¹ Sexti Properti Carmina,ed. J. S. Phillimore (Oxford, 1901). The antepenultimate paragraph of the introduction denounces transpositions of the sort proposed by Housman and introduced by Postgate in his Corpus Poetarum Latinorumtext, then continues: ‘Sed cuinam probari potest nostri archetypum ita dilaniatum et ut disticha passim inter se locum mutaverint, omnino superfuturum fuisse? Est profecto ut peccaverint in non nullis librarii; homines enim. At non beluae. Quod si in summa re codicum fidem respuerimus, quo denique stabitur? Vnus quisque enim in quolibet argumento proprium phantasiae tenorem propriam carminis deducendi inventionem sequitur, licet non semper optimam illam nee ceteris maxime arrisuram. Quid enim est aliud ingenium? Sin autem poetae cogitationem suo cuique arbitrio resarcire licet, non interpretari, quot editores tot Propertii.’

² Hallett, J. ‘The Role of Women in Roman Elegy: Counter-cultural Feminism’, Arethusa6 (1973), 103–24.Google Scholar

³ A. de Sanctis, Properzio: Saggio d′interpretazione psicologica(Bibliotheca Biographica 9; Rome, 1973).

⁴ This view has been advocated or assumed in a number of articles by a variety of (especially American) scholars, but its chief exponent has perhaps been J. P. Sullivan in Propertius: An Introduction(Cambridge, 1976).

⁵ Stahl, H.R.Propertius: Love and War. Individual and State under Augustus(Berkeley, 1985).Google Scholar

⁶ Papanghelis, T.Propertius: A Hellenistic Poet on Love and Death(Cambridge, 1987).Google Scholar

⁷ Benediktson, D. Thomas, Propertius: Modernist Poet of Antiquity(Carbondale and Edwardsville, 1989).Google Scholar

⁸ On 1.9.9 (1.10. [9.] 9 in his numbering) Lachmann (Leipzig, 1816) says of ducerein that line and of totisin 21 that ‘Immerito obtrusas Propertio elegantias hie ostentant exemplaria’, attributing the latter to Jan Dousa the younger, the former to Volscus′ edition of 1482; he notes that ‘Scripti omnes, ne uno quidem demto’ have dicereand totiens,and concludes, ‘Quis jure impugnet? quis tanti ducat defendere?’ Another conjecture that Lachmann rejected in a similar manner is furitat 4.6.56 (which I shall advocate later in this paper), of which he wrote, ‘Nempe critici poetas ubique aut furere aut magna verba aut elegantias effundere volunt. Liceat, quaeso, Propertio nostro ita loqui, ut eum locutum esse libri veteres testantur’. Here Lachmann′s attitude was anticipated in the commentary of C. Kuinoel (Leipzig, 1805) and apparently in that of Vulpius as well, which is not available to me; Kuinoel first quoted Burman′s note, in which fuitwas described as ‘languidius’, then continued, ‘Non nego, furitexquisitius esse: neque tamen scriptores vett. huius generis elegantias quouis loco sectati sunt, et cum vulgaris lectio commodum sensum pariat: nihil sine codicum auctoritate mutandum. Nee quaerendum, vt ad h. 1. notauit Vulpius, quid scriptor scribere potuerit, vel debuerit, sed quid vere scripserit’. Note that the unwillingness to alter ‘sine codicum auctoritate’ limits one effectively to conjectures made before the age of printing, and that both Kuinoel and Lachmann, instead of maintaining an open mind, were predisposed to assume without question that what the manuscripts gave was what Propertius wrote.

⁹ Postgate, J.P.Select Elegies of Propertius(London, 1881).Google Scholar

¹⁰ Sullivan, J.P.Ezra Pound and Sextus Propertius(Austin, 1964).Google Scholar

¹¹ It seems to have gone unnoticed that every printing of Select Elegiesfrom 1894 on contains the following ‘Publishers′ Note’: ‘The present issue is an exact reprint of the edition of 1884, and the Editor wishes it to be known that the book does not represent his present views on the text and interpretation of Propertius.’

¹² Taylor, G.Reinventing Shakespeare: A Cultural History, From the Restoration to the Present(New York, 1989), 408.Google Scholar

¹³ Reviewing Fedeli′s 1984 Teubner text and 1985 commentary on Book 3 at RFIC114 (1986), 212: ‘L′opera del critico del testo di Properzio, oggi, non pud che consistere nell′esercizio del giudizio critico entro i limiti di quella che si puo chiamare la ‘volgata moderna′ di Properzio.’

¹⁴ Hubbard, M.Propertius(New York, 1975), 2–3.Google Scholar

¹⁵ Goold has also discussed the text of Propertius in a series of articles, ‘Nodes Propertianae’, HSCP71 (1966), 59–106; ‘On Editing Propertius’, Papers in Honour of Otto Skutsch (BICSSupplement 5,1987), 27–38; ‘Problems in Editing Propertius’, in J. N. Grant (ed.), Editing Greek and Latin Texts(New York, 1989), 97–119; and ‘Paralipomena Propertiana’, HSCP94 (1992), 287–320.

¹⁶ Butrica, J.L.The Manuscript Tradition of Propertius (PhoenixSupplementary Volume 17; Toronto, 1984); Heyworth, S.J.The Manuscripts of Propertius: Toward a Critical Edition(diss. Cambridge, 1986). The principal manuscripts that will be discussed in the following pages are N (= Wolfenbuttel Gud. lat. 224, c 1200), A (= Leiden, Voss. lat. O.38, c 1240), F (= Florence, Bibl. Laurenziana pi. 36,49, c1380), L (= Oxford, Bodleian Library Holkham misc. 36, a. 1421), P (= Paris, Bibl. Nationale lat. 7989, a. 1423), and Z (= Venice, Bibl. Naz. Marciana Fondo antico 443 [1912], a. 1453); FLPZ are known collectively as the Petrarchan manuscripts because they derive from Petrarch′s lost copy of A. In addition, a group of fifteenth-century manuscripts appears to derive from the archetype independently of the other primary witnesses (Butrica calls their source X and argues that it shares an intermediate source with N, while Heyworth calls their source Aand thinks now, as he has pointed out to me, that it might have been the archetype itself); these are v (= Vat. lat. 3273, a. 1427), m (= Paris, B.N. lat. 8233, a. 1465), r (= Geneva, Bibl. Bodmeriana Cod. Bod. 141, a. 1466), u (= Urb. lat. 641, c1465–70), s (= Munich, Universitatsbibliothek Cim 22, c 1460–70), and c (= Rome, Bibl. Casanatense 15, a. 1470 or 1471). (Heyworth′s sigla for these six manuscripts are the upper-case equivalents of those given above, except that he uses T for v. In addition, he cites three descendants of a contaminated copy of X [A]—Parma, Bibl. Palatina 140; Wroclaw, Bibl. Uniw. AKC 1948 197 KN; and Bibl. Vaticana Capponianus 196—as JKW, collectively F,but their value remains to be established; the one case that he cites of an ‘apparently true, and not conjectural, reading not found in any other authoritative mss’ [92] could easily be accidental if not in fact conjectural.) The agreement of all or most of the witnesses to the text is here designated by the traditional symbol O.Google Scholar

¹⁷ Fedeli, P. ‘Properzio, fra culto del testo tradito e caccia alia corruttela’, Bulletin de la Faculte des Lettres de Mulhouse 15 (1987), 107–12. For the phrase ‘cautiously conservative’, see 108, where Fedeli writes, ‘Oggi la tendenza della critica testuale properziana, sopiti ormai i furori di tipo housmaniano, pud esser definita ‘accortamente conservatrice’: e questa una posizione che sostanzialmente condivido’.Google Scholar

¹⁸ Fedeli (above, n. 17), 107.

¹⁹ A. A.3.333 teneripossis carmen legisse Properti.

²⁰ Tr.2.465 inuenies eadem blandipraecepta Properti,5.1.17 blandique Propertius oris.

²¹ Hubbard (above, n. 14), 2.

²² 14.184.1 facundi carmen iuuenale Properti.

²³ 10.1.93 elegia quoque Graecos prouocamus, cuius mihi tersus atque elegans maxime uidetur auctor Tibullus. sunt qui Propertium malint. Ouidius utroque lasciuior, sicut durior Gallus.This important point, that Quintilian in no way implies disparagement of Propertius, was first made by Hubbard.

²⁴ Ep.9.22.1 uir est[sc. Passennus Paullus] optimus, honestissimus, nostri amantissimus; praeterea in litter is ueteres aemulatur, exprimit, reddit, Propertium inprimis, a quo genus ducit, uera suboles eoque simillima Mi in quo Me praecipuus. si elegos eius in manum sumpseris, leges opus tersum, molle, iucundum et plane in Properti domo scriptum.The translation of the last sentence given in the text has been taken from Hubbard, 2–3.

²⁵ To choose only one example from among many, I append the words of an Italian translator of Propertius: ‘Egli rimane un po′ come una divinita il cui linguaggio e troppo ermetico per la folia dei fedeli.... Egli e in certo modo il precursore dei nostri ermetici.... La sua oscurita non e mai fine a se stessa; attraverso le ambagi dello stile, egli mira piu a suggerire che a dire: meglio che un′immagine egli vuole destare in noi una sensazione: e crea intorno a se un alone musicale in cui non sempre si pud cercare un significato preciso’ (Sesto Properzio, Elegie, tr. G. Lipparini [Bologna, 1970], xix).

²⁶ Note, for example, the hiatus turn eruntin 2, the difficulty of ascertaining the precise meaning of legamin 5, and the awkward double modifier jixa... deiuitiorain the same line.

²⁷ The first advocate of the reading, however, seems to have been Hubbard in′Propertiana, CQ 18 (1968), 315–19, at 318–19, who correctly terms it ‘lively’ and suggests that noceat might have been interpolated after the corruption of adeo to deo.

²⁸ This figure was arrived at by counting transpositions, lacunae,places where an archetypal reading has been replaced by a conjecture, and obelized passages (an entire obelized line was counted as three corruptions, an obelized half-line as two, and, of course, a single obelized word as one).

²⁹ A comparable, though less extreme, result emerges from the lines cited by ancient grammarians: two or three of those nine lines are corrupt in the entire tradition, while two more were corrupted in one branch. (These are readily accessible in Butrica [above, n. 16], 30–2. The lines corrupt in the entire tradition are 2.3.24 {ardidusfor candidus),3.8.37 (tendistifor nexisti),and perhaps 2.14.1 (estadded at the end of the line); those corrupt in a single branch are 2.1.2 (orein N) and 2.33.37 (demissa... sertain the Petrarchan manuscripts).

³⁰ Gebhard seems to have derived the correction from his liber Commelinianus,which is Paris, BN lat. 8458 (written no earlier than 1474); I have not verified the presence of the reading in that manuscript, but it certainly does appear in Rome, Bibl. Casanatense 3227, copied around 1470 by Franciscus Maturantius.

³¹ Broekhuyzen′s own note runs, ‘vix apparet quo referri debeat istud timor.tentabam ego aliquando, Nee timor est ulli clausae tutela puellae,i.e. nee quisquam amator timet custodes dominae suae, quo minus ad earn accedere audeat palam’.

³² I had thought that this was an exclusively modern attitude until I consulted the commentary of F. G. Barth (Leipzig, 1777), which remarks sarcastically here, ‘Eximia Iaus fortitudinis Augusti!’ (but one can hardly imagine Propertius depicting the officially pious, and not yet officially deified, Augustus as greater than or even equal to a god).

³³ ‘Sometimes editors, both of classical and of modern works, argue that the most they are justified in doing is to attempt to purge the copy-text, or archetype, or paradosis, of errors—not to try to restore what the author wrote. But this argument cannot be praised for its respect of historical evidence; rather, it confuses two kinds of edition, both legitimate, neither of which, when done properly, disregards the evidence. If one is interested in a text as it appeared at a particular time to a particular audience, a diplomatic or facsimile edition of it serves the purpose best; correcting errors in it—editing it critically—would be out of place, for the errors, though unintended, were part of what the contemporary readers saw in the text in front of them. If, on the other hand, one wishes to correct errors—to try to repair the damage done to the text in transmission, however famous or influential its corrupt form may be—then one is producing a text that differs from any now extant (probably from any that ever existed), and the aim of the alterations is obviously not the preservation of a documentary form of the text but the construction of a text as close as possible (as close, that is, as surviving evidence permits) to the one the author intended’ (G. Thomas Tanselle, Textual Criticism and Scholarly Editing[Charlottesville/London, 1990], 301–2). For a classicist′s perspective on the same matter, see the entertaining preface to Hermann Koechly′s 1857 Teubner text of Nonnus, especially vi-vm, where he contrasts fashionably conservative modern editors, who boast that by reconstructing the archetype and removing only the most obvious and most trivial errors they have ‘emended, not interpolated’ their text, with the practice of men like ‘the ancient Hermann’, who ‘codicis archetypi—si quidem fuit—scripturam pro necessario quidem erigendi aedificii fundamento, sed non pro ipso aedificio habent’.

³⁴ I would argue that hochere is one of the associative errors demonstrating that N and X shared a common intermediate source; other possible examples of such errors include obcenisfor obscenis,found in N and v at 1.16.10; fletus iotflemus,found in N^pc(fletuN^ac) and v (in the form fletu⁹)and mru at 2.27.7 (note also fle followed by a lacuna in c); (a)eoifor Coi,found in N2 and sc at 3.1.1; and flaminefor flammae,found in N and sc at 3.13.21.

³⁵ F reads nee,but h,apparently representing haec,stands in the margin, written by correcting hand attributed to Salutati (though it is not always certain which of the four hands present in F has made a particular correction). The possibility that the correction could have been made originally in Petrarch′s lost manuscript, the exemplar of F and itself a copy of A (which reads nee),arises from the fact that other descendants of Petrarch′s copy also read h(a)ec.It is certainly the reading of P (whose scribe, however, introduced a substantial number of improvements as he copied). L is lost for this portion of the text, but Naples, Bibl. Nazionale IV.F.19, a descendant of it (or perhaps of its exemplar, as Heyworth has argued [above, n. 16, 30–2]), reads haecaccording to my own collation; Professor Heyworth, however, informs me that according to his own collation that manuscript offers in the text a reading that might be either Hecor Nee,with hocin the margin. Brussels, Bibliotheque Royale 14638, another descendant of L (or of its exemplar), reads hoc,but this is surely derived from v (for the suggestion that the Naples and Brussels manuscripts derive from v and L, see Butrica [above, n. 16], 110–12). Thus it remains at least possible that Petrarch had already introduced h(a)ecin his own copy and that the careless scribe of F corrupted it back to nee,but small words like these are interchanged with such frequency that certainty is impossible.

³⁶ It needs to be remembered that X (A)would certainly have been corrected by Poggio and perhaps by Niccoli as well before the first surviving transcript (v) was made, and that half of its descendants were copied by able scholars who surely introduced corrections of their own (Panormita copied v, Poggio′s son Jacopo copied s, Pomponio Leto copied c; for speculation that the common source of mru might have been a transcript made by Niccoli, see Butrica [above, n. 16], 70, with n. 16). For possible conjunctive errors of N and X, see n. 34 above.

³⁷ Smyth, W.R.Thesaurus Criticus ad Sexti Propertii Textum (MnemosyneSupplement 12; Leiden, 1970) offers in ‘Excursus I′ a conspectus of the wholesale reorderings of all four books by Gruppe, Carutti, and Richmond.Google Scholar

³⁸ In these lists of dislocations, corrections made by clearly later hands have been ignored;‘^ac;’indicates a correction that at least could have been made by the scribe himself, though many even of these were undoubtedly made by later hands as well.

³⁹ Of these two dislocations the second is certainly inherited, either from L or from its source, and the first may be as well; 1.6.15 was lacking in A and in Petrarch′s manuscript, so that the order 16/15 could have arisen from 15 being restored in an incorrect or unclear manner.

⁴⁰ In this case the intrusive line was originally written in an erasure over a version of 16 itself!

⁴¹ The scribe subsequently deleted this muddled version of 4.9 and began again from scratch.

⁴² Heyworth (above, n. 16), 67 makes the interesting observation that 4.7.31–82 occupy a complete folio in v (his T); it might be added that 4.7.83–8.36 do as well. For the descent of Cambridge 3394 and other manuscripts from a copy of v, see Butrica (above, n. 16), 100–3; it would appear, however, that the equivalency in 4.7.31–8.36 is purely coincidental, since the other blocks of lines shifted in Cambridge 3394 do not match the contents of folios in v.

⁴³ The manuscript originally read propinqui,but the hand that has introduced the transpositions discussed here has also introduced parentisas a correction.

⁴⁴ The poems collated were 1.1–2, 11–13, 17, 20–2; II.1–3, 8–13, 12–90, 24–34; III.1–5, 11, 13, 22; IV1–2, 6–9, 11, representing 2238 lines, or about 55 per cent of the text

⁴⁵ An amusing reflection of this occurs in the Patrologia Latinaedition of Petrus Cantor′s Verbum abbreviationat section 60 (= PLCCV 84–5). Here we find ‘Nunc impletum est poetae illud: aurea nunc vere sunl saecula, plurimus aurohenit honos.lauro perficilur, quidquid captatur iniquej nemoque praetenso munere tana rogat’.The editor attributes this to ‘PROPERT. 1. Ill, elegia 11’, having consulted, it seems, some edition which incorporated the Ovidian lines and numbered 3.13 as 11 (the Delphin edition of 1685 is the most likely candidate; I have not, however, identified the source of auro perficitur... vana rogat).

⁴⁶ Heyworth (above, n. 16), 76 indicates that v (his T) contains ‘A spurious couplet, written vertically in the far margin, for insertion between I ix 26 and 27’; the lines are in fact 1.9.27 (with tibirather than ubi nonand subcludererather than seducere,as in the version in the text) and 1.9.30 (with assiduaswhere the text has assuduas,corrected to assiduas),and it is uncertain whether they were intended as an interpolation or as corrections.

⁴⁷ For Propertius in the Middle Ages, see Butrica (above, n. 16), 20–30.

⁴⁸ See now the fine facsimile printed in Buonocore, M.Properzio nei codici delta Biblioteca Apostolica Vaticana(Assisi, 1995) as pi. XVIII, showing f. 21, the first of the two sides containing extracts from Propertius, where all the phenomena discussed here can be observed.Google Scholar

⁴⁹ Tarrant, R.J. ‘Toward a Typology of Interpolation in Latin Poetry’, TAPA 117 (1987), 281–98 and ‘The Reader as Author: Collaborative Interpolation in Latin Poetry’ in J. N. Grant (ed.), Editing Greek and Latin Texts(New York, 1989), 121–62.Google Scholar

⁵⁰ The word has a different sense in 1.15.24, tu quoque utifieres nobilis historia,where it means ‘legend’ or ‘myth’.

⁵¹ For this reading, see Butrica (above, n. 16), 77–8.

⁵² It should be noted that, in contrast to the circumstances that (according to Donatus) led to the creation of 65–6, Virgil has by now made sufficient progress in writing the Aeneidthat Propertius can allude to its opening lines (compare Troiani... arma... Lauinis moenia litoribusin Propertius 2.34.63–4 and arma... Troiae... Lauiniaque... litora... moenia in A.1.1–7).

⁵³ A remedy adopted by some editors has been to keep Vergiliowhile altering me iuuetto mi lubet(Housman); this creates an ungainly and awkwardly long sentence, and leaves the mystifying and irrelevant custodisintact.

⁵⁴ I do not know on what grounds C. Heimreich as reported in Smyth′s Thesaurus Criticus(above, n. 37), proposed the deletion of 61–80 as an interpolation, but that deletion would discard part of the genuine Propertian context (61–4) as well as breaking up the obvious unity of the ‘Virgilian’ insertion by detaching 81–4. (Smyth also reports that Heimreich proposed to transpose 83–4 to after 78; it is not clear whether this is part of the same proposal to delete 61–80 or a completely independent one.)

⁵⁵ Given the contrast between past and present here, emphasized particularly by at nuncin 47, one should perhaps emend 42 by reading ueterumfor uestris;i.e. ‘all the deities of the countryside offered propitious words at the altars of the folk of old; but now the groves are deserted and the shrines abandoned’, etc.

⁵⁶ In his Emendationes Propertianae,at JPh16 (1888) 11 (= J. Diggle and F. R. D. Goodyear [edd.], The Classical Papers of A. E. Housman[Cambridge, 1972], 1.36), Housman commented laconically, ‘III xiii 43–61 fear have no business here’, but he offered no reasons for his view and apparently accepted Propertian authorship.

⁵⁷ This point has been made forcefully and persuasively by S. J. Heyworth in ‘Propertius: Division, Transmission, and the Editor′s Task’, PLLS8 (1995), 165–85 at 171–5.

⁵⁸ So, for example, Hubbard (above, n. 14), 44–5, ‘The sad result is that of all the poems in a book of 1362 lines there are only eight, amounting in all to 276 lines, which both have a harmonious manuscript tradition about where they begin and end and which have not been linked with others or themselves split into two or more poems by editors from the fifteenth century on.’

⁵⁹ The debut of 2.7 as a unified poem seems to have occurred in Salamanca, Bibl. Universitaria 85 and Florence, Bibl. Laurenziana pi. 33,15 but only as an accident, to judge by the fact that it was ‘corrected’ in both places; the earliest occurrence of deliberate unification seems to be in Berlin lat. fol. 500, Pontano′s copy.

⁶⁰ Some sixteen manuscripts have incorporated this new division, while a number of others have it marked by later correcting hands; its earliest occurrence seems to be among the emanuscripts of shortly before the middle of the fifteenth century (London, British Library Harley 2574; Brescia, Bibl. Civica Queriniana A.VII.7; Leiden, Universiteitsbibliotheek Voss. lat. 0.13); for these manuscripts see Butrica (above, n. 16), 132–5.

⁶¹ This too occurs first in the e manuscripts (see n. 60 above).

⁶² These include Hamburg, Staats- und Universitatsbibliothek Serin. 139.4; Naples, Bibl. Nazionale IV.F.19; Florence, Bibl. Riccardiana 633; and Florence, Bibl. Laurenziana pi. 33,15.

⁶³ Leiden, Universiteitsbibliotheek Voss. lat. O.82; Salamanca, Bibl. Universitaria 245.

⁶⁴ These also include the emanuscripts (for which see n. 60 above).

⁶⁵ Leiden Voss. lat. O.82.

⁶⁶ This division, which seems to appear first in Berlin lat. fol. 500, could be another conjecture of Pontano; it is also found in some manuscripts, including the loosely related Vienna, Osterreichische Nationalbibliothek 3153, Parma, Bibl. Palatina 716, and Pesaro, Bibl. Oliveriana 1167, which are loosely related to Pontano′s copy.

⁶⁷ Vienna 3153 after correction.

⁶⁸ As the original reading of the manuscript in Florence, Bibl. Laurenziana pi. 38,37; Leiden, Universiteitsbibliotheek Voss. lat. O.81; and Oxford, Bodleian Library Canon, class, lat. 31. The commentary of Gaspar Manius in Vat. lat. 1612 has a note arguing against the division.

⁶⁹ In t h e two closely related manuscripts Bergamo, Bibl. Civica Angelo Mai 27.2.33 a n d British Library Harley 5246.

⁷⁰ In a group of related manuscripts of certain or probable Ferrarese origin, former Abbey 3242; Modena, Bibl. Estense a.T.9.17; and Berlin, Staatsbibliothek Diez B. Sant. 57; also as a late correction in Vienna 3153.

⁷¹ Laurenziana pi. 38,37; Leiden Voss. lat. O.81.

⁷² For the affiliations of the incunabula, see Butrica (above, n. 16), 159–69.

⁷³ Heyworth (above, n. 57), 173–5 records some instances where these divisions have been suspected.

⁷⁴ My suspicions are roused particularly by the poems late in Book 2 involving such Greek pseudonyms as Panthus, Demophoon, and Lynceus; why should such names be used only in one part of one book?

⁷⁵ Briefer versions of this paper were originally delivered at the annual meeting of the Atlantic Classical Association in St John′s, Newfoundland in 1991 and at the annual meeting of the Classical Association of Canada in Charlottetown, Prince Edward Island in 1992.1 should like to thank Dr Heyworth and the Anonymous Referee for their stimulating and constructive comments upon the original draft.