This essay discusses conceptual issues that arise from the study of human social change. The comparative and evolutionary world-systems perspective is explained as a theoretical research program for studying long-term social change. This approach employs an anthropological framework of comparison for studying world-systems, including those of hunter-gatherers. Problems of spatially bounding whole human interaction networks are addressed, and the utility of a comparative approach to the study of hierarchical relations among human polities (core/periphery relations) is examined. The hypothesis of semiperipheral development is explained, and criteria for empirically identifying semiperipheral regions are specified. World history and global history are the most important evidential bases, along with prehistoric archaeology, for the comparative study of world-systems. Getting the grounds of comparison right by correctly conceptualizing the spatial units of analysis and paying careful attention to core/periphery relations are crucial issues in the effort to comprehend and explain the development of world-systems.
The sociology of development, broadly construed, is the study and explanation of human social change in general, including the emergence of social complexity and hierarchy during, and since, the Stone Age. Most of the social science literature has focused on the transition from “tradition” to “modernity,” concepts that are usually understood in terms of characteristics of national societies such as urbanization, industrialization, and the demographic transition. The rise of globalization studies has shifted the focus to an emerging world society in which nation-states are understood as interacting organizations that claim sovereignty over a delimited territory and national societies are seen to be highly dependent on each other and on the larger global system. The world-systems perspective on modernity claims that this high degree of interdependence is not a recent phenomenon and that an important dimension of the global system has been, and continues to be, its stratification structure, which is organized as a core/periphery hierarchy in which some national societies have far more power and wealth than others.
The world-systems perspective emerged during the world revolution of 1968 and the antiwar movement. These produced a generation of scholars who saw the peoples of the global South (then called the “Third World”) as more than an underdeveloped backwater. It became widely understood that a global power structure existed and that the peoples of the noncore had been active participants in their own liberation. The history of colonialism and decolonization was seen to have importantly shaped the structures and institutions of the whole global system. A more profound awareness of Eurocentrism was accompanied by the realization that most national histories had been written as if each country were on the moon. World history courses had introduced high school and college students to the stories of non-European civilizations and global history. The nation-state as an inviolate, pristine unit of analysis was now seen to be an inadequate model for the sociology of development.
This awareness of a larger global context spread widely as globalization itself became a focus of public and scholarly discourse. Some versions claimed that the world was now flat and that international hierarchy was a thing of the past that had been transcended by instantaneous communication and the world market. Some of the global historians claimed, along with the theorists of a new global stage of capitalism, that the world had, in the last decades of the twentieth century, transitioned from a set of weakly linked national economies to a recently emerged single global economy. But from a world-systems perspective, waves of integration (globalization) have occurred throughout human history (Chase-Dunn 1999; Chase-Dunn and Lerro 2014), and the contemporary reproduction of global inequalities continues to make global collective action extremely difficult. We agree with most of Michael Mann's (2013:3) criticisms of the “hyperglobalizers” who see a recent radical transformation to a completely different kind of global social order. Mann's (2013) analysis of four interrelated but relatively autonomous strands of political, military, economic, and ideological globalization since 1945 is substantially accurate despite his metatheoretical presumption that there is no single integrated global system.
In this essay we intend to clear up several common misunderstandings about the world-systems perspective,1 and we introduce and discuss the conceptual basis for comparing the contemporary global system with earlier small and middle-sized regional world-systems. It is our contention that the world-systems perspective is a still-emerging theoretical research program that has great potential for enhancing our comprehension of the causes of long-term social change and also serving as a framework that will be useful to world citizens who are trying to deal with the problems that our species has shaped for itself in the twenty-first century.
The two main conceptual issues we shall consider are
the spatial bounding of whole world-systems
The comparative world-systems perspective is a strategy for explaining social change that focuses on whole interpolity systems rather than single polities. The main insight is that important interaction networks (trade, information flows, alliances, and fighting) have woven polities and cultures together since the beginning of human sociocultural evolution.2 Explanations of social change need to take whole interpolity systems (world-systems) as the units that evolve. But interpolity interaction networks were rather small when transportation was mainly a matter of carrying goods on one's back or in small boats. Globalization, in the sense of the expansion and intensification of larger interaction networks, has been increasing for millennia, albeit unevenly and in waves (Chase-Dunn 2006; Beaujard 2010; Jennings 2010).
World-systems are whole systems of interacting polities and settlements.3 Systemness means that these polities and settlements are interacting with one another in important ways: interactions are two-way, necessary, structured, regularized, and reproductive. Systemic interconnectedness exists when interactions importantly influence the lives of people and are consequential for social continuity or social change. All premodern world-systems extended over only parts of the earth. The word world refers to the importantly connected interaction networks in which people live, whether these are spatially small or large.
There is also the question of endogenous (internal) systems versus exogenous (external) impacts. The notion of systemness requires distinction between endogenous processes that are regularly interactive and systemic and exogenous impacts that may have large effects on a system but are not part of that system. The diffusion of genetic materials and technologies can have profound long-distance effects even though there are no regularized or frequent interactions. But single events that have such consequences should not be considered to be part of a sociocultural system. Climatic changes often have important impacts on human societies, but we do not try to include them as endogenous variables in our models of social systems until climate change becomes anthropogenic. Analogously, collisions with large asteroids have had huge consequences for biological evolution, but biologists do not include these as endogenous factors in biological evolution. Similarly, long-distance diffusion of ideas, technologies, plants, and animals is an important process that needs to be studied in its own right, and its impact on local systems must be acknowledged and understood, but models of sociocultural change must distinguish between endogenous processes and exogenous impacts. The sorting out of endogenous versus exogenous factors is the basic theoretical and empirical issue that must be addressed in specification of systemic wholeness.
Only the modern world-system has become a global (earthwide) system composed of national societies and their states. It is a single global economy composed of international trade and capital flows, transnational corporations that produce products on several continents, and all the economic transactions that occur within countries and at local levels. The whole world-system is more than just international relations. It is the whole system of human interactions. The world economy is now all the economic interactions of all the people on earth, not just international trade and investment.
The modern world-system is structured politically as an interstate system—a system of competing and allying states. Political scientists commonly call this the international system, and it is the main focus of the field of international relations. Some of these states are much more powerful than others, but the main organizational feature of the world political system is that it is multicentric. There is, as yet, no world state. Rather, there is a system of states. This is a fundamentally important feature of the modern system and of most earlier regional world-systems as well.
When we compare different kinds of world-systems it is important to use concepts that are applicable to all of them. Polity is a general term that means any organization with a single authority that claims control over a territory or a group of people.4 Polities include bands, tribes, and chiefdoms as well as states and empires. All world-systems are composed of multiple interacting polities. Thus we can fruitfully compare the modern interstate system with earlier interpolity systems in which there were tribes or chiefdoms but no states.5
So the modern world-system is now a global economy with a global political system (the modern interstate system). It also includes all the cultural aspects and interaction networks of the human population of the earth. Culturally the modern system is composed of several civilizational traditions (e.g., Islam, Christianity, Hinduism, Confucianism), nationally defined cultural entities or nations (which in turn are composed of class and functional subcultures, such as lawyers, technocrats, and bureaucrats), and the cultures of indigenous and minority ethnic groups within states. The modern system is multicultural in the sense that important political and economic interaction networks connect people who have rather different languages, religions, and other cultural features. Most earlier world-systems have also been multicultural.6
But the modern system also has a single geoculture that has been emerging since the late eighteenth century in the context of the multicultural situation depicted above (Wallerstein 2011; Meyer 2009). This geoculture is most importantly structured by the core, but it has also evolved in the context of a series of world revolutions in which the peoples of the noncore have contested the global power structure, and these have had important effects on the content of the geoculture.
One of the important systemic features of the modern system is the rise and fall of hegemonic core powers—the so-called “hegemonic sequence” (Wallerstein 1984; Chase-Dunn 1998). A hegemon is a core state that has a significantly greater amount of economic power than any other state and that takes on the political role of system leader. In the seventeenth century the Dutch Republic performed the role of hegemon in the Europe-centered system, while Great Britain was the hegemon of the nineteenth century, and the United States has been the hegemon in the twentieth century. Hegemons provide leadership and order for the interstate system and the world economy. But the normal operating processes of the modern system—uneven economic development and competition among states—make it difficult for hegemons to sustain their dominant positions, and so they tend to decline. Thus the structure of the core oscillates between unipolar hegemony and a situation in which several competing core states have roughly similar amounts of power and are contending for hegemony—that is, multipolar hegemonic rivalry (see figure 1).
So the modern world-system is composed of states that are linked to one another by the world economy and other interaction networks. Earlier world-systems were also composed of polities, but the interaction networks that linked these polities were not intercontinental in scale until the expansion of the Indian Ocean–centered system and then European expansion to the Americas in the long sixteenth century CE. Before that, world-systems were smaller regional affairs. But these had been growing in size with the expansion of trade networks and long-distance military campaigns for millennia (Bentley 1993; Beaujard 2005).
The notion of core/periphery relations has been a central concept in both the modern world-system perspective (Wallerstein  2011) and the comparative world-systems perspective (Chase-Dunn and Hall 1997). World-systems are systems of interacting polities, and they often (but not always) are organized as interpolity hierarchies in which some polities exploit and dominate other polities.7Chase-Dunn and Hall (1997) redefined the core/periphery distinction to make it more useful for comparing the modern world-system with earlier regional world-systems.
The intent of using the word core rather than center is to clearly signal the awareness that most interpolity hierarchies are multicentric (figure 2). There is a region or zone at the top layer of the hierarchy that is occupied by a set of allying and competing polities. All hierarchical world-systems go through a cycle of rise and fall in which a most powerful polity in the core grows and then declines. But most systems remain multicentric. The exceptions are what Chase-Dunn and Hall (1997; see also Scheidel 2009) call “core-wide empires,” in which a single polity conquers and rules an entire core region. These have been rare. Large empires that claim to control the universe rarely control all the core polities in their interaction system. Even the Roman Empire never conquered the Parthian Empire.
The modern world-system has been, and is still, importantly structured as a core/periphery hierarchy in which some regions contain economically and militarily powerful states while other regions contain polities that are much less powerful and less developed. The countries that are called “advanced,” in the sense that they have high levels of economic development, skilled labor forces, high levels of income, and powerful, well-financed states, are the core powers of the modern system. The modern core includes the United States and the countries of Europe, Japan, Australia, New Zealand, and Canada.
In the contemporary periphery we have relatively weak states that are not strongly supported by the populations within them and have little power relative to other states in the system. The colonial empires of the European core states dominated most of the modern periphery until waves of decolonization swept away the colonial empires starting in the late eighteenth century. Peripheral regions are economically less developed in the sense that the economy is composed of relatively less capital-intensive forms of agriculture and industry. Some industries in peripheral countries, such as oil extraction or mining, may be capital intensive, but these sectors are often controlled by core capital.
In the past, peripheral countries have been primarily exporters of agricultural and mineral raw materials. But even when they have developed some industrial production, this has usually been less capital intensive and has used less skilled labor than production processes in the core. The contemporary peripheral countries are most of the countries in Africa and many of the countries in Asia and Latin America—for example, Bangladesh, Senegal, Haiti, and Bolivia.
The core/periphery hierarchy in the modern world-system is a system of stratification in which socially and ecologically structured inequalities are reproduced by the institutional features of the system. The periphery is not “catching up” with the core. Rather, both core and peripheral regions are developing, but most core states are staying well ahead of most peripheral states. Though there is some upward and downward mobility, the overall structure of inequality has remained quite stable. There is also a stratum of countries in between the core and the periphery that is called the semiperiphery. The semiperiphery in the modern system includes countries that have intermediate levels of economic development or a balanced mix of developed and less developed regions. The semiperiphery includes large countries that have political/military power as a result of their large size (e.g., Indonesia, Mexico, Brazil, China, India) and smaller countries that are relatively more developed than those in the periphery (e.g., South Africa, Taiwan, South Korea; see figure 3).
The exact boundaries between the core, semiperiphery, and periphery are unimportant because the main point is that there is a continuum of economic and political/military power that constitutes the core/periphery hierarchy. It does not matter exactly where we draw lines across this continuum in order to categorize countries. Indeed, we could as well make four or seven categories instead of three. The categories are only a convenient terminology for pointing to the fact of international inequality and for indicating that the middle of this hierarchy may be an important location for processes of social change.
There have been a few cases of upward and downward mobility in the core/periphery hierarchy, though most countries simply run hard to stay in the same relative positions that they have long had. The most spectacular case of upward mobility in the modern core/periphery hierarchy is the United States. Over the last 300 years the territory that became the United States moved from being outside of the Europe-centered system (a separate continent containing several regional world-systems), to the periphery in the colonial era, to the semiperiphery in the first half of the nineteenth century, to the core by 1880, and then to the position of hegemonic core state in the twentieth century, and now its hegemony is slowly declining. An example of downward mobility is the United Kingdom of Great Britain—the hegemon of the nineteenth century and now just another core society.
The contemporary global stratification system is a continuum of economic and political-military power that is reproduced by the normal operations of the system. In such a hierarchy there are countries that are difficult to categorize. For example, most oil-exporting countries have very high levels of GNP per capita, but their economies do not produce high-technology products that are typical of core countries. They have wealth but not development. The point here is that the categories (core, periphery, and semiperiphery) are just a convenient set of terms for pointing to different locations on a continuous and multidimensional hierarchy of power. It is not necessary to have each case fit neatly into a box. The boxes are only conceptual tools for analyzing the unequal distribution of power among countries.
When we use the idea of core/periphery relations for comparing very different kinds of world-systems we need to broaden the concept and to make an important distinction (see below). But the most important point is that we should not assume that all world-systems have core/periphery hierarchies just because the modern system does. It should be an empirical question in each case as to whether core/periphery relations exist. Not assuming that world-systems have core/periphery structures allows us to compare very different kinds of systems and to study how core/periphery hierarchies themselves have emerged and evolved.
In order to do this it is helpful to distinguish between core/periphery differentiation and core/periphery hierarchy. Core/periphery differentiation means that societies with different degrees of population density, polity size, and internal hierarchy are interacting with one another. As soon as we find village dwellers interacting with nomadic neighbors we have core/periphery differentiation. Core/periphery hierarchy refers to the nature of the relationships between societies. This kind of hierarchy exists when some societies are exploiting or dominating other societies. Examples of intersocietal domination and exploitation would be the British colonization and deindustrialization of India, or the conquest and subjugation of Mesoamerica and the Andean region by the Spaniards. Core/periphery hierarchy is not unique to the modern Europe-centered world-system of recent centuries. Both the Roman and the Aztec empires conquered and exploited peripheral peoples as well as adjacent core states.
Distinguishing between core/periphery differentiation and core/periphery hierarchy allows us to deal with situations in which larger and more powerful societies are interacting with smaller ones but are not exploiting them. It also allows us to examine cases in which smaller, less dense societies may be exploiting or dominating larger societies. This latter situation definitely occurred in the long and consequential interaction between the nomadic horse pastoralists of central Asia and the agrarian states and empires of China and western Asia. The most famous case was that of the Mongol Empire of Genghis Khan, but confederations of central Asian steppe nomads managed to extract tribute from agrarian states long before the rise of Mongols (Barfield 1989; Honeychurch 2013).
The question of core/periphery status also needs to be considered with regard to different spatial scales of interaction. Chase-Dunn and Hall (1997) note that regional world-systems may have important interaction networks that have different spatial scales (see below). They adopt a “place-centric” approach to spatially bounding interaction networks that begins from a focal settlement or polity. The network is spatially bounded according to how many indirect links are needed to include all the interactions that have an important impact on the reproduction or transformation of institutions at the focal place. Bulk goods networks (BGNs), which were usually relatively small, and interpolity political-military networks (PMNs), which are usually somewhat larger, are important in all systems; and prestige goods networks (PGNs), which are often much larger, are important in some systems but not others. The issue of core/periphery status always needs to be asked for BGNs and PMNs, and also for PGNs if prestige goods played an important role in sociocultural reproduction or change.
SPATIAL BOUNDARIES OF WORLD-SYSTEMS
Whole interaction networks are composed of regular and repeated interactions among individuals and groups. Interaction may involve trade, communication, threats, alliances, migration, marriage, gift giving, or participation in information networks (INs) such as radio, television, telephone conversations, and e-mail. Conflict is also an important form of sociation both within and between polities. Warfare, ranging from ritual contests to ethnocide, has been an important player in the group selection process that produces sociocultural evolution (Morris 2014; Turchin 2011). Important interaction networks are those that affect peoples' everyday lives, their access to food and necessary raw materials, their conceptions of who they are, and their security from, or vulnerability to, threats and violence. World-systems are fundamentally composed of interaction networks.
One big difference between the modern world-system and earlier systems is the spatial scale of different types of interaction networks. In the modern global system most of the important interaction networks are themselves global in scale. But in earlier smaller systems there was a significant difference in spatial scale between networks in which food and basic raw materials were exchanged and much larger networks of the exchange of prestige goods or luxuries. Different kinds of important interaction had different spatial scales. Food and basic raw materials we call “bulk goods” because they have a low value per unit of weight. It is uneconomical to carry bulk foods very far under premodern conditions of transportation.
Imagine that the only type of transportation available is people carrying goods on their backs (or heads). This is a situation that actually existed everywhere until the domestication of beasts of burden. Under these conditions a person can carry, say, 30 kilograms of food. Imagine that this carrier is eating the food as she or he goes. So after a few days walking all the food will be consumed. This is the economic limit of food transportation under these conditions of transportation. This does not mean that food will never be transported farther than this distance, but there would have to be an important reason for moving it beyond its economic range.
Prestige goods are items that have great value and small size, or items that can easily be transported long distances intact and typically hold or increase their value in transit (e.g. spices, jade, jewels, or bullion). Prestige goods have a much larger spatial range than do bulk goods because a small amount of such a good may be exchanged for a great deal of food. This is why PGNs are normally much larger than BGNs. A network does not usually end as long as there are people with whom one might trade. Indeed, most early trade was what is called “down-the-line” trade, in which goods were passed from group to group. For any particular group the effective extent of its trade network is that point beyond which nothing that happens will affect the group of origin.
In order to bound interaction networks, we need to pick a place from which to start—the so-called “place-centric approach.” If we go looking for actual breaks in interaction networks we will usually not find them, because almost all groups of people interact with their neighbors. But if we focus upon a single settlement, for example the indigenous village of Onancock on the eastern shore of the Chesapeake Bay before the arrival of the Europeans in the seventeenth century CE (near the boundary between what are now the states of Virginia and Maryland in the United States), we can determine the spatial scale of the bulk goods interaction network by finding out how far food moved to and from our focal village.8 Food came to Onancock from some maximum distance. A bit beyond that were groups that were trading food to groups that were directly sending food to Onancock. If we allow two indirect jumps, we are probably far enough from Onancock so that no matter what happens (e.g., a food shortage or surplus), it would not have affected the supply of food in Onancock. This outer limit of Onancock's indigenous BGN probably included villages at the very southern and northern ends of the Chesapeake Bay.
Onancock's PGN was much larger because prestige goods move farther distances. Indeed, copper that was in use by the indigenous peoples of the Chesapeake may have come from as far away as Lake Superior. In between the size of BGNs and PGNs were the PMNs, the interaction networks in which polities made war and allied with one another. In the case of the Chesapeake world-system, at the time of the arrival of the Europeans in the sixteenth century Onancock was part of a district chiefdom in a PMN of multivillage chiefdoms. Across the bay on the western shore were at least two larger polities, the Powhatan and the Conoy paramount chiefdoms (Rountree 1993). These were core chiefdoms that were collecting tribute from a number of smaller district chiefdoms. Onancock was part of an interchiefdom system of allying and war-making polities. The boundaries of that network included some indirect links, just as the trade network boundaries did. Thus the PMN of which Onancock was the focal place extended to the Delaware Bay in the north and into what is now the state of North Carolina to the south. Information, like a prestige good, is light relative to its value. Information may travel far along trade routes and beyond the range of goods. Thus INs are usually as large as, or even larger than, PGNs.
A general picture of the spatial relationships between different kinds of interaction networks is presented in figure 4. The actual spatial scale of important interaction needs to be determined for each world-system we study, but figure 4 shows what is generally the case—that BGNs are smaller than PMNs and that these are in turn smaller than PGNs and INs.
As defined above, world-systems have grown from small to large over the past twelve millennia as polities and interpolity systems have gotten larger, more complex, and more hierarchical. This spatial growth of systems has involved the expansion of some and the incorporation of some into others. The processes of incorporation have occurred in several ways as systems distant from one another have linked their interaction networks. Because interaction networks are of different sizes, the largest ones come into contact first. Thus information and prestige goods link distant groups long before they participate in the same PMNs or BGNs. The processes of expansion and incorporation brought different groups of people together and made the organization of larger and more hierarchical societies possible. It is in this sense that globalization has been going on for thousands of years.
Using the conceptual apparatus for spatially bounding world-systems outlined above we can construct spatio-temporal chronographs for how the interaction networks of the human population changed their spatial scales to eventuate in the single global political economy of today. figure 5 uses PMNs as the unit of analysis to show how a “Central” PMN, composed of the merging of the Mesopotamian and Egyptian PMNs in about 1500 BCE, eventually incorporated all the other PMNs into itself.
Janet Abu-Lughod's 1989 study of the multicentric Eurasian world-systems of the thirteenth century was a very valuable and inspiring contribution. The most important theoretical issue it brought forth concerns differences between her approach and that of Immanuel Wallerstein ( 2011) regarding the spatial bounding of world-systems (see Boles 2012). Abu-Lughod used interaction networks, mainly long-distance trade, whereas Wallerstein used a hierarchical regional division of labor (see Wallerstein 1995). The upshot of this dispute is that both conceptual approaches have proven to have productive uses for explaining the causes of world-systems evolution. Wallerstein's ( 2011: chap. 6) fascinating analysis of why Russia was an “external arena” in the sixteenth century even though it was exporting the same goods to Europe as were being exported by peripheralized Poland, is a fascinating case in favor of his method of bounding. But Abu-Lughod's focus on trade, especially when combined with a consideration of geopolitical interaction among polities (see Wilkinson 1987; Chase-Dunn and Jorgenson 2003), is also a fruitful method that facilitates the comparative study of regional world-systems small and large. Abu-Lughod also helped to clear the way forward in world-systems analysis by rejecting the idea of the “ancient hyperglobalists,” such as Frank and Gills (1994), Modelski (2003), and Lenski (2005) that there has always been a single global (earthwide) system.9 She agreed with Wallerstein that as we go back in time there are multiple regional whole systems that should be studied separately and compared. Things would be much simpler if it made sense to use the whole earth as the unit of analysis since humans came out of Africa. The ancient hyperglobalists are correct that there has been a single global network for millennia because all human groups interact with their neighbors and so are indirectly connected with all others. But this ignores the issue of the fall-off of interaction effects discussed above. Frank and Gills (1994) contended that there had been a single global system since the rise of cities and states in Mesopotamia, though later they admitted that the Americas were largely disconnected from Afro-Eurasia before 1492 CE. But if we read Frank and Gills as studying the important continuities of the Central PMN and the Eurasian PGN, as discussed above, much of their analysis of core/periphery relations is quite valuable. They also raised the important issue of the evolution of modes of accumulation, claiming that there had been a “capitalist-imperialist” mode in the Bronze Age with alternating periods in which tribute taking and market-based profit making had been predominant (see also Ekholm and Friedman  1982). This debate is far from over.
WORLD-SYSTEM CYCLES: “RISE AND FALL” AND PULSATIONS
Comparative research reveals that all world-systems exhibit cyclical processes of change. There are two major cyclical phenomena: the rise and fall of large polities, and pulsations in the spatial extent and intensity of trade networks. Rise and fall corresponds to changes in the degree of centralization of political/military power in a set of polities—an “international” system. It is a question of the relative distribution of power across a set of interacting polities.
All world-systems in which there are hierarchical polities experience a cycle in which relatively larger polities grow in power and size and then decline. This applies to interchiefdom systems as well as interstate systems, to systems composed of empires, and to the modern rise and fall of hegemonic core powers (e.g., Britain and the United States). Though very egalitarian and small-scale systems such as the sedentary foragers of Northern California do not display a cycle of rise and fall, they do experience exchange network pulsations (Chase-Dunn and Mann 1998:140–41).
All systems, including even very small and egalitarian ones, exhibit cyclical expansions and contractions in the spatial extent and intensity of exchange networks. We call this sequence of trade expansion and contraction pulsation. Different kinds of trade (especially bulk goods trade vs. prestige goods trade) usually have different spatial scales. In the modern global system large trade networks cannot get spatially larger because they are already global in extent. But they can get denser and more intense relative to smaller networks of exchange. A good part of what has been called globalization is simply the intensification of larger interaction networks relative to the intensity of smaller ones. This kind of integration is often understood to be an upward trend that has attained its greatest peak in recent decades of so-called global capitalism. But research on changes in the level of trade and investment globalization shows that there have been two and a half recent waves of integration, one in the last half of the nineteenth century, a small one between 1900 and 1929, and a stair-stepped upward trend (half of a cycle) since World War II (Chase-Dunn, Kawano, and Brewer 2000).
The simplest hypothesis regarding the temporal relationships between rise and fall and pulsation is that they occur in tandem. Whether or not this is so, and how it might differ in distinct types of world-systems, are problems that are amenable to empirical research.
Chase-Dunn and Hall (1997) have contended that the causal processes of rise and fall differ depending on the predominant mode of accumulation. One big difference between the rise and fall of empires and the rise and fall of modern hegemons is in the degree of centralization achieved within the core. Tributary systems alternate between a structure of multiple and competing core states and a structure of corewide (or nearly corewide) empires. In the modern interstate system, hegemons rise and fall but the efforts that have been made to take over the other core states to form a corewide empire have always failed, mainly because modern hegemons are pursuing a capitalist rather than a tributary form of accumulation.
Analogously, rise and fall works somewhat differently in interchiefdom systems because the institutions that facilitate the extraction of resources from distant groups are less fully developed in chiefdom systems. David G. Anderson's (1994) study of the rise and fall of Mississippian chiefdoms in the Savannah River valley provides an excellent and comprehensive review of the anthropological and sociological literature about what Anderson calls “cycling”—the processes by which a chiefly polity extended control over adjacent chiefdoms and erected a two-tiered hierarchy of administration over the tops of local communities. At a later point these regionally centralized chiefly polities disintegrated back toward a system of smaller and less hierarchical polities.
Chiefs relied more completely on hierarchical kinship relations, control of ritual hierarchies, and control of prestige goods imports than do the rulers of true states. These chiefly techniques of power were all highly dependent on normative integration and ideological consensus. States developed specialized organizations for extracting resources that chiefdoms lacked—standing armies and bureaucracies. And states and empires in the tributary world-systems were more dependent on the projection of armed force over great distances than modern hegemonic core states have been. The development of commodity production and mechanisms of financial control, as well as further development of bureaucratic techniques of power, have allowed modern hegemons to extract resources from faraway places with much less overhead cost.
The development of techniques of power has made core/periphery relations ever more important for competition among core powers and has altered the way in which the rise-and-fall process works in other respects. Chase-Dunn and Hall (1997: chap. 6) argued that population growth in interaction with the environment and changes in productive technology and social structure produce sociocultural evolution that is marked by cycles and periodic jumps This is because both internal instabilities and environmental fluctuations cause each world-system to oscillate around a central tendency. Occasionally, on one of the upswings, people solve systemic problems in a new way that allows substantial expansion. We want to explain expansions, evolutionary changes in systemic logic, and collapses. That is the point of comparing world-systems.
The multiscalar regional method of bounding world-systems as nested interaction networks outlined above is complementary with a multiscalar temporal analysis of the kind suggested by Fernand Braudel's (1972, 1984) work. Temporal depth, the longue durée, needs to be combined with analyses of short-run and middle-run processes if we are to fully understand social change.
Figure 6 below depicts the coming together of the East Asian and the West Asian/Mediterranean systems. Both the PGNs and the PMNs are shown, as are the pulsations and rise-and-fall sequences. The PGNs linked intermittently and then joined. The Mongol conquerors linked the PMNs briefly in the thirteenth century, but the Eastern and Western PMNs were not permanently linked until the Europeans and Euro-Americans established Asian treaty ports in the nineteenth century.
The semiperiphery concept was originally developed to study the modern world-system (Wallerstein  2011). But it too has been expanded for use in comparing different kinds of systems. For Wallerstein the semiperiphery is a middle stratum in the global hierarchy that helps keep the system from breaking down because of polarization. But Chase-Dunn and Hall (1997) claimed that semiperipheral societies have often been agents of change in both the modern and earlier systems.
Hub theories of innovation have been popular among world historians (e.g., J. McNeill and McNeill 2003; Christian 2004) and human ecologists (Hawley 1950). These hold that new ideas and institutions emerge in central settlements where information crossroads are located. The mixing and recombination of ideas and information facilitates the emergence of new formulations. The hub theory is undoubtedly partly correct, but it cannot explain some of the long-term patterns of human sociocultural evolution, because if a large information crossroad were able to outcompete all contenders then the original information hub would still be the center of the world. But that is not the case. We know that cities and states first emerged in Mesopotamia around 5,000 years ago. Mesopotamia is now Iraq. Mesopotamia had 100 percent of the world's largest settlements and the most powerful polities on earth in the Early Bronze Age. Now it has none of these. All of the regional world-systems have undergone a process of uneven development in which the old centers were replaced by new centers out on the edge.10
According to Chase-Dunn and Hall (1997), it has most often been polities out on the edge (in semiperipheral regions) that have transformed the institutional structures and accomplished the upward sweeps. This hypothesis is part of a larger claim that it is people in semiperipheral locations who usually play the transformative roles that cause the emergence of greater sociocultural complexity and hierarchy within world-systems. This hypothesis of semiperipheral development is an important justification supporting the claim that world-systems rather than single polities are the right unit of analysis for explaining human sociocultural evolution.
The hub theory of innovation does not well account for the spatially uneven nature of sociocultural evolution. The cutting edges of power and scale move. Polities out on the edge that are able to conquer large territories or to rewire networks and to expand their spatial scale often transcend old centers. This is due to two things. Competitive success is not only about where new and adaptive technologies, ideas, and organizational forms are created. It is also importantly about which polities invest in and implement these innovations. Innovations do often emerge outside of large networks' nodes. New weapons, military techniques, and religions often emerge from peripheral or semiperipheral regions (e.g., Hamalainen 2008). But implementation is more important than innovation in explaining the phenomenon of semiperipheral development. Polities in semiperipheral locations often implement innovations that originated elsewhere. This is an important part of the explanation of semiperipheral development.
Semiperipheral development has taken various forms: semiperipheral marcher chiefdoms, semiperipheral marcher states, semiperipheral capitalist city-states, the peripheral and then semiperipheral position of Europe in the larger Afro-Eurasian PGN, modern semiperipheral nation-states that have risen to hegemony (the Netherlands, the United Kingdom, and the United States), and contemporary peoples in semiperipheral locations that are engaging in, and supporting, novel and potentially transformative movements.
Several possible processes might account for the phenomenon of semiperipheral development. Randall Collins (1981) has argued that the phenomenon of marcher states conquering other states to make larger empires is largely due to the “marcher state advantage.” Being out on the edge of a core region of competing states allows more maneuverability because it is not necessary to defend the rear. This geopolitical advantage allows military resources to be concentrated on vulnerable neighbors. Peter Turchin (2003) has argued that the relevant process is one in which group solidarity is enhanced by location on a “metaethnic frontier” in which the clash of contending cultures produces strong cohesion and cooperation within a frontier society, allowing it to perform great feats. Carroll Quigley (1961) distilled a somewhat similar theory from the works of Arnold Toynbee. Another factor affecting within-polity solidarity is the different degrees of internal stratification usually found in premodern world-systems between the core and the semiperiphery. Core polities develop old, crusty, and bloated elites who rely on mercenaries and “foreigners” as subalterns, while semiperipheral leaders are often charismatic heroes who are strongly supported by their soldiers and citizens. Less stratification often facilitates greater group solidarity. And this may be an important part of the semiperipheral advantage.
But Quigley also suggested another way in which the peoples of semiperipheral regions might be motivated to take risks with new ideas, technologies, and strategies. Semiperipheral polities are often located in ecologically marginal regions that have poor soil and little water or other geographical disadvantages. Patrick Kirch relies on this idea of ecological marginality in his depiction of the process by which semiperipheral marcher chiefs often are the conquerors that unify islandwide paramount chiefdoms in the Pacific (Kirch 1984). It is quite possible that all these features combine to produce what Alexander Gershenkron (1962) called “the advantages of backwardness” that allow some semiperipheral polities to transform and to dominate their world-systems.
As we have already said, the hypothesis of semiperipheral development claims that many of those innovations that make it possible for world-systems to get larger, more complex, and more hierarchical are created by peoples in semiperipheral locations and that some semiperipheral polities invest in, and implement, transformative innovations that are borrowed from core or peripheral societies.11 Some semiperipheral polities are involved in processes of rapid internal class formation and state formation and do not have large investments in, and commitments to, doing things the way they have been done in older core polities. They do not have institutional or infrastructural sunk costs. So they are freer to reinvent themselves, to implement new institutions and to experiment with new technologies.
There are several different important kinds of semiperipheries, and they not only transform systems but also often take over and become the new hegemonic core polity. We have already mentioned semiperipheral marcher chiefdoms. The societies that conquered and unified a number of smaller chiefdoms into larger paramount chiefdoms were usually from semiperipheral locations. Peripheral peoples did not usually have the institutional and material resources that would allow them to implement new technologies or organizational forms or to take over older core regions, though there were also some well-known “peripheral marcher polities” (e.g., the vast steppe confederacy organized by Genghis Khan that produced the Mongol Empire). It was mainly in the semiperiphery that core and peripheral social characteristics were likely to be recombined in new ways and where enough resources were available to allow significant investment in transformative instruments.
Much better known than semiperipheral marcher chiefdoms is the phenomenon of semiperipheral marcher states. Many of the largest empires were assembled by conquerors who came from semiperipheral polities. Well-known examples are the Achaemenid Persians, the Macedonians led by Philip and Alexander, the Romans, the Islamic Caliphates, the Ottomans, the Manchus, and the Aztecs (Alvarez et al. 2013).
But some semiperipheral peoples and polities transform institutions without taking over the interpolity system of which they are a part. The semiperipheral capitalist city-states operated on the edges of the tributary empires where they bought and sold goods in widely separate locations, encouraging peoples near and far to produce surpluses for trade. The Phoenician cities (e.g., Byblos, Sidon, Tyre, Carthage), as well as Malacca, Venice, Genoa, and the German Hanse cities, spread commodification by producing manufactured goods and trading them across great regions. Some cities even in the Bronze Age (e.g., Dilmun and Assur of the Old Assyrian city-state) specialized in long-distance trade. The semiperipheral capitalist city-states were agents of the development of markets and the expansion of trade networks, and so they helped to transform the world of the tributary empires without themselves becoming new core powers.12 These were the first capitalist states in which state power was mainly used to facilitate profit making rather than to extract taxes and tribute (Chase-Dunn et al. 2013).
Philippe Beaujard (2005:239) makes the point that core/periphery relations often involve coevolution. Even when exploitation and domination of the noncore by the core occurs, polities in both zones are altered and coevolve. In many systems in Afro-Eurasia and the Americas interactions between hunter-gatherers and farmers led to the emergence of polities that specialized in pastoralism (Lattimore 1940; Barfield 1989; Honeychurch 2013; Hamalainen 2008). Some of the pastoralists were exploited and dominated by core polities, but others turned the tables and were able to extract resources from agrarian states.
INDICATORS OF SEMIPERIPHERALITY
The Settlements and Polities (EmpCit) Research Working Group at the Institute for Research on World-Systems at the University of California–Riverside uses quantitative estimates of the population sizes of large cities and of the territorial sizes of polities to identify instances of “upsweeps” in which city and polity sizes significantly increased in scale ((Inoue et al. 2012, forthcoming). We also determine how many of the urban and polity upsweeps so identified were due to the actions of semiperipheral or peripheral marcher states. This task requires greater specificity about what is meant by semiperipherality. The core/periphery distinction is a relational concept. In other words, what semiperipherality is depends on the larger context in which it occurs—the nature of the polities that are interacting with one another and the nature of their interactions. The most general definition of the semiperiphery is “an intermediate location in an interpolity core/periphery structure.” The minimal definition of core/periphery relations, as mentioned above, is that polities with different degrees of population density and internal hierarchy and complexity are interacting with one another. This is what we have called core/periphery differentiation. We are looking for evidence that a polity that conquered other polities and was responsible for an upward sweep was semiperipheral relative to the other polities it was interacting with before it started on the road to conquest.
The alternatives to semiperipherality are coreness and peripherality. Core polities usually are older, are more stratified, have bigger settlements, and have had the accouterments of civilization, such as writing, longer. Peripheral societies are nomadic hunter-gatherers or pastoralists, hill people, or desert people. If they are sedentary, their villages are small relative to the settlements of those with which they are interacting. We have already noted that some conquest empires were formed by peripheral marcher states. Other polity upsweeps were caused by regime changes in old core states or by older civilizational cultures that made a comeback. David Wilkinson's (1991) survey of the core, peripheral, and semiperipheral zones of thirteen interpolity systems is helpful in suggesting criteria for designating these zones, but Wilkinson did not address the question we are asking here: Were the polities that produced empire and urban upsweeps semiperipheral before they did this?
We use four main empirical indicators to make such determinations:
the geographical location of the polity relative to other polities that have greater or lesser amounts of population density. Is it out on the edge of a region of core polities?
the relative level of development: population density, which is usually indicated by the sizes of settlements; the relative degree of complexity and hierarchy; and the mode of production: for example, foraging, pastoralism, nomadism versus sedentism, horticulture versus agriculture, the size of irrigation systems. Hunter-gatherers or pastoralists are usually peripheral to more sedentary agriculturalists
the recency of the adoption of sedentism, agriculture, class formation and state formation
relative ecological marginality
The Aztecs (Mexica-Culhua) are a prototypical example of a semiperipheral marcher state. They were nomadic hunter-gatherers who migrated into the Valley of Mexico and settled on an uninhabited island in a lake. There had already been large states and empires in the Valley of Mexico for centuries. The Aztecs hired themselves out to older core states as mercenary soldiers, developed a class distinction between nobles and commoners, and claimed to have been descended from the Toltecs, an earlier empire. Then they began conquering the older core states in the Valley of Mexico, strategically picking first on weak and unpopular ones until they had gathered enough resources to “roll up the system.” The Aztec story has most of the elements that we are using to examine our upsweep cases: marginal geographical location, recency of sedentism, class formation, and state formation.
Another indicator of semiperipheral location is relative environmental desirability. Core societies usually hold the best locations in terms of soil and water. Noncore polities hold ecologically marginal territories. The semiperipheral marcher chiefdoms of the Pacific Islands were typically from the dry side of the island where land was steeper, rainfall was less frequent, and soil was thinner.13
Another issue is “Semiperipheral to what?” A polity may have different relationships with other polities in the same interpolity network. For example, Macedonia had one kind of relationship with the other Greek states and a different kind of relationship with the Persian Empire. Semiperipherality is relative to the system as a whole but may also be affected by important differences between other states in a system and by the existence of different kinds of relations with those other states.
Philippe Beaujard (2005) makes good use of the semiperiphery concept in his study of the emergence of world-systems in the Indian Ocean. Beaujard (2005:442) mentions instances in which the emergence of regional settlements that connected hinterlands with core areas was facilitated by the presence of merchants and religious elites who were migrants from core regions. Beaujard's study of the emergence of unequal exchange between the coastal Swahili cities and the interior of the East African mainland notes that immigrants from the Arabian core helped to form commercial ties, intermarried with local elites, and converted locals to Islam, thereby promoting a process of class formation that led to the emergence of semiperipheral polities along the coast. Beaujard also affirms our point that innovations sometimes occur in semiperipheral polities (445).
RESULTS FROM STUDIES OF SCALE CHANGES OF POLITIES AND SETTLEMENTS
The EmpCit studies use estimates of city sizes and the territorial sizes of empires to examine and compare different regional interaction systems (e.g., Chase-Dunn, Manning, and Hall 2000; Chase-Dunn and Manning 2002; Inoue et al. 2012, forthcoming). We identified upsweeps, those instances in which the scale of polities and settlements has greatly increased (Inoue et al. 2012, forthcoming). We also identified downsweeps and systemwide collapses in which the largest polities or settlements declined below the level of the previous low point and stayed down for more than one typical cycle. We found that, while the decline of individual cities and empires is part of the normal cycle of rise and fall, there were few systemwide collapses in which a downsweep was not followed rather soon by a recovery.
We also found a greater rate of urban cycles in the Western (Central) PMN than in the East Asian PMN, which supports the usual notion that the West was less stable than the East. And our finding that the Central PMN experienced two urban collapses while the Eastern PMN experienced downsweeps but not collapses supports the idea of greater stability in the East. We also found that nine of the eighteen urban upsweeps were produced by semiperipheral development and that eight directly followed, and were caused by, upsweeps in the territorial sizes of polities.
We also identified twenty-two upsweeps of the largest polities in four world regions and in the expanding Central PMN since the Bronze Age, and we examined these to determine whether they were the result of semiperipheral marcher conquests (Álvarez et al. 2013).14 We found that over half of the polity upsweeps were produced by marcher states from the semiperiphery (10) or from the periphery (3). This means that the hypothesis of semiperipheral development does not explain everything about the events in which polity sizes significantly increased in geographical scale but also that the phenomenon of semiperipheral development cannot be ignored in any explanation of the long-term trend in the rise of polity sizes.
World-system scholars contend that leaving out the core/periphery dimension or treating the periphery as inert are grave mistakes, not only for reasons of completeness, but also because the ability of core elites and their polities to exploit peripheral resources and labor has been, and continues to be, a major factor in deciding the winners of the competition among core contenders. The resistance to exploitation and domination mounted by noncore peoples has played a powerful role in shaping the historical development of world-systems since the Bronze Age. Thus sociocultural development cannot be properly understood without attention to the core/periphery hierarchy, and world-systems are a fundamental unit of analysis for explaining long-term sociocultural evolution.
This said, the world-systems theoretical research program is yet in its infancy. Alternatives to the method proposed above for spatially bounding world-systems need to be operationalized and pitted against one another to see which methods of spatial bounding are more powerful for explaining the events that have caused the long-term trends toward greater complexity, integration, and hierarchy. And the project to accurately estimate the sizes of settlements and polities needs further work. Temporal resolution needs improvement, especially in the Americas, in order to make it possible to identify upsweeps in the way that has been done in Afro-Eurasia. More settlement and polity sizes will make it possible to study changes in size distributions within regions and to examine claims about the emergence of synchrony between regions. Geocoded data on climate change, warfare, polity boundaries, and trade networks will make it possible to examine the causes of “normal” upswings and downswings as well as the less frequent upsweeps and downsweeps. The scientific study of the development of world-systems will have important implications for issues such as human responses to climate change, ecological degradation, population density, the changing nature of the global city system, the rise and fall of hegemonic core powers, transitions from unipolar to multipolar power situations, and resilience and systemic collapse.
Still the best general introduction to the world-system approach for the general reader is that by Thomas Richard Shannon (1992). Chase-Dunn and Lerro (2014) have also written a textbook for upper-division undergraduates that uses the comparative evolutionary world-systems perspective on globalization. The data appendix for this text (online at http://www.irows.ucr.edu/cd/appendices/socchange/socchangeapp.htm) includes useful datasets on the population sizes of settlements, the territorial sizes of polities, colonization and decolonization, economic globalization, the rise and fall of US hegemony, transnational social movements, and a very small world-system in Northern California.
Use of the word evolution still requires explanation. We mean long-term patterned change in social structures, especially the development of complex divisions of labor and hierarchy. We do not mean biological evolution, which is a very different topic, and neither do we mean “progress,” a normative notion that is unnecessary for the scientific study of social change.
The term settlement includes camps, hamlets, villages, towns, and cities. Settlements are spatially bounded for comparative purposes as the contiguous built-up area.
How sovereignty and authority are constructed are, of course, cultural and institutional issues that need to be understood, but all polities, even nomadic foraging bands, try to regulate access to resources.
Our study of polity size upsweeps is presented in Inoue et al. (2012). The project is the Polities and Settlements Research Working Group at the Institute for Research on World-Systems at the University of California-Riverside. It is also known as the EmpCit (Empires and Cities) project. The web site is http://irows.ucr.edu/research/citemp/citemp.html.
The regional world-system in the Hawaiian archipelago before contact with Europeans was an exception in that it was based on a single ancestral Polynesian cultural heritage, but interesting cultural differences emerged out of ecological and social differences among the islands (Chase-Dunn and Ermolaeva 1994).
This is analogous to the study of commodity chains in the modern system except that the goods exchanged in some earlier systems were not commodities in the modern sense. Gift giving and tribute were the most frequent forms of exchange found in small-scale world-systems.
William H. McNeill (1963) has long been writing about shifting cores. McNeill contends that contact between sophisticated civilizational cores and frontier fringes was often welcomed and sought out by those in the periphery. His view is that the history of the human community, the “ecumene,” is seen in the patterns of contact, interaction, exchange, and the evolving and shifting of civilizational centers.
There may be an analogous phenomenon to interpolity semiperipheral development that occurs within polities. Organizations such as firms and political parties and social movements that are competing with each other may also exhibit aspects of the “advantages of backwardness.”
In ancient Southwest Asia and in Mesoamerica certain neutral territories were recognized as international trade enclaves. These “ports of trade” (Chapman 1957) allowed international exchange (what Karl Polanyi called “administered trade”) to go on even during periods of warfare between states. Most of these neutral territories were small cities near the boundaries of larger polities. Sabloff and Rathje (1975) found archaeological evidence that Cozumel (an island near the Yucatan coast during late postclassic Mesoamerica) oscillated between being a “port of trade” (a neutral territory that is used for administered trade between different competing states) and a “trading port” (an autonomous and sovereign polity that actively pursues policies that facilitate profitable trade). This latter corresponds to what we mean by a semiperipheral capitalist city-state. Sabloff and Rathje also contend that a trading port is more likely to emerge during a period in which tributary states within the same region are weak, whereas a port of trade is more likely to emerge during a period in which there are large strong states.
But ecological advantage and marginality are relative to the kinds of technologies that are available for appropriating resources. In the long run the marginal location of the Aztecs on an island in Lake Texcoco enabled them to develop large-scale and productive agriculture based on chinampas (raised fields built up in the shallow waters of the lake with a system of canals for transporting produce by canoe).
The data appendix for the Álvarez et al. (2013) paper (available online at http://irows.ucr.edu/cd/appendices/semipmarchers/semipmarchersapp.htm) contains a list of all the empire upsweeps and our rationale for classifying each of these as instances of semiperipheral development or not.