The Fall of the Roman Empire: A New History

by Peter Heather

  The evidence comes mostly from written sources. To start with, the annual volume of inscriptions known from Roman antiquity declined suddenly in the mid-third century to something like one-fifth of previous levels. Since chances of survival remained pretty constant, this massive fall-off was naturally taken as an indicator that landowners, the social group generally responsible for commissioning these largely private inscriptions, had suddenly found themselves short of funds. A study of the chronology also led the heavier tax burden imposed by the late Roman state to be seen as the primary cause, since the decline coincided with the tax hikes that were necessary to fight off the increased Persian threat. Such views were reinforced by other sources documenting another well known fourth-century phenomenon, commonly known as the ‘flight of the curials’. Curials (or decurions) were the landowners of sufficient wealth to get a seat on their town councils (Latin, curiae). They were the descendants of the men who had built the Roman towns, bought into classical ideologies of self-government, learned Latin, and generally benefited from Latin rights and Roman citizenship in the early imperial period. In the fourth century, these descendants became increasingly unwilling to serve on the town councils their ancestors had established. Some of the sources preserve complaints about the costs involved in being a councillor, others about the administrative burden imposed upon the curials by the Roman state. It has long been part of the orthodoxy of Roman collapse, therefore, that the old landowning classes of the Empire were overburdened into oblivion.14

  Other fourth-century legal texts refer to a previously unknown phenomenon, the ‘deserted lands’ (agri deserti). Most of these texts are very general, giving no indication of the amounts of land that might be involved, but one law, of AD 422, referring to North Africa, indicates that a staggering 3,000 square miles fell into this category in that region alone. A further run of late Roman legislation also attempted to tie certain categories of tenant farmers (coloni) to their existing estates, to prevent them moving. It was easy, in fact irresistible, to weave these separate phenomena into a narrative of cause and effect, whereby the late Empire’s punitive tax regime made it uneconomic to farm all the land that had previously been under cultivation. This was said to have generated large-scale abandonment – hence the agri deserti – as well as governmental intervention to try to prevent this very abandoning of the lands that the new tax burden had made uneconomic. Stripped of a larger portion of their production, the peasantry could not maintain their numbers over the generations, which further lowered output.15

  Into this happy consensus a large bomb was lobbed, towards the end of the 1950s, by a French archaeologist named Georges Tchalenko. As with many revolutionary moments, it took a long time for bystanders to realize that they had witnessed something earth-shattering, but this bomb set off a chain of detonations. Tchalenko had spent much of the 1940s and 1950s roaming the limestone hills in what is now a fairly obscure (and relatively peaceful) corner of the Middle East. In antiquity, these hills belonged to the rural hinterland of one of the great imperial capitals, Antioch: Antakya in modern Turkey. (The hills, by a quirk of fate, have ended up over the border in northern Syria.) In his explorations, Tchalenko came across the remains of a dense spread of villages, sturdily constructed from limestone blocks, which had been abandoned in the eighth to ninth centuries after the Arab conquest of the region.

  The villages showed that these hills had once been the home of a flourishing rural population, which could afford not only to build excellent houses, but to endow their villages with sizeable public buildings. This ancient population was much denser than anything the region has supported at any point since, and it clearly made its living from agriculture; Tchalenko believed it produced olive oil commercially. The really revolutionary bit was Tchalenko’s discovery that prosperity first hit the region in the later third and early fourth centuries, then continued into the fifth, sixth and seventh with no sign of decline. At the very moment when the generally accepted model suggested that the late Roman state was taxing the lifeblood out of its farmers, here was hard evidence of a farming region prospering.16

  Further archaeological work, using field surveys, has made it possible to test levels of rural settlement and agricultural activity across a wide geographical spread and at different points in the Roman period. Broadly speaking, these surveys have confirmed that Tchalenko’s Syrian villages were a far from unique example of late Roman rural prosperity. The central provinces of Roman North Africa (in particular Numidia, Byzacena and Proconsularis) saw a similar intensification of rural settlement and production at this time. This has been illuminated by separate surveys in Tunisia and southern Libya, where prosperity did not even begin to fall away until the fifth century. Surveys in Greece have produced a comparable picture. And elsewhere in the Near East, the fourth and fifth centuries have emerged as a period of maximum rural development – not minimum, as the orthodoxy would have led us to expect. Investigations in the Negev Desert region of modern Israel have shown that farming also flourished in this deeply marginal environment under the fourth-century Empire. The pattern is broadly similar in Spain and southern Gaul, while recent re-evaluations of rural settlement in Roman Britain have suggested that its fourth-century population reached levels that would only be seen again in the fourteenth. Argument continues as to what figure to put on this maximum, but that late Roman Britain was remarkably densely populated by ancient and medieval standards is now a given.17 The only areas, in fact, where, in the fourth century, prosperity was not at or close to its maximum for the entire Roman period were Italy and some of the northern European provinces, particularly Gallia Belgica and Germania Inferior on the Rhine frontier. Even here, though, estimates of settlement density have been revised substantially upwards in recent years.

  For the poverty of the latter two northern provinces, the explanation probably lies in third-century disruption. The Rhine frontier region was being heavily raided at the same time as so much energy was being poured into solving the Persian problem, and it may be that rural affluence in parts of the region never recovered. A methodological problem may also provide at least part of the explanation. Roman-period surveys rely on datable finds of commercially produced pottery to identify and date settlements. If a population ceased to import these wares, reverting to undatable locally made ceramics, especially if at the same time they were also building more in wood than in the traditional Roman stone, brick and tile – which surveys also find – then they would have become archaeologically invisible. This was happening in several areas of northern Europe by at least the mid-fifth century, so it is far from impossible that the seeming lack of fourth-century inhabitants in parts of the northern Rhine frontier region was caused not by substantial population decline, but by the first appearance of these new habits. The jury is still out.

  The case of Italy is rather different. As befitted the heartland of a conquest state, Italy was thriving in the early imperial period. Not only did the spoils of conquest flood its territories, but its manufacturers of pottery, wine and other goods sold them throughout the western provinces and dominated the market. Also, Italian agricultural production was untaxed. As the economies of the conquered provinces developed, however, this early domination was curtailed by the development of rival enterprises closer to the centres of consumption and with much lower transport costs. By the fourth century, the process had pretty much run its course; and from Diocletian onwards, Italian agriculture had to pay the same taxes as the rest of the Empire. So the peninsula’s economy was bound to have suffered relative decline in the fourth century, and it is not surprising to find more marginal lands there being taken out of production. But as we have seen, the relative decline of Italy and perhaps also of north-eastern Gaul was more than compensated for by economic success elsewhere. Despite the heavier tax burden, the late Roman countryside was generally booming.18 The revolutionary nature of these findings cannot be overstated.

  Looked at with this in mind, the literary evidence is f
ar from incompatible with the archaeology. The laws forcing labour to stay in one place, for example, would only have been enforceable where rural population levels were relatively high. Otherwise, the general demand for labour would have seen landowners competing with one another for peasants, and being willing to take in each other’s runaways and protect them from the law. More generally, the term ‘deserted lands’ (agri deserti) was coined in the fourth century to describe lands from which no tax was being collected. It carries no necessary implication that land so labelled had ever previously been cultivated, and certainly the large tract of North African territory referred to in the law of 422 consisted mostly of desert and semi-desert hinterland where normal agriculture had always been impossible. Nor is the late Empire’s more demanding tax regime incompatible with a buzzing agricultural economy. Tenant subsistence farmers tend to produce only what they need: enough to provide for themselves and their dependants and to pay any essential additional dues such as rent. Within this context there will often be a certain amount of economic ‘slack’, consisting of extra foodstuffs they could produce but which they choose not to because they can neither store them, nor, thanks to high transport costs, sell them. In this kind of world, taxation – if not imposed at too high a level – can actually increase production: the tax imposed by the state is another due that has to be satisfied, and farmers do sufficient extra work to produce the additional output. Only if taxes are set so high that peasants starve, or the long-term fertility of their lands is impaired, will such dues have a damaging economic effect.

  None of this means that it was fun to be a late Roman peasant. The state imposed heavier demands on him than it had on his ancestors, and he was prevented by law from moving around in search of the best tenancy terms. But there is nothing in the archaeological or written evidence to gainsay the general picture of a late Roman countryside at or near maximum levels of population, production and output.19

  There is, however, no doubt that most cities of the Empire appear to have suffered in one respect. The decline in inscriptions from the mid-third century reflects a fall in the number of new public buildings being commissioned. The only cities that continued to see public building on a large scale were the central and regional capitals of the Roman state. And even here, instead of local grandees endowing their towns with another memorial toilet block (or some such structure) in their own memory, buildings were being erected by state officials using state funds.20 The private funding of public building in one’s hometown belonged to the very early imperial period, when this constituted the prime route to self-promotion. Putting up the right kinds of public building was part of persuading some high official to recommend your hometown to the emperor for the grant of a Roman constitution. Once your town had Latin rights, then financing buildings became a strategy for winning power and influence within it. The towns of the Empire quickly built up endowments of publicly owned land (often from wills), and also acquired the right to levy local taxes and tolls, in itself a substantial annual income whose expenditure was controlled by the town council and particularly by its leading magistrates. Magistrates were voted into office by a town’s free citizens. Competitive local building in this context was all about winning elections and hence controlling the use of local funds.21

  The confiscation by the state of local endowments and taxes in the third century removed most of the fun from local government. By the fourth, there was little point in spending freely to win power in your hometown, if all you then got to do was run errands for central government. By this time, retired members of the expanding class of imperial bureaucrats (honorati) were being given all the interesting and prestigious tasks in local government, including the detailed allocation of their town’s tax burden. Nothing would be more guaranteed to generate invitations to dinner and other small marks of attention than the knowledge that, when the time came, you were going to be in charge of setting the new tax bills. Honorati also got to sit with the provincial governor when he was trying legal cases, and helped him arrive at verdicts. As the many surviving letters to local honorati make clear, this was another moment when great influence could be brought to bear, and, again, it tended to make the honoratus very popular in local society. What happened in the late Empire, in other words, was a major shift in local political power away from town councils to imperial bureaucrats. This did away with the whole point of the local displays of generosity recorded in the early imperial inscriptions.

  The stock image of the late Roman bureaucracy also needs revising. Much of its characterization as an oppressive alien force of ‘idle mouths’ sucking the vitality out of local society can be traced back to a speech of the rhetor Libanius, which catalogued the dubious social origins of some of the leading bureaucrats and senators of mid-fourth-century Constantinople. Three Praetorian Prefects (chief civilian executive officers) of the 350s and early 360s – Domitianus, Helpidius and Taurus – had fathers, Libanius tells us, who personally engaged in manual labour; the father of a fourth, Philippus, made sausages, and the governor of the province of Asia, Dulcitius, was the son of a fuller.22 The image conjured up of a bureaucracy dominated by new men from nowhere is very powerful, but in this speech Libanius had a very particular axe to grind. The Senate of Constantinople had just refused membership to one of his protégés, a certain Thalassius, on the grounds that Thalassius’ father was a ‘tradesman’ (he had owned an arms factory). As a vast body of other evidence makes clear (including endless letters of reference written by Libanius himself), however, the vast majority of the new bureaucrats and senators of the fourth-century Empire were actually drawn from the curial classes, not from further down the social scale. The language of this bureaucracy was the ‘correct’ Latin and Greek espoused by the traditional educational curriculum. This tells us instantly that its members had benefited from a lengthy and expensive private education. The bureaucracy of the late Roman period did not consist of outsiders or parvenus, then, but of town councillors who had renegotiated their position within the changing structures of Empire. Only a small hard-headed inner elite – called principales in Latin – stayed on the councils in order to monopolize the few interesting jobs left.

  Because bureaucratic positions were so attractive, emperors were flooded with requests for appointments. Many of these were granted. Emperors always liked to raise their popularity ratings by appearing generous, and these kinds of grants seemed, individually, pretty harmless. Despite the laws attempting to regulate bureaucratic expansion by forcing ex-town councillors back to their cities, by AD 400 large numbers of wealthy landowners were making the central imperial bureaucracy the main focus of their careers. At this date, the eastern financial office (the largitionales) had a staff of 224 officers, and a waiting list of 610 ready to take their places when they finished their stint. And, because of the delay involved in getting a post under these conditions, parents were appending their children’s names to waiting lists at birth. Thus, far from showing the power of a newly oppressive central state, the rise of the imperial bureaucracy demonstrates the continuation of the same kind of political relationship between centre and locality that we have already observed. Here again, as in the rescript system and in the whole process of Romanization itself, the state certainly started the ball rolling by setting up a new rule book, as it were. But the process was taken over by locals responding to the rule changes and adapting them to their own interests.

  Understanding bureaucratic expansion in this way makes it impossible to see the ‘flight of the curials’ as fundamentally an economic phenomenon, or, at least, as reflecting a decline in the private fortunes of the landowning class. It also takes much of the sting out of the argument that the bureaucracy were so many ‘idle mouths’. It is hard to suppose that these bureaucrats’ ancestors, as local landowners sitting on town councils, had been any less ‘idle’ – if one chooses to see them this way. They had always been essentially a rentier class, overseeing the labour of their peasants rather than engaging in the pri
mary work of agricultural production. But whereas before they had been ‘idling’ on their town councils, now they were idling in the offices of the central Roman state. Their salaries, paid by the state, were also very low. Bureaucratic expansion needed little extra taxation to fund.23 What made the jobs attractive, as we have seen, was the status that accompanied them and the chance to charge fees to those who needed your services.

  While these changes in upper-class career patterns certainly had some economic effects, there is nothing to suggest that upper-class life changed in any fundamental way. Written sources and archaeological excavation both confirm that the late Roman landowning elite, like their forebears, would alternate between their urban houses and their country estates. Fourth-century Antioch, for instance, boasted the hugely wealthy suburb of Daphne, and extensive investigations at the city of Sardis in modern Turkey have uncovered numerous wealthy private houses of the fourth and fifth centuries. There is no reason to suppose, therefore, that luxury urban trades, which depended on landowners coming to ‘town’ from time to time to spend their wealth, will have suffered very much. What may have happened is that the reorientation away from town councils to an imperial bureaucracy meant that larger landowners maintained houses in regional and provincial imperial capitals rather than in their hometowns. This would have increased the tendency – already noticed in patterns of public spending – for capitals to prosper at the expense of lesser towns.24