¹ Matthews, J. F., The Roman Empire of Ammianus (London, 1989), 462–4.Google Scholar

² Bowersock, G. W., reviewing The Roman Empire of Ammianus, in JRS 80 (1990), 244–50Google Scholar; Fornara, C. W., ‘Studies in Ammianus Marcellinus, I: The letter of Libanius and Ammianus' connections with Antioch’, Historia 41 (1992), 328–44Google Scholar; Barnes, T. D., ‘Ammianus Marcellinus and his world’, Classical Philology 88 (1993), 55–70, at 57–61CrossRefGoogle Scholar. Fornara's article is a revised version of the paper given at Oxford in February 1987, referred to at The Roman Empire of Ammianus, 478 n. 1.

³ The text and a translation of Libanius' letter are also provided by Fornara, 331–2, and a translation by Barnes, 58; see also, assuming the letter to be written to Ammianus, Norman, A. F., Libanius: Autobiography and Selected Letters (ed. Loeb, , 1992), No. 188 (Vol. II, pp. 423–33)Google Scholar. No serious difference of interpretation revolves on a question of translation as such. Fornara's text is also that of Foerster, with the exception of two points of punctuation, which I take to be misprints.

⁴ This phrase is to express the Greek μῖν (not translated by Fornara). Barnes has ‘I gather’.

⁵ Norman has a different nuance; ‘each published portion wins approval and invites another’.

⁶ Cf. Fornara, ‘has rendered the verdict’. Barnes' translation, ‘has formally recognized’, seems too explicit.

⁷ Or (Fornara 332, cf. 337), ‘bringing your inventions into public recitations’. Barnes, reading οἵκοθεν, has ‘from the study’.

⁸ Libanius could mean this literally; the next letter in the published editions, Ep. 1064 to Aristaenetus, begins with references to the deaths of Cimon and Calliopius, and to a weakness of the eyes, which ‘again cause tears to pour down upon my writing’.

⁹ Seeck, O., Die Briefe des Libanius (Leipzig, 1906; repr. Hildesheim, 1966), pp. 202Google Scholar (Marcellinus VII) and 463. The date 392 is accepted without query by Fornara and explicitly by Barnes. For early 393, see n. 11 below.

¹⁰ Seeck, 463; cf., on the death of Cimon, Sievers, G. R., Das Leben des Libanius (Berlin, 1868), 201Google Scholar; A.F., Norman, Libanius' Autobiography (Oration I) (London, etc., 1965), 231Google Scholar; Liebeschuetz, J. H. W. G., Antioch: City and Imperial Administration in the Later Roman Empire (Oxford, 1972), esp. 272Google Scholar; PLRE I, pp. 92f. (s. Cimon Arabius). The insult to Cimon referred to in Libanius' letter is apparently the last-minute withdrawal of the governorship of Cyprus, to which he had been nominated, cf. Or. I.283–4, again with the motifs of λπη and ὕβρις; Norman, , Libanius' Autobiography, 234Google Scholar, and in Autobiography and Selected Letters, Vol. I, pp. 334–7Google Scholar. The death of Calliopius is also referred to at Ep. 1051, 1064; PLRE I, p. 175 (Calliopius 4).

¹¹ Addaeus, Ep. 1062, cf. Die Briefe des Libanius, 463Google Scholar. At Regesten 100, cf. 282, Seeck gave reasons for amending the date of the law addressed to Addaeus (CTh 1.5.10 + 1.7.2) from 12 January to 12 June 393; he was still comes domesticorum on 31 December 392, cf. CTh 6.24.5 (where the MSS give 393); Regesten, 87, cf. PLRE I, p. 13.

¹² PLRE I, p. 779 (citing CTh 8.6.2 + 9.28.1), cf. p. 992 (s. Zeno 7). The exact references are to Libanius, Ep. 1052.3 and 1061.5.

¹³ Fornara, 333f., takes this phrase to refer to Marcellinus' ‘participation among the group of students taught by Libanius at Antioch’. This is a possible, but clearly not the only possible, reading of the text. It seems equally natural to take the first-person references throughout §§2, 3 & 4 of the letter (μῖν…λλ κα μς, ὡν στιν συγγραϕεὺς… μῖμ τοτο δδον) as all referring to Antiochene origin, which is made explicit in §4.

¹⁴ Priscian XI.51 (Keil, Grammatici Latini II, p. 487), on a question of orthography; ‘ut indulsi indulsum vel indultum. unde Marcellinus rerum gestarum XIIII: tamquam licentia crudelitati indulta’. The reference is to Amm. 14.1.4.

¹⁵ This point of chronology is not taken up by Fornara; see below, p. 262 with n. 52.

¹⁶ The Roman Empire of Ammianus, 8f.

¹⁷ Bowersock, G. W., JRS 80 (1990), at 245–6Google Scholar, records a suggestion that the destruction of the Serapeum was in 392, referring to Bauer, A. and Strzygowski, J., Eine Alexandrische Weltchronik (Denkschriften des kaiserl. Akademie der Wissenchaften in Wien 51; 1915), 66ffGoogle Scholar. (text at 74), which seems to place the destruction of ‘the temples of the Hellenes’ in that year. The conventional date of 391 for the destruction of the Serapeum is based on the involvement of the praefectus Augustalis Evagrius and the comes Aegypti Romanus, who were then in office (they are addressed together in CTh 16.10.11, of 16 June). The fragmentary ‘Alexandrian Chronicle’ has Evagrius in office from 389 to 392, but this is not correct either for the beginning or the end of his tenure; cf. PLRE I, pp. 268 (Evagrius 7) and 769 (Romanus 5) and the Fasti at p. 1085. According to Socrates, Hist. Eccl. 5.16, other temples in Egypt were destroyed after the Serapeum. The most notable were at Canopus, cf. Eunapius, Vit. Soph. pp. 471/2 (ed. Loeb, pp. 418–27). Also under 392 the ‘Alexandrian Chronicle’ locates the death of the usurper Eugenius, which was actually in 394. It does not seem wise to use this text to resolve difficulties in the other evidence. 391 is accepted by Fornara, 336 n. 16, and firmly supported by Barnes, 61–2.

¹⁸ Cf. Fornara, 338f.—a clear discussion, but not addressing all relevant passages, and (in my view, obviously) giving less than their full weight to some as indicative of an Antiochene background.

¹⁹ The Roman Empire of Ammianus, 412fGoogle Scholar. By the same token, the passage cannot be said, as it is by Fornara, Bowersock and Barnes, to disprove Ammianus' Antiochene origin. The context is the setting in authority over them of the disagreeable Alexander of Heliopolis as consularis (not ‘prefect’, as Fornara, 338 n. 21).

²⁰ The notion, criticised by Fornara, 328 n. 4, that Jovinianus was educated at Antioch, is speculative, but not arbitrary; it is based on the general experience of young hostages in such circumstances, cf. the excellent discussion (drawing material from an earlier period) in David, Braund, Rome and the Friendly King: the Character of the Client Kingship (London, etc., 1984), 12–17Google Scholar; and for the late empire Lee, A. D., ‘The role of hostages in Roman diplomacy with Sasanian Persia’, Historia 40 (1991), 366–74Google Scholar; see too my ‘Hostages, philosophers, pilgrims and the diffusion of ideas in the late Roman Mediterranean and Near East’, in Clover, F. M. and Humphreys, R. S. (edd.), Tradition and Innovation in Late Antiquity (Madison, Wisconsin, & London, 1989), 29–49Google Scholar, at 38–41. I suspect that Jovinianus was sent to Antioch for ‘acculturation’ but returned to Corduene to succeed to the satrapy.

²¹ The Roman Empire of Ammianus, 74–80Google Scholar; on Jovinianus, 44, 55–7. See also below, p. 268 with n. 76.

²² ‘ut in tenebrosis rebus confusione cuncta miscente, summatim quia nos penitissima gestorum memoria fugit, quae recolere possumus, expeditius absolvemus’.

²³ Cf. Fornara, at 339. My argument takes the passages in connection with the other references discussed above, especially the allusion to ‘our Hypatius’. A particularly vivid moment in Ammianus' narrative is the unusual direct address of the defendant Hiiarius to the judges at Antioch; ‘construximus’, inquit, ‘magnifici iudices’, etc. (29.1.29, with The Roman I Empire of Ammianus, 221).

²⁴ This is the famous episode recorded by John, Chrysostom, PG 9.273fGoogle Scholar, cf. The Roman Empire of Ammianus, 223.Google Scholar

²⁵ Chastagnol, Fastes, 205; cf. Libanius, Or. 1.179, on which see Norman's commentary (above, n. 10), pp. 201f. The episode is not picked up by Fornara, though it provides a clear indication of Hypatius' connections with Antioch.

²⁶ PLRE I, p. 338 (cf. Fornara, 339) proposed the origin of Eusebius and Hypatius (they were the brothers of Constantius' wife Eusebia) as Thessalonica (cf. Julian, Or. III, 106B, 107D), but it is not clear what this would mean of a family connected with government circles from the earlier fourth century; such people might attach themselves to any of a number of court centres. Barnes, at 60, is against Fornara's suggestion of a Macedonian origin for Hypatius and Ammianus, and takes Ammianus' expression to indicate ‘that both he and Hypatius resided in Antioch and faced the same perils there’. This is true, but I think that a reading of 29.2.16 will show that Ammianus meant much more than this.

²⁷ There is also an affinity with Antiochene sources (Libanius and Malalas) in the story told (at 23.5.3) of the third-century renegade Mareades, who led the Persians against his own city and was burned alive there by Sapor I; The Roman Empire of Ammianus, 170–1 and n. 84Google Scholar, and esp. Potter, D. S., Prophecy and History in the Crisis of the Third Century; a Historical Commentary on the Thirteenth Sibylline Oracle (Oxford, 1990), esp. 269fGoogle Scholar.; ‘The absence of allusion to the story [of the taking of Antioch by Sapor] outside of Antiochene authors suggests that it was not widely known elsewhere’.

²⁹ The Roman Empire of Ammianus, 12–13Google Scholar, drawing attention to the second person singular address of 14.6.12. On the digressions, ibid. 414–16.

²⁹ ibid., 23f.

³⁰ ibid., 25; Sabbah, G., La Méthode d'Ammien Marcellin; Recherches sur la construction du discours historique dans les Res Gestae (Paris, 1978), 228–30.Google Scholar

³¹ Below, p. 263.

³² I would argue that it is to the epilogue, if to any part of Ammianus' text, that Libanius would be referring, rather than to a new section perceived by some at the beginning of Book 26; cf. The Roman Empire of Ammianus, 204–5.Google Scholar

³³ Below, p. 262.

³⁴ contra Barnes, 58, who finds it ‘accurate and (sic) undeniable’.

³⁵ Fornara, 333. Nothing in the ‘traditional’ interpretation requires that Ammianus be an ‘old friend’ of Libanius; I myself think it unlikely that he was.

³⁶ ‘Imagine Ammianus being told this!’, exclaims Fornara (at 333). Why not?

³⁷ Famously dismissed by Alan, Cameron, ‘The Roman friends of Ammianus’, JRS 54 (1964), 15–28, at 15–18Google Scholar.

³⁸ Fornara, 330; ‘The inclination to place memorable contemporaries into mutual contact was not limited to the ancient scholars typified by Hermippus and the Alexandrians, and the psychology governing it need not be laboured’. I only wish that I knew what it was.

³⁹ JRS 80 (1990), at 247–8.Google Scholar

⁴⁰ West, M. L., ‘Magnus and Marcellinus: unnoticed acrostics in the Cyranides’, CQ n.s. 32 (1982), 480–1CrossRefGoogle Scholar. For the Cyranides, see Dimitris Kaimakis, Die Kyraniden (Beiträge zur klassischen Philologie 76; 1976); but the poems discussed here (pp. 50–1 and 96–7 Kaimakis) must be read with West's emendations and in his layout of the text.

⁴¹ On whom see RE VII, cols. 2416–17 (Harpocration 10); for his connection with the Cyranides, Kaimakis, p. 14.

⁴² PLRE I, p. 534 (Magnus 7); Eunapius, Vitae Sophistarum, pp. 497/8 (ed. Loeb, pp. 530–3).

⁴³ West, 481; ‘Without a firm identification of Magnus it is hardly profitable to speculate about the identity of Marcellinus. It may, however, be worth pointing out’, etc.

⁴⁴ I owe this observation to Peter Heather.

⁴⁵ 14.9.1; ‘a Nisibi, quam tuebatur, accitus Ursicinus’.

⁴⁶ The Roman Empire of Ammianus, 3f., 186fGoogle Scholar.

⁴⁷ See n. 39 above. Bowersock's arguments are referred to by Fornara, at 336 n. 16, and supported by Barnes, 59.

⁴⁸ I offer this reminder that Fornara, Bowersock and Barnes produce not one but two alternative identifications. If either is valid, it must be better than the other.

⁴⁹ Cf. also at 247; ‘the man—the young man—opted to do some readings of his own’. I see no warrant in Libanius' letter for the words added in parenthesis.

⁵⁰ The troubles over Cimon's status only arose acutely in the last years of Libanius' life; cf. Norman (above, n. 10), 231. Marcellinus might have known Calliopius, whose death Libanius introduces with no explanation; but it is possible that Libanius carried over sentiments from one letter to another (cf. above, n. 8), or even that he used the news about Calliopius to establish intimacy with a correspondent whom he did not know well.

⁵¹ Contrast the excellent Praetextatus at 22.7.6; ‘praeclarae indolis gravitatisque senator’; cf. Sabbah, , La Méthode d'Ammien Marcellin (above, n. 30), 230–1.Google Scholar

⁵² For some implications of this, cf. The Roman Empire of Ammianus, 446Google Scholar. This point is neglected by Fornara, who writes (336 n. 16) as if I had argued that Ammianus was still at work in 392. In The Roman Empire of Ammianus, 27Google Scholar, I make my position explicit; the history was brought to completion in 390 or 391. Barnes, at 60, is sceptical about the men returning to the east in 391, but Libanius makes clear that his informants had travelled from Rome, and I still think that the date and context suggested suit the terms of his letter.

⁵³ At 26.10.19 Ammianus describes some effects of the earthquake and tidal wave of 365. It had planted ships on the tops of buildings, as at Alexandria; Ammianus may have seen this himself, as he had explicitly a second such phenomenon, the stranding of a vessel two miles inland at Mothone in Laconia. He had certainly been to Egypt; cf. 22.15.1, ‘visa pleraque I narrantes’, said in introduction of an Egyptian digression.

⁵⁴ Fornara, 336f., questions this, partly on the basis of the similarity of vocabulary in Libanius, Or. 12.46, where Libanius promises one day to write a detailed account of Julian's German war, νν γρ δ πντα συνττμηται; ‘for now everything is compressed’ (sc. into a few summary paragraphs). This compression does not seem to me to be equivalent to a συγγραϕ divided, or divisible (εἰς πολλ τετμημνη), into parts—in fact, rather the opposite. Barnes (at 59) is against Libanius' correspondent having been an orator, citing the arguments put forward in The Roman Empire of Ammianus, 478 n. 1Google Scholar. For the slightly different reading offered by A. F. Norman, see above, n. 5.

⁵⁵ Fornara, 334–6.

⁵⁶ In Liddell, & Scott's, Lexikon (rev. 9th ed., 1968), s.v. (p. 1661)Google Scholar, the primary meaning given for συγγραϕες is ‘historian’.

⁵⁷ Photius, Bibl, cod. 80, cf. Müller, FHG IV, p. 58; Blockley, R. C., The Fragmentary Classicising Historians of the Later Roman Empire, II (Liverpool, 1983), p. 152.Google Scholar

⁵⁸ I am especially grateful to Hans-Ulrich Wiemer for suggesting and conducting this search, and more generally owe a great deal to his knowledge of Libanius and his works, and to his sensitivity to Libanius' manner of writing.

⁵⁹ The Ibycus search covered all Libanius' writings, but in general only the letters and speeches provide relevant evidence (see however n. 60 & 67 below). It also covered cognates of συγγραϕ and συγγραϕες (e.g. συγγράφειν, σγγραμμα), cf. n. 61 for an example.

⁶⁰ Transl. Norman, , Libanius, Selected Works I: The Julianic Orations, ed. Loeb, I, p. 165Google Scholar. Compare Progymnasmata 4 (‘Sententiae’) 3.10 (ed. Foerster, Vol. VIII, p. 120)Google Scholar for a similarly categorical distinction; ῥτορας κα συγγρϕας κα ποιητς ταὐτς εὑρσεις λγοντας– orators, historians and poets all agree that wealth is necessary for those who would accomplish their duty, citing Demosthenes, Pindar, Thucydides.

⁶¹ At Ep. 1508.6 the verb συγγρφειν (in the imperative σγγραφε) is again used in the immediate context of Thucydides. Libanius' correspondent, Seleucus, is encouraged to ‘write the history of [Julian's Persian] war as you promised to do’ (Norman, , Libanius: Autobiography and Selected Letters, II, No. 142; pp. 289–93)Google Scholar, like Thucydides making light of his exile.

⁶² These local histories, again described as συγγραϕαι, are also referred to at § 107 of the same speech.

⁶³ Eunapius, frag. 9 Müller/17 Blockley.

⁶⁴ Norman, , Libanius: Autobiography and Selected Letters, I, No. 38 (pp. 477–83)Google Scholar. At 3.8.2 Zosimus refers to an ἰδα συγγραϕ of Julian, referring apparently to his panegyric (or panegyrics) on the wars of Constantius; Paschoud, F., Zosime: Histoire Nouvelle, ed. Budé, Vol. II. 1 (1979), 82Google Scholar. In any case it is different from the συγγραϕ mentioned by Eunapius (previous note).

⁶⁵ Norman, , Libanius: Autobiography and Selected Letters, II, No. 120 (pp. 223–9)Google Scholar. See below, p. 267 for the significance of this reference to the ‘lips’ of a historian.

⁶⁶ This principle is enunciated by Ammianus, in his programmatic preface to Book 26, referring to pedantic details ‘praeceptis historiae dissonantia, discurrere per negotiorum celsitudines assuetae, non humilium minutias indagare causarum’ (26.1.1).

⁶⁷ Ignoring also Or. 59.59, in which the word is used in the sense of ‘contract’ (Constantine had a sort of ‘contract with Fortune’ which granted him eternal victory)—a common forensic usage, frequent in the Argumenta Orationum Demosthenicarum (information which I also owe to the Ibycus search) and of no relevance to the matter under discussion.

⁶⁸ It is part of a list of Julian's tireless activities at Antioch in the period preceding the Persian campaign; cf. Norman's note on the passage (Libanius, Selected Works, I, p. 263Google Scholar), comparing Or. 18.164ff., where the same topics occur in a different order.

⁶⁹ Libanius, Ep. 793 to Themistius, cited by Barnes as referring to the ‘products and activity of orators’, is similarly unspecific, the context implying philosophical prose but not actually J designating any form.

⁷⁰ Cf. Fornara, 335f.; Barnes, 59. It is not clear whether Libanius' correspondent (Gaius) had addressed his encomium to Libanius personally, or whether it was about oratory in general.

⁷¹ Fornara's statement (at 334), that ‘The usage is invariable. The verb in question, with its nouns, is neutral in meaning and requires an immediate supplement or latent context in which be properly understood’, is actually much more true of oratory than of history.

⁷² This is conceded, as a special case, by Fornara (337 n. 17).

⁷³ Cf. Fornara, 337f.

⁷⁴ Cf. Barnes, 60; ‘The consequences of discarding the traditional identification are enormous’, but then (having opposed Fornara's suggestion of a Macedonian origin), ‘this is not the place to attempt to resolve the problem’, allowing the possibility that ‘indirect and inferential evidence from his text’ may eventually reinstate Ammianus' Antiochene origin. In the meantime, ‘what matters…is that the question of where precisely in the Greek half of the Roman Empire Ammianus came from should be reexamined without preconception’. This is rather an all-embracing choice of attitudes.

⁷⁵ ‘miles et Graecus’, standing at the very end of the Res Gestae, is essentially a socio-cultural self-definition; The Roman Empire of Ammianus, 461fGoogle Scholar. The word ‘Graecus’ is not an ‘ethnikon’, and gives no grounds whatever for suspecting a Balkan origin, as Fornara, 339.

⁷⁶ Fornara, 328 n. 4, writes that I conceive of Ammianus as an exceptional type of protector domesticus who, at a comparatively late age, ‘took the high road to officer's rank’ perhaps ‘without serving in the ranks at all’, referring to The Roman Empire of Ammianus, 11 and 79Google Scholar. My text also says, between the two passages cited, ‘entering a fashionable regiment almost from his teens’; I do not say or imply that this was ‘at a comparatively late age’, nor is there any contrast or contradiction between my pp. 77 and 79. In general, Fornara's statement at p. 328 gives a misleading impression of my view of Ammianus' background. I do not think it can fairly be said that in my treatment the military career ‘recedes in importance’.

⁷⁷ The Roman Empire of Ammianus, 79–80Google Scholar.

⁷⁸ Cf. The Roman Empire of Ammianus, 71f., 80, 467fGoogle Scholar.

⁷⁹ The Roman Empire of Ammianus, 80, cfGoogle Scholar. Fornara, 330; ‘he displays far greater familiarity with the Latin classics than with the Greek’. This is of course largely inherent in his choice of the Latin language in which to write; he attributes to Julian a largely Greek culture, cf. for example 25.4.2f. (Plato, Sophocles, Bacchylides). Ammianus' Latin culture is the subject of Fornara's, further paper, ‘Studies in Ammianus Marcellinus, II: Ammianus' knowledge and use of Greek and Latin Literature’, Historia 41 (1992), 420–38.Google Scholar

⁸⁰ Cf. Bowersock, 247; ‘some boldness might have been productive’ (sc. in the abandonment of the traditional identification). For one possibly over-bold interpretation, on the nature of Ammianus' audience, see above, pp. 254 and 262.