《Settlement, Urbanization, and Population》笔记

-**Morley** sets out some of the theoretical debate about urbanization in the Roman world and its possible relationship to the economy, reminding us in particular of the city’s importance as a location of demand and of the concentration of political power.

-Historians’ tendency to take cities for granted as real and important social objects acting on their surroundings should be ‘replaced by a concern to understand towns as sites in which the history of larger systems—states, societies, modes of production, world economies—is partially, but crucially, worked out’. p153

-Urbanization can be understood as one of the products of the confluence of four different processes of social, economic, and cultural change: concentration, crystallization, integration, and differentiation. pp154-153

-**Wilson**’s contribution looks at some of the physical evidence for Roman towns to see how we might establish the parameters of the plausible in estimating population densities for Roman cities in different regions, and therefore creating a set of possible estimates for population sizes of towns whose physical extent can be measured. A rough estimate is then presented for how the aggregate total of the urban population living in centres of 5,000 people or more in the mid-second century ad might relate to guesses about the total size of the population of the Roman empire.

-Even the lower ends of some of the ranges for Roman Mediterranean cities are substantially higher than population densities for medieval European cities. J. C. Russell wrote of medieval cities: ‘The average population density of cities was about 100–120 persons to the hectare. The densely populated city might run past 200 to the hectare but this was rare.’ By contrast, the evidence of the Roman cities in the Mediterranean for which we can attempt correlations of household counts and areas suggests normal outer ranges of 100–400 people per hectare and likely ranges of 150–250 people per hectare; even more in some cities. p176

-**Marzano**’s chapter applies a rank-size analysis to datasets of physical areas for cities in Britain and the Iberian Peninsula, with results suggesting that the urban systems of these areas show a high degree of interaction with the outside world. Within the Iberian Peninsula, the urban system of Lusitania appears more selfcontained than those of the provinces of Hispania, where the deviation from the expected rank-size distribution suggests that we are not in fact looking at a complete urban system within the province, but at an urban system integrated into wider Mediterranean connections and whose primate city is Rome.

-Zipf’s law, or the rank-size rule, which, it should be stressed, is an empirical rule, says that ‘when ranks of cities, arranged in descending order, are plotted against their populations (rank 1 being given to the largest, and so on) in a doubly logarithmic graph, a rank-size distribution results’, or to put it much more simply: ‘In an ordered set of cities representing a given country, the product of the rank and size of a city is constant. p201

r = rank of a city, P = population of a city, and k and q = constants

D = population density, A = surface area of a city, r = rank of a city, and k and q = constants

-**Hanson** extends this approach to Asia Minor, showing both that the population estimates for some of the major cities of the region (Ephesus, Pergamon, Miletus, etc.) have been considerably exaggerated in previous literature, and that the model suggests that the region as a whole was closely integrated into a wider urban system focused on the Mediterranean.

- If, however, a concave deviation is displayed, the hierarchy is said to be primate. Here there is more emphasis on the forces of unification, leading to a pooling of resources in one city. This would indicate a complete settlement system with a leading central city, dominant over the rest of the territory. This can be related to the ideas of ‘hypercephalie’ noted above. If a convex distribution is displayed, this may be seen as the forces of diversification acting on the system, pointing to low levels of system integration. However, it has also been noted that such a distribution can also be the result of excluding the primate centre from the distribution, or the result of system partitioning, where only one part of a wider system, and in particular the periphery of a system, is examined. On the other hand such a system can also be interpreted as the result of system pooling, where several urban systems are mistakenly grouped as one. p260

-**Keay** and **Earl**’s chapter on cities and city territories in Baetica addresses the problem using multiple criteria and approaches to the definition of territories, thus enabling the attribution of a hierarchy of urban settlements to the territories of top-level settlements, something a simple Voronoi diagram or calculation of Thiessen polygons is unable to do.

-**Bowman** reviews the evidence for the population of Roman Egypt and its distribution among different kinds and sizes of settlement; possibly 20 per cent of a (high count) population of 7.5 million lived in (large) cities; and indeed, settlements in Roman Egypt were remarkably large by comparison with those elsewhere, some villages apparently being larger than major towns in other provinces. There is good evidence for population increase from the Ptolemaic period through to the mid-second century, when the Antonine Plague seems to have had an important impact, but there appears to have been some recovery by the third century; the evidence for what follows is far from conclusive and whether it is to be regarded as recovery, stagnation, or decline is still open to debate. Given the claim (now commonly accepted) that Egypt was very heavily populated by Roman standards, further claims about its broader significance in the context of the Roman Mediterranean highlight the counterpoint of regional idiosyncrasy versus generic patterns and need to be carefully formulated. In an empire composed of very diverse regions, the concept of ‘typicality’ is elusive and probably illusory, but the position here adopted is that analysis of Egypt’s population structure and the economic relationships between ‘units’ of population (cities, villages, households) is significant for patterns of human behaviour in the eastern Mediterranean in classical antiquity.