B

G

=

=

conspirantes lectiones Bernensis et Bambergensis.

codicis Bambergensis eae partes quae alia manu suppletae sunt. Introd. p. lviii.

b = manus secunda codicis Bambergensis.

H = codex Harleianus (2664) s. x-xi. Introd. p. Ixiv, sqq.

For the above (with the exception of H and Ioan. and a fresh collation of Bg and G) I have depended on Spalding, Halm, and Meister. In the same way I quote references occasionally to M (codex Monacensis s. xv), S (codex Argentoratensis s. xv), and L (codex Lassbergensis s. xv), the Gothanus, Guelferbytanus, Vossiani, &c.

A collation of the following has kindly been put at my disposal by M. Ch. Fierville, Censeur des études au Lycée Charlemagne (Introd. p. lxi, sqq.):—

Codex Pratensis (Prat.) s. xii.

Codex Puteanus (Put.) s. xiii.

Codex Parisinus (7231) s. xii.

Codex Parisinus (7696) s. xii.

Codex Salmantinus (Sal.) s. xii-xiii.

The readings of the Codex Vallensis (Vall.) are given from Becher's Programm des königlichen Gymnasiums zu Aurich, Ostern, 1891.

Other 15th cent. MSS., which I have specially collated for this edition, are the following (Introd. p. lxxiii, sqq.) :

Codex Harleianus 2662 (Harl. 2662). The inscription on this codex bears that it was finished 25th Jan., 1434.

Codex Harleianus 11671 (Harl. 11671), bearing date 1467.

Codex Harleianus 4995 (Harl. 4995), dated 5th July, 1470.

Codex Harleianus 4950 (Harl. 4950).

Codex Harleianus 4829 (Harl. 4829).

Codex Burneianus 243 (Burn. 243).

Codex Burneianus 244 (Burn. 244).

Codex Balliolensis (Ball.). This MS. is mutilated, and contains nothing after x. 6, 4 : there is moreover a lacuna from ch. ii to iii § 26.

Codex Dorvilianus (Dorv.), in the Bodleian at Oxford (codd. man. x. 1, 1, 13). Codex Bodleianus (Bodl.).

The readings of the Codex Carcassonensis (C-15th cent.) are given from M. Fierville's collation (De Quintilianeis Codicibus, Paris, 1874).

CHAPTER I.

§ 1. cognitioni, Harl. 4995: Burn. 243 (and so Gothanus, Spald.). Cogitationi G and most codd., probably mistaking a contraction in the ancient text.

§ 2. sciet G. The reading scierit (Harl. 4995 and many codd.) is probably due to H, which gives sciuit (so FT).

quae quoque sint modo dicenda. So GHFTL, and Halm. The alternative reading is quo quaeque s. m. d., S and all my 15th cent. MSS: Spalding and Meister, with the approval of Becher. See note ad loc. In the parallel passages i. 8. I Halm adopts Spalding's reading (ut sciat) quo quidque flexu... dicendum for quid quoque ABMS, and i. 6. 16 (notatum) quo quidque modo caderet for quid quoque BMS, and so Meister: Fierville returns to the reading of the MSS. In support of quo quaeque other exx. might be cited: v. 10. 17 quo quaeque modo res vitari vel appeti soleat, and vi. 4. 22 quo quaeque ordine probatio sit proferenda. But the parallel instances in the Tenth Book quoted in the notes (1 § 8: 7 §§ 5 and 6) seem to guarantee the correctness of the reading of the oldest MSS, though it is better to take quoque as the ablative of quisque than (as Halm) as the relative with -que.

:

tamen: GHFT Harl. 4950: tanquam Harl. 2662, 11671, 4995, 4829, L S Bodl. Ball. Burn. 243 Dorv. In Burn. 244 tanquam is corrected to tamen. Paratam explains in procinctu: so that tanquam is not so necessary as velut in xii. 9. 21.

ante omnia est: so all codd., and Halm. Hirt (Jahresb. des philol. Vereins zu Berlin viii. p. 69 sq. 1882: ix. p. 312 sq. 1883) conjectured ante omnia necessarium est, and this is approved by Kiderlin (Blätter f. d. bayer. Gymn. 1887, p. 454): cp. necessarium just above, and necessaria in § 1. Schöll (Rh. Mus. 34, p. 84) first challenged the MS. reading, and suggested that the original may have been ante omnia stat atque, corrupted into ante omniast [at] atque: for which use of sto, see Bonn. Lex. s. v. ii. y. As an alternative suggestion he put forward ante omnia necesse est, and this was adopted by Meister. Becher (Phil. Rundsch. iii. 14. 428) proposed ante omnia sciet, though more recently he has signified his adherence to the tradition of the MSS. Maehly suggested ante omnia opus esse. Perhaps the true reading may be ante omnia prodest.

The question depends to some extent on the treatment of the following passage. GH agree in giving proximam deinde inimitationem novissimam scribendi quoque diligentia. This Halm converted into proximum deinde imitatio est, novissimum . . . diligentia,-where the est is certainly superfluous (cp. i. 3. 1), while it may be doubted (comparing ii. 13. I and iii. 6. 81-Kiderlin 1. c.) whether proxima deinde imitatio, novissima &c. would not be a sufficient change: Kiderlin compares 'proxima huic narratio,' ii. 13. 1, and 'novissima qualitas superest,' and objects to the citation of 'proximum imitatio,' in 1. 31, in support of the neuter, on the ground that there 'signum ingenii' is to be supplied.

Kiderlin's proposed modification of Gemoll's conjecture (l. c. p. 454 note, cp. Rhein. Mus. 46 p. 10 note) proximum deinde multa lectio is adopted by Krüger (3rd ed.), who thinks that the sequence of thought makes the special mention of legere (alongside of dicere and scribere) a necessity: multa corresponds to diligentia in what follows: cp. multa lectione § 10. But legere has already been touched on in § 2, and moreover is included under imitatio (sc. exemplorum ex lectione et auditione repetitorum).

§ 4. iam opere. So Harl. 4995 and Regius: all other codd. iam opere iam. Becher reports iam opere also from the Vallensis.

qua ratione. For qua in oratione, the reading of all MSS., Hirt conjectured

qua exercitatione. Schöll proposed to reject in oratione as a gloss: but qua by itself (sc. via) is only used by Quint. with verbs of motion: see on 7 § 11.

In his latest paper (Rheinisches Museum, 46, pp. 10-13, 1891), Kiderlin subjects the whole of § 4 to a searching and destructive analysis. He translates: doch nicht darüber, wie der Redner heranzubilden ist, sprechen wir in diesem Abschnitte (denn dies ist genügend oder wenigstens so gut, als wir konnten, besprochen worden) sondern darüber, durch welche Art von Uebung der Athlet, welcher alle Bewegungen von seinem Lehrer bereits genau erlernt hat, für die Kämpfe vorzubereiten ist.' He doubts whether such passages as § 33 and 7 § 1 can be cited to justify the abrupt transition from orator to athlete, on the ground of the formal antithesis in which the two stand to each other,-'orator' coming in at the end of one clause, and ‘athleta' standing at the head of another, in front of 'quo genere exercitationis.' And yet it is just the 'orator' who is to be understood in the athleta.' As to the sentence introduced by 'Igitur eum,' if by' athleta qui omnes iam perdidicerit a praeceptore numeros we are to understand one who has mastered the whole theory of rhetoric, then it adds nothing to what has been said already, and is therefore altogether superfluous.

Kiderlin proposes to read: sed ut (so L and S,-also Harl. 2662, 4995) athleta, qui omnes iam perdidicerit a praeceptore numeros, multo (nonnullo?) varioque (numuro quae G,-also H: num muro quae T: numeroque F L: nimirum quo S) genere exercitationis ad certamina praeparandus erit (sit, the codd.) ita (so S,—also Harl. 2662, 4995 and Bodl.) eum, qui ... perceperit, instruamus, qua in praeparatione (qua in oratione, the codd.) quod didicerit facere quam optime, quam facillime possit. Ut may easily, he contends, have fallen out before at: and the running of three words into one (numeros multo vario-numero) is paralleled by such a case as § 23, where it will be found that Kiderlin sees ut duo tresque in utrisque. For 'multo varioque' he compares viii. 5. 28 multis ac variis: x. 5. 3 multas ac varias: xi. 3. 163 varia et multiplex: xii. 1. 7 totae tam variis; and, for varioque,' vii. 3. 16 latiore varioque, and xii. 10. 36 sublimes variique. Vario genere' actually occurs i. 10. 7, and multo may easily have been written in the singular, like nonnullus vi. 3. 11 (hoc nonnullam observationem habet) and elsewhere. The motive for changing que, quae, into quo and erit (est) into sit may have been the analogy of the foregoing quomodo sit. As for ut (sicut) ita (sic), it is so favourite a form with Quintilian that he uses it seven times in the first nineteen paragraphs of this chapter. Qua in oratione, the reading of all MSS., may have resulted from qua in praeparatione more probably than from qua ratione, which appears first in the ed. Col. 1527, and is not so appropriate to the context as qua in praeparatione (cp. praeparandus above, and parandae below). Quintilian is detailing in this Book on what preparation (cp. praeparant § 35, comparant § 67, praeparetur 6 § 6, praeparantur 7 § 19) the orator may best and most easily carry out in practice what he has learnt theoretically. For the preposition (in praeparatione) cp. viii. pr. 22: ut in hac diligentia deterior etiam fiat oratio.

The text of Quintilian, especially of this part of the Tenth Book, is admittedly very defective, and invites emendation: there is a great deal to be said for the theory that in many places several words must have dropped out. Kiderlin's attempts to remedy existing defects are always marked by the greatest ingenuity: they are all well worth recording as evidences of critical ability and insight, even though it may be that not all of them will be received into the ultimate text. Here there seems no reason why Quintilian, who was notoriously a loose writer, should not have said in the concluding sentence of the paragraph what he had already said, in the form of a metaphor, in the clause immediately preceding. Indeed the word igitur seems to suggest that after indulging in his favourite metaphor (sed athleta, &c.) he wishes to resume, as it were, and is now going on to say what he means in more ordinary language. It may not be artistic but it is Quintilian. If he had had some of his modern critics at

his side when preparing a second edition of the Institutio some of his angularities might have been smoothed away.

§ 5. Non ergo. Meister and 'edd. vett.': I find this reading in Harl. 4995, and Burn. 243. So Vall. Halm. has Num ergo, and so most codd. (including HFT Bodl. and Ball.).

§ 6. ex his. Qy. ex iis? so § 128: cp. Introd. P. xlix.

§ 7. quo idem, Meister and ‘edd. vett.': quod idem Halm, supported by Becher and Hirt, perhaps rightly. Nearly all my MSS. agree with GLS in quod: quo occurs in Harl. 4995 only.

§ 8. quod quoque GH Halm, Meister: quid quoque (as 7 § 5) occurs in L S, also in Bodl., Ball. For quid Zumpt cites also Par. 1 and 2: i. c. 7723 and 7724 (Fierville). Aptissimum (strangely mangled in most codd.-e. g. locis ita petissimum G) is given rightly in Dorv.

§ 9. omnibus enim fere verbis. This reading, ascribed by Meister to Badius, and by Halm to ed. Colon. (1527), I have found in Harl. 4995 (A. D. 1470): ferebis vel GH: fere rebus vel L S Harl. 2662, 4950, 4829. From the Vallensis Becher reports fere verbis vel.

intueri, ed. Col. 1527. In Harl. 11671 I find interim intueri : Harl. 2662 LS Ball., Dorv., Bodl., interim tueri.

quae nitidiore in parte occurs first in ed. Col. 1527: Vall.' Harl. 4995 Goth. Voss. ii. shows quae cultiore in p.: GH quaetidiorem in p.: LS Harl. 2662 Guelf. Bodl. quae utiliore in p.

§ 10. cum omnem, &c. cum omnem misermonem a. pr. accipiamus GH: cum omnem enim, most codd. Osann, followed by Gemoll and Krüger (3rd ed.), suggested omnem enim sermonem a. pr. accipimus.

§ 11. alia vero, Frotscher: aliave GH: aliaque Harl. 4995. This last Becher now prefers (alia que Vall.: alia quae Regius), comparing ix. 3. 89 and ix. 4. 87.

Tpomikās quasi tamen, Spalding, Zumpt, Meister and Krüger (3rd ed.): tropicos quare tam GH, quare tamen, later MSS. Halm obelized quare tamen: Mayor only quare. Becher recommends tamen by itself. Gensler (Anal. p. 25) reads tamen quasi, and is followed by Hild, who takes quasi with feruntur in the sense of referuntur μerapopá): Zumpt took it with eundem intellectum. Gemoll approves of the exclusion of quare, which he thinks must have arisen from a gloss figurate (either marginal or interlinear) on Tрomuŵs. Kiderlin adopts this and thinks the quare tam of GHL a mutilation of the gloss figurate: gurate and quare tã are not far apart.

§ 12. figurarum G (per compendium) : figuranus H. Kiderlin suggests mutuatione figurarum, sc, ostendimus: after which Quintilian continues 'sed etiam ex proximo mutuari licet.' Cp. Cic. de Or. iii. 156 translationes quasi mutuationes sunt. Kiderlin adds (Rhein. Mus. 46, p. 14 note) that in iii. 4. 14 all MSS. wrongly give mutantes for mutuantes, and in i. 4. 7 A1 has mutamur for mutuamur.

§ 15. hoc sunt exempla potentiora. Hoc is a conj. of Regius (also Vall.3), all the MSS. giving haec (hec). Hoc appears in the Basle ed. of 1555 and in that of Leyden 1665. It is challenged by Schöll (Rhein. Mus. 44, p. 85), who says quia stands too far away from hoc to allow of such a construction, and thinks the context has been misunderstood. According to him haec exempla (those derived from lectio and auditio) are set over against those which one gets in theoretical books and lectures: they are more telling, because they act directly on the mind, and are not served up as dry theory in the form of extracts ('quia quae doctor praecepit orator ostendit'). He therefore understands ipsis (exemplis) quae traduntur artibus,' but admits that ' etiam' is thus otiose, and would therefore read quam ipsis quae traduntur artibus.

Schöll is supported by Hirt (Jahresb. des philol. Vereins zu Berlin, 1882, p. 70), who thus gives the sense of the passage: Der Wortschatz wird durch Lektüre und vieles

Hören erworben. Aber nicht nur seinetwegen soll man lesen und hören; man soll es auch noch aus einem anderen Grunde. In allem nämlich, was wir lehren, sind diese Beispiele, d. h. diejenigen, welche uns die Lektüre und der Vortrag bieten, wichtiger selbst als die Beispiele welche die Handbücher und Vorlesungen darbieten, weil, was der Lehrer nur als Forderung aufstellt, bei dem Redner That geworden ist und sich durch den Erfolg bewährt hat.'

Iwan Müller (Bursian's Jahresb. vii. 1879, 2, p. 168) objects that if Quintilian had wished to convey this meaning he would have said, not haec exempla, but hinc ducta (petita) or quae hinc ducuntur (petuntur) exempla; and he rightly desiderates also quam quae (in) ipsis traduntur artibus. Meister also opposes Schöll (Philol. xlii. p. 149): the order quam ipsis quae traduntur artibus is in fact impossible.

On the whole it seems much better to keep hoc, and to understand: 'in all instruction, example is better than precept: the doctor relies only on precept, the orator on example.'

Gertz conjectures nam omnium quaecunque docemus hinc (cp. v. 10. 5: xii. 2. 31) sunt exempla, potentiora (i. e. quae potentiora sunt) etiam ipsis quae traduntur artibus. But with hinc, as Kiderlin observes, some other verb than sunt would be expected: v. 10. 15 is an uncertain conjecture, the MSS. giving nihil, and in xii. 2. 31 hine belongs to bibat and sumptam. Kiderlin himself at first proposed haec praestant exempla, potentiora: this he now withdraws, however, (Rhein. Mus. 46, p. 15) in favour of haec suggerunt exempla, potentiora, &c. By haec he understands legere and audire, and gives the sequence of thought as follows::-'Aber wenn auch auf diese Weise eine Fülle von Ausdrücken erworben wird, so ist das doch nicht der einzige Zweck des Lesens und Hörens. Denn von allem was wir lehren (nicht nur von den Ausdrücken) liefert dieses (das Lesen und Hören) Beispiele, welche noch wirksamer sind als die vorgetragenen Theorieen selbst (wenn der Lernende so weit gefördert ist, dass er die Beispiele ohne Beihilfe verstehen und sie bereits aus eigener Kraft befolgen kann), weil der Redner das zeigt, was der Lehrer nur vorgeschrieben hat.' For suggerere Kiderlin compares i. 10. 7 artibus, quae . . . vim occultam suggerunt, and v. 7. 8 ea res suggeret materiam interrogationi: cp. also § 13 quorum nobis ubertatém ac divitias dabit lectio, and ii. 2. 8 licet satis exemplorum ad imitandum ex lectione suppeditet.

rerum.

...

§ 16. imagine et ambitu rerum: so Harl. 2662 L S Ball. Burn. 243 and Bodl.: followed by Spalding, Frotscher, Herbst, and Bonnell. GH give imagine ambitu Halm (after Bursian) bracketed ambitu : but it is more probable that imagine is a gloss on ambitu than vice versa (so Hirt and Kiderlin), and Meister accordingly (followed by Krüger 3rd ed.) reads [imagine] ambitu rerum. It seems just as likely, however, that et has fallen out. Hertz suggested imagine ambituve rerum: Maehly thinks that ambitu was originally tantum.

nec fortuna modo. Gertz proposed nec forma modo: pro Mil. § 1 movet nos forma ipsa et species veri iudicii.

§ 17. accommodata ut: ed. Col. 1527, and so Meister and Krüger (3rd ed.): commodata ut Halm (after Bursian): commoda ut Spald., Frotsch., Herbst, and Bonnell. GHS give commoda aut: Land all my MSS commoda ut (except Burn. 243 which shows comendat ut).

et, ut semel dicam. Kiderlin would delete et, rendering' Stimme, Aktion, Vortrag ist, um es kurz zu sagen, alles in gleicher Weise belehrend.'

§ 18. placent-laudantur-placent: so Halm and most edd., following S, with which all my MSS. agree. The emphasis gained by the opposition of placent and non placent makes this reading probable. But GH give laudetur: and so Meister and Krüger (3rd ed.) prefer to follow Regius in reading placeant-laudentur—placent..

§ 19. e contrario. This reading, which Meister adopts from 'edd. vett.,' occurs in