II—
Gaïnas and Tribigild
The new prefect Caesarius did his best to salvage the situation. Synesius describes, with real or affected outrage, how even after the massacre Typhos "asked for negotiations with the barbarians, scheming again to admit the enemy army on the grounds that no irremediable evil had occurred" (121B). Synesius professes astonishment that even this manifest treachery did not bring immediate retribution down on Typhos's head, but Gaïnas himself had played no part in the massacre, and it was the Goths, for the most part unarmed civilians, who had been the victims. Synesius's evident anxiety not to offend Gaïnas in book 1, written in the days immediately after the massacre,[114] is enough to show that there was a short-lived fear that he would take revenge. Such fear tells against Socrates' claim that Arcadius declared Gaïnas a public enemy before the massacre.[115] There may have been several days of negotiations
[112] Socrates mentions the burning of the church but does not actually say that Goths were inside it at the time; Synesius refers callously to the Egyptians "smoking them out like wasps, together with their temples and priests" (121A). But according to Zosimus, "the more devout Christians considered that a grave defilement had been perpetrated in the midst of their city" (5.19.5).
[113] Cf. McCormick 1986, 195, on the standard use of scriptural quotation to introduce later Byzantine victory proclamations.
[114] See chapter 8, section III.
before Gaïnas finally made the irrevocable decision to follow the same path as Tribigild and Alaric.
Who was responsible for the declaration of war on Gaïnas and the dispatch of Fravitta? According to Demougeot, it was Caesarius, "looking for a compromise with the antibarbarian party."[116] They were happy with Fravitta because, though a Goth, he was "one of those rare foreigners converted to Greek paganism."[117] There is not a word about nationalists or compromise in any of our sources."[118] Which nationalists, anyway? Aurelian, Saturninus, and John were still in exile. This explanation inadequately and unnecessarily attempts to palliate the unwelcome, but obvious, fact that it was the supposedly pro-barbarian Caesarius who took prompt and decisive action against Gaïnas. In reality there is neither problem nor paradox. There was no pro-barbarian party. Caesarius had already enjoyed and still held the highest honors the Roman state could offer. Why should he have wished to collaborate with an incompetent and unpopular barbarian? As soon as the situation allowed, he did what he could to eliminate the problem.
Where did Fravitta and his army come from? Demougeot assumed that the army consisted of citizen soldiers and deserters from Gaïnas's army, an improvised force hastily trained by Fravitta.[119] There may well have been a certain number of deserters, and there were also the 3,500 guardsmen. But the core of the army must have been some portion of the troops already under Fravitta's command in his capacity as magister militum per Orientem .
It is hard to believe that in a matter of weeks Fravitta could have improvised and trained from deserters and raw recruits an army capable of defeating the army of the magister militum praesentalis .[120] On the other hand the Eastern army must have been somewhere. The Notitia dignitatum gives a complete list of the thirty-five units "sub dispositione viri
[116] Demougeot 1951, 259–60.
[118] Against the strange notion that these (nonexistent) nationalists were sympathetic to paganism, see above, chapter 3, section I.
[119] Demougeot 1951, 260.
[120] Indeed Gaïnas had probably absorbed in addition what remained of the forces of his late fellow praesentalis , Leo.
illustris magistri militum per Orientem."[121] It is true that the Eastern part of the Notitia as we have it dates from no later than ca. 394.[122] But when raising the army he took west with him, Theodosius will not have disturbed the troops on the eastern frontier,[123] essential to its protection against the very real threat of a Persian invasion. Just such a threat had in fact suddenly arisen in the course of 399.[124] Even if false, the alarm betrays an instinctive apprehension that Persia would exploit any deflection of the Roman war effort in the East, such as Tribigild's rebellion. Whatever the differences between Rufinus, Eutropius, Aurelian, and Caesarius, it is not easy to believe that any of the four would have dared to weaken the defense of the eastern frontier.
The civilian ministers of Arcadius may have feared the ambitions of the magistri militum praesentales , who were stationed at court: Eutropius seems not to have appointed one at all between ca. 395 (the dismissal of Timasius) and 399 (the appointment of Leo and Gaïnas). But there is solid evidence for magistri militum per Orientem continuously from 393 to 398: Addaeus from 393 to 396 and Simplicius from December 396 to March 398.[125] Zosimus does not give Fravitta's rank at the time of his appointment to the command against Gaïnas in the summer of 400 but reports that he had "freed the entire East from Cilicia to Phoenice and Palestine from bandits" (5.20.1). To have campaigned over so wide an area he must have been magister militum per Orientem . In confirmation, a fragment that is clearly extracted from the passage of Eunapius on which Zosimus here depends describes Fravitta specifically as "general of the East" (
[121] ND Or . 7, as reconstructed by Hoffmann 1969 (Beilage, pp. 4–5).
[122] Indeed it has been plausibly suggested that our text descends from a copy of the official Eastern Notitia in the possession of an Eastern bureaucrat in Theodosius's army when he marched west against Eugenius in 394. This copy then remained in the West, where Western lists were added and maintained to a somewhat later date: Hoffmann 1969, 52–53, 516–19; Barnes 1978, 82.
[123] Indeed there is good reason to believe that the defense of the eastern frontier had remained unchanged in its essentials since Diocletian: Jones 1964a, 3:357.
[124] Cameron 1970a, 140.
[125] PLRE 1.13, II.1013–14; Demandt 1970, 728.
with which Zosimus credits him suggest at least one campaigning season, and there is no reason to doubt that he was the direct successor of Simplicius, in office possibly from late 398, but certainly by 399.[128] The normal peacetime headquarters of the MVM per Orientem was Antioch; he is bound to have followed the movements of Tribigild and Gaïnas throughout 399–400, poised to take appropriate action.
One source actually refers to the existence of the eastern frontier army before the massacre. According to Socrates, one night Gaïnas sent his barbarians to burn down the palace.[129] They were repulsed by a band of angels in the form of gigantic warriors and reported the presence of this "large and noble army" to their leader. Gaïnas did not believe them, "for he knew that the bulk of the Roman army was some way away, stationed among the cities."[130] Of course, the context is not reassuring. Not to mention the angels disguised as soldiers, it is most unlikely that the barbarians attempted to burn the palace in the first place. Nonetheless, the army at the disposal of the magister militum per Orientem was indeed "some way away," dispersed among the various eastern frontier provinces.[131]
It is worth exploring why Fravitta did not act earlier. If Gaïnas had really seized Constantinople in April 400 and held it by force till 12 July, Fravitta could not have stood idly by. If Gaïnas had really joined forces with Tribigild the previous summer and helped him plunder Asia, as Zosimus alleges, Fravitta could hardly have ignored that either.[132] A simple and revealing explanation of such protracted inaction suggests itself.
Historians have always taken it for granted that Gaïnas really did join forces with Tribigild in the summer of 399, openly revealing himself as a rebel long before he "seized" Constantinople in April 400. Certainly contemporaries believed that the two were in league. It was natural, perhaps inevitable, that in retrospect Gaïnas should be thought to have planned every stage of his coup in advance. According to Zosimus, he
[128] So Demandt 1970, 728.
[129] HE 6.6.
[131] See map 4 in Jones 1964a and map 3 ("Kleinasien und Orient") in Hoffmann 1970.
[132] Though if there really was a Persian threat, it is possible that part at least of the eastern frontier army was being mobilized in case of invasion.
and Tribigild planned the whole thing in Constantinople early in 399 before Tribigild went out to Phrygia. The fullest and most recent modern discussion, by Paschoud, sees no reason to doubt this version.[133] But apart from serious internal improbabilities in Zosimus's account, analyzed below, both Claudian and Synesius tell against it.
In his narrative Synesius says only that the barbarian general and his troops "were waging an unsuccessful war against a rebellious contingent of their own people" (108B). But a few lines later Typhos's wife alleges that Osiris was planning to accuse the general of treason on the grounds "that he was fighting a collusive war, the barbarians pursuing a common policy with divided armies" (108C). This is another neat illustration of the way book 1 avoids accusing Gaïnas outright: the narrative is neutral, with the accusation coming in a speech by a character presumed to be lying. If called, the author can disavow it.[134] The status of the allegation is therefore unclear. Synesius no doubt believed it. He had freely predicted barbarian treachery in De regno . But his very evasiveness here suggests that it was less than established fact.
Early though the testimony of Synesius is, even book 1 was not written till after the massacre. Book 2 of Claudian's In Eutropium seems to have been written before Eutropius's fall. It is true that the elegiac preface and proem (lines 1–23) allude to his exile (August 399), but the rest of the book (24–602) presupposes that he was still in power. In particular, responsibility for the outbreak and success of Trigibild's rebellion is laid squarely on Eutropius. An elaborate section (376–461) describes how Leo's incompetence allowed Tribigild to destroy the Roman army and rampage unchecked through Pamphylia and Pisidia. The book closes with a denunciation of Eutropius that merges into an appeal to Stilicho for rescue (550–90). Lines 562–83 review Tribigild's revolt, claiming that Eutropius ignored it and thought only of dancing and feasting; to compensate for the revenue from the lost provinces he simply divides those that remain in half! The concluding appeal to Stilicho begs:
eripe me tandem servilibus eripe regnis.
(In Eutr . II.593)
Save me, save me at last from the servile kingdom.
[133] Paschoud 1986,124f., quoting earlier bibliography.
[134] This crucial distinction between narrative and speech is fatally blurred in Seeck's paraphrase: "Der Hauptmann der fremden Söldner führt Krieg gegen einem abgefallenen Theil seiner eigenen Genossen, und es regt sich der Verdacht, daß er mit dem Feinde im Einverständnis sei" (1894, 443).
The two leitmotifs of the poem are eunuch and slave. There are twenty-nine references to Eutropius's one-time servile status, some extended.[135] There can be no doubt that the "servile kingdom" means the East under the rule of Eutropius. In line 517, the Easterners are alleged to admit that they deserve punishment for "entrusting themselves to the governance of slaves" (qui se tradiderint famulis ). At 535 a suppliant Aurora, addressing Stilicho, refers to herself as "a plaything of slaves" (ludibrium famulis ). Usage and context put it beyond doubt that Stilicho is being asked to rescue the East from Eutropius because he is both unable and unwilling to stop Tribigild. It follows that the appeal was written before Eutropius's fall, when it could still be alleged that only Stilicho could stop Tribigild.[136]
Of course it did not happen this way. Gaïnas intervened instead, using the threat of Tribigild's superior strength to persuade Arcadius to depose Eutropius. Six months later he seized power himself. Claudian evidently had no idea of these developments when he wrote book 2 of In Eutropium .[137] Paschoud claims that Claudian's account is "not irreconcilable with the version of an early agreement between the two Gothic generals"[138] But the motive Claudian assigns for Tribigild's revolt is his own indignation at being treated worse than Alaric. It is true, as Paschoud observes, that Gaïnas had the same grievance, but the key fact is that Claudian applies it only to Tribigild. Moreover, Leo's defeat is attributed entirely to his own incompetence, which is used as another stick with which to beat Eutropius. Seeck argued that Claudian did not mention Gaïnas because he was Stilicho's secret agent in the East;[139] but it was for himself, not Stilicho, that Gaïnas seized power a few months later. And in any case it is not true that Gaïnas's part in the story is entirely suppressed. At 578f. Claudian describes how the Greuthungi lay waste Lydia and Asia, "relying not on their own valor or their numbers,"
sed inertia nutrit
proditioque ducum, quorum per crimina miles
captivis dat terga suis.
[135] 1.26, 30–44, 58–77, 83, 100–109, 122, 125, 142, 148–50, 176–77, 184–86, 212, 252, 276, 310–11, 478–81, 507–13; 2pr.3, 29–30, 62; 2.56, 69, 81, 132, 319, 351–53, 517, 535, 593. For the motif of the eunuch, see above, p. 135 n. 105.
[136] Cameron 1970a, 136f., unconvincingly disputed by Döpp 1978, 187f.; 1980, 161f.
[137] As observed long ago by Gibbon: "The conspiracy of Gainas and Tribigild, which is attested by the Greek historians, had not reached the ears of Claudian, who attributes the revolt of the Ostrogoth to his own martial spirit and the advice of his wife" (1909, 389 n. 27).
[138] Paschoud 1986, 125.
[139] Against this notion see Cameron 1970a, 148; Döpp 1980, 164.
But the treachery and feebleness of our leaders helps them;
it is through their crimes that our soldiers flee before their own slaves.
The feebleness obviously points to Leo, but the treachery can refer only to Gaïnas. Clearly for Claudian Gaïnas's treachery became an element in the story only after Leo's defeat. As even Zosimus's account makes clear (5.14.1–2), the two Roman generals divided their task; Leo marched into Phrygia, where Tribigild had been last reported, while Gaïnas stayed by the Hellespont, in case Tribigild marched north to cross into Europe. In the event he marched south, pursued by Leo, and Gaïnas did not march south until Leo had met with disaster. It cannot have been until then that the first suspicions arose of Gaïnas's collusion, prompted by his evident (and understandable) reluctance to engage the victorious Tribigild.
It is worth taking a closer look at Zosimus's account. At first sight, the sheer number of treacherous acts he details might seem to leave little room for doubt. In fact what we find is one after another of those all too familiar unfulfilled intentions—not surprisingly, seeing how little the early movements of Gaïnas and Tribigild suit the hypothesis of collusion. Instead of joining forces as soon as possible, Gaïnas waited by the Hellespont while Tribigild marched off in the opposite direction. At 5.14.3 Zosimus claims to know that Gaïnas began by ordering Tribigild to march north to the Hellespont, but Tribigild "was afraid of the troops stationed there" (Gaïnas's army?) and so marched south into Pisidia. If the plan had been carried through, "all Asia would have been taken." At 5.15.4 we are told once more that if Tribigild had marched east into Lydia instead of west into Pamphylia, all Ionia would have fallen, followed by the whole East as far as Egypt. Why then did he march west? The unfulfilled intentions even continue after the two Goths had allegedly joined forces: for example, he claims that they planned to take Sardis together but were foiled by unexpected spring rains (5.18.5).
The disproportionate detail in Zosimus's narrative of Tribigild's revolt derives from local information;[140] much of the action took place not far from the native city of his source, Eunapius of Sardis. Most of the incidents related are doubtless true enough, but the motives inevitably are all guesswork. We gain a lively insight into the fears and conjectures of people who suffered from "those aimless and destructive marches and countermarches,"[141] people who knew only that two Roman magistri
[140] 5.13–19 (nine pages in Paschoud's edition), "a copious and circumstantial narrative (which he might have reserved for more important events)," according to Gibbon, ed. Bury (1909), 7:387 n. 21.
[141] E. A. Thompson 1982, 42.
militum had failed to stop one rebel, and who suspected the worst. It was easy for such people to believe that even while far away by the Hellespont Gaïnas "secretly sent forces to assist Tribigild" (5.15.3) or sent "his barbarians" to corrupt and harass the various Roman units that were threatening Tribigild from all sides (5.17.1).[142]
Like the Tacitean Tiberius, even when Gaïnas does the right thing, it is from the wrong motive. For example, at one point Tribigild appeals to him in desperation; Gaïnas is distressed but, not wishing to reveal his hand, dispatches Leo against Tribigild (5.16.5)! In the event Tribigild inflicted a surprise defeat on Leo, allegedly with Gaïnas's secret help, but if Gaïnas had really wanted to protect Tribigild, he might better have sent Leo on some other errand and temporized.
Not only are all these secret acts and plans hard to believe; they would have been harder still to execute and involved Gaïnas himself in considerable risk.[143] Tribigild had only a very small force to begin with, and Gaïnas could not have counted on the revolt spreading in the way it did.[144] Nor was it a foregone conclusion that Eutropius would select Gaïnas, whom Zosimus admits he treated badly, to send against Tribigild. Furthermore, how far could Gaïnas trust xTribigild? Suppose that, having defeated Leo, Tribigild had turned on Gaïnas too? Suppose Alaric had intervened?[145] How could Gaïnas possibly have calculated on being able to manipulate to his own advantage so many imponderables?
Zosimus describes a meeting between Arcadius and Gaïnas in probably the fall of 399 at which "it was obvious to everyone that he was moving toward revolution." Once again, unfulfilled intentions, and Zosimus adds that Gaïnas himself behaved as though his intentions were still undetected (5.18.4). There is in fact no secure evidence that he ever joined forces with Tribigild. According to Zosimus (5.18.9), after reaching Chalcedon Gaïnas "ordered Tribigild to follow him." On the conspiracy theory, we should certainly expect to find Tribigild sharing in Gaïnas's triumph, but in fact Zosimus never mentions him again. Indeed, according to Philostorgius, who also drew on Eunapius, "after
[142] A claim incredibly taken quite seriously by Paschoud, comm. pp. 136–37.
[143] Ridley (1982) comments laconically on 5.15.3: "It is difficult to see how Gaïnas could have managed this."
[144] Zos. 5.13.4, 15.2. As Bellona assures Tribigild:
cunctaris adhuc, numerumque tuorum
respicis exiguumque manum? tu rumpe quietem;
bella dabunt socios.
(Claud. In Eutr . 2.220–22)
[145] See below, pp. 328–33.
suffering many losses, [Tribigild] escaped to the Hellespont and, crossing thence into Thrace, was killed soon afterwards."[146]
It was not till he marched into Thrace a few days after the massacre that Gaïnas first stepped outside the law. Up till then he had done nothing to warrant interference by Fravitta without instructions from Arcadius. To be sure he had exploited Tribigild's victories to his own advantage; he had prevaricated and bullied and blackmailed, but each time Arcadius had been intimidated and had acquiesced. It was Arcadius, not Gaïnas, who had dismissed first Eutropius and then Aurelian. When refused his Arian church within the walls, Gaïnas had meekly accepted defeat. And even when the people of Constantinople had risen up and massacred 7,000 of his fellow countrymen,[147] he did not take the immediate revenge the city expected.
It must have been an unbearably tense interval, through which Caesarius successfully stalled. To Synesius's highly tendentious representation of this period it is instructive to compare Claudian's account of Rufinus's behavior during Alaric's brief siege of Constantinople in 395:
Rufinus rejoices in the beleaguered city and exults in its misfortunes . . . . From time to time he laughs. He has only one regret: it is not his hand that strikes the blows. He watches the whole countryside ablaze by his own orders .[148] . . . He boasts that to him alone the enemy camp opens its gates and that he is allowed to parley with them. Whenever he goes forth to arrange some marvelous truce his companions throng around him, and . . . Rufinus himself in their midst drapes tawny skins about his breast.[149]
Since Alaric did leave, it is apparent that Rufinus arranged a satisfactory truce. Yet Claudian does not flinch from alleging that it was he who had arranged the siege in the first place. Synesius similarly makes Typhos's former allies testify "that Typhos had surrendered key positions and all but arranged the siege himself, so that the holy city might be gripped by a reign of terror" (123A). He is no more to be believed than Claudian, especially since this testimony is given at the wholly imaginary trial.
As to Caesarius's conduct when Constantinople was effectively "besieged"—that is, in the immediate aftermath of the massacre, when Gaïnas had withdrawn "a little distance away" (119C–D) from Constan-
[146] HE 11.8, p. 138.25–27 Bidez-Winkelmann; Eunap. frag. 75.7 (= 67.11 Blockley) may allude to Tribigild's death.
[147] For the figure 7,000, Zos. 5.19.4.
[148] praeceptis incensa suis, In Ruf . 1.71; for the text here, see Cameron 1968b, 392, accepted in the new text of J. B. Hall (1985).
[149] In Ruf . 2.61–85.
tinople, but before he retreated further into Thrace—Synesius charges that "he demanded that they negotiate with the barbarians, and was again working to admit the enemy army, claiming that no irreparable damage had been done" (121B). But the people stoutly and righteously resisted him, Synesius continues, "and in general his tyranny was as good as dead, since the force that sustained it had been driven from the city." Thus he implies that Caesarius's excuses were presented to the people of Constantinople—to whom indeed they might sound pretty feeble. But the claim "that no irreparable damage had been done" makes much better sense as palliation extended to Gaïnas. It was his fellow Goths who suffered what damage had been done, and he who, with an army encamped not far away, might be expected to seek revenge. Synesius himself, writing book 1 at just that time, took great pains to transfer all responsibility for his hero's fall from the Scythian general to Typhos: clearly he too feared to offend Gaïnas. It was Caesarius's responsibility as PPO to conciliate Gaïnas. He will have wanted to "readmit the enemy army" to Roman service, to prevent it from attacking the city as an enemy in deed. Naturally he had to minimize the effects of the riot and offer to repair what he could.[150]
In the event, these negotiations came to nothing. Gaïnas was not reconciled, but neither did he attack Constantinople. Even before the extension and reinforcement of the walls under Theodosius II, it would not have been easy for him to take the city by force. Alaric did not even try.[151] Nor is it likely that Gaïnas could have counted on the loyalty of his men if he had made the attempt. What finally drove him to turn his back on all his hopes and march off into Thrace was perhaps the news that Fravitta was on his way. It was not till then, surely, that Arcadius finally dared to pronounce him a public enemy.
It is significant that Synesius's narrative leaps straight from the massacre of July to Aurelian's return in September.[152] To be sure, he was writing before Fravitta's victory and Gaïnas's death at the hand of Uldin the Hun. But the campaign must have been under way by the time Aurelian
[150] It may be significant that Synesius mentions Typhos's negotiations immediately after his "indignant protest" over the burning of the church and "Scythianizing religious beliefs." Caesarius could hardly have conceded anything to Gothic Arianism (cf. below, p. 328), but Synesius is only concerned to represent his sympathy in the most unattractive light possible. It would not be unreasonable if his offers to repair damage had included something like a new Gothic Catholic church.
[151] Nor did the Goths who attacked the city in 378 immediately after their defeat of Valens at Adrianople: Amm. Marc. 31.16.4–7.
[152] On the date, see below, section III.
returned. He makes no attempt to assign Aurelian any credit even for his own return. The technique is reminiscent of Claudian. Naturally Synesius did not want to report that it was another Goth who had rescued the situation. Least of all did he want to allow Typhos any credit. But in all plausibility he could not allow Aurelian any either. So he said nothing. Aurelian's return is simply recorded, without explanation.
Yet it must have been Caesarius who claimed the political credit for suppressing the rebellion—and for bringing back the exiles. However much Synesius might muddy the waters with talk of treachery and malice, the fact remains that it was Aurelian's policies that provoked Gaïnas's revolt and Caesarius's that brought it to a rapid and successful conclusion. It was not Aurelian, but the brother Synesius caricatured as a barbarian-lover who sent the barbarian packing. The man who allows his opponent to play his only card does not deserve to get back into the game. It is not surprising that Caesarius remained in office for another two and a half years.