Hostname: page-component-89b8bd64d-72crv Total loading time: 0 Render date: 2026-05-07T03:09:27.340Z Has data issue: false hasContentIssue false
  • English
  • Français

AMMIANUS MARCELLINUS, LIBANIUS AND LATE ROMAN NAMES

Published online by Cambridge University Press:  27 April 2026

George Woudhuysen
George Woudhuysen*
Affiliation:
University of Nottingham
*
george.woudhuysen@nottingham.ac.uk
Rights & Permissions [Opens in a new window]

Abstract

This article considers the vexed problem of whether the letters of the Antiochene sophist Libanius provide evidence for the life and career of the historian Ammianus Marcellinus. After brief consideration of the debate on Ep. 1063, it turns to Ep. 233 of a.d. 360, which mentions an Ammianus in imperial service, setting out the various attempts to identify this individual as the historian. It then outlines the later Roman onomastic system, casting new light in particular on the role which names played in indicating social status and on the way in which Libanius used names. It shows that fourth-century conventions of naming preclude identifying the Ammianus of the letter with Ammianus Marcellinus.

Keywords

later Roman empireAmmianus MarcellinusLibaniusonomasticsLatin historiansnames

Information

Type
Research Article
Information
The Classical Quarterly , First View , pp. 1 - 12
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2026. Published by Cambridge University Press on behalf of The Classical Association

In memory of Wolf Liebeschuetz

LIBANIUS AND AMMIANUS MARCELLINUS

External evidence for Ammianus Marcellinus is in frustratingly short supply. He tells us a few things (evidently selected with some care) about his background and career in the Roman army. In passing remarks he perhaps reveals a little more about himself here and there, though less than has sometimes been thought.Footnote 1 There is, however, vanishingly little trace of him outside the pages of his Res gestae. The grammarian Priscian, writing probably in Constantinople in the first two decades of the sixth century, cited Ammianus once in his Ars (9.51 = Gramm. Lat. 2.487).Footnote 2 This is the only uncontroversial ancient reference to the historian. For the rest, debate has circled around two letters by the Antiochene sophist Libanius, who (it might reasonably be thought) surely ought to have known a fellow author who was perhaps a native of Antioch, and at any rate had spent a good deal of his life in the city.Footnote 3 The letters are those numbered 1063 and 223 in Foerster’s edition of the ample surviving correspondence of Libanius.

The first of these, Ep. 1063, has received the lion’s share of attention. It was written in a.d. 392 to a Marcellinus, with whom Libanius obviously had some prior acquaintance and who was engaged in some sort of literary activity in the city of Rome, where his work was being warmly received.Footnote 4 Ammianus was well acquainted with the metropolis and it seems all but certain that he finished his Res gestae there c.390.Footnote 5 Long simply assumed to be a reference to the historian, the identification of Libanius’ correspondent as Ammianus Marcellinus was challenged by C.W. Fornara in an incisive article published in 1992.Footnote 6 Though not accepted by all or immediately, Fornara’s case is a convincing one.Footnote 7 The coincidence of background, name, date, location and activity is tantalizing, but what the letter actually says about ‘Marcellinus’ and about what he was doing in Rome does not really suit Ammianus all that well.Footnote 8 Inferences from tone are always hazardous, but Libanius seems to be addressing someone rather younger and less experienced than the historian, plausibly a former student, who had only recently arrived in Rome. Read closely, the letter suggests that Marcellinus’ work was rhetorical, perhaps specifically epideictic, even delivered in a competition of some sort—moreover, it was in Greek. Ammianus was too old to have sat at Libanius’ feet, was not fresh off the boat to Rome, and his Res gestae were in Latin. For these reasons, Ep. 1063 unfortunately cannot give us a window onto Ammianus as he enjoyed at Rome the widespread acclaim his work surely deserved.Footnote 9 Without it, we lose our one certain foundation for thinking that the historian came from Antioch, however likely that continues to seem as an inference.

The second letter in question, Ep. 233, was written in (probably) a.d. 360 to Apolinarius and Gemellus, two former students of Libanius who resided in Cilicia.Footnote 10 In its closing section (4), the sophist says:

πειθέτω δὲ ὑμᾶς χρημάτων καταφρονεῖν ἄνευ τῶν ὑπ’ ἐμοῦ πολλάκις εἰρημένων ὁ τὰ γράμματα φέρων, ὃς ὑπὸ μὲν τοῦ σχήματος εἰς στρατιώτας, ὑπὸ δὲ τῶν ἔργων εἰς φιλοσόφους ἐγγέγραπται τὸν Σωκράτην ἐν μέσοις μιμησάμενος κέρδεσιν, ὁ καλὸς Ἀμμιανός.

Even without the things I have often said to you, be persuaded by the man who bears this letter to think nothing of money. By his dress, he is ranked amongst the soldiers, but by his deeds he is enrolled amongst the philosophers, having imitated Socrates in the midst of profits—the noble Ammianus!Footnote 11

The letter is potentially very useful for any student of Ammianus, because it helps to fill a significant gap in his biography. The general Ursicinus—on whose staff the historian had served for much of the 350s—was dismissed in the aftermath of the fall of the fortress city of Amida in a.d. 359.Footnote 12 We might reasonably assume that Ammianus was at something of a loose end thereafter. He next appears in the historical record during the Emperor Julian’s disastrous invasion of Mesopotamia in 363.Footnote 13 A letter that seems to show Ammianus still in service in 360 and active near Antioch provides a crucial bridge between these two points. The idea that this is a reference to Ammianus Marcellinus was aired on and off by scholars from the late nineteenth century onwards, though Otto Seeck, that great student of Libanius’ letters, was unpersuaded.Footnote 14 It was endorsed by both E.A. Thompson and Sir Ronald Syme in their influential mid-century monographs on Ammianus.Footnote 15 Following Seeck, John Matthews was sceptical in his landmark The Roman Empire of Ammianus, but (perhaps unsurprisingly) T.D. Barnes took the contrary view in his bracing Ammianus Marcellinus and the Representation of Historical Reality.Footnote 16 Barnes offered the fullest and most influential statement of the case for seeing Ep. 233.4 as a reference to Ammianus. The idea has continued to surface here and there subsequently.Footnote 17 While some scepticism has been expressed, there has been no sustained challenge to the arguments of Barnes.Footnote 18 Scholarship is thus at something of an impasse when it comes to whether Ep. 233.4 does provide precious evidence for Ammianus.

Fortunately, it is possible to resolve the question with some certainty, because there is a very simple reason that this cannot be a reference to the historian. The objections of Seeck and Matthews were primarily to do with the office held by the Ammianus of the letter. They both identified him not as a soldier but rather as a στρατιώτης in the extended fourth-century sense of a government official, a usage that grew out of the idea of the bureaucracy as a form of militia, or military service.Footnote 19 Though that is a very common meaning of the word in Libanius’ letters, it is not by itself an overwhelming argument, for he used στρατιώτης in both senses and there is nothing much in the broader context of the letter to help us decide which one was meant here.Footnote 20 Much more compelling is a simple fact. By the generally accepted naming practices of the fourth century, our historian would always have been Marcellinus, not Ammianus, if he were referred to by a single name.Footnote 21 Though common in modern scholarship (not least in this article) and by now of such long usage as to be irremovably established, references to the author of the Res gestae as bare ‘Ammianus’ are simply wrong in the onomastic practice of Late Antiquity. The letter cannot, therefore, refer to Ammianus Marcellinus.

THE LATER ROMAN ONOMASTIC SYSTEM

While the basic principles of later Roman onomastics are reasonably well understood, the subject remains surprisingly neglected, so it is perhaps worth laying out why we can reach this conclusion with such certainty.

The grammarians of the later Roman empire continued to describe the traditional Roman system of praenomen, nomen gentilicium and cognomen, with the agnomen often tacked on for fun, but actual practice in the period in which they lived and worked had diverged sharply from this.Footnote 22 In fact, in his commentary on the Ars of Donatus, the (probably) fifth-century North-African grammarian Pompeius was decent enough to admit that, if you actually asked someone what their cognomen was, you ran the risk of being laughed at.Footnote 23 The work of Pompeius seems to have been assembled from more or less uerbatim classroom notes, so the comment is perhaps particularly revealing.Footnote 24 Instead, a late antique individual was usually known by a single main name, sometimes referred to in the scholarship as the ‘diacritic’ or ‘diacritical name’.Footnote 25 This might be a traditional Latin cognomen, a Greek name, an innovative form in Latin ending in -ius (for example ‘Amantius’ and others like it), a Biblical name (for Christians and Jews), a name drawn from one of the many other languages spoken in the Empire, or something else entirely—there was considerable diversity.Footnote 26 In day-to-day usage, that single name was sufficient and it seems likely that the majority of the Empire’s inhabitants, particularly those lower down the social scale, used it alone. Though the onomastic habits of the rural masses are not well documented, the occasional glimpses we can catch of them strongly suggest this. The Kellis account book of (probably) the 360s opens up for us the world of small tenant farmers in an Egyptian oasis—they are all known by a single name.Footnote 27

This single name might, however, be supplemented on occasion by one more name, always placed before it. We see this fairly often both in more formal contexts and amongst those closer to the top of the social hierarchy. The additional name might indicate descent—a traditional gentilicium, for example, or a cognomen that had come to be used in similar fashion. It might also indicate status. The names Flavius and Aurelius were widely used in this way. Flavius, the gentilicium of Constantine, marked out someone in some official position.Footnote 28 Anyone sufficiently senior in government service might use it and, judging by the Egyptian evidence, the privilege extended quite far down the ranks. Aurelius, the gentilicium of Caracalla (M. Aurelius Antoninus), was promiscuously bestowed on the Empire’s free population by the universal grant of citizenship in a.d. 212. As a result, it was often a mark of a lack of social standing in the centuries thereafter, its recipients being the common people of the Roman world.Footnote 29 In general, it seems there was a sort of hierarchy of these additional names in the fourth century. Aurelius was the additional name of the masses (if they had one). Flavius marked out the upwardly mobile, who had risen to higher status through government service. Family names were borne by those of more elevated social position, who could probably trace their Roman citizenship and standing to before Caracalla. Though often technically entitled to the honorific Flavius, the upper classes tended to avoid it.

For an illustration of this hierarchy, we can turn to the list of those enrolled in the council of the city of Timgad in North Africa, inscribed on the walls of the Curia there in the 360s.Footnote 30 The councillors—men of some local position but hardly members of the Empire-wide elite—were overwhelmingly listed by two names, generally a traditional gentilicium and a main name, with some Flavii and a handful of Aurelii thrown in for good measure.Footnote 31 More evidence from Kellis helps to confirm the point.Footnote 32 Besides the account book, there are quite a few Greek papyri from the site, all seemingly dating to the end of the third century or the fourth. These documents—like the account book speaking to quite a humble slice of late Roman society—occasionally feature dual names. Here, in sharp contrast to Timgad, the documents reveal amongst the inhabitants two Flavii—both interestingly signing on behalf of someone illiterate—and a mass of Aurelii.Footnote 33 Only the praesides Thebaidos, the local provincial governors, diverge from this pattern in having family names: Domitius Asclepiades, Iulius Athenodorus, Septimius Eutropius, Valerius Victorinianus.Footnote 34 This is in itself revealing.

Much more rarely still than dual nomenclature, we find individuals with three, four, five or even more names, besides their main name (again, invariably placed before it). These occur almost always in very formal literary or epigraphic contexts, generally for members of the upper classes and especially the upper classes of the western part of the Empire. Typical places where we might find them are the titulus of the collected works of an author, or a cursus inscription giving the career of some grand senatorial magnate. So, for example, the famous fourth-century orator, epistolographer and senator was generally referred to simply as Symmachus, the main name he too seems to have used in the letters he sent.Footnote 35 Inscriptions and manuscripts of his works, however, reveal that he was known more formally as Q. Aurelius Symmachus.Footnote 36

Two points must be emphasized here. First, when an individual had more than one name, the additional names were invariably placed before the main name that was used more generally.Footnote 37 Second, however many names someone might possess, and some Roman aristocrats luxuriated in an implausible number of them, they were always referred to by at least their main name. The main name could not be omitted, so if only one name was used to refer to someone with multiple names, then (rather obviously) it was the main name. In normal circumstances, an individual who had multiple names was never referred to by more than one of them used in isolation: Q. Aurelius Symmachus was always Symmachus and never bare Aurelius (let alone Quintus). Cases where one individual is referred to by more than one main name—what has been termed ‘oscillation of the diacritic’—are so vanishingly rare as to be effectively non-existent.Footnote 38 The handful that exist are ones where an individual’s name had actually changed, either through the acquisition of a nickname that stuck or through adoption.Footnote 39 So, for example, the long-serving administrator of Constantine and Constantius II who finished his career as praetorian prefect of the East in the 350s must originally have been called Claudius Strategius.Footnote 40 Strategius acted as an assistant to Constantine in the investigation of various sects and the emperor was so impressed that he changed his name to ‘Musonianus’.Footnote 41 Henceforth, in official usage he seems to have been (Claudius) Musonianus, but Libanius at least stubbornly continued to call him Strategius. In a similar fashion, when Constantius II appointed his cousin Gallus as Caesar in a.d. 351, he changed his name to Constantius (a sign that he probably adopted him), which was official usage thenceforth.Footnote 42 Literary sources, however, continued to call him Gallus. It is noteworthy that in both of these cases the change of the main name was explicitly commented on by our sources—clearly, it was very unusual. It is also worth pointing out that in both cases the new name never seems to have been combined with the old: no Strategius Musonianus or Constantius Gallus in the fourth-century sources.Footnote 43

The names of Ammianus Marcellinus fit squarely within what we might call the ordinary operation of later Roman onomastics. In the citation of him, Priscian refers to the historian as simply ‘Marcellinus’: that was clearly his main name. In contrast, the explicits in Vat. lat. 1873—the only authoritative witness to the Res gestae—call the author ‘Ammianus Marcellinus’.Footnote 44 That was, we may surmise, his full nomenclature: a main name supplemented by another, plausibly a family name.Footnote 45 Given what we know about fourth-century names, mention of an Ammianus by that name alone cannot be a reference to Ammianus Marcellinus. Libanius, whose letters illustrate his onomastic practices in exhaustive detail, only ever used one name to refer to individuals.Footnote 46 Where we have fuller evidence for the nomenclature of his correspondents and acquaintances, we can see that this was always their main name: Q. Aurelius Symmachus was simply Symmachus, in the letter addressed to him (Ep. 1004).Footnote 47 With a sole exception, Libanius never referred to an individual by two different names in separate places.Footnote 48 Whoever the Ammianus of Ep. 233.4 was, therefore, he was not the author of the Res gestae.

When he advanced the argument that the letter mentions Ammianus Marcellinus, T.D. Barnes was aware that names might present a challenge to his case.Footnote 49 He argued, however, that ‘naming practices in the fourth century were not wholly consistent’ and furnished two examples to prove this: Julian’s praetorian prefect, sometimes referred to as Secundus and sometimes as Salutius, and the consul of 355, who appears sometimes as Mavortius and sometimes as Lollianus in our sources.Footnote 50 Let us take these points in turn, first the general characterization of later Roman onomastics and then the examples. To those familiar with the tria nomina of an earlier period of Roman history, the use of names in the later Roman empire can at first seem rather fluid and perhaps even confusing. The corpus of names in Late Antiquity is much larger, new names seem to be invented with greater freedom, and others are drawn promiscuously from Greek, or even Hebrew, Punic, Coptic and Syriac. Yet on closer inspection the late ancient onomastic system is much more rigid and standardized than it first appears. The principle of the main name was as close to an iron law as one is likely to find in something so personal and variable as the issue of names, much less flexible than earlier Roman naming practices.Footnote 51 Cicero could choose with exacting care the name by which he would address someone in a letter, or refer to them, to convey esteem or subvert pretension.Footnote 52 In contrast, Symmachus and Libanius—the most prolific epistolographers of the fourth century—had to essentially always use the single main name by which their correspondent was known.

Barnes’s examples similarly disappear on closer inspection. It is true that Julian’s prefect is Saturninius Secundus in an inscription, Secundus in the laws addressed to him, Secundus Salutius once in Ammianus, and generally Salutius elsewhere.Footnote 53 Equally, inscriptions show that the aristocratic consul of a.d. 355 had the full name Q. Flavius Maesius Egnatius Lollianus, but he is sometimes called Lollianus and sometimes referred to as Mavortius, not least in Ammianus who once uses each name alone (15.8.17 and 16.8.5).Footnote 54 Are these not examples of precisely what Barnes argues for Ammianus? They are not. Salutius and Mavortius are both signa, a type of nickname terminating in -ius that was reasonably common amongst members of the upper classes in the third and fourth centuries, especially, it would seem, the senatorial aristocracy.Footnote 55 signa, and nicknames more generally, were the sole significant exception to the ordinary rules of the single main name. Though not generally used in legal or administrative contexts, in literary works they might be substituted for the main name—as when Secundus is called Salutius and Lollianus referred to as Mavortius. They might equally be flexibly combined with the main name, even coming after it, as when Ammianus refers to the prefect as Secundus Salutius (22.3.1). Yet, if anything, the sheer rarity of signa—for they are not common—helps to throw into sharper relief the uniformity of later Roman onomastic practice more generally.Footnote 56 Indeed, the very fact that Barnes had to turn to nicknames to make his case rather disproves his point that fourth-century naming practices were fluid.

CONCLUSION

Deflating as the conclusion may seem, there is simply no way that Lib. Ep. 233.4 refers to the historian. The name Ammianus was fairly common, at least in the eastern part of the Roman empire, and so, despite the coincidence, there is not even any real reason to think that the Ammianus who was a philosopher in soldier’s dress was related to the historian.Footnote 57 The biography of Ammianus Marcellinus thus remains as elusive as ever, our knowledge of him resting on what he deigns to reveal in the Res gestae, the rather bare-bones testimonium that Priscian offers, and the not hugely informative paratexts furnished in Vat. lat. 1873.

It is possible, however, that this personal elusiveness matters less than it might seem at first glance. In 2008, Gavin Kelly showed brilliantly how a focus not on Ammianus’ life but on his literary context could illuminate the Res gestae.Footnote 58 Attempts to cull biographical details from the work, he suggested, had reached (and in some cases exceeded) the boundary that separates inference from speculation. In their place, he showed how careful attention to what Ammianus had read, and the way in which he alluded to it, might open up new vistas on his work and his world. There is still a great deal to be done in this line. The relationship of the Res gestae to the writings of other fourth-century historians perhaps especially merits further investigation, for Ammianus was writing in an era remarkably fecund in historiography. In particular, we might wonder about the ways in which he used and engaged with the work of his older contemporary, Sextus Aurelius Victor, a man whom he admired.Footnote 59 If, in other words, Ammianus the individual remains a frustratingly lonely figure, Ammianus the writer of history still offers rich avenues for further research.Footnote 60

APPENDIX: THE NAMES OF AMMIANUS MARCELLINUS

To my knowledge, the evidence for Ammianus Marcellinus’ nomenclature has not been gathered in one place.Footnote 61 In this Appendix, I present what we know and I briefly discuss what his names can (and cannot) tell us about the historian.

There are two sources of evidence for the names of Ammianus Marcellinus:

  1. 1. The explicits of individual books in Vat. lat. 1873, which give his name as Ammiani Marcellini (always in the genitive): Book XIV = 15v, XV = 26v, XIX = 67v, XX = 78r, XXI = 88v–89r, XXII = 102v, XXIII = 115r, XXIV = 124v, XXV = 135v, XXVI = 145v, XXVII = 155v, XXVIII = 168v, XXIX = 183r, XXX = 193r. The explicits for Books XVI = 38v, XVII = 50r, and XVIII = 57v seem originally to have called him Ammiani Marcelliani, but the intrusive -a- has been deleted. The explicit for Book XXII calls him Ammiani Marcelli, with a supralineal correction to Marcellini. The explicit of Book XXXI (fol. 208r) has Amiani Marcellini.Footnote 62

  2. 2. The citation by Priscian, Ars 9.51 (= Gramm. Lat. 2.487): … ut ‘indulsi indulsum’ uel ‘indultum’, unde Marcellinus rerum gestarum XIIII: tamquam licentia crudelitati indulta (14.1.4) …Footnote 63

Do these names tell us anything about Ammianus’ background?

First, Ammianus. Ammiani were reasonably widely distributed across the Greek world. The name is attested from the Early Imperial period onwards, apparently becoming rather more common after the later second century. On present evidence it was most frequently used in Asia Minor.Footnote 64 It was comparatively rarer in the Latin-speaking regions of the Empire, but is reasonably well attested there.Footnote 65 Barnes thought that the name had a ‘Levantine’ and specifically Semitic flavour.Footnote 66 This was useful to him, for it helped tie the historian to Syria or Phoenicia, from which he had been cut adrift ever since Fornara argued that Ammianus was not Antiochene. But there is no good evidence for it.Footnote 67 The name thus allows us to put the historian in the Greek-speaking East of the Empire, but not much more than that. It might just license the suggestion that his family (or part of it) came originally from Asia Minor.Footnote 68

Second, Marcellinus. This was a very common Latin name, attested from the late Republic onwards and found broadly distributed over the western empire.Footnote 69 It achieved reasonably wide distribution in the Greek world as well, with no particular geographical pattern—it is attested no later than the early first century a.d. Footnote 70 It seems unlikely that a name so widely used can tell us much about someone who bore it, but it might just hint at a more Latin background than Ammianus does. That need not necessarily mean in the Latin-speaking regions of the Empire, for Latin was of course the primary language of the Roman army.Footnote 71 Barnes has already argued that his father was a senior officer.Footnote 72

In short, the names of Ammianus Marcellinus do little more than point vaguely to an author with Greek and Latin connections, not such a surprise for the miles quondam et Graecus (31.16.9).

Footnotes

*

I would like to thank Justin Stover and Gavin Kelly for discussion of these issues as well as for comments on various drafts. I am grateful also to the two reviewers for CQ for their thorough and thoughtful comments. Abbreviations used: LGPN = Lexicon of Greek Personal Names database (https://www.lgpn.ox.ac.uk/); OPEL = B. Lőrincz, Onomasticon Provinciarum Europae Latinarum, 4 vols. (Vienna and Budapest, 1999–2005); PLRE = A.H.M. Jones, J.R. Martindale, J. Morris, The Prosopography of the Later Roman Empire, Volume I: A.D. 260–395 (Cambridge, 1971), references are given by name and number; Seeck, Briefe = O. Seeck, Die Briefe des Libanius (Leipzig, 1906).

References

1 See G. Kelly, Ammianus Marcellinus: The Allusive Historian (Cambridge, 2008), 104–58 for a masterful sketch of what we know, what we can infer, and what we ought not to guess at in the biography of Ammianus.

2 See R.A. Kaster, Guardians of Language: The Grammarian and Society in Late Antiquity (Berkeley, 1988), 346–8 and J.E.G. Zetzel, Critics, Compilers, and Commentators: An Introduction to Roman Philology, 200 BCE–800 CE (Oxford, 2018), 309–11 for basic details of Priscian’s life and writings. On the title of Priscian’s most famous work (not Institutio Grammatica), see M. De Nonno, ‘Ars Prisciani Caesariensis: problemi di tipologia e di composizione’, in M. Baratin, B. Colombat, L. Holtz (edd.), Priscien: Transmission et refondation de la grammaire, de l’antiquité au modernes (Turnhout, 2009), 250–9.

3 For orientation on Libanius, see L. Van Hoof (ed.), Libanius: A Critical Introduction (Cambridge, 2014).

4 For the date of Ep. 1063, see Seeck, Briefe 463 (cf. 446)—the case hinges on references to the death of Cimon (1063.5), the son of Libanius, and the death of Calliopius (1063.6), his subordinate teacher.

5 Kelly (n. 1), 109–14.

6 C.W. Fornara, ‘Studies in Ammianus Marcellinus I: the letter of Libanius and Ammianus’ connection with Antioch’, Historia 41 (1992), 328–44.

7 In particular, J.F. Matthews, The Roman Empire of Ammianus (London, 1989), 8 n. 1 offered a defence of the traditional interpretation—based on a seminar paper by Fornara, this actually appeared before the latter’s article was published. Matthews made his case again in ‘The origin of Ammianus’, CQ 44 (1994), 252–69. He certainly shows that not every argument made by Fornara carries full conviction, but he does not really challenge the fundamental point. Kelly (n. 1), 111–17 reviews the issue, inclining to Fornara and making also the important point that less depends on the letter than has been assumed. See now also S. Bradbury and D. Moncur, The Letters of Libanius from the Age of Theodosius (Liverpool, 2023), 342–4, who offer some important refinements, emphasizing particularly that the recipient of Ep. 1063 must have been well known to Libanius.

8 It does not help the cause of identification that Marcellinus was a very common name. See the Appendix at the end of this article.

9 G.W. Bowersock, ‘Review of J.F. Matthews, The Roman Empire of Ammianus (London, 1989)’, JRS 80 (1990), 244–50, at 247–8 instead identified the recipient as the Marcellinus to whom one Magnus dedicated some poems, as revealed by the acrostics discovered by M.L. West, ‘Magnus and Marcellinus: unnoticed acrostics in the Cyranides’, CQ 32 (1982), 480–1. These verses were later incorporated into the Cyranides, a late antique compilation of medico-magical material. Both West and Bowersock—followed by T.D. Barnes, ‘Ammianus Marcellinus and his world’, CPh 88 (1993), 55–70, at 59—suggested identifying this Magnus with the Magnus of Nisibis (PLRE ‘Magnus 7’) otherwise known to Libanius, active from the 360s to the 380s. Matthews (n. 7 [1994]), 260–1 was rightly sceptical: as G. Fowden, The Egyptian Hermes: A Historical Approach to the Late Pagan Mind (Cambridge, 1986), 87 and n. 57 pointed out, the poems of Magnus were already incorporated in the work of Harpocration of Alexandria, who can be securely dated to the second century a.d. Magnus of Nisibis, a rather composite figure, perhaps in any case needs some rethinking. He is usually identified with the subject of a poem by Palladas (Anth. Pal. 11.281), but that is incompatible with the Constantinian date for that poet convincingly advanced by K.W. Wilkinson, ‘Palladas and the Age of Constantine’, JRS 99 (2009), 36–60.

10 For the letter’s date, see Seeck, Briefe 372, 374. On the brothers, see PLRE ‘Apolinarius 2’ and ‘Gemellus 2’.

11 The precise sense of ἐν μέσοις … κέρδεσιν is rather murky: R. Cribiore, The School of Libanius in Late Antique Antioch (Princeton, 2007), 242 offers ‘with regard to profit’; T.D. Barnes, Ammianus Marcellinus and the Representation of Historical Reality (Ithaca, NY and London, 1998), 61 gives ‘despite having gainful employment’—neither seems right. Libanius used κέρδος for the profits of office (Ep. 604.5, almost paradigmatically) and in context the point should surely be that, although he had official position, Ammianus had not engaged (overmuch) in peculation. Foerster’s apparatus fontium is slightly misleading on the invocation of Socrates: E. Richtsteig, Libanius qua ratione Platonis operibus usus sit (Lignitz, 1918), 32 cites not Pl. Ap. 19d but rather Xen. Mem. 1.2.60 as the inspiration.

12 PLRE ‘Ursicinus 2’—see Amm. Marc. 20.2.2–5.

13 Ammianus re-enters his own narrative at 23.5.7: profecti exinde Zaitham uenimus locum. From then on, first-person plural verbs are frequent down to 25.10.1.

14 It seems to have originated with M. Büdinger, Ammianus Marcellinus und die Eigenart seines Geschichtswerkes. Eine universal-historische Studie (Vienna, 1895), 9. Cf. Seeck, Briefe 58.

15 E.A. Thompson, The Historical Work of Ammianus Marcellinus (Cambridge, 1947), 11. R. Syme, Ammianus and the Historia Augusta (Oxford, 1968), 45. Amusingly, Syme knocked Thompson for neglecting Ep. 233.4 (at 45 n. 4).

16 Matthews (n. 7 [1989]), 47 n. 26. Barnes (n. 11), 61–4, cf. 78. PLRE implicitly rejects the identification by entering the Ammianus of the letter as ‘Ammianus 2’.

17 D. Rohrbacher, The Historians of Late Antiquity (London, 2002), 19; T. Türel, ‘Decoding a narrative allusion: the death narratives of Ammianus’ Julian and Xenophon’s Cyrus’, Mnemosyne 74 (2021), 1015–33, at 1027.

18 Kelly (n. 1), 125 includes it in his list of ‘weakly competing cases’ for what Ammianus was doing in a.d. 360. Cribiore (n. 11), 241 identifies the recipient as ‘an official in charge of financial matters’.

19 A.H.M. Jones, The Later Roman Empire, 284–602: A Social, Economic and Administrative Survey, 3 vols. (Oxford, 1964), 1.566; TLL 8.962.27–963.2, 963.59–964.16 (Bulhart).

20 For στρατιώτης as bureaucrat in Libanius, see J.H.W.G. Liebeschuetz, Antioch. City and Imperial Administration the Later Roman Empire (Oxford, 1972), 114 and n. 3. Sampling in the letters suggests that it is by far the most common meaning there, when we can judge—it is used most often for agentes in rebus. In the orations, however, there are instances where it clearly refers to military men (e.g. Or. 2.37, an example picked almost at random).

21 A point well understood by Matthews (n. 7 [1994]), 254 in the context of Ep. 1063. In his discussion of Ep. 233, Seeck, Briefe 58 also objected that Libanius called the historian Marcellinus, but this was an inference from Ep. 1063.

22 For an example of simply outlining the traditional system, see Charisius, Gramm. 1.6 (page 193 Barwick). On the late Roman onomastic system, see, inter alia, Alan Cameron, ‘Polyonomy in the late Roman aristocracy: the case of Petronius Probus’, JRS 75 (1985), 164–82; B. Salway ‘What’s in a name? A survey of Roman onomastic practice from c. 700 b.c. to a.d. 700’, JRS 84 (1994), 124–45, especially 133–45; O. Salomies, ‘Réflexions sur le développement de l’onomastique de l’aristocratie romaine du Bas-Empire’, in C. Badel and C. Settipani (edd.), Les stratégies familiales dans l’Antiquité tardive (Paris, 2009), 1–26; I. Kajanto, Onomastic Studies in the Early Christian Inscriptions of Rome and Carthage (Helsinki, 1963).

23 Pompeius, Commentum artis Donati (Gramm. Lat. 5.140–1). Outside the grammarians, terminology showed signs of change: nomen was used generally in the later Roman period for the main name, so some authors in turn used praenomen for the gentilicium (because it came before the name). See TLL s.v. praenomen I.B.2 (Maslowski) and P. Veyne, ‘Le “prénom” de Naucellius’, RPh 38 (1964), 253–7, though note that in the Historia Augusta the manuscripts transmit pronomen, where Hohl printed praenomen.

24 On Pompeius, see Kaster (n. 2), 139–68, 343–4; Zetzel (n. 2), 308–9; A. Zago, Pompeii Commentum in artis Donati partem tertiam, 2 vols. (Hildesheim, 2017).

25 Logically, a diacritic or diacritical name ought to be the individuating or distinguishing name, as opposed to other names borne, the name of significance in formal (legal, administrative, ceremonial) situations perhaps being the most important. These might of course vary with context, but in modern western usage are perhaps roughly equivalent to the distinction between a first or Christian name and an initial with surname. Since at least Cameron (n. 22) (who is principally responsible for disseminating the usage, though he seems to have taken it over from A. Giardina, ‘Lettura epigrafica e carriere aristocratiche: il caso di Petronio Probo’, RFIC 111 [1983], 170–82), scholars of late ancient onomastics (the author included) have used ‘diacritic’ and ‘diacritical name’ for what I call here the single or the main name. This practice is perhaps unhelpful, because the single main name was normally both the diacritic and the name of significance.

26 Perhaps the best guide to this material is I. Kajanto, The Latin Cognomina (Helsinki, 1965), more wide-ranging than its title suggests. The origin of Latin cognomina in -ius/-ia is a tricky problem. Salway (n. 22), 136 states that Greek names formed from standard items of vocabulary with the suffix -ιος/-ια began to appear in the second century (theophoric names with the same suffix were obviously much older). These, he suggests, then influenced the development of Latin practice. Kajanto (n. 22), 70–86 offers the fullest discussion known to me of the issue, but comes to no firm conclusion on the question of what prompted the development in Latin. He denies (72–5) that the cognomina originated as nicknames and suggests that their spread owed something to the decline of the gentilicium (for which the -ius suffix had previously been reserved). The broad thrust of his examples and statistics might, however, support the primacy of Greek here.

27 See R.S. Bagnall, The Kellis Agricultural Account Book: P. Kell. IV Gr. 96 (Oxford, 1997), 62–70. Indications of family connections (‘son of’, ‘father of’, etc.) are more common and were presumably used for differentiation in day-to-day life. On Kellis in general, see now C.A. Hope and G.E. Bowen (edd.), Kellis: A Roman-Period Village in Egypt’s Dakhleh Oasis (Cambridge, 2022).

28 On Flavius, see Alan Cameron, ‘Flavius: a nicety of protocol’, Latomus 47 (1988), 26–33.

29 On the two names as markers of status, see the classic papers by J.G. Keenan, ‘The names Flavius and Aurelius as status designations in later Roman Egypt’, ZPE 11 (1973), 33–63 and 13 (1974), 283–304.

30 On the social history of Roman Timgad, see C. Witschel, ‘Die Entwicklung der Gesellschaft von Timgad im 2. bis 4. Jh. n. Chr.’, Klio 77 (1995), 266–331.

31 For the album of Timgad, see A. Chastagnol, L’Album municipal de Timgad (Bonn, 1978), especially 49–74 on its onomastics. Some of the Flavii may bear the name as a genuine family name, perhaps received from an ancestor enfranchized by a Flavius of an earlier era—it is perhaps too much to hope that it was not abbreviated to Fl. in those cases. It is interesting to note that only two of those in the album bear praenomina (Chastagnol [this note], 50). In the album of Canusium (of a.d. 223), a similar document, it seems that all individuals bear them. For the latter, see the edition with commentary by M. Chelotti, R. Gaeta, V. Morizio, M. Silvestrini, Le epigrafi romane di Canosa I (Bari, 1985), 45–68 and the brilliant article by B. Salway, ‘Prefects, patroni, and decurions: a new perspective on the album of Canusium’, in A. Cooley (ed.), The Epigraphic Landscape of Roman Italy (London, 2000), 115–71.

32 I base what follows on the index of names in K.A. Worp, Greek Papyri from Kellis: I (P. Kell. G.) Nos. 1–90 (Oxford, 1995), 230–42. I exclude from the analysis the names of consuls, who appear in dating formulae.

33 Ninety-seven Aurelii are listed in the index of names, though the precise number is probably somewhat lower (allowing for fragmentary cases, etc.).

34 Valerius Herculanus, who appears in P. Kell. G. 27.2, is very likely also a praeses of the Thebaid, though his office is not directly attested in the document. If he was not the praeses, then he was certainly some other high official.

35 The full and helpfully arranged index nominum to O. Seeck, Q. Aurelii Symmachi quae supersunt (MGH AA 6.1) (Berlin, 1883), 342–52 suggests that every letter of Symmachus which names him as the sender uses that name alone. Somewhat more surprisingly, the same is true of the Relationes, which, as formal reports to the emperors, one might expect to use the fuller nomenclature.

36 For the inscriptions, see PLRE ‘Symmachus 4’. For the manuscripts, to my knowledge the evidence has never been assembled. The full name is found scattered through the most important surviving witness to the letters, Paris lat. 8623 (B. Bischoff, Katalog des festländischen Handschriften des neunten Jahrhunderts. Teil III: Padua–Zwickau [Vienna, 2014], no. 4551: saec. IX2/4 from Ferrières), in the incipits and the explicits (which are not confined to the beginnings and ends of books alone): fol. 10r, 25v, 38v, 39v, 41v, 44v, 59r, 63v (where it is QR Aur Symmachi), 64r, 65v, 67v, 68r, 70r, 71r, 72r, 88v. The clustering is rather curious. The full name is also found as a running header in the Bobbio manuscript that transmits such fragments as we have of Symmachus’ Orationes. The main name is there given (in the genitive) as Summaci or Summachi (Seeck [n. 35], viii). The full name heads the first relatio in MS CLM 18787 (saec. X) page 2 (and then seemingly as a running header for some pages thereafter), which the apparatus criticus in Seeck (n. 35), 279 omits, and is found at the end of the collection in the manuscript as well (page 57), with the praenomen written out in full. It was also present at the start and at the end of the Relationes in MS Metz 500 (saec. X), destroyed in the Second World War: see Seeck (n. 35), 279 and 317.

37 Exceptions—where the main name was not the final one—are very few and can generally be explained by the desire to distinguish between closely homonymous members of the same family, who happened to share a final name: see Cameron (n. 22), 173.

38 ‘Oscillation’ once again goes back to Cameron (n. 22) and Giardina (n. 25).

39 Cameron (n. 22), 176–7.

40 See PLRE ‘Musonianus’. The gentilicium Claudius is furnished by an inscription from Hierapolis (F. D’Andria and M. Piera Caggia, Hierapolis di Frigia I. Le attività delle campagne di scavo e restauro 2000–2003 [Istanbul, 2007], 610, with C.P. Jones, ‘The institutional history of Hierapolis. Review of “T. Ritti, Hierapolis di Frigia IX: storia e istituzioni di Hierapolis [Istanbul, 2017]”’, JRA 31 [2018], 921–7, at 925) and is a plausible reading on another inscription from Paphos (J.B. Cayla, ‘Bassidius Lauricius gouverneur de Chypre? Une nouvelle hypothèse à propos de la restauration d’un monument incendié à Paphos’, CCEC 27 [1997], 71–6 + Pl. XXIII). In both cases, the inscriptions would date to his stint as praetorian prefect and in both the name is abbreviated.

41 Amm. Marc. 15.13.2—for the formula uoluit appellari, cf. 17.1.11 and 18.9.1, useful reminders that the emperor’s wishes were commands. On Musonianus, see J.W. Drijvers, ‘Ammianus Marcellinus 15.13.1–2: some observations on the career and bilingualism of Strategius Musonianus’, CQ 46 (1996), 532–7.

42 See PLRE ‘Constantius 4’, which omits the fact that Aur. Vict. Caes. 42.9 explicitly attests the change of name (cuius nomen suo mutauerat); cf. Ammon, Epistula de Pachomio et Theodoro 7 (J.E. Goehring, The Letter of Ammon and Pachomian Monasticism [Berlin and New York, 1986], 129): ἐρρέθη ἐνιαυτοῦ καὶ μικρόν τι πρὸς παρεληλυθότος ἀφ’ οὗ καῖσαρ ἀνηγορεύετο Γάλλος ὁ ἐπικληθεὶς νέος Κωνστάντιος. The change of name to Constantius is also recorded in the Chron. Pasch. (page 540 Dindorf) and Theophanes AM 5842. It is unclear whether Constantius also changed Gallus’ nomen gentilicium from Claudius to Iulius, to align it with his own. The name Fl. Iulius Constantius (some coins from Thessalonica, RIC 8.398) is attested for Gallus, but so is Fl. Claudius Constantius (most coins, RIC 8.189, 269, 292, 301, 373, 455, 477, 492; CIL 8.8475, 12.5560; AE 1989, 699a). Victor sometimes uses nomen to specifically mean the gentilicium (e.g. at 2.1), but more often as a general word for name.

43 In the Chronicle of 511 483 (page 644 Mommsen), occiso Gallo Constantio Histriae pro suspicione is probably meant to name two individuals (read occiso Gallo a Constantio …), not to give two names to an individual. Socrates, Hist. eccl. 2.32.8 and 2.33.2 does refer to Κωνσταντίου τοῦ Γάλλου (in a consular date) and Γάλλος ὁ καὶ Κωνστάντιος.

44 See the Appendix at the end of this article.

45 Ammianus was, of course, not a traditional gentilicium, but by the fourth century other names had long since taken on the same function of communicating descent. I.Aph. 12.21, a decree honouring Aelius Aurelius Ammianus Papias, son of M. Aelius Aurelius Ammianus, grandson of Papias, great-grandson of Papias Maron, etc., chances to show this process already happening to Ammianus in (probably) the third century. Clearly, Papias was a name handed down from the father’s side—perhaps Ammianus came in via a grandmother. In the fourth century, the distinction between gentilicia in -ius and other names in -us was probably already becoming hazy: it is noticeable at Timgad that a significant local clan bore the family name Pompeus, which had once presumably been Pompeius (Chastagnol [n. 31], 51).

46 I base this conclusion on the thorough index of names to Foerster’s edition (assembled by Richtsteig): in the enormous corpus of Libanius’ extant works, I do not believe that there is a single instance of an individual being referred to by two or more names. Remarkably, we do not know Libanius’ full nomenclature and we cannot even make inferences about it, for he tells us nothing of his family’s names.

47 See, besides, Publius Ampelius (PLRE ‘Ampelius 3’), Maternus Cynegius (‘Cynegius 3’), Aelius Claudius Dulcitius (‘Dulcitius 5’), Antonius Hierocles (‘Hierocles 3’), Claudius Iulianus (the Emperor Julian, ‘Iulianus 29’), Domitius Leontius (‘Leontius 20’), Ulpius Limenius (‘Limenius 2’), Vindaonius Magnus (‘Magnus 12’), Domitius Modestus (‘Modestus 2’), Claudius Musonianus (‘Musonianus’), Olympius Palladius (‘Palladius 18’), Aradius Rufinus (‘Rufinus 11’) and Eutolmius Tatianus (‘Tatianus 5’), not a comprehensive list. Given his aversion to using even family names when referring to people, it is hardly surprising that Libanius also shunned the honorific Fl(avius), as e.g. Flavius Philippus (‘Philippus 7’).

48 That exception is itself revealing. Libanius had an illegitimate son, whom he sometimes refers to as Cimon and sometimes as Arabius (PLRE ‘Arabius’ for details—the manuscripts tend to offer the form Ἀρράβιος: see Foerster’s apparatus criticus to Epp. 625, 678, 960 and 1000). Libanius struggled mightily to ensure his son could inherit from him (for an overview, see J. Evans Grubb, ‘Illegitimacy and inheritance disputes in the late Roman empire’, in B. Caseau and S. Huebner [edd.], Inheritance, Law and Religions in the Ancient and Mediaeval Worlds [Paris, 2014], 39–41). If the ‘oscillation’ is not simply from family intimacy (with one of the names as a nickname), then it may well be explained by a change of name associated with the formalization of Cimon’s status as heir; cf. Bradbury and Moncur (n. 7), 46, who arrive at a not dissimilar conclusion by a different route.

49 Barnes (n. 11), 62—there is a defensive note to ‘nomenclature is no obstacle to identifying Libanius’ philosophical soldier as the future historian’.

50 Barnes (n. 11), 62 and n. 38.

51 Cf. Cameron (n. 22), 171: ‘I would submit that in the period c. 300 to 500 the use of proper names among the Roman aristocracy was governed by far stricter and more uniform rules than in the early Empire.’

52 See J.N. Adams, ‘Conventions of naming in Cicero’, CQ 28 (1978), 145–66.

53 The evidence is assembled in PLRE ‘Secundus 3’ and in Seeck, Briefe 265–6. The name was Saturninius, not Saturninus (an easy slip, found quite often in the scholarship): it was most common in Gaul (OPEL IV.51) and Secundus was indeed of Gallic origin.

54 PLRE ‘Lollianus 5’, though see J.A. Stover and G. Woudhuysen, ‘The poet Nemesianus and the Historia Augusta’, JRS 112 (2022), 173–97, at 185 and n. 84 for an accurate account of Firmicus Maternus’ use of the names.

55 On signa, see I. Kajanto, Supernomina: A Study in Latin Epigraphy (Helsinki, 1966) and G. Woudhuysen, ‘Myrmeicus or Myrmecius?’, Mnemosyne 72 (2019), 840–60. There is always the chance that they are merely better attested for well-documented senators than for other members of the Roman elite.

56 PLRE I, which lists important officials (and some other categories) from 260–395, has (I believe) 4,154 entries. It gives signa for 56 of these individuals. Though PLRE both omits signa for some individuals who had them and includes as signa some things that are not (see especially I. Tantillo, ‘Memmius Vitratius Orfitus: signo Honorius?’, ZPE 190 [2014], 271–8, who shows that in some inscriptions Honori has been misinterpreted as a signum Honorius), its figures still give a sense of how comparatively rare these nicknames were, even amongst the well-documented upper classes, who were most likely to bear them.

57 See the Appendix at the end of this article.

58 Kelly (n. 1).

59 See, for a first approach, J.A. Stover and G. Woudhuysen, The Lost History of Sextus Aurelius Victor (Edinburgh, 2023).

60 Cf. A. Momigliano, ‘The lonely historian Ammianus Marcellinus’, ASNP 3rd series 4.4 (1974), 1393–407.

61 There is no discussion of it in PLRE ‘Marcellinus 15’.

62 The fragments of the Hersfeld manuscript of the Res gestae (Kassel, Landesbibliothek 2o Ms. philol. 27 + 4o Ms. chem. 31) are the only other independent witness to the work of Ammianus, but do not provide any evidence for his name. On the Hersfeldensis, see G. Kelly and J.A. Stover, ‘The Hersfeldensis and the Fuldensis of Ammianus Marcellinus: a reconsideration’, The Cambridge Classical Journal 62 (2016), 108–29.

63 Hertz records no variants for the name in his apparatus criticus.

64 I base these conclusions on the LGPN database, which has 105 entries for Ἀμμιανός, 21 for Ἀμμιανή, 13 for Ἀμιανός, and 1 for Ἀμιανή. Phrygia, with 58 attestations, and Lydia, with 30, lead the pack. I suspect that the name was developed from the somewhat more common female name Ἀμμία, which is also found predominantly in Asia Minor (Phrygia in particular) and has a slightly earlier chronological skew. Hesychius (α 3697) glosses the noun ἀμμία as μήτηρ or τροφός.

65 OPEL I2 lists four instances. To these add CIL 6.1058, 6.1600, 6.11557, 6.37078, 11.806.5, ICUR N.S. 5.13280, 5.14005, 6.16982, 6.17155, 6.17179b, 8.21684, 9.23789, 10.27058 (Rome), CIL 9.1207 (Italy, Aeclanum), CIL 12.2794 (Narbonensis, Massilia), CIL 13.6801 (Germania superior, Mogontiacum), CIL 8.4054 bis (Numidia, Lambaesis). The geographical distribution and the chronological skew to Christian material suggest that the name was an import from the Greek world (and note in CIL 6.1600 the parents C. Calpurnius Philokyrius and Septimia Ammias, with their son C. Calpurnius Ammianus). One of the reviewers suggested that Ammianus might on occasion be derived from the gentilicium Ammius (as Flavius gave rise to Flavianus) and an independent origin in Latin should perhaps not be excluded for some of these individuals.

66 Barnes (n. 11), 62–3. Barnes further suggests that Ammianus might have avoided using the name at Rome because it revealed his Eastern origins—given the vogue for Greek nomenclature amongst the senatorial aristocracy (many signa were Greek, for example), this seems most unlikely.

67 LGPN VI, covering Syria, Palestine and the regions beyond the Euphrates, has not yet appeared, but the name is already well enough attested that that does not really undermine the point. It is noteworthy that there are no Ammiani attested in the documents from Dura Europos (C. Bradford Welles, R.O. Fink and J.F. Gilliam, The Excavations at Dura-Europos, Final Report V, Part I. The Parchments and Papyri [New Haven, CT, 1959]), perhaps our best single body of evidence for names in the late Roman Near East.

68 Barnes (n. 11), 60 ruled out an Anatolian background for Ammianus because of a slighting reference to Asiatici mores in 16.7.6, but that is not in the historian’s own voice and in any case would not rule out a deeper family connection.

69 Kajanto (n. 26), 173 offers 533 instances, while OPEL III has 271 (their coverage partially overlaps, of course). There are 23 instances in PLRE I, which puts it even with Leontius and behind only Maximus (49), Eusebius (42), Iulianus (40), Severus (31), Theodorus (27) and Rufinus (25).

70 LGPN lists 32 instances of Μαρκελλῖνος and 21 of Μαρκελλῖνα.

71 J.N. Adams, Bilingualism and the Latin Language (Oxford, 2003), 599–608 argued persuasively that it was not the sole or the official language in the Roman army, but he acknowledged it had a position of primacy. See M.A. Speidel, ‘Learning Latin in the Roman army’, in A. Mullen (ed.), Social Factors in the Latinization of the Roman West (Oxford, 2023), 133–58, for an up-to-date overview.

72 Barnes (n. 11), 59, 63.