THE CLASSICAL QUARTERLY EDITED BY PROFESSOR E. V. ARNOLD, Litt.D. AND F. W. HALL, M.A. BOARD OF MANAGEMENT: J. W. MACKAIL, M.A., LL.D., Chairman. Prof. R. S. CONWAY, Litt.D., Hon, Treasurer. E. HARRISON, M.A., Hon. Secretary. » . With the Chairman, Tvea- | H. P. RICHARDS, M.A., Prof. en et and Secretary vepre- Representing the Oxford Philological Society. > DAN ) | senting the Council of the Prof. W. RIDGEWAY, LL.D., Litt.D., Sc.D., W. B.D. PANTIN, M.A. Classical Association. Representing the Cambridge Philological Society. With the co-operation of Prof. WM. GARDNER HALE, UNIversity oF CuicaGco; Principal W. PETERSON, LL.D., C.M.G., McGitt University, MontTrREAL; Prof. T. G. TUCKER, Litt.D., UNIVERSITY OF MELBoURNE; Prof. JOHN WILLIAMS WHITE, LL.D., Litt.D., HARVARD UNIVERSITY. VOLUME VIII - PUBLISHED FOR THE CLASSICAL ASSOCIATION BY JOHN MURRAY, ALBEMARLE STREET, LONDON, W. AND GINN & COMPANY, 29, BEACON STREET, BOSTON, MASS. 1914 TABLE OF CONTENTS No. 1. PAGE Notes on the Agamemnon. A. S. F. Gow , , , ' ' : ‘ I E mendations in Plato’s Phaidvos. H. G. VILJOEN . , , 7 On the Aristotelian use of Adyos: A Reply. J. L. Stocks . . , 9 "AOnvaiwv ToAureta, XXX. 3-4. J. A. R. Munro . ' , ‘ , ‘ 13 Note on Thucydides II. c. 48, § 3. W. Ruys RosBerts : ; , 16 Notes on Dares and Dictys. R. T. CLarK . ‘ ; ; ' , ' 17 On some passages of Ovid’s Tristia. S. G. OWEN . : , , , 21 The Grammatical Chapters in Quintilian I. 4-8. F. H. See : , 33 Salapantium disertum. H. W. Garrop , ' , : ' : 48 Etymologies and Derivations. Epwin W. Fay . : : , :, 50 Summamnies of Periodicals : Literature and General . , : ; ' , : 61 Language . ‘ , ; , . ; , ; , ' 71 No. 2. Thucydidea (continued). H. Ricuarps . , : ' ' 73 Theocritea. ArrHuR PuLaTr . ' , ; : . ' ‘ , 86 Boethus of Sidon. J. F. Dosson . ' , : , , ' 88 XPH and AEI. Tuomas D. GoopEL. . ' , , : , ; gI A neglected Manuscript of the Moretum. A. Souter . eee s: ‘i; 103 Note on Thucydides VII. 28.3. S.C. Booker . ; re, ast a 104 Notes on Lucan VIII. W. B. ANpDERsoN . ; : te. & has . 105 On the chronology of the Fronto Correspondence. C. R. Haikes.. A ssae + )6|6ERS Neaera asa Common Name. J. P. PostGaTe . : a . 21 The Caesura in Virgil, and its bearing on the authenticity of the — Vergiliana. W.G. D,. ButcHer . ; : ' , . es Claudius and the Primoves Galliae. H. J. CUNNINGHAM ' : ~~ « Summamnies of Periodicals : Literature and General . ' ' ' ; ‘ , 134 Language . TABLE OF CONTENTS No. 8. Notes on Horace. CHARLES E. BENNETT A Transposition in Propertius. A. E, Housman : , , : The Text of the Bellum Gallicum and the work of H. Meusel. T. Rice HoLmMEs The importance of the Veronese Palimpsest in the First Decade of Livy. Crara M. KNIGHT : On the value of Modern Greek ior + the study of Ancient Greek. ALBERT THUMB ' , , : A false quotation from Plautus. F. W. Har Verse-Weight. E. Harrison. , Rainbow, Sky, and Stars in the J/iad and ee a Chorizontic Argument. G. H. Macurpy Summarves of Periodicals : Literature and General Language . _ No. 4. On certain Fragments of Pindar. E. B. Capp Vindiciae Platonicae I. JoHN BURNET . , , . On the Text of the Stvomateis of Clement of eesitities J. P. PostGaTE The Codex Bambergensis of the First Decade of Livy (1.-VII. 17) with illus- tvations. A. H. Kyp In Memoriam—Alastair Hope Kyd. R. S. C. , ; Notes on Ovid’s Ibis, Ex Ponto Libri, and Halieutica. S. G. OWEN Notes on the Naturales Quaestiones of Seneca. H. W. Garrop , : Claudius and the Primores Galliae: A Reply and a Restatement. E.G, Harpy The Paean of Philodamos of Scarpheia. J. U. Powe. . Summanies of Penodicals : Litétatuze ‘and Geieral: eaters} INDICES TG. VoL; ViFr: ; rAGE 145 151 156 166 181 205 206 212 216 224 225 230 237 253 25+ 272 282 288 289 296 297 pre W fot ing dis Wl wl KNE Tro ERT ent. llus- RDY 181 205 206 212 216 224 225 230 237 248 253 25+ 272 282 288 289 296 297 THE CLASSICAL QUARTERLY -_—_ — JANUARY, 1914. NOTES ON THE AGAMEMNON. (Wecklein’s numeration.) 259 sqq. (The chorus have narrated the events up to what immediately preceded the actual murder of Iphigeneia.) vw 9 v7 9 \ > s > / ta 8’ évOev ov7’ eidov ovt’ evvétro* 260 réyvas dé Kddyxavtos ov« adxpavrTor. Aina &é tots pev trabod- ow pabeiv érippérret’ TO péAXov 0 3 \ / 9 XA 4 / ETEL YEVOLT AV KAVOLS* TPOYXALPETw* 265 icov 6€ T@ TpoaTéver. Topov yap H&et cvvopOpov avyais. / 9 9 , \ 4 v TwéXowTo 8 ovv Tat TOUTOLOLWW EvTTpA- gus" 263 6’ add. Elmsley, 264 ézet yévour’ a f h, émiyévorr’ M., 266 cuvopOpov Wellauer, cvvopfov M., avyaits Hermann, avtais M. Ll. 263 and 264 have been much vexed, and a string of conjectures will be found in Wecklein’s appendix. All of them?! produce roughly the same mean- ing—‘ it is useless to enquire into the future, which (I. 265) is bound to be disastrous.’ My complaint against these readings is that the whole antistrophe yields with them no reasonable sense. It is not true that the future ‘ will dawn clear with the light of day,’ and if we prefer to read cvywpov atais (Ahrens) or the like, still no reasonable meaning can be attached to 267 sqq. ‘ The future is bound to be disastrous, but at any rate may prosperity be ours henceforward ’ is an absurd prayer, and ‘the immediate future will be disastrous, but let us 1 Except Margoliouth’s 6 uéAAwr 8 érei yévnrac munc lugere, ueniet autem luce clavior, This also kNéos, mooxaipérw* | lrov 5¢ rov rpooréveww, which he produces no reasonable sense from the whole renders qui gauisurus est allato nuntio de excidio passage. Trotae is iam nunc gaudeat ; alium autem pay est iam NO. I. VOL. VIII. A 2 A. S. F. GOW hope that the remote future will be prosperous,’ if this meaning can be tortured from the Greek, is not much better. The landmarks in the sentence are, I think, two in number. One is that Agamemnon has to pay for his sin in sacrificing Iphigeneia (261 sq., cf. 228 sqq.),’ and the other is that tami tovrovow means ‘ from this point on,’ and that the prayer is for the immediate beginning of a new and prosperous era. It follows, therefore, that the period which the chorus know to be certainly lamentable must be the ten years which have elapsed since the sacrifice of Iphigeneia, during which the flower of Argos has been dying at Troy (cf. 435 sqq.); and of this period, if Clytemnestra’s preparations mean any- thing, they may reasonably expect to have information in the morning. What is to be done with the Greek I do not know. L. 259 must, I think, refer not to the ten years in question, but only to the actual sacrifice at which the narrative has stopped short; 1]. 260 alludes, therefore, to Calchas’s forebodings of that sacrifice (154 sqq.), not to his general prophecies as to the campaign (131 sqq.). In 1. 263, if ro wéAXov meant ‘ what was then in store for them,’ all would be well ;? but the neighbouring tenses forbid. Therefore TO wéAXov is Corrupt : To Aowrov would give the required sense after a fashion. In 264, whatever word should replace ro péAXov, the subject of yevolro is TO KAvewv, and the meaning is: ‘ you will hear < what has happened since > all in good time: till then let it be.” The general sequence of thought in the antistrophe is: ‘Agamemnon’s sin must have brought in its train disasters, of which we shall hear soon enough; let us hope that they are now ended and fortune once more favourable.’ 427. oppatov 8 év axnvias éppet tao’ ’Adpodita. Cf. Ach. Tat. vi. 6, parota yap év trois ofOarpois KaOntas TO KadXos, and see the whole passage. Eumathius iii. 7, rovs of@arpovs xaTepirnoa Kai OXov eis THY YruynY avipnoayny Tov Epwra* oPOarpos yap Epwros Tynyn. Both suggest that the eyes are those of Helen not of Menelaus. 536 tiec0ar 8’ akuwtatos Bpotav tav viv: Lapis yap ote cuyTEeAns TONS éfevyetat TO Spapua Tov maous Tréov. ‘Agamemnon,’ says the herald, ‘is the most august of men—that is of men now alive, for Paris. ...” Any unprejudiced person, asked to complete this sentence, would, I think, goon: ‘who was formerly more worthy of honour, is now dead.’ But that is not what Aeschylus says; on the contrary, he says 1 I take it to be certain that 261, 2embodythe attach ro péddopr to this sentence. doctrine of instruction by suffering (cf. 187 with 2 If ro wéAXov is taken with yévoiro alone, there Headlam’s note) and discount conjectures which is no subject for wpoyawérw. of lov as: pos Edi cor sta: def be tortured yne is that sq., cf. 228 ,’ and that us era. It e certainly sacrifice of y at Troy > mean any- ng. 59 must, I sacrifice at 9 Calchas’s ; as to the n in store Therefore a fashion. yévotTo 1S since > all eht in the lisasters, of ended and TO KdAXOS, bitnoa Kat 1yn- Both -that is of o complete of honour, ry, he says ro alone, there er NOTES ON THE AGAMEMNON 3 that Paris, so far from being tieo@ar aéios, was ofAwv aprayijs Te Kal KrOTIS dixnv and has deservedly paid the penalty for his crimes. I have translated Tay vov with emphasis, because if it were not emphatic it would not be there at all. “A&uwratos Bporay is a simple superlative (like, for example, Ch. 301 evkrecoTatous BpoTav) : THY voy an unexpected qualification of the superlative which demands an explanation.’ Since the qualification is not explained, and must in any case be irrelevant to this context, I take trav viv to be corrupt. If we consider further the relation between Agamemnon’s honour and the punishment of Paris, it is obvious that the relation is temporal rather than causal. Agamemnon has only been the most august of men since Paris paid the penalty ; so long as he remained unpunished Agamemnon’s reputation has been under a cloud. Write, therefore, To viv or tavov for rav viv. Nov does not occur elsewhere with the article in Aeschylus, but to mpiv appears more than once, to mpooGe in 1. 1437 of this play, and ta mpdcGe in 1. 19. The punctuation should be, I think, . a&twTatos Bpotav: tavov Ilapis yap «.T.X. 831. ‘ By the vague doxodvtas, “‘ certain ones in appearance,” he conveys to them that he is quite aware of Clytemnestra’s insincerity’ (Headlam). Verrall (p. 220) also holds that Agamemnon suspects Clytemnestra, and these are the only two opinions on the subject I have been able to discover. For the understanding of the Agamemnon as a whole it perhaps matters little whether Agamemnon knows what has been going on or not, but it makes a good deal of difference to our conception of this curious and ominous scene, and Aeschylus has, as it seems to me, been at pains to show that Agamemnon has no suspicions. The chorus have said in effect (779 sqq.) ‘appearances are deceptive and sympathy is often feigned, but to a discerning man the eyes betray those My congratulations are sincere, but time To this whose sympathy is hypocritical. will show who has been faithful, who unfaithful, in your absence.’ Agamemnon presently replies (821 sqq.): ‘I entirely agree with you, Tavpols yap avdpav €oTt aovryyeves TOOE, hirov Tov evtuxovvT’ avev POovov céPeww. I know from my own experience how empty appearances of friendship are, for of all those who came to Troy with me Odysseus alone was whole-heartedly loyal. For the rest, political and religious questions shall be discussed in the assembly, and we will take measures to correct anything amiss.’ 1 The emphasis is due not to the metrical dvip 6 dyaiés and 6 dvyp 6 dya@ds involve a con- position of the words rév viv (Headlam, On trast. Oi viv and oi viv dvOpwra are sometimes Editing Aé., pp. 5 sqq.) but to the order. Ac- barely distinguishable in sense from dv@pwmra, but cording to Kiihner-Gerth, G.G., p. 613, the sub- _to defend rév viv here we require examples of oi Stantive in such phrases as dviyp 6 dyabés is dvOpwro oi viv and dv@pwra oi viv so used; and defined only by the following phrase, and both I do not know of any. 4 A. S. F. GOW This conversation must be designed to show that Agamemnon suspects nothing. The chorus meant that Agamemnon must not be misled by the simulated warmth of Clytemnestra’s welcome, but Agamemnon, when he embarks on a disquisition on the effects of jealousy, and speaks of constitutional measures to combat the evil, shows clearly that he has misunderstood. In Clytemnestra’s attitude to her husband jealousy has no part, and her infidelity, if known, is surely matter for instant action by the king himself, not for leisured consideration and public debate. In Agamemnon’s subsequent con- versation with Clymnestra he addresses her as ‘guardian of my house,’ and shows, if I read him aright, some impatience at her preparations, but a willing- ness to do as she wishes (935, 947), and even to rally her on the extravagance of her welcome and her obstinacy in argument (g06, 931). Nowhere does he treat her as he would be bound to treat her if he knew the truth,! and surely much of the poignancy of the situation depends on the confident self- sufficiency with which he fails to understand the hints of the chorus. It is this failure, as much as the generally disquieting course of the subsequent dialogue, which arouses in the chorus the forebodings expressed in 966 sqq. 1028 of 8 ovrot’ éXricavtTes Huncay Kaos @mot Te OOVAOLS TaVTAa Kal Tapa oTdO un». Exers Tap npiv oldtep vouiteTat. 1030 e&e¢ Auratus. Opinion is divided as to whether wapa ord@unv means ‘ beyond’ the cra@un or ‘according to’ it, ‘unreasonable’ or ‘precise.’ The phrase, however, occurs elsewhere,” and always with the second meaning ; and since ord@yn is a plumb- line, TO omaptiov: otdOunv iOuvtevh porttBaxOéa (Suidas), mapa might perhaps be expected to have the force it has in wapa motauov, not that which it has in tapa vouov. But if apa ord@unv means ‘exact,’ ‘ precise,’ then @pot Kat Tapa oTd0unpv is a ridiculous anticlimax, and it is probable that apa ordOunv belongs to the next clause. Moved, I suppose, by these considerations, Wilamowitz writes : <(mixpot TroXtTals Etat yadv’ émnppévot > , @pot Te SOVAOLS. TaVTa K.T.X. but it is unnecessary to suppose a lacuna. We may accept ye from Thiersch® and write: > , A , \ 4 / @pmol ye SovAOLS TavTa* Kal Tapa cTaO un éEeus trap’ nutv olamep vopilerat. ‘Here you will have the regular allowances, and they will be regularly 1 Andas yéveOXov in 905 is no doubt, as Verrall Jon 1514 the metaphor seems to be different. says, a sinister address, but it is not necessarily 3 I have not seen Thiersch’s note, but I sup- so meant by Agamemnon, and the whole scene _ pose he wished to translate ‘ cruel even to excess.’ is packed with dramatic irony. The form of If the ye is to be classified, it may be called words is not in itself remarkable ; cf. 775. apodotic. Examples of this usage are given by 2 Theognis, 543, 945; Soph. fy. 421. In Eur. Neil in his edition of the Knights, pp. 199 sq. | aman meni el — Aan _ - — 1 suspects ed by the when he stitutional stood. In infidelity, lf, not for juent con- iouse,’ and a willing- ravagance re does he and surely ident self- rus. It is subsequent 66 sqq. oTtaOun or ver, occurs s a plumb- ht perhaps lich it has Nn wot Kal oa oTaOunv . Thiersch ? e regularly ‘different. ote, but I sup- ven to excess.’ nay be called > are given by Ip. 199 Sq. NOTES ON THE AGAMEMNON 5 distributed’: there will be, for example, no unfairness such as Sceledrus com- plains of at Plaut. M7/. 348. 1321 amak ét’ eitreiv phow 7 Opnvov 0érw €L0V TOV AUTHS. I agree with Professor Housman (J. Phil. 1888, p. 278) and Professor Platt (C.R. 1897, p. 96) that pjow 7 is intolerable, and with Professor Platt that it is probably patchwork. Remove pow 4, and the only natural and tolerable supplement seems to me to be mplv Oaveiv. Since, therefore, rpiv Qavety may very easily have fallen out owing to the repeated -esv, I conclude that that is what Aeschylus wrote. _ Professor Platt takes exception to elzeivy Ophvov, but it is defended by eye Tratava at 650. The paean there contemplated is a messenger’s speech, that is to say senarii, and Cassandra’s Opivos is also in senarii. It is a Opjvos in substance but in form a pjois, and eivreiv seems to me appropriate to its peculiar character. In any case the copyist who found a blank and filled it with pjow 7 did so because he had eizreiy before him. I would add that the word @pivos, which is certified here both by ama€ éru and €uov Tov avris, seems fatal to Headlam’s proposal to assign 1326-1329 to the chorus. The whole speech 1321-1329 will barely pass muster as a Opivos as it is, and it will only pass by virtue of 1326-1329, which are in some sense Cassandra’s epitaph. Give those lines to the chorus and she is left with only a curse, which is neither a Opfvos nor an adequate sequel to the weighty preface of 1321 sq. If something is lost after 1325, then the rest of the speech may go to the chorus; otherwise it must all be left to Cassandra. 1536 Aixa & én’ adrXo trpaypa Onyer BraBns mpos adrXas Onyavats moipa dixah Onyavec Hermann. What is required in this sentence is an object to Over or Onyaver. Musgrave’s Onyavais payaipavy or Housman’s @nydvatow daop would either of them do very well but for Ch. 643 Aixas & épeiderar muOpunjv- | mpoxadxever 8 aica dacyavoupyos, which defends potpa’ and, as a consequence, involves the dative Sica. Emperius’s 07yeTas does not seem to me satisfactory, nor Auratus’s dt«ny, nor Headlam’s tentative BAadBas. Bradys is not easily emended into any word meaning ‘ sword,’ it is in itself appropriate to the context, and BraBys Onyavais, if that is how the words are to be taken, receives at least a qualified support from Eum. 861. There is therefore no room in the existing sentence for the missing object? 1 Cf. also Ch. 909 sq. the unpleasant line Alxg 6’ é€x’ Addo wpayp’ dop 2 It can just be got in by writing, with Heusde, Ovex BAdBns. 6 NOTES ON THE AGAMEMNON and room must be made for it. Probably éz’ a\Xo mpayua is a gloss on mpos adras Onyavats, and, if this is so, Ones may well be right, for it is the form used by Aeschylus elsewhere, and syllabic correspondence is not required in lyric senarii.1 Aeschylus’s actual words are of course lost beyond recall, but he might, I imagine, have written something like: Aira S€ kawov dacyavov Onyer BAaBns pos adrats Onyavaior potpa. 1651 XO. cia 8 Eidos wpoKwrov Tas Tis evTpeTLETO. / , a AI. adda pny Kayo TpoKwTOS OVK avaivopat Oaveir. The repetition of wpoxwros in 1652, where it bears an entirely different and otherwise unknown sense,” has been received with unusual tolerance by editors, and Wecklein’s appendix records no objections to it. To me it seems exceed- ingly clumsy, and I suppose it to be due to a copyist’s reminiscence of the preceding line. The word which Aegisthus should echo is not mpoxwos but Eidos, and the most obvious correction is £u@ypns, which I find has already been proposed by Herwerden (Mnemosyne, 1892, p. 446). The word &udypns, how- ever, appears first in Euripides, and we should perhaps borrow from Eum. 595 the Aeschylean E:govdxos. A. S. F. Gow. TRINITY COLLEGE, CAMBRIDGE. ' Cf., e.g., Eum, 169=174. (cf. Suidas s.v.: cf. Eur. Or. 1477). It is still 2 The lexicons quote Herodian vii. 5, éywv far from the supposed Aeschylean use. mpoxwrov Thy defidv, ‘hand on hilt,’ which is an 3 Eum. 595 étpoudkw yxerpi: cf. Eur. Phoen. 363 extension of the ordinary @ywv mpdxwrov ro tipos = Ecgjpn yxetp’ Exwr. DSS ON mWpds is the form required in call, but he V. ferent and by editors, ms exceed- nce of the KwTrOS but ready been npns, how- | Hum. 595 Yr. Gow. ). It is still ise. ir. Phoen. 363 EMENDATIONS IN PLATO’S PHAIDROS. Phatdros, 230d. THE sentence @ozrep yap ot .. . ayovaewr, as it stands, seems to be unsatis- factory. The meaning is left too vague. We expect some such subject as Bov«onXor or Trowpéves after of, and the object (ta wewavtTa Opéupata) of ayovow is placed awkwardly next to that of wpoceiovres. We feel that wpoceiovres goes too closely with adyovow to have anything to do with the article. It may be entirely detached by taking it more or less as a parenthetical remark. Instead of the article we must read the relative, as its antecedent éxeivos can be dropped, according to other examples given in Ast’s Lexicon Platonicum. Cf. Gorgias, 511e: ElO@s OTL OVSEV AUTOVS BeXTioUs éEFeBiBacer 7 of evéBnoar, K.T.X. In these two passages the relatives corroborate the evidence of B,T,P, and editors without reason follow F,t, here in reading ofoz. So the sentence runs: @o7ep yap of Ta Tmewavra Opéupata, Padrov 7 Twa \ , KAPTOV TPOTELOVTES, AYOUCL, K.T.X. Phaidros, 236c. The absolute use of evraBnOnte. Cobet tries to solve the difficulty by cutting the knot. Schanz and Burnet follow him in deleting this verb. It seems that the only objection taken to its presence is that its correct meaning has been a puzzle to editors. When the eye of a scribe has been deceived, or when a copyist has disturbed a passage, it is only fair to suppose that the source of error is to be discovered on the two pages of the MS. open before him at the time of writing the sentence in which the corruption is found. Now, no other passage occurs on the two pages of the facsimile B before us in which the presence of this verb is in the least required to give sense; nor can we entertain the idea that it has found its way into the text from ascholiast’s marginal note. There is no occasion for annota- tion. We may add that this form of the imperative is too old and sound to be looked upon as more or less the usual thing a scholiast’s fancy hits upon for annotations. So the only way open to us is to leave it undisturbed in its present position and explain it as best we can. If we punctuate differently by putting a comma and a dash after aAAnAous, the sentence runs: pytéov perv yap . aAXHXOLs,—evAABNOnTL Kal un Bovrov x.t.rX. In Sophocles, O.R. 47, we 8 EMENDATIONS IN PLATO’S PHAIDROS. find this verb used absolutely: 76’, evAaBnOne’, os «.7.X. ‘Be on your guard’; or, as Jebb explains: ‘ Have a care for thy repute.’ Here, too, in our passage a warning note is sounded, as is shown by the anticipatory thought: ovTas dmrws olos TE El. Cobet’s method robs the sentence of its artistic effect, whereas its dramatic interest is fully sustained when we interpret the verb in the light of Sopho- clean usage. The final clause is to be linked to the first part of the sentence, and not as T blundered by inserting a 6é after va, making it dependent on evraBnOnrt. H. G. VILJOEN. THE UNIVERSITY, GRONINGEN, HOLLAND. ur guard’; ur passage geht: ovTws ‘s dramatic of Sopho- e sentence, yendent on VILJOEN. ON THE ARISTOTELIAN USE OF aorot: A REPLY. In the June issue of the Classical Review Professor Cook Wilson announces his conversion to the view that in ‘a well-defined group’ of passages in the Nicomachean Ethics Xoyos means Reason. While I cannot hope to re-convert Professor Cook Wilson, I feel that it is worth while to try to express the reasons for which it seems difficult to follow him. The most central of the passages in which he gives Adyos this meaning is the definition of 70:«7) adpery in 1106 b 36: eat dpa % apetn Ets mpoatpeTiKn, €V MECOTHTL OVTA TH Tpos Huds, wpLouévn AOYH Kal ws (Or ©) av oO Ppovipos opicevev. The fact that this is a definition precludes loose or popular phraseo- logy, and if Xoyos here means Reason, it must be admitted that such a use of the word is regular or technical in Aristotle’s philosophical writing. My chief objections to accepting the suggested translation ‘ prescribed by Reason’ are as follows: (1) There is, so far as I know, no list of ‘ parts’ or faculties of wux7 in the works of Aristotle which contains the term Aoyos. In the Ethics itself we are familiar with the phrase 70 AXoyor éyov, which implicitly denies that Aoyos is a faculty; and in the sixth book we have (a) a list of the three controlling principles of wpa&is and adnOeva (aicOnots, vods, peEis—1139 4 18); (5) a list of the five sources of true judgment (réxvn, ériotnun, Ppovnors, copia, vovs—1139 b 16): in neither list is Xdyos to be found. Its place, we shall have to say, is taken in both cases by vods. But surely it is unlike Aristotle to use in the actual defining formula of one of the central terms of the treatise an ambiguous synonym for vovs, without a word of explanation, instead of the unambiguous term voids itself. (2) The form of the phrase wpicpévn royo seems to me to give a presumption against the translation ‘prescribed bv Reason.’ When the passive voice of opifeww is joined with a noun in the dative, that noun is never the defining faculty, or judge, or legislator, but rather the standard, or measure, or mark, by which the definition is effected : it is, in fact, the dpos in question. The middle use is similar. “Opifec@ai te Twi means to distinguish something by a certain mark, the mark being in the dative. In the active the subject is properly the mark, but often there is a personal or quasi-personal subject, as in the concluding clause of the sentence under discussion. The following are typical instances—(a) passive: de An. 435 616 7H adh dpiota To Chv; Pol. 1235 a 23 wavta To épyy wpiora; N.E. 1115 b 22 opiterar Exactov TH TéAetL—(b) middle: N.E. 1166 a 10 rovtwy 6€ tiv Kal tHv diriav opitovrar: N.E. 1170 a 16 70 8€ Sv opifovras Tots Swous Suvvaper aicOncews—-(c) active: De Caelo 293 b 13 10 8 opitov To mépas: Ib. 297 6 34 10 J. L. STOCKS 0 opitwy Kixros; Phys. 219 a 22 Tov xpovov yvwpilouer, Stay opicwpev THY Kivnow, TO TMpoTepov Kal vatepov opitovres. I am unable to find in Aristotle any parallel to the phrase wpicpévn Oy if Aoyos is taken to stand for the faculty of Reason. The phrase ought to mean that, in one of the various applications of the term dpos, a Adyos is the dpos of the peodrns in which 70: apety is said to consist. If it means that, the translation ‘ reason’ seems to be out of the question. (3) Instead of following Professor Cook Wilson in detail into an examination of the various passages which he adduces to illustrate and support his contention, I will now try to state the interpretation of Aoyos which seems to me to be the true one. In so proceeding I follow the Aristotelian principle that the proofs of one theory are the difficulties of its rival; and the procedure in this case is the more necessary because Professor Cook Wilson has done less than justice to the view from which he is a convert. I start from the following passage of the De Partibus Animalium : ss . ¢ aA = , \ / mpos S€ TovToLs, émel mreElousS Op@pEV aiTias TrEpl THY yéverow THY HUaLKNY, +f 4 * \ ge \ - , / \ \ olov THhv Te ov Evexa Kal THY BOev apy? THS KWHoEws, SiopiaTEoY Kal TEpL , , , \ , , , 4 , A / ed A TOUTWY, TOLa TpwTH Kal SevTépa TréhuKEVv* HaiveTrar OE TMPwTNH, hv NEyouEV EvEKA e - \ / . 2 na TLvOS* AoOryos yap ovTOS, apy 5’ Oo AOYos Opmotws ev TE TOls KATA TEXVHV Kal EV TOUS A f x \ a , a a > , - , e ‘ + ‘ \ gvoe cuvestnKkoow. 1 yap TH Svavoia } TH atcOnoe opiadpevos O pev LAaTPOS THY ™ . e ; > \ > ) , \ , \ \ > _/ * vylecav, 0 8 oixddopos THY oikiav, amrodioacL TOUS OYOUS Kal Tas aLTLAaS OU Tovovaw ExaoTov, Kai ioTe TroLnTéov ovTwS (63g b II-I9Q). The passage deals, it is true, with téyvy not dpety, with roinovs not mpaéis, but the comparison of the two forms of activity given in N.E. II. iv. does not exhibit any distinction of principle between them on the intellectual side, while the closely parallel definitions of dpdvnois and téxvy given in VI. iv.” suggest an identity of principle. Of the various kinds of cause, says Aristotle, the primary is the final cause, which is a Adyos or plan. Such a plan is at the bottom equally of every fact in nature and of every production of art. With the aid of thought and sense the doctor comes to apprehend the health, the architect the building which he intends to create; and in this way each provides himself with a plan or ideal which is the justification or explanation of his actual procedure. Health in a particular person in the one case, a particular building in the other, is the product of the completed activity ; and it is of the essence of the activity that foreknowledge of this result is present in the agent. It is this that constitutes his art: for 7 téyvn Noyos Tod Epyou 1 Nothing in Professor Cook Wilson’s article is sO surprising as the last paragraph: ‘ The conclusion then is that in all the passages re- viewed \éyos means Reason in one of three senses ; either (1) reason as the faculty of reason, or principle of reason in the soul, or (2) reason as reasoning, or (3) reason as what is reasonable, in the sense of the deliverance of reason—reason as ordaining the moral law, reason as inculcating it, or the moral law itself as a form of reason.’ If by translating ‘ reason’ the translator commits him- self to nothing more than to one of the three senses of the word there given, the word is monstrously ambiguous, and had better vanish from our philosophical vocabulary. 2 réxvn is n wera Abyou (or pera Adyou ad7Oois) wounrixh es, while @pdévynors is H wera Adyou Fks mpaxTixyn Or (1140 6 5) &f&is aAnOns pera dvyou WPAKTiKY. anna’ aodl -_ —_-_, * Aa ,_ ler hUlUlCO fs @pev THV Aristotle d for the e various hich 70cKn bems to be n in detail istrate and 1 of Aoyos follow the ties of its . Professor a convert. y puorkny, 1 Kal Trepl omev EVEKA Kal €v TOUS taTpos THY oo G * ALTLAS OV roinats not ).E. II. iv. ntellectual in VI. iv.? s Aristotle, n is at the art. With health, the way each xplanation ne case, a ‘ivity ; and is present S$ TOU Epyou commits him- : of the three the word is better vanish Adyou adnBois) reTR Advyou Fkis ns peta Rdvyou ciaeeamet atl ON THE ARISTOTELIAN USE OF AOTO®: A REPLY II 0 avev THs Ans eotiv (640 a 31). The activity is thus the materializing (if the word will pass) of a form: for that which at the beginning of the activity presents itself as an end (ov &vexa) becomes finally the form of the matter operated upon. The same account, mutatis mutandis, holds good of action. Action originates (N.E. VI. 1139 a 31) in a mpoaipeous, or determination of will, which originates in a desire on the one hand and a conceived plan (Adyos o €vexd Tivos) on the other. IIpoaipecis is not itself the od évexa, but springs from the apprehension of a od évexa or end. This end is the explanation of the details of the action, and the achievement embodied in the completed action. Foreknowledge of the achievement is implied in the agent. It is this, indeed, that constitutes his intellectual virtue: for dp00s Adyos wept Tov Tovoutwy dpovnais éativ (1144 6 27) In action, as in art, the activity is the materializing of a form: the ‘end’ of deliberation becomes the form of the matter which the action is designed to affect. There is, it is true, a difficulty here and a failure of correspondence between réyvn and dpovnots : for in action the @povimos is in a sense both agent and patient, he is house as well as architect. The difficulty is stated, but not solved, by Aristotle when he gives as the matter of action 7a@n and wpaées, and when he bases the distinction of Totnows and mpa£€s on the fact that the former looks to a result beyond itself, the latter not. The problem raised by such statements is a profound one, not hitherto solved to general satisfaction by any philosopher; but it should not be allowed to obscure the fact that for Aristotle all distinctively human activity is one in principle. It is always the imposition of a certain preconceived form upon its appropriate matter; and his name for this form is Adyos. When the plans are drawn, the architect’s next business is to calculate and price the quantities. With a well-drawn plan for guide this process, though it may be long and laborious, is not difficult. It needs only care and accuracy, not insight. In action what corresponds is the process of BovAevars, and the name given to the quantitative rightness achieved is peoorns. Clearly the calculation must start from a plan and must be determined throughout by it: the peodrns is, as Aristotle says, @peopévn AOyw. If you ask ‘ What plan ?’ Aristotle can only answer that rightness of plan is the intellectual side of good- ness, the investigation of which is to come later in the treatise: to be Ppdvipos is to be consistently right in that respect.2, But without a right plan, whether ' This passage is cited by Professor Cook Wilson in connexion with N.E. II. 1103 & 31, which seems to him to imply that ‘the dp@ds Adyos meant in itself an dper7j.’ He concludes, ‘it cannot therefore be rule or definition, and must be Reason.’ I cannot admit the conclu- sion, and I do not think the translation ‘ Reason’ Suits the passages any better than it suits the parallel passage above quoted from De Part. An. Art and ¢pévnors are both thought of primarily as forms of knowledge, and what the agent has in mind is in each case a Aéyos—what Hoffding and other modern writers call ‘the idea of an end.’ Further, ‘Reason’ has never been in English the name of a virtue. Therefore, even if the 6900s Aéyos is an dpery, the translation ‘ Reason’ is not justified. 2? With regard to the difference of reading ws and @ in 1107 a 1, the latter (preferred by By- water) is impossible for Professor Cook Wilson and other champions of Vernunft: both are possible for me. I cannot understand, however, much hesitation as to which is preferable. With the reading @ the last clause is a direct answer to the very relevant question, ‘What plan?’ So Aspasius in his Commentary (ed. Heylbut, 48, 20): kal éweényeiras rol Adyw * TW TOD Ppoviuou kal @ av 6 ppdviuos dploeser. 12 ON THE ARISTOTELIAN USE OF AOTO®: A REPLY of the agent’s own construction or accepted from another, there can be no goodness at all. For the A\éyos, which is from one point of view a plan, or end, or ideal, or intention, is from another the form or principle, the soul of the action ; and every good act is a manifestation both of 7@c«y and of dsvavontixy apeTn. In conclusion I turn to Professor Cook Wilson’s difficulty with the phrase @S 0 Noryos TarTec and other similar phrases. It may be thought that to interpret Néyos as a form or plan is to make this difficulty more acute. My answer, summarily stated, would be—(1) that the Xoyos expounded by the philosopher is both a plan, realizable as a form, of life and an injunction: so much is implied in the theory that ethical speculation is meaningless unless it has a practical use (see esp. 1103 b 27): (2) that the Aoyos in the gPporvipos is similarly both something apprehended by him in virtue of his ¢povnots and a command which he accepts: (3) that the law of the state (which is a Xoyos) commands, exhorts, and encourages; and we do not therefore conclude that the law is either a person or a human faculty. We know that what is implied is that laws assume the imperative mood. It is the same in all three cases. In the passage already quoted from the De Part. An. the artistic Xoyos is said not only to explain the artist’s procedure but also to issue reasoned injunctions (dcoTe 7rotntéov oTws). So the agent’s plan comes to him as a command; and it is not very strange to regard the necessity with which it controls deliberation as speech in the imperative mood. J. L. Stocks. ST. JOHN’s COLLEGE, OXFORD. LY can be no lan, or end, soul of the Scavontiry the phrase to interpret My answer, ilosopher is 1 is implied a practical uilarly both nand which commands, the law is lied is that es. In the is said not injunctions mand; and leliberation STOCKS. AOHNAION IIOAITEIA, XXX. 3-4. A SIMPLE transposition in the text would, I venture to suggest, remove one or two of the many difficulties of this obscure chapter. We are told at the end of section 2 that the Hellenotamiae who are administering the funds, although elected é« trav aei BovXevovtwr, are not to sit in the Boule: tovds dé EAAnvoTapias of av diayeipifwor TA YpHwaTa pH cUpEBov- Aeverv. This statement is in itself clear and intelligible. I need not here go into the questions which have been raised as to its relation to the other clause about the Hellenotamiae farther back in the same section, and as to the language and interpretation of that clause. Section 3 deals with the constitution of the four councils: Bovras &é Toijoat TéeTTapas K.T.X. Each of these, in an order to be determined by lot, is to hold office in turn, and for one year: xai Ttovtwy To NaYov pépos Bovureveww 1 The intervening clauses, which provide for the distribution of the 100 commissioners and ‘the others’? to the four Anes, do not concern us. At the beginning of section 4 there is an obvious lacuna. The MS. text runs: els €viavtov Bovrevery | S€ 7 dv Soxh avtots apiota EEewv. Evidently a word has dropped out after SovAeverr. Mr. H. Richards (v. Class. Rev. 1891, p. 228) has convincingly supplied the missing word, BovAeveoOar (cf. BovrAevoacGa: just below, and XXIX. 3, XXXI. 2), which has been accepted by Sandys and by Kenyon in his Berlin edition. The omission is satisfactorily accounted for by the similarity of the words BovAevery BovrAeverOat, aided perhaps by the position of BovAevery at the end of a line in the papyrus. The first half, then, of section 4 (BovreveoPar d€ down to Tis avis HALKLaS) directs that the Council of the year are to consult as they think best for the safe custody and proper expenditure of the funds and for the discharge of their business in general to the best of their power; if they wish to consult with more persons, they are to call in each an assessor, whomsoever he will of those of the same age. \ > > | \ 4 .. KAL Els EviavTOV BovXeveLv. Two serious objections here present themselves. (1) The custody and administration of the funds have been already committed to responsible special boards, and can hardly be described as the main and peculiar province of the 1 I put a colon after the former fovXevew and to be ro Aaxov uépos. K-W insert rovs X\axdvras. after diaxAnpdoa, and take the subject of the 2 Probably, I think, the qualified citizens latter Bovdevew (if it needs any definite subject) serving in the fleet abroad. 14 J. A. R. MUNRO Council. We may contrast section 5, and in a broader comparison the later chapters which deal with the functions of the normal Boule. The Athenian Council’s control over finance was doubtless important, but was always rather a general supervision and direction than the particular duties here so specially assigned to it. Even the financial necessities of the moment do not justify this singular prominence. On the contrary, the purpose of the exclusion of the acting Hellenotamiae from the Council, as of the institution of the mpoBovrox two or three years before, was surely to rescue the finances from the less com- petent hands of the larger body and emancipate the expert managers from its dangerous influence. (2) The Council, if we may trust our author and our text, was to be a very large and clumsy assembly of about 1,000 members. Critics have fastened especially upon that weak point in itsconstitution. It is difficult to believe that this over-manned Boule should have needed, or been authorised, to double its number by calling in assessors ! But if we transpose section 3 and the first half of section 4, we at once get a natural and apposite sense. The text will then run: tovs 6€ éAAnvoTapias of av Sdiaxeipif~wor TA KXpHpaTta pr) TupBovreveww: BovrevecOar Sé 7 av SoxH avtois apiota e€eww Tepl TE TOV KpHUaTwV, OTws av Goa H Kal eis TO Sdéov avaniokntat, Kal Tepi TOV adXoV @s av SvYwYTaL apioTta* Kav TL OéXwow Bovrctvoac0ar peta TrELOVOY eTTELTKAXELY ExacTOV ETEioKANTOY Ov av EOéEAN THY eK THS aUTHS HAtKias. Povdras b€ ToLhoaL TéTTAPAS K.T.D. The duties here assigned to the Hellenotamiae are proper to them. The Tept T@Y aAdXwv need not cause any misgivings, for (1) the Hellenotamiae had always something more to do than guard and pay out money, e.g. the collection of the tribute; (2) their amalgamation with the other treasurers of State funds, especially the Colacretae, enlarged their functions, as inscriptions of the years A410 to 404 testify; (3) it is a practice frequent in all legal documents to add an ‘et cetera’ to cover possible omissions. Compare, in the first section of the next chapter, wepl T@v voywv Kal TOV EevovYaY Kal TOV ANNwY TPaTTEW 7 av HY@VTAL TULPEpeELy. The évreioxAnto to assist the Hellenotamiae would, of course, be the mdpedpot of the inscriptions. Nothing is known of the mode of their appoint- ment. Probably co-optation was the established practice, as with the trapedpor of the three chief Archons (LVI. 1). The connection with the context at the end of the transposed passages is as smooth as at the beginning: xai eis eviavtov BovrAevery. Tas 0 Edpas Toceiv ths BovAjs cata twevOnpepov K.T.rA. From the constitution of the Council and its tenure we pass to its meetings, procedure, and business. The dislocation is easily explained by the homoeoteleuton and homoeo- archon of the sentences, cvuuBovrevery and Bovrevery, BovreverOar and Bovras Sé. The two passages are of approximately equal length. It occurred to me that their metathesis might date from an original faulty copy of an inscription, and I have tried the experiment of putting them into epigraphic script of the period arranged orovyndov in such a way that those two pairs of words stand at Bearer the later Athenian ys rather specially ustify this on of the Tt poBovror less com- ; from its | our text, Critics is difficult ithorised, - once get nvoTapias ’ xv ~ ) av boxy \ , > TO d€éov — OéX\oow An TOV EK 2m. The miae had collection ate funds, the years to add an on of the TEW 7 QD e, be the appoint- | TapEedpot issages 1S aS ~TOLELY uncil and homoeo- d BovrXas ed to me scription, pt of the ;stand at A@HNAION TIOAITEIA, XXX, 3-4 15 the ends and the beginnings of lines respectively. But I cannot without recourse to Procrustean methods get them both quite to fit the spaces. We may, however, well believe that at some stage in the literary transmission of our text the cardinal words did so fall, and that an easy slip of the next copyist’s eye transposed them. It is worth noting that the first scribe of the London papyrus seems to have found some temptation to displace the words Bovras 6€ troujoa: térrapas, for he inserted them again in section 5 after évavtiov THs BovAms—(once more a word in fova-!)—but corrected himself ‘inter scribendum ’ (K.). J. A. R. Munro. NOTE ON THUCYDIDES II. c. 48, § 3. eyo dé otdv re eyiyvero AEEw, Kai ad’ Gv av tis oKoTaV, El TOoTE Kal adOis exuréecor, paruor’ av €xor Te Tpoedws pr) Ayvoetv, Tavta SnAWow ards Te VoOHTas Kai avTds idwv aAAOVS TAT KX OVTAS. May a reader of Mr. Herbert Richards’ Thucydidea briefly defend the manuscript reading in this familiar passage? In the October issue of the Classical Quarterly (vii. 245) Mr. Richards suggests xat roAXAovs idwv, in place of kat atrds idov. He says, ‘ The second avros is very pointless, and a zoAAov’s seems wanted. We do not need to be told that he saw them himself, and hardly that he did see them ; that he saw many is worth mentioning. Cf. Herod. I. 197 doa airds roujoas eepuye dpoinv vovoov 1) aAAov ede expvyovra.’ Would it not be truer to say that the words as they stand are full of point and of Thucydidean point ? Thucydides seldom refers to himself, but here he is frankly egotistic when his- torical truth demands it: éyw ... a’ros ... avtdos. He stresses, by repetition, his personal experience and his personal observation: he had himself fallen ill and him- self seen the sufferings of others. His delight in the first-hand evidence he can offer as a surviving victim and as an eye-witness is no less keen here than in I. c. 22, § 1 (Gv avrds 7jKovca, With context), and in I. c. 22, §§ 2, 3 (ois Te av7ds Tapyv, with con- text). He brushes aside such speculations as the plague must have aroused long after it was over among thinking men, whether doctors or not, who knew of it only by hearsay. He has experiences and observations of his own to record for posterity ; he is no mere reporter of what others told him they had felt or told him they had seen: he can show that he is imbued, in some degree, with the scientific spirit of his contemporary Hippocrates. The chapter of Herodotus (I. 197) from which Mr. Richards quotes a few words seems to confirm the traditional text of Thucydides, who may well have had these travellers’ tales in mind. In particular, a remembrance of the preceding clause ei tus Kal AUTOS ToLovTOV erable bkoiov dv Exyn 6 Kdpvwv 7) GAAov ede TaGOvTa May have led him deliberately to write aitos te voonoas kai avTos tdwy addous wéoxovras | Mr. Richards gives voootvtas against the manuscripts, probably by a slip of the pen]: ‘I, the selfsame person, was not only myself stricken but myself saw others suffering.’ ‘Others,’ be it noted, not ‘another,’ as Herodotus has it ; and this modest plural (aAAovs) is surely expressive enough without any addition whatsoever. From the critical notes in Hude’s text of Thucydides I see that Cobet would strike out the second avrés. For my own part (and I only wish I could carry with me so devoted a student of classical literature as Mr. Richards), I should be no less reluctant to omit the second airds here than to omit the second or the third or the fourth airés in the fragment of Aeschylus which Plato quotes at the end of the second book of the Republic. W. Ruys Roserts. October, 1913. SOc acc aga wit Th qui brir beli tha whc dro] EXCt pulc afte , S$ €miTEeroL, Sav adAovs nanuscript | Quarterly iswv. He We do see them ; Tas eLeuye e words as when his- etition, his and him- e can offer |. © 20,8 3 with con- oused long 7 of it only ‘ posterity : n they had pirit of his few words have had preceding afovra may dwv adAAovs y by a slip myself saw t; and this hatsoever. obet would carry with be no less third or the the second ROBERTS. Ce egg eR NOTES ON DARES AND DICTYS. DARES. C.i., p. 2,12 (Meister’s paging) dicit Peliae regi se eo uelle ire si uires sociique non deessent. Pelias ...Argum... iussit . . . nauim aedificaret. Considering the next sentence read perhaps naues for wives. C. il., p. 3, 25. Graeci aduentare nauibus. mittit ad portam. M reads nautbus utt. May this conceal et ita? cf. p. 11, 2. cf tense cf. opening lines of C. iii. C. ill., p. 5,12. Laomedonti regi nuntiatum est classem accessere et. quod ubi, on the analogy I suppose of C. iv., p. 5, 24, is added by Meister against the grammar of the sentence. It must be deleted; cf. C. ii., p. 3, 20 with C. ili., p. 5, 13. C. iv., p. 5. 24. Priamonuntiatum est . . . praedam euectam. euectam MG, deuectam Meister. Surely auectam is the word required. The MSS. reading arose from praedduectam. C. vill., p. 10, 7. Antenorem a Grais contumeliose tractatum neque ab eis quicquam impetrasse. There are four other passages similar—C. v., p. 7, 24; C. vi., p. 8, 11; C. xvii., p. 22, 7; C. xxxvill., p. 46, 6. Meister has thought it necessary to bring all these into harmony by making Antenor the subject. For change I can scarcely believe it necessary to ensure such harmony, but if it is it seems worthy of note that the passage quoted is the only one in which it is not Priam but Antenor who is contumeliously treated. Instead of Meister’s changes then in C. v. and C. vi., I would here transpose Antenorem to after neque, and suppose se has dropped out after contumeliose as it has done in C. vi. in all the MSS. except F. C. viii., p. 10, 18. dicere coepit Troianis extremum exitium futurum sed pulchrum in otio uitam degere. There is no antithesis but a preference: read et. C. xi., p. 14, 5. postea dictum est eos immortales factos, itaque Lesbios . . « Quaesitum isse. The Lesbians did not go .o seek Castor and Pollux because they were afterwards thought immortal: read itemque. C. xviil., p. 23, 3. de Aethiopia Perses et Memnon. 18 R. T. CLARK et must be deleted, There was no Perses at Troy, while in C. xl. Memnon is Persarum rector ; cf. also Strab. p. 728, Diod. 11. 22, iv. 75. C. xx., p. 25, 26. Hector Boetem Arcesilaum Prothoenorem occidit. There is no Boetes at Troy. Read Boeotos: there is a somewhat similar corruption in J/. Lat. 169; cf. also Dictys 111. Io. C. xxi., p. 26, 19. Hector . . . Schedium fcleo liel phenorem occidit. Corrected to Schedium Elephenorem. Perhaps cleo conceals Clonium. Cf. Il. Lat. 168; Dictys 1. 13, 17. | C. xxl, p. 26, 23. Agamemnon... pugnam reuocauit. laeti Troiani in castra reuertuntur. Transpose 1 castva to before the full stop. The Trojans always retreat in oppidum, in urbem, never 1 castra. C. xxvi., p. 32, I. succedit Pheres Admeti filius. The only son of Admetus at Troy is Eumelus. Pheres is probably a marginal note Pheris, and has ousted Eumelus from the text, or the original may have read Eumelus Pheris A. fiius. The name Eumelus must in any case, I think, be brought into the text. C. xxvi., p. 32, 18. interea (i.e., during the truce) naues onerandas curat Palamedes. ornandas is surely necessary. Why should he Joad the ships and with what ? C. xxvii., p. 33, 15. Agamemnon cum magno comitatu ad castra reuer- titur. There is no question of an imposing retinue. He has been deliberately insulted by being sent on the servile mission of helping other people to get provisions. Read commeatu and cf. Caes. B.G. 1. 34. C. xxvili., p. 34, Ig. cul exultanti et glorianti Alexander Paris collum transfigit. In both Dictys and Dares Paris is never called anything but Alexander : delete Parts here as a gloss. C. xxxvi., p. 44, 3. Menelaus ad Scyrum uenit arma Achillis Neoptolemo filio eius tradit quae cum sumpsisset, Argiuorum castris . . . lamentatus est. 1. Lycomedem has probably dropped out before Scyrum, cf. C. xxxv. 2. Something has gone before Argiuorum. Meister supplies wenit et in. It is surely simpler to delete Neoptolemo and read sumpsisset. Neoptolemus in. Neoptolemus got into the line above by attraction to filio and became dative. Meister’s wentt would still have Menelaus for subject. C. XXXvVil., p. 44, 21. ut consilium cogat. Meister changes to conuocet, but cf. C. xxxi., p. 37, 14. C. xli., p. 49, 9. ut hi suos omnibus . . . adducant. Meister’s ut is a needless refinement. Dares uses /z constantly; cf. p. 8, 8. C. xli., p. 49, 20. Hecuba dum fugit cum Polyxena Aeneam incurrit. Polyxenam tradit ei. SU sh eff 1. Memnon ‘cidit. somewhat occidit. nium. Cf. eti Troiani s retreat in probably a he original n any case, indas curat ps and with astra reuer- deliberately ople to get ‘aris collum Alexander : Neoptolemo ntatus est. XXXV. ; menit et in. optolemus in. came dative. constantly ; am incurrit. NOTES ON DARES AND DICTYS 19g SoG. L reads tradidisset whence Meister’s tradit se. It is surely easier to read Polyxenam with G. and understand Hecuba as subject of tradit. Polyxena has no mind of her own in any case though she has in Dictys. Aeneas occurrit L, but G.’s is the more vivid phrase. If occurrit is retained read A eneae. C. xlii., p. 51, 8. Agamemnonem poscit conqueritur exercitum accusat, Antenorem accersiri iubet. Neoptolemus is the subject of poscit, but it is not his place to give orders and later it is Agamemnon who manages the whole business. Can Agamemnon have dropped out before A ntenorem ? DICTYS. I., c. 1, p. 4, 9. Atrel gratia etsi ea quae exhibebantur cum laetitia accipiebant tamen . . . templi pulcritudine afficiebantur. For Aire: gratia Obrecht followed by Meister reads veges Graeciae,a phrase that is entirely unlike Dictys and has no point here at all. Could we read at rerum gratia? I., c. vill., p. 7, 19. reguli.. moribus malum minitantur. moribus seems to have incurred the censure of all the critics, though Meister retains it; perhaps we might read moti bus=‘ gestures ’, used of the gesticulation of an orator in Cic. Brut, 116. - concilium irrumpunt atque inconditis I., Cc. XXIll., p. 17, 31. Cum opportunum iam tempus nauigandi ingrueret. ingrueret can scarcely be right. I should prefer urg ueret, cf. V., c. xvi., p. 100, 22 urgente nawigit tempore. II., c. xxvill., p. 36, 15. Chryses praeferiens dei uultus. uultus can hardly be right. Even if intelligible the phrase is too bold for Dictys, and Chryses had none of the majesty of Apollo. I think we must read uittas, cf. Jl. Lat. 13, and Verg. Aen. vil. 237. II., c. xlil., p. 45, 17. interea Aiax adueniens cognito apud naues Hectore ... apparuit. Meister apparently takes Hectore with cognito. But cognito, etc., is an absolute phrase=cognito quod apud naues ageretur. The clause appears often— e.g. p. 49, 18, but the sense is clear without it. I would read Hectori dat. after apparutt. III., c. iii., p. 53, 33- scilicet quia fluxa fides et animi earum spectati. ? nec for et; but perhaps the MSS. reading originally was eariispectatt and su- dropped out before spectati (cf. p. 55, 12, tendiet for tendi uidet), and we should read su spectati. III., c. xxiii, p. 59, 26. instantes tela iaciunt crebra magis quam cum effectu. crebre is surely required to balance cum effectu, cf. in u. 25 temere et cum audacta. NOTES ON DARES AND DICTYS III., c. xxili., p. 66, 13. non enim Helenae... gratia exercitum... procul ab domo. . . militiam tolerare. surely ab domo. V., c. vii., p. g2, 18. auspicio tali monitis omnium animis Calchas.. . hortatur, ‘and explains.’ They were neither warned nor instructed till Calchas told them, but they were astounded like the enemy in u. 8. Read attonitis. V., C. XVi., p. 100, 10. ceterum post abscessum Ulixi Hecuba . .. multa ingerere maledicta imprecarique infesta omina in exercitum. Surely Ulixis. There is no sense in Hecuba heaping insults on a man who is not there, and it will not fulfil her plan of enraging the army since Ulixes has fled its hate. VI., c. ii, p. 103, 15. Menestheus cum Aethra Thesei et Clymena filia eius . . . recipitur. Meister’s index reads Aethra (?) filia Theset. But matey must have dropped out before e¢ if only to balance filia eius. There is no tradition I know of making Aethra Theseus’ daughter. This has surely been pointed out before, but I risk doing so here. R. T. CLark. MONICH. iichas... n, but they . multa on a man army since ymena filia ve dropped I know of out before, . CLARK. wert ON SOME PASSAGES OF OVID’S 7RIS7TTA. SINCE the publication of my critical edition in 1889 the Tristia of Ovid has received some attention. A paper in Hermathena, vol. vii. (1890) by Pro- fessor R. Ellis contains several conjectural emendations, and in a public lecture on The Second Book of Ovid’s ‘ Tristia’ (Clarendon Press, 1913), this veteran scholar analysed the intricate contents of Book II. Two learned pamphlets by Dr. R. Ehwald, Ad historiam carminum Ouidianorum recensto- nemque symbolae (I. Gotha, 1889; II. Gotha, 1892) deal with the history of the text, and the textual criticism and interpretation generally. Dr. Paul Vogel’s Kritische und exegetische Bemerkungen zu Ovid ‘Tristien’ (Festschrift, Schneeberg | Sachsen | 1891) is a valuable contribution to the understanding of the text and its interpretation. I have made much use of these works in writing this paper. Though little has emerged in the way of new manuscript material, the impor- tant discovery has been made by Dr. Ehwald of the fragments of a tenth century manuscript in the town library of Trier. These fragments are our earliest authority, since the best manuscript L, the Marcianus, belongs to the eleventh century. These leaves which have been recovered formed the bind- ing of a book. They contain, unfortunately, only I. xi. 1-31; I. xi. 33-II. 21; IV. iv. 35-65; IV. iv. 67-v. 9; their readings closely agree with our best manu- script L. The fragments are fully described by Dr. Ehwald, Symbolae, I1., PP- + Sqq- I. 1. 47, 48. da mihi Maeoniden et tot circumspice casus: ingenium tantis excidet omne malis. Here the word circumspice, if taken in its literal sense, is obscure. ‘Give mea Homer,’ says Ovid, ‘and look around upon my many misfortunes ; his talent will be entirely destroyed by such great troubles.’ One would naturally expect Ovid to say ‘ Give me a Homer and surround him with as many misfortunes as are mine,’ and that is what Heinsius made him say by his conjecture czrcwmuce, which is accepted by some editors, and apparently by Professor Housman, Manilius I. p. lii. But civcumice is unsatisfactory, since the verb, as far as I know, is nowhere used in Roman poetry. No alteration is necessary, for the meaning which ciycumice is intended to convey is actually the meaning of ciycumspice, which signifies ‘ provide for him.’ Czrcumspicere means ‘to look around for, look for,’ and so ‘ provide,’ as in the following passages: Verg. Georg. iii. 390, ‘plenoque alium circumspice campo’; Livy i. 30 ‘Sabini 22 S. G. OWEN circumspicere et ipsi externa auxilia’; Juvenal viii. 95 ‘ praeconem Chaerippe, tuis circumspice pannis’; Seneca Dial. i. 3. 3 ‘alius circumspiciatur cum quo conferre possimus manum.’ II. 77-80. a! ferus et nobis crudelior omnibus hostis, delicias legit qui tibi cumque meas, carmina ne nostris quae te uenerantia libris iudicio possint candidiore legi. I wish to modify the explanation given of this passage in my critical edition, p. xcv. It is difficult to decide between indicto LHPV and tudicio G. I now think that zudicio is right; the meaning is :—‘ Alas! he was a heartless enemy and more cruel than all to me whoever he be that read to you my trifles with the intent that no portions of my books extolling you may be read with more friendly judgement on your part. In my edition I accepted imdicio which has greater MS. evidence. But the meaning obtained is unsatisfactory. Vogel has pointed out that a special indicium would not be necessary for an unprejudiced reading of Ovid’s poems. Besides this, I object (i.) that one would expect indice not indicto, which cannot be equivalent to indice ; (ii.) though tndicio candidtore might be construed ‘ in con- sequence of more favourable information about my poems,’ taking indicio in a colourless sense, which is slightly supported by Pont. I. v. 76 where ‘ indicium studii mei’ means information about my poetry, and by Tvist. II. 379 where ‘indicio Homeri’ means ‘through information given by Homer,’ the word indicitum almost invariably has a bad connotation, standing for information or revelation of an unfavourable character, as Trist. III. 4, 71 ‘ nec meus indicio latitantes uersus amicos | protrahit.’ Thus to talk of a candidius indicium seems strange if not impossible. Vogel proposes to read ‘ne e nostris,’ not only for the sake of greater clearness, but because two of my MSS., a Bodleian and a Copenhagen MS., have an erasure of one letter after me. But as I stated in my apparatus the erased letter was c; confusions between ze and zec are frequent. As to the construction, I regard nostris libris as a poetical local ablative, like II. 437 ‘quorum libris modo dissimulata Perilla est’ [Perilla est is my correction for per illos MSS.], which means ‘those in whose books Perilla was lately a counterfeited name.’ So Ibis 639; ‘haec tibi tantisper subito sint missa libello.’ II. 83-86. cum coepit quassata domus subsidere, partes in proclinatas omne recumbit onus, cunctaque fortuna rimam faciente dehiscunt ipsa suoque eadem pondere tracta ruunt. This, I now think, is the true reading of these disputed lines. In line 86 I have ventured to conjecture suwogue eadem for suo guodam of the MSS. The SEER gE crete: Chaerippe, ir cum quo my critical ,HPV and is :—* Alas ! he be that ‘s extolling . In my 1e meaning al indicium s. Besides cannot be ed ‘in con- indicio in a 2 §indicium 379 where , the word ormation or leus indicio us imdicium . of greater hagen MS., paratus the As to the ike II. 437 rrection for as lately a sint missa n line 86 I MSS. The ON SOME PASSAGES OF OVID’S TRISTIA 23 meaning is: ‘When a tottering house begins to tumble, its whole weight subsides towards the leaning portions; and when chance causes a rift, all its structures crack; and moreover, dragged down by’ their own weight, collapse.’ IT. i. 89-92. at, memini, uitamque meam moresque probabas illo, quem dederas, praetereuntis equo: quod si non prodest et honesti gloria nulla redditur, at nullum crimen adeptus eram. The last couplet, pronounced ‘semibarbarus’ by Heinsius, and rejected by many editors, appears corruptly in L. Non quasi prodesset honesti gloria nulla Redditus est (redditur et m. 3) nullum crimen adeptus erit (eram m. 3). Ehwald Symb. II. 19 thinks it clear from the form in which it is given in L that the couplet is an interpolation of someone expressing his own view of the muster held by Augustus. But, as Vogel points out, the poet’s purpose is to bring into prominence the recognitio equitum, as being a formal decision given by Augustus which was creditable to Ovid himself. The expression crimen adipisct is unusual, but though adipisci generally denotes to secure something good, it may be used of securing something bad, as in Tac. Ann. I. 74 ‘ odium apud omnes adeptus.’ I am inclined to think that my conjecture ‘ et honos et gloria’ may be right ; but /onestus in the sense of ‘ having held honores (offices) ’ is Ciceronian. See pro Mur. § 87. The couplet would appear to be sound and to mean: ‘And if the fact that I passed muster before you brings me no advantage, and no repute for distinction is accorded to me, yet it must be admitted that I had incurred no charge.’ II. il. LOg-114. illa nostra die, qua me malus abstulit error, parua quidem periit, sed sine labe domus: sic quoque parua tamen, patrio dicatur ut aeuo clara nec ullius nobilitate minor, et neque diuitiis nec paupertate notanda, unde sit in neutrum conspiciendus eques. sit quoque nostra domus uel censu parua uel ortu, ingenio certe non latet illa meo. The order of these lines given in the MSS. I now think correct. In line 113 et neque, GPV and most MSS., must be accepted rather than wt neque LH, and in line 115 sit the reading of most MSS. is superior to st LV, with which the omission of the verb is awkward. The meaning is: ‘On that day on which a fatal mistake led me astray, a family was ruined which was stain- 24 S. G. OWEN less even if insignificant, though it is insignificant also with the reserve that it is described as distinguished in the time of our fathers, and as not inferior to the lineage of any, and as not remarkable either for riches or poverty, so that (unde = ut inde ortus eques) a knight sprung from it is noteworthy in neither of these respects. Also, even if my family is insignificant in property or extraction, my talent at least saves it from obscurity,’ that is, my talent cannot be blamed for its remaining unknown. II. 191-192. Ciziges et cholchi metereaque turba ietaeque Danubii mediis uix prohibetur aquis. So L. The difficulties of this couplet are no reason why it should be rejected. It is obviously designed to emphasize the remoteness of Tomi, which is too vaguely indicated by the expression egressus Histri (189); and it is designed to move the reader’s horror by the thought that the broad Danube scarcely protects the frontier post, the unhappy poet’s residence, from the incursions of wild migratory tribes who lived in the savage regions beyond it extending even beyond the Maeotis. As to the first of those tribes Iazyges, the reading of most MSS. will not do, both on account of the impossible synizesis involved (the word is always scanned elsewhere Lazygés; Ibis 133; Pont. 1.2.79 ; IV. 7. 9), and because this Sarmatian tribe lived on the lower Danube, too far west to be associated with the tribes mentioned close to it. The tribe is the Ciziges, whose name is correctly preserved by L. This Sarmatian tribe lived just beyond the Tanais on the Asiatic side of the Maeotis. It is hard to avoid supposing that they are identical with the Cizici of Pliny (N.H. VI. 19) ; May- hoff gives this apparatus—Czzicr DED ziztci a czict R'—with other variants. Mr. J. G. C. Anderson, whom I have consulted, remarks that the variation of terminations signifies little in a barbarousname. In that case the second tribe would seem to be the Colchi, who are a little further to the south. As to the enigmatical Meterea which contains the name of the third tribe, Ellis’s con- jecture Teretea (Hermathena VII. 193) comes near enough to the MSS. and fits in well with the other tribes. The Teretae or Toretae were a tribe who lived a little east of the Maeotis, north of the Colchi. The fourth tribe are the Getae, as given by most MSS. With this tribe Ovid moves nearer home, for they lived north of the Danube. Pyrohibentur should of course be read with most MSS. II. 253-254. at matrona potest alienis artibus uti, quoque trahat, quamuis non doceatur, habet. So Land most MSS. The meaning is ‘ But it may be objected that the married woman can make use of artifices intended for others, and she finds in your poems the means by which she can attract, though the lessons are not addressed to her.’ I am glad that Vogel supports me in this view. Many editors read quodque with the Gothanus and some inferior MSS., and this [On eee oeibiinin amet serve that it t inferior to poverty, so y in neither property or alent cannot t should be Tomi, which ); and it is oad Danube e, from the s beyond it Iazyges, the »le synizesis ont. 1. 2. 79 ; ube, too far tribe is the 1 tribe lived urd to avoid /19); May- er variants. variation of second tribe As to the Ellis’s con- MSS. and . tribe who ribe are the r home, for » read with d that the he finds in ns are not ew. Many .» and this ER RET ae ON SOME PASSAGES OF OVID’S TRISTIA 25 reading is defended by Ehwald, Symb. II. p. 19. Tyrahat is then taken in the medical sense of ‘contract by infection’—‘she finds in your poems some corruption that she can contract.’ This is less simple than the explanation which I have given, and having indifferent MS. support has little to recommend it. Though the medical sense of trahere is paralleled by Pont. III. 2. 14 ‘uicinum metuens ne trahat inde malum,’ the sense ‘to attract’ is equally well supported by A. A. III. 421 ‘se quoque det populo mulier speciosa uidendam: | quem trahat, e multis forsitan unus erit.’ Il. ii. 277-278. ~ ‘at quasdam uitio.’ quicumque hoc concipit, errat, et nimium scriptis arrogat ille meis. So L, which here shows its excellence. Other MSS. have ‘quasdam uitiat, quaedam uitia,’ and so forth. The meaning is ‘ But, say my critics, some women have turned your Ars to a corrupt use’—‘ quasdam uitio uertisse meos libros arguunt obtrectatores.. Ehwald makes it probable that the reference in quasdam is to both the elder and younger Julia (Sym. II. 20). Conjectures such as that of Vogel ‘ atque metum uitil’ are superfluous. II. il. 279-282. ut tamen hoc fatear, ludi quoque semina praebent nequitiae: tolli tota theatra iube: peccandi causam multis quam saepe dederunt, Martia cum durum sternit harena solum. The meaning is ‘ But even if I admit the charge just made, the pageants also provide incentives to immorality, so order wholesale suppression of the theatres: how often have the pageants afforded to many an occasion of sinning, when the sand is strown on the hard ground in the amphitheatre of Mars.’ The Judi were a combination of theatrical, gladiatorial, and other performances. It was at the Judi that comedies, and especially, under the Empire, the indelicate mimes were presented. The /udi were a time of public merry- making, and led to looseness especially through two means which Ovid mentions: (1) the theatres ; (2) the gladiatorial and other shows, which brought the sexes together as spectators. It is thus clear that /udz is the subject both of pracbent (279) and dederunt (281): there is therefore no reason for Ehwald’s conjecture mimi for multis, which involves the further alteration of cum to cur: nor is munera for mimi required, proposed by Vogel as it had been proposed by Heinsius before him. II. ii. 409-410. est et in obscenos commixta tragoedia risus, multaque praeteriti uerba pudoris habet. Here commixta is due to GHP. Commusta is found in V, and the third hand of L., the original reading of which is erased, but was probably commuixta. 26 S. G. OWEN The other MSS. have deflexa, accepted by most editors, in supposed allusion to the Satyric drama in which ‘tragedy was diverted to obscene jokes.’ But commixta, which I restored to the text, gives a better sense, and is approved by Vogel and Ehwald (Jahresbericht, 1902, p. 283), who compares Plaut. Amph. 59 ‘faciam ut commixta sit (haec) tragicomoedia.’ The meaning is: ‘ tragedy has been made composite in character (has not been kept pure) in order to provoke indecent laughter.’ The reference is probably to the (Aapotpaywdia: of Rhin- thon, which were burlesques of tragic themes. On this obscure writer some light has lately been thrown by Lejay in his edition of Horace’s Satires, p. xlii., and Michaut, Sur les Tréteaux latins, p. 128. III. 1. 47. causa superpositae scripto est testata coronae : seruatos Ciuis indicat huius ope. Thus the couplet is given in Lord Leicester’s MS. (H); the other MSS. omit est, which was added conjecturally by Ehwald after coronae. It has been proved from inscriptions that the legend OB.CIVIS.SERVATOS was placed on the ‘ querna corona’ itself (Ehwald, Jahresb. 1893, p. 100; B.P.W. 1890, p. 1624): the meaning is ‘ the reason for the crown fixed above the door is attested by the legend on it.’ Sak. 3 2f. si iam deficiam subpressaque lingua palato, So the MSS. The conjecture of Heinsius defictat is usually accepted. But no alteration is required; deficiat must be understood with lingua from deficiam. Iuu. xiil. 141 ‘quia tu (es) gallinae filius albae nos (sumus) uiles pulli.’ See below on III. 12. 1. III. 7. 27-28. forsitan exemplo, quia me laesere libelli, tu quoque sis poenae facta secuta meae. Ovid is addressing the young poetess Perilla, who he fears has abandoned poetry through terror at the disaster it has brought upon himself. Various conjectures have been made with a view to maintaining facta—retusa by Ellis (Hermathena vii. 211) and Zingerle (N. Phil. Rundschau, 1890, p. 70) indepen- dently, but this un-Ovidian word seems out of place ; quieta by Vogel, which is tame; soluta (slack) by myself, to which Vogel objects that so/utus means ‘set free from pressure or coercion.’ It does often, but Cic. Rwll. II. gt ‘urbem ipsam solutam ac debilitatam’ supports my meaning. Other conjectures in my edition. But I am now inclined to think that secuta is right and facta wrong, and to accept fata from the early printed editions, apparently an ‘in tr > Ff ta Fo a a ee ee ee.) | as d allusion to jokes.’ But approved by ut. Amph. 59 ‘tragedy has r to provoke tat of Rhin- writer some tives, p. xlii., other MSS. It has been ATOS was 00; B.P.W. ve the door y accepted. angua from 4mus) uiles abandoned f. Various isa by Ellis 0) indepen- 21, which is means ‘set git ‘urbem jectures in t and facta arently an oe es ON SOME PASSAGES OF OVID’S TRISTIA 27 Italian conjecture. The meaning is: ‘ perhaps in consequence of the lesson of my punishment, because my writings have been my ruin, you have sub- mitted yourself to destiny’; you have resolved to write no more since fatum seems to bring ruin on authors of poetry. Verg. Aen. I. 382 ‘data fata secutus.’ III. ro. 11-12. dum tpatet et boreas et nix habitare sub arcto, tum patet has gentes axe tremente prem. The patet of line 11 is clearly a corruption due to fatet in line 12. Various alterations have been proposed. I now suggest ‘dum prohibet Boreas,’ etc. For prohibere followed by infinitive see Pont. I. 5. 51 ‘hoc quoque me studium prohibent adsumere uires.’ Prohibet contracted may have easily become patet on account of the succeeding patet, and et following it is due to (metrical ?) dittography. The same idea is repeated in Pont. I. 3. 35; ‘quocumque aspicias campi cultore carentes, uastaque, quae nemo uindicat, arua iacent.’ The rather rare use of patet impersonally in line 12 is supported by Her. 16. 32 ‘juuenem facti paenituisse patet.’ III. 12. 1-2. frigora iam Zephyri minuunt, annoque peracto longior antiquis uisa Maeotis hiems. So the MSS. Editors regard Maeotis as corrupt, on account of the shortening of the diphthong ; thus Lachmann altered it to Tomztis, and Vogel, supposing it to conceal the finite verb, reads ‘ uim moderatur hiems.’ These are counsels of despair. I now think that Ovid wrote Macotis and scanned it Maéotis as a metrical experiment, imitating the shortening of diphthongs found sometimes in Greek poets. That by the law of Latin Phonetics such shorten- ing is permissible is shown by the prosody of Plautus, where we find Pellaeus as a dactyl in Asin. 333 (Lindsay, Captivi [1890], Introd., p. 19), by the scan- sions ‘ pracacutus,’ ‘praceunte,’ ‘pracustus’ (L. Miiller, De ve Metr. ed. 2, pp. 287 sqq.), and by the form ‘ prehendo,’ which we print although MSS. show ‘ praehendo.’ I believe that Ovid shortened ‘ Maéotis’ in imitation of Euripides Herc. 409 Maiétw audit, which words in the strophe correspond to éijA@e tév Te in the The choric ode in which this occurs, Herc. 348-450, appears to Besides line 400, Ma@rtuv, there antistrophe. be written in exact syllabic correspondence. are only two other uncertain places, line 374 (melas éSdpyafor in the anti- strophe corresponding to rots Odvodcw dyadya in the strophe; and line 350 Tav xadrid@rrov (MSS.) «:Odpav in the strophe corresponding to Eotpacév ToEois hovios in the antistrophe. As to (m7reias I believe that here again the diphthong is scanned short, as I have marked it, on the same principle, which removes the anomaly. As to the second case, Professor Gilbert Murray in his 28 S. G. OWEN edition follows Verrall in reading tov cadXe POirov, which violates the syllabic correspondence. The MSS. are corrupt; but if the emendation of Stephanus Tav Kadrripbiyyov be accepted, the syllabic correspondence is re-established. It seems hardly possible to resist the conclusion that Stephanus was right. There is a similar shortening of Nucaiwy in Soph. Antig. 1131. It is a fact beyond question that Greek poets did shorten diphthongs when standing before a vowel in certain cases, particularly where the words are adjectives. Thus we find éuzraiov (Od. XX. 379), xapaievvddes (Od. X. 243), otos (Il. XIII. 275), vids (Il. 1V. 473), and in tragedy maXaids (Soph. fr. 655 ; Eur. El. 497), yepaios (Eur. Hec. 64; Herc. 446, 901) wetpaia (Soph. Ant. 827), dethaios (Antig. 1310). In Soph. O.C. 200 and 238 yepaoy stands in Jebb’s text ; it is so found in the MSS. in 238, in 200 it is Dindorf’s correction for the yepavov of the MSS. This is interesting because it indicates that there was a tendency to slur over in pronunciation a diphthong followed by a vowel, a weakening of quantity due to conversational pronunciation. The same is indicated by the spellings tadads, madeds, for madaids, and the forms réXetos, Téde0s, aie, aet. Professor Gilbert Murray, who has given me his opinion on this point, says, ‘personally I don’t think there is any valid objection to Maia@ris from the Greek side. ... The syllable az was evidently in practical speech apt to be short: the spellings 7raXeos, etc., occur very early; again, -yaz -Tat in verbs is elided in Epic,' and in Menander, Efitrepontes, 138, 140, 226, and (me tudice) in the Iph. Aul. 407 BovXNop’ adrd. So xai is elided (without crasis) in womep ‘Apuddsos x’ ’Aptotoyeirwy and other cases.’ In Eur. Hipp. 1128 @ Spupos dpetos, 60 cvv@v, Professor Murray keeps the text scanned thus, with the approval, he tells me, of Wilamowitz, who formerly suggested dpéos on the analogy of rédAeos. This practically means that the spelling was indifferent, or rather that the forms téXeos 7raXeds, etc., were invented to suit the pronunciation, perhaps by the learned. I think this is the truth; and in the above passages of Sophocles Jebb would have done well to read yepacov in both lines. Such weakening of quantity is found with long vowels as well as diphthongs. Thus we find BéBrAyja (Il. XI. 380), Ilioa (Pind. Ol. II. 3, Bacchyl., V. 182), and Nypeis beside Nypnis. More difficult is the case of matpe@os found in Pind. Nem. IX. 14, and in Eur. Alc. 249, Tro. 162, Med. 431; in Euripides Professor Murray restores 7atpios. But it seems likely that here also the MSS. are correct, and present a further instance of weakening of quantity. At any rate #pwos (Od. VI. 303), code? Antimach. fy. 37 Kink., are undisputed. Cicero’s remarks on shortening in hiatus, which he regarded as archaic, are interesting in this connexion (Ovat. 152). The weakening of @ is illustrated by the shortened forms Britt6nes Vascones (Juv. XV. 124 and g3) beside Bpitrwves (Procop.) Ovacxwves (Strabo). In further illustration of this paradox that apparent false quantities are to be encountered in the best classical poets I take this opportunity of defending 1 And Aristophanes. See Kiihner-Blass, Gk. Gr., I, 237. made 5 ©} OO wees ee OS|lClCOMSN,C MNO Aa the syllabic f Stephanus -established. $ was right. thongs when e words are Jd. X. 243), ph. fr. 655; 1. Ant. 827), Is in Jebb’s ction for the there was a ry a vowel, a he same is rms TéXeELOS, opinion on ybjection to in practical LIN, -wat -TaL 40, 226, and shout crasis) Hipp. 1128 d thus, with rested dpéos pelling was nted to suit ith ; and in id yepacoy in s as well as . Ol. II. 3, the case of 2, Med. 431; ly that here eakening of 7 Kink., are regarded as ning of @ is [24 and g3) tities are to f defending Rep me wr tee ON SOME PASSAGES OF OVID’S TRISTIA 29 the following readings retained in my text of Juvenal, VI. 167 Venusinam, VII. 15 equitesque Bithyni, VII. 204 Thrasymachi, X1V. g ficedulas. Quantities, especially of proper names, were often unmanageable in verse. Greek authors complain of this: Thus Critias P.L.G., ed. 4, p. 282: ov yap Twas iv touvow’ éehappote édeyeiw ; Archestratus 39: txQvos av&nOévtos ov ev péTpw@ So Lucilius 228, Marx: ‘seruorum est festus dies hic, quem plane hexametro uersu non dicere possis.’ There are similar complaints in Horace (Sat. I. 5. 87), Ovid (Pont. IV. 12 on Titicanus), Martial (IX. rz on Marinus), Pliny (Epp. VIII. 4. 3 on Dacian proper names), Apollinaris Sidonius (Carm. 22. 485). Therefore Greek poets allowed themselves some latitude and tended to lengthen the first in a sequence of short syllables. Homer has [Ipiapos but Iptapidys, ’AdAXwv but ’AmdAXwvos (so Ap. Rhod. I. 403, Callim. Hymn Ap. 17), Callimachus (Hymn Del. 326) has ’AwoAXwv. The Romans followed them. Virgil has ‘ Priamus’ but ‘ Priamides,’ ‘ Siculus’ but ‘ Sice- lides,’ ‘Italia’ but ‘ Italus,’ ‘ Diana’ (Aen. I. 499, cf. Hor. Carm. IV. 7, 25, Ov. M. VIII. 353), but ‘ Diana’ (Aen. VII. 306, cf. Ov. M. VIII. 385, Grattius, 253). A violent instance of variation is Hor. Carm. III. 4.9 ‘me fabulosae Vulture in Apulo nutricis extra limen Apuliae,’ where no change should be made in the text. Ovid has ‘Io’ (Aen. I. 3. 21), but ‘lo’ (Her. 14. 103, Ibis 624). Virgil has ‘ Sithénius’ (Ec. 10. 66) with Lycophron 1357, though the 0 is long in Herodotus. Virgil has ‘ Fidenae’ (Aen. VI. 773), but Horace (Ep. I. 11. 6) and Propertius (IV. 1. 36) ‘ Fidenae.’ ‘ Siddnius’ is usual though the stem of ‘Sidon’ is long (Hor. Ep. I. 10. 26). Juvenal has ‘ Scantinius’ (II. 44), but Ausonius (Epigr. 92. 4, Peiper) ‘ Scantinius.’ ‘ Gradiuus’ is found forty-nine times against ‘Grddiuus’ four times (M. Warren, Am. Journ. Phil. IV. No. 13). Catullus (29. 3) has ‘ Mamurra,’ but Horace (Sat. I. 5. 37) and other poets ‘Mamurra.’ Propertius (I. 19. 7) and Ovid (A. A. II. 356) have ‘Phylacides,’ but Ovid (T. V. 14. 39) ‘Phylaceius.’ Propertius (III. 1. 29), Ovid (Her. 5. 94, M. XII. 547), and Persius (1. 4) have ‘ Polydamas.’ The Greek is IloAvdduas, but it is wrong with some editors to read in these passages ‘ Pulydamas’ (IlovAvédduas), since the long o probably came from a Doric or Aeolic form IlwAvéduas, as in the case of ‘pdlypus’ (Hor. Sat. ] 4 ? a ov Oéuis evtretv. I. 3. 40), where the dialectic form was introduced through the «ow recs. ‘Batauus’ and ‘ Batauus’ are both found (Mart. VI. 82. 6 Friedlander) ; Statius has ‘Neréis’ but ‘ Neréida’ (Achill. I. 24. 158); ‘ Palatinus’ is the adjective from ‘ Palatia’ (Mart. IX. 24. 1). (131) has ‘Daphne.’ ‘ Romuléus’ is usual, but we have ‘ Romulcus’ in Pervig. Ven. 74 (see my note and Clementi). Persius (3. 9; so Auson. Epigr. 5. 9) has ‘ridere,’ though the w is elsewhere short. Lucretius (IV. 641) has ‘cdturnix,’ Juvenal ‘ céturnix (xii. 97). ‘ Propino,’ has the initial o both short (Juv. V. 27, Mart. II. 15. I) in which case it is a Grecism, and long (Mart. [. 68. 3). Martial even has ‘ sc¢libra’ (I. 99. 15; V. 19. I1). enabled Manilius and Juvenal to shorten the penultimate of ‘ fortuitus,’ which is long in Horace (Carm. II. 15. 17): Manil. I. 182 ‘nam neque fortuitos ortus ‘Daphne’ is usual, but Petronius This same cause 30 S. G. OWEN surgentibus astris.’ Juv. xili. 225 ‘non quasi fortuitus nec uentorum rabie sed’: so probably Petron. 135 ‘fortuitoque luto.’ For in neither case can the word be scanned by synizesis, since this is never employed by Manilius or Juvenal (L. Miiller, de ve metr. ed. 2, p. 302). In Hor. Sat. II. 2. 76, Ep. I. I. 108 ‘pituita’ is probably not a case of synizesis, but should be scanned pitiiita ; just as Statius (S. I. 6. 16) has ‘ gratuitus,’ though ‘gratuitus’ is the proper quantity (Plaut. Czst. 740). These words had apparently two pro- nunciations: (1) -titus, the literary form; (2) uitus, the popular form. Further, in the Empire the diphthong ui became, at least in poetry, tii (e.g. citi, a pyrrhic). These being the facts I retain with confidence in the text of Juvenal ‘Venisinam’ in vi. 167, though Veniisia’ is usual; ‘equitesque Bithyni’ in vil. 15, though ‘Bithynus,’ is usual; ‘Thrasymachi’ in vii. 204, though ‘Thrasymachus’ is usual; ‘ficédulas’ in xiv. 9, though ‘ficédula’ is usual. Emendation is required in no case, neither Biicheler’s ‘ Venustinam,’ though this is confirmed by the Valenciennes MS. (Class. Q. VI. 31), nor Biicheler’s ‘faciantque equites Bithyni,’ accepted by Leo, nor Ritschl’s ‘ Tharsymachi ’ accepted by Leo, though he does not mention Ritschl, nor Lachmann’s ‘ficellas ’ (Lucret. III. 992). As to the construction in Ovid Trist. Ili. 12. 2, ‘ frigora minuit ’ must be supplied as complement to the subject ‘Maeotis hiems’ from ‘frigora minuunt’ line 1. The ellipsis is parallel to III. 3. 21 discussed above. IV. 10. 79-80. non aliter fleui, quam me fleturus ademptum ille fuit. matris proxima busta tuli. So most MSS., which reading I restored to the text in my edition. Most editors read ‘matri (V) proxima iusta (Cujacius) tuli’: wrongly, I think. Vogel, who approves my restoration, aptly quotes Stat. Siu. V. 1. 221 ‘nec enim fumantia busta clamoremque rogi potuit perferre. The meaning is ‘the next obsequies which I endured were my mother’s.’ This rare sense of busta ‘the burnt remains’ is found also in Stat. Theb. XII. 246 ‘ haud procul Ogygias . . « domos et egena sepulcri busta iacere reor.’ Lucan VIII. 849 ‘ transibis in urbem, Magne, tuam, summusque feret tua busta sacerdos.’ V. 1. 69-72. ‘at mala sunt.’ fateor. quis te mala sumere cogit ? aut quis deceptum ponere sumpta uetat ? ipse nec emendo, sed ut hic deducta legantur ; non sunt illa suo barbariora loco. So the passage should be read with the best MSS. Ovid is arguing with his critical reader, who found fault with the Trvistia on account of its doleful mn torum rabie ‘ase can the Manilius or 2. 76, Ep. I. be scanned litus’ is the ly two pro- m. Further, ( (e.g. cui, a of Juvenal Bithyni’ in 204, though a’ is usual. am,’ though - Biicheler’s larsymachi ’ Lachmann’s it’ must be ym ‘frigora ve. tion. Most ly, I think. I. 221 ‘nec ning is ‘the nse of busta cul Ogygias g ‘ transibis rguing with its doleful eee eee Tee ee Ip ON SOME PASSAGES OF OVID’S TRISTIA 31 character. It is right (replies Ovid) that one who is sad-hearted should express his sadness. ‘ But (the critical reader then objects) your Tvistia are poor stuff (mala).’ ‘I admit it (says Ovid). Who compels you to take them up if they are poor stuff? or who forbids you to lay them aside after taking them up, if you are disappointed? I do not even myself revise them. But you should read them with the thought before you that they were composed here in Tomi ; they are not more unpolished than the place of their origin.’ When ec is seen to be equivalent to me quidem as in Pont. 1. 1. 19, A.A. III. 108, /egantur construed as a jussive subjunctive, and wi taken as standing for welut, this disputed passage becomes intelligible. For this use of ut see Her. 9. 129 ‘ dat uultum populo sublimis ut Hercule uicto’ (where see Palmer) ; M. 1. 585 ‘ natamque miserrimus Io luget ut amissam’; Hor. Carm. III. 5. 42 ‘ut capitis minor.’ Editors usually accept the reading of the inferior MSS. ‘ipse nec hoc mando.’ This is evidently an interpolation to simplify the construction ot ut legantur, which words its author took to depend on mando. Those who keep this reading suppose it to mean ‘mando hoc i.e. ut sumas carmina siue deceptus sumpta ponas,’ which is intricate. V. 10. 41-42. utque fit, in me aliquid siquidem dicentibus illis abnuerim quotiens annuerimque putant. siquidem is found in GHPV: the rest of the MSS. have stquid, of which siguidem is a corruption. Lhere fails. Various alterations have been proposed, the best is that of Ellis (Hermathena, VII. 210) ‘in se aliquid fingi.’ Schenkl before him had proposed se, but I do not think me should be altered. The d of siquid is probably due to dittography, coming from the succeeding dicentibus. Thus siqgui was in the archetype. I think sense is best restored if we read fictt: utque fit, in me aliquid ficti, dicentibus illis abnuerim quotiens annuerimque, putant. Ovid is speaking of the inability of the Getae to understand Latin; he must converse with them by signs. ‘And, as often happens, whenever I nod dissent or agreement, they think that there is some deceit in me.’ Fictum means something made up, untrue. M. IX. 768 ‘sed iam consumpserat omnem materiam ficti’; XII. 57 ‘mensuraque ficti crescit.’ V. 12. II. des licet in ualido pectus mihi robore fultum. As fultus elsewhere, e.g. V. 13. 8, takes a simple ablative, Heinsius altered in to hic, i.e. ‘hic apud Getas’; and inthe Corpus Poetarum Tread en. But I am now convinced that im is right. The same variation on the simple ablative 32 ON SOME PASSAGES OF OVID’S TRISTIA construction is found with mztor in Lucretius V. gog ‘ nixus in hoc uno nouitatis nomine inani.’ Instances from Cicero are given by Kiihner, L. Gy. 1i. 293. V. 12. 49-50. nil mihi debebat cum uersibus amplius esse, cum fugerem merito naufragus omne fretum. For cum Heinsius introduced sed from a few inferior MSS. In the Corpus I printed guin, the conjecture of Dr. Postgate. I now think that cum should not be altered. The meaning is ‘I ought to have had no more dealings with poetry since, having been deservedly shipwrecked, I should have entirely avoided the sea.’ The language is metaphorical; the Avs Amatoria wrecked his fortunes and so he should embark no more upon the sea of poetry. S. G. OweEN. 10 nouitatis lil. 293. the Corpus cum should alings with ve entirely ria wrecked "y. + OWEN. > neil THE GRAMMATICAL CHAPTERS IN QUINTILIAN I. 4-8. A, THE five chapters which Quintilian has devoted to ‘Grammatica’ are in many ways the most valuable discussion of the subject which we possess. They are older than any other surviving account, except the remains of Varro De lingua Latina and the grammar of Dionysius Thrax, and this last, though far more complete than Quintilian in its examination of the parts of speech, has nothing that compares with the other chapters on analogy, etymology, etc., nor does it give so clear a view of ‘grammatica’ as a whole. These chapters do not appear to have been treated by any scholar since Nettleship published his article on ‘ Latin Grammar in the First Century’ in the Journal of Philology, vol. xv. If, in the course of this paper, I traverse on certain points the opinions of so distinguished a scholar, let it be understood that I do so with much diffidence, and only after much consideration of the facts as they appear to me. Quintilian divides ‘ grammatica’ into two parts, ‘ recte loquendi scientia,’ and ‘enarratio poetarum’ (i. 4, 2). The former, he tells us afterwards (1. 9, 1), is sometimes called ‘methodice,’ and the latter ‘historice. Here we have, of course, the distinction so familiar to ourselves between ‘grammar’ and ‘literature.’ The first question which I propose to examine in this paper is the history of this distinction. ‘Grammar,’ in our sense of the word, was no doubt in a way from the first an element in ypaypatixyn. Though first developed as a part of Aoyxn in the philosophical schools, the results there obtained were taken over and further developed by the Alexandrian Grammarians from the time of Aristophanes of Byzantium. That recognized ‘grammatici’ of various periods such as Aris- ‘archus, Dionysius Thrax, Nigidius Figulus, or Valerius Cato wrote on what we call ‘grammar,’ is a fact the evidence for which can be found in any hand- book to Greek or Roman literature. On the other hand, it is clear that the earlier conceptions of ‘ grammatica’ do not contain any clear or sharp distinc- tion between grammar and literature, such as we find in Quintilian. This appears both from formal definitions, as well as from general descriptions. Thus Cicero (de Oratore, i. 187) says, ‘in grammaticis poetarum pertractatio, 1 Afterwards reprinted in Lectures and Essays _ is given throughout. (Oxford, 1895), to the pages of which reference 34 F. H. COLSON historiarum cognitio, uerborum interpretatio, pronuntiandi quidam sonus.’ Philo, who has a great many references to the grammarians, always regards the study as literary. ‘Grammatice’ teaches us the toropia: twapa trountais Kai cuyypadedor.! It évtpépetac maraais Kai wyvyios Sofais;? it enables us apxaiav axony épywv Karav peTadioxe, GTEp LoTopiKol Kal GUpTrAaY TO TrOLNTLKOY yévos Trapaded@xace ;* although elsewhere‘ it is pointed out that the ypaypartixos borrows the explanation of the parts of speech and the nature of words from philosophy. But the short treatise of Dionysius Thrax shows us most clearly the relation which ‘ grammar’ is conceived of as bearing to ypaypatixn. The book really deals entirely with ‘grammar, but it opens with a definition of ypappatixyn and its constituent elements which is almost entirely literary. ypaupatixn itself is defined as €urepia tov Tapa Trontais Te Kal cvyypadedar @S €ml TO TOAD Aeyouevwmv. This is followed by the enumeration of its six pépn: (1%) avayveors ; (2) eEnynows Kata TpoTrous; (3) amodocts yAwoowy kal iatopiav ; (4) evpecis eruporoyias; (5) avaroyias éxroyiopos; (6) Kpiow. Of these the fifih, and possibly the fourth, are grammatical, but it is clear that the exposition of the parts of speech, etc., into which Dionysius plunges at once, has nothing to do with these, but is intended to serve as an introduction to dvayvwows. The doctrine evidently is that if we are to read properly and intelligently the things contained in the ‘poets and prose-writers,’ we must first understand the words of which they are composed. In fact, ‘ grammar’ stands tc ypaypatixn in much the same relation as the axioms, definitions and postulates of Euclid stand to geometry, a comparison which in fact is actually made by Philo.? It is interesting to compare with Dionysius the views of Varro so far as we know them. Varro defines ‘ grammatica’ itself in exactly the same terms,® but instead of the six yépn of Dionysius we have four ‘ officia ’—viz., ‘ lectio,’ ‘emendatio,’ ‘ enarratio,’ ‘iudicium.’’ I shall compare these two systems later ; here I will merely point out that all four belong to literature and not to grammar. For though ‘emendatio’ is defined by Diomedes as ‘ recorrectio errorum, qui per dictionem ® uel per scripturam fiunt,’ a definition which would seem to correspond to the ‘ recte loquendi et scribendi regula’ of Quintilian, he is probably mistaken in this. Varro’s definition of the four officia is merely a translation of the Greek ds0pOwots, avayvwors, €Enynots, Kpiois, which are declared by the scholiasts on Dionysius to be ta mada pépy Tijs ypappatixys, though they give them in varying order. When weremember that dcop@wars is the accepted term for correction of the text, there can, I think, be little doubt that the arrangement in which dcopf@wors comes first gives the right order, and 1 De Congressu, 4. This classification is discussed further on, p. 13. 2 De Sac, Ab., 22. Probably scribere, legerve, refer to the school of 3 ib. the ypaupuariorys (so Wilmanns), the other two 4 De Congressu, 26. to the school of the ypayparixés, and all four 5 ib. contemplate the work of the pupil, while the 6 Marius Victorinus K., vol. vi., p. 4. He other classification thinks of the teacher. adds ‘eius praecipua officia sunt quattuor, ut 7 Ap. Diomedes, K. vol i., p. 426, ipsi placet, scribere, legere, intellegere, probare.’ 8 Wilmanns proposed dictationem for dictionem. Nee — © Perec iw a yy Pow Po es a We th de ide of th tic thi am sonus.’ Lys regards rounTats Kal enables us ‘0 TTOLNTLKOV /PALMATLKOS words from nost clearly ruxn. The efinition of ly literary. ruyypagevar 1 of its six VWoo@V Kal kptow. Of ; clear that plunges at ntroduction roperly and s, we must ‘erammar’ nitions and t is actually ro so far as ame terms,°® iz., ‘ lectio,’ wo systems > and not to ‘ recorrectilo vhich would uintilian, he . is merely a , which are /PAPLLATLKNS, Su0pOwars is little doubt t order, and irther on, p. 13. » the school of the other two is, and all four upil, while the eacher. 20, m for dictionem. LOE DO rrr er rr ren, Se ~—— ee ee er ne. grr - THE GRAMMATICAL CHAPTERS IN QUINTILIAN I. 4-8 35 that a fairly correct explanation of the term is given by the scholiast who says mpo Tov aptacOar Tov véov avaytyvwoKe, 0 S10opOwris AapBavwv To BiBAtLov diwpOodTo avro.* On the whole, then, it would appear that ‘grammar’ made its way into ‘grammatica’ in an informal manner, asa necessary part of the equipment of the ‘grammaticus,’ but not as a separate branch or department. But at the same time it is probably true that from early times such a doctrine as Quintilian’s was occasionally held. The idea of ‘EAAnvcpos is as old as Aristotle, and the view that it is the first aper? of speech appears in the account of the Stoic system given by Diogenes Laertius (vii. 1. 40). In the face of the growing decay, or what seemed to be decay, of the language it was inevitable that an idea should spring up that the function of ypaypatixn was to preserve the laws of correct speech as well as to preserve the knowledge and understanding of the classical writers. Indeed, if we may trust a late authority, an early defini- tion was given in something like this sense. Marius Victorinus says, ‘ ut Aristoni placet, grammatice est scientia poetas et historicos intelligendi, formam praecipue loquendi ad rationem et consuetudinem dirigens.’? If this Ariston is the same as the Ariston whose definition of rhetoric is given by Quintilian (11. 15, 19), he belongs presumably to the second century B.c. But though this is a step towards Quintilian’s doctrine, it is not quite the same. ‘ Grammatica,’ it is true, has both the functions which Quintilian ascribes to it, but we have not arrived at the idea of two separate bodies of knowledge each with distinct functions. Two other considerations doubtless contributed to foster the conception In the first place the bulk of the subject grew immensely, and what in 150 B.c. might very well pass as a mere introduction to the study of literature, had by A.D. 50, or earlier, entirely outgrown such a conception. of ‘grammar’ as a science by itself. Seneca indeed*® seems to suggest that the ‘grammaticus’ might often find it necessary to confine himself to ‘cura sermonis.’* In the second place there was a growing feeling that the two subjects differed altogether in nature. The first was capable of being reduced to an ordered system; the second was essentially a mass of disconnected facts. Thus, says Quintilian, ‘ illam methodicen, hanc historicen uocant.’ At what date exactly this recognition of ‘ grammar’ as distinct from ‘ literature’ became thoroughly established, it is, I imagine, impossible to say. All we can tell is that it must be dated somewhere between Dionysius Thrax and Quintilian. It may be added that to judge from the discussion of the subject in Sextus E-mpiricus, the justice of the distinction was not questioned in later times. Professor Nettleship (Essay, p. 166) says ‘to Varro “ grammatica ” included ‘‘lectio,” “ enarratio,” ‘‘emendatio,” and “‘iudicium.” To Quintilian it is no more than “‘recte loquendi scientia,”’ and “ poetarum enarratio.”’ The school- 1 Gram. Graec., vol. iii., p. 12. * K. (vol. vi., p. 4). 3 Sen. Ep. 88, 3. # duédodos tX\n, Sext. Emp. Adu. Gram. 249. In Diomedes and elsewhere, the two are classed as opioTiky and e€&nyyriKy. 36 F. H. COLSON master has driven the philosopher away, or put him into the background.’ Whether the antithesis between ‘ philosopher’ and ‘schoolmaster’ is a sound one, or what there is especially philosophic in the Varronian classification, I will not stop to inquire. But the suggestion that the distinction the history of which I have been endeavouring to trace comes from the ‘schoolmaster’ is a point which deserves a little discussion. That the history of ‘ grammatica’ has in some ways been influenced by pedagogic considerations—considerations, that is, of what is easy and convenient for the teacher to teach, rather than of the natural order and real value of the things studied, is, I think, the truth. But the present case is not a case in point. The three considerations I have pointed out—(1) the supposed importance of preserving ‘EAAnvicpos ; (2) the rapid growth of scientific grammar ; (3) the intrinsic difference of what can be treated peOodixds, and what is dyéPodos, are matters which do not belong to the class-room but to the general trend of thought. Il. I now pass to Quintilian’s treatment of the ‘ recte loquendi scientia,’ which occupies chapters iv. to vi., and is followed bya chapter on the ‘ recte scribendi regula’ or orthography. This last chapter presents no special difficulties, but the scheme of chapters iv. to vi. is apparently very perplexed, and I cannot give the impression they produce at first sight better than by quoting Nettle- ship’s description : ‘That part of the first book which begins 1. 4. 1, and ends i. 5. 54, contains in an abridged form much of what might well have been found in an Ars Grammatica; i. 4 §§ 2-17 treats of letters, §§ 18-21 of the parts of speech, §§ 22-25 of nouns and cases, $§ 26-29 of the verb; 1.5 §§ 5-33 of ‘barbarisms,’ §§ 34-54 of ‘soloecismus’; and at this point Quintilian bids adieu to grammar. He then proceeds to consider the question of words—provincial, Gallic, Spanish, and Greek, simple and compound, literal and metaphorical. Thus the fifth chapter ends, and the sixth is a dissertation on sevmo or usage, considered under four heads, that of ratio or reason, including analogia and etymologia, antiquity, authority, and custom.”* This is, as I have said, a fair description of what appears at first sight, but the writer, I believe, has not observed the connecting-links, and the conclusions he bases on it seem to me erroneous. Assuming ‘as almost certain that Quintilian had some technical treatise before him the rules of which he throws into literary form,’ Nettleship proceeds to argue that iv. 1 to v. 54 comes from Palaemon,? and v. 55 to vi. 27 from Pliny. This assumption seems to me very far from certain. The points in 1 Essay, pp. 165, 159. Verrius show, I think, that Quintilian knew and 2 But as chapter vii. is supposed by Nettle- used Verrius, but the introduction of Palaemon ship to be ‘taken or adapted’ from Verrius, is pure hypothesis, at any rate in these sections. while iv. 6-17 exhibits the same characteristics as_ It is to be noted that some of the most note- vii., the theory has to be complicated by supposing worthy coincidences between Quintilian and that Palaemon borrowed these sections from Paulus occur in chapter vi., which Nettleship Verrius. As a matter of fact the resemblances does not ascribe to Verrius. between Quintilian and Paulus’ epitome of - —_ . wt a * te background.’ ‘> is a sound ssification, I the history of master’ is a grammatica ’ ynsiderations, ather than of 1k, the truth. ations I have ros; (2) the what can be belong to the ientia,’ which cte scribendi fficulties, but ind I cannot oting Nettle- . I, and ends ll have been 18-21 of the I. 5 §§ 5-33 of un bids adieu —provincial, netaphorical. ymo or usage, analogia and e said, a fair ieve, has not seem to me me technical a, Nettleship 7, 55 to vi. 27 Che points in itilian knew and ion of Palaemon n these sections. f the most note- QOuintilian and vhich Nettleship — . ~ — ne Pe Be THE GRAMMATICAL CHAPTERS IN QUINTILIAN I. 4-8 37 which Nettleship traces a correspondence between Quintilian on the one hand, and Palaemon and Pliny‘ on the other, appear to me most precarious, and even if they are sound they go very little way to show direct borrowing. Grammar was at that time a progressive science: each writer built upon the foundation of his predecessors, sometimes affirming, sometimes denying. Quintilian, no doubt, read and weighed his authorities. He actually quotes Caesar, Messalla, Asconius, and ‘nostri praeceptores.’ Many people seem to have a poorer opinion than I have of Quintilian’s ability. He may, as Rutherford says, ‘have read more than he can carry,’ but no one, I should have thought, could have failed to see that he always tried to think out things for himself. As to Palaemon, while it is true that the scholiast on Iuu. 6. 451 says that Quintilian had been his pupil, and that in that case he probably owes much to him, it should be noted that in the one passage from the part supposed to be adapted from Palaemon, in which he mentions that writer’s name, he says that ‘some authorities give nine parts of speech, but Aristarchus and in our own times Palaemon gave eight, as they did not distinguish the noun from the appellation, but whether this is right or not, as it matters little, I leave an open question.’ To suppose that this is a transcript from Palaemon seems somewhat perverse. . But Nettleship’s view involves a more serious charge against Quintilian. He is charged with having dovetailed Palaemon on grammar into Pliny on ‘sermo’ without any attempt at coherence or arrangement. This is entirely contrary to Quintilian’s habit, and it is a supposition which should only be adopted as a last resource. I believe the chapters, if read properly, show a very clear and coherent scheme. Quintilian begins with some remarks on the parts of speech and their sub- divisions, and of this chapter Nettleship’s analysis may be accepted; but it should be observed that there is no attempt to give any detailed account like that of Dionysius Thrax, but merely to note a few points which either are frequently neglected or present special interest. Thus he dwells at some length on the eight (or nine) parts of speech, because so many different opinions were held onthem. The importance of declining nouns and verbs is mentioned because so many teachers neglect it. Such questions as whether the ablative should be distinguished from the instrumental or the impersonal passive from the ordinary passive are noted. The whole of this chapter, in fact, is largely directed to meet the objection that grammar is ‘tenuis et ieiuna,’ and to show that in this introductory branch many difficult and interesting quaestiones arise. agreed with him; (d) Pliny is stated to have followed Varro; (c) is mentioned by Varro, L.L. viii. 35. All three were stock cases. Two other points are mentioned which Nettleship thinks may have been quoted by Charisius from Pliny. One of them, at any rate, is in Varro, the other, or a similar one, had been discussed both by Pliny and Caesar. ' He brings forward three points in which Quintilian discusses or mentions questions known to have been discussed by Pliny; (a) declension of Greek nouns in -o; (b) the coincidence in gender of the diminutive with the principal noun; (c) the two forms Albanus and Albensis from Alba. Of these (a) had been discussed by Caesar, and Pliny and Quintilian both dis- 38 F. H. COLSON When this work of classification and definition has been disposed of (so Quintilian continues in chapter v.), the principles of ‘emendate’ or ‘recte loquendi’ may be approached. These apply (1) to ‘singula uerba’; (2) to ‘plura uerba.’ ‘Singula uerba’ may be classified as (a) native or foreign; (b) single or compound; (c) literal or metaphorical; (d) in common use or made up (usitata aut ficta).. It is in choosing between the two classes in each pair that the ‘regula recte loquendi,’ so far as ‘ singula uerba’ are concerned, chiefly acts. But before we deal with these we must dispose of actually wrong or impossible words, (1) ‘ barbarismi,’ which belong to ‘uerba singula’; (2) ‘soloe- cismi,’ which belong to ‘ plura uerba.’ We accordingly deal with these. But barbarisms and solecisms are mere deformities (foeditas); the normal applica- of the ‘regula recte loquendi’ lies rather in the choice of words which are more or less dubia. We therefore now return to our main subject,® and say something about the lines to be followed in choosing ‘ singula uerba’ as classi- fied above. Having done this we should naturally proceed to deal with the ‘ regula recte loquendi in pluribus uerbis.’ Instead of this we have the dissertation on ‘sermo’ of Chapter vi. The solution of this difficulty lies, I believe, in sup- posing that ‘sermo’ is simply the same as ‘plura uerba.’ That Quintilian frequently uses ‘sermo’ as a contrast to single words may be shown by the following, (i. I, 31) ‘syllabis uerba complecti et his sermonem connectere,’ (i. 5, 2) ‘omnia (i.e. uerba) per quae sermo nectitur,’ (1. 5, 34, of the solecism) ‘in uerbo esse uitium non in sermone contendunt,’ (vill. 6, 1) ‘ tropus est uerbi uel sermonis ... It is not perhaps so easy to see how the examples which Quintilian treats in his discourse on ‘sermo’ fall under the head of ‘plura uerba.’ Why, for instance, should the question of domu or domo (i. 6, 5) or of audacter or audaciter be matters of ‘plura uerba’ any more than the choice of vaeda (i. 5, 57) (instead of some native Latin word), which is classed as ‘ unius uerbi’? The answer is, I think, that chapter vi. deals with formations or ‘declinatio,’ a term used very widely to describe not merely inflections, but derivatives, and that every such ‘ declinatum’ is considered in mutatio.’ relation to the word from which it is formed, and therefore raises a question not of ‘unum uerbum’” but of ‘plura uerba.’ This view that ‘sermo’ (in chapter vi.) = ‘declinata’=‘ plura uerba’ = ‘formations’ seems to be confirmed by i. 5, 16. Here Quintilian says that as ‘ barbarismi’ applies to single words, it may perhaps seem absurd to apply the term to cases where the erroneous form involves an error in a number or gender. He instances amongst others 1 Elsewhere called ‘ noua,’ 2 «Arte non egent, cuius in hoc opere non est usus nisi in dubiis,’ i. 6, 38; ‘ analogiae haec uis est, ut id, quod dubiwm est, ad aliquid simile de quo non quaeritur referat,’ i. 6, 3. 3 ‘hoc amplius wt institutum ordinem sequar.’ These are the words with which, on Nettleship’s theory, Quintilian ‘having bid adieu to Grammar’ (and Palaemon) proceeds to make a literary transcript of Pliny! No more unlikely place can be imagined for such a complete break. 4 One of the four pairs ‘ translata’ and ‘ pro- pria’ is passed over with a bare mention and no examples. Quintilian, no doubt, felt that it belonged rather to rhetoric—i.e., it was not so much a question ‘recte loquendi,’ as ‘bene dicendi.’ es a Se Ss Fe > _— -_~ ~~ fo ef 5S - Cf posed of (so e’ or ‘recte srba’; (2) to or foreign ; mon use or asses in each > concerned, tually wrong .’; (2) ‘soloe- these. But mal applica- is which are ect,’ and say a’ as Classi- the ‘ regula ssertation on lieve, in sup- t QOuintilian iown by the connectere,’ he solecism) Dus est uerbi he examples the head of uu or domo ’ any more vord), which 1. deals with not merely onsidered in S a question ‘sermo’ (in onfirmed by xle words, it oneous form ngst others likely place can oreak, lata’ and ‘ pro- nention and no t, felt that it , it was not so ndi,’ as ‘bene er Be — —_————_ a, —— THE GRAMMATICAL CHAPTERS IN QUINTILIAN I. 4-8 39 gladia for gladius. He himself apparently regards this as an error ‘ unius uerbi,’ for let gladia once be postulated, gladiam, gladiae follow normally, and are not really erroneous, but he is aware that the ‘ prima facie’ view is that all these forms in -a are erroneous, and therefore are not ‘ unius uerbi.’ Assuming then that this chapter gives us the ‘ regula recte loquendi in pluribus uerbis,’ what is this ‘regula’? I understand it as follows. The ‘declinatio’ of words is naturally and normally determined by ‘ratio,’ i.e. by ‘analogia’ and ‘etymologia.’ Thus by ‘analogia’ we should expect domus, domts, domu; senatus, senatus, senatu; ferueo, feruet: while by ‘ etymologia ’ we should expect medius to produce medidiem. The forces which militate against this natural tendency are ‘ uetustas,’ ‘ auctoritas,’ ‘consuetudo.” Thus ‘uetustas’ might be pleaded to justify senati, the ‘auctoritas’ of Lucilius for f:ruit, while ‘consuetudo’ declares for domo and perhaps for meridiem. It is true that in developing these heads Quintilian does not always confine himself to examples of ‘ declinatio.’! The question, for instance, of the correct- ness of topper is really a question ‘unius uerbi.’ But I do not think this is any objection to the view of ‘sermo’ and of this chapter taken above. T’e four forces ‘ ratio,’ ‘ uetustas,’ ‘ auctoritas,’ ‘consuetudo’ work principally in the sphere of ‘plura uerba,’ but they also work elsewhere. It is not very remarkable that Quintilian having introduced these four heads should discourse at large upon them, without confining himself strictly to the subject in connec- tion with which he introduced them. This may convict him to a certain extent of irrelevance, but irrelevance is a different thing from incoherence, of which under any other view of ‘sermo’ as used in this chapter, he must, I think, stand convicted. With regard to these four heads, Nettleship,? who holds (I think on quite insufficient grounds) that they come from Pliny, contrasts them with Varro’s ‘natura,’ ‘analogia,’ ‘ consuetudo,’ ‘ auctoritas,’ as reported in Diomedes (K. 1, 433). What Diomedes actually says is: ‘Latinitas est incorrupta loquendi obseruatio, secundum Romanam constat autem, ut asserit Varro, his quatuor, natura, analogia, natura uerborum nominumque immutabilis est, linguam. consuetudine, auctoritate. nec quicquam aut minus aut plus tradidit nobis, quam quod accipit. nam si quis dicit scrimbo quod est scribo, non analogiae uirtute, sed natura ipsius constitutionis conuincitur.’® Now passing over the point that Varro is stated to be discussing ‘ Latinitas ’ not ‘sermo,’ we may ask, Are the second and third sentences intended to give us Varro’s view of ‘natura’? If so, it corresponds exactly to Quintilian’s belongs to ‘ plura uerba.’ 2 Essay, p. 161. 3 In Charisius (K. 1, 50) the same statement with the same explanation of ‘natura’ and the same example of scrimbo is given of ‘sermo,’ but here it is not ascribed to Varro. 1 The fact is that it is very difficult to keep the two apart. Thus in v. 55 etc., the question of the choice between Greek and Latin words, which is a matter of ‘ singula uerba,’ inevitably leads to a discussion of Greek declension as opposed to Latin declension in names like Calypso, which, as a matter of formation, really 40 F. H. COLSON ‘singula uerba.’ Scvimbo is a ‘barbarismus.’ We decide between scrimbo and scvibo without reference to any other word. On this view of ‘ natura’ Quintilian, in deciding the principles on which his ‘sermo’ depends, would be bound to eliminate it, and his division is really much the same as Varro’s, or at least merely exhibits the development which might be expected during the time that elapsed between the two writers. I find it difficult, however, to bring Diomedes’ explanation into harmony with the ninth book of the De Lingua Latina. There Varro clearly identifies ‘natura’ with ‘analogia.’ If we follow analogy we follow nature: if we allow ‘ consuetudo’ to override it we go against nature. I am inclined on the whole to believe that what Varro really said was ‘natura, siue analogia.’ On either view, however, Quintilian will be brought into substantial agreement with Varro. The scheme then of chapters iv.—vii. in my view is as follows. The subject of the whole is ‘ recte loquendi et scribendi scientia’: Chapter iv. The grammarian must observe and classify the facts of the language. Chapter v. He must then apply this knowledge practically to establish the ‘ regula recte loquendi’ (a) in singulis uerbis, (0) in pluribus uerbis. (a 1) He will eliminate actual mistakes ‘in singulis uerbis ’ (barbarismus). (6 1) He will do the same ‘ in pluribus uerbis’ (soloectsmus). (a 2) He will deal with ‘dubia’ (i.e. with questions where there is fair room for discussion) ‘in singulis uerbis.’ Here the choice will be chiefly between (a) native and foreign words, (8) compounded and uncompounded, (y) literal and metaphorical, (6) familiar and new. Chapter vi. (6 2). He will deal with ‘dubia’ ‘in pluribus uerbis.’ Here his choice will lie between the conflicting claims of ‘analogia’ and ‘ etymologia’ on the one hand, and ‘ uetustas,’ ‘ auctoritas,’ and ‘ consuetudo’ on the other. It is to be understood that both in a 2 and 6 2 the final choice must depend on the ‘ iudicium’ of the grammaticus.? The subject is concluded by some remarks on the ‘ regula scribendi recte,’ Chapter vii. discussing various ‘ dubious’ cases, and deciding that the general rule must be to adhere as closely as possible to the sound of the word unless ‘consuetudo’ decides otherwise. Thus the chapters form an organic whole, though they do not attempt to give a complete ‘ Ars,’ but merely to sketch the principles on which the ‘ recte loquendi et scribendi scientia’ should be taught. They do not pretend to originality, but it is impossible to say on what authorities the writer relies, and for the theory that one part of them is drawn from one author, and another from another, there is no evidence whatever. [The words ‘est etiam sua loquentibus obseruatio, sua scribentibus’ at the beginning of Chapter vi. may seem to be against my view of the connection of that chapter. They look as if a new departure were now taken, and we were entering upon the question of spoken language in general, as opposed to questions of mere spelling, which are to be dealt with in Chapter vil. I am 1 i, 6, 3 ‘omnia (i.e. analogia, etc.) exigunt acre iudicium.’ een scrimbo of ‘ natura’ ends, would as Varro’s, cted during however, to of the De alogia.’ If override it what Varro , Ouintilian Che subject acts of the o establish ‘bis. vbarismus). here is fair be chiefly ym pounded, - Here his tymologia ’ the other. 1ust depend endi recte,’ he general vord unless inic whole, sketch the be taught. authorities n from one bus’ at the nection of 1d we were »pposed to vil. Jam THE GRAMMATICAL CHAPTERS IN QUINTILIAN I. 4-8 = 41 inclined to think that these words really belong to Chapter v., and refer to the subject with which that chapter closes—viz. ‘ onomatopoeic words.’ Quintilian does not here (though no doubt he does elsewhere) contrast ‘loqui’ with ‘scribere’ in the sense of words as they sound opposed to words as they are spelt, but in the sense of conversational language opposed to the more formal written language. Many onomatopoeic formations, he means, may sound all right in conversation, but you would hardly venture to use them in writing. We should probably say the same of slang in general. ] | ITT. I now turn to the consideration of the literary side of ‘ grammatica,’ that side which he describes in chapter iv. as ‘ enarratio poetarum,’ and in chapter ix. as ‘historice.’ But before doing this it will be well to compare Dionysius’ classification with the Varronian, which, as I have said, is not really Varro’s, but taken over by him from the Greek schools. For convenience, however, I will continue to use the term ‘ Varronian.’ We have: Dionysius. Varronian. | Kal” vroKpiow di0p8wous avayveaos 1 KaTa Tpotwodiay avayvoo.s = dtacToAny eEnrynots éEnynols KaTa TOUS évUTTaPKOVTAS TOLNTLKOUS TPOTFOUS Kplous A . & nw / yAwoowv Kal ioTopLay aTroboals ETUMLOAOYLAS EUPETIS 7 , » ] / avadoylas exXNoyLa pos / / KpPlals TTOLNMLATOV. We may fairly assume that the ‘ Varronian’ classification is the older, and we may note the following differences : Firstly, dsopA@wars is absent in Dionysius’ list, a phenomenon which we shal] find recurring in Quintilian. I do not think we should infer from this that the grammarians in their capacity of scholars had ceased or were ceasing to pay attention to textual criticism. The history of the school of Aristarchus proves the opposite, and there are several passages in Quintilian’ and other writers which testify to the attention paid to MSS. evidence. The omission is due to what may with justice be called a ‘ pedagogic’ reason, though at the same time I believe a perfectly sound one. It was thought wisest to put into the hands of boys the best text and expound that, without troubling them with various readings, a view in which probably most schoolmasters to-day would concur. Rutherford remarks on the paucity of textual notes in the scholia on Aristophanes.” Secondly, the vague term é&nynows has been developed into (1) é&yynous 1 Useful examples are given in Lehr’s Aristar- 2 A Chapter in the History of Annotation, chap- chus, p. 344. ter 1. 42 F. H. COLSON KATA TOUS evUTPaPXOVTAS TrOLNTLKOVS TpOTrOUS, (2) yAwoowY Kal ioTopLaY TpPOKELPOS amrodoots. This brings us to the very heart of the ancient theory of teaching literature. The underlying doctrine I take to be this: Obscurity, which it is the purpose of é&nynars to remove, arises in poetry from three causes: (1) The use of uncommon or obsolete words (yA@ooar), (2) allusions to facts unknown to the reader (‘oropiar), (3) the presence of ‘ tropes,’ i.e. ‘sermones a naturali et principali significatione translati,’ or ‘ figures,’ i.e. ‘ conformationes quaedam orationis remotae acommuni.’! To explain the yAdooa and (oropia: all that is needed is that the grammarian should know and state each point as it arises.? The difficulties that still remain must be solved by giving the student a know- ledge of the possible tropes and figures, and training him to trace each case to its particular lair. Beyond these no further é&7ynors is required. If we had not so completely lost the science of tropes and figures, we might possibly find that this doctrine, though far from being absolutely true, contains a considerable element of truth. At any rate, I am sure that a certain amount of é&yynous Kata Tpotous would be found by teachers of English Literature to impart both interest and method to their lessons. Thirdly, while the science of grammar, generally speaking, is not reckoned by Dionysius as a separate branch of the subject, but rather as an introduction, the grammatical treatment of the books is to some extent acknowledged under the heads of parsing (avanoyias éxXoyiopos) and etymology. This classification does not seem to have held its ground, in spite of the great authority of Dionysius, to judge from Sextus Empiricus,* who says that Dionysius atoms diatpettat, accuses him of raising to the position of wépn what are really popia and azrotedéopata, and does not seem to have found the classification in any other authority. Now to turn to Quintilian, though his terminology has a certain incon- sistency, his general principles are clear. In Chapters iv. and ix. he speaks of the whole literary department as ‘ enarratio poetarum,’ but in the first of these he hastens to add ‘ enarrationem praecedit emendata lectio.’ In chapter viii., where he deals with the subject in detail, he divides the course of instruction into (1) ‘ lectio,’ (2) ‘ praelectio’ (in praelegendo grammaticus debebit, etc.), and (3) ‘enarratio historiarum.’ With regard to the first of these, while his account is more discursive and less business-like than that of Dionysius, it covers practically the same ground. Thus 8, § 1: ubi suspendere spiritum debeat, quo loco uersum = Dionysius’ avayvwous distinguere, ubi claudatur sensus, unde in- KaTa OLlacToAny cipiat, quando attollenda uel submittenda 1 Quintilian ix. 1, 4. more ‘heuristic’ method has to be adopted, 2 For this reason perhaps the curious epithet Cf. Sext. Emp. Adu, Gram., 249, isropuxdv dé To mpdxeipos is attached to yAwoodv cal icropiadv epi rhv mpoxepdrnra THs aueOddou Uns. amédoc1s. With them the teacher has merely to 3 Adu, Gram., 250. State the facts; with the éf#yno.s cara rpbrovs, a V TpoXeELpos of teaching which it is Ss: (1) The ‘Ss unknown a naturali Ss quaedam call that is s it arises.? nt a know- ach case to we had not ly find that onsiderable of €Enynous mpart both it reckoned troduction, dged under assification uthority of 1US aTOT@S eally popva ‘ion in any fain incon- e speaks of ‘st of these lapter viil., instruction , etc.), and his account , it covers aVAYVWOLS LT TOANY | be adopted, isropucdv dé To nS. THE GRAMMATICAL CHAPTERS IN QUINTILIAN I. 4-8 43 sit uox, quo quidque flexu, quid lentius celer- ius, concitatius lenius dicendum. S 2: sit autem in primis lectio uirilis et cum suauitate = «ata mpooc@édiav quadam grauis, et non quidem prosae similis . non tamen in canticum dissoluta nec plasmate . . . effeminata. § 3: nec prosopopoelas .. . nuntiari uelim, esse tamen flexum quemdam, quo distinguantur ab iis in quibus poeta persona sua utetur. ad comicum morem pro- = «a@’ vroxpiow ' It may be noted here that Quintilian with his usual incisive common- sense adds, ‘unum est igitur quod in hac parte praecipiam: ut omnia ista facere possit, intelligat,’ from which we may infer that the ‘ praelectio’ was to some extent what its name implies. The teacher had first to explain sufh- ciently to make the general sense intelligible. Detailed comments no doubt followed the reading lesson. When we come to the ‘ praelectio’ the duties of the grammaticus are stated thus: 1. He must make his pupil give: (a) The ‘ partes orationis’ (avanXoyias éxXoyiop0s), (b) The ‘ pedum proprietates ’—1.e. scan. 2. Point out ‘barbara et impropria, quae contra leges loquendi sint posita.’ 3. Show the various senses in ‘which words are used—i.e. homonyms. 4. Explain ‘ glossemata’ or ‘ uoces minus usitatas’ (yAwoo@r a7rodocts.) 5. Show tropes and figures (é&nynows kata tpotovs). This is to be done ‘maiore cura.’ And then follows this sentence : Praecipue uero illa infigat animis, quae in oeconomia uirtus, quae in decore rerum, quid personae cuique conueniat, quid in sensibus laudandum, quid in uerbis, ubi copia probabilis, ubi modus. That is, he will group his criticisms of the poem under the following heads : (a) Structure. (b) Appropriateness and beauty of the incidents. (c) Characterization. (d) Notable thoughts. (ec) Beauty of diction. (f) Criticism of the style (at least from the point of view of full- ness or brevity). 44 F. H. COLSON This passage, which I venture to think contains much useful instruction for those who essay the difficult task of teaching English Literature, corresponds I believe to the Dionysian xpious, 6 KddXMOoTOY éoTL pépos THS TéxVNS. Finally there comes, with a somewhat depreciatory paragraph to itself, the ‘ enarratio historiarum.’ When we compare this sketch with the scheme of Dionysius we see that while the general principles remain the same, there are some important differ- ences. The grammatical element has come much more to the front. To the bare parsing or avadoyias éxXoyiopos is added scansion and the noting of ‘ barbara et impropria uerba,’ and the study of ‘homonyms.’ This, of course, was bound to be the consequence of the elevation of the ‘loquendi regula’ into one half of the whole ‘ grammatica.’ The ‘grammaticus’ might and did, no doubt, lecture separately on the ‘ loquendi regula,’* but he was also sure to use the literary ‘praelectio’ as a means of enforcing and illustrating the rules which he had expounded. Why érupodroryias eipecis finds no place in Quin- tilian’s scheme it is difficult to say. He has already discussed the value of etymology as a criterion of correct language, and though he thought much of the current work in the subject fanciful and ridiculous, he can hardly have intended to exclude it from the ‘praelection.. The omission is, perhaps, purely accidental. I have suggested above that the passage ‘ praecipue—modus’ 1. 8, 17 corresponds to the Dionysian xpiovs, but this point requires some discussion. Rutherford? holds that Dionysius rather refers to the art by which the critic detects the spuriousness of lines, passages or words, and it is true enough, as he says, that this was the part of Aristarchus’ work which attracted most notice in antiquity. It may be added that the scholiasts of Dionysius are unanimous in putting this interpretation upon the word, and further, perhaps, that Quintilian, in i. 4, 3,° implies that one of the functions of ‘iudicium’ was to declare that passages or whole books were unworthy of the supposed author and therefore spurious, or that the author as a whole was unworthy of a place in the ‘canon.’ It seems to me indeed most probable that in the formula d:0pOwors, avayvwors, €Enynots, Kpiows, the last term did originally bear this limited meaning. My reason for thinking that Dionysius uses it in a wider sense is that his purpose appears to be pedagogic, while the original formula thought of the ‘ grammaticus’ rather as a scholar. Just as Dionysius omits ds0pAwois because textual emendation was not a suitable study for schoolboys, so I take it he may be expected to have held the view that before you lay down to boys that such and such passages are genuine or spurious you must teach them to appreciate their merits or demerits. In what sense Varro used ‘iudicium’ is also uncertain, but such evidence as we have seems 1 Quintilian i. 2, 14, ut non uersus modo censoria quadam uirgula 2 A Chapter in the History of Annotation, chap- notare et libros, qui falso uiderentur inscripti, ter vi. tanquam subditos summouere familia permise- 3 ‘et mixtum his omnibus iudicium est, quo _ rint sibi, sed auctores alios in ordinem redegerint quidem ita seuere sunt usi ueteres grammatici, alios omnino exemerint numero.’ instruction orresponds s. Finally ‘ enarratio ve see that tant differ- it. To the > noting of of course, egula’ into ind did, no sure to use x the rules e in Quin- ne value of ht much of 1ardly have s, perhaps, s’ 1. 8, 17 discussion. 1 the critic enough, as acted most onysius are ar, perhaps, iclum ’ was 2 supposed inworthy of that in the vinally bear ses it in a the original s Dionysius > study for that before purious you what sense have seems jadam uirgula sntur inscripti, milia permise- nem redegerint THE GRAMMATICAL CHAPTERS IN QUINTILIAN I. 4-8 45 to point to the wider meaning. In the first place, both Marius Victorinus and Diomedes explain the term as equal to ‘ bene dictorum comprobatio,’ and, secondly, the other classification ascribed to Varro suggests that just as intelligere corresponds to ‘enarratio,’ so does probare to ‘iudicium.’! The teacher indicat, i.e., notes the merits of the writer; the pupil probat the same. Dionysius then having in view the requirements of the class-room, has, I think, widened the term xpiots, originally suggesting the sentence pronounced by the ‘grammaticus,’ to include the grounds of judgment on which such sentences were formed, to mean in fact ‘critical appreciations.’ Quintilian goes a step further: he keeps the substance of the Dionysian xpicus, but refuses to use the term (at least so I understand the passage i. 4, 3 quoted above) because he wants it for something else. His meaning is that ‘ iudicium’ cannot be limited to any particular department of ‘ grammatica.’ It must be remembered that Quintilian’s most fundamental educational principle is the right and duty of private judgment,” that there are no such things in any department of learning as ‘xafodrKa praecepta ’—rules which are universally and invariably true. Thus in iii. 3, 5 and vi. 5, 1 he deprecates the view that ‘judicium ’ can be regarded as a special province in rhetortc, for it is inseparable from every part of the work, and cannot be made the subject of the rules of art, any more than taste or smell. And so too with ‘grammatica,’ the word ‘judicium’ is wanted for the faculty of independent judgment, which is capable at any moment of overriding the dictates of general law and thinking for itself. We have seen that Quintilian reserves to the end the ‘ enarratio histori- arum ’—the exposition, that is, of questions arising out of the subject-matter of the book or allusions contained therein. The best account of ‘historiae’ is to be found in Sextus Empiricus,* making due allowance for his hostile bias. ‘ Histories’ may be classified as true (historical in the limited sense), or fictitious though possible as in comedies (7wAdopata), or fabulous (udGor). Or again they may be classified as dealing with places, dates, personages or actions.. Some examples are given. ‘Brilesus is a mountain in Attica;’ ‘Acamas is a promontory in Cyprus.’ ‘Pelops’ shoulder was of ivory, and was eaten by either Ares or Demeter.’ ‘Xenophanes was a Colophonian, and lived about the fortieth Olympiad.’ ‘Plato’s name was Aristocles, and he wore earrings.’ That information of this sort, which is frequently relevant and even necessary, should constantly tend to develop into laborious erudition is a necessity of human, or at any rate of a schoolmaster’s, nature. The sarcastic references to this are of course very frequent. I need not do more than refer to Mayor’s note on Juvenal vii. 234. Seneca in a passage® already quoted says ‘ grammaticus circa curam 1 Vide note on p. 3. 4 Adu, Gram. 250, and following sections. 2 Quintilian ii. 13, 14. 5 Ep. 88, 3. 3 Rutherford translates ‘ matters of fact.’ 46 F. H. COLSON sermonis uersatur, et si latius euagari uult, circa historias, iam ut longissime fines suos proferat, circa carmina.’ The natural meaning of this is that some grammarians taught ‘historiae’ without teaching anything else about ‘car- mina.’ In other words, ‘historiae’ in Seneca’s time were regarded as coming earlier in the literary curriculum than the other parts of ‘enarratio poetarum.’ Now it is no doubt possible that Seneca may be writing with the same dogmatic ignorance with which modern journalists and speakers often write and speak about school practice, but his words receive a strong confirmation from Sextus Empiricus. Sextus knows nothing of the Varronian classification, and only mentions the Dionysian to express his astonishment at it. With him the accepted classification is into teyvixov, iotopixov, idiaitepov. This last is described thus: ro Kata Tovs Tointas Kai cvyypadels emicKxoTovat, Kab’ 06 Ta acagas Neyoueva €Enyobvtat Ta TE Uyth Kal TA wn TOLAVTA KpivoveL, Ta TE yYNoLA amo TOV ToOLoOVTwY Stopifover.. Sextus does not suggest as Seneca does that there were schools where 70 iévaitepoyv was not taught at all, but he does suggest that the latter was felt to be the more advanced, and that there was a clear dividing line between it and the ‘historic’ treatment. When Quintilian reserves his remarks on the ‘ enarratio historiarum’ to the end he does not of course mean to reverse the order given by Seneca and Sextus.” He feels that the ‘historiae’ absorb so much time and attention that the question of their treatment demands special discussion, and he wishes to warn the teacher against exaggerat- ing their importance.® One other important difference between Quintilian and Dionysius remains to be noted. The latter includes prose writers (cvyypadeis) in the subject- matter of ypaupatixyn, while in Quintilian, although prose composition of a sort is learnt from the grammarians, the study of prose-writers is entirely reserved for the rhetoric school. So far as I know, in all Roman writers the practising sphere of the grammarian is invariably limited to poetry, though the view that originally ypapparixy) included at any rate the historians is put forward by Quintilian himself in 11.1, 4.4 What, then, led to their complete exclusion in later times from the Roman schools, if not from the Greek ? I have no doubt that we must largely ascribe it to the Roman conviction that poetry is an essential element in the formation of national character, a con- 1 Adu. Gram. 92, 93. I think this passage selves mainly to yAwoodyv kai loropidv damddoais, supports the view of xpisis taken above. He seems to use vy7js in the sense of ‘ valuable,’ ‘morally sound.’ I do not know whether tyj7s is ever used of ‘ sound text.’ Stephanus gives no examples of it. 2 Ini. 2. 14 he gives the same order as the others, ‘grammaticus si de loquendi ratione disserat, si quaestiones explicet, historias exponat, poemata enarret.’ 3 The view adopted by Quintilian and those who advocated the teaching of 7d ldiairepov as superior to 70 icropiuxév is Closely paralleled by modern views of teaching Shakespeare. The Clarendon Press editions, which restricted them- have now given place to the Pitt Press and other editions, in which xpicis plays a prominent part. 4 Quint. i. 4,4, ‘mec poetas legisse satis est: ex- cutiendum omne scriptorum genus non propter historias modo, sed uerba, quae frequenter ius ab auctoribus sumunt,’ does not contradict this. Quintilian is here speaking of the equipment of the ‘grammaticus’ himself. Just as he will require some knowledge of astronomy and music, so he must read prose-writers, both in order to master the rules of language and to enable him to deal adequately with the ‘historiae’ of the poets, longissime ; that some bout ‘ car- as coming poetarum.’ ie dogmatic and speak rom Sextus 1, and only th him the “his last is , Kal’ Oo Ta ‘a TE yVnoLA s that there suggest that ear dividing reserves his ourse mean 2 ‘historiae’ r treatment t exaggerat- ius remains the subject- iposition of ; is entirely | writers the _ though the clans is put ir complete the Greek ? \viction that icter, a con- Topiav amddoas, Press and other $ a prominent se satis est: ex- us non propter ‘equenter ius ab contradict this. e equipment of ust as he will omy and music, oth in order to | to enable him istoriae’ of the THE GRAMMATICAL CHAPTERS IN QUINTILIAN I. 4-8 = 47 viction so nobly expressed by Horace in the first epistle of the second book.! But it is also probably due to a pedagogic reason which remains for all time. Many teachers to-day, I fancy, find it exceedingly difficult to treat English prose-writers with any effect in school, and to produce a satisfactory examination paper upon them is as a rule still more difficult. The teacher may very possibly prefer Robinson Crusoe or Treasure Island to a play of Shakespeare, or at any rate think them more suitable and interesting for young boys, but he has not the slightest idea what to do with them in class. They have no yA@ooas or totopiat, and I do not think they would lend themselves to €&)ynots Kata tpotrovs, even if our schoolmasters knew anything of this last. We gather from Quintilian ii. 5 that when the orators and historians were read in the rhetorical schools they were merely read as models,? and little attempt was made to treat them in the same systematic way in which the poets were treated in the grammatical schools. Greek ‘ rhetors ’’ sometimes attempted something of the sort, though even they usually delegated it to their assistants. Quintilian himself advocated it, but only to the extent of giving literary appreciations, and even this he found was unwelcome to many of his pupils. The Classical Quarterly is not the place to discuss modern educational questions, but I may perhaps be allowed to say in conclusion that the ancient teachers of literature had thought out their business with a care and thorough- ness from which their modern successors have much to learn. There is a tendency to depreciate the schools of the Empire, and I am free to admit that in some ways they did their work too well. ‘They swamped the sacred poets with themselves.’ But I cannot think that their work was on the whole unsuccessful. It is true that neither ‘grammatica’ nor rhetoric taught people to write or speak with genius and power (what school ever did or could ?), but they did teach a very clear insight into the meaning, nature, and functions of language. Perhaps their work may best be seen in two products of these centuries, which have had at any rate a very remarkable permanence—the Roman Codes and the Creeds of the Christian Church. F. H. Cocson. | 1 Also admirably stated by Mr. Garrod in his ? And also for the value of the subject-matter introduction to the Oxford Book of Latin Verse. as ‘exempla.’ SALAPANTIVM DISERTVM. Catullus liil. 5. Di magni, salapantium disertum. Our best MSS. agree upon salapantium. But they also agree upon desertum for disertum. Seneca, who quotes the last two words of the line, has salaput(t)ium disertum ; and since he is right about disertum it is supposed that he is right about salaputium ; and salaputium stands in all our texts of Catullus. What it means nobody knows. It is mostly relegated to that numerous class of Latin words of which we conjecture that they are obscene and are content that they should remain obscure. A few commentators, aiming at respectability in their etymologies, connect it unconvincingly with the Greek cadmuyerns. But no one will say what is the matter with the salapantium of Catullus’ MSS.; and I begin to think that it has not occurred to anybody that it has got, or could have,a meaning. I say ‘to anybody’; yet I am not sure that the scribe of the Codex Ashburneri, or the scholar he served, did not long ago find the truth. The Codex Ashburnert, often cited by Ellis, is in the main a worth- less MS.: I have myself examined it for hours on end. But it has here a not uninteresting variant. It offers the form salafantium, 1.e. salaphantium. Sala- phantium is a diminutive of salaphanta. Someone called Calvus ‘an eloquent little salaphant. And why not? If he had called him ‘an eloquent little sycophant’ we should have understood him well enough. And there is not a whole world of difference between a sycophant and a salaphant. A salaphant is the same thing as a halophant (or halapant)—the one word is merely a hybrid variant for the other: and the halophant and sycophant occur together in literature more than once. WNonius p. 172, Lindsay, has halophantam aut sycophantam, hominum genus nequam quod ob suenda mendacia miserrima mercede conducitur, and he goes on to quote Plautus, Curculio 463, halophantamue aut sycophantam magis esse dicam , and from some unknown comedian halophantam mendacem uelit. agree upon the line, has apposed that ; of Catullus. merous Class are content espectability TANTUYKTNS. of Catullus’ at it has got, ure that the yng ago find ain a worth- is here a not tium. Sala- ‘an eloquent oquent little here is nota A salaphant rely a hybrid ' together in d ob suenda SALAPANTIVM DISERTVM 49 Festus, in the excerpt of Paulus (Lindsay, p. go), offers the form halapanta. ‘Significat omnia mentientem,’ he says (suggesting an inept derivation which need not detain us). Halophanta, halapanta, salap(h)anta, then, are one and the same word, a word formed humorously on the analogy of sycophanta, and meaning any liar on a large scale whose services may be hired. The ‘ nescio quis e corona’ who lifted his hands in astonishment and called Calvus a salapantium (he used the diminutive because, as Seneca says, Calvus was ‘ paruolus statura ") meant to call him plainly ‘a mighty little liar’; and that was what made Catullus laugh. H. W. GarRRop. Merton Coitiecre, Oxrorp. ETYMOLOGIES AND DERIVATIONS. I. Sanskrit med/n ‘comes’: pera. 1. In Skr. medin we have an Indo-Iranian -7z derivative of a proethnic start-form (s)met-sdos ‘ co-sedens,’ whose initial s may have been lost by haplology, but cf. Av. mat ‘pera.’ Homeric dofos ‘attendant’ (start-form sm-sodyos) is a like compound, meaning co-sedens and not ‘ mitganger’ (pace Johannson IF, III. 199), but has suffered psilosis. Out of composition, unless the ‘suffix’ conceals a posterius, we may have a further cognate in Lat. sodalis * boon-companion,’ wherein soda@- may have meant something like ‘session’ (the council or board of elders in the Scotch kirk). 2. I further hold that pera, before it became an adverb and preposition, must have functioned as a noun, *yvéra, meaning ‘ comes’; and as an adverb in Homer it has scarcely passed beyond the sense of ‘comes’ as used, with or without a dative regimen, in combination with zz ‘ goes’(=geht mit; cf. eg. Tibullus I. 4. 41; Aen. 6. 159; see a collection of the usage in Class. Journal 8, 253). The noun type in -d (? 2), excluding dubious cases like Homeric tr7rora, is best illustrated by Latin words like aduena (with secondary x for m). As ad-uena is a compound, so I analyze *uéra as a compound of sm, preaccentual reduction of som, +eta (eta?) ‘errans, currens, iens’: Skr. diati ‘errat.’ I have found this noun before now in Homeric éraiz ‘ comites,’ from sw- ‘ co-’ + éraz ‘ euntes’ (AJPh. 28, 166), as also, excep. exctp., in Latin com-et-. That in comes the e may be original and not represent an 7 (so already Ebel in KZ. 5, 187, anno 1856) seems to me demonstrated by the following cognation : II. Latin pedei-: Skr. padatis ‘ foot-soldier.’ 3. The Petersburg lexica give the following forms: paddti-s, pddat (lexical only), padata-s, padata-s (pronounced incorrect). The first form is a good epic word, and was analyzed in PW" (1865) as ‘ pad- oder pada- +4att “‘ gehend ” von at-,’ though Walter had in the meantime (1861) already advanced the now reigning explanation whereby equet- e.g. is equated with itmmorns (KZ. 10, 194 sq.); and the analysis of padati- as a compound was expressly rejected (p. 205). Walter’s conclusion has been held pretty much ever since, so that one would search in vain in the handbooks for a reference from pedes to paddait-s (e.g. in Walde’, p. 579). But, as Bartholomae saw (BB. 17, 114"), we cannot f a proethnic been lost by * (start-form singer’ (pace sition, unless rnate in Lat. ymething like d preposition, ;an adverb in ith or without . eg. Tibullus uynal 8, 253). wroTa, 1s best As ad-uena 1s tual reduction I have found -€rae ‘euntes’ - in comes the . 5, 187, anno , padat (lexical is a good epic att *‘ gehend ” anced the now otns (KZ. 10, ressly rejected ice, so that one des to padait-s 4'), we cannot ETYMOLOGIES AND DERIVATIONS 51 separate the two words, and Skr. paitis looks also to belong with them. When Bartholomae sweepingly declares that the division pad- + ati-s ‘iens’ is false, that is purely gratuitous on his part. Every objection to that analysis meets a ready answer. The root a*t- is alive in Sanskrit only: true, but it yielded before 2-t-: ¢y-. The IE. language probably had several words for ‘ going’ allocated to different sorts of goers, to different modes of motion. The unique survival of some of these is quite in the range of probability ; and so is their tautological combination as, say, in Skr. dt-ya-s ‘horse’ (§ 12). For the survival in composition of the root ct/ot, for this was its vocalism, I have brought a good deal of evidence in Classical Quarterly 3, 272, and I have collected a great deal more that I propose tu present in another paper. The objection that padati- is a tt- stem, Lat. pedei- a t- stem, needs no answer, but cf. lexical Skr. pddat. The objection that will weigh most is against the difference of quantity in -ef- and -4it-. 4. In arguing on this point, without denying that in paddti- we may have a prius pada (instrum.), I exclude it from consideration, and I shall write the vowel of the (supposed) posterius -of7s with 6 because that is the vowel appro- priate for the posterius. Also, though I believe that patizs is to be derived from pad-sthis: Skr. pada-sthas (see AJ Ph. 34, 39, § 95), I shall here consider it as derived from the root ef-. 5. Sanskrit compounds of the type of vathd-sith- ‘ curru-capax’ (sc. equus) are believed to contain ~ - ~ for ~ ~ ~ (see Wackernagel az. Gram. II. 1, § 56; cf. I. § 264 sq.), and it is current dogma to limit that lengthening to com- pounds of the type ~ —:~ and exclude it from compounds of the type ~ :— ~, words like Lith ap-emis ‘circumference’: ap-emimas ‘umnehmen,’ xuv-@7es ‘dog-eyed,’ Av. h>v-apah- ‘ bonum-sacrum-habens.’ In words of this class (cf. Lat. ambages) the long vowel is explained by — gradation, which is to put the effect before the cause. Thus Av. huv-dpah- is equated with Skr. (su-) das-, there being in the RV. a hapax dpas- over against a fairly common word apas- ‘ sacrum,’ quantitatively equal to Lat. Opus-, in the word operart, which means ‘sacruficare.’ There is no way to make people believe, I suppose, that the long-grade vocalism of the amb-dges type arose by rhythmic lengthening in compounds, as long as they prefer to believe that a long-grade vocalism arose —by long-grade vocalism and then found its place in compounds. For myself, I believe that Vedic dpas- is a sporadic use of an original composition-form with lengthened @- to suit metrical convenience. In quadrisyllabic forms like Vedic hart-saécas and Av. gatri-saé-o the posterius comes from sékw-es (n. pl.) ‘sequentes.’ Whether we call this type rhythmic (~ ~ = ~) or gradational is a matter of entire indifference when we extend the mere question of fact to the explanation of padd-tis, for the form *éd-ott-s is a perfectly legitimate con- struction from ped- (or pod-)+a derivative of the root ef- ‘ errare,’ and instead of a metrical lengthening we may reckon with weakening, and derive Skr. pattis from *ped(o)ii-s. 6. In Homeric aomdi-wrns ‘in scuto iens’ we have another military term 52 EDWIN W. FAY with posterius -dti- (-rns: Skr. ti-, cf. Seomorns: Skr. ddm-pati-s). The clear locative prius is also found in Homeric @ypou-wrns ‘in agro iens.’ Any reason that may be urged against taking pad-dti-s for a compound may be urged as validly against Sophocles’ 7rod-ayos ‘ attendant’:! Skr. pad-aji- (Panini), as to which see Wackernagel, I.c. II. 1. 62. In view of what has been said I believe no consistent reason can be advanced against taking Skr. pad-dtis as a com- pound, precisely as Bohtlingk and Roth did take it. Besides, we have a second Greek military word in otpati-wrns ‘ soldier,’ wherein I take the prius as a locative to a lost monosyllable strt-, whence orpat- in otpat-ayos. Besides the disyllable -ot#- in pad-dti- we have the root-noun of- in Lat. ped-et- (Bartholomae’s é in pedés, BB. 17, 114 fn., lacks, to the best of my knowledge, all metrical confirmation). Ili. Lat. equet-: Skr. dc¢va-pati-s. 47. The addiction of scholars for more than half a century to the equation of immotns with eques is most extraordinary. Thus (m7érns is an abstract, your ‘horseship ’—which seems incredible—and eque-t- is an abstract too, with -t@ (cf. senecta) reduced to -t. Now Ger. majestat, Ital. podesta, Eng. your highness, I understand ; and opera (1) auxilium, (2) auxiliator, I understand, but ‘ your horseshipb’—no. So as Greek deororns corresponds to Skr. dém-pati-s I feel constrained to believe that ¢ar7orns is from *imzo[7o |rns = Skr. dgva-pati-s ‘horse-lord ’ (VS.), also a proper name, cf.”Iw7ora Néotwp. 8. What would have become of proethnic *ékwo-potis in Italic? In certain of the dialects, assuming a stage before any sort of syncope existed, *ebo-potis, liable to haplologic shortening precisely as iw7orns. But in pre- Latin we should expect *equo-pfot(z)s, with possible loss of 7 as in hospes, if from *hosti-potis rather than *hostt-pa*-t ‘ stranger-protecting.’” How would qu- p- in a Latin word have been treated? Certainly p- gu- was assimilated to gu- qu- (see e.g. Walde?’, s.v. prope). If qu- p- (bear in mind a qu that may have gotten quite close to the # of the other dialects) was also assimilated to gu- qu-, then equet- from *equo-[quo|t(t)s is a simple case of haplology. What actually happened may have been quite different. As anceps is from ambi+caput and hospes from hosti+potis (or -pots), eques may represent something like *egu(o)- pot-s (disyllable) with assimilation (eventual loss) of after qu owing to the strong influence of eqguos ‘horse.’ Those who lay great weight on rigidity of equation symbols, formal regularity, will perhaps prefer to say that eques is for dialectic *epo[ po |ti-s affected by Latin equos. IV. Latin Cocles. g. The difficulty of accounting for Cocles as the borrowed word Ku«crkow seems to me insurmountable, but the two words are most closely cognate. To correspond with «v«dos, Skr. cakrd-m ‘ wheel, circle,’ we should expect 1 Herein mod- is quasi-prepositional: Aeolic in attendance on a chariot.’ wed-& ‘cum’; cf, pad- in Skr. pad-vatha-s ‘footman The clear Any reason be urged as Panini), as to said I believe is aS a com- , we have a ake the prius oT paT-ayos. n Lat. ped-et- y knowledge, the equation an abstract, ract too, with a, Eng. your - understand, kr. ddm-pati-s kr. dgva-patt-s Italic? In cope existed, But in pre- lospes, if from ould qu- p- in ted to qu- qu- y have gotten ) gu- qu-, then Vhat actually bi+caput and + like *egu(o)- owing to the on rigidity of lat eques is for word Ku«row sely cognate. hhould expect — ETYMOLOGIES AND DERIVATIONS 53 Latin quoglo- whence coclo-, and this, compounded with the monosyllabic stem oqu-, represented in the diminutive oculus, would yield (with elision of the stem -o) cocl0c-. The change from cocloc- to coclot- was due to avoidance of three successive c’s. V. Latin poples ‘knee, hamstring,’ etc. 10. I interpret popl-et- as dialectic from cocl’et-: coclo- ‘ wheel.’ Etymo- logically considered, the wheel was merely a ‘ goer,’ as we use ‘runner”™ to describe the sliding-iron of a sled. Latin coclo- designated, I take it, some part of the leg, and belongs semantically with Skr. cavdna-s=caritra-m (see § 14) ‘ foot, leg.’ Walde?, s.v. colo, cites O.Bulg. kolo ‘ wheel’: koléno ‘ knee.’ Here a special semantic closeness is shown by Ger. (knie-)scheibe, for the ‘patella’ or ‘knee-pan’ is a sort of disc or wheel. For the posterius of popl’-et we must compare xwA-n7- ‘the popliteal space ’=‘ knie-kehle,’ recalling that Prellwitz has already expressed the suspicion that -y7- belongs with Lat. apere ‘to bind,’ cf. Eng. ham-string, which is one sense of poples. This justifies us in deriving poplet- from k”ek”l’-ap (or -ap-), whence, with dissimula- tion of the third f, dialectic popl’et-. The startform of «®AnW may have been k™6l(0) + a*p-, with labial dissimilation of k”- from the following f-. But the influence of oxéXos is not to be excluded. VI. «cwrwrns ‘lizard’: lacerta. 11. The lizard, if one will examine the habits of the commoner kinds, is a limb-goer, putting his whole leg rather than his foot only upon the ground. Hence he is named from «@do-‘ leg’+-awrTns ‘ errans.? The same notandum in a-oxdnda-Bos= év-oxédet-Baivwv. By irradiation from cwAwTns we have a-cxada- Berns and yaXewrns. Similarly I explain lJacey-ta as syncopated (cf. Minerva, etc.) from an original laces- or lacer- ‘leg, limb’ [: O.Norse leggy, which Lidén (ap. Walde? s.v. lacerytus) has derived from a pre-Germanic lagiz-; or: Xéxpa-va' Tous ayxwvas (cf. Auep-titey: oxiptav)|+eta (as in p-erd, § 2) or 7-ta ‘iens,’ the whole again =‘ limb-going.’ The prius Jaces- | lacery- would be a suffixless locative (cf. Brugmann, Gr?. 2. 2 §§ 174-175 for parallels). For lacertus ‘arm,’ I write a startform Jacer|s|t(h)o-s (for [s] see Fay AJPh. 33, p. 383; and for the confix -sthos in names of parts of the body, 16. 34, p. 23, § 58 sq.), first of the out-standing lump of muscles onthe upper arm. Lat. -evtus may be, however, from rto- (or erto-): Skr. -r¢ in yékrt- ‘liver.’ Note that in the lacertus and oxéXos | x@Xov group a root /lék varies with s)kél-. VII. Sanskrit rivék (nom. sg.) ‘ sacerdos quidam.’ 12. The final k is of palatal provenance, and should have given -t. Wackernagel, at. Gram. I. 1, § 149 aa, remarks: ‘ Der guttural hielt sich hinter y (auch in rivik!), weil die nachbarschaft von r- laut und cerebral gescheut wurde.’ This is a truly curious effect for 7 to have at such a distance, 2 Cf. rpoxés * wheel '=7péxe: ‘runs.’ 54 EDWIN W. FAY when in rvaé and samrat ‘king,’ bhrat ‘gleam,’ v- has no effect. I rather think that in the sequence ¢ -k k was the result of preventive dissimilation in avoidance of the group -? -é. VIII. Avestan -Cardi-: xéXnr-. 13. Bartholomae defines the Avestan compound ravas-éarat- by ‘was sich im freien bewegt,’ i.e. ‘ ruri-cola,’ used in Ovid of an ox or a bull roaming at large (Met. 15, 1243 E1, 91), as ravas-carat- is used of wild animals. The equation of -Cardt- with Greek «éAnrt- ‘riding horse’ is perfect ; both being from the startform k”clét-, whence Greek «xéAnr-, with «- for r- by preventive dissimilation. As to its derivation, céX\-nT- is a tautological compound, quasi ‘ wander-roaming,’ with é in the oblique cases (~ ~ ~) for the reason that @ is long in padati- (§ 3). For tautological compounds of this type note should be made of the great variety of IE. verbs of motion, and for this particular com- pound of the fact that its separate elements were, as monosyllables, unstable (cf. Wackernagel, ap. JF. Anz. 24, 114). Skr. dt-ya-s ‘horse’ is also a com- pound, prius e/-, posterius -yo-s (: the root ez). IX. Sanskrit car-ttra-m, bhar-ctra-m ; @vperpa ‘ foris.’ 14. The usual division ts car-tram, but the conditions that allow of explaining car7-tram from car- allow as well of explaining *¢tram from 7- ‘ ire,’ and the complex car-¢iram (cf. car-ttd-m ‘itio’) as tautological. The tauto- logical principle of composition (see Fay, KZ. 45, 112; AJPh. 32, 408) exists and always has existed. Note an example like Germ. schell-hengst or English might-and-main. Given an old Greek *er-po- ‘door’ (: ef- ‘errare’ :: Lat. ia-nu-a ‘door’ [cf. Skr. yd-na-s ‘via’]: yd- ‘ ire’), @vp-erpa (Homer) becomes as transparent as schell-hengst. With car-ttram ‘leg’ cf. bhar-ttra-m ‘arm.’ In both of these alike old -2-tvo- ‘ goer, leg,’ whence ‘limb,’ may be contained. Then car-¢tram was the ‘ go-limb,’ and bhar-itram the ‘ carry-limb.’ In Greek hép-etpov ‘ bier’ -etpov will be still a ‘ goer,’ but in the sense of ‘ uehiculum’ or ‘go-cart,’ while gep- will indicate its use in carrying (cf. Eng. carriage). A collection of examples might show that the Greek words in -e-tpo- parallel with Skr. words in -ttra- all allow of explanation by irradiation from patterns in which the force of -er-po- and -t-tra- is still clear to sight. Note also Skr. car-isnuz- ‘se mouens, errans, instabilis’ with -zs22-: tsna@ti ‘ accelerat.’ X. Sanskrit divzt-, divztmant-: Latin divit-. 15. There are two competing definitions and analyses of divzt-. Grassmann defines as ‘splendour,’ but the Petersburg lexicon, followed by Wackernagel (at. Gram. 2, § 22a), as well as Ludwig and Griffith in their translations, analyze as div-it- ‘ad caelum iens.’ For Grassmann’s definition cf. Aufrecht in KZ. 2, 149, and Grassmann himself bid. 11. 9, though in his translation of RV. 10. 76. 6 (see below) he goes over to the other camp. In BB. 12, 114 fn. ai tl ather think milation in y ‘was sich roaming at mals. The both being r preventive ound, quasi on that @ is e should be icular com- es, unstable ilso a com- it allow of rom 7- ‘ ire,’ The tauto- , 408) exists or English ire’ :: Lat. sr) becomes farm.’ In . contained. In Greek uehiculum ’ riage). A ‘po- parallel ym patterns Note also celerat.’ Grassmann Vackernagel translations, cf. Aufrecht anslation of . 12, 114 fn. “eee ee nal ETYMOLOGIES AND DERIVATIONS 55 Bartholomae takes divét as a noun, but does not define it. Aufrecht compared divit- for its formation with harZt, a colour adjective =‘ bright ’—which does not keep the -t- of divzt- from being the ¢ of dydt- ‘splendour.’ In form divt-t- ‘bright’: di-divt-s ‘splendidus,’ is not unlike haré-t-: hari-s; cf. for the reduplication sarét- ‘flumen’ (cf. Eng. ‘ current’): sa-svi-s ‘currens’: sa-srz-t- ‘fluens,’ with further examples in Brugmann, Gr.* 2. 1, § 316. The analysis as div-ti- leaves div- hard to account for as a terminus ad quem, at least Sanskrit exhibits no parallel case with a prius div-=‘ ad caelum.’ 16. The words div#t- and divitmant are practically confined to RV., and divét- is a hapax. The examples are: 1. 26. 2 (Arnold’s A period) where Agni, invited to sit at the sacrifice, is addressed in the line ‘Agne divitmataé vacah’ (=vdcasé) and Grassmann’s rendering by ‘Agni durch unser glanzend lied’ is certainly not less apt than Griffith’s ‘through our heavenly word.’ 4. 31. 11 (Arnold’s A), mahé rayé divitmate. endem besitz des guts’; Griffith: ‘ for great celestial opulence.’ 5. 79. 1 (Arnold’s B*), twice repeated in SV., ‘mahé no adya bhodayatiso Grassmann: ‘zu grossem reichthum weck’ uns heut, o Grassmann: ‘zu glanz- rayé divitmati.’ morgenroth, zu glinzendem; Griffith: ‘O heavenly dawn, awaken us to ample opulence to-day.’ 10. 76. 6 (Arnold's B'), vaca divitd diviétmata. der zu dem himmel dringt,’ not substantially different from Griffith's ‘ with heavenly song that rushes up to heaven.’ Aufrecht (l.s.c.) had taken d. d. as ‘mit strahlendem glanze.’ I deviate from Aufrecht by rendering the words v. d. d. ‘uoce splendente splendidissima’ (see below), though there is a possi- bility of taking divétd d. as a very early instance of an instrum. qual. =uoce splendore splendidissimo. For the construction see Speyer, Skr. Synt. § 67, where we have the examples kosah svalpa-vyayena =aerarium parui-expenso and vividhaiy drumdih kdnanam ‘uariis arboribus siluam’ (see also Brugmann, Gr.2 2. 2. § 468; Delbriick, Gr.’ 3. § 106). Thus interpreted divtta divitmata is not unlike hAéritvaté varcasé ‘aureo splendore’ (instrum. of means) in RV. 10. 112. 3. Morphologically the ‘ possessive’ hdritvant- (hapax) would seem to be a mere irradiation from divét-mant- (with m in avoidance of the sequence v-v, cf. Schulze KZ. 39. 612; Bender, Suffixes -mant and -vanti, p- 39)-" 17. But diviét- is susceptible of defence as an adjective of the harit- type, while divéimant- ‘splendidus’ may be a blend, so to speak, of divit- ‘splendens * with dyu-mdnt ‘splendorem habens’ (cf. in Macdonnell’s Vedic Grammar S 246 fn. on the intrusion of d from ad-bhuta- ‘marvellous’ into an-atsbhuta- ‘unsurpassed ’). Grassmann: ‘mit schall may be as well to suppose that when nominalizedc in the possessive derivative a secondary accent of diacritic purport was given to the adj. stem harit (i.e. *hdrit- ‘ splendour’). 1 Bender's remarks on the general analogy that may have produced tdvasvant from taviis (p. 32 fn.) do not agree with his explanation of the accent of Adritvant- on p. 41, fn. 2, and *havituant- were as permissible as divétmant-. It 56 EDWIN W. FAY 18. In its usage divétmant once qualifies va@yé= Lat. ret (dat.) and once follows radyé, but is in agreement with wsas ‘Aurora.’ One wonders if here (5. 79. 1) divdtmatt, instead of being a nom. voc. is a dative in -ti, cf. ti (frequent), vitz (2°), see Macdonnell p. 232, a. Or is it a (masc.) dat.-loc. in -% of the type described, but not yet conclusively found (see, however, O.Bulg. 7- datives in Brugmann, Gy.” 2. 2. § 162. 1) by Bartholomae (Gr. Tran. Phil. 1. § 217)? It may rather be an emendation of the Vedic diaskeuast who, staggering at rd@y-é (usually feminine) with a masc. adjective divétmate, and ignorant of RV. 4. 31. 11, altered to divétmati, in construction with wsas. The remaining instances of divéimant agree with vaca ‘ voce’ or vdcas ‘ Férros’ (see Delbriick in KZ. 20, 227). That the sense of ‘splendidus’ is here admissible ought to be clear from Latin combinations like wox suauis et splendida, uerba splendida (Cicero), cf. also uerba luculentiora, scriptor lucu- lentus (Cicero), oratio luculenta (Sallust). That divétmant moves in the same semantic range as dyu-mdnt (see § 17) is manifest, for we have dyumdnt- with vayds ‘rei’ (gen. 1°), with the cognate word vayé-s (3°), and with their synonyms bhdgas (1°) and ¢dm (1°); also with vac-am ‘vocem’ (2°) and with vacas- (1°). 19. In view of the characteristic combination of dyumént | divitmant- with the words for ‘ wealth’ (ray- | ray¢-: Lat. v@s)—cf. also the parallelism of the adverbially used neuters vevdt || dyumdt in RV. 2. 9. 6; 5. 23. 4; 6. 48.7 (in every case qualifying the imperative d?-di-hi ‘ splende’)—I find myself quite unwilling to follow Schulze in KZ. 45. 190 and divorce Lat. diuet- ‘rich,’ diuitiae ‘ riches’ (with adj. luculentae in Plautus, cf. also luculenta hereditas ; in Cicero, secundas res splendidiores facit amicitia) from Vedic divit-. So far as the root goes, there is no issue to join with Schulze; but, in conformity with Skr. divit-, Lat. divet- must be explained from ‘shining’ and not from ‘sky,’ though if proethnic *diwit- had come into Latin with some of its original sense it may well have come under the influence of diuum (cf. sub dio ‘ under the bright sky’ with &-dtos ‘ meridianus’). What came into Latin may have been diw- (or deiw-)et, cf. et in apyeér- ‘white.’ XI. O.Bulg. bogatz ‘ Fortunatus.’ 20. Nor can I admit Schulze’s parallel instance, after Schroeder as cited, of O.Bulg. bogat# ‘ Fortunatus,’ and formed morphologically like Fortunda-tus, but its prius is to be correlated with Av. bayd- ‘ pars, portio’: Skr. bhdga- ‘share, wohlstand,’ and the connotation of ‘ fortuna’ or even ‘ diuitiae’ is clear from the hapax compound bhdga-bhaktasya (rayds)=quasi ‘ fortuna-diuisae (rei),’ cf. also bhdgavant ‘ gut-begabt.’ In Indo-Iranian bhdga is the Dis- penser (God), but also the dispensed (‘wealth’). In O.Bulg. we have bogu ‘god,’ but when Schulze writes ‘der arme heisst bei den Slaven “‘ von den géttern verlassen’”’; ubog#- und nebogii-’ we may admit that this is now the popular interpretation, but surely we must start with ‘ dyupopos, aporpos, at.) and once ynders if here in -ti, cf. ati asc.) dat.-loc. see, however, holomae (Gyr. dic diaskeuast tive divitmate, on with wsas. vdcas ‘ Férros’ idus’ is here uox suas et scriptoy lucu- ; in the same dyumdnt- with id with their (2°) and with tvitmant- with llelism of the 1; 6. 48. 7 (in _ myself quite diuet- ‘rich,’ hereditas ; in t-. So far as nformity with ot from ‘sky,’ of its original sub dio ‘ under itin may have yeder as cited, ce Fortuna-tus, : Skr. bhdga- litiae’ is clear ortuna-diuisae . is the Dis- ulg. we have Slaven ‘‘ von it this is now LOPOS, apoLpos, I OR ETYMOLOGIES AND DERIVATIONS 57 expers’ (as found, with some redundance, in Terence, Heaut. 652, ne expers partis esset de nostris bonis). Rather is bogat# like Latin nummatus. XII. Latin gurg-et- ‘ eddy.’ 21. In this complex I define gurg- by ‘ swallowing,’ a broken reduplication form, however we explain the abnormality of g, to the root of uorat. The combination gurg-et- meant ‘swallow-going,’ no bad description of an eddy for one who has seen the swirl and down-suck of it. The sense of ‘eddy’ is found in a derivative of ef, precisely in the word eddy: O.Norse ida. The combination of the notions ‘errare’ and ‘circulare’ is found in Eng. @roll : stroll (cf. Skeat or the Oxford Dictionary), and is as old as the time when the planets (‘ stellae errantes’) were observed to go round the earth, or when a ‘wheel’ was a ‘goer’ (see § 10). It is a fact, or a superstition, of wood-craft that a man when lost always wanders round and round in a circle. XIII. Sanskrit send@-ni: xoipa-vos. 22. The correspondence of send-ni ‘ general’ with Germ. her-zog cannot have escaped the notice of anybody whose attention has once been drawn to send-nt. Their Homeric synonym is xoipavos, and -vos=‘dux.’ The corre- lation of -vos: Skr. ni- ‘ducere’ is the same as the correlation of -¢d-s in girt-cds ‘in monte cubans’: -¢i- in syona-gi-s ‘in-molli-(lecto-) cubans.’ The shortening of the a in xoipa-: Germ. heer will have taken place as in potpa. Whether the shortening demonstrably took place before the composition Is a question of chronology I know not how to determine. XIV. Sanskrit syond- ‘ soft’: edv7 ‘ bed.’ 23. The aspirate in xaO-evdw ‘I sleep’ gives some warrant for supposing evvn to be for *eivy: syond- in the Sanskrit compound just cited. It is curious that the derivation of syond- has not been hitherto realized. It will have meant ‘pliant’ to start with, and belong with sy#man- ‘band’: the root siv- syti- ‘ plectere’ (earlier sense than ‘suere’). In -ed-dev the complex will have meant ‘ (se) lecto dare,’ and original *syéu- may have been a locative to a noun stem *syu-. In general, cf. déuxov ‘bed, bedding’: the root of Lat. domare ‘binden’ (earlier than ‘ bandigen,’ and also found in Lat. domus, the ‘wattled house’). In Wort. u. Sach. 3. 52 sq. Meringer would unite Lat. texit ‘ weaves’ with réxtwy ‘carpenter’ by defining their root as ‘ flectere,’ and I think this is right. The sphere in which syond- moves is seen In syona casi susada (VS.) which Bloomfield cites in AJPh. 33. 9 with a TB. variant. His rendering is ‘ pleasant art thou and a fair seat to rest on.’ XV. Vedic kridd-vasu- ‘ diuidens-bona.’ 24. The Rigveda has viddd-vasu =‘ inueniens-bona’ four times (three times in Arnold’s A hymns, once in his B*) and the hapax abhardd-vasu- ‘afferens-bona’ 58 EDWIN W. FAY (5. 79. 1), both perfectly clear compounds. The hapax ¢atdd-vasu- (1. 119. 1 = B?) will certainly be for *gatd-vasu-, as it follows another compound with sahdasra- ‘thousand.’ For krtad-vasu- (8. 31. 9), it would seem that we might satisfy ourselves with the plain etymological interpretation of the padas ab: vitchotra krtadvasé (% with the pada-text) | dacasyanta ’mrtaya kam= inviting-to-a-feast, cutting-goodies, bestowing verily on the [mmortal. With -vasz in the compound cf. bon-bons and goodies. XVI. Vedic pratad-vasi (RV. 8. 13. 27). 25- In this stanza Indra is supplicated to yoke his sorrels, his boon- companions (sadham@dya), his pratadvasté (dual) and rattle-along (abhé-svara) to the soma-drinking. In such a context we may credit the composer of the hymn, that ancient poet of wine, with looking upon the sorrels who are to fetch [Indra to the convivium as sharing in the extraction or preparation of that boon. The word preparation brings us to ask whether pratéd may have to do with Lat. parare, but its -t- would then have to be explained. The word extraction brings us to the root fret- as found in Lat. presst (see Fay, KZ. 43, 351), and to Skr. prt- ‘battle,’ i.e. ‘press’ (of battle). Note also prt-sutz-s ‘(proeli) impetus,’ where -sutz-s comes either from the root su- ‘ premere * or from su- ‘impel, hurl,’ not certainly different roots. The assumption precedent is that the root pret- was used to describe the extraction of sap or milk by pressure (or by boring). Here, in using pratdéd-, the poet may have had in mind the spirting, splashing, spattering of the drops as the soma- stalks were a-crushing and the like spatter of mud by the racing horses. One asks oneself whether ultimately pratdd is not ‘dashing’ (trans.), cognate with O.Norse spretta ‘ salire’ (intrans.), cf. Eng. sprints. I assume here no violation of Grimm’s law. In O.Norse spretia tt may have come from proethnic -in- as well as from a prior Norse sprenta, and the fixation of ¢ in this or some like formation may have analytically influenced the mt group in the sept of sprints. 26. For the relation of Skr. prt- ‘ battle’: Latin prem-zt, pf. pressit, note Dbartholomae’s third definition (Wtbch. 869) of parat- ‘jemandem (acc.) feindlich entgegen, zu leibe gehen’ (= Lat. aliquem petere), a sense not alien to Lat. premit ‘presses, chases, pursues’ (game, the enemy). Semantically, O.Norse s)pretia=‘salire,’ Av. parat-=‘insilire’ (in the late Latin sense of ‘adoriri’). Should we start with a compound per-ef- ‘ praeire ; adoriri’—or a tautological compound, with cumulative force, of the root per- ‘ to pierce, fare,’ and the root et- ‘to wander, run’ (cf. on Latin interpret- § 34)—the sense of ‘ squeeze, press,’ would have developed as a description of the concomitant or sequel action involved in ‘ piercing.’ XVII. Avestan fratat-éarata- ‘ pro-fluens, rapidus.’ 27. Passing over Bartholomae’s most complicated explanation as fra- tatéara-ta, and holding to the traditional division as a compound, I would explain fratat- as an adverbially used neuter of the present participle = prae- aa ., I1g. 1= B?) with sahasra- night satisfy ab: vitthotra e household- erily on the is, his boon- abht-svara) to poser of the » are to fetch yaration of | pratéd may lained. The esst (see Fay, Note also su- ‘premere ° - assumption on of sap or et may have as the soma- 1orses. One cognate with -no violation ethnic -in- as or some like ot of sprints. . pressit, note indem (acc.) ise not alien semantically, itin sense of jJoriri’—or a pierce, fare,’ -the sense of ncomitant or ition as fra- nd, I would ‘iciple = prae- a ee ETYMOLOGIES AND DERIVATIONS 549 currenter (from pro-+et- ‘currere’) combined with a ¢#- derivative of the root “kar (Bartholomae, 449) ‘uersari,’ viz.: cava-t- (for so in three of the four citations we find the stem): ¢ar-dt- (§ 13). Skr. at- is actually defined by ‘laufen ’ (PW.). XVIII. [pwrect-raogs ‘ qui praecurrit populum.’ 28. Every IE. name may be significant, a tell-tale name. The name of Protesilaos, if I have properly divined its meaning, tells us all about him that we learn from Homer, dro@pa@cKxovta rorv tpwTticTtov "Ayatov (B. 698 sq.). The quantity agreement of [Ipwreai[-Naos] and Av. fratat- raises the question whether @ represents an IE. contraction of o+e. There is very little evidence to settle this question (cf. Brugmann, Gr.? I. § g64), and it would require compounds that had lost touch with their primitives to remove the suspicion that the junction vowel owed its colour to the semantic, not the phonetic, preponderance of whichever vowel prevailed. In Ilpwreot-, though com- pounded of pro+eteti, w may be due to wparos ‘ first’ (? from pro + ofos ‘ goer’; then Doric mparos has secondary 4, cf. the adverb wpav and Latin pra-uos ‘ forward, pert).’ XIX. Vedic anv-ditat (A V. 18. 3. 40) ‘ consecutus est.’ 29. This is a blend of dit, impf. 3d. sg. of the root 7- ‘ire’ and dat, the corresponding form from the root at- ‘errare, currere.’ For the previous discussion of the form see Lanman’s note to Whitney’s AV. (Harv. Or. Series, 8, 860). XX. Avestan patarata- (- — ~) ‘ flying.’ 30. This, like Lat. lacerta (§ 11), is from a compound pa*ta’r-cta (or -tta) ‘in-penna-iens.’ Under what proethnic or pre-Avestan conditions the quadri- syllable (- ~ ~ ~) was reduced to a trisyllable (- — ~) cannot be made out. XXI. Hrop ‘heart’: Skr. dt-man- ‘ breath.’ 31. These words and their cognates (see e.g. Prellwitz*, s.vv. #rop and atmos) may be reasonably derived from the root eft-. The heart, bosom, lungs, stomach (abdomen) are all in perpetual motion, a fact much more in ocular evidence in early savagery than now. According to Fick-Stokes*, p. 29, O.Ir. i2-athar is from én-dtoro-. This would mean the ‘ in-goer.’ XXII. Old Welsh -znteredou. 32. The compound permed-interedou (glossed by ‘ilia’) will have come, in its posterius, from enter<->eto- (Fick-Stokes‘, p. 30). Is -eto- itself a posterius =‘ iens,’ and the whole a newer, but still IE., formation to make more precise en-tevo- (in év-repov) after -tevo- was losing its force of ‘ going’ (see lay, AJPh. 31, 407)? The @ of Skr. dntra- will be secondary. ,ETYMOLOGIES AND DERIVATIONS XXIII. Sanskrit k#-t-van- ‘machend, bewirkend.’ 33. I have recently noted (AJPh. 34.16, § 44 fn.) that the agent suffix -Zer- (cf. for its origin Prellwitz?, s.v. teipw) gives evidence in such derivatives as Lat. cul-tur-a nattiva of a by-form -tver- | -tvor- ‘ faciens’: O.Bulg. tvoriti ‘facere.’ If we had Skr. ky-tvar- it would be tautological, quasi ‘ make- maker,’ and ky-tvan- may have replaced the start-form *kriwor- with r- -n for r- -y, cf. the normal feminine k7-tvari- ‘magical,’ noting that po-tvori means ‘magic’ (cf. Jevons in Anthropology and the Classics, p. 98). The different treatment of kytvan- and krtvarvi- would be due to syllabification. Any root with vy would present the conditions for dissimilation. ‘The dis- similation was doubtless proethnic, for the feminine variation is found in the type of wiwv fem. zieipa, where the -twen | -twevt type had been imitated by the -wen- type. Perhaps in a word like Skr. *k-van- we have the reduction of *rk-tvay=‘ carmen-faciens,’ but a syncretic origin for -w-en- is also to be admitted. XXIV. atv ‘interpres, nuntius.’ 34. It is now (see Herwerden’s Lex Suppl. s.v.) not open to reasonable doubt that aty7jv is a genuine word. We may take its a- for an 2-, a reduced stage of a long grade word in ét-, cf. Prellwitz®? s.v. a@tuos ‘vapour.’ An aTunv was, in our modern parlance, a ‘runner,’ as was an tuter-py-el- (on pr-et- see Fay, Cl. Quart. 3. 272). XXV. Homeric és ‘ entgegen-kommend, sociable.’ 35- As a concluding testimony for the root et ‘errare, currere, ire,’ so often mentioned above, I cite éwnrns with its abstract é€znrvs. The sense of ém-ntes- (with protracted é as in § 5) is roughly ‘ approachable, accessible,’ terms denoting a passive sort of friendliness. But the original sense (see on active-passive reciprocals Fay in IF’. 29, 418) was rather active=‘ approaching, sociable, entgegenkommend.’ Comparable formations are Skr. vé-vacas— ‘variously speaking,’ sé-va@cas—‘ similarly speaking.’ ‘The accent of the Greek words is secondary, say after the synonymous pair oapiotyns (=apud socium stans, see Fay, AJ Ph. 34. 33) captors. Epwin W. Fay. AusTIN, TEXAS, gent suffix -/er- derivatives as O.Bulg. tvoriti quasi ‘ make- ov- with r- -n that po-tvorz p. 98). The syllabification. ion. The dis- is found in the sn imitated by the reduction - is also to be to reasonable 1 a-, a reduced ‘vapour.’ An inter-py-el- (on 9 cS. Irrere, ire,’ so The sense of le, accessible,’ sense (see on ‘approaching, kr. vé-vacas— t of the Greek =apud socium IN W. Fay. | | | | SUMMARIES OF PERIODICALS. LITERATURE AND GENERAL. American Journal of Philology. Vol. 34. No.3. 1913. Emory B. Lease, Neve and neque with the Imperative and Subjunctive. Summa- rizes with statistics the usage of these particles to express prohibitions in the prose and verse literature. Rarer varieties, et ne, ac ne, etc., are also dealt with. Roy C. Flickinger, The Accusative of Exclamation in Epistolary Latin. C. F. W.Miiller’s canon, that o must be inserted in Cicero’s letters to Atticus with names of things, is disproved. The idiom (0 continually gaining) becomes more rigid as we move from Plautus to Pliny. H.C. Hoskier, The Lost Commentary of Oecumenius on the Apocalypse. Roland G. Kent, Again Lucilins on El andl. Areply to Fay. Tenney Frank, Marginalia (a) Hor. Epod. 2. 26 (queruntuy of birds means ‘sing’; (b) Cic. ad Att. VII 2. 7; (c) Livy ap. Sen. Suas. VI 22; (d) Ennius Med. 259-61 (V.); (e) Cic. Verr. IV 163 (would read aequabilitey for aequaliter). Charles Knapp, De quibusdam locts primi Horati Sermonis. Reviews: Drerup’s Iliad V (W. K. Prentice), Mooney’s Argonautica of Apollonius Rhodius (E. Fitch). Athenaeum (Pavia). Vol. 1., Part IV. 1913. A. Calderini, The Epigrammi Cyzici 1 relation to tragedy. C. agrees substantially with Meyer of Ko6nigsberg that the Epigrams were written by a scholar in the fifth or sixth century a.p., describing certain bas-reliefs on the temple of Apollo at Cyzicus [I Cent. b.c.), the subjects for which were suggested by a pre-Alexandrine ‘ manual,’ drawing on Greek tragedies. The Efigrammi are valuable for the reconstruction of lost tragedies. Carlo Pascal, Ennius in the Middle Ages. Ennius was famed as scientist and poet, but no one later than Isidorus had any direct knowledge of him, and Cicero and Claudius were taken for the authority on his merits. M. Valgimigli gives notes on various passages in the recently discovered Satyric drama of Sophocles, discusses the question whether the coryphaeus and Silenus are identical personages in the drama, and quotes certain passages where the Chorus and Silenus seem to be acting or speaking independently. J. Procacci, Short notes on some passages of the Panegyrict Latin. P. comments on emendations by W. Baehrens and others, and adds some of hisown. V. Constanzi, Zeus Thaulios. Near Pharsalus an altar was found inscribed Atos @avAcov, an epithet only found in two other inscriptions. Hesychius’ lexicon gives Oatvyos 7) OatAos as a designation of Ares among the Macedonians. Constanzi discusses the views of Hoffmann and others, and considers that the forms with @avA- are Atticized out of Macedonian *AavdAos just as Hesychius gives us favOcxd, £av0~ = Maced. favdixd, favdp. He accepts Solmsen’s view placing Thessalian @avA-, Macedonian éavA- alongside of the Phrygian-Lydian xavéavAas= kuvayxa, and Phryg.-Lyd. ddos, ‘the wolf’ as ‘strangler,’ and @avAwy given as the Sovpoves among the Athenians. *OavAos among the Macedonians, meaning ‘ slayer,’ or even being the wolf himself, was then attached to Ares, thus taking us back to the ages of theriomorphism and of the common home of a race still undivided, dwelling in the valley of the Axius or Aliacmon. SUMMARIES OF PERIODICALS Berliner philologische Wochenschrift. 1913. Aug. 2. QO. Viedebantt, Quaestiones Eptphantanae metrologicae et criticae (Tittel). Throws some new light on ancient weights and measures. J. Paulson, Index Lucre- tianus (Kraemer). Trustworthy. H. Kiepert et Ch. Huelsen, Forvmae urbis Romae Antiquae (Graffunder). In the new edition one ‘ Hauptkarte’ and four ‘ Nebenkarten ’ have been added. Aug. g. A. Maidhof, Zur Begriffsbestimmung dey Koine besonders auf Grund des Attizisten Mois (Schwyzer).