THE CLASSICAL QUARTERLY JULY, 1914. NOTES ON HORACE. ODES, I. 3, I sqq.: Sic te diua potens Cypri, sic fratres Helenae, lucida sidera, uentorumque regat pater, obstrictis aliis praeter Iapyga, nauis, quae tibi creditum debes Vergilium: finibus Atticis reddas incolumem, precor, et serues animae dimidium meae. The sic of this passage is ordinarily taken as meaning, ‘on this condition,’ viz. the condition implied in reddas and serues. But du Mesnil (Bremer Rundschau, 1885, No. 258) urged that this interpretation was illogical. The fulfilment of the condition implied in veddas involves in itself the realization of the wish expressed in regat, and so makes that wish unnecessary. To this objection two answers have been made. Schiitz expresses the opinion that the prayer is for the perpetual enjoyment of the favourable conditions enumerated in verses I-4. But apart from the unnaturalness of this explanation, it involves the north-west wind as the perpetual attendant of the vessel. This would prevent its easy return to Italy. The other answer is that such ‘ matter of fact’ criticism should not be applied to poetry. But to the minds of many (myself among them) this seems a mere evasion of the question at issue. Most students of Horace, I venture to say, are far from satisfied with the traditional interpretation of the passage. Doederlein,! feeling the need of a clause of result after sc, proposed reading ut after Vergilium ; but this, while, of course, metrically possible, can hardly be considered the probable text, since result clauses with a parenthetic precor do not elsewhere occur. In my own edition of Horace (1gor) I sug- 1 Earle, Classical Papers, pp. 185sq.(=Proceed- has recently endorsed Doederlein’s proposal. ings of the Amer. Phil, Assoc, xxiv. pp. Xxil. sqq.), NO. III. VOL. VIII. L 146 CHARLES E. BENNETT gested that this idea of result might be secured in another way, viz. by regarding veddas as a jussive subjunctive explanatory of szc—‘ guide thee thus, viz. bring him safe to Attic shores!’! This interpretation is in perfect keeping with the sense of the passage. It also receives support from two other passages: Cato, De Agricultura, 70, 2, ‘per triduom de ea potione uni cuique boui dato. Ita diuidito, quom ter uni cuique dederis, omnem absumas,’ ‘distribute it in such fashion, (that) when you have given it three times to each, you shall use it all up’; Martial, ix. 99, I sqq.: Sic in gramine floreo reclinis, qua gemmantibus hinc et inde riuis curua calculus excitatur unda, exclusis procul omnibus molestis, pertundas glaciem triente nigro, frontem sutilibus ruber coronis ; sic uni tibi sit puer cinaedus et castissima pruriat puella: infamem nimio calore Cypron obserues, moneo precorque, Flacce ! The latter passage is precisely like the Horatian. The sic is explained by the following obserues—‘I hope you will enjoy the delights of Cyprus, bearing this in mind: avoid the dangers of the heated season!’ Wherever sic expresses a condition, the corresponding subjunctive or imperative invariably contains a request for some favour to be extended to the person speaking, in return for which the blessing contained in the stc-clause is entreated. There is a ‘quid pro quo.’ That any such notion is present in the Martial passage seems to me impossible. It may further be noted that the Horace and Martial passages are alike in that they both contain a parenthetic precor. I. 12, 45 sq.: crescit occulto uelut arbor aeuo fama Marcelli. Page renders this, ‘ Marcellus’s fame grows by the silent lapse of time.’ That this conception would have been consistent with the thought which the poet here wished to convey may be freely admitted, but it is impossible that occulto aeuo should mean, ‘by the silent lapse of time.’ Occulto? aeuo could mean only ‘secret time,’ ‘hidden time,’ ‘concealed time,’ neither of which yields a satisfactory sense. The passage is really one of great difficulty, as is indicated by the fact that it has given rise to no fewer than three conjectures, while those who accept the traditional text vary widely in their interpretation fact that several modern editors have assumed a meaning to fit their conceptions of the exigencies 1 Earle, op. cit. p. 18€, calls attention to the fact that my explanation was proposed by Hunter in his edition of Horaceasearlyas 1795. I have never seen a copy Of this work. 2 There are no other meanings of occultus. The of Latin usage. of this particular passage cannot alter the facts of the (mani the v variet expre by Pe best k victio. editio Lucre Ovid Epp. | retain intere: shoulc falsely to kee have ; stanti’ events ‘ grow meets of the I. B E pileg the re Bentle being author urged import accour Lingue after t! ut in n true, d than ] us, ing her que as, > to ined rus, Tr Sic ably g, In here sage rtial + ime. h the that could yhich as 1s ‘ures, ation imed a yencies ie facts ‘ NOTES ON HORACE 147 of the words occulto aewo. Some take these words as dative ; others as ablative (manner, separation, quality, time). It would be a gratuitous task to discuss the various renderings suggested for this phrase. Their very number and variety are a sufficient indication of the difficulty and uncertainty of the expression. Nor is help gained by Heinsius’s conjecture of aruo for aeuo, or by Peerlkamp’s Marcellis for Marcelli. Nevertheless, conjecture furnishes our best hope of securing an acceptable meaning for the passage. My own con- viction is that Horace here wrote occulte, and I have so printed in my own edition of 1901. The adverb occulte, ‘imperceptibly,’ occurs with decrescit in Lucretius I. 314: ‘ferreus occulte decrescit uomer in aruis’; with labitur in Ovid Met. x. 519: ‘labitur occulte fallitque uolatilis aetas.’ In Horace, Epp. I. 1, 80, we have ‘multis occulto crescit res faenore,’ where occulto, if retained, must go as adverb with crescit. The notion of‘ secret’ or ‘ hidden’ interest can hardly be taken seriously. Here also I should read occulte, and should attribute the error in both cases to the confusion of some copyist who falsely associated the word with aeuo and faenore. Another solution would be to keep occulto, the archaic form of the adverb; but surely the poet would have avoided the possibility of a false association of the word with the sub- stantive in each sentence, and would consequently have chosen occulte. At all events I desire to urge that the adverb be here recognized, since the meaning, ‘grows imperceptibly,’ ‘grows gradually,’ is the only one that adequately meets the demands of the context, and at the same agrees with the usage of the Latin language. I. 16, 5 sqq.: non Dindymene, non adytis quatit mentem sacerdotum incola Pythius non Liber aeque, non acuta sic geminant Corybantes aera, tristes ut irae. Barring 7’, all MSS. have the reading sic in line 8 of this ode, and Keller, E pilegomena, p. 66, is clearly right when he asserts that sc was undoubtedly the reading of the archetype of our extant MSS. of Horace. Nevertheless, Bentley’s conjecture sz (the reading also of 7’) has enjoyed no little favour, being read by Miiller, Kiessling, and Vollmer, not to mention editors of less authority. I shall not attempt to enumerate the telling arguments already urged against the acceptance of this change, but desire to call attention to an important consideration which, so far as I am aware, has never been taken into account. The Thesaurus Linguae Latinae shows no instance of aeque ut before Pliny the Elder and even after that time the usage is extremely rare, the variant et or ac alternating with ut in many of the passages for which some MSS. give ut. The Thesaurus, it is true, does seem to suggest the occurrence of three instances of aeque ut earlier than Pliny the Elder; but none of these will bear scrutiny. The three If we read si, aeque is then correlative with wt. 148 CHARLES E. BENNETT instances are: Varro, Sent. 26; Cic. Phil. ii. 94; ‘ quis cuiquam inimicior quam Deiotaro Caesar? aeque atque huic ordini, ut equestri, ut Massiliensibus, ut omnibus quibus rem publicam populi Romani caram esse sentiebat’; Seneca Rhetor, Cont. Exc. 4,8; ‘aeque dignus est poena qui ipse uim adhibet et qui ab alio admota ad lucrum suum utitur.’ No one of these citations, I repeat, will bear scrutiny. The Sententtae of Varro in the form in which they have descended to us are admittedly late. The Cicero passage is grossly misinter- preted. In this, wt does not go with aeque, but with an inimicus understood ; the correlative of aeque is atque; ut is here equivalent to stcut—‘ as hostile as he was to the Senate, just as he was to all who really loved the Commonwealth.’ In the passage from Seneca Rhetor, Kiessling’s text gives, not aeque dignus ut, but aeguus et, and this without any critical variant. If, therefore, Horace be credited with the employment of the locution aeque ut, it will be the first instance in Latin literature. It will also be the only instance in any poet, and (granting the soundness of the other instances cited in the Thesaurus) one of eight instances, at most, in the entire literature down to the time of Lactantius. The other seven instances given in the Thesaurus are attributed to Pliny the Elder, Pliny the Younger, Fronto (2), Gaius, Ulpian, Apuleius. That a search- ing examination of the textual authenticity of these passages might completely undermine the foundation of the locution aeque ut is not impossible. In the Apuleius passage (A pol. 14), aeque ac is apparently the better attested reading ; only one MS. has wi. At all events, the attribution of the usage to Horace seems, under the circumstances, most hazardous. II. 18, 7: nec Laconicas mihi trahunt honestae purpuras clientae. On this passage, Kiessling’ observes: ‘trahunt purpuras wird jetzt in der Regel von dem Tragen purpurner Schleppgewander erklart. Weshalb aber Horaz, um den Reichen zu charakterisieren, Ehrbarkeit und Putz gerade der Klientinnen hervorhebt, bleibt unklar. Einfacher ist die Erklarung, welche tvahere von dem Spinnen bez. Weben der Purpurwolle verstand.’ Despite Kiessling’s conclusion, I hope to show that the interpretation, ‘trail their purple robes’ is alone possible. In support of the meaning, ‘spin,’ ‘weave,’ Kiessling cites Varro, fr. 190, B; simul manibus trahere lanam. But it is a pure assumption to attach to trahere in this passage the meaning of either ‘weave’ or ‘spin.’ Harper’s Latin Dictionary, it is true, defines trvaho as meaning ‘spin,’ ‘manufacture ’ (I. B, 8), but without justification. tion of the passages in which the word occurs. all) of these are the following: The true meaning is shown by an examina- The most important (if not The Varro passage cited above. 1 ste Auflage, besorgt von Richard Heinze. Ji Ic B tion fr the wc on the S] ing ‘ in his fleeces in the Si Schiit: distinc The c nympl nor to shall kt cardin the lin In Me cardin and m eviden of the also. ‘spin.’ T regula quam us, ut eneca qui ab t, will ' have sinter- stood ; as he ealth.’ rus Ut, ice be > first t, and ynne of ntius. iy the 2arch- letely n the ding ; lorace nm der » aber de der velche espite their . 190, ch to rper’s ture ’ mina- if not a AR NOTES ON HORACE Juvenal il. 54 sqq.: uos lanam trahitis calathisque peracta refertis uellera, uos tenui praegnantem stamine fusum Penelope melius, leuiusque torquetis Arachne. Ovid. Met. xiv. 264: quae uellera motis nulla trahunt digitis nec fila sequentia ducunt. Ibid. xiii. 511 (Hecuba, speaking of Penelope) : quae me data pensa trahentem matribus ostendens Ithacis, ‘ haec Hectoris illast clara parens.’ Ibid. il. 411: lanam mollire trahendo. Id. Herotdes il. 75: nos humiles famulaeque tuae data pensa trahemus. In the cita- But in all these passages the word means ‘ card’ not ‘ spin.’ tion from Juvenal we clearly have a reference to two processes connected with the working of wool: (1) the carding; (2) the spinning. Schiitz, in his note on the Horace passage, cites the passage from Juvenal as supporting the mean- ing ‘spin’; but he cannot have intelligently examined the context. Weidner, in his note ad loc. correctly interprets, ‘‘ you card the wool and carry the carded fleeces in baskets.” The spinning is subsequent, and is obviously referred to in the words: tenui praegnantem stamine fusum torquetis. So, too, in the three passages from Ovid’s Metamorphoses (two of which Schiitz cites in further support of the sense of ‘spin ’) the sense is ‘card,’ as is distinctly recognized by Siebelis and Polle in their notes on these passages. The context of xiv. 264, in fact, makes the meaning ‘ spin’ impossible, for the nymphs are said neither to perform the operation indicated by wellera trahere nor to spin (fila sequentia ducunt). In Met. xii. 511, the idea is that Hecuba shall be compelled to perform the most menial and servile tasks, among which carding was certainly one of the humblest. The same notion is at the basis of the lines quoted from the Heroides, as indicated by the words humules, famulae. In Met. ii. 411, mollive could not apply to spinning, but does apply perfectly to carding, by which the matted and tangled fibres of the wool are disengaged and made soft and flexible. The Varro passage alone fails to give internal evidence of the meaning of trahere, but in view of the other established instances of the signification of the word, it presumably bears the same meaning here also. At least, the passage furnishes no evidence in support of the meaning ‘spin.’ The meaning ‘card’ is also a perfectly natural one for traho, as the word regularly connotes a certain inertia or reluctance on the part of the person or 150 NOTES ON HORACE thing to which it applies. An expression like fila sequentia trahunt (instead of ducunt, as above) would be impossible. The situation, then, is this: trahere can mean ‘ card’ and is found in this sense combined with Janam, pensa, uellera. Can it be taken in the passage from Horace in the same sense with purpuras ? This raises the question of the meaning of purpurae. Horace elsewhere uses this word of the dye or of the finished garment—c.g., Odes III. 1, 42; IV. 13, 13. Nor does any other writer seem to use it of purple wool—the sense that would be demanded here. Moreover, the Romans did not card the dyed wool, which would have involved needless waste. They dyed it after carding. See Marquardt, Das Privatleben der Romer, p. 504; Bliimner, Technologie und Terminologie der Gewerbe und Kiinste, 1. p. 221. It seems clear, therefore, that trahere cannot bear the sense of ‘card’ in our present passage, for the conception of carding either the purple dye or purple garments would be absurd. On the other hand, trahere can mean ‘ trail,’ ‘ draw after one, of a garment, and is so used in Hor. Ars Poet. 215; ‘trahere per pulpita uestem.’ Under the circumstances, this seems to be the necessary meaning in the passage in Odes, 11. 18, 7 sq. Porphyrio, to be sure, does take the phrase as meaning ‘work the purple’ (purpuram carpere), adding that the wife of a candidate for office was assisted by the wives of his clients in preparing the official robe in anticipation of the candidate’s election. But whether this is true or not, it can have no bearing on the passage under discussion. More likely than not Porphyrio’s explanation is pure invention devised to fit his false interpretation of the line. CHARLES E. BENNETT. CORNELL UNIVERSITY, IrHaca, N.Y. In ments He de differe race @ So far next V in ‘eq he, co will b imitat ] guess will. modes the p the ‘ Prop 2ad of n this issage of the of the other here. ‘olved tleben e und sense r the ment, ar the ge in aning te for be in it can n not ation TT. i i é eee A TRANSPOSITION IN PROPERTIVS. In the ninth elegy of his third book Propertius withstands the encourage- ments of Maecenas, who has urged him to attempt heroic themes. Maecenas eques Etrusco de sanguine regum, 1 intra fortunam qui cupis esse tuam, quid me scribendi tam uastum mittis in aequor ? He declares himself unequal and unsuited to the task: different minds have different aptitudes ; Lysippus excels in one craft, Calamis in another; some race afoot and some in chariots ; hic satus ad pacem, hic castrensibus utilis armis: naturae sequitur semina quisque suae. 20 So far his weapons of defence are taken from a common armoury ; but in the next verses he develops the argumentum ad hominem which was foreshadowed in ‘eques’ and ‘intra fortunam qui cupis esse tuam.’ Such promptings, says he, come strangely from Maecenas, whose own discreetness and self-repression will be famous in history, and whom he is resolved, so far as in him lies, to imitate. at tua, Maecenas, uitae praecepta recepi, 21 cogor et exemplis te superare tuis. cum tibi Romano dominas in honore secures et liceat medio ponere iura foro, .. . parcis et in tenues humilem te colligis umbras, uelorum plenos subtrahis ipse sinus. 30 crede mihi, magnos aequabunt ista Camillos iudicia, et uenies tu quoque in ora uirum, Caesaris et famae uestigia iuncta tenebis : Maecenatis erunt uera tropaea ~ -. I hold my hand over the last word of u. 34 and I challenge the reader to guess it. He cannot: remember it by chance he may, but guess it he never will. Its sense indeed he knows or thinks he knows: it is a word meaning modestia ; for nothing else would be coherent or even reconcilable with what the poet is saying. If the true trophies of Maecenas are anything other than the ‘ iudicia’ which place him beside Camillus and the ‘ uitae praecepta’ which Propertius has taken for his own, they must on no account be mentioned. 152 A. E. HOUSMAN The difficulty is to find a word of this sense which will suit the metre; but from passages like Plaut. Bacch. 613 ‘sine modo et modestia sum’ and Cic. de off. 1 15 ‘in omnium quae fiunt quaeque dicuntur ordine et modo, in quo inest modestia et temperantia’ it appears that modus sometimes approaches the abstract sense of modestia: probably then the hidden word is modus. I take my hand away and the word is fides. Out leaps the keystone from the arch, and the poet’s whole edifice of argument collapses into ruin. The true trophies of Maecenas are nothing related, however remotely, to Camillus in the past or Propertius in the present: they are his loyalty to his illustrious friend. From this scene of dissolution let us avert our gaze to II 1 17-42. quod mihi si tantum, Maecenas, fata dedissent 17 ut possem heroas ducere in arma manus, non ego Titanas canerem... . regnaue prima Remi aut animos Carthaginis altae 23 Cimbrorumque minas et benefacta Mari: bellaque resque tui memorarem Caesaris, et tu 25 Caesare sub magno cura secunda fores. nam quotiens Mutinam aut ciuilia busta Philippos aut canerem Siculae classica bella fugae. .. . te mea musa illis semper contexeret armis, et sumpta et posita pace fidele caput. Theseus infernis, superis testatur Achilles, hic Ixioniden, ille Menoetiaden. sed neque Phlegraeos Iouis Enceladique tumultus intonet angusto pectore Callimachus, 40 nec mea conueniunt duro praecordia uersu Caesaris in Phrygios condere nomen auos. aw wi The difficulty of the couplet 37 sq. has not, like that of III 9 33 sq., escaped the notice of interpreters: it has now been under discussion for more than a century and a half. First of all we must determine the meaning of the verb ftestatur. testart aliquem usually means festem adhibere aliquem, to take some one to witness. In what circumstances and for what purpose do Theseus and Achilles take to witness Pirithous and Patroclus? Translators add matter which is not in the Latin, and consequently differ in what they add. One explanation—that the two heroes cite their two comrades to convince the incredulous denizens of their two worlds that such a friend as Maecenas can actually exist—is best dismissed, as Vahlen Sztzb. d. kon. Akad. d. Wiss. Berlin 1882 pp. 279-80 dismisses it, without a word of comment; but here is the other: ‘Achill, dem Patroklus bis in den Tod treu geblieben, Theseus, der dem Pirithous bis in die Unterwelt gefolgt, rufen, jener vor den Gottern der Oberwelt den Patroklus, dieser vor den Gottern der Unterwelt den Pirithous zum Zeugen auf stets bewahrter Freundestreue.’ Achilles then, accord friend. cannot a true Patroc us, Co! vant n have k ever. III 12 lates v such 1 sense the na memo and tl icstant ning | not ql the mr ‘ Thes one of isle of keep 1 recall: T is the the se comp perple defen: a nun avOjo dédoph Scam to pre partic lishec neces * sper pleni poeta corru elenc to Ww 40 Sq., nore the take seus utter One the enas Viss. but ben, den den fe RON oe A TRANSPOSITION IN PROPERTIVS 153 according to Vahlen, cites Patroclus to testify that he, Achilles, has been a true friend. But Propertius does not say so, and would be foolish if he did; for he cannot imply that Caesar cites Maecenas to testify that he, Caesar, has been a true friend; while if it is Maecenas who cites Caesar, Caesar is compared to Patroclus and Maecenas to Achilles. The words of the author, as they lie before us, contain neither the absurd meaning which Vahlen rejects nor the irrele- vant meaning which he approves; indeed, if ‘ testatur’ is to have the sense we have hitherto been giving it, they seem to me to contain no meaning whatso- ever. But testart aliquem is capable of other senses; for instance in Luc. III 122 ‘ wictorem clara testatur uoce tribunus’ it means that Metellus expostu- lates with Caesar and protests against his action: it cannot of course have any such meaning here in Propertius, but I quote the example to show that the sense of the phrase is not rigidly fixed. Now testari c. acc., when the acc. is the name of a thing, often means to bear witness of something and to keep its memory alive, as in Prop. III 7 21 ‘sunt Agamemnonias testantia litora curas’; and the acc. may be the name of a person, as in III 13 51 ‘ torrida sacrilegum tcstantur limina Brennum’: the portal of Apollo’s temple, scorched by the light- ning launched in its defence, is a memorial of Brennus and his assault. This is not quite parallel to testatur Menoetiaden, for the presence of an adjective makes the meaning plainer; but it is analogous; and I understand our distich thus : ‘Theseus before the lower world, Achilles before the upper, bear witness, the one of Pirithous, the other of Patroclus.’ Theseus in hell and Achilles in his isle of Leuce are everlasting remembrancers of their less famous comrades and keep their character and story from oblivion: no one sees the knight but he recalls the squire; so indissoluble is the bond. The couplet therefore is quite germane in sense to the poet’s theme, which is the close tie between Caesar and Maecenas; but it has no relation to the sentences around it. The comparison which it evidently contains is not a comparison to anything present in the context; and hence the natural perplexity to which it has given rise. Vahlen opusc. I pp. 303-9 undertakes its defence without understanding what it is that he has to defend. He adduces a number of passages like Pind. Nem. IX 39 sqq. Aéyerae wav” Exrope ev Kr€os avOnoa A TRANSPOSITION IN PROPERTIVS 158 tenebis.” But in poetry the nominatives tw and ego, even where contrast or emphasis seem to demand them, are often omitted. A good parallel is fur- nished by Hor. serm. II 3 211-3. Aiax, immeritos cum occidit, desipit, agnos: cum prudens scelus ob titulos admittis inanis, stas animo? In prose it must have been ‘tu cum prudens’ etc. I add a few more examples out of many: Hor. epfist. II 2 g1 ‘carmina compono, hic elegos,’ Ouid. her. VI 135 ‘prodidit illa patrem, vapuc de clade Thoanta,’ XII 187 ‘si tibi swm uilis, communis respice natos,’ XVII gg ‘quam multos credis iuuenes optare, quod optas,’ XIX 185 ‘quod cupis, hoc nautae metuunt,’ Luc. II 292-5 ‘gentesne furorem | Hesperium ignotae Romanaque bella sequentur | diductique fretis alio sub sidere reges, | otia solus agam?’ Some- times even both of two contrasted pronouns may be omitted: Mart. V 13 g ‘quod sum, non potes esse.’ When they enquired of Solon why he had provided in his code no penalty for parricide, he answered that he thought nobody would ever do such a thing ; and there prevails a tacit assumption, no truer than Solon’s, that transposi- tions of this nature do not occur. They do occur, as anyone can ascertain who cares to know; but their occurrence is disquieting, so they are noticed with reluctance and rapidly forgotten. I will here hand over two specimens of the offence to be dealt with more maiorum: sew them up together in the usual sack and drown them according to precedent in Lethe. The first shall be Ouid. trist. III 11 21 sq. Ina thirteenth century MS, Paris. 7993, this distich is missing from its place and found between IV 7 18 and 19: ‘ Sphingaque et Harpyias serpentipedesque Gigantas | centimanumque Gyan semibouemque uirum. | causa facili cuiuis licet esse disertum | et minimae utrves frangere quassa ualent. | haec ego cuncta prius quam te, carissime, credam | mutatum curam deposuisse mei.’ This transposition has no apparent cause, but the second example is different. In one family of the MSS of Germanicus’ phaenomena the seven verses 387-393 are placed between 285 and 286. Their omission after 386 seems to be a pure accident, but the place of their insertion can be explained bya cause to which I have alluded in my notes on Manil. II 232 and 330. A reviser notices that certain lines have been left out and adds them at the end of the book. Then comes a copyist who wishes to put them back in their place: where is that? The first verse of the seven, 387, is ‘nec procul hinc dextra defundit Aquarius undas’: he turns over the leaves, lights upon 285 ‘dextra manus, latices qua fundit Aquarius, exit,’ and there he inserts them. The Propertian couplet III 9 33 sq. may have wandered to its present abode in the same way. It contains the names Caesavis and Maecenatis: the scribe who finds it astray looks about him for a context, and in III g he sees Maecenas at u. 21 and Caesar at u. 27. A. E. HOUSMAN. THE TEXT OF THE BELLVM GALLICVM: AND THE WORK OF H. MEUSEL. EVERY ONE who has used Dr. Meusel’s excellent edition of Caesar’s Bellum Ciuile will welcome the first volume, comprising Commentaries I.-IV., of his final edition of the Bellum Gallicum. Nominally, each of the two books is a revision of the work of other men,—in the former case of Kraner and Hofmann, in the latter of Kraner and Dittenberger ; really each, especially the latter, contains the fruit of so much independent research that the personalities of the older editors are obscured. That part of the joint work of Kraner and Dittenberger which corresponds with the revision consisted of 199 pages; the revision extends to 464. As Meusel remarks in the Preface, it is no longer intended for the use of schools, but of teachers and classical scholars. The historical introduction and the section on the Roman army, which retain much of Kraner’s work, have been carefully retouched; many of the footnotes have been corrected or rewritten; many, both on linguistic points and on the subject-matter, have been added. But what places this edition in a class by itself is its amazingly thorough Critical Appendix. Although it deals with less than half of the Bellum Gallicum, it fills 118 closely-printed pages ; and it would have filled half as many again if, in the case of all those variants and passages of doubtful authenticity which the editor had already discussed, he had not contented himself with merely referring to his published writings. Few English scholars have critically studied the text of the Bellum Gallicum, and very little is known here about Meusel’s critical work. In a valuable handbook, recently published, I read that for the constitution of the text seven MSS. only are of value, five of which belong to a group descended from ‘a common parent a,’ while ‘to the second group § belong two MSS. (TU) [now generally designated ah], presenting a text which has been polished at some period by an editor who has endeavoured to tone down Caesar’s terse and vigorous style by touches of Ciceronian elegance’;? and again, ‘ The first class was preferred by Nipperdey and others, while the second has found a champion in Meusel.’ This observation seems to suggest that 1 C,. Iulit Caesaris commentarii de bello Gallico Berlin, Weidmannsche Buchhandlung. 1913. erklart von Fr. Kraner und W. Dittenberger. 2 It is true that in certain passages the writer Siebzehnte, vollstandig umgearbeitete Auflage of the archetype of £p deliberately altered von H. Meusel. Erster Band. Mit einer Karte Caesar’s text; but even a is not free from von Gallien von H. Kiepert und drei Planen. gratuitous emendation. Meuse. nizes 1 T to 8 br former B, we | but A. belong group, Celsus Meuse celvab N which in his inform of 189 L begin Verein variou Nippe prefer the pt misses plete, Meuse — whi nume! and f know! from exami Thus to prc criticu 185 Ff gram! 1 Rh 2 Jal lin, 19 publish schrift | cover. 3 Th HE Sar’s -IV., ooks and y the lities and » the nger The nuch have . the s by . less ould ages not llum In ition roup long has own and cond that I 3. writer [tered from THE TEXT OF THE BELLVM GALLICVM 157 Meusel prefers the second class. It would be more true to say that he recog- nizes its independent value. The truth is that not seven only, but nine MSS. are indispensable, and that to 8 belong not two only but four, which are known as a, f, h, and /, the two former belonging to the family 7, the two latter to the family p. Both a and 8, we all know, are commonly regarded as derived from the same archetype, X ; but A. Klotz! has argued that the archetype of all the extant MSS. was a copy belonging to 8, and that a is descended from a copy belonging to the same group, in which readings from a MS. of the sixth century, published by Julius Celsus Constantinus and Flavius Licerius Firminus Lupicinus, were inserted. Meusel” has admitted that the pedigree which Klotz constructed may con- ceivably be right. Nipperdey not only preferred a, he rated 8 very low; but the evidence on which he based his judgement was inadequate. He was ignorant of h and /, and in his time a had not been accurately collated. Those who desire fuller information on this point should read the preface in Meusel’s critical edition of 1894. Let us see what Meusel has done for the constitution of the text. To begin with, he contributed two articles to the Jahresberichte des philologischen Vereins zu Berlin of 1885 and 1886, in which, after fixing the correlation of the various MSS., he proved that in many passages, besides those in which Nipperdey had been compelled to have recourse to #, its readings were to be preferred to those of a, and definitely established its worth.* Then followed the publication of his great Lexicon Caesarianum (1886-1893), which only just misses perfection because the collation of the best MSS. was not then com- plete, and therefore some of the readings are inaccurate. Meanwhile, however, Meusel himself collated the Ashburnham MS., known as S—one of the a group —which had been previously neglected, and also a, f, and h, correcting numerous errors which had been made by his predecessors, Diibner, Frigell, and Holder. Detlefsen furnished him with a collation of the Vatican MS. known as M, belonging to a, which in about four hundred passages differed from the collation of Frigell and from the one used by Diibner. Meusel examined these passages and found that Detlefsen was almost always right. Thus in 1894, soon after the last part of the Lexicon had appeared, he was able to produce an edition which for the first time offered a satisfactory apparatus criticus. In the same year he published in the Jahresberichte an article of 185 pages, in which 761 passages were minutely examined with regard to grammatical considerations, to the context, and to Caesar’s use of language.‘ 1 Rheinisches Museum, 1910, pp. 224-34. Neue Jahrbiicher fiir klassische Philologie of 1885, in 2 Jahvesberichte des philologischen Vereins zu Bey- which Meusel showed that, as regards the use lin, 1912, pp. 18-21. These Jahresberichte are of aand ab, B is much more trustworthy than a. published as supplements along with the Zeit- 4Mr. W. E. P. Pantin, in a letter which schvift fur das Gymnasialwesen and in the same_ I have just received from him, says ‘I found cover. that the ’94 article interested me much, though 3 These articles were preceded by one in the’ I had no special interest in the text of Caesar, 158 T. RICE HOLMES For several years his time was fully occupied by professional duties; but in 1910 he published in the Jahresberichte an article of 75 pages, dealing with passages which he regarded as spurious. I have briefly noticed his arguments, which often confirmed or reinforced those of other scholars, in my own modest edition. Of the valuable reviews of Caesarian literature which he has con- tributed to the Jahresberichte I need not speak. There could be no greater mistake than to suppose that Meusel takes sides with 8 against a. On the contrary, whenever he finds the two groups at variance and cannot discover a particular reason for preferring one to the other, he follows a. Independently I came to the same conclusion myself thirteen years ago.’ In the First Commentary Meusel decides in favour of a 100 times, in favour of 8 64; he decides in favour of one family or of one or more MSS. of a 20 times, of 8 19. In the Third Commentary the correspond- ing figures are 94, 76,7,andg. Those who have studied Meusel’s writings know the scrupulous impartiality with which he weighs the testimony of rival MSS. and the candour with which he welcomes well-reasoned correction. What he and his fellow-workers have done is to show that a does not deserve the preference which Nipperdey and his disciples ascribed to it, and that 8 deserves to be treated with respect. Indeed Nipperdey, although he was so imperfectly informed, tacitly admitted the claims of 8; and the editor of Caesar in the Bibliotheca Oxoniensis, who made no use of fand /,didthe same. In the Seventh Commentary Mr. Du Pontet adopted the readings of 8, or of MSS. belonging to 8, 153 times, although he never did so if any way of escape was open; and of these readings 84 are not noticed in his brevis adnotatio critica. What I propose to do now is to show how far the text has been purified by Meusel, Walther, Richter, Rudolf Schneider, Klotz, and others; how far it must, unless better evidence comes to light, remain uncertain. First, as to interpolations, real or supposed. In some cases Meusel’s arguments seem to me conclusive; in a small majority he has, I think, shown that the passages in question are at least suspicious ; but he is sometimes sure when his arguments do not seem to warrant more than doubt, and his attitude is generally that of a counsel for the prosecution rather than of a judge. I will give one instance. The paragraph? which describes the course of the Meuse and of the Rhine was long ago branded as spurious by von Géler and others. Meusel devotes to it three pages of destructive criticism ; and Klotz,® after an even more elaborate examination, independently came to the same conclusion. My own opinion is that they have discovered ground for mistrusting the authenticity of the passage, but that some of their arguments are more or less two earlier articles are equally valuable ; and if I can persuade even a few English scholars to study all three, this little paper will not have been written in vain, 1 Classical Review, 1901, p. 175. 2 iv. 10. 3 Cdsarstudicn, pp. 36-43, 135-38. because it gave meso much information as to the written Latin of the time of Caesar. Anyone who wants to know what is classical Latin (in that sense) should work at that article of Meusel’s. If it had been issued as a separate work, it would be as well known in England as Lebreton’s book on Cicero.’ I may add that the MOS ASE weak, do no says } the pl reason of the couple the res throug Triboc Nemet approz large i from t tribes Caesa flumin strean LXX)D have | writte: ® Rhine that e every war Ww that t Augus auxili< eviden forgot Vil. 77 it cert were 1 Caesa: Rhine self n positic evider writer Rhine sea’ (1 ti. 4 2 Mn ut in with ents, odest con- takes ps at >» the iyself ur of ne or yond- know MSS. at he 2 the erves fectly 1 the venth ging and rified far it usel’s hown ; sure itude [ will feuse hers. er an sion. + the ) r less and if lars to t have THE TEXT OF THE BELLVM GALLICVM 159 weak, and that there is hardly one which, taken by itself, could not be, I do not say refuted, but reasonably answered. The description of the Rhine, says Meusel, is irrelevant; and the Meuse rises not in the Vosges, but on the plateau of Langres, both of which regions Caesar had visited. Strong reasons have been given for believing that the Batavi did not occupy the delta of the Rhine before the time of Augustus; and in iii. 28, 1 Caesar does not couple them with the Menapii and the Morini,—the tribes which, ‘ while all the rest of Gaul was tranquillized, remained in arms.’ The Rhine did not flow through the country of the Nantuates; the geographical position of the Triboci is mis-stated; in the list of the Cisrhenane tribes the Rauraci, Nemetes, and Vangiones are ignored. The statement that the Rhine, as it approaches the sea, ‘branches off into several channels and forms numerous large islands’ is absurd. The Menapii occupied both banks of the Rhine ‘ not far from the sea’; therefore there was no room for the Batavi and the ‘ fierce rude tribes ’ which are said to have occupied ‘ many’ of the ‘ numerous large islands.’ Caesar would not have written parte quadam ex Rheno, but parte Rheni; caput fluminis elsewhere [in prose] means not the mouth, but the source of a stream ; Caesar would have written /ongius immediately before mulibus passuum LXXX instead of separating longius from milibus by ab Oceano; he would not have used the quasi-poetical expression citatus fertur, and he would have written partem magnam .. . imcolunt, not pars magna .. . tncolitur. Cannot something be said in reply? Irrelevant the description of the Rhine certainly is; but is vi. 24 strictly relevant, and is it not conceivable that even Caesar may occasionally have been tempted to digress? Could every general be trusted to state accurately in what part of the theatre of a war which he had waged this or that river rises? It is by no means certain that the Batavi did not occupy the delta of the Rhine before the time of Augustus ; Lucan,’ who aimed at accuracy, mentions them among Caesar’s auxiliaries; and J. H. Holwerda® argues from archaeological and literary evidence that the settlement took place in Caesar’s time. Caesar may have forgotten to couple the Batavi with the Menapii and Morini, just as in vii. 77, 14 he omitted to couple the Teutoni with the Cimbri; nor, indeed, is it certain that the Batavi, if they already occupied the delta of the Rhine, were in the same category with the Menapii and Morini.* Even if they were, Caesar would have left so inaccessible a tribe unmolested.* Certainly the Rhine did not flow through the country of the Nantuates; but Meusel him- self now adopts the reading of +,— Nemetum. The misstatement of the position of the Triboci and the omission of the Rauraci and Vangiones are evidence of carelessness, but not conclusive evidence of late authorship. The writer, whoever he may have been, was misinformed about the delta of the Rhine; why should not Caesar have been misinformed? ‘ Not far from the sea’ (non longe a mari) refers, as Meusel himself says, to the neighbourhood of 1 i. 431. 4 He did not attack even the Menapii in 2 Mnemosyne, 1913, pp. 1-9. 3 ib. 56 B.C. 5 iv. 1, I. 160 T. RICE HOLMES Xanten and Cleve, above the first bifurcation of the Rhine; the isula Batauorum was below. Nor is one disposed to yield unreservedly to the linguistic objections when Meusel himself admits that prognat: (ii. 29, 4) 1s poetical, and that breuttas (11. 30, 4) 1s used by Caesar in a sense which 1s else- where found only in poets and later writers. There is, however, if the text is right, one gross geographical mistake in the chapter, which I pointed out in 1899, and which Caesar would not have made. He would not have written per (fines), but praeter ; and this objection could only be countered by supposing that ptey, an abbreviation of praeter, was confounded with per. Moreover, though a chain is no stronger than its weakest link, the metaphor is not here applicable ; the authenticity of a paragraph which is assailable from so many sides is certainly open to grave suspicion; and I am only pleading that no passage should ever be condemned before due consideration has been given to all that can be said in its defence. Let us now attempt to classify the readings which Meusel’s apparatus presents. Everyone who systematically studies the text will observe a con- siderable number of passages in which obvious errors of the archetype were long ago corrected by emendations which have commanded universal assent. Thus in i. 1, 5 Oceano has been substituted for oceanum; in 1. 10, 5 Segustauos for sebusianos; in i. 16, 5 pracevat for pracerant; in i. 25, 3 tflexisset for inflixisset; in 1. 28, 5 parem for partem, and so on. The reader will also unhesitatingly cancel a large number of readings exhibited now by a, now by £, now by a family or MS. of either, which are so manifestly wrong that no editor would defend them. Next, perhaps, he will consider various passages, well known even to those who have not busied themselves much with textual problems, which are so corrupt that an editor must either adopt some one of the numerous emendations which they have evoked, or devise a new one, or merely obelize. A typical example is in i. 24, 2, where MB*8 have (ipse interim in colle medio triplicem aciem instruxit legionum quattuor ueteranarum [ueteranorum X]) zta uti supra se (in summo iugo duas legiones quas in Gallia citeriore proxime conscripserat et omnia auxilia conlocari ac totum montem hominibus compleri et interea sarcinas in unum locum conferri et eum ab iis {his X] qui in superiore acie constiterant muniri lussit); while in place of the words which I have italicized yB* have tta uti supra. sed. Meusel originally adopted what Klotz calls the desperate remedy of bracketing ita uti supra se and ac totum montem hominibus complert et interea. He has now adopted, as I have also done, a conjecture proposed by Klotz,? which is at all events far better than any other, not only yielding perfect sense, but being palaeographically sound :—(ipse interim in colle medio triplicem aciem instruxit legionum quattuor ueteranarum;) in summo iugo duas legiones quas in Gallia citertore proxime conscripserat et omnia auxilia conlocaut, ita utt supra se totum montem hominibus compleret ; tmpedimenta sarcinasque (in unum locum conferri et eum ab 2 Cdsarstudien, p. 239. 1 Caesar's Conquest of Gaul, p. 492 ; ed. 2, 1911, p. 452, 0, 7. nea Beane ime ~ RD be = lis qui 1 other b course « which sively \ style w been es authori Meusel furnishe not me! appeale majorit: moral | acknow has bee Caesar C. Wag give a horam ¢ Caesar, In v. 2: ibique — comprel that it i been pu have pr had asc context, ex ciuita able. ‘ Tasgeti importa also his report | _ identifie profictsct Therefo the spot all that in vi. 36 subsec adds, ‘ isula the 4) 1S else- xt 1S it in tten sing ver, here lany t no n to ratus con- were sent. AUOS t for also ry B, t no ges, tual e€ of e, Or ‘ipse rum allia item b iis the rally and have etter cally num v10ve ntein n ab nee eed THE TEXT OF THE BELLVM GALLICVM 161 iis qui in superiore acie constiterant muniri iussit). In my edition and in my other books I have noticed every passage of this sort; and Meusel has of course discussed all that occur in his first volume. Turn now to the passages which Meusel handled in his memorable article of 1894. They deal exclu- sively with orthographical and grammatical questions; questions relating to style were reserved for later treatment. The independent value of 8 having been established, it was impossible, in deciding between variants, to appeal to authority; to follow a blindly would have led to hundreds of mistakes. Meusel treated every case upon its own merits; and his great Lexicon furnished him with a useful instrument. But I need hardly say that he did not merely compare Caesar with Caesar; wherever an appeal was possible, he appealed to the context or to recognized grammatical principles. In the great majority of instances his decision seems to me to be sound. Often he achieves moral certainty; and when probability only has been attained, one must acknowledge that his method was the only method available. Occasionally he has been able to revise his judgements. For instance, in 1894 he believed that Caesar had written mensium (i. 5, 3; vi. 18, 2); but the researches of C. Wagener?! have since convinced him that the true form is mensum. give a few specimens of his work. Let me In v. 35,5 @ has (cum a prima luce ad horam octauam) pugnaretur ; B has pugnassent. The former is preferable because Caesar, in describing the duration of a battle, almost always uses the passive. In v. 25, 4 a has (L. Plancum cum legione . . . in Carnutes proficisci iubet ibique hiemare, quorumque opera) cognouerat (Tasgetium interfectum, hos comprehensos ad se mittere); while in 8 we find cognouerit. Meusel remarks that it is quite conceivable that Caesar, after stating in § 3 that Tasgetius had been put to death, and that the assassination had been reported to him, should have proceeded to imply, as he must have done if he wrote cognouerat, that he had ascertained the names of the assassins. Conceivable, but, considering the context, and particularly the words Hunc [| Tasgetium] inimict palam multis ex ciuitate auctoribus interfecerunt. Defertur ea res ad Caesarem, improb- able. ‘The names,’ says Meusel, ‘of the actual murderers—the inimici of Tasgetius—were perhaps reported to Caesar; but it was, of course, more important for him to render the instigators of this political murder, who were also his own antagonists; powerless for mischief. In the interval between the report [mentioned in § 3] and the dispatch of Plancus he could not have identified the ringleaders; for not only was there no time (Plancum celeriter proficisct iubet), but the distance (Caesar being at Amiens) was too great. Therefore it is almost certain that Plancus had to institute the investigation on the spot, and accordingly cognouerit is necessary.’ all that was good in the work of his predecessors. Meusel, of course, utilized Thus, after remarking that in vi. 36, 3—magna multitudo calonum, magna ws tumentorum quae in castris subsederat (B)—subsederant, the reading of a, is grammatically sound, he adds, ‘Aber mit feinem Gefiihl bemerkt Schneider, diejenigen, die subsederant 1 Neue philologische Rundschau, 1899, pp. 241-6. 162 T. RICE HOLMES schreiben, “‘iumentis voluntariae actionis partem maiorem, quam decet, attribuunt et quod ad vim seu multitudinem eorum apte refertur, id ut minus apte de singulis dici videatur, efficiunt. Lastly, we have to do with a large number of variants which it was not consonant with Meusel’s plan to examine in his long article, but which he has studied in his Critical Appendix, and with others which he will doubtless I have dealt with a few of them in my larger In many cases the divergence is between Thus in v. 3, 5 99 9 study in his second volume. books and in the Classical Review. words; in others between arrangements of the same words. we have to choose between famuliartiate (a) and auctoritate (8); in v. 12, 2, between permanserunt (a) and remanserunt (8); in v. 12, 6 between atque (a) and et (8); in v. 13, 1 between pertinet (a) and latus tenet (8); in v. 42, 2 between nobis (a) and nostris (8). Agatn, in v. 52, 2,a@ has legione producta and B pro- ducta legione ; in v. 54, 4 a has inferendi bell1 and 8 belli inferendi. ‘When the choice lies, as it often does, between wt and uti, one must generally adhere to the principle of following a. In questions of order especially the decision is often extremely difficult. Sometimes there is as much to be said for one read- ing as for the other. Ini. 44, 3 who can choose between a se uno proelto (a) and uno a se proelio (8)? The support which is forthcoming for the latter in i. g, 1 (velinquebatur una per Sequanos uta), vi. 5, 6 (duas ad eum legiones proficisct iubet), vii. 73, 8 (ternos inter se pedes distabant), etc. is not decisive ; and one can only follow a. So again in i. 50, 4 sortibus uaticitnationibusque (8) is as good as sortibus et waticinationtbus (a), but no better; and a claims the casting vote. Sometimes one is obliged in the last resort to trust to one’s feeling for style. I will cite two illustrations from modern literature. The late Poet Laureate once asserted that no sentence ought to begin with ‘And.’ Here is a case, taken from Macaulay’s History (i., 1848, p. 613), in which not only a sentence, but a paragraph begins with that conjunction. After describing the capture of the Duke of Monmouth, Macaulay ends a paragraph with the sentence ‘ The prisoner was conveyed under a strong guard to Ringwood.’ Then follows ‘And all was lost; and nothing remained but that he should prepare to meet death as became one who had thought himself not unworthy to wear the crown of William the Conqueror and of Richard the Lion-Hearted, of the hero of Cressy and of the hero of Agincourt.’ My other example shall be a stanza in which Matthew Arnold strengthened an impressive line by disregarding grammar : ‘In his cool hall, with haggard eyes, The Roman noble lay ; He drove abroad, in furious guise, Along the Appian way ; He made a feast, drank fierce and fast, And crowned his hair with flowers— No easier nor no quicker passed The impracticable hours.’ E00 PACT eta Fata A thou Dramat him. | of an el an eme! the His: the crit word. Me feeling are at vi be affect the sea- tissimae aduersae followed he has — proficisct precedes is in ear vol. 1., p that the two exc Italiam ( contra L justifies 1 is not t. harbour | I do not which th I wi the balan which th Meusel’s except in written t ‘ ‘ 1 Jahvesbes 2 Another method has 1, 12, 4. I reading of follows B, w which he g B.C. taken t oct oportunt passages a cet, inus 2, 2s )and ween pro- n the re to on is read- 10 (a) er 1n fictsct e can od as vote. le. I > once taken but a f the ‘The yllows meet crown ero of 1za in urding eg De dae _ mat pres: epee ed YU Yeates Wt a ‘~ - ¥ aoe THE TEXT OF THE BELLVM GALLICVM 163 A thousand years hence a critic may perhaps study the text of Arnold’s Dramatic and Lyric Poems. If he is a pedant, ‘nor’ (no quicker) will offend him. He will say ‘ Is it credible that a poet of the Victorian Age was ignorant of an elementary rule of English grammar ?’ and, confidently offering ‘ and’ as an emendation, he will ruin the line. Perhaps again an erroneous reprint of the History of England will omit ‘ And’ at the beginning of that paragraph. If the critic has a feeling for style, he will prefer the text which retains the word. Meusel has not only a fine sense of grammatical subtleties, but also a feeling for style. I will give one instance in which his and mine, such as it is, are at variance ; or perhaps he did not think that the passage in question could be affected by such criticism. In ili. 14, 2, when Caesar is about to describe the sea-fight with the Veneti he writes (circiter CCXX naues eorum para- tissimae atque omni genere armorum ornatissimae) profectae ex portu (a) (nostris aduersae constiterunt); 8 has e portu profectae. Meusel in his previous editions followed a: now he adopts the reading of 8, only changing ¢ into ex, which, as he has shown,! is certainly right. The reason which he gives is that with proficisct the statement of the place to or from which one goes almost always precedes the verb. Almost, but not quite. Here, as always, the reader who is in earnest will check Meusel’s Critica! Appendix by his Lexicon. Turning to vol. i., pp. 106, 1171, and to vol. 1i., pp. 96, 1240-1, he will perhaps conclude that the arithmetical method is not always satisfactory. He will note one or two exceptions, which are significant,—Caesarem [se] arbitrari profectum in Ttaliam (v. 29, 2) and ips profecti a palude ad ripas Sequanae e regione Luteciae contra Labtent castra considunt (vil. 58, 6). The nuance of meaning that justifies the exceptional order here seems to me to justify it also in iii. 14, 2: is not the deliberate and majestic movement of the Venetian fleet out of harbour better expressed by profectae ex portu than by ex portu profectae? But I do not expect every one to agree with me; this is one of several passages in which the personal equation must tell.? I will cite a few other instances in which, although it is not easy to decide, the balance of probability seems to incline to one reading, and, finally, two in which there is no room for doubt. In every instance but one, I agree with Meusel’s published texts ; but he has not yet stated the reasons for his choice, except in vii. 38, 3, and vii. go, 4: in both these cases I found, after I had written this paragraph, that my reasons were identical with his. In v. 12, 2 1 Jahvesbevichte, 1894, pp. 248, 9. well toss up as decide on this principle? Again 2 Another instance in which the arithmetical in iv. 31, 2, where Meusel follows 8 in reading method has perhaps been strained occurs in comparari instead of comportari, he remarks that iii, 12, 4. Here Meusel originally adopted the while comportave is used when the required reading of a, — oportunitatibus loci. He now materials are to be found in the neighbourhood, follows 8, which reverses the order. Thereason’ in this case they were to be fetched from the which he gives is that elsewhere in B.G. and _ continent. I am not arguing against comparari ; B.C. taken together oportunitas loci occurs once, butin B.C., ili. 42,2 we find frumentum ab Asia loci oportunitas four times. When the relevant atque omnibus vegionibus quas tenebat comportari passages are so few, might not one almost as imperat. 164 T. RICE HOLMES vemanserunt (8) seems to me more probable than permanserunt (a); for the writer is, I think, emphasizing the fact that the original Belgic invaders of Britain remained in the island, not the self-evident fact that their stay was permanent. In vii. 17, 6 I accept the reading of a—(hoc se ignominiae) laturos loco—which Meusel adopted in his critical edition, but which, in the reissue of his school edition (1908), he discarded in favour of loco laturos. When I read vi. 13, 1 (nam plebes paene seruorum habetur loco) and B.C. 1. 33, 2 (Pompeius enim . . . dixerat eodem se habiturum loco qui Romae remansis- sent atque in castris Caesaris fuissent), I conclude that the claims of a and § are here equally balanced, and accordingly follow a. In vii. 24, 5 my literary sense leads me to prefer the reading of 8—(tamen, quod instituto Caesaris) duae semper (legiones pro castris excubabant)—to that of a, semper duae. In vii. 38, 3 8 has (haec ab) Ais (cognoscite qui ex ipsa caede fugerunt); a has ipsis. Is it not likely that zfsts was written by a scribe whose eye had wandered to the following ipsa? In vii. 40, 3 Schneider gave a good reason for preferring (fratres Litauicci, cum comprehendi iussisset, paulo ante reperit ad hostes) profugisse (8) to fugisse (a). ‘ Mutati enim loci,’ he writes, ‘et liberi a periculo spatii significatio, quam praepositio addit, h.l. aptissima et verbum in his commentariis frequentissimum est.’ In vil. 32, 5 Sa—one MS. of a and one of 8—have (ciuitatem) omnem esse (in armis), while the remaining seven MSS. have esse omnem. Every one, I think, will agree that the latter is right, and that the meaning is ‘the people were all up in arms.’ Lastly, in vii. 90, 4—(T. Labienum) cum duabus (legionibus . . . in Sequanos proficisci iubet)—the evidence against duabus cum (a) furnished by Meusel’s Lexicon (i. 759-65) is overwhelming. Here the arithmetical method is sound.’ The nature of the evidence for the constitution of Caesar’s text is such that after the most diligent editor has done his utmost he is forced to acknow- ledge that in some passages there is no inherent reason for preferring one read- ing to another, and that in a considerable number, most of which, however, are comparatively unimportant, he can attain no more than probability, which is sometimes slight. But students who admit that each group, each family, each member of a family is entitled to consideration, will generally find them- selves agreed on which side probability lies. And for this result we are indebted to Meusel above all. I would only suggest that while he never falls into 1 I may say here that in the text of B.G. iv. 32,2. For sese read se. which I lately edited for the Medici Society I v. 1,2. Forsubductionesque read subductionisque. adopted in the following passages readings I also bracketed quam in i, 42,5, Romanorum in different from those of my other edition, which iii. 18, 8, and atque opinione timidiores in iii, 24, 5. was published in February for the Delegates of In the second and fourth of these cases I assent the Clarendon Press. The corrections, although to the arguments of Klotz, in the rest to those of they were too late for the list of Corrigendain Meusel. the complete edition, were made in separate Finally, in viii. 6, 3 I made an emendation,— editions (also issued by the Clarendon Press)of a T. (Labieno) instead of ab L. Nipperdey i., iii,, iv., and v. deleted ab L.; but in mentioning Labienus for iv. 3, 3. Ford read hi, the first time Hirtius would certainly have given iv. 5, 2. For enim read autem. the pruenomen, and T was often confounded iv. 10, 3. For Nantuatium read Nemetum. with L, le ae Tt the erro! mittunt, | man beha sind.1 G archetyp prefer to because Meusel,? 52-51 B.C day. A administ: amount < I suy with the worst te eighty ye him, said perfectio been to \V which M zu miiher sonst ror werden.’ Lexicon \ passage 1 will cont further 1 thencefor wane whe outlook 1: Swift, on same. And conquero! 1 TI owe t Companion t opportunity * Les auteur: au XX€ siéc - the rs of was niae) . the Uuvos. | oF om UNS1S- 3 are orary ALIS) In u has had ason perit ‘et la et MS. ining ter 1S ly, in iCiSCl xXICON such now- read- ever, rhich mily, hem- > are into nisque. yumi in 24, 5. assent ose of ion,— erdey us for given unded uth Pee al SS LE IML Se sje » Vier cae ATI Fo ln ESP a SS nina THE TEXT OF THE BELLVM GALLICVM 165 the error of those who, as Madvig said, rationem captiuam sub iugum codicum mittunt, he is sometimes prone to forget Boeckh’s warning, Im Allgemeinen kann man behaupten, dass von 100 Conjecturen, welche die Kritiker machen, nicht 5 wahr sind.» German criticism occasionally condemns a word or a phrase in the archetype as unworthy of Caesar. Unless it is manifestly indefensible, I prefer to let it stand, even when this or that emendation seems probable, because I believe that Caesar had no time to spare for minute revision. Meusel,? arguing that the Bellum Gallicum was written in the winter of 52-51 B.C., remarks that a practised journalist can easily write 1,500 words a day. A journalist, yes; but no man, least of all one who had to transact administrative business and to watch the course of politics, could write that amount and also polish it with the anxious care of a Flaubert. I suppose that in no department of human activity is the result, compared with the labour expended, so slender as in textual criticism. Even in the worst text one can see that the Bellum Gallicum was a great book. Nearly eighty years ago Macaulay, with a primitive edition of the Bellum Ciuile before him, said ‘ He is an admirable writer, worth ten of Sallust. His manner is the perfection of good sense and good taste.’* But to purify the text of Caesar has been to Meusel a labour of love; for he remembers the words of encouragement which Mommsen addressed to him—‘ Das edle Werk ist es werth, sich darum zu miihen. Der ungeheure Abstand dieser Commentarien zu allem, was sich sonst rémische Geschichte nennt, kann gar nicht tief genug empfunden werden.’ As soon as his second volume is published, readers who use his Lexicon will have the means of making up their minds about every single passage in which variants occur. Of course, within centain limits opinions will continue to differ; but the text will not be susceptible of much further improvement; and it is to be hoped that restless emendators will thenceforth keep within bounds their longing to correct it. If interest is to wane when finality has been reached in any branch of classical research, the Though there is no temptation to tinker at the text of May the destiny of Caesar be the outlook is not hopeful. Swift, one reads, admires, and enjoys. same. And now I bid a long farewell to the Belluim Gallicum, but not to the conqueror of Gaul. peared in tudes of August, 1913? It is unob- trusively learned, lucid, well written, and sane. 2 Jahvesberichte, 1913, p. 19. 3 Sir George Trevelyan’s Life and Letters of Lord Macaulay, 1881, p. 690. 1 I owe these quotations to Mr. F. W. Hall’s Companion to Classical Texts. May I take this opportunity of recommending an article called ‘Les auteurs classiques et la critique des textes au XX¢ siécle,’ by M. L. Laurand, which ap- T. RicE HOLMEs. THE IMPORTANCE OF THE VERONESE PALIMPSEST IN THE FIRST DECADE OF LIVY. I. INTRODUCTORY REMARKS. THE first excitement consequent upon the discovery of an ancient manu- script generally leads to an exaggeration of its importance. This was especially the case with the Veronese Palimpsest, first, because it is the only MS. of the first decade of Livy earlier than the ninth century, and secondly, because it is the only extant MS. for that part of Livy, which is not derived from the Nicomachian recension. Mommsen was naturally prejudiced in favour of the Veronese Palimpsest, which he collated and published, and, when the two families of MSS. differ from one another, he usually decides for the Palimp- sest. He used only MPL for his comparison, which is consequently very inadequate. It is possible now, however, with our wider and more accurate knowledge of the Nicomachian MSS.,! thanks to the patient research of Professor Flamstead Walters and Dr. Conway, more justly to estimate the importance of the Veronese Palimpsest in Livy’s first decade. My knowledge of the Nicomachian MSS. is entirely due to the kindness of Dr. Conway and Professor Flamstead Walters, who have given me the unique opportunity and honour of seeing in private the collations which they have made in preparation for their edition of Livy in the Bibliotheca Classica Oxontensis. The Veronese Palimpsest consists of portions of Books III., IV., V., and VI. (1-6). For details of the manuscript itself vide Th. Mommsen’s edition (T. Liuii ab urbe condita lib. III.-VI., quae supersunt in codice rescripto Veronensi descripsit et edidit Th. Mommsen). Il. CHARACTERISTIC ERRORS OF THE VERONESE PALIMPSEST. Before considering some of the readings, in which V differs [from the Nicomachian MSS., it is necessary to notice the chief errors to which the scribe of V was liable. It will be convenient to divide these into the following sections : 1 See Proc. of Camb. Phil, Soc., Nov. 1902, and Classical Quarterly, July, 1908, and Jan. 1910, Pe SSE a. Ser ‘ as 4 On Examp than th In som reader been cé rather 1 may be IT. IT. IV V. Tk IL. IV V. before : scribe 1 simply TE its addi of som a certal ordinur of V is this ca: occasio conside Aft words | instance IT] Th tegeritis It the MS to conf 1 This scribe of ACY Y lanu- clally f the it is 1 the f the » two limp- very urate oh of 2 the iness > the they ASSICA , and ition ripto 1 the cribe wing Oe THE VERONESE PALIMPSEST 167 (a) Perverted Order of Words. One of the commonest mistakes in V is a change in the order of words.’ Examples of this error are to be found scattered throughout the MS.; no less than thirty certain instances occur in the portions of the books included in it. In some cases it may be due to an attempt on the part of the scribe, or some reader of his exemplar to improve the order; but generally it seems to have been caused by the scribe’s habit of reading too fast and allowing his mind rather then his eye to guide his hand. Among examples of this, the following may be noticed : III. rr. 11 = multa cbt for 1bt multa. III. 42.6 ferre arma for arma ferre. IV. 25.8 fortunam Veis for Veis fortunam. V. 39.9 posse urbem for urbem posse. This practice of reading too fast is further illustrated in V: III. 62. 6 IV. 25. 14 V.2:.8 pontifices r soluendum religione is read in V, and the r inserted before soluendum can represent nothing but the beginning of religione ; but the scribe noziced his omission of soluendum before he had written more, and so simply left the v and wrote the correct order of words. The variation in the position of a word in different MSS. often points to its addition to the text from a marginal or interlinear explanatory note or gloss of some reader; but when perverted order is found to be a common failing in a certain MS., it can count for nothing. For instance, in III. 67. 6 (discordia ordinum et uenenum urbis huius, patrum ac plebis certamina) the huzus urbis of V is no proof that huwius should be deleted. It is true that V is supported in this case by U, but the latter is a MS. of very little consequence. It is only occasionally that V has helped to show glosses, and then it is due to other considerations more than the order of words. After this brief survey of the common error of change of the order of words in V, it may be possible to understand some of the more important instances in which this MS. differs from the Nicomachian MSS. III. 62. 3 This is Frigell’s conjecture. uictoriae for gloriae (probably introduced from the next line). studiis for antmis (introduced from studia in the next line). id consilit animique habiturus sum, quod uos mili, milites, feceritis. The readings of the MSS. are: ... hi feceritis...s V3; milites geritis M Vorm.; mth te geritis RL; .. feceritis P; mihi effeceritts P?FUD*; micht... geritis D. It is clear that gevitis is a variant which has displaced feceritis in some of the MSS. The true reading seems to have been feceritis; the effeceritis is due to confusion with the final syllable of mzlites after the first part of the word tegeritis H; mth. III. (71. 8) to IV. (21. 8). See Proc. of Camb. Phil, Soc., November, 1902. 1 This was also a common mistake of, the scribe of M, who was responsible for I. (1-9) and 168 CLARA M. KNIGHT had been omitted after mihi in minuscules. Hence Frigell’s conjecture seems to represent the original reading as far as it can be discovered from the corruptions in the MSS. V seems to show merely change of the order of milites feceritis, if the word partially wanting is milites. Mommsen, however, tries to prove that V has the true reading, though he then finds some difficulty with siilites (v. Codex Veronensis, p. 189). It is much more probable that V simply shows his common practice of perverting the order of words. IV.17.11 (The Fidenates and Veientines retired to the hills) mc ante in campos degresst sunt, quam legiones auxilio Faliscorum uenerunt. This is the reading of the Nicomachian MSS.; but V has Faliscoruim auxtliorum ; Faliscorum auxilio has been conjectured by Kiehl. The object of the conjecture was presumably to make the genitive depend on Jegiozes ; but it does not seem to be altogether necessary. Moreover, V’s auxiliorum points to Faliscorum as the second of the two words in his exemplar; he has accidentally interchanged the words, but retained the ending of the second. A genitive dependent on a predicative dative is not a very common constructicn (in this instance it is the so-called ‘Subjective Genitive’); but it is not without parallels—cf. Causae remotionis hoc nobis exemplo sit (Cic. Inv. 2. 29); ostentut scelerum (Sall. Jug. 24); indicio sui facti (Lucr. 4. 1019); etus ret testinonio sunt (Cic. Rose. C. 11); and perhaps also ut monumento area esset oppressaz nefariae spet. (Livy IV. 16. 1). VI. 1.8 is tribunos militum consulari potestate creat. For consulart V has pro...and HLA have proconsularit, which some editors retain on the ground that it occurs in both families of MSS. _ IV. 41. 10 (M. Postumius reus obiectus, qui tribunus militum pro consule ad Veios fuerat) is compared; but in that sentence pro consule means literally ‘instead of a consul,’ ‘with the power of a consul.’ No reader could take froconsulant potestate in any but the technical sense of ‘proconsular power. In the Palimpsest the corruption may be due to the initial letter of potestate having been written in the wrong place and not deleted in the exemplar. The scribes would also be more familiar with the phrase proconsulari potestate on account of its frequent occurrence in later history. (b) Additions. Another error to which the scribe of V was liable is the addition of words. Many of these are small words, probably originally in the margin or between the line of a MS., which have been incorporated into the text. Among these are the following : III. 12.7 Quod offendatineo .. . aetatem cottidie auferre is the reading given by the Nicomachian MSS.; but V adds magi<(s> before auferre. This is an obvious addition, and yet some editors on the authority of V alone accept it. For the omission of magis, even where the idea of comparison is present cf. III. 68. 11 (malae rei se quam nullius duces esse uolent). SE tie eal III. the Nic gloss. 4 news’ 01 Sex. Ros nuntias). ¥e4 In t wrong p above in copied s. Wed evant, 7 adds et ¢ VI. is the r compare?r have co compare extant. written probable compara compare XXXVI Lik usually times th from a ¢ III IV. IV. V adds. to whicl Bes whole w in some the thir Fal omissi0! It wast in uncia seems 1 the er of ever, culty at V ite in orunt ct of ut it ts to tally itive this hout entur ) sunt ariae some I. 10 ev at) of a ularr the ving ribes nt of THE VERONESE PALIMPSEST 169 III. 38.5 Legatt ab Tusculo praesidium orantes nuntiant is the version of the Nicomachian MSS.; but V adds ea after legati. Nuntiare is frequently used absolutely in the sense of ‘to carry the gloss. news’ or ‘ to act the part of the messenger.’ This, too, is a manifest (Cf. Cic. Rosc. Am. 34. 96 Occtso Sex. Roscio, qui primus nuntiat, and Ter. Hec. 4. 4. 20 Bene, tta me dt ament, nuntias). v.24 In this example nec seems to have been added from the margin in the wrong place. nec cum rep occurs in V for cum repente. It was probably meant to alter the corrupt seditzonis a few lines above into seditio nec (nec in urbe seditio nec in castris posset); but the scribe of V copied seditzonis and added nec in the wrong place. V. 40. 7 evant, velinquenda essent, consultantes is read in the Nicomachian MSS., but V quae sacrorum secum ferenda, quae, quia uires ad omnia ferenda de- adds et guae before relinquenda—an obvious addition. VI. I. 10 in primis foedera ac leges . . . conquirt, quae compararent, tusserunt is the reading of Nicomachian MSS. For quae compararent V has quae non comparerent. V has, no doubt, preserved the right verb which the other MSS. have corrupted into the commoner compararent ; but the addition of non to comparerent is impossible; they could not collect the laws, etc., which were not Mentioned in connexion with the Twelve Tables, it must also refer to extant. written laws and not to customs preserved by oral tradition. It is very probable that with the corruption of comparerent a variant reading quae non compararent arose. V’s non may have come from this. For this meaning of compareve cf. Livy XXVI. 30. 10; XXXII. 10. 3; AXAIV. 35. 6; and XAXVITI. 11. 5. Like most MSS. V has examples of dittographs. Simple dittography usually presents no difficulty, and it is needless to give examples; but some- times the scribe of V or of his exemplar attempts to emend an error arising from a dittograph, as: ITT. IV. IV. 31. 1 27.2 55- 2 vefectis his sequente for refectr, hi sequente. urbe est for urbe. enixe tribunt sibi is the reading of the Nicomachian MSS. ; but V adds enim after sibi, which is probably an emendation of ent, a part of entxe, to which the scribe had reverted. Besides the dittography of letters and syllables, V sometimes repeats a whole word or phrase; but this is not so common a fault in the Palimpsest as in some other uncial MSS. of the fifth century, as, for instance, Puteanus in the third decade of Livy. (c) Omissions. Far commoner and more important than the error of addition is that of omission in V. It occurs also in Puteanus with about the same frequency. It was much easier for a scribe to omit letters or words in copying from a MS. in uncial script, written continuously, than from a minuscule MS. in later times, 170 CLARA M. KNIGHT when the words, or at least some of the words, were separated from each other. Of this kind is the common error of haplography. To this class also belongs the error of what might be called ‘ crushed words,’ i.e. words of which the scribe wrote the beginning and end, leaving out the middle portion. If we were unable to supplement V by minuscule MSS. of the tenth century, derived from a better tradition, we should probably often experience great difficulty in discovering the true reading from these ‘crushed words,’ as the following examples show: ITI. 8.5 prodit for procedit ; III. 12. 4 neminum esse cum eius for neminem unum esse cuius; III. 57. 7 regiones for religiones; VI. 4. 7 gq. for quintum. (The same corruption occurs often in Puteanus—foll. 412a”, 416a’, 411a’, 4220’, 4230’, etc.) In addition to these omissions, V very often omits words, phrases, or whole sentences on account of homoeoteleuton. Undoubted examples of this are the following : ITI. 62. 5 IV. 12.5 v.92 V.27.7 Many more examples of omissions in V, due to homoeoteleuton, might be in aciem ... postero die is omitted after posteroque dte. neque belli for neque bello neque belli. Vetos omitted after Vetos. . aduersus omitted after aduersus. et supertovem ... eam aetatem .. given; but these are sufficient to show that it was an error to which the scribe of the Palimpsest, or of his exemplar, was liable. In spite of this there are clear instances of this mistake, by which important sentences are omitted, which editors accept as the original text on the authority of V, condemning the Nicomachian readings as ‘ glosses.” The most important of these are the following : IIT. 67. 10 Ecgquando unam urbem habere, ecquando communem hanc esse patriam licebit ? is the version of the Nicomachian MSS.; but V omits wnam urbem habere ecquando. Mommsen says of V’s reading ‘ fortasse recte’; but the repetition not only adds emphasis, but is thoroughly consistent with the whole character of this exhortation of the consul Quinctius, in which repetition of the same sentiment in different words is very common (cf. § 6 Discordia ordinum et uenenum urbis huius, patrum ac plebis certamina). Noone could omit ecquando unam urbem habere except on the unwarrantable ground that every- thing which is not indispensable is an addition. V. 3.7 Nisi forte hoc dicitis, quidquid patres faciunt, displicet is the reading given by the Nicomachian MSS.; but V. omits mtsz forte hoc dicitis after agitis, owing to homoeoteleuton, as it seems. Many editors have accepted V’s omission here. But, when it is considered in the light of the whole chapter, nist forte hoc dicitis appears an important link; it makes the transition from the suggestion of the Tribunes’ opposition to the interests of the Plebeians themselves to that of their unreasoned opposition to the Patricians smoother. It further emphasizes the fact that the Tribunes’ policy was preconcerted. V. 4.8 nos tamquam cum ciutbus agere uolumus agique tamquam cum patria nobiscul tamqua homoe« sentenc MSS. 1 really s much b alone <¢ Nicom: V. per agr V omi necessa Be words, the cas omissic Il. allo con V_ has reading defensi and (in ac nesct where - intectt wbebat abrupt. urdere, text. other two se] that he IT. version Nanniv editors. origina tpse sat aut bell uestram etiam p IV MSS. ; each ss also which If we derived ‘ulty in lowing emanene uIntum. , 4220’, ses, or of this ight be > scribe ere are mitted, 2mning are the WNC esse s unam but the » whole ition of iscordta d omit every- reading r agitis, ‘ed V’s hapter, n from ebeians oother. d. 2 patria OTE eee * } thks THE VERONESE PALIMPSEST I7I nobiscum aecum censemus is the reading of the Nicomachian MSS. V has Nos tamquam cum patria agt nobiscum, etc. The omission in the latter is due to homoeoteleuton, while agi has been added in an attempt to emend. The sentence is too ingenious to be regarded as an addition in the Nicomachian MSS. in general. Besides the palaeographical reasons against the omission, it really spoils the sentence; mos becomes unduly emphatic, while czuibus gives a much better contrast to mercennario milite in the preceding sentence than patria alone can give. Further, the idea of a mutual compact, expressed by the Nicomachian reading, is suited to the persuasive tone of the whole address. V.8.3 Minus militum periut, quia praeter aegros lixarum in modum omnes per agros uicinasque urbes negotiabantur is the version of the Nicomachian MSS. V omits per agros ... urbes; but this phrase seems to be important and necessary with negotiabantur in this sentence. Besides omissions due to homoeoteleuton, the scribe of V often left out words, of which it appears he did not see the exact force. This is especially the case with small words, like -gue, in, et, etc. A few examples of these omissions may be noticed : Ill. 24. 5 Adfirmantibus ... secum eum tum frequentemque ad signa sine ullo commeatu fuisse is the reading of all the Nicomachians. For frequentemque V has frequente, which Mommsen has emended to frequentem. The latter reading is accepted by most editors. The -que, however, seems not only defensible but necessary—‘ (they said) that he had been with them at that time and (in addition) that he had attended constantly at his post, etc.’ Cf. II. 2. 2 ac nescio an nimium undique eam minimisque rebus muntendo modum excesserint, where -que (omitted PFU) is retained by most editors. In III. 44. 6 manum intecit serva sua natam seruamque appellans sequique se tubebat, V gives sequt tubebat for sequique se iubebat ; but it makes the sentence intolerably harsh and abrupt. Again, at III. 43. 6 Stccium ... 1m medio tacentem armatumque.. . uidere, V’s omission of -que has been adopted by very many editors in their text. The -quwe seems more probably an omission in V than an addition in the other MSS,; it means that 72 medio iacentem and armatum are to be taken as two separate facts ‘(they saw) him lying in the middle and (they then noticed that he was) not despoiled of his arms.’ III. 8.7 urbi quoque Romae ingens pracbitus terror... in ve subita is the version of the Nicomachian MSS. V has ves subita for in ve subita, and from it Nannius has conjectured ve subita, which has been accepted by very many editors, and among them Madvig. Nevertheless, 71 seems to have been original and an omission in V. For the sense of zn cf. Consul perculsis omnibus ipse satis, ut in re trepida, inpauidus turbatos instruit (Livy XXII. 5): Cetert reges aut bello uicti in amicttiam a uobts receptt sunt, aut in suis dubiis rebus soctetatem uestram adpetiuerunt (Sallust, Jug. 14. 5); and Magno in aere alieno matores etiam possessiones habent (Cic., Cat. 2. 8). IV. 7.12 et forts... et domi otium fuit is the reading of the Nicomachian MSS.; V omits the first et. The omission of adverbial et is fairly common in 172 CLARA M. KNIGHT the Palimpsest. Cf. Et ab seditionibus urbanis otium fuit (IV. 35. 2); beatam urbem Romam et inuictam et aeternam illa concordia dicere (V. 7. 10); etiam praesidiumque et (=even) spem ultimam Romani nominis (V. 45. 4). In each of these sentences V omits et. IV.17. 3 Leuant quidam regis facinus is the reading of the Nicomachians ; but V has omitted guidam. Mommsen defends this omission on the ground that no diversity of opinion is recorded; but guidam seems appropriate here; for it was only those who had been present and heard the words of the king on that occasion, and not the Fidenates in general, as /euant alone would imply, who gave this evidence. V. 8.11 qui, ne quam opem ab inimico uideretur petisse, uinct ab hoste quam uincere per ciuem malutt is the reading of the Nicomachians; V omits quam before ofem, and this omission is accepted by Mommsen. Quam, however, seems necessary here in the sense of any (help), and it might very easily have been omitted in uncial script before opem. In this brief consideration of the characteristic errors of the Veronese Palimpsest, we see that many readings, in which that MS. differs from the Nicomachian MSS., are to be explained by common faults of the scribe of V or of his examplar. III. VARIANT READINGS AND GLOSSES. In a comparison of the readings of V with those of the Nicomachian MSS. we find many variant readings, which are both possible, as far as the gram- matical construction and the general sense of the passage are concerned. In these cases it is possible for the editor to distinguish subtle shades of meaning and the characteristics of Livy’s Latinity, and so to make his choice between the two. It is very likely that some of these variants go back to a very early date, perhaps even to the time of Livy himself. It may be convenient to divide them into three classes : (a) Passages in which V is not supported by any of the Nicomachtan MSS. III. 6.6 discessere socit pro tristt nuntio tristiorem domum referentes is the reading of the Nicomachian MSS.; but V has reportantes for referentes. It is difficult to decide which of these variants is the better. Referre is the usual prose word for report in the case of ambassadors, and is, on the whole, more probably the original verb. It seems unnecessary to suppose that Livy used veportantes here in imitation of Vergil’s Isque adytis haec tristia uerba reportat (Aen. II. 115), and that veferentes is a prose gloss. On the other hand, it is possible that reportantes arose from a marginal note of some reader who remembered the Vergilian passage. I{I. 44. 5 M. Claudio clienti negotium dedit, ut uirginem im seruitutem adsereret neque cederet secundum libertatem postulantibus uindicias is the reading of the Nicomachian MSS. V has petentibus for postulantibus. No one would hesi- tate to acknowledge that the Palimpsest has chosen the worse variant here. wr Ewe oe rene - "WF '- = i hy ea ee tia Pe. 3 bg 5 ce RA 4 oat 51 4 ef ues Pet ira “eg Ae bs = III. Nicoma not app (cf. nec | nec <1p adverbs IV. superesse V has There c here. ~ tion co Momms IV. Fo variant words. than ad Wa palatos f Thi of incur: change thought describe VI. of the N record 1 probabl which | favour 1 must be sentenc VI. dimisst | machiar sense he the follc also mo for b. or in simil: (VI. ro. atam ot1ain *h of ans; ound - for g on nply, juam quam ever, have ynese the V or ASS. ram- In ning ween early it to ; the It is isual more used ortat it is who ‘tutenr aig of hesi- > 6 oe Pee oe . - - — sa aaeaeidl = ene THE VERONESE PALIMPSEST 173 III. 66.2 «am non ultra discordia ciuisum reprimi poterat is read by the Nicomachian MSS.; but V has mec ultra for non ultra. This expression does not appear to occur elsewhere; but it cannot be said certainly to be wrong (cf. nec procul (I. 25. 10), where, however, Dr. Postgate has proposed to read nec procul); for the old negative nec is commonly used with some adverbs and adjectives, as necdum, necopinatus, etc. IV. 12. 10 profitert cogendo frumentum, et uendere, quod usut menstruo superesset, etc., is read by all the Nicomachian MSS. except U, which omits et. V has ut uenderet for et uendere, and Mommsen has emended it to ut uenderent. There can be no doubt that the Nicomachians have preserved the true reading here. The variant in the Veronese, which seems to be an attempted emenda- tion consequent upon the corruption of et to wt, cannot be right, and Mommsen’s conjecture seems, therefore, unnecessary. IV. 55.4 Altos recurrentis in arcem . . For in arcem V has ad arcem. It is not very easy to decide between these variant readings; but it has been seen that V has a tendency to confuse small words. In arcem is also more picturesque and more likely to have been changed than ad arcem. » CaeSOS. V. 45.3 Magna pars in agrum Antiatem delati incurstone ab oppidanis in palatos facta circumueniuntur. This is the version of the Nicomachian MSS. _ V. gives excursione instead of incursione and omits 1m palatos. The omission is probably the result of the change of iucursione to excursione, which was probably made by somebody who thought a separate ‘ sally’ was indicated, and not a continuaticn of the rout described in this chapter. VI. 2.3 mnouus quogue terror accesserat defectione Latinorum is the version of the Nicomachian MSS. V has adedefectionis for defectione. This seems to record two variant readings, defectionis or defectione ; the first two letters ad probably came, as Professor Walters suggests, from at which is aie, i.e. alit-e, which has been added from the margin. The reading defectionis has found favour not only with H. J. Miiller, but also with Madvig. If it is retained, it must be regarded as an objective genitive with terror ; but its position in the sentence makes it very awkward. VI. 3.8 quod accensum ab desperatione hostium fuisset, nt praecones per urbem dimisst pont arma et parci inermi tussissent is the reading of all the extant Nico- machian MSS. The Veronese has inermib.; but the singular gives the better sense here—‘ the unarmed man’ ‘ every person, who did not carry arms.’ Cf. The singular is also more likely to have been changed to the plural, while there is no reason for b. or bus having fallen out of the text. Cf. the Livian use of the singular in similar phrases: inernu cedetis hosti (1V. 33. 4), and parcique tussum inermt (VI. ro. 5). the following sentence nec praeter armatos quemquam utolari. 1 See Classical Quarterly, 1910, Pp. 269. 174 CLARA M. KNIGHT VI. 4.5 fremitus fut aspernantium imperium. contemnentium, which must be a gloss on the rarer word. ‘There can be little doubt that aspernantium is the more suitable word to express ‘ chafing at,’ while contemnentium would merely mean ‘ despising.’ In some instances the scribe of the Veronese Palimpsest has incorporated manifest glosses into the text. In most passages these have displaced the true reading ; for the Palimpsest, unlike some other MSS. for this part of Livy, especially M, rarely records two readings. In some cases, as will be seen, the error is further obscured byan attemptto emend. The following are examples of the incorporation of glosses in V: IlI.8.7 QO. Fabius praecerat urbt; is... tuta omnia ac tranquilla fecit. This is the reading of MOHRDL; PF have uwrbis for urbi is. V has praefectus evat urbis for praeerat urbt is—an obvious gloss. III. 64.2 Censum patrum causabantur, quo per contumeliam consulum tura plebis labefactata essent. In this passage V adds ?r. (i.e. tribunorum), after cura —an addition which is found in none of the Nicomachian MSS. No gloss For aspernantium V gives could be more manifest ; for the tribunes pretended to be the champions of the people, and, though they probably meant their own tura, they would be more likely to say that the ‘ iura plebis’ than the ‘ iura tribunorum plebis’ had been attacked. what was meant, and at the same time thought consulum needed some word to Tribunorum might easily have been added by a reader, who saw balance it. III. 65. 4 ¢tulit, ut... ts usque adeo rogaret dum decem tribunos plebi faceret is the reading of MOHD! or D°A; for usque eco RDL have usoeo, PFU usque, and V has adeo. Mommsen accepts the last as an archaism (cf. Merces uectatum undique adeo, dum, quae tum haberet, peperisset bona (Plaut. Merc. I. 1. 76); but it does not occur elsewhere in Livy, and does not seem to have been used in this sense later than Plautus. reading, or, as is more probable; adco is a gloss on eo, which V has substituted Perhaps usque adco is the true for the whole phrase wsque eo. V. 44. 7. St uobis in animo est tuert moenta uestra nec pati haec omnia Galliam fiert . . For Galliam, the reading of the Nicomachian MSS. V has a Gallis; this has led to various conjectures, as Gallorum by Cobet, a Gallis conferri by Zingerle, and a Gallis ferrt by Frigell. The reading of the Nicomachians exactly suits the sense—‘ not to allow all these places to become Gaul.’ V shows a weak corruption of this by somebody who did not understand Galliam ; for, if a Gallis had been the original reading, it is very improbable that it would have been altered to Galliam. The evil influence of the reading of the Palimpsest is seen in the many conjectures which have been . capite arma. based upon it. V. 46.4 Vets interim non anim tantum in dtes, sed etiam uires crescebant is the version of the Nicomachian MSS.; but V has numerus etiam uiresque for etiam uires, in which numerus is a manifest gloss on wives, betrayed by its position between sed and etiam, and que is an addition to emend the sentence. V.5 homines J quoqgue et factet qua et after fe version t by the ac that two Habitaues distinct i V. 5 recuperate For which ha VIL: Volscorun After antithesis moentbus, (0) III. adortus. HRL ara amens, an IV. placeant 1: For quam, F 4 are prese reading. though tl and expré V. 4 Romant b For. the rest c the introc tantis prt: pristinis C these con to beneficr gives little while rated . true Livy, , the nples fecit. fectus ura ura glos S f the more been Sa rd to plebi PFU lerces Merc. have true tuted mnie shian n by iding es to d not very ce of been mit 1S ve for yy its nce. I i a ms es 7 y E oh > mg 3 e: 7 % re 3 He THE VERONESE PALIMPSEST 175 V. 51.3 Cum... capta tota urbe Capitolium tamen atque arcem diique et homines Romani tenuerint habttauerint, wictortbus Romanis, recuperata urbe arx quoque et Capitolium deseretur, et plus uastitatis huic urbt secunda nostra fortuna factet quam aduersa fecit ? is the reading of the Nicomachian MSS.; but V adds et after tenuerint. It is clear, as Dr. Conway points out, from the Nicomachian version that habitauerint is a gloss on tenuerint—a fact which V has disguised by the addition of ef. It is unnecessary to suppose, as some editors suggest, that two verbs are required to balance deseretur and plus uastitatis . . . factet. Habitauerint is a mere repetition, while deseretur and plus uastitatis give two distinct ideas—the desertion and subsequent waste of the city. V. 51. 3 Qutd obsessam (sc. urbem) ex hostium manibus eripuimus, st vecuperatam tpst deserimus ? For ipst V has obstdione, which is clearly an adscript to explain recuperatam, which has displaced 7f/si—a word important for the emphasis of the sentence. VI. 2.11 1tta consternautt hostes, ut minor moles superantibus uallum in castra Volscorum Romants fuertt. After uallum V adds militibus munitum, which, besides a very un-Livian antithesis is grammatically impossible ; for munire is used with fossa, uallo, moentbus, Alpibus, muro, and the like, but not with mzlitibus. (b) Passages in which the Veronese Palimpsest 1s in agreement with some of the Nicomachian MSS. III. 44. 4 hance wrginem ... Appius amore amens pretio ac spe perlicere adortus. Thisisthe reading of VPFUO. For amens, MVorm. have ardens amens, HRL ardens, M® ardens et etus mentem. It is obvious that ardens is a gloss on amens, and V agrees with some of the Nicomachian MSS. in not recording it. IV. 24. 6 ut ve ipsa, inqut, sciatis, Quirites, quam miht diuturna non placeant 1mperia, dictatura me abdico. For placeant M has placeant re, and V has placeve with the omission of quam. Here we have a very clear instance of variant readings both of which are preserved in M. Mommsen considers that V has preserved the original reading. But it is difficult to say which is more likely to be the original, though there can be no doubt that the Nicomachian version is more forcible and expressive of the feelings of the dictator on this occusion. V. 44. 3 Vobts autem, Ardeates, fortuna oblata est et pro tantis popult Romant beneficus . . . gratiae referendae. For pro tantis popult Romani benefictis, which is the reading of V and H, the rest of the Nicomachian MSS. shows a variety of confusions produced by the introduction of glosses into the text: pro tantis pro latins piy RDL; pro tantis pristinis pir PFED*; pro tantis pristints p.r. U; pro tantis benefictis p.r. pristinis O ; pro tantis prolatis fr M‘A; pro tantis prolatis pir D®. The origin of these confusions seems to have been the incorporation of fristinis, an adscript to benefictis, into the text, and then in some MSS. latinis has displaced it by 176 CLARA M. KNIGHT confusion perhaps with Priscis Latinis, or, it may be, by a miswriting of tantis. VI. 3. 3 Sutrium is besieged by the Etruscans (paucitas oppidanorum) ... inevmis cum singults emissa uestimentis nuserabtli agmine penates relinqueret, eo forte tempore Camillus .. . interuentt. has inervmis and emissa; HT also have relinqueret, which is wanting in V. The other MSS. have confused the singular and plural in this passage, and there- fore do not affect the question as to whether we are to read the singular or the plural. The narrative indicates that all the citizens as a body left their homes. Hence the singular referring to paucitas oppidanorum is better than the plural. VI. 6.7 sibique destinatum in animo is the reading of all the Nicomachian MSS., except L, which omits zz. V has zd for in, and it is accepted by Momm- sen, who refers to IX. 16. 19 (Eum parem destinant animis magno Alexandro ducem) and others, who think it supports L in the omission of 7m. On the other hand, at XXI. 44.9 we find in Puteanus bene fixum omnibus destinatum in animo est, which is accepted by Heerwagen ; but Luchs reads here destinatum- gue animo, etc., with a later MS., as Professor Walters points out to me. In any case, in the passage under consideration 7d is very inelegant and unneces- sary, and probably one of the many confusions of small words, which are found in V, and really represents an original zz. This would put V on the side of the Nicomachians, which read in antsmo. In this passage V, in agreement with HT, (c) Passages 1n which all the Extant MSS. record Manzfest Glosses. sv. 87. Veientes defecere is the version of the Nicomachian MSS., while V omits ac It is quite clear, as the late H. J. Miiller observed, that all the MSS. V has Colonia Romana ad Lartem Tolumnium Veientium regem ac Veventes. have incorporated a gloss on Lartem Tolumnium, viz., regem Vetentium. made the further mistake of allowing it to displace ac Veientes. V. 4.6 patiatur se ab domo ab re familiart . . . abesse is the reading of all the Nicomachian MSS. except U. The latter in agreement with V gives ac re familiari for ab re familiart. Many editors accept the latter version; but it is very probable that ab domo is merely a gloss on ab re familiari, as Dr. Conway has suggested, which has been put into the text, and that in VU an attempt has been made to emend by changing ab to ac. Such an emendation would be very likely in U, and, as we have seen, not at all impossible in V; its only result is to conceal the error and so to render V and U less valuable than the other MSS. in this passage. V.5.5 wuallum fossamque, ingentis utrumque rem operts, per tantum spatiu duxerunt is the version of the Nicomachian MSS. V has utramque rem, which some critics accept. It is probable that the true reading is utrumque, as Drakenborch suggested, and that wtramque rem isa gloss on it. Hence the Nico- machian MSS., which record a combination of both, are more useful than V. V. 32. 3 fusa concursu primo actes, in fugam milia octo armatorum ab equiti- bus interclusa positis armis in deditionem uenerunt is the reading given by the SS TT TT: em ae ——- ey i Foe ae pert i Nicomac It is prol or, as Pr V has ex sentence MSS. in V. 3 Clusinum of all the is Is omit many ed original 1 then recc so often. parenthe IV. If tk preservec edition o: in numbe have bee: most cast Nicomac passages However primary | In s additions the follov III. the read ITT, 31. 2 V, to whi that the | these sho omitted t miliume sc profectus ; wrongly ; Sometim: as in the ng of sr ‘et, €0 . HT, The -here- or the omes. ral. chian omm- andro in the min atum- In 1eces- h are e side 7m ac its ac MSS. V has of all ; ac re it is nway em pt ld be only n the spati vem, ue, as Nico- 1 V. equiti- yy the a es LRT: (RES TL a eee. we eae THE VERONESE PALIMPSEST 177 Nicomachians. V adds wersa after fugam—a reading accepted by most editors. It is probable either that 7” fugam is a gloss on fusa, and should be excluded, or, as Professor Walters has suggested, we should read in fuga. In either case, V has expanded the corruption into im fugam uersa in order to fit it into the sentence better. Hence V again seems to be less useful than the Nicomachian MSS. in revealing a gloss, or guilty of an addition. V. 33. 3. tmuexisse in Galliam uinum inliciendae gentis causa Arruntent Clusinum tra corruptae uxoris ab Lucumone, cui tutor is fuerat ipse is the reading of all the Nicomachian MSS. except M, which has es, a mere slip for is. In V is is omitted, and zpse precedes fuervat—a version which has been accepted by many editors. It is very probable, as Dr. Conway has suggested, that the original reading was 7s, and that ipse is a marginal addition ; the Nicomachians then record both readings, while the Palimpsest has chosen the wrong one, as so often. Jpse makes the sentence too emphatic, whereas 1s makes it merely a parenthetical clause, as, in fact, is required by the sense. IV. PASSAGES IN WHICH THE VERONESE PALIMPSEST ALONE HAS PRESERVED THE TRUE READING. If the following passages, in which the Palimpsest alone seems to have preserved the true reading, be compared with the list at the end of Mommsen’s edition of the Codex Veronensis, it will be seen that they are considerably fewer in number. In some instances this is due to the fact that the readings of V have been found to be supported by some of the Nicomachian MSS.; but in most cases it is because further research in, and a better understanding of, the Nicomachian MSS. has put V into a less important position. In many of the passages in which V is the only MS. authority, it merely confirms conjectures. However satisfying such readings may be, they are not revolutionary, nor of primary importance. In some instances V preserves the true reading by omitting glosses and additions which occur in the Nicomachian MSS. The examples of this are the following : III. 29.6 eo dte L. Manilio Tusculano adprobantibus cunctis ciuttas data is the reading of V, but the Nicomachian MSS. add est after data. Again in III. 31. 1 annona propter em laboratum is the version in V, to which the other MSS. add est after laboratum. There can be little doubt that the Palimpsest has preserved the true reading in the omission of est in these short sentences. Many instances occur in Livy, in which es¢ is rightly omitted by all the MSS. as, for example, VI. 22. 8 quattuor legionibus quaternum milium scriptis, exercitu indicto ad portam Esquilinam in posteram diem, ad Satricum profectus ; but far more numerous are the passages in which est, etc., have been wrongly added, though without the authority of MSS. they cannot be omitted. Sometimes the addition is betrayed by its position varying in different MSS., as in the following examples: CLARA M. KNIGHT 178 IV. 6. 8 deducta est res PF'UODLA; deducta res est M; res is omitted by H, probably having been displaced by the addition of est. IV. 31.9 Rowmae terror ingens erat. H has erat before ingens ; erat should probably be omitted (cf. III. 22. 2 for the omission of erat). VI. 31.7 causa erat MHTLDA; erat causa PFUO. X. 18. 3 est Romam perlatus M* (or M'); Romam perlatus est P?FUO; est omitted by PDLA. In all these examples parts of the verb ‘to be’ seem to have been added wrongly. III. 56. 12 tribunt eodem foedere obligatos se fateantur tollendae appellationis is the reading of V; the Nicomachian MSS. add causa after appellationis. This is perhaps V’s most brilliant reading. This idiomatic use of the genitive for purpose was often misunderstood by scribes, who changed it to the dative case or added causa. IV. 11.7 wexationes ad podie dicta ab trix bunis re>manendo in cColonia>, quam testem i>1aeque habtauere is the reading of V. The Nicomachians add colont adscriptt (with some palaeo- graphical variation as M coloni adscribiti, and D colonis adscripti) before rema- nendo; this is an obvious gloss. IV. 14.6 Ahala Seruilius obtruncat respersusque cruore is the reading of V; the other MSS. add obtruncati after crwore—an obvious gloss. IV. 21.7 Fidenates, qui se primo aut montibus aut muris ten is the The Nicomachian MSS. add aut oppido after primo. Madvig had Hence V supports conjecture. . « domum reductus is version of V. observed that this was a gloss on aut muris. IV. 24.7 deposito suo magistratu, imposito fine altert . the reading of V; but the Nicomachians add modo aliorum magistratui after magistratu which is manifestly a gloss on fine altevi. In this passage V again confirms Madvig’s conjecture.. Gronov suggested that modo aliorum alone should be omitted. V. 31. 5 0b quae Volsinienses Sapienatibus (sc. Sappinatibus) adiunctis superbia inflati ultro agros incursauere is the reading of V. The Nicomachian MSS. add Romanos after agros. The latter, being plural, seems to mean ‘open fields,’ ‘open country’; Romanos is then a gloss added probably from ‘in agro Romano,’ which occurs a few lines above. Though not an impossible addition, it appears quite unnecessary to accept it as the original reading. V. 53. 3. etiamsi tum migrandum fuisset incolumt urbe, nunc has ruinas velinguendas non censervem is the version of V; the Nicomachian MSS. add {tote after incolumi, which seems to spoil the complete antithesis between ‘ in- columni urbe’ and ‘has ruinas.’ Further, if tota was in the original text, it must have been an emphatic word, and one would expect this to be made clear by its position in the sentence. Moreover, tota might very well be an addition by a scribe or reader, who remembered that the Capitol was still safe. In a few passages V supplies omissions in the Nicomachian MSS. : : . rrr cere es 4 ‘ 3 IV. 7. Monetae ec has ad mi POHLDA iv. 2 timentes is | is omitted, due to qu examples hort: for co quere. Cf. V. 43. raptum oni but it is ne Gronov. In oth error of th the minusc IIT. 1: have deuo, deulo emer Campanus ITI. 21 IIT. 2: machian M ture of Rut III. 61 In add or partly in III. 29 the other ) III. 34 lict priuatig quae for qui reject est, w to quae on z this passage ITI. 67 mina, etc., | V.9.2. For equis su O; equos st Even before Suls merert i tted yuld JO; lded L0NIS This for itive endo ; the aeo- ema- TVs ; the had MS 1S after gain lone wnctts eing gloss Dove. pt it unas | tota ‘ in- xt, it clear ition THE VERONESE PALIMPSEST 179 IV. 7.12 Luicinius Macer auctor est in foedere Ardeatino et in lintets libris ad Monetae ea inuenta. Ad Monetae ea is Mommsen’s conjecture from V, which has ad monetea. The Nicomachian MSS. have no trace of ea (ad monetae POHLDA; ad monete MF ; ammonet VU). IV. 25. 4 famem quoque ex pestilentia morbo implicitis cultoribus agrorum timentes is the reading of V. In the Nicomachian MSS. quogue . . . wmplicitis is omitted, which leaves an impossible sentence. The omission is probably due to guoltoribus being written for cultoribus in the archetype. Several examples of gu for c occur in the Nicomachian MSS., as in IV. 3. 2 M has quo hortt for coorti, in 1V. 15.7 MO have conquoquere and H has conquere for conco- Cf. pequnia, etc. on Latin inscriptions. V. 43-4 frumentum incendis urbis absumpserant, et ex agris per eos ipsos dies vaptum omne Veios erat is the reading of V; the Nicomachian MSS. omit eos ; but it is necessary to the sense. In this passage V supports the conjecture of Gronov. In other passages where V is the only authority for the true reading, the error of the Nicomachian MSS. is such as might easily have occurred in the the minuscule, but not in the uncial script. Among these are the following : III. 13. 10 deuto quodam tugurio is the reading of V. For deuio M Vorm. have deuo, and the rest of the Nicomachian MSS. have de ullo (i.e. probably deulo emended to de ullo). In this passage V supports the conjecture of Campanus and Rhenanus. III. 21. 12 fiunt wt in V; but wt is omitted by the Nicomachians. III. 23. 6 exercitu veducto castra locat is the reading of V. The Nico- machian MSS. have relicto for reducto. In this passage V confirms the conjec- ture of Rubenius and Crévier. III. 61. 12 For leuibus, the reading of V, the Nicomachians have lentbus. In addition to the above instances V has preserved the true reading wholly or partly in the following passages : III. 29.5 carmine triumphal et solemnibus tocis is the version of V. For et the other MSS. have ex, probably on account of the following ablative. III. 34.6 decem tabularum leges perlatae sunt, qui nunc . . . fons ommis pub- The other MSS. have quae for gut ; but it is very harsh, if est is retained, and there is no reason to reject est, which is found in all the MSS. Qu might easily have been changed to guae on account of the proximity of /eges and its long distance from fons. In quere. lict priuatique est turis is the version of the Palimpsest. this passage V supports the conjecture of Rhenanus. ITI. 67. 6 mina, etc., is the reading of V. discordia ordinum et uenenum urbis huius, patrum ac plebis certa- The Nicomachian MSS. have est for et. V. 7.13 tum primum equis suis merere equites coeperunt is the reading of V. For equis suis merere the Nicomachians show much confusion, as eguis merere O; equos si merere H; aeguos merere P; equos merere P*FU ; aequis merere L. Even before the collation of V equis suis merere had been conjectured from equts suis merert in the periocha. 180 THE VERONESE PALIMPSEST V. 24. 8 Ceterum (ceteram H) partem plebr (plebis H and U) partem (partim H) senatus destinabant habitandos Vetos. This is the reading of the Nicomachian MSS. except that P has a habitan- dos, and U has parte... parte. V has partem plebs, partim senatus, and habitando distina.... It is uncertain whether V had Vezos; but there seems to be sufficient space for it. In any case it is probable that Vezos is a gloss and the correct reading is habitando in V’s order. The order in the other MSS. is probably due to the addition of the gloss. There can be little doubt that V’s partim .. . partim, which also has some support from H, is right. For the rest, it seems best to read plebi and senatui with Rhenanus and Madvig, and destinabatury—an impersonal, which is also needed by the following words (duasque urbes communis rei publicae incoli a populo Romano posse). The destinabant of the Nicomachians is perhaps due to the error in partem. For the confusion of active and passive cf. VI. 4. 8 ne quid ad noua consilia relin- queretury uivium in which passage V has relinqueretur, but the Nicomachian MSS. have relinquerent. V. 53. 1 Camillus tries to persuade the Roman populace not to leave Rome and migrate to Veii—at enim apparet quidem >ma nec ullis piaculis expiari posse, sed res ipsa cogit uastam incendis ruinisque relinquere urbem et ad integra omnia Veios mgrare. This is the version of V except the two words, which are not clear, and which in the above are due to the conjecture of Madvig. All the Nicomachian MSS. omit at enim ... posse. There can be no doubt that V is correct here. There is, perhaps, no passage in which the Palimpsest has proved more useful ; for without it an irremediable lacuna must have remained in the text. CLARA M. KNIGHT. KinG’s COLLEGE FOR WOMEN ’ KENSINGTON SQUARE, W. 1 See Classical Quarterly, January, IgII. ON T THE s increasing]; Greek mor insisted on years, I ma to be indis acknowledg with Moder Thus, of Modern but Profess article appe been useful remains out will unfortu with this ne I think a good kno the charact methodical by some col not only to of future in whereas D1 Oxford, and to folklore a But be wish to say years of co selves with 1A lecture University of h 2 Gercke un Altertumswissen: 3 On the Valu rvtem ttan- ando o be the 5. is V’s the and ords The For ‘elin- {SS. eave ullis vbem and hian ere. oful ; iT. ON THE VALUE OF MODERN GREEK FOR THE STUDY OF ANCIENT GREEK. THE study of Hellenistic Greek, or of the «own, which has flourished increasingly since the beginning of the present century, has brought Modern Greek more and more within the view of classical philologists. As I have insisted on utilizing Modern Greek for Hellenistic philology for about twenty years, I may claim some credit if a knowledge of Modern Greek is now admitted to be indispensable to Hellenistic studies; but philologists only reluctantly acknowledge this new demand, and hesitate to acquire as full an acquaintance with Modern Greek as the study of the xown} demands. Thus, for instance, the linguist Paul Kretschmer? emphasizes the value of Modern Greek to students as an important source of linguistic information, but Professor Gercke, one of the editors of the work in which Kretscnmer’s article appears, states in the same volume that although Modern Greek has been useful in elucidating many problems of vulgar or Hellenistic Greek, it remains outside the interests of most classical students. Hence young students will unfortunately be inclined to conclude that they need not burden themselves with this new study, even if good opportunities should come in their way. I think I may claim to have proved by my own scientific work how much a good knowledge of Modern Greek contributes towards a right perception of the character of Hellenistic Greek, and I therefore venture to discuss the methodical value of Modern Greek philology to-day, and to illustrate its fruits by some concrete examples. The theme of my lecture will give me occasion not only to review the latest results of investigation, but also to trace the lines of future inquiry. I shall restrict myself to the linguistic side of the subject, whereas Dr. Menardos,® the Reader of Modern Greek in the University of Oxford, and the Dutch Neogrecist, Hesseling,* have extended their investigations to folklore and literature, treating the language itself but lightly. But before entering upon the discussion of my special subject, I should wish to say a few words as to the interest which the study of three thousand years of continuous linguistic development has for those who concern them- selves with the science of language in general. Only the Indian and Chinese 1A lecture delivered before the Victoria Hellenic Study, Oxtord, 1909. University of Manchester, October 13, 1913. 4 Det Betekenis van het Niew Grieks voor de 2 Gercke und Norden, Einileitung in die klass. Geschiedenis dey Griekse Taal en dey Griekse Letter- Altertumswissenschaft, 1? (1912), p. 474. kunde, Leiden, 1907. 3 On the Value of Byzantine and Modern Greek in 182 A. THUMB languages have a documentary history extending over so long a period. ‘The great gain from the study of contemporary Greek is the perception of the whole as a continuous and living language,’’ says Menardos concisely.! As in the development of human speech, transformation or change alone is constant —tavta pet, Says Heraclitus of Ephesus—it is of the highest importance for the knowledge of the laws of linguistic development that we should be able to observe the transformation of a language for as long a time as possible. For instance, the decay of the verbs in -y takes place before our eyes; we see it begin with Homeric forms such as devxvvover, continue through Hellenistic ioT@, TUM, KkaGouat, until Modern Greek is reached, in which eZywac alone remains from the rich Indo-European formation. In Latin the same process was completed before the beginning of written records. ‘Itacism,’ from its beginning with the transition of e into é in the ancient Corinthian dialect (6th century) to its completion with the transition of 7 and v to i, lies entirely within the textual tradition of Greek. The process was at work for about 1,500 years, 7.¢., from 600 B.c. to A.D. goo, as regards common Greek; but even to-day it is not entirely complete, for 7 and uv have still in some dialects their old values as e and wu. There are philologists who are accustomed to characterize all modifications of classical Greek as due to the decline or deterioration of the language. How wrong such an opinion is appears from the fact that some tendencies of classical Greek have only attained adequate expression in the modern language. So, for example, the old dualism expressed in the two forms éypadov éypawa only reaches completeness in the modern differentiation of the future into the two forms 0a ypadw and 0a ypdayrw. Modern Greek, accordingly, shows that the distinction between imperfective and perfective action, far from vanishing in post-classic times, has remained sufficiently vigorous to create new forms supplementary to those of classical Greek. If the development of a language can be traced through three thousand years, we may expect to find that in the succession of phenomena certain processes are at times repeated, and observations of this kind are of special importance for the study of the laws of linguistic development. We know, for instance, that in prehistoric Greek, long before the time of our textual tradition, the palatalization of q*:, g*e, has led through ci, ce to tu, Te (e.g., in tis, Té, wévte), and that this process repeated itself at the end of the classical period, spreading over a great part of the Modern Greek dominion to an extent which cannot at present be exactly delimited. In Crete and other islands, in the Peloponnesus and elsewhere, kz, ke has become ¢1, ¢e, tst, tse, and the like, exactly as in the Romance languages.” But whereas in the Romance languages this sound-change is almost universal, it is restricted in Modern Greek to particular dialects; and I mention it because just in the same way the prehistoric change did not establish itself in all dialects; for instance, the 1 The Value, etc., p. 19. Vernaculay (Edinburgh, 1911), § 17. 2 A. Thumb, Handbook of the Modern Greek i a Sr Rh Aeolic gr (Sidwor, T and the s the speec words lik also we fi interval probable (=old Ir a comple way the 3rd pl. é¢ more exa! occur nor has show in épope the form i.€., by t examples A de Old and | of the ger inquiry o about thi Such a ¢ and Gern over long Evenas transform much slo groups. The are less | accented language Greek ha declensio piror, pi) anuct ag 1 See Bn 2 See Ha 3 Korting (Berlin, 18q failure. K., xxxix, 87 sc The f the \s in stant e for able sible. e see 1istic ilone Ocess n its alect ‘rely bout but ects tions How sical So, only - two t the 1g In orms sand rtain ecial now, xtual y., In sical and logist, st not years, rreece Ve see battle 25 ac- t. A ifteen jovea, rmers living > long past, selves place ill used ypen to e their mitting test to : | ON THE VALUE OF MODERN GREEK 185 once held by Latin, as the language of monks and scholars. For us in ‘Europe’ this struggle for a really modern language belongs to the past. In Greece we see before our eyes how progress struggles with tradition, and how the national spirit is roused by an endeavour to introduce vernacular forms and words like tatépas for matnp, vepo for tdwp, omits for oixia, or to banish the dative from the written language. The study of this process teaches us not only how a new literary form of speech arises, but also what psychological factors, national or individual, prevent or promote such a It enables us to realize what difficulties the great Dante had to face in elevating the lingua uolgaris to the literary rank of Latin, and why the language of the great reformer Luther was so long in winning its victory over German dialects of North and South. And we see from the Greek parallel that such literary innovations came not only through academic discussion, but also through passionate strife. The present opportunity of observing the struggle for a new literary dialect illustrates in many details similar conflicts in the past, not everywhere perhaps so passionate as in modern Greece, where this literary question has incited men to political revolution, and even resulted in bloodshed. The new literary idiom is based on colloquial Greek, but forms also a higher unity, above the dialects. Greece had already solved this problem when Hellenistic Greek developed out of Attic, and united the whole of Hellenism in a common spoken and written language which finally superseded the old dialects. The study of this process, the rise of the so-called «ow, receives useful suggestions from the history of Modern Greek. So I come to my special theme to which I wish to call your attention, and proceed to inquire what is the value of Modern Greek philology for the study of Ancient Greek. Modern Greek research proceeded upon wrong lines until thirty years ago,’ and in con- sequence its importance for Ancient Greek was first sought in a wrong direc- It was believed that Modern Greek was a descendant of Aeolo-Doric dialects, and even that certain Modern Greek sounds and inflexions must be Thus, for instance, every a where Attic process. tion. derived from prehistoric Greek types. had » was thought to be Doric, the nom. and acc. plural tiwés = Tipat(-as) was identified with Aeolic tipais=tias, the old digamma was everywhere recog- nized where Modern Greek £8 or y or another consonant did not clearly corres- pond with identical sounds in the ancient language. In (dialectic) forms like ayaTraeis for ayaras the original uncontracted form was seen ; and in cases such as ép@Taya=ancient épwtwy the genuine Indo-European type was believed to survive. A scientific fantasy which was guided by no method regarded Modern Greek and its dialects as a help in recognizing unknown dialectic forms of old Greek, and even forms from prehistoric and pre-Hellenic times. Scholars affected by this Archdomanie, as it has been called, never examined or explained the absurd historical premisses on which those grammatical derivations and 1 See Thumb, Die neugriech. Sprache, Freiburg i. B., 1892. 186 A. THUMB speculations rested. European and Greek philologists revelled in the great antiquity of Modern Greek; and when the truth about its origin had already been stated, a Greek" still praised its Homeric character. Even to-day there are occasionally Greeks who surpass their predecessors, if possible, and make the most absurd assertions as to the archaic character of their mother tongue ; but they are amateurs, of no scientific importance, who are no longer taken seriously. For we have long known that Modern Greek with its dialects descended from the «own or Hellenistic common speech. This statement has changed the point of view from which the relation between Ancient and Modern Greek must be considered. In the first place, the present-day view of Modern Greek has a negative side; our knowledge of the ancient dialects cannot be enlarged by the study of the modern tongue, at any rate not on the basis of clear and undisputed facts. The traces of ancient dialectic elements are too unimportant, and their recogni- tion in Modern Greek depends too much upon the ancient traditional dialec- tology for us to draw any certain information from studies of this kind in the It is possible enough that a digamma should live on in the 8 of a rare dialectic word or place-name, but, apart from Tsaconian, only the place-name Boirvdro (a village in Maina) eludes critical examination ;? for Borxos ‘ kiddle’ (Western Peloponnesus) recently supposed to come from an ancient *Fodxos,* has its digamma disproved by the fact that oA«os and €\«cm in antiquity show no trace of it. On the other hand, an old digamma cannot be proved by the M.Gr. word because a 8 in other words also appears spontane- ously before an initial o.* It is therefore not worth our while to look in Modern Greek for traces of ancient dialect phenomena. The search would be a failure, and sheer waste of energy, for the direct way to the sources is shorter and surer. There is only one exception. The Tsaconian, spoken in the Parnon mountains, is the only descendant of an ancient dialect, and we may characterize it as a mixture of old Laconian and Hellenistic elements.° In this single case we can reconstuct an ancient dialect, starting from a modern one; and with the additional help afforded by tradition we can delineate the features of the later Laconian. As I remarked years ago,® it would be worth while to apply Tsaconian to the examination of the text and language of the Laconian psephism in Boethius, De Institutione Musica, I., 1, a forgery by a grammarian of the Imperial age. Professor von Wilamowitz, through neglecting the help of Modern Greek, has failed to advance the linguistic interpretation and reconstruction of this text.’ Further, Tsaconian gives us some knowledge of old Laconian characteristics modern language. 286, also expressed doubts.) 4 See Hatzidakis /.c. 5 See Thumb, Handbuch der gyriech. Dialekte (Heidelberg, 1909), 90 sqq. 6 Hellenismus, pp. 36sq.; most recently Indog. Forsch, xxxili. 298. 7 Timotheos, Die Persey (Leipzig, 1903), 69 sq. 1 Kovrémovnos, "APavacia ris EXXnvixys yAwoons, } avevipeois THs Ounpixkns yrAwoons év rats dnuwdece diaréxros, Athens, 1884 (2nd edition !), 2 See most recently Hatzidakis’ Meca:wyixa kai véa ‘EXXnvixa II. (1907) 316 sqq. 3 See Txiads in the ’Evernpis rov Ilavemiornuion, 1911, 217s8qq. (K. Dieterich, Byz. Zettschr. xxi. a 7 fe prese it not tradit 1u—a Laco ] paris conce the s gaps of th can fe ] resea Latin Greel follov repre rapid with ] in th almo: U3 al quant of the short. two a 7 to the Here unacc €l ani Greel since bet we about 4 The « 1 ku 2 Fo Greek de Chir Gramm great ready there make ngue ; taken alects it has - and rative idy of facts. -ogni- lialec- in the in the ly the 2 for m an K® In 1ot be itane- ces of ste of ; only only ure of istuct | help As o the thius, | age. c, has text.’ "istics Dialekte ' Indog. 69 sq. ea arama. ON THE VALUE OF MODERN GREEK 187 preserved in the later Laconian; thus Professor Hatzidakis has shown? that it not only proves an old pronunciation of v as u, and so confirms the ancient tradition, but also proves that under certain conditions Laconian made v into 1u—-a fact previously known only for the Boeotian dialect, and quite new for Laconian. But that which we may glean from Tsaconian is quite trifling in com- parison with the rich harvest of information that Modern Greek yields concerning Hellenistic. The modern language enables us, first, to deduce the spoken «own independently of the ancient sources, and so to fill up the gaps in our textual tradition. Then it renders possible a critical examination of this textual tradition. And so by the combination of these processes we can form an accurate conception of that important period of Greek. Beginning with the first point, we may observe that our scientific research has the same task and the same method as the study of Vulgar Latin by the aid of Romance philology. Since we may assume that Modern Greek dialects began to differentiate markedly between A.D. 500 and 1000, it follows that the deduced common basis of Modern Greek and its dialects represents Greek as spoken about the end of antiquity. Let us see now ina rapid review how Later Greek appeared about A.p. 500, if we reconstruct it with the help of Modern Greek alone, renouncing the aid of textual tradition. 1. The vocalism had already reached the stage known as ‘ Itacism,’ found in the language of to-day: es and 7 had become 7 everywhere, and so had almost everywhere; oc and v were pronounced alike, partly as z and partly as ii; at had the same sound as ¢, and ev, av were either ¢v, av, or ef, af. Vowel quantities were levelled: all accented syllables, whatever the original quantity of their vowel, became unifcrmly half-long, while all unaccented syllables were short. The tone difference between acute and circumflex had disappeared, the two accents denoting essentially stress, not pitch.? That the state of vocalism presupposed by Modern Greek really belongs to the close of antiquity is proved by the Northern Greek dialects of to-day. Here every unaccented 7 and « has been reduced or has disappeared, every unaccented e and o has become? and uw. In these changes the ancient 4, 7, v, ec and os are alike treated as 7,¢ and at as ¢,o and w aso. The Northern Greek changes of vocalism imply that Itacism had become complete. And since we may assume that these developments in Northern Greek started between A.D. 500 and 1000, the stage of Itacism must have been fully reached about the end of antiquity. 2. The consonantism too has the same aspect as that of Modern Greek. The old ¢, x, @ have become f, ci, p (except that of is or, x? is yt, and ¢@ is 3 See Hellenismus, p. 165, Prinzipienfragen, der Kown-Forschung, Neue Jahvbicher f. d. ki. Alt., 1906, pp. 258 sq. Kretschmer, however (Der heutig lesbische Dialekt, pp. 5 sq.), is not certain as to the antiquity of the Northern Greek phenomena, ! Kuhn's Zeitschrift, xxxiv. 81 sqq. 2 For a certain musical element in Modern Greek accent, see H. Pernot, Phonétique des parlers de Chio, 50 sqq.; also Brugmann-Thumb, Gy. Gramm., 176 sq. 188 A. THUMB or), and B, y, 6 are v, 3, @, except in the combinations pf, yy, v6, where the mediae remain unchanged. The tenues as a rule suffered no change, but they were softened after a nasal (as in €uzropos, évtepa, avayxn), and became fricatives before + (xAémtns becoming KrAeHrns, OxTH OTH, etc.); the nasals disappeared before ¢, y, o, as vidn, Kworavtivos; € and o before a voiced consonant were sounded =, as in Koopos. 3. In the inflexion of the nouns may be noted the loss of the dative, and such paradigms as yA@oo-a, -as, -av, -€s, -@v, -as, OF -e5; €ATIO-a, -as, -av, -€S, -WY, -asS, OF -€9; TONAL, TONS, TWOAL ; VAaUT-NS, -N, -NV, -ES, -@V, -ES, OF -as; TATEP-AS, A, AV, -€5, WY, -€5, OF -as; Wop-ds, ds, -G, Gdes; dreT-od, -ovs, -ovdES. The pronouns éyé and écé had become é€yév(a) and écér(a), jets and twas had been replaced by cets and cas, otros by tovros, and 6éd¢ had disappeared. 4. In verb-inflexion it is characteristic that the present active participle became indeclinable, the infinitive was limited (the substantival use must have retained its position unimpaired), reduplication was lost, and so were the middle aorist, the optative, and the perfect passive. The -y verbs were transformed into verbs in -w, except eiui, which was however conjugated as a middle (eiwar, eicar (€v1), eiueOa, eiobe (évc), with imperfect any, etc.). Subjunctive and indicative flexion coincided, as did that of imperfect, perfect, and aorist (€ypada, 70a, éypawa, evpnxa), the perfect tense receiving the syntactical force of the aorist.? It is however remarkable that édwxav and édo0cav, €Onxav and éGecayv lived on together. The difference between active and middle persisted, though middle and passive forms were levelled, and the subtleties of the ancient middle were lost. The total list of media tantuim verbs was not reduced, for such losses as wapait@ for -odmar, ‘ resign,’ were compensated by innovations like yatpouat. Such is the picture, in a few broad lines, of Hellenistic as presented to us by the comparative study of Modern Greek and its dialects. It is a hypothetical reconstruction, and we ought, if we had no other evidence, to place the asterisk before these assumed earlier forms, and write THs *yA@acas, ot *vadtes (*ndftes), *eiwar (*/me), *évi, *éypada, *7AGa. The comparative method however permits us to regard them as really existing: it is only on their relative age that we might entertain doubts, had we not direct textual evidence, so as to feel hesitation in ascribing them to the end of antiquity. But now the textual tradition, presented most conspicuously in the papyri and inscriptions, affords us immediate evidence that the deduced forms really existed in the Hellenistic age. Thus Itacism in its first stages can be distinctly observed from the third century B.c.; the Hellenistic change of 6 to p is proved for the second century A.D.; the accusative mavpiday and such forms as 70a, éypada have been attested from the beginning of our era; éyouoes=éyovoar from the first century A.D.; the genitive Tod arn\wTn from 1 This statement is made on the assumption that the modern « aorist continues the old perfect. | ”_ * 232erpanamnn..° the se from t the re with c latest claim third : epoch T is as deficie I ma modet the xc been Greek they a and 4 asper— us bas "A poi duly a Tépuo of tw this 11 MeaATi I furthe Helle: which the st Helle: gramr anoth been — Attic better Greek elda= develc testim with ; 1 Ind 2 Ha the but ame isals iced and -ay, -as ; OES. had iple nust vere vere 1 as Reale fect, the ived ugh ‘lent , for 1ons d to is a &, tO ‘Tas, itive only ‘tual uity. [pyr eally 1 be re of and era ; from * Rp ieee eS * 2 f.4 BES Re», ‘a ON THE VALUE OF MODERN GREEK 189 the second century; 7junv, ‘I was,’ from the third century B.c.; eiuas, ‘I am,’ from the fourth aA.p. And if we consider that the Hellenistic texts do not reflect the real condition of the spoken language—on this see below—we may assert with confidence that the xoww7 as deduced from Modern Greek was spoken at latest about the fifth century a.p. Indeed, we might safely go further, and claim that it shows us the condition of vernacular Greek as far back as the third: the speech of the common people had by the middle of the Imperial epoch a very Modern Greek appearance. This result, gained by a comparison of Modern and Hellenistic Greek, is a starting-point for further research, for it entitles us to supplement our deficient knowledge of the «oi from living speech. How we may do this I may illustrate by one or two examples. Years ago I concluded! from modern forms such as adXexdtTn, cpeSBarte (for nrAaxaTn, KkpaBPatiov) that in the «ow unaccented pa, Xa changed into pe, Xe; and these forms have since been discovered in Hellenistic texts. In the same way I attribute Modern Greek 0wpo, ypwota, swe (for Pewpa, xpewoTa, ciw7@) to the xown, although they are not yet attested by any text. The Modern Greek édéros, ‘ this year,’ and adovxpafouat, ‘1 listen,’ are proof for éros and axpodouat with spiritus asper—the first form being attested, the latter presumed. ModAifu, ‘lead,’ takes us back to an ancient worAzBos, which can be supported by written tradition.” ’Apxovda, ‘she-bear,’ proves an ancient apxos beside apxtos, and this form duly appears. Modern Greek axdov8H=dxorov0a, oxdpd0= cKdpodov, mépor= Tépuat, TepBort=TepiBorov are the product of a phonetic law by which one of two vowels flanking p or X falls out; and P. Kretschmer has recognized this in Hellenistic phonology from the occurrence of such words as cxopéon, MeAtivn, depvixn.® If Modern Greek provides us with new facts in Hellenistic, it serves further to establish forms for which there is no certain textual tradition. In Hellenistic texts we frequently find forms unknown to the classical language, which astonish a student who comes fresh from the schoolmaster’s Greek to the study of papyri and inscriptions. Philologists in consequence emended Hellenistic texts without scruple, according to the pattern of classical grammar. One day Polybius was treated in this manner by Dutch scholars; another day it was inscriptions or papyri; even the Byzantine authors. have been ‘corrected’ by the philologists of Bonn in accordance with the same For these philologists thought they understood Greek But one who knows Modern Attic prescriptions. better than the Greek authors themselves. Greek looks at such forms as aAXes= GAXas, euév = ue, yeyovay=yeyovact, el0a=etdooy with different eyes: he sees therein the earliest monuments of the developments that issued in the Greek of to-day, and collects them as valuable testimonies which he tries to explain, but refuses to correct like a schoolmaster with a pupil’s exercise. There are, no doubt, in papyri and inscriptions many Thumb, Gr. Gramm., pp. 156, 160. ’ Brugmann-Thumb, /,¢., p. 80. 1 Indog. Forsch., ii. 85. 2 Hatzidakis, Glotta, iii. 70 sqq.; Brugmann- 190 A. THUMB Not all that is written may be admitted as evidence of living speech. The occurrence of individual mistakes and momentary slips of the tongue must be acknow- ledged, and has of course its interest for the psychology of speech; but it is wrong spellings and grammatical blunders made by uneducated men. unconnected with the actual history of Greek. We must therefore first inquire whether such curious forms are individual or not. As a rule, the repeated occurrence of strange forms proves their linguistic value. If, how- ever, they are isolated, we must seek some criterion to determine whether they really belong to living speech. Here Modern Greek can help us, and it may prevent us from rejecting an isolated textual form. For that which actually lies in the direction of Modern Greek development cannot be a mere slip of the pen or an individual blunder: it must be evidence of the linguistic condition of its age. Thus, for instance, the nominative 7) yis for yA, or the adjective daoxemos for aoxertos—both of which are found in Graeco-Latin glossaries—are confirmed by Modern Greek 7 yijs, and formations like avéyyuyos for dvéyyixtos. It is wrong, therefore, to conjecture doxozros for aoxetros.” A gloss Bv&w is silently read Bufiv by a Neogrecist (not Biviov, as the editor has it), because he knows Modern Greek Avfi(v), ‘nipple.’ The aorist subjunctive épyw ° for €\@w reminds us of the Modern Greek aorist #pya. Philologists who are shocked by such ‘ barbarisms,’ and neglect them as simple blunders, have the old instinct surviving that only the classical and literary languages are really worth studying. They cannot appreciate Hellenistic Greek as a living development; and Modern Greek moves their contempt, ‘ because it has a7ro with the accusative !’ Of course it is not always easy to form a decisive judgment as to the linguistic value of dubious forms. Hence the need of a close acquaintance with the modern tongue, if we would estimate correctly the relations between the ancient and the modern. External resemblance or identity between a papyrus form and one found in Modern Greek does not necessarily imply a true historical connexion. The numerical téocepis on a Hellenistic inscrip- tion* will not be taken as prototype of Modern Greek réocepe:s; for the former is occasioned by the confusion of 7 and e, characteristic of the «ow of Asia Minor, whereas the modern form is produced by the analogy of tpeis. Again, at first sight the isolated and strange future ei7rocer, unearthed by Radermacher,® might be supposed to have a relation to the Modern Greek eiwOnka and eir@pévos. But a future like esroce having no analogy in the modern language, we must regard it probably as an individual aberration, starting from the Hellenistic 3rd plural e’vrocay, according to the analogy of édovAwoav : dovAwoet, and the like. General rules cannot be laid down for instances such as these; every case must be determined by special examination. 3 Papyri Iandanae (Leipzig, 1912), No. 22— between A.D. 619 and 629. 1 Nachmanson, Beitrdge zur Kenntnis der altgriech. Volkssprache(Uppsala, 1910), has not paid sufficient attention to this point of view. 4 Journal of Hell, Studies, xxii. 358, No. 119. 2 See Principienfragen, p. 249, where further 5 Neutestamentliche Grammatik (Tiibingen, rorr), examples of the same kind are given. p. 82. ; L / Sis ap a f # Sa 74 F 3 ; Bu well att either will ap) in the | As Hellen and int of sing! Greek. and acc not be I vy present exclude fore co! Greek ° to KaKc Testam preting with th Mark \ accordi mother enlight New Ti name i without the lan Kujeas' childre connex! express riddle 11 accomp leather water). Th Greek— least, | 1 In Cl gives anot for Ancie! 2 Pallis ——e TRE! AE SO" j ; 4 ie ON THE VALUE OF MODERN GREEK IgI But what is to be done if strange innovations in Hellenistic are at once well attested by textual tradition, and lacking in support from Modern Greek, either directly or by some indirect precedent? The answer to this question will appear later, when we try to describe the general character of Hellenistic in the light of the modern language. As Modern Greek affords us the means of enriching our knowledge of Hellenistic speech, we shall not be surprised to find it help us in understanding and interpreting Hellenistic texts. Above all it throws light on the meaning of single words. Thus Hellenistic lexicography gains by the study of Modern Greek. Exactly as we use the latter in reconstructing Hellenistic phonology and accidence, we may ask whether the meaning of certain modern words may not be applied to their ancient prototypes. I will first give some characteristic examples from Greek scholars of the present day. In Euripides Cyclops (v. 694) we find xax@s in a context which excludes the usual sense of the word. The Dutch philologist Cobet has there- fore conjectured aAXws ‘in vain.’ The Greek Pallis' reminds us of Modern Greek Tov xdx«ov ‘in vain,’ and the difficulty disappears if we give this meaning to xaxa@s in the passage of Euripides. Pallis, who has translated the New Testament into vernacular Greek, has also explained Mark vii. 19, by inter- preting the word Spepara not as ‘ food,’ ‘meat,’ but as ‘stench, ‘impurity,’ with the Modern Greek meaning.” Again, he has translated uépa evKaspos in Mark VI. 21, not as ‘convenient day,’ but as ‘an empty day,’ ‘a holiday,’ according to the Modern Greek meaning of evxatpos.* Thus a layman, whose mother tongue is Modern Greek, has been able, without scientific study, to enlighten philologists and theologians; for in the really good dictionary of the New Testament by Zorell* both interpretations are accepted, although Pallis’ name is not mentioned. If even a superficial knowledge of Modern Greek, without scientific research, promises such advantages, a thorough inquiry into the language will be still more profitable. A short paper by the Greek Kujeas® furnishes a good illustration. In the Characters of Theophrastus a children’s game is described, and the expression doxos 7réXexvs is used in connexion with the alternate lifting up and setting down of children. The expression has puzzled philologists, but Kujeas has found the solution of the riddle in a saying of his Peloponnesian home (Avia in Laconia). Those words accompany the action of lifting and setting down, and mean ‘light like a leather bottle’ (which floats) and ‘heavy like an axe’ (which sinks in water). | The student of Hellenistic texts must not be influenced by his classical Greek—he had better bend his eyes forward to the modern language or, at least, he must remain unprejudiced. Thus did Wellhausen in one of his 1 In Class. Rev., xix. (1905), 36; Rouse (ibid.) 1913), p. 14. gives another instance of Modern Greek as a help 3 Lb, p. 3%. for Ancient Greek. 4 Novi Testamenti Lexicon Graecum, Paris, 1911. 2 Pallis, A Few Notes on the Gospels (Liverpool, 5 Hermes, XLI., 478 sqq. 192 A. THUMB illuminating commentaries on the Gospels, translating Matthew XIV. 15, 7) wpa mapnrvev by ‘the time is advanced,’ or proposing for the verb vapdyew the meaning ‘to go on.’ The excellent orientalist was guided only by his natural feeling, for he did not know that the presupposed meaning of vapa is directly confirmed by Modern Greek expressions like wapazravo, ‘ higher up,’ rapaxato, ‘further down,’ wapapéca, ‘more inside.’ Again, if we observe, e.g., that the modern meaning of xazpos, ‘time,’ or ypovos, ‘year,’ is already attested in papyri,} we are entitled to project the Modern Greek meaning of words into the period of the xow7, even when Hellenistic texts do not oblige us to do so. The verb tpayovoe, for example, must have had the modern meaning ‘I sing,’ at the end of the classical period, although this sense is not attested by ancient texts; that the noun tpaywées had, in fact, the meaning ‘singer,’ is proved by a Greek loan-word in Syrian, which has this sense. Modern Greek vypov, ‘ water,’ must also have existed in antiquity; it is the contracted form of veapov, and the meaning ‘ fresh water’ is attested by a late-Greek schoolbook. For these reasons I see no difficulty in assuming that the meaning of o pufixos (sc. 00s), ‘reef’ (from pita ‘ root’), which is adopted by Joh. Schmidt in an inter- esting paper on Modern Greek ro pifixo (=Italian risico)*? belonged to the Hellenistic epoch, although at present no direct proof exists. Some groups of words may be mentioned in which Modern Greek can provide direct and decisive evidence towards determining the ancient meaning. The names of plants and animals are particularly suitable for investigation, because the modern vocabulary has been carefully examined in special treatises, and rendered available for research.* A few examples, taken at random, may show how the meaning of ancient names of plants, where it is doubtful, can be elucidated by an appeal to Modern Greek. Ancient Greek ayAaoda@tis, ‘a certain plant,’ according to the Greek dictionaries, is the Modern Greek dyXaodwrtns, Paconia officinalis ; Ancient Greek dpaxos, ‘a kind of lentil’ is Modern Greek apaxds, Pisum satiwum or Lathyrus satiuus, or Vicia dasycarpa ; éxworrovs, ‘a prickly plant,’ is Modern Greek aywvotrodas, Genista acanthoclada ; Ancient Greek Bpvwvyn or Bpvwvia, ‘a creeper or climbing plant,’ is modern Spvewviés, Smilax aspera; xovv§a, ‘a plant of strong smell,’ is modern axovvfa, Inula graueolens ; Naavy, ‘an eatable herb, is Aa(u)Wava, Sinapis aruensis or putescens ; mepdixvov, ‘a plant’ (!) is modern wepdicak, Parietaria; dudixn, ‘a species of Rhamnus,’ is modern durixt, Phyllyrea media or Rhamnus alaternus. the value of further investigation. But I do not know any enquiry into These examples suffice to prove 1 See K. Dieterich, Rhein. Mus., LIX., 234 sq. 2 See Krumbacher, Byz. Zschr., xi. 523. 3 Miscellanea linguistica in onore di G. (Torino, 1901). 4 Cf, Heldreich, Ta Snuwdy dvduara tov puTaov €xd.d. bro Sar. Mndapaxyn, Athens, rgto. A list of plant-names is also given by Wilski in Hiller v. Gartringen, Thera, iv., pp. 119-130. Lundstrom, Botaniska Lexika fvan meddeltiden, Minneskrift Goteborg, 1910, pp. 42 sqq. Krumbacher, Das neugriech. Fischbuch, Sitz.-Ber. d. Bayer. Akad., 1903, pp. 345 sqq. (Compare also G. Schmidt, Vremennik, xX. 603 sqq.) 5 I give the Modern Greek names and their botanical denotation from Heldreich, of. cit. Ascolt ee — ene ce ess Sk RN ee 68S RPE oF Ancient (¢ be pointe and dome With Medieval give us |] Krumbacl Greek wo surprise vu animals o observatio than look I give it i which en certainly © ‘* denté ’’). Philo. scarcity of But they” actually p changes o name of t name of a the burder meaning. this meani KOLVY) Mod Unfor to give us extends. accessible, tions given But di vocabulary Theoretica solution of dialects wk nese or C attribute t. 1A few int out by Kalitsu Spy. XIII. (191 2 Die antike same holds x Neuchatel, 19 a4 ypa the iral TO, the yri,} ‘iod rerb end chat eek ter,’ can ling. ‘ion, ises, cient dern reek cient m or dern a plant table (!) is »dern orove into eskriit r, Das Akad., hmidt, d their tt. eT ee ON THE VALUE OF MODERN GREEK 193 Ancient Greek plant-names in which Modern Greek is utilized ;1 only it may be pointed out that Victor Hehn, in his standard work on cultivated plants and domestic animals, does not quite neglect the Modern Greek plant-names. With regard to the names of animals, the conditions are similar. Medieval Greek works, such as the Sook of Fishes, edited by K. Krumbacher, give us links in the connexion between ancient and modern names, and Krumbacher’s treatise illustrates further how the interpretation of Ancient Greek words is assisted by the evidence of the modern language. It does not surprise us, however, when we find Professor O. Keller, in his book on the animals of the ancient world,? overlooking Modern Greek and Krumbacher’s observations. Manya classical scholar would leave a name uninterpreted rather than look for help in the most promising quarter. I give it in Krumbacher’s words.’ One example will suffice ; ‘The cvvaypis (an ancient name of a fish), which enjoys an unnecessary anonymity in our dictionaries, can almost certainly be identified. It is the dentex uulgaris or ‘‘ Zahnfisch” (French ‘‘ denté ’’).’ Philologists who neglect Modern Greek on these points cannot plead scarcity of relevant literature, for there are several useful works in existence.‘ But they might express doubt as to whether the words we have cited have actually preserved their ancient meaning to the present time. Of course, changes of meaning might occur, just as, for instance, the Indo-European name of the beech, *bhdgos, Old German buocha, Greek dnyos, became the name of a species of oak in a prehistoric epoch of the Greek language. But the burden cf proof lies on those who doubt the identity of ancient and modern meaning. Ifa Greek word has the same meaning in all the modern dialects, this meaning must hold for Hellenistic Greek, exactly as we attribute to the xowyn Modern Greek sounds and inflexions under the same circumstances. Unfortunately we do not yet possess the great Modern Greek Thesaurus, to give us information about the areas over which the use of the rarer words extends. Still, the lexicographic material, which is already known and easily accessible, will serve at least for a beginning, as the examples and investiga- tions given above have shown. But does Modern Greek afford us any possibility of adding to the ancient vocabulary quite new words, which are not attested at all in antiquity? Theoretically, the question must be answered in the affirmative, yet the solution of the problem is not so simple. For only if a word is found in dialects which are very distant from each other, for instance, in the Pelopon- nese or Crete and Cyprus, in Thrace and Asia Minor, are we authorized to attribute the word to the Hellenistic stock, according to the same method ' A few interesting equivalences are pointed out by Kalitsunakis, Mitteil d. Seminars fiir ovient.- spy. XIII. (1910), 96 sqq. 2 Die antike Tierwelt, Leipzig, 1909, 1913. The same holds good of Robert, Les noms des oiseaux, Neuchatel, rg1t. 3 Op. cit., p. 368. 4 See, besides the literature quoted above, p. 192, note 4, Krumbracher’s edition of the Book of Fishes. Add Stefani, Forsyth (Major), W. Barbey, Karpathos, Etude géologique, Lausanne, 1895. O A. THUMB 194 which permits us to infer an Indo-European word *g*ous ‘ox,’ from Old Indian gaus, Greek Bots, Old High German kuo. When we have the Modern Greek Thesaurus, it will be easy to find out the words which Modern Greek adds to the ancient vocabulary; at present such an inquiry is very toilsome, as we are obliged to collect our material from a great many dialect glossaries and monographs. But let us make an attempt with the scanty material which has been recently published as a specimen in an official paper of the Greek Com- mittee for the new Thesaurus.! There nine words are treated; four of them, which I give in the ancient form—viz., *)Aco7rupos, *Pappetos, *NaBow, *rapa- Gadapidiov, are missing in our Greek dictionaries. A priori, these four words could equally belong to Ancient Greek. But as the descendants of tapaf@ana- pidiov (BapOarapid:, Baptadapisd:) are found only in Crete, in the Aegean Sea, and Macedonia, it is dubious whether the word existed in antiquity, although its phonetical character would allow us to conclude that it did. With regard to the others, however, I do not hesitate to attribute them to the coun. For HALdtrupos (-ov) ‘ warmth of sun,’ ‘sunny place,’ does occur in Asia Minor, on the islands from Cyprus and Rhodes to Lesbos, in Euboea, in Epirus and the Peloponnese. @apperos, ‘ light-hearted,’ is found in Thrace, Bithynia, Mace- donia, Peloponnesus, and Zacynthus, and in the Tsaconian dialect, and it is moreover attested by medieval texts. Finally, *AaBow (Modern Greek A\aBovw) is in common use. Also the form of these words harmonizes with the character of Hellenistic Greek, and therefore I should not be surprised to find them one day in a papyrus, as on the other hand many a papyrus word, which is an addendum lexicis, or many a Greek loan-word from Latin, Hebrew, or Armenian (also lacking in the Greek dictionaries) is confirmed by Modern Greek.2, Sometimes internal (etymological) reasons lead us to assume that a modern word might be old; thus I myself* have concluded from Modern Greek ctudos, ‘ neighbour,’ and cvu7rda, ‘ near,’ although the form has only been found hitherto on some islands (Andros, Naxos, Syra, Melos, Carpathos), that there existed an Ancient Greek word *cvyu7Anyos, ‘ knocking together,’ ‘confining, which is, moreover, the basis of ancient cuvprdAnyds (LuptAnyddes ynoo).4 And Hatzidakis has declared Modern Greek ypizros to be old, on the authority of the Theocritean ypurevs*’—without knowing that ypizos is well attested in antiquity.® Finally, there is a whole group of modern words, which, for external historical reasons, must be attributed to the ancient vocabulary—viz., the Papyrusforsch., ii. 406 sqq. % Indog. Forsch,, xiv. 349 sqq. 4 The derivation from cuumrdjovos, cuurrAnoLA fw, which is tentatively suggested by K. Dieterich, Byz. Zschr., x. 595, and resolutely accepted by M. Vasmer, 1). xvi. 552 sq., seems to me hardly substantiated. 5 See ’AOnra, xiv. 523. 6 See Méya Aegixdy (Athen rgor-1906) s.v., and Lambertz, Glotta, v. 136. 1’Emirpor) Tod AeEcKoD THs “EAAnMiKIs yAwoons. "Avaxolvwois mpwrn. Athens, 1912. Compare, too, the lecture of Hatzidakis given before the Congress of Orientalists (Athens, 1912), ’A@nva xxiv. 373 sqq.,and the short but instructive article of Ilarayewpylov, Ilept trav cnuacidy rod piuaros dpuevitw, ‘'A@nva, xxiv. 459 sqq.; where the ancient dpuevov and dpuerif~w are elucidated by Modern Greek. 2 See the bibliography, Thumb, Arch, f. Latin | Greek | at the words castrum duty’; ‘pay hospitin obsequin afford a numbet words » Latin © Similar I may Byzanti loan-wo We manner extende: importa whole, z or spoke Greek t Similar | those of struct is influenc replaced language stituent but eve vernacul Greek t textual who, in popular | in the n copyists the man | 1 Cf. par ili. (Wien, : 2 Cf. fron Hahn, Rom Usten, Leipz Old ydern xreek 1e, as s and n has Com- them, Tapa- words GaXa- 1 Sea, nough ‘egard For or, on id the Mace- 1 it is Bovw) h the o find which ew, or [odern that a fodern is only athos), ether,’ nyaces on the is well xternal z., the rrAnoia gw, Jieterich, epted by ie hardly 1 B. Vee and pgs are Serre Sa 2 aoa hk Sor uel - pon ON THE VALUE OF MODERN GREEK 195 Latin loan-words which survive in Modern Greek. They must have come into Greek in the Latin epoch—i.e., before the formation of the Romance languages, at the latest at the end of antiquity.!. Our dictionaries know nothing about words such as *Bice€tos, bisextus, ‘leap year’; yovva, gunna, ‘fur’; xdotpor, castrum, ‘castle’; xiorépva, cisterna ; *xoppépKi(o)v, commercium, * customs duty ’; Aovpor, lorum, ‘ strap’; wavTirdu(o)v, mantile, ‘ handkerchief’; péya, *roga, ‘pay’; *oayirta, sagitta, ‘arrow’; oxovtédXa, scutella, ‘ plate’; ooits(o)y, hospitium, ‘house’; tovuma, tumba, ‘hill’; gaoxid, fascia. ‘tie’; *dwixcor, obsequium, ‘ wedding or funeral train’ ; and although the papyri and inscriptions afford a store of information on Latin loan-wards in Hellenistic Greek,” yet the number of such words is augmented when Modern Greek is drawn on. The words that have been found hitherto in Hellenistic documents prove that Latin words in Modern Greek represent an old Hellenistic inheritance. Similar evidence of antiquity can be found in other directions: for instance, I may call attention to the fact that we are justified in claiming wWi«e or Byzantine oixcov as an old element in Greek on the ground that it occurs as a loan-word in the New Hebrew literature of Midras (in the form apstkin). We have shown by a few details in grammar and lexicography in what manner our knowledge of Hellenistic Greek can be revised, confirmed, and extended by the help of Modern Greek study. Modern Greek is at least as important in enabling us to form correct views upon the ancient xowy as a whole, and as a link in the long history of the Greek language, whether literary or spoken. Our reconstruction of the xow7, which has led us from Modern Greek to the latest period of antiquity, results in a language which is very similar to Modern Greek. But the distance between the written texts, even those of the most vulgar character, and the spoken language we can recon- struct is very considerable. Although the speech of literary tradition is much influenced by the living language of the Hellenistic epoch, yet it is not replaced by it in any written document of Hellenism. For the written language shows throughout a mixed character. The proportion of the con- stituent elements drawn from the traditional and the living speech varies ; but even the most vulgar papyri stand at a certain distance from the vernacular language of their times. Without the knowledge of Modern Greek this fact would not appear so clearly. Hence arises a problem of textual criticism which has great importance for some Hellenistic authors, who, in opposition to the Atticising literature, avail themselves of a more popular language. Are all the vulgar (or Modern Greek) forms which occur in the manuscripts in reality elements of the original work? Uneducated copyists might spontaneously and involuntarily have inserted such forms into the manuscripts; conversely, educated copyists might have removed vulgar 1 Cf. particularly G. Meyer, Neugriech. Studien, Forsch. Anz., xxii. 39 sqq.); Der Sprachenkampf iii. (Wien, 1896). im vom, Reich, Philologus, Suppl. x. (1907) ; 2 Cf. from the most recent literature, especially Wessely, Die latein. Elemente in dey Grdzitdt der Hahn, Rom und Romanismus im griech,.-rémischen dgypt. Papyrusurkunden, Wiener Studien, xxiv. Osten, Leipzig, 1906; (thereon see Thumb, IJnmdog. (1902) 99 sqq., xxv. 40 sqq. 196 A. THUMB forms, on account of Atticising tendencies, which have dominated Greek literature from the beginning of our new era till to-day. If therefore (for example) the manuscripts of the Christian Apocrypha show more or less the occurrence of vulgar forms, we may suppose that the textual tradition has been infected and adulterated so as to bring it into harmony either with the classical or the vernacular language. It is not possible to reach a decision by general rules, but only by an exact examination of each literary work and of its textual tradition ;+ and here (a danger to be avoided) the knowledge of Modern Greek might be used in a mistaken direction: a form in the manu- scripts might remain unaltered just because it is modern. Modern Greek in the first place helps us to recognize the condition of the spoken language, which lies at the back of the written language. For it follows from the very existence and the character of Modern Greek that the xowvn Was a spoken language, used by all Greeks. Here we touch the highest problems, which concern the general character of the «oun. Modern Greek, together with its dialects, has arisen from the xow7 as a unity, and this fact presupposes the disappearance of the ancient dialects. It is the same process which took place in Vulgar Latin and in Armenian (not in the Germanic languages); Latin put an end to the old Italic dialects (Oscan, Umbrian, etc.), and thus became the mother of the Romance lan- guages. Literary Armenian, used as the common language of life, also caused the contemporary dialects to disappear, and became the starting-point of modern dialectical differentiation. So the Modern Greek dialects—those, for example, of Megara, Crete, Naxos, and Cyprus—as we said before, are not descendants of ancient Doric, Ionic, or Cyprian, but descendants of that cown Which in antiquity was spoken in the several districts. The ancient dialects, therefore, must have disappeared in the course of the Imperial period. . | Indeed, the ancient inscriptions present the counterpart of this. We observe how the dialects decreased more and more from the third century B.c. on, until they were everywhere displaced by the xown about the third century A.D. Some scholars have attempted to prove that this process, which is observed in the inscriptions, reflects only the usage of the official language, but is no illustration of the condition of the spoken language. This may be true in part. But on the whole the inscriptions certainly reflect the natural process, which ended with the disappearance of the old dialects; for Modern Greek imperatively demands this theory, and it seems to me sophistical to dispute the connexion between the evidence of the inscriptions and the evidence deduced from Modern Greek.* Thus the first epoch of the xouvy may be defined as a transition state between the classical epoch, where the dialects existed independently beside each other, and that basis of Modern des Hellenismus, Wien, 1907. 3 Cf. Hellenismus, pp. 28 sqq. 1 Cf. Hellenismus, pp. 12, 186 sq. 2 Cf,, for instance, Wahrmann, Prolegomena vu einer Geschichte der griech, Dialekte im Zeitalter Ms PT eed —v 2 RTS RTC sone RRR Tsien TTS > a *y > Pe he ah Greek comm only a Attic langu: illustr moder and £ perfor for mc Moder a disp KoWn, again, (dative transit these | innum in the diction As ancien M we rec (2) Cy Aegear (perhay Pelopo have b The gr on the 4and 5 of vows 1 OF 4, TI dialects which ¢ modern Helleni the nev mation. very gé Helleni on the rreek . (for s the 1 has h the yn by - and ge of lanu- f the ‘or it t the ehest asa s. It » (not alects > lan- aused nt of se, for ‘e not - that ncient perial We 'y B.C. third which ruage, 1ay be atural fodern cal to d the KOLW?) re the [odern =_— ee we ee x by 7 . S RL eee! Tei. See ee. jeanne He ON THE VALUE OF MODERN GREEK 197 Greek, where the old dialects had been quite absorbed by a unitary and common speech. Nevertheless, it would be incorrect to define the xowv7 as only a transition state: it has also an independent character, compared with Attic Greek, from which it originated in its essentials, as well as with the language which must be assumed as the basis of Modern Greek. This is well illustrated by the fact that certain Hellenistic formations are missing in the modern language. Forms such as ov@eis for ovdeis, xatyywp for Katnyopos and Ba@paké for Bdtpayos, ovyyevedou for ovyyevéot and yoveiot for yovedar, petCorepos for peiSwv and BéAtatos for BéAticTOS, HTw for éotw, morwy and dwn for zrovoin and doin, Svat for Sodvac are neither Attic nor predecessors of Modern Greek types. They are peculiar characteristics of the xown, showing a disposition to a new development, just like all the other innovations of the xown, only that they remained confined to the Hellenistic epoch and vanished again, partly because the grammatical categories to which they belonged (dative, optative, infinitive) have been lost. We may call them forms of transition, as belonging to Greek only for a time. We must lay stress upon these formations in order to obtain a clear conception of the notion «own ; innumerable are the lexical innovations, which have become obsolete again in the modern speech. A rapid comparison of the Hellenistic and the modern dictionary suffices to show this. As we have said, the modern dialects have nothing to do with the ancient dialects. Modern Greek is divided into numerous dialects. Disregarding Tsaconian, we recognize to-day the following main groups: (1) Pontus and Cappadocia ; (2) Cyprus and the south-eastern islands (Rhodes, etc); (3) Crete and the Aegean Sea; (4) the northern islands (Lesbos, etc.) ; (5) Thrace and Macedonia (perhaps with Thessaly); (6) Euboea, Attica, Megara (perhaps with the Peloponnesus); (7) Epirus and the Ionian islands (whose peculiar features have been insufficiently investigated) ; (8) the Greek dialects of Southern Italy. The greatest difference is that between the dialects of Pontus and Cappadocia on the one side, and all the rest on the other. The northern dialects (numbers 4 and 5, less distinctly numbers 1 and 7) differ from the southern by a number of vowel changes, involving the disappearance or reduction of every unaccented t or u, and the change of unaccented e and o into 2, w. The grouping of the modern has no relation to that of the ancient dialects. The former must accordingly have developed from a new basis, which can only be the xow7. The question now arises, to what extent do the modern dialects, at least in their embryo forms, pre-exist in the cow7 ? Had Hellenistic Greek dialectic variations which can be regarded as forerunners of the new dialect differentiation? The ancient texts give us practically no infor- mation. Hitherto we have not got beyond a few insignificant details, and some very general reflections. Probably an exact and minute examination of the Hellenistic sources will some day permit us to advance, especially, I believe, on the lexical side. But we have already seen that here also we cannot do 198 A. THUMB without the help of Modern Greek. As things are, the problem of Hellenistic dialects cannot be solved except by utilizing the modern language, or at least working in close connexion with it. Before examining this, however, it is impor- tant to define our task exactly and to put our question into the right form. When in its first epoch Hellenistic Greek began to eliminate the dialects, the xovvn naturally varied as spoken in Doric, Aeolic, or Ionic territory accord- ing to the various admixture of old dialectic ingredients. But these elements decreased continually, and the traces of the old dialects which have been locally preserved till the present day! are so trifling as to have had no influence in differentiating the modern dialects; for those traces—themselves equally insignificant—which are common to the «ow as a whole, do not concern our present problem. Moreover, Egypt and Asia Minor, a large part of Hellenistic territory, do not belong to the domain of the ancient dialects. And finally, if we speak of dialectic varieties of the «ov, we refer especially to such local differences as have developed from the new-born unity of speech. Taking into account these points of view, we must formulate our two questions, as we did above, guided by Modern Greek. We may now put the one questicn somewhat more exactly thus: In the Imperial age did dialect marks exist which may be regarded as germs of the modern dialects? The answer to this question depends again upon another: What ‘are the oldest dialectic pheno- mena of Modern Greek? Do they belong to the Hellenistic epoch? We have already touched on the conditions under which Modern Greek phenomena of a local or dialectic character can be attributed to the «ov. has received but little attention.’ That Modern Greek forms of local currency can be derived from analogous The problem forms of the «ow7, follows immediately from their occurrence in Hellenistic texts; and when we combine these forms of Ancient and Modern Greek, we deduce a number of facts which must be interpreted as signs of dialectic varia- tion in the xo. Compare, for example, the following features : 1. Phonology.—In Pontic the vowel » is preserved as é (1reyads= mnyaduor), whereas it has become 7 everywhere else. The change of 7 toz is well known from Hellenistic texts. If, now, the pronunciation of 7 as ¢ still exists in a modern dialect, we must infer that in antiquity there were two provinces of the xow7, one with n=2 and another with n»=e. The development of ¥y in Modern Greek dovrAevyw for dovrAevw is dialectal, but it is old, as the form evytNatou for eviAaTtov in a Hellenistic inscription of Amorgos clearly shows: and it is remarkable that the modern dialect of Amorgos belongs to the group of which this intrusive y is characteristic. In antiquity, therefore, there were two distinct provinces, one with dovAevw and one with dovrevyw. Geminata, or double consonants are, as a rule, simplified in modern speech, and this change was already taking place in the Hellenistic age. But geminata have some useful hints, but his methods are too constructive, and his results often leave us sceptical, 1 See Hellenismus, pp. 53 $qq. 2 K. Dieterich treated it with some minuteness in Kuhn’s Zeitschrift, xxxix. 81 sqq. He gives “fiat m2 n ES See Ji anserctbenanerant a Wh eta - f ee pit ; been pr! therefo which | a? peculia Cyprus the mc 3rd plu is also’ conseq! W existen side wi nize in KOLVN). where usual a several with é¢ this wa examin paralle: coverin a relati such fa local fe antiqui which I ventu tinuous varietie regions group ¢ from tk the Me Italy, n reason Euboez historic Pontic, Kon, t centurl nistic least npor- 1. lects, cord- nents been uence jually nh our nistic nally, local y into iS we ‘stion exist » this heno- We mena \blem gous nistic k, we yarla- dior), nown ina es of y in form OWS: ‘roup were nata, this have re too ve us ee ro err eee . - . cat Wadd z ‘ a ~T "SBS ile sattaied ~ we TEP eu, eee hk OR are ON THE VALUE OF MODERN GREEK 199 been preserved to the present day in the south-western islands. In the cow7 therefore there were districts which simplified double consonants, and districts which preserved them. 2. Inflexion.—The coincidence of nominative and accusative singular is peculiar to Pontic (e.g., o ddeAgov). It is witnessed by a ¢abella defixionis from Cyprus.! The phenomenon is accordingly ancient, but dialectic, for most of the modern dialects continue to distinguish 0 ddeddos, Tov adeAdov. The 3rd plural ¢épovv already appears in the Imperial epoch, but the type déepovare is also preserved to-day in the Aegean, in Cyprus, and elsewhere. There were consequently in Hellenistic times districts with -ovy and districts with -ovoc. We may then state the case generally thus: If we can establish the existence in antiquity of some local innovation in Modern Greek, and if side by side with this novelty the old form survives in another district, we may recog- nize in these differences the beginning of a dialectic differentiation in the xown. In this way we recognize also districts preserving final -y, and others where it was lost; a (small?) area showing palatalization of « beside the usual «; a district with dadepdos and one with aderdos ; districts with the several endings é\aBay and éAdBaou, and the now extinct €AaBocap, or again with édirouy and égirovcar, or the imperatives ypdyrov and ypawe. When in this way local varieties of the xovn have been indicated, I believe that a closer examination of Hellenistic Greek along such lines will be more successful than parallel researches in vulgar Latin. Thus some years ago I succeeded in dis- covering a real dialect in the tabellae defixtonum from Cyprus, and I formulated a relation between this and the modern dialects of the south-east.” Guided by such facts, we may suppose that certain marks of modern dialects were already local features of Hellenistic Greek, even if they are not directly attested from antiquity. But we cannct give a general rule as to the circumstances under which a modern dialecticism must be regarded as old. Normally, however, I venture to say, modern dialecticisms which are proper to a group of discon- tinuous modern dialects, can be derived with some probability from local varieties of Hellenistic Greek, unless spontaneous development in different regions may be expected. Thus, for instance, features common to the dialect group of Athens, Aegina, and Megara, which for several centuries was cut off from the other dialects by the Albanians, and common also to the dialect of the Maniates, as agreements between Epirus, Peloponnesus, and Southern Italy, may be considered as local elements of Hellenistic Greek. For this reason I believe that the « aorist, which appears in Aegina, Athens, Kyme (in Euboea), Epirus, and Maina, must be connected with the Hellenistic perfectum historicum. Further, it seems to me legitimate to deduce from agreements between Pontic, Cappadocian, and Cyprian characteristics of an ancient ‘ Eastern’ xo.vn, because these dialects have been separated from each other through five centuries, by Seljuks and Turks, and because on account of this isolation the 2 Ibid., p. 257. 1 See Principienfragen, p. 258. 200 A. THUMB Pontic and Cappadocian dialects have preserved their archaic character.! To the establishment of an Eastern coun? I attach particular importance for our views of Hellenistic Greek. For the linguistic diversity of the Greek mother country and Hellenized territories outside leads us by a priori reasoning on to some other problems, and in the first place to the question whether the old dialects in Hellas, or the speech of the Hellenized population in Asia Minor, influenced to any significant extent the development of Hellenistic and Modern Greek. For the purpose of these problems, Hellenistic must be divided into three main areas, the Greek mother country, Asia Minor (with Syria), and Egypt. But as Egypt and Syria have been lost again to Hellenism by the Arab conquest, only the linguistic contrast of Asia Minor and continental Greece concerns us for the purposes of our inquiry, which is based on Modern Greek. A comparative grammar of the dialects of Asia Minor would be indispensable for the problem we have touched upon. Hitherto we have only some preparatory monographs, such as the excellent sketch of the Cappadocian dialects which we owe to the English archaeologist Dawkins.® What we have so far learnt from Hellenistic sources about the «ovv2) in Asia Minor is but scanty. Some phonetic influences of the indigenous population can be distinctly recognized—viz., the confusion of tenues, mediae, and aspirates, and that of the vowels e and7z. But these very features (which are also found in Egypt) have remained barbarisms of the lower classes, and do not survive in Modern Greek. Other features, which are also probably due to the influence of the indigenous population of Asia Minor—viz., the change of v intoz* and the levelling of vocalic quantity (which had begun in Asia and in Egypt*®)—have become common in Modern Greek. Asia Minor initiated these innovations, and since their development required time, the xown Of Asia Minor (and of Egypt) must have differed for a period from that of continental Europe in these features, which slowly spread over the whole of Hellenism. If now Asia Minor was the birthplace of some common innovations, it follows of course that other innovations, arising in that country, might have failed to make their way through the whole domain of Hellenism. Perhaps the spontaneous nasalization of consonants, as in oadpfarov for caSPBarov, comes under this category. The last word has not yet been spoken about this phenomenon,® which appears both in the xow7 and in Modern Greek, but cannot yet be brought under a precise phonetic law. We need a more complete knowledge of the field of this nasalization. I therefore express myself with reserve. Examining in Buturas’ collection of material the instances of a spontaneous and purely phonetic nasalizing, in their local distribution, I found that the majority of dialectic (not common) examples 1 The archaic character of these dialects of Asia Minor has been emphasized by K. Dieterich, see Kuhn's Zeitschrift, xxxix., pp. 86 sqq. 5 See Hellenismus, pp. 143 sq. 2 Cf. also K. Dieterich, /.c., p. 87. 6 Not even by Buturas (Glotta, v. 170 sqq.), 3 Modern Greek in Asia Minor, in Journal of who treats it rather unmethodically. Hellenic Studies, xxx, (1910), pp. 109-132, 267-291. 4 Except the cases I have dealt with in Hellen- ismus, pp. 194 sq. (vu to 2). 7 rer —— ORE Pore belon; and I It see and e ultim: the « also | affect: popul again above I indig« remat Pontt dialec of Pc survi' Testa posse suppl whicl with to-da wortl the } Attic the c the p alrea two | and | and — latte: inste: infin! the « that who 17) is mor west. 2 As Pp. 13 rt 6 6To for our mother ‘on to the old Minor, fodern 2d into 1), and by the inental fodern uld be ve only docian o.v7) In genous nediae, (which es, and aly due change n Asia Minor 1e, the m that whole ymmon buntry, enism. TOV for spoken fodern e need erefore aterial r local amples n Hellen- 0 $qq.), —— a ; Le — OR mr z ON THE VALUE OF MODERN GREEK 201 belong to Eastern Greek (Asia Minor and Cyprus, Cos, Carpathus, Rhodes and Lesbos), and that they become less and less frequent as we go west. It seems as if the process had started in Hellenistic Asia (and in Egypt), and extended in such a manner that single words with nasalized consonants ultimately found their way into western Hellenism. If my theory is correct, the «own of Asia Minor was accordingly distinguished from that of Greece also by the nasalization of consonants. Moreover, the innovation seems to affected by the influence of the phonetic system of the Hellenized indigenous population,” whereas other traces of such an influence have been either lost again or become common to the whole language, as has been observed above. But the «own of Asia Minor was not only exposed to the influence of indigenous language: it lay within the sphere of the Ionic Greek. It is remarkable that Ionic traces have been preserved in the modern dialects of Pontus*® in a more characteristic manner than can be predicated of other old dialectic traces surviving in the new. In keeping with this archaic character of Pontic is the existence in Pontic and Cappadocian of two other archaic survivals, which have some importance also for the delineation of New Testament Greek.* Firstly, these dialects only have preserved the ancient possessive adjective éuos, cos, while in all other dialects those forms are supplanted by the enclitic genitives pov, cov, etc. Secondly, the infinitive, which in the New Testament had begun to be displaced by the construction with iva, has shown itself capable of resistance in Pontic, where it is still used to-day after several classes of verbs, as, for instance, verbs of motion. It is worth remarking that the infinitive of purpose, which appears distinctly in the New Testament, seems to be of Ionic origin: at any rate, it is but rare in Attic Greek. Thus we see how Ionic influence worked through the «ow upon the configuration of the modern dialects of Pontus. Since the disappearance of the pronouns és, etc., and the supersession of the infinitive by va can be seen already in the Hellenistic age, it follows again that in antiquity there were two dialect-areas, one more extensive, in which the pronouns pou, gov, etc., and the iva construction progressively superseded éyos etc. and the infinitive, and a smaller one where these processes were brought to a stop. To this latter area belongs the Greek of the New Testament: it has the infinitive instead of iva after the very classes of verbs which now in Pontic demand the infinitive, and in the Gospel of John éuos is used far more frequently than in the other New Testament writers. We can therefore draw the conclusion that New Testament Greek is a local variety of Hellenistic, written by men who spoke the Eastern «ow. The facts we have adduced are not the only marks of a local colour in 1 I have disregarded Tsaconian, Nasalization 3 See Hellenismus, pp. 87 sq. is more frequent in Epirus than elsewhere in the 4 Cf. Principienfragen, p. 259 (with biblio- west. graphy), and Moulton, Prolegomena, pp. 40 sq. 2 As I observed years ago in Hellenismus, 5 See Brugmann-Thumb, Griech. Gramm, § 580 PP. 133 sqq. (pp. 593 sq.). 202 A. THUMB this Eastern cow7. lees The the hief the htly ures sone, canara a Piri te ye - VERSE-WEIGHT 209 suggestive of the Euripides of its day: its trimeters have more trisyllabic feet, and they are end-heavier, than those of any other play of Sophocles. Four plays of Euripides—the A lcestis, Medea, Heraclidae, and Hippolytus— are constituted a compact group, distinct from all his other plays, by their sparing use of trisyllabic feet. In the matter of iambs and spondees also these four plays are very much alike; especially in the fifth foot, with 60 : 40, 60: 40, 61: 39, 61: 39. Then comes the deluge of trisyllabic feet, and at the same time the predominance of iambs in the fifth foot falls away; in the Bacchae it can scarcely be felt, and only by exception does it rise in the Helena to 58: 41. In the later manner of Euripides, as in the Philoctetes, the Cyclops, and the Acharnians, spondees in the fifth foot seem to serve as a make-weight against the trisyllabic feet. The Cyclops, by the way, does not differ in the present respect from some late tragedies. The fragments of the Ichneutae of Sophocles are hardly a quorum for this purpose, but as far as they go they show a higher proportion of iambs than any tragedy. Menander seems to agree with the Acharnians more or less. Lycophron comes under my four rules; and he is fonder of spondees even than the Rhesus, but distributes them more evenly between the third and fifth feet. Let us now return to the Rhesus. Under three of the four rules it marks the limit among tragedies. (I.) In the first foot its percentage of iambs is the lowest except the Hecuba’s, and its ratio of iambs to spondees is the lowest of all. (II.) In the third foot both percentage and ratio are distinctly the lowest. (III.) But in the fifth foot the Rhesus is not exceptional; at this point it comes within the range of variation displayed by Aeschylus and Sophocles, and it does not differ much either from the earliest plays of Euripides or from the Helena, a late play. Whereas in the first and third feet it marks extremes, in the fifth foot it is commonplace. (IV.) Yet in spite of the fifth foot its ratio of iambs to spondees in the three feet together is the lowest of all. mpoabe rNéEwv, Omriev Sé Spdxwv, pecan SE Xiparpa. What are we to make of this strange compound? Those who ascribe it to Euripides regard it as an early work, perhaps his earliest ; but by the present tests, whatever they may be worth, it should be later than the early group of four plays in respect of the first and fifth feet, and those four plays are all quite normal, where the Rhesus is very abnormal, in the third. As the iamb is lighter than the spondee, so is the tribrach lighter than the dactyl. As spondees outnumber iambs in the third foot in every tragedy, so, and more so, do dactyls outnumber tribrachs in the same place: by 3 to 1 in Aeschylus, 4 to 1 in Sophocles, 3 to 1 in Euripides. The Rhesus has no less 1 One might expect that the same would be Euripides three plays have only one such dactyl true of the first foot also, but it is not; at least between them, though the fourth, the Hippolytus, notin Aeschylus or Sophocles, in whom tribrachs has eleven, in deference to the hero’s name; but have the better of dactyls there by about 4to1 in the rest of Euripides dactyls prevail over and 3 to 1. Of the small and early group of __ tribrachs in the first foot by more than 3 to 2. P 210 E. HARRISON than 42 dactyls there against 6 tribrachs, or 7 to 1; another feature of the difference between the Rhesus and all other tragedies in the treatment of the middle of the line. These 42 dactyls, by the way, account for almost two-thirds of the play’s 65 trisyllabic feet. In every tragedy the dactyl in the third foot is by far the commonest of all trisyllabic feet, but in no other tragedy does it form so large a proportion of the total. This seems to me characteristic of the metrical style of the Rhesus, which is not exactly chary of all such means of varying the metre, but is rather shy, like modern composers, of the rarer kinds. Its total of trisyllabic feet offers another difficulty to those who would assign the Rhesus to the early days of Euripides. In this respect his plays fall into two unequal groups. The smaller and earlier group is close and con- sistent, and stricter in this matter than Aeschylus as a whole or Sophocles as a whole: the Alcestis (438 B.c.), the Medea (431 B.c.), the Heraclidae, the Hippolytus (428 8B.c.), with one such foot in 15°2, 13°8, 13°4, and 16°3 trimeters respectively. At one end of the larger and later group stands the Andromache, with one in 6°4; at the other end the Sacchae, the Iphigenia in Aulis, and the Orestes, with one in 2°3, 2°3, 2°I. To which group belongs the Rhesus, with one in 10°6? It is about midway between the Heraclidae and the Andromache. By this criterion, as by its spondees, it would not seem to be a very early play, certainly not his earliest. Anyone who cares to carry this verse-weighing further must consider what weight he is to allow to anapaests, dactyls, tribrachs. All would agree that the tribrach is lighter than the iamb. The anapaest would seem to have been felt lighter than the spondee, since tragedy admits it, though under strict limitations, even into the second and fourth feet, and in comedy it has the run of five. The dactyl is harder to weigh. If it was lighter than the spondee, and as light as the anapaest, why did tragedy use it so seldom in the first foot before the later manner of Euripides, and always debar it from the fifth? This last thing, however, may have nothing to do with weight. It may well be due, like the great rarity of tribrachs at that point, to a desire to keep the end of the line regular as well as light ; an instinct which asserts itself at the end of the Latin hexameter and pentameter and in other metrical forms. Some may think that the differences with which I have dealt are too small to be telling. But some at least of them are large. Compare Solon with the Acharnians in regard to the fifth foot, Simonides with the Rhesus in regard to the third. Inside the range of tragedy the differences are less striking, but still significant; and I was encouraged to proceed by the discovery that the four plays of Euripides which agree so closely in other respects agree closely in this, and that there is also close agreement in this respect between the three plays of the trilogy, the most compact group of all. Preferences for heaviness or lightness in the several parts of the line I should suppose to have been instinctive and scarcely conscious, not likely to be upset by large and sudden changes such as we can discern in other features of the metre, 4 ee — - TE ae VWiracs cing. aa, between Helena test of a myself t is by E content TRINI1 Whe Gesetz’ another ; bohmische: Ig0g, 1x. tragic tri explanatic who sees i The ( only all tr (in order t in which ¢ monosylla) fifth foot ? balance of and even i ratio very | lection for entitled, th predilectior f the f the lay’s r the large trical ig the vould ys fall | con- nocles , the neters mache, nd the , with mache. y play, r what e that e been ' strict as the yondee, ‘st foot fifth ? ay well eep the at the o small rith the gard to ng, but hat the closely een the nces for | pose to | oy large » metre, — RO NRE = —_ r ~ a a TREE Whose VERSE-WEIGHT 211 between the Electra and the Philoctetes, for example, or even between th Helena and the Orestes. This being so, I am inclined to put some trust in “ test of avoirdupois. But all such numerical tests are risky, and I do not flatt myself that any ‘cheesemongery’ will settle the nies whether th RA is by Euripides. The evidence on that question! is cumulative - “sa content to have added my tittle to the heap. _ ~ Trinity COLLEGE, CAMBRIDGE. E. HARRISON. POSTSCRIPT. When this was already in print, I came across an article on ‘ Por Gesetz’ by K. Witte in Hermes xlix. (1914), pp. 229-245, which led sae another article with the same title by J. Kral in Sitzungsberichte der kénigl- bohmischen Gesellschaft der Wissenschaften, Klasse fiir Philosophie usw., Jahr w 1909; IX. Kral reckons the percentage of spondees in the fifth foot or tragic trimeter at 43°4, and he has forestalled me in finding therein a partial explanation of Porson’s Law. This suggestion is curtly dismissed by Witt who sees in the predominance of iambs not a cause but a consequence stile Law The question is easily put to the test. From any tragedy subtract an only all trimeters that have a short syllable clearly exemplifying the Law, but (in order to avoid any questions about marginal cases) all trimeters wiicunnees in which the first syllable of the fifth foot is a short final syllable or a short monosyllable: in the remaining trimeters do iambs or spondees prevail in th fifth foot? Of my eight tragedies, three (Ag., Hipp., Rhes.) still show a ao balance of iambs, in two (Supp., Ant.) iambs are about as many as pe and even in the other three the ratio of iambs to spondees is about 45: : ratio very different from that of the first foot or the third. Thus the | ae lection for iambs in the fifth foot is not due to Porson’s Law We a entitled, then, to look at the matter the other way about, and to aos from this predilection at least a partial explanation of the Law. . E. Hi. 1 See W. H. Porter in Hermathena, xvii. (1913), pp. 348 sqq. RAINBOW, SKY, AND STARS IN THE /JZJAD AND THE ODYSSEY: A CHORIZONTIC ARGUMENT. THE opinion has been expressed frequently of late, notably by Professor Mackail and Miss Stawell, that the Odyssey may well be the work of the advanced years of the Homer of the Iliad. Miss Stawell remarks that one of an alert mind must feel that the Odyssey is the poem of an older man—one who has conceived and written a poem before. She suggests that that poem may have been the Iliad. So Professor Mackail argues that a ‘different mind may come to a poet with the lapse of years and with fresh experiences,’ and that ‘the poet who produced the J/iad in the early prime of his life was a poet capable of the artistic and poetical change that is felt in the Odyssey, among new surroundings, with an altered view of life, with an imaginative ardour burning less strongly and with increased constructional mastery.’ ? I cannot feel that the poet of the Odyssey with his delight in the tale for its own sake, his love for marchen and the marvellous, gives indication of a greater maturity than that of the poet who tells of the sorrows of Priam, of Hecuba, of Andromache, of Achilles. The passages in which, as Professor Mackail says, ‘a whole criticism of life is concentrated,’ are as many in the Iliad. Indeed the two? which Professor Mackail cites as pre-eminent in this are taken by him from the Iliad. The thoughts put in the mouth of Glaucus and of Paris in these places are those of a man who has known the contempla- tive life as fully as the Odyssey poet. In no poem do we find old men repre- sented with greater sympathy and understanding. I believe that no changed circumstances or experiences in the life of one man can explain the great difference in the temper of the poems. More than one generation of men has passed away between the epic of the J/iad and the enchanted tale of the Odyssey. A different age, a different taste, a changing faith produced that highly civilized poem. The old weather-gods of the Achaeans are fading out, and Athena, no longer a Valkyrie, but the patron saint of Odysseus, and rather more honourable, though less vigorous and vivid than the Athena of the J/iad, and on her way to be goddess of city life, is the divinity most fully in evidence. In the following discussion of a point of difference which I hold to be of 7 1 Lectures on Greek Poetry, p. 12. 3 J]. VI. 145-9; III. 65, 6. 2 Ibid., p. 73. on or = eremnemePaa asec in k the gre emphas and the an abso ‘quam ¢ Odyssey Sky-, th heaven: The Sky-Goc implicat called be by Zeus second 11 (7roppupé comes so down in as a hear The and realr in the Oa than in 1 poems. | sky, and | seen by r with the 1 apytKképau vepns (8-: pavvos (8- name of recession away by t mentionec ‘OrvptrL05 Light in Dante « The splenx the Odysse for the lis lightning the gleam 1 ‘Lurid’? Rh Leaf, J/ dnp, RAINBOW, SKY, AND STARS IN ILIAD AND ODYSSEY 213 the greatest importance, and not yet, so far as I am aware, sufficiently emphasized, I shall, for the sake of convenience, speak of the poet of the Iliad and the poet of the Odyssey, without intending to commit myself to a belief in an absolute singleness of authorship for each poem, but realizing with Wolff ‘quam egregie Carmini utrique suus color constet.’ My argument is that the “ Odyssey differs from the Iliad not only in knowing Zeus less in his aspect of Sky-, that is, aether-God, but in its lack of sensibility to all the phenomena of the heaven :—‘ Fire and hail, snow and vapours, wind and storm fulfilling his word.’ The rainbow and the Rainbow Goddess, who acts as messenger for the Sky-God, are found only in the J/iad. The Rainbow Goddess appears by implication once in the Odyssey in the jest at the expense of Irus the beggar, so called because he ran on errands. The bright rainbow itself, as a thing sent ngs | by Zeus as a portent (tépas) to mortals, is spoken of twice in the Iliad. In the — | second instance it is Athena who comes from heaven wrapped in a gleaming ee (7ropgupén)* cloud like the gleaming rainbow. In the Odyssey Athene never a may comes so meteorologically as in this case and in J/iad Book IV., where she shoots 1 may | down in the form of a meteor into the Greek camp. In the Odyssey she comes 1 that as a hearty ship-captain, or a shepherd, or a little girl carrying a pitcher. : pest The word ai@np ‘ the scene of meteorological phenomena’? and ‘ the home mane and realm of Zeus as raingiver,’ appears twenty-one times in the I/iad and five re in the Odyssey. The words odpavos and anp occur much oftener in the Iliad than in the Odyssey, though both, especially ovpavés, are usual words in both ale for #§ Poems: In two places* in the Iliad the genitive Acos is used in the sense of iil sky, and the preponderance of that aspect of Zeus in the older poem may be a al _ seen by referring to Geddes’,* where a list of physical epithets of Zeus is given ae with the number of instances in Iliad and Odyssey. The epithets ATTEPOTNTNHS, ee apyiKépavuvos, epiBpeméerns, oTEpotnyeperns, occur only in the Iliad, and xedar- sie | vedns (8-3), vepernyeperns (27-7), épiySoutros (7-3), UYuBpeuérns (4-2), Tep ure- ae | pauvos (8-6) occur more frequently in that poem. It may be argued that the nail name of Zeus himself occurs only half as often in the Odyssey. But this very regu recession of the Sky-God has its significance, which has never been explained away by those who advocate the single authorship. His épos onpavtixor is aa mentioned about five times as often in the J/iad, and in that poem the adjective | ‘Odvptr0s is far more common.? , oa : Lightning and fire have a far greater part in the Iliad. ‘ Nowhere except ‘ : : | in Dante does fire so penetrate the whole structure,’ says Professor Mackail. angins | The splendour of fire, céXas, is used eleven times in the Iliad, and but twice in a ™ the Odyssey. The Iliad poet uses the word for the glory of the full moon and P — | for the lightning of Achilles’ angry eyes, as well as for the flashing of the . — | lightning bolt and for blazing fire. Radiance, aly, is used but twice, and of _— * the gleaming of metal in the sun. The phrase am’ aiyAnevtos ’OdAvprov appears o be of bs 1 *Lurid’? So Leaf. 3 Iliad, XIII, 837; XIX. 357. a 2 Leaf, Iliad, II1., App. H, on ovpavds, aldyp, * Problem of the Homeric Poems, p. 137. anp. 5 According te Gehring’s Index, 32-12. 214 G. H. MACURDY twice, and gives the starting-point for that line of magical beauty in the famous Olympus passage of the Odyssey—Xevkn 8 émidédpopev aiyrn. The whirling snowflakes, vefades, which give such splendid similes to the poet of the Jliad, are never mentioned in the Odyssey. The Iliad poet seems even to have imagined an avalanche. sented in motion. At least his mountain of snow? is repre- The poet of the Odyssey speaks of snow (xv#v) three times? —once of its absence from shining Olympus, again of the alleged suffering of Odysseus at Troy on the night of the freezing snowstorm, and third, the com- parison of Penelope’s tears to snow on the mountain tops melted by the east wind. This last has been well compared by Professor Geddes with the splendid falling snow of Iliad, XII. 280, as a ‘ contrast between two landscapes under snow.’ The flight of Iris in J/iad, XV. 170, is compared to snowflakes or hail from the clouds. Hail, snow, and ice are mentioned in J/iad, XXII. 151. Hail does not occur in the Odyssey. non-Odyssean word. And opixdAn mist, ‘ vapours,’ is another Though rain is mentioned in the Odyssey on occasion, there is nothing to match the magnificent description of the autumn floods, when Zeus pours down his raging waters.°® It is in his awe and wonder before the stars and moon—Aaympods Suvactas é€umpérovtas aibép, that the poet of the J/iad shows most poignantly the splendour of his genius and his divergence from the quality of the poet of the Odyssey with his love for ‘ flat bits’ and a background of sea. The passage in the eighth book of the Jliad translated by Tennyson :—@s 8’ 67’ év ovpave adotpa haewny audi cernvnv | paiver’ apitpetréa, OTE T’ EwAETO vnvEwos alOnp K.T.X. conveys that sense of great mysterious beauty that overwhelms us and makes us breathless before the loveliness and the awfulness of the stars themselves. The exquisite passage in which Hector’s babe is said to be fair as a star, the comparison of Diomed to the star of summer, the simile of Iliad XI. 62, where the star emerges in brightness from the clouds, and is again hid in their dark- ness, will occur to everyone, together with other splendid passages. The poet of the Odyssey has the word aornp but twice. Once in a line repeated from Iliad, V. 294, we have the short simile aornp 8 os améNaprtre Of a wrétros. In Odyssey, XIII. 93, to mark the time of the boat’s reaching the island, it is said that at that moment the star most bright arose, which comes to herald the light of early dawn. The passage in which the constellations represented on the shield are enumerated (I/iad, XVIII. 483-89) is copied in part in the Odyssey. ‘The verses in the Iliad,’ as Miss Clerke* well says, ‘have the strong spontaneous ring of originality, while the Odyssean lines betray excision and interpolation.’ The epithet starry for the heaven occurs four times in the Odyssey and nine times in the Iliad. The moon, cednvy, appears four times in the Odyssey—three times in a brief stereotyped simile, and again (IX. 144) ina statement that the moon was not shining on a certain dark night (Odyssey, IX. 144). In the Iliad besides the wonderful passage in Book VIII., there is the magnificent line 1 Iliad, XIII. 754. 2 Odyssey, VI. 44; XIV. 476; XIX. 204. 3 Iliad, XVI. 765-69, * Studies in Homer, p. 45. . + 4 - 3 . ? $ 3 r 8 2 2 ee ty I NEALOD to pnv7 to eart soldier TI passage mist, < times < cloud s both p vedea ¢ the Ilt four tit times <4 sixth b spot, n y and ra Athene seen 0} Iliad s and m: glory spirit « the Od a jest aiBépe giimps MV tion o North: poems ‘} among of a pli sea—fi Northe stone, Harlec fringe T Vas i in the to the seems repre- times? ring of » COM- ie east th the scapes ikes or I, 151. nother casion, floods, LTT pous nantly 0et of assage oUpaVve@ p K.T.X, makes selves. ar, the where r dark- e poet 1 from ss In is said ight of > shield ‘The aneous lation.’ id nine —three hat the e Iliad nt line een i) “< x 7 ? i q . fa «. RAINBOW, SKY, AND STARS IN ILIAD AND ODYSSEY 215 néduov T akadpavTa cednvny Te TANGoveay, another reference to ceAnvyn and two to unvn. The Iliad has also a meteor in IV. 75-79, where Pallas Athene comes to earth in the form of ‘a star, a portent for mariners or a wide host of soldiery, and many sparks are scattered from it.’ The poet of the Odyssey took ‘ready-made’! similes from the ‘lion’ passages of the Iliad, but comparisons to stars, lightning, meteor, floods, hail, mist, and driving snow do not appeal to him. Cloud, végos, appears three times as often in the J/iad as in the Odyssey, and vedéAn twice as often. The cloud simile in I/iad V. 523 sqq. is of especial beauty. Fondness for shadows both poems show. The Ji/iad set the fashion with its lovely ovpea oxtoevta and vedea oxioevta. The words expressing brightness are much more frequent in the Iliad. In the Odyssey Xapmpos occurs but once, and forms of Adurw but four times. The Iliad uses daecvos five times as often as the Odyssey. True, we have the gleaming of Olympus in the sixth book of the Odyssey, but that is an unearthly splendour of an enchanted spot, not, as in the Iliad, the living splendour of the very stars of heaven. ‘The counsellor and cloud-compeller of classical Athens was the wizard and rain-maker of earlier times.’? The Zeus of the Iliad and his daughter Athene are nearer that earlier stage in which the phenomena of the heavens seen on their semantic heights revealed them to mortals. And the poet of the Iliad sees the splendour of the phenomena of nature caused by Zeus with awe No other Greek poet save Aeschylus gives such a sense of the glory of them. ‘@ dios ai@np!’ the cry of Aeschylus’ great hero, expresses the spirit of the J/iad. There is no such thrill of awe and worship in the poet of the Odyssey before the grandeur of the elements. So we have in his poem only a jest at the Rainbow Goddess, a Zeus who lives in a shining heaven, not ev aidépc kal vedédXnot on the stormy top of an dpos onuavtixdy, and barely a giimpse of the stars. With all the fresh light on Greek origins that has come with the recogni- tion of the fact that Zeus? is the patriarchal mountain-god of the invading Northmen, the following passage in Ruskin has a meaning for the Homeric poems which he himself did not know: These are characteristic J/iad words. and marvelling. ‘How different must the thought about nature have been of the noble who lived among the bright marble porticos of the Greek groups of temple or palace—in the midst of a plain covered with corn and olives, and by the shore of a sparkling and freighted sea—from those of the master of some mountain promontory in the green recesses of Northern Europe, watching night by night, from amongst his heaps of storm-broken stone, rounded into towers, the lightning of the lonely sea flash round the sands of Harlech, or the mists changing their shapes for ever among the changeless pines that fringe the crests of Jura.’ The Iliad is essentially a Northern, the Odyssey a Mediterranean poem. GRACE HARRIET MACURDY. VASSAR COLLEGE. 1 Murray, Rise of the Greek Eptc, p. 215. 2 Murray, Four Stages of Greck Religion, p. 68. SUMMARIES OF PERIODICALS. LITERATURE AND GENERAL. American Journal of Philology. Vol. 35. No.1. 1914. E. G. Sihler, Ciceyvo. An Appreciation. A highly favourable estimate of Cicero’s powers and career from a forthcoming book. Charles Knapp, Notes on Plautus and Terence. Practically a detailed criticism of E. Hauler’s revision of Dziatzko’s Phormio of Terence. He notes the editor’s failure to take proper account of American scholars’ work within his field. Walton B. McDaniel, Some Gveek, Roman, and English Tityvetus. On the doings of the aristocratic rowdy, for whom this is an appellation, in Greek, Roman, and English history. The word is insufh- ciently explained by the reference to V. Ecl. |. 1, ‘Zzityve tu patulae,’ etc. The original reference is to ‘tityrus’=satyrus, i.e. ‘lecher.’ The allusion to V. Ecl. would be an afterthought. Tenney Frank, A Rejected Poem and a Substitute. Catullus 68a was written to decline a request from Mallius for the consolation of love poems in his desertion, but not sent. Later on, when the conditions were somewhat altered, Catullus complied with M.’s desire, thinly disguising M.’s name in the ‘me Allis’ of line 1 and incorporating 21-24 in the poem. W. Peterson, in Move about the Dialogus of Tacitus replies to Dr. Gudeman’s ‘rejoinder’ on the pagination of the ‘ Hersfeldensis,’ vol. 35, No. 2. A. C. Johnson adds supplementary notes to his paper on the Tribe Ptolemais, vol. 34. R. H. Tukey reviews Norden’s ayvworos Geos (Teubner). Athenaeum (Pavia). Vol. II. No. 2. C. Morelli, Nevone poeta e 2 poett intorno a Nerone. Seneca was the dominating figure in the literary circle that gathered around Nero. He seems to have inspired the frequent references in the poets (e.g. Calpurnius Siculus and Lucan) to the inauguration of a new Golden Age under Nero; see ’Amvoxod, 4, Tac. Ann. XIII. 3, (speech composed for Nero by Seneca), etc. ‘The Meliboeus of Calp. Eel. I. and IV. is Seneca. The Iiias Latina seems to allude to the early years of Nero. Nero’s Tvoica, on a theme not only national but specially connected with the Julian house from Caesar to Nero, shows marked resemblances to Seneca’s T7voades, and both the emperor’s poem and Lucan’s early J/iaca were probably inspired and encouraged by Seneca. The famous passage, Luc. IX. g80-986, is unintelligible if the poet is thinking merely of Julius Caesar; he must allude to Nero, his now jealous rival. The ascription of Pers. I. gg-102 to Nero is without real authority and is most improbable. PP. Fossataro in notes on the text of De optimo geneve ovatovum discusses various passages with emendations by himself or others. S&S. Pellini publishes a codex found by himself at Soliera, near Modena, consisting of three leaves of 4 pages each, with about 300 lines of the Curculio of Plautus. The writing points to the beginning of the fifteenth century. P. gives a complete list of variant readings, taking as base the edition of Ritschl and Goetz, and concludes that the Soliera Codex is most nearly akin to the Codex Britannicus of the twelfth century, with which it agrees more than 100 times in 300 lines. P. quotes some of the most “y a = en - —. aes —_ ime of ¢ Aus Cor the mal rea edit spec and N. ord Lit enl: not whi (Ba are Att gTO Sct enl: (v. wit tyat W. stu not coll sta Ges Ex; tim Noi der imate of Notes on vision of r account ne Greek, or whom is insufhi- tc. The > V. Eel. Catullus ye poems t altered, ie Allius’ about the yn of the 2s to his aLyVoorrTos minating - inspired n) to the , XIII. 3, -and IV. Nero’s ian house and both couraged e poet is us rival. is most discusses blishes a f 4 pages ts to the readings, e Soliera ury, with the most ey PP rerere—n or H-. . eg rr SUMMARIES OF PERIODICALS 217 important variants peculiar to this codex—about 200 in number—and adds a fac- simile of a page. A. Parravicini enumerates and analyses the metric ‘ prefaces’ of Claudian, written in elegiacs, whereas the poems which they introduce are in hexameters. These prefaces combined a real introduction with a special appeal to the benevolence of a courtly audience, allusions to patrons, or the Emperor. Ausonius, his contemporary, alludes to the example of Catullus’ dedication to Cornelius Nepos and P. discusses the origin of these ‘ prefaces,’ concluding that they go back to drama, Claudian being original only in using a different metre and making them into a vehicle for conveying special meanings to his hearers and readers. Berliner philologische Wochenschrift. 1914. March 14. G. Wissowa, Religion und Kultus dey Romey (Samter). The second edition has been thoroughly revised. H. Kaden, Quaestionum ad Cicevonis Balbianam spectantium capita tria (Klotz). Treats of the life of Balbus, the legal question, and some rhetorical points. March 21. E. H. Minns, Scythians and Greeks (Philipp). Very warmly praised. N. Schneider, De vevbi in lingua Latina collocatione (Schmalz). A good study of the order of words in Caesar. March 28. M. Wellmann, A. Cornelius Celsus. Eine Quellenuntersuchung ( (ind), An important study in the history of medicine. W. v. Christ, Geschichte der griech. Literatur, bearb. von W. Schmid (K. F. W. Schmidt). Sixth edition of Teil I, enlarged to 771 pp. W. W. Tarn, Antigonos Gonatas (Swoboda). Highly praised. April 4. Fr. Niedzballa, De copia verborum et elocutione Pyomether Vincti q. f. Aeschyleae (Aly). In P. V.a peculiarly large number of words is found which do not occur in other Aeschylean plays. The reviewer, however, concludes that anyone who compares the three parts of the Oresteia will find no difficulty in assigning P. V. to Aeschylus. Beschreibung vimischer A ltertiimer, gesammelt von C. A. Niessen (Barthel). Specially valuable for glass and pottery. April 11. F. Windberg, De Hevodoti Scythiae et Libyae descviptione (Philipp). G. Rauschen, Tertulliani apologetici vecensio nova. Ed.11 (Hoppe). Both text and notes are good. J. Handel, De lingua communi in titulos Ionicos ivvepente (Kallenberg). Atticisms first appear in Ionic inscriptions in the second half of the fifth century, gain ground in the fourth, and almost drive out Ionic forms in the third. Max C. P. Schmidt, Realistische Stoffe im humanistischen Unterricht (Tittel). Third edition, much enlarged. April 18. E. M. Walker, The Hellenica Oxyrhynchia, its authorship and authority (v. Mess). The reviewer considers W.’s study of great value, but does not agree with his conclusion that Ephorus is the author. L. Parmentier, Recherches sur le tvaité d’Isis et d’Osivis de Plutarque (Gruppe). Valuable for the history of religion. W. Holtschmidt, De Culicis cayminis servmone et de tempore quo scriptum sitt(Helm). A study of the vocabulary which in the opinion of the reviewer proves that Virgil was not the author. O. Kern, Inscriptiones graecae (Hiller v. Gaertringen). A good collection, showing the inscriptions as they appear on the actual monuments. April 25. H. Fohl, Tvagische Kunst bet Hevodot (Nestle). Helps us to under- stand Hdt.’s conception of history. A. Steier, Avistoteles und Plinius. Studien zur Geschichte dey Zoologie (Scharold). A scholarly book. W. W. Fowler, The Religious Experience of the Roman People (Richter). Useful to the specialist and at the same time a good introduction to the subject. E. Fraenkel, Geschichte dey griechischen Nomina agentis auf -rijp, -tTwp, -rns (-t-). Zweiter Teil. Entwicklung und Verbreitung der Nomina im Attischen, Entstehung und Accentuation der Nomina auf -7ys (Schwyzer). Highly praised. There is a good index. 218 SUMMARIES OF PERIODICALS May 2. E. Kallés, Noten und Exkurse zu Archslochos [in Hungarian] (Lang). Contains much that is new and valuable. J. G. Frazer, The Belief in Immortality (Gruppe). Makes some interesting criticisms. I. Schaefer, De Joue apud Caves culto (Wide). Collects the evidence (literary, archaeological, and numismatic) for the different cults. May g. M. Dieulafoy, La batatlle d’Issus (Lammert). A critical analysis of. the MS. on the battle of the Issus presented to the French Academy by Commandant Bourgeois, in which he identifies the Pinarus with the Pajas. E. Windisch, Das keltische Brittannien bis zu ivaisey Avthuy (Anthes). Good on the etymology of the proper names. On the history and archaeology the writer is unfamiliar with much recent work. M. Schonfeld, Worterbuch dev altgermanischen Personen- und Volkernamen nach dev Uberlieferung des klassischen Altertums (K. Fr. W. Schmidt). Excellent. May 16. F. Wendorff, Die avistohvatischen Sprecher dev Theognis-Sammlung (Sitzler). Develops further the view set forth in his dissertation, ‘ Ex usu convivali Theognideam syllogen fluxisse demonstratur.’ T. Hudson-Williams, The Elegtes of Lheognis (Sitzler). ‘He has not observed that my edition appeared thirty-four years ago, and that since then my views have changed on many points.’ O. Kraus, Platons Hippias Minor. Versuch einer Erklarung (Pavlu). S.H. Ballou, De clausulis a Flavio Vopisco adhibitis (Bitschofsky). H. Endres, Die offiziellen Grundlagen dey Alexander- tibevlieferung und das Werk des Ptolemédus (Philipp). May 23. A. E. Kalischek, De Ephovo et Theopompo Isocvatis discipulis (Ammon). Car. Hosius, M. Annaet Lucani belli ciuilis libri decem, tertium ed. C. H. (Tolkiehn). For the new edition H. has collated the Montepessulanus and used Bick’s revision of the Vienna palimpsest ; he has added an Index grammaticus and an Index metricus. H. Bulle, Dey schéne Mensch im Altertum. Zweite Auflage (Herrmann). Practically a new book; it now has 320 plates, 370 pp. of text. Warmly praised. May 30. C. Clementi, Bibliographical and other Studies on ihe Perurgilium Venenss (Hosius). A useful piece of work. A. Rostagni, Swill’ autenticita dell’ tdillio VIII dt Teocrito (Sitzler). A thorough examination of the question, useful whether one agrees with the conclusion (that the poem is genuine) or not. G. Semeka, Piole- miiisches Prozessyecht (Beseler). E. Pais, Storia critica di Roma. 1. (Soltau). Unfavour- able. J. B. Bury, A History of the Eastern Roman Empive (Gerland). Warmly praised. Classical Weekly (New York). 1914. | March 7. (1) H. E. Butler, Propertius with an English Translation. (2) J. S. Phillimore, Properitus Tvanslated (B. L. Ullman). ‘ P.’s translation is more faithful and gives a far better idea of Propertius’s genius, while B.’s is much freer, and at the same time smoother and more even.’ March 14. F. Plessis, Virgile: Les Bucoliques (W. P. Mustard). The first instalment of a new edition of V. which is to replace that of Benoist. ‘The commentary is ciear and concise.’ March 28. J. Wight Duff, A Literary Htstory of Rome (E. 5. McCartney). ‘ Brings out the fact that the literature is but a reflection of the character of the people.’ April 4. (1) R. C. Seaton, Apollonius Rhodius, the Argonautica, with an English Translation ; (2) G. W. Mooney, The Argonautica of Apollonius, with Introduction and Commentary. (1) ‘Anexcellent translation, more accurate than the poetic but literal version of Way.’ (2) ‘A first-rate school-boy’s edition.’ April 25. Fr. Leo, Geschichte der vomischen Literatur (J. W. D. Ingersoll). ‘ Not a collection of monographs, but a coherently woven narrative.’ May 2. M. Niedermann, Historische Lautlehve des Lateinischen. Zweite Auflage (E. H. Sturtevant). ‘The improvements are so numerous and important that this second German edition will prove more serviceable to American teachers than the English translation of the first (French) edition.’ — ntl = tw ee - ——-* mer Cas Bru Juli the Phi Bae bea (N. ores Rui tem Six con coll of 3 ed. of | for: Hor the of t Lang). vovtality ves culto for the ; of. the andant =h, Das of the 1 much eynamen a mmlung nvivali legies of r years Platons t Flavio -xander- nmon). kiehn). ‘evision Index ‘mann). ised. Venerss 10 VIII er one ., Ptole- favour- raised. ) I. S. faithful and at he first ‘ The irtney). people.’ English ion and t literal ‘Not Auflage 1at this han the & Py the am HF — | | | SUMMARIES OF PERIODICALS . 219 May g. C. L. Brown, Latin Songs : Classical, Mediaeval and Modern, with Music. Edited by C. L. B. (H. L. Cleasby), ‘An attractive and comprehensive collection.’ May 16. Kirby Flower Smith, The Elegies of Albius Tibullus: The Corpus Tibullianum edited with Introduction and Notes on Books I, II, and IV 2-14 (B. L. Ullman). Warm praise. D. B. Durham, The Vocabulary of Menander con- sideved in its Relation to the xowy (E. H. Sturtevant). Examines the question, ‘ Did M. write Attic or Hellenistic Greek ?? ‘The result is a complete vindication; such variations from the norm as M. shows are due to the colloquial style of comedy rather than to non-Attic influence.’ May 23. C. H. Weller, Athens and its Monuments (W.N. Bates). ‘ His general plan is to translate passages in Pausanias as far as they concern the monuments and then to describe these monuments as fully as the extant remains and other evidence permit.’ Deutsche Literaturzeitung. i914. March 7. M. Wellmann, A. Cornelius Celsus (H. Westenberger). Celsus was merely the translator of a popular Greek hand-book, probably written by one Cassius, a physician of Tiberius. J. Geffcken, Kaiser Julianus (R. Asmus). Brings out strikingly the common atmosphere of the Hellenists and Christians of Julian’s time. March 21. C. Barwick, De Platonts Phaedri temporibus (W. Nestle). Places the Phaidyos before the Symposton, and regards as its object the recommendation of Philosophy in preference to Khetoric. April 4. René Cirilli, Les prétves danseuvs de Rome (F. Geiger). Disappointing. L. Friese, De praepositionum et pronominum usu qui est im titulis africants latinis (W. A. Baehrens). Shows good judgment. April 11. W. Riezler, Wetssgrundige attische Lekythen (G. Karo). The most beautiful work on vases which we possess. April 18. M. H. Swindler, Cretan Elements in the Cults and Ritual of Apollo (N. M. P. Nillson). Disapproved. May 2. W. Larfeld, Griechische Epigraphik (F. Hiller v. Gaertringen). ——EEE > -——_ X L An el Sappt Book Also | the C Wrede inscri Hdt. x i work. accoU the hi from of th justifi the p I ] in A. and | collec five A. BE deali G. k prob. Amst lanec litere Kea. to A: the \ and Tact emel signa ZUY Verc versl | Hern admi The A ychi disco adilis this testo LEN. on of Ne in imen- ve, Cf. 495- text tarch text ued). bsol. e Tay \. fil. Con- tbulls ecoris rthy To mn of lents ited. ‘ifies S10S, , An it as f VOY 10ns 2 of n of rues hen’ s of but ogy. the ome etler %. --_____.._ 4 , “> ——_. ! | SUMMARIES OF PERIODICALS XXXIII. 4. 1914. U. von Wilamowitz-Moellendorff, Neue lesbische Lyvik (Oxyrhynchos-Papyri X). An elaborate discussion of substance, metre, and language of the new Alcaeus and Sappho. The Marriage of Hector (1232, col. 2) is not Sappho’s : it was put last in Book II. as dubious, and introduced by a lost note (in the latter part of col. 1). Also some discussion of the Menander, Cicero, and Thucydides fragments, and of the Chrestomathy (1241), especially in its Alexandrian bearings. V. Gardthausen, Wiedergefundene Oviginale histovischey Inschriften des Altertums. A survey of Greek inscriptions preserved both in originals and in literary citations, e.g. Thuc. VI. 54. 6, Hdt. V. 77. XXXIII. 5. 1914. M. Pohlenz, Friedrich Leo. A long and detailed appreciation of that scholar’s work. W. Capelle, Die Nilschwelle. An extraordinarily learned and interesting account of ancient observation and theory about the rising of the Nile, examined in the light of modern knowledge. E. Bethe, Die Einhett unsevey Ilias. ‘A chapter from the book now appearing: Homer, Dichtung und Sage.’ A defence of the unity of the Iliad from a broad analysis of its structure. Books B-H in particular are justified artistically. But detailed analysis may yet force us to the conclusion that the poet incorporated older poems. Philologus. LXXIII. Heft 1. 1914. M. Boas, Die Sylloge Rufiniana. Discusses the source of the Rufinus epigram in A. P. V. 1-101. 1 (to be assigned to Rufinus) and 102 were originally the first and last poems in a ‘Sylloge Rufiniana.’ Cephalas used this as he used Straton’s collection in XII. Results for Rufinus’ date. J. Baunack, Hesychiana II1Il. Twenty- five more glosses from Hesychius (aAefeSav—eretrov) emended and explained. A. Berger, Stveifziige duvch das vom. Sklavenvecht. A long article in three parts dealing with inheritance, donatio, etc., in case of the sevuus fyuctuarius and dotalts. G. Kafka, Zu ]. Adams Evklivung dev Platonischen Zahl. A new review of the problem based on Adam, with one or two corrections. Th. O. H. Achelis, De Avistophanis Byzantit argumentis fabularum III. Discusses a large number of miscel- laneous points, e.g., mention of choregi and actors, bibliographical information, and literary criticism contained in the arguments, the difference between vrodeors and keharaov, etc. Finally gives a list of arguments that may reasonably be assigned to Aristophanes. A. Miiller, Zur Verwendung dey Musik 1m vomischen Heeve. Disputes the view recently advanced, that music was used exclusively for signalling purposes, and cites passages where march tunes seem to be mentioned. K. Lincke, Aeneas Tacticus 40. 2. Read tov pev év t7 moAe instead of tov ev TH TOAE THV pev (Hercher’s emendation). G. Esau, Zu dem lat-griech-kopt. Gesprachsbuch (Kho XIII). Alters signatus et to signatam . et, comparing Colloquium Montepessulanum. L. Steinberger, Zuv Tabula Peutingeviana. Identifies Coveliacas of the Tabula with Cavaglia, near Vercelli. W. Anderson, Die Meleagrossage bet den Tschuwaschen. A contemporary version current among this Turco-Tartar tribe. Revue des Etudes greeques. XXVI. Nos. 118, 119. F. Greif, Studies on Ancient Music. An elaborate article criticising the views of Hermann, Rosbach, and Westphal. A. Reinach, Macedonian Trophies. It is generally admitted that the Macedonians were the only Greeks who did not raise trophies. The writer examines the origin and change of this custom. A. de Ridder, Aychaeological Repovt. An article of 40 pp., covering a wide field of the most recent discoveries. : SUMMARIES OF PERIODICALS Rheinisches Museum. 69. 2. 1914. B. Schmidt, Die Lebenszett Catulls und die Hevausgabe seiner Gedichte. Jerome’s date for Catullus’ birth and death may be disregarded. From the evidence of the poems, etc., the most likely dates are 82-52 B.c. Catullus published at least two collections of poems. C. 1 is introduction to the first of these. The collection we have was made and arranged after his death. F. Ruhl, Dre Interpolationen in Prokops Anekdota. The Anecdota is not a contamination as Ranke would have it, but is a work by Procopius with well-defined interpolations (c. 12, 14-13, 1: c. 18, I-2: Cc. 12, 17: c. 18: c. 30, 34). The passage 20, 16 sqq. is no proof against Haury’s theory that the thirty-two years of misfortune are dated from the accession of Justin I. P. Friedlander, Avitische Untersuchungen zur Geschichte der Heldensage. I. Avgonautensage. The motives come from diverse places and are woven into an epic in Asia Minor under Milesian influence. To this is added a beginning and an end at Corinth. In this form the saga then appears in Eunelus. The saga was not made poetry in Thessaly because Jason is Thessalian. To suppose that is a fault in method. II. Der Krieg um Theben. A finished epos of Thebes is presupposed by the Homeric poems. This epos is not the work of Boeotia, but of the poetic inspira- tion of Asia Minor. Anhang zur Alkmatonis. More belonged to this than Bethe in his Heldenliedey will allow. Bethe’s ’Apduiapdov e£eAacia does not exist. The Alkmaionis belongs to Corinthian poetry and is not earlier than 600 B.c. III. Oiya- Aias GAwois. States more clearly the argument of PAzl. Unteys. XIX. ch. i. The deeds of Heracles, which must belong to this saga, were most probably worked up into an Epic in Samos, not in Aetolia, as Bethe would have it. Th. Birt, “Ayvworo: Oeoi und die Aveopagrede des Apostels Paulus. Seeks to show that the altars of ayvworo Gcoi bore only the inscription Jew or Geots, sometimes OEOI, which was for Gew, but might be read Geot. They were ayvworo only on the lips of ciceroni or guide-book writers. Hence there is no reason to postulate Gnostic influence. The ayvworos Geos of the Gnostics belonged to dogma and esoteric teaching, not to cult. The speech, then, is not the child of the Gnosis, nor is it the words Paul spoke. It is an artistically-composed speech, using the teaching of the Stoa to convert the Stoics, composed about 100 a.p. The speeches which Norden compares with it are not its model, but are, like it, based on models which were common property already. The attitude of the Acts to the Roman government and to the Jews fits the time of Trajan better than that of Hadrian. F. Jacoby, Drei Gedichte des Properz. Analyses and discussion of i. g (to be concluded). MiszeELtten. W. Aly, Zu Aeschylos Prometheus, v. 480. Read zaorov (cf. Theocr. 11.2). H. Mutschmann, Sext. Emp. Adv. log. 1. 339 (p. 263. 19 Bekk.) on text. G. Mercati, Un codice non veconosciuto dello Ps-Filopono sull’ Isagoge di Porfirio: namely, ff. 1-19, Cod. Vatic. gr. 309. A. Klotz, Czc. Phil. 11. 64. For infelix or hostis read infensus, which explains variant. E. Bickel, Zum chnstliche Fischsymbol. Read accipencerem (late Latin form of acipenserem) for MSS. accipiens acvem : the fish of the gourmet set against the IXOY2 symbol of Christ. Zu Mant. 1. 285. For e: read aever: fricticulae means ‘baked’ bread or fish. V. Gardt- hausen, ofvpvyxos und o€vypados. Criticism of Mentz’ theory (Rh. M. 68, p. 610). A. Brinckmann, Liickenbiisser. 13. On the fire on the Lycian Olympus. 14. On the text of Methodius on St. Nikolaos. Rivista di Filologia e d’Istruzione Classica. Vol. 42. No.2. 1914. S. Consoli, Juvenal’s Second Sative in the Litevavy Tradition to the end of the Middle Ages. Collection of quotations and references in late and medieval Latin writers to the close of the 14th century. V.Costanzi, The Relations between the Aurunct and Rome. Sceptical examination of the traditional account of the events of 504, 345, and 314. F. Stabile, Studies on the Text of the Rule of S. Benedict. On the number of e's the two 1 we kops t is [-2: ry’s . of age. an an not ault by ira- the The Xa- up TOL "TOL Dut 0k TOS its he eS us, IT’ tO 1d 5% SUMMARIES OF PERIODICALS 223 the chapters and on their headings and on points of vocabulary (e.g. occasio, ‘ motive’) and grammar. F. Garin gives from a Florentine MS. Notes of A. Politian’s Lectures on Theocritus in the ‘Studio Florentino,’ 1482-3. The notes, which are almost entirely confined to the first seven idylls, are of a superficial character. C. Cessi argues that in the Scholium on Ap. Rhod. 4. 1141 a real poem of Philetas on the Mysian Telephus is referred to. J. E. Harry, Philoctetes Inermis. On Soph. Phil. 1153-5. A. Beltrami, Euy. Hipp. 468 sq. A. Cosattini, Notes on Aeschylus. In Cho. 59 sqq. 1. erurxoret with trois pév. P. Rasi disputes inunctis in sense of ‘non unctis’ in Hor. S. I. 3.25. P. Bellezza illustrates this paradox of expression (‘enantiosemia’) from several languages. L. Valmaggi in Orthographical Tnfles points out that the arbitrary way in which the scribe of the Brussels MS. (10847) alters the spelling of the Paris MS., of which it is otherwise a faithful copy, is significant of the habits of copyists. U. Moricca, On the rdoiov 7) evyai of Lucian. On the purpose of this work (a satire on the extravagance of human desires, with no special reference, e.g., to Alexander) and its composition. Wochenschrift fur klassische Philologie. 1914. Mar. g. L. Mitteis und U. Wilcken, Grundziige und Chrestomathte dey Papyrius- kunde, 1. 1, 2; IJ. 1, 2 (Helbing). These are the historical and juristic parts. K. Burk, De Chionis epistulis (Helbing). The writer was a pupilof Plato. Th. Stein- wender, Die vomische Taktik zuy Zeit der Manipularstellung (R. Oehler). R. B. Steele, Case usage in Livy. IV. The ablative (H. Blase), March 16. E. Meyer, Dev Papyrusfund von Elephantine (C. Fries). H. W. Litchfield, Cicevo’s Judgment of Lucretius (F. Harder). L.’s interpretation declined. A. Kurfess, De Sallustii in Cticeronem et inuicem inuectiuis (Th. Stangl). ‘Thoroughly satisfactory.’ " March 23. F. Baumgarten, F’. Poland, R. Wagner, Die hellenistisch-vimische Kultur (Liebenam). ‘No less excellent than the Hellenische Kultur’ Th. Birt, Rimische Chavakterkipfe (H. Philipp). A kind of Roman history by biographies. March 30. Windberg, De Hevodoti Scythiae et Libyae descniptione (E. Juhre). F. Haverfield, Ancient Town-Planning (E. Hohl). ‘Gives a clear account.’ W. Gerhausser, Dey Protreptikos des Posetdontos (H. Mutschmann). ‘ The foundation for all further research.’ April 6. A. Meillet, Apergu d'une histoive de la langue grecque (K. Wagner). ‘Most strongly to be recommended.’ A. Miller, Asthetischer Kommentar zu den Tvagidien des Sophokles (F. Adami). Theban Ostraca: I. H. Gardiner, Hieratic Texts ; II. H. Thompson, Demotic Texts ; III. G. Milne, Greek Texts; IV. H. Thompson, Coptic Texts (A. Wiedemann). ‘ Done with great care.’ April 13. N. Wecklein, Uber Missverstindnisse dlterer Wendungen und Ausdniche bet den gviechischen Dichtern, insbesondere bet den Tragtkern (J. Sitzler). A. Ruppel, Konzeption und Ausarbeitung dev Avistophanischen Komodien (E. Wust). Not all the conclusions correct. C. H. Haile, The Clown in Greek Literature after Aristophanes (E. Wiist). The Swpoddxos is restricted to the old comedy. April 20. L.H. Galiart, Bettraiige zur Mythologie bet Bakchylides (H. Steuding). On the Meleager legend, the Heracles-Deianeira legend, and the Zeus-Io legend. G. de Sanctis, ’"ArOis. Storia della vepubblica ateniense dalle origini alla eta di Pericle. 2. ed. (F. Cauer) I. April 27. K. Brugmann, Vergleichende Laut-, Stammbildungs- und Flexions- lehve nebst Lehve vom Gebvauch dev Wortformen dev indogermanischen Sprachen. II. Lehre von den Wortformen und ihrem Gebvauch. 3 Teil, 1 Lief. (G. Gintert). E. Michon, Un décvet du deme de Cholargos velatif aux Thesmophories (WW. Larfeld). G. de Sanctis, "ArOis. Storia della vepubblica ateniense dalle ovigini alla eta di Pericle 2. ed. (F. Cauer) II. 224 SUMMARIES OF PERIODICALS May 4. P. Masqueray, Bibliogvaphie pratique de la littévature grecque des ovigines a la fin de la période Romaine (R. Wagner). ‘An ever valuable work.’ A. Clausing, Knittk und Exegese dey homevischen Gletchnisse im Altertum (Chr. Harder). J. S. Reid, The Municipalities of the Roman Empive (M. Gelzer). ‘A pity that what is right and reasonable in the book will not be fully appreciated owing to its defects.’ May 11. Schneidewin and Nauck, Sophokles, erkl. von S. und N. 5. Bandchen. Elektra. 10. Aufl. Von E. Bruhn (F. Adami). R. Nichard, Le probléme des Bacchantes d’Euvipide (K. Busche). Directed against the views of Norwood and Verrall. W. Kohler, Dze Versbrechung bet den gviechischen Tvagikern (Reinert). Th. Birt, Praefatio zu Propertii codex Guelferbytanus phototypice editus (G. Friedrich). ‘A model introduction.’ May 18. Greek Liutevatuve. A Series of Lectures delivered at Columbia University (R. Wagner). Lectures by P. Shorey, H. W. Smyth, E. D. Perry, J. R. Wheeler, E. Capps. J. Petri, Poetische Dokumente griechischen Lebens und Denkens(R. Wagner). ‘The work of an expert.’ C. Morawski, De contentionibus litterartis apud Romanos, imprimis apud Ciceronem (W. Isleib). May 25. F. M. Bennet, Religious Cults associated with the Amazons (H. Steuding). ‘A rich collection of material.’ S. Kriegbaum, Dey Uysprung dev von Kallikles in Platons Gorgias vertvetenen Anschauungen (H. Gillischewski). E. Bevan, Stoics and Sceptics (M. Dibelius). ‘ Impressionist in the best sense.’ June 1. F. W. Hall, A Companion to Classical Texts (W.Schonack). ‘ Will find many readers in Germany.’ F. Strenger, Stvabos Evdkunde von Libyen (H. Philipp). L. Volkmann, 7. Lucretius Carus, der Jiingey Eptkurs (J. Tolkiehn). LANGUAGE. Glotta. VI. Band, 1 Heft. 1914. M. Lambertz, The Etymology of SovAos, which the writer regards as a word borrowed by the Greeks from the non-Greek inhabitants of Asia Minor. Witkowski, Contributions to Greek Syntax, rejects the ‘prescriptive’ optative proposed by Del- briick and accepted by Brugmann, regards Aafe yovvwy and the like as a partitive genitive (against Brugmann-Thumb, who treat it as a variety of gen. loci); the genitive after épxw depends on the implied substantive (I am ruler of) and should not be taken as an adverbial genitive with Brugmann-Hirst; sections on the genitive after ov, peta- to denote change, and parataxis in Homer. R.,. Methner in a long article deals with The Development of the Qualitative Ablative, which he derives from (1) an instrumental of the accompanying circumstances, (2) enallage of an adjective with an ablative of limitation, e.g., «iv magno ingenio originated in wiv magnus ingenio. A. Sonny, Demonstratives as Indefinites ; e.g. onpepov 7 avprov wopevoopeba eis THvdE THV médw, Epist. Jac. 4.13; ‘expers consilli, inquies, haec atque illa temptans,’ Sall. or. Phil. 11. Persson on Latin u in Unaccented Open Syllables. Indogermanische Forschungen. XXXIII. Band, 3, 4 Heft. W. Frhr. v. Osten-Sacken contributes a series of comments on Walde’s Lat. Etymologisches Worterbuch chiefly with reference to parallels in the Slavonic lan- guages. V. Grienberger discusses in detail the text and exposition of the Old Latin inscription from Lucera. Thumb once again defends Late Laconian against the oft-repeated charge of artificial origin. Brugmann on Greek Desideratives in -cetwy and xeiwy, dyelovres= der idvTes, Kelovres=Ke[e]e tovTes; Of pev KaKkeiovTes EPav oixovde Exaoros (Il. 1. 606), ‘they went home, occupying their thoughts with lying down,’ xe:- being an instrumental or local dative. que des ovigines A. Clausing, ). J. S. Reid, at is right and S. . 5. Bandchen. e probleme des Norwood and keyvn (Reinert). G. Friedrich). nbhia University J. R. Wheeler, s(R. Wagner). apud Romanos, (H. Steuding). on Kalltkles in an, Stoics and ). © Will find ¢ (H. Philipp). ds as a word . Witkowski, oosed by Del- as a partitive ven. loct); the nd should not the genitive ner in a long » derives from f an adjective magnus nmgento. a eis THVOE THV tans,’ Sall. or. Walde’s Lat. Slavonic lan- the Old Latin n against the sideratives in LKKELOVTES EDaY ts with lying