This book has been awaited since Kelly’s contribution s on the subject to the Cambridge Ancient History and Bowersock, Brown and Grabar’s guide to late antiquity. Very much a book of two parts, Ruling the Later Roman Empire deals chiefly with the administrative and bureaucratic elites of the late empire, first through the prism of the De magistratibus of John Lydus, then through three long chapters on the symbolic and actual meaning of late Roman bureaucracy and its contrast with the early empire. Kelly’s sources are familiar, at least until the short epilogue in which he adduces sermon literature envisaging heaven in imperial terms. His approach is expository and analytical, resolutely traditional: the shrine of Bourdieu is by-passed without modish genuflection, and a topic seemingly made for thick description does not get it. All of which is to say that, to those for whom such things matter, Ruling the Later Roman Empire will seem woefully under-theorized.
But the book is conceptually modern in other ways. Kelly is laudably unwilling to surrender administrative history to German dissertations and gigantic French thèses exhaustively cataloguing the legal minutiae of individual departments. Instead, he searches to find the personal and affective side of Roman bureaucracy, with great success: this is a fundamentally humane book. Kelly can show imaginative feeling both for an invented ageing bureaucrat, gasping with relief when the emperor insists on promotion strictly by seniority, and for canny operators slowly weaving intricate webs of shadowy power and perquisites. Time and again, the book conjures the image of a late Roman Sir Humphrey Appleby, ranging self-interestedly through bureaux every bit as tiresome as the Department of Administrative Affairs, lacking only a sense of humour to complete the picture.
Kelly begins from the proposition that the rancour, the misanthropy, and the partis-pris of John Lydus qualify his De magistratibus not as a dispassionate account of Roman bureaucracy but rather as a privileged insight into what it might have been like to be a late Roman bureaucrat. The first two chapters therefore revolve around John and the dashing of his high hopes when John the Cappadocian tears down the judicial side of the praetorian prefecture’s administrative apparatus. Both chapters adumbrate, through the figure of John, themes that will be given a more systematic analysis in the book’s second half. After an expository first chapter, “All the Prefect’s Men,” chapter two on “The Competition for Spoils” expands on the themes of Kelly’s elegant OCD entry on corruption. Again centring on the Justinianic evidence of John Lydus, Kelly shows how important fee income was to late Roman bureaucrats, and by how much it dwarfed their salaries. For emperors, fees paid to officals redistributed the cost of government away from the state and onto those who made use of its services. Reorganization and administrative reform were resisted because they were unpredictable and because those who stood to lose out from reform were neither litigants nor the emperor but rather the staff of the affected bureau. Thus a man like John Lydus viewed as a vile innovation any reduction of his own bureau’s competencies, however well-intentioned or useful in promoting efficient government the change might have been.
John’s autobiographical perspective is abandoned in the second part of the book. Throughout its three chapters, this section deals in contrasts between early and late imperial government — between the “economy of effort” with which the early empire was administered through local elites and ramified patronage networks, and the bureaucratic state of the later Roman empire. Kelly puts the numerical difference in the scale of government between early and late empire at something on the order of two or three times, which may be an underestimate, but he would regard the substance of that change as momentous even if the smallest number obtains. For Kelly, the introduction of accepted fees for government services, as much as a new self-consciousness among bureaucrats, helped create a praxis of government which not only supplemented but also partly supplanted traditional government through extended networks of patronage. Even though money had always played a part in Roman government and access thereunto, the early empire had seen money as at best a supplement, and never a substitute, for the exercise of personal bonds of friendship and patronage. With the widespread extension of fee-paying, an inherently impersonal transaction, the fourth and later centuries witnessed a change to the way the state’s power was organized and distributed, if not always a direct change in the social groups who exercised that power.
Chapter 3, “Standing in Line,” deals with the difficulties of carrying out the sort of government services and supervision that late Roman emperors would ideally have wanted, particularly the slowness and unpredictability of communication and the physical difficulties of accessing such information as might be collected and compiled. Kelly contrasts the ratio of government administrators to population in early and late empires, but illustrates how even in the later empire there were very strict limits to what administrators could do. Kelly would argue that these limits — which we regard as inefficiencies or inadequacies — were actually systemic and not wholly undesirable, for they added to the mechanisms of imperial control by never making the benefits of imperial government wholly predictable.
Chapter 4, “Purchasing Power,” studies how the balance between patronage and money in organizing access to government shifted in favour of the latter during late antiquity. Kelly usefully attempts to determine where boundaries of government accessibility were drawn socially in order to discover who might feel that imperial government was capable of access and who simply would not. He shows, through the admittedly impressionistic set of figures we possess from the period, that many poorer and rural citizens were effectively priced out of the market, while the fees required to initiate and prosecute legal actions also priced many disputes out of the official judicial market altogether. The nexus of cost and exclusion, like fee-paying itself, worked to the overall advantage of imperial government: it shifted the cost of litigation onto the consumer, saved government the cost of redistributing judicial fees to officials as salary, and by excluding some people and some actions also eased pressure on the system as a whole. It is this rational calculation of cost and benefit that lies behind the later empire’s multiple impulses to arbitration, as discussed recently by Jill Harries’ Law and Empire in Late Antiquity.
Chapter 5, “Autocracy and Bureaucracy,” looks at the late antique shift in the modes of government as experienced from above — at the way in which the extension of bureaucracy opened up for the emperor the possibility of more complete and more penetrating government. For all the tantalizing prospects of control which the extension of bureaucracy held out to an emperor, it could also lead to trouble if officials became too independent. It was hard for emperors to balance the measure of independence necessary for officials to accomplish the tasks required of them against the need to maintain enough control to drive, rather than be driven by, their bureaucrats. To prevent the delegation of authority turning into its haemorrhage, emperors might turn to deliberately unpredictable restructuring or to occasional terrorism within the ranks of a particular department in order create to just enough uncertainty to enforce their primacy of power. Lines of control within late Roman bureaucracy were kept intentionally complex for the same reason. Complexity increased the chance that the emperor’s will would be carried out, because officials with overlapping responsibilities maintained a rivalry and suspicion that increased surveillance, and hence their fear of being seen not to do their jobs. There was, as Kelly puts it, “a deliberate imperial reluctance to demarcate clearly areas of responsibility” (209). If imperial officials were kept uncertain, they would be unwilling to make many decisions without first referring them to the emperor himself. That, in turn, kept the emperor central to the administration of the empire.
As one would expect from Kelly’s earlier treatments of late Roman bureaucrats, there is much of value here. His readings of John Lydus and analogous sources show clearly how the identity of an imperial bureaucrat invariably saw individual ambition mediated through corporate pride in bureau, which in turn mitigated and legitimized the expression of ambition. Similarly good are his explications of how ordinary objects like inkwells, and ordinary actions like writing, could be transformed and, in a sense, ennobled by their bureaucratic context. Insofar as the limited evidence allows, Kelly demonstrates the complexity of relations between the bureaucratic establishment and the shifting tableau of high-officials, so much more vulnerable to imperial displeasure than their subordinates but so much more capable of harming those beneath them — and thus calling forth the sort of shrewish memoir which a man like John could produce.
With its unflinchingly authoritative tone, and standing as it does in a field of one, Kelly’s book is destined for much approving citation in the coming years, and deserves it. But it deserves even more than that to be debated, and widely. In that spirit of debate let me close with three points of discussion which, to my mind, Kelly’s analysis begs.
First, a question of method. Kelly is surely correct to understand the bureaucratic corps of the later empire as a deliberately exclusive group which imposed itself on others in large part by consciously mystifying its activities. John Lydus, however, comes very late in the process by which bureaucratic power and exclusivity were created, and by setting him at the start of his story, Kelly may risk creating an implicit teleology of late Roman bureaucracy. Thus, to explicate the ceremony celebrating John’s retirement with the rank of cornicularius Kelly turns for help to the famous missorium of Theodosius. The bestowal of codicils portrayed in the missorium is said to have “an eternal quality” (20), and so it was no doubt meant to. But to accept the aspiration to eternity inherent in the scene raises an old problem, the same one that beset Jones’ great Later Roman Empire : is a synchronic approach necessary to make sense of the fragmentary late antique evidence, or ought we to take a diachronic tack to avoid the elision of progressive change made inevitable if we take chronologically diverse evidence as a piece? The problem may not be soluble, but by declining to address it we risk endorsing the late Roman bureaucracy’s image of its own continuous existence, rather than explaining the necessary reality of change over time. In the present case, at any rate, starting from John has the effect of foregrounding Constantinople too soon and making too much of its centrality in the fourth century. That, in turn, deprives Kelly of the chance to explore how the rise of the very bureaucratic outlook he documents was imbricated in the transition from a mobile imperial court to one rooted in a stationary imperial capital.
Second, the vexed question of the connection between late imperial bureacracy and its early imperial antecedents. Kelly is comfortable with the traditional sharp division between early and late imperial modes of government, and, as he himself has done so much to show, a wealth of evidence exists to document the self-consciousness of late imperial administration. And yet the traditional division is by now perhaps a little too comfortable. That early imperial government was passive and reactive does not mean that the empire was systematically under-administered, and it is possible that by looking for direct prototypes of late antique administrators in the early empire we are asking the wrong question. Nelis-Clément’s study of the beneficiarii, for instance, does much to dispel the view of the early imperial provinces as under-administered; indeed, some of the governmental processes it describes are distinct parallels to those performed in the later empire by Kelly’s bureaucrats.1 To be sure, the beneficiarii were military officials, while late antique bureaucracy was resolutely, almost agressively, civilian. But it is no less plausible to look for the antecedents of late antique bureaucracy in the specialist military administrators of the high empire, than it is to accept traditional accounts of late antique administrative autocthony. Similarly, the Wirgefühl of late Roman officialdom is by no means uniquely late antique. Its affective and exclusive language, its advertisement of both competitive and communal experience, has an unmistakeable forerunner in early imperial collegia, those epitomes of Rome’s reflexive corporatism. Again, the late imperial regulation of the number of, and intervals between, tenures of lucrative bureaucratic posts, magisterially dissected by Kelly, has clear parallels in every period of Roman history and is simply one more expression of the Roman impulse towards oligarchy, regardless of the ostensible governmental dispensation. While documenting the uniqueness of late antiquity is useful, indeed an indispensable part of the historian’s task, it is just as important to see continuities with the past where those exist.
Third, and finally, the Banaji effect. Kelly’s book — like many others, not least my own study of Spanish late antiquity — ignores Jairus Banaji’s arguments about the economic transformation of the later Roman empire (in Agrarian Change in Late Antiquity [Oxford, 2001]). The reason is not far to seek: Banaji’s sociology and his economics are monuments of opacity, painfully inaccessible to non-specialists. But a weakness shared is no less a weakness, and Kelly’s decision not to engage with Banaji’s argument must stand as an opportunity missed, especially because Kelly makes great play with the ordo salutationis from Timgad — where charges are expressed in modii of wheat — in his discussion of the shift from patronage to fee-paying in organizing access to Roman government. If Banaji is correct in seeing the Constantinian solidus as fundamental to the creation of a specially late antique social hierarchy, because it created a stable form of wealth to which government employees had privileged access, that affects Kelly’s arguments about the transformative role of fee-paying in government: it would imply that the decisive shift to fee-paying took place not just because of factors inherent to government, but rather because fourth-century bureaucrats found themselves by accident on the winning side of an economic transformation that gave them a leg up on municipal and landed elites who had hitherto dominated the patronage networks of government. Fee-paying came to supplant personal patronage relations because the new monetary situation favoured precisely those who were in a position to demand fees, and whose understandable enthusiasm for this good fortune drove forward the economic transformation of the larger empire. As it stands, Kelly’s account of bureaucracy’s political economy is masterly; in dialogue with the empire’s wider economic history it would have been more persuasive still.
That Ruling the Later Roman Empire can elicit questions fundamental to understanding the trajectory of Roman history as a whole is testament to its success. Rather than the last word on its subject, the book should be the first on many others.
1. Jocelyne Nelis-Clément, Les Beneficiarii: militaires et administrateurs au service de l’empire. Bordeaux, 2000.