² For a study of the significance of the late panegyrics against their distinctive political and religious background see Béranger, J., ‘L'Expression de la divinité dans les panégyriques latins’, MH 27 (1970) 242–54Google Scholar; MacCormack, S., ‘Latin prose panegyrics: tradition and discontinuity in the later Roman Empire’, Revue des études augustiniennes 22 (1976) 29–77CrossRefGoogle Scholar; Rodgers, B.S., ‘Divine insinuation in the Panegyrici latini’, Historia 35 (1986) 69–104Google Scholar; also more generally MacCormack, S., Art and ceremony in late antiquity (1981)Google Scholar.

³ For the ideological difference made by the Greek language see Price, S.R.F., ‘Gods and emperors: the Greek language of the Roman imperial cult’, JHS 104 (1984) 79–95, esp. 83–5CrossRefGoogle Scholar; for differences of technique between Greek and Latin panegyrics more generally, cf. Russell, D.A. and Wilson, N.G., Menander Rhetor (1981) xviiGoogle Scholar.

⁴ Clearly speeches of praise can have other objects: not only the dead, but also gods, places, animals or even abstractions; see e.g. Aristotle, , Rhetoric 1366a28–32Google Scholar; Quintilian 3.7.6; Menander Rhetor (I) 331–2, with Russell and Wilson (n.3) ad loc.

⁵ Some in the ancient world distinguished even the praise of the dead from that of the living; but they admittedly do not mention the specific context of religion, and hence regard the practical difference as relatively small; see e.g. Plato, , Laws 801e–2aGoogle Scholar (on which see Buchheit, V., Untersuchungen zur Theorie des Genos epideiktikon von Gorgias bis Aristoteles (1960) 87–8)Google Scholar; Theon, Progymnasmata 109.22–6 (Spengel).

⁶ See e.g. Rhet. Her. 3.15; Cicero, De oratore 2.349; Quintilian 2.1.11,3.4.11,3.7.2; ps.-Dionysius, Rhetoric 347.14–16 (Usener-Radermacher).

⁷ E.g. Quintilian 3.4.14; Pliny, Panegyric 4.1, Letters 3.18.2; Seneca, Letters 94.39; ps.-Dionysius, Rhetoric 347.1–11 (Usener-Radermacher); cf. Aristotle, , Rhetoric 1367b37–8a9Google Scholar. See Pernot, L., La Rhétorique de l'éloge dans le monde gréco-romain (1993) 72–3, 710–24Google Scholar; also Griffin, M., Seneca: a philosopher in politics ed. 2 (1991) 137–8Google Scholar; Moles, J.L., ‘The Kingship Orations of Dio Chrysostom’, PLLS 6 (1990) 297–375 (at 303–5)Google Scholar.

⁸ E.g. Demetrius, , On style 292, 295Google Scholar; Pliny, Panegyric 3.4; Philostratus, , Lives of the sophists 493–4Google Scholar; ps.–Dionysius, , Rhetoric 295–358Google Scholar (Usener-Radermacher). On the theory of ‘figured speech’ that underlies this see Ahl, F., ‘The art of safe criticism in Greece and Rome’, AJP 105 (1984) 174–208Google Scholar; Bartsch, S., Actors in the audience (1994) 65–71, 93–7CrossRefGoogle Scholar. Quintilian 9.2.65–95 discusses ‘figured speech’ without mentioning panegyrical contexts, only declamation and forensic speeches; but this, as Ahl (190) and Bartsch (95–6) suggest, may well be his own ‘figured’ way of avoiding offending Domitian, by recommending the device without mentioning the precise context in which his readers would feel it most appropriate.

⁹ For this basic distinction, see e.g. Aristotle, Rhetoric 1358b2–8, where the listener to an epideictic speech is said to be a θεωϱός, the listener to a forensic or a deliberative speech is a ϰϱιτής – though note also 1391b15, where even the former is said to be a ϰϱιτής after a fashion. Note also Quintilian 3.4.13–14, who objects to the name ἐπιδειϰτιϰόν for the third genre of oratory on the grounds that it would include speeches that formam suadendi habent et plerumque de utilitatibus Graeciae locuntur; also ps.-Dionysius, Rhetoric 347.5–16 (Usener-Radermacher). See Pernot (n.7) 28–30.

¹⁰ For the circumstances of publication see Pliny, Letters 3.18.1; cf. Durry, M., Pline le Jeune: panégyrique de Trajan (1938) 5–6Google Scholar; more generally on the extent of the revision see Dierauer, J. in Büdinger, M. (ed.), Untersuchungen zur römischen Kaisergeschichte (1968) 187–217Google Scholar; Mesk, J., ‘Die Überarbeitung des plinianischen Panegyricus auf Traian’, WS 32 (1910) 238–60Google Scholar.

¹¹ Cicero describes the delivery of the Pro Marcello at Fam. 4.4.3–4; on his own account, his decision to speak at all was a last-minute one, which probably suggests that the published version was at best a later construct on the basis of what he remembered he had said, rather than a precise reproduction of it. Some, however, have argued that the speech we have derives from a stenographic record, and hence closely represents the words he used on the day: see Gotoff, H.C., Cicero's Caesarian speeches: a stylistic commentary (1993) xxxiiGoogle Scholar.

¹² For reasons of space I shall be discussing these only briefly: see below p. 77.

¹³ Pliny refers to his speech as a gratiarum actio (Panegyric 1.6,90.3; cf. 4.2). In two of the three Cicero speeches the term gratiarum actio, while not mentioned directly, is alluded to as the form of the speech: see Red. sen. 1, 30, 31; Marc. 33, 34 (cf. Fam. 4.4.4); cf. also Red. Quir. 24.

¹⁴ Dyer, R.R., ‘Rhetoric and intention in Cicero's Pro Marcello’, JRS 80 (1990) 17–30Google Scholar.

¹⁵ See also Earl, D.C., The moral and political tradition of Rome (1967) 60Google Scholar; however, the evidence he cites is all from well into the imperial period.

¹⁶ Cf. Weinstock, S., Divus Julius (1971) 234–7Google Scholar.

¹⁷ Fam. 5.4.2; Att. 1.2.1; Rab. perd. 13; Fam. 5.1.2. See Hellegouarc'h, J., Le Vocabulaire latin des relations et des partis politiques sous la république (1963) 261–3Google Scholar.

¹⁸ See Dyer (n. 14) 21 nn.27 and 29; the references are Att. 14.22.1, Ad Brut. 2.5.5, 1.15.10. For this standard theme see e.g. Sallust, Histories 1.55.1.

¹⁹ For example, he regards the use of iste in the speech as derogatory (29 with n.70); but it is frequently used in neutral, or even laudatory contexts; in all the examples he cites it is employed adjectivally to qualify abstractions, rather than to refer to an individual. See Kühner-Stegmann, (rev. Thierfelder, ) Ausführliche Grammatik der lateinische Sprache: Satzlehre I 621–2Google Scholar; TLL VII 2, 497–500Google Scholar; OLD s.v. iste esp. A 1. Many of Dyer's alleged ironies are taken out of context. For example, his claim (29) that Marc. 8 nulla est enim tanta uis quae non ferro et uiribus debilitari frangique possit looks to Caesar's fall, ignores the way in which Cicero develops the argument (see further below p.71): uis may be vulnerable, but Caesar, as he goes on to say, has demonstrated uirtus, which is not.

²⁰ In this and subsequent discussion I shall be assuming that attributions and implications of divinity in speeches are in some sense intended literally: that they are not simply metaphors or hyperbolic comparisons. Even under the Republic, while no official cult of any living Roman was practised at Rome, the idea that mortals might be worshipped as gods was a recognized phenomenon (contra Classen, C.J., ‘Gottmenschentum in der römischen Republik’, Gymnasium 70 (1963) 312–38)Google Scholar. At least since the early second century Roman leaders had accepted cultic worship in the Greek-speaking world (Price, S.R.F., Rituals and power: the Roman imperial cult in Asia Minor (1984) 42–7Google Scholar); and even by Romans divine honours were from time to time paid unofficially to leading figures: see Weinstock (n.16) 294–6, 300; Fishwick, D., The imperial cult in the Latin West I (1987) 54–5Google Scholar. Against such a background, the language of divinity in a panegyrical context could not be divorced from its literal significance: see above all Le, I.M.Quesnay, M. Du, ‘Vergil's First Eclogue’, PLLS 3 (1981) 29–182Google Scholar (at 97–113). Moreover, Caesar himself has been argued to have claimed divinity in a more systematic fashion, and to have sought to create for himself a central role in the official cult: see esp. Weinstock (n. 16) (though the matter is controversial: note Yavetz, Z., Julius Caesar and his public image (1983) 33–53)Google Scholar. In this case religious language in Pro Marcello will have carried especial point; see further below p. 77.

²¹ On the general importance of sapientia in the speech see Rochlitz, S., Das Bild Caesars in Ciceros Orationes Caesarianae (1993) 111–15Google Scholar.

²² E.g. Man. 10, 33, 36, 42; Mil. 91; Phil. 3.3, 5.43.

²³ Cf. Gotoff (n.11) 16.

²⁴ See Weinstock (n. 16) 112–27, though his account is controversial; for a more recent study that is balanced but tentative see Champeaux, J., Fortuna: recherches sur le culte de la Fortune dans le monde romain des origines à la mart de César (1982–1987) 259–91Google Scholar. In addition to the evidence they cite, note the references to Fortune on two Caesarian coins (Crawford, , RRC 464/3 and 465/8Google Scholar), both, like the Pro Marcello, from 46 B.C.

²⁵ See Hellegouarc'h, J., ‘La Fortune du prince’ in Bibauw, J. (ed.), Hommages à Marcel Renard, I (1969) 421–30Google Scholar (at 424–6); Wistrand, E., Felicitas imperatoria (1987) 35–43Google Scholar.

²⁶ In the year after this speech, the Senate decreed that a temple be built to Clementia Caesaris (Plutarch, Caesar 57.4; Appian, Civil War 2.106; Dio 44.6.4). See Weinstock (n. 16) 233–43; also more generally Treu, M., ‘Zur Clementia Caesaris’, MH 5 (1948) 197–217Google Scholar. On the role of the concept in Pro Marcello see Rochlitz (n.21) 40–58, 103–10.

²⁷ praesens is frequently used of gods; see TLL X 2, 839Google Scholar; OLD s.v. praesens 3. For the specific phrase praesentem intueor used of observing a god see Livy 45.28.4; cf. Curtius 3.6.17; Pan. lat. 10.2.1; also note Virgil, Aen. 3.174. See Gotoff (n. 11) 37.

²⁸ On Caesar and the cult of Victoria see Weinstock (n. 16) 91–103.

²⁹ A standard panegyrical motif: see the passages cited at Oakley, S. P., A commentary on Livy books VI–X 1 (1997) 451Google Scholar.

³⁰ Cf. Cicero's earlier claim (Marc. 11–12) that the monuments of Caesar's military victories, being purely physical, would one day decay to nothing – unlike his virtuous mercy, which thus endures forever.

³¹ For the centrality of this ambiguity in the ruler-cult of Asia Minor, see Price (n.20) esp. 133–233; for similar ambiguities in the Latin-speaking western empire, see Fishwick (n.20) e.g. 83–93, though his analysis is somewhat different.

³² See Scheid, J., Religion et piété à Rome (1985) 23–6Google Scholar; cf. Levene, D.S., Religion in Livy (1993) 8–9Google Scholar.

³³ This theme, of the ruler as the especial gift of the gods to men, is standard in ancient panegyrical writing and plays, as will be seen below, an especially important role in Pliny's Panegyric. See Fears, J.R., Princeps a diis electus (1977)Google Scholar; but note also Brunt, P.A., ‘Divine elements in the imperial office’, JRS 69 (1979) 168–75Google Scholar.

³⁴ On the multiple significance of salus more generally in a religious context see Fears, J.R., ‘The cult of Virtues and Roman imperial ideology’, ANRW II 17.2, 827–948Google Scholar (at 859–61).

³⁵ Weinstock(n. 16) 167–71.

³⁶ An analogy may be drawn with Greek, where ἀθάνατος is employed with similarly ambiguous overtones; see Price (n.3) 87–8.

³⁷ On the post reditum speeches generally see Nicholson, J., Cicero's return from exile (1992)Google Scholar, though he discusses religion only briefly (104–5).

³⁸ Cults of the Roman People were by time of the late Republic widespread in the East, and they were paralleled in Republican Rome by cults of the Genius Populi Romani and Genius Publicus. See Fears, J.R., ‘ – Genius Populi Romani’, Mnemosyne 31 (1978) 274–86CrossRefGoogle Scholar; also Oakley, S.P., A commentary on Livy books VI–X IIGoogle Scholar (forthcoming), note on Livy 7.30.20.

³⁹ For a more extended discussion of the religious themes in the Panegyric and its relationship to religious policy under Trajan see Schowalter, D.N., The emperor and the gods (1993)Google Scholar.

⁴⁰ For this standard topos cf. Woodman, A.J., Velleius Paterculus: the Tiberian narrative (2.94–131) (1977) 238–9Google Scholar.

⁴¹ For the use of ‘prayer’ language simultaneously to assimilate the emperor to and distance him from the gods cf. Price (n.3) 88–93; id. (n.20) 232–3.

⁴² On felicitas in Roman thought see Erkell, H., Augustus, Felicitas, Fortuna (1952) 43–128Google Scholar; on its role in the imperial cult see Fishwick, D., The imperial cult in the Latin west II (1991) 466Google Scholar; Wistrand (n.25) 44–62.

⁴³ Wistrand (n.25) 69 rather oddly wishes to distinguish the ‘felicitas of the ruler cult’, which he thinks is not at issue in this passage, from the ‘true felicitas’ which he sees Pliny as referring to here. Such a distinction, however, cannot be sustained: the language of the passage will inevitably evoke the ruler cult, even if Pliny is at the same time arguing that Trajan is especially worthy of it.

⁴⁴ Fishwick (n.42) 466.

⁴⁵ Also cf. Schowalter (n.39) 40–4, suggesting that the use of the phrase optimus princeps in 74.3 may have carried connotations of Jupiter Optimus Maximus (though Pliny also has provided it with more ambiguous associations, as he had previously (2.7) suggested that its meaning was purely human).

⁴⁶ For the terminology see Appel, G., De Romanorum precationibus (1909) 138Google Scholar; on this passage cf. Schowalter (n.39) 61, with 55–8 on the use of prayer language. On the general significance of prayers to the Emperor in the imperial cult see Price (n.3) 90–3.

⁴⁷ On this theme see Mause, M., Die Darstellung des Kaisers in der lateinischen Panegyrik (1994) 220–2Google Scholar.

⁴⁸ This is an especially significant theme: in ancient eyes not all gods necessarily possessed these qualities, and a god who heard everything was seen as the greatest of all. See Versnel, H.S., ‘Religious mentality in ancient prayer’, in Versnel, H.S. (ed.), Faith, hope and worship: aspects of religious mentality in the ancient world (1981) 1–64Google Scholar (at 28–9, 35–7).

⁴⁹ As do e.g. Wistrand (n.25) 69–70; Price (n.20) 246–7. Schowaiter (n.39) 71–80 sees something of the balance with regard to Pliny, but argues that it adds up to a consistent overall picture of the Emperor as specially favoured but strictly human; this conclusion can be reached only because Schowalter neglects much of the more extravagantly divine language attributed to Trajan, such as Panegyric 80 (which he barely mentions in this context).

⁵⁰ E.g. Livy 7.38.1, 22.37.5.

⁵¹ E.g. Caesar, Gallic War 1.30; Livy 28.9.11–15, 30.15.11–12, 30.17.7–9, 30.17.12–13; Tacitus, Annals 12.38.1. One does find occasional brief references to people being praised that are introduced with phrases like laudare ad caelum (e.g. Livy 2.49.1, 9.10.3; cf. 7.36.7, 22.30.7). These, however, do not appear in contexts that imply formal speeches, but simply refer to general talk; moreover the phrase appears to be a regular metaphor for the intensity of praise, rather than providing any indication of the content of what is spoken.

⁵² Briscoe, J., A commentary on Livy books XXXIV–XXXVII (1981) 359Google Scholar.

⁵³ As there is a lacuna in the opening of Scipio's speech in Polybius 21.17.1–2, we cannot be entirely certain that Polybius included no such reference; but the opening words survive, and correspond to a later section of the argument in Livy, which suggests that this comment is indeed Livy's addition. See Engel, J.-M., Tite-Live histoire romaine tome XXVII, livre XXXVII (1983) 147Google Scholar.

⁵⁴ See Walsh, P.G., Livy: book XXXVII (1992) 171Google Scholar.

⁵⁵ Polybius 21.16.9: , ‘above all they asked what it was necessary to do to obtain peace and alliance with Rome.’

⁵⁶ Polybius 30.18.7: , ‘appearing entirely contemptible, he received for that very reason a kindly answer’.

⁵⁷ The same story appears in Suetonius, Tiberius 27, where the term adulatio is also used, but the offending phrase is sacras occupationes.

⁵⁸ Curtius 8.5.8: hi tum caelum illi aperiebant, Herculemque et Patrem Liberum et cum Polluce Castorem nouo numini cessuros esse iactabant, ‘these then began to open heaven to him, boasting that Hercules and Father Liber and Castor along with Pollux would give way to the new god’.

⁵⁹ Justin 12.7.1–3; Plutarch, , Alexander 54Google Scholar; Arrian, Anabasis 4.9–12 (esp. 4.10.1).

⁶⁰ Arrian, Anabasis 4.10.6–7; in his version the speech is given by Anaxarchus, not Cleo.

⁶¹ Arrian referred to Anaxarchus earlier (Anabasis 4.9.9) as one of (‘those giving in to Alexander in flattery with regard to [proskynesis]’); but this is specifically connected with proskynesis itself, rather than with the speech in support of deification, as in Curtius; in Arrian the speech comes slightly later.

⁶² Levene (n.32) 184–9. Contrast Plutarch, Camillus 10.5, where the Falerians refer to Camillus as , words that have no parallel in Livy.

⁶³ See Briscoe, J., A commentary on Livy books XXXI–XXX11I (1973) 313Google Scholar.

⁶⁴ Briscoe (n.63) 312. For the religious significance of the word in Polybius see Walbank, F.W., A historical commentary on Polybius II (1967) 613–14Google Scholar. It is true that the parallel passage in Livy comes before the start of the panegyric. However, in having a panegyric here at all Livy has transferred into the mouths of the Greeks sentiments that in Polybius had been expressed in the authorial voice (see Tränkle, H., Livius und Polybios (1977) 98Google Scholar). Since Livy was creating the panegyric in the first place, he could have included the Greeks’ words here as part of it. Instead, in inventing his panegyric he made the significant choice to ignore Polybius’ account of the Greeks’ own panegyrical statements, and to confine himself to the views of Polybius himself.

⁶⁵ See Kraus, C.S., Livy: Ab urbe condita book VI (1994) 175Google Scholar.

⁶⁶ Oakley (n.29) 478–9.

⁶⁷ Kraus (n.65) 176.

⁶⁸ See Kraus (n.65) 146–219 passim.

⁶⁹ It is worth observing further that the whole episode is introduced as an example of ‘sedition growing by the day’ (6.14.1 gliscente in dies seditione). Moreover the overtones of the centurion in the context of the work as a whole are such as to make his words look especially unreliable – he closely recalls earlier ‘injured parties publicly speaking out’ at 3.13.1–4 (cf. 3.24.3–5) and 3.71–2, who, despite their solid Roman pedigree, were shown by Livy to be lying. This adds to the sense that his words here are to be read as an example of pernicious flattery.

⁷⁰ For this double-edged treatment of Scipio, which itself recalls the panegyrical treatment of the ruler as both human and divine see Levene (n.32) 18–19, 61–2, 67–76.

⁷¹ E.g. Annals 4.153, 4.55–6.

⁷² On the way in which Tacitus slants this passage so as to endorse the negative view of Augustus see Goodyear, F.R.D., The Annals of Tacitus I (1972) 156Google Scholar.

⁷³ For this idea cf. Suetonius, Augustus 98, where Augustus' acceptance of divine honours is treated by Suetonius as part of a sequence anticipating and foreshadowing his posthumous divinity.

⁷⁴ Note the precise phrase in 15.74.4: deum honor, ‘the honour due to gods’. For other examples cf. Annals 13.8.1; Pliny, Panegyric 52.3; Suetonius, Tiberius 26.1. See Price (n.20) 146–56; Fishwick (n.42) 540–50.

⁷⁵ Martin, R.H. & Woodman, A.J., Tacitus: Annals IV (1989) 192–3Google Scholar argue that Tacitus presents this criticism in terms that undermine its validity; but, even on this view, this does not add up to a rejection of emperor worship per se, for which even Tiberius in his own speech rejecting the cult was prepared to allow a limited role (4.37.3). The more favourable reactions to this speech that Tacitus reports are in terms of Tiberius' modestia (‘moderation’) or quia diffideret (‘because he distrusted himself), thus locating the grounds for the objection in Tiberius’ own character, rather than a fundamental rejection of the cult.

⁷⁶ See Ullmann, R., La Technique des discours dans Salluste, Tite-Live et Tacite (1927)Google Scholar (though note the reservations of e.g. Briscoe (n.63) 17–22); also more recently Luce, T.J., ‘Structure in Livy's speeches’ in Schuller, W. (ed.), Livius: Aspekte seines Werkes (1993) 71–85Google Scholar.

⁷⁷ This also derives from Polybius (21.23.4), though the idea of the divine there depends on Ursino's (very plausible) conjecture: . Moreover, Livy has transferred the thought from a later point in Polybius' version of the speech.

⁷⁸ Polybius 21.23.13: , ‘they seemed to everyone to have spoken moderately and well’. Cf. Hoch, H., Die Darstellung der politischen Sendung Rams bei Livius (1951) 15Google Scholar.

⁷⁹ This aspect of the passage, and many other similar sections of the speech, are extensively discussed by Bartsch (n.8) 148–87; she focuses especially on their potential for being self-defeating, as they draw attention to the possibility that his own language may be comparably insincere.

⁸⁰ The self-contradiction is observed by Bartsch (n.8) 162–4. However, she discusses it from a very different standpoint, focusing on its specific effect within the Panegyric; she does not address it in the context of the intrinsic ambiguities of ruler-cult as a whole, nor against the background of the wider treatment of these particular themes in panegyric as a genre.

⁸¹ For the first, see TLL VII 1, 2132Google Scholar; OLD s.v. inter 2, 3: this appears to be the interpretation of Sherwin-White, A.N., The Letters of Pliny: a historical and social commentary (1966) 452Google Scholar. For the second, see TLL VII 1, 2127–8Google Scholar; OLD s.v. inter 1b (for a similar ambiguity cf. Plautus, , Rudens 6–7Google Scholar). I owe this suggestion to Tony Woodman.

⁸² Schowalter (n.39) 74– 5 regards this passage as humorous or possibly sarcastic: neither seems a natural reading of its tone.

⁸³ In Topica 90 religion is part of aequitas – which is explicitly said to be part of judicial oratory.

⁸⁴ For some early examples see Thucydides 3.67.6; Euripides, , Medea 580–5Google Scholar, Hecuba 1187–94; Aristophanes, , Acharnians 370–4Google Scholar M. See Pernot (n.7) 499–500; cf. Mossman, J., Wild justice: a study of Euripides Hecuba (1995) 133–4Google Scholar; Buxton, R., Persuasion in Greek tragedy (1982) 184Google Scholar.

⁸⁵ The claim was especially associated with Protagoras: see Aristotle, , Rhetoric 1402a23Google Scholar. The best-known attack upon it is perhaps the extended and hostile satire in Aristophanes, , Clouds (889–1114Google Scholar; note the specific allusion to the Protagorean doctrine at 112–15). See also Plato, , Apology 19b5–clGoogle Scholar; Isocrates, , Antidosis 15Google Scholar; Quintilian 2.16.3.

⁸⁶ For a general, though unsympathetic, discussion of Plato's attitudes to oratory see Vickers, B., In defence of rhetoric (1988) 83–147Google Scholar; Yunis, H., Taming democracy (1996) 117–236Google Scholar is more sympathetic.

⁸⁷ Plato, , Gorgias 452d–e, 456a–7cGoogle Scholar.

⁸⁸ Esp. Gorgias 471e7–2a1: , ‘this sort of proof is worthless with respect to the truth’. Compare Apology 17b1–8.

⁸⁹ Plato, , Gorgias 502d10–4e4Google Scholar. See however Irwin, T.H., Plato: Gorgias (1979) 215Google Scholar, who argues that this ‘rhetoric’ (unlike that of the Phaedrus) does not involve normal rhetorical techniques at all: it is not that Plato here regards standard forms of rhetoric as acceptable if one is aiming at moral ends; cf. also Vickers (n.86) 109–10. On the central role of morality in validating rhetoric for Plato see Yunis (n.86) 120–1, and cf. Politicus 296d7–7b3, 304al–2.

⁹⁰ Plato, , Phaedrus 273e5–a5Google Scholar.

⁹¹ Plato, , Phaedrus 262a–bGoogle Scholar; the same paradox of deception is explored at greater length in Hippias Minor.

⁹² Plato, , Phaedrus 260d7–9Google Scholar.

⁹³ See for example the rejection of arguments from probability at Phaedrus 272d–3e on the grounds that they appear like the truth to the masses (Phaedrus 273d4: ), but are not truth itself.

⁹⁴ See Republic 389b2–9; also Politicus 304c10–d2.

⁹⁵ Note the actual example of such an acceptable falsehood that Plato gives at Republic 414b8–15d2, a long myth whose object is to persuade the citizen body to accept his proposed constitution.

⁹⁶ On the significance of Plato for Cicero's rhetorical theory see Michel, A., Rhétorique et philosophie chez Cicéron (1960) 86–100Google Scholar. On the great influence of Phaedrus in particular upon the literature of the Empire, both Greek and Latin, see Trapp, M.B., ‘Plato's Phaedrus in the second century’, in Russell, D.A. (ed.), Antonine literature (1990) 141–73Google Scholar.

⁹⁷ See Vickers (n.86) 167–70.

⁹⁸ Cicero, , Orator 11–15, 39, 41–2Google Scholar. However, this is part of his paradoxical enlistment of Plato on the side of oratory, suggesting that the ideal orator whom he describes is a Platonic Form (Orator 10, 101), as well as that the true orator must be philosophically educated (Orator 11–19, 118–19). For other uses of this latter argument see e.g. De inventione 1.1; De oratore 1.50–7, 3.137–42; Quintilian 2.15.28–9, 2.20.5–10, 12.2.1–28; Tacitus, Dialogus 31.1–2. At De oratore 1.219–33 Cicero places in the mouth of Antonius arguments for the irrelevance of philosophy to oratory; however, these are undermined by Cicero in his own voice at 2.1–6 (see also the acute comment on this passage by Quintilian 2.17.5–6).

⁹⁹ Cf. De oratore 3.60–1, 3.72–3, 3.122, 3.129; also 1.85–93 where the Academic philosopher Charmadas is reported as attacking rhetoric in very similar terms to those used by Plato in Gorgias. Indeed the whole setting of De oratore is modelled on Phaedrus (see 1.28 and cf. Trapp (n.96) 146, 165–6).

¹⁰⁰ De inventione 1.3; De oratore 1.47–50; cf. 1.61–3, 1.229–33, 3.129, 3.142–3; Quintilian 12.2.9.

¹⁰¹ Note, however, the arguments of Quintilian 2.20.5–10, seeking to show that rhetoric is not a morally neutral tool, but is intrinsically virtuous.

¹⁰² See further Quintilian 2.16.1–10.

¹⁰³ Note De oratore 3.55: ‘the greater the eloquence, the more it must be united with morality and the greatest wisdom; if we give skill in speaking to people lacking these virtues, we shall not have made them orators, but shall have given some weapons to madmen’. Similarly Quintilian 2.15.32, 12.1.1–2. For the more general idea that rhetoric should be employed only for moral ends cf. De oratore 1.202, 2.35, 2.85; Quintilian 1. pr. 9–20, 2.15.30–2, 2.16.7–11, 12.1, 12.7.1–7. In fact Quintilian not only argues that it should be so used, but even follows Plato in arguing that no one who is not virtuous can truly be an orator at all: see Winterbottom, M., ‘Quintilian and the vir bonus’, JRS 54 (1964) 90–7Google Scholar.

¹⁰⁴ Cf. further Quintilian 5.14.29, 6.1.7, 12.1.33–45. In particular defending the guilty, even by lies, is often justified on the grounds that a wider morality is thereby served: see Cicero, De officiis 2.49–51; Quintilian 12.1.40–3.

¹⁰⁵ See e.g. Gorgias, , Helen 21Google Scholar; Isocrates, , Busiris 9Google Scholar. Also Plato, , Phaedrus 265c8–9Google Scholar; cf. Yunis (n.86) 199–201.

¹⁰⁶ Note Cicero, De oratore 2.341, where epideictic oratory for the purposes of entertainment is seen as a fundamentally Greek activity that a Roman need not concern himself with; also Orator 65 (on which see further below p. 97). On the relative coolness towards epideictic at Rome in this period see Pernot (n.7)50–3, 106–11.

¹⁰⁷ See Buchheit (n.5) 96–108; at 105 he argues that this passage is intended by Plato as a response to Isocrates (esp. Busiris 4); cf. the ironic account of epideictic at Menexenus 234c1–5b2. On Plato's hostility to epideictic in general and its implications for his own paradoxical adoption of encomiastic techniques see Nightingale, A. Wilson, Genres in dialogue (1995) 93–132CrossRefGoogle Scholar; also 154–8, where she argues that Plato saw the epideictic genre as fundamentally incompatible with truth. Note, moreover, the attacks on lying panegyric by Democritus 86DK63, 86DK192 (see Buchheit (n.5) 90–1); also Philodemus, , Rhetorica I 285–6Google Scholar (Sudhaus).

¹⁰⁸ On the political function of epideictic see Pernot (n.7) 607–21. Although funeral orations stand outside the scope of this paper, Loraux, N., The invention of Athens, tr. Sheridan, A. (1986) 223–5Google Scholar, is revealing: she denies that the funeral oration is an epideictic speech proper, and does so largely on the grounds of its political importance.

¹⁰⁹ Note Cole, T., The origins of rhetoric in Greece (1991) 152Google Scholar, who suggests that the elevation of persuasion as the object of rhetoric and the consequent problematizing of epideictic, is a Platonic innovation. Conversely see Yunis (n.86) 181–8 on Plato's attempt to assimilate all corrupt rhetoric to epideictic, on the grounds that oratorical display is the only end aimed at in most ostensibly deliberative speeches. On the problematic status of epideictic in general see the extended discussion in Pernot (n.7) 493–605.

¹¹⁰ See note 8 above on ‘figured speech’, many examples of which assume precisely such an approach. Ahl's article, cited there, is fundamental; he does not, however, consider the possibility that the accounts of ‘figured speech’ in the rhetorical theorists are at least in part governed not by the reality of speech under a tyranny, but by the necessity of defending panegyrical rhetoric itself, associating it either with covert persuasion or with a hidden statement of the truth. His own example (198–9) is that to compare an Emperor to a god would by its very nature be an instance of ‘figured speech’; but this anachronistically ignores the religious significance of ruler-cult in the ancient world.

¹¹¹ Cf. Isocrates, , Evagoras 9–10Google Scholar, distinguishing poetic from prose panegyrics partly on the grounds that the latter are more bound to truth: . Alexander Numenius 2.31–3.20 (Spengel) distinguishes ἔπαινος, where the speaker is sincere, from ἐγϰώμιον, where he is not (a distinction based on Aristotle, , Rhetoric 1367b28–36Google Scholar, who, however, presents it in terms of the topic of praise alone). See Pernot (n.7) 219–20.

¹¹² Aristotle, by contrast, does seem to treat this move as more actively designed to mislead: note the word παϱαλογιστιϰόν – ‘mistaken reasoning’ (Rhetoric 1367b4). On this theme in the Rhetoric, and the attempts by later theorists to face up to the moral difficulty see Pernot (n.7) 517–18. For the part that the concept of ‘vices close to virtues but not identical with them’ plays in Aristotle, 's moral philosophy, compare Nicomachean Ethics 1108b30–9a19Google Scholar.

¹¹³ On Quintilian's attempt to escape the moral problem here cf. Bartsch (n.8) 171. Mause (n.47) 23–4 cites Quintilian's recommendation of amplificatio in panegyric (3.7.6) as proof of his belief that truth should be distorted there; however, this derives from a misunderstanding of amplificatio, which generally refers to the use of linguistic and rhetorical devices to stress a particular theme and enhance its importance: see Lausberg, H., Handbuch der literarischen Rhetorik (1960) 220–7Google Scholar; Ueding, G. (ed.), Historisches Wörterbuch der Rhetorik I (1992) 445–9Google Scholar. I shall be exploring the question of amplificatio further in a separate article.

¹¹⁴ Brutus 62 (cf. his praise of archaic Athenian funeral orators at De legibus 2.63 for their strict regard for truth). For the same complaint see Livy 8.40.4–5. See Woodman, A.J., Rhetoric in classical historiography: four studies (1988) 91 with nn. 89–90Google Scholar.

¹¹⁵ Contra Brunt, P.A., ‘Cicero and historiography’, in Φιλίας χάϱιν: Miscellanea in onore di Eugenio Manni I (1979) 311–40, at 329–32Google Scholar (= Studies in Greek history and thought (1993) 181–209)Google ScholarPubMed, who argues that disregard for the truth was seen as a fundamental feature of the epideictic genre. However, he fails to see that such an attitude to epideictic at Rome (and indeed often at Greece) was evident only in passages where it was being condemned: all his relevant examples fall into this category.

¹¹⁶ Above n.31; cf. Beard, M., ‘Cicero and divination: the formation of a Latin discourse’, JRS 76 (1986) 33–46Google Scholar; Levene (n.32) 16–30.

¹¹⁷ This is obvious for direct theoretical discussions of rhetoric; but it is no less true of panegyrics embedded within narratives. For a discussion of how all reporting of speeches in narrative involves incorporating the perspective of the narrator see Sternberg, M., ‘Proteus in Quotation-Land: mimesis and the forms of reported discourse’, Poetics Today 3:2 (1982) 107–56CrossRefGoogle Scholar; id., ‘How indirect discourse means’, in Sell, R.D. (ed.), Literary pragmatics (1991) 62–93Google Scholar (esp. 87–9); cf. Vološinov, V.N., Marxism and the philosophy of language, tr. Matejka, L. & Titunik, I.R. (1973) 115–23Google Scholar.

¹¹⁸ On this speech see Moles, J.L., ‘The date and purpose of the Fourth Kingship Oration of Dio Chrysostom’, Classical Antiquity 2 (1983) 251–78CrossRefGoogle Scholar; more generally id. (n.7).

¹¹⁹ Which emperor is more controversial; but Trajan is the most likely recipient, as argued by Moles (n. 118).

¹²⁰ Or. 4.19; Alexander, at any rate, takes what Diogenes says here as a complimentary address, before being deflated by what follows (4.20).

¹²¹ Or. 4.139; see Moles (n. 118) 255–61.

¹²² See especially Menander Rhetor (II) 368–70.

¹²³ For the sources and origin of Hellenistic ruler-cult and its influence upon the development of the imperial cult see Price (n.20) 23–59; also n.20 above.

¹²⁴ For what follows see above all Price (n.20) 7–11; Phillips, C.R., ‘The sociology of religious knowledge in the Roman Empire to A.D. 284’, ANRW, II 16.3 (1986) 2677–773Google Scholar.

¹²⁵ Price (n.20) 9–11 argues that with the imperial cult one should focus not on individual belief, but on the ‘public cognitive system’ (9) as expressed in ritual; merely to address the question of ‘belief’ in ancient religion he regards as the imposition of Christian assumptions; see also Phillips (n. 124) 2697–711. Nevertheless the question of belief in religion could concern the ancients themselves in certain contexts (see e.g. Moles (n.7) 356 with n. 149; Levene (n.32) 10–13). Price himself recognises that ancient writings can reveal beliefs on religious matters, and at 114–17 briefly considers some of these, concluding that no element of them can be identified with true scepticism about the imperial cult.

¹²⁶ On the gap between theory and practice in panegyric more generally see Pernot (n.7) 251–3; Russell and Wilson (n.3) xviii–xix.

¹²⁷ More generally on the divergences between Greek and Latin rhetoric under the Roman Empire, see Russell, D.A., ‘Greek and Latin in Antonine literature’ in Russell, D.A. (ed.), Antonine literature (1990) 1–17, esp. 10–17Google Scholar.

¹²⁸ The question of the nature of historical truth in the ancient world, and to what extent ancient historians were expected to pursue it, is highly controversial. See e.g. Wiseman, T.P., Clio's Cosmetics (1979)Google Scholar; Brunt (n.115); Woodman (n.114) esp. 70–116; Pelling, C.B.R., ‘Truth and fiction in Plutarch's Lives’ in Russell, D.A. (ed.), Antonine literature (1990) 19–52Google Scholar; Moles, J.L., ‘Truth and untruth in Herodotus and Thucydides’ in Gill, C. and Wiseman, T.P. (eds.), Lies and fiction in the ancient world (1993) 88–121CrossRefGoogle Scholar. According to Woodman, the nature and expectation of truth in history is closely comparable to that in rhetoric; but cf. Pelling 42–3 with n.65; Moles 114–18; one may add in general that, as this paper has shown, the role of truth in rhetoric is itself more complex than the debate has often appeared to assume.

¹²⁹ Cf. Pernot (n.7) 661–4.

¹³⁰ Cf. Bartsch (n.8) 148–87, who reaches a similar conclusion from a different perspective. Her focus is on the unavoidable slippage between the language of praise and that of mendacious flattery; but with religion it is only, as I have argued, in certain generic contexts that such a slippage matters, as only here does the question of truth or sincerity necessarily impose itself.