¹ Barnes, T.D., ‘The significance of Tacitus' Dialogus de Oratoribus’, HSPh 90 (1986), 225–44Google Scholar, at 225: ‘the most problematical of Tacitus' works’; Köves-Zulauf, T., ‘Reden und Schweigen im taciteischen Dialogus’, RhM 135 (1992), 316–41Google Scholar, at 341 speaks of its fundamentally puzzling nature: ‘originäre Rätselhaftigkeit der Schrift’.

² Luce, T.J., ‘Reading and response in the Dialogus’, in id. and Woodman, A.J. (edd.), Tacitus and the Tacitean Tradition (Princeton, 1993), 11–38CrossRefGoogle Scholar and Mayer, R., Tacitus: Dialogus De Oratoribus (Cambridge, 2001)Google Scholar, for example, represent the nearly universal assumption that the dialogue assumes the decline of oratory. Contra: Goldberg, S., ‘Appreciating Aper: the defence of modernity in Tacitus' Dialogus de Oratoribus’, CQ 49 (1999), 224–37CrossRefGoogle Scholar. Brink, C.O., ‘Can Tacitus' Dialogus be dated? Evidence and historical conclusions’, HSPh 96 (1994), 251–80Google Scholar, at 276–7 remarks that the dialogue ‘pursues no thesis, single or composite, nor does it answer a particular question, not even very fully the question posed at the outset’. Fantham, E., The Roman World of Cicero's De Oratore (Oxford, 2004)CrossRefGoogle Scholar, at 321: ‘it is far from obvious that Tacitus identifies with any of the points of view expressed by his interlocutors’. Gowing, A., Empire and Memory: the Representation of the Roman Republic in Imperial Culture (Cambridge, 2005)CrossRefGoogle Scholar, 109. Already Hirzel, R., Der Dialog. Ein literarhistorischer Versuch (Leipzig, 1895)Google Scholar, considered it a trait of the genre to expand the ostensible theme beyond that announced in the preface. Levene, D.S., ‘Tacitus' Dialogus as literary history’, TAPhA 134 (2004), 157–200Google Scholar, fruitfully sidesteps the question of decline by focussing on continuities in the picture of literary history that develop in the course of the dialogue.

³ ‘Pro-Aper’: Champion, C., ‘Dialogus 5.3–10.8: a reconsideration of the character of Marcus Aper’, Phoenix 48 (1994), 152–63CrossRefGoogle Scholar and Goldberg (n. 2), both with details of the ‘anti-Aper’ bibliography. Mayer (n. 2) and Strunk, T.E., ‘Offending the powerful: Tacitus' Dialogus de Oratoribus and safe criticism’, Mnemosyne 63 (2010), 241–67CrossRefGoogle Scholar, bring on board much from the anti-Aper line.

⁴ Some perceived attitudes from the later works, the Histories and Annals, have been projected back onto the Dialogus; e.g. Mayer (n. 2), 7–8 sees the Dialogus as the first flower of disenchantment – only later does the full force of Tacitus' real opinions as a historian make itself felt. Cf. the initial framing by Syme, R., Tacitus (Oxford, 1958)Google Scholar, 465. On the complexities of Tacitean biography see also Goldberg, S., ‘The faces of eloquence: the Dialogus de Oratoribus’, in Woodman, A.J. (ed.), The Cambridge Companion to Tacitus (Cambridge, 2009), 73–84Google Scholar, at 78–80, who points out the extent to which Tacitus' oratorical renown continued well into the 100s c.e.

⁵ Gallia, A., ‘Potentes and potentia in Tacitus's Dialogus de oratoribus’, TAPhA 139 (2009), 169–206Google Scholar, sees Maternus' attacks as part of a larger tendency against the delatores (‘prosecutors’). Rutledge, S., Imperial Inquisitions: Prosecutors and Informants from Tiberius to Domitian (London and New York, 2001)Google Scholar places the actions of the delatores and the prejudices against them in the broader social and political contexts of Rome in the Early Empire.

⁶ Winterbottom, M., ‘Returning to Tacitus' Dialogus’, in Wooten, C.W. (ed.), The Orator in Action and Theory in Greece and Rome (Boston, 2001)CrossRefGoogle Scholar, 148, summarizing Luce (n. 2), 21. Cf. Hirzel (n. 2), 2.56 n. 1.

⁷ Luce (n. 2), 18.

⁸ The bibliography is considerable. I select the following examples because they not only grapple with the themes of the work but also with its design and literary devices: Luce (n. 2); Bartsch, S., Actors in the Audience: Theatricality and Doublespeak from Nero to Hadrian (Cambridge, MA, 1994)CrossRefGoogle Scholar; Goldberg (n. 2); Levene (n. 2).

⁹ Luce (n. 2) and Lier, H., ‘Rede und Redekunst im Diskurs. Tacitus' Dialogus de oratoribus als Schullektüre’, AU 39 (1996), 52–64Google Scholar emphasize the declamatory elements and the speakers' skills. Cf. Mayer (n. 2), 40 n. 96. Allison, J.W., ‘Tacitus' Dialogus and Plato's Symposium’, Hermes 127 (1999), 479–92Google Scholar, at 482 dubs the Dialogus a ‘literary declamatio’. Maternus, drawing on the language of the bar, remarks that he regularly defends poetry against Aper: ego, cui desidiam advocationum obicis, cotidianum hoc patrocinium defendendae adversus te poeticae exerceo (‘You criticize me for neglecting court duties and here I am defending the art of poetry against you on a daily basis’, 4.1). On arrival Messalla (rightly) suspects a private session of a practice case: interveni secretum consilium et causae alicuius meditationem tractantibus? (14.1). Experience has taught him what to expect.

¹⁰ Winterbottom (n. 6), 151, for example, prefers a philological solution to the problem of ‘late’ versus ‘early’ Maternus, that is, to the apparently contradictory attitudes in Maternus' two speeches. Bartsch (n. 8), 113 and 255 n. 38 and Strunk (n. 3), 247, also reject Luce's rhetorical explanation.

¹¹ Luce (n. 2), 37.

¹² Viewing the speeches as unconnected set pieces breaks up the dialogue into discrete units that need not have any relevance to one another. Luce (n. 2) did not explicitly make this last point – his argument for a composite message would preclude it in some sense – but this is essentially the logic of Lier (n. 9), the only other article emphasizing the declamatory nature of the speeches. Both Winterbottom (n. 6) and Strunk (n. 3), 247 take issue with this implicit conclusion.

¹³ Especially Goldberg (n. 2), Winterbottom (n. 6) and Levene (n. 2). The first two essays note correspondences among the second speeches of both Aper and Maternus; additional parallels deserve to be examined in greater depth. Levene (n. 2) details how the picture of literary history in the Dialogus evolves over the course of the six speeches and how this relates to the work's (failed) attempts to slot itself into the picture of literary history outlined in its speeches. Significantly, Goldberg argues against the thesis of decline, while Levene (n. 2) foregoes the question entirely (as noted above), which frees him to explore a single subject across all the interlocutors' statements. It would appear that a major hindrace to reading the dialogue as a coherent piece is the assumption that Tacitus sought to demonstrate decline.

¹⁴ Murgia, C., ‘The date of Tacitus' Dialogus’, HSPh 84 (1980), 99–125Google Scholar, at 111.

¹⁵ Brink, C.O., ‘History in the “Dialogus de Oratoribus” and Tacitus the historian: a new approach to an old source’, Hermes 121 (1993), 335–49Google Scholar, at 338. How to define and assess unity is a problem for the study of many ancient dialogues, most famously for Plato's Phaedrus, on which see, most recently, Werner, D., Myth and Philosophy in Plato's Phaedrus (Cambridge, 2012), 236–58CrossRefGoogle Scholar with bibliography and discussion of different modern notions of unity. Levene (n. 2), 196–7 also discusses the issue for the Dialogus and, briefly, ancient approaches to unity.

¹⁶ The two main exceptions to the tendency to focus on the speeches are: 1) the opening question, though it is rarely ‘interpreted’ but rather taken as a face-value statement that declares the subject of the dialogue: what has caused decline? (though see Goldberg [n. 2] on complexities in the preface); 2) the offence caused by Maternus' poetry (which is typically compared to the attitudes he seems to express and to the titles of his plays).

¹⁷ Octavius 39: cum Octavius perorasset, aliquamdiu nos ad silentium stupefacti intentos vultus tenebamus, et quod ad me est, magnitudine admirationis evanui, quod ea, quae facilius est sentire quam dicere, et argumentis et exemplis et lectionum auctoritatibus adornasset et quod malevolos isdem illis, quibus armantur, philosophorum telis rettudisset, ostendisset etiam veritatem non tantummodo facilem sed et favorabilem (‘After Octavius finished his speech we were amazed to the point of silence and stood with fixed gaze, and in my case, I was overcome with immense wonder at the fact that Octavius had adorned those things, which it is easier to feel than to say, with arguments, examples, and literary precedents, and that he struck back at the malevolent with those same weapons of the philosophers with which they are armed, and he demonstrated even a truth that is not only comprehensible but also attractive’). Cf. Macrob. Sat. 6.1.4: Rufius Albinus cites Afranius' borrowings from Menander in order to defend Virgil's imitatio, discussed by Kaster, R.A., ‘Macrobius and Servius: verecundia and the grammarian's function’, HSPh 84 (1980), 219–62Google Scholar, at 232.

¹⁸ Merit: Syme (n. 4), 526 n. 9, though with a broader view of Tacitus' oeuvre. Framework: O'Hara, J., Inconsistency in Roman Epic: Studies in Catullus, Lucretius, Vergil, Ovid and Lucan (Cambridge, 2007)CrossRefGoogle Scholar, 5, focussing on poetry. On intratextuality see Sharrock, A. and Morales, H. (edd.), Intratextuality: Greek and Roman Textual Relations (Oxford, 2000).Google Scholar

¹⁹ S. Rutledge, ‘The literary, cultural, and historical background of Tacitus' Dialogus de Oratoribus' (Diss., Brown University, 1996), 67 defines ‘apprenticing’ (the tirocinium fori): ‘the practice whereby a young man, upon assuming the toga virilis, would attach himself to a well-known orator, and attend the cases and meetings at which he would speak in order to absorb better a knowledge of law, speaking style, and whatever else might be learned from daily attendance at court. The practice flourished in Cicero's day and beyond.’ See also Leeman, A.D., Pinkster, H. and Wisse, J., M. Tullius Cicero: De Oratore Libri III (Heidelberg, 1981–2008), 1.23–5Google Scholar; Hall, J., ‘Social evasion and aristocratic manners in Cicero's De oratore’, AJPh 117 (1996), 95–120Google Scholar, at 100. Cicero was also a role model for others, cf. Fam. 1.9.23, 9.16.7; Off. 1.1; and of course Cicero memorably notes that Caelius' father had entrusted Caelius to Cicero for training (Cael. 9). Quint. Inst. 10.5.19 attests to the endurance of the tirocinium in his praise of it.

²⁰ Although I cannot find much meaning to the specific differences between the two passages, it is hard not to read at least some playful, possibly useful, irony in these two phrases: princepem … locum is ascribed to the Republic, while ingenia is pluralized in the description of contemporary circumstances, a point in contrast with the preface's suggestion that past ages, in contradistinction to the present, flourished in ingenia.

²¹ Mayer (n. 2), 93 notes ‘fabulas “conversations” (OLD 1, but the word is not found with this sense before the first century ad)’. Bennett, C.E., Tacitus: Dialogus de Oratoribus (Boston, 1894)Google Scholar, 62 lucidly explains altercatio: ‘the altercatio was opposed to the continua or perpetua oratio; the former consisted of questions and answers, the latter allowed no interruptions. An altercatio might naturally become so heated in character as to degenerate into wrangling (iurgium) which was doubtless often the case.’ On altercatio cf. Cic. Brut. 44 and 164, Quint. Inst. 6.3.4 and 4.1.28.

²² To my knowledge the details of the intratext have yet to be examined in the scholarship. Lier (n. 9), 53 and Rutledge (n. 19), 68–9 come closest to noting the parallels. To remark on the similar content of the passages has become a commonplace in the scholarship, typically as a disqualification of Messalla's arguments. Lier sees it as an indiciation of Tacitus' intellectual affinity with Messalla (‘geistige Nähe zu Messalla’, 53). The contradiction is further sharpened by Aper's claim that young men (iuvenes) seek to improve themselves by following about practised orators (oratores sectantur, Dial. 20.4). On sectari the tiroconium see further Pliny's report of Quintilian's statement that he would follow about Domitius Afer: adsectabar Domitium Afrum (Ep. 2.14.10). Cf. Quint. Inst. 8.5.21; Plin. Ep. 4.16.

²³ Ep. 4.13.10: ex copia studiosorum, quae ad te ex admiratione ingeni tui convenit (‘from among a group of eager followers, who flock to you out of regard for your talent’, 104/5 c.e.). Cf. Pliny's (long) letter to Tacitus on brevitas (Ep. 1.20), which emphasizes Tacitus' authority on questions of oratorical style. An earlier date for the writing of the work (usually placed at or around 102 c.e., the suffect consulship of the addressee, Fabius Iustus; see Mayer [n. 2], 22–7 for summary) would not preclude this self-memorializing function. On dating see also Murgia (n. 14) and Brink (n. 2). It may have been a happy coincidence that Tacitus achieved the following he seems to have cultivated in the Dialogus. R. Saller, ‘Status and patronage’, CAH ² vol. 11 (2000), 817–54, at 848 remarks on the general situation under the emperors: ‘in the absence of any system of credentials, aspiring young advocates continued to rely on a senior figure to introduce them into the courts through a sort of apprenticeship’. The (later) role of Q. Mucius Scaevola ‘the Augur’ as Cicero's mentor should also not be overlooked.

²⁴ Cf. Cic. Orat. 41–2, whose translation of Plato demonstrates close knowledge of the Phaedrus. Cicero does indulge in a bit of indirect self-praise by lauding Hortensius, since a reader would be likely to compare Cicero's achievements to those of his rival.

²⁵ Ahl, F., ‘Homer, Vergil, and complex narrative structures in Latin epic: an essay’, ICS 14 (1989), 1–31Google Scholar outlines a model of rhetorical indirection and the pragmatic function of speeches in epic, demonstrating how speakers strategically aim statements towards various audiences. A speech's function is reconstructed from the context of the speaker's situation, which can generate additional meanings. These meanings, however, are neither coincident with – nor necessarily contrary to – the literal utterance.

²⁶ For this mode of reading, Hinds, S., Allusion and Intertext: Dynamics of Appropriation in Roman Poetry (Cambridge, 1998), 34–47Google Scholar invaluably discusses how topoi have considerable consequences for texts, despite their apparent resistance to analysis: ‘[t]he so-called commonplace, despite our name for it, is not an inert category … but an active one, with as much potential to draw poet and reader into, as away from, engagement with the specificities of its history’ (40).

²⁷ Fantham (n. 2), 321. Cf. Luce (n. 2), 20–1; Levene (n. 2), 195.

²⁸ A pithy version of the idea is captured by Cato's famous vir bonus dicendi peritus (‘good man skilled in speaking’), which Quintilian (Inst. 12.1.1) adopted as a slogan for the true orator. Winterbottom, M., ‘Quintilian and the vir bonus’, JRS 54 (1964), 90–7Google Scholar, discusses Quintilian's moral conception of the orator, but should be read with Goldberg (n. 2) and Rutledge (n. 5).

²⁹ On the background of the censorial edict, see Kennedy, G.A., The Art of Rhetoric in the Roman World 300 BC–AD 300 (Princeton, 1972), 90–5Google Scholar; Gruen, E.S., Studies in Greek Culture and Roman Policy (Leiden and New York, 1990), 179–92Google Scholar; Leeman, Pinkster and Wisse (n. 19), 4.304–6; Narducci, E., Cicerone e l'eloquenza romana: Retorica e progetto culturale (Rome, 1997), 25–9Google Scholar; Habinek, T., The Politics of Latin Literature: Writing, Identity, and Empire in Ancient Rome (Princeton, 1998), 60–1Google Scholar; Mayer (n. 2), 196; for the primary material, Kaster, R.A., Suetonius. De Grammaticis et Rhetoribus (Oxford, 1995), 273–4Google Scholar and 291–4 with Cic. De or. 3.93–5; Suet. Gram. et rhet. 1; Gell. NA 15.11.2. Messalla's odd claim is further discussed below.

³⁰ J. Murphy, ‘Tacitus on the education of the orator’, ANRW 2.33.3 (1991), 2284–97, at 2291: ‘[d]espite Messalla's claim, it is not clear … how the Roman youth was to exercise his oratorical skills if not by way of declamation or something like declamation’.

³¹ On the development of declamation, see Bonner, S.F., Roman Declamation (Liverpool, 1949), 27–31Google Scholar; Fairweather, J., Seneca the Elder (Cambridge, 1981), 124–31CrossRefGoogle Scholar. Winterbottom (n. 6), 149: ‘The rhetores were influential in the first half of the first century b.c., just as they were under Vespasian’.

³² Hirtius and Dolabella: Fam. 9.16.7 and 9.18.1 (46 b.c.e.) with Quint. Inst. 8.3.54 and 12.11.6. Hirtius and Pansa (coss. 43 b.c.e.): Att. 14.11.2, 14.12.2, 14.20.4, 14.22.1 (44 b.c.e.); Quint. Inst. 12.11.6; grandes praetextatos: Sen. Controv. 1 praef. 11; Suet. Gram. et rhet. 1.

³³ Bonner (n. 31), 22–5 details many of the topics covered by these exercises, noting their connections to Roman law and their similarities to themes found in Seneca the Elder. Quintilian would devote a lengthy chapter (Inst. 2.4) to the progymnasmata.

³⁴ The language includes what would eventually become known as declamation: in exercitationibus commentationibusque multum operae solitum esse ponere (‘He would typically exert much effort on practice speeches and compositions’, Cic. Brut. 105). The same language resurfaces in Cicero's narration of his own youthful training: (commentans oratoriis … exercitationibus, Brut. 305) and (commentabar declamitans, Brut. 310).

³⁵ Quintilian also adamantly supported declamation as an exercise, e.g. at Inst. 2.10. There is a further possible irony in Messalla's claims, since Quintilian says (2.4.42) that he discussed the inveterate origins of declamation in another writing, often thought to be the treatise (now lost) De causis corruptae eloquentiae. See C.O. Brink, ‘Quintilian's De causis corruptae eloquentiae and Tacitus' Dialogus de oratoribus', CQ 39 (1989), 472–503.

³⁶ Tacitus has Messalla state that they ordered the schools closed (cludere … iussi sunt). Crassus uses the verb sustuleram. Neither adequately reflects the fact that censorial disapproval was non-binding. It would appear that Messalla's (Tacitus'?) rendering of the historical facts is drawn from Crassus' statements in De oratore. As an anonymous reader points out, Messalla's failure to distinguish the Greek and Latin rhetores may ultimately reflect the ascendency of the rhetores Latini at Rome. Understandably, Quint. Inst. 2.4.42 glosses over Crassus' rejection of the Latin orators by merely noting that the schools began towards the end of Crassus' life and by pointing out the renown of L. Plotius Gallus.

³⁷ Scholarship on the censors' motivations long held that members of the elite objected to instruction in Latin because it might popularize rhetorical instruction by making it available to a larger segment of Rome's (non-Greek speaking?) population. Gruen (n. 29), 190–2 sees the edicts rather as an affirmation of ‘the Hellenic character of Roman higher education’, at 190. By Tacitus' time the distinction was no longer pedagogically useful: Greek and Latin rhetorical education went hand in hand.

³⁸ The emphasis on ingenia also comes up in Cicero's early letter to Titinius (Suet. Gram. et rhet. 26.1 = Cic. Ep. fr. 1 Watt).

³⁹ Through Crassus Cicero indulges in some retrospective self-memorialization, since he was foremost among the figures of the Late Republic to naturalize Greek rhetoric (and philosophy) in Latin. Leeman, Pinkster and Wisse (n. 19), 4.313 emphasize especially Cicero's philosophical encyclopedia, but it is hard to see how the immediate emphasis is not also on the Latinizing of rhetorical material (the prudentia which reunites the fields of rhetoric and philosophy).

⁴⁰ Hirzel (n. 2), 2.58–9 compares the passage to metacritical remarks on writing and composition in the Phaedrus. Cf. Cic. Brut. 231.

⁴¹ Despite immense recent interest in declamation and a virtual volte-face in opinion about this rhetorical form (cf. e.g. Gunderson, E., Declamation, Paternity, and Roman Identity: Authority and the Rhetorical Self [Cambridge, 2003])CrossRefGoogle Scholar, Messalla's statements about it have yet to see extensive treatment in the scholarship on the Dialogus. Hömke, N., Gesetzt den Fall: Ein Geist Erscheint (Heidelberg, 2002), 60–73Google Scholar outlines the attitudes towards declamation and rhetoric in the Dialogus. At the very least, it should be clear from this essay that scholars should be chary of citing Tacitus as an ancient authority against declamation (even if Messalla echoes common criticisms of declamation).

⁴² Bonner (n. 31), 80–2 details the copious examples.

⁴³ Sen. Controv. 9 pr. 3; cf. Quint. Inst. 10.5.18. Cassius Severus, a watershed in oratorical development according to Aper (Dial. 19.1), had nearly the opposite problem: best in person, he did not translate well to the page.

⁴⁴ Again, Bonner (n. 31), 72–4 collects the material.

⁴⁵ A trace of this idea has left itself in Tacitus' remarks on Helvidius Priscus' defeat at the hands of Eprius Marcellus; the Stoic philosopher was no match for the skilled orator, and Tacitus may have incorporated this piece of information as a nod to Cicero's remarks about Publius Rutilius Rufus for his refusal (inability?) to defend himself adequately against charges of extortion. Cf. De or. 1.227–30 and Brut. 115. On Rutilius see Kallet-Marx, R., ‘The trial of P. Rutilius Rufus’, Phoenix 44 (1990), 122–39.CrossRefGoogle Scholar

⁴⁶ Brink (n. 35), 476 summarizes: ‘let declamation be an exercise and not an end in itself’. Quintilian makes declamation identical with forensic activity provided that declamation advances the orator's real-world skills: nihil ergo inter forense genus dicendi atque hoc declamatorium intererit? si profectus gratia dicimus, nihil (Inst. 2.10.9). He provides an excellent example: the ‘disinherited son’ motif from declamation could be fruitfully applied to inheritance disputes in the centumviral courts (Inst. 7.4.11 and 7.4.20). Pliny is a rare voice in suggesting at one point that declamation may be ‘better than the real thing’, because speakers also acquire bad habits in the forum (Ep. 2.3.5).

⁴⁷ Similar dynamics are found in Petronius' Satyrica, in which Encolpius criticizes declamation while himself declaiming (declamare, Petron. Sat. 3). See Kennedy, G.A., ‘Encolpius and Agamemnon in Petronius’, AJPh 99 (1978), 171–8Google Scholar, Gunderson (n. 40), 9–12, and Schmeling, G., A Commentary on the Satyrica of Petronius (Oxford, 2011)Google Scholar. Although the point cannot be elaborated in this article, criticism of declamation appears to be a central topos of declamation. The reading presented here also addresses a hitherto unissued objection to Luce's and Lier's claim that the speeches are like declamations, specifically, how to reconcile Messalla's disapproval of declamation with Tacitus' incorporation of declamatory elements into the dialogue's speeches.

⁴⁸ Murphy (n. 30), 2290: ‘Tacitus, through Messalla, calls for a similarly broad education of the orator in his day as well’. The present analysis demonstrates what allows us to make this inference.

⁴⁹ Brink (n. 15) and (n. 35) partly aligns Tacitus with Messalla, but reads the work as a rejection of Quintilian's neo-Ciceronianism. Brink (n. 35) is still the most accessible study of the scanty material on De causis, but I remain fundamentally sceptical towards a major assumption which allows him to connect both works, namely, that both Tacitus in the Dialogus and Quintilian in De causis seek to offer a historical explanation for the decline of oratory in the imperial era.

⁵⁰ Messalla complained that students speak and are heard (et dicant et audiantur) with equal negligence. The second verb emphasizes the need for an attentive and critical audience.

⁵¹ Securitas/securus occurs in a wide range of applications at 3.2, 5.2, 5.5, 10.7, 11.4, 13.1, 35.3 and 37.8, including political and pedagogical contexts. The etymological nuance of ‘free from care’ (se-cura) can be read in Secundus' remarkable suggestion (pun?) that Maternus render his Cato safer (securiorem) through greater attention (diligentius retractares 3.2). Maternus responds in kind by referring to the play as a cura (3.3), ‘the trouble taken over a literary work’; see Mayer (n. 2), 96.

⁵² The triad ars, exercitatio and ingenium were fundamental to rhetorical teachings, going back at least to Protagoras and Plato's Phaedrus (269d). See Leeman, Pinkster and Wisse (n. 19), 1.209–11 and Mayer (n. 2), 94 for discussion and bibliography.

⁵³ Hall (n. 19) on this aspect of De oratore.

⁵⁴ Compare Aper's resounding praise of the other interlocutors at 23.6.

⁵⁵ Crassus non tam existimari vellet non didicisse, quam illa despicere (‘Crassus did not so much wish to be judged not to have studied so much as to scorn these things’, De or. 2.4); Antonius autem probabiliorem hoc populo orationem fore censebat suam, si omnino didicisse numquam putaretur (‘Antonius, however, would think that his speaking would be more believable with the people, if he were thought never to have studied at all’, De or. 2.4). I examine the parallels at greater length in C.S. van den Berg, The World of Tacitus' Dialogus de Oratoribus (Cambridge, forthcoming).

⁵⁶ Thus, we no longer need to think that Tacitus criticizes Quintilian's neo-Ciceronian programme through Messalla. The relationship to Quintilian must be revised entirely.

⁵⁷ Compare Antonius' repeated claim that he had never seen someone who was truly eloquens (e.g. De or. 3.54), though he eventually accords the title to Crassus: inveni iam, quem negaram in eo, quem scripsi, libello me invenisse eloquentem (‘I've finally found an eloquent man, which in the book that I wrote, I had said I'd never come across’, 3.189). In the first preface to De oratore Cicero wonders at the paucity of orators: quaerendum esse visum est quid esset cur plures in omnibus rebus quam in dicendo admirabiles exstitissent (‘it seemed necessary to ask why more admirable men have existed in all things but in speaking’, 1.6).

⁵⁸ Luce (n. 2), 37.

⁵⁹ For example, the ‘bad evidence’ of the first debate (noted in §I), Aper's delatores and Messalla's mythical models, could also be read as a means to add nuance to their arguments. These ‘internal’ concessions prove more valuable when regarded as a supplement to argument. They add limits to the positions advanced (not all oratory is beneficial, nor all poetry unqualified quietism).