World history might be viewed, from one vantage point, as a history of cities, if only because urbanization has been coextensive with that portion of human experience on earth – the past five-six millennia – that is commonly portrayed in world histories. Most prominently representatives of urban development at any given time have been world cities. That is why a minimum of acquaintance with, if not the actual study of world cities and their interrelationships cannot but offer useful insights not just into the growth of a central form of human organization, but into world history itself.


This overview proposes, and sketches out the answers to, three questions:  (1)   how do we define world cities, and what are the criteria used for identifying them?   (2)   over the entire time span of world history,  which actually were those cites, and can they meaningfully be regarded as the centers of the world system of their time, and (3)  what can a systematic survey tell us about the major outlines and trends in the experience of the system of world cities?   



1.         World cities


We define ‘world cities’ as “cities of world importance, actually and potentially”.   The source of that importance might be twofold:  either (1) size, and/or (2) role or position in the functioning of the world system.   Size will be reckoned principally in terms of population numbers, and as compared with other cities of that same era, for we know that over historical time cities have grown larger by orders of magnitude.   We therefore distinguish among world cities of the ancient, classical, and modern periods.   World system position might, in the most obvious case, refer to a place in the global economy, either in a commercial or in a productive capacity.   The capitals of powerful or strategically important states, and the military garrisons they might harbor, will be politically important.   Cult and/or religious centers attract interest and visitors from afar.   Finally, there have been cities noted for their cultural assets and the excellence of the learning they afford.  The large cities that we are interested in have been most often multifunctional, and large populations, in turn, made multi-functionality possible, and might also provide fertile soil for innovative enterprises.   As actual importance is often hard to measure, we regard size and position to be indications of the probability of importance.


In this survey, we place less emphasis on individual cities and their fortunes than upon the ensemble of world cities viewed as an interdependent system.   The search is not for a single, ‘dominant’, world city because such is in fact hard to find over  the past several millennia  but rather for sets of urban centers, their form and composition, and the connections among them, that might be identified at various points in world time,.   One question that will arise will concern the imperial or non-imperial, autonomous, character of these systems.


The present overview draws its empirical data from two quantitative censuses of urban growth, and contextualizes that data for the several eras, and regions.   Until quite recently, the prevailing view held that a statistical description of urbanization prior to about 1800 was an “impossibility”.   More recently new sources have opened up, as in archaeology, or social and economic history, that make that task less difficult.   The pioneering effort in this regard has been Tertius Chandler’s “Four Thousand Years of Urban Growth” (1987), a product of decades of research that brought together a mass of material on cities between 2250 and 1975.   Complementing and extending that work has been George Modelski’s “World Cities:-3000 to 2000” (2003) that provides fuller coverage for the first four millennia but deals more lightly with the data from 1000 onward;  it also reports on world trends in urbanization.


In the present survey, world cities will be identified as such basically by the use of the criterion of size.   For the ancient era (about 3000 to 1000 BCE) these will be settlements whose population may be estimated (e g. on the basis of the archaeologists’ site reports) to be in excess of 10,000.    Most such cities will be found to be in the range of 10,000 to 100,000.   For the classical era (1000 BCE to 1000) we shall be looking at urban centers with 100,000 or more, and in fact basically at those in the range of 100,000 to one million.   For the modern era (since 1000) we focus principally on “millionaire” conurbations, those in the 1-10 m range.  The population estimates that underlie these selections are documented in the works just cited but given the rarity of census data prior to 1800 most must be regarded as little more than ‘educated guesses’, whose levels of precision leave much to be desired but which may be adequate for purposes of outlining the broad picture of urban development.


Why only size?   In attempting a survey on such a large scale, both spatial, and temporal, an empirical, detailed, and documented assessment of world system position and participation is not really practicable. Without doubt, some of names that will come up purely on account of size do not merit ‘world city’ status, and there are other, smaller, places that have been omitted here and did exert significant influence at various times in the past.   All we can say is that this broad sweep through world history nets many significant big fish, but maybe not all of them.


2.         Backbone of world history


Which were the world cities of the ancient, classical, and modern era, and what was their role in world history, that is in organizing social life on earth?  As city life has often been treated as synonymous with civilization, and as the fate of cities has tracked the course of civilization, the story of urbanization is also in part an account of the civilizing process.


Ancient world    “History begins  at Sumer” was the tile of a popular account by S.N. Kramer that applies equally well to the emergence of the first city system, in Southern Mesopotamia in the fourth millennium BCE, in what  Gordon Childe dubbed the “Urban Revolution”.   The reference here is to a group of urban settlements centered on Uruk, clearly a major cult center but also a focus possibly of political activities but certainly also of regional exchanges whose reach has been shown to have extended from Iran in the east, to the upper Euphrates in the north, and evidently also to Egypt in the west.   By 3000 BCE we find here (and nowhere else) some half-dozen units that satisfy our criteria, hence an incipient world city system.   Uruk is at that time obviously the largest among them, with a population possibly reaching 40,000, without doubt the largest city in the world at this time, and that is just one reason for calling it the first world city.  For we also know, from archaeological and literary evidence, that that was also the likely locus of the invention of writing, and of calendars, innovations that proved to be of epochal significance.


This was the Uruk nucleus of an emerging system of world cities.  The first basic trend is the emergence, by mid-third millennium, of a viable and productive center in Summer, the “heartland of cities”, then organized in the form of some two dozen autonomous city-states.   An increasingly costly competition for regional leadership animated those states (by about 2300 that between Umma and Lagash), such that Sargon of Akkad, from outside the Land of Sumer was able step in and subdue it.   The reign of Akkad and Sumer came and went, and was followed by a native dynasty based on Ur.   As late as about 2000 something of a numerical parity existed between Sumer, and non-Summer cities, but not much later the former land of cities completely dropped out of sight.  By contrast important cities rise in Egypt (Memphis,  Thebes, Heliopolis), in north Mesopotamia (Mari), and in the Indus Valley (Mohenjo-Daro and Harappa),


The second basic trend therefore was the experience of dispersal, or more precisely the spread of urban practices throughout Eurasia that coincided with what in several areas had been described as “Dark Ages”:  in Sumer, in the Harappan region, and in post-Mycenean Greece.    By the end of the ancient era (and the Bronze age) three of the four major regions of the ‘Old World’ had been fertilized by the Urban Revolution:  West Asia (e.g. Babylon),  the Mediterranean (Mycenae), and East Asia (Yin, near Anyang, a major Shang capital).  The less than successful experiments in the Indus Valley, in the Ukraine, and even Peru would ultimately bear fruit too.  This dispersal was in fact a form of redistribution because while Sumer lost cities and was virtually de-urbanized, urbanism rose elsewhere and the number of world cities remained about the same it was a millennium earlier (22 in 2000 became 23 in 1200).  In other words, the story of the ancient era was rapid urban expansion at the center in its first half, followed by deceleration and dispersal in the second.


Classical world        The principal tendency of the classical era was the rapid formation, and subsequent consolidation, of strongly regional but also interconnected urban systems in the four main regions of Eurasia:  East Asia, South Asia, the Mediterranean, and West Asia.  A separate development also occurred in the Americas.  In the first three of these regions we first observe a thriving system of independent city-states, one that then succumbs to imperial rule; but in West Asia, the sequence is in fact reversed, and in Mesoamerica, the Mayan system of city states collapses of its own into incoherence.


In East Asia, virtually all of the principal urban growth occurred in China.   Haoqing (near Xi’an) was the Western Zhou capital and ceremonial center that bridged the ancient and classical periods, but after its destruction in 771 BCE the political center shifted to Luoyang, and in this Eastern Zhou era that followed, urbanization took off with considerable flourish.   One report credits 91 cities as likely founded pre-771, the number rising to 466 in the Spring and Autumn of Eastern Zhou (Wheatley 1971:173).   City-building was apparently part of state-building, with the term for “state” (guo) denoting a walled city.  While many of these cities were quite small, in no time a system of independent and flourishing states arose, each anchored in a major city, each with merely nominal, and ceremonial links to Luoyang.  That, in turn, left the field open to immensely destructive wars, in what came to be known as the period of Warring States.   The most ruthless and war-like of these states, Qin, conquered all others and founded the First Empire.   But that proved short-lived, and the Han empire that replaced it quickly built a great new capital, Changan (Xi’an), while continuing Luoyang in a secondary role.  Changan and Luoyang each had their good and bad times (the former was “largely destroyed” in 189, sacked in 315, plundered in 756 and 763, and subjected to a bloodbath in 881; the latter was “utterly destroyed” in 189-190, sacked in 311, and declined after 907)  but they continued to alternate as China’s leading cities right until the end of the classical era, when Kaifeng assumed a central place in the Northern Song.    By that time Kyoto (Japan) and Sorabol (in Silla, Korea) also joined the ranks of world cities in East Asia that accounted in the classical era, for about one third of the world’s major urban potential.


The South Asian experience paralleled that of East Asia, albeit on a smaller scale, in that the first millennium saw a flourishing of cities, in a context of autonomy, in North India in particular, a process that was then set back by the founding of imperial structures, and by external incursions.  Out of a cluster of tribal units an urban landscape emerged in the Ganges Valley than in turn coalesced into a system of independent polities,  that also became the seedbed of Buddhism.   But a recent study of the sites of the capitals of 16 of these Early Historic states showed them to be in the range of 50 to 200 hectares, suggesting populations smaller than their counterparts in China of that time and falling below our threshold.   Over time, one of these became the dominant power and formed the core of an empire centered on Pataliputra, on the Ganges, a large city that greatly impressed a Greek ambassador c. 300 BCE.   In the second half of the classical era the focus of North India shifted to Kanauji, that became the capital of the Guptas, and continued as the center piece of regional politics until sacked by  Moslem armies in 1018, and then destroyed.   In the south, Buddhism gained an early foothold in Sri Lanka at Anuradhapura,   that came to radiate its influence to Southeast Asia, including  Sri Ksetra in Burma, Palembang in Srivijaja, and Angkor in Cambodia.


The Mediterranean was the focus of the other major urban network of the classical world, equal in weight to the East Asian.   The Mediterranean network began to form later in the ancient era and soon expanded via three great waves of urbanization:  the Phoenician, the Greek, and the Roman.   In about 1000 BCE Tyre sent a first colony to Cyprus, and soon its settlers founded Carthage that in short order became the power house of the Western Mediterranean.  Pliny the Elder, the encyclopedist, put its population, prior to the start of the wars with Rome, at 700,000, probably too high but indicating its reputation.   The second wave came from the Greek world, under the sponsorship of by individual cites.  For example. Corinth initiated the foundation of Syracuse, on Sicily, whose defeat of the Athenian expedition tipped the scales in the Peloponnesian war,  and that may have been, c. 400, the largest of the Greek world.   But the greatest impetus to Hellenization came from the conquests of Alexander the Great, the founder of numerous cities.  The most important of these was Alexandria, that became the capital of the post-Alexandine rulers of Egypt, the Ptolemies who fashioned it not only into a center of political power, and of trade and shipping, but equipped it, too with a great lighthouse, on Pharos island, a beacon not only to ship-farers  but also a symbol of enlightenment for a city boasting of a museion and a huge library.   A city that was home to many nationalities became the first exemplar, in Stoic thought, of a cosmopolis, a city of the world, a higher formation than a mere polis.   The last wave was that of the Romans, builders of empire, but also builders of cities.  Rome annihilated Carthage  (but within a century put a new city in its place), and conquered Alexandria, to become the imperial capital of this Greco-Roman world.  It grew to huge proportions, to become the world’s most populous city, its citizens, untaxed, living off free bread, slave labor, and other spoils of empire. But the sack of Rome in 410 marked the start of the collapse of Western Rome, and the definitive onset, in this part of the world, of the second Dark Age.   


Which was the city whose population was the first to attain one million?   The estimates for Alexandria, c.100 BC, tend to put it in the 500,000 range, with some scholars claiming that it might have reached 1 m. by 200-100 BCE, and that would make it first.  But the more conservative guess would probably point to Rome that at the turn to the new millennium likely reached that figure, and held onto, and exceeded, it for some two-three centuries.    The next to reach ‘millionaire’ status was Tang era Changan, c.700-800.


Early in the classical era, powerful West Asian empires, in particular the Assyrian and the Persian, pressed upon the Mediterranean world, probably pushing the Phoenicians out to sea, and impressing the Greek world.   But the collapse of the Persian realm diminished the vitality of that region, and reduced its urban potential, and  it was not until the Moslem conquests of the 7th century that new political and urban space opened up to become the ‘Moslem world’.   Arab cavalry armies overthrew the Sassanian empire, and overrun large portions of the Eastern Roman empire based on Constantinople.   Urbanization became one of the hallmarks of the Moslem world.   Many cities would be conquered, such Alexandria or Antioch, others would be destroyed such as the Sassanian capital, Ctesiphon, and yet others would also be founded, among them Fustat, and later Cairo, al Kufah, and Basrah, as well as Baghdad, and Rayy (later to become the seed of Tehran), together with Kairouan  and  Cordova (as a capital) in the west.   By 900, the Moslem world had the densest urban network and on the eve of the modern era was a principal precinct of the world system.




The four regional processes in Eurasia just described for the classical era each had their own developmental trajectory but they were not in fact isolated phenomena but rather were in several ways entangled, though seldom in a complete fashion.   They are best thought of as a “world city system” lined by two main lines of communication:  the overland Silk Roads, via Central Asia, and the maritime Spice Roads, via the Indian Ocean, and both in effect bound East Asia to the Mediterranean.   That means that the world cities basically formed one system, and that the chief routes were links among world cities.    These were not just trade routes but also the paths taken by new ideas and social innovations such as Buddhism.


The one area of significant urban development that stands apart was in the Americas, between 400 and 800 in particular, when we find cites seemingly meeting our criteria in Mexico (Teotihuacan), in the Mayan lands (Tical, Caracol) , and possibly even in Peru, a conceivable nucleus of a regional city system.   But the system was short-lived and largely collapsed after 800.  David Webster (2002:208)  questions the “urban” character of the Mayan cities in particular, and suggests that they were “king-places” or “regal-ritual” cities whose single-functionality was one reason for the fragility of the system when exposed to environmental stress and persistent warfare.


Modern world     For the modern era, we return to a unitary vision because our threshold criterion rises to one million, in other words, to “millionaire cities” that, at least initially, were few in number. Table 1 summarizes this information.


Table 1:  Modern “millionaire” world cities

             (cities of one million inhabitants or more)























































New York

St. Petersburg





























The table depicts a process that in the course of one millennium has raised the population of modern world cities from one to some 300, a rate of urban expansion never seen before.   What is more, most of that expansion occurred in the last one-two centuries.  


To start with, the urban landscape at the turn to the second millennium continued as it was in 900, with a central role for the Moslem world, and with strong Chinese participation.   But then, soon after 1200, disaster struck.   In a space of two-three generations the Mongols captured all the then ‘millionaire’ cities, and seized control of the Silk Roads even while laying waste to North China and Central Asia, destroying, and massacring the inhabitants of,  i.a. Beijing, Merv, Samarkand, Herat, and Baghdad.   When they faded away, one century later, this was still the “Asian age” but the spirit has gone out of it, and the Silk Roads withered on the vine.


Viewing  the table of modern world cities we might still consider it showing an “Asian age” right up to 1800 because all the world’s major cities were then Asian, if not actually Chinese.   But on closer inspection that is less of an indicator of wealth and power than a symptom of stagnation  because the table shows, before 1800, no evidence of growth, only some form of musical chairs.   The growth that mattered was happening at the other end of Eurasia, in Western Europe but it was, for a while, still under the radar.   The growth factor initially were the city states, Genoa and Venice and in the Low Countries whose experience was by 1500 translated into that of nation-states, Portugal, Spain, and the Dutch Republic, assuming increasingly important global roles spearheaded by their enterprising cities:  Lisbon, Antwerp and Amsterdam  (Braudel 1984; Tilly and Blockmans 1994).   These were experiments in new forms of global organization that reached maturity with the experience of Britain, and that is when London breaks through to the “millionaire” league, by 1800.  They would also suggest that innovation is not just a product of mere city size.


It is the list for 1900 that offers a clear picture of the urban structure that shaped what we tend to think as the world of the 19th and 20th centuries:  the strong British stake earned in the Industrial Revolution, with London, Manchester, Birmingham, Glasgow;  the then new, and rising, United States’ challenge, with New York, Chicago, Boston and Philadelphia;  the ‘traditional’ great powers of ‘old’ Europe, with Paris, Berlin, Vienna, Moscow, and St. Petersburg, and somewhat marginally, Tokyo, Beijing, and Calcutta.


But that was then, and by 2000 the urban landscape was much changed.   The figure reported in Table 1, some 300 cities for that year, means that the population of world cities became too numerous to be shown in that small table.   In that year, all the world’s regions were now fully represented.   The largest single share was that of East Asia that, at over 1/3, was not really surprising; for it is about what it was in the classical era.   Overall, the world city system had moved into a new configuration and scholars are now at work to clarify the nature and composition of that new system (see i.a. King and Taylor 1995).


3.         Major trends


The powerful process presented in this overview has been the steady and unrelenting urbanization on a world scale.   That is, over the past several millennia, cities have emerged and have grown bigger and better, and their weight in the social-make-up of the human species has kept on rising.   Such that, at the start of the 21st century, one half of the world’s population was living in cities, and ten per cent in world cities alone (as compared to about one per cent in the ancient, and two per cent in the classical worlds).   What is more, because of the concurrent rise, in the same time frame, of the overall population of this planet, from maybe six million to over six billion, the absolute number of people living in cities is of course the highest ever.


The world cities just described have composed, over the millennia, the center of the emerging world system, but that also means that their story can never constitute the entire picture, for such a story would neglect all those living outside the great urban centers, those in the peripheries and in the ‘hinterlands’ of  world history.   In fact, the process of interaction between the centers and the peripheries has been a  major and recurrent feature of the world system.   Thousand-year long sequences of ‘concentration’, that brought city building, may be documented to have alternated with periods of “redistribution” that show stable if not stagnant urban conditions. and these, in the past two instances, gave rise to Dark Ages (Modelski and Thompson 2002).   This raises the question whether the urban expansion and concentration recently and so prominently underway might not be heralding the onset of a new ‘age of redistribution’.


The world cities here presented might usefully be thought of as making up a complex, interlinked system, in a condition that might be termed “urban entanglement”.   In approaching this system, the useful starting point is the presumption of connectivity; in other words, what needs to be demonstrated is isolation and lack of contact rather than the reverse that is currently the standard.   If world cities have formed a system, then that system might be thought of as having emergent properties over and above the characteristics of its members, such as an alternation of phases of concentration, and redistribution, just mentioned.


More generally, the world city system might be viewed as a component of world system evolution.   The eras of world history (ancient, classical, modern) that helped us to organize the presentation of our material might be viewed as phases of that process (Modelski 2000), and we have shown that each such era displays quantitative features (such as changes in magnitude) as well as qualitative changes (such as scope of the network).   Our study therefore suggests that world urbanization that is a key component of globalization might also usefully be studied as an evolutionary process.


George Modelski


Prepared for the  Encyclopedia of World History”.



Further Reading


Abu-Lughod, J.   (1989)   Before European Hegemony:  The World System A.D.1250 -1350


New York:  Oxford University Press.               


Adams,  R. McC.   (1981)   Heartland of Cities, Chicago:  Chicago University Press.


Bairoch,  P.   (1988)   Cities and Economic Development,  Chicago:  Chicago University Press.


Braudel,  F.   (1984)   Civilization and Capitalism 15th to 18th Century;Volume 3:   The Perspective of the


World,  London:  Collins;    Parts 2 and 3: The City-centered Economies of the European Past.


Chandler, T.  (1987)  Four Thousand Years of Urban Growth: An Historical Census,  Lewiston: 


St. Gavid’s.


Childe, G.V.   (1950)  “The Urban Revolution”,   Town Planning Review, 21(1),3-17.


Hourani, A.  (1991)    A History of the Arab Peoples,   Cambridge:  Harvard University Press,


Part II:  Arab-Muslim Societies.


------- and  S.M. Stern eds.   (1970)   The Islamic City,  Oxford:  Oxford University Press.


Kenoyer,  J.   (1998)   Ancient Cities of the Indus Valley Civilization,  Karachi:  Oxford University Press.


King, P.L  and P. Taylor eds. (1995)   World Cities in a World-System,   Cambridge:  Cambridge


University Press.  


Modelski, G.  (2000)   “World System Evolution” at pp.24-53 of R. Denamark,  R., J. Friedman, B. Gills,


and G. Modelski eds.  World System History:  The Social Science of Long-term Change,


New York:  Routledge.


-----          (2003)   World Cities: -3000 to 2000, Washington:  Faros 2000.


-----      and  W. R. Thompson  (2002)   “Evolutionary Pulsations in the World System” at pp.177-196


of  S. C. Chew and J.D. Knotterus eds.   Structure, Culture, and History:  Recent Issues in Social


Theory,   Lanham: Rowman and Littlefield.


Mumford,  Lewis   (1938,1970)   The Culture of Cities,  New York:  Harcourt Brace.

Nichols, D. and Y.H. Charlton eds. (1997)   The Archaeology of City-States,  Washington:  Smithonian.

Toynbee,  A.  ed.  (1967)  Cities of Destiny,   New York:  McGraw Hill.

Tilly, Ch. and W.P. Blockmans eds.   (1994)   Cities and the Rise of States in EuropeA.D.1000 to 1800,   

Boulder:  Westview Press.

Webster,  D.   (2002)   The Fall of the Ancient Maya,   New York: Thames and Hudson.

Wheatley,  P.  (1971)  The Pivot of the Four Quarters:  A Preliminary Inquiry into the Origins and the

 Character of the Ancient Chinese City,  Chicago:  Aldine.