Hostname: page-component-5d59c44645-klj7v Total loading time: 0 Render date: 2024-02-21T05:49:53.579Z Has data issue: false hasContentIssue false

Sabinus, the Heroides and the Poetnightingale. Some observations on the authenticity of the epistula Sapphus*

Published online by Cambridge University Press:  11 February 2009

Gianpiero Rosati
Università di Udine


Of all the works attributed to Ovid but of disputed authenticity, the epistle of Sappho to Phaon is notoriously the one which has most perplexed scholars. Most philologists at the end of the 19th century asserted the Ovidian paternity of the epistle; but in recent years the discussion has flared up once again, especially following an important contribution, tending in the opposite direction, by R. J. Tarrant, and today, above all in Anglo-American studies, the pendulum seems to be swinging more in the direction of inauthenticity, according to the movement typical in debates of this kind. The present article obviously does not intend to discuss the whole question once again nor to reaffirm tout court the attribution to Ovid, but brings to the attention of scholars certain arguments which should not be neglected in the discussion (and which point in the direction of authenticity). I do not mean to underestimate the linguistic, stylistic, and metrical anomalies which scholars up to Tarrant and beyond have imputed to the epistula Sapphus, but rather to indicate some characteristics, above all of compositional technique, which have not been considered but which I think have a not insignificant weight in the debate on authenticity.

Research Article
Copyright © The Classical Association 1996

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)


1 ‘The Authenticity of the Letter of Sappho to Phaon (Heroides 15)’, HSPh 85 (1981), 133–53.

2 Besides Tarrant and Murgia (n. 3), this position is shared by e.g. E. J. Kenney, Philologus 111 (1967), 213 n. 2 and CQ 29 (1979), 430 n. 124; P. E. Knox, HSPh 90 (1986), 207–8; McKeown, J. C., Ovid: Amores. Text, Prolegomena and Commentary.Vol. I Text and Prolegomena (Leeds, 1987), 86 n. 32. Less clear are the positions of S. Hinds (n. 7), 44, and of A. Barchiesi (Aevum[ani\ 5 [1992], 236 n. 37, but cf. P. Ovidii Nasonis Epistulae heroidum 1–3 [Florence, 1992], 52 n. 2).Google Scholar

3 The recent article by E. Courtney, ‘Ovid and an epigram of Philodemus’, LCM 15 (1990), 117–18, also moves in this direction: he intends to deny the claim of Murgia, C. E., ‘Imitation and Authenticity in Ovid: Metamorphoses 1.477 and Heroides 15’, AJPh 106 (1985), 471, that in the epistula Sapphus ‘not a single line shows internal evidence of earliness’ with respect to the amores and to Ovid's other erotic works.Google Scholar

4 It is upon this catalogue that depend both the hypotheses concerning the dating of the heroides (this issue is still quite controversial) and those relating to their very physiognomy and dimensions. As is well known, a restrictive interpretation of Ovid's testimony (which goes back to Lachmann but has continued to enjoy a certain favour ever since and still finds adherents) limits the number of epistles to be attributed to Ovid to only the nine (or, if Sappho's is excluded, to the eight) which he lists (unless one adds to them, as a substitute for the epistle of Sappho, that of Briseis: so Tarrant [n. 1], 152). With regard to the dating (full discussion in Jacobson, H., Ovid's Heroides [Princeton, 1974], 300–18;Google Scholar J. C. McKeown, pp. 86–8; see too the clarification of Citroni, M., Poesia e leltori in Roma antica [Bari, 1995], 467–8 nn. 22 and 24, pp. 470–1 n. 32), I shall limit myself here to pointing out the argument proposed again most recently by A. Primmer (‘Datierungs- und Entwicklungsfragen bei Vergil und Ovid’, WS 16 [1982], esp. 254ff.), but in fact already hinted at by W. Kraus, in Ovid, M. v. Albrecht and E. Zinn (eds.), (Darmstadt, 1968), p. 83 n. 11 (revised version of the article Ovidius in the RE). This argument (apparently not taken account of by either McKeown or Hinds [n. 7] or C. E. Murgia, ‘The Date of Ovid's Ars amatoria 3’, AJPh 107 [1986], 74–94), which seems to me decisive for interpreting am. 2.18.19 as referring to the ars amatoria and hence for assigning the elegy to the second edition of the amores, points out that v. 20 (cf. below) must be alluding to the third book of the ars, the one Ovid wrote for the benefit of women (this is the motif of ars 3.590 nee dubito telis guin petar ipse meis; Murgia 91 n. 28 takes it instead to refer generally to amores, ars and heroides). No serious objection to this interpretation, which dates the third book of the ars before the second, three-book edition of the amores, is provided by the famous and much tormented verse ars 3.343, which–according to a controversial reading but one accepted in various modern editions (deve tribus libris)—would in fact presuppose the three-book edition of the amores: first of all because there are very good reasons to read, with the Hamiltonensis, deve tener (so now, after E. Pianezzola in his ed. of the ars [Milan, 1991], Kenney too in the OCP); and furthermore because one cannot exclude, at the limit, the possibility that the second edition of the amores and that of the ars (including the third book) were being prepared simultaneously (cf. also Citroni 470 n. 32).Google Scholar

5 Bardon, H., La litte'rature latine inconnue (Paris, 1956), ii. 6061Google Scholar 60–61 is entirely misleading on Sabinus: it is obviously false to assert that Ovid attributes to him only a letter of Ulysses to Penelope (the reference is to Pont. 4.16.13–14, whereas the more important testimony, am. 2.18, is ignored) and that it is up to us ‘d'imaginer tout le recueil, avec les epitres de Jason, de Protesilas, de Paris, d'Hercule, et avec les responses de leurs amantes’ (Bardon evidently supposes that Sabinus wrote first letters from the heroes to their women and then letters of reply on the part of the latter). Likewise without foundation is the hypothesis of Delia Corte (‘L'annunzio delle Heroides’, GIFn.s. 3 [1972], 315 [=Opuscula, vol. XI (Genua, 1988, p. 102]) ‘ che, via via che Ovidio scriveva una lettera delle heroides, e la mandava a Sabino, questi, che stava allora compiendo un viaggio per tutto il Mediterraneo, prowedesse a prepararne la risposta (w. 27–8)’ (cf. also 316); far more plausible is the interpretation of the two verses as witty, as a literary circle's joke, proposed by C. Neumeister, A&A 28 (1982), 100.

6 The text is that of Kenney's edition (cf. also n. 12 below). Tarrant's hypothesis that verses 26 and 34 are both fruits of an interpolation designed to justify the illegitimate penetration of the epistula Sapphus (according to Tarrant, a product of the Neronian or Flavian era) into the corpus of the heroides is too contrived to seem credible. The very implausibility of such a theory induces other scholars, convinced as they are of the epistle's spuriousness, to prefer the hypothesis that an original, genuinely Ovidian epistle of Sappho (accepting the evidentiary value of am. 2.18) was replaced by the transmitted epistula Sapphus: along these lines cf. most recently Murgia (n. 3), 471–2, and Booth, J., Ovid. The Second Book of Amores (Warminster, 1991), 189. Against this ‘rather unlikely set of coincidences’, cf. already Jacobson [n. 4], p. 278.Google Scholar

7 ‘Medea in Ovid: Scenes from the Life of an Intertextual Heroine’, MD 30 (1993), 9–47, esp. 31–4.Google Scholar

8 The point is well put by Goold, G. P., HSPh 69 (1965), 43: ‘the fact that Penelope's mail is mentioned first and Sappho's last I interpret as a poet's intimation that replies were sent to all. But Ovid has tastefully not repeated the whole list and, equally tastefully, has varied the order’ (but cf. already Jacoby, F., RhM 60 [1905], 71 n. 2). The only other reference to Sabinus, in Pont. 4.16.13–14 (et qui Penelopae rescribere iussit Ulixem / errantem saevo per duo lustra mart), is significant in this sense: here the allusion solely to Ulysses' epistle (the reply to the very first epistle, Penelope's) is enough to indicate the whole collection. If we lacked am. 2.18.29–34, who would not believe that that was the only letter Sabinus wrote?Google Scholar

9 Cf. the opposition miserae / crudelis in the last distich of the epistle of Sappho.

10 The claim that 23 refers to Hypsipyle's epistle (on the basis of 33, i.e. the explicit mention of a letter of reply to her) is refuted by Hinds (n. 7) 32–3. Certainly male gratus of 23 is a significant hint, and hence either refers to the epistle of Medea, as Hinds suggests, or, one might perhaps suggest as a second possibility, comprises both letters, that of Hypsipyle as well as that of Medea, indicating in Jason's ingratitude his tendency towards ‘recidivism’ (the parallelism of the two episodes which the sixth epistle so much insists upon). On the other hand, with regard to the list of Sabinus' letters, if the reasons of poetic economy proper to an Alexandrian catalogue illustrated by Hinds suggested that only one reply of Jason's could be cited, this could only be the one to Hypsipyle, the first of the two women, abandoned for the second one: combining the effect with the cause, the mention of a single woman could designate both. An analogous taste for parallelism (Jason's recidivism as the ‘family destiny’ which links Theseus and his son: on this motif, but in relation to another son, Demophoon, cf. also n. 24) can be recognized in v. 24 Hippolytique parens Hippolytusque…Google Scholar

11 According to Booth, J., si modo vivit (v. 32)‘ perhaps also reflects Demophoon's anxiety and renewed promise of return’ (p. 189 ad loc.; cf. also p. 86). It cannot be excluded that Demophoon might have tried to oppose Phyllis' suicidal plans, but in any case his reply could not have modified the course of events.Google Scholar

12 An evident superposition of v. 34 upon 26 seems to have produced the corruption in this latter. Like Kenney, Munari, F. (Florence, 1951) too recurs to cruces, but excludes from them Aoniae Lesbis, limiting the corruption to the last two words. I too believe that, even if the origin of the corruption seems evident, this is no reason to exclude the possibility that there really was a partial coincidence between the two verses (which could have facilitated the corruption), so as to mark in parallel, perhaps precisely with the mention of the lyre, the close of the two catalogues, and that the corruption (it has most obviously affected amata, which is meaningless here, and, precisely because of the very mode of the corruption, offers no hope of conjectural restoration) might therefore be more limited than the sequence Aoniae…lyrae (cf. now also Booth ad loc). The variant arnica is not convincing (cf. e.g. Tarrant [n. 1], 151, and Booth), though it is accepted by Goold (who nonetheless in his revision of the Loeb edition [Cambridge, MA and London, 1977] leaves in 26 the text of Showerman, et Aoniae Lesbis amata lyrae) and most recently by McKeown, too.Google Scholar

13 This observation is attributed to Loers (1829) by Jacobson (n. 4), p. 278, but it is already found in Domizio Calderini's introduction to his commentary on the epistle (1475).Google Scholar

14 On the intense discussion about the relation between the meaning of am. 2.18.34 and that of w. 181ff. of the epistula Sapphus, cf. the bibliography supplied by Dorrie, H., Ovidius Naso, P.. Der Brief der Sappho an Phaon (Munich, 1975), 187Google Scholarn. 1 (add Comparetti, D., Suitautenticita delTepistola ovidiana di Saffo a Faone [Florence, 1876], 15–18 and 30, whose interpretation, however, is based in w. 169–70 upon a minority reading, tetigit lentissima Pyrrhae / pectora, which presupposes that the woman loved by Deucalion fell in love with him in her turn).Google Scholar

15 ‘Denn daB Sappho nicht mit Phaon gliicklich wurde, das war in der Uberlieferung fest verankert’: Dorrie [n. 14], p. 189.Google Scholar

16 Cf. Dorrie [n. 14], p. 188: ‘Keiner der iibrigen Sabinus-Briefe bringt eine Wende in der Situation der Heroinen’. As we have seen, some degree of uncertainty may remain in the case of Demophoon's reply to Phyllis; but it seems certain anyway that, regardless of this latter (whether it was negative or arrived too late), the destiny of death which the myth regularly assigns to the heroine was not modified in the end: cf. Barchiesi, A., ‘ Narrativita e convenzione nelle Heroides’, MD 19 (1987), 63 n. 1 (repr. as introd. to A.B., P. Ovidii Nasonis Epistulae heroidum 1–3 cit.):‘ al v. 32 […] sembra di cogliere una giocosa riserva di Ovidio, basata proprio sulla stretta scansione temporale che e tipica delle epistole 1–15: la situazione in cui Fillide scrive non lascia tempo per una risposta’. In any case, here too Ovid is presupposing familiarity with the heroides on the part of the reader of the catalogue (and hence not only at v. 34, as maintained by Tarrant [n. 1], 150, who uses this as a further argument to declare it an interpolation).Google Scholar

17 As proposed by Dorrie [n. 14], pp. 189–90, who therefore hypothesizes, in contradiction to what is generally thought, a corruption produced by the superposition of v. 26 upon 34 (and suggests that amata is a gloss which has penetrated mistakenly into the text; but this hypothesis is hardly credible, because the glossator would demonstrate himself to be scarcely familiar with the text he recalls).Google Scholar

18 The point is well put especially by Barchiesi, A. (n. 16): Tiniziativa dell'amico Sabino […] puo aver avuto il carattere scherzoso di una violazione intenzionale, un po‘ da guastafeste (comunque in spirito ovidiano, data la passione di Ovidio per il disvelare tongue-in-cheek le convenzioni letterarie e la loro arbitrarieta)’ (an allusion to Sabinus' witty distortion of Ovid's intention can also be detected in the iussit of Pont. 4.16.13, cf. n. 8).Google Scholar

19 There is merely a hint in this sense in Jacobson (n. 4), p. 335, and, as I now see, in Heldmann, K., ‘Ovids Sabinus-Gedicht {Am. 2, 18) und die Epistulae Heroidum’, Hermes 122 (1994), 188–219, at 198, who moves, however, in a direction different from my own, and in any case is not concerned with the authenticity of the epistle of Sappho (on which 211 n. 78).Google Scholar

20 In fact it must not be forgotten that the position of the epistula Sapphus concluding the first cycle of epistles is a conjecture of Daniel, Heinsius (1629) based precisely upon the catalogue of am. 2.18; hence the danger of a circular argument is apparent (but a piece of evidence for such a collocation already in the 15th century is pointed out by Pulbrook, M., ‘The original published form of Ovid's Heroides’, Hermathena 122 [1977], 29–45, at 44 n. 24). We shall return to this issue below (n. 22).Google Scholar

21 The witty, provocative character of Sabinus' ploy probably explains another fact as well. In Sappho's last words, Ovid had posed an alternative between return and desertion (this latter to be communicated with a farewell letter); Sabinus does indeed have Phaon write a letter to Sappho, but so as to announce to her his return, thereby jumbling the alternatives. To be sure, I am not unaware that this ‘violation’ of the base-text has not only prompted suspicions concerning the text of am. 2.18.34 (cf. above n. 17), but can also nourish doubts concerning just what epistle (one different from the transmitted epistula Sapphus?) Sabinus was replying to.

22 As was pointed out above, the location of the epistula Sapphus concluding the first series in modern editions of the heroides is conjectural. But that conjecture, based upon the correct observation that in both catalogues of am. 2.18 the same epistles, Penelope's and Sappho's, occupy—probably not by chance—the extreme positions and thereby indicate the limits of the work, can on my view count upon another good argument too, besides finding corroboration (pace Tarrant [n. 1], 148) in the Florilegium Gallicum, which transmits its excerpta in an intermediate position between those of Hypermnestra's letter and Paris'. The exquisitely poetological character of Sappho's epistle (for good reason, hers is the only voice of the heroides endowed by so marked a literary consciousness, as we shall see again in the final pages of this article) authorizes the hypothesis that it must have had a prominent location, such as the final one, so as to be able to construe the whole work in an elegiac perspective, writing its poetics and offering an authorized interpretation for it. It is quite evident that, for a poet as concerned as Ovid was to indicate the borders of texts, Sappho was an ideal figure with whom to conclude his work—just like, at the opposite end, Penelope, the woman of waiting, the very emblem of the woman far from her husband (in his turn, lenlus par excellence, cf. 1.1). On the contrary, it would be difficult to assign an analogously significant function to Hypermnestra (the final epistle according to the manuscript tradition). True, the final distich of her epistle (scribere plura libel, sedpondere lassa catenae / est manus et vires subtrahit ipse timor, 131–2) can lend itself to being read symbolically (on the motif of fatigue in epistolary conclusions cf. Kirfel, E.-A., Untersuchungen zur Briefform der Heroides Ovids [Bern-Stuttgart, 1969], pp. 7980), but this seems more pertinently motivated by the features which distinguish this particular character (Hypermnestra is lamenting an unjustly inflicted punishment—cf. catena—and is terrorized by a violent paternal figure). Another argument against such a hypothesis is the anomaly of a collection of 14 epistles, given that it is well known that «within individual libelli the Augustan poets cultivated structures based on multiples of fiveGoogle Scholar (Kenney, E. J., Apuleius. Cupid & Psyche [Cambridge, 1990], p. 3). In this sense it is hardly accidental that the catalogue of epistles which Ovid lists at am. 2.18 (nine of his own and six of Sabinus') adds up to fifteen (Tarrant's hypothesis [n. 1], 152 n. 39, that a single edition comprises the nine Ovidian epistles indicated in the catalogue of the am. and Sabinus' six responses, is not persuasive).Google Scholar

23 ‘ein wenig prezios’, as Dorrie puts it ([n. 14], pp. 107–8 and n. 29). He feels the need to defend Ovid for his use of so unusual a term against possible accusations of a taste for enigmas and appeals to the clarity of the context, which leaves no doubt about the reference to Sicily.Google Scholar

24 Exactly as Theseus himself (associated with his son Demophoon) is used in the ars as a paradigm of deceptae… crimen amantis (3.454) whom the Athenian women should guard against (here too the exemplum is selected with a view towards the addressee, as belonging to the very same city): parcite, Cecropides, iuranti credere Theseo: / quosfaciel testis, fecit et ante, deos. / et tibi, Demophoon, Thesei criminis heres, / Phyllide decepta nulla relicta fides (457–60).Google Scholar

25 There is an evident contamination, here as in other passages in Ovid and in other Augustan poets (Lyne, cf. R. O. A. M. in his commentary to the Ciris [Cambridge, 1978], on w. 54–7;Google ScholarTimpanaro, S., Nuovi contributidifilologia e storia della lingua latina [Bologna, 1994], pp. lOlff.), between Scylla the daughter of Nisus and the homonymous nymph transformed by Circe into the marine monster girded at the waist by barking dogs.Google Scholar

26 On this verse see now, after Hinds (n. 7), 15, also S. Casali, ‘Ancora su Medea e Scilla (Ovidio, Heroides 12, 124’, MD 32 (1994), 173–4; further discussion in the forthcoming commentary by Federica Bessone on the epistle of Medea.Google Scholar

27 Codified in elegy in passages like Prop. 4.4.39–40 (together with Ariadne) and 3.19.21–4 (where, analogously, the exemplum is accompanied by a warning to unmarried women: at vos, innuptae, felicius urite taedas, 25).

28 A full collection of material on the nightingale is provided by Sauvage, A., Etude de themes animaliers dans la poesie latine. Le cheval-Les oiseaux (Bruxelles, 1975), pp. 192206 (on its association with groves, pp. 193–5; on the maestitia of its song, esp. pp. 198–201 and 204–6); for the Greek sources, cf. above allGoogle ScholarKannicht, R. in his comm. to Eur. Hel. (Heidelberg, 1969), ii. 281ff.Google Scholar and Bulloch, A. W. on Call. Lav. Pall. (Cambridge, 1985), pp. 205–6.Google Scholar

29 The difference in age between the very young Phaon and the more mature Sappho (of which there is no explicit trace in the epistle, but which seems to be presupposed in the references to his extraordinary youth in w. 21, 85, and esp. 93: o nee adhuc iuvenis, nee iampuer) might well sound slightly ironic in this connection (cf. also the assimilation of Sappho's pain to that of a mother at the funeral of her son in w. 115–16).Google Scholar

30 The analogies between the two myths have been pointed out more than once (e.g. M. Pohlenz, Hermes 48 [1913], 7ff.; H. Trankle, Hermes 91 [1963], 465ff.), and explained with reference to the presumed Hellenistic epyllion which recounted the story of Hero and Leander and served as a model to Musaeus and the two Ovidian epistles, leaving conspicuous traces in other Latin literary texts (cf. recently Papanghelis, T. D., Propertius: A Hellenistic Poet on Love and Death [Cambridge, 1987], pp. 103ff.; further discussion in my forthcoming commentary on heroides 18 and 19).CrossRefGoogle Scholar

31 Cf. e.g. AP 9.262.5–6 [Gow-Page, GP w. 2831–2, with the note ad loc.], where a mother's lament for the death of her six children, three by disease and three at sea, is assimilated respectively to the nightingale's lament and the halcyon's; Prop. 3.10.9–10 alcyonum positis requiescant ora querelis; / increpet absumptum nee sua mater Ityn (here too the value of a dire omen is attributed to their song); Ov. trist. 5.1.60 hie querulam Procnen Alcyonenque facit; [Ov.] epiced. Drus. 105–8 (cf. n. 40); but cf. e.g. also Sen. Agam. 669ff. and 68Iff.; [Sen.] Oct. 7–8; Stat. silv. 3.5.57ff., etc. Cf. also Sauvage (n. 28), pp. 198–9 and 283–4.Google Scholar

32 The position of Courtney, E., Hermathena 119 (1975), 83 (and already BICS 12 [1965], 63–6), who thinks all the double epistles are spurious, seems isolated.Google Scholar

33 The material is gathered by Steier, A., RE XIII.1864.42ff.Google Scholar

34 Cf. 7.49 Powell.

35 Wiseman, Cf.T.P., Catullan Questions (Leicester, 1969), pp. 1718; most recently,Google ScholarSpoth, F., Ovids Heroides als Elegien (Munich, 1992), pp. 2930Google Scholar (esp. p. 30 n. 11 for the correspondences with the epistle of Sappho), and A. Barchiesi, ‘Riflessivo e future Due modi di allusione nella poesia, ellenistica e augustea’, Aevum(ant) 5 (1992), 241.

36 The reuse of Catullus (ignored in the comm. of Palmer and Dorrie; merely pointed out, but not evaluated, in that of Vries, S. G. De [Berlin, 1888], p. 93; cf. also I. Cazzaniga, La saga di his nella tradizione letteraria e mitogrqfica greco-romana [Varese-Milan, 1951], vol. i, p. 83) is rendered certain not only by concinit (cf. also am. 3.12.32 concinit Odrysium Cecropis ales Ityn) but above all by the epithet Daulias, which Thucydides says was widespread in Greek poetry (iroXXois Si Kai Toliv noirjTwp kv ar)86vos (J.vrnj.r) Aavkias ij opvis enwvofiaoTai), but which in Latin is never attested elsewhere before the epistula Sapphus and [Ov.] epiced. Drus. 106 (cf. n. 40); then cf. [Sen.] Here. Oet. 192 Daulias ales and [Verg.] Ciris 199–200 puellae / Dauliades (a curious feature shared by numerous texts which are spurious or whose authenticity is at least suspect): cf. Lyne ad loc; Daulis in Sen. Thyest. 275.Google Scholar

37 According to an identification widespread above all in Roman poetry: indications in Hinds, S., The Metamorphosis of Persephone. Ovid and the Self-conscious Muse (Cambridge, 1987), pp. 103–4. It should also be noted that the same Catullan poem contains the word maeror (15 in tantis maeroribus), which recurs in epist. Sapph. 117 but whose rarity in Augustan poetry (only in Hor. ars 110, then also in epiced. Drus. 294) has naturally provoked scholarly suspicions (Tarrant [n. 1], 140).Google Scholar

38 For a full discussion of the epigram and the problems connected with it, cf., after Gow- Page, HEn. 191–2, at least N., Hopkinson, A Hellenistic Anthology (Cambridge, 1988), p. 249, and now (with further bibliographical indications) R. Hunter, ‘Callimachus and Heraclitus’, MD 28 (1992), 113–23; on the echoes of the epigram in Ovid, cf. Williams, G., CQ 41 (1991), 169–77.Google Scholar

39 I am thinking, for instance, of the fragment of Sappho (150 V.) on the incompatibility between threnos and the Muses (motivated by a death in the family, as Maximus of Tyre, who transmits the fragment, seems to attest), which suggests a certain analogy with the opening verses of Catullus' poem, on the pain that makes it impossible for the poet to venerate the Muses (etsi me assiduo confectum cura dolore / sevocat a doctis, Ortale, virginibus, / nee potis est dulcis Musarum expromere fetus / mens animi…). Another indication might come from Thucydides' report (cf. n. 36) of the diffusion in poetry of the epithet AavXids to designate the nightingale: in fact no secure attestation seems to survive in Greek (according to a hypothetical supplement by Pfeiffer, Callimachus may have used the epithet in frg. 113.2), while in Latin poetry, as was mentioned above, we have no occurrence before Catullus. For possible relations (in a different sense) between Sappho and Catullus 65, cf. P. A. Johnston, ‘An Echo of Sappho in Catullus 65’, Latomus 42 (1983), 388–94.

40 The evident fact that epist. Sapph. 154 Ismarium Daulias ales Ityn and epiced. Drus. 106 Threicium Daulias ales Ityn cannot be entirely independent of one another raises the familiar and quite difficult question of the relation between the two texts (cf. e.g. Purser in Palmer, p. 421). If the hypothesis proposed here, that epist. Sapph. 153–4 is directly dependent upon Catullus 65.12–14, is accepted, then it seems likely to me that epiced. Drus. 105–8 italis in umbrosis, mitis nunc denique, silvis / deflet Threicium Daulias ales Ityn; / alcyonum tales ventosa per aequora questus / ad surdas tenui voce sonantur aquas depends in t u rn upon the epist. Sapph. (mitis nunc denique seems to provide a corrective clarification of virum non ultapie of epist. Sapph. 153). Of course, if we accept the posteriority of the consolatio ad Liviam (the date of which is uncertain, but the likeliest proposals do not go beyond the beginning of the Neronian period: Schoonhoven, cf. H., The Pseudo-Ovidian Ad Liviam de morte Drusi [Groningen, 1992], pp. 37–8; Butrica, J. L., CR 43 [1993], 265, does not even find anything in it incompatible with the Augustan age), it follows that the attribution of the epistula Sapphus to the Neronian or Flavian period, as suggested by Tarrant (n. 1), 134, is to be excluded (the termini post and ante quern set by Murgia [n. 3], 466, are instead respectively Pont. 2.10 and the tragedies of Seneca).Google Scholar

41 On mollis as a technical term of elegy cf. Fedeli on Prop. 1.7.19 and Spoth (n. 35), p. 72 n. 45.

42 Ov. am. 2.4.9–10 non est forma meos quae certa invitet amores: / centum sunt causae cur ego semper amem also depends directly upon this Propertian passage. There are some observations on the Sappho epistle's reuse of Ovid's erotic poetry in Jacobson (n. 4), p. 298 (who however— beside interpreting the whole poem, quite improbably, as parodic—speaks indistinctly of poetic activity, without noting the specificity of Sappho's discourse on elegy).

43 On the relation of the heroides to elegy see now above all the specific monograph of F. Spoth (n. 35).

44 So much so that Tarrant's judgement (n. 1), 135–6, seems too severe: ‘It is my private opinion that the ES is a tedious production containing hardly a moment of wit, elegance, or truth to nature, and that its ascription to Ovid ought never to have been taken seriously’.

45 Murgia (n. 3), 466 n. 24.