INTRODUCTION

Very soon after I began working on the identity of magic-workers in classical antiquity, I realized that it was necessary to come to terms with a thesis about depictions of erotic magic-working in Greek and Roman literature. It asserted that male writers engaged in a systematic misrepresentation of the realities of magic-working in portraying erotic magic as an exclusively female preserve; the reality was that men were the main participants in this form of magic-working. The thesis is based on the supposition that the truth about erotic magic and the people who performed it is to be found in the formularies or spell-books preserved in papyrus and in defixiones. These two sources of information are said to show us that erotic magic was performed by men and not by the women who are the persons depicted engaging in love-magic in literature. The scholar who first presented the thesis was the late John Winkler. A version of it is to be found in Fritz Graf's general account of Greek and Roman magic. There is agreement over what are taken to be the facts, but views diverge over their interpretation. Winkler appeals to the Freudian notion of denial and transference to offer an explanation not only of the discrepancy between life and literature, but of what he took to be the belief held by the young men who cast erotic spells that the girls who were the objects of their spells were as sexually eager as they were: men, when overwhelmed by sexual desire for unattainable women, through a process of denial transfer that feeling to women, whether old or young, whom they fondly imagine suffer the same intense sexual longings as themselves.

A neglected source for the text of Marcus Aurelius’ Meditations is to be found in the writings of the Byzantine theologian Joseph Bryennius, who seems to have been born about 1350 (details of his early life are obscure) and to have died before the Council of Florence (1438), probably in 1430/1. He was a monk who was also a scholar, a theologian, and an ecclesiastical diplomat. He spent the years 1382–1402 in Crete (then under Venetian rule), and was sent in 1406 on a mission of ecclesiastical diplomacy to Cyprus. Otherwise the greater part of his life was spent in Constantinople; from about 1402 to 1406 he lived at the monastery of Stoudios, from 1416 to 1427 at that of Charsianeites. He was a court preacher, and as a theologian upheld the claims of the Greek Church against the Roman; among his published works are twenty-one Discourses on the Trinity maintaining the Greek Orthodox position.

In 1023ff. the poet explains that he has not been spoiled by success. The verb ༐κτελσαι in 1024 has been suspected, and though recent editors accept it, taking it as absolute, I am far from convinced that it is what the author wrote. Blaydes, in his usual fashion, records conjectures and makes some of his own, but though he hits the mark quite often in Aristophanes as he does in Sophocles, in this passage his efforts, e.g. ༐κγελσαι, fail to satisfy. I propose instead ༐κχαλσαι, ‘relax’. It might be transitive, but I slightly prefer to take it as intransitive. It is a rare word and all the more likely to be corrupted. The best parallel I can find is Hippocrates, De Octimestri Partu 1.2 (ed. Joly, 5 ed. Grensemann): οἱυες ༐υ oἶσι τυ ρχυ ༐τρøη, ὣσπερ τυ σταχωυ, ζεχλασαυ πρσθευ υαγκαζώμευου 祴 τελεως ζαδρυυθυαι τυ καρπυ.

The verb is correctly transmitted in the best MS, Marc.gr. 269, while the other main source of the treatise, Vat.gr. 276, offers ζεκλεσαυ9, which is clearly due to an error by a scribe who did not recognize the unusual word required by the context. As to the date of this treatise, it is believed by one of the best modern authorities to be no later than the end of the fifth century and is therefore nearly contemporary with Aristophanes.

A much-discussed fragment of Phrynichos’ comedy Monotropos, together with the comments of the scholiast on Aristophanes who preserves it, have often been taken to indicate that at some point before the production of the play, in spring 414 B.C., the Athenian politician Syrakosios moved a decree that restricted the right of comic playwrights to satirize individual Athenians. The relevant passage (schol. ad Arist. Aves 1297 = Phrynichos F 27 K–A) reads as follows (I reproduce the text of Kassel and Austin):

This is James Diggle's OCT (1994), with a modified apparatus. The lacuna at 186 prompted two conjectures by Willink: υτ∊ῖθ∊υ and the ingenious αὐλθ∊υ (‘from the palace’; unknown in extant Greek). I wonder, however, whether an adverb is what we want: the anadiplosis of ἓλακ∊υ (giving a characteristic Euripidean jingle) would not have come amiss to lend a touch of hysterical urgency to the cry of the rapist's victim; but that would not give us the rhythm we need (––). Something approaching the effect of the suggested anadiplosis might be obtained by the supplement πoλλοῖσιυ. Although it lacks the palaeographical plausibility of αὐλθ∊υ, it gives us an expression that is characteristically Euripidean in contexts pertaining to souls in torment: Ale. 185 πoλλυ δακρωυ, 938 πολλυ δ μχθωυ Su. 1119 ἂλγ∊σι πoλλοῖς, Tr. 28 πολλοῖς δ κωκυτοῖσιυ, 38 δκρυα χουσα πολλ, etc. The adjective applied to sound is not unknown:Rh. 290–1 πολλι γᾰρ ἠχι Θρικιοςῥωυ στρατς1ἓστ∊ιχ∊. It can also be used to describe the frequency of sung laments accompanied by hysterical self-mutilation: S. El. 88–90 πολλς μυ θρυωυ ὡιδς, 1 πολλᾰς δ’ υτρ∊ις ισθxυ 1 στρυωυ πλαγᾰς αἱμασσομυωυ.

Stephen Tracy's neat demonstration that IG I3 35—authorizing the building of a temple and appointment of a priestess for Athena Nike—was cut by the man responsible for the Promachos accounts (IG I 435) at first seemed decisive for the traditional c. 448 B.C. against my radical down-dating. Ira Mark then argued that this decree provided for the naiskos and altar of his Stage III in the 440s: the marble temple belonged to Stage IV over twenty years later. Despite these two powerful interventions the matter is not closed. David Gill has, I fancy, convincingly refuted Mark on archaeological and architectural grounds. And there is still more to be said from the epigraphic angle.

IG I 36, cut on the back of the stele, looks like a delayed rider to 35. But just how delayed was it? It arranged for the regular payment of the priestess's salary by the kolakretai in office in the month Thargelion. On the traditional view the gap would be close to a quarter of a century, since 36 is firmly dated 424/3 B.C. This is quite extraordinary, though reasons have been found for it. More serious perhaps is some neglected epigraphic evidence. We have eighteen other examples in fifth-century Attic epigraphy where decrees are followed on the same stone by other texts; but virtually all the gaps are short, never more than a few years. The relevant texts are IG I 4, 11/12, 41, 42/43, 52 A–B, 59, 61, 66, 68, 71, 72, 73, 89, 93, 101, 127/II1, 156, 1454. It is true that 42/43 are dated c. 445–442 and c. 435–427 B.C. in IG I, but this is quite arbitrary.

Besides the direct tradition, these verses are cited by Stephanus of Byzantium, s.n. Aῖγυς, who explains: λγος π∊ρ ‘Eλυης Λακωυικς οὒσης καῷ ἂρρ∊υ μτ∊κοσης κα τ ’Aλ∊ζυδρω κα Δηϊøβω γαμηθ∊σης Commentators have followed him both as to the identity of the three husbands and the sense of θηλπαις: ‘female-childed’.

Did the Carthaginian cavalry general Maharbal really urge Hannibal to march on Rome after Cannae, and then comment bitterly Vincere sets, Hannibal; victoria uti nescis, when his leader refused? There are two main objections: Maharbal may not have been there, and anyway Cannae was too far away to justify such a march. The whole story has been seen as one of those well-known Roman historiographical inventions. But there may well be more to the story than that, illuminating both Hannibalic history and some Roman historiographical techniques.

‘Excitement struggles with the restraint of form and language and the artifice of verbal repetition… runs riot.’ The repetition is more pronounced and personal here than in another Lesbia epigram, no. 70, where ‘the repetition dicit…dicit makes it certain that Catullus had [Callimachus, Ep. 25 Pf.] in mind’. Poem 70 illustrates how Catullus might allude to and adapt a Hellenistic model in expressing his personal feelings; while the longer elegiac poems in particular (and 66, the translation of Coma Berenices) show the depth of his engagement with Callimachean literary technique. We should not be surprised to find Callimachean elements here too, given the demonstrable correspondences with poem 68 in particular, a composition noted for its use of Alexandrian artifice. But while there are close echoes of the high emotion, the doctus poeta of 68 seems to be largely missing from 107. Here Catullus exults ipsa refers te / nobis (5–6); there his mistress se nostrum contulit in gremium (132).

Juturna drives Turnus’ chariot now here now there, hoping to throw off Aeneas’ pursuit, but he follows the twisted circles (tortos orbes, 12. 481) of her course. Virgil compares her to a black hirundo flying through a rich man's house out into the colonnades and then round the pools or fishtanks. Hirundo can mean swallow, martin, or even swift. All these birds eat insects and air-borne spiders; they do not eat human food. The common swallow chiefly eats flies, and feeds the nestlings on flies; it also eats wasps and bees. Its average prey size is much greater than the house martin's. Virgil's hirundo gathers pabula parua for the nestlings. W F. J. Knight in the Penguin translation writes ‘tiny scraps of food’; C. Day Lewis translates ‘crumbs of food’. If Virgil meant scraps of meat or crumbs of bread, stolen from the rich man's dinner table, then Virgil did not know what these birds eat.

Names of Romans in Josephus are notoriously liable to corruption. Two minor characters in his account of the assassination of Caligula have so far defied plausible emendation, ‘Timidius’ in A.J. 19.33–4 and ‘Bathybius’ in 19.91. The sources of Josephus’ account of this dramatic episode were unquestionably high class—two, rather than one, Latin historians, as Wiseman has demonstrated, the main one (rather than the only one) being Cluvius Rufus, the other possibly Fabius Rusticus.

The most recent and by far the best edition of this work is that of H. M. Hine (Stuttgart/Leipzig, 1996), to which I refer for full bibliographical information. Many passages of the text are most helpfully discussed in the same scholar's Studies in the Text of Seneca's Naturales Quaestiones (Stuttgart/Leipzig, 1996).

ut nubes infici possint, … sol ad hoc apte ponendus est; non enim idem facit undecumque effulsit, et ad hoc opus est radiorum idoneus ictus.

Seneca is dealing with rainbows. Hine (Studies, 24–5) shares Axelson's suspicion of ictus, but is unhappy both with Axelson's situs (‘seems too static a word for the sun's rays’) and with my tractus, ‘direction’ (‘in this sense seems to be used only of concrete physical objects’); very tentatively he suggests angulus. Much more credible palaeographically and still yielding good sense and a good clausula would be i<mp>etus, a noun which is very common in this work of Seneca's.

Hostius fuit Quadra obscenitatis in scaenam usque perductae.

The process of giving greater credence to the codices has been constant in editions of Apuleius from the last century to the present day. Accordingly, the following suggestions are proposed.

I

et statim miser ut cum ilia adquieui, ab unico congressu annosam ac pestilentem con<dicionem> contraho.

And at the very moment when I—wretch that I am!—lay with her, from that one relationship, I contracted this interminable and miserable condition.

condicionem contraho scripsi: con contraho Fø: coniunctionem c. Chodaczek: consuetudinem c. van der Vliet: contagionem c. Lütjohann: cladem c. Helm: luem c. Heinsius: contraho adfectionem Novák con dittographiam putans, cf. 8.14.1 (187.22): noxam ante annosam Giarratano, post annosam Birt

In this case, where a word is obviously missing and the mysterious con appears, there are two general possibilities for correcting the text: con may be taken as a dittograph for contraho or considered to be part of the missing word. Another approach would be to interpret contraho differently: the verb can apply to a disease (as proposed by Lütjohann and Heinsius—followed by Hildebrand—no doubt swayed by pestilentem), harm or misfortune (meaning: ‘draw misfortune upon oneself’, as in the later sentence nequam tibi lingua intemperante noxam contrahas [Helm, Giarratano, and Birt]), a custom (more or less following van der Vliet and Novák) or marriage (Chodaczek). The solutions offered by van der Vliet, in particular, and Novák are reasonable: other commentators have been led astray by the sentence I have quoted, assuming that what is missing is something like ‘harm', ‘misfortune’.

It is usually assumed that Apuleius gave one of his characters the name ‘Thrasyllus’ because of its etymological connection with θρασ. Indeed it is singularly appropriate and Apuleius himself draws attention to the fact: Thrasyllus, praeceps alioquin et de ipso nomine temerarius… (Met. 8.8). However, it does not follow that a name with such an etymological significance can have no other connotations: in this note I suggest that there is a further frame of reference behind ‘Thrasyllus’ and that Apuleius may have expected his readers to realize this.

Right so far as Homer is concerned, and Paulus, a poet of Justinian's court best known for his epic poem (modelled after Homer) composed on the occasion of the rededication of the Church of St Sophia, clearly evokes Callinus. But the commentators have overlooked the pointed use of μχρι(ς) τυος + the present indicative in Achilles Tatius’ τᾰ κατᾰ Λευκππην κα κλειοøντα. Examination of the examples there suggests that Achilles Tatius could make greater demands on his readers than is sometimes generally assumed for the Greek novelists.

Little, if anything, in Justin scholarship has been as controversial as the dating of the so-called Epitome of the Philippic History of Pompeius Trogus. Suggested dates have varied from the time of Antoninus Pius through the third century to the end of the fourth. The latter was proposed in 1988 by Sir Ronald Syme, but has in fact received little support in subsequent literature on Justin, which has tended to accept the earlier dating (late second/early third centuries). An exception is T. D. Barnes, who has voiced support for a later dating based on a linguistic parallel to Justin in the Historia Augusta.

Barnes observes that Oscar Hey, author of the article ducatus in TLL (5.2129.30–2131.42), drew attention to the similarity between Justin 30.2.5 Agathocles regis lateri iunctus civitatem regebat, tribunatus et praefecturas et ducatus mulieres ordinabant and HA Heliog. 6.2 militaribus… praeposituris et legationibus et ducatibus venditis. Hey, he notes, refers at the head of that particular section of the article (Section 2 [2130.5–63]) to Seeck's article on dux. Seeck had demonstrated—and this is now taken for granted by scholars—that from the time of Diocletian dux is used technically as a formal title, and, in drawing attention to the parallel between HA Heliog. 6.2 and Justin, Hey must have been intimating that Justin is, like the author of Heliog. 6.2, using ducatus as a (post-Diocletianic) formal title. He then seems to suggest that Hey was right and Justin is to be dated to at least some time after 260.