The Historians of Late Antiquity

by David Rohrbacher

  Indeed, given what we can reconstruct of Ammianus’ early career, it is not unlikely that he was raised in a Latin-speaking family (Matthews 1989: 71–80). Ammianus served as protector domesticus, a military staff assistant (Trombley 1999). Some men reached this position through long years of service, but Ammianus was still a young man when he became protector, which suggests that he received his rank through family connections. His father had probably been a soldier (Barnes 1998: 58–9), and was perhaps even the Marcellinus who served in the powerful position of comes Orientis in 349 (Gimazane 1889: 24–7). Ammianus several times complains about the unfair burdens placed upon the municipal elite, who were increasingly compelled to perform onerous duties as members of the curia, or city council (22.9.12, 25.4.21; Thompson 1947a). Ammianus may have derived exemption from such service from his father’s position in the imperial service as well as from his own military profession (Barnes 1998: 58–9). The general under whom Ammianus served as protector, Ursicinus, had a home in Antioch (18.4.3), and Ammianus’ service as protector may have been an apprenticeship under a family friend. Appointment to the position of protector required the ritual of adoratio purpurae, prostration before the emperor and the kissing of his purple robe (Avery 1940), and Ammianus probably performed this ritual while Constantius II was still in Antioch, and therefore before 350.

  We can extract some suggestive information about Ammianus’ youth from remarks scattered throughout his work. He describes himself as ingenuus and therefore accustomed to ride rather than walk (19.8.6); the word implies a reasonably high social status. In 359, he tells us, he stayed at the home of a certain Jovinianus, who was the satrap of the Armenian province of Corduene (18.6.20–1). Jovinianus had developed a love of Roman literature during his youth, which he spent as a hostage in the eastern empire. Perhaps the two met in school. The acquaintanceship reveals a bit of the cosmopolitan nature of the Antioch of Ammianus’ childhood.

  This traditional account of Ammianus’ birthplace and background, outlined by E.A. Thompson (1947a) and augmented by John Matthews (Matthews 1989, esp. 67–80), has come under attack in recent years (Barnes 1993a, 1998; Fornara 1992a; Bowersock 1990b). Some scholars have denied that Ammianus was born in Antioch and have suggested in its place some other city of the Greek east. Most revisionists pay particular attention to a letter of the Antiochene orator Libanius to “Marcellinus” (ep. 1069), which had long been believed to be addressed to the historian. The letter congratulates its recipient on the success of his recent readings at Rome and was written around 392. In reviews of Matthews’ work, Bowersock (1990b) and Barnes (1993a) both suggest that this letter addressed to Marcellinus was not written to our Ammianus Marcellinus but rather to another holder of that common name. Bowersock goes so far as to claim that “the assumption [of Ammianus’ Antiochene origins] is based solely upon the identification of the recipient of [the] letter … as the historian Ammianus Marcellinus” (1990b: 247). Both reviewers were influenced by the work of Fornara (1992a), who similarly argued that Libanius’ Marcellinus cannot be the historian, and that our understanding of Ammianus and his background should be reconsidered, as this letter is the “singular pillar” holding up the traditional account.

  It is, however, incorrect to suggest that the letter of Libanius is the sole evidence linking Ammianus with Antioch. Matthews and Sabbah respond to the revisionists with numerous other pieces of circumstantial evidence (Matthews 1994; Sabbah 1997). Several passages praise the city (14.8.8, 22.9.14). Ammianus identifies himself with the Antiochenes during the reign of Valens several times (29.1.24, 29.2.4). In addition, there is his acquaintance with Jovinianus (18.6.20–1), the home which his patron Ursicinus had in Antioch (18.4.3), and the reference to the Antiochene Hypatius as “our” Hypatius, “praiseworthy from his youth” (29.2.16). Ammianus also knew Syriac, the native language of Antioch, a point made by Matthews that is accepted by one of his critics (Barnes 1998: 56). Barnes also recognizes Ammianus’ “profound knowledge” of the city (1998: 60). None of this evidence is decisive, but it is very suggestive.

  Fornara argues that the tone and diction of the letter of Libanius to Marcellinus would be appropriate if Libanius were addressing a young orator, but would be wholly unsuitable in a letter to an old historian (1992a: 331–8), and Barnes agrees that “Libanius’ tone precludes his [addressee’s] identification as the historian Ammianus Marcellinus” (1998: 57–8). Both Barnes and Fornara provide a Greek text and English translation of the words of Libanius, as does Matthews (1994), to which the interested reader may turn. This writer is simply not convinced that the tone of the letter demands any such conclusion. Moreover, altering the identity of the letter’s recipient demands one to accept a very impressive coincidence. The letter is dated to 392 because of its position in Libanius’ collection and its mention of the recent death of his son and of a former student. The latest datable events alluded to in Ammianus’ work include the consulship of Neoterius (26.5.14) in 390 and the death of Petronius Probus (27.11.2) around 390. In addition, Ammianus lavishly praised the Serapeum of Alexandria as a building which would last eternally (22.16.12), a comment unlikely to be made after its destruction in 391 (Matthews 1994: 254). Ammianus Marcellinus, then, who might have been circumstantially considered to be a native of Antioch, completed his history around 391. Within a year, critics suggest, Libanius was writing a letter to a completely different Marcellinus of Antioch, giving readings from a prose work to great acclaim. Surely it is most economical to identify the recipient as the historian.

  Toward the beginning of the surviving text of the Res Gestae, Ammianus reveals that in 353 he had been a member of Ursicinus’ staff and was traveling with the general from Nisibis in Mesopotamia to Antioch, where Ursicinus was to serve as a judge for treason trials being orchestrated by the Caesar Gallus (14.9). The emperor Constantius had decided to remove Gallus from power, Ammianus tells us, and was persuaded by his advisors of the dangers of leaving the popular Ursicinus alone in the east, unchecked by a member of the imperial family (14.11.1–3). In 355, Ammianus was at the court of Constantius II at Milan, where, he claims, constant intrigue was being directed against Ursicinus (15.2.1–6). Trouble arose in Gaul, where the general Silvanus had been forced into revolt against the emperor (Hunt 1999; Drinkwater 1994; den Boer 1960; Balducci 1947). The emperor selected Ursicinus and Ammianus for the unenviable task of gaining the confidence of the usurper and then assassinating him. They succeeded in these goals, befriending Silvanus and then paying soldiers to murder him as he sought refuge in a church (15.5.27–31). After Silvanus’ death, Ursicinus took over his position as general in Gaul, although technically he remained assigned to the eastern frontier (Frézouls 1962). This allowed Ammianus to witness the early successes of the young Julian described in the sixteenth book of his history. Although Marcellus was sent as a successor to Ursicinus in 356 (16.2.8), Ursicinus was ordered to remain, and was only summoned by the emperor to Sirmium in northern Italy in the summer of 357 (16.10.21), with Ammianus remaining in his retinue.

  Leaving Sirmium, Ursicinus returned to his old position as magister equitum in the east. In 359, however, the general Sabinianus arrived bearing imperial letters which demanded the recall of Ursicinus to court (18.6.1). As Ursicinus and Ammianus were returning to Italy, they were met in Thrace by other imperial messengers ordering their immediate return to the eastern front (18.6.5). Ammianus saw these events as manifestations of the incompetence and malice of Constantius and his court, and blamed the increase in Persian aggressiveness on their discovery that Ursicinus had been replaced by the inferior Sabinianus (18.6.3–4). Ammianus also considered the return of Ursicinus to the east to be part of a cunning plot by the general’s enemies at court, allowing them to blame him for any failures of a campaign directed by Sabinianus, while exempting him from praise should the campaign be successful (18.6.6). The truth is surely more charitable to the emperor. Ursicinus’ recall appears to have been demanded in order to replace the recently executed west
ern general Barbatio, and his return to the east a necessity given the increasing Persian hostilities (Matthews 1989: 40–1).

  Ammianus’ involvement in the Persian invasion of 359 makes for an exciting story (18.6.8–19; Paschoud 1989b). Ursicinus and Ammianus set out for Nisibis, a city in Mesopotamia, to help prepare it for a siege, but found that they had arrived too late. Smoke from enemy encampments was already visible along the horizon. Two miles from Nisibis, the party came upon an 8-year-old boy who had been abandoned by his panicked mother in the flight from the city. Ammianus was ordered to return the boy to the city, and, surrounded by Persian cavalry, he barely missed being caught up in the siege himself. The swiftness of his mount enabled him to outrun his pursuers and to reach Ursicinus and his companions, who were pasturing their horses, unaware of the proximity of the enemy. Ammianus urged them to flee, but the risk of detection remained high due to the full moon and the level terrain. As a decoy, they attached a lantern to a mule and sent it to their left while they themselves headed for the mountains on the right. The party entered a wooded area which the inhabitants had abandoned. There they discovered a Roman soldier who admitted, under harsh questioning, to being a deserter and Persian spy. After providing the party with information about the enemies’ movements, he was put to death. From there the party hastened to the city of Amida, where scouts brought them information about the Persian advance hidden in the scabbard of a sword and couched in deliberately obscure language, which has only recently been fully understood (Blockley 1986).

  It was then decided that Ammianus would go to Corduene to gather more information about Persian troop movements from Jovinianus. At Corduene he claims to have seen the massive Persian army on the move (18.6.20–3). After determining the course of the army, Ammianus returned to Roman territory and made his report. Orders came down to remove peasants from their land and to burn the fields to limit the enemies’ fodder (18.7.1–3). As Ammianus and his party quickly destroyed several bridges to prevent the enemy from crossing, the carelessness of certain Roman cavalrymen allowed a Persian contingent to attack. Ammianus narrowly escaped and took refuge in Amida, which was filled with refugees and with farmers who had been participating in an annual fair nearby (18.8).

  Ammianus was then shut up inside the walls of Amida for seventy-three days while the Persian king Shapur besieged the city. Ammianus’ account of the siege (19.2–8) “is one of the high points of his narrative and a classic passage in Roman historical narrative,” according to Matthews (1989: 58). The city finally fell to Shapur, but the delay forced by the protracted siege left the Persians unable to capitalize on the victory. Ammianus himself slipped out of a gate with two companions when he recognized that the fall was imminent (19.8.5). Their escape was not without incident, but eventually, with the help of a captured runaway horse, they made their way to the Armenian town of Melitina, where they met up with a general whom they accompanied to Antioch (19.8.6–12).

  The fall of Amida had repercussions for Ursicinus’ career, and, one must imagine, for Ammianus’ as well. An investigation at court into the reasons for the fall led to intrigue against the general, who from frustration blamed his troubles on the emperor’s excessive deference to the palace eunuchs (20.2). As a result, Ursicinus was forced into retirement, and Ammianus disappears as an actor in his history for several years. He may have left the military when his sponsor did, or performed more mundane duties. A letter from Libanius to former students in Tarsus, written in 360, could possibly refer to Ammianus. In Barnes’ translation (1998: 61), Libanius writes, “to judge from the dress [of the man bearing this letter] he is enlisted in the army, but in fact he is enrolled among philosophers; he has imitated Socrates despite having gainful employment – the fine Ammianus” (ep. 233).

  After the revolt of Julian and the death of Constantius, the new emperor headed east to prepare for a renewed Persian campaign. Ammianus frequently discusses Julian’s recruitment of personnel for his new administration, and presumably Ammianus himself came out of retirement to rejoin the military when the emperor arrived at Antioch in 363. Ammianus reappears in the narrative just before the Roman army invaded Persian territory. He describes the crossing of the river Abora and then includes himself among the men who saw, ominously, the tomb of the third-century emperor Gordian, a military hero who was treacherously killed (23.5.7). The invasion of Persia was a dreadful failure, and Julian was killed. Ammianus records the difficult and dangerous retreat of the army, and the cession of the city of Nisibis to the Persians by the new emperor, Jovian. Ammianus’ last use of the first person occurs at 25.10.1 (“we came to Antioch”), and this suggests that he did not proceed on to Constantinople with Jovian (Matthews 1989: 13).

  During the next twenty years Ammianus must have traveled to gather information for his history and continued to compose his work. He claims personal familiarity with a few places which he must have visited during this time: Greece (26.10.19), Egypt (22.15.1), and the Black Sea (22.8.1), for example. He was in Antioch during the treason trials which Valens held in 370 and 371, and he describes the terror that gripped the city in book 29 of the work. Material for the earlier books of the Res Gestae could most easily be collected in the east, but proper coverage of Valentinian’s reign required the historian to head west. Ammianus writes with satirical anger about the expulsion of foreigners from Rome during a food crisis (14.6.19). If he was among those expelled, he would have had to be in the city by 383 or 384 at the latest.

  At Rome, Ammianus probably collected information about Valentinian’s campaigns on the Rhine, events in Rome itself, and affairs in North Africa. His complaints directed against a few senators, whom he accused of being unworthy of their proud heritage, have often led modern scholars to imagine him as bitter and alone. Such an analysis misinterprets the satirical persona which the historian has created in these passages and ignores the non-senatorial circles in which Ammianus probably traveled (Cameron 1964; Matthews 1989: 465–6). He integrated the new information he gathered into the work up to around 390. This suggests that Ammianus gave public readings from his work in 390 or 391, in time for details to reach Libanius in the east within a year. No further notices of his life or work remain.

  Work

  The extant text of the Res Gestae stretches from book 14 to book 31. The first books were lost at an early stage in the transmission of the manuscript. Book 14 begins in 353 and describes the last year of the life of the Caesar Gallus. Books 15 and 16 describe the rise of Julian as Caesar and his successful campaigning in Gaul, despite the attempts of the emperor Constantius II to obstruct his nephew. Books 17 and 18 alternate between Julian’s military success in Gaul and Constantius’ failure in the east. Book 19 is largely devoted to the siege of Amida by the Persian king Shapur, which ends in Roman defeat.

  In book 20, Julian is pronounced Augustus by his troops, and Constantius refuses to recognize his claim of equality. In book 21, the two armies begin their march and prepare for war, but Constantius dies of illness in October 361. In book 22, Julian, now sole ruler of the empire, continues east to Constantinople and Antioch, distributing patronage and establishing his position. He withdraws state support from Christianity and flaunts his paganism. At the end of book 22, Julian prepares for a major invasion of Persia, the subject of books 23 and 24. In book 25 the emperor is killed and the army manages its hasty retreat under the leadership of the newly chosen emperor Jovian, who is forced to sign an unfavorable peace treaty with Shapur. The book ends with the death of Jovian from smoke inhalation in 364.

  Book 26 begins with a preface that suggests that Ammianus is now moving up to “the boundaries of our own time.” The book covers the selection of Valentinian as emperor, his selection of his brother Valens as co-emperor, and Valens’ suppression of the revolt of Procopius, a cousin of Julian. Book 27 describes the military activities of Valentinian in the west and Valens in the east. Book 28 treats western events in Rome, Britain, and North Africa, particularly cases of unfair prosecutions and
corruption, and book 29 begins with parallel events in the east, where Valens oversaw a number of trials for treason and magic. The narrative then turns back to the west to treat the campaigns of Valentinian and the successful suppression of the African rebel Firmus by the general Theodosius, the father of the future emperor. Book 30 begins with eastern affairs as Valens and Shapur vie for control of Armenia. The scene shifts west, to the successful campaigning of Valentinian, who then dies of a stroke in 375. His son is elevated to the throne, despite his youth (he is 4 years old). Finally, book 31 is an almost continual narrative of how the admission of a group of Goths into the Roman empire turned into a disaster, culminating in the Battle of Adrianople and the death of Valens in 378.

  We know from the last sentence of the Res Gestae that the work began with the accession of Nerva in 96. The extant eighteen books cover the period from 353 to 378, for an average of less than a year and a half for each book. The earlier books clearly must have been narrated in far less detail than the books which survive. The first thirteen covered 257 years, for an average of about twenty years per book. The point at which the work shifted from severe compression to full and detailed narrative cannot be known, but Matthews (1989: 27) estimates that if it occurred at book 11 with the accession of Constantius in 337, the ten earlier books would have covered an average of twenty-five years per book, and if it occurred earlier, say in the seventh book with Constantine’s defeat of Licinius in 324, the first six books would have covered an average of forty years per book.

  The disparity between the lost portion and the surviving portion of the work, both in their scope and in the research methods necessary for their writing, once led some scholars to suggest that Ammianus was the author of two entirely separate works. Adherents to this theory argued that the first work covered the period from Nerva to perhaps the accession of Constantine, and the second work, of which we possess the second part, would have covered the period from Constantine to Adrianople with the sustained level of detail found in the surviving books. This theory is generally not accepted by scholars today. Careful study of the parts of the extant books which refer back to events of the lost books (Barnes 1998: 213–17; Frakes 1995; Gilliam 1972) reveals almost no information about that ill-recorded period that we do not already know from other sources, suggesting that the coverage could not have been especially detailed. Similarly, some important information about the second and third century, of the sort which would presumably have been covered in a full and detailed account, is provided in the surviving part of the work apparently for the first time. Moreover, some of the formulas used to refer back to lost material that would have been in the supposed first work are identical to formulas used to refer to material that would have been in the supposed second work. One might think that different terminology would be required when directing the reader to an entirely separate work than would be used to remind the reader of an episode related earlier in the book at hand.