An Urban Geography of the Roman World, 100 BC to AD 300

In view of the prominence of ancient urbanism in the scholarly literature during the last couple of decades, Hanson’s study is both highly relevant and well timed. The debates over the nature of ancient towns, the spatial and quantitative properties of urban systems, and the potential and limitations of various approaches to these topics have hitherto either been based on a small number of unrepresentative case studies or on uncritical general surveys of the towns of the ancient world. This is especially true for the Roman period, for which the sheer size of the territory in question, the variable extent of research carried out in the different corners of the empire, and the very large number of towns have discouraged systematic studies on an empire-wide level. Hanson’s study is the first attempt at a quantitative analysis based on what he claims to be an exhaustive inventory of the Roman towns in the period between the Late Republic and the High Empire (6–8, 41–3).

The territorial scope of the study is colossal. At its peak, the Roman empire stretched over 3.5 million km² and might have included some 2,000 towns. This vast data universe was tackled by a single researcher in a limited period of time. Factual errors, even in relatively high numbers, are therefore fully understandable. It is not the reviewers’ intention to dwell on the incorrect figures for or categorizations of individual towns, as these will be evident to specialist readers. As the author rightly argues in the introduction (8), one of the strengths of the big-data approach is its capacity to absorb individual erroneous observations. Instead, we examine the coherence and strength of the study’s main arguments, beginning with the structure of the book and some of its main points.

The study has two basic components: a discussion, illustrated by figures, tables, graphs, and thematic maps; and an alphabetical catalogue of the Roman towns grouped according to provinces. Chapter 1 introduces the reader to the general background, main objectives, and approach. It highlights the relevance of the topic, linking it to debates that have marked the fields of Roman social and economic history and urban studies in general: the nature of economic growth and the degree of economic integration in the Roman empire, the nature of the Roman and post-Antique city, and the urban-rural divide. Chapter 2 briefly presents the main theoretical approaches borrowed from the field of urban geography, such as central place theory and rank-size analysis.

Chapter 3 presents the main sources for the new inventory: the ancient writers, Strabo, Pliny, and Ptolemy; modern atlases of the ancient world; and studies of individual regions and provinces (33–45). Discussion of the results begins in the same chapter (45–8) with an overview of the total number of sites and their spread in the period between the first century B.C.E. and the third century C.E.

Chapter 4 discusses the size distribution of the urban settlements included in the analysis. Particular attention is devoted to population estimates, with a separate section on the literary and archaeological evidence for the population of the five largest cities of the Roman empire (49–55). This is followed by a lengthy section on the method by which estimates have been derived from the sizes of built-up areas (55–66). The author multiplies the built-up areas of various urban sites by a sliding range of fixed population densities, increasing from 100 persons per hectare for the smallest sites (<50 ha) to 500 persons per hectare for sites larger than 400 ha. The results for the individual cases and the integral urban system are discussed in the last section of the chapter (66–74). Special emphasis is placed on estimates for the total urban population and the urbanization rate. Chapters 5 and 6 discuss the nonquantitative parameters considered in the study, the spread and distribution of public buildings (75–80) and statuses (81–7). The author looks at the spatial dimensions of the urban system in chapter 7 (88–93), analyzing three basic aspects: degree of clustering of the urban centers (i.e., nearest neighbor analysis); possible extents of hinterlands and total urban settlement; and proximity to coastlines, rivers, and major roads.

The results of the study are briefly summarized in chapter 8 (94–104), followed by comparisons between the urban systems of the Roman empire, the Hellenistic world, and Late Medieval and early modern Europe. The catalogue of sites, the index of place names, and a select bibliography constitute the second part of the book (193–818). The catalogue contains information about size, status, public buildings, foundation dates, evidence for earlier settlement, and references. Unfortunately, the data for particular size estimates or juridical status are not referenced. A short bibliography for each town is listed at the end of the catalogue unit. Plans are omitted for reasons of copyright.

One of the central observations of Hanson’s study is that from a comparative perspective the Roman period was marked by urban growth and economic expansion unprecedented until the early modern period. The evident decrease in the urbanization rate between the Classical-Hellenistic and Roman periods (100–1) is downplayed by arguing that the absolute number of towns increased in the Roman period and that most of the Greek poleis were essentially large villages. These are weak arguments, as the increase in the number of towns in the Roman period may be attributed to the simple fact that the Roman empire was much larger than the classical and Hellenistic koine. The claim that the classical and Hellenistic polis was more agrarian than the Roman town is not supported by concrete evidence. As most towns in the east existed prior to the Roman conquest, one has to admit either that there was a structural difference between cities in the western and eastern halves of the Roman empire (contradicting the author’s conclusion that the urban system was unified and highly integrated [99]) or that the economies of the pre-Roman poleis changed radically after conquest.

Based on the convex shape of the rank-size graphs (figs. 71–80), and especially on an estimate for the average-sized town (67), the author asserts that a large segment of the population lived in average- to large-sized towns, or in his words: “A number of secondary and intermediate sites are more developed than we would have anticipated” (73). The use of an average estimate of 54 ha is hardly warranted, because this figure is of little relevance when the distributions are highly clustered with relatively few distant outliers. Of Hanson’s measurements, 596 out of 885 (67% [120–21]) of the cities were smaller than 50 ha. This is the very minimum, as most of the cities missing from the catalogue would have been smaller than the 50 ha threshold (see below). But the point is best illustrated by concrete examples. The large cities of Pisidia (Antioch, Kremna, Sagalassos, and Adada [the size of the last is not estimated by Hanson]) are all smaller than 50 ha. The same can be said for the Lycian towns, where only the largest cities of Kadyanda and possibly Telmessos are 50–60 ha. Only the capital and largest town of Roman Dalmatia is larger than 54 ha, and cities of that size would rank among the largest towns in most of the provinces of the Balkan and Iberian Peninsulas.

There are also problems with the interpretations of the rank-size graphs (since when is convexity a symptom of a highly integrated system? [73–4]), but we would rather focus on the data set. The author admits that the smallest cities, conventionally seen as the most numerous category, are most likely missing (41–2). This is easily checked. For example, the evidence for minting (available through the Roman Provincial Coinage project [http://rpc.ashmus.ox.ac.uk/]) is a good indicator for the civic status of the towns in the east (see P. Weiss, “The Cities and Their Money,” in C.J. Howgego et al., eds., Coinage and Identity in the Roman Provinces [Oxford 2005] 58–60). For the region of Asia Minor, Hanson lists some 180 cities (Asia [299–335], Bithynia et Pontus [354–58], and Cappadocia et Galatia and Cilicia [371–92], excluding cities on Cyprus and in Lycia et Pamphylia [662–84]), whereas the Roman Provincial Coinage database yields 275 cities that minted. Since many of the missing may have belonged to the smallest size category, it appears that Hanson’s catalogue is skewed toward the larger cities. The author justifies his approach by arguing that most of these towns either were in decline by the Roman period or failed to qualify as urban by the standards set out by urban geographers (41–3). This does not make a very convincing argument, as independent numismatic and epigraphic evidence unambiguously show we are dealing with self-governing poleis. Hanson’s approach not only obscures one of the most apparent features of ancient urbanism but also conveys a very pessimistic picture of the degree of urbanization in the Roman East by severely diminishing the total number of towns.

The author claims that the clustering of large towns implies a higher level of economic integration because the shortage in land and resources had to be compensated for by intensified production and trade between the different parts of the empire (98–9). This conclusion is not easily reconciled with the idea of balanced growth in all segments of the urban system. The only possible path to sustainable growth would have been the exchange of higher-value goods and services for natural resources and labor, inevitably leading to differential developments in the two zones. The author comes close to reaching this conclusion at the end of the book (99), but in his attempt to emphasize the optimality of the system he resorts to extending the catchment radii of individual towns to 80 km (fig. 141) instead of trying to explain the gaps in the urban network and the striking variations in urban density.

The conclusions drawn from the observed patterns in monumentality are equally problematic. The assertion that theaters were used for dramatic displays and amphitheaters were used for gladiatorial combat (79) ignores the fact that theaters (many predating the Roman conquest) were often converted to allow for gladiatorial displays (F. Sear, Roman Theatres: An Architectural Study [Oxford 2006] 17, 43), and the popularity of gladiatorial combat in the east was established long ago (L. Robert, Les gladiateurs dans l’Orient grec [Paris 1940]). Similarly, Hanson’s conclusion that towns were gradually losing their autonomy during the High Empire since the number of public buildings with administrative functions is comparatively small is surprisingly naive (78). It is difficult to understand why the author ignores the possibility that not all his building categories are equally represented in the archaeological record before drawing far-reaching conclusions on the changing nature of the ancient town under the High Empire.

The very structure of the book does not allow the disengaged reader to follow the author’s arguments and interpretations, contributing to the impression that he has invested more time in putting together his data set than in exploring the archaeological evidence with the aim of answering any particular research questions. Viewed in this light, it is particularly regrettable that a considerable portion of the book is devoted to discussions of well-trodden topics of ancient urbanism and that only some of the geographic models referred to by the author have been applied to the collected data set. While we accept that the implementation of big-data projects may leave limited room for methodological reflection and interpretation, we strongly feel that the author should have done a better job in trying to make sense of the large body of evidence he has managed to collect.

Damjan Donev and Rinse Willet
Institute of History
Leiden University
d.donev@hum.leidenuniv.nl
r.willet@hum.leidenuniv.nl