In lieu of an abstract, here is a brief excerpt of the content:

One of the most elusive aspects of the demography of Roman Egypt—the ancient society for which we have the best documentary data—has been the extent to which the results of the analysis of the census declarations from Middle Egypt (Bagnall and Frier 1994) can be considered typical of other parts of Egypt, let alone other provinces of the Roman Empire. The data explored in this chapter, coming from a papyrus edited after the appearance of Bagnall and Frier’s (1994) work, may make a contribution to understanding some of the ways in which the population characteristics of Roman Egypt differed not only between urban and rural areas but also between different regions of Egypt, particularly different cities. The study of the demography of the ancient Mediterranean world was long marked by a lack of methodological rigor, either in devising models or in assessing the ancient evidence. These de¤ciencies were scathingly catalogued in Tim Parkin’s (1992) book on the subject. In recent years there have been substantial advances concerning both evidence and models. An excellent and detailed summary of this work can be found in Scheidel (2001b). There has been an enormous amount of critical work on the value of the ancient sources, the most salient element of which has been a more realistic valuation of the limits of our largest single body of evidence, the ages given on gravestones (see generally Parkin 1992:5–19). There is now a consensus that the ages in gravestones do not, in general, provide a pattern recognizable in any historical population and that commemoration therefore must be affected by one sort of bias or another. It has, to be sure, been argued that the patterns displayed by some bodies of inscriptions may be consistent with a population exhibiting catastrophic mortality (see Paine and Storey 1999, this volume). Although such catastrophes certainly occurred, most notably with the plague under Marcus Aurelius (Bagnall 2002; Duncan-Jones 1996; Scheidel 2002), this argument does not seem to me capable of explaining the persistence of distortion in population structures through a body of material ex7 An Urban Population from Roman Upper Egypt Roger S. Bagnall tending over a long period of time. Nonetheless, it is certainly true that the various epigraphic data sets deserve individual valuation and examination. Scheidel (2001a), following up earlier work by Shaw (1996) and himself (1996, 1998), has used Egyptian funerary epigraphy to try to make the case for regional variation in the causes of death underlying seasonal patterns of mortality, although he rejects most other uses of the inscriptions for basic demographic questions. Demographic modeling based on the more widespread and nuanced application of life tables to the ancient evidence has been even more central to the discussions of the last decade. In particular, it has been possible to show that the Egyptian census declarations on papyrus, which date to the period from Augustus to the mid-third century CE, reveal a population with mortality and fertility characteristics broadly consistent with those of the Coale-Demeny Model West tables with life expectancy at birth in the low twenties, probably between 22 and 25 (Bagnall and Frier 1994). Scheidel (2001a) has strongly reinforced the concerns expressed in our study of the census data about the extent to which any particular model life table derived from modern evidence is a fair representation of ancient (or even modern) high-mortality populations beset by endemic diseases . The problem is particularly acute in accounting for infant mortality, as very young children are poorly represented in both gravestones and the census data. Scheidel’s work has, however, neither shown conclusively that the endemic diseases he discusses would collectively alter the models in a particular fashion nor provided a basis on which to construct a model more faithful to ancient realities. The Egyptian census returns come partly from rural villages and partly from regional urban centers. There is thus an inherent possibility that these data can shed some light on the question of the demography of the rural/urban continuum. (This potential is, however, seriously limited in its range by the absence of any returns from Egypt’s one very large city, Alexandria.) The data are not, however, so abundant as to allow for as much deepening of the analysis as one might wish. Because the number of published returns as of 1994, when Bruce Frier and I published our book on the demography of Roman Egypt, was only about 300, and the number of known individuals...

Share