He used to play jokes on his slaves, even ordering them to bring him a thousand pounds of cobwebs and offering them a prize; and it is said that he collected ten thousand pounds’ worth, and then remarked that one could realize from that how great a city Rome was.
Historia Augusta, Elagabalus 26
R
ome was exceptionally large for a pre-modern, pre-industrial city. The widely accepted estimate that its population at the time of Augustus was around 800,000–1,000,000 implies that it had only a handful of rivals for size, all in China, before the nineteenth century. Since 1800, dramatic improvements in technology and far-reaching changes in the organization of society and economy as a whole have meant that cities of several million are scarcely rare across the globe. For such a city to develop and thrive under far more restrictive conditions was a remarkable achievement, a reflection above all – as ancient commentators recognized – of the power and wealth of the Roman empire.The size of Rome is not merely a symbol of greatness or a trump card in inter-cultural rivalry; it is directly significant for understanding the dynamics of Roman society. The fact that the city could grow so large, despite all the practical, ecological and technological imped-iments, provides a sense of the capacities of the Roman economy;
scarcely comparable with modern economic performance, of course, but to be ranked highly among pre-industrial societies. Moreover, the growth of the city was itself one of the major sources of change, as its demands for people and raw materials transformed economic and social structures throughout its Mediterranean-wide hinterland. Rome’s role as the capital of the empire constituted the main basis for its growth, as the ruling elite invested the spoils of imperialism in the urban envi-ronment and migrants flocked to service their needs and gain a share
of the empire’s wealth; but the elite made this investment precisely because of the importance of the city in establishing and maintaining their power – Rome’s greatness was itself a crucial element of the ide-ology that sustained Roman rule. Further, the size of the city shaped the nature of urban life, politics and society. As will be discussed in later chapters, the concentration of a large population in a limited area gave rise to a distinctive demographic regime, leaving the city heavily dependent on inward migration just to maintain its size and so creating a society dominated by change, fluidity, insecurity and the interaction of many different cultures. In contrast to the relatively face-to-face social relations of the countryside or the small town, social interaction in the metropolis was largely impersonal and anonymous, with a gaping divide between the political elite and the masses. The maintenance of social order depended not on personal relationships between patron and client but on large-scale rituals, ceremonies and grand public occasions, and on the power of public architecture and imagery, to maintain the symbolic order as a means of binding society together.
It is clear from the ancient sources, from Latin poets to Greek orators, that Rome was far larger than any other city in the ancient Mediterranean; as Pliny the Elder concluded from his brief comments on the physical size of the city, ‘a very fair estimate would be formed that would bring us to admit that there has been no city in the whole world that could be compared to Rome in magnitude’ (Historia Nat-uralis 3.5.67). For many historical purposes, that general impression of exceptional size may be sufficient. However, in order to grasp prop-erly the crucial issues of Rome’s demographic regime and its demand for food and other resources, we need a more specific idea of the size of its population. This is particularly important if we wish to com-pare Rome with other great pre-industrial cities, in order to gain a broader understanding of the nature and impact of urbanization. In some historical contexts, such as western Europe in the Middle Ages, a ‘primate’ city might eclipse its rivals and dominate its society with a population of a few tens of thousands; the fact that it was, from a comparative perspective, relatively small might not lessen its political and cultural significance in that context, but it suggests that the city’s potential impact on economy and society might be limited. If, on the other hand, Rome did indeed reach 1 million or more inhabitants, as many historians have argued, then its impact on the empire must have been significantly greater.
The problem is that there is no direct evidence for the size of the urban population; the figure of a million people is a widely accepted
P o p u l a t i o n s i z e a n d s o c i a l s t r u c t u r e
hypothesis, not a well-established fact. Historians find themselves in a similar position to the emperor Elagabalus, seeking to extrapolate from ‘proxy’ evidence, and returning time and again to debate the possible interpretations of the few figures (often of doubtful reliability) and measurable attributes at their disposal. One consequence has been the enormous range of estimates for the size of Rome developed over the last few centuries, from a mere 150,000 (still impressive compared with the cities of medieval Europe, admittedly) to a truly extraordinary 4million, three or four times the size of any city known before the nine-teenth century. Ten thousand pounds’ worth of cobwebs is undoubtedly a very large quantity, the harvest of a great many rooms and buildings;
but in the absence of reliable information about, for example, the typical number of spiders in each apartment block and their average produc-tivity, deriving a sense of the size of the population from such data will always be a matter of fierce debate.
T h e p h y s i c a l c i t y
Ancient sources rarely make direct reference to the size of the urban population. One exception was Cassiodorus, looking back at the city’s past greatness from the sixth century:
The vast numbers of the Roman people in the past are evidenced by the extensive provinces from which their food supply was drawn, as well as by the wide circuit of their walls, the massive structure of their amphitheatre, the marvellous size of their public baths, and the enormous multitude of mills, which could only have been made for use, not ornament.
(Varia 11.39) Earlier commentators eulogized Rome’s greatness in more general terms, rather than focusing on the size of the population, but they considered similar themes: the city’s physical extent, both in area and height, the magnificence and size of its principal monuments, and the sheer number of other, more mundane buildings. The Greek orator Aelius Aristides claimed that it might be considered presumptuous to attempt to praise the city; Rome was too great to be taken in by the eyes – only all-seeing Argus could adequately survey its extent – let alone to be encompassed in words.
Like the snow, she covers mountain peaks, she covers the land intervening, and she goes down to the sea . . . And indeed she is poured out, not just over the level ground, but in a manner with which the simile cannot begin to keep pace she rises great distances into the air, so that her height is not to be compared to a covering of snow but to the peaks themselves.
(Oration 26.6–8) Other ancient writers celebrated the size and wealth of the city through the enumeration of its public buildings, especially the ‘veritable rivers’ of its aqueducts. Rome simply contained more of everything. According to one of the texts collected in the Talmud, ‘the great city of Rome has 365streets, and in each street there are 365 palaces. Each palace has 365 stories, and each story contains enough food to feed the whole world’
(Pesahim 118b). Fourth-century descriptions of the city, known as the Notitia and the Curiosum or the Regionary Catalogues, set out in great detail the different buildings to be found in each of the fourteen regions of the city, not only the temples and monuments known by name but the total numbers of bath houses, warehouses, bakeries, fountains, domus and insulae. These documents have been described as ‘panegyrics in statistics’, designed to impress the reader through sheer weight of numbers.1
It is understandable, therefore, that some historians have sought to estimate the population of Rome on the basis of evidence for its physical size and the number of residential buildings. The Regionary Catalogues, which differ slightly in their figures, give a total of around 1,800 domus, the familiar town houses of the Roman elite, and around 46,500 insulae, clearly defined in legal sources as self-contained, inde-pendent buildings divided into separate apartments or coenacula and well known as a result of the excavations at Ostia.2 If one assumes that the typical elite familia contained around thirty persons, including slaves (though the richest senators might have several hundred of those), and that the typical insula had three floors, each containing a number of apartments inhabited by individual families, giving perhaps thirty to
1 The text of these documents is in Nordh (1949); the best discussion of their date and purpose, and the problems involved in using them to reconstruct the urban population, as discussed in the following paragraphs, is Hermansen 1978.
2 See Chapter 9 for a more detailed discussion of the nature of Roman residences.
P o p u l a t i o n s i z e a n d s o c i a l s t r u c t u r e
thirty-six persons in each building, then the population of Rome in the fourth century was around 1.5–1.7 million, along with an unknown number of soldiers, the slaves and freedmen of the imperial household and other public buildings, and the homeless, perhaps another 80,000–
100,000 individuals in total. Different estimates for average household size will push the figure up or down, but the sheer number of insulae means that the total is invariably well over a million.
The obvious problem with this calculation is highlighted when it is related to the physical area of Rome, the 13.86 km2that lay within the third-century Aurelian Wall, which may be rounded up to 15 km2 to take account of the fact that the fourteen regions established by Augustus covered a larger area. If the population of the city was at least 1.5 million, this implies an average population density of 100,000 persons per km2, a figure which is higher than any known historical example besides a few specific localities within proverbially crowded modern cities like Mumbai or Hong Kong. It seems scarcely conceivable that this could be a realistic average density for the entire city, given the amount of land taken up by public monuments, some of them – like the Circus Maximus – covering very large areas. Further, if only half of the total area of Rome was occupied by residential buildings, and the average domus is estimated to have occupied 600 m2, then the ground area of the average insula was barely 150 m2, hardly compatible with the remains of the insulae known from Ostia. This raises serious doubts about the credibility of the figure of 46,500 insulae.
One way of dealing with this problem and rescuing the credibility of the source is to conclude that the insulae listed in the Regionary Cat-alogues are not separate buildings, as the term is used in the legal texts, but the different floors of those buildings. If only ten to twelve people lived in each of these ‘insulae’, the fourth-century population would be a much more reasonable 580,000–680,000. A more extreme inter-pretation, put forward on the basis that the contemporaneous Notitia of Constantinople does not list insulae at all, is that the entire pop-ulation of Rome was accommodated within the 1,790 domus, while the numbers given for insulae represent the number of door open-ings of these residences; that bropen-ings the population down to barely 150,000 all told, with a population density of only 10,000 per km2. However, there is no evidence at all to support the interpretation of insula as either ‘floor’ or ‘apartment’, or as anything other than a self-contained block of flats, even if not necessarily identical to those found in Ostia. The absence of insulae from the Regionary Catalogue for
Constantinople may be explained by changes in residential fashion in the fourth century, but is more likely to be a simple omission; it certainly does not support the idea that the insulae of Rome were merely door openings.
The final attempt at saving the testimony of the Regionary Cata-logues has been to recalculate how much of Rome’s total area may have been taken up by monuments, producing a higher estimate of the space occupied by residential buildings; this in turn allows a new calculation of the average size of each insula, which then implies a different estimate for the inhabitants of each building and a new population total – around 1.2–1.3 million. The problem is that this new calculation works only by ignoring the detail of the Regionary Catalogues’ description of each region; it may be possible to accommodate 45,000 insulae within the total area of the city, but it is impossible to accommodate the numbers specified by the texts as belonging to particular regions, especially in the highly monumentalized centre of the city. The only possible con-clusion is that the numbers in the text are grossly inflated if not simply corrupt; the purpose of the Catalogues is not to provide an accu-rate description of the city but to represent and celebaccu-rate its greatness through the power of number. The statements about the numbers of residential buildings offer no secure basis for estimating the population;
indeed, the way in which some historians have desperately sought to redefine the insula in order to find space for 45,000 of them within the city shows that the argument has shifted from attempting to estimate the population to attempting simply to defend the credibility of the figures.
If the figures for the number of insulae are set aside, there remains the possibility of deriving a more plausible estimate from the evidence of archaeology and the Forma Urbis Romae, the Severan marble plan of the city.3 One reconstruction of the buildings identified as residen-tial on fragments of the Forma Urbis suggests that the average domus covered 675 m2 and the average insula 250 m2. If half the city’s built-up area was occbuilt-upied by housing, 1,790 domus (the figure from the Regionary Catalogues, though why this should be considered more reliable than that for insulae is unclear) occupied 1.2 km2 and the insu-lae covered the remaining 6.25 km2; that suggests a total of about 25,000 insulae at 250 m2each, and a total population of around 880,000 to one million. This implies an average population density of about
3 Developed by Hermansen 1978 and, drawing on more recent archaeological work at Pompeii and Ostia, Storey 1997. On the Severan marble plan, cf. Chapter 10.
P o p u l a t i o n s i z e a n d s o c i a l s t r u c t u r e
60,000–70,000 people per km2, still considerably higher than the aver-age for cities like Mumbai or Hong Kong but considered by many to be credible. Studies of housing in Pompeii and Ostia, however, suggest that the figure is much too high; house-by-house counts from these sites, with computer-aided reconstructions of the unexcavated areas, produce estimates of population densities of 16,615 persons per km2 and 31,700 persons per km2 respectively, figures which imply much smaller household sizes and a total population for Rome of 250,000 or 475,000.
These population totals for Pompeii and Ostia tally well with recent estimates for those cities derived by other means, and that seems to confirm the validity of the figures used in the calculation for house-hold size (3 to 6 individuals for each apartment, 13 to 17 for each domus).
The crucial question is whether such household sizes were also typical of Rome, which was manifestly a very different sort of city from either Pompeii or Ostia, or whether a higher multiplier, yielding a higher population total and a higher density, should be used. The same issue is raised by cross-cultural comparisons; the rejection of estimates of a million on the grounds that they imply higher population densities than those known from other cities begs the question as to whether Rome should be assumed to be comparable to those other cities, or whether it might have been very different. The risk of circularity is clear; in order to obtain a sense of the overall population size to inform their understanding of the city and urban life, historians evaluate different estimates according to their preconceptions of what the city was like.
Thus, those who see Rome as an over-crowded slum, the Calcutta of the ancient Mediterranean, are happy to accept exceptionally high estimates of average population density, whereas those who see it as a more pleasant and civilized place seek to reinterpret the meaning of insula or to draw on what they see as more appropriate comparative evidence.4Different interpretations rest on quite different assumptions, and so cannot be directly compared.
Consideration of the physical city does at least establish plausible maxima and minima for the urban population and exclude the most extreme estimates. An average density across the whole 15 km2of more than 80,000 persons per km2, as high a figure as found in some of the most densely settled districts of more modern cities, is entirely implausible; older estimates of a total urban population of 1.5–2 million
4 Compare Laurence 1997 and Morley 2005 on different traditions of interpretation of the healthfulness or otherwise of the Roman city.
can thus be rejected, and a figure of just over a million taken as a likely maximum. At the other end of the scale, it is difficult to imagine that Rome was less crowded than Pompeii, so it must have contained at least 250,000 people; given that we are dealing with the capital of the empire rather than a regional centre, the estimated population density for Ostia might seem a more likely comparison, implying a total of just under half a million inhabitants. We may then consider reasons why Rome may have been more crowded than Ostia: a greater capacity for drawing in migrants and keeping them, despite crowded living conditions, combined with a reduced scope for the city to spread further outwards because of the ring of gardens and the estates of the rich around it, so that a growing population could be accommodated only by increasing building height and population density. More importantly, we need to compare these arguments with the estimates obtained by looking at alternative sources of evidence.
F o o d s u p p l y a n d d i s t r i b u t i o n
The alternative set of proxy data for urban population relates to the city’s imports of grain from the provinces and the distributions to recipients of the annona, the corn dole.5 Several ancient sources make reference to the size of Rome’s imports from Africa and Egypt or to the volume of its total consumption; unfortunately, the figures are not compatible with one another. Two schools of thought have developed; the com-bination of two of the four relevant sources produces an estimate of 60 million modii of imported grain per year, the combination of the other two suggests only 27 million. Quite apart from the risks of relying
The alternative set of proxy data for urban population relates to the city’s imports of grain from the provinces and the distributions to recipients of the annona, the corn dole.5 Several ancient sources make reference to the size of Rome’s imports from Africa and Egypt or to the volume of its total consumption; unfortunately, the figures are not compatible with one another. Two schools of thought have developed; the com-bination of two of the four relevant sources produces an estimate of 60 million modii of imported grain per year, the combination of the other two suggests only 27 million. Quite apart from the risks of relying