The dating of the composition of books XXIX to XXXI of Ammianus’s Res Gestae has been greatly debated between scholars. It varies between 389 and 398 CE, that is before and after Eugenius’s usurpation (22th of August 392 CE - 6th of September 394 CE), and even before and after Theodosius’s death (17th January 395 CE) (for a survey of the bibliography see Sabbah, Ammien Marcellin, p. xxxiii-xxxiv, n. 95). Guy Sabbah has defended the idea that the atmosphere of books XXIX to XXXI fit in with the troubled period of Eugenius’s usurpation (see Sabbah, Ammien Marcellin, p. xxxiv-xxxvii). He has even suggested that the last book of the Res Gestae, from which the extract presented here stems, may have been composed at the time when the civil war between Eugenius and Theodosius I was officially declared, that is from the end of 394 CE up to the beginning of 395 CE (see Sabbah, Ammien Marcellin, p. xlvi-xlix). If that proposal is correct, the events narrated in book XXXI, in particular the conciliatory policy of Valens towards some Gothic groups and the terrible defeat of the Roman army against the Goths at Adrianople, may have been narrated by Ammianus also to warn Theodosius I against some dangers of his own politics. Moral reflections on Rome’s destiny pervade the Res Gestae, and more precisely with how Rome had always had to fight against all the external and internal elements that could cause its disintegration (see Sabbah, Ammien Marcellin, p. xli-xlii). In book XXXI, Ammianus deals with this theme and the way this book ends – with the narration of the massacre of numerous Goths after they had not been able to take advantage of the tremendous defeat they inflicted upon the Romans (on the context of which see later) – shows that he may have wanted to end the Res Gestae with a less pessimistic tone than if he had ended it with the defeat of Adrianople (Boeft, Drijvers, Hengst and Teitler, Philological, p. xv). At the beginning of the work, in Res Gestae XIV.6.3, Ammianus writes that Rome is vowed to live as long as there will be men (victura dum erunt homines Roma). He adds that the alliance of virtus and Fortuna have guaranteed her a glorious growth (sublimibus incrementis), and that without this alliance Rome would have never succeeded in reaching her ultimate greatness (ad perfectam summitatem) (Boeft, Drijvers, Hengst and Teitler, Philological, p. xi). Some years after, when he composed book XXXI, Ammianus was older, probably less optimistic about Rome’s future, and judged the events of 377-378 CE as being a serious challenge to Rome’s stability. However, Ammianus may have seen in the terrible anecdote described in the text presented here a glimmer of hope.
Book XXXI covers the period from the end of 376 CE to the summer or fall of 378 CE. It deals principally with the settlement of Goths in Thrace and with its consequences. In 376 CE, due to Hunnic pressures, some Gothic groups (essentially Goths Thervingi) asked the emperor of the East, Valens, to cross the Danube, and thus to enter into Roman territory, which Valens accepted. The Goths were granted food and lands to cultivate under the condition of remaining in peace and of furnishing some military contingents if necessary. The situation then deteriorated when, in 376 or 377 CE, another Gothic people, the Greuthungi, after having seen their request to be settled inside the Empire had been rejected, entered into Roman territory without any imperial permission. The Goths then fell out of Roman control and began to plunder the Thracian lands and cities. The imperial power had to react. In the first half of the year 377 CE, various fights opposed Roman troops against the Gothic groups. The tensions reached their peak during the summer of 378 CE, when Valens and the Roman army waged a war against the Gothic troops and their Hunnic and Alanic allies at Adrianople on the 9th of August (the battle is narrated in Res Gestae XXXI.12-13). Adrianople had been a double disaster for the Romans as it led to one of the most important slayings of Roman soldiers in a battle in the whole history of Rome (Ammianus compares it to the losses during the battle of Cannae; see Res Gestae XXXI.13.19; losses in the Roman Eastern field army have been estimated as between 10 000 and 20 000, see Halsall, Barbarian, p. 179). Moreover, it also led to the death of the emperor Valens. Book XXXI ends with two chapters (one of them corresponds to the text presented here) that narrate bloody events showing that even if the Goths severely defeated the Roman army, they were unable to maintain this success. Indeed, they failed to make the siege of Adrianople and to capture Constantinople. The text presented here appears after these two failures, when Julius, the commander-in-chief of the troops located trans Taurum, that is south and east of Mount Taurus, having heard about the disasters that happened in Thracia ordered the garrison commanders to gather their Gothic soldiers in the same place, with the promise that they would receive their payment (stipendium), and to kill them all (about Julius as magister equitum et peditum per Orientem see PLRE I, Iulius 2, p. 481; about the expression trans Taurum see Boeft, Drijvers, Hengst and Teitler, Philological, p. 291). This passage is interesting for various reasons. Ammianus presents it as an exemplum, and through it he conveys the idea that if the Romans took preventive and harsh measures against the troublesome barbarians, they could succeed in pushing them out of the Roman provinces (see Sabbah, Ammien Marcellin, p. xxxi; Boeft, Drijvers, Hengst and Teitler, Philological, p. xv-xvi; about the fact that Julius’s massacre was a preventive one see Colombo, “Annotazioni,” p. 262). It is also important to note that Julius knew the rhetor Libanius and probably also Ammianus. As recalled by John Matthews, Julius had his official residence at Antioch on the Orontes, a city that may have also been Ammianus’s place of origin, or at least a place in which he had been resident. Thus, the fact that Ammianus welcomes Julius’s actions against the Goths so positively – Antioch was actually part of the cities located at the south-east of Mount Taurus – can be explained by the fact that Julius freed his native region from barbarian danger (see Matthews, The Roman Empire, p. 227). Another interesting element is Ammianus’s remark about the origin of these rectores to whom Julius sent his orders. He writes that they were all Roman and that this happened rarely at that time (quod his temporibus raro contingit). This short remark has been variously interpreted and we will question it in order to see what it may reveal from Ammianus’s opinions about the increasing influence of the barbarians into the affairs of the Empire.
Before analysing Ammianus’s remark about the fact that Romans were rarely rectores of Goths it is necessary to return to two debated elements of the passage presented here: first, its dating, and second the identity of the Goths present in Asia Minor here described. To answer this question many scholars have compared Ammianus’s version of the events with the one that appears in Zosimus, New History IV.26. Zosimus narrates that when the Goths crossed the Danube they were allowed to settle in Thrace provided that they gave children as hostages. Valens ordered that they had to be sent in the East. There Julius became their educator and their guardian, and decided that they should be distributed in various cities. Having heard the news of the disappointments of the Goths in Thrace, these young hostages revolted in various Eastern cities. Julius decided not to reveal the on-going troubles to the emperor Theodosius, who was then near Macedonia. After the Senate of Constantinople told him to do what he considered necessary, Julius conveyed this to all the men who led the armed units (Zosimus does not mention their origin) and ordered them to spread the message that he wanted to grant to the Goths gifts and lands. The barbarians were gathered in the cities in which they had been ordered to go, and were killed. Scholars are divided between those who follow Zosimus’s version and who consider that the events took place under Theodosius’s reign at the beginning of 379 CE, and others who consider that it took place just after the battle of Adrianople, that is, during the late summer or autumn of 378 CE. The second option seems to be more plausible (see Colombo, “Annotazioni,” p. 262-263; for the bibliography Boeft, Drijvers, Hengst and Teitler, Philological, p. 290, 296). Another debated element is related to the identity of the Goths present in the East. Zosimus’s version of the events presents one major discrepancy: the young children became, in just one year, men able to foment revolts. In addition, it has been shown that Zosimus must have distorted his source, Eunapius, who does not relate that the young hostages were sent to the East, but rather speaks of Thrace, Macedonia and Thessaly (see Boeft, Drijvers, Hengst and Teitler, Philological, p. 293-294). Zosimus thus seems to have mixed various elements, and Ammianus’s testimony of this event may be more reliable (see Boeft, Drijvers, Hengst and Teitler, Philological, p. 296). Moreover, the fact that Ammianus takes the trouble to specify that the Goths had been dispersed through cities and “military camps” (castra) show that they may have been auxiliary soldiers, that is, Goths who served in military units mostly composed of barbarians (Boeft, Drijvers, Hengst and Teitler, Philological, p. 194). In addition, while the term rectores has a quite general meaning and refers to leaders, Ammianus uses it most of the time to refer to “military commanders” in regular military units (Colombo, “Annotazioni,” p. 263-264). These elements show that the Goths slain by Julius in the East were soldiers who served Rome, possibly in auxiliary units.
Scholars have also significantly debated how Ammianus’s remark about the fact that his temporibus, “at this/our time,” should be interpreted, as it was pretty rare to find Romans as military commanders of barbarian soldiers. The expression his temporibus is ambiguous as Ammianus could be referring to Julius’s time or to the period during which he was writing. Some scholars have interpreted this short remark as revealing Ammianus’s opposition to the barbarisation of the Roman army, and especially to the growing numbers of officers of barbarian origin (see Paschoud, Roma aeterna, p. 43-46). The enlisting of barbarians into the Roman army was not a phenomenon limited to the fourth century. Already during the third century BCE some Hispanic, Numidian, Gallic, or Germanic allies of Rome fought alongside Roman generals. Then, Augustus reformed the participation of peregrines to the Roman army. Many of them were organised in permanent units and served as professional soldiers under the command of equestrian officers. During the third century, more and more barbarians were enlisted in the army. The sources of this enlisting varied: some barbarians voluntarily proposed their service, others were part of peoples who had been settled in the Empire and who, in return, had to provide military help to Rome when needed, or on a regular basis. The increasing enlisting of barbarians continued during the fourth century, and while an old historiographical tradition has overestimated the proportion of barbarians enlisted in the Roman army, it is now generally admitted that the barbarian contingents may have represented no more than 25% of the global force. Yet, during the fourth century CE there was a clear overrepresentation in the high-command of officers of barbarian origins (Janniard, “Armée romaine,” p. 244). By then, it is interesting to note that in some Jewish texts it is the heterogeneous origins of the soldiers enlisted by Rome which is highlighted to characterise the Roman power negatively. For instance, in Genesis Rabbah 42:4, the “evil kingdom,” that is Rome, “enlists its soldiers from all the nations of the world” (this idea is also expressed in Pesiqta de Rav Kahana 5:7 (ha-ḥodesh ha-ze, piskah 7)). A similar idea appears in a third century Christian text very critical towards Rome, namely the Commentary on Daniel. The author of this text writes that Rome “is a collection of all languages and of all human races and it is a raising of troops in preparation for the war, who are all called Romans, but do not originate from one country” (Commentary on Daniel IV.8). So, in this Christian text as in the later Jewish text, Rome is characterised by the fact that its army is a collection of peoples from various nations, even if the author of the Commentary insists more on the fact that they are integrated into a common entity, as they are called Romans. In Ammianus’s text the partition is clear between the Gothic soldiers and their Roman leaders.
To argue that through this sentence Ammianus expressed a radical opposition to the presence of barbarians within the army and in high-command positions may nevertheless be erroneous. Himself a former army officer, Ammianus must have risen up through the ranks alongside soldiers and officers of barbarian origins. Moreover, in the Res Gestae he repeatedly praises the qualities of some barbarians who had a high position in the Roman military hierarchy (see for instance the example of the Sarmate magister equitum Victor in Res Gestae XXXI.12.6). Ammianus was probably not opposed to the presence of barbarians in the Roman army, nor in high-command, yet he expected these barbarians to serve Rome loyally (in that perspective, see Mary, “Ammien Marcellin,” p. 190). We thus rather suggest that through the remark quod his temporibus raro contingit, Ammianus expressed a judgement about evolutions of the high-command of barbarian units that occurred at the time he composed the Res Gestae, that is under Theodosius’s reign. Numerous scholars have connected this remark with the treaty of 382 CE that was concluded between Rome and various Gothic peoples after three years of confrontation (on whether this treaty was a deditio rather than a foedus see Halsall, Barbarian, p. 180-185; about the debate related to this treaty, see Delaplace, La fin de l’Empire, p. 91-94). In 382 CE, the Goths were granted lands, and in return they had to provide troops to the Empire and they might have been allowed to serve under their own leaders. It might be this authorisation granted to the Goths to serve under their own leaders that was targeted by Ammianus through this short remark (in that perspective, see Sabbah and Angliviel de la Beaumelle, Ammien Marcellin, p. 293, n. 587; Boeft, Drijvers, Hengst and Teitler, Philological, p. 294). Another possible explanation might be that to wage war against Eugenius, Theodosius went in the West with military contingents of Huns and Goths which were commanded by barbarian commanders (expedition narrated in Claudian, Panegyricus de tertio consulatu Honorii 68-74; De Consulatu Stiliconis I.155-160). Actually, in 394 CE, two of the most important generals of Theodosius, Saul and Gainas, were of barbarian origin, and, together with Bacurius, they commanded the barbarian contingents during the campaign against Eugenius (see Zosimus, New History IV.57.2-3). Yet, due to their number and the troubles they may have caused on their way, these barbarian troops must have been a source of concern and fear for the provincial populations (see Sabbah, Ammien Marcellin, p. xlvi). Ammianus may thus have wanted to implicitly warn Theodosius against the dangers of not controlling these important barbarian contingents, but also against the dramatic consequences of a politic excessively conciliating towards the barbarians. By then, immediately after the victory of the troops of Theodosius against Eugenius at the Frigidus on the 6th of September 394 CE, some Gothic contingents who had not been paid as they hoped went back to Thrace where they revolted. Then, in February 395 CE, they besieged Constantinople and compelled the praetorian prefect of the East, Rufinus, to reach an agreement with them. It is thus highly possible that through the narrative of events of the end of 378 CE Ammianus wanted to warn against some aspects of Theodosius’s policy regarding the relationships that the Roman power should keep with the various barbarian groups (in that perspective, see Sabbah, Ammien Marcellin, p. xlvi-xlvii). Finally, Ammianus may have also inserted this anecdote about the general Julius who severely repressed the troublesome barbarians present in the cities of this Eastern region in order to indirectly criticize the fact that under Theodosius some Roman generals who had accomplished similar actions to save some Roman provinces from troublesome barbarian groups had been punished by the imperial power. This is the case with the Roman general Gerontius, who, in Thrace in 386 CE, had been punished for a similar massive slaughter of Gothic groups who had been settled in the region by the imperial power and who caused some troubles. Theodosius punished Gerontius severely for having killed the Goths that had been settled by him in the region (episode narrated in Zosimus, New History IV.40; see Sabbah, Ammien Marcellin, p. xlvii-xlviii, and n. 139).
To conclude, it would be wrong to interpret this short remark of Ammianus concerning the few Romans who were military commanders of barbarian soldiers as being a piece of evidence of Ammianus’s radical opposition to the fact that many barbarians served Rome militarily and that many officers were of barbarian origins. What Ammianus may have wanted to implicitly criticise through this historical narrative was the fact that under Theodosius I the imperial power granted more and more freedom to the barbarian groups settled in the Empire; the imperial policy was thus too soft and thus inadequate. On Ammianus’s choice to end his Res Gestae with the story of Julius, Hagith Sivan rightly concludes: “By concluding with a bloody massacre rather than with a story of peaceful settlements, the aged soldier expressed his dissatisfaction with current imperial policies” (Sivan, “Ammianus’ Terminus,” p. 118). Finally, this passage and many others of the Res Gestae dealing with barbarian groups settled in the Empire and/or serving Rome militarily shows that an aged soldier like Ammianus characterised first and foremost the protagonists of his history through the opposition Romans/foreigners, the various barbarians being characterised not by their status regarding Rome but first and foremost by their ethnic origins. The legal statuses of the peoples of barbarian origins present in the Balkans and in the Eastern regions at the end of the fourth century varied a lot according to the way they had been settled in the Empire, according to the way they served Rome militarily, and according to the length of time they had been present inside the Empire. All this complexity is denied by Ammianus, who, in this episode, presents the dichotomy Romans/non-Romans and the severity of the Romans as being the key elements explaining the rescue of the Eastern provinces from the barbarian threat.
For a presentation of Ammianus’s life and of his work, the Res Gestae, see Ammianus Marcellinus, Res Gestae XIV.6.3-6.
The dating of the composition of books XXIX to XXXI of Ammianus’s Res Gestae has been greatly debated between scholars. It varies between 389 and 398 CE, that is before and after Eugenius’s usurpation (22th of August 392 CE - 6th of September 394 CE), and even before and after Theodosius’s death (17th January 395 CE) (for a survey of the bibliography see Sabbah, Ammien Marcellin, p. xxxiii-xxxiv, n. 95). Guy Sabbah has defended the idea that the atmosphere of books XXIX to XXXI fit in with the troubled period of Eugenius’s usurpation (see Sabbah, Ammien Marcellin, p. xxxiv-xxxvii). He has even suggested that the last book of the Res Gestae, from which the extract presented here stems, may have been composed at the time when the civil war between Eugenius and Theodosius I was officially declared, that is from the end of 394 CE up to the beginning of 395 CE (see Sabbah, Ammien Marcellin, p. xlvi-xlix). If that proposal is correct, the events narrated in book XXXI, in particular the conciliatory policy of Valens towards some Gothic groups and the terrible defeat of the Roman army against the Goths at Adrianople, may have been narrated by Ammianus also to warn Theodosius I against some dangers of his own politics.
Moral reflections on Rome’s destiny pervade the Res Gestae, and more precisely with how Rome had always had to fight against all the external and internal elements that could cause its disintegration (see Sabbah, Ammien Marcellin, p. xli-xlii). In book XXXI, Ammianus deals with this theme and the way this book ends – with the narration of the massacre of numerous Goths after they had not been able to take advantage of the tremendous defeat they inflicted upon the Romans (on the context of which see later) – shows that he may have wanted to end the Res Gestae with a less pessimistic tone than if he had ended it with the defeat of Adrianople (Boeft, Drijvers, Hengst and Teitler, Philological, p. xv). At the beginning of the work, in Res Gestae XIV.6.3, Ammianus writes that Rome is vowed to live as long as there will be men (victura dum erunt homines Roma). He adds that the alliance of virtus and Fortuna have guaranteed her a glorious growth (sublimibus incrementis), and that without this alliance Rome would have never succeeded in reaching her ultimate greatness (ad perfectam summitatem) (Boeft, Drijvers, Hengst and Teitler, Philological, p. xi). Some years after, when he composed book XXXI, Ammianus was older, probably less optimistic about Rome’s future, and judged the events of 377-378 CE as being a serious challenge to Rome’s stability. However, Ammianus may have seen in the terrible anecdote described in the text presented here a glimmer of hope.
Book XXXI covers the period from the end of 376 CE to the summer or fall of 378 CE. It deals principally with the settlement of Goths in Thrace and with its consequences. In 376 CE, due to Hunnic pressures, some Gothic groups (essentially Goths Thervingi) asked the emperor of the East, Valens, to cross the Danube, and thus to enter into Roman territory, which Valens accepted. The Goths were granted food and lands to cultivate under the condition of remaining in peace and of furnishing some military contingents if necessary. The situation then deteriorated when, in 376 or 377 CE, another Gothic people, the Greuthungi, after having seen their request to be settled inside the Empire had been rejected, entered into Roman territory without any imperial permission. The Goths then fell out of Roman control and began to plunder the Thracian lands and cities. The imperial power had to react. In the first half of the year 377 CE, various fights opposed Roman troops against the Gothic groups. The tensions reached their peak during the summer of 378 CE, when Valens and the Roman army waged a war against the Gothic troops and their Hunnic and Alanic allies at Adrianople on the 9th of August (the battle is narrated in Res Gestae XXXI.12-13). Adrianople had been a double disaster for the Romans as it led to one of the most important slayings of Roman soldiers in a battle in the whole history of Rome (Ammianus compares it to the losses during the battle of Cannae; see Res Gestae XXXI.13.19; losses in the Roman Eastern field army have been estimated as between 10 000 and 20 000, see Halsall, Barbarian, p. 179). Moreover, it also led to the death of the emperor Valens.
Book XXXI ends with two chapters (one of them corresponds to the text presented here) that narrate bloody events showing that even if the Goths severely defeated the Roman army, they were unable to maintain this success. Indeed, they failed to make the siege of Adrianople and to capture Constantinople. The text presented here appears after these two failures, when Julius, the commander-in-chief of the troops located trans Taurum, that is south and east of Mount Taurus, having heard about the disasters that happened in Thracia ordered the garrison commanders to gather their Gothic soldiers in the same place, with the promise that they would receive their payment (stipendium), and to kill them all (about Julius as magister equitum et peditum per Orientem see PLRE I, Iulius 2, p. 481; about the expression trans Taurum see Boeft, Drijvers, Hengst and Teitler, Philological, p. 291). This passage is interesting for various reasons. Ammianus presents it as an exemplum, and through it he conveys the idea that if the Romans took preventive and harsh measures against the troublesome barbarians, they could succeed in pushing them out of the Roman provinces (see Sabbah, Ammien Marcellin, p. xxxi; Boeft, Drijvers, Hengst and Teitler, Philological, p. xv-xvi; about the fact that Julius’s massacre was a preventive one see Colombo, “Annotazioni,” p. 262). It is also important to note that Julius knew the rhetor Libanius and probably also Ammianus. As recalled by John Matthews, Julius had his official residence at Antioch on the Orontes, a city that may have also been Ammianus’s place of origin, or at least a place in which he had been resident. Thus, the fact that Ammianus welcomes Julius’s actions against the Goths so positively – Antioch was actually part of the cities located at the south-east of Mount Taurus – can be explained by the fact that Julius freed his native region from barbarian danger (see Matthews, The Roman Empire, p. 227). Another interesting element is Ammianus’s remark about the origin of these rectores to whom Julius sent his orders. He writes that they were all Roman and that this happened rarely at that time (quod his temporibus raro contingit). This short remark has been variously interpreted and we will question it in order to see what it may reveal from Ammianus’s opinions about the increasing influence of the barbarians into the affairs of the Empire.
Before analysing Ammianus’s remark about the fact that Romans were rarely rectores of Goths it is necessary to return to two debated elements of the passage presented here: first, its dating, and second the identity of the Goths present in Asia Minor here described. To answer this question many scholars have compared Ammianus’s version of the events with the one that appears in Zosimus, New History IV.26. Zosimus narrates that when the Goths crossed the Danube they were allowed to settle in Thrace provided that they gave children as hostages. Valens ordered that they had to be sent in the East. There Julius became their educator and their guardian, and decided that they should be distributed in various cities. Having heard the news of the disappointments of the Goths in Thrace, these young hostages revolted in various Eastern cities. Julius decided not to reveal the on-going troubles to the emperor Theodosius, who was then near Macedonia. After the Senate of Constantinople told him to do what he considered necessary, Julius conveyed this to all the men who led the armed units (Zosimus does not mention their origin) and ordered them to spread the message that he wanted to grant to the Goths gifts and lands. The barbarians were gathered in the cities in which they had been ordered to go, and were killed. Scholars are divided between those who follow Zosimus’s version and who consider that the events took place under Theodosius’s reign at the beginning of 379 CE, and others who consider that it took place just after the battle of Adrianople, that is, during the late summer or autumn of 378 CE. The second option seems to be more plausible (see Colombo, “Annotazioni,” p. 262-263; for the bibliography Boeft, Drijvers, Hengst and Teitler, Philological, p. 290, 296).
Another debated element is related to the identity of the Goths present in the East. Zosimus’s version of the events presents one major discrepancy: the young children became, in just one year, men able to foment revolts. In addition, it has been shown that Zosimus must have distorted his source, Eunapius, who does not relate that the young hostages were sent to the East, but rather speaks of Thrace, Macedonia and Thessaly (see Boeft, Drijvers, Hengst and Teitler, Philological, p. 293-294). Zosimus thus seems to have mixed various elements, and Ammianus’s testimony of this event may be more reliable (see Boeft, Drijvers, Hengst and Teitler, Philological, p. 296). Moreover, the fact that Ammianus takes the trouble to specify that the Goths had been dispersed through cities and “military camps” (castra) show that they may have been auxiliary soldiers, that is, Goths who served in military units mostly composed of barbarians (Boeft, Drijvers, Hengst and Teitler, Philological, p. 194). In addition, while the term rectores has a quite general meaning and refers to leaders, Ammianus uses it most of the time to refer to “military commanders” in regular military units (Colombo, “Annotazioni,” p. 263-264). These elements show that the Goths slain by Julius in the East were soldiers who served Rome, possibly in auxiliary units.
Scholars have also significantly debated how Ammianus’s remark about the fact that his temporibus, “at this/our time,” should be interpreted, as it was pretty rare to find Romans as military commanders of barbarian soldiers. The expression his temporibus is ambiguous as Ammianus could be referring to Julius’s time or to the period during which he was writing. Some scholars have interpreted this short remark as revealing Ammianus’s opposition to the barbarisation of the Roman army, and especially to the growing numbers of officers of barbarian origin (see Paschoud, Roma aeterna, p. 43-46). The enlisting of barbarians into the Roman army was not a phenomenon limited to the fourth century. Already during the third century BCE some Hispanic, Numidian, Gallic, or Germanic allies of Rome fought alongside Roman generals. Then, Augustus reformed the participation of peregrines to the Roman army. Many of them were organised in permanent units and served as professional soldiers under the command of equestrian officers. During the third century, more and more barbarians were enlisted in the army. The sources of this enlisting varied: some barbarians voluntarily proposed their service, others were part of peoples who had been settled in the Empire and who, in return, had to provide military help to Rome when needed, or on a regular basis. The increasing enlisting of barbarians continued during the fourth century, and while an old historiographical tradition has overestimated the proportion of barbarians enlisted in the Roman army, it is now generally admitted that the barbarian contingents may have represented no more than 25% of the global force. Yet, during the fourth century CE there was a clear overrepresentation in the high-command of officers of barbarian origins (Janniard, “Armée romaine,” p. 244). By then, it is interesting to note that in some Jewish texts it is the heterogeneous origins of the soldiers enlisted by Rome which is highlighted to characterise the Roman power negatively. For instance, in Genesis Rabbah 42:4, the “evil kingdom,” that is Rome, “enlists its soldiers from all the nations of the world” (this idea is also expressed in Pesiqta de Rav Kahana 5:7 (ha-ḥodesh ha-ze, piskah 7)). A similar idea appears in a third century Christian text very critical towards Rome, namely the Commentary on Daniel. The author of this text writes that Rome “is a collection of all languages and of all human races and it is a raising of troops in preparation for the war, who are all called Romans, but do not originate from one country” (Commentary on Daniel IV.8). So, in this Christian text as in the later Jewish text, Rome is characterised by the fact that its army is a collection of peoples from various nations, even if the author of the Commentary insists more on the fact that they are integrated into a common entity, as they are called Romans. In Ammianus’s text the partition is clear between the Gothic soldiers and their Roman leaders.
To argue that through this sentence Ammianus expressed a radical opposition to the presence of barbarians within the army and in high-command positions may nevertheless be erroneous. Himself a former army officer, Ammianus must have risen up through the ranks alongside soldiers and officers of barbarian origins. Moreover, in the Res Gestae he repeatedly praises the qualities of some barbarians who had a high position in the Roman military hierarchy (see for instance the example of the Sarmate magister equitum Victor in Res Gestae XXXI.12.6). Ammianus was probably not opposed to the presence of barbarians in the Roman army, nor in high-command, yet he expected these barbarians to serve Rome loyally (in that perspective, see Mary, “Ammien Marcellin,” p. 190). We thus rather suggest that through the remark quod his temporibus raro contingit, Ammianus expressed a judgement about evolutions of the high-command of barbarian units that occurred at the time he composed the Res Gestae, that is under Theodosius’s reign. Numerous scholars have connected this remark with the treaty of 382 CE that was concluded between Rome and various Gothic peoples after three years of confrontation (on whether this treaty was a deditio rather than a foedus see Halsall, Barbarian, p. 180-185; about the debate related to this treaty, see Delaplace, La fin de l’Empire, p. 91-94). In 382 CE, the Goths were granted lands, and in return they had to provide troops to the Empire and they might have been allowed to serve under their own leaders. It might be this authorisation granted to the Goths to serve under their own leaders that was targeted by Ammianus through this short remark (in that perspective, see Sabbah and Angliviel de la Beaumelle, Ammien Marcellin, p. 293, n. 587; Boeft, Drijvers, Hengst and Teitler, Philological, p. 294). Another possible explanation might be that to wage war against Eugenius, Theodosius went in the West with military contingents of Huns and Goths which were commanded by barbarian commanders (expedition narrated in Claudian, Panegyricus de tertio consulatu Honorii 68-74; De Consulatu Stiliconis I.155-160). Actually, in 394 CE, two of the most important generals of Theodosius, Saul and Gainas, were of barbarian origin, and, together with Bacurius, they commanded the barbarian contingents during the campaign against Eugenius (see Zosimus, New History IV.57.2-3). Yet, due to their number and the troubles they may have caused on their way, these barbarian troops must have been a source of concern and fear for the provincial populations (see Sabbah, Ammien Marcellin, p. xlvi). Ammianus may thus have wanted to implicitly warn Theodosius against the dangers of not controlling these important barbarian contingents, but also against the dramatic consequences of a politic excessively conciliating towards the barbarians. By then, immediately after the victory of the troops of Theodosius against Eugenius at the Frigidus on the 6th of September 394 CE, some Gothic contingents who had not been paid as they hoped went back to Thrace where they revolted. Then, in February 395 CE, they besieged Constantinople and compelled the praetorian prefect of the East, Rufinus, to reach an agreement with them. It is thus highly possible that through the narrative of events of the end of 378 CE Ammianus wanted to warn against some aspects of Theodosius’s policy regarding the relationships that the Roman power should keep with the various barbarian groups (in that perspective, see Sabbah, Ammien Marcellin, p. xlvi-xlvii). Finally, Ammianus may have also inserted this anecdote about the general Julius who severely repressed the troublesome barbarians present in the cities of this Eastern region in order to indirectly criticize the fact that under Theodosius some Roman generals who had accomplished similar actions to save some Roman provinces from troublesome barbarian groups had been punished by the imperial power. This is the case with the Roman general Gerontius, who, in Thrace in 386 CE, had been punished for a similar massive slaughter of Gothic groups who had been settled in the region by the imperial power and who caused some troubles. Theodosius punished Gerontius severely for having killed the Goths that had been settled by him in the region (episode narrated in Zosimus, New History IV.40; see Sabbah, Ammien Marcellin, p. xlvii-xlviii, and n. 139).
To conclude, it would be wrong to interpret this short remark of Ammianus concerning the few Romans who were military commanders of barbarian soldiers as being a piece of evidence of Ammianus’s radical opposition to the fact that many barbarians served Rome militarily and that many officers were of barbarian origins. What Ammianus may have wanted to implicitly criticise through this historical narrative was the fact that under Theodosius I the imperial power granted more and more freedom to the barbarian groups settled in the Empire; the imperial policy was thus too soft and thus inadequate. On Ammianus’s choice to end his Res Gestae with the story of Julius, Hagith Sivan rightly concludes: “By concluding with a bloody massacre rather than with a story of peaceful settlements, the aged soldier expressed his dissatisfaction with current imperial policies” (Sivan, “Ammianus’ Terminus,” p. 118). Finally, this passage and many others of the Res Gestae dealing with barbarian groups settled in the Empire and/or serving Rome militarily shows that an aged soldier like Ammianus characterised first and foremost the protagonists of his history through the opposition Romans/foreigners, the various barbarians being characterised not by their status regarding Rome but first and foremost by their ethnic origins. The legal statuses of the peoples of barbarian origins present in the Balkans and in the Eastern regions at the end of the fourth century varied a lot according to the way they had been settled in the Empire, according to the way they served Rome militarily, and according to the length of time they had been present inside the Empire. All this complexity is denied by Ammianus, who, in this episode, presents the dichotomy Romans/non-Romans and the severity of the Romans as being the key elements explaining the rescue of the Eastern provinces from the barbarian threat.