Specialists in gender issues, economics, papyrology, and a number of other fields of research on antiquity have begun to develop an appreciation for the potential benefits of utilizing modern methods of population study. Thus few will object when this book opens with a celebration of the last decade's explosive growth of interest in ancient demography (p. vii). All of the contributors to the volume under review have already established themselves as major players in this field. Their individual contributions in this book, the last four chapters of which were originally presented in a colloquium at Cambridge in 1997, offer illuminating examples of how demographic study can advance our understanding of the Roman world. When viewed together, these essays provide an outstanding introduction to the cutting edge of research on Roman demography.
In the first chapter, "Progress and Problems in Roman Demography," Walter Scheidel offers a balanced overview of the history of research on Roman demography (pp. 1-13). In this survey Scheidel justifiably attributes watershed status to the work of Karl Julius Beloch in the late nineteenth century and Keith Hopkins in the 1960s. Scheidel then presents a rich and detailed commentary on research on current problems in the field, including mortality, fertility, migration, population size, economic issues, and the impact of ecology on demographics (pp. 13-81). Among the many works that he singles out for special treatment are Bruce Frier's studies demonstrating the predominance of a natural fertility pattern in the Roman world (p. 34). Building on the research of Bagnall and Frier on Roman Egypt, Frier showed that deliberate efforts to limit procreation were not employed widely enough or with enough success to have a significant impact on the population of the empire (pp. 34-46; cf. 67-72). Scheidel's resulting dismissal of moralizing theories that posit a major decline in population as a result of birth control, abortion, or infanticide is just one of a number of important conclusions that is reached in the course of this chapter.
In chapter two, "The Seasonal Birthing Cycle of Roman Women," Brent D. Shaw argues that epitaphs from early Christian tombstones suggest a peak in the number of births among women in Rome and the rest of Italy during the winter months of December through February (pp. 83-110). This birthing cycle coheres with seasonal patterns of infant mortality and abandonment of infants in most pre-modern Mediterranean countries. Together these patterns suggest a corresponding cycle of intensive reproductive activity in the period of April through June. These birthing and reproductive cycles were "meta-patterns" consistent over long periods of time and wide geographical areas. Shaw concludes that these cycles probably were shaped by environmental factors, although disruptions in the basic patterns often are best explained by cultural factors.
In chapter three, "Recruitment and the Size of the Roman Population from the Third to the First Century BCE," Elio Lo Cascio challenges the prevailing notion (originally championed by Beloch) that the Augustan census figures refer to the whole citizen population of Italy (pp. 111-37). Following Tenney Frank, Cascio argues that, like the Republican figures, the Augustan figures refer only to adult males. Comparative data on ratios between numbers of men under arms and the respective total populations in preindustrial modern Europe indicate that the ratio of Roman men under arms to the total citizen population would be implausibly high if the Augustan figures referred to the entire citizen body. Comparative data and a reevaluation of the Polybian figures for 225 BCE suggest that Republican census figures were much too low. The total population of Italy during both the Republic and the empire was thus higher than suggested by the consensus position of Beloch. While it has been argued against Frank's position that it brings the population of Italy too close to what it attained in the fourteenth century, Cascio proposes that this actually serves to strengthen this position because it implies more reasonable rates of growth in the various periods from the Republic onward.
In chapter four, "More is Worse: Some Observations on the Population of the Roman Empire," Bruce W. Frier raises the possibility that the population of the Roman empire may have closely paralleled a "Chinese" mortality model (pp. 139-59). In this scenario, high fertility rates produce a population that outstrips its resources, thereby depressing the economy and thus contributing to high mortality rates that ultimately achieve an equilibrium with the high fertility rates. Definitive evidence that large parts of the empire were caught up in this kind of overpopulation trap is lacking, but such a model would explain the coincidental patterns of high fertility and high mortality that Frier has discovered in his earlier work.
In chapter five, "Urban Population in Late Roman Egypt and the End of the Ancient World," Richard Alston presents evidence for a demographic decline in the major cities of Egypt beginning perhaps as early as the fifth century and reaching archaeologically significant levels by the eighth century (pp. 161-204). Comparative data from medieval and early modern bubonic plague outbreaks indicates that the impact of plague varied widely from city to city and often left isolated settlements relatively unaffected. This suggests that the impact of plague epidemics on the overall population of late antique Egypt would not explain the decline of the cities in this period. Like plague, wars cannot explain the decline of the cities of Egypt. Alston suggests that the data for the decline in the cities in the fifth to eighth centuries are best explained by juxtaposing them with the data suggesting the growth of new villages that begins in the fourth century and peaks in the sixth century. Alston argues that this evidence indicates a demographic shift from cities to villages in Egypt as a result of a transfer of power structures away from the traditional urban elite to the village elite, the state, and the church with its rural monasteries.
Scheidel's masterful overview of research on Roman demography in chapter one provides an excellent framework in which to analyze the shorter essays that make up the remaining chapters in the book. The latter chapters in fact may be treated as case studies in the methodological problems with which Scheidel persistently wrestles.
A recurrent theme in chapter one is the problem of how to resolve the tension between data derived from ancient sources of questionable reliability and projections derived from models based on observable modern societies (esp. pp. 10-13). Scheidel insists on the need to apply comparative modeling because this approach is the only way to assure the production of quantifiable results, i.e., statistics, which are for him the very core of demography (pp. 2, 7). The chapters by each of the other four contributors are exemplary in this respect because they devote considerable space to both comparative models and quantitative modes of analysis. The reason why comparative and quantitative methods are so crucial to demography is that they offer the only way to counter the tendency to read the ancient evidence through the prejudicial filters of a modern ideology or the value judgments of classical authors, such as the notion that Rome's moral failures contributed to its decline (pp. 3-5, 7, 37-46; cf. 67-70). The articles of Cascio, Frier, and Alston illustrate this point quite well because they require a radical reevaluation if not outright rejection of the very notion of a demographic "decline" (as does other research discussed by Scheidel, pp. 67-72). The success of most of the articles in this volume justifies Scheidel's assertion that the use of projections derived from comparative models can sometimes help overcome the limitations of unreliable or missing ancient evidence (pp. 10-13, 80). Shaw's article exquisitely demonstrates this by showing how the cogency of a hypothesis formulated on the basis of fragmentary epigraphic evidence can be enhanced by comparative data from more recent periods. Cascio's appeal to degrees of militarization in early modern preindustrial societies and Alston's use of mortality rates during Medieval and early modern plagues provide excellent examples of how comparative models can force us to reevaluate ancient evidence for various demographic phenomena.
Scheidel emphasizes that such parametric modeling offers the most secure footing among possible approaches to demographic study because it is grounded in human reproductive capacity, the activity of disease agents, and other biological factors that place cross-cultural constraints on human behavior (pp. ix, 25-27). Thus Frier, for example, places great stress on the predictive power of carrying capacity because of its origins in something as fundamental as animal biology (pp. 140-41; but see Scheidel's cautions on carrying capacity, pp. 50, 57-60). In the same way Cascio's criticism of Beloch's model derives much of its persuasive power from Cascio's observation that the implausibly high proportion of males engaged in warfare in Beloch's model of the size of the Roman population would pose incredible difficulties for the simple biological necessity of food production (p. 129; also Scheidel on pp. viii, 49-59, though he rejects Cascio's view).
One of the major advantages of concentrating on the cross-cultural factors of biologically-based behavior is its generative potential for sensitizing researchers to cultural distinctives. For example, Shaw's demonstration that reproductive behavior follows meta-patterns conditioned by environmental factors opens an entirely new direction for research by pointing out that deviations from those patterns often have causes unique to specific cultures. Explaining the cultural origins of deviations thereby becomes potentially just as interesting as the initial discovery of the patterns (pp. 101-102). Efforts to distinguish environmentally induced patterns from culturally induced behavior thus can help bring cultural factors into higher resolution. Alston's argument for limiting the impact of factors in the physical environment (specifically, plague) on the decline in the population of the cities of Roman Egypt plays a central role in his effort to construct an alternative model that gives more weight to social and cultural factors that may have contributed to this decline. Such attempts to distinguish patterns rooted in environmental factors from those rooted in cultural factors also have the benefit of forcing researchers to take more seriously the systemic impact of physical and environmental phenomena (cf. Scheidel on pp. 28-29).
All of the contributors to this volume recognize that the great potential of quantitative and comparative methods does not make them a panacea for all the ills of demographic research. These methods cannot eliminate all the ambiguities in our ancient sources. One might plausibly assert, for example, that some of the archaeological evidence cited for urban decline in Alston's article resulted not from urban contraction throughout whole cities but rather from simple shifts in activity centers away from areas that became gentrified to areas that cannot be archaeologically evaluated because they are covered by modern buildings. This possibility remains especially strong in the case of Alexandria (pace pp. 170-71). Nevertheless, the uncertainty that this produces in assessing a few elements of the ancient data does not undercut the overall strength of Alston's article because he convincingly demonstrates localized variations between apparent decline and apparent growth. Such observable variations in patterns even within the same ancient society reinforce Scheidel's insistence that if demographic patterns in ancient societies vary widely, they also can be expected to differ radically from modern analogues (pp. 16-17). Scheidel recommends that this slippage between ancient societies and modern analogues be given formal recognition in the way that conclusions of demographic research are expressed. Rather than making naive claims for a precise fit between ancient cultures and statistical projections formulated by comparative modeling, he emphasizes the need to settle for statistical ranges in estimates of demographic phenomena (pp. 22-24).
Scheidel also does not shrink from grappling with the greatest problem with comparative modeling. This is that in some cases, such as in the use of carrying capacity as a basis for estimating population size, the application of models can be downright misleading (pp. 50, 57-59; cf. 18-22). The tentative tone of Frier's article, which relies heavily on comparative studies of the dynamics of carrying capacity, demonstrates a clear awareness of this potential problem. But Scheidel's point is not just that greater circumspection is needed in the application of a model. His point is rather that ideologies or presuppositions can easily lead the researcher to use a model in an illegitimate way or to select the wrong model in the first place (pp. 18, 20-21, 55, 59-60, 80). The very models that are employed to curb researcher bias thus can too easily become instruments in its advancement. To avoid the problem of circularity in the use of models, Scheidel demands more attention not so much to the ancient cultures that are the targets of analysis as much as to the modern analogues and the methodology by which the ancient and the modern are brought into comparison with each other (pp. 59, 80). The avenue that he suggests for charting a way out of what he calls the "seriously undertheorised" state of ancient demography is to pay greater attention to other disciplines (p. 80).
Although Scheidel does not elaborate on this point, an abundance of literature is available for exploring these issues. For example, the need to sensitize the researcher to the unavoidable circularity in the operation of presuppositions and unconscious models derived from one's own culture has long been a theme in research on hermeneutic philosophy.1 Prehistorians, for whom the deceptive luxury of appealing to written primary source texts is unavailable, have produced an extensive body of theoretical research on ethnographic analogy that can be fruitfully employed to inform the methodology of comparative demographic modeling.2
The contributors to this volume have demonstrated that the benefits to be derived from future attempts to develop the methodological integrity of ancient demography will be more than worth the effort. It remains to be seen, however, whether the value of the type of research exemplified in this book will receive enough recognition that more programs in Classics and ancient history will at last begin to formally require students to master the quantitative methods needed to operate in the new research environment created by modern technology and the modern interest in social history (cf. pp. 8-10). Perhaps this book will help to inspire curricular changes that will prove Scheidel wrong in his dismal assessment of the future of quantitative methods in Roman demography (p. 10). Whether or not it does, the book's clarity, scope, and methodological rigor commend it to a wide audience. If not for the steep price, it would serve as an excellent textbook for students at both the undergraduate and the graduate level.
1. E.g., Hans-Georg Gadamer, Truth and Method (2nd revised edition; New York: Crossroad, 1990).
2. E.g., Richard A. Gould and Patty Jo Watson, "A Dialogue on the Meaning and Use of Analogy in Ethnoarchaeological Reasoning," and the response by Alison Wylie, "An Analogy by Any Other Name Is Just as Analogical," Journal of Anthropological Archaeology 1 (1982) 355-81; 382-401.