City systems and world-systems: 

Four millennia of city growth 

and decline

Christopher Chase-Dunn


University of California, Riverside

Susan Manning

Sociology, Hofstra University


This is a study of the growth of cities in four regions over the past 4000 years. We discuss changes in the relationship between political/military power, economic power andcity systems with special attention to the rise of European hegemony and the subsequent rise of East Asian world cities.We compare East Asian urban growth with the original heartland of cities in West Asia and North Africa, as well as Europe and the subcontinent of South Asia. This reveals the trajectories of city growth and decline and the relative importance of the different regions over time.And we re-examine the hypothesis of synchronicities of city growth and decline across distant regions as the Afro-eurasian world-system became more and more integrated. 

V.3-7-01 (6547 words).Submitted to International Studies Quarterly.

The comparative study of systems of cities, and settlement systems more generally, is an important basis of our understanding of human social evolution.The processes by which a world inhabited by small nomadic hunter-gatherer bands became the single global political economy of today involved the growth of settlement sizes and the expansion of interaction networks. These processes of growth and expansion were uneven in time and space.Settlements and cities did not always get larger. There were cycles of growth and decline. And those regions that originally developed larger settlements and cities were, in later epochs, no longer the leading regions in terms of the sizes of their largest cities.

This paper is a continuation of our studies of patterns of long-term social change.Our earlier studies have used data on both city sizes and the territorial sizes of empires to examine different regional interaction systems and the hypothesis that regions distant from one another were experiencing synchronous cycles of growth and decline (e.g. Chase-Dunn and Willard 1993; Chase-Dunn, Manning and Hall 1999).Our early study of city-size distributions in Afro-Eurasia (Chase-Dunn and Willard 1993; Chase-Dunn and Hall 1997: 222-223) found an apparent synchronicity between changes in city size distributions and the growth of largest cities in East Asia and West Asia-North Africa over a period of 2000 years.That led us to examine data on the territorial size of empires for similar synchronicity, which we found (Chase-Dunn, Manning and Hall 1999). This paper is a re-examination of the city size data which will enable us to address claims about the relative importance of China in the Afro-eurasian system advanced by Andre Gunder Frank (1998), and to more thoroughly examine the synchronicity hypothesis regarding city growth.

The relationship between power and demography has changed in important ways over the last 4000 years. Archaeologists often assume a direct correspondence between population density and societal power in intersocietal interactions, and they also assume that the settlement size hierarchy (e.g. Kowalewski 1982) indicates stratification within a polity.

The relationship between stratification and the sizes of settlements needs to be considered both within societies and between them.It is generally assumed that societies that have larger cities will also have greater power than societies with smaller cities, and it is similarly assumed that a society that is internally more stratified will have a steeper city-size distribution – the relative population size of settlements within the society. 

Though there was never a simple correspondence of population density and the relative power of societies vis a vis each other, there was a general correspondence.Societies that could concentrate greater numbers of people generally had an advantage in warfare. Exceptions to this are semiperipheral marcher chiefdoms and semiperipheral marcher states that conquered older core powers (Chase-Dunn and Hall 1997: Chapter 4). These less dense, typically less hierarchical and upwardly mobile semiperipheries used geopolitical advantages as well as superior organizational and military techniques to defeat older core societies and to form larger regional polities.But it was still generally the case that societies with greater numbers of people and with larger cities had greater power over other societies than those with lower population densities. 

In the modern world-system the relationship between population density and power has become even more complicated.While it is still true that the existence of large cities indicates the ability of a society to produce and acquire the great resources necessary to support huge populations living densely, the largest cities are even less likely than before to be in the most powerful countries.As of 1988 Mexico City was the second largest urban agglomeration on Earth, and Sao Paolo was the fourth largest. 

Power among societies is now much more directly a function of technology than it has been in the past. Machines controlled by a few people are capable of exercising power over great distances, and huge bodies of armed men are much less important than they have been in the past. Also economic power based on the ability to produce profitable high technology commodities and the ability to control financial resources has become a much more important source of power than it was in the systems of the tributary empires. In tributary empires military power was itself predominant, and it was greatly dependent on the ability to mobilize and supply large armies.We would not argue that economic power has replaced military power, but only that economic power has become a much more important basis of predominance in the modern capitalist world-system than in earlier tributary systems.Economic power also has a demographic basis and population density can be an advantage, but it is an advantage that is more strongly conditioned by technology and organizational features than ever before.[1]

The problem of synchronicity – changes of important social structural features that are simultaneous in time -- is germane to our understanding of the emergence of the modern world-system out of the formerly separate regional systems.It is plausible that synchronous processes in distant locations indicate systemness – the interaction of important processes that are influencing local developmental. The emergence of an integrated global system has been a long-term process that has been characterized by pulsation cycles (Chase-Dunn and Hall 1997).Long-distance trade and long-distance military campaigns have expanded and then contracted, but the overall trend has been greater and greater integration of larger and larger territories.After our re-examination of city synchronicity we will discuss possible causal explanations for synchronicity in the Afro-eurasian macro-system.

The questions we will try to answer in this research are: 

1. what can patterns of urban growth and decline tell us about the relative trajectories of development and the relationships among different regions in the emerging Afroeurasian system?This problem we shall label “Regional Importance.” And,

2. are there statistically significant amounts of synchronicity in urban growth and decline across distant regions that are similar (or different from) the synchronicities that we have found in the growth and decline of the territorial sizes of empires? This we shall call “City Synchronicity.”

Regional Importance

Andre Gunder Frank’s (1998) provocative study of the global economy from 1400 to 1800 CE[2] contends that China had long been the center of the global system. Franks also argues that the rise of European hegemony was a sudden and conjunctural development caused by the late emergence in China of a “high level equilibrium trap” and the success of Europeans in using bullion extracted from the Americas to buy their way into Chinese technological, financial and productive success.Frank contends that European hegemony was fragile from the start and will be short-lived with a predicted new rise of Chinese predominance in the near future. He also argues that the scholarly ignorance of the importance of China invalidates all the social science theories that have mistakenly understood the rise of the West and the differences between the East and the West. In Frank’s view there never was a transition from feudalism to capitalism that distinguished Europe from other regions of the world. He argues that the basic dynamics of development have been similar in the global system for 5000 years (Frank and Gills 1994).

Frank’s model of development is basically a combination of state expansion and financial accumulation, although in Reorient he focuses almost exclusively on financial centrality as the major important element.His study of global flows of specie, especially silver, is an important contribution to our understanding of what happened between 1400 and 1800 CE.Frank also uses demographic weight, and especially population growth and growth of the size of cities, as an indicator of relative importance and developmental success. 

It is our intention to systematically examine the growth of cities in order to shed more light on Frank’s claims about the relative development of East and West. Our study will begin in 2000 BCE when we first have data on the population sizes of cities in different regions.Though we understand the spatial nature of world-systems in terms of the sizes of different kinds of interaction networks (Chase-Dunn and Hall 1997: Chapter 3), in this study we will compare the same regions over time despite the fact thatinteraction nets grow from being rather small to being global over the period we are studying. Thus the unit of analysis in this study is the region, and regions are held constant over the whole period.The regions we will study are:

1. Europe, including the Mediterranean and Aegean islands, that part of the Eurasian continent to the west of the Caucasus Mountains, but not Asia Minor (now Turkey).

2. West Asia- North Africa, including Asia Minor, Egypt, Mesopotamia, Syria, Persia, the Levant, and Bactria.

3. The South Asian subcontinent, including the Indus river valley.

4. East Asia, including China, Korea and Japan and Southeast Asia, but not Indonesia.

These regions are defined for purposes of examining the claims made by Frank and others about relative development and possible synchronicity.

Measures of importance based on city data

Wilkinson (1992,1993) compared East Asia with West Asia using data from Chandler (1987) on the number of large cities in each region.Wilkinson used political-military interaction networks (which he calls “civilizations”) as his unit of analysis. Political-military interaction networks (PMNs) are a good unit of analysis because the alliances and enmities of polities are an important systemic feature of all world-systems. But PMNs change in size and location over time.[3] For purposes of our present study we will use constant regions (described above)as the unit of analysis.

We have improved upon Wilkinson’s study by weighting the cities by their population sizes.Using only the number of cities ignores differences in the sizes of cities.Figure 1 shows the population-weighted percentage that our four regions held of the twenty largest cities[4] on Earth from 2000 BCE to 1988 CE.[5] We also used the Chandler data, but we interpolated city sizes from Chandler’s tables in order to estimate the populations of cities.[6]This provides only a rough guess, but is still an improvement over Wilkinson’s simple count of the number of cities in each region. 

Figure 1: Regional Urban Population as a % of the World’s Largest Cities

In Figure 1 we see the emergence of the world’s first cities in Mesopotamia and Egypt represented here by the designation West Asia/North Africa. In the upper left hand corner of the graph the dashed line shows that this region had 100% of the largest cities on Earth in 2000 BCE.As other regions developed large cities this monopoly necessarily diminished and 4000 years later only a very small percentage of the world’s largest citiesare in this region. This is strong evidence of the notion of uneven development and the movement of the cutting edge of societal complexity.

The relative size importance of European cities (indicated by the solid line) shows a long oscillation around a low level, indicating Europe’s peripheral and semiperipheral location in the larger Afro-eurasian world-system. The long history of the incorporation ofthe very small systems of Europe into the expanding Central 

System of West Asia/North Africa is portrayed in Chapter 9 of Chase-Dunn and Hall (1997). Europe had been firmly incorporated into the Central System during the bronze and iron ages. Figure 1 indicates that by around 1450 CE Europe began a long rise. It passed East Asia in 1825 CE and peaked in 1850, and then underwent a rapid decline in importance as indicated by the relative size of large cities. 

Recall that all the largest cities on Earth, including those in the Americas and Oceania, are in the denominator of our measure of importance. So in the decades of the 20th century the percentages shown in Figure 1 do not add up to 100% because some of the largest cities are in none of the regions tracked (e.g. New York,Mexico City, etc.).When we use changing interaction network boundaries, as we have done in earlier research, these new large cities are included within the Central System, but the contribution of this study is to see what happens when we hold regions constant. Thus Figure 1 indicates that the relatively smaller and older European cities (e.g. London and Paris), were surpassed by the much largerAmerican and Japanese cities in the 20th century.

The trajectory of Europe displayed in Figure 1 supports part of Gunder Frank’s (1998) analysis, but contradicts other parts. The small cities of Europe in the early period indicate its peripheral status vis a vis the core regions of West Asia/North Africa, South Asia and East Asia. As Frank argues, Europe did not best East Asia (as indicated by city sizes) until the eighteenth century.But the long European rise, beginning in the fifteenth century, contradicts Frank’s depiction of a sudden and conjunctural emergence of European hegemony.Based on relative city sizes it appears that the rise of Europe occurred over a period of 500 years.

For East Asia we see in Figure 1 a rapid rise that began in 1200 BCE with the emergence of the first states in the Yellow River valley.This was followed by a small decline and then another burst of relative urban growth that began in 361 CE and that rose to a peak in 800 CE, another decline, and then a further rise to the highest peak of all in 1350 CE.Then there was a small decline and another peak in 1800. Not until 1825 was East Asia bested by the European cities after a decline that started in 1800 and continued until 1914, when a recovery began. The European cities were bested again by the East Asian citiesbetween 1950 and 1970 during the rapid decline of the European cities in terms of their size importance among the world’s largest cities. This most recent rise of the East Asian cities is related to the upward mobility of Japan and the East Asian NICs in the global political economy.Shin and Timberlake (1999) have demonstrated the contemporary rising importance of East Asian cities in the global airline transportation network.

Frank’s depiction of a sudden and radical decline of China that began in 1800 CE is supported in Figure 1.His analysis, which focuses on the period from 1400 to 1800 CE, does not examine the relative decline of East Asian predominance that began in 1350 and the rise to a new peak that began in 1650 as indicated in Figure 1.
The South Asian cities indicate how this region has fared during the long integration of the Afro-eurasian system.In Figure 1 the South Asian cities are indicated by the line with triangles.The early emergence of cities in the Indus river valley can be seen, as well as their demise, and then the rise of the Gangetic states that peaked in terms of city size importance in 200 BCE.The Indic cities disappeared completely from the world’s 20 largest cities in 1200 CE, but then rose to another peak in 1500 corresponding with the Mughal empire.In 1988 the South Asian cities had risen once again to a level as high as they had had between 1650 and 1700 CE.

The Rise of Europe

Figure 2 shows the largest cities in each region from 1400 to 1988 CE. This and the following figures differ from Figure 1 in that they are not percentages of the world’s largest cities, but are just graphs of the city sizes for each region. The most striking feature is the geometric growth rate of city sizes that began in Europe in the 19th century spread to East Asia, South Asia and West Asia/North Africa. Again, the big cities of the Americas are not included in Figure 2. 

Figure 2: Largest Cities in each Region Since 1400 CE

New York, which became the largest city in the world by 1925, beating out London, is thus not included because we are studying constant regions. Tokyo, the third largest city in 1925, had become the largest city on Earth by 1970, and Osaka held third place in that year. By 1980 Tokyo was still first, but Mexico City held second place, and Sao Paolo was in fourth place.

The geometric growth rate in the last two centuries obscures, in Figure 2, important fluctuations in the period from 1400 to 1850. These are germane to Frank’s argument about the relative centrality of China and Europe.Figure 3 excludes the period after 1850 to show these fluctuations.

Figure 3: Largest Cities in each Region, 1400-1850 CE

During the period between 1400 and 1850 the largest cities in South Asia and in West Asia/North Africa did not increase in size. Rather they fluctuated around a level that was smaller that the largest East Asian cities at the beginning of the period but larger than the largest cities in medieval Europe. Beginning in 1500 the largest European city, Constantinople, began a rapid period of growth that achieved the size of the largest city in East Asia (Peking) by 1550.European and Chinese cities were similarly large until 1700, when Peking began another period of rapid growth. European growth experienced another upsurge after 1750 with the mushrooming of London, but the size of London did not equal that of Peking until 1825. Within East Asia, Tokyo did not become larger than Peking until 1900.

The patterns shown in Figure 3 are quite similar to those to be found in Figure 4,a graph of the sum of the three largest city populations in each region, except that the most recent European rise is shown to have begun earlier, in 1600.

Figure 4: Three Largest Cities in each Region, 1400-1850 CE

Our examination of the largest cities in Europe and East Asia further reflects upon Frank’s (1998) characterization of the centrality of China and the rise of European hegemony. While we have contended above that the “European” rise began much earlier than Frank describes (shown in Figures 1-4 above), this early rise appears to have been due to the growth of Ottoman Constantinople.Though Constantinople was within the continent of Europe as wehave defined it, Frank might contend that crediting the Christian Europeans of later fame with the success of the Ottoman Turks is unfair, and that his hypothesis of the conjunctural nature of European hegemony is not challenged by this.

But there are some other facts that need to be taken into account here. The second and third largest cities in Europe in 1500 were Paris and Venice, followed by Naples and Milan.From 1500 to 1600 Paris grew from 185,000 to 245,000 and the other large cities of Christian Europe grew at a similar pace.So the early upsurge was not due only to the growth of Constantinople.Christian Europe was also experiencing a sixteenth century boom period. This does not dispute the relatively greater centrality of China in this period, but it does suggest that Christian Europe did not remain a peripheral backwater until it finally sprang to hegemony at the last minute in the 18th century. 

Constantinople’s size leveled off at 700,000 in 1600 and it stayed at that size until 1700, after which it began to decline. In this same period the largest cities of Christian Europe were growing rapidly. London grew larger than Constantinople by 1750. 

The Standardized Primacy Index (SPI) is a statistic that was invented for comparing city size distributions (Walters 1985).It calculates a single number from the population sizes of the largest cities in a region based on deviations from the rank-size rule.[7]

The SPIs of Europe and East Asia are uncorrelated in the period from 1400 CE to 1988.East Asia has a rank-size distribution from 1400 to 1600 while Europe dips through a period of flatness during this same period, recovering a rank-size distribution in 1550 (see Figure 5). 

Figure 5: East Asian and European SPIs, 1400-1988 CE

After 1600 East Asia drops to a very flat distribution in 1650, and then slowly recovers a city-size hierarchy that peaks in 1850, and then declines and rises again with the predominance of Tokyo. The European distribution declines also after 1600, but not to a similar depth. It recovers by 1850 and stays rank-size until 1914 when it again drops toward flatness. 

City Synchronicity

This section replicates our study of synchronous changes in West and East Asian city systems. We have excluded the South Asian and European regions from this analysis because they do not reveal synchronicity with the other regions or with each other. There are two main differences between what we have done in this study and what we did in earlier work. The first is that we have grouped the cities according to constant regions rather than spatially changing political/military interaction networks (PMNs), as described above. So the West Asia/North Africa region contains cities in this region over the whole time period rather than the expanding Central PMN studied by Chase-Dunn and Willard (1993) and Chase-Dunn and Hall (1997). The main difference due to this change is that emerging cities in the Aegean and points west are not included in the West Asian/North African region, whereas they were included in the Central PMN.This might reduce the degree of synchronicity found because most of the Greek and Roman cities are in the European region, and an important component of the synchronicity found in earlier work was that between the rise and fall of the Roman and Han empires. 

The other difference from earlier research is the use of a five-city SPI rather than a three-city SPI. This uses more of the cities in each region in the calculation of the SPI, and so should be a superior indicator of the city-size aspect of regional city systems. In practice the values are probably not greatly different from those obtained using only the three largest cities.

In addition to the SPI, we used two other indicators of changes in regional city systems.These are:

·the population size of the largest city, and

·the sum of the populations of the three largest cities.

Figure 6 shows our replication of the Eastern-Western comparison based on the five-city SPI and using spatially constant regions. Visual inspection of Figure 2 suggests a definite correlation between the rise and fall of steep and flat city size distributions in these distant regions from 800 BCE to 1600 CE. Before and after this period there is no correspondence between East and West Asian city size distributions.The synchronous Pearson’s r correlation coefficient produced by the values in Figure 6 is .32, and this approaches, but does not reach, statistical significance. 

Figure 6: West Asian and East Asian 5-city Standardized Primacy Indices

Figure 10.7 in Chase-Dunn and Hall (1997:217), based on a three-city SPI and comparing the Central and East Asian PMNs, finds a synchronous correlation of .44 and this is statistically significant at the .03 level.We suspect that the somewhat weaker correlation produced in this study is due to the fact that Rome is not included in the West Asian/North African region, whereas it was included in the Central PMN.The correlation between Roman and Han events and development, studied closely by Teggart (1939), was an important piece of the synchronicity of the Afro-eurasian system.That we finda substantial degree of synchronicity even when Rome is excluded is additional support for the hypothesis that East and West Asian growth/decline processes were linked in some way.

Visual inspection of Figure 6 also suggests a lead-lag relationship between East and West Asian city size changes. The West Asian/North African city size distribution seems to be leading the ups and downs by a lag of one or two hundred years.This may be an important clue to the nature of the causal relations that link the two regions, but this suggestedlag needs to be investigated using more temporally fine data. The long periods between measured city sizes in the Chandler data, especially before 1000 CE, make any statements about synchronicity or lagged relationships open to a good degree of doubt.On the other hand, our replication of the synchronicity found in empire size data using temporally finer time points increased the correlation found between the East Asian and Central PMNs (Chase-Dunn, Manning and Hall 1999).City population data for more time points and shorter intervals needs to be assembled to see how city synchronicity fairs with temporally finer data.

Figure 7 shows the trajectories of changes in the population sizes of the largest cities in East and West Asia over the same period as examined in Figure 6. The Pearson’s r synchronous correlation coefficient of .55, significant at the .003 level, further supports the hypothesis of East-West Asian city synchronicity. 

Figure 7: Largest Cities in West Asia/North Africa and in East Asia

Visual inspection of Figure 7 also suggests a lead-lag relationship, but it is more complicated than indicated in Figure 6. The temporal lead of changes in direction (growth and decline) seems to shift back and forth between West Asia/North Africa and East Asia. For example, the decline in the size of the largest city that began in East Asia in 800 CE did not start in West Asia until 900 CE, and the rise that began in East Asia in 1000 CE did not start in West Asia until 1150 CE. On the other hand the earlier rise and decline seems to have been led by West Asia. Again, these lag structures require further investigation that uses more temporally fine data with measurement points that are closer together.

Figure 8 includes data on the three largest cities in each region. The correlation coefficient further supports the hypothesis of city synchronicity. Visual inspection suggests another complicated lead-lag relationship with both regions leading during different periods.

Figure 8: Three Largest Cities in East Asia and West Asia/North Africa

The question of causal explanations of the East/West synchronicity of city and empire growth-decline sequences in this period has been considered in earlier studies (Chase-Dunn and Willard 1993; Chase-Dunn and Hall 1997; Chase-Dunn, Manning and Hall 1999). Basically these boil down to the following possibilities:

1.Long-distance trade between these distant regions transmitted economic forces great enough to bring them into a synchronous pattern of growth and decline. 

2.The cycle of the rise and fall of semiperipheral (or peripheral) states in Central Asia – the steppe nomad confederations that periodically attacked the agrarian empires of the East and the West – brought the East and West Asian systems into synchronicity.

3.World-systems may undergo similar sequences of expansion and contraction due to internal dynamics. If the Central and East Asia systems just happened to get going at about the same time, then their internal processes could have created a synchronicity that has little to do with causal relationships linking the two regions.

4.Some force exogenous to both regions created a synchronous pattern of expansion and contraction.The only candidate is climate change.

Of possible relevance is the fact that the South Asian region, spatially in between the Eastern and Western Asian regions, was not brought into either city or empire synchronicity with the others (Chase-Dunn and Hall 1997, Chase-Dunn Manning and Hall 1999).And the apparent lack of any synchronicity before 800 BCE or after 1600 CE is also an important clue. The possibility of a Western temporal priority suggested by visual inspection of Figure 6 above, would also be important because causality is assumed to go forward, but not backward, in time.


We can report additional support for the hypothesis of city synchronization between 800 BCE and 1600 CE for the distant regions of East Asia and West Asia/North Africa. The causality that resulted in this phenomenon is still unknown. What is needed is the operationalization for the relevant regions and time period of measures of the main hypothesized causal variables: climate change, warfare, and long-distance trade (e.g. Chase-Dunn 1995). Further research should also improve upon the existing city and empire size data, and should investigate the relationship between these within and across regions.

Our examination of the problem of the relative importance of regions relies exclusively on the population sizes of cities, a less than ideal indicator of power and relative centrality as discussed in the introductory section of this paper. Nevertheless, our results suggest some possible problems with Frank’s characterization of the relationship between Europe and China before and during the rise of European hegemony. Frank’s contention that Europe was primarily a peripheral region relative to the core regions of the Afro-eurasian world-system is supported by the city data, with some qualifications.Europe was for millennia a periphery of the large cities and powerful empires of ancient West Asian and North Africa. The Greek and Roman cores were instances of semiperipheral marcher states that conquered important parts of the older West Asian/North African core. After the decline of the Western Roman Empire, the core shifted back toward the East and Europe was once again importantly peripheral.

The synchronicity findings support the idea proposed in Frank and Gills (1994) that there was an integrated Afro-eurasian world-system much earlier than most historians and civilizationists suppose. But we cannot yet be certain that interaction networks were important causes of the synchronicity, and if they were, we do not know which kind of interaction was most important. 

Counter to Frank’s contention, however, the rise of European hegemony was not a sudden conjunctural event that was due solely to a developmental crisis in China. The city population data indicate that an important renewed core formation process had been occurring within Europe since at least the 14th century.This was partly a consequence of European extraction of resources from its own expanded periphery. But it was also likely due to the unusually virulent form of capitalistic accumulation within Europe, and the effects of this on the nature and actions of states. The development of European capitalism began among the city-states of Italy. It spread to the European interstate system, eventually resulting in the first capitalist nation-state – the Dutch Republic of the 17th century as well as the later rise of the hegemony of the United Kingdom of Great Britain in the 18th century. This process of regional core formation and its associated emphasis on capitalist commodity production further spread and institutionalized the logic of capitalist accumulation by defeating the efforts of territorial empires (Hapsburgs, Napoleonic France) to return the expanding European core to a more tributary mode of accumulation.

Acknowledging some of the uniquenesses of the emerging European hegemony does not require us to ignore the important continuities that also existed as well as the consequential ways in which European developments were linked with processes going on in the rest of the Afro-eurasian world-system. 

The more recent emergence of East Asian cities as again the very largest cities on Earth occurred in a context that was structurally and developmentally distinct from the multi-core system that still existed in 1800. Now there is only one core because all core states are directly interacting with one another. While the multi-core system prior to the 18th century was undoubtedly systemically integrated to an important extent, it was not as interdependent as the global world-system has now become.

A new East Asian hegemony is by no means a certainty, as both the United States and German-led Europe will be strong contenders in the next period of hegemonic rivalry (Bornschier and Chase-Dunn 1999). In this competition megacities may be more a liability than an advantage because the costs of these huge human agglomerations have continued to increase, while the benefits have been somewhat diminished by the falling costs of transportation and communication.Nevertheless megacities will continue to be an indicator of predominance because societies that can afford them will have demonstrated the ability to mobilize huge resources.


Abu-Lughod, Janet Lippman 1989. Before European Hegemony The World System A.D. 1250-1350 New York: Oxford University Press. 

Algaze, Guillermo 1993 The Uruk World System: The Dynamics of Expansion of Early Mesopotamian Civilization Chicago: University of Chicago Press.

Barfield, Thomas J. 1989 The Perilous Frontier: Nomadic Empires and China. Cambridge, MA.: Blackwell 

Bornschier, Volker and Christopher Chase-Dunn 1999 The Future of Global Conflict. London: Sage.

Bosworth, Andrew 1995 "World cities and world economic cycles" Pp. 206-228 in Stephen Sanderson (ed. ) Civilizations and World Systems. Walnut Creek,CA.: Altamira.

Chandler, Tertius 1987 Four Thousand Years of Urban Growth: An Historical Census. Lewiston,N.Y.: Edwin Mellon Press 

Chase-Dunn, Christopher 1985 "The system of world cities: A.D. 800-1975." Pp. 269-292 in Michael Timberlake (ed.) Urbanization in the World-Economy, New York:Academic Press. 

______ 1992 "The changing role of cities in world-systems." Pp. 51-88 in Volker Bornschier and Peter Lengyel (eds.) Waves, Formations and Values in the World System World Society Studies, Volume 2, New Brunswick, N.J.: Transaction Publishers. 

__________ 1995 “ Simultaneities of world-system development: cities, empires and climate change.”An unfunded proposal submitted to the National Science Foundation.

Chase-Dunn, C and Thomas D. Hall 1997 Rise and Demise: Comparing World-Systems Boulder, CO.: Westview Press.

Chase-Dunn, C. , Susan Manning and Thomas D. Hall 1999 “Rise and fall: East-West synchronicity and Indic exceptionalism reexamined,” Social Science History.

Chase-Dunn, C.and Alice Willard,1993 "Systems of cities and world-systems: settlement size hierarchies and cycles of political centralization, 2000 BC-1988AD" A paper presented at the International Studies Association meeting, March 24-27, Acapulco.

Chaudhuri, K.N. 1985 Trade and Civilisation in the Indian Ocean: An Economic History from the Rise of Islam to 1750. Cambridge: University of Cambridge Press.

Chi, Ch'ao-ting 1963 [1935] Key Economic Areas In Chinese History: As Revealed in the Development of Public Works for Water-Control, New York: Paragon Books. 

Cioffi-Revilla, Claudio 1991 "The long-range analysis of war," Journal of Interdisciplinary History 21:603-29.

Collins, Randall 1992 "The geographical and economic world-systems of kinship-based and agrarian-coercive societies." Review 15,3:373-88 (Summer). Westport, CT.:

Elvin, Mark 1973 The Pattern of the Chinese Past, Stanford: Stanford University Press. 

Fitzpatrick, John 1992 "The Middle Kingdom, the Middle Sea and the geographical pivot of history." Review 15,3:477-522 (Summer). 

Frank, Andre Gunder 1992 "The Centrality of Central Asia" Comparative Asian Studies Number 8.

______ 1998 Reorient: Global Economy in the Asian Age. Berkeley: University of California Press.

Frank, Andre Gunder and Barry Gills 1994 The World System: 500 or 5000 Years? London: Routledge.

Kowalewski, Stephen A. 1982 "The evolution of primate regional systems" Comparative Urban Research 9,1:60-78.

Lattimore, Owen 1940 Inner Asian Frontiers of China. New York: American Geographical Society. 

Mann, Michael 1986 The Sources of Social Power: A History of Power from the Beginning to A.D. 1760. Cambridge: Cambridge University Press.

McNeill, William H. 1976 Plagues and People Garden City, NJ: Anchor Books. 

Modelski, George and William R. Thompson 1994 Leading Sectors and World Powers: The Coevolution of Global Economics and Politics. Columbia, S.C.: University of South Carolina Press.

Nissen, Hans J. 1988 The Early History of the Ancient near East, 9000-2000 BC Chicago: University of Chicago Press. 

Pomeranz, Kenneth 2000 The Great Divergence. Berkeley: University of California Press.

Schneider, Jane 1991 "Was there a pre-capitalist world-system?" Pp. 45-66 in C. Chase-Dunn and T.D. Hall (eds.) Core/Periphery Relations in Precapitalist Worlds. Boulder, CO.: Westview. 

Shin, Kyoung-Ho and Michael Timberlake 1999 “The global city network, 1977-97” A paper to be presented at the annual meeting of the American Sociological Association, Chicago, August 10.

Smith, David A. and Michael Timberlake 1995 “Conceptualizing and mapping the structure of the world system’s city system,” Urban Studies 32,2:287-302.

Taagepera, Rein 1978a "Size and duration of empires: systematics of size" Social Science Research 7:108-27.

Teggart, Frederick J. 1939 Rome and China: A Study of Correlations in Historical Events Berkeley: University of California Press.

Walters, Pamela Barnhouse 1985 "Systems of cities and urban primacy: problems of definition and measurement." Pp.63-86 in Michael Timberlake (ed.) Urbanization in the World-Economy, New York: Academic Press.

Wilkinson, David 1987 "Central Civilization." Comparative Civilizations Review 17:31-59. 

______1991 "Core, peripheries and civilizations." Pp. 113-166 in Christopher Chase-Dunn and Thomas D. Hall (eds.) Core/Periphery Relations in Precapitalist Worlds, Boulder, CO.: Westview. 


______1992a "Decline phases in civilizations, regions and oikumenes." A paper presented at the annual meetings of the International Studies Association, Atlanta, GA. April 1-4. 

______ 1992b "Cities, civilizations and oikumenes:I." Comparative Civilizations Review 27:51-87 (Fall).

_______ 1993 “Cities, civilizations and oikumenes:II” Comparative Civilizations Review 28


[1]It is possible to empirically examine the relationship between city systems and power over the long run by studying the relationship between urban growth, city size distributions and the territorial sizes of empires.Sucha study would require the gathering of additional data on city populations, because the city data now available (Chandler 1987) has time points that are so widely spread out in earlier centuries that comparison with the more temporally fine empire size data would be inconclusive. Thus this study looks only at data on cities.
[2]We use the designations now employed by most world historians. BCEmeans “before common era,” and CE means “common era.”
[3] Wilkinson has chronographed the expansion of both political/military networks and trade networks. He uses the term “Central Civilization” for the PMN that was formed by the merger of the Mesopotamian and Eqyptian PMNs in about 1500 BCE. We call this entity the Central PMN.
[4] Before 650 BCE the number of largest cities reported in Chandler is fewer than twenty, as follows:

Number of largest cities
2000 BCE
1800 BCE
1600 BCE
1360 BCE
1200 BCE
1000 BCE
800 BCE
600 BCE

[5] Note that these regions do not include the Americas or Oceania. Europe means cities on the continent of Europe, not cities populated by people who migrated from Europe.
[6] Only 11 % of the city populations we studied were interpolated, and all of these were before 1300 CE. The city population data we used are available from
[7] The rank-size rule hypothesizes that the largest city will be twice as large as the second largest city, three times as large as the third largest, and etc.The SPIcalculates the average deviation from this standard in a manner similar to the Chi squared statistic, and the resulting value is “standardized” by dividing by the number of cities used to calculate the SPI. This makes it possible to compare city size distributions with data on different numbers of cities. The SPI takes a value of 0 when a city size distribution meets the rank-size rule. Negative values indicate a flatter distribution and positive values indicate a steeper one.