¹ For an introduction to the history of reception, see Tomkins, J. P., ‘The reader in history: the changing shape of literary response’, in Tomkins, J. P. (ed.), Reader-Response Criticism from Formalism to Post-Structuralism (1980), 201–32Google Scholar. See also the collection of essays edited by Eco, U., Interpretation and Overinterpretation (1992).CrossRefGoogle Scholar

² On the use of these terms in relation to Ovid, see the discussion and bibliographic material of Nugent, S. G., ‘Tristia 2: Ovid and Augustus’, in Raaflaub, K. A. and Toher, M. (eds), Between Republic and Empire: Interpretations of Augustus and his Principate (1990), 239–57Google Scholar, at 241; Barchiesi, A., Il poeta e il principe. Ovidio e il discorso augusteo (1994), 34–6Google Scholar; Williams, G. D., Banished Voices: Readings in Ovid's Exile Poetry (1994), 154–8Google Scholar; Habinek, T., The Politics of Latin Literature (1998), 3–14Google Scholar. For a theoretical treatment of the issues, see Kennedy, D. F. ‘“Augustan” and “Anti-Augustan”: reflections on terms of reference’, in Powell, A (ed.), Roman Poetry and Propaganda in the Age of Augustus (1992), 26–58Google Scholar, while Galinsky, G. K., Augustan Culture: an Interpretive Introduction (1996), 225, 244–6Google Scholar, draws attention to the need to see ‘Augustan’ as a term with a wider frame of reference than that of agreement (or disagreement) with the views of the princeps. Note also the important article of Ahl, F., ‘The art of safe criticism in Greece and Rome’, AJPh 105 (1984), 174–208Google Scholar.

³ The reciprocal relationship between Augustus' edict of relegation, described as ‘tristibus … uerbis’ (Trist. 2.133) and Ovid's exile poetry, Tristia, is noted by Habinek, op. cit. (n. 2), 155–6.

⁴ For this view see most recently Galinsky, op. cit. (n. 2), 269.

⁵ On the issue of ‘Augustanism’ in the Metamorphoses, see e.g. B. Otis, Ovid as an Epic Poet (1966), 145, 302–5, 329; Galinsky, G. K., Ovid's Metamorphoses: an Introduction to the Basic Aspects (1975), 210–17Google Scholar; Hardie, P., ‘Questions of authority: the invention of tradition in Ovid Metamorphoses 15’, in Habinek, T. and Schiesaro, A. (eds), The Roman Cultural Revolution (1997), 182–98.Google Scholar

⁶ On 559–60 see Bailey, D. R. Shackleton, ‘Notes on Ovid's poems from exile’, CQ 32 (1982), 390–8CrossRefGoogle Scholar, at 393, who construes surgens as neuter. It is, however, perfectly possible to take surgens as masculine, referring to Ovid himself; for the sliding relation between an author and his text, compare the discussion of Tristia 2.5 below, at p. 21. On the relation between this passage and the opening of Ovid, Metamorphoses 1, see Barchiesi, A., ‘Voci e istanze narrative nelle Metamorfosi di Ovidio’, MD 23 (1989), 55–97Google Scholar, at 91, who notes the subtle change from ‘ad mea perpetuum deducite tempora carmen’ (Met. 1.4) to ‘in tua deduxi tempora, Caesar, opus’ (Trist. 2.560).

⁷ On this passage see e.g. Otis, op. cit. (n. 5), 303–4; Galinsky, op. cit. (n. 5), 259.

⁸ Contrast however Millar, F. G. B., ‘Ovid and the Domus Augusta: Rome seen from Tomoi’, JRS 83 (1993), 1–17Google Scholar, at 8, who regards this passage and the references to the Metamorphoses as straightforward panegyric: ‘Looking back in Tristia II on his poetic achievement before his exile, Ovid, if anything, rather underestimates how profoundly shaped by Augustan loyalism this work had been (555–62).’

⁹ cf. Trist. 2.61–2 (on the Ars Amatoria): ‘quid referam libros, illos quoque, crimina nostra, / mille locis plenos nominis esse tui?’ and the discussion of Barchiesi, op. cit. (n. 2), 22–3; Williams, op. cit. (n. 2), 172.

¹⁰ Note especially Trist. 2.66: ‘inuenies animi pignora multa mei’. Galinsky, op. cit. (n. 5), 219 regards Trist. 2.63, ‘inspice maius opus’, as an echo of Virgil, , Aen. 7.44Google Scholar: ‘maius opus moueo’. This argument is even more convincing if one compares Ovid's use of maius opus at Am. 3.1.24 to refer to the possiblity of writing tragedy. At Trist. 2.63 Ovid is also referring to Met. 15.750–1, ‘nequeenim de Caesaris actis / ullum maius opus, quam quod pater exstitit huius’, where Julius Caesar's greatest achievement is his (adoptive) paternity of Augustus; on this passage of the Metamorphoses, see further S. E. Hinds, ‘Generalising about Ovid’, in A. J. Boyle (ed.), Ramus: Critical Studies in Greek and Roman Literature 16. Imperial Roman Literature I (1988), 4–31, at 24–6. Note also Fasti 5.567–8, ‘spectat et Augusto praetextum nomine templum, / et uisum lecto Caesare maius opus’, where Mars is looking at Augustus' temple to Mars Ultor. Other occurrences of maius opus in Ovid are Ars Am. 3.370, Rem. 109, and Met. 8.328.

¹¹ Heyworth, S. J., ‘Notes on Ovid's Tristia’, PCPhS 41 (1995), 138–52Google Scholar, at 146 n. 39: ‘… less than 2 pages reveal the Metamorphoses as shaped by Augustan loyalism, without mention of any episode between 1.205 and Aeneas in book 13!’

¹² Evans, H. B., Publica Carmina. Ovid's Books from Exile (1983), 11Google Scholar: ‘Yet the problem of Ovid's attitude to Augustus cannot be ignored in any examination of the exile poetry. As proponents of the non-political Ovid have observed, the poet did not give major emphasis to imperial themes in his earlier works. The Ovidian concordance reveals that by far the largest number of references to Augustus appear in the books for [sic] Tomis. This is not surprising when we remember the main themes of the exile poetry, Ovid's defense of his conduct and appeals for imperial mercy.’ Contrast however Williams, op. cit. (n. 2), 162 (on Tristia 2): ‘To take sides with the self-caricature of the poet against his caricature of the emperor may be to enter into the spirit of the poem, but it is not criticism.’

¹³ On the textual difficulties of this passage, see Luck, G., P. Ovidius Naso. Tristia Band I (1967), 14Google Scholar; Heyworth, op. cit. (n. 11), 139–40.

¹⁴ Compare, for instance, Propertius 1.7.13: ‘me legat assidue post haec neglectus amator’, where me stands for the text of Propertius. Note also Ovid, , Ars Amatoria 1.2Google Scholar, ‘hoc legat et lecto carmine doctus amet’, where me is found as a variant for hoc in some manuscripts. Regardless of whether me or hoc is the text Ovid wrote, the variant strikingly illustrates a hesitation as to whether or not to equate a text with the author.

¹⁵ Nugent, op. cit. (n. 2), 251, speculates that Ovid's claim to personal virtue may glance at Augustus' own immoralities.

¹⁶ On the complexities of Catullus 16 see Selden, D. L., ‘Ceveat lector: Catullus and the rhetoric of performance’, in Hexter, R. and Selden, D. L. (eds), Innovations of Antiquity (1992), 461–512Google Scholar.

¹⁷ See e.g. Trist. 1.9.59–60, 3.2.5–6, Ex Ponto 2.7.47–50, 4.8.19–20, Martial 1.4.8, ‘lasciua est nobis pagina, uita proba’, with Citroni's commentary ad loc. and Luck, G., P. Ovidius Naso. Tristia Band 2 (1977), 131–2Google Scholar.

¹⁸ See Barchiesi, A., ‘Insegnare ad Augusto: Orazio, Epistole 2,1 e Ovidio, Tristia II’, MD 31 (1993), 149–84Google Scholar, at 176–8.

¹⁹ For comparison with the opening of Horace, Epist. 2.1, see Barchiesi, op. cit. (n. 18), Barchiesi, op. cit. (n. 2), 20–1, and Williams, op. cit. (n. 2), 180–1. See also Cutolo, P., ‘Captatio ed apologia in Tristia II’, in Gallo, I. and Nicastri, L. (eds), Cultura, poesia, ideologia nell' opera di Ovidio (1991), 265–86, at 277–8Google Scholar.

²⁰ John Moles suggests to me the possibility that falli may evoke the Dios Apate in Iliad 14.

²¹ The issue of misreading is discussed by Eco, op. cit. (n. 1), 45–88. Eco argues (52) ‘that we can accept a sort of Popperian principle according to which if there are no rules that help to ascertain which interpretations are the “best ones”, there is at least a rule for ascertaining which ones are “bad”’. This positivism is challenged by R. Rorty in Eco, op. cit. (n. 1), 89–108.

²² See Habinek, op. cit. (n. 2), 151–69, who sees Ovid's representation of Tomis and the Roman frontier in the exile poetry as a discourse of colonization.

²³ Nugent, op. cit. (n. 2), 250–1, argues that the issue of whether Augustus read the Ars Amatoria is ‘a nowin situation proposition for Augustus’.

²⁴ Barchiesi.op. cit. (n. 2), 22–3, argues that ‘Il punto è che, se Augusto avesse avuto tempo, avrebbe trovato le parole “nullum […] crimen” nell' Ars, il carmen che per lui è un crimen: inque meo nullum carmine crimen erit (1,34 “e nella mia poesia non ci sarà alcun capo d'accusa”). L'argomento è circolare (e serpentino). Questo testo non è incriminabile perché dice a chiare lettere: “Io non sono un testo incriminabile”.’ See also Barchiesi, op. cit. (n. 18), 166–7.

²⁵ Nugent, op. cit. (n. 2), 251, detects a different emphasis: ‘Thus does Ovid prescribe Augustus’ reading and, with the extended revisionist reading of earlier texts, foist his own readings upon Augustus. Again Ovid recommends a specific reading of his own works to Augustus: “Just open my books and you'll see what a role you play there, how I really value you”.’

²⁶ Commenting on the alteration of Ars. Am. 1.33, ‘nos Venerem tutam concessaque furta canemus’ to ‘nil nisi legitimum concessaque furta canemus’ (Trist. 2.249), Barchiesi, op. cit. (n. 18), 166, remarks that: ‘L'interpolazione nil nisi legitimum mostra che non si è mai sicuri abbastanza: l' Ars si era protetta contro le accuse, ma è stata condonnata.’ As Barchiesi notes on the same page, the alteration to the text of the Ars reinforces the earlier implication that Augustus was not an attentive reader; one might make the further point that such an alteration itself illustrates the independence of text from author: a text cannot only be misunderstood, but even altered (although, paradoxically, it is here the author, Ovid, who is altering his own text). See also Williams, op. cit. (n. 2), 206–9.

²⁷ The couplet referring to Lucretius also tellingly illustrates the unstable nature of signs and meaning. ‘Aeneadum genetrix’ in the first line signifies the works of Lucretius, here represented by the opening words. In the second line, the same pair of words literally refers to the ‘mother of the Aeneadae’ (itself a paradoxical idea, since Aeneas is her son). The shift in meaning between the first and second lines of the couplet mirrors the fluidity and uncertainty of a text's reception; Ovid imagines a reader rebelling and asking awkward questions right at the inception of the Lucretian text. Cf. Barchiesi, op. cit. (n. 2), 22–3, on ‘nullum … crimen’ in Trist. 2.240 and 2.247–50. On first words of literary works, see S. J. Heyworth, ‘Horace's Ibis; on the titles, unity, and contents of the Epodes’, in Cairns, F. (ed.), Papers of the Leeds International Latin Seminar 7 (1993), 85–96Google Scholar, at 85–6; Brown, P. G. McC., ‘An interpolated line of Terence at Cicero, De finibus 2.14’, CQ 47 (1997), 583–4CrossRefGoogle Scholar with n. 3.

²⁸ The Stoic implications of recta mens (recalling ὀρθὸς λόγος) are noted by Luck, G., P. Ovidius Naso. Tristia Band 2 (1977), 123Google Scholar and Williams, op. cit. (n. 2), 164. For Ovid's argumentation here, compare also the discussion of good and bad speech at Plato, Phaedrus 258d, which commences with a recognition that the writing of speeches is not in itself shameful, and Gorgias' defence of rhetoric as a morally neutral skill at Plato, Gorgias 456C6–457C3; see further E. R. Dodds' 1959 commentary on the Gorgias, and Vickers, B., In Defence of Rhetoric (1988), 84–120Google Scholar.

²⁹ On this passage see Barchiesi, op. cit. (n. 2), 23–4; Williams, op. cit. (n. 2), 165, 201–4.

³⁰ Williams, op. cit. (n. 2), 202.

³¹ Ovid's recusatio of an epic on Augustus is discussed by Williams, op. cit. (n. 2), 190–3, who notes the link with Amores 2.1 where Ovid renounces epic (and Jove) in favour of love. See also S. Stabryla, ‘In defence of the autonomy of the poetic world (some remarks on Ovid's ‘Tristia’ II)’, Hermes 122 (1994), 469–78Google Scholar, at 473–4.

³² In 357, though other manuscripts read uoluntas, two manuscripts (EV) have uoluptas, a reading favoured by Williams, op. cit. (n. 2), 170 n. 39 and Diggle, J., ‘Notes on Ovid's Tristia, Books I–II’, CQ 30 (1980), 401–19CrossRefGoogle Scholar, at 417–18. uoluptas seems preferable, since it accords well with ‘mulcendis auribus’ in 358; for the use of such language in contexts describing the pleasure of literature cf. e.g. Apuleius, Met. 1.1: ‘auresque tuas beniuolas lepido susurro permulceam’. Compare also the textual problems of Lucretius 2.257–8 where successive line endings in the manuscripts OQ are uoluptas and uoluntas. In Trist. 2.357–8 the manuscripts also offer variants for feres: feret, ferens, refert, and fores. For a full account of variants in this passage see Hall's 1995 Teubner apparatus. Luck's 1967 text is as follows:

nec liber indicium est animi, sed honesta uoluntas

plurima mulcendis auribus apta feret.

Hall's own text is more radical, linking the couplet with the succeeding lines on Accius and Terence:

si liber indicium est animi nec honesta uoluptas,

plurima mulcendis auribus apta dare,

Accius esset atrox, conuiua Terentius esset,

essent pugnaces qui fera bella canunt.

In this text, si, nec, and dare are all Hall's own conjectures.

³³ For the legal flavour of indicium, note such idioms as indicium postulare (to seek pardon by informing) and indicium profiteri and indicium offerre (to offer information): see OLD s.v. indicium 2b.

³⁴ Nugent, op. cit. (n. 2), 253: ‘In the context of Ovid's exilic work, however, the assertion is devastating, for it directly contradicts Ovid's stance throughout the entire corpus of his exilic poetry — namely, that his poetry of exile is a direct reflection of his life in exile. More specifically, the assertion undermines the claim to credibility that this apologia itself might have.’ Cf. Williams, op. cit. (n. 2), 171: ‘But Ovid's defence leaves him with a new problem. He defends the Ars by appealing to the benefits of a reading which is alive to the disjunction between poet and poetic persona; but he invites us to believe that in lines 353–8 poet and poetic persona are one. His defence can only stand if it is read without the kind of literary sophistication which that defence calls for to vindicate the Ars.’ The point is also made by Barchiesi, op. cit. (n. 2), 18: ‘Un testo così coinvolto nei problemi dell'interpretare e nella ricerca di letture sdoppiate, che forse ci sta invitando a una lettura non univoca di se stesso.’

The problems of interpreting a text whose author is still living are discussed by Eco, op. cit. (n. 1), 72–88. Typical of his approach is the following observation (73): ‘At this point the response of the author must not be used in order to validate the interpretations of his text, but to show the discrepancies between the author's intention and the intention of the text.’

³⁵ Note that an unpunctuated text of 365 allows equal priority to the alternative meaning: ‘What did Lesbian Sappho teach except to love girls?’

³⁶ Williams, op. cit. (n. 2), 193, suggests that Ovid portrays Homer's epics ‘as if they were Hellenistic love-romances’.

³⁷ Compare Propertius 2.8.29–36 for a similarly erotic treatment of the Iliad; see also Propertius 2.1.49–50; Benediktson, D. T., ‘Propertius’ elegiacization of Homer’, Maia 37 (1985), 17–26Google Scholar.

³⁸ cf. e.g. Propertius 3.12, Horace, , C. 3.7 and 10Google Scholar, Ovid, , Am. 3.4.23–4Google Scholar.

³⁹ Note that Agamemnon, at Odyssey 24.196–8Google Scholar predicts that the immortals will ensure Penelope's lasting fame in a song. There have been several recent treatments of the Odyssey centred on Penelope: see e.g. Katz, M. A., Penelope's Renown. Meaning and Indeterminacy in the Odyssey (1991)Google Scholar, Felson-Rubin, N., Regarding Penelope. From Character to Poetics (1994)Google Scholar.

⁴⁰ Compare Hermesianax, fr. 7.27–34Google Scholar Powell for an erotic interpretation of Homer's biography on the basis of his poems.

⁴¹ Williams, op. cit. (n. 2), 193–4, offers a different emphasis, seeing these rereadings of Homer as simply being Ovid's reply to Augustus' reading of Ovid's poetry: ‘If Augustus has been critically naive and one-sided in his evaluation of the Ars Amatoria, then Ovid can be equally one-sided and simplistic in his assessment of the Homeric poems, as well as of Greek tragedy and the poets he mentions in lines 363–470.’

⁴² Williams, op. cit. (n. 2), 198, interprets this list as a reminder to Augustus that ‘poetry can immortalise persons other than the poet himself. Ovid's characteristic overstating of his case in Tristia 2 with sheer weight of examples is noted by Williams, op. cit. (n. 2), 194.

⁴³ The ironic aspects of Ovid's presentation of Ennius and Lucretius are perceptively discussed by Barchiesi, op. cit. (n. 2), 15–18.

⁴⁴ Barchiesi, op. cit. (n. 2), 16–17.

⁴⁵ Note also Ovid's earlier account of his balanced and moderate lifestyle (Trist. 2.89–116), discussed by Williams, op. cit. (n. 2), 162–3.

⁴⁶ cf. Nisbet, R. G. M., ‘“Great and Lesser Bear” (Ovid Tristia 4.3)’, JRS 72 (1982), 49–56Google Scholar, at 56: ‘Most curiously of all, and surely deliberately, he [Ovid] professes an Augusta n ideal of marriage, even if the celestial pattern is marred by the imperfections of earth.’

⁴⁷ On Gallus' gossip about Augustus, see Dio Cassius 53–23–5, Boucher, J.-P., Caius Cornélius Gallus (1966), 49–54Google Scholar.

⁴⁸ On Ovid's use of Tibullus 1.6 see Barchiesi, op. cit. (n. 18), 171–2; Williams, op. cit. (n. 2), 195–6. In a curiously biographical moment Williams argues (195) as follows: ‘It is not too fanciful to believe that Tibullus’ telling complaint heu heu nunc premor arte mea (1.6.10) is what first led Ovid to select the poem as eminently suitable for reproduction in Tristia 2.’

⁴⁹ On the didactic elements in Propertius, see Wheeler, A. L., ‘Propertius as praeceptor amoris’, CPh 5 (1910), 28–40Google Scholar.

⁵⁰ On the link between games and poetic lusus see Williams, op. cit. (n. 2), 204–5, who notes the appearance of gambling in the Ars Amatoria. On the whole passage see also Pallarès, J. Gómez, ‘Sobre Ovidio, Tristia II, 471–492’, Latomus 52 (1993), 372–85Google Scholar, who interprets (380–1) the reference to poems written on ‘fucandi cura coloris’ (Trist. 2.487) as an allusion to Ovid's own Medicamina faciei feminae.

⁵¹ For lentus in a sexual (but metaphorical) context, cf. Catullus 28.9–10: ‘o Memmi, bene me ac diu supinum / tota ista trabe lentus irrumasti’. Cutolo, op. cit. (n. 19), 281–2 with n. 38, sees a link with Virgil's description of Tityrus as ‘lentus in umbra’ (Ecl. 1.4), with the intriguing possibility of a contrast between Augustus, linked to Tityrus, and Ovid, compared implicitly with the exiled Meliboeus.

⁵² For Ovid's use of the Aeneid in the Metamorphoses, see e.g. Galinsky, op. cit. (n. 5), 217–51; Solodow, J. B., The World of Ovid's Metamorphoses (1988), 110–56.Google Scholar

⁵³ cf. Barchiesi, op. cit. (n. 2), 18: ‘L'Eneide prediletta dal principe e appropriata dal discorso augusteo (tuae) ha portato fortuna a Virgilio, felix perché opposto a Ovidio che scrive tristia per colpa dell' Ars amandi; eppure anche lì c' è una storia d'amore di un certo tipo. La legge della pertinenza, il decorum, è stata violata perché l'epica, fattasi impura, potesse aprirsi a un tema erotico che dona successo e popolarità a Virgilio.’ Cutolo, op. cit. (n. 19), 283, sees in this passage of the Tristia an allusion to Horace, , Epist. 2.1.245–7Google Scholar: ‘at neque dedecorant tua de se iudicia atque / munera, quae multa dantis cum laude tulerunt, / dilecti tibi Vergilius Variusque poetae’.

⁵⁴ For reworkings of arma uirumque in Ovid's works, see now Barchiesi, op. cit. (n. 2), 5–14, 25–6. Ovid's ‘toros’ recalls Aeneas speaking his narrative of his wanderings ‘toro … ab alto’ (Virgil, , Aen. 2.2Google Scholar), and Dido's dying words, spoken ‘os impressa toro’ (Aen. 459).

⁵⁵ For the popularity of Dido, see Galinsky, op. cit. (n. 5), 248, who also notes that Ovid treats the episode only briefly in the Metamorphoses.

⁵⁶ Note that Augustus himself is recorded as having heard readings of portions of the Aeneid (Vita Donati 32); cf. Ovid's own suggestion (Trist. 2.557–8, discussed above) that Augustus arrange to have the pan-egyric passage at the end of the Metamorphoses read to him as an excerpt. Note also Masters, J., Poetry and Civil War in Lucan's Bellum Civile (1992), 222Google Scholar, on Vacca's account of Lucan's recitation to Nero: ‘Lucan may indeed have published or recited three books in advance of the rest, precisely because he was conscious of the fact that Virgil had done similarly with the Aeneid.’ On Augustus' literary tastes, see Suetonius, Div. Aug. 89; Syme, R., The Roman Revolution (1939), 460, 484–5Google Scholar; Galinsky, op. cit. (n. 5), 211–12; Williams, op. cit. (n. 2), 181.

⁵⁷ On the audience of Tristia 2 itself, see Wiedemann, T., ‘The political background to Ovid's Tristia 2’, CQ 25 (1975), 264–71CrossRefGoogle Scholar, at 271, who makes the important point that Tristia 2's readership extends beyond the notional addressee of the poem, Augustus, and suggests that the poem ‘was not intended for Augustus' eyes at all; it was meant to influence the circle of educated Roman aristocrats to whom Ovid's other poems from Tomoi were addressed, and Ovid hoped that they would be the ones who, recognizing the absurdity of Augustus' grounds for exiling Ovid, would do their best to see that he was recalled’.

⁵⁸ For discussion of the idea that a text can simultaneously be read in two ways, see Demetrius of Phalerum, De elocutione 291, with Ahl, op. cit. (n. 2), 195; Otis, op. cit. (n. 5), 305; Barchiesi, op. cit. (n. 2), 21; Williams, op. cit. (n. 2), 157–8; Hinds op. cit. (n. 10), 25: ‘If he [Ovid] was subversive in his writing (as I believe he was), how could he possibly proceed but by indirection and nuance? In any but the most powerful or the most reckless of Romans, publicly voiced anti-Augustanism must needs be a rhetoric of ambiguity and innuendo. Every passage ever written by Ovid about Augustus admits of a non-subversive reading: but that is not in itself a refutation of Ovidian subversion.’

⁵⁹ Note that Ovid even countenances the possibility that the Tristia themselves may give offence to a reader at Trist. 1.1.22 (addressed to his book): ‘ne, quae non opus est, forte loquare, caue.’ Ovid's despatch of his book to Rome without him in Trist. 1.1 can be compared with the discussion in Plato, Phaedrus 275d of the defencelessness of the written word. Note also the language of control used by Eco, op. cit. (n. 1), 83, in discussing his failure to prevent interpretations of the title of his novel, Foucault's Pendulum, as a reference to Michel Foucault, despite the fact that the pendulum of the title was the work of Léon Foucault: ‘But the pendulum invented by Léon was the hero of my story and I could not change the title: thus I hoped that my Model Reader would not try to make a superficial connection with Michel. I was to be disappointed; many smart readers did so. The text is there, and perhaps they are right: maybe I am responsible for a superfical joke; maybe the joke is not that superficial. I do not know. The whole affair is by now out of my control.’

⁶⁰ Trist. 3.7.47–8: ‘ingenio tamen ipse meo comitorque fruorque: / Caesar in hoc potuit iuris habere nihil.’ Compare Tacitus' comments on the folly of imperial book-burning (Ag. 2.1–2, Ann. 4.35.5). For censorship during Augustus' reign, see Seneca, Contr. 10 pr. 4–5 (on T. Labienus), Dio Cassius 56.27.1 (anonymous pamphlets), Suetonius, Caligula 16, Tacitus, , Ann. 1.72Google Scholar (Cassius Severus), Suetonius, Div. Aug. 36 (ending of publication of the acta senatus). Though Syme, op. cit. (n. 56), 486–7, refers to ‘stern measures of repression against noxious literature’ towards the end of Augustus' reign, there is perhaps a danger of overestimating the nature and extent both of such literary and intellectual opposition to Augustus and of the Emperor's responses; see now Raaflaub, K. A. and Samons, L. J. II, ‘Opposition to Augustus’, in Raaflaub, K. A. and Toher, M. (eds), Between Republic and Empire: Interpretations of Augustus and his Principate (1990), 417–54Google Scholar, at 436–47, who rightly draw attention to the ancient evidence for Augustus' lenience in such matters. As Syme, ibid., notes, the Ars Amatoria was not suppressed. See also Citroni, M., Poesia e lettori in Roma antica (1995), 440–2, 431–5.Google Scholar

⁶¹ One might, however, contrast the opening poem of the third book of the Tristia, where his book describes its failure to gain admission to the temple libraries of Rome (Trist. 3.1.59–80). The book, however, ends up in private hands: perhaps a more dangerous form of reception? Cf. Kermode, F., ‘Freedom and interpretation’, in Johnson, B. (ed.), Freedom and Interpretation: the Oxford Amnesty Lectures 1992 (1993), 46–68Google Scholar, at 46: ‘There is obviously a close relation between liberty of interpretation and political liberty in general.’

⁶² Ancient rhetorical theory (in Demetrius of Phalerum and in Quintilian) of ambiguity as a mode to be used when addressing tyrants is usefully discussed by Ahl, op. cit. (n. 2), 186–92; Williams, op. cit. (n. 2), 159–60. See also Dewar, M., ‘Laying it on with a trowel: the proem to Lucan and related texts’, CQ 44 (1994), 199–211CrossRefGoogle Scholar, for a sceptical response to ironic readings of Lucan's opening address to Nero.

⁶³ See further S. G. Owen, P. Ovidi Nasonis.Tristium Liber Secundus (1924), 55; Wiedemann, op. cit. (n. 57)), 271; Nugent, op. cit. (n. 2), 243–4. Note however the verdict of Syme, R., History in Ovid (1978), 222Google Scholar: ‘… a fine piece of work, lucid, coherent, and forceful, worthy of a great orator or a good historian.’ For analysis of the rhetorical structure of the poem see Owen, op. cit., 48–54; Focardi, G., ‘Difesa, preghiera, ironia nel II libro dei Tristia di Ovidio’, SIFC 47 (1975), 86–129Google Scholar, at 87–105; Stabryla, op. cit. (n. 31), 471.

⁶⁴ Williams, op. cit. (n. 2), 168, sees the shift between author and reader in terms of responsibility, rather than power.