PART I
2
The guilty thief is produced, is interrogated as he deserves; he is tortured, the torturer strikes, his breast is injured, he is hung up . . . he is beaten with sticks, he is flogged, he runs through the sequence of tortures, and he denies. He is to be punished; he is led to the sword. Then another is produced, innocent, who has a large patronage network with him; well-spoken men are present with him. This one has good fortune: he is absolved.
This is an extract from a Greek-Latin primer for children, probably of the early fourth century. It expresses, through its very simplicity, some of the unquestioned assumptions of the late Roman empire. Judicial violence was normal, indeed deserved (in fact, even witnesses were routinely tortured unless they were from the élite); and the rich got off. The Roman world was habituated to violence and injustice. The gladiatorial shows of the early empire continued in the fourth-century western empire, despite being banned by Constantine in 326 under Christian influence. In the 380s Alypius, a future ascetic bishop in Africa, went to the games in Rome, brought by friends against his will; he kept his eyes shut, but the roar of the crowd as a gladiator was wounded made him open his eyes and then he was gripped by the blood, ‘just one of the crowd’, as his friend the great theologian Augustine of Hippo (d. 430) sympathetically put it. Augustine, an uncompromising but also not a naive man, took it for granted that such a blood lust was, however sinful in Christian eyes, normal. Actually, all the post-Roman societies, pagan, Christian or Muslim, were equally used to violence, particularly by the powerful; but under the Roman empire it had a public legitimacy, an element of weekly spectacle, which surpassed even the culture of public execution in eighteenth-century Europe. There was a visceral element to Roman power; even after gladiatorial shows ended in the early fifth century, the killing of wild beasts in public continued for another hundred years and more.
As for the rich getting off: this was not automatic by any means, as the senatorial victims of show trials for magic in Rome in 368-71 knew. But the powerful did indeed have strong patronage networks, and could very readily misuse them. Synesios, bishop of Ptolemais in Cyrenaica (modern eastern Libya) in 411-13, faced a brutal governor, Andronikos, at his arrival as bishop. Andronikos, Synesios complains in his letters, was particularly violent to local city councillors, causing the death of one of them for alleged tax offences. Synesios got him sacked, which shows that only a determined bishop with good connections in Constantinople could properly confront abuse of power - or else that a local official, whether good or bad, could fail to survive a frontal attack by a determined political opponent with his own ecclesiastical and central-government patronage network. But the patronage was crucial, and most of our late Roman sources (as, indeed, early Roman sources) lay great emphasis on it. One could not be a success without it. The Roman world was seriously corrupt, as well as violent. What looks like corruption to us did not always seem so to the Romans, at least to those who formed the élite: it had its own rules, justifications and etiquette. But corruption and its analogues did privilege the privileged, and it was, at the very least, ambiguous; an entire rhetoric of illegal abuse of power was available to every writer.
I begin with these comments simply to distance us a little from Roman political power. The Roman state was not particularly ‘enlightened’. Nevertheless, nor was it, around 400, obviously doomed to collapse. Its violence (whether public or private), corruption and injustice were part of a very stable structure, one which had lasted for centuries, and which had few obvious internal flaws. Half the empire, the West, did collapse in the fifth century, as a result of unforeseen events, handled badly; the empire survived with no difficulty in the East, however, and arguably reached its peak there in the early sixth century. We shall follow how this occurred in Chapter 4, which includes a political narrative of the period 400-550. In this chapter, we shall see how that stable structure worked before the western empire broke up, and, in the next, we shall look at religious and other cultural attitudes in the late Roman world. Fourth-century evidence will be used in both chapters, extending into the early fifth in the West, a period of relative stability still, and into the sixth in the East, for the state did not change radically there until after 600.
The Roman empire was centred on the Mediterranean - ‘our sea’ as the Romans called it; they are the only power in history ever to rule all its shores. The structure of the empire was indeed dependent on the inland sea, for easy and relatively cheap sea transport tied the provinces together, making it fairly straightforward for Synesios to move from Cyrenaica to Constantinople and back again, or for Alypius to move from Thagaste (modern Souk Ahras in eastern Algeria) to Rome and back. By 300 it was recognized that the empire could not easily be ruled from a single centre, and after 324 there were two permanent capitals, Rome and Constantine’s newly founded Constantinople. The empire thereafter had, most of the time, an eastern (mostly Greek-speaking) and a western (mostly Latin-speaking) half, each with its own emperor and administration. But the two halves remained closely connected, and Latin remained the official legal and military language of the East until well into the sixth century.
Rome was a huge city, with a million people at its height in the early empire, and still half a million in 400, when it was no longer the administrative capital of the western empire (which was, in the fourth century, Trier in northern Gaul, and after 402 Ravenna in northern Italy). Constantinople started much smaller, but increased in size rapidly, and may have reached half a million, by now more than Rome, by the late fifth century. Cities of this size in the ancient or medieval world were kept so large by governments, who wanted a great city at their political or symbolic heart for ideological reasons. Rome and Constantinople both had an urban poor who were maintained by regular state handouts of grain and olive oil, from North Africa (modern Tunisia) in the case of Rome, from Egypt and probably Syria in the case of Constantinople, Africa and Egypt being the major export regions of the whole empire. These free food-supplies (annona in Latin) were a substantial expense for the imperial tax system, making up a quarter or more of the whole budget. It must have mattered very much to the state that its great cities were kept artificially large, and their populations happy, with ‘bread and circuses’ as the tag went - though the circuses (including games in the amphitheatres of Rome) were paid for in most cases by the privately wealthy. The symbolic importance of these cities was such that when the Visigoths sacked Rome in 410 the shock waves went all around the empire, as we shall see in Chapter 4.
This concern for the capitals was only the most obvious aspect of the lasting Roman commitment to city life. The whole of the world of culture was bound up in city-ness, civilitas in Latin, from which come our words ‘civilized’ and ‘civilization’, and which precisely implied city-dwelling to the Romans. The empire was in one sense a union of all its cities (some thousand in number), each of which had its own city council (curia in Latin, boul in Greek) that was traditionally autonomous. Each city also had its own kit of impressive urban buildings, remarkably standard from place to place: a forum, civic buildings and temples around it, a theatre, an amphitheatre (only in the West), monumental baths, and from the fourth century a cathedral and other churches replacing the temples; in some parts of the empire, walls. These marked city-ness; one could not claim to be a city without them. And the imagery of the city and its buildings ran through the whole of Roman culture like a silver thread. The Gaulish poet Ausonius (d. c. 395) wrote a set of poems in the 350s called the Order of Noble Cities, nineteen in number, from Rome at the top to his own home town of Bordeaux at the bottom (he uses the word patria, ‘fatherland’, of both Rome and Bordeaux); he enumerated his cities by their buildings, and, in so doing, he was in effect delineating the empire itself.
Political society focused on the cities. Their traditional autonomy had meant in the early empire that being a city councillor (curialis in Latin, bouleuts in Greek) was the height of local ambition. This was less so by the fourth century, however, as the centralization of imperial government meant cities finding that more decisions were taken over their heads; the expansion of the senate and the central administration also meant that the richest and most successful citizens could move beyond their local hierarchies, and the curia thus became second best. City councillors became, above all, responsible for raising and also underwriting taxes, a remunerative but risky matter. Slowly, the formal structures of such councils weakened, above all in the fifth century, and by the sixth even tax-raising had been taken over by central government officials. These processes have often been seen in apocalyptic terms, for it is clear from the imperial law codes that curiales often complained of their tax burdens, and that some (the poorer ones, doubtless) sought to avoid office; emperors responded by making such avoidance illegal. Put that together with the trickle of literary evidence for local élites in the West preferring rural living to city life, and an archaeology which increasingly shows radical material simplifications after 400 or so on western urban sites, and the tax burden on city councillors starts to look like a cause of urban abandonment, maybe in the context of the fall of the empire itself.
Such an interpretation is over-negative, however. First of all, it does not fit the East. Here, city councillors were indeed marginalized, and are documented less and less after around 450 (except in ever more hectoring imperial laws), but political élites remained firmly based in cities. What happened was that city government became more informal, based on the local rich as a collective group, but without specific institutions. Senators who lived locally, the local bishop, the richest councillors, increasingly made up an ad-hoc élite group, often called prteuontes, ‘leading men’. These men patronized city churches, made decisions about building repairs and festivals, and, if necessary, organized local defence, without needing a formal role. Nor did cities lose by this; the fifth and sixth centuries saw the grandest buildings being built in many eastern cities. Once we see this post-curial stability in the East, it is easier to see it in the West too. Sidonius Apollinaris (lived c. 430-85), whose collection of poetry and letters survives, was from the richest family of Clermont in Gaul, son and grandson of praetorian prefects, and son-in-law of the emperor Eparchius Avitus (455-6). He did not have to be a curialis, and largely pursued a central government career. But he ended up as bishop of Clermont, enthusiastically supporting local loyalties in his letters, including city-dwelling; and his brother-in-law Ecdicius, Avitus’ son, defended the city with a private army. So this sort of commitment to urban politics did not depend on the traditional structure of city councils. Essentially, it went on as long as Roman values survived; this varied, but in many parts of the empire it continued a long time after the empire itself fell. The presuppositions of civilitas achieved that on their own. In the West, urban élites also had rural villas, lavish country houses where they spent the summer months (in the East, these were rare, or else concentrated in suburbs, like Daphne in the cooler hills above Antioch); but cities remained the foci for business, politics, patronage and culture. Few influential people could risk staying away from them. And where the rich went, others followed: their servants and entourages, but also merchants and artisans who wanted to sell them things, and the poor who hoped for their charity; the basic personnel of urban life.
It is possible to see the network of cities as the major element of Roman society, more important even than imperial central government. By modern standards, indeed, the empire was lightly governed, with at the most some 30,000 civilian central government officials, who were concentrated in imperial and provincial capitals (though this excludes lesser state employees, such as guardsmen, clerks, messengers, ox-drivers of the public post, who could have been ten times as numerous). When we add to this all the evidence we have for the inefficiency and poor record-keeping of Roman government, plus the time needed to reach outlying provinces of the huge empire (to travel from Rome to northern Gaul took a minimum of three weeks; an army would take much longer), we might wonder how the Roman world held together at all. But it did; a complex set of overlapping structures and presumptions created a coherent political system. Let us look at some of its elements in turn: the civil administration, the senate, the legal system, the army, and the tax system which funded all these. The shared values and rituals of the Roman political élite will then be discussed in Chapter 3, along with the growing importance of a new political structure, the church hierarchy.
The administration of each half of the empire was controlled by the emperor, the central political figure of what was, in principle, an uncompromising autocracy. Some emperors, indeed, imposed themselves politically: in the fourth century Constantine (306-37) and Valentinian I (364-75 in the West) are the most obvious examples, to whom we should add Julian (360-63), whose dramatic and failed attempt to reverse the Christianization set in motion by Constantine has fascinated historians ever since; fifth-century emperors were less impressive, but Justinian in the sixth (527-65 in the East) was as dominant as any of his predecessors, as we shall see in Chapter 4. But not all emperors wanted to do much ruling; they could simply live their lives as the embodiment of public ceremonial, as did, for example, the emperors of the first half of the fifth century. Even if they were active, aiming at an interventionist politics and choosing their major subordinates, they could find themselves blocked by poor information and the complex rules of hierarchy from making a real impact (the most active emperors usually had a military background, without direct experience in civil government). Not that most of the major officials of the empire were full-time bureaucrats, either; even the most assiduous politicians were only intermittently in office. The empire, in a sense, was run by amateurs. But the group of amateurs at least had shared values, and family experience in many cases as well, particularly in the West, where there were more old and rich senatorial families, who were often active in politics in the fourth and fifth centuries. And their subordinates were real career officials, who committed themselves to the administration for life. It is that network of office-holders which gave government its coherence. That, and the stability of the offices themselves. The four praetorian prefectures, each with responsibility for a quarter of the empire (and with a hierarchy of provincial governors beneath them), the six major bureaux of central government and the urban prefectures of Rome and Constantinople all had their own traditions and loyalties, going back in some cases for centuries. John Lydos, who wrote an account of government in the 550s, described the praetorian prefecture of the East in which he had served, tracing the office back, impossibly, to Romulus the founder of Rome; he was very loyal to his department, for all its inadequacy and inconsistency, and he saw the whole of imperial history through its ups and downs. One had to put a good deal of effort in to change the entrenched practices and rituals of bureaucracies like these, and not many people did (one was Justinian’s right-hand man, the praetorian prefect John the Cappadocian (531-41), who was thus predictably John Lydos’ bête noire).
One instance of a leading career politician was Petronius Maximus (lived 396-455), from the powerful senatorial family of the Petronii. He seems to have entered the senate of Rome with the ceremonial office of praetor in 411, with particularly lavish praetorian games; he was a tribune in 415, and comes sacrarum largitionum for the West, one of the main financial officials of the empire, in 416-19, starting that is to say at the age of twenty - young, given the importance of the post. He was urban prefect of Rome in 420-21 and again at some point in the next couple of decades (most of these dates are approximate); in 439-41 he was praetorian prefect for Italy, probably for the second time. He was twice consul, a major honour but without formal duties, and had the coveted title ofpatricius by 445. Unusually, for a career administrator, he was briefly emperor, in 455, for two months before he was killed. In a letter a decade or so later, Sidonius Apollinaris speculates about how much Maximus must have regretted the hourly regulated rituals and responsibilities of imperial office, given the contrast with the ‘leisure’ (otium) of being a senator. This seems surprising at first sight, but ‘leisure’ is partly just a manner of speaking: Maximus had long been a major political dealer, with a huge clientele (as Sidonius himself says) and imperial ambitions. We must nonetheless recognize that in the four decades of his political career he only seems to have held formal office for around ten years; he had plenty of time for otium as well, which indeed contemporary authors, time and again, describe as one of the characteristics of senatorial élites.
The senate had its own identity, partly separate from the imperial bureaucracy; indeed, in the West it was even physically separate, for the government was no longer in Rome. It was the theoretical governing body of the empire, as of the Roman republic four centuries before, and although the senate was by now no longer a reality, it still represented the height of aspiration for any citizen. It brought with it many fiscal and political privileges, although it was expensive to enter and participate in, given the games and other ceremonies senators had to fund. It had no formal governmental function, but high officials became senators as of right; furthermore, by the early fifth century, only the highest of the three grades of senator, the illustres, were regarded as full members of the senate, and the title of illustris was only available to officials and direct imperial protégés. The senate was thus tightly connected to govern ment, and expanded as the administration expanded in the fourth century; but it was nonetheless separate, with its own rituals and seniority. It represented aristocratic wealth, privilege and superiority, and, although membership of it was not technically heritable, in practice the same families dominated the senate, in Rome at least, throughout the fourth and fifth centuries. All the male heirs of anillustris were anyway at least clarissimi, the lowest senatorial grade, which involved at least some privileges even after full senatorial eligibility contracted. And all the grades seem to have been regarded as nobilis, ‘aristocratic’, in late Roman parlance. This close but sideways relation to government has some parallels with that of the House of Lords in modern Britain, both before and after the reforms of 1999.
The existence of this effectively hereditary aristocracy was a key feature of the empire. Not because it dominated government; most leading bureaucrats were not of senatorial origin, even if they became senators later (Maximus was in that sense atypical) but rather because it dominated the tone of government. The Roman empire was unusual in ancient and medieval history in that its ruling class was dominated by civilian, not (or not only) military, figures. Only China’s mandarinate offers any real parallel. Senators regarded themselves very highly, as the ‘best part of the human race’ in the well-known words of the orator Symmachus (d. 402); their criteria for this self-satisfaction did not rely on military or physical prowess, but on birth, wealth and a shared culture. Birth was important (Sidonius could be contemptuous of a powerful rival, Paeonius, the praetorian prefect for Gaul, because he was ‘of municipal origin’, that is, from a curial, not a senatorial family), although very long ancestry was less vital; even the Anicii, by far the leading Roman family in the fourth and fifth centuries, only traced their family back to the late second century. Wealth went without saying: no one was politically important in the Roman world (apart from a few high-minded bishops) without being rich. One needed wealth to get anywhere in the civil administration, as both bribes for appointments and the maintenance of a patronage network cost money, but once one was important, the perks of office, both legal and illegal, were huge. In the army, too, although it was more open to merit, all successful generals ended up rich. And the independently wealthy families of the senate of Rome, the Anicii, Petronii, Caeonii and half a dozen others, had estates throughout southern Italy, Sicily, North Africa and elsewhere, ‘scattered across almost the whole Roman world’, as the historian Ammianus Marcellinus said of the leading politician Petronius Probus in the 370s: these may have been the richest private landowners of all time. When two Roman aristocrats, Melania and Pinianus, got religion around 405 and sold off all their land, which provided 120,000 solidi (around 900 pounds of gold) a year in rents, it wrecked the property market, according to Melania’s saint’s life. The senatorial hyper-rich were only in Rome, however; in Constantinople senators were from the provincial élites of the East, and operated on a smaller scale. Throughout the empire, in fact, there were provincial élites, the leaders of which had senatorial status and were in line for public office; they were locally powerful, but could not match the Anicii. Sidonius was an example, and indeed the élites of Gaul seem to have been a particularly coherent group.
A shared culture perhaps marked the Roman senatorial and provincial aristocracies most, for it was based on a literary education. Every western aristocrat had to know Virgil by heart, and many other classical Latin authors, and be able to write poetry and turn a polished sentence in prose; in the East it was Homer. The two traditions, in Latin and Greek, did not have much influence on each other by now, but they were very dense and highly prized. There was a pecking-order based on the extent of this cultural capital. Ammianus reports scornfully that senators in Rome, the supposed crème de la crème, only really read Juvenal, a racy and satirical poet, so by implication not the difficult texts; whether or not this was true, it was a real insult. Conversely, literary experts, such as Ausonius in the West and Libanios (d. c. 393) in the East, could rise fast and gain imperial patronage and office simply because of their writing - in Libanios’ case so fast that he was accused of magic - although both were already landowners of at least medium wealth. The emperor Julian in his attempt to reverse Christianization tried to force Christian intellectuals to teach only the Bible, not the pagan classics, thus enclosing them in a ghetto of inferior prose. This failed, but the assumptions behind such an enactment clearly show the close relationship between traditional culture and social status. Some Christian hard-liners responded by rejecting Virgil, but this failed too: by the fifth century the aristocracy knew both Virgil (or Homer) and the Bible, and might add to these some of the new Christian theologians too, Augustine in the West or Basil of Caesarea in the East, both of whom were good stylists.
It is this culture which makes the late Roman empire, or at least its élites, unusually accessible to us, for the writings of many of these aristocrats survive: elegant letters or speeches for the most part, but also poetry, theology, or, in the case of the fifth-century senator Palladius, an estate-management manual. Roman literary culture used to be regarded as the high point of civilization; this belief, inherited from the Renaissance, perhaps reached its peak in the late nineteenth- and early twentieth-century English public-school tradition, in which Virgil (and indeed Juvenal, by now seen as a more difficult author) was regarded as a basic training even for the government of India, not to speak of an academic career. This belief is less strong now; few academics know enough Latin to read Virgil (outside Italy), and even fewer politicians. It is thus easier today to see Roman literary culture as an attribute of power, rather than virtue; Roman politicians were at least as cynical and greedy as their successors, and not obviously better at ruling. But it is important to recognize its all-pervasiveness; in all the cities of the empire, even local office was linked to at least some version of this education. The shared knowledge and values that it inculcated was one of the elements that held the empire together, and indeed made the empire remarkably homogeneous, as not only its literature but its surviving architecture and material culture show. It must finally be said that, although the Roman world left a dense legacy of institutions and assumptions to its early medieval successors, a literary education was not part of that, except in the increasingly separate career structure of the church. The culture of post-Roman aristocracies instead became military, based on the use of arms and horses, and as a result we know much less about it from the inside.
Roman law was another intellectual system that was, in principle, the same everywhere, and it acted as a unifying force. It consisted of imperial legislation, which was very extensive in the fourth to sixth centuries, and a network of tracts by earlier Roman jurists, which represented a distillation of case-law precedents and the workings-out of legal principles. To master this properly required a special training, at the law-schools of Rome, Beirut or (after 425) Constantinople, although all education involved an element of rhetoric, essential for court advocacy. Alypius spent time training at the Rome law-school in the early 380s before going with Augustine to work in Milan (where both were converted to a more thoroughgoing Christianity, and switched their career path to the church); Augustine, by contrast, although trained in rhetoric, makes it clear in his writings that he did not feel himself to be a legal expert, for his education was not specific enough. Law was not in fact at all easy to master before Theodosius II had imperial laws collected into the Theodosian Code in 429-38. Justinian revised and expanded the code (twice) in 528-34, and had juristic literature of the second and third centuries excerpted and systematized in the Digest in 530-33 as well. The Theodosian Code remained a point of reference in the post-Roman West, even though the laws of the post-Roman kingdoms were different; Justinian’s corpus survived as the law of Byzantium, and was separately revived in the West in the twelfth century. We must, however, be careful about what such a commitment to law means. The complexity of this legal system was such that experts (iurisconsulti) were needed in every court, and sometimes just to draw up documents, but they may not always have been available or been fully reliable if they were. Even if legal help was accessible, courts did not necessarily judge justly, and the rich often benefited from judicial corruption and patronage, as we saw at the start of this chapter and as many sources confirm. In Egypt, papyrus documents recording the settlement of civil disputes in the fourth to sixth centuries show a strong tendency to avoid courts altogether, given their huge expense and danger, and to go directly to private arbitration.
It would be tempting to reduce the law to its criminal dimension, with its recourse to torture, and conclude that the legal system was in practice simply an instrument of heavy-handed state coercion, the work of a public power that relied on terror because it did not have the personnel to dominate daily life in any detail. Such a temptation would be largely justified, but all the same the law was important. Egyptian arbitrations may have avoided the courts, but they refer frequently to legislation and legal terminology. Augustine was not expert in the law, but he sought to know it, for example writing to the iurisconsultus Eustochius for rulings. An interesting letter survives from Africa of around 400 in which an unnamed landowner chides a neighbour and former friend, Salvius, for tyrannizing the former’s tenants: ‘Is there one law for advocates, another for ex-lawyers? Or one equity for Rome, another for Mateur?’ Salvius, an advocate from (we assume) Mateur, would presumably have thought so, and his illegalities are standard. But his correspondent had been a lawyer too; Salvius had taught him the law of tenancy, and it was this, together with the law of inheritance and possession, which the letter invokes in detail, before offering a deal. Law and its imagery were all-pervading in the empire, and we could indeed suppose that the setpiece denunciations of judicial corruption in our sources at least showed high expectations.
The Roman army was much larger than the civil administration, and was always the empire’s major expense: in 400 there were some half a million soldiers, give or take a hundred thousand. These were mostly on the northern Rhine and Danube frontiers, and on the eastern frontier with Persia (the long southern border faced the Sahara, and was less vulnerable), but there were detachments in every province, acting as garrisons and as ad-hoc police. It was of course their existence that made it possible for provincial élites to remain civilian; private armies were very rare before the empire broke up. Conversely, armies were capable of imposing their own candidates for emperor, all the more easily because they held most of the weapons. This had been common in the third century, but was much rarer in the fourth; it revived in the West in the final years of empire in the fifth, but in the East there were no successful coups until 602. Even without coups against the emperor, however, army leaders remained important in politics, and several weak emperors (such as Honorius, western emperor 395-423) had military strongmen ruling for them, who could succeed each other by violence. There was a sense in which the office of emperor was more military than the civilian bureaucracy around him, and emperors were closer to the military than to the civilian hierarchy. Generals were more likely than senior administrators to have risen from nowhere, especially if they came from frontier regions, as was very common; the Rhine frontier and the Balkan frontier in particular were heavily militarized societies, with less and less social distance between the Roman and the ‘barbarian’ sides of the border, as we shall see later in this chapter. This did not make them so very different from the civilian élites, as long as they were successful, as they could end up with senatorial position, civilian clients and a literary education for their children. But military leaders were less dedicated to expensive prestige buildings or the patronage of games, and senators regularly looked down on them for their lack of culture. Soldiers also moved around more than civilians did. The historian Ammianus (d. c. 395), a Greek-speaker who wrote in Latin, the language of the army, was an ex-soldier who had served on both the Persian and the Rhine frontiers, as well as spending much time in Rome.
The scale of the army and its presence everywhere, and the need to keep it properly provisioned and equipped, made it the major concern of the whole Roman state. The state had a developed system of frontier fortifications and its own food-supply lines: the distribution of oil amphorae along the lower Danube, for example, shows that the army there was supplied from the Aegean into the late sixth century. It also had its own factories for military equipment, of which thirty-five are listed, distributed all across the empire, in the Notitia Dignitatum, an account of the imperial military structure dating to the end of the fourth century. Perhaps a half of the entire imperial budget went on feeding and paying the army, and the logistics of army supply were the single most important element that linked all the imperial provinces together, along with the permanent need to feed the imperial capitals.
Underpinning all these structures, and making them possible, was the imperial tax system, which was based above all on a land tax, assessed on acreage, though also buttressed by a much lighter tax on merchants and artisans, by the revenues from imperial lands and by a variety of smaller dues. In recent years some historians have reacted against an earlier image of the ‘coercive state’ of the late empire, taxing so heavily that land was abandoned and the economy began to break down; this revision is correct, but they seem to me to have gone too far in their arguments. Taxation does seem to have been very heavy overall: in the sixth century a small number of sources, mostly from Egypt, converge in showing that a quarter of the yield of land could go in tax, and it was more in times of extra taxation (superindictiones) which was assessed on top of the main tax burden. This is a very high figure for a precapitalist, agrarian society, with a relatively simple technology. But the high taxes were needed to pay the salaries of all those soldiers, bureaucrats and messengers, and to feed the capitals; they were needed to fund the enormous scale of Roman public buildings and state wealth. They also connected the different parts of the empire together physically, as grain moved northwards from Africa, Sicily and Egypt, and olive oil moved out of Africa, the Aegean and Syria, in ships themselves commandeered by the state (shipowners moved goods for the state as part of their tax liability). This movement of goods was essentially Mediterranean-based, as it was far easier and cheaper to transport in bulk by water than by land; Gaul, the Rhineland and Britain formed a smaller and separate network, and inland Spain, far from both sea and frontiers, seems to have been somewhat marginal. The core of the empire remained Mediterranean, and it, at least, or, rather, its two halves, were unified by the fiscal movement of goods.
A land tax cannot work properly, especially when it is high, unless assessment is accurate and collection systematic. This takes work. The state has to have up-to-date records about who owns the land; these are not easy to obtain systematically (and no easier to keep in order for easy reference), and establishing them requires a considerable amount of personnel and intrusive information-gathering. Land sales had to be publicly registered in the late empire for this reason, and such registrations can sometimes be found in the rare collections of private documents from the late empire, usually papyri from Egypt, although a few texts do survive elsewhere. And, most important, from the fourth century onwards the government issued laws to tie the peasantry, who were actually paying the taxes, to their place of origin, so that they would not move around or leave the land, thus making tax-collection more difficult. These laws were part of a general legislative package aimed at ensuring that people essential to the state stayed in their professions, and that their heirs would do so too. Curiales were tied to their offices, as we have seen; so were soldiers, and the workers in state factories; so were shipowners and the bakers and butchers of Rome, who were essential for the annona of the capital. Even if this network of laws was regularly obeyed, which we can doubt, they make up a large proportion of the imperial codes, and they were generated by the need to stabilize the tax infrastructure of the empire. Add to that the actual collection of taxes, which could be a tense and violent moment, and was certainly undertaken by armed men, and the impact of the imperial fiscal system was continuous, capillary and potentially coercive of nearly everybody in the empire.
This intrusiveness was made worse by illegality. The rich could buy immunity corruptly; assessors and collectors certainly got rich corruptly. The victims were almost always the poor. They responded by fleeing the land (hence the laws tying them down), or by seeking protection from the powerful against having to pay taxes to the state. There are also laws against such patronage, although we have seen that patronage, too, was a stable part of the Roman political system. Most taxes were, it is true, probably paid regularly and even legally; it is striking that the Egyptian papyrus archive of the sixth-century Apion family, then one of the richest families of the Greek East and overwhelmingly dominant in their home town, the city of Oxyrhynchos (modern Bahnasa), shows them paying taxes in a very routine manner. But given the weight of tax, and the endemic injustice that marked the Roman system, it is not surprising that corruption should focus on it. Social critics, more numerous as the empire went Christian and a radical fringe of moralists gained a voice, very frequently stress fiscal oppression in their invective; only judicial corruption and sexual behaviour were as prominent. This would last as long as the empire.
Taxation thus underpinned imperial unity itself, for it was the most evident single element in the state’s impact on the population at large, as well as the mainstay of the army, the administration, the legal system and the movement of goods throughout the Mediterranean and elsewhere, all the elements which linked such a large land area together. If it failed, the empire would simply break up. But in fact the empire broke up for other reasons, as we shall see in Chapter 4. After it did so, taxation was a casualty in the West, but survived in the East. This contrast cannot be underestimated, and it underpins many of the events described in later sections of this book. All the same, fiscal breakdown was not yet predictable in 400, or even 500 in some places. In 400 the stability, and relative homogeneity, of the imperial system was not yet seen by anyone to be at risk.
So far, we have focused on the state, and the imperial political system in general. Local differences have been downplayed, and our vision has been top-down, seen from the viewpoint of administrators and the rich. Let us now look at the rest of the population, and at some of the regional differences which we can pin down in the late Roman empire.
The first thing to be stated is that the population of the empire consisted overwhelmingly of peasants: families of cultivators, who worked the land they owned or rented, and who lived off the food they themselves produced, as well as giving surpluses to landlords (if they had them) in rent, and in tax to the state. Many of them were servi, unfree with no legal rights, particularly in parts of the West, but the plantation slavery of early imperial Italy and Greece had almost entirely vanished by the late empire, and free and unfree peasants by now all lived their lives in similar ways. (This book will as a result not use the word ‘slaves’ for unfree peasants, as it is misleading; the word will be used only for unfree domestic servants, who were fed and maintained by their masters as plantation slaves had been.) In the early Middle Ages, peasants made up 90 per cent or more of the population; the proportion must have been less in the late empire, as more people lived in towns - in Egypt, exceptionally, up to a third of the total population - but could have been as much as 80 per cent, still an extremely high proportion.
Most peasants were probably the tenants of landlords. Legislators certainly assumed so, for their laws tying peasants to the land were directed to coloni, the standard Latin word for tenant. The huge estates of the emperor and of Roman senators, and the even greater collective landed wealth of all the provincial and curial élites, also presupposed the existence of millions of dependent tenants who supplied their rents. This was often through middlemen, conductores, who leased whole estates from the great landowners; but some of the latter paid considerable attention to managing their own estates for profit, such as the Apions in sixth-century Egypt, and Palladius, the estate-management manualist, in fifth-century Italy. Unfortunately, our evidence is not good enough to tell us how often, and where, peasants owned their own land. Egyptian papyri show that some city territories were dominated by owners of large estates, but others had a substantial landowning peasantry and much more autonomy. A good example is the territory of the large village of Aphrodito (modern Kom Ishqaw), from which many sixth-century documents survive, as we shall see shortly. The still standing late Roman villages of Syria and other parts of the eastern Mediterranean show in the best preserved cases (such as in the Limestone Massif of northern Syria: see below, Chapter 10) an architectural ambition and a homogeneity of house types that is difficult to square with tenurial dependence; there are few visible estate centres, in particular. It is generally thought, therefore, that these villages mostly belonged to independent owners.
Overall, it seems that there were more peasant owners in the East than the West, which also fits the fact that fewer hyper-rich landowners are known of in the East. In the West, by contrast, much of Italy and Africa in particular and parts of Gaul were probably dominated by landowners, and we know of more estates which included large areas; one of Melania and Pinianus’ estates in Africa was ‘larger than the city itself’, that is to say, the city territory of the nearest town, Thagaste. (In Africa, where not all dioceses were based in towns, some estates were so substantial that they had their own bishops.) But in both West and East, even large estates were normally highly fragmented and scattered, and many consisted of hundreds or thousands of separate land parcels; there was plenty of space for peasant owners and village-level élites to exist in between them. Some tenants owned land as well, and the laws on tax-paying distinguish between coloni who owned some land, who paid taxes directly to collectors, and coloni who owned none (calledadscripticii), who paid taxes through their landlords. The latter were much more dependent, more similar to unfree tenants (who did not pay tax: their lords paid it directly); Justinian, indeed, in one of his laws, wondered what real difference there was between serviand adscripticii. The answer probably varied regionally: tenure was certainly more flexible in Egypt, where leases were shorter, more peasants owned land, there was more wage labour and rural unfreedom was very rare; in Italy, by contrast, there were whole estates with only unfree tenants, and rural subjection was probably greater overall.
One real difference between East and West was that peasants lived in villages much more often in the East. Some of the villages still stand, as just noted, at least in marginal areas where the land has since been abandoned to pasture or desert. But documents and archaeology both show that villages (komai or chria) were normal in most of the Greek- speaking world, and they could be tightly organized, with their own headmen, as in particular in Egypt. Owners and tenants lived side by side in these villages, and peasant society was, simply for that reason, relatively coherent and autonomous (eastern landed aristocrats, as we have seen, normally lived in towns), as well as potentially more fraught, as village factions fought over pasture and water rights, or over the pecking-order between the successful and the less successful that existed in every village. We know so much about the Egyptian village of Aphrodito because we have the papyrus archive of Dioskoros, son of Apollos (lived c. 520-85), who was a fairly well-off village leader there: he was sometimes its headman, as his father had been. Dioskoros had a literary and legal education, probably in Alexandria, and became a local notary when he returned; more unusually, he was also a poet, and wrote praise poems to local dukes and other officials. He is interesting for a variety of reasons. He is the best-documented village-dweller of the whole late empire; but his personal character comes across in the sources as well. Although he was certainly from the local élite, he felt threatened on all sides: by the governor of the nearest city, Antaiopolis, jealous of Aphrodito’s autonomy; and by neighbours, tenants, shepherds and creditors in his own village. We have some of his lawsuits; his poems, too, often end with pleas for help; they were transactions in his extensive patron-client network. Aphrodito was not a peaceful village. We even have a double-murder investigation by a senior military official, in which the senatorial aristocrat Sarapammon and his associate, the soldier Menas, defend themselves and accuse the villagers themselves of the crime. It is clear, however, that no single person could control it, and keep down its tensions. Aphrodito was only united when it faced off other villages and threats from Antaiopolis. These fractious societies were typical of the East.
The West was different. Here, villages were rarer, except in some mountain zones; instead, as much archaeology shows, the countryside was scattered with isolated farms and the rural villas or estate-centres of major landowners. Even the concept of the village territory was hardly present in most places; land was simply identified by its owner, and most estates had their own names. We do not have Egyptian levels of documentation here, so it is hard to tell how rural societies worked, but it is likely that they were less coherent than in the East, for there was less to bind them together. Probably the tenants of single estates had something to link them, the common experience of paying rent to a landlord or conductor; this did not match the coherence of village life, but it could increase local tensions. The gap between the powerful and the poor was in general wider in much of the West, in fact, and we can sometimes see its results.
One example comes from Augustine’s Africa. Augustine, as bishop of Hippo, appointed his monk Antoninus in the 410s to be bishop of a subordinate diocese at Fussala, one of Africa’s relatively few villages, in the hills of what is now eastern Algeria. Antoninus turned out to be a bad man - he was young and from a poor family, he was promoted too fast - and he terrorized his village, extorting money, clothing, produce and building materials. He was also accused of sexual assault. Augustine removed him, but did not depose him, and tried to transfer him to the nearby estate of Thogonoetum. Here, the tenants told Augustine and their landowner that they would leave if he came. Antoninus caused no end of trouble, even appealing to the pope in Rome (this being the context in which two surviving letters were written about him by Augustine, in 422-3). Augustine was very embarrassed, as indeed he should have been (‘I did not dare look the people of Fussala in the eye’). It is interesting, however, how scared the peasants were: in their angry and bitter witnessing, even after Antoninus’ removal, they would not give their names. The people of Fussala included tenants (who were interrogated without their conductores being present, to try to get them to relax), but probably not all of them were dependent; it is interesting, conversely, that the coloni of Thogonoetum were more prepared to resist Antoninus than were the villagers - illegally, too, for they were of course tied to the land by law. All the same, peasant protagonism here seems largely negative, marked by bitterness, fear and rejection. There was too much separation in this part of Africa between peasants and landlords, and more hostility between them as a result; there was no Dioskoros to mediate between the peasants and the authorities. It is not surprising that Augustine’s main fear was that the peasants would revert to the Donatist church (see Chapter 3), abandoning Catholic Christianity altogether.
Another element that was very different from place to place were the patterns of commercial exchange and artisan production. Three decades of archaeology have led to a major revaluation of late Roman commerce, which as late as the 1970s was thought to be marginal to the economy. On archaeological sites, the density of finds of amphorae (which carried wine, oil and fish sauce above all, that is, food products) and fine pottery (a guide to other large-scale artisanal products such as cloth and metalwork) allows us to say which areas of the empire were major exporters, and where their products typically went. North African Red Slip tableware is found all over the late Roman Mediterranean; similar tableware from Phocaea on the Turkish Aegean coast and Cyprus matches it in the eastern Mediterranean as well. It evidently travelled by sea, but can be found quite far inland in Italy and in Syria and Palestine. In northern Gaul and Britain and in inland Spain it was not available in more than tiny quantities, but large-scale local production is found instead; for this reason above all we can say that those areas, although active, were separate from the main Mediterranean economic network. Cloth, always the main artisanal product, is not easy to identify archaeologically, but literary sources (including the detailed lists in the imperial Price Edict of 301) show that Italy, Gaul, Egypt and Syria were among the major exporters. Amphorae allow us to add African, Syrian and Aegean oil, and south Italian, Palestinian and Aegean wine. These were large-scale distribution networks, and the commodities concerned were evidently produced on a large scale as well. Indeed, the African (that is to say, above all, Tunisian) and coastal Syrian/Palestinian economies probably depended substantially on exports for their prosperity. Internally, too, the complexity of the economies of southern Italy, the Aegean, Egypt and Palestine in particular, seems to show a dense network of inter-city and city-country exchange.
We have already seen that some parts of the empire sent much of their surplus in tax to other areas: Africa, Egypt and to a lesser extent Syria, Palestine and the Aegean. These provinces were probably in agricultural terms the richest in the empire (the climate was then much as it is today, global warming apart); and they are mostly prominent in these commercial networks as well. It would certainly be wrong to see the archaeological distributions as signs of the tax network only; they extend to too many insignificant places for that to be the case, such as tiny settlements in central Italy or eastern Palestine. But it is likely, all the same, that commercial exchange was underwritten by the tax network. Ships left Africa for Italy every autumn, bringing state grain and oil to Rome as annona; doubtless they took commercial goods as well, ceramics and once again oil, the transport costs of which were thus covered by the state, and which could be sold on the other side of the Mediterranean more competitively, whether in Rome or in other ports. Egypt’s commercial exports are less well known, but they probably consisted above all of cloth and papyrus, which archaeology does not pick up (Egyptian wine production was enormous in the late empire, but was of low quality, and was for consumption within Egypt only). The tax network made commerce easier, and also contributed to the commercial prominence of certain regions. When the empire began to lose its fiscal homogeneity in the West, which was when the Vandals seized the heartland of North Africa in 439, breaking the Carthage-Rome tax spine, western Mediterranean commerce began two centuries of steady involution; but the East remained politically and fiscally strong, and eastern Mediterranean commerce was as active in 600 as in 400.
The late Roman world always maintained a double face, local and imperial. Latin and Greek were far from its only languages. Proto-Welsh was spoken in Britain, Basque in parts of Spain, Berber in Africa, Coptic in Egypt, Hebrew, Arabic and Aramaic/Syriac in the Levant, Isaurian and Armenian in Anatolia, and there were doubtless other languages too. Coptic, Hebrew, Syriac and Armenian had their own literatures. Local societies were at least as different then as they are now, in the range of realities that stretch from the Welsh mountains to the Egyptian desert, both as a result of their necessary adaptations to the huge differences in local ecology, and as a result of the more human-made contrasts discussed in the last few pages. On the other hand, the Roman world not only held together but increased many aspects of its cohesiveness with time. Christianization swept away many local religious traditions, as we shall see in the next chapter. Cities looked remarkably similar, in their public buildings and their layout, in different parts of the empire. The administration and the army had the same overarching structure everywhere, and the tax system affected everybody. Some cultural differences were lessening: Gaul, for example, lost its local language, Gaulish, perhaps in the fifth century. Egypt, in particular, was much less atypical in its society and culture in the fourth and fifth centuries than it had been in the first and second; it had ceased to use its huge temple complexes and had abandoned their Pharaonic architectural style, and had even deserted its traditional beer-drinking in favour of wine. People felt themselves to be part of a single Roman world, an awareness which extended not only to city élites but even into villages, for Antoninus of Fussala had appealed to the pope in Rome for support against Augustine, and the villagers of Aphrodito appealed to the empress Theodora herself for support against the governor of Antaiopolis.
This awareness of a wider community is linked in our sources, over and over again, with patronage. The patron-client relationship has existed in most societies (the lord-vassal bond of the central Middle Ages is an example), but Roman culture laid immense stress on it. Seeking help from a patron, alongside official channels, was normal. It could be stigmatized as corrupt, but often only by extreme moralists, or else by victims; most people, however, accepted its day-to-day logic. Actually, even the official channels were often expressed in patron- client terms, as with personal or collective appeals to the emperor, which were commonplace, or as with the endless, and legal, personal payments (sportulae) which were expected by low- and medium-level bureaucrats who might either facilitate or obstruct tax registration or a court case. The point about a patronage system of this kind is that in the end it involves everybody, and everybody can feel they somehow have a stake in the social system. They will often not get anything out of it, as with the average peasant, but they feel that they can get an element of protection from patrons, if not this time then the next. Everyone except the emperor and his most powerful subordinates needed a patron, and sometimes many. They boasted about it, too, as when John Lydos was fast-tracked as a trainee administrator by the praetorian prefect Zotikos, who was from the same province as him, and did not even have to buy his appointment. Similarly, everyone with even a modicum of local power, from Dioskoros upwards, had clients. Abinnaios, a medium-level soldier stationed in southern Egypt in the 340s, whose archive also survives, preserved requests for special favours from his subordinates, but also from friends and clients who were city councillors, priests, artisans or peasants; he was asked to arbitrate disputes, and to apprehend robbers. Little of this was in his official remit, but it was totally normal. The Antiochene intellectual Libanios was outraged in the 390s when his tenants sought a military patron to protect them against paying him rent; he claimed that their main patron should be their landlord, but anyone in his audience would have known that was specious. A great part of the elegant letters that the educated élite wrote to each other consisted of or included recommendations for clients or requests for help. So did Dioskoros’ poetry, as we have seen. Far from ‘corrup tion’ being an element of Roman weakness, this vast network of favours was one of the main elements that made the empire work. It was when patronage failed that there was trouble. Peasants in Africa who felt that the Catholic church’s patronage was unavailable to them could turn to Donatism. When peasants in Egypt who had used patrons to lift some of their tax burdens in difficult years felt that this did not work, they would flee; and when the new Arab government after 640 excluded traditional rural patrons from political influence, as we shall see later, in Chapter 12, they could revolt. Above all, perhaps, when local élites in the fifth-century West ceased to believe that their traditional patrons in central and provincial government were capable of helping them, they could turn to the new military leaders of ‘barbarian’ tribes in their localities instead, and a major political shift resulted. We shall look at the causes and consequences of that shift in Chapter 4.
The Roman world was surrounded by ‘others’, whom Romans regarded with varying degrees of contempt and incomprehension, but who interacted with them in complex ways. To the east, there was always Persia, the great sister empire of west-central Eurasia, ruled between the 220s and the 640s by the Sassanian dynasty. This was a permanent threat, but a stable one: it involved only border wars, at most extending into Syria, for the two hundred and fifty years between Julian’s disastrous invasion of what is now Iraq (then Persia’s economic and political heartland) in 363 and the temporary Persian conquest of the Roman East in 614-28, which culminated in the siege of Constantinople in 626. The Persian state was almost as large as the Roman empire, extending eastwards into central Asia and what is now Afghanistan; it is much less well documented than the Roman empire, but it, too, was held together by a complex tax system, although it had a powerful military aristocracy as well, unlike Rome. The militarization of Persian culture extended west into Armenia, which Romans and Persians fought over but which remained partly independent and culturally separate. The Armenians converted to Christianity in the fourth century, which separated them further from the Persians, who were Zoroastrian for the most part (although with sizeable Jewish and Christian minorities, and also local traditional religions). Zoroastrianism certainly contributed to Persian ‘strangeness’ in the eyes of the Romans; for example, its priests, called magoi in Greek or magiin Latin, gave their name to ‘magic’ in both languages, even though Zoroastrian religion favoured an abstract theology and public rituals, just as Christianity did. But it was arguably Persia’s military culture and enormous respect for ancient dynastic tradition that marked it out as most culturally different from Rome, for the Roman sense of kinship could link far-flung cousins and cousins-in-law in patronage networks, but even ‘old’ families rarely had more than a century or two of prominence. The dynastic element helped Persian traditions survive better than Roman traditions when both were swept away, from Carthage to Samarkand, by the Arabs in the seventh century.
Rome’s other borders were shared with far less organized political groups, all of which the Romans called barbari, ‘barbarians’, a conveniently vague term which I shall adopt (keeping the inverted commas) as well. To the south they faced nomadic and semi-nomadic tribes in the Sahara and its fringes, mostly speaking Berber languages; for a long time these were not taken very seriously as military threats, but such groups were gaining in social and military coherence, largely as a result of Roman influence, and one tribal alliance, the Laguatan, was very aggressive at the start of the fifth century, as Synesios in Cyrenaica, among others, complains; the Vandals in Africa had trouble with Berbers later, too. The Picts and the Irish to the north and west of Britain were also a potential threat, although only to the already militarized British borderlands, especially around Hadrian’s Wall (they staged a substantial invasion in 367-8). The long Rhine and Danube frontier faced tribal communities, mostly speaking Germanic languages, which historians since Tacitus in the first century had seen as a whole as Germani, although there is no evidence whatsoever that these peoples recognized any common bonds. The main groups along the frontier were by the fourth century the Franks on the lower Rhine, the Alemans on the middle and upper Rhine, and the Goths on the lower Danube and north-eastwards into the steppes of what is now Ukraine. Further back were Frisians in the modern Netherlands, Saxons in modern north Germany, and Vandals and Longobards or Lombards to their east. These were the main groups, but there were dozens of others. The Quadi in what is now Slovakia and Hungary are perhaps worth mentioning, if only because, after they fought a small war against Valentinian I in 374-5, they met the emperor and argued (correctly, in fact) that their own attacks were a justified and largely defensive response to Roman aggression: this was seen by Valentinian as so insolent that he had an apoplectic fit and died. One might have a soft spot for the Quadi as a result, but they vanish from history soon afterwards: they must have been absorbed into the Hunnic empire in the early fifth century, which was based in the same area, and their probable descendants in the fifth century were called Suevi and perhaps also Rugi.
The transformation of the Quadi is only one example out of many of one crucial feature of all these tribal communities: they were very changeable. For a start, none of them were united ethnic groups; they all consisted of smaller tribes, each with a separate leader (as with the half a dozen Gothic groups, even though the Goths were among the most coherently organized of the Germanic peoples). Historians have indeed sometimes argued that some Germanic tribes had no permanent leadership at all, only generals in times of war. This latter pattern seems less likely (if only because war was pretty common); more plausible is that war encouraged the temporary development of alliances or confederations of separate tiny tribes, each with its own permanent leader, but choosing a temporary leader for that confederation. This at least fits the Alemans of the 350s-370s described by Ammianus, whose seven kings (reges) united under Chnodomar and his nephew Serapio to fight Julian in 357, but the latter were also flanked by ten lesser leaders, regales, and aristocrats as well, ‘from various nationes’. Did all of these nationes even think of themselves as ‘Aleman’, or is this, like ‘German’, just a Roman term for a much more inchoate reality? We cannot be sure, but, if the latter was so, this would at least explain the frequent name changes of the major peoples the Romans described. The problem is, of course, that the Romans wrote our only written sources (the only certainly Gothic source is Ulfilas’ Gothic translation of the New Testament, although the Passion of Saba, about an early Christian martyr in the Gothic lands who died in 372, may have been written by a Goth too). Roman ethnography was never reliable, and was usually highly moralized, with ‘barbarians’, naturally inferior but often noble in their savagery, acting as a mirror for the faults of the Romans themselves. It is highly unlikely that even Ammianus, although present on the Rhine in 357, had more than second-hand information about Aleman society and practices, and other observers were further removed still.
Certain things can nevertheless be said about the ‘barbarian’ groups, partly thanks to written sources, partly thanks to archaeology. The northern and southern neighbours of Rome were all mixed-farming peasant societies (except for the Sahara nomads), living for the most part in villages, with élites generally living side by side with cultivators. They were settled and stable societies; they did not normally move about. They seem, however, in all cases to be better organized by the fourth century than they had been in the early empire. The archaeology shows the slow development of material cultural differences between regions (unfortunately, we have no way of knowing if these mapped onto the ethnic distinctions between Franks, Alemans, Goths, etc., and this is in my view unlikely), and, most important, increasing concentrations of wealth: the rich in the Germanic world, and we can add the Berber world as well, were becoming richer, thus presumably showing that power was slowly becoming more stable too. This was largely the simple result of contact with the Roman empire, which was vastly more wealthy and powerful than any ‘barbarian’ group. A substantial proportion of the artefacts in rich graves beyond the frontier in the fourth century are of Roman manufacture, as far north as Denmark. The Romans traded beyond the frontiers; they also employed ‘barbarians’ as paid soldiers, in every century. As the ‘barbarians’ became better organized, they also became more dangerous, and the Romans had to defend themselves more carefully against them. A long frontier region developed on the northern boundaries of the empire, in which militarization was capillary, affecting much wider strata of society than was the case elsewhere: northern Gaul and the Balkans were the largest such frontier regions, but there were smaller ones elsewhere too. As ‘bar barians’ were used in the army and often settled in the empire, at the same time as hierarchies developed under Roman influence beyond the frontiers, society on each side of the frontier slowly became more similar: there may not have been so very much difference on one level between Valentinian, himself from the Pannonian frontier in modern Hungary, and the leaders of the neighbouring Quadi whose bold reply killed him.
This type of observation has been used by some recent historians as the basis for an argument that nothing really changed when the ‘barbarians’ entered the Roman empire in the fifth century and replaced its western half with their own kingdoms. Emperors had long been drawn largely from military families on the frontier; the successor states had kings of a similar type, only from just beyond the frontier. This is a better argument than the traditional one that waves of migrating Germans overbore the weakened (because barbarized) Roman army and state; but it does go too far, all the same. There was a major political difference between each side of the frontier: on one side Romans ruled, on the other they did not. Julian and Valentinian could attack Alemans and Quadi precisely because they were not under Roman rule, and the latter saw themselves as structurally different from Romans, something that did not change when they invaded. Conversely, the soldiers of ‘bar barian’ origin largely deracinated themselves when they joined the army. Take Silvanus, a Frank by origin according to Ammianus, who was a Roman general in the 350s, as his father had been. Silvanus was falsely accused of treason in a piece of palace intrigue in 355, when based at Cologne on the Rhine frontier. He wondered what to do. Should he flee to the neighbouring Franks, his kin? He was dissuaded from this, on the grounds that the Franks would kill or betray him; he claimed the empire instead, as army leaders had often done in the past. This failed, and Ammianus was himself instrumental in having him killed. It would have been easy for Ammianus to depict Silvanus as an untrustworthy and perhaps savage outsider (he does so on other occasions, as with the Romanized Berber aristocrat Firmus, who becomes ‘barbaric’ when he revolts in 373). But Ammianus was instead sympathetic to Silvanus’ plight, and paints him simply as a Roman soldier, and as both politically and culturally separate from the Franks beyond the Rhine; Silvanus’ army training had seen to that. The major military politicians of ‘barbarian’ extraction who were important in late fourth-century politics, such as the Frank Arbogast (d. 394) and the half-Vandal Stilicho (d. 408), both of whom were de-facto heads of state, were similar: they were career soldiers, and operated in an entirely Roman political arena. This was normal in fourth-century politics. It was the politics of the fifth century, when some ‘barbarian’ military leaders fought for Rome at the head of substantial bodies of troops from their own communities, and who called themselves Goths or Franks rather than Romans, that was often different.
In the 370s the Huns appeared in the East, a nomadic people from central Asia. Ammianus depicts them in very hostile and impossibly schematic terms, as hardly human, eating raw flesh, never entering houses, living on horseback, and without rulers: the classic uncivilized ‘others’. They were good fighters, all the same. They may not have been a single political group in the 370s (although they became one, for a generation under Attila, between the 430s and 454). But they destroyed the rule of at least one of the Gothic tribes, Ermenric’s Greuthungi, in or before 375, and menaced others. As nomads, they were as alien to the Goths as to the Romans. As a result, the majority of another Gothic tribe, the Tervingi, sought entry to the Roman empire in 376, and so did other sections of the Goths, although others stayed north of the Danube and slowly accepted Hunnic hegemony. ‘Barbarian’ tribes had invaded the empire often enough in the preceding two centuries; usually they ravaged sections of one of the military zones, the Balkans and northern Gaul, and were then defeated and enslaved, absorbed or driven back. Submissive requests for entry were rarer, and the Romans, including the eastern emperor Valens (364-78), Valentinian’s brother, were not sure how to handle this. They accepted the request, and the Goths, immigrating into the eastern Balkans, became in the following decades ‘Arian’ Christians, the variant Christianity of both their early missionary Ulfilas and, to a lesser extent, Valens himself. But Roman suspicion remained. The Goths were deprived of supplies, and soon revolted under their leader Fritigern; and Valens, underestimating them, was defeated and killed at Adrianople (modern Edirne in European Turkey) in 378. The Goths did not manage to build on this, for they were too few and in a strategically weak position, and they accepted peace in 382. By 394 they were fighting in the east Roman army, against a western usurper put up by Arbogast. But they did not become ‘Roman’, and remained as a separate ethnic grouping, the first group inside the empire to do so.
This sort of interpenetration became steadily more common, in particular after a larger number of ‘barbarian’ groups invaded the empire in 405-6, probably as a result of the steady development of Hunnic power. This did not by any means have to be inimical to Roman power structures and in the East was not; but political errors in handling ‘barbarians’, like those of Valens, continued after his death, and these would be more problematic. We shall see in Chapter 4 how strategic ineptness in the face of a steadily changing political situation in the end helped to sink the western half of the empire. But the stability discussed in this chapter was not illusory, all the same, and many of the political and social patterns described here lasted long into the early medieval world.