not to give it (in Books 34- 39) as much attention as he gave to the more striking variants of Mg. Thirdly, there is no instance of his using the plural ‘ archetypa,’ etc., where the reading is conspicuously different in length from that of B and the ‘ deteriores,’ and in most cases of his plural use B and Mg are in substantial agreement. It appears then a tenable hypothesis that Gelenius’ ‘S’ and B had a parent or grandparent in common, and that a different fate befel the two. While B never completed copying his archetype, ‘S’ reached 40. 37. 3 (where, indeed, the archetype may have ended), and that while Book 33 had com- pletely disappeared from ‘S’ at an early stage, all the ‘ deteriores’ known to us were derived ultimately from ‘S.’ From this short survey of the work of Carbachius and Gelenius two further reasons for the importance of the ‘dett.’ appear: (i.) The uncertainty about Mg making research into further MSS. evidence more essential, and (ii.) the need to confirm or refute the above-mentioned hypothesis as to the common origin of the ‘ dett.’ in ‘S’ and the avuncular relation of B. Lastly, how many varying traditions are included among the ‘ dett.,’ and do any of them owe anything to the Moguntine tradition? For instance, if the work even of a corrector of some later MS. showed signs of being based on Mg it would be of obvious importance. These MSS. have not been systematically examined or classified by any editors. There were eleven ‘deteriores’ to which Drakenborch had access? (though he does not regularly quote them all in every instance). After the collation of B by Kreyssig editors have mainly ignored the ‘ deteriores,’ but have quoted from Drakenborch’s readings. The frequent approximation of the readings of the few known ‘dett.’ to the text of the early editions has made a systematic examination of these MSS. appear unnecessary. But if their assistance is to be obtained, as I have attempted to show it should be, some knowledge of which of the many we ought to turn to and how they are associated with each other will be essential. I selected, therefore, some 1 E.g., cf. the large number of readings in are conspicuously different, then the reading of which Gelenius’ statement of reading is suffi- Gel. is not far removed from B—e.g. 37. 23. 10 ciently approximate to that of Carbachius to (cf. elenchus). The hypothesis is thus confirmed. suggest that Gelenius was quoting from Mg 2 Cf, Drak., p. 327, Vol. VII. rather than from any other MS. Where they 180 pass far the t eighteen gestion | applied | examine guided i1 The full Tho in the p and Gel (where I these pa contrary readings After so may the passages, I. T a readin; Some of a large | for doub varies. | frequentl to lines | Frequent 36. 1 et plures the angul (The omi the word ‘metatus 36. : praesidio ablatives. In e signs tha: that thes that the ] ' See belc 2 Hencefc by Carbachiu. 2adings 1 out— » of the hetical. at ‘ S 9 ults of is that S. was nce for nough ks 34- hirdly, eading and in x 2 parent While where, | com- wn to work of the urther te the >’ and , and tance, based y any ccess? sr the , but on of s has ut if d be, y are some ling of 23. 10 firmed, LIVY’S FOURTH DECADE 71 180 passages to use as an ‘elenchus’ for testing the value of these MSS. So far the test has been applied by Miss N. M. Holley, of Newnham College, to eighteen MSS. in Rome, and to the Agenensis* by myself—both at the sug- gestion of Professor R. S. Conway. But it is not yet possible (before it is applied to more MSS.) to make any attempt in this article to classify the examined MSS. _I propose merely to give the main principles by which I was guided in making these selections, adding a few examples by way of illustration. The full ‘elenchus’ will be found at the end of the article. Though readings of these MSS. have been seen to be especially important in the part of the decade between 38. 46. 4 and 40. 37. 3 (where Carbachius and Gelenius are our only other authorities), and also in Books 31 and 32 (where B and ‘S’ are the only other authorities), this was no reason why these parts of the decade should be chosen for a preliminary test. On the contrary, it seemed best to let this first examination cover passages where the readings of both Mg and B were available (i.e. from Books 34 to 38. 46. 4). After some estimate is obtained of the value of the MSS. to be examined, it may then be possible to weigh the evidence of the best of them in those passages, where their evidence is even more important. PRINCIPLES OF SELECTION OF TEST PASSAGES. I. The first series of examples chosen were passages in which Mg2 provides a reading of which there is no trace either in B or in the early printed text. Some of these passages are suspected to be emendations or glosses in Mg, but a large proportion of them are of such a character as to leave little room for doubt that they are genuine readings. The length of these omissions varies. Frequently a line (probably of uncial script) had fallen out, and frequently the omission was clearly due to ‘corruptio ex homoeoteleuto’ or to lines and sentences beginning with the same words, syllable, or letters. Frequently more than one line had dropped from a similar cause. 36. 10. 12: ‘Ibi (M. Baebius) castra latius <(quam pro copiis et plures> quam quot satis in usum erant ignes cum accendisset ... Here the angular brackets indicate words omitted by B and supplied by Carbachius. (The omission of seventeen or eighteen or twenty letters and the reiteration of the word guam makes it most likely that a line had dropped. The omission of ‘metatus ’ may have been due to the us of ‘ latius.’) 36. 13. 17: ‘cum quingentis peditibus in praesidio erat.’ (Omission here assisted by numerals and by reiteration of ablatives.) In everyone of these cases* the ‘ deteriores’ so far examined gave no more signs than B of the Moguntine intrusion. Provisionally it may be concluded that these later MSS. are descended from some MS. closely akin to B, and that the parent whom they shared with B had the failing of too often dropping ' See below. 3 Cf. in elenchus : 34. 4. 3, 34. 16. 1, 34. 34. 2, 2 Henceforth ‘Mg’ will mean ‘Mg as quoted 35. 2. 7, 35. I. II, 35. 20. 4, 36. 10, I, 37. 16. 9, by Carbachius’ unless otherwise stated, 37+ 30. 7, 37- 41. 10, 38. 21. 6, 38. 25. I, etc. 72 S. K. JOHNSON words and lines from his archetype. The fact that these omissions are often the length of a line normal in uncial script is a further piece of evidence that the errors of omission shared by B and the ‘ deteriores’ are to be traced to a considerably early source. I have not found a single example of these longer omissions on the part of B whitch ts also supplied by the ‘ deteriores.’ Omissions due to B’s own carelessness are most common, but they are usually omissions of single words and syllables, or at the most of two or three words. II. A second series of examples was chosen from passages where B and Mg give two different readings and the early printed text gives a third. Obviously such passages are of importance for an estimation of the ‘ dett.’! In 36. 43. 1 Gelenius gives as ‘ antiqua lectio’ (possibly Mg, possibly ‘S.’ Carbachius says nothing) ‘ Liutum Delt per aliquot dies . . . (aduersi uenti tenuerunt),’ a sensible and no doubt correct reading. The early printed text gives ‘ Liuium deliberantem aliquot dies . . ., etc.; Band one of the examined ‘dett.’ supply ‘ Liuiwm deliberanti quod dies . . .. Of the remaining later MSS. examined two Roman codices agree with the early printed text,? and the rest read ‘ Liuium deliberantem quot dies.’ It seems probable that the parent shared by B and the archetype of the examined ‘ deteriores’ had already corrupted the reading to something like ‘deliberaliquod(t) dies’ (the p having been changed to 6 at an earlier date). 2B and the archetype of the ‘deteriores’ con- verted this into ‘ deliberantiquod (or -quot) dies.’ The reception given to this honest but un-Latin reading by the various later MSS. is important. One class does not understand the reading, but copies it faithfully. Another class does not like the dative participle, which appears to agree with ‘ Liuium,’ and changes it accordingly to ‘ dtliberantem,’ and a third gives ‘ deliberantem,’ and also attempts sense by changing ‘quod’ or ‘quot’ to ‘aliquot.’ [Such examples lead us to suppose that those MSS. which in this case appear (in reading ‘aliquot ’) nearest to the reading of Mg are the least reliable, since they owe more to intelligent emendation and less to faithful copying of errors.*] In 35. 16.6 we have an example where the intelligent correction can be traced more certainly to the parent of the examined ‘deteriores.’ ‘ Bello superatas a maioribus stipendiarias ac uectigales factas in antiqguum ius repetit.’* The italicized words are supplied by Gelenius, who introduces them with a ‘lege.’ The early printed editions gave for them ‘ Antiochus repetit.’ Whether Gelenius’ reading is taken from Mg (or from ‘S’ ?) or whether it is an emendation (less likely alternative), it is most probably what Livy wrote. This is partly confirmed by B, who gives us ‘Antiochum repetit,’ and who may himself have dropped the ‘ius’ as he did in 35. 7. 3. The parent of the ‘deteriores,’ either because his eye fell on the ‘us’ of ‘ius,’ or because he 1 Cf. 34. 8. 2, 34. 31. 18, 34. 44. 3, 35. 8. 7, these instances are of the type that occurred at a 35. 20. 3, 36. 7. 4, 36. 23.10, 37. 18. 4, 37. 20. 4, period when corruption according to sense was 38. 23. 6, 38. 38. 8, etc. common. 2 So, too, on Drak.’s authority, four of his * For phrase cf., with Drakenborch, 32. 33 ‘in MSS. antiqui formulam iuris restituere.’ 3 It may be worth noticing that the errors in wanted : ‘deterior In 3 gives us gloss or « (alone) i: Drakenb: MSS. giv archetype and that ingenious In 3 early prir Romanis Romanus some suc! corrupted the secon: parent of so forme misissent.? (In a evidence : thesis of ‘deteriore examined evidence i III. ] sisted of against th by L. Val according (though an of Draken and suitab Moguntint latter is s occurred a 1 Zingerle does not men 2 As for the appears likely gloss in Mg o singular ‘ Ror into the text of e often ce that ed toa longer 1issions lissions B and _ third. tt.’! ossibly st uentt ed text amined r MSS. he rest shared ‘rupted hanged ’ con- to this One r class n, and n,’ and amples eading ey owe can be ‘ Bello sbetit.’ 4 with a hether is an This O may of the use he rred ata nse was 2. 33 ‘in LIVY’S FOURTH DECADE 73 wanted a nominative subject, gave ‘ Antiochus repetit,’ and all the examined ‘deteriores’ follow him. In 38. 4. 9 the early editions read ‘cum ceteris ...,’ etc. Mg (Carbachius) gives us ‘cum cetera manu, a reading which does not seem likely to be due to a gloss or emendation, since none is needed. But when we find that ‘cum cetera’ (alone) is read by B, and by four of the late Roman MSS., and by three of Drakenborch’s MSS., and that the Agenensis and the rest of Drakenborch’s MSS. give the vulgar reading ‘ cum ceteris, we may provisionally infer that the archetype of B and the ‘deteriores’ parent gave ‘cum cetera’ (omitting ‘manu’), and that the ‘ deteriores’ may again be divided into the honest copiers and the ingenious emendators. In 34. II. 12, ‘mist praesidium Romant misissent’ is the reading of the early printed editions, and of the examined ‘ deteriores,’ B has ‘nist praesidium Romanis missts et; and, according to Carbachius, Mg read ‘mtst praesidio Romanus miles esset.’ Possibly here the parent of B and the ‘ deteriores’ had some such reading as ‘ nist praesidium Romants misisset’ (which had itself been corrupted from ‘ nist praesidium Romanus misisset’). B changed the position of the second s in ‘ misisset.’ But when that archetype was in the hands of the parent of the ‘ deteriores’ the s of Romanis dropped, and the plural Romani so formed obviously needed a plural verb, and therefore misisse¢ became misissent.” (In all the examples that I have selected in this second category, the evidence about the ‘ dett.’ so far obtained tends rather to confirm the hypo- thesis of the collateral relation between B and a MS. from which the ‘deteriores’ are all ultimately derived, and tends to show that none of the examined ‘dett.’ are derived from an entirely different tradition. But the evidence is not yet sufficiently complete for any certain statement on the point.) III. From the third category no lengthy citations are necessary. It con- sisted of passages in which B and the early printed text agree positively against the quoted reading of Mg.* In 34. 5. 2 Cato is called ‘ uir grauissimus ’ by L. Valerius according to B and the early printed text, ‘uir clarissimus’ according to Mg. The examined ‘dett.’ agree with the common reading (though an obvious corruption ‘ gratissimus’ occurs in one Roman MS. and two of Drakenborch’s). Probably the reading‘ grauissimus’ is the more forcible and suitable, and in any case the following point is worth remembering. The Moguntinus was probably an older text than the Bambergensis, but where the latter is supported by a consensus of the ‘dett.,’ it is likely that the reading occurred at least as early as in the archetype of B, and it is this archetype which 1 Zingerle cites B as reading ‘et,’ Kreyssig been introduced and a predicative dative seemed does not mention it. to be expected, or Carbachius himself may have 2 As for the reading of Carbachius from Mg,it misread what he saw. appears likely that miles had been written as a > Ch 26. % 32, 96 Fo Be She MD 2, 93 G G, gloss in Mg or a predecessor to account for the 35. 26.9, 35. 27. 2, 36. 14. 10, 36.17. 5, 36. 19, 2, singular ‘ Romanus’; ‘ praesidio’ may havecrept 37. I. 6, 37. 11. 6, 37. 48. 4, 37. 48. , 38. 45. 8, into the text of Mg, when (‘miles’) ‘misesset’ had etc. F 74 S. K. JOHNSON has to be set up against Mg. In most of these instances, disagreement between the ‘dett.’ and the early printed text appears to be very small. IV. I chose my fourth series of examples from passages where B omits something supplied by the early printed text. Here, of course (vid. sup.), the ‘dett.’ are of most obvious value. B’s omissions are most frequent. A certain proportion of these passages are accompanied by blank spaces in B, but a larger number are without such spaces.1_ Examples were chosen of both these varieties. [In very few? of such passages do the citers of Mg give any reading (since they only gave readings of Mg when it differed from the text they were correcting), but in some cases such a reading is quoted to complete the sense when attention is being drawn to a variant in another adjacent word.*] Among cases where B leaves no blank space, but makes important omissions supplied by the early reading, may be mentioned in 34. 44 ‘ uilla [publica] ’ and ‘ scelera [in carcerem]’ (both examples showing B’s distaste for low life!). In the majority of chosen instances of this class the examined ‘ deteriores ’ appear to vary but little from the early printed text tradition. Among cases where B leaves a blank space may be mentioned 37. 5. 6, where ‘eminisci’ is offered by all the ‘deteriores’ examined. ‘Comminisci’ is read by modern editors. If our hypothesis of B’s relation to the ‘ deteriores’ is correct, then does not the fact that B found the word difficult to read suggest that the parent of the ‘ deteriores ’ misread an abbreviation for ‘ com- **? V. Closely associated with this category are the examples in which there is a positive difference of reading between B and the early printed text, while our authorities say nothing of Mg.5 In 37. 60. 3, all the ‘ deteriores ’’ examined agree in reading ‘Italict . . . per totam insulam,’ while B reads ‘Italict... per totam Italiam,’ his eye having been attracted by ‘ Italici.’ VI. A sixth category was taken from a number of cases where B confirms wholly or partially the quoted reading from Mg, and the two are at variance with the early printed text. In 37. 4. 1 ‘hi omnes’ is read by B and Mg, ‘ duo homines’ by four of the examined ‘ dett.,’ and ‘duo omnes’ by one of them. The parent of the ‘dett.’ surely interpreted ‘ii’ (frequently confused with ‘hi’) as the figure ‘ii,’ and when ‘duo’ had thus crept into the text, those later MSS. that demanded some sense altered ‘duo’ to ‘ biduo,’ which was incorporated into the early printed text (‘ Homines’ and ‘omnes’ are of course frequently confused). In 35. 3. 3, quingentos fere passus’ is given by Car- syllable ‘con’ isin question. In the speech of Cato about repeal of Oppian Law B leaves a gap after ‘feminas’ (they were too much for him), All the examined ‘dett.’ give ‘ concitandas,’ 34. 2.7; Cf. also 37. 54. 28. 5 Cf. 34. 4. 18, 34. 19. II, 37. 24. 3, 38. 2. 10. 6 Cf. 34. 22. 2, 34. 27. 10, 34. 58. 5, 35. 7. 2, 35. 39. 5, 35- 44. I (‘aliis’), 36. 1. 1, 37. 10. 5, 37. 13.1, etc. The importance of this category for testing the soundness of Carbachius’ and Gelenius’ readings has already been suggested. 1 Cf. 34. 2. 2, 34. 13. 4, 34. 28. 8, 34. 41. 4, 36. 16. I, 36. 33. 5, 37- 17. 10, 37. 21. 4, 38. 8. 10, etc. 2 Cf., however, 35. 10.4, where Gelenius (‘anti- quior lectio’) reads ‘petebant,’ B omits verb, and ‘ tenebant’ occurred in some of early printed texts, while others (with all Roman dett.) read the same as Gelenius, and four of Drakenborcit's codices read ‘ pendebant.’ 3 E.g. 34. I. 6, 34- 35. 4, 37- 20. 4. 4 It may be worth noting that in another passage where B leaves a blank space the bachius, ‘fere ’). examine display | passus ’ No dout sign and * spectatc * perpauc Usir scattered Bamberg by Mg or of an earl No ¢ MSS. tov other can So far th hypothesi itself deri to vary cc text, and displayed. A not of the ear] decade it \ with the ss close thou; Roman cc 38.24. Ir most hone: it varies fri corruption between th If it is 1 In the m ‘deest numer of Valla. 2 E.g. 34. 4! idoneis’ Mg (: ‘aptis’); and dett. : ‘uanum B dett.: ‘opule 3 E.g. 34. 44 ‘sociorum’ J barous word ‘ ‘consociorum ’ between B omits sup.), the A certain B, but a oth these y reading hey were the sense Among supplied | ‘ scelera In the appear to es where is offered 1 editors. does not nt of the ich there xt, while examined ee confirms - variance Mg, ‘ duo of them. ised with xt, those vhich was of course n by Car- 1e speech of leaves a gap ch for him). -oncitandas,’ 3, 38. 2. 10. 3. 5y 35+ 7+ 2; I, 37. 10. 5, his category yachius’ and | suggested. ‘consociorum’ B; and 35. 14. 7 (where Carba- LIVY’S FOURTH DECADE 75 bachius, and his reading is confirmed by B (who, however, gives ‘ferme’ for ‘fere’). The early printed text reads ‘ferme mille passus.’ Three of the examined late MSS. read ‘ ferme passus,’ omittingia numeral,! and the others display a less honest tradition, giving ‘ferme passus mille’ or ‘ferme mille passus ’ (ignoring, of course, the fact that Livy would say ‘ mille passuum’). No doubt the parent of the examined ‘dett.’ misunderstood the numerical sign and omitted it. In 35. 36. 3 B confirms Carbachius’ reading from Mg of ‘spectatorum,’ while the examined ‘dett.’ read, with the early printed text ‘perpaucorum ’ (a reading presumably due to an incorporated gloss). Using these six categories for selection, and taking a representatively scattered number of readings from each (including passages covered by the Bamberg Uncial fragments), I also included instances where a reading given by Mg or B was most probably due to a gloss,” and instances where a reading of an early printed text appears at first sight closer to Mg than to B.® No general statement can yet be made concerning the evidence of the MSS. to which this test has at present beenapplied. Fortheir relation to each other cannot be surely determined until many more ‘ dett.’ have been examined. So far there has been no evidence to refute and not a little to confirm the hypothesis that they are all ultimately derived from a single MS. which was itself derived from a parent or grandparent of B. But they have been found to vary considerably in their approximation to the established early-printed text, and in the stage of corruption reached and the honesty and ingenuity displayed. A note may be added on the subject of the Agenensis.4 Though in parts of the earlier decades it was written in a thirteenth-century hand, in the fourth decade it was entirely written by a fourteenth-century scribe. After testing it with the selected passages, I found that its likeness to the early printed text is close though by no means invariable. Two points separate this MS. from the Roman codices which have been examined. (1) The Agenensis ends at 38. 24. 11 (earlier even than B).° (2) It is not to be classed either with the most honest or the most ingenious of the Roman MSS. Most commonly when it varies from the early printed reading, it is to point to the way by which the corruption has come, giving some readings that may be described as ‘ halfway’ between the printed text and B.® If it is possible to test the remaining late MSS. from the passages printed 1 In the margin of the Agenensis is written ‘deest numerus’ by a hand that may be that of Valla. 2 E.g. 34. 46, 4, ‘locis apertis’ B dett.: ‘*locis idoneis’ Mg (where one of Drak.’s MSS. gives ‘aptis’); and 35. 47. 2 (‘in naue’ B: ‘inane’ dett,.: ‘uanum’ Mg) and 37. 32. 13 (‘ potentiora ’ B dett.: ‘opulentiora’ Mg. 3 E.g. 34. 44. 8, ‘consciorum’ majority of dett. : ‘sociorum’ Mg (probably correction of bar- barous word ‘consociorum’) and one late MS. : chius probably saw ‘ultimas’ in Mg, but did not quote it) ; cf. also 37. 7. 15, 37. 54. 28 (vid. sup.). ‘ British Museum, Har/. 2493, described by Professor C. F, Walters in Class. Rev. XVIIL., Pp. 392. 5 An attempt to connect this ending with that of Gelenius’ ‘S’ proved fruitless, 6 E.g, 36. 4.2, AB ‘etiam’: X ‘et’; 36. 3. 7, ‘nuntiaret’ B: ‘nuntiare’ X : ‘ nuntiare. et’ 4; 36. 43. 1 (cf. above) ; 37. 13.1, ‘ praeteruehémtis ’ AB: ‘ praeteruehentes’ X ; 38. 23. 6, ‘ late roma- nis’ B: ‘later ois’ A: ‘ late inter omnis’ X, 76 S. K. JOHNSON below, then some idea of their reliability may be formed ; and after a collation of some of the better MSS., it is to be hoped that a ‘stemma’ of these may be rendered possible, and that our equipment will be improved for the study of the Fourth Decade. The preliminary work of testing the many MSS. will be facilitated by the Elenchus, which is printed below in the briefest possible form. Abbreviations are used occasionally owing to lack of space, but only where there can be no ambiguity after reference to any text of Livy. The following ‘sigla’ are employed: B=Bambergensis (on the evidence of Kreyssig); Mg = citation from Moguntinus by Carbachius; Gel. =reading given by Gelenius, probably coming from Mg or S; X=reading of early printed texts or reading of those ‘ dett.’ that have yet been examined (for the sake of conciseness, these MSS. are not named, nor are their variants all given). Except in the case of B, all readings have been taken direct from the earliest source. THE TEST PASSAGES. 34: : 34: : 1. 6 (conciliabulisque?) conueniebant Mg:c— 37. 3 affinxisset BMg: adiunxisset X. isconuenerant BX:c—isquec—rantX, 40. 5 oOppugnarent tyrannum Mg: oppugnatur 2. 2 sustinere X : om. B (post singulas). tyrannus BX. 2. 7 ad f. concitandas X: ad f. (spatiorel.) B: 41. 4 quantum X: om, B. f.adc. X. 43. 9 censuisset Mg: uoluisset BX. 2. 11 quoque BX: prope Mg. 44. 3 pridie Calendas Mg: id. Maias (sic) B: 2. 14 feceritis Mg : faciatis X: facietis BX. idus Maias X. 4. 3 quo magis imperium Mg: imperiumque 44. 5§ publicaab X: om. B. 44. 6 incarcerem X: om. B. 4.18 exorabilis es B : exorabiles (-is) X. 44. 8 consciorum X: sociorum Mg: consocio- 5. 2 grauissimus BX: clarissimus Mg. rum B. 7. 2 hic—uicorum MgGel : om. BX. 46. 2 and 3 bigati—oscensis Mg: om. BX. 7, 2 permittemus Mg: est BX. 46. 4 apertis BX: idoneis Mg. 8. 2 collegarum rogationi B: rogationi col- 49. 8 effrenatam et pr. Mg: effr. X: pr. et legae Mg: collegarum orationi X. effr. B. 9. 5 semper (aliquis ?) Mg: aliquis BX (prae- 58. 5 eadem BMg: om. X. bent X uel ante assiduus: om, B). 58. 12 Graecas Asiae urbes BX: Asiae urbes II. 2 praesidium Romani misissent X : prae- quae Graii nominis sint Mg. sidium Romanis missis (et?) B: prae- 35.: sidio Romanus miles esset Mg. z. 6 aut praetores Mg: om. BX (scriberetur 13. 4 satis X : om, B. X: scriberent BX). 14. 8 uerteritque in hostem BX: et aduersos 2. 7 scribi—milites Mg: om. BX. in h. uerterit Mg. 2. 7 eligerentur Mg: legerentur BX. 15. 5 hostium oppugnabantur Mg: oppugna- 3. 3 quingentos ferme passus BMg (hic fere) : bantur BX : oppugnantur X. ferme passus X : ferme mille passus X. 16, 1-2 tria—iussit Mg: om. BX, 6. 6 niis BX: uelle ut is Mg. 16. 3 suos X: milites Mg: suis B. 6. 9 fuerit elapsum BX: fuerat non deletum 16. 5 uinoque—curatos Mg (qui remisit: dimi- Mg. sit Gel): om. BX. 7. 2 ita libero foenore BMg (hic liberos ¢t 18. 4 perse BMg: om. X. obruebant) : in tali uero f. X (uel itali). 19. Ir regressus X : progressus B, 7. 3 ius MgX: om. B, diceretur Mg: redde- 22. 1 manu BX: exercitu Mg. retur BX. 22. 2 omisso bello MgGe/l: missoB: omnes X. 8. 1x gerebantur Mg: agerentur BX. 27. 10 suspensis BMg: suspectis X. 8. 7 arguere B: argueretque X: aut argue- 28.11 infida ac Gel: om. BX. ret Mg. 31. 18 fortunae ac dignitatis fore MgGel (ad pro 10. 4 petebant XGel: om. B: pendebant X. ac Mg): foreadd. fore B : foreadd. X. 11. 6 hostium Mg: om. BX. 34. 2 illud—fallat Mg: om, BX. 11. 10 hostium X : om, B. 35. 4 educeretur MgX (sed hi post priuatum, 11.11 propinqua—omnia Mg: om. BX. ille ante sine): om, B. g Romanae amicitiae X : Romae a—as B. 35-: 14. 17. 18, 20. 39» ? 44. 44- 40. 47+ 48. 49. 50. 51. 36. : 7 4 6 in 4 le 2 sil Q Spi I cu 2 pe I ad 4 ele 3 et 8 uel 5 ue 9 reg I alii I rex 6 (et ( 2 uar 7 alte Io tut 12 pot 2 €OS i 9 adp 1 Cn, 7 an: I t 2 etia 4 ut « C - 13 qua q I prof 8 nun U . II met . Ir qua 6 Stra 7 et e . Io renc . 13 cons e€ collation - may be ly of the will be »le form. ere there g ‘sigla’ on from 7 coming e ‘ dett.’ . are not readings X. oppugnatur ias (sic) B: : CONSOCIO- . BA. X: pr. et Asiae urbes 8s (scriberetur RX. g (hic fere) : le passus X. ion deletum c liberos ét X (uel itali). Mg: redde- X. aut argue- lebant X. BX, ae a—as B. LIVY’S FOURTH ot 7 quas uisere Mg: om. BX, esset Mg: om. BX (qui ultimas post quod praebent, om. Mg ?). 16. 6 in antiquum ius Gel (e Mg ?): antiochum B: antiochus X. 17. 7 a paruis B: per ius X. 18 6 ferarum Mg (qui eum fost iras): ferum ferarum BX. 20. 3 esset ducere Mg: esse duceret X: du- ceret B, 20. 4 legiones—ducis Mg: om, BX (qui et scri- beret om.), 26. 2 simulacris naualis Mg: simul uel Achaeis nauibus B: simul cum Ach. nauibus X, 26. 9 speculatoria Mg : piscatoria BX. 27. I cum prospere latus Mg: om, BX, tum etiam BMg: tum spem X. 27. 2 periculi timens BX : periculo fore Mg. 28. 1 ad id maxime animum MgGel: se ad id max. BX, 32. 4 elephantos accitos Mg: elephantos X : elephantes B. 36 3 et spectatorum Mg: sp. B: perpauco- rum X, 38. 8 uenirent ante ad Mg, post oppugnandos BX, 39. 5 uenisse se B: sese uenisse Mg : uenire X. 2. g regem conspici regem Mg: regem BX. 44. I aliis introductus facto silentio Mg: aliis in silentio facto B : inde silentio facto X, 44. 1 rex dicere orsus Mg : dicere orsus rex BX. 46. 6 (et ?) re liberandae Mg: liberandare B : et liberandae re X (sed om. aliq. et). 47. 2 uanum Mg: in naue B: inane X. 48. 7 altero ciuitatis altero Mg: alt. unius ciu, alt. B : altero altero rege X. 48.10 tutissima Mg: iustissima BX. 49. 12 potius Mg: tocius B: toties X. 50. 2 eosdem genti Achaeorum hostes Mg: gentes fidos BX. 51. 9 adportari B: comportari X. 36. : 1. 1 Cn. filium Mg: Cn. B: om. X, 3. 7 am satis esset ad praesidium aliquod eius nuntiaret B: an ad praesidium nun- tiaretur Mg: aliquod eius nuntiare X. 4. 2 etiam B: om. Mg?: et X. 7. 4 ut quibus Mg: utrisque B: cum utris- gue X, 9. 13 quattuor milia X: mille trecentos Mg: quattuor B. Io. 1 profectus Mg: om. BX. 10. 8 nunc uires X: nunc cuius res B: tunc uires Mg. Io. 11 metatus MgX : om, BX. 10, II quam pro copiis et plures Mg: om. BX. 11, 6 Stratum X: magistratum BX. 13. 7 et equitibus quadraginta Mg: om. BX. 14. 10 renouato Mg: reparato BX, 14. 13 consul primo misit Mg: primo missi (uel emissi BX), 36. : 15. 16, 17. 17. 17, 37. 17. 18, 19g. 22, 23. 24, 24. 27. 31. 33. 35+ 36. 42. 42. 43- 37: : Z. NS DAs Io. 10, II, II. 12. 12. 13. 16. 16, 16. 17. 17. 18, 20. DECADE a= 7 et achaei Mg: om BX. I portas X: om, B. 5 ferocissimae BX: bellicosissimae Mg. 5 seruitutis digna Mg: seruituti nata BX. 7 omittam Mg: sileam BX, 8 ingratissima BX : leuissima Mg. 9 uenerunt frequentes BX : frequentes con- uenerunt Mg. 7 suis ad pugnandum B: ad p. suis Mg: ad p. X. 2 in castris fuerant regiis Mg: in c. regis erant BX. 7 arcem extra muros quae Mg: om. BX, Io ex stationibus Mg: ex statione X : expec- tatione B. 5 in qua aedificia extra urbem erant Mg: qua progressi erant BX. 9g descensurus B: ascensurus AX: exces- surus (ante consul) Mg. 5 iura Mg: om. BX. 9 aut in posterum Mg: in p. aut BX. 5 regi X: om. B, 7 Eleis MgX: aetolisBX. 7 eo Mg: om. BX. 4 maturandum ratusomnium Mg: om. BX. 5 samum Mg: tamen B: tum X. 1 Deli per aliquot Gel: deliberanti quod B: deliberantem aliquot X. 6 infecta BX : incerta Mg. I2 eaeque proximae Mg ?: eo anno BX. rt ut hi omnes BMg: ut duo homines X: ut biduo omnes X, 5 uiginti triremes Mg: om. BX. 2 dato X: om. B. 6 machinas X: machinationes B. I metuebant uenire (fost saeuiretur) BX: m. redire (ante ne) Mg. . 15 prosecutus MgX : ipsos secutus B. 5 regiam BX : aperit regiam Mg. 5 aut maiorem eius B: aut maiorem Mg: om. X. 6 parte Mg: classe BX. 13 duabus Cois Gel: cois duabus Mg: dua- bus copiis BX: duabus Cypriis X. 4 inflatus BX: elatus Mg. 7 fuerat X: om. B. I praeteruehentis classis si quas Gel (sed Carb. nil nisi si quas citat): praeter- uehentes quas X: praeteruehémtis classis quas B. 9 magis quam conserebatur Mg : om, BX. II remigum turbam Mg: naualiumr. t. B: nauium r. t, X. 14 in Italiam X: om B. 5 perire sinerent Mg: perisse liceret BX (perire X). 10 non posse X: om, B. 4 excursionibus Mg: excursione B : excur- satione X. 4 postquam Mg: quicquam B: ne quic- quam X. 78 37-: 21. 23. 24. 24. 28. 30. 31. LIVY’S FOURTH 38. : 3- 4 uenisse X: om. B. Io nec sic in altum Ge/, : Eudamus in altum Mg: Eudamusinalto X; necin altoB. 3 inclinabat X: declinabat B. 7 trepidationem Mg: expugnationem BX. 1 oculis Mg: agris BX, 7 et—uentus Mg: ephesum BX, 9 ubi Gel. : urbs BX: urbi X. 32. 13 potentiora BX : opulentiora Mg. 34. 34. 35. 39+ 4I. 48. 48. 50. 54- 54. 6o. 38. : I. 2. 2. 6 in eo delapsum tumultu ex Mg: delap- sum B: in eo t. delapso X. 7 illud satis X: om. B. 3 quod su. p. fuerint BX : quaes. p. f. X: quas s. p. ediderant Mg. Ir uerebatur obiectis BX: uidebatur egere Mg. Io partim—coniectorum Mg: om. BX. 4 imperata Mg: imperia BX. 6 subtexit BX: adiecit Mg. 6 ueniret BX : rediret Mg. 8 in seruitutem BMg : om. X (ei ante in Mg, post traderetis BX). 28 comprobaritis X : comprobaueritis Mg: probabitis B. 3 insulam X: italiam B. 4 ab eo nuntiant principibus BX: nuntii ab eo ad principes Mg. 10 flumen X: om. B. (traiecerunt ? B: traicerent X). 14 postea BX : postera die Mg. ON 23. 25. 28. 29. 33: 38. 45: DECADE 4 Aetolorum BX: in Aetolorum imperio Mg: Aetolorum inde X. 9 cetera manu Mg: cetera B: ceteris X, 7 receptu BX : regressu Mg. . 10 argenti X : om. B. 2 ut castellum quod Mg: quod castellum BX. 3 et refertum—copia Mg: om. BX. 1 Galli BX : hi Galli Mg. . 15 adse non X : ad senatum B., . 17 omnis illa BX: ea Mg. 5 inuiis X : inuisa X : inuises B, 6 temere Mg: om. BX. . 14 instare Mg: stare BX. 2 praedaque Mg: praedaq’b B: praeda- quibusque X. 5 quantum equis Mg: quidem quantum equi in colles B: quantum equi in colles BX, 6 per omnes Mg: romanis B: inter om- nes X, 1 ab Ancyra castra Mg: om. BX. 4 aquae BX: pluuiae Mg. 5 certiore BX : certiusque Mg. 7 tumultu concitato turba B: turba tu- multu concitato Mg: tum. concitata turba X. 8 actuarias Mg: om. BX, 8 monerem ex Geil, (minorem ex Mg): mi- nore ea B : minorem ea X, 8 assumpseras BX : desumpseras Mg. S. K. JOHNSON. THE from tha but arise from 7r0é KWNAKPE zur gr. | Gramm. « between | In F which is The cuvaryeipe letters w! of letters the line. above th hitherto which it palayper result of e into o. The far from divides th parts anc times, or m imperio ‘eteris Y, castellum X. : praeda- quantum n equi in inter om- turba tu- concitata Mg): mi- Mg. ISON. VARIA. VITOTUTTTO. THE meaning of brotvmra, to plunge or dip into a liquid, is so different from that of rumrw, as to suggest that the word does not really contain rurro, but arises from tzro-d0mtw (dvmTw = dive, plunge), in the same way as Tora7ros from mrodamrés (Schweizer, Gramm. der Pergamenischen Inschriften, p. 107, note 1), xcwraxpetas from xwraypérat, Ilerayeitvios from Iledayeitvios (Solmsen, Bettr. zur gr. Wortforschung, p. 234), Sexatapxos from dexddapyos (cf. Thackeray, Gramm. of Old Testament in Greek, p. 103), in all of which a voiced stop situated between two unvoiced stops becomes unvoiced. cuvayaypeTat. In P. Oxy. 1788, Fr. 15. ii. 18 (Plate II.) a papyrus of Alcaeus has— cotper|. |apy|. . |. wepBéBal . |acxpol Gesal [ . Jatcaprrococe| . |onoovvarpet[ which is thought to mean— gol pev yap 45n uTepBéBatat ypovos Kal xaptros Gaocos 7s cuvayéppeTat. The sense seems to need the perfect rather than the present passive of cuvayeipw (-ayéppw). The hole in the papyrus has removed the tops of the letters which are conjectured to have been azp, as well as an unknown number of letters (probably three or any smaller number) which were written above the line. The scribe perhaps wrote cvvayperac and then added another ay above the line, meaning cuvayayperar. This I take to be ovvaydyperas, a hitherto unknown Lesbian form, synonymous with Att. cuvarynyepta, to which it seems to be related as the second element of immaypérns and palaypérns to that of ottayéptns. ocuvayaypota:, which would be the Lesbian result of cuvayayrraz (supposing this to have existed), would involve altering € into o. ° @S EXACTOL. The translation pro se quisque given by Liddell and Scott is at any rate not far from the truth. A. G. Laird, AJP 27 (1906), p. 33, says that ‘@s éxacrot divides the noun to which it refers (generally the subject of the sentence) into parts and implies that the different parts act in different ways, at different times, or in different places; or otherwise differently if the point of difference 80 VARIA is implied by the context.’ This is fully borne out by the passages discussed by him. This ws (I am not sure that the suggestion is entirely my own) is the adverb of és (Cret. Fos, I.-Eur. svos) the possessive reflexive adjective, and meant, when the phrase was first coined, in his (or their) own way. Being an ordinary adverb and not a conjunction or relative adverb, it is normally not followed by a subordinate verb, but belongs to the principal verb, and the usual assumption of an ellipse of the verb is unnecessary ; see, e.g., Thuc. 5. 57. It was, however, only natural that this as (Fws) should be confused in the Greek mind with the word for as. The existence of this confusion is proved by the cases (they are a minority) in which it is followed by a subordinate verb, e.g. ws ExacrTos amuxvéotro, Herodotus I. 29. The position of ws before éxaaros is idiomatic, as in Lat. suum quisque and Leg. Gort. 1. 18: g@wviovres Fov Fexdrepos juny, each declaring that (the slave) ts his own. EMTTETT ANY LEVNV. At Soph. O.T. 1264 a number of MSS., including the Laurentian, offer— TEKTALS EWPals EuTreTANY LEVY, which appears to be correct, éumremAnypévnv (entangled) being the perfect parti- ciple passive of éumaXaccopuat, formed like rérpnya the perfect of tapacow or Opattw. This mardocopar is found in wepimadraocecOa, which is used of atoms which hook on to one another, Diels Vorsokr. II.* 44. 19 (Theophr. Sens. 66), cf. mepemrdrakis 1b. 1. 16: both words are corrupted in the MSS., cf. Diels, Hermes 40, 306-310. It can hardly be related to the Homeric taXtdocw (-ouat) in either of its senses (bespatier and cast lots), or to mAéxew. On the other hand it has been noticed (v. Steph. Thes., Leaf on Il. 12. 72) that évrAnEwpev, -wor, -avtes in Homer mean ‘to be entangled’; are they in reality forms of éuTarkaccouat ? R. McKENZIE. CHA edition b (the num Professo forschung the celeb: But why bination Festus 11 Roman < I giv reader n dpovrida in this p abbrevia: Cf. Fest ‘contend quote a some suc cannot { (not sé). Festus 3: discussed own) is adjective, own way. erb, it is ipal verb, see, €.2., ed in the 'S proved ordinate nN quisque that (the offer— ‘ct parti- \aoow Or used of ir. Sens. f. Diels, » (-ouat) er hand EV, -WOl, rms of NZIE. ENNIVS, ANN. 508. CHARISIVS, in his chapter on Adverbs, cites for Hispane (p. 260 of the new edition by Barwick; p. 200 of the old edition by Keil) a line of Ennius’ Annals (the number of the book is not stated ; see Barwick’s note) : Hispane, non Romane memoretis loqui me. Professor Norden (Ennius und Vergil, p. 114), the apostle of Combinations- forschung, combines this fact with another fact mentioned by Livy (XXI. 19, 6), the celebre per Hispaniam responsum of a Spanish town to a Roman embassy. But why does he ignore a third fact which must be brought into any com- bination that can be convincing—the quotation of this line of Ennius by Festus in a paragraph (of the R-chapter) which apparently deals with Graeco- Roman affinities (362, 10) ? I give my attempt at a restoration of the paragraph in the hope that some reader may suggest a better, or even that Professor Norden, if he has devrépas ppovridas auelvovas, may communicate them to this journal (the lines of Festus in this part of our unique MS. average thirty-three letters and employ some abbreviation-symbols) : <(Romanos priscos Graeco>s appellat Enni cos, Graixcos> memo ngua longos per ‘ Hispane, non Ro <‘mane memoretis loqui me.’ Gra>ecum Romulus ur yin[a]Je locutus sit, etc. Cf. Fest. 386, 32: ‘Sos pro eos antiqui dicebant, ut Ennius.. .’ Lib. XI. ‘contendunt Graecos, Graios memorare solent sos.’ It is Festus’ habit to quote a complete line rather than a complete sentence. We must suppose some such words as historict Romanos fuisse to have preceded, since contendunt cannot play the part of imitantur, and sos is expressly stated to mean ¢os (not se). For details of the ancient theories of the name Roma (‘Pwpn) see Festus 326, 28-330, 18 (e.g. 328, 13). W. M. Linpsay. THE AUTHORSHIP OF THE A/PPIAS MATOR. In his article in a recent issue of the Classical Quarterly! Mr, G. M. A. Grube attempts to vindicate the authenticity of the Hippras Mator, concentrating his attention upon an article? by me as ‘embodying most of the objections of earlier scholars’ against the genuineness of the dialogue. I believe that I am still ‘the latest exponent,’ in any detail, of this view; but I may claim at least some good company in the heresy, and to the names already cited by Mr. Grube or myself may be added in particular that of Wilamowitz,? who also believes the Hippias Maior to have been written in the Academy within Plato’s lifetime. I was not aware when writing my previous article of this noteworthy support in a work just then appearing. Continued and detailed study of the Hippias Maior, while it has naturally led me to correct or modify some of the opinions stated in my article, has confirmed me in the general view that the dialogue is not Plato’s work, but was written during his lifetime, and by one in close touch with his literary and philosophic activities ; and, further, that it bears some relation to the development of Plato’s own thought. Obviously the case on either side is likely to remain strictly not proven, however strong personal conviction may be. The close similarity, in some respects, between this dialogue and undisputed work of Plato is admitted by all the heretics, and constitutes alike the chief difficulty and the chief interest of the problem. At the same time, I still believe (I trust not merely as Geow dvapvAdrrovew) that it is possible to demonstrate, by the usual canons of comparison, its unplatonic qualities both of language and of thought. It will be convenient, in replying to Mr. Grube’s arguments, to follow the order of his article : 1. Evidence in Aristotle—The reference (Met. 1025a 6) to a passage in Hippias Minor as 6 év tw ‘Irrig. Adyos is not proof that Aristotle knew only one dialogue of the name, But neither is the similarity of Topica 146a 21 sqq. with Hippias Maior 300-303 proof that Aristotle had read Hippias Mator ; nor again, if he had, would it follow that the dialogue was the work of Plato. Aristotle’s use of the disjunctive 7 in first stating the formula (ofov 7d xaddv 7d de’ dYews 7d Se axons 750) goes directly against the argument in Hippias Mator 299c sqq., where the application to ‘both’ as well as to ‘each’ is emphasized. It seems probable that, as Zeller* suggests, the formula may have been known to Aristotle as propounded by some sophists, independently of its appearance in Hippias Maior, or else as current in the Academy (cf. the discussion of true pleasures in Phtlebus 51B sqq.), and therefore equally familiar to Aristotle and to the author of Hippias Maior. Mr. Grube himself legitimately points out, in a later connection, that the similarity between Protagovas 347A and Hippias Maior 286a need not prove copying in the latter dialogue. The fact, which Mr. Grube says I ‘admit,’ that the Hippras Mator ‘ can hardly have been written by an intelligent person who had read the Philebus,’ is of course integral to the view I take of its authorship; and this point will recur later. I believe that the dialogue was written by an intelligent person; I believe that he had mung platonische Art, noch bei Lebzeiten des 2 Journal of Philology XXXV. (1920). Meisters in der Akademie entstanden,’ 3 Platon (1919) II. 326 sqq.; p. 328: ‘ Der 4 Plato and the Older Academy (Eng. trans.), Dialog diirfte ziemlich alt sein, keine Falschung, _p. 71, n. 61. sondem eine unfreie und unzulangliche Nachah- 1 July-October, 1926, pp. 134-138. not read he was P The be made « 2. In individua. to its ger compares Laches, Ci no one el general t< character they rese close verl is questio use of pl Phaedo. considera habit of v the simp]. Wit! the writer have here The spee may be ir have mad Style manner,’ ; is slightly The mot interestin precise li fall under distinct fr A. Non > WwW ND HH “I In al character » 9 Grube tention holars’ latest mpany : added e been ng my led me | me in ng his ; and, ought. ywever tween ‘Ss, and At the ossible oth of - order lippias of the Mawr would f the To Oe where ybable tle as 10¥, OF iS 51B VU aior. larity n the ardly ourse ar. I e had 2n des rans.), D. TARRANT 83 not read the Philebus, because it was not yet written; but I do not believe that he was Plato. The evidence from Aristotle is at the most very slight, and not much can be made of it on either side. 2. Internal Evidence.—As a literary work the Hippias Mator shows more individuality and interest than any other disputed or rejected dialogue. With regard to its general scope and plan, it may be remarked that, while in length the dialogue compares most nearly with some early ‘Socratic’ works of Plato (Hippras Mator, Laches, Charmides, are each 23 pp., Lysis 20), and in its form of simple duologue (with no one else present) only with Euthyphro and Crito (both still shorter), in style and general tone its affinity is much more with the longer works of a later period. The characterization, the use of poetic words and general exuberance of tone, so far as they resemble Plato, resemble the Plato of Euthydemus, Republic, or Phaedo. The close verbal parallels are mainly with Republic and Phaedo. Purely stylistic argument is questionable ; but it may be noted that H. von Arnim,! investigating specially the use of phrases of assent, finds a close similarity of style between Hippias Mator and Phaedo. If the dialogue is Plato’s own work, we have to suppose that at a stage of considerable maturity of literary and dramatic power, and when he had formed the habit of writing at considerable length, he here reverted suddenly to the brevity and the simplicity of setting characteristic of his earliest works. With regard to the figure of Hippias, it is in the main a clever caricature, But the writer’s zeal in some places overreaches itself ; and my impression still is that we have here not Plato, but a close imitator with less subtlety and delicacy of touch The speeches of Hippias at 284c, 2858, 295a, and those of Socrates at 292A, 292C-D, may be instanced as cruder in satire or in bluntness than Plato’s own hand would have made them. But these are, of course, ultimately matters of opinion. Style-—(1) The imaginary speaker. Mr. Grube says this ‘is quite in Plato's manner,’ and cites a number of instances in support; he says that ‘everyone of them is slightly different from all the rest, and no less so than the speaker in our dialogue.’ The motive for every use of this kind of device is clearly the same; it is an interesting and dramatic way of introducing ancther opinion. But as regards the precise literary form, it will be found that (i.) the examples given by Mr. Grube fall under a limited number of headings ; (ii.) the Hippias Maior instance is of a type distinct from any of them. I keep Mr. Grube’s numbering : A. 1. Theaet. 165a-168c. Protagoras imagined as speaking. . Symp. 201D sqq. Diotima described as having taught Socrates. . Euthyphro 5a, B. Meletus imagined as intervening. . Laws X. 8938. An imaginary atheist personifying the drarnAds Aoyos. . Crito 50a sqq. The Laws personified and imagined as upbraiding Socrates. . Republic V. 4794 sqq. The man who denies absolute Beauty (éxeivos o ptAobedywv xtX.) imagined as answering questions. . (2). Gorgias 452A. The various Sypsovpyot imagined as defending their claims. In all these examples, so far, the interlocutor is either named or definitely characterized as a personification or type. B. 7 (1). Gorgias 451a. An indefinite person (r:s) imagined as putting questions. Gorgias 506c, p. Again the imagined disputant would be t:s—or possibly Callicles, in which case this goes under group A. Add Gorgias 455D, and numerous other instances, where an indefinite and imaginary questioner appears. Nn —& WwW ND “SI 1 Sprachliche Forschungen zur Chronologie der platonischen Dialoge. 84 D. TARRANT Note that in all these examples but one the situation is frankly hypothetical. ‘Suppose Protagoras .. .,’ or ‘ Suppose somebody .. .,’ is Plato’s regular formula. The only exception is Symp. 201D sqq.; and here, while no doubt the whole setting is to be taken as fictitious, Diotima appears by name as an identifiable person. In the Hippias Maior, on the other hand, we have an indefinite r1s, never named or satisfactorily described, mentioned not hypothetically but categorically as having actually attacked Socrates on the subject of 73 xaAdv (286c, évayyxos ydp tis... els dzropiav pe kateBadev xrA.). This introduction is not like anything in Plato. Socrates’ impersonation of the objector later on (2874, prpovpevos éy® Exeivoy) is of course on the lines of the hypothetical examples quoted. The use of the unknown tis is also, as Horneffer! pointed out, spasmodic and inconsistent. From 293E to 2984 he is dropped out altogether; 2984-300, he reappears; then again he is ignored from 300, to 303D. (2) The reference by Socrates to himself in the third person. Mr. Grube quotes several passages in which a person present speaks of himself, or is referred to, in the third person ; and hesays that Euthyphro 5a is ‘ an exact parallel.’ There Euthyphro Says ovde Tw dv Siadépor EvOvdpwv trav rorAXAGv avOpwrwv, ei pi) TA TOLADTA TadVTA aKpLBas eideinv. In so far as the speaker uses of himself in one sentence first the third and then the first person, there is a superficial similarity. But in Hippias Maior 2988 we have something much more pointed in sense. Socrates says the objector he really fears is tov Zwdpovioov, ds euol oddév dv paddAov tadtra éritperor avepetvnra dvrTa padiws A€eyew 7) Ws ciddra & px ofa. Here ‘the son of Sophroniscus’ is brought in as a sort of inward monitor, a wiser self, who will not allow Socrates to be slovenly in argument. This suggestion of a conflict within the personality is unlike any of the merely formal instances of third-person reference used by Plato for playful effect. Here again, as with the unknown 71s, we seem to have an attempt to imitate Plato’s manner—pushing the device a little beyond any of the authentic examples. Words and Expressions: Echoes of Platonic Phrases—To the instances previously noted by me as suggesting faulty imitation rather than mere cross-references the following may be added : 285B: 7a ovpavia radOy. An apparently unique use of the adjective as equivalent to a subjective genitive, and perhaps suggested by Phaedo 96c, 7a wepi tov ovpavody Kat THY ynv TaOn. 286D : dvapayovpevos tov Adyov. dvapdxouar is elsewhere (in classical writers) intransitive. The only other Platonic instance is Phaedo 89¢, mpiv av vixijow avapaxo- pevos Tov Lippiov Te kai KéBnros Adyov, where of course Adyov is governed by vixjow. The present construction suggests a misunderstanding or imperfect recollection of that passage. 2QIE: xatayeAdceros . .. movnpov y’'... yéAwra xtA. This seems to be a rather laboured reminiscence of Republic VII. 518s. Vocabulary.—(a) Poetic or burlesque expressions. It may be noted that all my examples criticized by Mr. Grube are spoken by Socrates in his own person, and therefore cannot (except perhaps 293A, dvopnpov) be explained either by reference to any tendency on Hippias’ part to ‘archaic and bombastic language,’ or as suited to the imaginary speaker. The following, also all spoken by Socrates, may be added to my previous list : 283B: ri0ev ... Twv TOAEwv. . .; 213D: THv ony optriay., 295B: orep£w TH Eun TUXD. 303E: ao.weoratas, 304A: ov Oeus. 304C: Satpovia tes TUyN . . =KaTEXEL. 1 De Hippia Maiore qui fertuy Platonts, p. 30. : (b) ¢ of the im attempt | 2865 Crito 53} presence 2881 collectiv: Review, |] Aga (c 300 is surely 301 ‘ beating use of tl 3035 Socrate: and ‘ bo i.e. in objectio additior To eivat, C Parm. I As dismiss pointed this or express numbe1 if ever’ admitte here sp quasi-c 3. not a : ordinar this sta thetical. formula, > setting 1, r named having » €is ocrates’ e on the dic and ooa, he > quotes )», in the thyphro aKptpsas iird and .98B we e really L padiws S a sort enly in y of the l effect. Plato’s viously ces the livalent ovpavov writers) VOpLaX O- ViKnTW. lection » be a all my yn, and once to suited . added THE AUTHORSHIP OF THE HIPPIAS MAIOR 85 (b) Colloquialisms. My point in citing these was not simply that they ‘smack of the imitator,’ but that they would naturally be characteristic of a young student’s attempt to write in easy conversational style. 286r. The use of rd éudv wpaypa is not, as I understand it, ‘ exactly similar ’ to Crito 53D as interpreted by Adam—‘ Socrates and everything about him.’ The presence of idwwrixdv limits the meaning here. 288p : cipperos. See for further instances of the later-developed use of such collective words in an individual sense my note on Catullus LXXII. 3, 4, in Classical Review, February-March, 1925. Again, I may add some further examples of comic or colloquial terms: 288B: ws yAvkts ef. A variant on ndvs in the same sense. 29OD: ETvovs. 290D: “Hpakdets. 290E: yevvaiov. Cf. Rep. II. 3728, also clearly colloquial. 2Q9I1B: amadAdfas cavrov. 2Q1E: peyadeiws. 292C: érws; echoing a question. 295B: apede, 297E: mas TL Gap’ av aywvifoipeda ; 3008: terov@vias Te TOV OvTwY OTLOUY. 300C: 7d€ws ye. I find no other instance of the adverb in the derisive meaning. 301D: aPeArepor, (c) Philosophic jargon : 300C: Aé~is Adoywv. The use of Aegis here in the verbal sense of ‘ signification’ is surely distinct from that in Rep. III. 400p, 301B: kpovere 5€ aroAapPavovres. The meaning of Theaet. 1548 is, I take ‘it, ‘beating argument with argument,’ not, as here, tapping in order to sound; and the use of the simple verb in the latter sense is not paralleled in Plato. 303B: appyra, pyra. (Bya slip I attributed to Hippias these words spoken by Socrates.) In this context, where the whole point is the question between ‘ each’ and ‘ both,’ surely the meaning of cvvaydorepa can only be ‘ taken both together ’— i.e. in addition. The idea of multiplication would be quite alien here. The objection therefore stands that irrational numbers cannot be made rational by addition. To the examples of jargon already given may perhaps be added 300E: zérov@a evar, Cf. 302B; Tvyxdve... elvat werovOvia. The only similar instances are Parm. 140a and 1484, and the construction is very curious. As regards the evidence of style and vocabulary in general (which Mr. Grube dismisses as ‘a few epic words, one or two slightly unusual expressions’), it must be pointed out that the argument against authenticity rests not so much on the fact that this or that word is poetic, colloquial, or otherwise unusual, or that any given expression is nowhere found in Plato, as on the cumulative effect of a considerable number of such abnormal expressions occurring within one short dialogue. Even if everything found here could be paralleled somewhere in Plato’s writings (and admittedly many of the words and phrases cannot), why should he have concentrated here specimens of all his capabilities elsewhere shown for writing in a quasi-poetic, quasi-colloquial, and sometimes ungrammatical style ? 3. The Metaphysical Implications of the Dialogue.—Mr. Grube says that ‘there is not a single metaphysical expression in the Hippias Mator that goes beyond what is ordinarily found in the other early dialogues.’ But his own examples do not support this statement. Two types of expression in particular may be re-examined : 86 D. TARRANT (1) 287c: Sixatoovvy Sixasot ot Sixaroi, , € / 2 ‘ \ = *§ ~ , > ‘ > / codia ot copoi ciot copoi kai TH ayady ravra tayaGe ayada. 4 ”~ a Kat Ta KaAG ravra T) KaA@ eote KaAd. With this instrumental use of the dative Mr. Grube compares Euthyphro 6p, £: GAN’ éxeivo avtd 7d eidos @ mdvTa Ta Goa Gord eoTW; EpnoOa ydp mov pia idea Ta Te avoo.a avéc.a elvat Kai Ta Gova dora, Now in this example, while the use of the dative is found, the use of the cognate word, characteristically repeated in the Hippias Mator passage, is not. That formula is worked out step by step, and as I believe for the first time, in Phaedo 100D, E, to express what Socrates gives as his most satisfactory theory of causation: ov« dAAo te roret avtd Kaddv 7 H éxeivov Tov KaXov Eire TapovTia ... ov yap éTe TovTo Sucxrupifoua, GAA’ Ste TH Kady Ta Kara ylyverar KaAG... Hyobpar. .. agpaXes civat . . . amoxpivacGat, 6te TO KaAw Ta KaAG KaAd, The formula thus arrived at is repeated with numerous other words in the paragraph which follows, Its introduction in Hippras Mator 287¢ sqq. as an established form and without comment, and with the repeated emphasis on the self-existence of 7d xaddv, etc., given by €OTL TL TOUTO . . « OGL YE TUTL TOUTOLIS . . . OVTL yé TLve TOUTw, Seems far more likely to succeed and derive from the Phaedo passage, and the whole argument in that dialogue, than to precede it. (2) 289D: eredav tporyéevntat TovTo 7d efdos. 292D: 7d kaAov avTd . . . @ dv mporyévntar.. . cf. wapayeveoOar and rapeivar, 294A, C. This semi-concrete phraseology describing the ‘ arrival’ of the avrd xaAddv is like nothing in Mr. Grube’s quotations or elsewhere, so far as I can find, in the early dialogues. This conception, again, is developed and used at length in the Phaedo (r00p sqq.). The nearest parallels outside that dialogue are the uses of verbs compounded with mapa at Euthydemus 301A (rdpeotw péevtoe Exdotw avtov KdAAos Tt. "Eav ovv, pn, tapayevntai aor Bots, Bots ef; xrA.) and Gorgias 497E (xaXeis .. . Tous Kadovs oiorep av xadAos apy) and 498D, E; also possibly Lysis 2178 sqq.; and in none of these passages is it likely that the theory of Ideas is in any sense implied. With regard to the passage in Euthyphro 68, where the idéa is referred to as a Tapdderypa, the expression is strange in so apparently early a dialogue; but it does not really touch the type of Ideal theory implied in the Hippias Maior, where, as in Phaedo 100D sqq., peOeErts alone is assumed, without any suggestion of pipyons, as the relation between particular and Idea. The avrd rd xaddv transcendentally described in Sympostum 211D does, as Mr. Grube says, ‘exist in a different world altogether.’ But so does the €xaorov 6 €ott, povoedées dv avtd Kal’ aitd KrdA. of Phaedo 78D; and yet in 100D sq. no hint beyond the simple avrd xa airé is given of such transcendental qualities for the Idea now conceived as ‘present’ in particulars. If Plato could so detach himself from the one conception and con- centrate on the other within the same dialogue, why is it beyond possibility that someone writing after the Phaedo, aware of both conceptions, but only interested in the peGeErs theory, should similarly detach himself from a mystical transcendentalism which did not appeal to him ? The confusion and vagueness of the use of «fds, id€a, ovoia, rdfos is as likely to be characteristic of an aspiring but not altogether metaphysically-minded disciple as of an immature Plato. The writer had indeed precisely ‘ enough ability to write the Hippias Mawr’—that is, a good deal of literary ability (though unevenly displayed), but much less of philosophical power, in spite of the acquaintance which he shows with the Phaedo and earlier dialogues. Mr. Grube’s a forttort reference to the Sophist, « Hippias « % 2 same na instance spurious an admi in sharp whole th himself | follow fr way ast 4. 1 wed tov Still mo (rd de oy separate! Mawr, 2 section ¢ the gen introduc argumenr of its he as I poir problem solution 5. J a legitin by some the Phil Parmenia appears Plato’s r more lik doctrine with yo as I put philosop 6. ( Platonic evidence large pr phraseol sometim the Pha tendency of Ideas Parmenta was WTI of the A BEDF 1 Pp, Fr 0 OD, E: EL Th TE e dative as Matoy > for the sfactory UTtM we Uae... arrived vs. Its mment, iven by ikely to ialogue, is like e early Phaedo f verbs KaAXos me ««s-% }-; and r sense oasa it does where, LiLNT US entally world Tr. of given nt’ in d con- y that sted in italism asa whole, implies that the editor (whoever he was) of the ‘Calliopian’ text was responsible for the modernizing. Thus (a) Eun. 582 haec A for hae 2; (b) Eun. 632 puto A for veputo = (rather the substitution of the uox propria in the ‘ Calliopian ’ text than of a modern word); (c) Phorm. 877 tnaudws A for audius 2; (d) Eun. 998 mecessus A for necesse &. But it is puzzling how the first hand of E and of » alone has conserved, at Andr, 368, the genitive Chrems, for which the regular Chvemss was substituted in all the other MSS.; or, again, how Jovialis, the corrector of A, is the sole witness (at Phorm. 1015) to the reading sed... gus, which Lindsay compares with the familiar atgut, In this last case Donatus’ commentary, as we know it, has nothing to say. Eun. 274 falsus ansmt does not raise the same problem, for the correct reading, anims, of D?, with the suprascript note alster animt secundum Donatum, is an indication that D, or rather the archetype of the 6-family of MSS., the family of which D is a member, was corrected at a comparatively late date with the help of Donatus’ commentary (vid. Lindsay, Class. Quart. XIX. 32). We need not be surprised that no change has been made in Haut. 177. Charisius says that sincere there is an adjective, so that the last syllable would be short. To understand stmcervé (adv.) would make no great difference to sense Or scansion. I summarize in order the remaining examples of archaism mentioned by Donatus and preserved in Terence MSS. : Andy. 42 aduorsum te for apud te; 433 gratia for causa; 443 licitum for licust (Don. compares placitum and puditum); 608 nulls for nuliius (Don.’s note suggests a compound mulliconsili. He also adds sols); 750 msror for nescto (cf. Eun. 290, Phorm. 235); Haut. 207 atque for atqut (Don. on Andr. 482); Eun. 56 proin for proinde (Don. compares extn); 259 gratulor with acc. ; 262 dtscere de for discere a; 269 hisce for ht (was it metrical considerations that saved the old form here?); 401 habes for intellegis (usd. Lindsay’s suggestion habent for putant in Haut. 798); 847 cognoscere for agnoscere; Phorm. 225 noxta for noxa (cf. Hec. 312, on which there is no remark in Donatus. At Eun. 852, also, Don. refers to noxta, and seems to say that in Haut. 298 Terence wrote oxa, though our MSS. have soxzsa, which is rendered necessary by the metre); 91 #it for siltc; Hec. 4 stupsdus for stupens; 160 atque for atqui; 424 aufugerim for fugertim; 755 explere ansmum for emunire* (?) an. (Donatus’ remark is Teventianum est); Ad. 27 tevant (producta I Don. It would have been easy for a corrector to write iwerant); 49 im co me oblecto for co me obl.; 116 alls for tht; 272 hoc 1 Schoell’s emendation for et inantre of the MSS. 92 J. D. CRAIG mtht dolet (rp atrixurp pro ‘doleo’ Don.); 452 pater is (Pleonastic use of ts with the Subject or Object of the sentence. For Plautine examples see Lindsay, Syntax of Plautus, p. 47); 550 etiam taces ? for etiam tace ; 841 primo luct for prima luce (Don. explains quite wrongly) ; 874 slum optant ut uiuat (ovAAnpyis, nam sic ueteres ; nos dicimus “slit optant,’ non ‘ sllum optant’ Don..). I repeat that of some items in the above list we can say at least that Donatus thought they were archaisms. (8) More interesting still are the few archaisms which occur in Terence MSS., and which are confirmed by no external authority—at least, no external authority referring definitely to the passages where they appear. It is natural that they are mostly in Haut., for which we have no commentary of Donatus, though the Bembine scholia perhaps preserve shreds of it—e.g. exercivent in Haut. 143. (a) Andry. 328 and 438 haec for hae (cf. Hec. 762, 790); (6) Hec. 735 quaests (gen.) A for guaestus = (Don. seems to imply gquaestus, but his note may have been garbled) ; (c) Haut. 1002 fide (here shortened by the Brevis Brevians law to fidé) for fidei (gen.); (d) on Andr. 763 adposistt p' for adposuisti, compare my remark on Eun, 274 above; (¢) Haut. 693 apts A for adepti =, and 1065 Archonids (gen.) A for -dis =, show that the ‘ Calliopian’ edition has dropped the archaism in both cases, as I have pointed out on Eun. 582. The other instances are: Andr. 700 haec for hae (C?7DGLEV); Haut. 8 quota (= whose); 133 ad haec utenda, the old usage of utoy with acc.; 195 and 461 atque for atqut; 210 face (D*L*E*n*) for fac; 511 comgyuére, unusual scansion only, a matter which does not appear to have concerned the old grammarians much. I have said nothing of the Infinitive Passive in -tery. The MSS. have without exception transmitted it faithfully everywhere, and the form is not thought worthy of remark by Donatus. These are most of the occurrences: Andr. 500, 573; Haut. 206, 268, 300, 716, 923; Eun. 164, 225, 510; Phorm. 92, 206, 305, 306, 406, 535, 589, 603, 632, 697, 931, 978, 1021 ; Hec. 316; Ad. 273, 535. The ‘ Calliopian ’ edition appears to have actually foisted this form on the text in Andr. 689 sollicitariey = (unmetrical). The evidence of A is wanting here in Andria. Il. But from the point of view of establishing a principle of emendation, the greatest weight must be attached to such archaisms as have been completely effaced from the text tradition of Terence, and which are restored on external authority, chiefly the authority of Donatus. In the subjoined list it is to be understood that Donatus is the authority unless otherwise stated : Andy. 656 haec for hae; 778 tuvandum for iustuvandum; 832 tetulst for tulit ; 849 etiam vespondes ? for etiam responde (A is unfortunately not in evidence for these four passages): Haut. 471 techinis for technis, cf. Eun. 718 (the evidence is gi. IJ.) ; 699 mits for meis (the witness is Velius Longus): Eun. 267 Thainits for Thazdts (see Lindsay on the ‘ legitur et’ formula in Donatus, Class. Quart. XX. 103-5); 493 poste for post (based on gl. II.) ; 314 gractlae for graciles: Phorm. 82 ardere (with acc.) for amare (Arruntius Celsus seems to attest the archaism. Donatus’ note, as we have it, is not clear. It seems, however, to imply not amare but another verb); 190 protinam for protinus (Don. of course says nothing of the short 0): Hec. 817 attoli (?) (Lindsay) for attult (A reads attolli, = attuls. Donatus’ remark is ‘ move suo,’ which would seem to imply a characteristic Terentian spelling. If so, it suggests that attul- of our 1 In the new Oxford text gi. I. and gi. IJ. two lost MSS. of the 6é-family of Terence (vid. Lindsay, Class. Quart. XIX. 101-2) denote (minuscule) MSS. MSS. s to be vemmoti Infiniti for inc 994 sect Fi with un is a cle is (see t An pos| u jis Haut. ; ut cett.; 2 suppos < Hec. 21 for virg educfa]t Thi the rem: the wor any whet Wh emendat we show For world (A (and this unfortun possess corrector attestatic Very from a : clearer b hot-head If this we For it w ancient wv had omit: end of tk place). | Terence. Same errc So I with the Syntax ¢ (Don. s dicimus Donatus 5S., and uthority they are 3embine ; guaests ve been fidé) for lark on .) A for ‘ases, as 8 quota utque for , matter without - worthy 9, 573: 06, 406, the text here in ion, the . effaced thority, od that or tulit ; or these rl, II.+); tdts (see 93 poste cc.) for have it, brotinam indsay) Id seem of our Terence ARCHAISM IN TERENCE 93 MSS. should always be changed to attol-): Phorm. 1028 sum for eum (but this appears to be a quotation from some old tragedian—e.g. Livius Andronicus): Hec. 22 vemmotum, with double m ; 94 tls for illic; 314 the construction of fortasse with the Infinitive, which is implied by g/. JJ. and, apparently, by Donatus : Ad. 206 occeperis for inceperis; 264 pote for potest; 674 illim for illinc (the authority is gi. J1J.); 994 secundare for obsecundare. III. Finally we turn to the restorations of modern scholarship, which, confronted with unmetrical lines, proceeds on the principle that possibly the secret of corruption is a Classical and familiar form which has ousted an unfamiliar archaism. The list is (see the new Oxford text for details) : Andy. 439 hospitai; 483 post; 521 tamenfi]dem ; 729 adpos[ulisse; 742 pos[uJisti; 789 atti[n]gas ; 814 grandi{us|cula (gi. JJ.); 898 [a]mittere ; 957 fors[itan] : Haut. 129 soli{us]; 165 inpulerim; 515 Cliniai; 693 opinofr]: Eun. 117 ut ; 197 fors[itan]; 197 paruom (?) for parum CD!GL'$!, paruam (sc. fidem) cett.; 268 and 312 re[s e]st; 356 mage; 364 ut; 624 puer; gI2 supposuit; 1010a ut: Phorm. 566 it[in]ere ; go5 ut ; 993 credas: Hec. 217 illi[c]; 719 puer; 729 possim: Ad. 146 sem; 346 virgini for virgine (cf. Lindsay, Early Latin Verse, pp. 81 sqq., especially p. 85); 495 educfa]ti; 940 fac. That the smooth surface of our Terence, as compared with Plautus, is deceptive the remark of Festus (p. 11) must bring home to us. Festus says that Terence used the word affatim in the sense of ad lassitudinem. And yet affatim does not occur anywhere in the six comedies as we know them. What is the moral of all this? That new tactics should be adopted by emendators of Terence; that Terence is not to be emended in the same fashion as we should emend other authors. For witnesses to the text of Terence we have (1) a manuscript of the ancient world (A); (2) a number of transcripts (=) from another MS. of the ancient world (and this other showed a different edition of the text from A); (3) the commentary, unfortunately very scrappy and corrupt, of Donatus, who, we must remember, would possess all or most of the previous commentaries on Terence; (4) Jovialis, the corrector of A in the sixth (?) century; (5) for single lines here and there, the attestation of this or that grammarian. Very different conditions from, e.g., Phaedrus, all of whose MSS. are transcripts from a single lost MS. of the ancient world. But the argument will be made clearer by an example. Take the iambic senarius, Haut. 58g (spoken by the more hot-headed of the two ephebi): di te eradicent, Syre, qui me hinc extrudis. (w) If this were a line of Phaedrus, no exception could be taken to the emendation : di te eradicent qui me hinc extrudis, Syre. For it would be a quite legitimate assumption that the single lost MS. of the ancient world (or even some later MS., an archetype codex of ali our existing MSS.) had omitted Syve at its proper place and put it, when the error was discovered, at the end of the line (or in the margin, with or without a mark to indicate its proper place). But this assumption is impossible, or at least unlikely, for the text of Terence. Can we reasonably assume that different ancient editions fell into the Same error, unless that error was of the ‘ inevitable’ type? So I prefer to suppose that Terence wrote : di te eradicent, Syre, qui me hinc extruditis, 94 ARCHAISM IN TERENCE and that the bizarre expression te .. . gus extruditis would be corrected (in the interests of schoolboys) to te .. . gus extrudts. (Of course the Indicative is natural in colloquial Latin.) The expression does not seem too bizarre for Terence. It is not a use of wos for tu. Clitipho is in a rage, not merely with the slave Syrus, but with his own father Chremes. Terence could not indeed make him say ds wos evadicent. The pretas of a Roman audience would not tolerate that. But it suits the unbridled temper of Clitipho that he should show his anger with his father as well as with the slave. And the substitution of extruditis for extrudis gives just that delicate touch which is characteristic of Terence’s diction. Not quite a parallel, but worth mention, is the plural at Haut. 473: Syrus cum illo uostro consusurrant, or at Ad. 774: in ipsa turba atque in peccato maxumo, quod uix sedatum satis est, potatis, scelus. In these two passages the apparatus critscus of the new Oxford Terence shows how MSS. have got rid of the plural—in the second completely (A, however, by accident) ; but that Donatus had a plural, whether Present or Perfect, in the text before him is certain from his note (and Jovialis, who corrected the miswriting potts of A, probably got fotatis from Donatus). The plural fotatis is natural. Demea, exasperated, and taking a gloomy view of the Micio household, makes an inference from the particular to the general. The point is unobtrusive, like many another in Terence. Take another example—Haut. 596 (a trochaic septenarius) : Cu. repperisti tibi quod placeat an nondum etiam? Sy. de fallacia dicis? est: inueni nuper quandam, etc. Editors are conceivably right in cutting out dum etiam, It is conceivable that the sense was made clearer than daylight to the schoolboy reader by an nondum etiam instead of an non. But I should like to point out another possibility. Fallacia is the word usually employed in this play for the trick which Chremes urged Syrus to carry out. But not invariably (e.g. fabrica appears in line 545). If, then, some archaic or bizarre word was employed here it would inevitably be changed to fallacia. The true restoration of the metre may be the substitution of an iambic word for the quadrisyllable. What iambic word? Best of all would be the Old Latin word falla or fala, if we could be sure that fala was its real form. I do not, however, dare to propose it, since the evidence—such as it is—favours the trochaic form. Plaga, in the Singular, may also be challenged. The best word I can think of is stropha, but it, too, is not convincing. It will be clear that I have done no more than emphasize a feature of the new Oxford Terence, and that my obligation to Professor Lindsay, privately, is great. J. D. Craic. East ScorgEs, St. ANDREWS. T Tu are fort hithertc without but ext though Th guardia: mention I. Republ: and the 2. | others ( to be dc . . children five to f official « 5. songs (4 6. ° controll vary it | . are also to be se not the 1 Theg with tha’ seem not quences Review 1\ sistent a word, S of Aristot 2 Rep. 3 That Trav Te Wa does not auxiliarie yuvatk@y 464A and use of é1 plained i1 The wor émixoipor } times in (in the natural f wos for n father tas of a mper of e slave. which is . shows ver, by he text ng potts Demea, iference other in hat the m etram Fallacta d Syrus n, some iged to iambic he Old do not, rochaic 1 think he new eat. RAIG. THE MARRIAGE LAWS IN PLATO’S REPUBLIC. Tue difficult and apparently inconsistent regulations by which certain marriages are forbidden in the Republic have not, it would seem, been consistently explained hitherto.!. It is the purpose of this article to prove that—if we read Plato’s text without prejudice—marriages between brothers and sisters are nowhere prohibited, but expressly allowed; and that there are in the ideal city certain family groups, though I do not contend that any very great importance is to be attached to these. The passage where Plato explains what is and what is not incest among his guardians offers the chief difficulty. But before we deal with that it may be well to mention briefly the regulations that Plato has already made clear : 1. We must remember that the communism of wives and children in the Republic (which is also communism of husbands) concerned only the guardian class, and the whole of that class.? 2. Only the legitimate children of the fittest people are to be reared. The others (whether born out of state wedlock to anyone or to unfit parents within it) are to be done away with‘ (459D-E, 460c, 461B-c). 3- No one under or over the age fixed for official unions will be allowed to have children,’ The marriageable age is fixed at twenty to forty for a woman and twenty- five to fifty-five for a man.° 4. No man or woman within this age limit is to have any unions apart from the official ones (457D, 458D). 5. These official unions will take place at special festivals with sacrifices and songs (459). 6. The number of these marriages’ that take place at any particular festival is controlled by the archons in order that they may regulate the number of births, and vary it according to the needs of the state (4604). 7. Further, in order to ensure the health of the coming generation, the archons are also to decide what particular people are to be united on each occasion. This is to be secured by a faked ballot, so that those who are not chosen will blame chance, not the rulers (460a).°® 1 The general verdict seems to be in accordance with that of Herbert Richards: ‘ Plato would seem not to have thought out all the conse- quences that would or might ensue’ (Classical Review IV., p. 8). Even so, no glaringly incon- sistent account ought to be taken as the last word, See also I. A, Loos, Studies in the Politics of Aristotle, and the Republic of Plato (1899). 2 Rep. V. 461B-E. 3 That the expression 7 cowwvla rots émixodpos rav te waldwy xal ray yvvaxdy in 4648 and 466A does not imply that the system is meant for auxiliaries only is proved by the expression 7 rév yuvaikay re xal waidwy xowwvla rots Pidaiw in 464a and the use of ¢0Aaxes throughout. The use of érixofpo to include all guardians is ex- plained in various ways (see Adam's note, ad loc.). The word ¢v\axes is used immediately before émixodpo: in both passages (once in 464 and three times in 466a). Plato may be using ém:xoipa to avoid wearisome repetition, and because they, being younger, are chiefly concerned. 4 It is beyond doubt that Plato recommends infanticide. See Adam’s App. IV. to Book V., and Jowett and Campbell’s note on 460c. Even scholarly attempts to prove the contrary are due to Christian prejudice. 5 461A. 6 460£ and Adam’s note. 7 So Jowett and Campbell take 7d rA7O0s Trav yduwv in their note ad loc. It is obvious that, unless the rulers can control all child-bearing unions (and not only the majority), their control is of no avail, 8 It is surprising that Plato cannot trust his guardians, carefully educated as they were, to obey the rulers without this rather mean trick. But we must remember the bigger portion of them—the ém:xofpo—do not share in the higher intellectual training later to be described. se Wiki" = Wee hg 96 G. M. A. GRUBE 8. Parents must not know who are their children (457D). Mothers may suckle babies, but will not know whether it is their own or another’s they are nursing (460c-D). 9. All children are taken to a public créche immediately after birth and looked after by public nurses (460c-p). It is important to realize, with Richards,! that these marriages or unions do not last longer than the festivals themselves. This is clear from Plato’s analogy with breeding animals (459a-B), and from the fact that in no other way can the rulers effectively control the numbers of the population. Besides, the fittest? people are to have many such unions, and Plato did not mean his warriors (whether man or woman) to live in a harem, but in a camp (4608, 468c). A good deal of licence is allowed to those over age. They can have irregular (but of course childless) unions with anyone they choose, with the following exceptions: For a man—his mother, mother’s mother, daughter, and daughter’s daughter ; for a woman—her father, father’s father, son, and son’s son (461c). This includes all grandparents and grandchildren, for, as Richards says: ‘If a woman cannot marry her father’s father, a man cannot marry his son’s daughter.’ After these restrictions have been made for those over age, Glaucon asks how fathers and daughters, mothers and sons, and others just mentioned, will know each other. Socrates’ answer must be quoted in full, for it is the crux of the whole question : ovdapos, hv & éyw: aGAXd’ ad’ Fs dv juépas tris avrav vupdios yévnrac, per’ exeivny Sexadry pnvi «at EBSépo Sy a av yévnta Exyova, ratra mdvta mporepe; 7a pev Gppeva teis, ra Sé OyrAca Ovyarépas, Kal éxeiva éxeivov marépa, Kai ovTw oy Ta TOUTWY Exyova Taidwy waidas, Kal éxeiv’ ad éxeivous mdmrmous Te Kai THOds, Ta S Ev Exeivy TY XpPOvw yeyovora, év G ai pyrépes kai of marépes aiTav éyévvwy, adeAdds te Kai adeAgots, Gore, 5 vuvd% EAEyopev, GAAHAwY py GrrerOar, adeAgods de Kai adeAgas ducer 6 vouos ovvorxely, €av 6 KAHpOS TavTy Cuprirry Kai » [[vOia mporavaipy. The end of this passage, from ra 8’ év éxeivy, is usually taken to mean: ‘And those born within the time during which their parents were engaged in child-bearing are called sisters and brothers, so that they can have no unions together, as we were saying just now. But the law will allow brothers and sisters to live together if the lot should so fall and the Pythian agree.’ adeAdpds re kai ddeAdovs are then taken to be subjects of pw) drreoOa, and brother and sister unions are taboo except in special cases. But this makes nonsense of o vuvdy éA€youev,? for brothers and sisters have not been mentioned before. I will take the doubtful points in order ; vuppios must mean bridegroom ; see 459E, VI. 495£. So Jowett and Campbell ; the time referred to is a particular festival. Translate: ‘From the day on which one of them is granted a union.’ 1 Classical Review, Vol. 1V., pp. 6-8; Table of Affinity in Plato’s Republic, 2 Is the standard of fitness purely physical ? I think Plato is uneasy on this point, and betrays this by such expressions as roils dyaGois trav véwv €v wohéup % &ddoGi wou (4608), and airy dxuh Twpards Te Kal ppoviwews (461A) ; for it is contrary to Plato’s usual attitude to age to say that a man’s dxuh gpovjcews is from five-and-twenty to fifty-five. He would begin dialectic at thirty, and reaches his highest level of theoretical and practical knowledge at fifty. See VII. 5394 sqq. 3 As Adam admits in his note ad loc. ; see also Richards : ‘Although in 461p brothers and sisters are Curiously omitted from the list of persons forbidden to form irregular unions, we seem obliged by these words to include them.’ The same mistake is made by Bekker (1825 edition), by Jowett (in his translations and in introduc- tion, Vol. III., xxiii), by Ernest Barker (preface to Methuen’s translation, 1906), and in the translations of Sydenham and Taylor (1804), Davis and Vaughan: ‘all these shall refrain’ etc., where all must include brothers and sisters, following Schleiermacher (edited by Kirchmann, 1870), H. Spens (Everyman), A. D. Lindsay. Massey (1713) and Schneider (1850) preserve the ambiguity of the original. Grote is sound as far as he goes, but does not mention brothers and sisters. for this points o avtwv to to the ch to refer t the state months particula the moth and mot! adeX: together. no irreg all-impor they can: has been fact that parents— relation 1 grasped is here n to adeAgdc other, wl expresse: groups c taken in not have one that refer to adeAors class tha between adeAdds 7 them as 1 we discus the other they are | adeXc the same that the «¢ are to be would th relations. people w! brothers | say SO, ar 1 It was, an eight | correct rec. this was th the quotati y suckle nursing d looked is do not gy with e rulers ople are ler man rregular lowing ughter’s . This woman sks how 2w each 2 whole Zt, per’ iT POT Epet » b) Ta KEeiV® adeAas > b€ Kai tp) wed in gether, to live yrother nse of I will ipbell ; which fe seem .. The sdition), itroduc- ‘preface in the (1804), refrain ’ sisters, hmann, indsay. reserve ; sound rothers THE MARRIAGE LAWS IN PLATO’S REPUBLIC 97 Sexdty pnvi xat éBdoup: Adam rightly takes this as ‘in the seventh and tenth month,’ and rejects Jowett and Campbell’s rendering ‘from seven to ten months.’? Ta 8’ év éxeivy . . . éyévvwv: This does not refer to ‘the whole time of life during which father and mother were allowed, if the lot fell upon them, to take part in regular unions’ (Richards, followed by Adam). Nor is Jowett’s version correct: ‘all who were begotten at the time when their fathers and mothers came together’; for this mistake the sense of é¢v, and does not account for the imperfect (as Adam points out). Jowett is right, however, in understanding ai pnrépes xal of rarépes avrov to refer to the whole class of ‘ parents’ who came together at the festival, not to the child’s real father and mother. But €v éxeivw tw ypovw must, I think, be taken to refer to the period mentioned before—namely, the seventh and tenth month after the state marriage—and the imperfect is used because it refers to the whole two months when this class of parents were bringing children to birth, not to the particular day on which a certain child was born. It is true that, strictly speaking, the mother alone is concerned, but I do not think it unnatural to make both ‘ fathers and mothers’ subjects of éyévvwy in this sense. adeAdds re wai ddeAdots . .. py GrrecGac: These words are usually taken together. We then understand that these so-called brothers and sisters must have no irregular unions when over age, but that the taboo does not apply to the all-important state marriages. This seems somewhat extraordinary, especially since they cannot be blood-relations at all, except in the case of twins. The point that has been missed is that they are called brothers and sisters merely to emphasize the fact that they are all to bear the same filial relationship to the same group of parents—in fact, that they all are ‘children of the same parents.’ It is the parental relation that this whole passage is meant to keep clear. This will be more easily grasped if we print a colon after adeAdovs instead of the usual comma.? There is here no reference to unions between brothers and sisters at all, for dAAjAwv refers to ddeAdds re xai dSeAghovs on the one hand and their parents and grandparents on the other, which was being discussed just before (6 vuvdn eAeyouev). The consequence, expressed by dere dAAnAwY py arteoGa, does not follow from the fact that certain groups of children are to be called brothers and sisters, but from all the precautions taken in this paragraph ‘so that they (i.e. the children and their parents, etc.) shall not have unions.’ In support of this interpretation I would point out: that it is the only one that makes sense of 6 vwvé;) éAéyouev. That Plato was not quite so careless as to refer to something he did not say less than ten lines before. That adeAdgds re xai adeAgovs are grammatical objects of rpowepe? ris. It is the members of the parent class that ‘will address them as brothers and sisters,’ and it is the relationship between him and them that we are concerned with. That the locution re xai binds ddeAdds te wai ddeAdovs into one group, and that this makes it difficult to think of them as two, as we must if dAAnAwv refers to their mutual unions. Note that when we discuss unions between brothers and sisters the words are not so joined. That the otherwise unnecessary repetition adeAdots cai adeAdas makes it at least likely that they are not sole subjects of arreoOau. adeAdoris 5é cal addeAdds: These words are perhaps most easily taken to refer to the same artificial groups of ‘brothers and sisters’ as above. I feel, however, that the expression 6 vuvdy éAéyopev rounds off the question of how parental relations are to be kept clear, and that a new point is made in the following sentence. I would therefore take ‘ brothers and sisters’ here at least to include the real blood relations. For the sanction of the oracle would surely not be required to unite people who are not really kindred. But whether implied or expressed, the fact that brothers and sisters can be married is clearly intended. There is, indeed, no need to say so, and this last sentence merely forestalls a possible objection. Indeed, no one 1 It was, and is, a common superstition that | whose mode of expression is the same as Plato’s an eight months’ child does not live. The here; in Diels, Doxogr. Graec., pp. 427-429, on correct reckoning should be seven and nine, and __ the question d:a ri ra éwraunmaia youima. this was the usual form of expression. But see 2 The only text that puts a longer stop at the quotations from Polybius and oi uadnuarixol, ddedpovs is Hermann’s Teubner text, 1887. 98 G. M. A. GRUBE will know who real brothers and sisters are.1 I would then translate the whole passage as follows: ‘ They will not know each other at all, said I. But all the children born in the seventh and tenth month from the day on which any one of them is married he will call his sons if male, his daughters if female, and they will call him father ; and the children of these last will call the first group grandfathers and grandmothers, while all the children born in the above-mentioned period when their mothers and fathers were bringing children into the world he will call sisters and brothers; so that, as we were Saying just now, they (i.e. those in parental line) will have no unions with each other. But brothers and sisters the law will allow to live together if the lot so fall and Pythian also agree.’ The usual theory is that Plato intends a// children of the same generation to be brothers and sisters, and to call the older generation their parents throughout the guardian class. This is unwarranted by anything we find in the text. Even we admit Adam and Richards’ interpretation of éxeivy to yxpdvy as the whole period of possible child-bearing—a view I hope to have proved mistaken—it does not tally with Plato’s quite unambiguous words just before that, when anyone is married he will consider as his sons and daughters the children born in the seventh and ninth month after his marriage day. For this sentence, which has been conveniently passed over by commentators, quite clearly implies the division of the guardians into groups of parents and children, though individual parents will belong to as many of these as he or she has been allowed official marriages. It is interesting to inquire what the size of these groups would be. The number of Plato’s guardians may be taken at 5,000.2, The duration of human life in the ideal Republic is 100 years. The number of the population is to be static.* It follows that in a normal year the number of people born to live to maturity is about 50. Infant mortality and exposure of unfit babies would be low under ideal conditions, nor would losses in war be very high. Allowing 20 per cent. for these and similar causes (a liberal estimate), 60 children will be born per year on the average, but the number might of course be exceeded, and, again allowing for accidents, about 128 parents would be required.® Now it is quite clear that marriage festivals would only take place at definite times. The frequency of these festivals does not really effect the argument, for the closer we bring them together the smaller our groups will become. If they happen twice a year, we get 64 fathers and mothers to at most 30 children; if four times a year, 32 parents to 15 children, and so on. To know that your child is one of a given fifteen is a different thing from not knowing it at all. 1 Adam, thinking that unions between brothers and sisters are forbidden, has to suppose that the rulers would know the real blood relation- ship, and that they would see that a real brother and sister union would never take place. But we are not the people that are to be deceived, so why does Plato tell us that these unions are to be allowed? Nor is there anything to show that any records are to be kepi of how people are really related. I am sure Plato intends this secret to be kept by nobody. Without this knowledge Adam is forced by his rendering of éy éxelyy tT» xpévw into flagrant absurdity. He tells us that ‘a son, for example, who is born when his mother is twenty-one and his father is twenty-six cannot marry till he is forty-nine.’ The impossibility of this is evident if for ‘a son ’ we substitute ‘adaughter.’ Her future husband could not be born till she is twenty-nine, and therefore cannot reach marriageable age till she is too old to have an official union atall. Besides, how is a child to know its parents’ age when it does not know who its parents are. 2 This is a maximum. Plato nowhere gives figures, and his fighting force of 1,000, mentioned in IV. 423A, is meant as a minimum, not to be taken literally. 3 See Rep. X. 6158 and Adam's note. @ 460A. 5 Clearly these figures are only an approxima- tion. They are quite near enough, however, to serve as illustrations. Note that exposure and infant mortality would tend to make the children group smaller. 6 459E, VI. 546a, and Adam’s note. Ther marriage: and it onl and fifty- avoid as unions wi! after one regulatior groups.* Did | that it wa whether s stones to kind of fi That not be ig did imag achieve u Nor incompat a fellow- meeting |: of his for] know his of his ‘fr inquiry, | between of this kin Timaeus, | some suc reference: in detail, UNIVE S 1 Aristot TrouTwv vio exaggeratic as we have 2 Rep. V $s E.g. i seems to — earlier the erroneous Rep. X. 5¢ excluded a do (as Ada X. 603A tc 1e whole ‘n in the d he will - and the rs, while d fathers at, aS we vith each ot so fall ion to be hout the Even we le period not tally urried he nd ninth veniently ians into many of ye. The in life in tic.* maturity w under cent. for yer year wing for - definite , for the ’ happen if four ur child nine, and ze till she Besides, e when it ere gives nentioned not to be )proxima- wwever, to sure and e children THE MARRIAGE LAWS IN PLATO’S REPUBLIC 99 There was never any danger of incest between mother and son in official marriages, because the mother is at least forty-five when her son reaches twenty-five, and it only exists between father and daughter when the man is between forty-five and fifty-five. But in any case the existence of these groups makes it very easy to avoid as it keeps the filial relations clear, and all a man has to do is not to have unions with a member of any group of those born in the seventh and tenth month after one of his own official marriages. He may marry anyone else. These regulations become very much simpler when we realize the probable size of these groups.’ Did Plato intend these groups to take the place of the family? Did he realize that it was impossible to expect parents to love the whole of the next generation, whether selfishly or unselfishly ? and did he mean these groups to act as stepping- stones to facilitate the passing of the unsocial family love he wished to avoid into tf kind of filial patriotism that he obviously had in view ? That is, of course, pure conjecture ; but the groups are there, and they should not be ignored. They were clearly present to Plato’s mind, and I believe that he did imagine that by making the limits of the family wider and vaguer he would achieve unity by means of relationships so intertwined as to be inextricable. Nor are references to the communism of wives and children in the Republic incompatible with this interpretation. Plato says that: ‘No one will look upon a fellow-guardian as a stranger. Whomever a man meets he will think he is meeting a brother or a sister, a father or a mother, a son or a daughter, or someone of his forbears or of his descendants.’2 Indeed, this is strictly true, for he will never know his real brothers and sisters at all, nor will he know by sight all the members of his ‘parent’ and ‘grandparent’ class. All he can find out, even on careful inquiry, is whether there is any possibility of a direct parental relation existing between them. I also feel that there is more point in comparing a community of this kind to a living organism. It is true that elsewhere in the Republic, and in the Timaeus, he refers to his marriage laws in the most general terms, and seems to imply some such an arrangement as has been fathered upon him.* But Plato’s cross- references are almost always vague. They are meant to call to mind, not to repeat in detail, what has gone before. G. M. A. GRUBE. UNIVERSITY COLLEGE, SWANSEA. 1 Aristotle’s yiyvovra 8’ éxdorw xii Tov mokitay viol (Pol. II. t261b 11) is not only an exaggeration, but quite contrary to Plato’s text as we have it. 2 Rep. V. 463c-D. 3 E.g. Rep. VIII. 543 (where the reference seems to be to IV. 4238). For reference to earlier theories which would give an entirely erroneous view of what has gone before, see Rep. X. 5954, where he speaks as if he had excluded all munrixh before, a thing he did not do (as Adam notes). See also the reference in X. 6034 to his earlier psychological theory of Book IV. 439D, as if he had never made @duos the third part of the soul. The same applies to the brief recapitulation of Timaeus 18D, though Scomep dv ris mperovtons Evros jdxlas ylyvwrrac might be taken as those born in the seventh or tenth month after a festival, for the groups would not overlap unless there were more than four festivals a year. In any case, the Timacus directly contradicts the Republic on the question of infanticide. On this question see Adam’s App. IV. to Book V., where he comes to the conclusion that ‘in any case we must interpret the Republic by itself.’ VIRGIL AND THE SIBYL. Ir the Bucolics as a whole ‘look on us with dark enigmatical eyes,’ for long no more impenetrable darkness and no more compelling enigma could be found than that of the fourth poem in the collection. But Skutsch, the authors of Vtrgil’s Messtamc Eclogue, and more recently Mr. Royds in his Virgil and Isatah, have done much to solve its many mysteries. Their efforts, however, were mainly directed towards the problems of fact contained in the Eclogue, and so far little attention seems to have been given to an equally interesting point, that of its metrical arrange- ment, It is plain that we must consult the Sibyl! for help: if it is not plain, that is not Virgil’s fault. I hope to be able to show here that Virgil himself has left in this poem certain unmistakeable signs for those who have eyes to see and ears to hear, besides his own rather casual reference to the Cumaeum carmen, and incidentally to dispose anew of the theory which reads in that expression an allusion not to the Sibyl but to Hesiod. In the first place, it is certainly not fanciful that in the metrical arrangement of this Eclogue the reader is conscious of a new note, quite different from that of the rest of the Bucolics. When Virgil wrote paulo maiova canamus, he was not indulging in airy boasting—he never did. But I think that these words refer not only to the subject but to the manner of the poem, and that he quite deliberately changed his usual methods to suit his unusual theme. The time has surely gone by when we are to regard the Bucolics as mere juvenilia: the late Professor Phillimore has most sympathetically and conclusively shown in his pamphlet, Pastoral and Allegory, that we must think of them as at least the work of a mature poet if not of one who had reached his full maturity. And if Virgil is here not merely trying the ground, but is definitely sure of himself, then it is not incongruous to attribute to him a purpose which might certainly seem ludicrous in a youth coming now for the first time before the literary circles of Rome. I think that just as in the other Eclogues he takes as his model Theocritus, so in this particular poem he has set himself to produce a Sibylline atmosphere, or if not Sibylline, something very like it. He was writing a prophecy, not a very common subject. Where would be the obvious place to look for the proper tricks of the trade? Where but in the ‘Sibylline’ literature? Virgil was well-read and receptive: these oracles must have been quite familiar to the Graeco-Roman world, as is shown by the numbers that kept coming in when Augustus not so very much later ordered an official collection of them to be made: I do not therefore think it too improbable that when Virgil wished to prophesy he turned to them for the manner of his new art, in much the same way as any modern balladist—e.g. Andrew Lang—might turn to the vast anonymous collection of Berder ballads to see the way to set about it. Sir W. M. Ramsay, in the Expositor for 1907, noticed that the versification of the poem was peculiar, and he quoted Landor’s remark: ‘ The lines afford one another no support . . .: toss them in a bag and throw them out, and they will fall as rightly in one place as another.’ We may even go a step further and say that if this were an ordinary poem by an ordinary writer much of the versification would be positively bad.2 Most people have noticed how frequently the end of a line coincides with a 1 Throughout this paper I have used the term or Christian origin. | ‘Sibyl,’ ‘Sibylline Books,’ ‘ Sibylline Oracles’ 2 Landor absolutely condemned it without in the sense of the collection of Greek oracles __ realizing that the peculiarities are deliberate. usually so called, which were chiefly of Jewish pause in | acteristic: that Catu revealed finished < without < pointed o poem. I they appée poem gre There in beginnin; while in : The rhyr following suberunt p 37-40 en not only | two grot Again, lo That isa This is complete This survey W sonantal ‘long no und than £ Virgil's ave done directed attention arrange- n, that is ft in this to hear, ntally to it to the sment of f the rest ilging in y to the nged his n we are as most ory, that who had d, but is purpose ie before takes as oduce a writing : to look Virgil r to the n when nade: | hesy he modern ' Berder n of the another rightly lis were sitively with a without orate. ROLAND G. AUSTIN IOI pause in the sense or the actual end of a sentence. This is one of the chief char- acteristics in Catullus’ Peleus and Thetis, and it has to be excused there on the ground that Catullus was a pioneer, a young writer for whom the hexameter had not as yet revealed its possibilities. Now when we find the same trait in Virgil, the most finished artist of his or any age, what are we tothink? Did Virgil write anything without a purpose? But there is much more than this. No one seems to have pointed out the enormous amount not only of assonance but of actual rhyme in this poem. It is true that here and there in the Georgics rhyming passages occur, but they appear to be accidental and isolated, whereas in this Eclogue the shortness of the poem greatly emphasizes the phenomenon. Take the opening lines (1-10): Sicelides Musae, paulo maiora canamus. non omnes arbusta iuuant humilesque myricae ; si canimus siluas, siluae sint consule dignae. ultima Cumaei uenit iam carminis aetas ; magnus ab integro saeclorum nascitur ordo. iam redit et Virgo, redeunt Saturnia regna; iam noua progenies caelo demittitur alto. tu modo nascenti puero, quo ferrea primum desinet ac toto surget gens aurea mundo, casta faue Lucina; tuus iam regnat Apollo. There in ten lines are twenty-seven assonances of o and ae alone, a significant enough beginning. In line 12 imcipsent occurs in the same position as manent in line 13, while in 14 and 15 soluent and acctpiet similarly correspond. Take next lines 20-23: mixtaque ridenti colocasia fundet acantho. ipsae lacte domum referent distenta capellae ubera, nec magnos metuent armenta leones. ipsa tibi blandos fundent cunabula flores. The rhymes here are marked with singular clearness. In lines 28-31 we have the following endings: flauescet campus artista, pendebit senttbus uua, sudabunt rosctda mella, suberunt priscae uestigia fraudis. Lines 32-3 end in cimgere murts, tnfindere sulcos. Lines 37-40 end thus: fecenit aetas, nautica pinus, omnia tellus, uinea falcem. Here we have not only assonance but four consecutive lines ending with the same rhythm (in the two groups just previously quoted there are five consecutive similar endings). Again, look at lines 48-52: Adgredeve 0 magnos, aderit iam tempus, honores, cara deum suboles, magnum Iouis incrementum ! adspice conuexo nutantem pondere mundum, terrasque tractusque maris caelumque profundum, adspice, uenturo laetantur ut omnia saeclo. That is almost more striking than lines 20-23. Finally, there are the two lines (58-59): Pan etiam Arcadia mecum si iudice certet, Pan etiam Arcadia dicat se iudice uictum. This is almost like the tongue-twisting sentences of our childhood—the nearly complete similarity makes one blink to discover where the actual difference lies. This is only a rapid summary of the more self-evident passages: a more careful survey would doubtless bring out many more points, especially in the way of con- sonantal assonance. The results are of course the more remarkable owing to the po bs i es hae Oke 07 ant sina Ag Aten 5 S40." Saag via * . e are rm nem ee oe eo Oe «idee Wael Pa pre Tae | ie ER RR sg stl Sinan ae: Se ae en Ries 102 ROLAND G. AUSTIN shortness of the poem. Why should there be so much rhyme? I must confess that I am led to one conclusion only—namely, that Virgil must have found the same phenomena in the Sibyl. Now acursory examination of the extant Sibylline books will reveal that in all alike, whether in the pre-Christian parts or the later additions made right up to the time of Justinian, similar rhymes and assonances are to be found, though with the far greater mass of material they are not always quite so obvious. As definite examples let me quote from the oldest parts the following (using Rzach’s text) : III. 275-7, mdvra xapat revéovrar, Ste ppeviv ov éréOnxas aBavarto.o Geov ayviv vopov, dAAG rAavnbeis eidwAois eLarpevoas decker ovde poBnOeis .. . III. 780-3, maca yap eipnvn adyabav ei yaiav ixnrat’ poppainv & adeAovor Geov peydroro rpopyras ° avTot yap kpitai ciot Bpotav BacrAcis re Sixacor. ca ‘ ‘ on > > , , ; eoras dy Kat mXovros €v avOpwrous Sixacos. ITI. 788-90, 5€ AvKor re Kai dpves év ovperty Gupey’ ESovrat xoprov mapddXsés 7’ Epihos dua Booxynrovra. ° GpkTo. crv poryxotrs vouddero” avrAw Pycovrac. III. 827-9, ov pev €y® vipdn xai ad’ aiparos atrov érvx Any, T) Ta Tpwr’ eyévovTo, Ta 8 Evyara wavr’ aredeix On * as 7 am’ euov oropatos Tad GAnOiva ravra AcA€Ex Ow. Now these are only isolated instances of what can be found throughout all the books on almost every page. Equally marked are assonances of the type of Acpads cai Aoipds and strings of homoioteleuta such as— II. 9-10. xai POopai avOpwrev tre, Bowy Sé re pvKopevdwv TETpATOOWY KTHVOV TE Kai OUpywy TaAaEpyov. It will of course be objected that all these things are due simply to the grammatical exigencies of an inflected language. But though it is unreasonable not to expect future tenses from a prophet, it is surely legitimate to demand from one who combines prophecy with poetry some avoidance of such jingles. May we not there- fore assume that their occurrence is deliberate, and, in fact, part of the prophetic manner, just as for example ‘They hadna sailed a league, a league, A league but barely three. . .’ is typical of the ballad style? Nor can it be altogether coincidence that these assonances tend to manifest themselves not so much in single couplets (though these are frequent enough) but in little groups of three or four lines at a time. Now if we find the same characteristics in Virgil, as I think is apparent from my quotations above, there is but one conclusion, that in this particular poem he has much more than dipped into some of the Sibylline Oracles as a source, but is definitely trying to suggest the Sibyl’s manner. For even though of our extant Sibylline books only parts of Book III date from before Virgil’s time, we know that what we have only represents a stabilized tradition of the masses of these oracles that must have been current ; and I think it is reasonable to assume that since these metrical peculiarities are in evidence throughout all the books, late and early alike, they were definitely the proper thing to do, and so whatever was available for Virgil’s own reading would necessarily have conformed more or less to the same type. There are, however, a few other smaller points in which the actual language of the fourtt can stres: Book III simul . . frequent in ‘iam re ueneni O. several pa The stron common i is alsoac which at | actual ver comparisc with Oy. ; of Ecl. 32 of Ecl. 38, or Or. Sb Although echoes of least suffic my purpo: noted by | themselve appendix), Saturn, E may perhz of Babel a reference maneat, etc 1 For this Aristophane fess that he same 1e books \dditions re to be quite so ollowing ie books at Actos imatical » expect ne who t there- rophetic at these ‘rh these ww if we otations +h more rying to ks only ve only ve been liarities tely the x would ‘uage of VIRGIL AND THE SIBYL 103 the fourth Eclogue shows a fellow-feeling with that of the Sibyl, though naturally one can stress them only where they correspond with the pre-Virgilian passages in Book III. But, as a point of style, does not the double ‘At tibi . . 2 and ‘At simul . . .’ of lines 18 and 26 recall the dA2’ érav and the avrip érei of the Sibyl, so frequent in adducing some new stage of prophecy?! The use of anaphora as in ‘iam redit et Virgo, redeunt Saturnia regna,’ or ‘ occidet et serpens, et fallax herba ueneni Occidet,’ though doubtless a mannerism of Virgil’s own (Conington quotes several passages), is at least to be noted more than once in the Sibylline books, e.g. : III. 378, pevger’ ax’ dvOpumrwv revin cal pevger’ dvd-yxn. VI. 9, SeiSer & dvOpwmrowrs dors, Seige 5¢ xeAevOovs. VII. 9, mrAeioe vq, rAcioere 8 Spy, rACioes SE Kai aiPyp. The strong pause after the first foot, as in occedet in line 25, though not of course un- common in Virgil, is also a noticeable feature of the Sibylline hexameter. There is also a couplet in Or. Ssb. III. 394-5— Sv 5 rep yevenv avrds Géres eLaroreccan, éx tov dn yevens Keivov yévos eLaroAcirat, which at least recalls the mechanism of Ecl. 53-4 (Pam deus Arcadia, etc.). As regards actual verbal similarities, besides the instances quoted by Mayor, I would suggest a comparison of Ecl, 27 sqq. (of the child who now knows to choose righteousness) with Ov. Stb. V. 281 sqq.: evoeBéwv Sé povwv ayin xOuv ravra rad’ oices, vapa pedurtayéos amd mrérpyns 78’ ard rnyns Kai yAdyos auSpdc.ov pevoes mavreroe Sixaious—: of Ecl. 32 (of walled cities the fruit of prisca fraus) with Or. Ssb. VIII. 209 sqq.: yaia & ion ravrwy, ov teixertv, ov Tepippay pois ExT popepi(opevn Kaprrovs TOTE TACiovas Oi~E—: of Ecl. 38, 41 (of the end of seafaring and of ploughing) with Or. Sd. II. 210-11: > ~ aa = , , OV vaus EuopTos ert KUpAaTL TOVTOTOpPHCEL, ov Boes iOuvrnpes dpotpetoovety apovpay—:; or Or. Ssb. 1X. 348: ‘ ~ , ” 9 39 , > w , Kai vavs poprov €xovo” ext Kvpacw ovK ett rAEvoEL, Although these passages do not occur in the pre-Virgilian parts, they may well be echoes of older oracles which Virgil may have read, and the reminiscences are at least sufficiently striking for them to be pointed out, though I do not stress them for my purpose here. It should be noted in passing that the coincidences with Hesiod, noted by all editors, are easily accounted for by the fact that the Sibylline books themselves are full of such imitations, Hesiodic as well as Homeric (see Rzach’s appendix). Further, for the indifferent mythology of the Eclogue, where Apollo and Saturn, Etruscan and Stoic theories of the golden age are inextricably confused, I may perhaps cite the passage in Or. Sib. III. 105 sqq., where references to the Tower of Babel and the Greek myths of Kronos, etc., are similarly interwoven. Lastly, the reference to the prophet himself at the end of the Eclogue (line 53, 0 mths tum longae maneat, etc.) may possibly be parallel with the customary Sibylline allusion to the 1 For this dda imperantis regular in \éya, see in which are quoted instances from oracles and Aristophanes, Eguites 197, and Neil’s note ad/oc., oracular parodies. 104 ROLAND G. AUSTIN prophetess herself at the end of each book or completed part of a book. All these points may in themselves be trifling or even improbable enough, but the cumulative effect suggests tendencies that we can hardly regard as altogether negligible. Now if all this is not mere fancy (and it is easy enough to see resemblances if one is determined to do so), what is the ultimate origin of these peculiarities which appear to me at least to be common to both Virgil and the Sibyl? The solution seems to lie in the special characteristics of Hebrew poetry.1 As one entirely ignorant of Hebrew I write with much diffidence, and am wholly dependent on the recognized authorities of the subject. These Sibylline oracles are generally taken to be the work of first Jewish and then Christian writers. Mayor’s account of them is easily accessible,” and does not need recapitulation here. Without assuming that Virgil had actually read Hebrew, it goes without saying that in the Oracles he found a very strong tinge of Hebrew language and methods, and this will explain not only the verbal similarities to Isaiah upon which all the editors comment (Mr. Royds’ book gives the fullest account) but all these points of versification which I have here noticed. In the first place, the stichic character of Hebrew poetry is well known. ‘In Hebrew, especially, the end of the line uniformly coincides with a break in the sense,’ So says Prof. K. Budde, the author of the article on the subject in Hastings’ Dictionary of the Bible. But this is precisely what Sir W. M. Ramsay remarked in the fourth Eclogue, and what strikes, or should strike, every reader. And the Sibylline books show the same peculiarity to a marked extent. Secondly, as regards rhyme and assonance, this too is a well-defined if subordinate feature of Hebrew poetry, especially the sort of ‘grammatical rhyme’ that is apparent both in the Sibyl and in Virgil. ‘Where the same logical relations govern a series of lines, rhyme must come in with more or less regularity’ (D,B., .c.). Thirdiy, the same article in Hastings, dealing with the Hebrew farallelismus membrorum, states that ‘this parallelism gives to Hebrew poetry in general the impression of aphorisms linked together.’ Is not this precisely what we feel about the fourth Eclogue from its very opening lines? And is not the same atmosphere apparent in the Oracles? But the most fascinating thing of all lies in the matter of verse-formation. The Hebrew parallelismus mem- brorum is briefly a connexion by means of the sense. ‘Individual stichoi, which themselves each form a unit of sense, combine in the verse to form a larger unit (D.B., l.c.). I have tested the fourth Eclogue, with most interesting results. It, too, falls quite naturally into a stanza-formation, each stanza consisting of a small group of lines, each forming its own sense-group. What is more, without unduly straining the divisions, these stanzas appear to be not of haphazard arranging, but of regular alternating lengths. They are as follows: Prelude to the Muses. . Lines 4-7. A new age is to begin on the birth of a child. Lines 8-10. Invocation to Lucina. . Lines 11-14. Address to Pollio. . Lines 15-17. The child shall initiate the heroic age. Lines 18-20. The earth of itself shall change its nature. Lines 21-23. Animals and plants of themselves shall change their nature. . Lines 24-25. (Summary) Poisonous animal and plant alike shall perish. Lines 26-30. The child is advanced to youth. Fresh natural changes. . Lines 31-33. Yet there will be traces of former guilt (general instances given). . Lines 34-36. Particular instances of such evils. . Lines 1-3. a HOD ON AM PW DD Lam nothing more than remark on the stichic nature of the poem and the paratactic arrangement of the lines (Expositor, 1907). 2 Virgil’s Messianic Eclogue, pp. 87 sqq. 1 Mr. Royds himself hints at a similar conclu- sion: see Appendix B of Virgil and Isaiah, p. 115. Sir W. M. Ramsay thought too that Virgil was experimenting in Hebrew metre, but he did 12. | 13. | 14. ] 15. I 16. |] 17. I 18. | 1g. I I think t sented t stanzas. stanzas < single gr (6) an alt arrangen single se; the last : ABCBC. pause wl Virgil ha of proph almost a This which to of their appropria all reader these ma oracles. by sense | pleasant 1 itself, whi derived fr to sO sym a poem or artificial, : itself, My b Oxford, fe here my p the Eclogu but also fc Way, and | UNIVER 1 For oth ‘Umrahmur All these umulative blances if ties which e solution e entirely nt on the y taken to yf them is ming that ; he found n not only ir. Royds’ have here ll known. sak in the Hastings’ narked in And the is regards f Hebrew the Sibyl es, rhyme article in arallelism ther.’ Is ing lines? ascinating mus mem- oi, which rger unit . It, too, 1all group - straining of regular ature. ish. es. es given). chic nature ngement of 7 sqq. VIRGIL AND THE SIBYL 105 12. Lines 37-39. The child is now a man. shall bear all things. 13. Lines 40-41. Agricultural toil shall cease. (Compare Group 6.) 14. Lines 42-45. Fleeces shall not need to be dyedany more. (Compare Group 7.) 15. Lines 46-47. Such were the words of the Fates. 16. Lines 48-52. The time isat hand. Come to the world that awaits thee with joy. 17. Lines 53-54. Would that I may then be alive to see. . 18, Lines 55-59. None of the great singers of myth shall then vanquish me in song. 1g. Lines 60-63. Come, little child, be worthy of thy parents. Seafaring shall end. All the earth I think that it may fairly be claimed for these little stanzas that in each is repre- sented the elaboration of a complete thought, with occasional inter-connexion of stanzas. It will be seen that the line-groups resolve themselves into (1) a group of stanzas of alternately 3, 4, 3, 4 lines each; (2) a group of 3, 3, 3, 2 lines each; (3) a single group of 5 lines; (4) another group of 3, 3, 3, 2; (5) a single group of 4 lines; (6) an alternation of 2, 5, 2, 5-—14 in all, just as in group (1), the other alternating arrangement ; and (7) another single group of 4. This regularity of groups, with the single separate buttress-groups interspersed, cannot be accidental. If we take away the last four lines, we have for the other groups an arrangement in the pattern of ABCBCA—a regular and finished formation. May this not partly account for that pause which editors unanimously feel must be made before those last four lines? Virgil had completed his artistic structure at line 59: he then dropped his character of prophet, and carried away by the reality of his own creation he added what is almost a postscript in the person of Virgil the man.' This, then, is my conclusion. Virgil had before him the Sibylline Oracles on which to model his prophetic poem. These oracles had certain stylistic mannerisms of their own, which he adopted for this particular Eclogue. In this way he wrote in appropriate and recognizable oracular form on a subject whose obscurity has puzzled all readers from the very beginning, a characteristic of every good oracle. Further, these mannerisms were the inevitable result of the Jewish origin of many of the oracles. Virgil either noticed for himself in them the method of grouping the lines by sense (which is fairly well defined in the Sibylline books also), or, as it would be pleasant to think, he had actually some knowledge of the principles of Hebrew poetry itself, which, without necessarily being able to read Hebrew, he could easily have derived from some learned Jew eager to communicate all the enthusiasms of his race to so sympathetic a listener. Then, with his own incomparable artistry, he produced a poem on structurally similar methods, a work of poise and balance and rhythm— artificial, no doubt, but full of that special kind of art whose essence it is to conceal itself. My best thanks are due to my friend Mr. C. J, Fordyce, Fellow of Jesus College, Oxford, for much helpful and suggestive criticism ; and I should like also to record here my profound debt to the late Professor Phillimore, not only for the interest in the Eclogues and indeed in Virgil’s poetry as a whole with which he first inspired me, but also for the countless personal kindnesses and help which he gave me in every way, and which he knew so well how to give to all. | Roitanp G. AUSTIN. UNIVERSITY OF GLASGOW. 1 For other speculations as to the metrical ‘Umrahmung’ of the poem, see C, Brakman in Mnemosyne 1926 (Part i). H SUMMARIES OF PERIODICALS LITERATURE AND GENERAL. American Journal of Philology. XLVII. 3. 1926. M. Bloomfield, On Ovganized Brigandage in Hindu Fiction. Illustrates the distinction between organized ‘ brigands’ and unorganized ‘thieves’ by a detailed account of the habits, customs, methods, and characteristics of the Bhillas, G. L. Hendrickson, Cicero’s Correspondence with Brutus and Calvus on Oratorscal Style. Defends the genuineness of this correspondence, known to us by references in Quintilian and Tacitus, and seeks to show (1) that it arose from certain hostile criticisms passed on the De Ovatore, (2) that its general subject was the use and abuse of ‘ rhythm’ and the periodic structure, and (3) that it represents a preliminary debate, out of which grew both the Brutus and the Oratory. R. G. Nisbet, Justae quibus est Mezentius Irae. Argues, despite the authority of Macrobius, Amsianus Messius, and (possibly) Aemilius Asper, that this famous line contains not an anomalous predicative dative, but a simple genitive of quality. Adds a list of phrases from various authors containing ablatives or genitives of quality which express a meaning closely akin to that of the predicative dative. E. M. Steuart, Some Notes on Roman Tragedy. Suggests that those Roman tragedies without a known or accepted Greek ‘ original’ may represent (1) direct inspiration from Homer (e.g. The Andromache of Ennius), or (2) a ‘complementary’ principle by which minor personages were grouped around the great figures (e.g. The Zecmessa of Julius Caesar Strabo), or (3) deliberate ‘ selection,’ by which particular portions of familiar stories were chosen for treatment (e.g. The Jliona of Pacuvius). Athenaeum. III. 2-3. P. Fraccaro on G. Gracchus maintains that of the two chief sources for the years of his tribunate, Appian and Plutarch, Appian is the surer guide, not only for the tribunate of B.c. 133, but also for the events of 123-121. Appian’s first book isa condensed account of the period 133-70, and serves as an introduction, this arrange- ment going back to Appian’s source itself, and being made with skill and an eye to essentials. The events of three years are crowded into a few pages. F. then examines the reason for the selections made, and the logic of their sequence. He shows Appian’s substantial accuracy as to the passing of the ‘lex frumentaria,’ the judiciary laws, the franchise laws (‘ rogatio de sociis’), and the results of all these on the popularity of Gracchus. Plutarch, in his biography, shows a good deal of con- fusion, perhaps due to his sources, but some due to the method of enumerating the laws, interpolating anecdote and fiction, and then returning to the laws. This has in several instances caused subsequent writers to treat the second mention as a record of new laws. Vol. III. 3 (continued). He deals with the efforts of Gracchus to secure citizenship for the allies, the opposition of Senate, the alienation of Fannius, and the veto of Livius Drusus. Here Plutarch gives valuable information, but confuses the sequence of events, and there are serious gaps in the story of Gracchus’ absence in Africa and his return to Rome. Appian and Plutarch vary in their accounts of the events in the ‘area Capitolina,’ and of the exact share of respon- sibility to be charged to Gracchus in the murder of Antullius. Plutarch as a Philogracch underlines all the details which may tend to clear Gracchus of the actual murder. Plutarc and has traditio sources, Livy’s | whether hoc ut) ut,’ but IV, | belongs loses its IV. A, suggests husband. But A. p of her vc the home IV. & 2 identifica incidenta. works of. Mne: 5. F. ‘families accidental rectors: ¢ adds criti Vana, he may be aj of ruber, , Fibra gen evidence t it were c1 P. would | nauclevus. Seneca, in the death: occurred a a contemr corded, F Thyestes, D Tacitus, by a man of the year 7 Regulus. ates the detailed Bhillas, cal Style. ences in n hostile use and liminary et, Justae \msianus ; not an a list of ty which , Steuart, rithout a n Homer ch minor is Caesar ar stories s for the t only for book is a arrange- in eye to F. then nce. He faria,’ the these on il of con- ating the 1is has in a record cchus to Fannius, ition, but zracchus’ yin their f respon- rch as a he actual SUMMARIES OF PERIODICALS 107 murder-charge, and he or his sources do some violence here and there to chronology. Plutarch lays the blame chiefly on Flaccus. Appian’s account is more consistent and has often the support of other authorities. He gives us the serious and solid tradition, while Plutarch’s biography is full of alterations and reserves. The other sources, though miserably scanty, show less signs of tampering, doubtless because of Livy’s honesty. Some of these contain fragments of some value. C. Pascal queries whether the genuinely classical, though rare, use of credo ut (or as in Martial credo hoc ut) may not be a colloquial usage, perhaps derived from the more normal ‘ credibile ut,’ but possibly also having given rise to the latter. IV, 1. C. Pascal supports his suggestion that the name of ‘The Latin Menander’ belongs to Terence, by Macrobius VI. 4. Caesar’s phrase ‘dimidiate Menander’ loses its force unless Terence was commonly known as the Latin Menander. IV. 2. A, Annaratone on the epitaph published in Notizet degli Scavi, 1923, p. 357; suggests that it is not a duologue, but the lament of a wife who has died before her husband. ‘ Mater sanctissima’ has been referred to Cybele, or to the ‘magna mater.’ But A. prefers to refer it to the high-priestess of Cybele, the speaker having been one of her votaries. The allusions to her husband’s ‘ pristina cura,’ and the allusion to the homestead points to the couple having been /iberti. IV. 3. C. Zacchetti asserts the ‘prior rights’ of Denys Lambin (born 1516) to the identification of M. Damasippus, of Horatian satire, with the D. of Cicero’s letters, incidentally quoting fully passages to show that D. was known to both as a buyer of works of art, houses, and estates. Mnemosyne. LIV. 4. 1926. J. P. Postgate, Ad Caecilii Plint Secund: Epistulas, points out that there are two ‘families’ of MSS. of Pliny’s letters. The differences between them are sometimes accidental, sometimes the result of the ill-timed labours of ancient editors or cor- rectors: often it is difficult to know which preserves the preferable reading: He adds critical notes on a large number of passages. In a second contribution, entitled Vana, he brings further evidence in support of Vollgraff’s contention that xvAcvépos may be applied to round objects which can be rolled. He discusses the quantities of ruber, fibva, ciniflo. In ruber the u is generally long, but short in Lucr. 4. 42. Fibra generally, fibya Manilius 1. 92, and three times in the Hercules Oetaews—an evidence that it is not Seneca’s work. Cuzintflones was pronounced cini-flones, not as if it were cinif-lones. In Lucr. V. 1006 ‘improba navigit ratio tum caeca iacebat,’ P, would read with Rouse navivegt, which, however, he takes as gen. of navivegus = naucleyus. P. J. Enk, De Octavia Praetexta. The author of the tragedy cannot be Seneca, in view of the prophecies which the play contains of the suicide of Nero and the deaths of Poppaea and Crispinus before Nero, all of which events actually occurred after the death of Seneca. The author, however, imitated Seneca: he was a contemporary of Nero, as he relates incidentally certain details elsewhere unre- corded. Praetextae were written shortly after Nero’s time—e.g. Maternus’ Cato, Thyestes, Domitius. H. Wagenvoort, De Regul in Tactti Dialogo Partsbus, argues that Tacitus, by representing Messalla, who was a brother of Regulus the informer, but a man of very different character, as taking part in a dialogue on eloquence in the year 78, and criticizing the prevailing style, was making a veiled attack on Regulus. Men of like character, Eprius, Marcellus and Vibius Crispus, are openly 108 SUMMARIES OF PERIODICALS censured, but the name of Regulus is left in silence, a proof that R. was still alive. Hence the dialogue was written between the years 93, when Vibius Crispus died, and 106; but in view of the reference in Pliny IX. 10 to poemata ... quae tu inter nemora et lucos commodissime perfict putas (cf. Dal. g-10), W. thinks the date may be fixed to 105 or the beginning of 106. H.J. Rose, De Virginibus Vestalibus, thinks the Vestals originally were the daughters of the King. Their attire does not indicate, as has been supposed, that they were symbolically married, but that they were engaged in a religious rite: the bride wore a similar dress because she also was engaged in a religious rite. Originally the period of office (involving celibacy) was as short as five years (Dion. Hal. Ant. Rom. I. 76, 3). Afterwards it was lengthened to thirty years. Celibacy in matters of ritual and magic was of great importance, and the celibacy of a maiden of marriageable age was of more ritual value than that of a younger girl. Neue Jahrbucher fiir Wissenschaft und Jugendbildung. 1. 6. 1925. Fr. Boértzler, Das wahve Gesicht des Sokvates. Socrates’ insistence on ignorance was meant to clear the ground for such constructive systems as his followers produced. F. Kuntze (t+), Ez Plautinisches Lustspielmotiv in der Weltittevatur. Dhis- cusses the motive of the pierced party-wall in the Miles gloriosus, and its connexion with similar Eastern and medieval European stories. R. Petsch, Etim Mosedvama aus hellenistischey Zeit. Discusses, after Kappelmacher, the Exagoge of the Jew E-zechiel, a Hellenistic play dealing with the Exodus, partly preserved by Alexander Polyhistor, through Eusebius : II. 1-5. 1926. 1. C. Watzinger, Die gviechische Hevoenzeit und Homer. A severe criticism of Dorpfeld’s recent heresies: the Phoenician-Sabaean origin of Cretan art, and the shortening of early vase-chronology. C.Clemen, Der gemeinindogermanische Totenkult. An attempt to show that certain practices with regard to the dead are a common heritage of the Indo-Germanic stock. 2. E. Groag, Neuere Literatur tiber Caesay und Augustus. Chiefly a review of E. Meyer’s Caesars Monarchte and H. Dessau’s Geschichte dev vimischen Kaiserzett. G. Neckel, Regnatoy omnium deus. Identifies the mysterious god of Tacitus Germ, 39 with Wodan, and discusses the origin and diffusion of the Wodan cult, and various problems of German and Scandinavian ethnology and culture. J. Geffcken, Kingsley’s ‘Hypatia’ und thr geschachtlicher Hintergrund, Criticizes much detail, but praises highly the general truth of the picture. O. Schiitz, Eine gvriechische Lauttafel. A reasoned plea for the scientific reform of the pronunciation of Greek in universities and schools. 3. M. Pohlenz, Stoa und Semitismus. Finds many detailed traces of Semitic influence in Stoicism (e.g. in grammatical classification and in the importance attached to divination), and also in their eiyappévn, and in their conception of moral duty. W. Schur, Dive ovientalische Frage im vémischen Reiche. Sketches the history of the Eastern question down to the Arab conquest. Pompey’s Euphrates frontier was a_ historical accident, but it led to a cultural interplay between East and West of inestimable importance for both. 4. F. Studniczka, Von dey griechischen Plastik des V. Jahrh. v. Chr. (Mit 22 Abbildungen auf 10 Tafeln). A severely critical examination of the theories of H. Schrader’s Phidias. 5. W. Schubart, Hellenismus und Weltreligion. An attempt to distinguish Eastern and Western influences in Hellenistic and early Imperial religious thought, ending in a discussion of the causes of the triumph of Christianity. E.Salzmann, Kaiser Hadrian und das Problem seiner Personlichkeit. A sympathetic study of Hadrian in relation to his times. Phil =o Parallels Anstotle’s Phil R. F regard A represent the actua Savpoviov G. Herb The coin textbook. other so biichern ¢ Odes IV. and sum Excursu: Misi Rothstei: Lesbia p Propertit IV. 6, 3. Suggests example: the signi F. Kred Mulomedi LX: K. v Ideenlehve is genera probably absence, article, 1 (Leipzig, (especial. Philitas | II. Date Wilamoy 280-75 B. on coins’ For Ber Fr. 244 | marriage graduall: was not | firmed b also wro: 245 B.C. G. Meye ill alive. died, and tu inter > may be 1inks the licate, as engaged ged ina short as to thirty and the 1 that of 1925. znorance followers iy. Dis- onnexion losedvama the Jew lexander ‘icism of and the Totenkult. common wesay und Dessau’s tifies the igin and dinavian achtlicher n of the scientific Pohlenz, Stoicism on), and hur, Die question listorical stimable ’. Jahvh. n of the tveligion. nd early umph of Ret. A SUMMARIES OF PERIODICALS 10g Philological Quarterly (Iowa). V. 3. 1926. R. B. Steele, Variation in the Latin Dactylic Hexameter. A. S. Cook, Greek Parallels to Certain Features of the Beowulf. M. T. Herrick, The Early History of Avistotle’s Rhetoric in England. Philologus. LXXXI. 4. 1926. R. Frese, Dte ‘ avistophantsche Anklage’ in Platons Apologie. Plato did not really regard Aristophanes as responsible for Socrates’ fate, but uses him as convenient representative of the nameless accusers, with whose charges he unfairly confounds the actual indictment brought by Meletus, etc. This latter was concerned with the da:yoviov and corruption of the young, but Plato construes it as a charge of atheism. G. Herbolzheimer, Cicevos rhetorict libvt und die Lehyschrift des Auctoy ad Hevennium. The coincidences in the two treatises derive from a common source—viz., a Latin textbook. The Auctor ad H. followed this closely, Cicero used it to supplement his other source—viz., the avs of Hermagoras. W. Port, Die Anovdnung in Gedicht- biichern augustetschey Zeit. In continuation of previous article examines Horace, Odes IV., Tibullus and the other poets of the Corpus Tibullianum, Ovid’s Amores, and sums up results, concluding with conjectures about the Hellenistic practice. Excursus on the date of publication of Tibullus, Book I., which is fixed at 25-24 B.c. MIscELLANEOuS.—R. Foerster: The Zittau MS. of Plato has no value. M. Rothstein: The Caelius of Catullus’ 58 and 100 is mot M. Caelius Rufus. Catullus’ Lesbia poems belong to post-Bithynian period. M. Schuster: Defends the MSS. at Propertius III. 18, 31 sqq. (animae tuae=‘breezes that impel Charon’s boat’); IV. 6, 3; IV. 11, 37 sqq.; in II. 29, 21 suggests miserunt=‘released.’ S. Eitrem: Suggests that in Plato, Epp. 344a, Lynceus is name of eye-salve, and cites other examples of such nomenclature in Greek and Latin. C. Fensterbusch: Examines the significance of entry from the right and left in the Greek and Roman theatres. F. Kredel: Discusses the system of proportions in horses given in Vegetius, Mulomedicina Chivonis III, 1. LXXXII. 1. 1926. K. v. Fritz, Die Ideenlehve des Eudoxos von Kuidos und thy Verhdltnis zuv platonischen Ideenlehve. E.’s rather crude theory, which is further from Anaxagoras’ views than is generally supposed, must have been put out before Plato developed the doctrine, probably about 367-5 B.c., when E. was acting head of the Academy in Plato’s absence, and Aristotle a young student there. R. Herzog, Hevondea. An important article, which supplements the author’s revision of Crusius’ edition of Herondas (Leipzig, 1926): I. Textual and exegetical notes, particularly on Mimes VII. (especially on currency-problem) and VIII. [Herzog still identifies the yépwv as Philitas and suggests that Philitas, Fr. 17 (Powell), refers to a bucolic masquerade]. II. Dates of composition. Herondas wrote under Philadelphus, not Euergetes (as Wilamowitz still declares!), compare especially Mime IV., which was written 280-75 B.c., and Mime I., written soon after 270 B.c. The Oecot adeAdoi of I. 30 (as on coins) are the two pairs—viz., Soter and Berenice, Philadelphus and Arsinoe. For Berenice as half-sister of Soter compare = on Theocritus, 17, 34 and Alexis, Fr. 244 (Kock) ; the latter may be dated about 290 B.c. Philadelphus covered his marriage with a full sister by establishing the combined cult. Ordinary usage gradually separated the two pairs, but an independent priesthood of the Oeoi cwrjpes was not established till 216-5 B.c. by Philopator. This dating of the Mimes is con- firmed by Herondas’ relations with Theocritus, Callimachus, etc. Wilamowitz is also wrong about the date of Callimachus’ elegy for Sosibius, which is certainly about 245 B.c. III. Author criticizes Knox’s views about Attic connexions of Herondas. G. Meyer, Zu Minucius Felix und Tertullian. Refutation of Gudeman’s article in IIO SUMMARIES OF PERIODICALS Nos. 1-4 of Philolog. Wochenschr. 44, 1924, 90-2. J. Balogh, Voces Paginarum. First part of article reviewing the evidence for reading aloud as the usual method of antiquity. MIscELLANEOUS.—W. Dorpfeld: Severe criticism of Birchner’s article on Leucas Leucadia in Pauly-Wissowa. F. Stahlin: Attacks the view that the battle of Pharsalus was fought on the right bank of the Enipeus. J. Moor: Points out that Plut. Marius, 14, recalls Xen. An. II. 6. Supplement Band XVIII., Heft 2. 1926. Dr. Kurt von Fritz, Quellen-untersuchungen zu Leben und Philosophie des Diogenes von Sinope. Pp. 97. Revue de Philologie. L. 2-4. 1926. B. Haussouillier, Inscriptions de Didymes (concluded). E, Gallettier, A propos du Catalepton. Holds the view that many of the poems are not by Virgil, but may have been discovered among his papers after his death. F. Butavand, L’Odyssée dans le Texte étrusque de la Momie d Agvam II. W. Deonna, Les Poémes figurés. Thinks these may have been symbolic or magical in origin. M. Holleaux, Rome et la Gréece au IIJe Siecle. A continuation of a criticism of Walek. R. Waltz, Ego et Nos. Nos can represent ego, nos ambo, nos omnes, and homines omnes. G. Mathieu, Le Myrte des Tyrnnnicides. The myrtle was a branch and not acrown. G. L. de la Vida, La Lettre suv la Vautour. Illustrations of the article ‘ Bothvos ou Avétas’ in L. 1 from Oriental sources. L. Laurand illustrates Cic. Pvo Muvena g. 22 te gallorum, illum bucinavum e.g s. by Polyb. XII. 26. 1 and g. Parts 2 and 3 contain summaries of classical journals and of reviews of classical works published during 1925. Rheinisches Museum fiir Philologie. LXXV.2. 1926. M. Bose, Das dlteste Catozitat. ‘The oldest express quotation of the Disticha Catonts (in the Epsistola Vindiciant comitis avchiatrorum ad Valentinianum tmpevatorem) gives exigua for auxilium, by confusion on the writer’s part with II. 9, in the sense ‘the maladies of.’ Exigud, with a lengthened im arst, has parallels in the Disticha. O. Viedebannt, Warum hat Seneca die Apokolokyntosis geschrieben? Seneca’s official position in 54 would not permit of his writing such a tract out of personal animosity. His political aim is to glorify Nero’s régime by vilifying his predecessor and dis- crediting the apotheosis which Agrippina had secured for him. The funeral of Claudius must have preceded his consecration. Seneca’s laudatio funebris was the necessary expression of the continuity of Nero’s rule with that of Claudius. R. Holland, Battos, On the story of the untrustworthy shepherd Battos, in Antoninus Liberalis Fab. 23. This Greek treatment, its modification by Ovid, and the different burlesque version in which Hermes is a child (Hom. H. III. and [Apollodor.]} ITI. 112-5: in 114 H. suggests das for Boas) are exhaustively discussed by H., who thinks that the version in the Eoae is later than the Homeric Hymn and in conscious opposition to it. The Barrov cxori. in Arcadia is so-called on the analogy of a Barrov oxomia in Cyrenaica. W. Bannier, Zu den Beschliissen I.G. 1.2 91/92 (= Ditt. Syll$ QI). dmodovva: means here ‘pay moneys due to’ not ‘repay moneys borrowed from.’ erevoy) . . . avevijveyaras Means ‘now. . . have been paid in,’ and Anon. Arg. is to be restored accordingly. 3B. proposes a restoration of some of the mutilated text on the back of the first text, and argues that it is not a continuation. The front probably belongs to Ol. 87. 2, the back to the end of Ol. 87 or the beginning of O/. 88. A. Busse, Der Wortsinn von Adyos bet Hevakltt. tov S€ Aoyou tovd’ éovros is the opening of the treatise, which had no title other than the author’s name, and means ‘of the doctrine here announced, ae goes with afvvero: yivovrar, B. gives a valuable analysis of the me (D). Re and a cc Protagora: C. Mras, in 197. shows it Hackert’: Misz the fact 1 and discu A. Klotz, LX2 A. ] Columbu its size r revival o: Geograph regions. English | also his theory th rejected. possibly I, Od. 4: is the poe Iliad, L I. In§4_ as were th difficulty gained fi periegetic accept H our MSS are most understat Kunitische | VI. 14; LX?) E. NV gloss on " Max. Tyr Kunstlehr 399-407, | ascribed ance witk An exhai W. Enss. A new te 127, with . First sthod of ticle on yattle of out that genes von bvopos du ay have > dans le Thinks la Greéce et Nos. ,¢ Myrte ‘ida, La . I from m, illum classical Disticha vatovem) e sense Misticha. official imosity. ind dis- 1eral of was the us. R. itoninus lifferent wr. ] III. 1., who scious 1 Barrov t. Syll8 d from.’ is to be t on the robably 88. A. opening ‘of the analysis SUMMARIES OF PERIODICALS III of the meanings of Adyos. H. Oppermann, Plotin-Handschniften I. Codex Darmstadiensts (D). Results of a new collation of this manuscript, dated between 1470 and 1492, and a copy of A. P. Keseling, Justins, Dialog gegen Trypho (c. 1-10) und Platons Protagovas. Argues for direct literary dependence. W.Uhdet, Zu Plutarchs Moralza. C. Mras, De Culice Vergilit. In 198 read tavdo somni, keeping nescius in 199 and cristam in 197. In 243 read with Voss stmul ite; in 364 keep denotum bellis. The poem shows its author’s youth. F. Marx, Vetus Italia in Italia nova. A parallel from Hackert’s visit to Sicily for that of Lucilius. MiszeLuan. Th. Birt, Uber vas argenteis und Verwandtes. B. draws attention to the fact that he anticipated the thesis put forward by Klotz, Ri.M. LXXV. 98 sqq., and discusses illustrations, Th. Birt, Nachtvag zu Catull.,c.17. Postscript to p. 115 sq. A. Klotz, Berichtigung. LXXV. 3. 1926. A. Eltert, Das Altertum und die Entdeckung Amernthas. The expedition of Columbus was inspired by the idea that the earth is spherical and by an estimate of its size made by the Greeks, and brought to light again by the fifteenth century revival of interest in the Geography of Ptolemy. R. Hennig, Neue Evkenninisse zur Geographie Homers. Od. 10. 82-86 and 11. 13-19 do not show knowledge of Polar regions. The Cimmerians are to be placed in Cornwall; the Styx is perhaps the English Channel. Dérpfeld’s localization of Phaeacia in Corfu is untenable, as is also his identification of Ogygia with Cape Santa Maria di Leuca. Herrmann’s theory that the former is Tunis, the latter a small island off Chios is also to be rejected. Scheria is Tartessos, Ogygia one of the Western Canary islands or possibly Madeira. Plato’s Atlantis is also Tartessos. C. Fries, Homerica II. 1, Od. 4 is an accurate representation of conditions at Sparta. 2. The pvnornpodovia is the poetry of popular indignation with the aristocrats and is in contrast with the Iliad. L. Weber, Pausanias’ Beschrethung des Kevramethos-Friedhofes. Paus. I. 29. I. In § 4 keep the reading pvjya as meaning state-cemetery and explain rpora erapynoav as weve the first to be buried (instead of making zporo: local). II. The chronological difficulty of rporo.. An accurate picture of the arrangement of the graves cannot be gained from P. III. Individual grave-monuments and their dates. IV. The periegetic source of P. E. Bickel, Dey Sirius bet Manilius. In Manilius I. 409 accept Huet’s wentve for uultu; in 407 read wirtusque for cursusque. The archetype of our MSS. often has corruptions which give a sense; the Lipsiensis and Gemblacensis are most interpolated. In 398 read nunc torrida prima resurgit; in 399 keep sols, understanding wacuum in an absolute sense; in 414 read aucta for acta. M. Schuster, Kritische und evklivende Beitvige zu Martial. Notes on spect. 13; III]. 13, 26; IV. 21; VI. 14; VII. 41; XI. 87; XII. 34; XIII. 69; XIV. 158. LXXV. 4. 1926. E. Maass, Psaphon und Sappho. After notes on the termination -ayos and on the gloss on Wada in Hesychius, M. discusses Védwv-Lardw and the story of Psaphon in Max. Tyr. XXIX. 4, which he localizes in Provence. O. Schlissel, Die rhetorische Kunstlehve des Rufus von Pevinth. The réxvyn edited by C. Hammer, Fiet. gr. I. 2, 399-407, is a tour de force in brevity, of the end of the second century A.pD., and may be ascribed to Rufus of Perinthos. An analysis shows the author’s extensive acquaint- ance with rhetorical theory. E. Peterson, Engel- und Damonennamen Nomina barbara. An exhaustive catalogue raisonné of these names in magical and astrological texts. W. Ensslin, Ein Prozessvergleich unter Klevikern vom Jahve 481 (Zu Papyrus Princeton). A new text of this document, first published by H. B. Dewing, 7.A4.P.A. LIII. 113- 127, with full comments on points of legal and economic interest. 112 SUMMARIES OF PERIODICALS MiszeELLe.—E,. Schwyzer, Zu Kallimachos (Epigr. 41 . 44) und zu Lykophvron (vs. 1391). In Call. 41. 5 read ovxi cvveip’”"Hpwv; in 44 keep 6 otyapyys (accent un- certain) as a new word meaning os giya dpyvrac; in Lyc. 1391 read Anxnrpia for Anxrnpia, as an epithet derived from Aaxnrjp, the S.W. promontory of Cos, LANGUAGE. Indogermanische Forschungen. XLIV. 2. 1926. W. Brandenstein summarizes and criticizes recent theories of the dependent sentence. A. Debrunner sees in Kvda@jvacov an analogical extension (cf. ’A@nvaios) of *Kvéa@nvos from Kvdafov (with a ‘pre-Greek’ suffix -afov), IF. Muller explains I. Eu. *oket5u- (dual) ‘eight’ as ‘two sets of (four finger-)points (of the two hands),’ with the -to- formant as in *hm-tom, root *ak., *ok- “sharp, pointed’; and ofuétudr, whence g#etudr- ‘four,’ from oket(o)-udvo- (cf. Skt. vara- ‘series’) as ‘a set of (four finger-)points (of one hand). For counting with the fingers, cf. (e.g.) Gr. rewrdgev. A. Hillebrandt assembles passages in which Vedic dhaman occurs to justify his previous account of its meaning (1. Statte, Sitz. 2. Schar=das was diese Statten enthalten). A. Debrunner quotes from a ‘ Kirchenbuch’ of Miinchenstein (near Basle) an entry dated 1727 with the form movixit (‘mortuus est’), evidently a jingle on uixit, and suggests that O.H.G. morixlen ‘die’ had its origin in such a riming jingle. M. Runes, virgo from *uirigo (: uiridis, *uivo: utveo like feruo: feruco), cf. ovigo: ovioy. J. Whatmough, Lat. Jarix, shows that the word was known in Latin before the importation of larch timber from the Alps by Tiberius (Lariscolus, 43 B.c.). F. Slotty writes a long account of the singular use of the first person plural in German (apparently independent of Conway’s study of the use in Latin). F. Holt- hausen cites (from Lidén) Armenian cognates to some fifteen Teutonic words. Jd. Mid. Low Germ. vitke f. ‘net for hunting, fish basket, petticoat’: Skt. piga-s, G. vy}, etc., the root idea being that of rounding. Jd, Parallels in Low Germ. to Eng. ‘ half seas over.’ A. Mayer finds the name and site of ancient Acruvium in mod, Grbalj. Book reviews. ykophron cent un- ntpia for ependent A @nvaios) explains » hands),’ okuétuor, of (four reptracev, stify his > Statten ein (near a jingle a riming cf, ovigo: in before 43 B.C.). plural in F. Holt- yrds. Id. . puga-s, Germ. to t in mod,