Chapter 34

Theaut ... autimmediately below is very much against this conjecture, which however Krüger (3rd ed.) has received into the text: we should expect rathernescio an illa quisquam, ornullus poeta, or keepingillaas nominativenescio an illa poeta ullo. Quintilian’s use ofnescio an(like that of post-Augustan writers generally) is vague: it is usually an expression of doubt, theanmeaning either ‘whether,’ or ‘whether not’ indifferently. Cp. ix. 4. 1: vi. 3. 6: viii. 6. 22: xii. 10. 2: i. 7. 24. (Mayor cites also Plin. Ep. i. 14. 9: iii. 1. 1: iv. 2. 1: v. 3. 7: vi. 21. 3: vii. 10. 3: 19. 4: viii. 16. 3: ix. 2. 5; and adds ‘In all these instancesnescio an(dubito an) is ‘I doubt whether’; in Cicero the meaning is always ‘I rather think.’’)Andresen proposednescio an ulla poeseos pars. The passage closely resembles§28, and must be emended on the same lines.§66.tragoedias. Thurot (Revue de Phil. 1880, iv. 1, p. 24) conjecturedtragoediam: cp.§67hoc opus. He is followed by Dosson, against all MS. authority. Becher points out that we must supply withhoc opusin§67the words ‘tragoedias in lucem proferendi,’ so thatopusandtragoediassquare well enough with each other.§68.quod ipsum reprehendunt, Meister, Krüger (3rd ed.) and Becher. This reading also occurs in the Codex Dorvilianus. Other readings arequod ipsum quodGHT Burn. 243, Bodl.:quo ipsumMS Harl. 2662, 4995, 4950, Ball. Halm conjecturedquem ipsum quoque, and was followed by Mayor and Hild. But as no fault has been found with Euripides in the foregoing,quoqueseems out of place.Founding on the reading of GHT, &c., also on that of F (which givesquod ipsum qui) Kiderlin (Hermes 23, p. 165) proposes to readquod ipsum quidam, comparing§98, where forquem senes quem(GT) Spalding rightly conjecturedquem senes quidem, and7, §21, where Bn, Bg givequodforquosdam. He then goes on, in an interesting paper, to reconstruct the whole passage, which is open to suspicion, especially in respect thatsublimiorstands as predicate withgravitasandcothurnus, as well as withsonus. The admirers of Sophocles consider his elevation of tone more appropriate than the strain of Euripides.Sublimioris therefore perhapsnotthe predicate of the sentence, however suitable it may be as the attribute ofsonus. The predicate may have dropped out, andsublimiormay have been transferred from its real place to supply it. It is striking that GFTM (also H and Bodl.) all givesublimior erit. Kiderlin imagines that a copyist who missed the predicate wrote in the margin ‘sublimior erit ponendumpost esse’: and then another insertedsublimior eritafteressein the text. For the predicate,magis accommodatusmight stand: in copying, the eye may have wandered frommagis accommodatustomagis accedit: formagis accomm.cp. ii. 5. 18 and x. 1. 79. Kiderlin therefore boldly proposes to make the parenthesis run, ‘quod ipsum quidam reprehendunt quibus gravitas et cothurnus et sublimior sonus Sophocli videtur esse magis accommodatus’: ‘was gerade manche tadeln, welchen das Würdevolle, der Kothurnus, und der erhabenere Ton des Sophokles angemessener zu sein scheint.’et dicendo ac respondendo7231, 7696:dicendo ac respondoGH:in dicendo et in respondendoPrat. Put. S (et respondendoM).praecipuus. Hunc admiratus maxime est. This is Meister’s reading, except that foreumI give (with Prat. Put. 7231, 7696 Harl. 2662 and 4995)hunc, which is commoner in Quint. at the beginning of a sentence (§§46, 78, 91, 112). The following are the readings of the MSS.: GHpraecipuus et admiratus miratus: M Bodl. Harl. 4950, 4829, Burn. 244, C, Burn. 243 Ball. Dorv.praecipuus et admirandus: Spraecipuum. Nunc admiratus et: Prat. Put. Harl. 2262 and 11671praecipuus hunc admiratus et maxime est ut saepe test. et sec. quamvis: Harl. 4995,hunc admiratus max. ut s. test. et eum secutus quamquam. Halm givespraecipuus est. Admiratus maxime est: Kiderlin insists on theestafterpraecipuus, to correspond withaccedit, though it seems better to take all that comes afteracceditas an explanation of the statementmagis accedit oratorio generi: he also retains theetof most MSS. and readspraecipuus est. hunc et admiratus(Blätter f. d. bayer. Gymn. 24, p. 84). Wölfflin (partly followed by Krüger 3rd ed.) proposed a more radical change (Rhein. Mus. 1887, 2 H. p. 313)praecipuus. Hunc imitatus, quoting in support of the conjunctionimitatus ... secutus§122, eos iuvenum imitatur et sequitur industria:5 §19, deligat quem sequatur, quem imitetur: Ovid, Fasti v. 157, ne non imitata maritum esset et ex omni parte secuta virum. But Kiderlin (l.c.) aptly remarks that if Quintilian had writtenimitatus, he would not have saidut saepe testaturbutut ex multis locis patet(apparet, videmus): while vii. 4. 17 (on which Wölfflin relies) is not really to the point. Moreover Quintilian, would never have separated such synonyms asimitatusandsecutusbyut saepe testatur.Charisi nomini addicuntur, Frotscher:Charis in homine adducturaGH:Charisii nomine edunturPrat. Put. 7231, 7696 Harl. 2662 Dorv.§70.aut illa iudiciaPrat. Put. 7231, 7696 Harl. 4995. GH Harl. 4950 giveaut illa mala iudicia: Bodl. Burn. 243aut alia mala iud.S Harl. 2662 Dorv. and Ball.aut alia iudicia.The edd., following Gesner, have generally given (with Harl. 4950)aut illa mala iudicia(so Halm and Meister), and have takenmalaas predicate, though the order of the words makes that impossible. Becher approves of Andresen’s deletion ofmala. Krüger (3rd ed.) printsmala [illa] iudicia, thinking thatillaarose by dittography, and that then the order was changed in the codd. toilla mala iudicia. Kiderlin (in Hermes 23) gives as an alternative to deletingmalathe conjectureilla simulata iudicia(‘jene erdichteten nachgemachten Gerichtsverhandlungen’; cp. xi. 1. 56: cum etiam hoc genus simulari litium soleat). A similar mutilation occurs, e.g., xi. 1. 20, where b givessecumMsecusinstead ofconsecutum.§71.filiorum militum, most codd.:filiorum maritorum militumPrat. Put. 7231, 7696 S.§72.si cum venia leguntur. The reading of the MSS. is upheld by Iwan Müller, Meister, and Kiderlin. Spalding suggestedcum verecundia: Schöllcum iudicio: Bechercum ingenio. Becher points out (Bursians Jahresb. 1887) that the expression is meant to coverdecerpereas well aslegere, anddecerpereindicates careful and intelligent reading (cp.§69,diligenterlectus):cum ingenio= ‘mit Verstand’: cp. Cic. ad Fam. xiii. 10. 2 quod versabatur in hoc studio nostro .. et cum ingenio .. nec sine industria: Ulp. Dig. 1. 16. 9 patientem esse proconsulem oportet, sed cumingenio, ne contemptibilis videatur. Finally, Krüger (3rd ed.) proposescum acumineorcum vigilantia(cp. v. 7. 10).—Prat. Put. 7231, 7696 S Harl. 2662 all give Osann’s conjecturelegantur.praveGH Harl. 4995, 4950 Burn. 243 Bodl.:pravisRegius, Halm, Meister, Becher draws attention to the parallelism between the clauses:ut prave praelatus est sui temporis iudiciis, ita merito creditur(= meruit credi)secundus consensu omnium.§76.nec quod desit ... nec quod redundet: H Burn. 243 and Bodl. givequod .. quod: Prat. Put. MS Harl. 2662, 4995, 4950, Burn. 244, Dorv. C, and Ball,quid .. quid. The latter reading is supported by Becher (Phil. Rund. iii. 434). Forquodcp. xii. 10. 46: (xii. 1. 20 where forquod adhucBM givequid adhuc): on the other hand, in vi. 3. 5 the MSS. are in favour ofquid, though Halm readsquod(followed by Meister). Forquidcp. Cic. pro Quint. §41, neque praeterea quid possis dicere invenio.§77.grandiori similis. So all MSS.: Halm and Meister. Several conjectural emendations have been put forward. Comparing2 §16(fiunt pro grandibus tumidi), Becher suggestsgrandi oratori,—an easy change, if the copyist used contractions, but without point: above in§74, ‘oratori magis similis’ is appropriate enough in speaking ofhistorians, but ‘oratori’ would be inappropriate here. This is accepted, however, by Hirt (Berl. Jahr. ix., 1883, p. 312; cp. P. Hirt, Subst. des Adjectivums, p. 12). Schöll proposes to readgladiatorisimilis, in view of the close connection with what follows, strictus ... carnis ... lacertorum: butpleniorandmagis fususare a bad introduction togladiatori, and if Aeschines hadplus carnisandminus lacertorum, he cannot really have resembled a gladiator. This reading is, however, adopted by Krüger (3rd ed.). Finally, Kiderlin (Hermes 23, p. 166 sq.) has conjecturedet grandi(orgrandiori)organo similis, and applies the figure throughout: ‘voller und breiter lässt Aeschines den Ton hervorströmen, einem grossen Musikinstrumente gleich’: ‘einer Orgel gleich,’—he isgrandisonus. The translation appears to limit unnecessarily the meaning ofplenusandfusus: though the former is used of tone i. 11. 6 (cp. xi. 3. 15 of the voice: ib. §§42, 62: and §55 of the breath): whilefususis used of the voice xi. 3. 64. For such a use ofgrandiscp.§58(cenae):§88(robora): xi. 2. 12 (convivium): 3. 15 (vox): 68 (speculum): and fororganum, i. 10. 25: ix. 4. 10: xi. 3. 20 (where there is a comparison between the throat and a musical instrument): probably also i. 2. 30. There is an antithesis in the two parts of the sentence between fulness and breadth, on the one hand, and real strength on the other; and for the transition to the second figure Kiderlin compares§33.§78.nihil enim est inane: perhaps ‘nihil enim estin eoinane’ (Becher), ornihil enim inest.§79.honesti studiosus. Becher’s proposal to alter the punctuation of this passage is discussed in the notead loc.—Forauditoriisandcompararat, see ontenuia atque quae§44, above.§80.quem tamen. Kiderlin, in Hermes (23, p. 168), raises a difficulty here.Tamenshows that the clause cannot go with the main statement (fateor), and its position forbids us to take it with thequamquam is primumclause: it can only go withquod ultimus est, &c., ‘though Demosthenes isultimus fere, &c.,yetCicero, &c.’ To prevent so awkward a joining of the clauses, Kiderlin proposes to readeumque tamen: pointing out that thequaeof the MSS. (GH) may have arisen out ofque, and that Quintilian may have writteneumque; cp. vi. 2. 13, where Halm makesutqueout ofquae(G), and xi. 2. 32, where Meister readsestque. The meaning will then be: Demetrius is worthy of record as being about the last, &c., and yet Cicero gives him the first place in themedium genus.—It seems better, however, to givetamena general reference: ‘yet, in spite of all that can be said on the other side’ (e.g., inclinasse eloquentiam dicitur). Cp.§99quae tamen sunt in hoc genere elegantissima.§81.prosam(prorsam)orationem etall MSS.; Halm, Meister, Krüger (3rd ed.) omitet. I find that Becher supports the view stated in the notead loc.: he would however writeprorsam, which the best MSS. give also in Plin. v. 31, 112 D.quodam Delphici videatur oraculo dei instinctus: so Frotscher, followed by Krüger (3rd ed.). On the other hand Claussen (Quaest. Quint., p. 356) and Wölfflin (followed now by Meister, pref. to ed. of Book x., p. 13) propose to deleteDelphici, of which Becher also approves. But the MS. evidence cannot be disregarded. The following are the various readings: GHquaedam Delphico videatur oraculo de instrictus, and so FT, the former giving also (by a later hand)de instinctus, the latterdei instructus. Bodl. givesquodam delphico videatur oraculo dei instructus. The most frequent reading is that of Prat. Put. 7231, 7696 Harl. 2662, 4995, 4829, 11671, Ball. and most edd.,quodam delphico videatur oraculo instinctus: S agrees, but is reported to havedelphicoafteroraculo: Harl. 4950 and Burn. 244 have the same reading, withinstitutuscorr. toinstinctus: Burn. 243 givesinstructus.Delphicowas originally deleted by Caesar: Phil xiii, p. 758. Halm readtamquam Delphico videatur oraculo instinctus: but Quintilian would take no trouble to avoid the repetition ofquidam(cp. divina quadam, above).—For the arrangement of words, Krüger (3rd ed.) compares§41qui ne minima quidem alicuius certe fiducia partis memoriam posteritatis speraverit.§82.quandam persuadendi deam. Nettleship (Journ. of Philol., xxix, p. 22) conjecturesSuadam[persuadendi deam], comparing Brutus, §59, quotedad loc. Persuadendi deamwould thus become a gloss onSuadam: but the expression in the text is quite in Quintilian’s style.§83.eloquendi suavitate:eloquendi usus(orusu)suav.GH and all codd. except Harl. 4950, and Dorv., both of which give simplyeloq. suav.Halm admitted into his text Geel’s conj. forusus, ‘eloquendivi acsuavitate,’ and this has met with some acceptance (Iwan Müller and Becher). But the parallel from Dion. Hal.,Ἀρχ. κρ.4 is hardly conclusive:τῆς τε περὶ ἑρμηνείαν δεινότητος ... καὶ τοῦ ἡδέος. Hirt properly remarks that the agreement between the two is not so great as to allow of correcting the one by the other. Kiderlin conjectureseloquendi vi,suavitate,perspicuitate.tam est loquendi. See notead loc.for Kiderlin’s conj.tam manifestus est. Though Meister’stam est eloquendiis probably a misprint, it is found in some MSS.—Harl. 4950: Burn. 244.§84.sane non affectaverunt. Bodl. and Vall. (verusubpunctuated in the latter:affectantPrat. Put. 7231 MS Ball. Dorv. Harl. 2662, 4995, 4829, 11671:sene non adfectitacueruntGH Burn. 243:adfectarunt7696:adfectitantHarl. 4950, and so Burn. 244 (corrected fromaffectant).§85.haud dubie proximus. Halm insertedeiafterdubie, though it is not found in any MS.: Regius had suggestedilli. Kiderlin (Hermes 23, p. 170) points out that ifpropiores aliiin§88is allowed to stand without a dative,eiis not necessary here. He suggests, however,illibeforealiiin§88: both passages must be dealt with in the same way.—Forhaud(Vall.), GHS haveaut: Mhaut. Cp. on3 §26.§86.ut illi ... cesserimus:cum illiGHFT Harl. 4995 Burn. 243:ut illiPrat. Put. 7231, 7696: and so S Harl. 4950 (withcaelesti atque divinae):ut illeM Harl. 2662. Kiderlin (Hermes, p. 170) proposes to go back to the reading of the older MSS.cum illi, and instead ofcesserimusto readcesserit, so as to make Vergil the subject throughout.Cumcannot, he contends, be a copyist’s error, motived byita; and it is probable, therefore, that at firstcesserit awas inadvertently written forcesserit; then (in G or some older MS.)cesserimus itawas made out of that, to correspond withvincimurbelow: and then in the later MSS.cumwas changed tout, because ofita. For the transition, with this reading, from cesserit to the plural (vincimur, pensamus), hecompares§107, where, after speaking of Demosthenes and Cicero, Quintilian passes tovincimus.§87.sequenturMS Halm and Meister:sequenterGseq̅nt’H:sequunturPrat. Put. 7231, 7696.φράσινid est. These words are omitted in the Pratensis, which is Étienne de Rouen’s abridgement of theBeccensis, now lost. This is an additional proof thatφράσινwas originally written in Greek: cp. on§42.§88.propioresH Prat. Put. Vall. Harl. 2662, 4495, 11671, Burn. 243. Bodl., Halm:proprioresGMS 7231, 7696, Harl. 4950, C, Burn. 244, Dorv., Meister. In Cicero and Quintilianmagis propriiwould be more usual for the latter.§89.etiam si sit. This conjecture of Spalding’s (foretiam sitGH Bodl. &c.:etiam siM Harl. 4950 Dorv.:etiam sicPrat. Put. S Harl. 2662) I have found in the Balliol codex. 7231 and 7696 giveetiam si est. Cp. note ontenuia atque quae§44, above.ut est dictum. These words were bracketed as a gloss by Halm, and are now omitted altogether by Krüger (3rd ed.): see however notead loc.Döderlein proposed to place them afterpoeta melior, Fleckeisen afteretiam si.Serranumis Lange’s conjecture forferrenumGHM:farrenum7231, 7696 Harl. 2662, 11671:PharrenumPrat. Put. Some MSS. (e.g. Vall. Harl. 4995, Burn. 243 and 244) givesed eum, but it is obvious that the criticism of Severus stopped with the wordlocum.§90.senectute maturuited. Col. 1527 and so 7231, 7696 (Fierville):senectutem maturbitGH:senectute maturumPrat. Put. MS Harl. 2662, 4995, 4950, Burn. 244, Dorv. and Ball.:senectus maturavitBodl., Burn. 243.et, ut dicam. Halm’ssedinstead ofethas been rejected by later critics. Cp. Claussen (Quaest. Quint., p. 357 note):sed‘sententiam efficit ab hac operis parte alienam. Nam cum oratori futuro exempla quaerantur oratoria virtus in quovis scriptore laudi vertitur (§§46, 63, 65, 67, 74, &c.). Itaque propter huius censurae consilium Quintilianus Lucani elocutionem oratoriam laudat, sed ingenium poeticum una reprehendit.’§91.propiusH Prat. Put. Burn. 243, Harl. 2662 and other codd.: Bodl. Ball. Harl. 4950proprius. Reisig conjecturedpropitius, which also is apt; but in spite ofindustrius,necessarius, cited in its support (cp. iv. 2. 27: vii. 1. 12), it is too uncertain a form to be received into the text. Iwan Müller thinks it would have to bemagis propitiae. Halm givespromptius: Wölfflinpronius: while Schöll now suggestspropitiae potius(cp. iv. pr. §5: 2 §27: vii. 1. 12).§92.feresG Prat. Put. 7231, 7696 S Harl. 2662, 4829, Dorv., Ball., Halm.:ferasH, Harl. 4950, Burn. 243, Bodl. C and M, Meister and Krüger (3rd ed.). Harl. 4995 hasfere: from Vall. Becher reports feras, ‘probably at firstferes.’elegeaGH 7696, and so A2BN Put. S at i. 8. 6.§94.abunde salisG Prat. Put. M and all my MSS. except H, Burn. 243, Bodl. which haveabundantia salis.multum est tersior. The variety of MS. readings seems to point to anetwrongly inserted aftermultum, perhaps from a confusion with ‘multum et ver gloriae’ below. GH givemultum et est tersior: M Harl. 4950, Bodl. Ball. C Dorv. Burn. 243 and also Harl. 4829multum etiam est t.: Prat. Put. 7231, 7696 S Harl. 2662, 11671multum est tersior: while Harl. 4995 (and Vall.) hasmulto et est tersior. Osann proposedmulto eo est tersior: Wölfflinmulto est tersior: Halm and Meister printmultum eo est tersior. Formultum, cp. multum ante xii. 6. 1: and see Introd.p. li.non laborGH Burn. 243 Bodl. and Meister:nisi labor7231, 7696 S Harl. 2662, 4995, 4950, 4829, 11671, Burn. 244, Dorv. Ball. C, and Halm. Prat. and Put. havemihi labor.hodieque et qui: H, Prat., Put., 7231, 7696, Harl. 2662, 4829, Bodl. Dorv.:hodie et quiBurn. 243:hodie quoque et quiVall. Harl. 4995, 4950:hodie quod et quiS.—Becher is of opinion that the text will not bear the explanation given in the note, and would readhodie quoque et qui: ‘es giebt auch heute noch berühmte Satirendichter, die einst &c.’Et quihe takes withclari, not withhodie quoque, theetbeing omitted in translation: clari (hodie quoque) qui (olim) nominabuntur.§95.etiam prius. Founding on the classification given in Diomedes (see notead loc.), according to which thesaturaof Pacuvius and Ennius preceded and was distinct from that of Lucilius, Horace, and Persius, Claussen (Quaest. Quint., p. 337) thinks that the true reading here may beAlterum illud et iam priusEnnio temptatumsaturae genus, &c. For the satura of Ennius, cp. ix. 2. 36. Iwan Müller points out that Ennius is not mentioned below (§97), beside Attius and Pacuvius, probably because neither in tragedy nor in satire did Quintilian consider him to have produced anything helpful for the formation of an oratorical style. Other unnecessary conjectures areetiam posterius, Gesner:etiam proprium, Spald.:etiam amplius, L. Müller:etiam verius, Riese:alterum illud Lucilio prius sat. genus, Krüger (3rd ed.).sola:solumPrat. and Put.collaturus quam eloquentiae. These words, omitted in GHS Bodl. Burn. 243, occur in all my other codd.§96.sed aliis quibusdam interpositus: sc. carminibus, Christ. In H the reading isquibusdam interpositus: so 7231, 7696 Bodl. and Burn. 243: but M Harl. 4950, 4829 Burn. 244 Dorv. and Ball, givea quibusdam interpositus: Scuiusdam: Prat. and Put.opus interpositus. Osann conjecturessed quibusdam, and so Hild. In the margin of Harl. 4995 is the variantaliquibus interpositis.In Hermes, vol. 23, p. 172, Kiderlin makes a fresh conjecture. Recognising that something must have fallen out beforequibusdam, but dissatisfied with Osann’ssedand Christ’ssed aliis, he proposes to readut proprium opus, quibusdam aliis tamen carminibus(orversibus)a quibusdam interpositus. The eye of a copyist may easily, Kiderlin thinks, have wandered from the first to the secondquibusdam: cp. v. 10. 64, ut quaedam a quibusdam utique non sunt, &c., and for quibusdam aliis xi. 3. 66, et quibusdam aliis corporis signis.intervenit, which is a conjecture of Osann, I have found in Harl. 2662, 11671 Prat. Put. 7231, 7696.lyricorum. Kiderlin thinks there may be something wrong in the text here. The last sentence (sed eum longe, &c.) shows clearly that Quintilian had a high opinion of the lyrists of his day: if Bassus waslegi dignus, they were even more so. Would he then have said ‘of the Roman lyrists Horace is almost the only one worth reading’? Perhaps we should readlyricorum priorum: after-ricorum,priorummight easily fall out, and it gives a good antithesis toviventium. Bassus (quem nuper vidimus) forms the transition: and the next paragraph beginsTragoediae scriptores veterum, &c.§97.clarissimi. This reading is stated by Halm to be ‘incerta auctoritate,’ and is referred by Meister to the Aldine edition. It occurs in Prat. 7231, 7696 Harl. 2662 (A.D.1434) Vall. 4995, 4829, 11671, Dorv. and Ball.: Put. givesclarissime: G hasgravissima: HFTSgravissimus, and so also Harl. 4950, Burn. 243, Bodl. and C. Halm printsgrandissimi: Ribbeck (Röm. Trag. p. 337, 3) inclines to accept the sing.grandissimus, M, of Pacuvius alone.Kiderlin (in Hermes 23, p. 173) rejects all the above readings.Gravissimusandgravissimaare obviously due, he says, togravitatefollowing: but the word beforegravitatemust have begun with the same letter, and soclarissimicannot stand, especially as it is inappropriate to the context. Forceterumshows that the sentence before it must have contained some slight censure: some defect, or quality excluding others equally good, must have been mentioned. He therefore conjecturesgrandes nimis, in preference tograndissimi, which in tragedy would hardly be a fault. Attius and Pacuvius, Quintilian says, are ‘zu grossartig, sie kümmern sich zu wenig um Zierlichkeit (Eleganz) und die letzte Feile (d.h. Sauberkeit im Kleinen); doch daran ist mehr ihre Zeit schuld als sie selbst.’ He evidently thinks more of the ‘Thyestes’ of Varius and Ovid’s Medea: cp. Tac. Dial. 12. With this judgment Kiderlin compares§§66,67tragoedias primus in lucem Aeschylus protulit, sublimis et gravis et grandiloquus saepe usque ad vitium, sed rudis in plerisque et incompositus ... sed longe clarius inlustraverunt hoc opus Sophocles atque Euripides, and is of opinion that the parallelism cannot be mistaken. For the position ofnimishe compares ix. 4. 28 longae sunt nimis: v. 9. 14 longe nimium: xii. 11. 9 magna nimium.§98.quem senes quidem parum tragicum. So Spalding, Bonnell, Halm, Meister, and Krüger.Quidemoccurs in no MS.: GH havequem, M Vall., Harl. 4995, Burn. 244, Ball, omit it: Bodl. Burn. 243 and Dorv. show the corruptionPindarum. Becher would excludequidem, regardingquemin G as an instance of the tendency of copyists inadvertently to repeat, after a particular word that by which it has been immediately preceded, e.g.§68quod ipsum quod (G): ix. 4. 57 ut cum ut (G): iv. 1. 7 ipsis litigatoribus ipsis (b): iv. 2. 5 aut ante aut (bT): x. i. 4 iam opere iam (G).—But here the authority of the Pratensis and its cognates may be invoked. In the archetype from which they are derived something must have stood beforeparum, as Prat. Put. 7696, 7231 all givequem senes non parum tragicum: so Harl., 2662 (A.D.1434), and 11671. Above in§96, G Prat. Put. 7231, 7696 havesi quidemforsi quem.§100.linguae suae. So Köhler (v. Meister pref. to Book x. p. 13):suaesupplies an antithesis to ‘sermo ipse Romanus’: GH givelinguae quae: so Harl. 4950: S Burn. 243, Bodl.linguae: while Harl. 2662, 4995, 4829, 11671, Dorv. and Ball. omit it altogether: M hasligweque.§101.Titum: GH Prat. Put. M. 7231, 7696.commendavit: Halm and Meister givecommodavit, which is approved also by Hirt. Halm compares§69where Menander is said to be ‘omnibus rebus personis adfectibus accommodatus.’ But this would require the meaning ‘appropriately treated,’ and there is no instance in Quintilian of the verb used absolutely in this sense. Nor is there any example to support Hild’s interpretationpraestitit, which would be moreover extremely weak. The recurrence of the word so soon afteraccommodatatells against Halm’s reading, though Quintilian is negligent on this head.—On the other hand, in vi. 3. 14 the reading ‘ad hanc consuetudinem commodata’ is rightly accepted against ‘commendata’ most edd.§102.immortalemGS Meister:illam immortalemPrat. Put. M Halm:immortalem illamVall.velocitatem. So all MSS, except S, Burn. 243, and Bodl., which havecivilitatem. Kiderlin (in Hermes 23, p. 174) thinks that we might have expectedideoque immortalem gloriam quam velocitate Sallustius consecutus est: ‘und darum hat er dievelocitasdurch (von der velocitas) verschiedene Vorzüge erreicht.’Consequicannot mean ‘to supply the place of’: andimmortalisis inappropriate as an attribute ofvelocitas: besides, Quintilian has not spoken of Sallust’svelocitas, even indirectly. Schlenger conjecturedclaritatem: Andresenauctoritatem(‘klassisches Ansehen,’ cp. iv. 2. 125: xii. 11. 3): Kiderlin now proposesdivinitatem, which in Cicero = Vortrefflichkeit, Meisterschaft: cp. xi. 2. 7. Judged by the previous sentences the expression is not too strong. Forimmortalem divinitatemcp.§86illi ... caelesti atque immortali: and forconsecutus estiii. 7. 9 quod immortalitatem virtute sint consecuti.clarus vi ingenii. This is a conjecture of Kiderlin’s, which I find has been adopted also by Krüger (3rd ed.). GHFT giveclarius ingenii: Prat. Put.clari ingenii vir: 7231, 7696clari vir ingenii: MS Harl. 4995, 4950, 4829, Burn. 243 and 244, Dorv.C and Ball,clarus ingenio; Harl. 2662 and 11671clarus(?) orclaret vir ingenii. Spalding had already pointed out thatclarusis not found withingenium, except whereingeniumis used of a person: e.g.§119erant clara et nuper ingenia: he therefore wroteelati vir ingenii(following Goth.elatus ingenioand Bodl.elatus ingeniis). Kiderlin compares§70sententiis clarissimus, and forvis ingeniii. pr. 12: ii. 5. 23: x. 1. 44: xii. 10. 10. The readingclarus vi ingeniipoints the contrast to what follows in ‘sed minus pressus,’ &c.: it was hisstylethat did not altogether suit the dignity of history.§103.genere ipso, probabilis in omnibus, sed in quibusdam. Till Kiderlin made this happy conjecture (see Hermes 23, p. 175)generehad always been joined withprobabilis, and the text was twisted in various directions. GHS, Burn. 243, Bodl. givein omnibus quibusdam: M Harl. 2662, 4995, 4950, Burn. 244, Dorv.in omnibus sed in quibusdam, and so apparently Prat. Put. 7231, 7696. Out ofomnibusHalm gives on Roth’s suggestion,operibus: afterwards he decided forpartibus, and this (thoughomnibustopartibusis not an easy transition) is adopted by Meister. Kiderlin’s punctuation makes everything easy: ‘Anerkennung verdienen seine Leistungenalle,manchestehen hinterseinerKraft zurück.’ Even these last, Quint. means, areprobabiles(cp. viii. 3. 42 probabile Cicero id genus dicit quod non plus minusve est quam decet); but they do not show the great powers that distinguish his other writings. It is uncertain whether Quintilian wrotein quibusdamorsed in quibusdam(M). The easiest explanation of the omission in the other MSS. is to suppose that he wrotein omnibus in quibusdam: perhaps the copyist of M saw thatomnibusandquibusdamwere antithetical, and insertedsed. Kiderlin notes Quintilian’s liking for chiasmus, without any conjunction: cp.§106in illo, in hoc (where in hoc is wanting in M).suis ipse viribus: ed. Col. 1527 (Halm), and so (Fierville) 7231, 7696. In Harl. 2662 and 11671 (A.D.1434 and 1467)suisalready appears, corrected fromvisGH. The Juntine ed. (1515) hassuis viribus minor: so Prat. and Put.§104.et exornat. Vall. and (apparently) Prat. Put. 7231, 7696, and most edd.:et ornatM Halm, Meister, Krüger:exornatGHS. Becher remarks thatet exornatmight easily pass intoexornat.nominabitur: Weber and Osann proposednominabatur(which appears in Harl. 2662, but corrected to-itur). Krüger at first accepted this in support of his theory that the whole passage refers to Cremutius, who ‘in former days (olim), while his works were under a ban, was only named (i.e. was a mere name, but now is known and appreciated).’ The parallel passage (§94) is sufficient to dispose of any such interpretation: sunt clari hodieque et qui olim nominabuntur.Cremuti. Nipperdey, Philol. vi, p. 193, Halm, and Meister:remutiH Prat. Put. 7231, 7696remremutiG,rem utiliBurn. 243:remittiS. Bodl.:nec imitatores utiHarl. 2662, 4995, 4950, 4829, 11671. A review of the various explanations of the whole passage (Superest—quae manent) will be found in Holub’s Programm ‘Warum hielt sich Tacitus von 89-96 n. Chr. nicht in Rom auf?’—Weidenau, 1883: but his conjectureremoti(i.e. relegati) forremutiis not to be thought of.dividendi: first in the Aldine edition: all MSS. havevidendi, except M (indicendi) and Prat. Put. Harl. 4995 (vivendi). Cp. i. 10. 49, where the case is the same.§105. In the Aurich Programm, Becher gives a more recent statement of his views: ‘wie zucumcausale, so tritt praesertim auch zucumconcessivum, in diesem Falle wiedenzugeben mit, “was um so auffallender ist, als.” Der Sinn ist also: “Ich weiss sehr wohl, welchen Sturm des Unwillens ich gegen mich errege, und dies (dieser Sturm) ist um so auffallender, als ich jetzt gar nicht die Absicht hege, meine (in Potentialis gesprochene) Behauptung (fortiter opposuerim) wahr zu machen, resp. comparando durchzuführen. Ich lasse ja dem Demosthenes seinen Ruhm—in primis legendum vel ediscendum potius.”’§106.praeparandi. For Kiderlin’s conj.praeparandi,narrandi,probandiseead loc.[omnia]denique, GH, Burn. 243, Bodl. omitomnia(which is in all my other MSS.), and Meister now approves (following Spalding, Osann, and Wölfflin), on the ground that Demosthenes and Cicero werenotalike ineverythingthat belongs toinventio. Halm thinks thatomniais to be found inracioniof the older MSS.: but Kiderlin points out that this error may have arisen from the carelessness of a copyist who, after thrice writing the terminationi, gave it also to the fourth word.illi—huicPrat. M, S Vall. Harl. 4995, 2662 Bodl. &c.:illic—hicGH Put. 7231, 7696, Halm.§107.vincimus, H, G2, and most MSS.: (cp.§86):vicimusG.§109.ubertateHarl. 4995. This is also the reading of codd. Vall. and Goth.:all the other MSS. giveubertas.totas virtutesBn Bg N Prat. Ioan. 7231, 7696:totas viresM b.§112.ab hominibusHalm and Meister:ab omnibusBn Bg HFT Ioan. Prat. 7231, Sal. and most codd.:hominibusS Harl. 4995 Bodl.§115.urbanitas. Kiderlin proposes to readet praecipua in accusando asperitas et multa urbanitas: cp.§117:§64:2 §25: ii. 5. 8.Ciceroni, forCiceronemof the MSS. In the Rev. de Phil. (Janv.-Mars, 1887) Bonnet quotes from the Montpellier MS. a note of the sixteenth century deleting the name as a gloss (oninveni). Certainly all codd. giveCiceronem, notCiceroni. Bonnet thinks that the insertion does not accord with Quintilian’s habitual deference towards Cicero: ‘Quintilien se trouvant dans le cas de contredire Cicéron ne le nomme pas.’—Becher reportsCiceroni, a correction in the Vallensis.castigata, B (i.e. Bn and Bg) Ioan. Prat. 7231, 7696 Harl. 2662, 4995, 11671:custoditaH M b F T Alm. Harl. 4950, 4829, Burn. 243, 244, Bodl. Dorv. and Ball. Forgravis(bH M Vall. and seemingly Prat.) B Sal. 7231, 7696 and Ioan. givebrevis.si quid adiecturus sibi non si quid detracturus fuit, Vall. Harl. 4995. For the repetition, see on haud deerit3 §26. Halm and Meister printsi quid adiecturus fuit—(sc.virtutibus suis, cp.§§116,120)—the reading of B (i.e. Bn and Bg), which is also that of Ioan. Prat. N 7231 Harl. 2662, 11671: while M Harl. 4950, 4829, Burn. 244 havesi quid adiecturus fuit, non si quid detracturus. The reading of H issi quid adiecturus sibi non si quid detracturus[Sulpicius insignus]fuit ut servius sulpicius insignem&c.: so also T, Burn. 243, Bodl. The brackets in H are by a later hand, indicating a gloss which arose from a mistake made by the copyist of H. In Bg the passage stands:—sibinon siquiddetracturussi quid adiecturus:fuitetserviussulpiciusThe words added above the line are by the hand known as b.In copying H wrote:si quid adiecturus sibi non si quid detracturus(then omittingfuitcontinues)et Serv. Sulp.(then goes back and resumes)fuit et servius&c. This is the origin of the confusion which exists in all the MSS. of this family.§117.et fervor. This is Bursian’s conjecture, adopted by Halm and Krüger (3rd ed.), and now approved by Becher. BM haveet sermo, which is also the reading of N Prat. Sal. 7231, 7696 Ioan. Harl. 2662, 4950 and Ball.: Hbet summo: Harl. 4829, 11671, Burn. 244et smo: while Bodl., Dorv., and Burn. 243 give the correction in Teius summa, out of which the second hand in the Vallensis (Laurentius Valla) madeet vis summa, a reading which occurs also in Harl. 4995. Meister readset sermo purus; while Kiderlin proposeset simplex sermo(cp. iv. 1. 54: viii. 3. 87: ix. 3. 3: 4. 17: viii. pr. 23: x. 2. 16).ut amari sales. Francius conjecturedut amantur sales, but this loses the antithesis betweenamariandamaritudo ipsa. Kiderlin’sut amantur amari sales(viii. 3.87: vi. 1. 48) is an improvement; but ifridiculais taken in a good sense it seems impossible that after censuring Cassius for giving way unduly tostomachus, Quintilian should go on to say, ‘moreover, though bitter wit gives pleasure, bitterness by itself is often laughable.’ Is it possible that we ought to readut amari sales risum movent ita amaritudo ipsa ridicula est? Such an antithesis might have been written ‘per compendium,’ and the wordsrisum moventmay then have dropped out. See the notead loc.: and cp. especially vi. 1. 48fecit enim risum sed ridiculus fuit, andοὐ γέλωτα κινεῖ μᾶλλον ἢ καταγελᾶται, quoted in the note on1 §107.—Krüger (3rd ed.) adoptsfrequentiorforfrequenter, which gives a good sense, except thatfreq. amar ipsais awkward.§121.leneHalm and Meister:leveB Prat. N 7231 M 7696 C. Here again Becher prefersleve, comparing Cic. de Orat. iii. §171, quoted on§44above: levitasque verborum 1. 52: and levia ... ac nitida, v. 12. 18.§123.scripserint. So Bn Bg H Ioan. Prat. 7231, 7696 Vall. Harl. 4995, 2662, 11671, Bodl., Dorv., Spalding, and Bonnell. Becher compares among other passages2 §14(concupierint), and points out that Quintilian is not thinking of individual writers on philosophy, but of the class, as opposed to the class of orators, historians, &c.—Halm, Meister, and Krüger havesupersunt(Put. M, Ball. Burn. 243 Harl. 4950).§124.Plautus, Prat. N, 7231 Ioan. Harl. 2662, 4829, 11671:plantusM Harl. 4950:PlantatusSal.:plaustusHb:Plancusedd. vett. and Harl. 4995.Catius. The name is rightly given in Harl. 4995.§126.iis quibus illi.Iisis the conjecture of Regius, followed by Halm, Meister, and Krüger. Becher would retainin quibus illi,—the reading of BN Prat. Ioan. Vall. M Harl. 4995, 2662, 4950, 11671, Burn. 244 Dorv. Ball. The difficulty of construing probably led to the omission ofinin bH Bodl. Burn. 243, 7231, 7696, Spalding and Bonnell.ab illoB Ioan. 7231, 7696 Sal. Harl. 2662, 4950, 4829:ab eobHM Burn. 243.§127.foret enim optandum:fore enim aliquid optandumbHFT. Spalding conjecturedalioqui optandum, which Kiderlin approves.ac saltemall MSS.: Meister hasaut saltem, probably relying on a wrong account of the Bambergensis: see Halm vol. ii, p. 369.illi viroB:illi virusbHM:illi virtutibusHalm:illi viro eos(orviro plurimos) Kiderlin.§128.multa rerum cognitio: so all codd. except Ioannensis and Harl. 4995, which havemultarum rerum cognitio. b omitscognitioand is followed by HFT.§130.si obliqua contempsisset, si parum recta non concupisset. I adopt the reading recently proposed for this vexed passage by Ed. Wölfflin in Hermes, vol. xxv (1890), pp. 326-7, though it is right to note that he was partly (as will be seen below) anticipated by Kiderlin.Obliquaseems thoroughly appropriate in reference to Seneca’s unnatural, stilted, affected style,—‘jene unnatürliche, durch unmässigen Gebrauch von Tropen und Figuren auf Schrauben gestellte Ausdrucksweise, welche statt der Klarheit ein Schillern zur Folge hat.’ Wölfflin compares ix. 2. 78rectum genusadprobari nisi maximis viribus non potest: haec diverticula et anfractus suffugia sunt infirmitatis, ut qui cursu parum valent flexu eludunt, cum haec quae adfectatur ratio sententiarum non procul a ratione iocandi abhorreat. Adiuvat etiam, quod auditor gaudet intellegere et favet ingenio suo et alio dicente se laudat. Itaque non solum si persona obstaretrectae orationi(quo in genere saepius modo quam figuris opus est) decurrebant ad schemata ... ut si pater ... iacularetur in uxoremobliquissententiis. This passage supplies (what is indeed suggested byobliquaitself) the antithesisparum recta: cp. ii. 13. 10 si quis ut parum rectum improbet opus.In theJahrbücher f. Philologie(vol. 135, 1887: p. 828) Kiderlin had previously dealt with the passage on similar lines. The traditional readingsi aliqua contempsisset(b) he considers too indefinite, though not impossible: in point of authority, though preferable to thesi nil aequalium cont.of the later MSS., it cannot rank so high as the reading of Bn and Bg, which givesimile quamwithout any attempt at emendation. This Kiderlin thinks must be nearest the original: he therefore rejects such conjectures as Jeep’ssi antiqua non, on the ground that it is improbable thatsimile quamarose out ofantiqua. He introduces his own conjecture by referring to ix. 2. 66 and 78 (see above), and to the contrast betweenschemataandrectum genus,recta oratio; the former are calledluminaorlumina orationis(xii. 10. 62). Cp. viii. 5. 34. He would read:nam si mille ille schemata(orillas figuras)similiaque lumina contempsisset, si parum rectum genus(orsermonem)non concupisset, &c.Similiaqueoccurs ix. 4. 43:mille(forsescenti) is used v. 14. 32: forcontempsissetcp. ix. 4. 113.Si mille illaandsimiliaquemay easily have run together, whenschemata(orfiguras) would fall out:quamin the older MSS. may representque lumina, which again reappears in thequaliumof the later codd. (si nil aequalium). As an alternative forparum rectumgenns(orsermonem) Kiderlin suggests Wölfflin’s readingparum recta: and compares ix. 2: ii. 5. 11: v. 13. 2: ix. 1. 3; 3. 3: x. 1. 44; 89: ii. 13. 10.Of the MSS. Prat. 7231 Sal. 7696 N Ioan. Harl. 2662 and 11671 agree with Bn and Bg in givingsimile quam: b hassi aliqua: HFT, Burn 243, Bodl.aliqua: M Harl. 4995, 4950, 4829, Burn. 244, Dorv. Csi nil aequalium. Among previous conjectures aresi multa aequalium, Törnebladh:si ille quaedam, Halm (whereilleis surely superfluous):si antiqua non, Jeep. Meister accepts the readingsi aliqua non: Becher thinks thatsi nil aequaliummay be right.It is generally admitted that a word must have fallen out afterparum: the codd. all givesi parum non concupisset. Jeep proposedsi pravum(=corruptum: cp. ii. 5. 10)non conc.: on which Halm, comparingomnia sua, remarks, ‘debebat saltemprava.’ Butpravaseems too strong a word for Quintilian to have used in a criticism where he is so studiously mixing praise and censure. Halm suggestedsi parum sana, and is followed by Meister: cp. Fronto’s ‘febriculosa’ of Seneca, p. 155n. Sarpe proposedsi pravaorparvaorplura: Buttmannsi parum concupiscenda(orconvenientia): Herzogsi parvum: Madvigsi partimorpartem(i.e.paulo plus quam aliqua, and in opp. toomnia sua, below): Hoffmannsi opiparum: Seyffertsi garum: Kraffertsi non parum excussisset(cp.§101,§126: v. 7. 6; 7. 37; 13. 19: xii. 8. 13, &c.): Gustaffsonsi parva(cp. i. 6. 20 frivolae in parvis iactantiae): Andresensi similem ei quem contempsit se esse(sc.concupisset; cp. Tac. Ann. xiii. 56: xii. 64: Hist. i. 8: Livy xlv. 20. 9)si parem non concupisset(i.e.si Ciceronianum genus dicendi imitari quam diverso genere gloriam eius aemulari maluisset): or,nam si similem ei quem contempsit se esse, non parem concupisset: Krüger (3rd ed.)si parum arguta: Hertz (who argues that the word which has fallen out must, withparum, correspond tocorruptaabove)si parum pura.utrimqueMeister and Becher, following old edd., Spalding, and Bonnell:utrumqueB N 7231, 7696:virumqueM:utcumqueHalm, ‘in every way,’ ‘one way or another,’—proposed by Gesner at6 §7.CHAPTER II.§2.atque omnis. Kiderlin (Blätter f. d. bayer. Gymn. 1887, p. 454) proposes to put commas atsequiandvelimus, and make this clause also subordinate.§3.aut similes aut dissimiles. Andresen suggestsaut similes aut non dissimilesoraut similes aut certe haud dissimiles.§6.tradiderunt(BNM Harl. 2662, 4995, 4829, Burn. 243, and Dorv.) is powerfully supported by Becher in his latest tractate (Programm des königlichen Gymnasiums zu Aurich, p. 13) againsttradiderint, the reading of b Prat. Bodl. and Vall. (corrected in the last fromtradiderunt), Burmann, Spalding, Bonnell, Halm, Meister, and Krüger. Becher holds that in Quintilian, as frequently in Cicero,cumwith the indicative is often used in such a way (quoting from C. F. W. Müller) ‘ut non prorsus idem sit, sed simillimum ei, quod barbare dicere solemus identitatis. Nam ut “cum tacent clamant” non est “si tacent,” multo minus “quo tempore” aut “propterea quod” aut “quamquam,”—sed “tacent idque idem est ac si clament,” sic “cum hoc facis qui potes facere illud?” et sim., German, item “wenn du dies thust” valet: “hoc facis ex eoque per se efficitur, non ratione, sed ipsa natura, ut illud non possis facere.” Ut pro Q. Roscio 3. 9 quam ob rem, cum cetera nomina in ordinem referebas, hoc nomen in adversariis relinquebas? non significat nec “quamquam” nec “quando,” sed “wenn.”’ Becher adds the following parallel passages: Cic. pro Cluent. 47. 131 id ipsum quantae divinationis est scire innocentem fuisse reum, cum iudices sibidixeruntnon liquere, and Verg. Ecl. 3. 16 quid domini facient, audent cum talia fures? (Cp. Madvig de Fin. p. 25.) In the same way he treatscum ... sunt consecuti7 §19below, which seems, however, to be somewhat different. Here there is an antithesis, and in such casescum(‘whereas’) may very well take the indicative: there the clause ‘cum sint consecuti’ is added to show the reasonableness (cum= ‘since’) of the demand that extemporary facility shall be made fully equal tocogitatio—seead loc.Neither instance can be explained on the analogy ofcumwith the indic. used of ‘identity’ (as ‘cum tacent, clamant,’ quoted above): in such cases the subject is generally the same in both clauses. And in such a passage as pro Cluent. §131cumis usually explained as =quo tamen tempore.eruendasM Harl. 4995: all other codd.erudiendas.mensuris ac lineis. Krüger (3rd ed.) quotes with approval the conjecture of Friedländer (Darst. aus der Sittengesch. Roms iii. 4. p. 194. 4)eisdem mensuris ac lineis, and recommends the insertion ofeisdemin the text,—afterlineis, where it is more likely to have fallen out. But this is unnecessary.§7.turpe etiam illud est. Hild puts a comma aftersciant, and by supplying beforeturpe estanitato correspond withquemadmodum, makes out a comparison of whichquemadmodum, &c., is the first clause andturpe etiam illud estthe second. This is certainly to misunderstand the passage. Thequemadmodumclause goes with what is before, not with what follows, so that a comma afteralieniwould be enough, were it not for the necessity of having the mark of interrogation (cp.§9below). Thenturpe etiam illud estcomes in, resumingpigri est ingeniiin§4, just as immediately afterwardsrursus quid erat futurum§7resumesquid enim futurum erat§4. The whole passage is an elaboration of the dictum with which§4opens, ‘imitatio per se ipsa non sufficit.’ Quintilian first says that we, as well as those who have gone before us, may make discoveries (cur igitur nefas est reperiri aliquid a nobis quod ante non fuerit?). Surely we arenotto confine ourselves to hard and fast lines like servile copyists.Then he goes on to add in§7that we must surpass our models (plus efficere eo quem sequimur), instead of resting content with mere reproduction (id consequi quod imitamur): otherwise Livius Andronicus would still be the prince of poets, we should still be sailing on rafts, and painting would still be nothing more than the tracing of outlines. The necessity for progress is first shown (§§4-6) by an appeal to the example of the past, and by the unfruitful work of such painters as are mere copyists: then in§7poetry, history, navigation, as well as painting are put in evidence for the argumentex contrario.§8.mansit, Meister:sitcodd.:estFleckeisen (and Halm):fuitGensler.§9.adpetentBg HFT:appetentPrat. Ioan. Harl. 4995 Bodl. &c.:appetuntN Harl. 2662, 11671, Burn. 243.hoc agitHalm, followed by Meister (cp.7 §4):hoc aitb H,om. Bn Bg N Ioan. Prat. Harl. 2662, 11671:agit(sine hoc) Harl. 4995, 4950 M, and most codd.§10.quaeque pares maximemay be a gloss: it is found only in those MSS. which givesimplicissimaeforsimillimae: b H Harl. 4950 M Burn. 243 Bodl.utique(b M Vall. Harl. 4995, 4950, Burn. 243 Bodl. Dorv.) may also be suspected: it does not occur in Bn Bg N Ioan. Prat. Harl. 2662, 4829, 11671.§11.orationibus, Bg: Ioan, givesoratione: so also Voss. 1 and 3 (Zumpt).accommodaturb H Ioan. Harl. 4995, 4950, 4829, Bodl. Dorv. and Meister:commodaturBn N Prat. Harl. 2662, 11671, and Halm.§12.inventio visB Harl. 2662, 11671:inventionisb H Harl. 4495, 4950, 4829, C, Burn. 243, Bodl., Dorv.§13.cum et, ed. Colon. 1527:et cumB H Ioan. Prat. N (et quum) M:cumVall. Harl. 4995. On the usual interpretation of this difficult passageut quorum ... collocata suntforms one parenthesis: but this is an unnecessary extension of the explanation ofintercidant invalescantque temporibus. Seead loc.accommodata sit, codd. except Harl. 4995, which omitssit:acc. estHalm, followed by Hild (depending onprout, notcum: see notead loc.). Madvig’s conjectureaccommodanda sitis approved by Kiderlin (cp. ix. 4. 126 adeoque rebus accommodanda compositio). But the correctness of the reading in the text (and also of the explanation given in the notead loc.) will be evident to any one who considers the whole sentence carefully. Tocum et verba intercidantcorresponds exactly the double clauseet compositio ... rebus accommodata siton the one hand, andet compositio ... ipsa varietate gratissima(sc.sit—repeated fromaccommodata sit) on the other. This double clause is rather awkwardly joined bycum ... tum. To takeaccommodata sitas depending on thecumwhich followscompositiois to destroy the balance of the sentence. In this case an independentsitwould have to be supplied withgratissima(to makeet compositio ... gratissima sitcorrespond toet verba intercidantabove): and the translation would then be: ‘it is just when (cum ... tum), or exactly in proportion as, it is adapted to the sense (rebus accommodata) that the very variety (thereby secured) gives the arrangement its greatest charm.’ But if this had been Quintilian’s meaning he would surely have writtencum rebus accommodatur(or—ata est)tum ipsa varietate sit gratissima.§14.quos imitemur. The D’Orville MS. givesquos eligamus ad imitandum,—probably an emendation by the copyist, though it may explain the origin of the reading of b and Hquos at imitandum.quid sit ad quod nos. Theadis due to Regius: most codd. havequid sit quod nos, except Harl. 4995, which is again in agreement with Goth. Vall. Voss. 2 and the second hand in Par. 2:quid sit quod nobis.§15.et a doctis, inter ipsos etiam. The explanation given in the notes is due to Andresen (Rhein. Mus. 30, p. 521), who, however, wished to insertetbeforeinteripsos. The comma makes that unnecessary. So Kiderlin (Berl. Jahrb. XIV, 1888, p. 71 sq.).dicunt, Harl. 4995:dicantall codd.: ‘emend. Badius’ (Halm).ut sic dixerimVall. (Becher): cp. pr. 23: i. 6. 1: ii. 13. 9: v. 13. 2. BM Prat. haveut dixerim. Halm wroteut ita dixerim, comparing i. 12. 2: ix. 4. 61: butut sicis more common in the Latinity of the Silver Age.§16.compositis exultantes. Kiderlin (Berl. Jahrb. XIV, 1888, p. 72) would prefercompositis rigidi(cp. xi. 3. 32: xii. 10. 7: ix. 3. 101: xii. 10. 33),comptis(cp. i. 79: viii. 3. 42)exultantes= ‘statt wohlgeordnet steif, statt schmuckliebend putzsüchtig.’ Another unnecessary emendation islaetis exultantes, compositis corrupti(Lindau): orcompositis exiles(Düntzner).§17.quidlibet, most codd.:quamlibetM, Vall. Harl. 4995, 4950:qui licetbH. Iwan Müller (Bursian’s Jahresb. 1879, p. 162) condemnsillud, and would read eitherquamlibet frigidum(cp.3 §19and ix. 2. 67: quamlibet apertum), orquidlibet frigidum, which latter is approved by P. Hirt. Eussner suggests the deletion ofillud frigidum et inane, thinking that these words may be the remains of a gloss on§16.Attici sunt scilicet. Spalding’s reading seems on the whole to be preferred. The retention ofsunt(represented in some MSS. by a simples,—hence the readingAtticis scilicet) makes it less necessary to follow Meister in inserting asuntafterqui praec. concl. obscuri: in so loose a writer as Quintilian the firstsuntwould do duty for both. Halm follows Bn and Bg, which apparently (as also N Harl. 2662, 4829, and 11671) haveAttici scilicet: Meister (with bHM and Harl. 4950) givesAtticis scilicet. In the Ioannensis I findAttici s(forsunt): Dorv. and Burn. 244 giveAtticis s. Scilicet(om. Prat.) may be a gloss, and the true reading may beAttici sunt. Some codd. (Bodl. Burn. 243) giveAtticos scilicet(AthicosHarl. 4995): qy.Atticorum similes? (cp. Cic. Brut. §287).—Becher now prefersAtticis(sc.se pares credunt).§22.proposito. This conjecture by Gertz (Opuscula philol. &c., p. 134) I have found in the Ioannensis (*ppo) and in Harl. 2662 and 11671. It is approved also by Kiderlin. BNHb Prat. Sal. givepropositio: all other codd.proposita. Perhaps we should read (with Ioan.)sua cuique proposito est lex, suus decor est. Prat. omits the secondest.§23.tenuitas aut iucunditas, Halm and Meister:tenuitas ac iucunditasb H, Burn. 243, Bodl.:tenuitas aut nuditasN Ioan. M Harl. 2662, 11671:tenuitas ac nuditasPrat. Harl. 4995, 4950, 4829, C, Burn. 244, Dorv.:aut iuditasBg.§25.quid ergo? non est satis, &c. Gertz proposes to read, shortly afterwards,mihi quidem satis esset; set si omnia consequi possem, quid tamen noceret vim Caesaris ... adsumere?(=sed etiam si satis mihi esset, tamen nihil noceret vim Caesaris ... adsumere, si omnia haec consequi possem).§28.deerunt, Francius:deerant(derant) all codd. Becher defendsdeerant: ‘der Rhetor meint dassqui propria bona adieceritöfter Veranlassung gehabt haben wird, Fehlendes zu ergänzen als zu beschneidensi quid redundabit.’oporteatbHFT Bodl. M Harl. 4950 Burn. 243:oportebatB Prat. N Sal. Ioan. Harl. 2662, 4995, 4829, 11671 Burn. 244 Dorv. The latter (which is adopted by Halm) would indicate (cp. viii. 4. 22) a condition which ought to have been and may still be realised: the former (adopted by Meister and approved by Becher) is the conjunctive potential, and is quite in Quintilian’s manner (cp. xi. 2. 20): it conveys the expression of a present duty and obligation, the realisation of which may now be expected, and it connects also more intimately witheritin the following sentence.CHAPTER III.

Theaut ... autimmediately below is very much against this conjecture, which however Krüger (3rd ed.) has received into the text: we should expect rathernescio an illa quisquam, ornullus poeta, or keepingillaas nominativenescio an illa poeta ullo. Quintilian’s use ofnescio an(like that of post-Augustan writers generally) is vague: it is usually an expression of doubt, theanmeaning either ‘whether,’ or ‘whether not’ indifferently. Cp. ix. 4. 1: vi. 3. 6: viii. 6. 22: xii. 10. 2: i. 7. 24. (Mayor cites also Plin. Ep. i. 14. 9: iii. 1. 1: iv. 2. 1: v. 3. 7: vi. 21. 3: vii. 10. 3: 19. 4: viii. 16. 3: ix. 2. 5; and adds ‘In all these instancesnescio an(dubito an) is ‘I doubt whether’; in Cicero the meaning is always ‘I rather think.’’)Andresen proposednescio an ulla poeseos pars. The passage closely resembles§28, and must be emended on the same lines.§66.tragoedias. Thurot (Revue de Phil. 1880, iv. 1, p. 24) conjecturedtragoediam: cp.§67hoc opus. He is followed by Dosson, against all MS. authority. Becher points out that we must supply withhoc opusin§67the words ‘tragoedias in lucem proferendi,’ so thatopusandtragoediassquare well enough with each other.§68.quod ipsum reprehendunt, Meister, Krüger (3rd ed.) and Becher. This reading also occurs in the Codex Dorvilianus. Other readings arequod ipsum quodGHT Burn. 243, Bodl.:quo ipsumMS Harl. 2662, 4995, 4950, Ball. Halm conjecturedquem ipsum quoque, and was followed by Mayor and Hild. But as no fault has been found with Euripides in the foregoing,quoqueseems out of place.Founding on the reading of GHT, &c., also on that of F (which givesquod ipsum qui) Kiderlin (Hermes 23, p. 165) proposes to readquod ipsum quidam, comparing§98, where forquem senes quem(GT) Spalding rightly conjecturedquem senes quidem, and7, §21, where Bn, Bg givequodforquosdam. He then goes on, in an interesting paper, to reconstruct the whole passage, which is open to suspicion, especially in respect thatsublimiorstands as predicate withgravitasandcothurnus, as well as withsonus. The admirers of Sophocles consider his elevation of tone more appropriate than the strain of Euripides.Sublimioris therefore perhapsnotthe predicate of the sentence, however suitable it may be as the attribute ofsonus. The predicate may have dropped out, andsublimiormay have been transferred from its real place to supply it. It is striking that GFTM (also H and Bodl.) all givesublimior erit. Kiderlin imagines that a copyist who missed the predicate wrote in the margin ‘sublimior erit ponendumpost esse’: and then another insertedsublimior eritafteressein the text. For the predicate,magis accommodatusmight stand: in copying, the eye may have wandered frommagis accommodatustomagis accedit: formagis accomm.cp. ii. 5. 18 and x. 1. 79. Kiderlin therefore boldly proposes to make the parenthesis run, ‘quod ipsum quidam reprehendunt quibus gravitas et cothurnus et sublimior sonus Sophocli videtur esse magis accommodatus’: ‘was gerade manche tadeln, welchen das Würdevolle, der Kothurnus, und der erhabenere Ton des Sophokles angemessener zu sein scheint.’et dicendo ac respondendo7231, 7696:dicendo ac respondoGH:in dicendo et in respondendoPrat. Put. S (et respondendoM).praecipuus. Hunc admiratus maxime est. This is Meister’s reading, except that foreumI give (with Prat. Put. 7231, 7696 Harl. 2662 and 4995)hunc, which is commoner in Quint. at the beginning of a sentence (§§46, 78, 91, 112). The following are the readings of the MSS.: GHpraecipuus et admiratus miratus: M Bodl. Harl. 4950, 4829, Burn. 244, C, Burn. 243 Ball. Dorv.praecipuus et admirandus: Spraecipuum. Nunc admiratus et: Prat. Put. Harl. 2262 and 11671praecipuus hunc admiratus et maxime est ut saepe test. et sec. quamvis: Harl. 4995,hunc admiratus max. ut s. test. et eum secutus quamquam. Halm givespraecipuus est. Admiratus maxime est: Kiderlin insists on theestafterpraecipuus, to correspond withaccedit, though it seems better to take all that comes afteracceditas an explanation of the statementmagis accedit oratorio generi: he also retains theetof most MSS. and readspraecipuus est. hunc et admiratus(Blätter f. d. bayer. Gymn. 24, p. 84). Wölfflin (partly followed by Krüger 3rd ed.) proposed a more radical change (Rhein. Mus. 1887, 2 H. p. 313)praecipuus. Hunc imitatus, quoting in support of the conjunctionimitatus ... secutus§122, eos iuvenum imitatur et sequitur industria:5 §19, deligat quem sequatur, quem imitetur: Ovid, Fasti v. 157, ne non imitata maritum esset et ex omni parte secuta virum. But Kiderlin (l.c.) aptly remarks that if Quintilian had writtenimitatus, he would not have saidut saepe testaturbutut ex multis locis patet(apparet, videmus): while vii. 4. 17 (on which Wölfflin relies) is not really to the point. Moreover Quintilian, would never have separated such synonyms asimitatusandsecutusbyut saepe testatur.Charisi nomini addicuntur, Frotscher:Charis in homine adducturaGH:Charisii nomine edunturPrat. Put. 7231, 7696 Harl. 2662 Dorv.§70.aut illa iudiciaPrat. Put. 7231, 7696 Harl. 4995. GH Harl. 4950 giveaut illa mala iudicia: Bodl. Burn. 243aut alia mala iud.S Harl. 2662 Dorv. and Ball.aut alia iudicia.The edd., following Gesner, have generally given (with Harl. 4950)aut illa mala iudicia(so Halm and Meister), and have takenmalaas predicate, though the order of the words makes that impossible. Becher approves of Andresen’s deletion ofmala. Krüger (3rd ed.) printsmala [illa] iudicia, thinking thatillaarose by dittography, and that then the order was changed in the codd. toilla mala iudicia. Kiderlin (in Hermes 23) gives as an alternative to deletingmalathe conjectureilla simulata iudicia(‘jene erdichteten nachgemachten Gerichtsverhandlungen’; cp. xi. 1. 56: cum etiam hoc genus simulari litium soleat). A similar mutilation occurs, e.g., xi. 1. 20, where b givessecumMsecusinstead ofconsecutum.§71.filiorum militum, most codd.:filiorum maritorum militumPrat. Put. 7231, 7696 S.§72.si cum venia leguntur. The reading of the MSS. is upheld by Iwan Müller, Meister, and Kiderlin. Spalding suggestedcum verecundia: Schöllcum iudicio: Bechercum ingenio. Becher points out (Bursians Jahresb. 1887) that the expression is meant to coverdecerpereas well aslegere, anddecerpereindicates careful and intelligent reading (cp.§69,diligenterlectus):cum ingenio= ‘mit Verstand’: cp. Cic. ad Fam. xiii. 10. 2 quod versabatur in hoc studio nostro .. et cum ingenio .. nec sine industria: Ulp. Dig. 1. 16. 9 patientem esse proconsulem oportet, sed cumingenio, ne contemptibilis videatur. Finally, Krüger (3rd ed.) proposescum acumineorcum vigilantia(cp. v. 7. 10).—Prat. Put. 7231, 7696 S Harl. 2662 all give Osann’s conjecturelegantur.praveGH Harl. 4995, 4950 Burn. 243 Bodl.:pravisRegius, Halm, Meister, Becher draws attention to the parallelism between the clauses:ut prave praelatus est sui temporis iudiciis, ita merito creditur(= meruit credi)secundus consensu omnium.§76.nec quod desit ... nec quod redundet: H Burn. 243 and Bodl. givequod .. quod: Prat. Put. MS Harl. 2662, 4995, 4950, Burn. 244, Dorv. C, and Ball,quid .. quid. The latter reading is supported by Becher (Phil. Rund. iii. 434). Forquodcp. xii. 10. 46: (xii. 1. 20 where forquod adhucBM givequid adhuc): on the other hand, in vi. 3. 5 the MSS. are in favour ofquid, though Halm readsquod(followed by Meister). Forquidcp. Cic. pro Quint. §41, neque praeterea quid possis dicere invenio.§77.grandiori similis. So all MSS.: Halm and Meister. Several conjectural emendations have been put forward. Comparing2 §16(fiunt pro grandibus tumidi), Becher suggestsgrandi oratori,—an easy change, if the copyist used contractions, but without point: above in§74, ‘oratori magis similis’ is appropriate enough in speaking ofhistorians, but ‘oratori’ would be inappropriate here. This is accepted, however, by Hirt (Berl. Jahr. ix., 1883, p. 312; cp. P. Hirt, Subst. des Adjectivums, p. 12). Schöll proposes to readgladiatorisimilis, in view of the close connection with what follows, strictus ... carnis ... lacertorum: butpleniorandmagis fususare a bad introduction togladiatori, and if Aeschines hadplus carnisandminus lacertorum, he cannot really have resembled a gladiator. This reading is, however, adopted by Krüger (3rd ed.). Finally, Kiderlin (Hermes 23, p. 166 sq.) has conjecturedet grandi(orgrandiori)organo similis, and applies the figure throughout: ‘voller und breiter lässt Aeschines den Ton hervorströmen, einem grossen Musikinstrumente gleich’: ‘einer Orgel gleich,’—he isgrandisonus. The translation appears to limit unnecessarily the meaning ofplenusandfusus: though the former is used of tone i. 11. 6 (cp. xi. 3. 15 of the voice: ib. §§42, 62: and §55 of the breath): whilefususis used of the voice xi. 3. 64. For such a use ofgrandiscp.§58(cenae):§88(robora): xi. 2. 12 (convivium): 3. 15 (vox): 68 (speculum): and fororganum, i. 10. 25: ix. 4. 10: xi. 3. 20 (where there is a comparison between the throat and a musical instrument): probably also i. 2. 30. There is an antithesis in the two parts of the sentence between fulness and breadth, on the one hand, and real strength on the other; and for the transition to the second figure Kiderlin compares§33.§78.nihil enim est inane: perhaps ‘nihil enim estin eoinane’ (Becher), ornihil enim inest.§79.honesti studiosus. Becher’s proposal to alter the punctuation of this passage is discussed in the notead loc.—Forauditoriisandcompararat, see ontenuia atque quae§44, above.§80.quem tamen. Kiderlin, in Hermes (23, p. 168), raises a difficulty here.Tamenshows that the clause cannot go with the main statement (fateor), and its position forbids us to take it with thequamquam is primumclause: it can only go withquod ultimus est, &c., ‘though Demosthenes isultimus fere, &c.,yetCicero, &c.’ To prevent so awkward a joining of the clauses, Kiderlin proposes to readeumque tamen: pointing out that thequaeof the MSS. (GH) may have arisen out ofque, and that Quintilian may have writteneumque; cp. vi. 2. 13, where Halm makesutqueout ofquae(G), and xi. 2. 32, where Meister readsestque. The meaning will then be: Demetrius is worthy of record as being about the last, &c., and yet Cicero gives him the first place in themedium genus.—It seems better, however, to givetamena general reference: ‘yet, in spite of all that can be said on the other side’ (e.g., inclinasse eloquentiam dicitur). Cp.§99quae tamen sunt in hoc genere elegantissima.§81.prosam(prorsam)orationem etall MSS.; Halm, Meister, Krüger (3rd ed.) omitet. I find that Becher supports the view stated in the notead loc.: he would however writeprorsam, which the best MSS. give also in Plin. v. 31, 112 D.quodam Delphici videatur oraculo dei instinctus: so Frotscher, followed by Krüger (3rd ed.). On the other hand Claussen (Quaest. Quint., p. 356) and Wölfflin (followed now by Meister, pref. to ed. of Book x., p. 13) propose to deleteDelphici, of which Becher also approves. But the MS. evidence cannot be disregarded. The following are the various readings: GHquaedam Delphico videatur oraculo de instrictus, and so FT, the former giving also (by a later hand)de instinctus, the latterdei instructus. Bodl. givesquodam delphico videatur oraculo dei instructus. The most frequent reading is that of Prat. Put. 7231, 7696 Harl. 2662, 4995, 4829, 11671, Ball. and most edd.,quodam delphico videatur oraculo instinctus: S agrees, but is reported to havedelphicoafteroraculo: Harl. 4950 and Burn. 244 have the same reading, withinstitutuscorr. toinstinctus: Burn. 243 givesinstructus.Delphicowas originally deleted by Caesar: Phil xiii, p. 758. Halm readtamquam Delphico videatur oraculo instinctus: but Quintilian would take no trouble to avoid the repetition ofquidam(cp. divina quadam, above).—For the arrangement of words, Krüger (3rd ed.) compares§41qui ne minima quidem alicuius certe fiducia partis memoriam posteritatis speraverit.§82.quandam persuadendi deam. Nettleship (Journ. of Philol., xxix, p. 22) conjecturesSuadam[persuadendi deam], comparing Brutus, §59, quotedad loc. Persuadendi deamwould thus become a gloss onSuadam: but the expression in the text is quite in Quintilian’s style.§83.eloquendi suavitate:eloquendi usus(orusu)suav.GH and all codd. except Harl. 4950, and Dorv., both of which give simplyeloq. suav.Halm admitted into his text Geel’s conj. forusus, ‘eloquendivi acsuavitate,’ and this has met with some acceptance (Iwan Müller and Becher). But the parallel from Dion. Hal.,Ἀρχ. κρ.4 is hardly conclusive:τῆς τε περὶ ἑρμηνείαν δεινότητος ... καὶ τοῦ ἡδέος. Hirt properly remarks that the agreement between the two is not so great as to allow of correcting the one by the other. Kiderlin conjectureseloquendi vi,suavitate,perspicuitate.tam est loquendi. See notead loc.for Kiderlin’s conj.tam manifestus est. Though Meister’stam est eloquendiis probably a misprint, it is found in some MSS.—Harl. 4950: Burn. 244.§84.sane non affectaverunt. Bodl. and Vall. (verusubpunctuated in the latter:affectantPrat. Put. 7231 MS Ball. Dorv. Harl. 2662, 4995, 4829, 11671:sene non adfectitacueruntGH Burn. 243:adfectarunt7696:adfectitantHarl. 4950, and so Burn. 244 (corrected fromaffectant).§85.haud dubie proximus. Halm insertedeiafterdubie, though it is not found in any MS.: Regius had suggestedilli. Kiderlin (Hermes 23, p. 170) points out that ifpropiores aliiin§88is allowed to stand without a dative,eiis not necessary here. He suggests, however,illibeforealiiin§88: both passages must be dealt with in the same way.—Forhaud(Vall.), GHS haveaut: Mhaut. Cp. on3 §26.§86.ut illi ... cesserimus:cum illiGHFT Harl. 4995 Burn. 243:ut illiPrat. Put. 7231, 7696: and so S Harl. 4950 (withcaelesti atque divinae):ut illeM Harl. 2662. Kiderlin (Hermes, p. 170) proposes to go back to the reading of the older MSS.cum illi, and instead ofcesserimusto readcesserit, so as to make Vergil the subject throughout.Cumcannot, he contends, be a copyist’s error, motived byita; and it is probable, therefore, that at firstcesserit awas inadvertently written forcesserit; then (in G or some older MS.)cesserimus itawas made out of that, to correspond withvincimurbelow: and then in the later MSS.cumwas changed tout, because ofita. For the transition, with this reading, from cesserit to the plural (vincimur, pensamus), hecompares§107, where, after speaking of Demosthenes and Cicero, Quintilian passes tovincimus.§87.sequenturMS Halm and Meister:sequenterGseq̅nt’H:sequunturPrat. Put. 7231, 7696.φράσινid est. These words are omitted in the Pratensis, which is Étienne de Rouen’s abridgement of theBeccensis, now lost. This is an additional proof thatφράσινwas originally written in Greek: cp. on§42.§88.propioresH Prat. Put. Vall. Harl. 2662, 4495, 11671, Burn. 243. Bodl., Halm:proprioresGMS 7231, 7696, Harl. 4950, C, Burn. 244, Dorv., Meister. In Cicero and Quintilianmagis propriiwould be more usual for the latter.§89.etiam si sit. This conjecture of Spalding’s (foretiam sitGH Bodl. &c.:etiam siM Harl. 4950 Dorv.:etiam sicPrat. Put. S Harl. 2662) I have found in the Balliol codex. 7231 and 7696 giveetiam si est. Cp. note ontenuia atque quae§44, above.ut est dictum. These words were bracketed as a gloss by Halm, and are now omitted altogether by Krüger (3rd ed.): see however notead loc.Döderlein proposed to place them afterpoeta melior, Fleckeisen afteretiam si.Serranumis Lange’s conjecture forferrenumGHM:farrenum7231, 7696 Harl. 2662, 11671:PharrenumPrat. Put. Some MSS. (e.g. Vall. Harl. 4995, Burn. 243 and 244) givesed eum, but it is obvious that the criticism of Severus stopped with the wordlocum.§90.senectute maturuited. Col. 1527 and so 7231, 7696 (Fierville):senectutem maturbitGH:senectute maturumPrat. Put. MS Harl. 2662, 4995, 4950, Burn. 244, Dorv. and Ball.:senectus maturavitBodl., Burn. 243.et, ut dicam. Halm’ssedinstead ofethas been rejected by later critics. Cp. Claussen (Quaest. Quint., p. 357 note):sed‘sententiam efficit ab hac operis parte alienam. Nam cum oratori futuro exempla quaerantur oratoria virtus in quovis scriptore laudi vertitur (§§46, 63, 65, 67, 74, &c.). Itaque propter huius censurae consilium Quintilianus Lucani elocutionem oratoriam laudat, sed ingenium poeticum una reprehendit.’§91.propiusH Prat. Put. Burn. 243, Harl. 2662 and other codd.: Bodl. Ball. Harl. 4950proprius. Reisig conjecturedpropitius, which also is apt; but in spite ofindustrius,necessarius, cited in its support (cp. iv. 2. 27: vii. 1. 12), it is too uncertain a form to be received into the text. Iwan Müller thinks it would have to bemagis propitiae. Halm givespromptius: Wölfflinpronius: while Schöll now suggestspropitiae potius(cp. iv. pr. §5: 2 §27: vii. 1. 12).§92.feresG Prat. Put. 7231, 7696 S Harl. 2662, 4829, Dorv., Ball., Halm.:ferasH, Harl. 4950, Burn. 243, Bodl. C and M, Meister and Krüger (3rd ed.). Harl. 4995 hasfere: from Vall. Becher reports feras, ‘probably at firstferes.’elegeaGH 7696, and so A2BN Put. S at i. 8. 6.§94.abunde salisG Prat. Put. M and all my MSS. except H, Burn. 243, Bodl. which haveabundantia salis.multum est tersior. The variety of MS. readings seems to point to anetwrongly inserted aftermultum, perhaps from a confusion with ‘multum et ver gloriae’ below. GH givemultum et est tersior: M Harl. 4950, Bodl. Ball. C Dorv. Burn. 243 and also Harl. 4829multum etiam est t.: Prat. Put. 7231, 7696 S Harl. 2662, 11671multum est tersior: while Harl. 4995 (and Vall.) hasmulto et est tersior. Osann proposedmulto eo est tersior: Wölfflinmulto est tersior: Halm and Meister printmultum eo est tersior. Formultum, cp. multum ante xii. 6. 1: and see Introd.p. li.non laborGH Burn. 243 Bodl. and Meister:nisi labor7231, 7696 S Harl. 2662, 4995, 4950, 4829, 11671, Burn. 244, Dorv. Ball. C, and Halm. Prat. and Put. havemihi labor.hodieque et qui: H, Prat., Put., 7231, 7696, Harl. 2662, 4829, Bodl. Dorv.:hodie et quiBurn. 243:hodie quoque et quiVall. Harl. 4995, 4950:hodie quod et quiS.—Becher is of opinion that the text will not bear the explanation given in the note, and would readhodie quoque et qui: ‘es giebt auch heute noch berühmte Satirendichter, die einst &c.’Et quihe takes withclari, not withhodie quoque, theetbeing omitted in translation: clari (hodie quoque) qui (olim) nominabuntur.§95.etiam prius. Founding on the classification given in Diomedes (see notead loc.), according to which thesaturaof Pacuvius and Ennius preceded and was distinct from that of Lucilius, Horace, and Persius, Claussen (Quaest. Quint., p. 337) thinks that the true reading here may beAlterum illud et iam priusEnnio temptatumsaturae genus, &c. For the satura of Ennius, cp. ix. 2. 36. Iwan Müller points out that Ennius is not mentioned below (§97), beside Attius and Pacuvius, probably because neither in tragedy nor in satire did Quintilian consider him to have produced anything helpful for the formation of an oratorical style. Other unnecessary conjectures areetiam posterius, Gesner:etiam proprium, Spald.:etiam amplius, L. Müller:etiam verius, Riese:alterum illud Lucilio prius sat. genus, Krüger (3rd ed.).sola:solumPrat. and Put.collaturus quam eloquentiae. These words, omitted in GHS Bodl. Burn. 243, occur in all my other codd.§96.sed aliis quibusdam interpositus: sc. carminibus, Christ. In H the reading isquibusdam interpositus: so 7231, 7696 Bodl. and Burn. 243: but M Harl. 4950, 4829 Burn. 244 Dorv. and Ball, givea quibusdam interpositus: Scuiusdam: Prat. and Put.opus interpositus. Osann conjecturessed quibusdam, and so Hild. In the margin of Harl. 4995 is the variantaliquibus interpositis.In Hermes, vol. 23, p. 172, Kiderlin makes a fresh conjecture. Recognising that something must have fallen out beforequibusdam, but dissatisfied with Osann’ssedand Christ’ssed aliis, he proposes to readut proprium opus, quibusdam aliis tamen carminibus(orversibus)a quibusdam interpositus. The eye of a copyist may easily, Kiderlin thinks, have wandered from the first to the secondquibusdam: cp. v. 10. 64, ut quaedam a quibusdam utique non sunt, &c., and for quibusdam aliis xi. 3. 66, et quibusdam aliis corporis signis.intervenit, which is a conjecture of Osann, I have found in Harl. 2662, 11671 Prat. Put. 7231, 7696.lyricorum. Kiderlin thinks there may be something wrong in the text here. The last sentence (sed eum longe, &c.) shows clearly that Quintilian had a high opinion of the lyrists of his day: if Bassus waslegi dignus, they were even more so. Would he then have said ‘of the Roman lyrists Horace is almost the only one worth reading’? Perhaps we should readlyricorum priorum: after-ricorum,priorummight easily fall out, and it gives a good antithesis toviventium. Bassus (quem nuper vidimus) forms the transition: and the next paragraph beginsTragoediae scriptores veterum, &c.§97.clarissimi. This reading is stated by Halm to be ‘incerta auctoritate,’ and is referred by Meister to the Aldine edition. It occurs in Prat. 7231, 7696 Harl. 2662 (A.D.1434) Vall. 4995, 4829, 11671, Dorv. and Ball.: Put. givesclarissime: G hasgravissima: HFTSgravissimus, and so also Harl. 4950, Burn. 243, Bodl. and C. Halm printsgrandissimi: Ribbeck (Röm. Trag. p. 337, 3) inclines to accept the sing.grandissimus, M, of Pacuvius alone.Kiderlin (in Hermes 23, p. 173) rejects all the above readings.Gravissimusandgravissimaare obviously due, he says, togravitatefollowing: but the word beforegravitatemust have begun with the same letter, and soclarissimicannot stand, especially as it is inappropriate to the context. Forceterumshows that the sentence before it must have contained some slight censure: some defect, or quality excluding others equally good, must have been mentioned. He therefore conjecturesgrandes nimis, in preference tograndissimi, which in tragedy would hardly be a fault. Attius and Pacuvius, Quintilian says, are ‘zu grossartig, sie kümmern sich zu wenig um Zierlichkeit (Eleganz) und die letzte Feile (d.h. Sauberkeit im Kleinen); doch daran ist mehr ihre Zeit schuld als sie selbst.’ He evidently thinks more of the ‘Thyestes’ of Varius and Ovid’s Medea: cp. Tac. Dial. 12. With this judgment Kiderlin compares§§66,67tragoedias primus in lucem Aeschylus protulit, sublimis et gravis et grandiloquus saepe usque ad vitium, sed rudis in plerisque et incompositus ... sed longe clarius inlustraverunt hoc opus Sophocles atque Euripides, and is of opinion that the parallelism cannot be mistaken. For the position ofnimishe compares ix. 4. 28 longae sunt nimis: v. 9. 14 longe nimium: xii. 11. 9 magna nimium.§98.quem senes quidem parum tragicum. So Spalding, Bonnell, Halm, Meister, and Krüger.Quidemoccurs in no MS.: GH havequem, M Vall., Harl. 4995, Burn. 244, Ball, omit it: Bodl. Burn. 243 and Dorv. show the corruptionPindarum. Becher would excludequidem, regardingquemin G as an instance of the tendency of copyists inadvertently to repeat, after a particular word that by which it has been immediately preceded, e.g.§68quod ipsum quod (G): ix. 4. 57 ut cum ut (G): iv. 1. 7 ipsis litigatoribus ipsis (b): iv. 2. 5 aut ante aut (bT): x. i. 4 iam opere iam (G).—But here the authority of the Pratensis and its cognates may be invoked. In the archetype from which they are derived something must have stood beforeparum, as Prat. Put. 7696, 7231 all givequem senes non parum tragicum: so Harl., 2662 (A.D.1434), and 11671. Above in§96, G Prat. Put. 7231, 7696 havesi quidemforsi quem.§100.linguae suae. So Köhler (v. Meister pref. to Book x. p. 13):suaesupplies an antithesis to ‘sermo ipse Romanus’: GH givelinguae quae: so Harl. 4950: S Burn. 243, Bodl.linguae: while Harl. 2662, 4995, 4829, 11671, Dorv. and Ball. omit it altogether: M hasligweque.§101.Titum: GH Prat. Put. M. 7231, 7696.commendavit: Halm and Meister givecommodavit, which is approved also by Hirt. Halm compares§69where Menander is said to be ‘omnibus rebus personis adfectibus accommodatus.’ But this would require the meaning ‘appropriately treated,’ and there is no instance in Quintilian of the verb used absolutely in this sense. Nor is there any example to support Hild’s interpretationpraestitit, which would be moreover extremely weak. The recurrence of the word so soon afteraccommodatatells against Halm’s reading, though Quintilian is negligent on this head.—On the other hand, in vi. 3. 14 the reading ‘ad hanc consuetudinem commodata’ is rightly accepted against ‘commendata’ most edd.§102.immortalemGS Meister:illam immortalemPrat. Put. M Halm:immortalem illamVall.velocitatem. So all MSS, except S, Burn. 243, and Bodl., which havecivilitatem. Kiderlin (in Hermes 23, p. 174) thinks that we might have expectedideoque immortalem gloriam quam velocitate Sallustius consecutus est: ‘und darum hat er dievelocitasdurch (von der velocitas) verschiedene Vorzüge erreicht.’Consequicannot mean ‘to supply the place of’: andimmortalisis inappropriate as an attribute ofvelocitas: besides, Quintilian has not spoken of Sallust’svelocitas, even indirectly. Schlenger conjecturedclaritatem: Andresenauctoritatem(‘klassisches Ansehen,’ cp. iv. 2. 125: xii. 11. 3): Kiderlin now proposesdivinitatem, which in Cicero = Vortrefflichkeit, Meisterschaft: cp. xi. 2. 7. Judged by the previous sentences the expression is not too strong. Forimmortalem divinitatemcp.§86illi ... caelesti atque immortali: and forconsecutus estiii. 7. 9 quod immortalitatem virtute sint consecuti.clarus vi ingenii. This is a conjecture of Kiderlin’s, which I find has been adopted also by Krüger (3rd ed.). GHFT giveclarius ingenii: Prat. Put.clari ingenii vir: 7231, 7696clari vir ingenii: MS Harl. 4995, 4950, 4829, Burn. 243 and 244, Dorv.C and Ball,clarus ingenio; Harl. 2662 and 11671clarus(?) orclaret vir ingenii. Spalding had already pointed out thatclarusis not found withingenium, except whereingeniumis used of a person: e.g.§119erant clara et nuper ingenia: he therefore wroteelati vir ingenii(following Goth.elatus ingenioand Bodl.elatus ingeniis). Kiderlin compares§70sententiis clarissimus, and forvis ingeniii. pr. 12: ii. 5. 23: x. 1. 44: xii. 10. 10. The readingclarus vi ingeniipoints the contrast to what follows in ‘sed minus pressus,’ &c.: it was hisstylethat did not altogether suit the dignity of history.§103.genere ipso, probabilis in omnibus, sed in quibusdam. Till Kiderlin made this happy conjecture (see Hermes 23, p. 175)generehad always been joined withprobabilis, and the text was twisted in various directions. GHS, Burn. 243, Bodl. givein omnibus quibusdam: M Harl. 2662, 4995, 4950, Burn. 244, Dorv.in omnibus sed in quibusdam, and so apparently Prat. Put. 7231, 7696. Out ofomnibusHalm gives on Roth’s suggestion,operibus: afterwards he decided forpartibus, and this (thoughomnibustopartibusis not an easy transition) is adopted by Meister. Kiderlin’s punctuation makes everything easy: ‘Anerkennung verdienen seine Leistungenalle,manchestehen hinterseinerKraft zurück.’ Even these last, Quint. means, areprobabiles(cp. viii. 3. 42 probabile Cicero id genus dicit quod non plus minusve est quam decet); but they do not show the great powers that distinguish his other writings. It is uncertain whether Quintilian wrotein quibusdamorsed in quibusdam(M). The easiest explanation of the omission in the other MSS. is to suppose that he wrotein omnibus in quibusdam: perhaps the copyist of M saw thatomnibusandquibusdamwere antithetical, and insertedsed. Kiderlin notes Quintilian’s liking for chiasmus, without any conjunction: cp.§106in illo, in hoc (where in hoc is wanting in M).suis ipse viribus: ed. Col. 1527 (Halm), and so (Fierville) 7231, 7696. In Harl. 2662 and 11671 (A.D.1434 and 1467)suisalready appears, corrected fromvisGH. The Juntine ed. (1515) hassuis viribus minor: so Prat. and Put.§104.et exornat. Vall. and (apparently) Prat. Put. 7231, 7696, and most edd.:et ornatM Halm, Meister, Krüger:exornatGHS. Becher remarks thatet exornatmight easily pass intoexornat.nominabitur: Weber and Osann proposednominabatur(which appears in Harl. 2662, but corrected to-itur). Krüger at first accepted this in support of his theory that the whole passage refers to Cremutius, who ‘in former days (olim), while his works were under a ban, was only named (i.e. was a mere name, but now is known and appreciated).’ The parallel passage (§94) is sufficient to dispose of any such interpretation: sunt clari hodieque et qui olim nominabuntur.Cremuti. Nipperdey, Philol. vi, p. 193, Halm, and Meister:remutiH Prat. Put. 7231, 7696remremutiG,rem utiliBurn. 243:remittiS. Bodl.:nec imitatores utiHarl. 2662, 4995, 4950, 4829, 11671. A review of the various explanations of the whole passage (Superest—quae manent) will be found in Holub’s Programm ‘Warum hielt sich Tacitus von 89-96 n. Chr. nicht in Rom auf?’—Weidenau, 1883: but his conjectureremoti(i.e. relegati) forremutiis not to be thought of.dividendi: first in the Aldine edition: all MSS. havevidendi, except M (indicendi) and Prat. Put. Harl. 4995 (vivendi). Cp. i. 10. 49, where the case is the same.§105. In the Aurich Programm, Becher gives a more recent statement of his views: ‘wie zucumcausale, so tritt praesertim auch zucumconcessivum, in diesem Falle wiedenzugeben mit, “was um so auffallender ist, als.” Der Sinn ist also: “Ich weiss sehr wohl, welchen Sturm des Unwillens ich gegen mich errege, und dies (dieser Sturm) ist um so auffallender, als ich jetzt gar nicht die Absicht hege, meine (in Potentialis gesprochene) Behauptung (fortiter opposuerim) wahr zu machen, resp. comparando durchzuführen. Ich lasse ja dem Demosthenes seinen Ruhm—in primis legendum vel ediscendum potius.”’§106.praeparandi. For Kiderlin’s conj.praeparandi,narrandi,probandiseead loc.[omnia]denique, GH, Burn. 243, Bodl. omitomnia(which is in all my other MSS.), and Meister now approves (following Spalding, Osann, and Wölfflin), on the ground that Demosthenes and Cicero werenotalike ineverythingthat belongs toinventio. Halm thinks thatomniais to be found inracioniof the older MSS.: but Kiderlin points out that this error may have arisen from the carelessness of a copyist who, after thrice writing the terminationi, gave it also to the fourth word.illi—huicPrat. M, S Vall. Harl. 4995, 2662 Bodl. &c.:illic—hicGH Put. 7231, 7696, Halm.§107.vincimus, H, G2, and most MSS.: (cp.§86):vicimusG.§109.ubertateHarl. 4995. This is also the reading of codd. Vall. and Goth.:all the other MSS. giveubertas.totas virtutesBn Bg N Prat. Ioan. 7231, 7696:totas viresM b.§112.ab hominibusHalm and Meister:ab omnibusBn Bg HFT Ioan. Prat. 7231, Sal. and most codd.:hominibusS Harl. 4995 Bodl.§115.urbanitas. Kiderlin proposes to readet praecipua in accusando asperitas et multa urbanitas: cp.§117:§64:2 §25: ii. 5. 8.Ciceroni, forCiceronemof the MSS. In the Rev. de Phil. (Janv.-Mars, 1887) Bonnet quotes from the Montpellier MS. a note of the sixteenth century deleting the name as a gloss (oninveni). Certainly all codd. giveCiceronem, notCiceroni. Bonnet thinks that the insertion does not accord with Quintilian’s habitual deference towards Cicero: ‘Quintilien se trouvant dans le cas de contredire Cicéron ne le nomme pas.’—Becher reportsCiceroni, a correction in the Vallensis.castigata, B (i.e. Bn and Bg) Ioan. Prat. 7231, 7696 Harl. 2662, 4995, 11671:custoditaH M b F T Alm. Harl. 4950, 4829, Burn. 243, 244, Bodl. Dorv. and Ball. Forgravis(bH M Vall. and seemingly Prat.) B Sal. 7231, 7696 and Ioan. givebrevis.si quid adiecturus sibi non si quid detracturus fuit, Vall. Harl. 4995. For the repetition, see on haud deerit3 §26. Halm and Meister printsi quid adiecturus fuit—(sc.virtutibus suis, cp.§§116,120)—the reading of B (i.e. Bn and Bg), which is also that of Ioan. Prat. N 7231 Harl. 2662, 11671: while M Harl. 4950, 4829, Burn. 244 havesi quid adiecturus fuit, non si quid detracturus. The reading of H issi quid adiecturus sibi non si quid detracturus[Sulpicius insignus]fuit ut servius sulpicius insignem&c.: so also T, Burn. 243, Bodl. The brackets in H are by a later hand, indicating a gloss which arose from a mistake made by the copyist of H. In Bg the passage stands:—sibinon siquiddetracturussi quid adiecturus:fuitetserviussulpiciusThe words added above the line are by the hand known as b.In copying H wrote:si quid adiecturus sibi non si quid detracturus(then omittingfuitcontinues)et Serv. Sulp.(then goes back and resumes)fuit et servius&c. This is the origin of the confusion which exists in all the MSS. of this family.§117.et fervor. This is Bursian’s conjecture, adopted by Halm and Krüger (3rd ed.), and now approved by Becher. BM haveet sermo, which is also the reading of N Prat. Sal. 7231, 7696 Ioan. Harl. 2662, 4950 and Ball.: Hbet summo: Harl. 4829, 11671, Burn. 244et smo: while Bodl., Dorv., and Burn. 243 give the correction in Teius summa, out of which the second hand in the Vallensis (Laurentius Valla) madeet vis summa, a reading which occurs also in Harl. 4995. Meister readset sermo purus; while Kiderlin proposeset simplex sermo(cp. iv. 1. 54: viii. 3. 87: ix. 3. 3: 4. 17: viii. pr. 23: x. 2. 16).ut amari sales. Francius conjecturedut amantur sales, but this loses the antithesis betweenamariandamaritudo ipsa. Kiderlin’sut amantur amari sales(viii. 3.87: vi. 1. 48) is an improvement; but ifridiculais taken in a good sense it seems impossible that after censuring Cassius for giving way unduly tostomachus, Quintilian should go on to say, ‘moreover, though bitter wit gives pleasure, bitterness by itself is often laughable.’ Is it possible that we ought to readut amari sales risum movent ita amaritudo ipsa ridicula est? Such an antithesis might have been written ‘per compendium,’ and the wordsrisum moventmay then have dropped out. See the notead loc.: and cp. especially vi. 1. 48fecit enim risum sed ridiculus fuit, andοὐ γέλωτα κινεῖ μᾶλλον ἢ καταγελᾶται, quoted in the note on1 §107.—Krüger (3rd ed.) adoptsfrequentiorforfrequenter, which gives a good sense, except thatfreq. amar ipsais awkward.§121.leneHalm and Meister:leveB Prat. N 7231 M 7696 C. Here again Becher prefersleve, comparing Cic. de Orat. iii. §171, quoted on§44above: levitasque verborum 1. 52: and levia ... ac nitida, v. 12. 18.§123.scripserint. So Bn Bg H Ioan. Prat. 7231, 7696 Vall. Harl. 4995, 2662, 11671, Bodl., Dorv., Spalding, and Bonnell. Becher compares among other passages2 §14(concupierint), and points out that Quintilian is not thinking of individual writers on philosophy, but of the class, as opposed to the class of orators, historians, &c.—Halm, Meister, and Krüger havesupersunt(Put. M, Ball. Burn. 243 Harl. 4950).§124.Plautus, Prat. N, 7231 Ioan. Harl. 2662, 4829, 11671:plantusM Harl. 4950:PlantatusSal.:plaustusHb:Plancusedd. vett. and Harl. 4995.Catius. The name is rightly given in Harl. 4995.§126.iis quibus illi.Iisis the conjecture of Regius, followed by Halm, Meister, and Krüger. Becher would retainin quibus illi,—the reading of BN Prat. Ioan. Vall. M Harl. 4995, 2662, 4950, 11671, Burn. 244 Dorv. Ball. The difficulty of construing probably led to the omission ofinin bH Bodl. Burn. 243, 7231, 7696, Spalding and Bonnell.ab illoB Ioan. 7231, 7696 Sal. Harl. 2662, 4950, 4829:ab eobHM Burn. 243.§127.foret enim optandum:fore enim aliquid optandumbHFT. Spalding conjecturedalioqui optandum, which Kiderlin approves.ac saltemall MSS.: Meister hasaut saltem, probably relying on a wrong account of the Bambergensis: see Halm vol. ii, p. 369.illi viroB:illi virusbHM:illi virtutibusHalm:illi viro eos(orviro plurimos) Kiderlin.§128.multa rerum cognitio: so all codd. except Ioannensis and Harl. 4995, which havemultarum rerum cognitio. b omitscognitioand is followed by HFT.§130.si obliqua contempsisset, si parum recta non concupisset. I adopt the reading recently proposed for this vexed passage by Ed. Wölfflin in Hermes, vol. xxv (1890), pp. 326-7, though it is right to note that he was partly (as will be seen below) anticipated by Kiderlin.Obliquaseems thoroughly appropriate in reference to Seneca’s unnatural, stilted, affected style,—‘jene unnatürliche, durch unmässigen Gebrauch von Tropen und Figuren auf Schrauben gestellte Ausdrucksweise, welche statt der Klarheit ein Schillern zur Folge hat.’ Wölfflin compares ix. 2. 78rectum genusadprobari nisi maximis viribus non potest: haec diverticula et anfractus suffugia sunt infirmitatis, ut qui cursu parum valent flexu eludunt, cum haec quae adfectatur ratio sententiarum non procul a ratione iocandi abhorreat. Adiuvat etiam, quod auditor gaudet intellegere et favet ingenio suo et alio dicente se laudat. Itaque non solum si persona obstaretrectae orationi(quo in genere saepius modo quam figuris opus est) decurrebant ad schemata ... ut si pater ... iacularetur in uxoremobliquissententiis. This passage supplies (what is indeed suggested byobliquaitself) the antithesisparum recta: cp. ii. 13. 10 si quis ut parum rectum improbet opus.In theJahrbücher f. Philologie(vol. 135, 1887: p. 828) Kiderlin had previously dealt with the passage on similar lines. The traditional readingsi aliqua contempsisset(b) he considers too indefinite, though not impossible: in point of authority, though preferable to thesi nil aequalium cont.of the later MSS., it cannot rank so high as the reading of Bn and Bg, which givesimile quamwithout any attempt at emendation. This Kiderlin thinks must be nearest the original: he therefore rejects such conjectures as Jeep’ssi antiqua non, on the ground that it is improbable thatsimile quamarose out ofantiqua. He introduces his own conjecture by referring to ix. 2. 66 and 78 (see above), and to the contrast betweenschemataandrectum genus,recta oratio; the former are calledluminaorlumina orationis(xii. 10. 62). Cp. viii. 5. 34. He would read:nam si mille ille schemata(orillas figuras)similiaque lumina contempsisset, si parum rectum genus(orsermonem)non concupisset, &c.Similiaqueoccurs ix. 4. 43:mille(forsescenti) is used v. 14. 32: forcontempsissetcp. ix. 4. 113.Si mille illaandsimiliaquemay easily have run together, whenschemata(orfiguras) would fall out:quamin the older MSS. may representque lumina, which again reappears in thequaliumof the later codd. (si nil aequalium). As an alternative forparum rectumgenns(orsermonem) Kiderlin suggests Wölfflin’s readingparum recta: and compares ix. 2: ii. 5. 11: v. 13. 2: ix. 1. 3; 3. 3: x. 1. 44; 89: ii. 13. 10.Of the MSS. Prat. 7231 Sal. 7696 N Ioan. Harl. 2662 and 11671 agree with Bn and Bg in givingsimile quam: b hassi aliqua: HFT, Burn 243, Bodl.aliqua: M Harl. 4995, 4950, 4829, Burn. 244, Dorv. Csi nil aequalium. Among previous conjectures aresi multa aequalium, Törnebladh:si ille quaedam, Halm (whereilleis surely superfluous):si antiqua non, Jeep. Meister accepts the readingsi aliqua non: Becher thinks thatsi nil aequaliummay be right.It is generally admitted that a word must have fallen out afterparum: the codd. all givesi parum non concupisset. Jeep proposedsi pravum(=corruptum: cp. ii. 5. 10)non conc.: on which Halm, comparingomnia sua, remarks, ‘debebat saltemprava.’ Butpravaseems too strong a word for Quintilian to have used in a criticism where he is so studiously mixing praise and censure. Halm suggestedsi parum sana, and is followed by Meister: cp. Fronto’s ‘febriculosa’ of Seneca, p. 155n. Sarpe proposedsi pravaorparvaorplura: Buttmannsi parum concupiscenda(orconvenientia): Herzogsi parvum: Madvigsi partimorpartem(i.e.paulo plus quam aliqua, and in opp. toomnia sua, below): Hoffmannsi opiparum: Seyffertsi garum: Kraffertsi non parum excussisset(cp.§101,§126: v. 7. 6; 7. 37; 13. 19: xii. 8. 13, &c.): Gustaffsonsi parva(cp. i. 6. 20 frivolae in parvis iactantiae): Andresensi similem ei quem contempsit se esse(sc.concupisset; cp. Tac. Ann. xiii. 56: xii. 64: Hist. i. 8: Livy xlv. 20. 9)si parem non concupisset(i.e.si Ciceronianum genus dicendi imitari quam diverso genere gloriam eius aemulari maluisset): or,nam si similem ei quem contempsit se esse, non parem concupisset: Krüger (3rd ed.)si parum arguta: Hertz (who argues that the word which has fallen out must, withparum, correspond tocorruptaabove)si parum pura.utrimqueMeister and Becher, following old edd., Spalding, and Bonnell:utrumqueB N 7231, 7696:virumqueM:utcumqueHalm, ‘in every way,’ ‘one way or another,’—proposed by Gesner at6 §7.CHAPTER II.§2.atque omnis. Kiderlin (Blätter f. d. bayer. Gymn. 1887, p. 454) proposes to put commas atsequiandvelimus, and make this clause also subordinate.§3.aut similes aut dissimiles. Andresen suggestsaut similes aut non dissimilesoraut similes aut certe haud dissimiles.§6.tradiderunt(BNM Harl. 2662, 4995, 4829, Burn. 243, and Dorv.) is powerfully supported by Becher in his latest tractate (Programm des königlichen Gymnasiums zu Aurich, p. 13) againsttradiderint, the reading of b Prat. Bodl. and Vall. (corrected in the last fromtradiderunt), Burmann, Spalding, Bonnell, Halm, Meister, and Krüger. Becher holds that in Quintilian, as frequently in Cicero,cumwith the indicative is often used in such a way (quoting from C. F. W. Müller) ‘ut non prorsus idem sit, sed simillimum ei, quod barbare dicere solemus identitatis. Nam ut “cum tacent clamant” non est “si tacent,” multo minus “quo tempore” aut “propterea quod” aut “quamquam,”—sed “tacent idque idem est ac si clament,” sic “cum hoc facis qui potes facere illud?” et sim., German, item “wenn du dies thust” valet: “hoc facis ex eoque per se efficitur, non ratione, sed ipsa natura, ut illud non possis facere.” Ut pro Q. Roscio 3. 9 quam ob rem, cum cetera nomina in ordinem referebas, hoc nomen in adversariis relinquebas? non significat nec “quamquam” nec “quando,” sed “wenn.”’ Becher adds the following parallel passages: Cic. pro Cluent. 47. 131 id ipsum quantae divinationis est scire innocentem fuisse reum, cum iudices sibidixeruntnon liquere, and Verg. Ecl. 3. 16 quid domini facient, audent cum talia fures? (Cp. Madvig de Fin. p. 25.) In the same way he treatscum ... sunt consecuti7 §19below, which seems, however, to be somewhat different. Here there is an antithesis, and in such casescum(‘whereas’) may very well take the indicative: there the clause ‘cum sint consecuti’ is added to show the reasonableness (cum= ‘since’) of the demand that extemporary facility shall be made fully equal tocogitatio—seead loc.Neither instance can be explained on the analogy ofcumwith the indic. used of ‘identity’ (as ‘cum tacent, clamant,’ quoted above): in such cases the subject is generally the same in both clauses. And in such a passage as pro Cluent. §131cumis usually explained as =quo tamen tempore.eruendasM Harl. 4995: all other codd.erudiendas.mensuris ac lineis. Krüger (3rd ed.) quotes with approval the conjecture of Friedländer (Darst. aus der Sittengesch. Roms iii. 4. p. 194. 4)eisdem mensuris ac lineis, and recommends the insertion ofeisdemin the text,—afterlineis, where it is more likely to have fallen out. But this is unnecessary.§7.turpe etiam illud est. Hild puts a comma aftersciant, and by supplying beforeturpe estanitato correspond withquemadmodum, makes out a comparison of whichquemadmodum, &c., is the first clause andturpe etiam illud estthe second. This is certainly to misunderstand the passage. Thequemadmodumclause goes with what is before, not with what follows, so that a comma afteralieniwould be enough, were it not for the necessity of having the mark of interrogation (cp.§9below). Thenturpe etiam illud estcomes in, resumingpigri est ingeniiin§4, just as immediately afterwardsrursus quid erat futurum§7resumesquid enim futurum erat§4. The whole passage is an elaboration of the dictum with which§4opens, ‘imitatio per se ipsa non sufficit.’ Quintilian first says that we, as well as those who have gone before us, may make discoveries (cur igitur nefas est reperiri aliquid a nobis quod ante non fuerit?). Surely we arenotto confine ourselves to hard and fast lines like servile copyists.Then he goes on to add in§7that we must surpass our models (plus efficere eo quem sequimur), instead of resting content with mere reproduction (id consequi quod imitamur): otherwise Livius Andronicus would still be the prince of poets, we should still be sailing on rafts, and painting would still be nothing more than the tracing of outlines. The necessity for progress is first shown (§§4-6) by an appeal to the example of the past, and by the unfruitful work of such painters as are mere copyists: then in§7poetry, history, navigation, as well as painting are put in evidence for the argumentex contrario.§8.mansit, Meister:sitcodd.:estFleckeisen (and Halm):fuitGensler.§9.adpetentBg HFT:appetentPrat. Ioan. Harl. 4995 Bodl. &c.:appetuntN Harl. 2662, 11671, Burn. 243.hoc agitHalm, followed by Meister (cp.7 §4):hoc aitb H,om. Bn Bg N Ioan. Prat. Harl. 2662, 11671:agit(sine hoc) Harl. 4995, 4950 M, and most codd.§10.quaeque pares maximemay be a gloss: it is found only in those MSS. which givesimplicissimaeforsimillimae: b H Harl. 4950 M Burn. 243 Bodl.utique(b M Vall. Harl. 4995, 4950, Burn. 243 Bodl. Dorv.) may also be suspected: it does not occur in Bn Bg N Ioan. Prat. Harl. 2662, 4829, 11671.§11.orationibus, Bg: Ioan, givesoratione: so also Voss. 1 and 3 (Zumpt).accommodaturb H Ioan. Harl. 4995, 4950, 4829, Bodl. Dorv. and Meister:commodaturBn N Prat. Harl. 2662, 11671, and Halm.§12.inventio visB Harl. 2662, 11671:inventionisb H Harl. 4495, 4950, 4829, C, Burn. 243, Bodl., Dorv.§13.cum et, ed. Colon. 1527:et cumB H Ioan. Prat. N (et quum) M:cumVall. Harl. 4995. On the usual interpretation of this difficult passageut quorum ... collocata suntforms one parenthesis: but this is an unnecessary extension of the explanation ofintercidant invalescantque temporibus. Seead loc.accommodata sit, codd. except Harl. 4995, which omitssit:acc. estHalm, followed by Hild (depending onprout, notcum: see notead loc.). Madvig’s conjectureaccommodanda sitis approved by Kiderlin (cp. ix. 4. 126 adeoque rebus accommodanda compositio). But the correctness of the reading in the text (and also of the explanation given in the notead loc.) will be evident to any one who considers the whole sentence carefully. Tocum et verba intercidantcorresponds exactly the double clauseet compositio ... rebus accommodata siton the one hand, andet compositio ... ipsa varietate gratissima(sc.sit—repeated fromaccommodata sit) on the other. This double clause is rather awkwardly joined bycum ... tum. To takeaccommodata sitas depending on thecumwhich followscompositiois to destroy the balance of the sentence. In this case an independentsitwould have to be supplied withgratissima(to makeet compositio ... gratissima sitcorrespond toet verba intercidantabove): and the translation would then be: ‘it is just when (cum ... tum), or exactly in proportion as, it is adapted to the sense (rebus accommodata) that the very variety (thereby secured) gives the arrangement its greatest charm.’ But if this had been Quintilian’s meaning he would surely have writtencum rebus accommodatur(or—ata est)tum ipsa varietate sit gratissima.§14.quos imitemur. The D’Orville MS. givesquos eligamus ad imitandum,—probably an emendation by the copyist, though it may explain the origin of the reading of b and Hquos at imitandum.quid sit ad quod nos. Theadis due to Regius: most codd. havequid sit quod nos, except Harl. 4995, which is again in agreement with Goth. Vall. Voss. 2 and the second hand in Par. 2:quid sit quod nobis.§15.et a doctis, inter ipsos etiam. The explanation given in the notes is due to Andresen (Rhein. Mus. 30, p. 521), who, however, wished to insertetbeforeinteripsos. The comma makes that unnecessary. So Kiderlin (Berl. Jahrb. XIV, 1888, p. 71 sq.).dicunt, Harl. 4995:dicantall codd.: ‘emend. Badius’ (Halm).ut sic dixerimVall. (Becher): cp. pr. 23: i. 6. 1: ii. 13. 9: v. 13. 2. BM Prat. haveut dixerim. Halm wroteut ita dixerim, comparing i. 12. 2: ix. 4. 61: butut sicis more common in the Latinity of the Silver Age.§16.compositis exultantes. Kiderlin (Berl. Jahrb. XIV, 1888, p. 72) would prefercompositis rigidi(cp. xi. 3. 32: xii. 10. 7: ix. 3. 101: xii. 10. 33),comptis(cp. i. 79: viii. 3. 42)exultantes= ‘statt wohlgeordnet steif, statt schmuckliebend putzsüchtig.’ Another unnecessary emendation islaetis exultantes, compositis corrupti(Lindau): orcompositis exiles(Düntzner).§17.quidlibet, most codd.:quamlibetM, Vall. Harl. 4995, 4950:qui licetbH. Iwan Müller (Bursian’s Jahresb. 1879, p. 162) condemnsillud, and would read eitherquamlibet frigidum(cp.3 §19and ix. 2. 67: quamlibet apertum), orquidlibet frigidum, which latter is approved by P. Hirt. Eussner suggests the deletion ofillud frigidum et inane, thinking that these words may be the remains of a gloss on§16.Attici sunt scilicet. Spalding’s reading seems on the whole to be preferred. The retention ofsunt(represented in some MSS. by a simples,—hence the readingAtticis scilicet) makes it less necessary to follow Meister in inserting asuntafterqui praec. concl. obscuri: in so loose a writer as Quintilian the firstsuntwould do duty for both. Halm follows Bn and Bg, which apparently (as also N Harl. 2662, 4829, and 11671) haveAttici scilicet: Meister (with bHM and Harl. 4950) givesAtticis scilicet. In the Ioannensis I findAttici s(forsunt): Dorv. and Burn. 244 giveAtticis s. Scilicet(om. Prat.) may be a gloss, and the true reading may beAttici sunt. Some codd. (Bodl. Burn. 243) giveAtticos scilicet(AthicosHarl. 4995): qy.Atticorum similes? (cp. Cic. Brut. §287).—Becher now prefersAtticis(sc.se pares credunt).§22.proposito. This conjecture by Gertz (Opuscula philol. &c., p. 134) I have found in the Ioannensis (*ppo) and in Harl. 2662 and 11671. It is approved also by Kiderlin. BNHb Prat. Sal. givepropositio: all other codd.proposita. Perhaps we should read (with Ioan.)sua cuique proposito est lex, suus decor est. Prat. omits the secondest.§23.tenuitas aut iucunditas, Halm and Meister:tenuitas ac iucunditasb H, Burn. 243, Bodl.:tenuitas aut nuditasN Ioan. M Harl. 2662, 11671:tenuitas ac nuditasPrat. Harl. 4995, 4950, 4829, C, Burn. 244, Dorv.:aut iuditasBg.§25.quid ergo? non est satis, &c. Gertz proposes to read, shortly afterwards,mihi quidem satis esset; set si omnia consequi possem, quid tamen noceret vim Caesaris ... adsumere?(=sed etiam si satis mihi esset, tamen nihil noceret vim Caesaris ... adsumere, si omnia haec consequi possem).§28.deerunt, Francius:deerant(derant) all codd. Becher defendsdeerant: ‘der Rhetor meint dassqui propria bona adieceritöfter Veranlassung gehabt haben wird, Fehlendes zu ergänzen als zu beschneidensi quid redundabit.’oporteatbHFT Bodl. M Harl. 4950 Burn. 243:oportebatB Prat. N Sal. Ioan. Harl. 2662, 4995, 4829, 11671 Burn. 244 Dorv. The latter (which is adopted by Halm) would indicate (cp. viii. 4. 22) a condition which ought to have been and may still be realised: the former (adopted by Meister and approved by Becher) is the conjunctive potential, and is quite in Quintilian’s manner (cp. xi. 2. 20): it conveys the expression of a present duty and obligation, the realisation of which may now be expected, and it connects also more intimately witheritin the following sentence.CHAPTER III.

Theaut ... autimmediately below is very much against this conjecture, which however Krüger (3rd ed.) has received into the text: we should expect rathernescio an illa quisquam, ornullus poeta, or keepingillaas nominativenescio an illa poeta ullo. Quintilian’s use ofnescio an(like that of post-Augustan writers generally) is vague: it is usually an expression of doubt, theanmeaning either ‘whether,’ or ‘whether not’ indifferently. Cp. ix. 4. 1: vi. 3. 6: viii. 6. 22: xii. 10. 2: i. 7. 24. (Mayor cites also Plin. Ep. i. 14. 9: iii. 1. 1: iv. 2. 1: v. 3. 7: vi. 21. 3: vii. 10. 3: 19. 4: viii. 16. 3: ix. 2. 5; and adds ‘In all these instancesnescio an(dubito an) is ‘I doubt whether’; in Cicero the meaning is always ‘I rather think.’’)Andresen proposednescio an ulla poeseos pars. The passage closely resembles§28, and must be emended on the same lines.

§66.tragoedias. Thurot (Revue de Phil. 1880, iv. 1, p. 24) conjecturedtragoediam: cp.§67hoc opus. He is followed by Dosson, against all MS. authority. Becher points out that we must supply withhoc opusin§67the words ‘tragoedias in lucem proferendi,’ so thatopusandtragoediassquare well enough with each other.

§68.quod ipsum reprehendunt, Meister, Krüger (3rd ed.) and Becher. This reading also occurs in the Codex Dorvilianus. Other readings arequod ipsum quodGHT Burn. 243, Bodl.:quo ipsumMS Harl. 2662, 4995, 4950, Ball. Halm conjecturedquem ipsum quoque, and was followed by Mayor and Hild. But as no fault has been found with Euripides in the foregoing,quoqueseems out of place.

Founding on the reading of GHT, &c., also on that of F (which givesquod ipsum qui) Kiderlin (Hermes 23, p. 165) proposes to readquod ipsum quidam, comparing§98, where forquem senes quem(GT) Spalding rightly conjecturedquem senes quidem, and7, §21, where Bn, Bg givequodforquosdam. He then goes on, in an interesting paper, to reconstruct the whole passage, which is open to suspicion, especially in respect thatsublimiorstands as predicate withgravitasandcothurnus, as well as withsonus. The admirers of Sophocles consider his elevation of tone more appropriate than the strain of Euripides.Sublimioris therefore perhapsnotthe predicate of the sentence, however suitable it may be as the attribute ofsonus. The predicate may have dropped out, andsublimiormay have been transferred from its real place to supply it. It is striking that GFTM (also H and Bodl.) all givesublimior erit. Kiderlin imagines that a copyist who missed the predicate wrote in the margin ‘sublimior erit ponendumpost esse’: and then another insertedsublimior eritafteressein the text. For the predicate,magis accommodatusmight stand: in copying, the eye may have wandered frommagis accommodatustomagis accedit: formagis accomm.cp. ii. 5. 18 and x. 1. 79. Kiderlin therefore boldly proposes to make the parenthesis run, ‘quod ipsum quidam reprehendunt quibus gravitas et cothurnus et sublimior sonus Sophocli videtur esse magis accommodatus’: ‘was gerade manche tadeln, welchen das Würdevolle, der Kothurnus, und der erhabenere Ton des Sophokles angemessener zu sein scheint.’

et dicendo ac respondendo7231, 7696:dicendo ac respondoGH:in dicendo et in respondendoPrat. Put. S (et respondendoM).

praecipuus. Hunc admiratus maxime est. This is Meister’s reading, except that foreumI give (with Prat. Put. 7231, 7696 Harl. 2662 and 4995)hunc, which is commoner in Quint. at the beginning of a sentence (§§46, 78, 91, 112). The following are the readings of the MSS.: GHpraecipuus et admiratus miratus: M Bodl. Harl. 4950, 4829, Burn. 244, C, Burn. 243 Ball. Dorv.praecipuus et admirandus: Spraecipuum. Nunc admiratus et: Prat. Put. Harl. 2262 and 11671praecipuus hunc admiratus et maxime est ut saepe test. et sec. quamvis: Harl. 4995,hunc admiratus max. ut s. test. et eum secutus quamquam. Halm givespraecipuus est. Admiratus maxime est: Kiderlin insists on theestafterpraecipuus, to correspond withaccedit, though it seems better to take all that comes afteracceditas an explanation of the statementmagis accedit oratorio generi: he also retains theetof most MSS. and readspraecipuus est. hunc et admiratus(Blätter f. d. bayer. Gymn. 24, p. 84). Wölfflin (partly followed by Krüger 3rd ed.) proposed a more radical change (Rhein. Mus. 1887, 2 H. p. 313)praecipuus. Hunc imitatus, quoting in support of the conjunctionimitatus ... secutus§122, eos iuvenum imitatur et sequitur industria:5 §19, deligat quem sequatur, quem imitetur: Ovid, Fasti v. 157, ne non imitata maritum esset et ex omni parte secuta virum. But Kiderlin (l.c.) aptly remarks that if Quintilian had writtenimitatus, he would not have saidut saepe testaturbutut ex multis locis patet(apparet, videmus): while vii. 4. 17 (on which Wölfflin relies) is not really to the point. Moreover Quintilian, would never have separated such synonyms asimitatusandsecutusbyut saepe testatur.

Charisi nomini addicuntur, Frotscher:Charis in homine adducturaGH:Charisii nomine edunturPrat. Put. 7231, 7696 Harl. 2662 Dorv.

§70.aut illa iudiciaPrat. Put. 7231, 7696 Harl. 4995. GH Harl. 4950 giveaut illa mala iudicia: Bodl. Burn. 243aut alia mala iud.S Harl. 2662 Dorv. and Ball.aut alia iudicia.The edd., following Gesner, have generally given (with Harl. 4950)aut illa mala iudicia(so Halm and Meister), and have takenmalaas predicate, though the order of the words makes that impossible. Becher approves of Andresen’s deletion ofmala. Krüger (3rd ed.) printsmala [illa] iudicia, thinking thatillaarose by dittography, and that then the order was changed in the codd. toilla mala iudicia. Kiderlin (in Hermes 23) gives as an alternative to deletingmalathe conjectureilla simulata iudicia(‘jene erdichteten nachgemachten Gerichtsverhandlungen’; cp. xi. 1. 56: cum etiam hoc genus simulari litium soleat). A similar mutilation occurs, e.g., xi. 1. 20, where b givessecumMsecusinstead ofconsecutum.

§71.filiorum militum, most codd.:filiorum maritorum militumPrat. Put. 7231, 7696 S.

§72.si cum venia leguntur. The reading of the MSS. is upheld by Iwan Müller, Meister, and Kiderlin. Spalding suggestedcum verecundia: Schöllcum iudicio: Bechercum ingenio. Becher points out (Bursians Jahresb. 1887) that the expression is meant to coverdecerpereas well aslegere, anddecerpereindicates careful and intelligent reading (cp.§69,diligenterlectus):cum ingenio= ‘mit Verstand’: cp. Cic. ad Fam. xiii. 10. 2 quod versabatur in hoc studio nostro .. et cum ingenio .. nec sine industria: Ulp. Dig. 1. 16. 9 patientem esse proconsulem oportet, sed cumingenio, ne contemptibilis videatur. Finally, Krüger (3rd ed.) proposescum acumineorcum vigilantia(cp. v. 7. 10).—Prat. Put. 7231, 7696 S Harl. 2662 all give Osann’s conjecturelegantur.

praveGH Harl. 4995, 4950 Burn. 243 Bodl.:pravisRegius, Halm, Meister, Becher draws attention to the parallelism between the clauses:ut prave praelatus est sui temporis iudiciis, ita merito creditur(= meruit credi)secundus consensu omnium.

§76.nec quod desit ... nec quod redundet: H Burn. 243 and Bodl. givequod .. quod: Prat. Put. MS Harl. 2662, 4995, 4950, Burn. 244, Dorv. C, and Ball,quid .. quid. The latter reading is supported by Becher (Phil. Rund. iii. 434). Forquodcp. xii. 10. 46: (xii. 1. 20 where forquod adhucBM givequid adhuc): on the other hand, in vi. 3. 5 the MSS. are in favour ofquid, though Halm readsquod(followed by Meister). Forquidcp. Cic. pro Quint. §41, neque praeterea quid possis dicere invenio.

§77.grandiori similis. So all MSS.: Halm and Meister. Several conjectural emendations have been put forward. Comparing2 §16(fiunt pro grandibus tumidi), Becher suggestsgrandi oratori,—an easy change, if the copyist used contractions, but without point: above in§74, ‘oratori magis similis’ is appropriate enough in speaking ofhistorians, but ‘oratori’ would be inappropriate here. This is accepted, however, by Hirt (Berl. Jahr. ix., 1883, p. 312; cp. P. Hirt, Subst. des Adjectivums, p. 12). Schöll proposes to readgladiatorisimilis, in view of the close connection with what follows, strictus ... carnis ... lacertorum: butpleniorandmagis fususare a bad introduction togladiatori, and if Aeschines hadplus carnisandminus lacertorum, he cannot really have resembled a gladiator. This reading is, however, adopted by Krüger (3rd ed.). Finally, Kiderlin (Hermes 23, p. 166 sq.) has conjecturedet grandi(orgrandiori)organo similis, and applies the figure throughout: ‘voller und breiter lässt Aeschines den Ton hervorströmen, einem grossen Musikinstrumente gleich’: ‘einer Orgel gleich,’—he isgrandisonus. The translation appears to limit unnecessarily the meaning ofplenusandfusus: though the former is used of tone i. 11. 6 (cp. xi. 3. 15 of the voice: ib. §§42, 62: and §55 of the breath): whilefususis used of the voice xi. 3. 64. For such a use ofgrandiscp.§58(cenae):§88(robora): xi. 2. 12 (convivium): 3. 15 (vox): 68 (speculum): and fororganum, i. 10. 25: ix. 4. 10: xi. 3. 20 (where there is a comparison between the throat and a musical instrument): probably also i. 2. 30. There is an antithesis in the two parts of the sentence between fulness and breadth, on the one hand, and real strength on the other; and for the transition to the second figure Kiderlin compares§33.

§78.nihil enim est inane: perhaps ‘nihil enim estin eoinane’ (Becher), ornihil enim inest.

§79.honesti studiosus. Becher’s proposal to alter the punctuation of this passage is discussed in the notead loc.—Forauditoriisandcompararat, see ontenuia atque quae§44, above.

§80.quem tamen. Kiderlin, in Hermes (23, p. 168), raises a difficulty here.Tamenshows that the clause cannot go with the main statement (fateor), and its position forbids us to take it with thequamquam is primumclause: it can only go withquod ultimus est, &c., ‘though Demosthenes isultimus fere, &c.,yetCicero, &c.’ To prevent so awkward a joining of the clauses, Kiderlin proposes to readeumque tamen: pointing out that thequaeof the MSS. (GH) may have arisen out ofque, and that Quintilian may have writteneumque; cp. vi. 2. 13, where Halm makesutqueout ofquae(G), and xi. 2. 32, where Meister readsestque. The meaning will then be: Demetrius is worthy of record as being about the last, &c., and yet Cicero gives him the first place in themedium genus.—It seems better, however, to givetamena general reference: ‘yet, in spite of all that can be said on the other side’ (e.g., inclinasse eloquentiam dicitur). Cp.§99quae tamen sunt in hoc genere elegantissima.

§81.prosam(prorsam)orationem etall MSS.; Halm, Meister, Krüger (3rd ed.) omitet. I find that Becher supports the view stated in the notead loc.: he would however writeprorsam, which the best MSS. give also in Plin. v. 31, 112 D.

quodam Delphici videatur oraculo dei instinctus: so Frotscher, followed by Krüger (3rd ed.). On the other hand Claussen (Quaest. Quint., p. 356) and Wölfflin (followed now by Meister, pref. to ed. of Book x., p. 13) propose to deleteDelphici, of which Becher also approves. But the MS. evidence cannot be disregarded. The following are the various readings: GHquaedam Delphico videatur oraculo de instrictus, and so FT, the former giving also (by a later hand)de instinctus, the latterdei instructus. Bodl. givesquodam delphico videatur oraculo dei instructus. The most frequent reading is that of Prat. Put. 7231, 7696 Harl. 2662, 4995, 4829, 11671, Ball. and most edd.,quodam delphico videatur oraculo instinctus: S agrees, but is reported to havedelphicoafteroraculo: Harl. 4950 and Burn. 244 have the same reading, withinstitutuscorr. toinstinctus: Burn. 243 givesinstructus.Delphicowas originally deleted by Caesar: Phil xiii, p. 758. Halm readtamquam Delphico videatur oraculo instinctus: but Quintilian would take no trouble to avoid the repetition ofquidam(cp. divina quadam, above).—For the arrangement of words, Krüger (3rd ed.) compares§41qui ne minima quidem alicuius certe fiducia partis memoriam posteritatis speraverit.

§82.quandam persuadendi deam. Nettleship (Journ. of Philol., xxix, p. 22) conjecturesSuadam[persuadendi deam], comparing Brutus, §59, quotedad loc. Persuadendi deamwould thus become a gloss onSuadam: but the expression in the text is quite in Quintilian’s style.

§83.eloquendi suavitate:eloquendi usus(orusu)suav.GH and all codd. except Harl. 4950, and Dorv., both of which give simplyeloq. suav.Halm admitted into his text Geel’s conj. forusus, ‘eloquendivi acsuavitate,’ and this has met with some acceptance (Iwan Müller and Becher). But the parallel from Dion. Hal.,Ἀρχ. κρ.4 is hardly conclusive:τῆς τε περὶ ἑρμηνείαν δεινότητος ... καὶ τοῦ ἡδέος. Hirt properly remarks that the agreement between the two is not so great as to allow of correcting the one by the other. Kiderlin conjectureseloquendi vi,suavitate,perspicuitate.

tam est loquendi. See notead loc.for Kiderlin’s conj.tam manifestus est. Though Meister’stam est eloquendiis probably a misprint, it is found in some MSS.—Harl. 4950: Burn. 244.

§84.sane non affectaverunt. Bodl. and Vall. (verusubpunctuated in the latter:affectantPrat. Put. 7231 MS Ball. Dorv. Harl. 2662, 4995, 4829, 11671:sene non adfectitacueruntGH Burn. 243:adfectarunt7696:adfectitantHarl. 4950, and so Burn. 244 (corrected fromaffectant).

§85.haud dubie proximus. Halm insertedeiafterdubie, though it is not found in any MS.: Regius had suggestedilli. Kiderlin (Hermes 23, p. 170) points out that ifpropiores aliiin§88is allowed to stand without a dative,eiis not necessary here. He suggests, however,illibeforealiiin§88: both passages must be dealt with in the same way.—Forhaud(Vall.), GHS haveaut: Mhaut. Cp. on3 §26.

§86.ut illi ... cesserimus:cum illiGHFT Harl. 4995 Burn. 243:ut illiPrat. Put. 7231, 7696: and so S Harl. 4950 (withcaelesti atque divinae):ut illeM Harl. 2662. Kiderlin (Hermes, p. 170) proposes to go back to the reading of the older MSS.cum illi, and instead ofcesserimusto readcesserit, so as to make Vergil the subject throughout.Cumcannot, he contends, be a copyist’s error, motived byita; and it is probable, therefore, that at firstcesserit awas inadvertently written forcesserit; then (in G or some older MS.)cesserimus itawas made out of that, to correspond withvincimurbelow: and then in the later MSS.cumwas changed tout, because ofita. For the transition, with this reading, from cesserit to the plural (vincimur, pensamus), hecompares§107, where, after speaking of Demosthenes and Cicero, Quintilian passes tovincimus.

§87.sequenturMS Halm and Meister:sequenterGseq̅nt’H:sequunturPrat. Put. 7231, 7696.

φράσινid est. These words are omitted in the Pratensis, which is Étienne de Rouen’s abridgement of theBeccensis, now lost. This is an additional proof thatφράσινwas originally written in Greek: cp. on§42.

§88.propioresH Prat. Put. Vall. Harl. 2662, 4495, 11671, Burn. 243. Bodl., Halm:proprioresGMS 7231, 7696, Harl. 4950, C, Burn. 244, Dorv., Meister. In Cicero and Quintilianmagis propriiwould be more usual for the latter.

§89.etiam si sit. This conjecture of Spalding’s (foretiam sitGH Bodl. &c.:etiam siM Harl. 4950 Dorv.:etiam sicPrat. Put. S Harl. 2662) I have found in the Balliol codex. 7231 and 7696 giveetiam si est. Cp. note ontenuia atque quae§44, above.

ut est dictum. These words were bracketed as a gloss by Halm, and are now omitted altogether by Krüger (3rd ed.): see however notead loc.Döderlein proposed to place them afterpoeta melior, Fleckeisen afteretiam si.

Serranumis Lange’s conjecture forferrenumGHM:farrenum7231, 7696 Harl. 2662, 11671:PharrenumPrat. Put. Some MSS. (e.g. Vall. Harl. 4995, Burn. 243 and 244) givesed eum, but it is obvious that the criticism of Severus stopped with the wordlocum.

§90.senectute maturuited. Col. 1527 and so 7231, 7696 (Fierville):senectutem maturbitGH:senectute maturumPrat. Put. MS Harl. 2662, 4995, 4950, Burn. 244, Dorv. and Ball.:senectus maturavitBodl., Burn. 243.

et, ut dicam. Halm’ssedinstead ofethas been rejected by later critics. Cp. Claussen (Quaest. Quint., p. 357 note):sed‘sententiam efficit ab hac operis parte alienam. Nam cum oratori futuro exempla quaerantur oratoria virtus in quovis scriptore laudi vertitur (§§46, 63, 65, 67, 74, &c.). Itaque propter huius censurae consilium Quintilianus Lucani elocutionem oratoriam laudat, sed ingenium poeticum una reprehendit.’

§91.propiusH Prat. Put. Burn. 243, Harl. 2662 and other codd.: Bodl. Ball. Harl. 4950proprius. Reisig conjecturedpropitius, which also is apt; but in spite ofindustrius,necessarius, cited in its support (cp. iv. 2. 27: vii. 1. 12), it is too uncertain a form to be received into the text. Iwan Müller thinks it would have to bemagis propitiae. Halm givespromptius: Wölfflinpronius: while Schöll now suggestspropitiae potius(cp. iv. pr. §5: 2 §27: vii. 1. 12).

§92.feresG Prat. Put. 7231, 7696 S Harl. 2662, 4829, Dorv., Ball., Halm.:ferasH, Harl. 4950, Burn. 243, Bodl. C and M, Meister and Krüger (3rd ed.). Harl. 4995 hasfere: from Vall. Becher reports feras, ‘probably at firstferes.’

elegeaGH 7696, and so A2BN Put. S at i. 8. 6.

§94.abunde salisG Prat. Put. M and all my MSS. except H, Burn. 243, Bodl. which haveabundantia salis.

multum est tersior. The variety of MS. readings seems to point to anetwrongly inserted aftermultum, perhaps from a confusion with ‘multum et ver gloriae’ below. GH givemultum et est tersior: M Harl. 4950, Bodl. Ball. C Dorv. Burn. 243 and also Harl. 4829multum etiam est t.: Prat. Put. 7231, 7696 S Harl. 2662, 11671multum est tersior: while Harl. 4995 (and Vall.) hasmulto et est tersior. Osann proposedmulto eo est tersior: Wölfflinmulto est tersior: Halm and Meister printmultum eo est tersior. Formultum, cp. multum ante xii. 6. 1: and see Introd.p. li.

non laborGH Burn. 243 Bodl. and Meister:nisi labor7231, 7696 S Harl. 2662, 4995, 4950, 4829, 11671, Burn. 244, Dorv. Ball. C, and Halm. Prat. and Put. havemihi labor.

hodieque et qui: H, Prat., Put., 7231, 7696, Harl. 2662, 4829, Bodl. Dorv.:hodie et quiBurn. 243:hodie quoque et quiVall. Harl. 4995, 4950:hodie quod et quiS.—Becher is of opinion that the text will not bear the explanation given in the note, and would readhodie quoque et qui: ‘es giebt auch heute noch berühmte Satirendichter, die einst &c.’Et quihe takes withclari, not withhodie quoque, theetbeing omitted in translation: clari (hodie quoque) qui (olim) nominabuntur.

§95.etiam prius. Founding on the classification given in Diomedes (see notead loc.), according to which thesaturaof Pacuvius and Ennius preceded and was distinct from that of Lucilius, Horace, and Persius, Claussen (Quaest. Quint., p. 337) thinks that the true reading here may beAlterum illud et iam priusEnnio temptatumsaturae genus, &c. For the satura of Ennius, cp. ix. 2. 36. Iwan Müller points out that Ennius is not mentioned below (§97), beside Attius and Pacuvius, probably because neither in tragedy nor in satire did Quintilian consider him to have produced anything helpful for the formation of an oratorical style. Other unnecessary conjectures areetiam posterius, Gesner:etiam proprium, Spald.:etiam amplius, L. Müller:etiam verius, Riese:alterum illud Lucilio prius sat. genus, Krüger (3rd ed.).

sola:solumPrat. and Put.

collaturus quam eloquentiae. These words, omitted in GHS Bodl. Burn. 243, occur in all my other codd.

§96.sed aliis quibusdam interpositus: sc. carminibus, Christ. In H the reading isquibusdam interpositus: so 7231, 7696 Bodl. and Burn. 243: but M Harl. 4950, 4829 Burn. 244 Dorv. and Ball, givea quibusdam interpositus: Scuiusdam: Prat. and Put.opus interpositus. Osann conjecturessed quibusdam, and so Hild. In the margin of Harl. 4995 is the variantaliquibus interpositis.

In Hermes, vol. 23, p. 172, Kiderlin makes a fresh conjecture. Recognising that something must have fallen out beforequibusdam, but dissatisfied with Osann’ssedand Christ’ssed aliis, he proposes to readut proprium opus, quibusdam aliis tamen carminibus(orversibus)a quibusdam interpositus. The eye of a copyist may easily, Kiderlin thinks, have wandered from the first to the secondquibusdam: cp. v. 10. 64, ut quaedam a quibusdam utique non sunt, &c., and for quibusdam aliis xi. 3. 66, et quibusdam aliis corporis signis.

intervenit, which is a conjecture of Osann, I have found in Harl. 2662, 11671 Prat. Put. 7231, 7696.

lyricorum. Kiderlin thinks there may be something wrong in the text here. The last sentence (sed eum longe, &c.) shows clearly that Quintilian had a high opinion of the lyrists of his day: if Bassus waslegi dignus, they were even more so. Would he then have said ‘of the Roman lyrists Horace is almost the only one worth reading’? Perhaps we should readlyricorum priorum: after-ricorum,priorummight easily fall out, and it gives a good antithesis toviventium. Bassus (quem nuper vidimus) forms the transition: and the next paragraph beginsTragoediae scriptores veterum, &c.

§97.clarissimi. This reading is stated by Halm to be ‘incerta auctoritate,’ and is referred by Meister to the Aldine edition. It occurs in Prat. 7231, 7696 Harl. 2662 (A.D.1434) Vall. 4995, 4829, 11671, Dorv. and Ball.: Put. givesclarissime: G hasgravissima: HFTSgravissimus, and so also Harl. 4950, Burn. 243, Bodl. and C. Halm printsgrandissimi: Ribbeck (Röm. Trag. p. 337, 3) inclines to accept the sing.grandissimus, M, of Pacuvius alone.

Kiderlin (in Hermes 23, p. 173) rejects all the above readings.Gravissimusandgravissimaare obviously due, he says, togravitatefollowing: but the word beforegravitatemust have begun with the same letter, and soclarissimicannot stand, especially as it is inappropriate to the context. Forceterumshows that the sentence before it must have contained some slight censure: some defect, or quality excluding others equally good, must have been mentioned. He therefore conjecturesgrandes nimis, in preference tograndissimi, which in tragedy would hardly be a fault. Attius and Pacuvius, Quintilian says, are ‘zu grossartig, sie kümmern sich zu wenig um Zierlichkeit (Eleganz) und die letzte Feile (d.h. Sauberkeit im Kleinen); doch daran ist mehr ihre Zeit schuld als sie selbst.’ He evidently thinks more of the ‘Thyestes’ of Varius and Ovid’s Medea: cp. Tac. Dial. 12. With this judgment Kiderlin compares§§66,67tragoedias primus in lucem Aeschylus protulit, sublimis et gravis et grandiloquus saepe usque ad vitium, sed rudis in plerisque et incompositus ... sed longe clarius inlustraverunt hoc opus Sophocles atque Euripides, and is of opinion that the parallelism cannot be mistaken. For the position ofnimishe compares ix. 4. 28 longae sunt nimis: v. 9. 14 longe nimium: xii. 11. 9 magna nimium.

§98.quem senes quidem parum tragicum. So Spalding, Bonnell, Halm, Meister, and Krüger.Quidemoccurs in no MS.: GH havequem, M Vall., Harl. 4995, Burn. 244, Ball, omit it: Bodl. Burn. 243 and Dorv. show the corruptionPindarum. Becher would excludequidem, regardingquemin G as an instance of the tendency of copyists inadvertently to repeat, after a particular word that by which it has been immediately preceded, e.g.§68quod ipsum quod (G): ix. 4. 57 ut cum ut (G): iv. 1. 7 ipsis litigatoribus ipsis (b): iv. 2. 5 aut ante aut (bT): x. i. 4 iam opere iam (G).—But here the authority of the Pratensis and its cognates may be invoked. In the archetype from which they are derived something must have stood beforeparum, as Prat. Put. 7696, 7231 all givequem senes non parum tragicum: so Harl., 2662 (A.D.1434), and 11671. Above in§96, G Prat. Put. 7231, 7696 havesi quidemforsi quem.

§100.linguae suae. So Köhler (v. Meister pref. to Book x. p. 13):suaesupplies an antithesis to ‘sermo ipse Romanus’: GH givelinguae quae: so Harl. 4950: S Burn. 243, Bodl.linguae: while Harl. 2662, 4995, 4829, 11671, Dorv. and Ball. omit it altogether: M hasligweque.

§101.Titum: GH Prat. Put. M. 7231, 7696.

commendavit: Halm and Meister givecommodavit, which is approved also by Hirt. Halm compares§69where Menander is said to be ‘omnibus rebus personis adfectibus accommodatus.’ But this would require the meaning ‘appropriately treated,’ and there is no instance in Quintilian of the verb used absolutely in this sense. Nor is there any example to support Hild’s interpretationpraestitit, which would be moreover extremely weak. The recurrence of the word so soon afteraccommodatatells against Halm’s reading, though Quintilian is negligent on this head.—On the other hand, in vi. 3. 14 the reading ‘ad hanc consuetudinem commodata’ is rightly accepted against ‘commendata’ most edd.

§102.immortalemGS Meister:illam immortalemPrat. Put. M Halm:immortalem illamVall.

velocitatem. So all MSS, except S, Burn. 243, and Bodl., which havecivilitatem. Kiderlin (in Hermes 23, p. 174) thinks that we might have expectedideoque immortalem gloriam quam velocitate Sallustius consecutus est: ‘und darum hat er dievelocitasdurch (von der velocitas) verschiedene Vorzüge erreicht.’Consequicannot mean ‘to supply the place of’: andimmortalisis inappropriate as an attribute ofvelocitas: besides, Quintilian has not spoken of Sallust’svelocitas, even indirectly. Schlenger conjecturedclaritatem: Andresenauctoritatem(‘klassisches Ansehen,’ cp. iv. 2. 125: xii. 11. 3): Kiderlin now proposesdivinitatem, which in Cicero = Vortrefflichkeit, Meisterschaft: cp. xi. 2. 7. Judged by the previous sentences the expression is not too strong. Forimmortalem divinitatemcp.§86illi ... caelesti atque immortali: and forconsecutus estiii. 7. 9 quod immortalitatem virtute sint consecuti.

clarus vi ingenii. This is a conjecture of Kiderlin’s, which I find has been adopted also by Krüger (3rd ed.). GHFT giveclarius ingenii: Prat. Put.clari ingenii vir: 7231, 7696clari vir ingenii: MS Harl. 4995, 4950, 4829, Burn. 243 and 244, Dorv.C and Ball,clarus ingenio; Harl. 2662 and 11671clarus(?) orclaret vir ingenii. Spalding had already pointed out thatclarusis not found withingenium, except whereingeniumis used of a person: e.g.§119erant clara et nuper ingenia: he therefore wroteelati vir ingenii(following Goth.elatus ingenioand Bodl.elatus ingeniis). Kiderlin compares§70sententiis clarissimus, and forvis ingeniii. pr. 12: ii. 5. 23: x. 1. 44: xii. 10. 10. The readingclarus vi ingeniipoints the contrast to what follows in ‘sed minus pressus,’ &c.: it was hisstylethat did not altogether suit the dignity of history.

§103.genere ipso, probabilis in omnibus, sed in quibusdam. Till Kiderlin made this happy conjecture (see Hermes 23, p. 175)generehad always been joined withprobabilis, and the text was twisted in various directions. GHS, Burn. 243, Bodl. givein omnibus quibusdam: M Harl. 2662, 4995, 4950, Burn. 244, Dorv.in omnibus sed in quibusdam, and so apparently Prat. Put. 7231, 7696. Out ofomnibusHalm gives on Roth’s suggestion,operibus: afterwards he decided forpartibus, and this (thoughomnibustopartibusis not an easy transition) is adopted by Meister. Kiderlin’s punctuation makes everything easy: ‘Anerkennung verdienen seine Leistungenalle,manchestehen hinterseinerKraft zurück.’ Even these last, Quint. means, areprobabiles(cp. viii. 3. 42 probabile Cicero id genus dicit quod non plus minusve est quam decet); but they do not show the great powers that distinguish his other writings. It is uncertain whether Quintilian wrotein quibusdamorsed in quibusdam(M). The easiest explanation of the omission in the other MSS. is to suppose that he wrotein omnibus in quibusdam: perhaps the copyist of M saw thatomnibusandquibusdamwere antithetical, and insertedsed. Kiderlin notes Quintilian’s liking for chiasmus, without any conjunction: cp.§106in illo, in hoc (where in hoc is wanting in M).

suis ipse viribus: ed. Col. 1527 (Halm), and so (Fierville) 7231, 7696. In Harl. 2662 and 11671 (A.D.1434 and 1467)suisalready appears, corrected fromvisGH. The Juntine ed. (1515) hassuis viribus minor: so Prat. and Put.

§104.et exornat. Vall. and (apparently) Prat. Put. 7231, 7696, and most edd.:et ornatM Halm, Meister, Krüger:exornatGHS. Becher remarks thatet exornatmight easily pass intoexornat.

nominabitur: Weber and Osann proposednominabatur(which appears in Harl. 2662, but corrected to-itur). Krüger at first accepted this in support of his theory that the whole passage refers to Cremutius, who ‘in former days (olim), while his works were under a ban, was only named (i.e. was a mere name, but now is known and appreciated).’ The parallel passage (§94) is sufficient to dispose of any such interpretation: sunt clari hodieque et qui olim nominabuntur.

Cremuti. Nipperdey, Philol. vi, p. 193, Halm, and Meister:remutiH Prat. Put. 7231, 7696remremutiG,rem utiliBurn. 243:remittiS. Bodl.:nec imitatores utiHarl. 2662, 4995, 4950, 4829, 11671. A review of the various explanations of the whole passage (Superest—quae manent) will be found in Holub’s Programm ‘Warum hielt sich Tacitus von 89-96 n. Chr. nicht in Rom auf?’—Weidenau, 1883: but his conjectureremoti(i.e. relegati) forremutiis not to be thought of.

dividendi: first in the Aldine edition: all MSS. havevidendi, except M (indicendi) and Prat. Put. Harl. 4995 (vivendi). Cp. i. 10. 49, where the case is the same.

§105. In the Aurich Programm, Becher gives a more recent statement of his views: ‘wie zucumcausale, so tritt praesertim auch zucumconcessivum, in diesem Falle wiedenzugeben mit, “was um so auffallender ist, als.” Der Sinn ist also: “Ich weiss sehr wohl, welchen Sturm des Unwillens ich gegen mich errege, und dies (dieser Sturm) ist um so auffallender, als ich jetzt gar nicht die Absicht hege, meine (in Potentialis gesprochene) Behauptung (fortiter opposuerim) wahr zu machen, resp. comparando durchzuführen. Ich lasse ja dem Demosthenes seinen Ruhm—in primis legendum vel ediscendum potius.”’

§106.praeparandi. For Kiderlin’s conj.praeparandi,narrandi,probandiseead loc.

[omnia]denique, GH, Burn. 243, Bodl. omitomnia(which is in all my other MSS.), and Meister now approves (following Spalding, Osann, and Wölfflin), on the ground that Demosthenes and Cicero werenotalike ineverythingthat belongs toinventio. Halm thinks thatomniais to be found inracioniof the older MSS.: but Kiderlin points out that this error may have arisen from the carelessness of a copyist who, after thrice writing the terminationi, gave it also to the fourth word.

illi—huicPrat. M, S Vall. Harl. 4995, 2662 Bodl. &c.:illic—hicGH Put. 7231, 7696, Halm.

§107.vincimus, H, G2, and most MSS.: (cp.§86):vicimusG.

§109.ubertateHarl. 4995. This is also the reading of codd. Vall. and Goth.:all the other MSS. giveubertas.

totas virtutesBn Bg N Prat. Ioan. 7231, 7696:totas viresM b.

§112.ab hominibusHalm and Meister:ab omnibusBn Bg HFT Ioan. Prat. 7231, Sal. and most codd.:hominibusS Harl. 4995 Bodl.

§115.urbanitas. Kiderlin proposes to readet praecipua in accusando asperitas et multa urbanitas: cp.§117:§64:2 §25: ii. 5. 8.

Ciceroni, forCiceronemof the MSS. In the Rev. de Phil. (Janv.-Mars, 1887) Bonnet quotes from the Montpellier MS. a note of the sixteenth century deleting the name as a gloss (oninveni). Certainly all codd. giveCiceronem, notCiceroni. Bonnet thinks that the insertion does not accord with Quintilian’s habitual deference towards Cicero: ‘Quintilien se trouvant dans le cas de contredire Cicéron ne le nomme pas.’—Becher reportsCiceroni, a correction in the Vallensis.

castigata, B (i.e. Bn and Bg) Ioan. Prat. 7231, 7696 Harl. 2662, 4995, 11671:custoditaH M b F T Alm. Harl. 4950, 4829, Burn. 243, 244, Bodl. Dorv. and Ball. Forgravis(bH M Vall. and seemingly Prat.) B Sal. 7231, 7696 and Ioan. givebrevis.

si quid adiecturus sibi non si quid detracturus fuit, Vall. Harl. 4995. For the repetition, see on haud deerit3 §26. Halm and Meister printsi quid adiecturus fuit—(sc.virtutibus suis, cp.§§116,120)—the reading of B (i.e. Bn and Bg), which is also that of Ioan. Prat. N 7231 Harl. 2662, 11671: while M Harl. 4950, 4829, Burn. 244 havesi quid adiecturus fuit, non si quid detracturus. The reading of H issi quid adiecturus sibi non si quid detracturus[Sulpicius insignus]fuit ut servius sulpicius insignem&c.: so also T, Burn. 243, Bodl. The brackets in H are by a later hand, indicating a gloss which arose from a mistake made by the copyist of H. In Bg the passage stands:—

The words added above the line are by the hand known as b.

In copying H wrote:si quid adiecturus sibi non si quid detracturus(then omittingfuitcontinues)et Serv. Sulp.(then goes back and resumes)fuit et servius&c. This is the origin of the confusion which exists in all the MSS. of this family.

§117.et fervor. This is Bursian’s conjecture, adopted by Halm and Krüger (3rd ed.), and now approved by Becher. BM haveet sermo, which is also the reading of N Prat. Sal. 7231, 7696 Ioan. Harl. 2662, 4950 and Ball.: Hbet summo: Harl. 4829, 11671, Burn. 244et smo: while Bodl., Dorv., and Burn. 243 give the correction in Teius summa, out of which the second hand in the Vallensis (Laurentius Valla) madeet vis summa, a reading which occurs also in Harl. 4995. Meister readset sermo purus; while Kiderlin proposeset simplex sermo(cp. iv. 1. 54: viii. 3. 87: ix. 3. 3: 4. 17: viii. pr. 23: x. 2. 16).

ut amari sales. Francius conjecturedut amantur sales, but this loses the antithesis betweenamariandamaritudo ipsa. Kiderlin’sut amantur amari sales(viii. 3.87: vi. 1. 48) is an improvement; but ifridiculais taken in a good sense it seems impossible that after censuring Cassius for giving way unduly tostomachus, Quintilian should go on to say, ‘moreover, though bitter wit gives pleasure, bitterness by itself is often laughable.’ Is it possible that we ought to readut amari sales risum movent ita amaritudo ipsa ridicula est? Such an antithesis might have been written ‘per compendium,’ and the wordsrisum moventmay then have dropped out. See the notead loc.: and cp. especially vi. 1. 48fecit enim risum sed ridiculus fuit, andοὐ γέλωτα κινεῖ μᾶλλον ἢ καταγελᾶται, quoted in the note on1 §107.—Krüger (3rd ed.) adoptsfrequentiorforfrequenter, which gives a good sense, except thatfreq. amar ipsais awkward.

§121.leneHalm and Meister:leveB Prat. N 7231 M 7696 C. Here again Becher prefersleve, comparing Cic. de Orat. iii. §171, quoted on§44above: levitasque verborum 1. 52: and levia ... ac nitida, v. 12. 18.

§123.scripserint. So Bn Bg H Ioan. Prat. 7231, 7696 Vall. Harl. 4995, 2662, 11671, Bodl., Dorv., Spalding, and Bonnell. Becher compares among other passages2 §14(concupierint), and points out that Quintilian is not thinking of individual writers on philosophy, but of the class, as opposed to the class of orators, historians, &c.—Halm, Meister, and Krüger havesupersunt(Put. M, Ball. Burn. 243 Harl. 4950).

§124.Plautus, Prat. N, 7231 Ioan. Harl. 2662, 4829, 11671:plantusM Harl. 4950:PlantatusSal.:plaustusHb:Plancusedd. vett. and Harl. 4995.

Catius. The name is rightly given in Harl. 4995.

§126.iis quibus illi.Iisis the conjecture of Regius, followed by Halm, Meister, and Krüger. Becher would retainin quibus illi,—the reading of BN Prat. Ioan. Vall. M Harl. 4995, 2662, 4950, 11671, Burn. 244 Dorv. Ball. The difficulty of construing probably led to the omission ofinin bH Bodl. Burn. 243, 7231, 7696, Spalding and Bonnell.

ab illoB Ioan. 7231, 7696 Sal. Harl. 2662, 4950, 4829:ab eobHM Burn. 243.

§127.foret enim optandum:fore enim aliquid optandumbHFT. Spalding conjecturedalioqui optandum, which Kiderlin approves.

ac saltemall MSS.: Meister hasaut saltem, probably relying on a wrong account of the Bambergensis: see Halm vol. ii, p. 369.

illi viroB:illi virusbHM:illi virtutibusHalm:illi viro eos(orviro plurimos) Kiderlin.

§128.multa rerum cognitio: so all codd. except Ioannensis and Harl. 4995, which havemultarum rerum cognitio. b omitscognitioand is followed by HFT.

§130.si obliqua contempsisset, si parum recta non concupisset. I adopt the reading recently proposed for this vexed passage by Ed. Wölfflin in Hermes, vol. xxv (1890), pp. 326-7, though it is right to note that he was partly (as will be seen below) anticipated by Kiderlin.Obliquaseems thoroughly appropriate in reference to Seneca’s unnatural, stilted, affected style,—‘jene unnatürliche, durch unmässigen Gebrauch von Tropen und Figuren auf Schrauben gestellte Ausdrucksweise, welche statt der Klarheit ein Schillern zur Folge hat.’ Wölfflin compares ix. 2. 78rectum genusadprobari nisi maximis viribus non potest: haec diverticula et anfractus suffugia sunt infirmitatis, ut qui cursu parum valent flexu eludunt, cum haec quae adfectatur ratio sententiarum non procul a ratione iocandi abhorreat. Adiuvat etiam, quod auditor gaudet intellegere et favet ingenio suo et alio dicente se laudat. Itaque non solum si persona obstaretrectae orationi(quo in genere saepius modo quam figuris opus est) decurrebant ad schemata ... ut si pater ... iacularetur in uxoremobliquissententiis. This passage supplies (what is indeed suggested byobliquaitself) the antithesisparum recta: cp. ii. 13. 10 si quis ut parum rectum improbet opus.

In theJahrbücher f. Philologie(vol. 135, 1887: p. 828) Kiderlin had previously dealt with the passage on similar lines. The traditional readingsi aliqua contempsisset(b) he considers too indefinite, though not impossible: in point of authority, though preferable to thesi nil aequalium cont.of the later MSS., it cannot rank so high as the reading of Bn and Bg, which givesimile quamwithout any attempt at emendation. This Kiderlin thinks must be nearest the original: he therefore rejects such conjectures as Jeep’ssi antiqua non, on the ground that it is improbable thatsimile quamarose out ofantiqua. He introduces his own conjecture by referring to ix. 2. 66 and 78 (see above), and to the contrast betweenschemataandrectum genus,recta oratio; the former are calledluminaorlumina orationis(xii. 10. 62). Cp. viii. 5. 34. He would read:nam si mille ille schemata(orillas figuras)similiaque lumina contempsisset, si parum rectum genus(orsermonem)non concupisset, &c.Similiaqueoccurs ix. 4. 43:mille(forsescenti) is used v. 14. 32: forcontempsissetcp. ix. 4. 113.Si mille illaandsimiliaquemay easily have run together, whenschemata(orfiguras) would fall out:quamin the older MSS. may representque lumina, which again reappears in thequaliumof the later codd. (si nil aequalium). As an alternative forparum rectumgenns(orsermonem) Kiderlin suggests Wölfflin’s readingparum recta: and compares ix. 2: ii. 5. 11: v. 13. 2: ix. 1. 3; 3. 3: x. 1. 44; 89: ii. 13. 10.

Of the MSS. Prat. 7231 Sal. 7696 N Ioan. Harl. 2662 and 11671 agree with Bn and Bg in givingsimile quam: b hassi aliqua: HFT, Burn 243, Bodl.aliqua: M Harl. 4995, 4950, 4829, Burn. 244, Dorv. Csi nil aequalium. Among previous conjectures aresi multa aequalium, Törnebladh:si ille quaedam, Halm (whereilleis surely superfluous):si antiqua non, Jeep. Meister accepts the readingsi aliqua non: Becher thinks thatsi nil aequaliummay be right.

It is generally admitted that a word must have fallen out afterparum: the codd. all givesi parum non concupisset. Jeep proposedsi pravum(=corruptum: cp. ii. 5. 10)non conc.: on which Halm, comparingomnia sua, remarks, ‘debebat saltemprava.’ Butpravaseems too strong a word for Quintilian to have used in a criticism where he is so studiously mixing praise and censure. Halm suggestedsi parum sana, and is followed by Meister: cp. Fronto’s ‘febriculosa’ of Seneca, p. 155n. Sarpe proposedsi pravaorparvaorplura: Buttmannsi parum concupiscenda(orconvenientia): Herzogsi parvum: Madvigsi partimorpartem(i.e.paulo plus quam aliqua, and in opp. toomnia sua, below): Hoffmannsi opiparum: Seyffertsi garum: Kraffertsi non parum excussisset(cp.§101,§126: v. 7. 6; 7. 37; 13. 19: xii. 8. 13, &c.): Gustaffsonsi parva(cp. i. 6. 20 frivolae in parvis iactantiae): Andresensi similem ei quem contempsit se esse(sc.concupisset; cp. Tac. Ann. xiii. 56: xii. 64: Hist. i. 8: Livy xlv. 20. 9)si parem non concupisset(i.e.si Ciceronianum genus dicendi imitari quam diverso genere gloriam eius aemulari maluisset): or,nam si similem ei quem contempsit se esse, non parem concupisset: Krüger (3rd ed.)si parum arguta: Hertz (who argues that the word which has fallen out must, withparum, correspond tocorruptaabove)si parum pura.

utrimqueMeister and Becher, following old edd., Spalding, and Bonnell:utrumqueB N 7231, 7696:virumqueM:utcumqueHalm, ‘in every way,’ ‘one way or another,’—proposed by Gesner at6 §7.

§2.atque omnis. Kiderlin (Blätter f. d. bayer. Gymn. 1887, p. 454) proposes to put commas atsequiandvelimus, and make this clause also subordinate.

§3.aut similes aut dissimiles. Andresen suggestsaut similes aut non dissimilesoraut similes aut certe haud dissimiles.

§6.tradiderunt(BNM Harl. 2662, 4995, 4829, Burn. 243, and Dorv.) is powerfully supported by Becher in his latest tractate (Programm des königlichen Gymnasiums zu Aurich, p. 13) againsttradiderint, the reading of b Prat. Bodl. and Vall. (corrected in the last fromtradiderunt), Burmann, Spalding, Bonnell, Halm, Meister, and Krüger. Becher holds that in Quintilian, as frequently in Cicero,cumwith the indicative is often used in such a way (quoting from C. F. W. Müller) ‘ut non prorsus idem sit, sed simillimum ei, quod barbare dicere solemus identitatis. Nam ut “cum tacent clamant” non est “si tacent,” multo minus “quo tempore” aut “propterea quod” aut “quamquam,”—sed “tacent idque idem est ac si clament,” sic “cum hoc facis qui potes facere illud?” et sim., German, item “wenn du dies thust” valet: “hoc facis ex eoque per se efficitur, non ratione, sed ipsa natura, ut illud non possis facere.” Ut pro Q. Roscio 3. 9 quam ob rem, cum cetera nomina in ordinem referebas, hoc nomen in adversariis relinquebas? non significat nec “quamquam” nec “quando,” sed “wenn.”’ Becher adds the following parallel passages: Cic. pro Cluent. 47. 131 id ipsum quantae divinationis est scire innocentem fuisse reum, cum iudices sibidixeruntnon liquere, and Verg. Ecl. 3. 16 quid domini facient, audent cum talia fures? (Cp. Madvig de Fin. p. 25.) In the same way he treatscum ... sunt consecuti7 §19below, which seems, however, to be somewhat different. Here there is an antithesis, and in such casescum(‘whereas’) may very well take the indicative: there the clause ‘cum sint consecuti’ is added to show the reasonableness (cum= ‘since’) of the demand that extemporary facility shall be made fully equal tocogitatio—seead loc.Neither instance can be explained on the analogy ofcumwith the indic. used of ‘identity’ (as ‘cum tacent, clamant,’ quoted above): in such cases the subject is generally the same in both clauses. And in such a passage as pro Cluent. §131cumis usually explained as =quo tamen tempore.

eruendasM Harl. 4995: all other codd.erudiendas.

mensuris ac lineis. Krüger (3rd ed.) quotes with approval the conjecture of Friedländer (Darst. aus der Sittengesch. Roms iii. 4. p. 194. 4)eisdem mensuris ac lineis, and recommends the insertion ofeisdemin the text,—afterlineis, where it is more likely to have fallen out. But this is unnecessary.

§7.turpe etiam illud est. Hild puts a comma aftersciant, and by supplying beforeturpe estanitato correspond withquemadmodum, makes out a comparison of whichquemadmodum, &c., is the first clause andturpe etiam illud estthe second. This is certainly to misunderstand the passage. Thequemadmodumclause goes with what is before, not with what follows, so that a comma afteralieniwould be enough, were it not for the necessity of having the mark of interrogation (cp.§9below). Thenturpe etiam illud estcomes in, resumingpigri est ingeniiin§4, just as immediately afterwardsrursus quid erat futurum§7resumesquid enim futurum erat§4. The whole passage is an elaboration of the dictum with which§4opens, ‘imitatio per se ipsa non sufficit.’ Quintilian first says that we, as well as those who have gone before us, may make discoveries (cur igitur nefas est reperiri aliquid a nobis quod ante non fuerit?). Surely we arenotto confine ourselves to hard and fast lines like servile copyists.Then he goes on to add in§7that we must surpass our models (plus efficere eo quem sequimur), instead of resting content with mere reproduction (id consequi quod imitamur): otherwise Livius Andronicus would still be the prince of poets, we should still be sailing on rafts, and painting would still be nothing more than the tracing of outlines. The necessity for progress is first shown (§§4-6) by an appeal to the example of the past, and by the unfruitful work of such painters as are mere copyists: then in§7poetry, history, navigation, as well as painting are put in evidence for the argumentex contrario.

§8.mansit, Meister:sitcodd.:estFleckeisen (and Halm):fuitGensler.

§9.adpetentBg HFT:appetentPrat. Ioan. Harl. 4995 Bodl. &c.:appetuntN Harl. 2662, 11671, Burn. 243.

hoc agitHalm, followed by Meister (cp.7 §4):hoc aitb H,om. Bn Bg N Ioan. Prat. Harl. 2662, 11671:agit(sine hoc) Harl. 4995, 4950 M, and most codd.

§10.quaeque pares maximemay be a gloss: it is found only in those MSS. which givesimplicissimaeforsimillimae: b H Harl. 4950 M Burn. 243 Bodl.

utique(b M Vall. Harl. 4995, 4950, Burn. 243 Bodl. Dorv.) may also be suspected: it does not occur in Bn Bg N Ioan. Prat. Harl. 2662, 4829, 11671.

§11.orationibus, Bg: Ioan, givesoratione: so also Voss. 1 and 3 (Zumpt).

accommodaturb H Ioan. Harl. 4995, 4950, 4829, Bodl. Dorv. and Meister:commodaturBn N Prat. Harl. 2662, 11671, and Halm.

§12.inventio visB Harl. 2662, 11671:inventionisb H Harl. 4495, 4950, 4829, C, Burn. 243, Bodl., Dorv.

§13.cum et, ed. Colon. 1527:et cumB H Ioan. Prat. N (et quum) M:cumVall. Harl. 4995. On the usual interpretation of this difficult passageut quorum ... collocata suntforms one parenthesis: but this is an unnecessary extension of the explanation ofintercidant invalescantque temporibus. Seead loc.

accommodata sit, codd. except Harl. 4995, which omitssit:acc. estHalm, followed by Hild (depending onprout, notcum: see notead loc.). Madvig’s conjectureaccommodanda sitis approved by Kiderlin (cp. ix. 4. 126 adeoque rebus accommodanda compositio). But the correctness of the reading in the text (and also of the explanation given in the notead loc.) will be evident to any one who considers the whole sentence carefully. Tocum et verba intercidantcorresponds exactly the double clauseet compositio ... rebus accommodata siton the one hand, andet compositio ... ipsa varietate gratissima(sc.sit—repeated fromaccommodata sit) on the other. This double clause is rather awkwardly joined bycum ... tum. To takeaccommodata sitas depending on thecumwhich followscompositiois to destroy the balance of the sentence. In this case an independentsitwould have to be supplied withgratissima(to makeet compositio ... gratissima sitcorrespond toet verba intercidantabove): and the translation would then be: ‘it is just when (cum ... tum), or exactly in proportion as, it is adapted to the sense (rebus accommodata) that the very variety (thereby secured) gives the arrangement its greatest charm.’ But if this had been Quintilian’s meaning he would surely have writtencum rebus accommodatur(or—ata est)tum ipsa varietate sit gratissima.

§14.quos imitemur. The D’Orville MS. givesquos eligamus ad imitandum,—probably an emendation by the copyist, though it may explain the origin of the reading of b and Hquos at imitandum.

quid sit ad quod nos. Theadis due to Regius: most codd. havequid sit quod nos, except Harl. 4995, which is again in agreement with Goth. Vall. Voss. 2 and the second hand in Par. 2:quid sit quod nobis.

§15.et a doctis, inter ipsos etiam. The explanation given in the notes is due to Andresen (Rhein. Mus. 30, p. 521), who, however, wished to insertetbeforeinteripsos. The comma makes that unnecessary. So Kiderlin (Berl. Jahrb. XIV, 1888, p. 71 sq.).

dicunt, Harl. 4995:dicantall codd.: ‘emend. Badius’ (Halm).

ut sic dixerimVall. (Becher): cp. pr. 23: i. 6. 1: ii. 13. 9: v. 13. 2. BM Prat. haveut dixerim. Halm wroteut ita dixerim, comparing i. 12. 2: ix. 4. 61: butut sicis more common in the Latinity of the Silver Age.

§16.compositis exultantes. Kiderlin (Berl. Jahrb. XIV, 1888, p. 72) would prefercompositis rigidi(cp. xi. 3. 32: xii. 10. 7: ix. 3. 101: xii. 10. 33),comptis(cp. i. 79: viii. 3. 42)exultantes= ‘statt wohlgeordnet steif, statt schmuckliebend putzsüchtig.’ Another unnecessary emendation islaetis exultantes, compositis corrupti(Lindau): orcompositis exiles(Düntzner).

§17.quidlibet, most codd.:quamlibetM, Vall. Harl. 4995, 4950:qui licetbH. Iwan Müller (Bursian’s Jahresb. 1879, p. 162) condemnsillud, and would read eitherquamlibet frigidum(cp.3 §19and ix. 2. 67: quamlibet apertum), orquidlibet frigidum, which latter is approved by P. Hirt. Eussner suggests the deletion ofillud frigidum et inane, thinking that these words may be the remains of a gloss on§16.

Attici sunt scilicet. Spalding’s reading seems on the whole to be preferred. The retention ofsunt(represented in some MSS. by a simples,—hence the readingAtticis scilicet) makes it less necessary to follow Meister in inserting asuntafterqui praec. concl. obscuri: in so loose a writer as Quintilian the firstsuntwould do duty for both. Halm follows Bn and Bg, which apparently (as also N Harl. 2662, 4829, and 11671) haveAttici scilicet: Meister (with bHM and Harl. 4950) givesAtticis scilicet. In the Ioannensis I findAttici s(forsunt): Dorv. and Burn. 244 giveAtticis s. Scilicet(om. Prat.) may be a gloss, and the true reading may beAttici sunt. Some codd. (Bodl. Burn. 243) giveAtticos scilicet(AthicosHarl. 4995): qy.Atticorum similes? (cp. Cic. Brut. §287).—Becher now prefersAtticis(sc.se pares credunt).

§22.proposito. This conjecture by Gertz (Opuscula philol. &c., p. 134) I have found in the Ioannensis (*ppo) and in Harl. 2662 and 11671. It is approved also by Kiderlin. BNHb Prat. Sal. givepropositio: all other codd.proposita. Perhaps we should read (with Ioan.)sua cuique proposito est lex, suus decor est. Prat. omits the secondest.

§23.tenuitas aut iucunditas, Halm and Meister:tenuitas ac iucunditasb H, Burn. 243, Bodl.:tenuitas aut nuditasN Ioan. M Harl. 2662, 11671:tenuitas ac nuditasPrat. Harl. 4995, 4950, 4829, C, Burn. 244, Dorv.:aut iuditasBg.

§25.quid ergo? non est satis, &c. Gertz proposes to read, shortly afterwards,mihi quidem satis esset; set si omnia consequi possem, quid tamen noceret vim Caesaris ... adsumere?(=sed etiam si satis mihi esset, tamen nihil noceret vim Caesaris ... adsumere, si omnia haec consequi possem).

§28.deerunt, Francius:deerant(derant) all codd. Becher defendsdeerant: ‘der Rhetor meint dassqui propria bona adieceritöfter Veranlassung gehabt haben wird, Fehlendes zu ergänzen als zu beschneidensi quid redundabit.’

oporteatbHFT Bodl. M Harl. 4950 Burn. 243:oportebatB Prat. N Sal. Ioan. Harl. 2662, 4995, 4829, 11671 Burn. 244 Dorv. The latter (which is adopted by Halm) would indicate (cp. viii. 4. 22) a condition which ought to have been and may still be realised: the former (adopted by Meister and approved by Becher) is the conjunctive potential, and is quite in Quintilian’s manner (cp. xi. 2. 20): it conveys the expression of a present duty and obligation, the realisation of which may now be expected, and it connects also more intimately witheritin the following sentence.

Back to IndexNext