¹ Flavii Sosipatri Charisii Artis grammaticae libri V, ed. Barwick, K. (Leipzig, 1925, repr. with corrections by F. Kühnert, Leipzig, 1954)Google Scholar. This edition is a great improvement on that of Keil, H. in Gramm. Latini, i.1–296 (Leipzig, 1857, repr. Hildesheim, 1981).Google Scholar

² I am preparing a monograph on Romanus' introduction and other parts of his work.

³ A quotation in Greek would heighten the admiration for Romanus' erudition – a third implication. Also at other places Romanus has quotations in Greek, e.g. 149.22ff. but these are more to the point.

⁴ The correction substantia〈m〉 is mine. Et caetera quae Graece sequuntur is found in one MSS family and is, of course, an indication that the scribe left out Greek words, perhaps because he could not read them.

⁵ In the margin of the Greek text stand the following characters ΜΑΣ ΕΙΝ Τϒ ΤΑ ϒΠΟ ΤΩΝ ΝΕΩΝ (Corr. ΝΕΡΙΩΝ) ΤΟϒ ΛΟΓΟϒ ΙΑΝΑΟϒΜΕΝΑ, which Keil (GL i.609) interprets from τ onwards as: τ ὑπ τν νερων το λγον διαοομενα. David Runia (private communication) suggests νομζεται δ' ὑπ τν νον (i.e. recent commentators) .τ το λγον διαροὑμενα (or: διανοοὑμενα) or something like τ ὑπ τν μερν το λγον διανοοὑμενα. I shall be grateful for further suggestions. For the second definition Runia suggests κατ τν ϕαιν δς πἱ τν οὐσα ϕέρονσα πρτερον 〈ἒπειτα π τ σνμβεβηκτα〉 and compares Alcinous/Albinus, Didasc. v init. τς διαλεκτικς δ στοιχειωδστατον γεῖται πρτον μν τ τν οὺσαν πιβπειν …, ἕπειτα περ τ⋯ν σνμβεβηκτων

⁶ Cf. e.g. Dillon, J., The Middle Platonists (London, 1977), pp. 49–51, 121f. and 233–6.Google Scholar

⁷ Froehde, O., De C. Iulio Romano Charisii auctore, Jhb. f. class. Philol., Suppl. xviii (Leipzig, 1892), p. 650.Google Scholar

⁸ The absence of discussion is not surprising, for in his edition (p. 193.25) Keil did not yet give the Greek words. They were printed in his Addenda from the annotations written by Ioannes Cauchius (c. 1540) in the margin of his copy of the editio princeps (cf. pp. xxiv–xxvii). Cauchius based these annotations on a MS. which disappeared during the sixteenth century. Barwick was the first editor to make full use of Cauchius' notes. See his praef. pp. xiiff. and ‘Zur Geschichte und Rekonstruktion des Charisius-Textes’, Hermes 59 (1924), 322ffGoogle Scholar. When Banvick's edition appeared the study of ancient grammar had started a decline which was not halted before the sixties.

⁹ Correspondence with the editor of the Dict, des phil. antiques in 1988–9 led to the inclusion of Romanus' Aquila under the relevant entry (see note 19 below).

¹⁰ Tolkiehn, J., ‘Die Lebenszeit des Grammatikers Charisius’, BPhW 30 (1910), 1054fGoogle Scholar., cf. BPhW 35 (1915), 188f.Google Scholar

¹¹ Stein, A., Hermes 63 (1928), 480–1Google Scholar. Cf. Parsons, P., JRS 57 (1967), 134ffGoogle Scholar. and Prosop. Imp. Rom. (PIR)² v. no. 247 (I owe the latter reference to Dr K. Worp).

¹² PLond. 3.1157, PWisc. 2.86, POxy. 17.2123, POxy. 33.2664, POxy. 1.78 and Bodl. Gr. lnscr. 3018. See also Kaster, R. A., Guardians of Language. The Grammarians and Society in Late Antiquity (Berkeley–Los Angeles, 1988), pp. 427f.Google Scholar

¹³ In Prosopogr. of the Later Roman Empire (PLRE), i.800 it is left open that the one may, be a son of the other, but in PIR v. no. 247 the identification is accepted sine dubio.

¹⁴ Romanus is not in the habit of adding titles, and Stein therefore supposes that he knew Marcius Salutaris personally.

¹⁵ Kaster, op. cit. (n. 12), p. 428 neglects this fact and, therefore, thinks it also possible that the Salutaris of Romanus was a son of the Salutaris of the papyri, or even an ancestor.

¹⁶ See also Kaster, op. cit. (n. 12), pp. 424ff. The statement in PIR iv. no. 520 paulo post medium saeculum tertium is based upon the inference by Stein. In this connection it is unimportant that Jeep, L., Rhein Mus. 51 (1989), 401ff.Google Scholar, argues that the parts from Romanus were added later after Diomedes had used Charisius. For even if he should be correct (but see Barwick (above n. 8) 335ff.), it would be wrong to date Romanus later than Charisius, as Tolkiehn, J., BPhW 28 (1908), 1168 rightly remarks.Google Scholar

¹⁷ Gloeckner, St., Quaestiones rhetoricae. Historiae artis rhetoricae qualis fuerit aevo imperatorio capita selecta, Breslauer philol. Abhandl. viii, 2 (Breslau, 1901), pp. 64ffGoogle Scholar. (p. 69: ‘Quod attinet ad Euagorae et Aquilae aetatem, non ante annum 250 eos floruisse coniecerim’); Schilling, L., Quaestiones rhetoricae selectae, Jhrb. f. class. Philol., Suppl. 28 (1905), pp. 693ffGoogle Scholar. On George and Nilus see Kennedy, G. A., Greek Rhetoric under Christian Emperors (Princeton, 1983), pp. 115 and 313f.Google Scholar

¹⁸ ‘Zwei Identificationen’, Hermes 42 (1907), 548–61.Google Scholar

¹⁹ Dictionnaire des philosophes antiques (Paris, 1989), vol. i, s.vGoogle Scholar. Aquila (Akylas). In the final version of his article Pernot discusses an earlier draft of my article, in which I had put Aquila earlier than A.D. 270 -Kennedy, op. cit. (n. 17), pp. 74f. says ‘they [Evagoras and Aquila] cannot be dated securely before or after Porphyry’, but does not seem to be aware of Keil's discussion.

²⁰ Cf. Kaster, op. cit. (n. 12), p. 425.

²¹ Application of the principle that persons of the same name ought not to be multiplied beyond necessity cannot prove, of course, that I am right. Moreover, as the referee of my paper reminds me, in the third century there are numerous cases of two contemporaries possessing similar traits and identical names, e.g. two Origens, both Platonists and both pupils of an Ammonius; two men named Dionysius, both bishops.

²² But without a reference to Evagoras, although, to all appearances, both texts have as their source the same text, Evagoras' περ τ⋯ν στσεων (Syrianus 3.24).

²³ E.g. Julian us. Or. v. 162c says τοιατα γρ γὡ μμνημαι το Ξενρχον λγοντος κηκοώς. Xenarchus lived in the time of Augustus. I borrow this example from Anne Sheppard, D. R., Studies on the 5th and 6th Essays of Proclus' Commentaries on the Republic (Goötingen, 1980), p. 31 n. 29, but there are many more examples. I treat this matter in a forthcoming article.CrossRefGoogle Scholar

²⁴ See Russell-Wilson on Menander Rhetor, pp. 295 and 300 and Russell, D. A., Greek Declamation (Oxford, 1983), pp. 77ffCrossRefGoogle Scholar. for more information on extempore speeches in late antiquity and their links with philosophical διατριβα.

²⁵ Because of the many details Evagoras gives, the same scepticism cannot be extended to his declaration of having personally witnessed the sophist's performance. But in that case it would be easier to identify the sophist with the grammarian Phrynichus, an identification defended by Rabe, H., RhM 62 (1907), 260 n. 1Google Scholar. The latter Phrynichus dedicated a praeparatio sophistica to the Emperor Commodus. However, to date Evagoras as early as c. 190 seems quite wrong.

²⁶ Thus PLRE i.284 on Evagoras: ‘…he was probably a Neoplatonist, and therefore [my italics] at earliest late third century and probably early fourth century’. Similarly on Aquila (i.90), who because of his being a pupil of Evagoras has to be dated accordingly.

²⁷ Cf. Kennedy, op. cit. (n. 17), p. 78.

²⁸ Under the entry ‘Aquila E’ Pernot discusses the possibility that Syrianus' Aquila and his teacher Evagoras are mentioned by Libanius, who in 363–4 addressed three letters (Epist. 809, 1132 and 1440; cp. 137 and 1351 Förster) to a certain Evagoras, a younger colleague of Libanius, and in 392 a letter (ibid. 469, 4 and 1130) to an Aquila, who at that time is a young man and is encouraged to finish the long enterprise he has started. Pernot himself concedes, however, that Libanius neither says that Evagoras and Aquila know each other nor indicates the nature of Aquila's work.

²⁹ See e.g. Dillon, op. cit. (n. 6), pp. 180, 228, 235 and 262, and Graeser, A., Plotinus and the Stoics (Leiden, 1972).CrossRefGoogle Scholar

³⁰ In the same way as the Aquila mentioned in the Suda (above) and the Aquila referred to by Proclus have been identified by others with Syrianus' Aquila. See Pernot s.v.

³¹ I am grateful to David Runia and the anonymous referee of this journal for their comments on an earlier draft of this paper.