The Romans’ own perception of the first-century Republic was that a stable society, whose different elements had been established by legendary figures like Romulus and Servius, was now being undermined and transformed by an array of malign forces: corruption, luxury, ambition, faithlessness, violence. As Dionysius of Halicarnassus argued, summarizing the institution of patronage for a Greek audience, ‘‘the practices instituted by Romulus established so great a consensus amongst the People of Rome that there was no bloodshed or murder amongst them for six hundred and thirty years’’ (Ant. Rom. 2.11.2). The modern perspective, taking the long view over decades or centuries, is that societies are never static. ‘‘Social structure’’ shapes and influences the behavior of individuals, rather than determining their actions, and the cumulative effect of individual social behavior can transform society over time. Romans’ pursuit of their own interests brought about movement between social groups, changes in the ways that the boundaries of those groups were defined (the development of new criteria to identify the elite, for example) and the rise of new groups (slaves, freedmen, and the urban poor). External factors might either change an individual’s behavior or give them more power to pursue their interests; thus the influx of wealth and slaves from Rome’s conquests prompted far-reaching changes in the Italian countryside, with consequences for society as a whole (see also Chapters 27 and 28).27 The changes of the final century of the Republic were particularly dramatic, at least in the political sphere, but it would be misleading to assume that Roman society was stable and unchanging until the Gracchi undermined its foundations.
There were, after all, other kinds of structures, besides the framework of social groups and relationships, that influenced and constrained social behavior by setting the ‘‘limits of the possible.’’ One such set of limits on human action is established by the environment, the combination of climate, geography, and the distribution of natural resources that favored certain sorts of agriculture (and thus influenced the class structure of antiquity) and encouraged the development of particular patterns of settlement and communication (see Chapter 5). 8 Environmental change in the preindustrial era was generally slow, so that its effects were scarcely perceptible at the human level; far more significant in the medium term were the limits set by human reproduction. The size of a population in relation to the availability of resources, its rates of mortality and fertility, its age structure and sex ratio, and the average life expectancy, all have far-reaching implications for the fate of individuals and the society in which they live.
It should be obvious that if we have no conception of the numbers of peoples about whom we write and read we cannot envisage them in their concrete reality. What does a statement about the Romans mean if we do not know roughly how many Romans there were? Without such knowledge even politics and war cannot be understood. For instance, a description of Roman political institutions in the third century B. C. could only be misleading if we did not know that the citizen body was so numerous and scattered that in the absence of the representative principle the democratic features which they seem to manifest were bound to be illusory in practice.29
Clearly this affects more than politics; the size of the population has implications for military activity (the size of the pool of potential recruits), economic structures (the availability of labor, the degree of poverty and inequality, the level of malnutrition) and social relations (the relative numbers of slaves and masters, elite and masses). In turn, political, military, economic, and social behavior has unforeseen consequences for the population, as the war effort reduces the pool of men who will produce the next generation, and poverty and malnutrition reduce the fertility of the mass of citizens.
Demography is not only, or even mainly, concerned with population size; such absolute figures can be simplistic and misleading, since populations never remain static. More often, the focus is on demographic structures and processes, the ways in which the population changes over time. Study of rates of mortality and fertility can provide vital insights into the workings of the family: we can see the complex interrelation between the average age at marriage of men and women, the average numbers of children (both in total, and those who survive infancy), and the likelihood of the family reproducing itself in the next generation (see also Chapter 15). Demography affects our view of both the frequency and the motives for infanticide, as well as the traditions of patria potestas and adoption.30 Thinking beyond the family, it raises questions about the age structure of society. Recalling Cicero’s stress on men ‘‘of every order, of all origins and indeed of all ages’’ (Cat. 4.14), should we think of Rome as divided between age groups with different interests and patterns of social behavior, with the young sent off to fight wars on behalf of the old men who held power (since eligibility for different magistracies depended, at least in theory, on age)? Certainly rates of birth and death might affect relations between different social groups; thus the oldest senatorial families proved unable to reproduce at a rate which would replace their numbers in the next generation, and so always had to draw new recruits from the wider equestrian order.31
The importance of demography, with its implications for every aspect of ancient history, has been increasingly acknowledged, especially since the mid-1990s.3 However, there is a fundamental problem; we do not have adequate evidence for a proper study of ancient demographic structures, either in quantity or quality. There is no ancient equivalent of the parish register of births, marriages, and deaths; under the Republic, the census, because its original purpose was to police the citizen body and to establish Rome’s military strength, recorded only absolute numbers, not the age structure of the population, and ignored women, children, slaves, and other noncitizens. Historians have therefore been faced with a choice: to make the best of such ‘‘proxy’’ data as is available - tombstones, for example, and tax registers from Egypt - or to focus on comparative evidence, on the assumption that the Roman population cannot have been too dissimilar in its structure to the populations of other preindustrial societies about which we know rather more.
These two approaches are frequently in direct conflict. Demography suggests questions that might be asked of evidence that was never intended to be used for demographic purposes. For example, Roman epitaphs often include the age of the deceased; given a sufficiently large sample, information about what proportion of the population is dying at different ages can be extrapolated to produce a model ofthe age structure of society and its average life expectancy. However, if this data is considered in relation to our knowledge of the demography of other societies, it seems clearly flawed. Epitaphs tend to record ages in multiples of five and ten (a phenomenon known as ‘‘agerounding’’), and many of the ages recorded seem implausibly high for a pre-industrial society (but perfectly explicable if one recalls that old age was greatly respected in antiquity, and the lack of records meant that many people might be quite uncertain of their real age). Above all, there are far too few infant burials, whereas in a typical premodern society with life expectancy at birth of 25 years (e0=25), mortality in the first year oflife might be as high as 30 percent, with 50 percent of a ‘‘birth cohort’’ dying by the age of 10. Either the Romans were much healthier than we thought, or there is a problem with the evidence; most likely another manifestation of the ‘‘epigraphic habit,’’ such that inscriptions provide information not about the reality of demographic structures but about the attitudes and assumptions of those who chose to spend money on commemoration - in this case, evidence that many infant deaths were not commemorated in the way that adult deaths were.33
Similar objections can be made to other attempts at reconstructing Roman demography from ancient evidence; if it is assumed that the appropriate comparisons for antiquity are underdeveloped societies with high levels of mortality and low life-expectancy, the picture of Roman demographic structures offered by the sources cannot possibly be correct, and so the evidence must be at fault. However, it is clearly a radical step to reject all ancient sources out of hand; it can be argued instead that a different comparison should be made, and that, if the ancient evidence suggests that average life expectancy at birth was 35 rather than 25, Rome should be thought of as having a more modern demographic structure, with lower mortality rates. This fits with the beliefofsome historians that Rome was a more sophisticated and developed society than the labels ‘‘pre-modern’’ or ‘‘pre-industrial’’ would suggest; others argue that ‘‘the burden of proof is firmly on those who wish to assert that the Roman population in general had a lower mortality than other pre-industrial populations with similar technical achievements or towns; they must show that there were present in the Roman Empire factors which could have led to a general diminution of mortality.’’
It should also be noted that there are problems with the ‘‘model life tables’’ from which historians draw their impressions of pre-industrial population structure.35 Life tables offer models of the age structures of different populations, showing the complex ways in which life expectancy and rates of mortality and fertility interact; however, they are based not on actual pre-industrial populations but on mathematical extrapolation from modern populations, and they deliberately exclude the distorting effects of diseases such as malaria on age structure. They are idealized models that offer a sense of how the ancient world might have been, not how it must have been; they need to be employed sensitively, reintroducing factors such as the effects of disease and of ‘‘culturally specific’’ behavior like infanticide and the limited reproductive opportunities available to slaves. Above all, the historian has to decide which model to employ, based on prior assumptions about the nature of antiquity. It is generally, but not universally, accepted that Rome is best understood as a ‘‘high-pressure’’ pre-industrial society, with high rates of mortality and fertility and an average expectation of life at birth of 25.
There are still more problems with the reconstruction of absolute population numbers for any period. We have no idea of the number of slaves in Italy, for example, beyond the general impression that it increased significantly over the last two centuries of the Republic (see Chapter 28). It was in no one’s interest to attempt to count them; our evidence is limited to occasional impressionistic comments in the literary sources, such as the remark of the medical writer Galen that there was one slave to every two free men in the city of Pergamum (5.49K). It is at best possible to exclude some of the wilder estimates through consideration of the capacity of Italy to support a large population (though that of course depends on the figure assumed for the free population) and by considering the rate of imports necessary to sustain a particular number of slaves, given that not all slaves were able to reproduce and mortality rates were high.3 We are ignorant of the numbers of freedmen, besides noting that they dominate the body of inscriptions from the city of Rome. Estimates for the urban population are based on the city’s grain supply and the figures for recipients of the corn dole; even more speculative estimates for the populations of other towns of Italy are offered on the basis of such evidence as the size of the built-up area, the length of the walls, and the number of citizens who benefited from the generosity of local notables.37 Estimates of the population of Pompeii, the city for which the most archaeological evidence is available, vary from 10,000-12,000 to 25,000-30,000, depending largely on the historian’s prior assumptions about the city’s society and economy.
The fiercest debate has been over the size of the total population of Italy, partly because there is some, apparently useable, evidence - the figures produced by the Roman census, especially those collected under Augustus - and partly because demographic change seems to be central to the whole process of the transformation of the rural economy in the late Republic (see Chapters 27 and 28).39 The Augustan census in 28 recorded a figure of just over 4 million; the census under the Republic had always counted only adult males, and so this figure suggests that the total free population, including women and children, was about 10 million. This can be compared with the population figure of 4.5 million that has been estimated for 225, on the basis of Polybius’ account of the military strength of Rome and its allies at that date. Over two centuries, therefore, the free population of Italy more than doubled. However, the idea of a population expansion seems incompatible with historical accounts of the crisis of the Italian peasantry and the depopulation of the countryside in this period, and so this interpretation of the Augustan total has often been rejected. If the Augustan census is interpreted instead as having included all citizens, not just adult males, it indicates a total population of just 4 million, a slight decline since 225 which can be attributed to the effects of constant warfare and the displacement of peasants to make way for slave labor into the countryside.4 This ‘‘low’’ estimate of the Augustan population, and the ‘‘decline’’ theory of Italian demography, has been dominant since Brunt’s 1971 study of the Roman population, and underpins most historical accounts of the period; if the ‘‘high’’ figure were accepted, history - political history, not just economic and social - would have to be rewritten.41
Neither interpretation can be proved beyond doubt on the basis of the literary evidence, and so the proponents of each view have turned to comparative arguments. Once again, however, this depends on prior assumptions about the ancient world, which determine what is chosen as the most appropriate comparison. The ‘‘high’’ figure implies that Augustan Italy was more densely populated than nineteenth-century Italy: is this grounds for rejecting it, since pre-industrial technology was inadequate to support so many people, or grounds for taking a more positive view of the efficiency of Italian agriculture? The rate of population increase seems implausibly high, especially taking into account the ‘‘population sink’’ effects of high mortality rates in the city of Rome, if Roman Italy is compared to early modern Europe; but comparable rates of increase are known from nineteenth-century America. In other words, comparative evidence and modern scientific knowledge can suggest what might have been possible, but they cannot say how things must have been. The latest twist in the debate is the suggestion that the population figure for 225, which all participants have hitherto taken for granted, should be revised downward, giving a better fit with evidence that implies a rising population (such as growing competition for land) without implying that Italy was grossly overpopulated by the time of Augustus (see Chapter 27).
‘‘Scarce source references are interpreted through the lens of conflicting samples of comparative data.’’42 This might indeed be said of all approaches to the study of Roman social structure, not just the demographic perspective. Roman social structure looks significantly different depending on one’s choice of analytical categories; and the historian’s preference for particular concepts and particular historical comparisons often depends on prior assumptions, not just about Rome but about ‘‘society’’ in general. From almost every perspective, the late Republic is a period of major changes (not necessarily to be considered as ‘‘decline’’) in the composition of social groups and the dynamics of social relationships; but the nature of those changes, and the best tools for understanding them, remain matters of fierce debate.
World History
