Historical demography (ancient)

Comp. by: ps0001 Stage : Proof ChapterID: sh076 Date:30/6/11 Time:09:16:51 Filepath://spiina1004z/Production_III/WILEY_BLACKWELL/3B2/Bagnall/Application/Batch13/sh076.3d

DEMOGRAPHY

C

O

R

R

EC

TE

Populations are the object of study of demography. The discipline aims to understand the size, the distribution, the composition (structure), and the development of populations. Historical demography, in focusing on populations of the past, is tightly intertwined with the discipline of history. Asking truly transdisciplinary questions, demography often finds answers in knowledge provided by neighboring disciplines: anthropology, sociology, biology, climatology, geography, archaeology, and economy are of particular importance in this. Many questions asked by demographers ultimately pertain to the three main demographic parameters that determine macro-demographic developments: mortality, fertility, and migration.

PR

SASKIA HIN

levels of uncertainty, and can reach only reasonable approximations at best. The most useful material transmitted to us consists of census documents from Roman Egypt. A little over three hundred household records provide us with quantifiable information on, for example, LIFE EXPECTANCY, household composition, and marriage patterns (Bagnall and Frier 2006). Data of this kind are not available for other regions and time periods in the ancient Mediterranean. Salt-tax registers from Ptolemaic Egypt provide data at the household level, but generally include adults only (Clarysse and Thompson 2006a; 2006b). A series of census totals are known for Roman Italy covering, with mostly five-year intervals, most of the Republic and the Early Imperial period. One census is recorded for Classical Athens. Beyond that, figures relate mostly to military forces, to manpower available or to members of certain (political) institutions or organizations, such as ephebes (see EPHEBE, EPHEBEIA) or members of the BOULE. Inscriptions provide a second body of evidence. Although FUNERARY INSCRIPTIONS are not of use in reconstructing life expectancies due to recording biases (see EPIGRAPHIC HABIT), avenues for studying ages at first marriage and family ties on the basis of the information they provide are more promising. Population simulations in which ancient data are combined with demographic modeling allow ancient historians to explore interactions between variables and get a better sense of the validity and implications of certain hypotheses. This is demonstrated in Saller (1994),

D

Demography, historical (ancient Mediterranean)

O O

F

76

METHODS AND SOURCES

U

N

The study of ancient demography is marked by severe limitations. Reliable quantifiable sources, the core materials with which demographers of the modern and early modern periods work, are scarce for societies as remote as the ancient Mediterranean. Data collections on the main demographic variables of mortality, fertility, and migration are limited in number and in quality. As a result, the discipline of ancient demography is confronted with high

The Encyclopedia of Ancient History, First Edition. Edited by Roger S. Bagnall, Kai Brodersen, Craige B. Champion, Andrew Erskine, and Sabine R. Huebner. © 2012 Blackwell Publishing Ltd. Published 2012 by Blackwell Publishing Ltd.

Comp. by: ps0001 Stage : Proof ChapterID: sh076 Date:30/6/11 Time:09:16:51 Filepath://spiina1004z/Production_III/WILEY_BLACKWELL/3B2/Bagnall/Application/Batch13/sh076.3d

2

DEMOGRAPHY, HISTORICAL (ANCIENT MEDITERRANEAN)

C

O

R

R

EC

The best data on population size and development within the ancient Mediterranean are come from Italy, where we have a series of census figures covering most of the Republican period and the Early Empire. But even these are problematic to interpret because targeted groups may have shifted over time, and the definitions of who were registered are evasive. Debates have been and still are intense. Two widely divergent population estimates have dominated scholarly discussion: one of ca. 12–14 million, proposed by Lo Cascio (e.g., 1994); another of 4–6 million (see esp. Brunt 1987). Hin (2008) now suggests an intermediate population total that implies slow demographic growth for the Republican period, based on a reading of the Augustan census as the population SUI IURIS. The population in Greco-Roman Egypt probably reached higher numbers than during the early nineteenth century, and fluctuated between ca. 4 and 7 million (Scheidel 2001). Where the Greek mainland and the Aegean Islands are

N

F

O O

PR

TE

POPULATION SIZE AND DEVELOPMENT

U

concerned, there is more convergence among ancient historians in population estimates: the region had about 2–2.5 million inhabitants according to most scholars, though Hansen (2006) regards a total of 3 to 3.5 million more realistic. Guesstimates of the total population of the Roman Empire tend to be in the order of 50–70 million. It is generally, though not universally, agreed upon that the population of the city of Rome in the first century CE was around 800,000. With perhaps around 120,000 free inhabitants (cf. Ath. 6.272C), late-fourth-century Athens was significantly smaller. Over the long term, from ca. 1500 BCE to CE 200, the Mediterranean as a whole experienced population growth at a slow rate of perhaps ca. 0.1 percent per annum. In the short and medium term, rates could deviate substantially from that average, depending upon time- and place- specific conditions (see Scheidel 2007). Exemplary in this regard is the period of the “Antonine” PLAGUE (second and third centuries CE), which led to demographic contraction. Subsequently, demographic divergence followed. While the Roman West experienced ongoing stagnation and population decline, the Greek East regained vitality.

D

which explores the demographic reality of PATRIA POTESTAS. Archaeology offers data on land use and settlement patterns, although linking these to population developments has proven to be problematic. In recent years, some headway has been made in ageing and sexing bones. Developments in bioarchaeology have, however, been most rapid where biological standards of living are concerned: dietary intakes, gender differences in health, and pathology are now increasingly being studied (e.g., King 2005). Lastly, comparative evidence not only frames ancient demography within the wider context of historical demography, but also sheds light on underlying causes and constraints that are invisible in the available sources.

THE THREE MAIN DEMOGRAPHIC PARAMETERS Mortality

Mean life expectancy at birth (known as e0) generally fell within the range of twenty to thirty years. The low survival rates implied were caused by, among other factors, high levels of neonatal and infant mortality, deaths from infectious and epidemic diseases, and – with varying rates – warfare. Little is known about the actual shape of the population pyramid, or the age structure of ancient populations. The census records from Egypt comply well with age distributions in Coale and Demeny’s Model Life Tables pertaining to life expectancies between ages

Comp. by: ps0001 Stage : Proof ChapterID: sh076 Date:30/6/11 Time:09:16:51 Filepath://spiina1004z/Production_III/WILEY_BLACKWELL/3B2/Bagnall/Application/Batch13/sh076.3d

DEMOGRAPHY, HISTORICAL (ANCIENT MEDITERRANEAN)

PR

O O

F

from their advantageous position on the preventive side of healthcare. Third, gender is associated with differential mortality. Frequent childbirths under difficult conditions put women at risk. INFANTICIDE and EXPOSURE OF CHILDREN were most likely directed disproportionately against girls, who were perhaps also more likely to suffer from malnutrition. For these reasons, it may have been that female life expectancy was somewhat lower than male life expectancy, although these factors should be weighed against war mortality risks for men. Even our best evidence for sex ratios (men to women) in the ancient world, the Egyptian census data, are problematic, but they show no evidence for strongly biased overall sex ratios.

D

FERTILITY AND THE FAMILY Populations with high mortality levels need to produce accordingly high numbers of offspring to ensure the reproduction of the population. With life expectancies between age twenty and thirty under normal circumstances, an average of four to six children per woman surviving to MENOPAUSE age would suffice. High marriage rates in combination with low marriage ages for women created demographic preconditions for high fertility rates. Evidence of ages at first marriage for women across the ancient Mediterranean invariably suggests that the average woman was married by her mid teens (Greece) or late teens (Rome) (see AGE). This was fairly early when compared with later preindustrial populations with life expectancies in a similarly low range, and implies that slow population growth was achieved by checks on reproduction of a different nature than marriage practices. For stopping behavior, a mechanism in place in recent times, we find no evidence: Egyptian census documents show a natural fertility pattern, with women bearing children until age forty to forty-five. But ancient parents spaced their sons and

U

N

C

O

R

R

EC

TE

twenty-two and twenty-five. But part of this fit is a construct, because infant mortality data are missing. Also, the distribution of death by age in the Model Life Tables for populations with life expectancies as low as these is an extrapolation. Known death rates of longer living, more recent populations have been artificially heightened for all ages in a non-discriminatory way. But disease environments in the ancient world may well have been different, and consequentially may have killed off people in different veins and following different age patterns of mortality. Most recently, it has been suggested that death rates among young adults were perhaps higher than ancient historians have tended to believe, and neonatal and infant mortality rates lower; the latter perhaps as a result of relatively favorable breastfeeding practices (but see WET-NURSE, WET-NURSING) and longer periods between births. Age-specific mortality rates, as well as life expectancy itself, likely also varied within the ancient Mediterranean. We can think of variation along three lines: geography or “locality,” socioeconomic position, and gender. There is little doubt that mortality risks were partly dependent upon one’s place(s) of residence. Variation in environmental conditions made certain places unhealthier than others. Low-lying marshy areas were notorious for being malarial killing fields; hills were healthier in this respect (Livy 5.54.4). Population density also played a major role: people living in urban centers ran higher risks of dying because concentrations of people – and waste – were susceptible to the outbreak and spread of infectious diseases. When broken down according to locality, the census data from Egypt indeed show higher mortality among metropolite populations. It has been suggested that location was in fact all that mattered, and that mortality among pre-transitional populations lacked a socioeconomic gradient. This line of argument, however, deserves reconsideration. It is true that little could be done to cure conditions of ill-health after their occurrence, but elites may have profited

3

Comp. by: ps0001 Stage : Proof ChapterID: sh076 Date:30/6/11 Time:09:16:51 Filepath://spiina1004z/Production_III/WILEY_BLACKWELL/3B2/Bagnall/Application/Batch13/sh076.3d

4

DEMOGRAPHY, HISTORICAL (ANCIENT MEDITERRANEAN)

assumptions on family patterns derived from inscriptional evidence.

MIGRATION

D

PR

O O

F

Migration tends to be the “stepchild” of (historical) demography. The subject is complex, and data are patchy – even more so for the ancient world. The phenomenon of colonization, bringing about migration in an institutionalized context, is best described. It is less clear how much migration across regions and borders we can expect to have occurred without sponsoring and organization by Greek or Roman states. There has been a strong belief among demographers and (ancient) historians that historical agrarian populations were characterized by immobility, but this view is increasingly undermined by comparative evidence. For the ancient world, Erdkamp (2008) has recently made a case for thinking in terms of spectra of mobility, with movements backand-forth in accordance with labor market cycles. Athenaeus (6.272C) claimed that in the small city-state of Athens, a third of all free adult men were metics, but these need not have been first-generation immigrants. On the basis of inscriptional evidence from Rome, it has been suggested that at least 5 percent of its third-century CE population consisted of free provincial immigrants (Noy 2000). The demographic composition by sex and age of migration flows likely depended on distance and destination. Migrants to a big city such as Rome seem mostly to have been young males. The demographic impact of migration toward urban environments, or urbanization, was negative. Whether migration negatively affected fertility inasmuch as it did survival chances is questionable, given that labor migration by males tends to disrupt reproduction less than does female migration. War, conquest, and slave trading created flows of forced migration. Here we find migration at its most disruptive in a demographic

U

N

C

O

R

R

EC

TE

daughters well. Protracted periods of breastfeeding and concurrent taboos on sexual intercourse encouraged relatively long birth intervals. Abortion, infanticide, and abandonment also played a role. Concerns about low fertility levels among upper classes, who had less social and economic motivations to bear children, motivated the Roman emperor Augustus to install a family policy directed at increasing birth rates (see MARRIAGE LEGISLATION OF AUGUSTUS). Family patterns were affected by high mortality levels, frequent divorce, remarriage, and age gaps between marriage partners. Evidence available suggests that men on average married substantially later than women, in their twenties rather than in their teens. As a result, WIDOWS were more common than widowers. But the duration of marriage unions need not always have been negatively affected by wide age gaps: records cover far from everyone, and do not foreclose the possibility that other men may have married young, depending on class and economic circumstances. Households consisting of more than two generations were rare and mostly short-lived. It is generally believed that in the Latin West, small, nuclear families were remarkably predominant (Saller and Shaw 1984). However, it must be emphasized that the inscriptional evidence is indirect, and tells us more about the strength of family ties than about actual residence. The contrast with what is observed for several cities in the Greek East on the basis of similar evidence is sharp. Here, ties between members of the extended family seem to have been more intense (Scheidel in press). Census records show that patterns of co-residence varied in Roman Egypt, where nuclear families were more prevalent in the cities than in rural environments. This may well have held true for other Mediterranean regions, but we lack relevant quantifiable sources for rural areas. On the basis of her findings from a comparison between Egyptian commemorative inscriptions and Egyptian census data, Huebner (2011) now criticizes the validity of

Comp. by: ps0001 Stage : Proof ChapterID: sh076 Date:30/6/11 Time:09:16:51 Filepath://spiina1004z/Production_III/WILEY_BLACKWELL/3B2/Bagnall/Application/Batch13/sh076.3d

DEMOGRAPHY, HISTORICAL (ANCIENT MEDITERRANEAN)

REFERENCES AND SUGGESTED READINGS

U

N

C

O

R

R

EC

F

O O

TE

Bagnall, R. S. and Frier, B. W. (2006) The demography of Roman Egypt. Cambridge. Bradley, K. R. (1991) Disovering the Roman family: studies in Roman social history. Oxford. Brunt, P. A. (1987) Italian manpower. Oxford. Clarysse, W. and Thompson, D. J. (2006a) Counting the people in Hellenistic Egypt, vol. 1: Population registers. Cambridge. Clarysse, W. and Thompson, D. J. (2006b) Counting the people in Hellenistic Egypt, vol. 2: Historical studies: see esp. ch. 7. Cambridge. Erdkamp, P. P. M. (2008) “Mobility and migration in Italy in the second century BC.” In L. de Ligt and S. J. Northwood, eds., People, land, and politics:

PR

SEE ALSO: Birth control; Birth, registrations of; Census; Childbirth; Colonialism; Colonies, Roman and Latin (Republican); Colonies, Roman Empire (East); Colonies, Roman Empire (West); Colonization, Greek; Death; Death, registrations of; Divorce, Greek; Divorce, Roman; Disease and health, Greece and Rome; Epidemic disease; Family, Greek and Roman; Famine and food shortages, Greece and Rome; Fertility, natural; Infant diseases and mortality; Marriage, Greece and Rome; Migration; Nutrition and malnutrition; Paleopathology; Public health, Egypt; Public health, Greece and Rome; Standards of living, wealth.

demographic developments and the transformation of Roman Italy 300 BC–AD 14: 417–49. Leiden. Hansen, M. H. (2006) The shotgun method: the demography of the ancient Greek city-state culture. Columbia, MO. Hin, S. (2008) “Counting Romans.” In L. de Ligt and S. J. Northwood, eds., People, land, and politics: demographic developments and the transformation of Roman Italy 300 BC–AD 14: 187–238. Leiden. Huebner, S. R. (2011) “Household and the family in the Roman West.” In B. Rawson, ed., The ancient family. Oxford. King, H., ed. (2005) Health in antiquity. London. Lo Cascio, E. (1994) “The size of the Roman population: Beloch and the meaning of the Augustan census figures.” Journal of Roman Studies 84: 23–40. Noy, D. (2000) Foreigners at Rome: citizens and strangers. London. Saller, R. P. (1994) Patriarchy, property, and death in the Roman family. Cambridge. Saller, R. P. and Shaw, B. D. (1984) “Tombstones and Roman family relations in the Principate: civilians, soldiers and slaves.” Journal of Roman Studies 74: 124–56. Scheidel, W. (2001) Death on the Nile. Leiden. Scheidel, W. (2007) “Demography.” In W. Scheidel, I. Morris and R. Saller, eds., The Cambridge economic history of the Greco-Roman world. Cambridge. Scheidel, W. (in press) “Epigraphy and demography: birth, marriage family, and death.” In J. Davies and J. Wilkes, eds, Epigraphy and the historical sciences. Oxford.

D

sense, with families driven apart, and slaves being precluded from any formal family rights.

5