¹ On these considerations see in general Mylne, Vivienne, The eighteenth-century French novel: techniques of illusion (2nd ed., Cambridge, 1981), pp. 149–55, 234–6Google Scholar. Most recent research on the literary potentialities specific to the epistolary form has concentrated on the epistolary novel and on Laclos in particular; cf. Rousset, Jean, ‘Une Forme littéraire: le roman par lettres’ in Forme et signification: essais sur les structures littéraires de Corneille à Claudel (Paris, 1962), pp. 65–108Google Scholar, Jost, Francois, ‘Le Roman épistolaire et la technique narrative au XVIIIe siècle’, Comparative Literature Studies 3 (1966), 397–427Google Scholar, Todorov, Tzvetan, Littérature et signification (Paris, 1967)Google Scholar, Altman, Janet Gurkin, Epistolarity: approaches to a form (Columbus, Ohio, 1982Google Scholar).

² Cf. Her. 16. 129 f., 217 ff., 275 f., 299 f., 17. 7 ff., 159 f.

³ Les Liaisons Dangereuses provides a close parallel to this successful transgression of the logical epistolary convention of separation of writer and addressee. The arch-seducer, the Vicomte de Valmont, does not address a letter to the virtuous object of his attentions, Madame de Tourvel, a fellow-guest staying under the same roof in the absence of her husband, until (Letter XXIV) she has rejected in horror his first approach to her, an event he describes in Letter XXIII. He thus loses the social intimacy he previously enjoyed and must henceforth press his suit by letter. Her reluctant willingness to reply (Letter XXVI), like that of Helen, marks the beginning of her downfall. A coincidence? I know of no treatment of Ovid's possible influence on Laclos (Laurent Versini's otherwise encyclopaedic Laclos et la tradition (Paris, 1968), pp. 242–6Google Scholar, mentions Ovid only in the most vague terms), but Valmont's description of his attempted seduction presents us with an unmistakably Ovidian scene and sentiment; cf. Letter XXIII: ‘Aussi, en descendant de voiture, elle [Mme de Rosemonde] passa dans son appartement, et nous laissa tête ả tête, ma belle et moi, dans un salon mal éclairé; obscurité douce, qui enhardit l'amour timide', and Ov. Am. 1. 5. 3 ff., esp. 7 f. ‘ilia verecundis lux est praebenda puellis|qua timidus latebras speret habere pudor’.

⁴ Cf. also the bad omens mentioned at Her. 18. 81 f., 141 f., 19. 195 ff. The outcome of the story would have been familiar to Ovid's readers at least from Verg. G. 3. 258 ff., and probably also from Hellenistic antecedents; cf. Page, D. L., Select Papyri, III (London, 1941), pp. 512–4Google Scholar.

⁵ Cf. Anderson, W. S., ‘The Heroides’ in (ed.) Binns, J. W., Ovid (London, 1973), pp. 70–4Google Scholar. We are reading the letter, or, to be more accurate, we are peeking over the woman's shoulder as she is writing it. Sometimes the fiction is transparent, or even breaks down. Penelope's letter to Ulysses is apparently one of those she gives to almost every passing sailor, in the hope that it will reach her husband; but Ulysses never receives it.

⁶ Cf. Her. 4. 7 ff. ‘ter tecum conata loqui ter inutilis haesit|lingua, ter in primo destitit ore sonus.|qua licet et sequitur, pudor est miscendus amori;|dicere quae puduit, scribere iussit amor.’ Whether this is Ovid's idea or is derived from one of his numerous possible sources for the story of Phaedra remains uncertain; cf. Jacobson, H., Ovid's Heroides (Princeton, 1974), pp. 142 ff., esp. 146 n. 11Google Scholar.

⁷ Commentary on Her. 7 init.

⁸ Paradoxically, the amusement that such a flirtation with verisimilitude arouses in exotic contexts seems to aid the suspension of disbelief by emphasizing the playful, literary quality of the text. Zilia, , the heroine of Mme de Grafigny's Lettres d'une Péruvienne (1746)Google Scholar, abducted from Peru, tells how she recorded her first seventeen letters to her betrothed Aza, a Peruvian prince, on quipos, knotted cords of various colours used by the Peruvians for sending messages, until her French was good enough to translate them; cf. Mylne, op. cit. n. 1 above, p. 154.

⁹ Cf. Ars 3. 345 ‘vel tibi composita cantetur epistula voce’; also Am. 2. 18. 21 ff., and Kirfel, E.-A., Untersuchungen zur Briefform der Heroides Ovids (Noctes Romanae Band 11, Bern/Stuttgart, 1969), esp. pp. 11–36Google Scholar.

¹⁰ Op. cit. n. 5 above, p. 66.

¹¹ All of Ulysses' men were, of course, killed at one stage or another on the journey home.

¹² ‘nos Pylon, antiqui Neleia Nestoris arva,|misimus; incerta est fama remissa Pylo.|misimus et Sparten; Sparte quoque nescia veri.’ If these lines refer to the mission of Telemachus, they contradict not only Homer but 99–100 ‘ille per insidias paene est mihi nuper ademptus,|dum parat invitis omnibus ire Pylon.’ However, it is surely to complicate matters unduly to postulate a separate mission to these same places not attested in Homer.

¹³ Notably the infinitive of purpose quaerere after misso in 37; 39–40, linked in sense to 37–8, contain further difficulties. Excision is the path of least resistance, but leaves an impossible transition between 36 and 41. Cf. the next note.

¹⁴ Indeed against her wishes, as she says to him on his return (⋯μεû ⋯έκητι, Od. 17. 43). The editors have passed on to me the suggestion of reading iusso for misso in 37, comparing the vv. ll. at Juv. 3. 78. This would immediately account for the infinitive quaerere. I would hesitate to alter the transmitted text on two grounds: (i) misso gains some support from Penelope's insistence (cf. the emphatic repetition of misimus at the beginning of 64 and 65) that it was she who initiated the journey of Telemachus. (ii) iusso immediately reminds us that Telemachus was indeed ordered to go to see Nestor and Menelaus by Athene disguised as Mentes (Od. 1. 284 f.), but when Penelope learns of his mission from Medon (4. 701 f.), she is puzzled at what impelled him to go (707 ff.), and she is nowhere told, not even by Telemachus on his return (17. 108 ff.), of the divine instigation of his journey. If the Odyssey functions as an ‘objective’ account of the events which lie behind Her. 1 (the central interpretative problem of the poem which I discuss below), Ovid's Penelope, however deviously she may distort those events for her own purposes, strictly speaking should not have access to information which the Odyssey clearly indicates she has not been given.

¹⁵ The transmitted text Antilochum…ab Hectore victum was questioned as early as 1489, by Politian (Miscellanea, i. 76, a reference I owe to the editors), who suggested either Amphimachum (cf. Il. 13. 185 ff.) for Antilochum (necessitating a similar change in line 16), or Memnone (cf. n. 16 below). Housman (CR 11 [1897], 102 f. = Classical Papers, i. 381 f.) suggested ab hoste revictum. He adduced the two main arguments in favour of emendation: (i) the deviation from the canonical Homeric account; this is part of a larger question bearing on the poem as a whole which I shall deal with below; (ii) the inelegance of the repetition of Hector's name directly after ‘nomine in Hectoreo pallida semper eram’ (14). It could be argued, however, that ‘sive…sive…’ (15–18), with the death of Patroclus at the hands of Hector forming the second instance, should form two illustrations of Penelope's fear of Hector before she goes on to include all Greek deaths at the hands of Trojans as a cause of fear for her husband's safety (19–22).

¹⁶ 4. 187 f. This is the normal version; cf. Pind. Pyth. 6. 28 ff., Dio Chrys. 11. 352, Diet. 4. 6, Quint. Smyrn. 2. 244 f. I cannot satisfactorily account for the deviation in Ovid. Hector is found as the killer of Antilochus elsewhere only in Hyg. Fab. 113, but nothing can be inferred from this, as in Fab. 112 his killer is said to have been Memnon. Textual corruption in Hyginus has been suspected; cf. Housman, loc. cit. n. 15 above. Another variant, in which Paris killed Antilochus, is recorded in Dares of Phrygia, De excidio Troiae historia 34.

¹⁷ Cf. Her. 3. 30–6 and Il. 9. 122–30; Her. 3. 37–8 and Il. 9. 144–7; Her. 3. 47–50 and Il. 19. 291–6.

¹⁸ Unless Jacobson (op. cit. n. 6 above, p. 36 n. 56) is correct in suggesting that Ovid meant us to understand that Achilles had told Briseis of the incident.

¹⁹ By contrast, Achilles is represented as addressing her at this point in Ovid (Her. 3. 54).

²⁰ It is possible that foreshadowing might help to account for the deviation from Homer in Penelope's description of the night attack in Her. 1. 41–4 ‘ausus es, o nimium nimiumque oblite tuorum,|Thracia nocturno tangere castra dolo,|totque simul mactare viros, adiutus ab uno!|at bene cautus eras et memor ante mei!’ Is Ovid inviting us to look ahead to events Penelope necessarily cannot know of at the time she is writing this letter, when against large odds Ulysses will play the major part in killing a great number of men at one and the same time, the suitors, and will keep Penelope totally in the dark as to what is to happen? The only snag is ‘adiutus ab uno’ (43). In Od. 22, Odysseus is helped in the slaying of the suitors not only by Telemachus but also by the swineherd Eumaeus, the cowherd Philoetius and, at a later stage, Athene disguised as Mentor, as seems to be foreshadowed in 103 f. ‘hac faciunt custosque boum longaevaque nutrix,|tertius immundae cura fidelis harae’. But if it is not to be obtrusive and spoil the epistolary illusion, foreshadowing must be partial and allusive. Foreshadowing was a recognised literary technique in the ancient world; cf. Duckworth, G. E., ‘Пροαναϕώνησις in the Scholia to Homer’, AJPh 52 (1931), 320–38Google Scholar, esp. 326 on foreshadowing by analogy and 328 on foreshadowing of action outside the narrative of the epic itself.

²¹ Op. cit. n. 6 above, pp. 243–76. Jacobson traces in detail the freedom, not to say irreverence, with which the character of Penelope was treated by post-Homeric writers.

²² That the epistle was the form which above all others revealed the character of its writer was recognised by Demetrius, On Style 227: πλεῖστον δ⋯ ⋯χέτω τό ἠθικòν ή ⋯πιστολή, ⋯σπερ κα⋯ ό διάλογος σχεδόν γ⋯ρ ε⋯κόνα ἔκαστος τ⋯ς έαντο⋯ ψυχ⋯ς γράϕει τ⋯ν ⋯πιδτολ⋯ν. κα⋯ ἔστι μ⋯ν κα⋯ ⋯ξ ἄλλου λόγου παντ⋯ς ἰδεῖν τ⋯ ἧθος τοû γράϕοντος, ⋯ξ οὺδεν⋯ς δ⋯ οὗτως, ώς ⋯πιστολἦς.

²³ Blatant inconsistency within a single poem occurs elsewhere in the Heroides and must, I think, be seen as a function of characterization. Jacobson (op. cit. n. 6 above, p. 148) draws attention to Her. 4, where Phaedra describes herself as a married woman (17 ff.), yet immediately afterwards presents herself to Hippolytus through images which suggest virginity (21 ff.), a manifestation of her desire to see herself in relation to Hippolytus as a puella (cf. 2), and not as the noverca (cf. 129) she really is. Cf. also Pearson, C. S., ‘Simile and Imagery in Ovid Heroides 4 and 5’, ICS 5 (1980), 110 ff., esp. 113Google Scholar.

²⁴ Cf. Schol, . Od. 4. 69, 113, 8. 43, 489Google Scholar, Eust. 1487. 15 ff., 1489. 35 ff. I owe the references in this and the following note to the kindness of Dr Richardson, N. J.. On Aristotle's influence on the scholia see his article ‘Literary Criticism in the exegetical Scholia to the Iliad: a Sketch’, CQ n.s. 30 (1980), 265–87CrossRefGoogle Scholar.

²⁵ Cf. Schol, . Vd. Od. 21. 208Google Scholar (Schrader, H., Porphyrii quaestionum Homericarum ad Odysseam pertinentium reliquiae [Leipzig, 1890], p. 123Google Scholar), Eust. 1873. 45 ff.

²⁶ Cf. Rem. 364 f. ‘ingenium magni livor detractat Homeri; | quisquis es, ex illo, Zoile, nomen habes.’ In spite of a number of attractive practical demonstrations of the relevance of the Homeric scholia to the criticism of Virgil byHeinze, R., Vergils epische Technik (Leipzig, 1915Google Scholar) andSchlunk, R. R., The Homeric scholia and the Aeneid (Ann Arbor, 1974)Google Scholar, the methodology of their application to Roman poetry remains to be explored.

²⁷ The recognition scene in the IT obviously made a deep impression on Aristotle, as he refers to it also in Poetics 1452b 5–8 and 1454a 7.