Keywords

JEL Classifications

Ancient demography covers the population history of early civilizations from the third millennium BCE to the seventh century CE.

Due to the uneven distribution of relevant evidence, scholars have focused primarily on Middle Eastern, Greek and Roman populations. This survey deals primarily with the demography of the Greco-Roman world. Demographic conditions in antiquity are generally only dimly perceptible, and attempts to reconstruct them inevitably entail considerable uncertainty and conjecture. Information is provided by tombstone inscriptions, census documents on papyri, skeletal remains and literary accounts.

Ancient birth and death rates were extremely high by modern standards. Mean life expectancy at birth is commonly estimated to have been around 20 to 30 years. The distribution of ages recorded in census returns from Roman Egypt is consistent with model life tables that posit a mean life expectancy at birth of 22 to 25 years. This estimate receives additional support from a variety of other data samples including funerary inscriptions from Roman North Africa, a Roman schedule used to calculate annuities known as ‘Ulpian’s Life Table’, and the age structure of a few cemetery populations. Roman emperors who died of natural causes had a similarly low life expectancy. This suggests that socio-economic standing had little effect on longevity. Mortality regimes were highly localized and determined primarily by the prevalence of particular infectious diseases. In parts of the Roman empire, seasonal variation in mortality can sometimes be reconstructed with the help of dates of death reported in tombstone inscriptions. These seasonality patterns also allow inferences about the underlying disease environments. According to these datasets, seasonal spikes in adult death rates were much stronger than in the more recent past, suggesting that even the most resilient age groups were susceptible to fatal infections. The main causes of death can be inferred from ancient medical texts and literary sources. Gastro-intestinal diseases, malaria and tuberculosis were particularly important. Both malaria and leprosy expanded during the Greco-Roman period. Smallpox epidemics occurred, possibly in Athens in 430 BCE and probably throughout the Roman empire in the 160/180 s CE. Plague spread from 540 to 750 CE in a pandemic that foreshadowed the medieval Black Death.

High levels of mortality required correspondingly high birth rates. The average woman surviving to menopause had to give birth to five or six children to ensure reproduction at replacement level. Birth rates within marriage were higher still: it has been calculated that in Roman Egypt, a woman who had been continuously married between menarche and menopause would on average have given birth eight or nine times. According to census records from the same region, 95 per cent of freeborn children were born to married parents. These documents also allow us to reconstruct the maternal age distribution of childbirths, which implies what is known as a ‘natural fertility’ regime in which fecundity was a direct function of a woman’s age, peaking around age 20 and gradually declining over time. Signs of stopping behavior – that is, the cessation of reproduction in response to family size or composition – are absent from these data.

At the same time, early and near-universal marriage for women and high birth rates went hand in hand with fertility control within marriage. Census returns from Roman Egypt indicate mean birth intervals of 3–4 years. Birth-spacing may have been achieved by prolonged breastfeeding or by other means. Ancient medical texts discuss a variety of putative contraceptives and abortifacients. More drastic intervention in the form of child exposure and infanticide was often socially condoned, although the actual scale of these practices remains unknown. The extent to which parents discriminated against female offspring is particularly controversial. While Greek and Roman sources sometimes refer to femicide, and evidence of male-biased sex ratios has been taken to reflect this custom, we are usually unable to determine whether ‘missing’ females had been killed or exposed after birth or were merely omitted from written records.

Among Greeks and Romans, (serial) monogamy was the norm. Polygamous unions were largely confined to ruling and elite families in Middle Eastern societies. At the same time, sexual access to slave women facilitated resource polygyny even in formally monogamous settings. In ancient Greek culture, women often appear to have been married off in their mid-teens while men took wives considerably later, around the age of 30. Funerary inscriptions from the western half of the Roman empire point to typical marriage ages of about 20 years for women and 30 years for men. Roman aristocrats generally married at younger ages. For Roman Egypt, the census records reflect mean marriage ages of 17 or 18 years for women and 25 years for men. They also show that whereas almost all women had married by their late 20 s, it was only by age 50 that most men had married at least once. This pattern of moderately early female and late male marriage resembles the so-called ‘Mediterranean marriage pattern’ which prevailed in the more recent past, suggesting a measure of long-term continuity in that region. Divorce could be initiated by both husbands and wives, and commonly lacked strong stigma. Remarriage was much more common for men than for women, especially after age 30: according to the Roman Egyptian census returns, two-thirds of men but only one-third of women were still married at the age of 50. In the pre-Christian period, celibacy was not normally considered desirable.

Marriages were mostly virilocal or neolocal. Bridal dowries were common but are best attested for elite circles. Slaves could not legally marry but were able to enter informal unions, primarily (but not only) with other slaves. Consanguineous unions were more widespread in the eastern Mediterranean, and especially in the Middle East, than in western Europe. Thus, first-cousin marriage occurred mostly in the East, and occasional half-sibling unions are known from the Greek world. Scholars still debate whether references to married couples of full siblings found in Roman Egyptian census documents reveal genuinely incestuous unions or record the unions of cousins who had legally become siblings through adoption. However, instances of brother–sister and parent–child marriage are credibly attested for ancient Middle Eastern rulers, and more generally for members of the Zoroastrian community in Mesopotamia and Iran.

The Greek and Latin languages lack specific terms for what we would call the nuclear family. Notions of family and household were more inclusive: next to parents and their offspring, the Greek oikos and the Roman familia or domus routinely encompassed co-resident kin and slaves. At the same time, Roman funerary inscriptions tend to privilege commemorative ties within the nuclear family, showing it to have been the principal locus of familial sentiment and obligation, of inheritance, and probably also of residence. More complex households were common in the eastern Mediterranean and the Middle East. In Roman Egypt, for example, the majority of the rural population belonged to households comprised of extended or multiple families. High death rates offset high fertility, thereby limiting family size, which averaged 4.3 in the same region.

Owing to unpredictable mortality and the desire to preserve male lineages, adoption of relatives appears to have been common. Partible inheritance rather than primogeniture was the norm. Daughters either received dowries as a substitute for an inheritance or inherited alongside their brothers. The social effects of high death rates undermined the formally patriarchal character of ancient households. A significant share of Greeks and Romans must have lost their fathers as minors and were assigned guardians, while many widows were unable to remarry. For these reasons, family units in which women and children were under the control of fathers and husbands were less common and more fragile than modern observers have often imagined.

Population numbers are very poorly known and continue to generate controversy. Statistical documents survive only from parts of Egypt, and literary references to population size are commonly vitiated by rhetorical stylization, ignorance or indifference. Archaeological data help to fill this gap but pose their own problems of interpretation. What we do know is that the Mediterranean regions and its hinterlands underwent significant population growth in the Greco-Roman period. In the Aegean, the collapse of late Bronze Age civilization around 1200 BCE coincided with strong demographic contraction. Population recovered from the early first millennium BCE onward and peaked in the classical period, in the fifth and fourth centuries BCE, when Greece may have been more densely populated than at any other time prior to the 20th century.

During this growth phase, Greek settlers established hundreds of colonies in Sicily, South Italy and the Black Sea littoral. By the fourth century BCE, up to 1000 Greek city-states were inhabited by 7 million people or more. Most of these communities were very small. The conquests of Alexander the Great in the late fourth century BCE triggered Greek emigration to Egypt, Syria, Mesopotamia, Iran and Central Asia. Large-scale state formation under his successors led to the creation of capital cities in excess of 100,000 residents, most notably Alexandria in Egypt. Meanwhile, populations expanded farther west in Italy, where this process drove the conflicts that eventually resulted in Roman regional hegemony, and more generally in western Europe and the Maghreb. A series of Roman census tallies from the last three centuries BCE offers insight into demographic change on the Italian peninsula. Even so, the total size of its population cannot be established with precision: depending on different interpretations of the extant census counts, by the beginning of the Common Era Italy may have been inhabited by no more than 6 million people (including slaves) or by two or even three times as many. These uncertainties interfere with modern assessments of Roman economic performance.

For a variety of reasons, an estimate between the extreme ends of spectrum seems appropriate: with a peak population of perhaps 10 or 12 million people, Roman Italy may well have matched the population densities of the high medieval and early modern periods. The population of the Roman empire as a whole is necessarily even more difficult to ascertain: while a total of 60–80 million seems realistic, a higher figure cannot entirely be ruled out. Maybe 10–20 per cent of these individuals lived in some 2000 cities. The capital city of Rome appears to have grown to a million residents, an urban population unparalleled in Europe prior to London around 1800. Starting in the late second century CE, epidemics reduced population numbers, although settlement densities remained high into late antiquity. Massive population losses finally accompanied the disintegration of the western half of the Roman empire in the fifth century CE and the onset of recurrent plague pandemics in the 540 s CE.

Despite its overall paucity and numerous shortcomings, demographic information from the Greco-Roman world is of considerable relevance to our understanding of ancient economic history. Centuries of continuous population growth, first in the eastern Mediterranean and later farther west, highlight the scale and persistence of an economic expansion which was driven by the spread of farming, technological innovation and gains from trade. Concurrent urbanization reinforces our impression of dynamic economic development. In the long run, however, ancient economies resembled other premodern economies in their inability to overcome Malthusian pressures through ongoing technological innovation. As population continued to expand, per capita economic growth eventually abated, first in Greece and later in the western Mediterranean. Judging by a variety of archaeological proxies of economic performance, by the time exogenous shocks in the form of plague and invasions began to affect the Roman empire in the second and third centuries CE the economy had already ceased to grow in real terms.

There is no indication that the Greco-Roman economic-demographic expansion significantly improved health or longevity: accretions to the stock of knowledge proved insufficient to mitigate the impact of the main causes of death, and potential gains from infrastructural provisions (such as aqueducts) may well have been offset by the demographic burden of urbanization and rising population densities which increased exposure to infection. In a number of skeletal samples, average body height was smaller in the Roman period than both immediately before and after, which likewise speaks against the notion of improvements in physiological wellbeing. Moreover, widespread skeletal evidence of deficiency diseases points to pervasive morbidity which would have curtailed productivity. High death rates discourage investment in education and impede human capital formation. Correspondingly high fertility depresses female labour participation and the status of women. In this environment, sustainable economic growth, let alone a fertility transition, was not feasible.

See Also