1. For a general discussion of this literary background, see e.g. Nagle, B.R., The Poetics of Exile, Latomus 170 (Brussels 1980), 32-70Google Scholar, who also discusses the relations between Ovid’s exile poetry and his earlier erotic poetry books. In this paper I focus primarily on the generic connections between Ovid’s epistolary exile poetry and his Heroides, a topic not widely studied.

2. In G.D. Williams’s words, ‘Ovid exploits the creative possibility of exile to pursue a form of psychological investigation which he had previously conducted through his literary characters.’ J. McKeown points out to me that the Heroides themselves are already adapting the conventions of ‘personal’ love elegy, i.e. the genre in which Ovid purports to be writing about himself, so we should be wary of assuming a straightforward chronological development from Heroides to Tristia, literary to ‘personal’.

3. The double epistles (Her. 16-21) will, however, enter my discussion at points.

4. Ovid does refer to carmen et error (Tr. 2.207), but does not equate these with actual ‘crimes’.

5. These adjectives are culled from the works of Conte, G.B., Latin Literature: A History, tr. J.B. Solodow (Baltimore 1994), 350Google Scholar; Kenney, E.J., Ovid Heroides XVI-XXI (Cambridge 1996), 3Google Scholar; Anderson, W.S., ‘The Heroides’, in J.W. Binns (ed.), Ovid (London 1973), 49-83Google Scholar, esp. 66f.; Hine, D. (tr.), Ovid’s Heroines (New Haven 1991), ixGoogle Scholar; Verducci, F., Ovid’s Toyshop of the Heart (Princeton 1985), 4Google Scholar.

6. While I refer occasionally to passages from the Epistulae ex Ponto, the main focus of this paper will be relations between the Heroides and the Tristia.

7. This argument was presented by G.D. Williams in his original panel paper.

8. Kirfel, E.-A., ‘Untersuchungen zur Briefform der Heroides Ovids’, Nodes Romanae 11 (1969), 28fGoogle Scholar. and 33f., traces conversational style in the exilic works through word searches: e.g. precor, rogas, scis, credo. Knox, P., Ovid Heroides (Cambridge 1995), 25fGoogle Scholar., also mentions ‘word choice and syntax intended to suggest everyday speech’.

9. Kirfel (n.8 above), 25: ‘Diese Hinwendung zum Adresssaten ist der wichtigste Unterschied zwischen der reinen Elegie und der Briefelegie.’ See also ibid. 29: ‘Die Form des Briefes hängt einzig und allein von den Anfangs- und Schlussformeln ab.’ This issue is also discussed intelligently by Davisson, M.H.T., ‘Tristia 5.13 and Ovid’s Use of Epistolary Form and Content’, CJ 80 (1985), 238-46Google Scholar, who acknowledges that Ovid’s letters had no real practical function (i.e. they were not sent as letters) but nevertheless retained many aspects of epistolary style and content. I discuss epistolary form more thoroughly in my current project on the subject.

10. See Rahn, H., ‘Ovids elegische Epistel’, A & A (1958), 105-20Google Scholar, esp. 105.

11. See Nicolai, W., ‘Phantasie und Wirklichkeit bei Ovid’, A & A (1973), 107-16Google Scholar, esp. 108.

12. This trend has been supported by the work of such scholars as A. Barchiesi, S. Hinds, E.J. Kenney and P. Knox. New commentaries on individual Heroides have also recently appeared, by S. Casali, T. Heinze and G. Rosati.

13. See, for example, Showerman, G., Ovid Vol I: Amores and Heroides, 2nd ed., rev. G.P. Goold (Cambridge MA 1977), 8Google Scholar: ‘The Heroides are not a work of the highest order of genius. Their language, nearly always artificial, frequently rhetorical, and often diffuse, is the same throughout—whether from the lips of barbarian Medea or Sappho the poetess.’ Jacobson, H., Ovid’s Heroides (Princeton 1974), 74Google Scholar, labels them ‘repetitive’. Even Kenney (n.5 above, 1) states regretfully that it ‘is difficult to rescue them, especially if they are read sequentially, from the charge of monotony’. But to accuse Ovid’s heroines of continuous harping on their misfortunes is to ignore the constant variations in pitch, tone, motivation, justification, and levels of paranoia that make each case unique.

14. Thus W.S. Anderson (n.5 above, 81) sustains the idea that the exile poems reflect a narrowing of options, a reduction in quality: ‘Fate and unappreciative Augustus would reduce him to the despair of the women in the Heroides I-XV which he had once imagined so brilliantly.’ B.R. Nagle (n.l above, 132f.) concludes that ‘monotony is the unavoidable product of constantly unpleasant surroundings and a constant desire to improve them’, but she does acknowledge that it is in Ovid’s best interest to portray his life in exile as being as miserable as possible. See also Nicolai (n.11 above, 108f.), who implies that Ovid’s personal situation ‘forces’ him to write such dismal exile poetry. Cf., however, Davisson (n.9 above), 246: ‘Ovid’s integration of epistolary form and content was neither an inevitable result of his situation nor merely a practical strategy. Rather, he was attracted by the artistic challenge of creating a form which was both poetic and epistolary.’

15. Williams, G.D., Banished Voices (Cambridge 1994Google Scholar).

16. G.D. Williams points out to me the comparable situation of Seneca in exile, who presents one self-image (deferent and desperate to win Claudius’ pity) in the Consolatio ad Polybium, and a very different one (cheerful and optimistic) to his mother in the ad Helviam.

17. Tears on the page are, of course, a conventional epistolary trope: in addition to the examples to follow, see e.g. Ovid Ex Pont. 1.9.1f.; Prop. 4.3.3f.; Cic. ad fam. 14.2.1 or ad Q. fr. 1.3.3. For a discussion of the heroines’ carefully designed self-presentation, see Flaherty, S., The Rhetoric of Female Self-Destruction: A Study of Homer, Euripides, and Ovid (Diss. Yale Univ., 1994Google Scholar).

18. The debate on the authenticity of Sappho’s epistle continues to rage: for the most recent opinions, see P. Knox (n.8 above), 12-14 and bibliography.

19. The personified book-letter is adapted from Horace; see Rahn (n.10 above), 108-10.

20. This point was suggested to me by G.D. Williams.

21. Dido also sits with drawn sword, and predicts that blood will soon mingle with her tears (Her. 7.184-86), although she does not specifically say that the drops will fall on the page.

22. Certain letters set up contexts that the reader mistrusts: thus in Tr. 1.11, we do not really believe that Ovid wrote the actual letter in the midst of a storm (cf. e.g. Petr. Sat. 115: Eumolpus writes a poem at sea), although we may be convinced by his basic sentiments expressed therein; similarly in Her. 10 Ariadne writes her letter on a deserted island, with no messenger present to deliver it.

23. In the spirit of the conceit, Cydippe begins her letter with a disclaimer: quam tibi nunc gradient uix haec rescribere quamquelpallida uix cubito membra leuare putas? (‘How thin and delicate do you think I am, with difficulty writing this answer to you? How pale as I scarcely prop my body up with my elbow?’, 21.15f); but she does somehow manage to continue for roughly another 230 lines.

24. The phrasing of the passage to follow recalls similar outbursts of self-pity in the Heroides: see variations on me miseram, a demens, or ei mihi in Tr. 5.2.39, 5.10.51; Her. 2.103, 3.61, 5.149, 7.98; see also the common complaint that the unfeeling and unresponsive interlocutor was born of flint, suckled on wild beasts: Tr. 3.11.3f.; Her. 7.37, 10.132. In this passage, uereor leaves a certain amount of ambiguity as to whether Augustus' house is to be respected or simply feared.

25. Sappho is a particularly tempting parallel because she is the only heroine who is a poet in her own right, but Ovid’s situation of exile and abandonment suggests parallels with the shared experiences of all the heroines, as I argue throughout this paper.

26. See e.g. Tr. 1.9.59-65; 2 passim; 3.5.43-54; 4.4.35-54.

27. On abandonment as a specifically ‘female’ condition, see Lipking, L., ‘Aristotle’s Sister: A Poetics of Abandonment’, Critical Inquiry 10 (1983), 61-81CrossRefGoogle Scholar, and his book on the same subject, Abandoned Women and Poetic Tradition (Chicago 1988Google Scholar).

28. Note the potentially subversive erotic undertones of all these nouns for Ovid’s ‘condition’. Williams (n.15 above, 84f.) discusses Ovid’s poetic madness as part of the general unhappiness of his exile, but also as a mental state traditionally associated with poetic inspiration. He sees an erotic presence, however, in Tr. 5.13 (ibid. 124ff.), where Ovid transfers symptoms of love-sickness to his exilic state. He draws some important conclusions from these observations (ibid. 127): ‘And yet the allusive presence of so many erotic motifs in Tr. 5.13, with no attempt made to conceal their literary origins in their new exilic context, suggests that Ovid’s break with his erotic past is not quite as complete as he would have us believe in Tr. 5.1 and elsewhere.’

29. This is the interpretation accepted by most critics: see e.g. Nagle (n.l above), 71; Williams (n.15 above), 89. See the similar approach to letter-writing as a cure for grief in Cicero, e.g. ad Att. 9.10.1; 12.39.2.

30. His stated inability to stop himself is yet another poetic stance, of course. This kind of pathology also appears in the Ibis, that ‘exercise in compulsive frustration’; see Williams, G.D., ‘On Ovid’s Ibis: A Poem in Context’, PCPS n.s. 38 (1992), 171-89Google Scholar, and also PCPS suppl. vol. 19.

31. See Tr. 3.14.43-47; 5.7.53-58; 5.12.55-58.

32. It is a telling moment when Phyllis tries to find something ‘real’ or stable to swear by: she tries to swear by Demophoon’s grandfather, but then catches herself short and exclaims nisi fictus et ille est (‘unless he too is a fiction’, Her. 2.37).

33. This is discussed briefly in Rahn (n.10 above), 119. But Rahn goes on to argue that Ovid’s style shifts gradually from a distanced or ironic treatment of themes in earlier works (e.g. Heroides) to a more direct artistic self-portrayal in the exilic works; I disagree with this interpretation, as I will make clear later.

34. Further references to the wandering letter: Tr. 3.7 (to Perilla); Ex Pont. 1.1; 3.9; 4.5. For a discussion, see Nagle (n. 1 above), 82-90.

35. Again, in the case of the many different ways in which Ovid portrays his books as forlorn and powerless, Ovid projects a miserable situation but offers elaborate poetic variations in his descriptions of that desolation. As G.D. Williams argues (n.15 above), there is a tension between Ovid’s claim of declining poetic powers in bad circumstances and the dynamism of his poetic imagery.

36. This point was suggested to me by Sara Lindheim.

37. The following line begins (5.2.3) pone metum (‘lay aside your fear’), a neat allusion to Acontius’ reassurance to Cydippe in Her. 20.1: pone metum!

38. Rahn (n.10 above), 112 calls this poem ‘eine Art Testament’, and reminds us that Propertius also played with such type scenes (1.17; 1.19; 2.13; 3.6).

39. The next logical question would be whether the humour generated by the Petronian treatment is at all present in Ovid’s rendition. The answer will vary with the individual reader.

40. On this point see Rahn (n.10 above), 112-14.

41. Equally unrelenting is Hypsipyle’s lengthy vision of revenge on her rival Medea (Her. 6.147-51). On a lighter note, Ovid creates a dialogue in Tr. 5.1 with a reader who objects to the bad quality of his verse.

42. Note the flexibility (self-contradictory nature?) of the argument: his erotic poetry does not accurately reflect his chaste life, but here his sad life in exile is indeed reflected in his equally sad poetry, see also Ex Pont. 3.9.49: musa mea est index nimium quoque uera malorum (‘My Muse is only too true a witness of my sufferings’). I will return to this issue towards the end of this paper, but it is worthy saying here that this line of flexible argument certainly implies a parallel flexibility in the self-presentation of the writer. Ovid manipulates his persona to suit his literary needs at any given moment.

43. Nagle (n.l above), 28-32, offers some interesting observations on Ovid’s use of language which equates exile with civic, physical, and also poetic death, i.e. the end of his career as a major poet. Ovid stages a funeral both for himself (Tr. 1.3) and for his poetry (Tr. 1.7: the Metamorphoses), which he wishes to suppress on the model of Vergil’s Aeneid.

44. On this subject, see in particular Williams (n.l5 above), 50-99, on e.g. Tr. 3.14.37-40; 5.12.53-58. As he says, ‘it is in a context which, paradoxically, laments the decline of his talent that Ovid’s creative facility and allusive technique are most fully active’ (57).

45. This passage is discussed in Williams (n.15 above), 60; see also S. Hinds, ‘Booking the Return Trip: Ovid and Tristia 1’, PCPS 31 (1985), 13-32, esp. 14.

46. On such conventions see Kirfel (n.8 above) passim, but esp. 29-32, and Rahn (n.10 above), 112f.; see also Williams (n.15 above), 122f., who notes, on the conventional opening greeting of salutem mittere, that ‘the epistolary format of Tr. 5.13 lends itself to creative exploitation of the language of sickness and health’. On salutem mittere vs. dicere as a way to emphasise the writer’s distance from his addressee, see Davisson (n.9 above), 240. For further play with salus in the opening lines, see e.g. Her. 4 (Phaedra), and 16 (Paris), discussed in Williams (1994), 124f., and Ex Pont. 1.10. Epistolary uale occurs in the last lines of Her. 9.168; 20.242; 21.248. The word occurs internally at moments of emotional leavetaking: Her. 5.51; 12.56; 13.14; 15.100; there is also an internal uale in Tr. 1.8.26.

47. This is discussed briefly in Kirfel (n.8 above), 27f.

48. On this connection between letters and amicitia, see Williams (n.15 above), 116-28.

49. Rahn (n.10 above, 111f.) compares Ovid’s behaviour in this farewell scene (Tr. 1.3) to that of heroines in the Heroides: ‘Hatte in den Heroides eine Laodamia, Ariadne, Penelope geklagt wie eine vom Leid getroffene Romerin, so klagt nun ein vom Leid getroffener Römer wie die Heroen und Heroinen der troianischen Zeit.’

50. In Tr. 5.5.43, Ovid writes that the day of his wife’s birth brought forth a character comparable in chastity and prudence to that of illis heroisin (Salmasius’ correction of heroibus).

51. Kenney, E.J., ‘The Poetry of Ovid’s Exile’, PCPS n.s.l1 (1965), 37-49Google Scholar, reminds us that Tr. 1.6 is not a personal poem, and that we should not blame Ovid for what we perceive to be a boasting tone, praising his wife as a reflection of his own poetic reputation. He also comments on the choice of comparanda (40): ‘In comparing his wife to Andromache or Penelope Ovid was not paying her an empty and formal compliment but bestowing on her the highest praise that he, a poet, could conceive.’

52. Cf. Kenney (n.51 above), 39, who puts it slightly differently: ‘The exempla are chosen because the men in question were famous poets, for the women were almost certainly not their wives but their mistresses.’ I think the point here is not marital status, but rather historical vs. fictional identity.

53. There is some confusion on the ordering of these lines; I adopt the transposition given by modern commentators, e.g. Luck, G., P. Ovidius Naso Tristia (Heidelberg 1967), 52Google Scholar. On the first book of the Tristia in general, see also Hinds (n.45 above).

54. His affiliation to Odysseus is facilitated by his favourite comparison of Penelope to his wife; see Tr. 5.5.3-4, where he imagines himself as Odysseus, far away at the edges of the earth on his wife’s birthday.

55. Besslich, S., ‘Ovids Winter in Tomis. Zu trist. 111 10’, Gymnasium 79 (1972), 190Google Scholar, asserts another view: ‘Wenn die Spätdatierung dieser Briefe richtig ist, so sind sie die letzten erotischen Gedichte vor der Verbannung und damit besonders geeignet, wenn der Dichter Beispiele sucht, um glaubhaft zu machen, dass von ihm fernerhin nichts dergleichen mehr zu erwarten ist.’

56. I quote Maria Wyke, ‘Taking the Woman’s Part: Engendering Roman Love Elegy’, in Boyle, A.J. (ed.), Roman Literature and Ideology (Bendigo 1995), 110-28Google Scholar, esp. 116. See also Edwards, C., The Politics of Immorality in Ancient Rome (Cambridge 1993), 63-97CrossRefGoogle Scholar.

57. This line of thought was suggested to me by Sara Lindheim.

58. On this see Fränkel, H., A Poet Between Two Worlds (Berkeley 1945), 125ffGoogle Scholar. Ovid also refers to Acontius and Leander in A.A. 1.455-58, 2.249f. For a detailed comparison of Tr. 3.10 with Her. 18 (Leander), see Besslich (n.55 above), 186-89.

59. See Mariotti, S., ‘La Carriera Poetica di Ovidio’, Belfagor 12 (1957), 609-35Google Scholar, esp. 633: ‘Ovidio diventa un “personaggio” della propria poesia come le dolenti eroine delle epistole amorose.’ Similarly Williams (n.15 above), 49: ‘An exile who creates an “unreal” picture of his circumstances in exile by manipulating his “facts” to creative advantage,’ or, on the Ars (ibid. 170): ‘As soon as Ovid argues that the material of the Ars is not a valid depiction of his personal character, the extent to which any of his material can be taken as truly reflecting a real inner self is immediately open to question.’ On this theme, see also Tr. 2.353-58.

60. Ovid never has a heroine resort to this threat, i.e., that she has the power through writing to tarnish the reputation of her beloved. In this way, although otherwise their situations are so similar, he distinguishes his own poetic voice in the Tristia from those of his female characters in the Heroides.

61. Rahn (n.10 above, 115) reads these great names from classical mythology as vivid embodiments of human emotional situations and relationships.

62. Rahn (n.10 above, 116-18) argues that Odysseus is different in nature from Ovid’s other models of friendship and courage, and actually resembles a ‘Leitmotiv’ in Ovid’s exile poetry. Williams (n.15 above, 107-11) also discusses this syncrisis and its relationship to the first part of the poem on friendship. Odysseus appears as a character in love elegy also in Prop. 4.8 and Tib. 1.3.

63. He is, of course, already following his own advice and, as the best of the docti poetae, writing his own mala.

64. Williams (n.15 above, 113) notes the programmatic language in these passages about hardness (epic) and softness (elegy), war and love.

65. Williams (n.15 above), 108.

66. Williams (n.15 above), 109.

67. Rahn (n.10 above, 145) argues for a logical artistic progression from the earlier epistolary experiments to the later, stating that the Tristia and Epistulae ex Ponto are the telos of the genre begun with the Heroides. While I agree that there is no huge gap between the pre- and post-exilic works, and that many of the same attitudes are present in both works, I hesitate to interpret them ideologically, for this again would automatically and unfairly denigrate the earlier material.

68. In labelling these stories ‘literary’ instead of ‘mythical’, I accept the argument of P. Knox (n.8 above, 18-25), who suggests a literary point of reference behind each of the single Heroides.

69. Rahn (n.10 above, 144f.) puts it well: ‘1st er so heillos routiniert und rhetorisiert, dass alles nur Mache, unecht, Gefühlssurrogat, Kitsch ist?’ I particularly like the use of the word ‘Kitsch’ to describe some of Ovid’s practice.

70. On issues of truth and lying in ancient poetry, see Puelma, M., ‘Der Dichter und die Wahrheit in der griechischen Poetik von Homer bis Aristoteles’, MH 46 (1989), 65-100Google Scholar, and Pratt, L. H., Lying and Poetry from Homer to Pindar (Ann Arbor 1993Google Scholar).

71. This paragraph is based on a larger project on epistolary fictions which I hope to publish in book form. Epistolary technique in the context of Ovid’s Heroides is also discussed by Kauffman, L.S., Discourses of Desire: Gender, Genre, and Epistolary Fiction (Ithaca 1986), 25Google Scholar.

72. I thank S. Flaherty, S. Lindheim, J. McKeown and G.D. Williams for their very helpful comments and criticisms. I am also grateful to A. Walker for the original invitation to contribute to this collection.