¹ Precise dates are few and far between, but approximate ones tell a sufficient tale. For evidence and discussion the following in particular may be consulted: for Quintilian, Peterson, W., Quint. I.O. 10 (1891)Google Scholar; Schwabe, L., RE 6 (1909)Google Scholar, Fabius 137, cols. 1845ff.; Colson, F. H., I.O. 1 (1924)Google Scholar, Introd. ch. 1; Stein, A., PIR ², Pars 3 (1943)Google Scholar, Fabius 59, pp. 109–10; also Kennedy's, G. brief but well-documented and critical study, Quintilian (New York, 1959), pp. 15 ffGoogle Scholar. (although we disagree on the evidence for De causis; cf. below n. 48); the Tacitean dates are very economically presented by Stein, in PIR ², Pars 2 (1936)Google Scholar, Cornelius 1467, pp. 365–7. Syme's, R.Tacitus (1958)Google Scholar contains carefully pondered views also on matters of chronology.

² The conventional date for Quintilian's birth is ‘c. 35’, though with the qualifications noted, e.g. by Peterson, iii–iv, Schwabe, col. 1845, Colson, ix–x; Syme, p. 108 n. 6 even suggests c. 33. I incline towards 40 +, cf. Kennedy, p. 15, ‘he was studying rhetoric in Rome about A.D. 57… (being) somewhere between fourteen and twenty at the time. This suggests that he was born a few years before or after A.D. 40, probably not so early as 35, which is often given as the date of his birth.’ In turn ‘the birth of Tacitus may be assigned to 56 or 57’, Syme, p. 63, cf. Stein, PIR ², Cornelius 1467, p. 365Google Scholar, ‘natus non post a. 58, quoniam a. 88 praetor fuit.’ On any of these reckonings Quintilian belonged to the older generation of contemporaries from Tacitus’ point of view; Julius Secundus, one of Tacitus’ teachers (Dial. 2.1), was Quintilian's coeval (I.O. 10.3.12 aequalem meurri).

³ Hier. Chron. a. 2104 = A.D. 88 (p. 190 Helm) ex Hispania Calagurritanus, Auson. Prof. Burdig. 1.7 (p. 49 Peiper) adserat… Fabium Calagurris alumnum.

⁴ The best account of the brittle evidence I am aware of is given by Syme, in Tacitus 2, ch. 45Google Scholar ‘The origin of Cornelius Tacitus', especially p. 619 on Plin. Ep. 9.23.2 ‘Italicus es an prouincialis?’ and p. 622 on the case for regarding Cisalpine Gaul or Gallia Narbonensis as his palria.

⁵ Romanilas in that sense is obvious in Tacitus' case. As for Quintilian's Romanitas, it is expedient to remember Syme's, question, Tacitus, p. 618Google Scholar: ‘But who would have surmised the origin of Quintilian (in contrast to Martial) if it were not attested? That author mentions Spain only once – and there it is with the curious affectation of not knowing much about a certain local word’.

⁶ Quintilian's talent was first officially noticed by Galba in the sixties. Under Vespasian his standing as a barrister and teacher must have been sufficient to justify appointment as holder of the first public Roman chairs of Latin and Greek rhetoric, the stipend to be defrayed from the imperial purse (Hier. Chron. a.2084 = A.D. 68, and a.2104 = A.D. 88 [from Suet.], pp. 186, 190 Helm; cf. Suet, . Vesp. 18)Google Scholar. The chronology is not without snags; for discussion, see Schwabe (above n. 1), col. 1849, and Kennedy (above n. 1), 19, 142.

⁷ Classicism it is occasionally called, thus in the instructive survey by Leeman, A. D., Orationis Ratio (Amsterdam, 1963)Google Scholar, ch. 12 ‘The classicist movement’, ch. 13 ‘Classicist oratory’, ch. 14 ‘Classicist historiography’. For attempts at greater precision of terminology, see T. Gelzer, ‘Klassizismus, Attizismus and Asianismus’ in Le classicisme a Rome etc. Fond. Hardt, , Entret. 25 (1978), 1–55Google Scholar, K. Heldmann, below n. 48.

⁸ So in a famous pronouncement Tacitus himself put on record to avow impartiality ‘despite benefits owed to the dynasty’ (Syme, p. 63). Hist. 1.1.3.

⁹ To mention only the honores on record, held or designated certainly or possibly under the Flavians: praetor (and xvuir sacr.fac.) A.D. 88: Ann. 11.11.1; cos. (suff.) in 97: Plin. Ep. 2.1.6, for the possibility of designation by Domitian, and Tacitus' earlier posts, see Syme, p. 70 and App. 17.

¹⁰ Quint. 1.0 1 pr. 1 post impetratam studiis quietem, quae per uiginti annos erudiendis iuuenibus impenderem. The retirement, quies, cannot have been very quiet. He suffered great personal misfortunes – the loss of his young wife and his two sons (I.O. 6 pr.). He was appointed tutor to Domitian's great-nephews (the emperor's intended heirs, Suet. Dom. 15.1), scarcely a sinecure.

¹¹ A.D. 89 according to Kennedy, p. 22 with n. 26; p. 88, Schwabe, col. 1852.

¹² The date is discussed by Schwabe, cols. 1885–7, Kennedy, pp. 26–8, al.

¹³ This may be inferred from Plin. Ep. 2.14.9 with Sherwin-White's note, cf. also 6.6.3; Colson, xviii, unconvincingly doubts the inference.

¹⁴ Tac. Agr. 3.1 nunc demum redit animus, et quamquam primo statim beatissimi saeculi ortu Nerua Caesar res olim dissociabiles miscuerit principatum ac libertatem, augeatque felicitatem lempomm Nerua Traianus etc. But at 44.5 Trajan is called princeps. Cf. Syme, p. 19. The view of the Agricola as Tacitus' first literary publication is criticized by C. E. Murgia – see n. 15.

¹⁵ Murgia, C. E. strongly argues in favour of dating the Dialogus before the Agricola in A.D. 97:Google Scholar‘The Date of Tacitus' Dialogus’, HSCP 84 (1980), 99–125Google Scholar, and ‘Pliny's Letters and the Dialogus’, HCSP 89 (1985), 171–206Google Scholar. I do not go into this question here.

¹⁸ Thus Bornecque, H., Les Déclamations el les rhéteurs d'après Sénèque le père (1902)Google Scholar (also his text and translation, 2nd ed. 1932); Bardon, H., Le Vocabulaire de la critique littèraire chez Sénèque le rhéteur (1940)Google Scholar; Parks, E. P., The Roman Rhetorical Schools as a Preparation for the Courts under the Early Empire (1945)Google Scholar; and especially Bonner, S. F., Roman Declamation in the Late Republic and Early Empire (1949)Google Scholar; A. D. Leeman (1963) (above n. 7), ch. 9 ‘Orators and rhetoricians in the Early Empire’; Winterbottom, M., Roman Declamation (1980)Google Scholar (also introd. to his text and translation of Seneca's Controversiae and Suasoriae, 1967), and the same author's papers, ‘Quintilian and the Vir Bonus’, JRS 54 (1964), 90–7Google Scholar, and ‘Cicero and the Silver Age’, Fond. Hardt, , Entret. 28 (1981), 237–74Google Scholar, ‘Ars Rhetorica antica e nuova’, Univ. di Genoa, Fac. di lettere (1983), 57–76, ‘Quintilian and Declamation’ in Hommages a Jean Cousin, Inst. Gaffiot, F., vol. 1 (n.d., apparently also 1983), 225–35Google Scholar; D.A.Russell, Greek Declamation (Cambridge, 1983)Google Scholar; Kennedy, G. A., The Art of Rhetoric in the Roman World (Princeton, 1972)Google Scholar, Greek Rhetoric under Christian Emperors (Princeton, 1983)Google Scholar.

¹⁷ Oxford Classical Dictionary ² (1970), F. A. G. Beck, s.v. Education; C. J. Fordyce, s.v. Declamation; but see next note.

¹⁸ Or perhaps more than a modicum. For S. F. Bonner (above n. 16), p. iii, makes a case for the thesis that ‘beneath the follies and extravagances of their (sc. the declaimers') oratorical exhibitions lay a far closer acquaintance with the Roman Law than had commonly been suspected’. Yet we would be well advised to remember M. Winterbottom's salutary warning against inferring from what he calls ‘declaimers 'laws’ any particular relationship to real laws: Winterbottom (1983) (above n. 16), 72.

¹⁹ Above n. 6.

²⁰ The first subjects assigned to the rhetor are still to be in some connexion with those taught by the grammaticus (2.4.1). All the exercises in Quintilian 2.4 are summed up by his at the beginning of this Extract. Among these there is an introduction to what is called thesis (2.4.25–6).

²¹ The order of words in Quintilian's reference to Demetrius may seem to facilitate the view, which is occasionally expressed, that ut alio quoque libro sum confessus relates to ab ipso… inuentum, so that a change of the author's mind from the De causis to the Institutio would be indicated by himself: ‘on whether he himself invented this kind of exercise (as I acknowledged in another book), I have insufficient information.’ Wilamowitz, Thus, Glaube d. Hell, ii (1932)Google Scholar, repr. 1955, ii. 537 n. 1 – inadvertently, as K. Heldmann (1982) pointed out (below n. 48), p. 119; correct, e.g. as Heldmann noted, in Spalding's ed. (1798); so more recent translations, also A. Reuter (1887) (below n. 38), 8. For quoque surely aligns the earlier book with the present: ‘as I acknowledged also in another book.’ For thesis and controuersia see Appendix 1.

²² For the provenance of declamation see Appendix 2.

²³ The similarity is often stressed, e.g. by H. Bornecque (1902) (above n. 16), ch. 6, esp. pp. 118ff. on Cassius Severus (Sen. Contr. 3 pr.) and Votienus Montanus (Sen, . Contr. 9 pr.)Google Scholar; by S. F. Bonner (1949) (above n. 16), Ch. 4 ‘Declamation and its ancient critics’, p. 73 on Cassius Severus, 73–4 on Votienus Montanus, and especially 80ff. on Quintilian's own views; and by later scholars. Thus too writers outside the special rhetorical field, as Petron, . Sat. 1.2Google Scholarui, cum in forum uenerinl, pulent se in alium orbem terrarum delatos, et al.

²⁴ Thus De or. 3.214, on impassioned delivery; the orators are adores ueritatis, not ‘agents in accordance with truth’ but ‘those who enact real life’ – a kind of delivery said to be now no longer in vogue and to be taken over by the stage ‘actors’, who are merely imitalores… ueritatis. (215) ac sine dubio in omni re uincit imitationem ueritas. Cf. Wilkins ad loc, Leeman-Pinkster, , De or. 1.77nGoogle Scholar. This is one of the Ciceronian notions that much impressed Quintilian in very different political circumstances. Though much favoured by Cicero, this idiomatic usage of ueritas appears before him, applied to actual pleading, at Ad Her. 4.32 cum in ueritate dicimus, ‘when we speak in an actual case’, as opposed to the oratory of mere display, epideixis. This, according to the TLL archives, is the earliest Latin instance of the special idiom on record, but the noun ueritas happens to be infrequent before Ad Her. and Cicero anyway. The adj. uerus, however, is common in the sense ‘real, actual’; see OLD s.v. uerus. The idiomatic usage is paralleled by Greek ⋯ληθ⋯ʗ, ⋯λ⋯θεια, as the dictt. show. Although a direct borrowing from the Greek in a rhetorical context cannot be excluded, direct dependence of the general Latin idiom, ‘real, actual’, is unlikely. (The special rhetorical application of ⋯ληθ⋯ʗ, ⋯λ⋯θεια is noted by Ernesti, J. C. T. in Lex. tech. Gr. rhet., 12Google Scholar, with reference to Dion. Hal. Demosth. ch. 32.1056.) It is not surprising that ueritas often appears in passages of Quintilian where he argues what above I have called his compromise: declamation is to be approved only if it prepares for the ‘reality’ of forensic or deliberative oratory. Thus above in the text of 5.12.22, and 2.20.4, 8.3.23, 10.2.12 in illis (sc. orationibus) uera, in his (sc.declamationibus) assimulata materia est, 10.5.14, 17 in falsa rerum imagine in contrast with 21 in declamando quam simillimum esse ueritati, et al. Occasionally however Quintilian uses uerum and ueritas for ‘truth': 2.10.12, al.

²⁵ At 7.10.6–7 the word licentia is not actually used and ignorantia is but alluded to, although he talks of wilfulness in the arrangement of a subject, e.g. cum fuerit de prima quaestione dicendum, passim et ut quidque in mentem ueniet miscuerit, etc. But in the treatment of simile and comparison, 8.3.76, he says, quod quidem genus a quibusdam declamatoria maxime licentia corruptum est.

²⁶ Masculine strength in contrast with (objectionable) female softness of style amounts to a motif in Quintilian. Thus 8.3.6 sed hie ornatus (repetam enim) uirilis el fortis et sanctus sit nee effeminatam leuitatem etfuco ementitum colorem amet: sanguine et uiribus niteat, cf. 9.4.142, al. This tone is familiar from Augustan Atticism. Dion. Hal, . Comp. 23.112Google Scholar (on polished or smooth composition, γλαøupsiv;ρ⋯ν εἔøων⋯ τε εἰναι βο⋯λεται π⋯ντα ⋯ν ⋯ματα κα⋯ παρθενωπ⋯, τραϰε⋯αιʗ δ⋯ ʗɛλλαβαβαῖʗ κα⋯ ⋯ντιτ⋯ποιʗ ⋯π⋯͐θεται πο, κτλ. Cf. Isocr. 13.560. The metaphor in this kind of context is certainly much older than Augustan Atticism. Thus Philod. Rhet. i. 198 (Sudhaus) citing Hieronymus of Rhodes, a Peripatetic of the 3rd century B.C. (Die Schule des Aristoteles, Hieron, fr. 52a and b, Wehrli), criticizing Isocrates' unvaried smoothness: τ⋯ δ⋯ κεκλαʗμ⋯νον… ⋯ποβεβληκ⋯ναι, τ⋯ι δ⋯ λει⋯τητι δι⋯ʗ δολε⋯ειν.

²⁷ Wilamowitz, op. cit. (above n. 21), ii. 537Google Scholar.

²⁸ op. cit. 442. It is of some interest to notice how much Wilamowitz's appreciation of Quintilian had grown from, say, the famous paper, ‘Asianismus und Attizismus’ published in 1900. There (Kl. Schr. 3, 246) Quintilian is a rhetorician ‘well intentioned but flat’ – ‘dem wohlmeinenden aber flachen Rhetor’.

²⁹ Wilamowitz op. cit. and Reuter (below n. 38), 57, though qualified 60, narrowly restrict Quintilian's causation to style. This view was rejected, quite rightly, by H. Drexler in his important survey of earlier Tacitean research, Jahresber.f. Altertumsw. 229, Suppl. 1929, 271–3, also by Giingerich, R., Gn. 27 (1955), 440Google Scholar, and Brugnoli, G., Orpheus 6 (1959), 36–7Google Scholar; they did not, however, draw appropriate conclusions from these observations, nor did they persuade many scholars that the observations are in fact just.

³⁰ Quint. I.O. Books 8–9, the former on diction and style, the latter on figures of thought and style, also on structure and rhythm.

³¹ At 186 ⋯νタμ⋯ζω is Gale's necessary correction for -ει cod.P.

³² For κακοζηλ⋯α see Appendix 3.

³³ Two famous Augustan instances may be cases in point – Augustus ap. Suet. Aug. 86 and the Donatus Vita Verg. 44 – but are too brief and epigrammatic to be much use in this context. On the other hand it has been noted (thus by Wilamowitz, Rhys Roberts, and Grube) that, surprisingly, Dionysius of Halicarnassus does not use the term.

³⁴ I refer particularly to Horace on Poetry ii. 105–13.

³⁵ Quint. I.O. 8.6.76.

³⁶ Hor, . A.P. 311Google Scholaruerbaque prouisam rem non inuila sequentur.

³⁷ A. D. Leeman (1963) (above n. 7), i. 278, with nn. 123 and 124 in ii.477–8. Others, in the last three decades or so, who have seen a reference to the De causis in Quintilian's remark on Seneca, although varying in their opinions on other features, are G. Brugnoli (1959) (above n. 29), Gelzer, T., MH 27 (1970), 212–23Google Scholar, K. Heldman n (1980) (below n. 48), pp. 12–19. Heldmann also raises strong points against Gelzer's treatment of Quintilian's passage as a case of rhetorical simulatio.

³⁸ Reuter, A., De Quintiliani libro quifuit De causis corr. eloq. (Diss., Göttingen, 1887)Google Scholar, which owed much to the teaching of Wilamowitz. The specific reference is to pp. 30–1. Schanz-Hosius, Gesch. d. röm. Lit. 2” (1935), p. 748, and Trillitzsch, W., Seneca im lit. Urteii d. Antike i (1971), p. 62Google Scholar, follow suit.

³⁹ The first passage comes from Wilamowitz, , ‘Asianismus und Attizismus’, Hermes 35 (1900), 25Google Scholar n. 1, repr. Kl. Schr. iii.246 n. 1, the second from D. Glaube d. Hell. (cit. above n. 21), ii.537. Wilamowitz was a man of strong intuitions; his perception of intellectual and artistic development was acute and often unquestionably right. Not always, however. An interesting case, also as it happens concerned with the development of historical sense (Lorenzo Valla's in that case), where biographical simplification overrode attention to evidence, was pointed out by Pfeiffer, R., Hist, of Class. Scholarship i. 39Google Scholar, Ausg. Schr. (1960), p. 164, Wilamowitz, criticizing, Reden u. Vortr. ⁴ (1926), p. 115Google Scholar, Gesch. d. Phil. p. 12.

⁴⁰ What little detail was mentioned by Wilamowitz desirably widened the subject. He did not, however, face the problem of the Institutio as a possible source for Tacitus in addition to the De causis; the matter is to be discussed below. And he was too hasty in claiming style as Quintilian's only causa for corrupta eloquentia. We have seen that there were several causae – an important point for what I believe was Tacitus' reaction to his aetiology. On the other hand he was by no means unaware of the difference between the rhetorician's corrupta eloquentia in the De causis, and the reduced eloquence for the causes of which Tacitus professes to search in the Dialogus.

⁴¹ K. Barwick (1) (1913), RhM 68, 279–85, largely, though not entirely, succeeded in determining the size of the lacuna at chs. 35–6 of the Dialogus; he returned to the question later to defend his position against protesters. The result is important, because a lacuna of this size would not accommodate a set speech by Secundus, on which in turn the layout of the speeches depends. (2) 1929, in Festschrift W. Judeich (Weimar), pp. 81–108, his best contribution to this subject, and indispensable for further consideration. Two results stand out: his comparison of the Messala speeches, Dial. chs. 28ff. on rhetorical education, theory, and practice, with Quintilian, especially De causis, and, secondly, his analysis of the layout of all the speeches in the Dialogus. But, as noted above, the work is not faultless. His statements on the Tacitean side of the account, which resulted in the simplistic formula ‘Messala-Quintilian’, followed by ‘Maternus-Tacitus’, are erroneous and have caused much confusion. (3) 1954, ‘Der Dialogus de oratoribus des Tacitus’, in Berichte… der sächs. Ak. d. Wiss. zu Leipzig, phil.-hist. K.1. 101.4, p. 42. This essay usefully places his earlier findings in a wider framework, but the same simplistic tendency mars it in certain places.

⁴² Syme, R., Tacitus i (1958), p. 114Google Scholar, and bibliography in n. 9, beginning with Dienel's, R. long paper, WS 37 (1915), 239–71Google Scholar, which still, painstakingly, had to prove a post-Quintilianean date for the Dialogus. Later scholars, from Bardon, H., REL 19 (1941), 113–31Google Scholar, onwards, had a n easier row to hoe; they could confront Institutio and Dialogus in more straightforward terms; cf. references in G. Kennedy, op. cit. (above n. 1), ch. 6 n. 13, to which add Kiihnert, F., ‘Quintilians Stellung zud. Beredsamkeit seiner Zeit”, Listy Fit. 57 (1964), 33–50Google Scholar, based on his Jena dissertation of 1951, Die Tendenz in Quintilians Institutio Oratorio, which remained unpublished, an d which I have not seen in typescript. Th e relation between Institutio and Dialogus was discussed by Güngerich, R. in a brief and acute paper, ‘D. Dial, des Tac. und Quint. I.O.’, CP 46 (1951), 159–64Google Scholar. Giingerich however is no t there concerned with the De causis.

⁴³ Murgia, C. E., HSCP 84 (1984), 100–1 with n. 10Google Scholar.

⁴⁴ I find it hard therefore to accept Murgia's suggestion, loc. cit. (preceding note), that the parallels between Institutio and Dialogus by their quantity and distribution render ‘untenable a theory that the Dialogus and the Institutio simply share a common source (such as an earlier work of Quintilian)’. Since the evidence as I see it points to the De causis as a source for Tacitus, the other two works (I mean Dialogus and Institutio), pace Murgia, do share a common source, but the sharing is not simple, since Tacitus is not unlikely to have drawn on the Institutio as well.

⁴⁵ The point that caused Quintilian here to mention his first work concerns Demetrius of Phalerum: above pp. 475–6. But the rhetorical context renders it likely that the conjunction of the Greek fictae materiae with the early Latin rhetores figured in the first work as well. Also there is no reason to guess that he refrained from muffling Cicero's condemnation of those early rhetores in the De causis (Extract 2), but did so muffle it in the corresponding passage, I.O. 2.4.42. After all he had a prejudice in favour of declamation then as well as later, although he well knew the drawbacks of the genre. Contrast Tac. Dial. 35.1 with Quintilian, De causis Extract 2. and consider Giingerich's observation, loc. cit. (above n. 42), 163.

⁴⁶ I here select Reuter (above n. 38), 40–1, 58–60; Wormser, G., Rev. Phil. 36 (1912)Google Scholar, especially 187–9; Dienel, R.W.S. 37 (1915)Google Scholar, e.g. 235; Drexler (above n. 29), 227; K. Barwick (above n. 41, 1929), 81–90, the basic discussion on Messala an d Quintilian, an d 106–8, o n the layout of the Dialogus; Keyssner, K., Wiirzburger St. z. Altertumsw. 9 (1936), 107Google Scholar; Bardon, H., R.E.L. 19 (1941)Google Scholar, especially 119–27; Barwick (above n. 41, 1954), which sets the author's observations of 1929 in a wider context. The formula ‘Messala-Quintilian’ appears e.g. at Wormser, loc. cit., Barwick, (n. 41, 1954), 15, 17.

⁴⁷ Thus e.g. den Boer, W., Mnem. 3. 7 (1939), 209–11Google Scholar; Gungerich (above n. 29), loc. cit., and on Tac. Dial. 30.1; Haussler, R., Philol. 113 (1969), 49–50CrossRefGoogle Scholar (but, at p. 47, he convincingly extols Barwick's merit in clarifying the outlines of the Dialogus as a whole); Bringmann, K., MH 27 (1970)Google Scholar, especially pp. 172–4.

⁴⁸ Kennedy, (above n. 1), pp. 22–4, 136–9Google Scholar, especially p. 138; even more emphatic, Heldmann, K., Poetica 12 (1980), 1–23Google Scholar, especially 19–21, and (1982) (Miinohen), Antike Theorien iib. Entwick. und Verfall d. Redekunst, especially 213–54, 255–86. At Poetica 12, 19 Heldmann says that Quintilian's ‘indications are so scanty that there is no sense in any attempt at reconstruction’. I agree, the evidence is insufficient for a ‘reconstruction’; see above p. 474. But I seek to demonstrate in this paper that it is possible nevertheless to know something substantial about the De causis, without reconstructing the lost work. Nor is the basic disagreement of Tacitus' Messala (who seems to regard the decline of oratory and the other arts as irretrievable) with Quintilian (who sees the decline but hopes to check it) a sufficient reason for excluding Tacitus' use, for largely polemical purposes, of Quintilian, as Heldmann appears to argue; cf. below n. 72.

⁴⁹ Barwick, 1929 and 1954, above n. 46. For Barwick on the relation between the Messala speeches and the rest of the Dialogus, see below p. 498.

⁵⁰ Quint. l.O. 1.1.1 falsa esl enim querela paucissimis hominibus uim percipiendi quae Iradantur esse concessam, plerosque uero laborem ac tempora tarditate ingeniiperdere; etc.; Tac. Dial. 28.2 quis enim ignorat et eloquentiam et ceteras artes desciuisse ab ilia uetere gloria non inopia hominum, sed etc. This agreement between the teacher, Quintilian, an d the idealizer of republican oratory, Messala, is on the surface only, since their aims differ. The arguments on the decline of oratory in the first imperial century have often been discussed. K. Heldmann's work of 1982 (above n. 48) offers the evidence most fully although I have expressed some disagreement with its analyses, Gn. 57 (1985)Google Scholar; cf. Appendix 2.

⁵¹ For reasons that will become apparent below, Tacitus' Messala speaks of obliuio moris antiqui (Dial. 28.2) whereas uitia in general terms (I.O. 1.2.2) an d likewise corrupti mores (1.2.4) are censured by Quintilian.

⁵² In this area the most significant likeness between Tacitus' Messala and Quintilian is discovered in the fault named last in the text, namely ignorance on the part of teachers. Thus Messala at Dial. 28.2 inscientia praecipientium, Maternu s at 33.2 (a noticeable extension from ‘teacher’ to ‘us’) nostrae desidiae et inscientiae aduersus acerrima…studia; Quint. I.O. 2.10.3 ut inter praecipuas quas corrumperent eloquentiam causas licentia atque inscitia declamantium fuerit, with my remarks, above p. 477, on likely provenance from the De causis. The likeness is not equally marked when it comes to Dial. 28.2 desidia inuentutis and neglegentia parentum. Contrast l.O. 1.2.5, 1.3. A teacher of Quintilian's stamp is unlikely to talk about his disciples an d their parents in Messala's terms.

⁵³ Quintilian and Tacitus concur with Cicero himself in regarding the subjects they advise the orator to study, philosophy in particular, not as so many sciences, but as elements of a general civic culture. Thus Quint, . I.O. 12 ch. 2Google Scholar, especially §§ 6–7 hinc etiam illudest quod Cicero pluribus et libris et epistulis testatur, dicendi facultatem ex intimis sapientiae fontibus fluere… quapropter haec exhortatio mea non eo pertinet ut esse oratorem philosophum uelim… (7)… atqui ilium quern instituo Romanum quendam uelim esse sapientem, qui non secretis disputationibus sed rerum experimentis atque operibus uere ciuilem uirum exhibeat. Tac, . Dial. 31.7Google Scholar (Messala) neque enim sapientem informamus neque Stoicorum comitem sed eum qui quasda m artes haurire, omnes libare decet. Cicero, while assenting to the universal approach which necessitates ‘sampling’ of knowledge, offers a varied range of opinions in the De or.: contrast 1.21 (with A. D. Leeman's and H. Pinkster's comm. vol. 1 [1981], pp. 58–60), 1.218, 3.54, 76, al.

⁵⁴ For Cicero's encyclopaedic approach in Quintilian and Tacitus, see Appendix 4.

⁵⁵ Cornelia at Quint. I.O. 1.1.6, and in Messala's speech, Tac, . Dial. 28.5Google Scholar; Laelia and Hortensia at Quint, ibid., Aurelia and Atia at Tac. ibid. The juxtaposition of these chapters in Quintilian and Tacitus strikes me as one of the best features of A. Gudeman's introduction to his commentary, 2nd ed. (1914), pp. 93–6. In both these authors alike we find an articulate and markedly similar theory of nurture and early education. But, since Gudeman was not explicit on this point, it is worth noting that, unlike Tacitus, Quintilian enables us to divine the origin of the theory at issue. For, in spite of all his color Romanus (which is even more marked in Tacitus), the author, at I.O. 1.1.4, 3.14, cites a Greek discussion by the famous Stoic Chrysippus, most probably Περ⋯ πα⋯δων ⋯γωγ⋯Η. Gudeman too traced parallels in ps.-Plutarch's essay identically titled, and noted Quintilian's reference to the Stoic philosopher Diogenes of Babylon.

⁵⁶ Thus Tac. Dial. 30.1. Cf. Quint. I.O. 1.1.35 inter prima elementa, 1.2.1 exire de gremio.

⁵⁷ cf. Appendix 4 for Barwick's comparison of Quint. I.O. 1 chs. 4–10 with Tac. Dial. chs. 30–2 as regards the teaching of artes, and Pfeiffer for a wider view of Hellenistic grammatice.

⁵⁸ Quint. I.O. 1.4.5 (on literary teaching) quo minus suntferendi… qui hanc artem… cauillantur, quae nisi oratoris futuri fundamenta fideliter iecit, quidquid superstruxerit corruet, al. Tac, . Dial. 30.1Google Scholar (transeo prima dicentium elementa, in quibus et ipsis parum laboratur:) nee in auctoribus cognoscendis nee in euoluenda antiquitate nee in notitia uel rerum uel hominum uel temporum satis operae insumitur.

⁵⁹ Quint. I.O. 10.5.19 quare iuuenis qui rationem inueniendi eloquendique a praeceptoribus diligenter acceperit…, exercitationem quoque modicam sibi fuerit consecutus, oratorem sibi aliquem, quod apud maiores fieri solebat, deligat, quern sequatur, quern imiletur, etc. Tac, . Dial 34.1Google Scholarergo apu d maiores nostros iuuenis ille, qui foro et eloquentiae parabatur, imbutus iam domestica disciplina, refertus honestis studiis, deducebatur a patre uel a propinquis ad eum oratorem qui principem in ciuitate locum obtinebat, hunc sectari, hunc prosequi, etc. An occasional touch of colour differs characteristically, but nevertheless it is clear that the same advice is given in either case.

⁶⁰ Thus Quintilian frequently, but with especial vigour at I.O. 5.12.22, an extract (no. 3) from the De causis: igitur et ille quern instituimus adulescens componal se ad imitationem ueritatis…, iam in schola uictoriam speclet, etc. Cf. above pp. 476–7, and Barwick (1954), 11–13. Thus too Tac, . Dial. 31.1Google Scholarnon ut in rhetorum scholis declamarent nee ut fictis nee ullo modo ad ueritatem accedentibus conlrouersiis…uocem exercerent etc., cf. 35.4.

⁶¹ den Boer, W. (above n. 47), p. 209Google Scholar, argues on these lines.

⁶² Dial. 28.2–3, quae malaprimum in urbe nata, moxper Italiam fusa, iam in prouincias manant. quamquam uestra uobis notiora sunt; (3) ego de urbe et his propriis ac uernaculis uitiis loquar, quae etc.

⁶³ This, however, as Dessau, H., PIR 31Google Scholar, Messala 468, points out, does not necessarily make him a descendant of the renowned orator Messala Corvinus. Indeed one feels tempted to go a step further, and wonder whether the absence of any reference to this connexion is not likely to exclude such a guess. In a dialogue de oratoribus it would have been natural to say so, had it been the case.

⁶⁴ cf. Tac. Hist. 4.42.1.

⁶⁵ I refer especially to den Boer (above n. 47), pp. 209–10, Güngerich, (above n. 29), p. 440Google Scholar, K. Bringmann loc. cit. (above n. 47).

⁶⁶ See preceding note.

⁶⁷ Cicero im Wandel d. Jahrhunderte ³ (1912), p. 36. There is still muc h to be said about Cicero in the literature of the first imperial century. Valuable hints for such a discussion are found in vol. 28 of the Eritretiens Fond. Hardt, 1982; Éloquence el rhé'torique chez Cicéron, especially in M. Winterbottom's contribution, ‘Cicero an d the Silver Age’.

⁶⁸ Below n. 71 and p. 494.

⁶⁹ Th e contrast between Cicero's less moralistic approach to oratory an d Quintilian's moralism has been discussed by Classen, C. J., ‘Ciceros orator perfectus ein vir bonus dicendi peritusT in Commemoratio, Studi…R. Ribuoli (1986), 43–55Google Scholar.

⁷⁰ Perhaps the clearest, though not the only, indication of Tacitus' independence is the reference to the text of Cic, . Or. 12Google Scholar, at Dial. 32.6 in contrast with Quint. I.O. 12.2.23, on which see Giingerich's acute discussion, CP 46 (1951), 163–4Google Scholar.

⁷¹ Barwick (1929) (above n. 41), 85, rightly says of the (neo-) Ciceronian Bildungsideal that Messala represents it ‘unter Umbiegung der tatsachlichen Verhaltnisse, als das in alter Zeit allgemein gültige’, referring to Dial. 30.4, 31.1. Talking about the various aspects of Cicero's eloquence, Tacitus shows himself by no means unaware of the differences between his oratory and the oratory of his time. For he makes Aper say at 22.1 ad Ciceronem uenio, cui eadem pugna cum aequalibus/uif quae mihi uobiscum est. illi enim antiquos mirabantur, ipse suorum temporum eloquentiam anteponebat. nee ulla re magis eiusdem aetatis oratorum praecurrit quam iudicio etc. This of course differs from Bildungsideal, but any rhetorical work of Cicero's, and particularly the De oratore, illustrates a similar point. For the relation of Ciceronianism proper and Quintilian's neo-Ciceronianism, see above p. 489.

⁷² So especially Boer, den (above n. 47), p. 210Google Scholar, but also others. I argue above that optimism and pessimism so called are notions too crude to impinge on Quintilian's argument which is partly optimistic and partly pessimistic. Even less do they impinge on Tacitus' firm and perceptive modes of thinking. They do not affect Barwick's reasoning (1954) (above n. 46), pp. 17–18, who considered Tac. Dialogus as a whole directed precisely against Quintilian's ‘optimism’ in De causis and Institutio alike, but also, within the Dialogus itself, against the optimism of ‘Messala-Quintilian’. My own argument above is even less affected by them, as it does not operate with the notion ‘Messala-Quintilian’. Heldmann (1980) (above n. 48), and also 1982 passim, suggested that the doctrinaire classicism of Messala in the Dialogus excluded almost by definition the use of Quintilian's reforming classicism. I doubt nevertheless that the similarity of Quintilian's and Messala's educational doctrines can be set aside, whatever the application, optimistic or pessimistic, reforming or restrictive; cf. above n. 48.

⁷³ This difference between doctrines and their presentation applies to various controversies in Tacitean scholarship, including Giingerich's attempt, in Gn. (1955) (above nn. 29 and 47), 440–1, to remove the Messala section of the Dialogus from the proximity of Quintilian by stressing certain secondary, though not unimportant, differences. Such are the functions variously attributed to philosophy by these two writers on oratory. Again Messala advises the orator to make occasional use of Epicurus and the Epicureans (Dial. 31.6, Giingerich ibid.), another difference of that order, for the advice would have been unacceptable to Quintilian or Cicero. But neither disagreement would have rendered Quintilian's remarks on philosophy and rhetoric unusable for Tacitus' own purposes.

⁷⁴ The distinction between the subject of Quintilian's first work – corrupla eloquentia, decadent eloquence – and that of Tacitus' Dialogus – the decline of oratory – has been noted recently, as though it had not in fact been seen as early as Reuter (above n. 38), p. 57. Reuter did not allow that distinction to interfere with establishing a close connexion between these two aetiological works (as I have called them above) and I see no reason why that should be otherwise. Several scholars, however, have recently demurred for that reason; thus e.g. Kennedy (above n. 1), p. 23 and in general p. 138, and, within a larger argument on Gattungsgeschichle, Heldmann, (1980), pp. 8, 20Google Scholar, also (1982) (both above n. 48).

⁷⁵ For illustration we could cite much in Messala's speeches from Dial. 28 onwards. See especially 28.2 desidia iuuentutis…et inscientiapraecipientium, 30.1 nee…satis operae insumilur, 30.2 sed expetuntur quos rhetoras uocanl etc. referam necesse est animus ad earn disciplinam qua usus esse eos oralores accepimus quorum infinitus labor et cotidiana meditatio el in omni genere studiorum adsiduas exercitationes ipsorum continentur libris, 30.3 ref. to Cic. Brutus, 31, 32.1ipsa multarum artium scientia etc., 32.3 ignorance of contemporary diserti, 32.5–6, 35.

⁷⁶ The possibility of some critical reflections by Messala on diction and style, near the beginning of the lacuna in ch. 35, cannot be wholly excluded; see 35.5 ingentibus uerbis-a context that recalls, as Dr M. Winterbottom reminds me, the well-known criticism of declamation, not excluding style, at Petron. 1. Yet I doubt if that will have amounted to much since the text before the lacuna ends cum ad ueros iudices uentum, and thus moves away from the topic of style. For these words involve the antithesis to the scholastic situation of the declaimers much noticed by various authors, not least by Quintilian himself, e.g. I.O. 5.13.42ff., especially 45 in scholis permittendum semper, inforo rarum, ibid, ubi uera res agitur, et al. Cf. De causis, Extract 3 (from I.O. 5.12.22–3).

⁷⁷ Winterbottom, M., ‘Quintilian and Rhetoric’, in Empire and Aftermath, Silver Latin lied. Dorey, T. A. (1975), pp. 79–97Google Scholar, moved by some impressive examples of non-Ciceronian ethos in Quintilian's style, seems to be inclined largely to discount the notion of neo-Ciceronianism in first-century prose.

⁷⁸ Above n. 46.

⁷⁸ Above n. 46.

⁸⁰ See above nn. 47–8.

⁸¹ These points caused Kennedy and, more recently, Heldmann (above n. 48) to dissociate not only Messala's presentation but also the contents of his speech from Quintilian.

⁸² An equivocation of Ciceronian and late republican was detected by Barwick (1929), p. 85; cf. above n. 71. Tacitus, however, was quite aware of that; the emphasis with which M. Aper at Dial. 22.1 dissociates Cicero from his contemporaries has also been remarked at n. 71: ad Ciceronem uenio, cui eadem pugna cum aequalibus suis fuit quae mihi uobiscum est etc. The disputants hold fascinatingly diverse views of historical periods.

⁸³ The bibliography is too large to be recorded here, nor do I intend to go into the difficulties remaining in the layout of the set speeches and in the MSS. evidence. A good statement on some of the strong arguments against a set speech by Secundus on grounds of composition is made by U. Hass von Reitzenstein (1970) (Diss. Cologne), pp. 106–11 citing earlier literature. For the MSS. evidence see now Murgia, C. E., ‘The Length of the Lacuna in Tac. Dial.’, Cal. St. in Cl. ant. 12 (1981), 221–40Google Scholar, Haussler, R., Philol. 153 (1986), 69 ffGoogle Scholar, especially pp. 73–7. They agree that the lacuna must be small, but are at variance about the remaining aspects of the problem.

⁸⁴ For speculations on the author's motive, see Hass von Reitzenstein (preceding n.), pp. 142–3, and, with a different approach, W. Deuse, ‘Zur advocatus-diaboli Funktion Apers im Dialogus und zur Methode ihrer Deutung’, Grazer Beitrage 3 (1975), 51–68Google Scholar.

⁸⁵ Diverse certainly their ways were. Yet it was the same imperial protector (Vespasian) of the, idealistic rhetorician Quintilian, and of the young orator Tacitus, whom the Tacitean Aper extols / / (at 8.3) as uenerabilis senex et patientissimus ueri relying on the arts of Eprius and Vibius. The realistic and amoral Aper is the only disputant in the Dialogus who defends a full modern role for oratory. He allows that neither Eprius nor Vibius was moribus egregius, and yet lauds them as follows: per multos iam annos potentissimi sunt ciuitatis ac, donee libuit, principes fori, nunc principes in Caesaris amicitia agunt feruntque cuncta atque ab ipso principe cum quadam reuerentia diliguntur, etc. Tacitus' outspokenness concerning modern oratory seems to put a large question-mark against the brief aside of Kennedy, G. (above n. 1), p. 138Google Scholar, that the Dialogus ‘does not touch upon the most idealistic aspects of Quintilian's orator as an advisor to states and emperors'.

⁸⁶ Eprius, , Dial. 5.7, 8.1Google Scholar (cf. RE 6, 261–4, PIR 3Z, 82–4); Vibius, , Dial. 8.1Google Scholar (cf. RE 8 A2, 1968–70, PIR 3², 420–1; Bosworth, A. B., Ath. 51 (1973), 72–4Google Scholar.

⁸⁷ As regards Marcellus Eprius, the tone of Hist. 4.42–3 amounts to stern censure; such [touches as 43.2 Marcellus minacibus oculis cannot escape notice, and are not meant to do so. One lean only regret that the Tacitean summary in the Histories of Eprius' career is lost together with fthe reign of Titus and Eprius' demise. As regards Vibius, the celebrated epigram, Hist. 2, 10.1, leaves no doubt: pecunia potentia ingenio inter claros magis quam inter bonos, cf. para. 4 in the same chapter.

⁸⁸ The kind of amoral oratory implied in Aper's praises of Eprius and Vibius remains by no means unchallenged in the Dialogus. Maternus makes no secret of harsh general disapproval: 2.2 nam lucrosae huius et sanguinantis eloquentiae usus recens et ex malis moribus natus atque, ut tu dicebas, Aper, in locum teli repertus. But Tacitus allows him even personal censure of Aper's admired delatores: 13.4 nam Crispus iste et Marcellus, ad quorum exempla me uocas, quid habenl in hac sua for tuna concupiscendum ? quod timent, an quod limenlur? etc.

⁸⁹ M. Winterbottom (1964) (above n. 16), p. 94.

⁹⁰ Above nn. 41, 83.

⁹¹ (1) Dial. 5.3–10.8, 11.1–13.6; (2) 16.4–23.6, 25.1–26.8; (3) 28.1–32.7 and 33.4–35.5 (plus part of lacuna), 36.1 (lacuna)–41.5.

⁹² Barwick (1954) (above n. 41), p. 17 ‘In dem Gegensatz Messala-Quintilian auf der einen und Maternus-Tacitus auf der anderen Seite fassen wir eins der Hauptmotive, die Tacitus zur Abfassung seines Dialogs veranlasst haben.’ ‘Maternus-Tacitus’ again p. 18, and implied in the rest of the paper.

⁹³ A perceptive remark by R. Häussler may be compared with this notion. The remark appeared in 'De causis corruptae eloquentiae. Variationen eines röm. Themas’, Actes de la XII^e Conférence internat. d'étudesclassiques. Eirene (1972), Hakkert (Amsterdam, 1975), p. 314: ‘Die taciteische Originalität liegt also nicht nur in der matemischen Erkenntnis, sondern in der Form des Ganzen.’

⁹⁴ Wilamowitz (1900), Kl. Schr. iii.246, Wilamowitz (1932), repr. 1955, ii.538 (above n. 39), denied that Tacitus, if judged by his own style in the large historical works, could have regarded the modern oratory for which Aper makes a case as corrupta eloquentia. Aper represents an aspect of Tacitus’ personality according to Keyssner (above n. 46), 113, Syme, , Tacitus i. 109Google Scholar, Maternus another, conflicting, aspect, as again Keyssner argued.

⁹⁵ Messala's final rejoinder, 42.1, might here be noted.

⁹⁶ The writer gratefully acknowledges some acute and helpful comments by Dr M. Winterbottom on an earlier draft of this paper.