I thank Prof. P. A. Brunt, Mr E. L. Bowie and an anonymous referee for comments on an earlier draft.

¹ Abbreviations are as follows:

Bosworth 1972: Bosworth, A. B., CQ xxii (1972) 167 f.Google Scholar, 174 f.

Bosworth 1980: id., A historical commentary on Arrian's History of Alexander (Oxford 1980) 11, 104 ff.

Breebaart: Breebaart, A. B., Enige historiografische aspecten van Arrianus' Anabasis Alexandri (Leiden 1960) 23–7Google Scholar

Brunt 1976: P. A. Brunt, Loeb Arrian i (1976) 53

Brunt 1983: id., Loeb Arrian ii (1983) 534–41

Schepens: Schepens, G., Ancient Society ii (1971) 254–68Google Scholar

Stadter 1980: Stadter, P. A., Arrian of Nicomedia (Chapel Hill 1980) 61 ffGoogle Scholar. (cf. my review in JHS cii (1982) 254–5Google Scholar, which anticipates some of the arguments of the present paper)

Stadter 1981: id., Ill. Cl. St. vi 1 (1981) 157–71

Wirth: Wirth, G., Historia xiii (1964) 224Google Scholar.

I cannot accept the new idea of Bosworth 1980, 7 f., that Anabasis may not even be the work's correct title, though this hardly matters here.

² On the text see Bosworth 1980, 102–7.

³ τῶν πρώτων is certainly neuter: Bosworth 1980, 107.

⁴ Cf. Bosworth 1980, 100 f.; Stadter 1981, 166.

⁵ Cf. Breebaart 25; Schepens 263; Stadter 1980, 63.

⁶ Cf. Bosworth 1980, 104.

⁷ Cf. Wirth 225; Stadter 1980, 63.

⁸ μακαρισμός: cf e.g. Isocr. Evag. 70 ff, Xen., Ages 10.3–4Google Scholar; ‘worthy commemoration’: Evag. 2, 40, Ages. 1.1; ‘making clear’: Evag. 33, 65, Ages. 1.6; σύγκρισις: Evag. 37 ff. Ages. 1.6, Men. Rhet. ii 377. General discussion: Bosworth 1980. 14 ff., 30 f.

Arrian does not invoke the tradition of Pindar and Bacchylides (pace Stadter 1980, 63; Bosworth 1980, 104): i 12.2 οὐδὲ ἐν μέλει is dismissive (à la Thucydides i 20?)

⁹ Metaphorical use: LSJ s.v. A II 2 and B 2; note that ὁρμηθη̑ναι is picked up at vii 30.3 (ring composition: Breebaart 27). Arrian's play on ὁρμάομαι perhaps owes something to Herodotus' analogy (i 5.3) between the ‘progression’ of his History and Odysseus' ‘travels’.

¹⁰ Cf. e.g. Xen. Ages. 1.1 and often in λόγοι ἐπιτάφιοι.

¹¹ Stadter 1980, 64; Brunt 1976; Bosworth 1980, 106; a survey of earlier versions in Schepens 258 ff.

¹² Cf. below ἐκει̑νο ἀναγράφω and Dem. i 19 περὶ . . . τη̑ς βοηθείας ταυ̑τα γιγνώσκω(ταυ̑τα referring back).

¹³ Cf. Schepens 265 f., Bosworth 1980, 106; Stadter 1980, 64. Arrian is presumably also polemicising against the excessive self-advertisement of some contemporary historians (Lucian Quomodo hist. 14 ff.), perhaps specifically Appian (below). He can hardly suppose personalia irrelevant when a historian records the remote past (pace Breebaart 17; Schepens 266): he cites them in the Bithyniaca proem (cf. n. 37); nor is he in this respect imitating Homer and Xenophon (pace Breebaart 24 f): their matter-of-fact omission of personalia is quite different from his ostentatious recusatio. Nor is he proudly silent because his Roman career ended abruptly (pace Schwartz, E., RE ii [1895] 1231Google Scholar): he boasts, implicitly, of his political eminence (below), no Roman allusion is required in context at all (below), nor is there evidence of such an abrupt end: Bosworth 1980, 4; Syme, R., HSCP lxxxvi (1982) 206Google Scholar.

¹⁴ Rome: Stadter 1980, 181, 212 n. 19; Brunt 1983, 538 f; Athens: Wirth 224; Nicomedia: Bosworth 1972, 174; 1980, 106.

¹⁵ Cf. my discussion in LCM viii 9 (1983) 130 fGoogle Scholar.

¹⁶ Teles 23, 9 Hense with O'Neil, E., Teles: the Cynic teacher (Missoula 1977) xivGoogle Scholar.

¹⁷ Brunt 1983, 538.

¹⁸ Alex. 48; Quomodo hist. 5, 17, 29.

¹⁹ Bosworth 1972, 174 n. 6, 1980, 11, argues that τ̢η̑ ἐμαυτου̑ cannot be Rome because no contemporary Greek would describe Rome as his πατρίς, but this ignores the fact that most of those who argue for a reference to Rome distinguish τ̢η̑ ἐμαυτου̑ from Arrian's πατρίς.

²⁰ Brunt's view (1983, 538 f.) that not only τ̢η̑ ἐμαυτου̑ but also πατρίδα denotes Rome is obviously untenable: while a Greek might perhaps describe Rome as his πατρίς in certain circumstances, he would certainly not do so in answer to the question: ‘who are you, what is your πατρίς and γένος?’ That question is about ‘roots’.

²¹ Bosworth 1972, 174 n. 6.

²² Brunt 1983, 539 emphasises this point. Bosworth's argument (1980, 106) that Arrian's 'name is perfectly familiar. He assumes that the reading public already knows his work and the simple fact of the name Ἀρριανός upon the title will recommend the book without further need for self-advertisement’ does not meet the difficulty, since (a) though ἄγνωστον formally only applies to ὄνομα, its implications clearly carry over to οὐδὲ πατρίδα . . . ἡ̑ρξα and (b) ‘not at all unknown among men’ entails ‘celebrated for rank’ (Brunt 1983, 539).

²³ Cf. Stadter 1980, 186 n. 6; Bosworth 1980, 106; Breebaart, , Gnomon liv (1982) 26Google Scholar.

²⁴ To later Greeks, that is. Arrian's historical narrative properly distinguishes Greeks and Macedonians (Brunt 1976, xxxvii and n. 33).

²⁵ E.g. in the (lost) σύγκρισις of Plutarch's Alexander-Caesar (for which cf. Appian, BC ii 149–54Google Scholar).

²⁶ Pace Bosworth 1972, 175.

²⁷ Bosworth 1972, 170 n. 7, 175 n. 2 (not seeing that this undermines his argument).

²⁸ Naturally I do not deny that in certain contexts Greeks might parade their Roman qualifications, even above their Greek ones (Brunt 1983, 539, citing C. Iulius Severus, OGIS 543/544). But some contexts make such behaviour inappropriate, just as in Britain today patriotic Scots, Irishmen and Welshmen who have won both local and national distinctions will often not mention the latter when appealing for favour on home ground.

²⁹ Colourless term: Praef. 1; LSJ s.v. ll 1; ‘inscribe’: LSJ s.v.

³⁰ Thuc. i 22.4; Pind. Pyth. 6.6 ff.; Hor., C. iii 30.1Google Scholar ff.

³¹ Stadter 1980, 64 f. and 212 n. 18; Diss. ii 22.16: Bosworth 1980, 106 f.

³² Bosworth, , Phoenix xxxvi (1982) 178Google Scholar.

³³ Bowie ap. Brunt 1976, 53 n. 4.

³⁴ Note also how the verbal paradox γένος . . . οἵδε οὶ λόγοι . . . ἀπὸ νέου ἐγένοντο further emphasises Arrian's disavowal of γένος in the usual sense.

³⁵ (i) Bosworth 1972, 168 (less dogmatic in 1980, 107); Stadter 1980,64 f., 212 n. 21; Brunt 1983, 538; (ii) Bowie, , P & P xlvi (1970) 26 f.Google Scholar; Schepens 264.

³⁶ ξυγγραφή: An. Praef. 3, i 12.4, vi 28.6, vii 3.1, 30.3; λόγος: 19.9, 32.1, 43.14; Cyn. 1.4; Tact. 32.3.

³⁷ Phot. cod. 93 p. 73b11 ff., ἤδε αὐτῷ ἠ συγγραφὴ ἐξεπονήθη καὶ ἐξ ἀρχῆς μὲν ἀφ᾿ οὖ γράφειν ἴσχυσε ταύτην ἐνστήσασθαι καὶ συντάξαι τὴν ὑπόθεσιν βουληθέντι.

³⁸ Cyn. 1.4, ἀμφι ταὐτὰ[τω̨̑ Ξενοφω̑ντι] ἀπὸ νέου ἐσπουδακώς, κυνηγέσια καὶ στρατηγίαν καὶ σοφίαν makes a similar claim to both Bithyniaca and Anabasis.

³⁹ Note also that the wording ὄνομα . . . οὐδὲ ἄγνωστον ἐς ἀνθρώπους(i 12.5) seems to be picked up at vii 30.2 (referring to Alexander).

⁴⁰ Stadter 1981.

⁴¹ Bosworth 1980, 106.

⁴² 30: Xen., Mem. i 2.35Google Scholar; 40: Pl. Leg. 951e, cf. 666b.

⁴³ Cyn. 1.4(n. 38); 140 and 55 are round figures; on the chronology of Arrian's life see now Syme (n. 13) 181–211.

⁴⁴ Bosworth 1972, 168 n. 1 (especially as ἀπὸ νέου does not mean ‘right from childhood’).

⁴⁵ Bosworth's contention (1972, 168) that it ‘was only a parergon, one of the works he undertook for practice in handling non-contemporary material’ is wholly untenable.

⁴⁶ As Mr E. L. Bowie suggests to me.

⁴⁷ Literary polemic: general discussion in Bosworth 1980, 12; date of Appian's Emphylia; Gabba, E., Appiani Belhrum Civilium Liber Primus (Firenze 1958) x–xiGoogle Scholar; Bosworth 1972, 178 (c. 161–3); Bowie prefers an earlier date.

⁴⁸ Cf. Bosworth 1972, 176 ff. (Bowie disagrees).