Scholars often assume that historians in late antiquity ended their narratives with the previous regime because it was too dangerous for them to write candidly about the current imperial reign. While this was generally true, there are some notable exceptions to the pattern which, when studied together, show that this rule was not iron-clad and cannot safely be relied on to date the composition of texts and the scope of their coverage (in the case of lost works). Both the rule and the exceptions illuminate the nuanced play of “truth” and “power” in late antiquity. In assessing our evidence we must often read between the lines, but not all historians were as timid as we might assume. The Roman empire's “secret history” could sometimes be aired.
“We write what is untrue to our disgrace, and what is true at our peril.”1sidonius apollinaris on writing history (475/6 C.E.).
It is generally and correctly assumed that late antique historians—our most important narrative sources for their age—were writing in an unsafe political environment that hindered them from revealing compromising truths about the current regime or engaging in overt criticism of it, at least not without suffering severe repercussions. Prokopios’ Secret History (550/1) is the poster-child for this impression.
I could not at that time [when he published the Wars] give a candid report concerning certain events so long as the people who were responsible for them were still alive. It would have been impossible either to evade detection by the legions of their spies (πλήθη κατασκόπων) or, having been caught, not to suffer a most cruel death. I could not even trust my closest relatives. Moreover, I was forced to conceal the causes (ἀποκρύψασθαι τὰς αἰτίας ἠναγκάσθην) of many of the events that I narrated in the earlier books. It is therefore incumbent on me here to reveal what had previously remained concealed as well as to disclose the causes of those events that I did report there.2
In terms of methodology, the existence of these dangers has led scholars to assume that historians could not speak freely about the current emperor and, as a rule, that they ended their works with the previous reign. This axiom is used not only to determine where a lost or partially preserved work ended—i.e., with the end of the previous reign—but also to date the publication of any historical work after the end of any regime which it criticized.3 For example, “it is not easy to believe that Eunapius published his extraordinarily hostile account of Theodosius in the emperor's lifetime… it would be unparalleled for any historian to continue a full-scale narrative right up to the moment of publication.”4 Some scholars go further and assume that a historian could not have published his history under a certain regime if the work was not sufficiently panegyrical when it came to that emperor (I discuss the case of Herodian below). This method of dating, premised on assumptions about what could or could not have been said, or what had to be said, shapes our interpretation of texts, their author's politics, and the climate of opinion under which they wrote. These assumptions are brought out to support specific conclusions (about dates, e.g.) but they themselves remain fuzzy. It is unclear, for example, if it was sufficient for the historian to refrain from criticizing a reigning emperor, or if one was required to be panegyrical; whether previous members of the same dynasty were also off-limits, or only the reigning ruler; or if such rules were absolute, or subject to exceptions.
The present study argues that these methodological axioms are not always reliable. Testing them against the evidence, we find that they need to be moderated. Prima facie evidence for caution is provided by none other than Prokopios: he wrote precisely “a full-scale narrative right up to the moment of publication” (the Wars) that is neutral-to-critical when it comes to Justinian, and his vivid description of the dangers facing historians in the preface of the Secret History comes from a polemic that he wrote against that emperor and circulated in some form. Eunapios too may have done exactly what is alleged above to have been impossible. We should soften our rules: what historians felt that they could say safely varied by circumstance, and some were willing to take greater risks than is often supposed. The brave in this tale, it turns out, were the old—by ancient standards, the really old.
The present study is the first part of a triptych, whose other leaves present separately the impact of the religious transformations of late antiquity on the perils of authorship, and, finally, the dangers of rhetoric. Here we look at historians, and specifically political historians. The first part lays out their basic legal and sociopolitical context, outlines the circumstances that made their work alternately profitable and risky, and explores what consequences the historian might face for overstepping certain (always ill-defined) bounds. The second part discusses individual authors who seem to have flouted our conventions about what could or could not have been written, or to whom we have applied the conventions too rigidly. I focus on Greek and Latin authors from the Severan period to the seventh century C.E. The topic, which in modern thought falls under the rubric of “freedom of speech,” is not without broader significance. The later empire—or “Dominate,” as previous historians have characteristically called it—has been viewed as fundamentally unfree, governed by an oppressive, intolerant, militaristic, and wrathful imperial order. Many scholars might now dispute that description and call its government responsive rather than oppressive. Yet Roman history is still frequently figured in terms of declining freedom, from the libertas of the Republic to Byzantine theocracy,5 though this alleged decline has never been investigated systematically in relation to the perils of authorship.
The later empire, it turns out, was not more dangerous for historians than, say, the Julio-Claudian or Flavian periods, when many authors were accused of treason.6 This was not because the later emperors were more permissive, though they were perhaps more secure: the storm of trials in the early empire reflected the insecurity of a newly established monarchy that did not yet know where it stood in relation to the rest of elite society. By contrast, only one later historian (Pelagios) was executed, and even his case is highly ambiguous. But this does not make for a more rosy late antiquity: historians had simply adjusted to a perennially dangerous imperial order, learning to navigate its ideological minefields. Thus, whereas the case-studies presented below argue for greater flexibility in the “rules” that we apply, historians still took protective precautions, even if some of them were prepared to go further than others. They shielded themselves through discriminating self-censorship. This, however, transfers part of their burden onto us, as they sometimes presented their politics in oblique ways to mitigate the peril of their circumstances. The argument for flexibility therefore rebounds onto us, their readers, since we must be more sensitive to what historians carefully chose not to say and to what they said indirectly. Their peril creates wider interpretive pitfalls.
THE FRAMEWORK OF PERIL: LEGAL, SOCIAL, LITERARY
Freedom of speech was not legally protected as such in the empire, yet while powerful men expected deference from their lessers, brave denunciations of injustice and truth-telling in the face of power (parrhesia) were valorized in Greek and Roman culture.7 The Theodosian Code contains an entire chapter (9.34) that bans anonymous documents defaming private individuals and scurrilous “secret histories.” The theory was that a citizen who had damaging information should make a formal accusation and take responsibility for his words. Emperors, however, generally expected to receive “unanimous” praise from subjects, especially from writers who had the power to move elite opinion, and they were watchful against plots, discontent, and dissent. The law behind many of the prosecutions of senators and writers in the early empire was the Julian Law on Treason (maiestas).8 Most of its surviving provisions are about treasonous actions and say nothing about authorship or criticisms of the emperor. From the many references in Ammianus to treason trials involving courts and judges, it seems that this law lay behind the fourth-century prosecutions too.9 In 393, Theodosius I issued a remarkable law stipulating that men who had cursed the emperor or maligned the regime should not be immediately punished if their actions stemmed from levity, insanity, or malice, which deserved tolerance, pity, and forgiveness (respectively). The facts of the case were to be reported to the emperor, who would decide whether punishment was warranted.10
There is some evidence that imperial agents sought out seditious elements. From the early empire comes the famous passage in Epiktetos about the secret agent who accosts you in civilian clothes and wins your confidence by complaining about the regime, encouraging you to incriminate yourself. Epiktetos was likely referring to the frummentarii.11 The agentes in rebus, who replaced the frummentarii in ca. 300, also acted as informants and spies, especially against imperial officials, though they performed other functions too.12 Ammianus explains how agentes would infiltrate dinner parties and report on potentially seditious talk, including “disgust at the existing regime,” prophesies, and hopes for change.13 Libanios took rhetorical revenge on these informants by exposing the protection-rackets they ran under Constantius II.14 The materials that these men investigated sometimes cut close to historians’ concerns. Ammianus reports that during the trial of the ex-governor Hymetius for embezzlement and nefarious sacrifices, Hymetius’ papers were seized from his house, including a document in which he criticized the greed and cruelty of Valentinian I (364–375), traits which Ammianus too would highlight in his account of the reign. The emperor wanted Hymetius executed, but referred the matter to the Senate, who exiled him to Dalmatia, angering Valentinian. Ammianus says that such incidents “made men fear that what was happening in a few cases would be the fate of all.”15 The fate of Hymetius, along with that of others who were caught up in Valens’ persecutions of alleged magicians in the east, caused “whole libraries to be burnt by their owners for fear of a similar fate. We all crept about at that time in a Cimmerian darkness … with swords hung over [our] heads suspended by horse-hairs.”16 Their own written words could be used against them.
Our most detailed sources, Ammianus and Prokopios, give the impression that these dangers were typical under Constantius II, Valens, and Justinian, but they were likely endemic to imperial society. In the early fourth century, a certain North African deacon Felix was summoned to court for writing a seditious letter against the emperor (possibly Maxentius, 306–312).17 A fragment of Eunapios’ History refers to “the prying ears of Eutropios,” a high minister of Arcadius (395–408) from whom all shrank in fear.18 So people did feel watched and could be formally prosecuted, and historians were often closely connected to the court. Yet we know of only one late antique historian who was executed possibly because of his history, the silentiarius Pelagios. He was hanged by the emperor Zeno (474–491), ostensibly because he was a pagan but in reality (or so our sources say) because the emperor was told that his successor would be a silentiarius, or because Pelagios had criticized him. He was the author, among other poems, of a verse history of Rome after Augustus, though it is not clear that his fatal criticisms of Zeno were contained there.19
Historians belonged to the social elite, but with only few exceptions they were not “super-elite” types, i.e., emperors, powerful generals, or the highest rank of senators. Rather, they were secretaries, lawyers, teachers, and mid-level officers and officials.20 They too had to tread carefully around controversial issues and easily-offended personages. It is important to consider why they wrote history at all, given the apparent danger. History was primarily a private literary pursuit, engaged in by individuals for personal reasons. It was not produced by institutions in the form of “official” versions, though it could reflect the outlook of powerful men whom an author wished to please.21 Otherwise, it was a sophisticated and prestigious pastime of cultured men that could serve various moral, political, intellectual, and personal aims simultaneously. Among them were careerist motives, as some sought the patronage of the court through works of history. In these cases, the present reign or dynasty had to be treated in a panegyrical way and, therefore, its beginning created a rupture in historiographical time: the standards of critical history could be applied to most events before it, but current power-holders were praised, if they were covered at all. These historians thus had to switch genres at the end of their narratives. This practice worked for emperors too. To appear cultivated and supportive of elite culture, they wanted histories to be dedicated to them, and presumably wanted those works to meet high critical standards, as their names would be associated with them. Besides, criticism of past emperors made them look good by contrast.
For example, the imperial secretary Eutropius dedicated his Breviarium of Roman history to Valens (364–378) in 369, apparently prompted by the emperor himself, and was soon afterward attested as governor of Asia. The flattering dedication seems to have paid off in terms of career advancement. The work ends with the death of the previous emperor, Jovian, in 364, with a note that in order to continue it would have to switch registers to “a more elevated style”—in other words, to panegyric. Eutropius avoided going there, but the polite hint was apparently enough.22 Menandros the Protector dedicated his history to Maurikios (582–602), his patron. He praises and thanks the emperor in his preface, but ends the work in 582, the year of his accession. Still, wherever Maurikios appears within it as a general, before his accession, he is praised panegyrically, even if in brief bursts.23 We consider other examples of this phenomenon below. It also seems that some histories were recited at literary gatherings. This is believed for Ammianus (based on a letter probably addressed to him by Libanios),24 and is certain for the History of Theophylaktos (640s). The latter received some kind of public performance in Constantinople and is addressed to the patriarch Sergios, who is praised in the preface.25 These authors had to ensure that their works were safe, not bland like modern textbooks, but they had to be discreet, tactful, and allusive regarding sensitive issues.
The shift to a panegyrical mode made some historians uncomfortable, however. Many insisted that panegyric (and invective) were separate genres, whereas the duty of the historian was to record the truth impartially. The tenacity with which they repeated this banal claim is striking, suggesting that it was a serious matter for them.26 Sensitive to the distinction between the genres, historians policed the boundaries and integrity of their work and asserted its intellectual independence from genres that catered to political interests, including the praise that emperors required for themselves and the blame that had to be cast upon their rivals or predecessors. The historian's need to establish independence sometimes went so far that Tacitus had to rebuke excessive criticism of past emperors, which was presumably done ostentatiously by some to establish their objectivity.27 Agathias of Myrina, who wrote a history of the years 552–559 C.E., asserts the distinction vigorously in his preface. Writing twenty years after the events he records, he differentiates between history and panegyric, protesting that too many so-called historians of his times approach their subjects “with flagrant disregard for the truth” because they are “intent on flattering and fawning upon many influential people,” especially emperors. He calls them charlatans because they label their works “histories” but “eulogize living men during their lifetimes.” Agathias claims that he will write an impartial history, but adds also that he will be writing primarily about the dead.28 This way he could be both candid and safe, though at the cost of avoiding recent history.
Therefore, the horns of the dilemma facing imperial historians were, first, panegyric, which elicited potential censure from peers in the future, and, second, the risk of offending the powerful. The main strategy by which they avoided this dilemma was by writing only about previous emperors, often by ending their narratives at the end of the last reign. This was a form of partial self-censorship: silence-or-panegyric (as opposed to freedom-or-death). Thus, what is generally missing from late antique historiography is “live” coverage of imperial history and especially critical accounts of the present reign. I consider interesting exceptions to this pattern below, but they must be appreciated against the general rule. The photonegative image of the perils of history was silence (but a silence, perhaps, about which we can draw conclusions).
The work of Aurelius Victor offers an interesting twist on the familiar dynamic and is worth presenting here as typical of the careerist type of historian. Victor was a senator who, around 360, wrote an epitome of imperial history in Latin, with a paragraph devoted to each emperor down to Constantius II (337–361). He wrote it toward the end of that reign and so presented Constantius panegyrically, hiding his many faults. In 360, however, a war broke out between Constantius and his cousin Julian. On his way to confront Constantius, almost immediately after Victor had finished writing his epitome, Julian met Victor and made him governor of Pannonia II.29 Julian was apparently untroubled by the praise that Victor had just publicly lavished on his enemy—more proof, if we needed any, that panegyric of living emperors was understood to be pro forma and did not entail political convictions or candor. Moreover, the epitome as we have it ends with the following tagged-on indictment of Constantius’ regime:
These qualities, so great and renowned, have been tarnished by his inadequate attention in approving provincial governors and military commanders and at the same time by the incongruous character of most of his officials and, furthermore, by his neglect of all the best sort of men. And, to end on a brief note of truth (verum), just as no one is more outstanding than the emperor himself, so nothing is more frightful than the majority of his subordinates.30
These closing remarks have plausibly been seen as an addition made after the historian joined Julian, or even after the death of Constantius later that year.31 We see here the kinds of truths that Victor believed that he could not reveal about the regime of Constantius while he was still in power; they were kept in storage and brought out only after the regime changed. It is significant that Victor signals his intention to speak the truth (verum) precisely when he discards his panegyrical façade. In other words, the panegyrical mutations of historical works, the points at which they brush up against the present, are revealed as concessions to the facts of power and not as integral parts of the history. They are dispensable. But it is not clear that, even in this hostile addendum, Victor had worked free of the constraints of power: he still defends Constantius’ character. This may have been a concession to dynastic continuity, and welcomed by Julian, who broadcast the view that his cousin had been misled by wicked advisors. Victor simply switched from Constantius’ view of Constantius to Julian's view of Constantius. He was in the shadow of a dynasty. Note that Victor says nothing about how Constantine the Great had executed his son Crispus and wife Fausta. Less than a decade later, however, and under a different dynasty, the epitomator Eutropius could mention those executions.32
This section has explored that ambiguous space between discrete silence and panegyric that imperial historians entered when they reached the present. It represents the “rule” that scholars typically apply in studying their works. But the discussion above rests on a hidden premise that we must now draw out: the nexus of silence or/and panegyric is most emphatic in careerist historians who were addressing emperors and hoping for patronage. The rule is therefore a rule only if we take a socio-determinist approach to historiography. But other authors may have written for intellectual, literary, or political-theoretical reasons. I now turn to assess how free they were when they were not actively seeking favors from the regime.
NON-PANEGYRICAL, CONTEMPORARY HISTORY
Some authors appear to deviate from the general pattern of silence-or-panegyric, and these require us to adjust our methodological axioms. I begin with three later and more striking cases (Eunapios, Jordanes, and Prokopios) and then, in light of their experience, reconsider three earlier ones (Herodian, Cassius Dio, and Ammianus).
Eunapios of Sardeis was a pagan intellectual trained in medicine and rhetoric who also had an interest in philosophy. His History, which is mostly lost, reached, in its final form, from 270 to the early fifth century and idolized the emperor Julian (360–363).33 Its panegyrical aspect would, therefore, have been untimely and obnoxious to the Christians ruling the empire in Eunapios’ time. A fragment of the work, preserved in the tenth-century Byzantine Excerpta (albeit out of context and obscure in its wording), discusses the courage required to be a truthful historian.
The written record is incredible (ἀπίθανον). But if anyone else is able to write these things down, I admire his bravery and judge him a brave man for his patient endurance. It was proper that those who both made their accounts as accurate as they could, considering the times when and the persons about whom they were writing, and proposed to write in safety (ἀσφαλῶς), should incline their narrative to favor and disfavor (χάριν καὶ ἀπέχθειαν). However, as I record these events, my path has led not in that direction, but rather to approach and stay as close as possible to the truth.34
In other words, authors who write “safely” inevitably end up writing panegyric or invective, presumably because they must curry favor with powerful interests. By contrast, Eunapios believes it is an act of “courage” to tell the truth, especially about contemporary events, and includes himself among truth-telling historians.
On one level, such declarations established a historian as morally superior to his peers, even if in reality he was not taking a particularly serious risk. But in this case there was more than just rhetoric at work. Eunapios was highly critical of the Theodosian dynasty and was openly anti-Christian in an age of increasing Christian violence.35 We cannot know if he was ever in danger because we do not know the circumstances in which his work was released. So much attention has been paid by scholars to his paganism that we sometimes overlook his politics. The History apparently ended in 404 and so was finished and published during the reign of Arcadius (395–408) or early in the reign of his son Theodosius II (408–450). Eunapios was too old to have been writing much later. The History was read and attacked by Christian authors in the first half of the fifth century, so it was not a “secret” work, at least eventually. Eunapios was hostile to the dynasty's religion and its founder, Theodosius I; to many of Arcadius’ officials (though they had meanwhile fallen from power); and also to Arcadius’ empress Eudoxia, who died in 404. Eunapios presents her as corrupt and arrogant, and the period generally as one of failure and decline.36 As he planned to organize his work by reign,37 404 is a strange year at which to end, suggesting that he was possibly writing up to the moment, therefore under Arcadius. Perhaps, as we will see with Prokopios, Eunapios was waiting for the empress to die. Like Theodora, Eudoxia was known to be more vindictive than her husband. Eunapios knew the fate of the patriarch John Chrysostom, whose deposition the empress demanded after he criticized her publically.38
It is also possible that Eunapios published the work under Theodosius II (408–450), Eudoxia's son, albeit a child-emperor for years. But even in this case, his risk was only slightly smaller: he was criticizing the dynasty and the emperor's mother. In light of the “rules” discussed above, such a History should have been impossible, or foolhardy. In interpreting the decision of Eunapios, we could speculate that he had inside information that the regency of Theodosius II would not move against dissidents, or that he was taunting the regime. But a different scenario is more likely: Eunapios did not have a political career, was not seeking any favors from the dynasty, and was interested only in promoting his own religious-historical ideology among similarly alienated pagan thinkers. He was not linked to anyone whom the regime might regard as a credible threat and could expect to be disregarded. He may have been both more brave, and more distant and insulated from the politics of the court. Moreover, in 410, Eunapios would have been about sixty-five years old.
Turning to the mid-sixth century, Jordanes wrote two interrelated Latin chronicles that reached from their origins to the moment he was writing, in 551, likely in Constantinople.39 Unfortunately, due to the topical nature of his work, he has not attracted much attention outside the specialized subfield of early Germanic history. But Jordanes intended his two works to be read together. The first is the Getica, a sweeping if condensed narrative of Gothic history from its mythical origins in Scandinavia down to the contemporary and ongoing war of Justinian against the Ostrogoths in Italy. The second, the Romana, is a breviary of Roman history, most of it derivative but not the final part, which also reaches down to the present. Jordanes is therefore among the few imperial historians who wrote up to the minute, and is one of two who did so in two simultaneous, interlinked works that deliver different verdicts on the reign (the other is Prokopios, writing in the same year and same city; both men, moreover, were former secretaries to Roman generals).
Jordanes reveals at the end of the Getica that he was of Gothic descent. He was, therefore, in the curious position of recounting the history of a barbarian people—“his” people even, at least in part—while living in the capital of an emperor bent on the destruction of their kingdom in Italy. He asserts there that he wrote the Getica in order to praise not the Goths but “him who conquered them,” i.e., Justinian, a flattering if premature claim in 551. Jordanes avoids mentioning the ongoing Gothic resurgence and treats the war as finished by Belisarios in 540. This suggests that he was toeing the official line. Many interpreters of the Getica have accordingly read it as Justinianic propaganda.40 But this may not be the case, as Brian Swain argues.41 Jordanes’ conclusion is ostensibly panegyrical because the Getica does in fact praise the Goths throughout and does not say anything that redounds to Justinian's glory (unless we assume that his future victory will be the greater for having conquered so noble a foe). Moreover, just as interpreters have privileged the last words of the Getica over the rest of its contents, so too have they privileged the Getica over its sister-text, the Romana. Jordanes was a Roman too, after all, despite his Gothic origin. The Romana reaches down to the present as well, but its end is dramatically pessimistic. Alongside the victories, Jordanes does not hide the defeats, revolts, and domestic troubles of Justinian's empire, including the resurgence of the Goths in Italy that is absent from the Getica. There is no panegyric of Justinian in the Romana, which ends as follows:
Such was the fall of the Roman res publica, apart from the daily invasions by Bulgars, Antae, and Slavini. If someone wishes to know them, even a cursory glance at the annals of the consuls will show that the res publica of our time (nostri temporis) is worthy of a tragedy. He will understand where it came from, how it grew, how it subjected all lands to itself, and then how it lost them through ignorant rulers. All this we too have covered briefly.42
This is a scathing indictment of the current state of the empire, casting Justinian's reign as the nadir of a “tragedy.” Justinian is not presented as a potential savior but as one of the “ignorant rulers” who has brought Rome to its knees. This was not a “small… distasteful truth” uttered by an (alleged) partisan of Justinian.43 It is the critical climax and negative verdict of the work.
Unfortunately, we know nothing about the release and dissemination of the Romana. The individuals whom Jordanes addresses in the prefaces of his two works are unknown (if they were real). For whatever reason, Jordanes felt safe to express a pessimistic verdict at the end of the Romana. The reign of Justinian was rife with dissident groups and heated debates, especially in the religious sphere, which may have provided critical authors with receptive audiences.44 Jordanes, however, does not appear to be linked to any domestic opposition that worried Justinian, and he wrote his histories in “retirement,” possibly after taking up a religious vocation. He too was old, possibly over seventy.45
We come now to Prokopios of Kaisareia, to whom we owe the most explicit statement by any ancient historian (quoted at the start of this study) on how power constrained the writing of history. This statement was generated by the intersection of two novelties that mark Prokopios’ work as exceptional: his decision to write primarily contemporary history and to do so in two works, one a secret invective supplement to the more “neutral” public one. In the preface of the Secret History, Prokopios mentions the regime's domestic “legions of spies (κατάσκοποι),” the same word that he uses for agents sent to spy on Persia.46 Here he is probably using the word in loose sense for any “informers,” including the historian's own relatives, who might turn him in. This points to Theodora as the chief danger, for later in the Secret History she is depicted as suborning relatives and prying into domestic affairs. He uses language that echoes the preface: “It was impossible for anyone who opposed her to escape detection. An army of spies (πλῆθος κατασκόπων) informed her of everything that was both said and done in public and within private households.”47 It seems, then, that Theodora took a greater interest in monitoring opinion than Justinian. This is suggested by Prokopios’ claim in the preface that the people he most feared were now dead. In 550/1, when he finished the Secret History,48 that applied only to Theodora, who died in 548. Justinian and Belisarios were still alive, as were many other officials maligned in the work.
It would be impossible to resolve here the problems posed by the parallel production of the Wars and Secret History in 550/1, an insignificant year for both politics and war. A likely scenario is the following.49 Possibly as early as the mid 530s Prokopios, a secretary of sorts to the leading general Belisarios, decided to write a history of contemporary events, and he had made considerable progress toward it by the mid 540s. By then, he hated the regime and was alienated from his former employer, the general Belisarios. He surely would have preferred to write a single narrative that integrated the good and the bad but was terrified, especially of Theodora. Her death in 548 opened up opportunities. Justinian himself was almost seventy and had no sons; everyone, therefore, would have been thinking of the succession. Theodora's death allowed the star of Germanos, Justinian's cousin and a victim of her machinations, to rise. Prokopios presents Germanos favorably, suggesting that he may have supported him, and the Wars-plus-Secret History would have been enabled by—and possibly supportive of—his succession and the ensuing change of policies. Germanos, however, died suddenly in 550. With no end to Justinian's reign in sight, Prokopios had to move forward on a dual track with two separate but interlinked works, though he did integrate some of the material from his Secret History file into the Wars (especially scandals concerning the disgraced prefect John the Cappadocian). The two works as we have them, therefore, were not the product of a faction that supported Germanos—or any other faction for that matter—but the result of Prokopios’ own decisions about how to present the reign to posterity. He was not burdened by links to any subversive group, and was free of the watchfulness of Theodora.
The Wars thus reflects what Prokopios believed he could get away with under Justinian and his remaining courtiers, in the absence of Theodora's spies. Therefore, it is significant that, while the Wars always refers to the emperor in a respectful way, it does not support his propaganda and contains no panegyric;50 it is critical of many of Justinian's appointments, officials, and policies; it breaks with many of his pronouncements on matters of Roman history and imperial ideology;51 and offers damning allusions and insinuations. In sum, the Wars is as critical of Justinian and his high officials as the author of the Secret History could make it, and he opens various “conduits” from one text to the other. A few years after its release, Prokopios published a supplement and boasted of the success of the first edition.52 There is nothing like this in previous imperial historiography, or anything like the pair of works Wars and Secret History. The problem of this uniqueness is compounded by the existence, in two versions (an original and an expansion), of the Buildings, a work by the same author which is panegyrical of Justinian but also contains some negative insinuations that point to the criticisms of the Secret History.53
For all his uniqueness, Prokopios still leaves us with a usable standard against which to evaluate other historians, even though it too is derived from a specific moment in time. It reveals that a semi-critical, up-to-the moment work such as the Wars could safely be published, even if only after Theodora's death and under the cover of a simultaneous proclamation of loyalty in the Buildings, whereas the inflammatory material in the Secret History could be circulated only among a small circle.54 The stance of Justinian himself is more difficult to evaluate. On the one hand, Honoré has observed that Justinian's learning and intellectual preoccupations were bureaucratic, legal, and theological, with “no evidence of classical culture.”55 His regime was preoccupied with wars, religious dissidence, plague, and declining revenues. In his many purges and persecutions, he never seems to have targeted literature. Authors were punished (mostly bishops) not because of what they had written but because they opposed imperial policies in concrete ways. Justinian, moreover, was less concerned with political than ideological opposition:56 he pardoned men who had tried to kill him, along with kings who opposed him in war, but was harsh on bishops who disagreed with him on obtuse points. This, perhaps, allowed the Wars to slip past. There is no evidence that Justinian or his top courtiers knew of its existence. On the other hand, Justinian had a strong view of his place in history,57 which he expounded with detailed references to previous Roman history in the prefaces to his laws,58 and he issued public proclamations recounting and framing important events as they occurred.59 He also personally commissioned a classical history of his war with Persia, though that was twenty years before the publication of Prokopios’ Wars.60 It boggles the mind to imagine how Justinian might have read the Wars, but somehow his regime either allowed it or ignored it. It was not dedicated to the emperor, nor was Prokopios seeking favors based on it. He was, moreover, over fifty years old when it appeared, or possibly much older.
Let us now take what we have learned from the aforementioned authors and apply it to historians whom scholars have rigidly subjected to the standard rules of silence-or-panegyric. Herodian was the third-century author of A History of Events after Marcus Aurelius, ending with the accession of Gordian III in 238. He gives away nothing about himself but his name, yet scholars agree that Herodian could not have published his work under Gordian III (238–244) because he does not give a panegyrical account of the prior Gordians, I and II (the grandfather and uncle of III), who were emperors for a month in 238; does not seem to have liked child-emperors in general, which can be read as an indirect critique of Gordian III, who was thirteen in 238; and records Gordian III's accession flatly, with no panegyrical flourish. His last words are: “Gordian, aged thirteen, was acclaimed emperor and took over the Roman empire.”61 C.R. Whittaker comments: “This somber conclusion… can hardly have been written during the reign.”62 In short, the work was not panegyrical enough to be contemporary, even though it was not overtly hostile.
One wonders whether a (probably) provincial author, otherwise unknown, would run into trouble merely for recording the current emperor's accession in such a neutral way, and what kind of trouble that might be. The persona of the author, as it emerges from the text, is that of an educated cipher: he literally has no political or biographical presence, and this might have been a protective device. There was, at any rate, no question of offending the emperor, who was too young, but rather his regency. Herodian may have highlighted the problem of child-emperors not in order to criticize Gordian but to draw his handlers’ attention to what needed to be done. His striking emphasis on the need for imperial paideia may have been offered as a moral guide. There would have been no reason to go after the author of this text. And the fact that he ends his work with the very start of Gordian's reign suggests that he was indeed writing under him, because if he was writing after him he could have continued the narrative down to 244. In sum, in this case we have perhaps been too quick to pull the trigger and enforce the “rules.”
Cassius Dio was a prominent senator and consular of Greek origin who, in the early third century, wrote a long Roman History in Greek that survives only in excerpts and fragments. Dio had previously written a pamphlet on the omens and dreams that predicted the rise of Septimius Severus (193–211), which he sent to Severus. The latter was pleased and responded with a complimentary letter. This led Dio to write a history of the wars that brought Severus to power, which also pleased the emperor. This history, born of the politics of imperial patronage, was eventually expanded into the Roman History.63 There is debate over when the various editions of this work were finished after that point,64 but the first edition covered events down to 211 and so it conformed to the pattern by not discussing the current reign (whichever that was). Dio is more critical of Severus in the Roman History than he was, presumably, in the works that he published during Severus’ reign.65 With hindsight, he blamed Severus for much that had since gone wrong. The emperor's death, then, offered the opportunity for a candid, revised account. Dio, moreover, retroactively exposed the pressure to dissimulate. During any given reign, senators were obliged to publicly uphold a more favorable view of the emperor than the one they held privately. Many of them (himself included) had acclaimed Commodus (180–192) without conviction, even with shame,66 while others only pretended to side with Severus, sometimes with such theatricality that they gave themselves away.67 Dio did not disguise the constraints on the free expression of opinion and he shreds the sincerity of imperial panegyric.
The final books of Dio's history have two interesting features from our standpoint. First, they offer a critical narrative of the Severans down to 229, but were certainly written under that dynasty.68 Second, Dio continued to update and revise the work down to 229, the year he was consul with the emperor Severus Alexander (222–235). He then retired to his native Bithynia for the rest of his life, where he expanded the work down to 229. But for the years 222–229 he says that he will give only notes, because he was not resident in Rome.69 In the fragments that survive of these “notes,” Dio says little directly about the emperor, though he presents a troubling picture of unrest which he sought to flee by returning to Bithynia. It is unclear when he published these final books, and it is here that assumptions about what could not have been said come into play. One school of thought has it that the history is so critical of the dynasty, and its portrait of the emperor so insufficiently panegyrical, that “it would have been politically dangerous… to publish during his lifetime. Even in his retirement… in Asia Minor, he would not have been safe.” Instead, Dio must have revised the work in ca. 230 and made arrangements for its publication only after his death.70 But not everyone believes that the work was later revised, or even that in its final state it was fatally critical of the Severans.71 In other words, we are on subjective grounds here for dating, reconstruction, and interpretation.
There is no proof to resolve the dilemma either way. But let us imagine a (perfectly factual) scenario in which it was possible to publish the Roman History under Alexander. Dio was old—about seventy—and retired. He had personally alienated the armies that he had recently been posted to command, so he did not pose a threat as a rebel general.72 In fact, one might argue that his retirement was forced on him by the praetorians and that his career was over.73 Alexander was a young emperor (fourteen at accession), surrounded by senatorial advisors who may not have regarded Dio's history as threatening (they may even have agreed with it). The bitter memoirs of a retired elderly senator, who was apparently still on good terms with the emperor, might not have elicited repression. Dio would have been better able to gauge this than we are. Moreover, another pattern may be emerging: three of the critical histories we have considered (of Dio, Herodian, and Eunapios) may have been published during the reigns of young emperors, whose senatorial handlers may have been less interested in literary repression, and probably did not want to leave a trail of murdered writers for their charges to inherit. Such reigns may have constituted a known window through which to slip a less-than-panegyrical work. During his career as a historian, therefore, Dio seems to have moved from the model of Eutropius to that of Prokopios, from panegyrical to somewhat critical (albeit always in full view of the regime), yet not perceived as threatening, especially given his prior loyal service. The decisions of historians about what they could or could not say were formed subjectively and in response to a unique set of circumstances.
We come finally to Ammianus Marcellinus, a self-described Greek and former soldier, writing in Rome in his retirement. His Latin Res gestae, whose first part is lost, reaches to 378, i.e., the aftermath of the battle of Adrianople. The work seems to have been finished between 390 and 391 (when the Serapeion in Alexandria was destroyed, which Ammianus describes as still standing).74 This is the consensus view regarding its date of completion, and it is reinforced by a number of forward-references by Ammianus to events in the late 380s and 390, and to none after that. Some scholars defend a later date in the 390s, but I follow the dominant view here (and discuss one alternative argument below).75 The senior emperor in 390–391 was Theodosius I (379–395), who was in Italy in 388–391 after defeating his western rival Magnus Maximus (383–388). Yet there had been another child-emperor in the west all along, Valentinian II (375–392), the son of Valentinian I (364–375). While Ammianus was writing his history, he reigned (in theory) from Milan and then Vienne, in Gaul. Theodosius supported Valentinian II, treating him as a junior emperor, and married his sister, Galla.
These relationships and their history are crucial for appreciating the extent of Ammianus’ perilous frankness. Following conventional “rules,” all comments that Ammianus makes about Theodosius and his father (also called Theodosius) are panegyrical and read like an intrusion by a different genre into his narrative. A recent study claims that “Ammianus could not have” blamed Theodosius for anything.76 Accordingly, Ammianus does not mention the trial for treason and execution of Theodosius’ father (in ca. 375).77 When he introduces the reign of Valentinian I (364–375) and Valens (364–378), during which that trial occurred, he says cryptically that he will avoid well known matters, partly to escape “the dangers that often attend upon truth,” but then he quickly changes the subject to explain why he will not divulge trivial matters—which is not the same thing as the “dangerous” matters to which he has just alluded.78 When he then criticizes Valentinian I, he does so “now that I am free to speak freely,” suggesting that previously he was not and, by implication, that he may not now be free to criticize the reigning emperor Theodosius.79
In the next book, Ammianus echoes Tacitus when he refers to the better moral climate that prevails now, i.e., under Theodosius, but in practice (again as in Tacitus) this apparently makes it easier for him to recount only previous reigns, not the current one.80 His praise of Theodosius is, then, formulaic: he does not stake his safety on it. This claim leads Ammianus to digress on the dangers that authors sometimes face when they tell the truth, which he illustrates through the ancient (and thus safe) story of Phrynichos, who was punished by the Athenians for inadvertently reminding them of their failure to assist Miletos when it was sacked: “They punished the author, whose object, they thought, was not to console but to reproach.”81 Ammianus thereby reminds us that authors can be punished, even when they do not intend to “reproach.” A mere mention of a sensitive topic is enough. In the next book, Ammianus cryptically advises his readers to consider what he is passing over in silence, though he does not say why he is reminding us of this.82 He is alluding to contemporary knowledge that was unfit or unsafe to write. Finally, in summing up Valentinian's character, he explicitly notes that only posterity can truly be “an impartial judge of the past, being free from the constraints of fear and shameful servility.”83
In other words, while constantly affirming the benevolence of Theodosius, Ammianus reminds us repeatedly of the perils of writing in a way that suggests that they still hang over him. Overall he offers a terrifying picture of the empire under Constantius, Valentinian, and Valens, who account for twenty-two of the twenty-five years covered in the surviving portion of his work. This picture is lifted only for the brief reign of Julian, whom Ammianus presented favorably, if not uncritically. Other than his panegyrical references to Theodosius, which are pro forma, Ammianus gives the reader no substantive reasons for thinking that matters had improved after 378, especially as Theodosius was the opposite of his hero Julian in most respects. Theodosius is said elsewhere to have been interested in history,84 but there is no evidence that he read the Res gestae.
Let us also consider the other side of the merged dynasty. Ammianus does not paint a flattering portrait of Valentinian I, who, we must remember, was the father of Valentinian II, a reigning (albeit young and relatively powerless) emperor at the time when Ammianus was writing. Valentinian I was also the father of Galla, Theodosius’ second wife and so a current empress. Ammianus says little overall about Valentinian I's elder son, Gratian (367–383). He presents him favorably, although he does compare him at one point to Commodus in frivolity. Finally, all that he says about Valentinian II himself is that he was elevated to the throne when he was a child, basically the same kind of bland report that Herodian gives regarding Gordian III.85 In other words, Ammianus brought the whole apparatus of silence-or-panegyric to bear on Theodosius I, managing nevertheless to leave us with a striking impression of terror, if we read between the lines.86 By contrast, he treated the still-reigning dynasty of Valentinian as if it were extinct, with bland neutrality about its current members, with criticism of their father Valentinian I, and with considerable hostility toward their uncle Valens.
Based on these facts, and through a rigorous application of our a priori “rules” about historiography, one scholar has tried to date the last six books of the Res Gestae, which cover the Valentinian dynasty, to (at least) after the death of Valentinian II in 392.87 But as we have seen, this was by no means necessary (and poses other problems of dating that we need not go into here). Ammianus was either brave—he was, after all, about sixty years old—or else one could treat the dynasty of a child-emperor in this way, or both. And we are told explicitly by another historian that Valentinian II had virtually no power and was entirely in the hands of his advisors.88 As we have seen repeatedly, these “advisors” may have always been more lax toward frank authors.
CONCLUSION: PARRHESIA, PANEGYRIC, AND SELF-CENSORSHIP
From the preceding survey, it is possible to conclude, first of all, that not all late antique political histories ended with the previous reign or dynasty, nor did they all cover the present regime panegyrically. A minority of historians ventured critically into the present, though under circumstances that they must have deemed relatively safe. Second, such frankness was for the old. In addition to the cases discussed above, the historian Zosimos was a retired fiscal officer when he finished his New History in ca. 500. The work was daring not because it criticized the emperor but because it was so anti-Christian.89 Younger authors at the start or middle of their careers were, by contrast, more eager to please (e.g., Aurelius Victor, Eutropius, and Menandros). This is an interesting sociology. Third, the accession of a child-emperor seems to have opened a brief window for more frank approaches to his dynasty. And fourth, we should not assume that the imperial system was interested in cracking down on each and every author who did not gush with praise, as we seem to expect it to, or that any given regime was even able to do so. Even if the court was made aware of every published work, there were individual circumstances that placed some authors beyond serious concern. Methodologically, therefore, we must treat each case on its own terms and not impose the same template on all based on a priori assumptions about what they could or could not have said.
On a more general level, we may reposition the late antique historians in relation to two complementary ideals of their culture, parrhesia and panegyric. Parrhesia was a social role performed by a man of virtue who, at personal risk, exposed unpleasant truths to those in power or sought to moderate their excesses for the greater good.90 Panegyric was the exaltation of the current emperor as an ideal monarch who would lead humanity into a golden age. These two seem like opposites, but could be combined: a bold intervention could rest on panegyrical premises, while a panegyrical address could also offer corrective advice.
Obviously, historians exemplify neither type perfectly and fall somewhere in between, albeit in idiosyncratic ways. Parrhesia is not a good model for them because it was ideally practiced in person, in the flesh, as it entailed putting one's self on the line. It was the totality of the speaker's moral presence that earned him the credit to speak freely. This did not carry over well to the sphere of writing. While a few authors donned a parrhesiastic persona,91 historians were not among them, even though they did comment on the dangers of speaking the unvarnished truth and sometimes on the moral kudos owed to them for doing so. Among the latter, Eunapios did attack the current dynasty, but not in order to reform it: he thought it was hopelessly corrupt. Ammianus exposed many of the evils of the imperial system, but set his critique in the past. He admitted that the dangers of speaking about the present were too great. If he hoped that the powerful would improve their ways after reading his work, they would have to do it by drawing their own conclusions. Prokopios revealed some failings of Justinian's regime in the Wars (e.g., corrupt officials), but he did not adopt an overtly parrhesiastic persona, and he disclosed the full awful “truth” in a secret work. Moreover, the more critical historical works were not addressed to the regime in the first place but to a wider community of readers, and to posterity. Thus, while some historians were braver than we assume, they were also prudent and even evasive. Parrhesia does not capture their authorial projects; it was for monks and philosophers.
Panegyric did not sit well with the historians either. Few of them wrote works panegyrically dedicated to emperors that also promoted those emperors’ propaganda (Eutropius for Valens may be an exception); usually it was a tacked-on concession. Many disdained such praise as an obnoxious and compromising imposition upon the objectivity of history, and they avoided it at the cost of sacrificing contemporary history. When they had to include it, they made it seem generically alien to their project or treated it as a mask that could be replaced with another when circumstances or the context changed (e.g., Aurelius Victor on Constantius II). Some, especially Dio Cassius and Ammianus, exposed how insincere panegyric was in Roman political life, implying the same for their own occasional (textual) concessions to it. We here run up against methodological problems that divide disciplines as we move from classical to late antique history. Classicists are more likely to view the ancient historians as exposing subversive truths about the political order, including the hollowness of imperial panegyric.92 But the later we travel into late antiquity, the more scholars treat imperial panegyric as the key to unlocking both social mentalities and authors,93 including authors who are as problematic as Prokopios. Full and unconditional acceptance of the ideological order is taken as a given, even when it is granted that these authors were not entirely at liberty to speak their mind. This, in turn, requires methodologies for containing the dangerous deviation that is implied by the very existence and contents of the Secret History.94 The methodological divide is appreciable, and may stem from the difference between literary scholars who view texts as constructing their own realities and social historians who view texts as reflecting wider social ideologies and mentalities.
The present study gives more comfort to the former approach. The limits of safe speech were always narrower than those of thought, and the gap between them was, in our texts, expressed in two ways: silence and possible indirection, including allusions. Early imperial culture was rife with such subversive classical allusions,95 and there is no reason to assume they lapsed in late antiquity, which also had biblical texts from which to draw. This area has not been systematically explored in the study of late antiquity, yet we know some emperors who grasped that esoteric allusions in texts pointed covertly at them.96 We have to invest more in learning to read between the lines, but also in reading silences. We should not fill those in with the official ideology, to treat the latter as the default content for anything “missing.” As Tolstoy said, “What matters is not what the censor does to what I have written but to what I might have written.” Silences born of peril swallow up whole parts of life, but they were nonetheless real. For example, just because pagan authors did not complain explicitly about Christian persecution of pagan intellectuals does not mean that intellectual debates were being carried out in a courteous and safe environment of polite disagreement.97 Self-censorship challenges modern readings that stick “safely” to the surface of what is reported out of fear of “overinterpreting” them. Silences do not reliably prove absences. They are rather useful photo-negatives of territory that has yet to be mapped. How many potential Secret Histories were not written?