Back to Contents

K. Ekholm and J. Friedman

The above title may appear provocative to those who would maintain that capital is, by definition, wage-labor capital, that imperialism is the highest stage of capitalism, and that before the industrial revolution there were only "embedded" economies whose goals were related to the gaining of prestige, conspicuous consumption, and the maintenance of alliances for "social reasons" (Polanyi 1947; Finley 1973: 130, 158). This is because our argument is aimed at a tendency in anthropology and anthropologically influenced history and archaeology to divide the world's history into distinctive market/nonmarket or capitalist/precapitalist systems. We feel that such "substantivist" and "historical-materialist" categorizations are based on false abstractions from reality that obscure some of the essential continuities of social evolution from the rise of the first civilizations. Our own point of view is that there exists a form of "capitalism" in the ancient world, that there are "world economies," and that many properties of the dynamics of such systems are common to our own world economy. This is not to take a stand for the "formalist" approach in economic anthropology. The entire substantivist/formalist debate, which centers on the question of market rationality versus socially prescribed nonoptimizing behavior in precapitalist society, is very much a distortion of the original primitivist/modernist debate. The latter was not concerned with general models of individual behavior but with the macrostructure of ancient (not primitive) economies (Biicher 1893; Meyer 1910; Rostovtzeff 1957; for a discussion see Humphreys 1970). The opponents of the primitivists did not stress the praxeology of individual maximizing agents but rather the substantive existence of a kind of capital accumulation in the ancient world. Their argument was, in Polanyi's own definition, a substantivist one insofar as they attempted to characterize a specific social system and not to assert a basic propensity of the human species.

We do not deny that there are important differences between industrial capitalism and the ancient systems. It is clear that the modern system in which industrial capitals compete for survival by direct investment in the productive forces implies a kind of dynamic unknown in the past. The accumulation of capital as a form of abstract wealth, however, is a truly ancient phenomenon. To say that this ancient "capital" played a fundamental economic role is not to say that it functioned directly in the production process, but that its accumulation and control were dominant features of those economies. The system to which we refer is characterized not only by an accumulation of capital, but by the emergence of an imperialist pattern: center/periphery structures are unstable over time; centers expand, contract, and collapse as a regular manifestation of the shift of points of accumulation. These phenomena are, we think, more general than modem capitalism. Similarly, the world economic crisis that we are experiencing today can be understood in terms of processes more general than a capitalist mode of production, processes that constitute a disastrous dynamic that has been the driving force of "civilized" history.

Our point of departure is that the forerunner of the present kind of world-system first emerged in the period following 3000 bc in southern Mesopotamia. Here we can describe the first example of the rise of a center of accumulation within a larger economic system and the development of an imperialist structure.

We refer repeatedly to "larger economic systems," to center/periphery relations, etc. Our object of analysis here is not the institutional structures of society, but the processes of reproduction within which such local structures are formed and maintained. To the extent that a society is not a self-corftained unit of production and consumption, it becomes necessary to take up the larger system within which that society, in conjunction with others, reproduces itself. It is at this level that we can grasp the total economic flow, the dynamic, and the conditions of existence of the society in question.

1 Supralocal exchange systems existed long before the rise of the first civilizations, and, when considered as systems in evolution, they are crucial to an understanding of the emergence of civilization. A great many examples from late Palaeolithic and early Neolithic Europe and the Middle East demonstrate the existence of trade over rather wide areas. The obsidian trade of the Near East predates the formation of urban settlements by several thousand years, and the total process of exchange which accompanied it is clearly linked to economic growth in the area (Wright 1969; Lamberg-Karlovsky 1975). Early "tribal" systems may have differed from later systems insofar as the exchange was not absolutely necessary to the maintenance of local productive forces, but socially necessary to the maintenance of internal group relations in the local population - i.e. as when prestige goods necessary for local transactions (brideprice and other services) and defining social position were imported objects. Ethnographic examples of such systems can be found throughout Africa, Melanesia, and Indonesia (Ekholm 1972, 1977a; Friedman 1975; Friedberg 1977).

2 With the rise of civilization we have a new situation in which we may speak of a technologically integrated system. The emergence of a developed center of "high culture" depends on the accumulation of resources from a wide area so that the very economic base of the locally developed society is likely to be the result of its center position within a larger system. Civilization is here coterminous with the existence of a center/periphery relation (Ekholm 1976, 1977b).

Generally speaking, the center is the center of most advanced industrial production based on raw materials and semifinished products imported from the periphery, which in exchange obtains some of the manufactures of the center. The very maintenance of the center depends on its ability to dominate a supralocal resource base. Mesopotamia is the clearest example of the extent to which a center's industrial base can be imported. To insist, as is usually done, that the evolution of high cultures is based on the agricultural surplus of intensive irrigation is systematically to avoid the problem that surplus grain cannot be locally transformed into bronze, cloth, palaces (of imported stone), fine jewelry, and weapons -hallmarks of the great civilization. Even stone and wood were imports in the case of Mesopotamia.

Center/periphery relations are not necessarily defined in terms of their import-export pattern. Thus, it is unnecessary that a center be the sole locus of industrial manufacturing in the system, or that the periphery be the sole supplier of raw materials. A relation based on a technical division of labor does not correspond to either the mechanisms of development or the functioning of global systems. Center/periphery relations refer, rather, to different structural positions with respect to total accumulation. The possession of extremely "valuable" commodities such as silver (Athens) makes it possible to accumulate a disproportionate pan of the production of the larger system. If Athens only imported and exported goods, it would never have become a great center. Its accumulation of wealth was, in the first instance, a result of large-scale military tribute-taking and plunder, mercantile profit, and the export of silver. This primary accumulation laid the foundation for a formidable expansion of industrial production. Generally speaking, industrial production is not the means by which centers accumulate initially, and accumulation of wealth from the larger system proceeds well in advance ui nome production. i ms initial and oiten continual "primitive accumulation" has always taken the form of tribute, booty, and enormous mercantile profits.

It is usually argued that capitalism can be rto the production of industrial capital to the exclusion of all other accumulative activities. Imperialism in such a system is a logically secondary phenomenon related to the needs of self-expanding industrial capital. This construct is opposed to the ancient economy where the struggle for prestige and political power predominated, thus where industrial growth, imperialism, and profit were marginal phenomena. It is assumed that capitalism is a self-igniting and self-accumulative process while the ancient economy was a more "embedded" system in which production for specific social uses determined the degree and form of growth. This distinction builds upon the subjectivity of the industrial capitalist in one case and on that of the classical Greek and Roman aristocrat in the other. In neither case is the structure of the total reproductive cycle taken into account in the definition. In the modem capitalist mode of production, for example, the accumulation of money capital is not a dependent function of production but rather, operates parallel and in contradiction to production. The purpose of production here is the accumulation of money and it is certainly not the only means although it establishes the limit conditions of that accumulation. Large portions of the total liquid wealth of capitalist society are invested in "nonproductive" and even noncommercial activities. Similarly, while it is clearly the case that the landed aristocracy and, later, the imperial bureaucracy may have been the dominant class faction in ancient society, their power depended upon the enormous wealth and profit gained in commercial agriculture, and their direct involvement in urban and international commerce, as well as their access to imperialist tribute. This situation is not different in kind from the medieval Arab economy or early modern Europe. It is often overlooked that mercantile Europe operated very much like Rome in its expansion, that it accumulated and "squandered" great amounts of wealth, not primarily by producing, but by pillaging large parts of the globe, and that capitalist production only began within this larger imperialistic process upon which it was, materially speaking, entirely dependent. That a capitalist mode of production became dominant in Europe is, of course, related to specific local structures. The emergence of wage labor in one place and slavery in another is dependent on a difference in initial conditions, but the resultant social forms may be worlds apart.

Our argument is that the general properties of imperialist mercantilist expansion are common to ancient and modem worlds irrespective of specific local forms of accumulation. The growth of industry and commercial agriculture by whatever form of exploitation occurs within an already constituted imperialist structure and is not a local and closed process.

3 Center/periphery systems are, by definition, imperialistic insofar as the center of a system accumulates wealth based on the production of a wider area. While the existence of larger exchange systems is linked to and reinforced by the emergence of local hierarchy and class domination, the center/periphery relation further integrates such local class structures into a differentiated pattern where the central class becomes increasingly elaborated into factions: landed aristocrats, bureaucrats, merchants, etc. - exploiting by direct taxation, slavery, and even contract and wage labor, while the peripheral class structure is more or less restricted to a chiefly or feudal elite (that may become more elaborate in the development process) which mediates the export of raw materials and controls all imports. The kinds of structure that develop internally depend, as we shall suggest later, upon the regional structure of the larger system. The center need not be a single political unit which would, in fact, require an extraordinary degree of direct control over the accumulation process. More often it tends to consist of a number of competing/exchanging political units, one of which may exercise hegemony within the center. Galtung (1971) represents a generalized imperialist structure (see Figure 2.1).

4 Center/periphery structures are drastically unstable because of the vulnerability of centers in the external (supply/market) realm which is so difficult to control. The existence of a production/resource area wider than that of the political unit which must be maintained by it is the fundamental weakness of such systems. Evolution is, as a result, a necessarily discontinuous process in space. Centers collapse and are replaced by other areas of high civilization. The development of total systems is not equivalent to the development of individual societies. On the contrary, the evolution tends to imply the shift of centers of accumulation over time. The "rise and fall" phenomenon is thus a manifestation of a more continuous process as a higher level of organization.

5 Imperialism is the characteristic of a center/periphery process that tends to reproduce simultaneous development and underdevelopment within a single system. "Empire" is a political mechanism, the control over a larger multisociety region by a single state. Empire, in functional terms, is a political machine for the maintenance and/or direct organization of imperialistic economic processes. In other words, it consists in the direct control over an already existing larger economic network.

1 The Early Dynastic period in Mesopotamia represents the earliest emergence of a center/periphery structure which is at all well documented.

The import of wood, copper, stone, and later tin as well as increasing quantities of precious metals and stones (silver, gold, lapis lazuli) is an indication of the degree to which Mesopotamian civilization must be denned in terms of a larger system.

2 The Early Dynastic period is also the period of the formation of large walled cities in the center,² emerging from a period in which population is more evenly distributed in smaller settlements of varying sizes probably organized in larger regional political hierarchies (Wright 1972). The urban implosion leads to the formation of many compact, warring city states.' If we compare Mesopotamian development generally with that, for example, at Susiana in the fourth millennium bc there is a striking contrast between the settlement hierarchy in Susiana, where a single center apparently controlled the external trade for a whole local region, and the final situation in the plain where a number of neighboring centers all competed autonomously in the same larger network.⁴

3 The decentralized or city-state organization of a center results from the impossibility of monopolizing the relations between the center as a whole and the rest of the system, so that instead of a local control hierarchy there is competition among equals. Mesopotamia emerges as a dense trade network linked to Anatolia and the Mediterranean in the West and to Iran and India in the East.⁵ There are no local resources within Mesopotamia that can be monopolized and no single trade route out of the area. From the start, the main export is food, then textiles and manufactures produced from imported raw materials; this is possible for all cities.⁶

4 Egypt, as opposed to Mesopotamia, is an isolated area with respect to external trade networks. Trade can move up and down the Nile, and there are several points where raw materials, especially gold, might conceivably be controlled, although, as these areas are parallel to the Nile, local access is not clearly determined. There are, however, two areas that are crucial in terms of the larger system. In the north control over the delta area means control over the only real access to the Mesopotamian-Mediterranean trade area. In the south, control over the raw material and labor resources of Nubia is crucial as a means of economic power - as a supply zone for the larger area. What is crucial here is the absence of a multitude of points of access either to important raw materials, or especially to external trade.⁷ Egypt is not, like Mesopotamia, located in the midsof a vast trade network. Rather it opens out at only a very few points, thus permitting the maintenance of a monopoly over the area's relation to the larger economy. As a result, internal competition and decentralization do not occur - no city states emerge. Theocratic and bureaucratic structures are increasingly elaborated. The redistributive structure based on various forms of direct taxation is extended. No private economic sector develops outside the state sphere and the upper class continues to be identified with the state itself.⁸

5 The different evolutions of Egypt and Mesopotamia, perhaps from similar original structures, are of great significance. The former is based on central monopoly and territorial annexation; the second on political fragmentation, competition, warfare, and empire. The second is apparently the more dynamic: the early and explosive developments of bronze, advanced weaponry, commercial techniques, abstract-formal writing all emerge in Mesopotamia long before they are introduced in Egypt.

6 Without here describing the precise nature and social categories of functioning of the Early Dynastic economic system we can link together some fundamental external characteristics.

The economy is clearly expansionist in nature. The form of expansion implies continuous increase in the work force by the mechanisms of slavery (captive or in other ways imported). Agricultural production is intensified, both to supply export needs and to support an increasing division of labor linked to the elaboration of textile, metal, and stone production, much of which is clearly for export. Such economic growth and differentiation lead to territorial expansion and a resultant conflict between political units of similar type.' Warfare leads to the intensification of weapon production and to its technological development, instrumental, perhaps, in the very evolution of bronze technology (bronze is harder and clearly superior to copper as material for weapons).¹⁰ The growth of weapon production implies a further division of labor entailing a double demand to increase subsistence goods to support specialized labor power, and to increase exports in order to obtain copper and tin. This, in turn, leads to a further effort to increase the area of land in agriculture, to territorial expansion, and interstate conflict. As a result, the very survival of the city state in this system becomes totally dependent upon the secure control over the external resource base necessary for the maintenance of the production apparatus, especially the growing military sector, which is the foundation for the defense of all other economic and political activities."

7 Externally, i.e. without considering the social form of the system, it can be said that the center as a whole comes increasingly to be the major exporter of manufactured goods, final consumption goods, and arms in exchange for the necessary materials for the very production of those goods. The principal effect of this development is the increased need for control over the external conditions of local reproduction.

The expansion of the center as a whole is based on widening rather than deepening of the production/consumption area. As the system is directed by an upper class that remains the principal consumer, there is no room for "market" expansion except in the realm of long-distance trade. This is partially offset by the expansion of wage sectors in the military and the bureaucracy and by the increase in monetary circulation in urban areas. The principal tendency is expressed in the fact that the expansion of the Early Dynastic system eventually incorporates the entire region from the Indus to the Mediterranean in a regular trade network.

As it is difficult, on the basis of our familiarity with existing data, to analyze the specific categories of the operation of the economies of Mesopotamia, we can only attempt to offer some tentative characterizations of the way they may have developed.

1 To contrast Egypt and Mesopotamia again, we venture to say that from the start they are both temple-dominated, conical clan-type structures. This implies that the upper class is a theocratically defined, hierarchically organized group of aristocratic lineages that dominate a population distributed into a scalogram of larger and smaller centers.

a) The upper class is identical with the state.

b) The accumulation of wealth is centralized at the top of the hierarchy.

c) The top of the hierarchy is thus in a position of control over external trade, export production, and the local distribution of imports (Johnson 1973).

d) The forms of exploitation in this predynastic structure consist of direct taxation of "commoners" and the use of varying forms of slavery -internal debt-slaves and imported - in expanding sectors of production, i.e. in temple agriculture and some craft sectors.

2 The strategy in the early system was one based on a temple economy fulfilling ritual functions, where, within the larger region, control depended upon the accumulations of "means of circulation," prestige goods - obsidian, metals, etc. - necessary for the social transactions of all local kinship or corporate units. The centralized control over the flow of such goods implies control over aristocrats inhabiting the hinterland and secondary centers, as well as over real wealth and labor power.¹³ The control over such circulation by a central group depends upon a monopoly over external relations. Wealth measured in prestige goods is equivalent to control over labor whose surplus product reproduces the ability of the ruler to maintain his nodal position between local production and the production of other areas.

3 This kind of structure breaks down throughout the Early Dynastic period, but not in Egypt where continued control over interregional trade enables the conical structure to become increasingly elaborate. The same kind of phenomenon that occurs within the region occurs within the evolving city state. With the whole aristocracy now in the same enclosed area and with increasing economic competition, possibilities for a more decentralized accumulation of wealth emerge within the ruling class. The increasing acquisition of formerly monopolized means of circulation by other members of the upper class is part of a crucial differentiation in the economic system. While previously, religious, political, and economic power were one, they now begin to be differentiated.

a) A class of wealthy aristocrats (great families) develops alongside the temple. Their access to liquid wealth enables them to buy land and set up estates, apparently as early as 2700 bc (Diakonoff 1959: 9).

b) A secular palace sector emerges out of the temple, specializing in political-administrative, trade, and warfare functions. This corresponds to a sector of the upper class with access to a share of the total wealth and, in this case, direct control over industrial production, as with the temple.

c) A merchant class increasingly monopolizes the interregional circulation process and is able to accumulate substantial portions of the total wealth through mercantile profit.

d) The differentiation of wealth accumulation in the upper class is reflected in the division of production into private and public sectors. The forms of exploitation, about which there is much debate, include labor which is directly dependent on the state ("helots" in Diakonoff 1977), private and state slaves, and a "free" population exploited by taxation. There is also evidence, perhaps in the free labor sector, of contract or wage labor connected with skilled or more specialized tasks. The unfree classes are only different from one another with respect to the degree to which their members exist as property.¹⁴

4 The significant internal phenomenon here is the development of a process of wealth accumulation as a private or, at least, as an extrastate process. While it is, in fact, the members of the state class who first engage in such accumulation, the end product of internal differentiation is the development of several potentially opposed upper-class fa."

5 We use the word "capital" to refer to the form of abstract wealth represented in the concrete form of metal or even money that can be accumulated in itself and converted into other forms of wealth; land, labor, and products. It is this abstract form that increasingly becomes the economic organizational basis of control in the system, competing with the older direct forms of state control, taxation, and slavery. "Capital" is not tied to a specific form of exploitation. It is, rather, the forerunner of, or perhaps identical to, merchant capital in its functioning. What is important here is that it is an independent form of economic wealth and power and not merely a means of circulation or a measuring device.

6 Private capital is probably not a major driving force in itself. It is not directly invested in production nor linked in a necessary cycle of production and realization. The state is in fact the major investor as well as the major producer and consumer, and, while the accumulation of abstract metallic wealth by the state is necessary to pay for the maintenance of the system, the accumulation includes direct taxation in kind as well as the proceeds of import-export activity. We may, however, speak of a kind of state capital in this system to the extent that the accumulation by the state of abstract wealth becomes increasingly necessary to maintain import flows and the internal system of payments. This fund of wealth reveals its importance when there is a decline in other forms of production for export (e.g. in the Roman empire).¹⁶

7 In decentralized center/periphery systems where competition and warfare are necessary components of the functioning of the system, arms production tends to become the leading sector in the economy. The state-palace sector in such a phase of development tends to dominate the rest of the local economy, and it may be in severe conflict with the temple and private sectors.

8 A significant dialectical mechanism in this economy consists in the complementarity/contradiction among the economic goals of the different sectors.

a) Private accumulation removes necessary wealth from the state sector, striving to reproduce this privilege. But the conditions of private accumulation depend on the successful functioning of the state sector which is the source of manufactures and military power.

b) The merchant class is necessary to the realization of state-produced goods in the larger system, but its increased accumulation removes wealth from the state sector which, in turn, is necessary for the political and economic maintenance of mercantile activity.

9 With the development of the city-state economy, various forms of exploitation accrue to the former tax/state-slavery combination: private slavery on private estates, forms of contract and wage labour in craft and military sectors, and some form of corporate guild structure in the mercantile sector are common developments.

1 Dense trade networks correspond to competitive centers. Sparse networks correspond to hierarchical territorial structures. In Mesopotamia, the Aegean, and eastern Mediterranean, territorial hierarchies break down due to the impossibility of monopolizing trade in the larger region. The increasing density of trade may also have led to the transition from Western Zhou hegemony to the Warring States period in China.'⁷ Similar kinds of phenomena have been observed at various levels of economic organization. Thus, recent analyses of the evolution of Melanesian social systems in terms of trade structures would seem to indicate a fragmentation of larger political structures with the increasing density of trade (Alien 1977; Ambrose, n.d.).

2 Compact city states correlate with the loss of monopoly over the local economy by a state class and the development of a structure of class factions. In Phoenicia and Greece, there is an apparent transition from a palace-based (or royally controlled) economy to an oligarchic class structure. While there are conflicts between state and private sectors in Athens, the private sector is clearly in command in the classical period (this is also true for Phoenicia).¹⁸ In Mesopotamia, the state sector and its bureaucratic membership remain a powerful force, maintaining a monopoly over industrial if not agricultural production.

3 There is a clear differentiation between agricultural-based production states and trade states within the larger system. The latter are, by definition, later developments that necessarily have to import subsistence goods to support their specialized labor. Mesopotamia (southern) represents an agricultural-based production center surrounded by trade states such as Assur, in the north, and Phoenicia, in the west. Trade states specialize in specific forms of industrial production, the carrying trade, and middleman activity between other production areas. They must necessarily control the larger market network to which they belong. Dense local network conditions generate competitive expansion in the form of colonization. Empires in such systems as Athens, are, again, political-military devices for maintaining control over the larger network. It would appear, however, that the absence of a previous state-bureaucratic sector in such systems implies the continued dominance of the private, oligarchic, class and the continued furthering of their interests. The establishment of empire can take on a more or less bureaucratic form. In southern Mesopotamia the empire is the work of the state bureaucracy, and their interests, in terms of their own level of consumption and military needs, take precedence over the interests of private accumulators. The result is a great bureaucratic machine and a strictly controlled economy, centralized at the expense of private interests.

1 Empires that develop in the core/periphery systems that we have described are political mechanisms that feed on already established forms of wealth production and accumulation. Where they do not overtax and where they, simultaneously, maintain communication networks, they tend to increase the possibilities of production and trade in the system, i.e. the possibilities for all existing forms of wealth accumulation.

2 Empires maintain and reinforce core/periphery relations politically, by the extraction of tribute from conquered areas and peripheries. But insofar as empires do not replace other economic mechanisms of production and circulation, but only exploit them, they may create the conditions for their own demise.

3 This occurs where the revenue absorbed from the existing accumulation cycles increases more slowly than total accumulation itself. In such a case an economic decentralization sets in, resulting in a general weakening of the center relative to other areas. The classic example that is well documented is the decentralization that occurs in the Roman empire, resulting in the virtual bankruptcy of the center itself, which becomes a net importer, and where its costs of imperial maintenance far outstrip its intake of tribute. In the case of Rome, where the class of private capitalists and landowners is dominant, the decentralization process is probably much more rapid and uncontrolled than in Mesopotamia, where the state sector can maintain stricter surveillance of the private sector and where a large portion of production is state monopoly. It is the Roman upper class itself, after all, that is largely responsible for the decentralization of the economy."

4 Grossly stated, the balance of empire is determined by: (booty + tribute [tax] + export revenue) - (cost of empire + cost of imports)

The maintenance of the center position in an empire depends on the mechanisms that determine the net economic flows in the larger system. When the empire does not organize those flows directly, there will be a tendency in the "capitalist" structure, referred to above, for the center eventually to decline.

Our aim has been to present the sketch of an argument and not a definitive analysis of any kind. Our point has been to stress the fundamental continuity ancient and modern world-systems. We have repeatedly stressed the larger-system aspect in opposition to models that take society as the sufficient unit of analysis, thereby, as we see it, implicitly removing the question of dynamics, evolution, and devolution. In such a short space we have not attempted to take up more specific aspects of the available data, choosing instead to state a number of points that we feel are important areas of further discussion.

The general properties that we have discussed apply, at a sufficiently high level of abstraction, to all systems of "civilization." We are, perhaps, talking about the same system. The forms of accumulation have not changed so significantly. The forms of exploitation and oppression haw all been around from the earliest civilizations although, of course, they have existed in different proportions and in varying combinations. Even in our own recent past, a major form of the accumulation of capital was by means of slavery, and it might be argued that today's eastern-bloc economies are but another (ancient) form of exploitation linked to the same process of accumulation (Frank 1977). There are, to be sure, a great many differences, but the similarities are, perhaps, a more serious and practical problem.

1 The center need not, of course, be the originator of new technology. It need only concentrate that technology within its boundaries. While a good deal of military technology has been invented by peripheral nomadic groups, it is only in the centers that it has been mass-produced on a large scale.

2 Before the Jemdet Nasr period there are virtually no walled settlements. By the Early Dynastic (ED) period all cities are walled (Adams 1971: 581).

3 Young (1972: 832), for example, states that the "concentration of population into the larger urban centers appears to take place at the expense of the countryside."

4 Extensive discussions of the predynastic situation in southern Mesopotamia can be found in Johnson (1973, 1976) and Johnson and Wright (1975), where a case is made for the existence of hierarchical control over trade. We feel that the evidence they present can be interpreted in this way in spite of the perhaps unwarranted use of central-place theory.

5 The import of raw material is already evident in the Ubaid (Stigler 1974: 114).

6 Crawford (1973) points out that there is little clear evidence of large-scale production for export in the ED period in spite of the massive import of raw materials. In her argument for the existence of invisible exports from Mesopotamia which do not leave clear material traces, she stresses the importance of grain and dried fish in the initial periods. Spindle whorls are found at least as early as the Ubaid period (Kramer 1970), but it is difficult to determine exactly when textiles become an exported item. Later in the ED period, at Lagash there is mention of 200 slave women engaged in the various moments of textile production (Deimel, cited in Adams 1971: 5114). Adams (1971: 583) also mentions the export of tools and weapons.

7 It is, in fact, only in the delta area, where there are several possible outlets to the Mediterranean, that there was any serious competition between centers.

8 A long line of scholarship has stressed this specific property of Egyptian civilization. Weber (1976), for example, emphasizes the long-term maintenance and development of a bureaucratic state-class as opposed to the more mercantile and decentralized structure of Mesopotamia. Notable in Egypt is the lack of an urban civilization and its expression in walled towns, the very restricted use of a monetary medium up until a relatively late date (Janssen 1975a), the state monopoly over external trade, also until very late, and the lack of an independent class of "great families" or a merchant class. For further discussion see Janssen (1975a, 1975b), Morenz (1969), Kemp (1972), O'Connor (1972), Smith (1972) and Uphill (1972).

9 It might be hypothesized that territorial expansion was the direct outcome of competitive export of food (possibly the first major export) before ED in exchange for raw materials, and that this export agriculture was a key factor in the territorial conflicts that led to the formation of walled city states.

10 Metal was relatively rare in ED I, but there is a clear development of metallurgy in ED II and "lost wax process reached a climax in ED III when armourers made heavy and efficient weapons of war, axes, spears, and adzes" (Mallowan 1971: 239 f.).

11 The importance of warfare is stressed in Childe (1952), who refers to the first example of "organized homicide" (30) in c. 3000 bc.

12 In the Jemdet Nasr period and earlier in Susa, there are large storerooms attached to the central temple (Frankfort 1971: 1; 2: 84). "It has been said that the larger Ubaidian settlements were probably market towns as well as religious centers, both for the local satellite communities and for wider trade" (Stigler 1974: 114). See also Adams and Nissen (1972).

13 Long before bronze technology, copper makes its appearance in the form of copper beads in Iran at Ali-Kosh from 6750-6000 (Hole et al. 1971: 278) and in Egypt c. 4000 (Stigler 1974: 135).

14 Diakanoff claims, contra Gelb, that state sector "slaves" or helots are the equivalent of patriarchal slaves at the state level and not of serfs. They are, in any case, quite unlike the classical slaves of Greece and Rome. It is not clear that his attempt at differentiating ancient Mesopotamia from the Mediterranean classical period is theoretically valid, especially in light of the massive expansion of the slave sector following 2000 bc.

15 The existence of a sphere of private capital accumulation is well documented for areas that are traditionally thought to be characterized by theocratic centralized economies. Thus, Larsen (1967, 1976) has shown that Assur, for example, in the Old Assyrian period, is very much a private commercial economy. His analysis of caravan-procedure texts, of costs and methods of trade, demonstrates the existence of a considerable profit-oriented private sector. Similar evidence exists for Ur (Oppenheim 1954) and for Lagash (Lambert 1953). Adams (1974) has discussed how the damgar, merchant, who begins as a state official, becomes an increasingly private operator (1974: 248). Farber's material (1978) demonstrates the very high degree of commercialization of the Babylonian economy (the existence of parallel movements of prices of different goods and services and wages, a general price level, and economic cycles).

16 Hopkins' (1978) work represents a sophisticated attempt at developing a model of the dynamic of the Roman economy that shows the fundamental importance of the imperialist drive of the state in the fueling of the other forms of accumulation in the system. The classical discussion of the relation between the decentralization of capital accumulation in the Roman empire as linked to the crisis of the state can be found in Rostovtzeff (1957).

17 A general discussion of the breakdown of trade monopolies and the development of city and territorial states can be found in Friedman and Rowlands (1977). Hsu (1965) discusses the internal transformations in China in the Zhou period, especially the emergence of private accumulation and the breakdown of the earlier aristocratic bureaucracy.

18 There is, however, a large-scale regional oscillation between private and state-controlled economies in the long run. Thus, the development of the Hellenistic states marks the emergence of state control after a period of more "democratic" oligarchic rule in the Mediterranean.

19 A pattern in the Roman imperialist system which may be common to all such systems is one that we can clearly see in modem imperialism. The first period would consist in military or military backed expansion where large areas are taxed at exorbitant rates. Such areas, e.g. Greece in the Roman period, being short of means of payment, borrow from Roman capitalists, a phenomenon reminiscent of the nineteenth and early twentieth centuries. The result is a double flow of tribute and debt service payments back to the center aa deindustrialization of the periphery. The enormous accumulation of capital in the center leads to a great deal of unproductive as well as productive investment. But the generally high rate of accumulation compared to the increase in production leads to increasing costs at home, increasing taxation, and conditions generally unfavorable for home producers on the world market. At this point the pattern shifts from a center/periphery structure toward a decentralization of production as well as accumulation. Rome's relative monopoly over certain mass-production items for the empire is lost in the imperial period, and its ability to maintain its increasingly expensive social complex at Rome becomes a serious problem leading to a more or less continual crisis of the state. It has been observed by some (Ekholm 1977b; Friedman 1978; Froebel et al. 1977) that a similar deindustrialization of capital accumulation is occurring today and that it, too, is directly linked to the present general economic crisis and to the crisis of the state in the West.

This article was published for the first time in 1979 in pan as the result of work carried out in the framework of an interdisciplinary seminar on exchange systems in a historical perspective. It was also a product of the application of a global systemic anthropology that was emerging in our cooperative work in that period.

Just as there was heated debate concerning our general approach, there was quite some debate about this particular presentation. Moses Finley was one of those with whom we, quite understandably, had several heated discussions in the late 1970s. All this was a reflex of the dominant society-centered focus of the social sciences at the time. Since that time we have had little chance to return to the issues of ancient civilizations although many of our colleagues have continued to develop the global approach in ancient history and prehistory (Spriggs 1988; Rowlands et al. 1987; Thomas 1989). Since the publication of that article other research has applied the global approach to the prehistory of Oceania (Friedman 1981, 1982, 1985) and to the relation between global processes and the emergence in the nineteenth century of social and cultural forms in central Africa that are today regarded as traditional (Ekholm 1991). Given the initial skeptical reception of the 1979 article as well as a number of similar endeavors (Ekholm 1975; Ekholm and Friedman 1980), it is indeed gratifying to discover that it has now been incorporated into a growing field within anthropology and "archaeology and related subjects." After quite a few years of global anthropology in the field (Ekholm 1991; Friedman 1992a, b) we have been prompted to begin looking again into the data of ancient societies and of prehistory in general. At present we can do no more than confine ourselves to several brief remarks concerning the status of the article today. A more systematic summary of our work is forthcoming in Review (Friedman 1992c).

The general argument of this article consisted in the assertion that the dualist division of world history that was so prevalent in anthropologically inspired economic history following Polanyi and disciples, in Marxist economic history, and even in the then emergent world-systems theory of Wallerstein (also partially informed by Polanyi), one that envisioned world history as divided into pre- and post-European Renaissance eras, was an ethnocentric misunderstanding. Frank and Gills have recently developed a similar theme, perhaps more extreme, that the past 5,000 years of world history can be understood as the continuity of the same system. In another article published at about the same time (1980) we emphasized the degree to which global systems are multistructural, i.e. that they contain numerous articulations among local and global processes and that system refers to the systemic properties of globally open processes rather than to an operationally definable empirical entity. This has been further elaborated in a senes of articles dealing with the relation between global processes and cultural processes (Friedman 1988, 1989, 1990, 1991). In another article ^from the early phase of our work it was suggested (Friedman 1979) that the so-called transition from feudalism to capitalism was essentially a shift I of capital accumulation from East to West in a declining Middle-East- dominated world system in the Middle Ages. This argument was taken from the pioneering work of Maurice Lombard, and since then Abu-Lughod has brilliantly made this kind of argument definitive, thus effectively forcing other researchers to rethink if not eliminate previous discussions of the internal transition in Europe.

While clearly sympathetic to the gist of the argument advanced by Frank, we think it should be reframed in more hypothetically interesting terms; lest the notion of system become so diluted as to be nearly useless as a; theoretical tool. Frank has claimed that we have sometimes shied away from the assumption of a single world-historical system. In the article reprinted here we do not make it a central issue since the empirical foundations are quite lacking. We suggest it as an avenue of research and concentrate instead on the issue of continuity. We have attempted to understand the degree to which such continuity exists. This is a more complex problem than the enduring appearance of empires, world markets, the accumulation of capital, and cycles of hegemony. It is related to the nature of the local structures as well, and the degree to which they are products of the same larger system. It is necessary to understand how different kinds of local organization arise, why caste here and ethnic pluralism there, why class in one situation and region and ranked estates in another. Now we have gone so far as to suggest that different kinds of social-identity formation are related to differences in the degree of capitalization of social relations, differences in the degree of resultant social individualization; that western ethnicity is, as a result, a product of what might be called the culture of "modernity" understood as a general phenomenon, as opposed to the kinds of cultural identity that we find today in India and south-east Asia (Friedman 1991). That so-called modern ethnicity may have appeared in earlier periods, i.e. the Hellenistic, to a greater or lesser extent might have significant implications for the continuity hypothesis. If Protestant ethics are more general than Christianity, as they would appear to be, this is also an important argument for continuity. There is more to the configuration of global systems than simply the general economics of wealth accumulation. One might, for example, take up the issue of the emergence of feudal structures as opposed to what we have described elsewhere as kin-based prestige-good systems on the peripheries of commercial civilizations. One might ask whether different proportions of different dominant forms of accumulation of wealth, mercantile (abstract wealth) vs. control over products and/or labor, affect the dynamic properties of the overall "system." We are not saying all this in order to play a Tilley-like role of devil's advocate, but because these are absolutely fundamental issues for the understanding of how global systems or the system function(s) and change(s).

The question of system vs. systems is not as simple as Frank defines it, although we are sympathetic to his point, since we have ourselves argued that the similarities are "a more serious and practical problem" (Ekholm and Friedman 1980) in political terms. It is necessary to come to grips with the questions of transformation even while assuming a continuity. Now we have suggested on several occasions that the similarities between empirically delimited systems are great enough to warrant a continuist hypothesis. But this does not eliminate the need to understand the conditions for the emergence of modem industrial capitalism and with those aspects of modernity that diverge from earlier social orders. If the family resemblances we discover enable us to talk of variwithin the same system instead of the emergence of new systems this remains a complex empirical and theoretical problem of boundaries (not geographical). That China need not have been dynamically connected to the India-Mesopotamia-Mediterranean nexus in the third millennium bc for us to say that their regional histories can be seen as dependent on one another ought not to dismay us. The New World was not, in all probability, systemically connected to the Old in any strong sense until quite late. But it clearly displays global systemic processes of a similar order. In our argument the continuity of global systems concerns the continuity of global processes and not of a physically delimited portion of the globe. Contractions in such systems have not infrequently created "isolates," "untouched primitives," and Shangrilas. These are, of course, global products, but they are not globally connected in the sense demanded by Frank's model. Conditions and structures of reproduction must be examined in order to ascertain the way in which empirical global systems are constituted, the way in which regions and populations are linked, .the degree to which the local is produced rather than simply connected to the larger whole and the degree to which its historical trajectory is tied to that whole. We are still operating with a working hypothesis and there is no obviously established 5,000-year-old system even if there might be much to suggest the existence of a system variable in extent and connectivity over time.

On the other hand there is also much to suggest that world history is like a Kafkaesque nightmare of repetition compulsions, more a scenario of imprisonment in larger systems than one of self-control led and empowering evolution.

This chapter first appeared in 1982 as " 'Capital' imperialism and exploitation in ancient world-systems," in Review 4 (1) (summer): 87-109.

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

Adams, Robert (1971) "Developmental stages in ancient Mesopotamia," in Prehistoric Agriculture, ed. S. Streuver, New York: Natural History Press, 572-90.

覧 (1972) "Patterns of urbanization in early southern Mesopotamia," in Man, Settlement and Urbanism, ed. P. Ucko, G. Dimbleby, and R. Tringham, London: Duckworth, 735-50.

覧 (1974) "Anthropological perspectives on ancient trade," Current Anthropology 15 (3) (September): 229-58.

Adams, Robert and Nissen, H. (1972) The Uruk Countryside, Chicago: University of Chicago Press.

Alien, Jim (1977) "Sea traffic trade and expanding horizons," in Sunda and Sahul, ed. J. Alien, J. Golson, and R. Jones, London: Academic Press, 187-417.

Ambrose, W.R. (n.d.) "Obsidian and its prehistoric distribution in Melanesia," (manuscript).

Bucher, K. (1893) Die Enststehung Volk der Volkwirtschaft, Tubingen: Laupp. Childe, Gordon (1952) New Light on the Most Ancient East, New York: Praeger. Crawford, H.E.W. (1973) "Mesopotamia's invisible exports in the third millennium bc", World Archaeology 5: 232-41.

Diakanoff, I.M. (1959) "The structure of society and state in early Sumer," translated from Sumer: Society and State in Ancient Mesopotamia, Moscow: Nauka.

覧 (1977) "Slaves, helots and serfs in early antiquity," Soviet Anthropology 15 (2-3) (autumn/winter): 50-102.

Duncan-Jones, Richard (1974) The Economy of the Roman Empire, Cambridge:

Cambridge University Press.

Ekholm, K. (1972) Power and Prestige: The Rise and Fall of the Kongo Kingdom, Uppsala: Skrivservice.

覧 (1975) "On the limits of civilization: the dynamics of global systems," Dialectical Anthropology 5 (2): 155-66.

覧 (1976) "Om studiet av det globala systemets dynamic," Antropologiska Studier 20: 5-32.

覧 (1977a) "External exchange and the transformation of central African social systems," in The Evolution of Social Systems, ed. J. Friedman and M. Rowlands, London: Duckworth, 115-36.

覧 (1977b) Om studier av riskgenerering och av hur risker kan awarjas, Goteborg: Samrbets Kommitten for langsiktsmotiverad forskning, rapport 11.

覧 (1991) Catastrophe and Creation: The Formation of an African Culture, London: Harwood.

Ekholm, K. and Friedman, J. (1980) "Towards a global anthropology," in History and Uriderdevelopment, ed. L. Blusse, H. Wesseling, and C.D. Winius, Leiden and Paris: Center for the History of European Expansion.

覧 (1985) New introduction to "Towards a global anthropology," Critique of Anthropology 5 (I): 97-119.

Farber, Howard (1978) "A price and wage study of northern Babylonia during the 77

Barry K. Gills and Andre Gander Frank

WORLD SYSTEM ORIGINS

The designation in time of the origin of the world system depends very much on what concept of system is employed. We may illustrate this problem by analogy with the origins of a major river system. For instance, look at the Missouri-Mississippi river system. In one sense, each major branch has its own origin. Yet the Mississippi river can be said to have a later derivative origin where the two major branches join together, near St Louis, Missouri. By convention, the river is called "the Mississippi" and it is said to originate in Minnesota. Yet the larger and longer branch is called "the Missouri," which originates in the Rocky Mountains in Montana. Of course, all these also have other larger and smaller inflows, each with their own point(s) of origin. The problem in how to set a fixed point of origin when in fact no such single point of origin exists for the river system as a whole. In the case of the world system it would be possible to place its origins far up stream in the Neolithic period. However, it may be more appropriate to discuss the origins further down stream, where major branches converge.

By the river-system analogy, we may identify the separate origins of Sumer, Egypt, and the Indus as at some time in the fourth to the third millennia bc. The world system begins with their later confluence. David Wilkinson (1989) dates the birth of "Central civilization," through the political-conflictual confluence of Mesopotamia and Egypt into one overarching states system, at around 1500 bc. Wilkinson's work is of very great value to the analysis of world system history. Essentially, the confluence of "Mesopotamia" and "Egypt" gave birth to the world system. However, by the criteria of denning systemic relations, spelled out below, the confluence occurs considerably earlier than 1500 bc. By economic criteria of "interpenetrating accumulation," the confluence included the Indus valley and toe area ot byna and the Levant. Thus, the confluence occurred some time in the early or mid-third millennium bc, that is by about 2700-2400 bc.

Historical-materialist political economy begins with the recognition that "getting a living" is the ultimate basis of human social organization. The ultimate basis of "getting a living" is ecological, however. The invention of agriculture made possible the production of a substantial surplus. Gordon Childe (1951) made famous the term "Neolithic Revolution" to describe the profound effects on human social organization brought about by the production of an agricultural surplus. The subsequent "Urban Revolution" and the states that developed on this basis contributed to the formation of our world system.

From the outset, this social organization had an economic imperative based on a new type of relationship with the environment. The alluvial plains of Egypt, Mesopotamia, and Indus are similar in that their rich water supply and fertile soil make possible the production of a large agricultural surplus when the factors of production are properly organized. However, all three areas were deficient in many natural resources, such as timber, stone, and certain metals. Therefore, they had an ecologically founded economic imperative to acquire certain natural resources from outside their own ecological niches in order to "complete" their own production cycles. Urban civilization and the state required the maintenance of a complex division of labor, a political apparatus, and a much larger traor economic nexus than that under the direct control of the state. Thus, the ecological origins of the world system point to the inherent instability of the urban civilizations and the states from which it emerged. This instability was both ecological-economic and strategic. Moreover, the two were intertwined from the beginning.

Economic and strategic instability and insecurity led to efforts to provide for the perpetual acquisition of all necessary natural resources, even if the required long-distance trade routes were outside the direct political control of the state. This was only possible through manipulated trade and through the assertion of direct political controls over the areas of supply. The internal demographic stability, and/or demographic expansion, of the first urban centers depended upon such secure acquisition of natural resources.

However, in a field of action in which many centers are expanding simultaneously, there must come a point when their spheres of influence become contiguous, and then overlap. As the economic nexus of the first urban civilizations and states expanded and deepened, competition and conflict over control of strategic sources of materials and over the routes by which they were acquired tended to intensify. For example, control over certain metals was crucial to attaining technological and military superiority vis-a-vis contemporary rivals. Failure to emulate the most advanced technology constituted, then as now, a strategic default.

The ultimate rationale for the origins of the world system were thus embedded in the economic imperative of the urban-based states. A larger and larger economic nexus was built up. Specialization within the complex division of labor deepened, while the entire nexus expanded territorially "outward." In the process, more and more ecological niches were assimilated into one interdependent economic system. Thereby, the world system destroyed and assimilated self-reliant cultures in its wake.

By the third millennium bc, the Afro-Eurasian economic nexus, upon which the world system was based, was already well established. Thereafter, the constant shifts in position among metropoles in the world system cannot be properly understood without analysis of the ecological and technological factors "compelling" certain lines of action. The rise and decline of urban centers and states can be made more understandable by placing them within the world systemic context. This also involves paying attention to their role in the economic nexus, particularly with regard to the sources and supply of key commodities and natural resources. The logic of the political structure of the world system is one in which the security of the member states, and their ability to accumulate surplus, is perpetually vulnerable to disruption. This situation created a dynamic of perpetual rivalry, ftus, attempts are made to extend political control over strategic areas of supply in the overall economic nexus.

New historical evidence suggests that economic connections through trade and migration, as well as through pillage and conquest, have been much more prevalent and much wider in scope, than was previously recognized. They have also gone much farther back through world history than is generally admitted. By the same token, manufacturing, transport, commercial, and other service activities are also older and more widespread than often suggested. The long history and systemic nature of these economic connections have not received nearly as much attention as they merit (Adams 1943). Even more neglected have been these trade connections' far-reaching importance in the social, political, and cultural life of "societies" and their relations with each other in the world system as a whole. Even those who do study trade connections, as for instance Philip Curtin's (1984) work on cross-cultural trade diasporas, often neglect systematic study of the world systemic complex of these trade connections.

Historical evidence to date indicates that economic contacts in the Middle East ranged over a very large area even several thousand years before the first urban states appeared. The Anatolian settlement Catal Huyuk is often cited as an example of a community with long-distance trade connections some 7,000-8,000 years ago. Jericho is another of ten cited example. Trade or economic connections between Egypt and Mesopotamia were apparently somewhat intermittent before 3000 bc, and therefore possibly not systemic. However, both Egypt and Mesopotamia very early on developed economic connections with Syria and the Levant, which formed a connecting corridor between the two major zones. The putative first pharaoh of unified Egypt, Narmer, may have had economic connections to the Levant. Certainly by 2700 bc, Egypt had formal political and economic relations with the city of Byblos on the Levantine coast. Byblos is probably the earliest port of economic contact mentioned in both Egyptian and Mesopotamian historical sources.

For both Egypt and Mesopotamia, war and trade with Syria and the Levant involved the search for access to strategic and other materials, such as timber, metals, oils, and certain luxury consumption goods. The apparent goal of Akkadian imperial expansion was to benefit from putting all the most strategic routes in one vast corridor from the Mediterranean to the Persian Gulf under its sole control. There is evidence that Akkad maintained maritime economic connections with the Indus, known as "Meluhha," via ports in the Persian Gulf. Thus, Akkad consolidated a privileged position in the overall economic nexus. The city states of Syria and the Levant became the objects of intense rivalry between Egypt and Mesopotamia. Oscillation occurred in the control of these areas: from the first and second dynasties, of Egypt, over to Akkad, then to the third dynasty of Ur. By the nineteenth century bc, Egypt again exercised influence over most of the Levant as vassal states. It is clear that throughout a considerable historical period, even to the time of the Assyrian and then the Persian empires, Syria and the Levant played a crucial role as logistical interlinkage zones and entrepots within the. world system. They linked the Mesopotamian, Egyptian, and Indus zones in one world system.

Accumulation is a major incentive for, and the ultimate cause of, economic, political, and military expansion by and interlinkage within the world system. Therefore, the process of accumulation and its expansion is also importantly related to the extension of the boundaries of the world system. Two additional analogies of expansion may be useful in understanding the process: the glacier analogy and the ink-blot analogy. By analogy to a glacier, the world system expanded along a course of its own making, in part adapting to pre-existing topology and in pan itself restructuring this topology. By analogy to an ink blot, the world system also spread outward, beyond its area of early confluence. Probably the most spectacular single instance of this expansion was the "discovery" of the New World and later Oceania. David Wilkinson (1987) also sees Central civilization as expanding into other areas and societies and incorporating them into itself. In one sense, the process is one of simple incorporation of previously unincorporated areas, by analogy with the expansion of an ink blot.

However, the incorporation of some regions into the world system also involved processes more like merger than mere assimilation, as when two expanding ink blots merge. For instance, the incorporation of India, and especially of China, appear to be more merger than assimilation. Mesopotamian trade with the Indus was apparently well established at the time of the Akkadian empire. Repeated evidence of economic contact with India exists, though with significant periods of intermittent disruption. These disruptions make it difficult to set a firm date for the merger of India with the world system. Chinese urban centers and states appear to have developed essentially autonomously in the archaic Shang period. However, the overroutes to the central world system to the west were already opened by the end of the second millennium bc, particularly as migratory routes for peoples of Central and Inner Asia. The actual historical merger of Chinese complexes into the world system comes only after state formation in China reached a more advaed stage, in the late Zhou period. A series of loose hegemons began with Duke Huan of Qi (685-43 bc) and a process of unification of smaller feudatories into larger territorial states occurred. According to Wolfram Eberhard (1977), the eventual victory of the state of Qin and the creation of the first centralized empire in China was influenced by Qin's strong trade relations with Central Asia. These economic connections allowed Qin to accumulate considerable profit from trade. The Wei and Tao valleys of the Qin state were "the only means of transit from east to west. All traffic from and to Central Asia had to take this route" (Eberhard 1977: 60).

The maintenance of maritime and overland trade routes, and the peoples located in the areas between major zones, play key logistical interlinkage roles in the process of merger. In the formation of the world system, the interaction of high civilization with tribal peoples, especially in Inner and Central Asia, but also in Arabia and Africa, played a crucial but largely neglected role, to which we shall return below.

WORLD SYSTEM ROUTES AND NEXUSES

The advertising blurb of The Seacraft of Prehistory by Paul Johnstone (1989) reads: me iiduuciu uimension or prehistory has not receivea the attention it deserves.... Recent research has shown that man travelled and tracked over greater distances and at a much earlier date than has previously been thought possible. Some of these facts can be explained by man's mastery of water transport from earliest times.

Generally the sea routes were cheaper and favored over the overland ones. Some particularly important maritime routes are discussed below.

The silk roads formed a sort of spinal column and rib cage - or, more precisely perhaps, the circulatory system - of the body of this world system for some 2,000 years before 1500 ad. These "roads" extended overland between China, through Inner and Central Asia, to the "Middle East" (west Asia). From there, they extended through the Mediterranean into Africa and Europe. However, this overland complex was also connected by numerous maritime silk-"road" stretches through the Mediterranean, Black Sea, Red Sea, and Persian Gulf, and along many rivers. Moreover, the predominantly overland silk-road complex was complemented by a vast maritime silk-road network centered on the Indian Ocean through the Arabian Sea and Bay of Bengal, and on the South China Sea. These maritime silk roads in turn were connected by overland portage across the Kra isthmus on the Malay Peninsula, as well as by ship through the Malaccan Straits between it and Sumatra, etc.

The silk roads of course derive their name from China's principal export product to the West. However, the trade of items and peoples extended far beyond silk alone. Indeed, the silk had to be paid for and complemented by a large variety of other staple and luxury goods, money, services, and enslaved and other people who performed those services. Thus, the silk roads also served as the trade routes, urban and administrative centers, and military, political, and cultural sinews of a vast and complex division of labor and cultural diffusion.

If one looks at a map of Eurasia, it becomes clear that Central Asia (in present Afghanistan and former Soviet Central Asia) was well positioned to act as the ultimate nodal center. Central Asia was the crossroads of a world system in which China, India, Persia, Mesopotamia, the Levant, and the Mediterranean basin all participated. For instance. Central Asia played a key role in the joint participation in the world system of Han China, Gupta India, Parthian Persia, and the Roman empire.

However, Central and Inner Asia were also more than the meeting points 01 others. Inner and central Asia also originated meir own cycles of outward invasory/migratory movements in all directions. These cycles lasted an average of approximately two centuries and occurred at roughly half-millennium intervals. For instance, there were waves of invasions in 1700-1500 bc, in 1200-1000 bc, around 500 bc, around 0, in 400-600 ad; around 1000-1200/1300 ad. Each inner wave pushed out outer waves, except the last one of Genghis Khan and his successors to Tamerlane, who overran all themselves.

Whether or not all these invasions responded to climatic changes, presumably they were both cause and effect of changes in rates of demographic growth and decline, which may in turn have climatic causes. However, they were also caused by - and in turn had effects on - the ecological, socioeconomic, and political relations with their civilized neighbors. Thus, Inner and Central Asia and its pulse require special attention in world system history. How central was Central Asia to world system history? To what extent was Central Asia, and not primarily the other civilized areas, something of a motor force of change in the whole system? How was the rise and decline of various cities (Samarkand!) and states in this area related to system-wide developments in trade?

The place and role of Central Asia are as important as they are neglected. The entire development of the world system has been profoundly affected by the successive waves of invasion from the Eurasian steppes on the perimeter of the agroindustrial zones. This "system implosion" is such a major phenomenon that cries out for systemic study and explanation. These system implosions were not deus ex machina, but integral to the overall developmental logic of the world system's expansionary trajectory. In particular, the invasions and migrations from Inner and Central Asia were always instrumental in transforming the economic, social, political, and cultural life of their neighboring civilizations - and in forming their racial and ethnic complexions. Nor has the enormously important role of Central Asia as an intermediary zone in the world system received the systematic analysis which its functions merit. Other nomadic and tribal peoples, for instance on the Arabian Peninsula before Mohammed and in much of Africa, also participated in world system history and world accumulation in ways which have not been acknowledged except by a very few specialists.

Three magnets of attraction for political-economic expansion stand out. One is sources of human (labor) and/or material (land, water, raw materials, precious metal, etc.) and technological inputs into the process of accumulation. The second is markets to dispose of one zone's surplus production in exchange for more inputs, and to capture stored value. The third, and perhaps most significant, are the most privileged nexuses or logistical corridors of interzonal trade. Bottleneck control over the supply routes of raw materials, especially of metals and other strategic materials, plays a key role in attracting powers to such areas. This may also provide a basis upon which to make a bid for expansion of "Imperial power. Especially here, economic, political, and military conflict and/or cultural, "civilizational," religious, and ideological influence all offer special advantages for tapping into the accumulation and the system of exploitation of other zones to benefit one's own accumulation. Therefore, it is not mere historical coincidence that these three nexus areas have recurrently been the fulcra of rivalry, commerce, and of religious and other cultural forms of diffusion.

Certain strategically placed regions and corridors have played such especially important roles in world system development. They have been magnets which attracted the attention of expansionist powers and also of migrants and invaders. Major currents of thought also migrated through them. This attention is based on their role in the transfer of surplus within the world system, without which the world system not exist. Certain metropoles have become attractive in and of themselves due to their positions along trade corridors, the growth of a-market within the metropolitan city, and the accumulated wealth of the metropole itself. The rise and fall of great regional metropolitan centers and their "succession" reflects extra regional changes in which they participate. For example, the succession of metropoles in Egypt from Memphis to Alexandria to Cairo reflects fundamental underlying shifts in world system structure. So does the succession in Mesopotamia from Babylon to Seleucia to Baghdad.

Three nexus corridors have played a particularly pivotal and central logistical interlinkage role in the development of the world system.

1 The Nile-Red Sea corridor (with canal or overland connections between them and to the Mediterranean Sea, and open access to the Indian Ocean and beyond).

2 The Syria-Mesopotamia-Persian Gulf corridor (with overland routes linking the Mediterranean coast through Syria, on via the Oromes, Euphrates, and Tigris rivers, to the Persian Gulf, which gives open access to the Indian Ocean and beyond). This nexus also offered connections to overland routes to Central Asia.

3 The Aegean-Black Sea-Central Asia corridor (connecting the Mediterranean via the Dardanelles and Bosporus to the overland silk roads to and from Central Asia, from where connecting routes extended overland to India and China).

The choice between the two primarily sea-route corridors mostly fell to the Persian Gulf route. It was both topographically and climatically preferred to the Red Sea route. Moreover, the Persian Gulf corridor had connecting routes overland to Central Asia, which came to serve as a central node in the transfer of surplus among the major zones of the world system.

These three nexus corridors represented not only mere routes of trade. Repeatedly, they were integrated zones of economic and political development and recurrently the locus of attempts to build imperial systems. As the world system expanded and deepened, attempts were made by certain powers to place either two or all three corridors under a single imperial structure. Thus, such a power would control the key logistical interlinkages which have been central to the world system. For instance, the Assyrian empire attempted to control both the Syria-Mesopotamia corridor and the Nile-Red Sea corridor, but succeeded only briefly and sporadically. The Persian empire likewise controlled both these corridors for a time, and it also had partial control over the Aegean-Black Sea-Central Asia corridor. Thus the Persian empire is the first historical instance of a "three-corridor hegemony." Alexander the Great's grand strategic design for a world empire, or "world system hegemony," included plans to control all three corridors, plus the Indus complexes and the west Mediterranean basin. His successors split the Macedonian conquests almost precisely into realms parallel to the three corridors. They allowed the Indus to fall from Seleucid influence to the Mauryan empire and the west Mediterranean basin to control by Carthage and Rome. Du^g the Hellenistic period, the recurrent rivalries between the Ptolemaic and Seleucid dynasties are indicative of continued struggles between the corridors for a privileged position in the world system's accumulation processes. Even the Roman imperium did not entirely unify the three corridors, however, since Mesopotamia was denied to Rome first by the Parthians, and later by the Sassanian Persians. They used their control of thi^ area to extract considerable profit from the trade among Rome, India, and China.

Of course, each of these three main corridors had competing/complementary alternative variants and feeder routes of its own. For instance, there were several silk roads between East and West and different feeder routes in east and Central Asia and to/from south Asia. There were also routes connecting northern and western Europe through the Baltic Sea via the Dnieper, Don, Volga, and other Ukrainian and Russian routes. There were routes connecting the Adriatic to continental Europe, and the east Mediterranean to the west Mediterranean. Similarly, topological and other factors also favored some locations and routes as magnets of attraction and logistic nexuses in and around Asia. They deserve much more attention than they have received in world history. As the Afro-Eurasian nexus expanded and deepened, the number and role of these routes and choke-points increased. At the same time, their relative importance changed vis-a-vis each other as a result of world system development. Locations such as the Straits of Malacca and of Ceylon had significant logistical roles for very long periods of world system development.

The three overland and sea-route corridors and their extensions were the most important nexuses between Europe and Asia for two millennia before the shift to transoceanic routes in the fifteenth and sixteenth centuries. This historic shift from the centrality of the three corridors to that of transoceanic logistical interlinkages was probably the single most important logistical shift in world histocy and world system development. However, rather than creating it as Wallerstein (1974) argues, the shift occurred within the already existing world system.

Accumulation implies infrastructural investment and technological development. Infrastructural investment takes many forms in many sectors, such as agriculture, transportation, communications, the military, industrial and manufacturing infrastructure, and bureaucratic administration. There is investment even in ideological (symbolic) infrastructure, both of the cult of the state and of religion. In the state form of accumulation, the state seeks to create social wealth in order to extract it. By laying the basis for increases in production and facilitating accumulation, the state increases its own access to surplus and therefore its potential capabilities vis-a-vis rival states. This in turn helps it to protect "what we've got" and to get more. In the private form, the propertied elites likewise create wealth in order to extract it and invest in infrastructure to facilitate production and thereby accumulation. The ultimate rationale of such investment would in all cases be to preserve, enhance, and expand the basis of accumulation itself. The development of infrastructure and the technology it embodies feed back into the generation of surplus and accumulation. This growth of surplus in turn feeds back into further growth and development of infrastructure and technology in a cumulative fashion. The pattern is spiral, whereby the world system itself grows and becomes more firmly "established" via infrastructural investment and accumulation.

Technological progress in techniques of production, organization, and trade, both military and civilian, has long played an important, and often neglected, role in the history of the world system and in the changing relations among its pans. Technological advance and advantage have been crucial throughout history in armaments, shipping, and other transportation as well as in construction, agriculture, metalworking, and other manufacturing methods and facilities. Progress, leads, and lags in all of these have had significant contributory it not causative enecis on ^ano aiｻu some derivative effects from) the regional and other relations of inequality, within the world system. Some examples were examined by William McNeill (1982) in The Pursuit of Power.

Infrastructural investment is linked to technological change and to organizational innovation. Technological change in archaic and ancient periods, and even in medieval periods, was mostly slower than in modern industrial rimes. However, the essence of patterned relationships among technological innovation, infrastructural investment cycles, and the cycles of accumulation and hegemony (discussed below) probably have existed throughout history. When and what were the most significanttechnological innovations in world system history? Which innovations brought about restructuring of accumulation and of hegemony in the world system and which altered the logistical interlinkages? The diffusion of technology across the world system is another major area for systematic and systemic analysis.

In the general period of the contemporaneous Roman/Byzantine, Parthian/Persian Sassanian, Indian Mauryan/Gupta, and Chinese Han empires, cumulative infrastructural investments integrated each of these empires into a single world system. This high level of systemic integration was achieved via the well-developed logistic nexuses and the simultaneity of imperial expansion. At the end of that period, the entire world system experienced a general crisis. Hinterland peoples from Inner and Central Asia invaded Rome, Persia, India, and China. They caused (or followed?) a decline in infrastructural invesmnent and (temporary) serious disruption of the world system's logistical interlinkages compared to the previous era.

How is infrastructural investment linked to productivity, and increases in productivity to the processes of accumulation in the world system? Technological innovation and technological change have been pervasive in world system development. Gordon Childe (1942) pioneered a materialist analysis of the effects of technology on the ancient economy. Logistic capabilities, for instance those of maritime trade, depend on technological capability. So does the dynamic of military rivalry. Indeed, the expansion of the world system depended from the outset on technological capabilities. Invasions from the "barbarian" perimeter to the civilized centers depended upon the technological and military superiorities of the barbarians. Such invasions did not cease until "civilized" technological developments made the attainment of military superiority by the barbarians virtually impossible. By asserting a new military-technological superiority, the Russian and Manchu empires finally put an end to the strategic threat of Inner Asia in the seventeenth and eighteenth centuries ad.

The industrial revolution gave European powers the military capability to destroy or subordinate contemporary empires in the world system such as the Mughal in India, the Qing in China, and the Ottoman in the three-corridors region.

Technology has always been intimately associated with the ecological interface of the world system and its natural resource base. For instance, the technologies of farming created a secular trend to place more and more areas under agricultural production, thus to increase the sources of agricultural surplus. Particular technological innovations have dramatically affected the ecological interface, particularly those of industrialized production. Since the introduction of these technologies, the trend has been their extension across more and more of the world system, often with devastating ecological consequences.

There have been instances when environmental conditions brought about major changes in world system development. For instance, the salination of soils and silting up of irrigation works affected the relative economic strength of certain zones. For example, already before and even more after the sacking of Baghdad by the Mongols in 1258, Mesopotamia experienced relative decline. This was partly due to such environmental factors, and partly to shifts in logistical interlinkages in the world system.

Certain areas have been extremely difficult to incorporate into the world system for primarily ecological and/or topographical reasons. These difficulties (still) characterize, for instance, the Tibetan plateau, the Amazon-ian basin, the Great Northern Arctic of Canada and the former Soviet Union, and Antarctica. The social ecology of the peoples of Inner Asia, which Owen Lattimore (1940) contrasted to that of sedentary agricultural peoples, was a major factor in the world system's development for most of world history. The present ecological crises of industrial civilization remind us that ultimately ecology and the natural environment set limits on the expansion of the world system and on sustaining production and accumulation. If there have been any ecological cycles, rhythms, or trends, we should investigate what they are and how they have affected world system development.

The capture, say, by elite A here (with or without its redistribution here) of part of the economic surplus extracted by elite B there means that there is "interpenetrating accumulation" between A and B. This transfer or exchange of surplus connects not only the two elites, but also the economic, social, political, and ideological organization of their "societies". That is, the transfer, exchange, or "sharing" of surplus connects the elite A here not only to the elite B there. Surplus transfer also links the "societies' " respective processes of surplus management, their structures of exploitation and oppression by class and gender, and their institutions of the state and the economy. Thus, the transfer or exchange of surplus is not a socially "neutral" relationship, but rather a profoundly systemic one. Through sharing sources of surplus, the elite A here and the classes it exploits are systemically interlinked to the "mode of production," and even more important, to the mode of accumulation in B there. By extension, if part of the surplus of elite B here is also traded, whether through equal or more usually unequal exchange, for part of the surplus accumulated by elite C there, then not only B and C but also A and C are systemically linked through the intermediary B. Then A, B, and C are systemically connected in the same overarching system of accumulation.

This means that surplus extraction and accumulation are "shared" or "interpenetrating" across otherwise discrete political boundaries. Thus, their elites participate in each others' system of exploitation vis-a-vis the producing classes. This participation may be through economic exchange relations via the market or through political relations (e.g. tribute), or through combinations of both. All these relations characterize the millenarian relationship, for instance, between the peoples of China and Inner Asia. This interpenetrating accumulation thus creates a causal interdependence between structures of accumulation and between political entities. The structure of each component qapty of the world system is saliendy affected by this interpenetration, and empirical evidence of such interpenetrating accumulation through the transfer or exchange of surplus is the minimum indicator of a systemic relationship. Concomitantly, we should seek evidence that this interlinkage causes at least some element of economic and/or political restructuring in the respective zones. For instance, historical evidence of a fiscal crisis in one state or a zone of the world system (e.g. in third-century Rome) as a consequence of an exchange of surplus with another zone would be a clear indicator of a relationship at a high level of systemic integration. Evidence of change in the mode of accumulation and the system of exploitation in one zone as a function of the transfer of surplus to another zone would also constitute evidence of systemic relations. Evidence of political alliances and/or conflict related to participation in a system of transfer of surplus would also be considered evidence of a systemic relationship. According to these criteria, if different "societies," empires, and civilizations, as well as other "peoples," regularly exchanged surplus, then they also participated in the same world system. That is "society" A here could and would not be the same as it was in the absence of its contact with B there, and vice versa.

Trade in high-value luxury items, not to mention precious metals in particular, may, contra Wallerstein (1974, 1989), be even more important than lower-value staple trade in defining systemic relations. This is because the high-value "luxury" trade is essentially an interelite exchange. These commodities,besides serving elite consumption, or accumulation, are typically also stores of value. They embody aspects of social relations ot production, which reproduce the division of labor, the class structure, and the mode of accumulation. Precious metals are only the most obvious example, but many "luxury" commodities have played a similar role (Schneider 1977). Thus, trade in both high-value "luxury" items and staple commodities are indicators of interpenetrating accumulation.

Center-periphery-hinterland (CPH) complexes and hierarchies among different peoples, regions, and classes have always been an important pan of world system structure. However, the occupancy of musical-chair places within this structure has frequently changed and contributed to the dynamics of world system historical development. To what extent (and why) have the world system and its pans been characterized by center-periphery and other structural inequalities? Wallerstein (1974 and other works) and Frank (1978a, b, 1981), among others, have posed questions and offered answers about the center-periphery structure of the world system since 1500. Ekholm and Friedman in chapter 2 above, Chase-Dunn (1986, 1989), and others are trying to apply similar analyses to world systems before 1500. The "necessity" of a division between center and periphery and the "function" of semiperipheries in between are increasingly familiar, not the least thanks to the widespread critiques of these ideas. Chase-Dunn (1986, 1989) surveys the propositions and debates. Wilkinson (1989) examines center-periphery structures all over the world for 5,000 years. Rowlands, Larsen, and Kristiansen (1987) analyze center and periphery in the ancient world. Chase-Dunn and Hall (1991) examine precapitalist center-periphery relations.

Chase-Dunn (1989) and Wilkinson (1989) have already made the argument that center-periphery hierarchies characterize systemic development much funher back in world-historical development than 1500 ad. In fact, center-periphery relations characterize development since the origins of the state and systems of states. However, we need a more comprehensive CPH concept than most other scholars have used. The hinterland is not directly penetrated by the extracting classes of the center, but nevertheless it has systemic links with the center-periphery zone and its processes of accumulation. Wallerstein's use of the term hinterland to mean external to the world system is insufficient because it neglects the structural and systemic significance of zones which are "outside," but nonetheless related to, the center-periphery complex. These CPH relationships have been insufficiently analyzed.

The CPH complex does not refer to mere geographical position, nor only to unequal levels of development. CPH also refers to the relations among the classes, peoples, and "societies" that constitute the mode of accumulation, me ** complex is me oasic social complex upon wnicn hegemony, as discussed below, is constructed in a larger systemic context. More research is necessary on how "geographical" position in a hegemonic structure affects class position in the CPH complex. We could expect to find that the class structure of a hegemonic state may be significantly altered by the surplus that this state accumulates from its subordinates in the CPH complex. For example, the subsidy to the plebeian class of Rome may be taken as an example of such systemic effects. Conversely, we might expect a CPH complex to give rise to increased exploitation of producers in subordinate positions.

The "hinterland" contains natural resources, including human labor, which are tapped by the center-periphery. However, what distinguishes the hinterland from the periphery is that the peoples of the hinterland are not fully, institutionally, subordinate to the center in terms of surplus extraction. That is, they retain some degree of social autonomy. If a hinterland people come under political means of extraction by the center, then the process of "peripheralization" begins. Nevertheless, despite a degree of social autonomy from the center, the hinterland is in systemic relations with the center. The frequency of center-hinterland conflict is one indicator of such systemic relations. The hinterland may also have functional roles in logistical interlinkage. In this sense, the hinterland may facilitate the transfer of surplus between zones of the world system. These roles of hinterlands merit as much theoretical attention in determining positional shifts an a system change as those of semiperipheries.

The center- (or core-)periphery-hinterland concept is not intended to replace, but to extend, Wallerstein's (1974 and elsewhere; Arrighi and Drangel 1986) core-semiperiphery-periphery formulation. However, the semiperiphery has always been a weak and confusing link in the argument. The hinterland "extension" may confuse it still further and may counsel reformulatiqn of the whole complex. For instance at a recent conference (with Wallerstein, Arrighi, and Frank among others), Samir Amin suggested that the semiperiphery has functionally become the real periphery, because it is exploited by the center; while the "periphery" has been marginalized out of the system, because it no longer has anything (or anybody) for the center to exploit for its own accumulation. As argued above, however, historically the hinterland has also contributed to core accumulation in the CPH complex.

Thus, CPH complexes are integral to the structure of the world system in all periods. They must be studied, not only comparatively, but also in their combination and interaction in the world system. It is important to examine how center-periphery zones expanded into the hinterland in order to understand the way in which accumulation processes were involved. The rationales of expansion and assimilation in the hinterland appear to be related to the "profitability" of such expansion, in terms of tapping new sources of surplus. They also help resolve internal contradictions in the center-periphery complex brought about as a result of exploitation and demographic pressure. Class conflict in the center-periphery complex is affected by the expansion of accumulation into the hinterland. Demographic trends are an important factor; the hinterland provides new resources to sustain the growing population of the center-periphery zone. The physical geographical limits of hinterland peripheralization by the center seem to be set by both logistical capabilities and by a cost-benefit calculus. Areas are occupied primarily if they can be made to pay for the cost of their own occupation or are deemed to be strategically necessary to protect another profitable area. Conversely, such areas are again abandoned if, or when, their occupation proves too costly. Fortification at such systemic boundaries has a dual function of keeping the barbarians out and keeping the producers in. That is, such fortification impedes military disruption of the zone of extraction and also impedes the escape of dependent/subordinate producers into the "free" zone.

It is important to examine how systemic links between center and hinterland are formed. How does the hinterland interact over time with the center-periphery complex and thereby affect changes in the structure of that complex itself, and vice versa? A particularly important aspect of this question is the nature of the historical relations between the so-called tribal "barbarians" and the so-called "civilized" "societies." How are the barbarians "assimilated" into civilization and yet also able to transform civilization? Throughout most of world history, this barbarian-civilization relationship has been crucial to the territorial expansion of the state, imperialism, and "civilization."

The work of Arnold Toynbee (1973), Tom Hall (1986), Eric Wolf (1982), William McNeill (1964), and Owen Lattimore (1940, 1962) illuminates many aspects of how these center-periphery-hinterland hierarchies are created, deepened, and systemically transformed. Toynbee's "systemimplosion" is of particular interest. Robert Gilpin (1981) follows Toynbee to show how an older center is eventually encircled and engulfed by new states on the periphery, which implode into the center. Thus, a "center shift" takes place by way of an implosion from the former periphery to the center of the system. For instance, this occurred with the creation of the Qin empire at the end of the Warring States period in China. It also happened with the creation of the Macedonian empire at the end of the classical period in Greece. In even earlier examples of such hinterland impact, the "tribal" Guti, the Amorites, the Kassites, and the Akkadians were intimately involved in the political cycles of archaic Mesopotamia. Each of these peoples made a transition from hinterland roles to that of ruling class in the center. Moreover, these invasions of the center by the hinterland took place for systemic reasons, not just gratuitously. Eberhard (1977) and Gernet (1985) analyze how Inner Asian nomads repeatedly invaded China to appropriate its productive structure and economic surplus. Frederick Teggart's (1939) study of correlations of historical events in Rome and China analyzes the systemic causal connections across the whole Afro-Eurasian economic nexus, which caused hinterland-center conflict in one zone to affect relations in another zone. The sequencing of conflicts follows a logic that corresponds to both logistical elements in the nexus, struggles over shares of accumulation, and social tensions due to the expansionary pressure of the center-periphery complex into the hinterland.

If we are to study any "modes" at all, we might better study the modes of accumulation, instead of the "mode of production." In the world system, production is the means to an end. That end is consumption and accumulation. It may be useful to study the differences, and also the mutual relations and combinations, of the "articulation" of "public" (state) and "private," "redistributive" and "market" modes of accumulation. It is doubtful that any of these modes, or other modes, have ever existed alone in any pure form anywhere. However, we should study not only how modes of accumulation differ and combine with each other "locally," but also how they interconnect with each other throughout the world system as a whole. Thus, world system history should both differentiate and combine modes of accumulation: horizontally through space as well as vertically through time. The "articulation" of modes is a way of analyzing how the mode(s) of accumulation in one zone of the world system is (are) affected by systemic links with other zones' mode(s) of accumulation. Can the overall world system be characterized by a single mode of accumulation? If not, why not?

Shifting the focus of analysis from production to accumulation need not mean abandoning analysis of the class structure. In fact, a focus on the relations of accumulation should sharpen the analysis of class relations. Geoffrey de Ste Croix (1981) argues that the key to every social formation is how the "propertied classes" extract the surplus from the working classes and ensure themselves a leisured existence. He defines a mode of production based on the means by which the propertied classes obtain most of their surplus. This approach is an akemativAto trying to determine what form of relations of production characterize the entire social formation. That is, he focuses on the dominant mode of accumulation. Ste Croix delineates several means of extracting surplus: wages, coerced labor (in many variants), rent, and through the state (via taxes and corvee labor, and through "imperialism"). Interestingly, Ste Croix explains the fall of the late Roman empire as due primarily to gross overextraction of surplus, overconcentration of wealth in the hands of the upper classes, and the overexpansion of the bureaucratic and military apparatus (1981: 502-3). The latter is similar to Paul Kennedy (1987) argument about military-economic overextension in The Rise and Fall of the Great Powers. This analysis implies a link between cycles of accumulation and cycles of hegemony, to which we will return below.

Equal, or perhaps even greater, analytical emphasis must be placed on horizontal interelite conflicts over apportioning "shares" of the available social surplus. This struggle has its focus in the ultimate political determination of the mode of accumulation. To say that the elites of different zones of the world system share in each others' system of exploitation and surplus extraction through interpenetrating accumulation is not to deny possible differences between these zones in terms of the mode of accumulation. The exchange or transfer of social surplus both affects and is affected by class structure. However, interpenetrating accumulation affects both the producing strata and the extracting/accumulating strata, though in different ways.

Perhaps the single greatest weakness in historical materialism to date has been the failure to theorize transitions between modes in a world systemic context. Traditional Marxist interpretations of world historical development relied heavily on a schema of transitions between modes of production in a predetermined unilinear progression. This oversimplistic framework of analysis has long since been abandoned and revised by most historical-materialists. We propose instead to study transitions between modes of accumulation. However, they did not occur merely within each "separate" zone of the world system. Rather they were the key determinants of transition in both the "parts" and especially the whole of the world system. Therefore, the research task is not to search solely or even primarily for indigenously generated determinants of transition between modes, but rather to analyze the overall interactions of each zone of the world system with the dynamic of the entire world system. This is true of both the economic and the political aspects of modes of accumulation.

It would also be a mistake to attempt too strict an analytical separation between "agrarian" and "industrial" modes of accumulation in the world system. Even in very archaic phases of the world system, the economic nexus included nonagricultural sources of production and accumulation. The roles of industry and commerce before the onset of "industrialization" in the modern world system require much more study than they have received. The associated social and political relations of accumulation have changed very significantly across world-historical time, but not in any predetermined or unilinear progression of modes of accumulation. The precise nature and timing of such transitions is still an open empirical question.

In principle, there are four possible permutations of private and public accumulation:

1 Dominant private accumulation (the state "facilitates" private accumulation).

2 Dominant state accumulation (private accumulation "facilitates" slate accumulation).

3 All private accumulation.

4 All state accumulation.

Type 1, dominant private accumulation, may correspond to mercantile states and to modern democratic states. Type 2, dominant state accumulation, may characterize a number of bureaucratic states and empires as well as certain modem authoritarian regimes. Type 4, all state accumulation, might be characterized by states such as ancient Sparta, the Inca empire, and some modern (state) "socialist" states. Type 3, all private accumulation, raises the theoretical question of whether private accumulation is in fact possible at all without the state, or at least without the presence of the state somewhere in the overall economic nexus. There may be niches in the world system's economic nexus where all private accumulation may occur, but it has been difficult to identify instances of this.

State accumulation is typically characterized by a much larger scale and much greater potential capabilities to extract surplus than any sole private accumulator is capable of organizing. That is why "im" is such an attractive means of accumulation. State accumulation centralizes accumulation more than private accumulation. For this reason, these two modes of accumulation and their respective elites are locked into a perpetual conflict over apportioning the shares of the surplus. Both private accumulating classes and the state elite, as a "state class," struggle to form a coalition of class fractions. Such a "hegemonic bloc" of class fractions allows them to cooperate to utilize the political apparatus to establish the dominant mode of accumulation. The oscillation between predominance by the private accumulators and the state class in a social formation is a key dimension of the cycles of accumulation, discussed below.

There is a contradiction between a relatively unbounded economic nexus and a relatively bounded political organization of this economic nexus in world system development. The total economy of the major states and centers of the world system is not under their sole political control. This tension is universally recognized today as affecting the structure of modern capital accumulation. However, this phenomenon is not new. This economy/polity contradiction is characteristic not only of the so-called contemporary age of "interdependence," but has in fact always been a factor in world system development.

Even though the world system has since its origin developed logistical interlinkages that create a single overarching economic system, the political organization of the world system has not developed a parallel unity. Why is that? For the modern world system, Wallerstein (1974 and other works) argues that the capitalist mode of production structurally inhibits the creation of a single "world empire." That is, in this view the resolution of the economy/polity contradiction in the modern world system by a single overarching political entity is inhibited by its capitalist mode of production. However, it appears that even in other modes of accumulation it has not been possible to create a single political structure for the entire world system. Attempts to do so have been failures. The Mongol attempt in the thirteenth century perhaps came closest to success. The question of why the world system has never successfully been converted into one political entity should be seriously posed. The answer may be structural, or simply a matter of logistical and organizational limitations. Whatever the answer to this question about politics in the world system, it need not deny and may even strengthen the thesis of its essential economic unity.

Hegemony is a hierarchical structure of the accumulation of surplus among political entities, and their constituent classes, mediated by force. A hierarchy of centers of accumulation and polities is established that apportions a privileged share of surplus, and the political economic power to this end, to the hegemonic center/state and its ruling/propertied classes. Such a hegemonic structure thus consists schematically of a hierarchy of center-periphery-hinterland complexes in which the primary hegemonic center of accumulation and political power subordinates secondary centers and their respective zones of production and accumulation.

The rise and decline of hegemonic powers and cycles of hegemony and war have lately received increasing attention, e.g. by Modelski (1987), Thompson (1989), Wallerstein (1974, 1988), Wight (1978), and Goldstein (1988), and even bestseller status (Kennedy 1987). Most of these studies confine themselves to the world system since 1500. However, we argue that the world system began earlier and was previously centered outside Europe. Therefore, the same, and more questions, about hegemonic rise, decline, cycles, and shifts apply - even more interestingly - to the larger and older world system, prior to Europe's rise to super-hegemonial economic and political power within it. Where and when were there hegemonic centers in the world system before 1500, and in what sense or how did they exercise their hegemony? David Wilkinson (1989) has made a systematic study of world states and hegemonies that could serve as the starting point for an answer.

The following are some other important questions. As one hegemonic center declined, was it replaced by another, and which and why? Were there periods with various hegemonic centers? Did they "coexist" side by side, or with how much systemic interconnection? In that case, did they complement each other, or did they compete with each other, economically, militarily, or otherwise until one (new?) center achieved hegemony over the others? Rather than continuing to look merely comparatively at contemporary hegemonic structures in different zones of the world system or to investigate the dynamic of each region separately, we must look at systemic links among all the constituent political organizations of the world system. Of course, these especially include contemporaneous hegemonic structures,

Hegemony takes a variety of historical forms. They vary from highly centralized integrated bureaucratic empires, to very loosely structured commercial or maritime hegemonies. In the latter, much of the surplus is captured not via direct political coercion, but via commodity exchange, albeit via unequal exchange. How and why do these various forms of hegemony occur at particular times and places? How do they reflect the interests of the actors which choose them and the prevailing conditions in the world system at the time?

Given the absence in the historical record of any single "world system hegemony," we must look to the rise and decline of hegemonies in each of the major zones of the world system in order to construct an overall picture of the hegemonial cycles, rhythms, and trends in the various regions and their possible relations. For instance, the oscillation between unitary hegemonies and multi-actor states systems has already been recognized as a key pattern of world-historical development (Mann 1986; Wilkinson 1989). These oscillations and the succession of hegemonies in each part of the world system should not be analyzed only on a comparative basis, but from a world systemic perspective. Only in this way can the dynamics of the world system's economy/polity contradiction be more fully understood.

All this suggests that the primary object and principal economic incentive; of a bid for hegemony is to restructure the overarching system of accumulation in a way that privileges the hegemon in capital/surplus accumulation. Simply put, hegemony is a means to wealth, not merely to "power" or "order." That is, "power" in the world system is both economic and political at all times. In fact, economic power is political power, and vice versa. Turning Mann (1986) on his head, the ends of power are above all control over accumulation processes and the determination of the dominant mode of accumulation. The processes of accumulation are more fundamental to world system history than Mann's forms of social power per se. The political and economic processes in the world system are so integral as to constitute a single process rather than two separate ones. Success in accumulation plays a critical role in success in a bid for hegemony. This is true not only of modem states, but even of archaic ones. For instance, the victory of the state of Qin in the Warring States period in Chinese history depended greatly on its innovations in tax structure, infrastructural investments, bureaucratic administration, and trade links to the world system. All of these gave the Qin very real advantages in accumulation and in military capabilities over its more traditional, "feudal" rivals.

The perpetual "symbiotic conflict" between private accumulating classes and state accumulating classes is indicative of cycles of accumulation. The oscillation between unitary hegemonies and multi-actor states systems is indicative of cycles of hegemony in the world system. Cycles of accumulation and cycles of hegemony are probably causally interrelated. This causal interrelationship appears to date frvery early in world system history in various parts of the world system.

These cycles and their interrelationship are the central phenomena of the world system's longest cumulative patterns. These cycles have partly been analyzed by Gills's (1989) analysis of synchronization, conjuncture, and center shift in the cycles of east Asian history. Briefly, prior to the industrialization of production, the phase of accumulation in which private accumulating classes become dominant seems to be closely associated with the decline of hegemonies and their political fragmentation. That is, decentralization of accumulation affects the decentralization of political organization. These processes may be called "entropic." Phases of accumulation in which the bureaucratic state elite is dominant seem to be associated with the consolidation of hegemonies. That is, the centralization of accumulation affects the centralization of political organization, and vice versa. However, rising and declining hegemonies also call forth opposing (and also temporarily supporting) alliances to thwart existing and threatening hegemonial powers. Shifting alliances seem to promote some kind of "balance of power." All this may seem obvious, but the cyclical dynamic of hegemony (also through political conflict and shifting alliances) in relation to the process of accumulation has not previously been given the attention it, deserves.

Implosion from the hinterland upon the center appears to be most likely to occur in entropic phases of the system. The hinterland, and perhaps the periphery, take advantage of weakness or entropy in the center to restructure the structure of accumulation. This may occur by usurping political power at the center, or by "secession" from the center altogether.

Too much attention has been given to the political and strategic aspects of long cycles of war and leadership with the exclusion of the underlying dynamics of accumulation. General war, as Modelski (1987) argues, does indeed produce new sets of victors who go on to establish a new order. However, one should not merely examine the political and military aspects of these cycles. The new victors, without exception, also proceed to restructure world accumulation. This and not mere political realignments or "order" alone is the ultimate end of such general conflict. The intense military rivalry that proceeds hegemony may stimulate production, but much of the economic benefit is consumed in the process of rivalry and war. Typically, a new hegemony is followed by a period of infrastructural investment and economic expansion, which is "the hegemonic prosperity 'phase" of accumulation. A unified hegemony usually reduces or even eliminates previous political obstructions to the greater integration of the economic nexus. This has a tremendous impact on the process of accumulation.

We must contemplate the existence, and study the development, of a wider world system farther back in world history to find answers to a host of questions about the dynamics of states systems and cycles of accumulation and hegemony. Particularly important are questions about the existence of world system wide accumulation processes and shifts in the centralization of accumulation from one zone of the world system to another. How do such shifts affect cycles of hegemony? What are the real patterns and "laws" of the world system's overall expansion, transformation, and decay?

The historical process of economic surplus management and capital accumulation is so interregional and inter-"societal" as to lead to the conclusion that it constituted a process of world accumulation in the world system over the millennia. A privileged position therein, in which one zone of the world system and its constituent ruling and propertied classes are able to accumulate surplus more effectively and concentrate accumulation at the expense of other zones, could be called "super-hegemony."

Thus, super-hegemony is also a class position in the overarching world-accumulation processes of the world system. A research agenda would be to examine the causes of possible super-hegemony, positional shifts from one zone to another, and the degree to which super-hegemony is transformed into further economic and political power within the world system. While hegemony is built up of center-periphery-hinterland complexes, super-hegemony occurs in the largest field possible, that of the entire world system and all its constituent hegemonic structures.

Super-hegemony links all the constituent hegemonies into one overarching systemic whole. Of course, the degree of institutional integration among distinct hegemonies is not as great as the degree of integration within each hegemony. Nevertheless, contemporary and/or contiguous hegemonies are not autonomous if interpenetrating accumulation exists. In the entire class structure of the world system, in whatever mode of accumulation, the super-hegemonic class position is the most privileged and the ultimate "center of centers" in the world-accumulation process.

To what extent did this overarching super-hegemony rest or operate on more than the mere outward exercise of political power and the radiation of cultural diffusion? In particular, to what extent and through what mechanisms did such overarching super-hegemony include centralized (super-hegemonic) capital accumulation? Was accumulation fed through the inward flow and absorption of economic surplus generated in and/or transferred through other (sub)-hegemonic centers? The answers to both questions are in general affirmative, for which we can find ample historical evidence if we only look for it. For instance, William McNeill (in conversation with Frank) suggests that China itself accumulated capital by absorbing surplus and capital from the West in the several centuries before 1500 ad. Was China therefore super-hegemonic? Prior to China, India was possibly super-hegemonic in the world system. In the period of the eighth and ninth centuries ad, the Abbassid caliphate, with its great metropole at Baghdad, may have been super-hegemonic. The development of European domination over the Mughal, Qing, and Ottoman empires should however also be understood in terms of the conjuncture of European expansion and these regions' entropic phases of accumulation and hegemony. In the nineteenth century, Great Britain is a candidate for super-hegemonic status, followed by the United States in the mid-twentieth century, and possibly Japan in the very late twentieth and early twenty-first centuries.

Thus, super-hegemony need not be limited only to the capitalist world economy, but may have existed at other times in the history of world system development. Super-hegemony is more flexible than empire, or imperialism. Super-hegemony operates not only through political and interstate levels of diplomacy, alliance, and war, but also and maybe more importantly, through super-accumulation.

If super-hegemony existed before recent times, how, when, and why did the super-hegemonic center of the world system, the most favored locus of accumulation, shift around the world system? What effects did such shifts in super-hegemonic centers have upon, and what "functional" role, if any, did they play in, the world system's development? For instance, the super-hegemony of the Abbassids in the eighth century was reflected in their ability to defeat Tang China at Talas in 751, their treaty of alliance with the Tang in 798 ad, and their continued ability to control Central Asia. Perhaps the super-hegemony of Britain contributed to its ability to arbitrate the balance of power on the continent of Europe and to defeat bids to impose a unitary hegemony, such as that by Napoleon? The super-hegemony of the United States after 1945 allowed it to restructure the international order and greatly expand its economic and military influence in the world system. It remains to be seen whether or how Japan might translate super-hegemonic status in world-accumulation processes into further political and economic power in the world system in the twenty-first century.

How long, then, has there been an overarching and interpenetrating world system process of capital accumulation, which affected the structure of the structures of which it is composed? In other words, how long has there been a cumulative process of capital accumulation on a world system scale? The (occasional and temporary) existence of super-hegemony also implies super-accumulation at those times, as noted above. Even in the absence of super-hegemony, however, the process of accumulation in one zone of the world system would not have been the same without the linkages to the process of-accumulation in another zone or zones of the world system. Therefore, even competing hegemonies and linked structures and processes of accumulation could have contributed to the world system wide cumulation of accumulation. Indeed, such an overarching structure of accumulation and the resulting process of cumulation of accumulation implies that there may be a unitary "logic" of systemic development.

The cumulation of accumulation in the world system thus implies not only a continuous, but also a cumulative, historical process of ecological, economic, technological, social, political, and cultural change. Cumulation of accumulation involves or requires no uniformity among these processes throughout the system or its parts, no unison among its parts, no unidirectionality of change in either the pans or the whole, and certainly no uniformity of speed of change.

On the contrary, both the historical evidence an our analysis suggest unity in diversity (to use the phrase Mikhail Gorbachev used at the United Nations). The unity of the world system and its cumulative process of accumulation are based on the diversity of center-periphery-hinterland, mode-of-accumulation, and hegemonic differences we have emphasized. Of course they also rest on the variety of social, gender, racial, ethnic, cultural, religious, ideological, and other differences, which characterize humankind. Historical change in both the whole (system) and its parts takes place in many "progressive" and "retrogressive" directions, and not unidirectionally or even in unison between here and there.

For this reason among others, historical change also takes place and even cumulates, not uniformly, but at changing rates, sometimes fast, sometimes slowly, sometimes (degenerating) in reverse. Indeed, as in physical transformations and in biological evolution, historical change suddenly accelerates and/or bifurcates at critical junctures. More than likely, contemporaries are rarely aware that they are living and acting in such "special" periods - and many at other times who think they are, are not. Hindsight seems to throw more light on history than foresight or even contemporary side-sight or introspection. Yet even historical hindsight has a long way to go, especially in grasping the dynamics and variability of historical change. We briefly return to these problems below under the subtitle "dynamics."

In this chapter we have made three key arguments. The first is that the world system predates the development of modern capitalism, perhaps by several thousand years. The second is that accumulation processes are the most important and fundamental processes of the world system throughout its development. The third argument is that, though the mode of accumulation underwent many historical transformations, there has been a continuous and cumulative process of accumulation in the world system. Therefore, we argue that a new research agenda is needed to focus more analysis on these cumulative processes of accumulation over the entire historical development of the world system - of some 5,000 years. The secular trends, cycles, and rhythms of the modern capitalist world system thus become contextually more understandable within the much longer cycles, trends, and rhythms of the historical world system, and particularly within its process and cycles of accumulation.

We base our argument upon a new set of criteria for defining what constitutes a "systemic" interaction. The transfer or exchange of economic surplus is the fundamental criterion of a world systemic relationship. Diplomacy, alliances, and conflict are additional, and perhaps derivative, criteria of systemic interaction. Thus, we introduce the criterion of "interpenetrating accumulation" into the definition of the world system. By applying these criteria we saw the origins of the world system recede by several millennia. The world system had its ultimate origins in the development of an archaic Afro-Eurasian economic and political nexus, which first developed in the area now known as west Asia, the Middle East, and the eastern Mediterranean about 2500 bc. Once in existence, this world system continued to develop and expand and deepen. It eventually assimilated and/or merged with all other center-periphery-hinterland zones to form our modern world system. Its relatively unbounded economic nexus is perpetually in contradiction with a more bounded political organization of the economic nexus. Cycles of accumulation and cycles of hegemony, like center-periphery-hinterland relations, have characterized the world system and its subsystems from its inception.

World system history forms a genuine continuum within which cycles of accumulation and cycles of hegemony are the two most fundamental phenomena. These two cyclical phenomena are causally systemically interrelated to one another. They are the basis of our assertion that there are cumulative accumulation processes in the world system over such an extended time frame.

Significant aspects of our argument were anticipated by Kasja Ekholm and Jonathan Friedman over a decade ago (Ekholm and Friedman, chapter 2 in this book). We agree that the "forerunner of the present kind of world system" emerged in ancient Mesopotamia. However, in our view this original formation of the world system was more the result of interregional relations between Mesopotamia and other regions in the "Middle East" and the Indus Valley, rather than their "incorporation" by the Mesopotamian Early Dynastic system. We agree that the world system then expanded and took on certain "general properties", which still define it today. We concur that such general properties "are common to ancient and modern worlds irrespective of specific local forms of accumulation."

Ekholm and Friedman and we agree on the centrality of capital accumulation in. this long historical process and system(s) and that "capital" exists not only under "capitalism" and "is not tied to a specific form of exploitation." However, our concept of capital and its accumulation is broader than theirs. They define capital as abstract wealth represented in the concrete form of metal or money that can be accumulated in itself and converted into other forms of wealth. We stress the existence and combination of both state and private capital, as does Chase-Dunn (1989), and we include nonmonetary forms of the production, extraction, transfer, and accumulation of surplus. We also pay more attention than they do to the interregional dimensions of accumulation and supra-regional super-accumulation. Moreover, we stress the cumulative, albeit cyclical, process of capital accumulation - which also contributes to continuity in the world system.

Ekholm and Friedman argue, as we do, that the system is also characterized by center-periphery structures that are unstable over time and that centers expand, contract, and sometimes collapse as regular manifestations of shifts in the locus of accumulation. We have extended this to include the hinterland, which in our view also contributes to accumulation in the center and to transformation in the system as a whole. Moreover, again, we stress the systemic relations among various center-periphery-hinterland complexes, which make up the world system as a whole.

We agree with Ekholm and Friedman that systemic economic relations tend to be more extensive than political ones, and that this is a fundamental "weakness" orcontradiction of the world system(s). This contradiction gives rise to instability in and transformation of the system. Yet as Ekholm and Friedman point out: "The development of total systems is not equivalent to the development of individual societies." We discuss these relations and transformations as cycles of hegemony. We also relate hegemony to the center-periphery complex and to accumulation within it. However, we also urge the study of possible overarching system-wide super-accumulation and super-hegemony.

Our debate with Ekholm and Friedman is primarily over the issue of continuity within the world system. They stress "the fundamental continuity between ancient and modem world-systems" and admit the possibility of the same world system. We completely concur with their view that the forms of accumulation have not changed so significantly and that the forms of exploitation and oppression have all existed from very ancient times, though in different proportions and in varying combinations. However, we wish to stress a fundamental similarity and continuity not so much of ancient with modern world systems. We are definitely talking about common characteristics and continuity within the same world system.

Ekholm and Friedman continue to argue that "the similarities between empirically delimited systems are great enough to warrant a continuity hypothesis." For them this continuity is of global processes and not of a "physically delimited portion of the globe," and "system" refers to "systemic properties of globally open processes rather than to an operationally definable empirical entity." The distance between their position and ours may not be as great as it might appear and is certainly bridgeable. We recognize and endorse their attention to the nature of local structures and the attempt to establish to what degree they are products of the same larger system, rather than simply connected to it. Because Ekholm and Friedman view global systems as multistructural, that is, containing many articulations among local and global processes, they argue that there is "no obviously established 5,000 year old system." Nevertheless, the gap between our positions is narrowed by their final conclusion that world history is marked by "repetition compulsions and a scenario of imprisonment in larger systems." Their global systemic anthropology is a necessary complement to our world system hypothesis and clarifies issues of the articulation of the local and global processes within a context of global systemic continuities.

Therefore, there is good reason, justification, and merit in constructing a historical-materialist political economy of world system history. Almost all historical and (other) social-scientific analysis of the world and its pans before 1500 ad (and most analyses of the time since then also) have neglected these systemic aspects of world-historical political-economic processes and relations. Some scholars (e.g. Tilly 1984) have considered undertaking such a world system history and have rejected the task as inadvisable or impossible. Others, like Farmer (et al. 1977, 1985), Chase-Dunn (1986, 1989), Ekholm (1980), and Ekholm and Friedman, in chapter 2, have started down this road, but have apparently taken fright and stopped or even turned back. A few scholars, especially Childe (1942), McNeill (1964, 1990), Stavrianos (1970), and most recently Wilkinson (1987, 1989) have made pioneering advances toward writing a world system history. Frank (1990) examines their and many other theoretical and historical considerations and rejects their reservations as unfounded. He then proposes why and how these and other pioneering works should be extended and combined for a history of the world and its world systemic historical-materialist political economy along the present lines.

A historical-materialist political economy of cumulation of accumulation in world system history does not exclude or even downgrade social, political, cultural, ideological, and other factors. On the contrary, it relates and integrates them with each other. Nor need such a study be "economic-determinist." On the contrary, this study would recognize the interaction and support of at least three legs of the social stool, without which it could not stand, let alone develop. These three legs are: organization of political power; identity and legitimation through culture and ideology; and management of economic surplus and capital accumulation through a complex division of labor. Each of these is related to the other and all of them to the system as a whole and its transformation.

A historical-materialist political-economic analysis of the historical development of this world system should incorporate ecological, biological, cultural, ideological, and of course political factors and relations. Thus, there is justification and merit in also seeking to explain many political institutions and events and their ideological manifestations through the ecological and economic incentives and limitations that accompany if not determine them. In particular, we should pay much more attention to how the generation and capture of economic surplus help shape social and political institutions, military campaigns, and ideological legitimation. Economic institutions, such as Polanyi's (1957) famous reciprocity, redistribution, and market, appear mixed up with each other and always with some political organization. Many political institutions and processes also have economic aspects or "functions."

The three component aspects, the three legs of the stool, are embedded in the mode of accumulation. No mode of accumulation can function without a concomitant ideology of accumulation; an economic nexus founded on a complex division of labor m which class relations facilitate extraction of surplus; and finally a political apparatus, which enforces the rules and relations of accumulation through the ultimate sanction of "legitimate" coercion. The ideology and political apparatus are integral aspects of the mode of accumulation. They are not super-structurally "autonomous" from each other or from the characteristics of the economic nexus. However, ideology and political competition and emulation sometimes appear to take on at least a semi-autonomous character. Even if we grant this, it does not invalidate the alternative assertion that overall they are not autonomous from the economic nexus.

We reject any vulgar unidirectional schema of causality whereby the economic nexus must necessarily determine the ideology and political apparatus of a mode of accumulation because they are not in fact separate. We suggest an alternative concept of the mutual intercausality among the three aspects of a mode of accumulation which is historically specific to each case. Indeed, particularly in periods of transition between one mode of accumulation and another, ideological and political forces can play an extremely significant role in determining the structure of the economic nexus that emerges from the transition. It is in these periods especially that broad-based social movements intercede in world (and local) history. These social movements are often neglected altogether, or they are considered but not sufficiently analyzed in their structural and temporal world systemic context. We can well depart from vulgar economism, but not necessarily from a form of "economic" determinism, if by economic we mean giving the political economic processes of accumulation their due.

Most important perhaps are the dynamics of the world system, that is how the world system itself operates, behaves/functions, and transforms (itself?). Are there trends, cycles, internal mechanisms of transformation in the pre-(and post-)1500 world system? When and why does historical change accelerate and decelerate? What are the historical junctures at which quantitative turns into qualitative change? What are the bifurcations at which historical change takes one direction rather than another? And ? Perhaps general-systems theory offers some answers or at least better questions also for this (world) system. For instance, Prigogine and Sanglier (1988) analyze how order is armed out ot chaos and how at critical times ana places small changes can spark large alterations and transformations in physical, biological, ecological, and social systems.

Recent studies by, for instance, Ekholm and Friedman (chapter 2 in this book) and Chase-Dunn (1989) are looking into both structural and dynamic properties of partial "world" systems before 1500. However, it may be possible to trace long (and within them shorter) cycles of accumulation, infrastructural investment, technological change, and hegemony much farther back in world system history. Not only may they have existed, but they may often have had considerable relative autonomy from policy and politics per se. Indeed as in more recent times also, much of this policy was and is instead more the effect of and response to largely uncontrolled cyclical changes. Moreover, policy tends to reinforce more than to counteract these cycles and trends. This cyclical process and policy response may be seen in the decline of various empires, including the present American one.

In particular, to what extent has the process of capital accumulation and associated other developments been cyclical? That is, were there identifiable subsystemic and system-wide acceleration/deceleration, up/down, swings in structure and process? And were any such swings cyclical, that is endogenous to the system, in the sense that up generated down and down occasioned up again? This kind of question has been posed and some answers have been offered for the world system (or its different economic and political interpretations) since 1500. For instance, Wallerstein (1974) and Frank (1978a, b) find long cycles in economic growth and technology. Modelski (1987) and Goldstein (1988) find long cycles in political hegemony and war. Wallerstein also posits a life cycle of expansion and foreseen decay of the system. Toynbee (1973), Quigley (1961), Eisenstadt (1963), and others have made comparative studies of the life cycles of individual civilizations before 1500. So have archaeologists like Robert M. Adams (1966). But to what extent were there also world system wide fluctuations and cycles, and what role have they played in the transformation and development of the world system?

Infrastructural investment apparently occurs in cyclical or phased patterns, and in direct correspondence with the cycle/phase of accumulation and of hegemony. Newly formed hegemonic orders are usually associated with a subsequent intense phase of infrastructural investment, followed by general economic expansion and a concomitant increase in accumulation. Therefore, it could also be fruitful to search for a long-lasting continuous up-and-down cycle of super-hegemony.

Thus, infrastructural investment cycles would be related to cycles of accumulation and cycles of hegemony in the world system. Are there also cumulative aspects of infrastructural investment that affect subsequent world system development? An affinative answer does not imply we take the position of a single "capital-imperialist" mode of production based on the use of imperial political power as a political apparatus of accumulation throughout world history as posited by Ekholm and Friedman (1982). How did private and state infrastructural investment interact in world system development? For instance, what is the role of private infrastructural investment in creating and sustaining the complex logistical interlinkages of the world system? To what extent does state infrastructural investment create and sustain the logistical interlinkages of the world system? How does the conjuncture and synchronization of phases among contemporary hegemonies affect the respective cycles of infrastructural investment?

If we view the entire five-six millennia development of the world system as a unified cumulative continuum and seek to explain its most significant trends, cycles, and rhythms, based on a historical-materialist political economy, then a "world system history" should follow. Such a world system history should not merely be a comparative history of the world or even a comparative history of world systems. A historical-materialist world system history would regard class formation, capital accumulation, state formation, and hegemonic construction throughout the world system as being integral aspects of the one, cumulative, process of world-historical world system accumulation and development. This history would not be Eurocentric, and should avoid any other form of centricity. A comprehensive world system history would be humanocentric.

NOTE

This chapter first appeared in 1990/1 as "The cumulation of accumulation: theses and research agenda for 5000 years of world system history," in Dialectical Anthropology (New York/Amsterdam), 15 (1) (July 1990): 19-42. An expanded version was published as "5000 years of world system history: the cumulation of accumulation," in Precapitalist Core Periphery Relations edited by C. Chase-Dunn and T. Hall, Boulder, CO: Westview Press, 1991: 67-111.

REFERENCES

Adams, Brooks (1943) The Law of Civilization and Decay: An Essay on History, New York: Alfred A. Knopf.

Adams, Robert M. (1966) The Evolution of Urban Society: Early Mesopotamia and Prehistoric Mexico, Chicago: Aldine.

Amin, Samir (1989) "Le Systeme mondial contemporain et les systemes amerieurs" (manuscript).

Arrighi, Giovanni and Drangel, Jessica (1986) "The stratification of the world-economy: an exploration of the semiperipheral zone," Review 10 (1) (summer): 9-74.

Chase-Dunn, Christopher (1986) "Rise and demise: world-systems and modes of production" (manuscript) (Boulder: Westview Press, forthcoming).

覧 (1989) "Core/periphery hierarchies in the development of intersocietal networks" (manuscript).

Chase-Dunn, Christopher and Hall, Thomas D. (eds) (1991) Precapitalist Core/Periphery Relations, Boulder: Westview Press. Childe, Gordon (1942) What Happened in History, Harmondsworth, Pelican.

覧 (1951) Man Makes Himself, New York: Mentor.

de Ste Croix, G.E.M. (1981) The Class Struggle in the Ancient Greek World, London: Duckworth.

Curtin, Philip (1984) Cross-Cultural Trade in World History, Cambridge: Cambridge University Press.

Eberhard, Wolfram (1977) A History of China, 4th rev. edn, London: Routledge & Kegan Paul.

Eisenstadt, S.N. (1963) The Political Systems of Empires, Glencoe, IL: The Free Press.

Ekholm, Kajsa (1980) "On the limitations of civilization: the structure and dynamics of global systems," Dialectical Anthropology 5 (2): 155-66.

Farmer, Edward L. (1985) "Civilization as a unit of world history: Eurasia and Eurasia's place in it," The History Teacher 18 (3) (May): 347-63.

Farmer, Edward L. et al. (1977) Comparative History of Civilization in Asia, Reading, MA.: Addison-Wesley.

Frank, Andre Gunder (1978a) World Accumulation 1492-1789, New York:

Monthly Review Press; London; Macmillan.

覧 (1978b) Dependent Accumulation and Underdevelopment, New York:

Monthly Review Press; London: Macmillan.

覧 (1981) Crisis: In the Third World, New York: Holmes & Meier; London:

Heinemann.

覧 (1990) "A theoretical introduction to five thousand years of world system history," Review 13 (2) (spring): 155-248.

Gernet, Jacques (1985) A History of Chinese Civilization, Cambridge: Cambridge University Press.

Gills, Barry K. (1989) "Synchronization, conjuncture and center-shift in east Asian international history," paper presented at the joint International Studies Association, British International Studies Association Conference, London, 1 April.

Gilpin, Robert (1981) War and. Change in World Politics, Cambridge: Cambridge University Press.

Goldstein, Joshua (1988) Long Cycles: Prosperity and War in the Modem Age, New Haven: Yale University Press.

Hallｻ Thomas D. (1986) "Incorporation in the world system: toward a critique," American Sociological Review 52 (3): 390-402.

Johnstone, Paul (1989) The Seacraft of Prehistory, London; Routledge.

Kennedy, Paul (1987) The Rise and Fall of the Great Powers, New York: RHouse.

Lattimore, Owen (1940) Inner Asian Frontiers of China, Boston: Beacon Press.

覧 (1962) Studies in Frontier History: Collected Papers 1928-1958, Oxford: Oxford University Press.

Mann, Michael (1986) The Sources of Social Power, vol. 1: A History of Power from the beginning to ad 1760. Cambridge: Cambridge University Press.

McNeill, William H. (1964) The Rise of the West. A History of the Human Community, Chicago: University of Chicago Press.

覧 (1982) The Pursuit of Power: Technology, ArmedJ'orce and Society since ad 1000, Chicago: University of Chicago Press.

覧 (1990) "The Rise of the West after twenty-five years," Journal of World History 1 (1).

Modelski, George (1987) Long Cycles in World Politics, London: Macmillan.

Polanyi, Karl (1957) The Great Transformation: The Political and Economic Origins of Our Time, Boston: Beacon Press.

Prigogine, Ilya and Sanglier, Michele (eds) (1988) Laws of Nature and Human Conduct. Sfecifidttes and Unifying Themes, Bruxelles.

Quigley, Carrol (1961) The Evolution of Civilisations. An Introduction to Historical Analysis, New York: Macmillan.

Rowlands, Michael, Larsen, Mogens, and I^istiansen, Kristian (eds) (1987) Center and Periphery in the Ancient World, Cambridge: Cambridge University Press.

Schneider, Jane (1977) "Was there a pre-capitalist world system?". Peasant Studies 6 (1); 30-9.

Stavrianos, L.S. (1970) The World to 1500. A Global History, Englewood Cliffs: Prentice-Hall.

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

Thompson, William (1989) On Global War: Historical-Structural Approaches to World Politics, Columbia: University of South Carolina Press.

Tilly, Charles (1984) Big Structures, Large Processes, Huge Comparisons, New York: Russell Sage Foundation.

Toynbee, Arnold (1973) A Study of History, Oxford: Oxford University Press.

Wallerstein, Immanuel (1974) The Modem World-System, vol. 1, New York: Academic Books.

覧 (1988) The Modem World-System, vol. 3, New York: Academic Books.

覧 (1989) "The West, capitalism, and the modem world-system," prepared as a chapter of Joseph Needham, Science and Civilization in China, vol. 7: The Social Background, part 2, sect. 48: "Social and economic considerations" (forthcoming).

Wight, Martin (1978) Power Politics, New York: Holmes Meier. Wilkinson, David (1987) "Central civilization," Comparative Civilizations Review 17 (Fall): 31-59.

覧 (1989) "The future of the world state: from civilization theory to world

politics," paper presented at the annual meeting of the International Studies

Association, London, 28 March-1 April. Wolf, Eric (1982) Europe and the People Without History, Berkeley: University of California Press.

Barry K. Gills

The purpose of this chapter is to present an argument in favor of a new general organizing concept for the study of world history and the central role within it of the world accumulation process and hegemonic power. Rather than viewing transitions between discrete modes of production as the general organizing concept, world history can be analyzed as a series of hegemonic reorganizations or "hegemonic transitions" entailing shifts in the locus of accumulation in the world economy. These hegemonic transitions are very far-reaching in their overall economic, social, and political consequences, composing what might be called a transition in the "mode of hegemony" and thus affecting the character of world order as well as the composition of all the "societies" in that order. The hegemonic transitions reflect the underlying rhythm of competition in the world system and especially the cycles of accumulation in the world economy. However, I argue that the conventional single-hegemon model is seriously misleading and would better be replaced by a new concept of "interlinking hegemonic powers" which more often characterizes the world system.

In addition, I will argue that the conventional sharp dichotomy imposed between premodem and modern states and economies is unwarranted and misleading. In my view, the world system has always, for thousands of years, been characterized by a mixture of modes of capital accumulation involving both private capital and the state. Most importantly, it should be accepted that trade and commerce have always played a crucial determining role in the world accumulation process. This is true even in historical periods supposedly dominated by so-called "bureaucratic empires" or "world empires."

Finally, I argue that beyond the many changes in mixed modes of production, there is perhaps a more fundamental patterning in the world system. This patterning is both economic and political, touching both accumulation and hegemonic power at the same time. It is the cyclical concentration and subsequent diconcentration of control over surplus and capital, by classes within states, classes between states, and between states themselves. This pattern is paralleled in the rise and decline of hegemonic powers and in the occurrence of periodic world crises. This brings us to a tentative conclusion implying a general theory of the causes of periodic economic and hegemonic crises which combines class struggle with cyclical change and conjunctural moments of historical transformation.

The notion of hegemony has probably been the most debated term in international-relations literature in recent years (Higgott 1991: 97). The word derives from the Greek "hegemon," which simply means "leader." Two ideas seem to dominate current thinking on hegemony in the international system. First, hegemony is usually regarded as being more about political and military power than economic power, and secondly it is usually held that hegemony passes from one power to another in a succession from "like to like" insofar as the attributes of each hegemonic power are held to be very similar. Conceiving hegemonic succession as long historical patterns of the "rise and fall" of empires or great powers has been a very common way in which scholars have acknowledged it to be a fundamental pattern of change in the international system (Toynbee 1946; Eisenstadt 1963; Wight 1977, 1978; Gilpin 1981; Doyle 1986; Mann 1986; Chase-Dunn 1989; Modelski 1987; Kennedy 1987).

While it is true that the conventional understanding of hegemony focuses on a hierarchy of power, most scholars who study hegemony and long historical patterns of ascendance and decline recognize that hegemonic power is not simply or solely a matter of military and political power. Recently, Paul Kennedy examined the relationship between economic power and military-political power in the rise and fall of the great powers over the past five centuries (Kennedy 1987). Kennedy's thesis emphasizes the very close interrelationship between these dimensions of power: in the long run military-imperial power is unsustainable without a sufficient economic base. Michael Mann (1986) defines social power as being broader than military power alone, yet remains traditional in his orientation, i.e. he dubs his magnum opus "a history of power," thereby declaring his interest is primarily in political institutions. George Modelski and W.R. Thompson (1988) are even more traditional, since they focus explicitly on military, and specifically naval, indices of power in their study of successive "world leaders" in the past five centuries. Nevertheless, they also recognize a very complex and critical relationship between technological change and the pursuit of power.

Alternatively, some scholars suggest that we should have a definition of hegemony that focuses explicitly on economic processes, while not separating these completely from the realm of politics and military power. Wallerstem (1974, 1980, 1988) defines hegemony in a way that depends primarily on very specific economic criteria, whereby a core achieves supremacy sequentially in the spheres of production, commerce, and finance. Keohane (1980, 1984) defines hegemony using both economic and power-political criteria. His formula relates economic capabilit, expressed as a high concentration of economic power, to attainment of hegemonic political power. Keohane's work sparked off much debate on "hegemonic stability," which implicitly recognizes the importance of economic processes to hegemony and vice versa. Frank (1978) views the world capital-accumulation process, with its characteristic cycle of expansion and contraction, as the underlying context of hegemonic competition in the world system from 1492 to 1789, and indeed to the present. Braudel (1982) focused not on the familiar succession of political-military hegemonic states, but rather on a succession of key cities identified by their primary role in capital accumulation in a world economic, or at least regional European, framework. His departures from a state-centric analysis to one based on shifts in the locus of accumulation in the world economy is indicative of a fundamental change of the unit of analysis.

Gills (1987, 1989a, b) and Gills and Frank (chapter 3 above) explicitly define hegemony as a hierarchical structure of accumulation between classes and states, mediated by force. In this definition of hegemony, economic and political dimensions are inseparable. The essential feature of hegemony therefore is not formal political domination per se, but rather a hierarchy of centers of accumulation, as well as polities. This hierarchy apportions a privileged share of surplus, and the political-economic power to this end, to the hegemonic center/state and its ruling/propertied classes (Gills and Frank, chapter 3). Force is always an element in the exercise of hegemonic power, though other "economic" means of attaining surplus are also at work. This formulation does not require that either military power explains the attainment of economic power or vice versa. Rather it assumes that both are always employed in the pursuit of hegemonic power.

Hegemony is more than just a hierarchy of power among states. It is a complex pyramid of actors operating at many levels of social organization. At the apex of the hegemonic pyramid are the elite classes in the hegemonic coalition, classes located both in the center and in the periphery, i.e. dispersed throughout the pyramid at key points. These classes are themselves composed of elite families and individuals. Inter-elite relations within a hegemonic pyramid combine elements of competition, cooperation, and subordination, whatever the modes of production through which accumulation is occurring. This way of understanding hegemony is intended to synthesize two dimensions of analysis: military-political competition in systems of states, and economic processes of surplus transfer and its centralization (i.e. accumulation).

By contrast, realist models of international relations that assume that all states are intrinsically equal or similar because they are states either ignore or at least do not adequately account for the hierarchy of accumulation.

Nor do they recognize the effects on each actor of its structural position within the hierarchy of accumulation. They admit a hierarchy of power based on unequal distribution of capalysities between states, but the paradigm treats these states as if each is a completely separate discrete actor in the economic sphere. In short, the sociological-economic dimension of realism is too crude. In reality, no state is as impermeable or discrete an entity as they represent it. This definition of hegemony is also quite distinct from that employed by Immanuel Wallerstein. In particular, a distinction must be drawn between the Gills and Frank conception of how hegemonies operate in the world economy and Wallerstein's concepts of the world-empire and the world-economy. Wallerstein holds these two to be not only distinct, tut really opposite types of political economy. A world-economy is distinguished by a multiplicity of states, whereas a world-empire is an economy presided over by one overarching imperial state. The capitalist mode of production requires and at the same time sustains a multiplicity of states and therefore remains a world-economy. According to this concept, if a world-empire were to emerge, the capitalist world-economy would cease to exist � having become something quite different.

Wallerstein's terms perpetuate a sharp dichotomy between the idea of a politically determined mode of accumulation (the world-empire) and the economically determined form of the (capitalist) world-economy. Wallerstein has devised them in such a way as to emphasize the supposed sharp historical break in forms around 1500 ad. In our terms a world-empire has probably never actually existed on the scale of the world system as a whole. The only "world-economy" we recognize is the sole world economy of the entire world system. Therefore we prefer to drop the use of Wallerstein's terminology and we speak of hegemonies, denned above, rather than world-empires. The world economy we recognize has virtually always been characterized by a multiplicity, not only of states, but of interlinked hegemonic powers. Thus there was no sharp historical break in hegemony in the world economy around 1500 ad in the sense that Wallerstein would have it. There was an important hegemonic reorganization in the world economy at that time however, entailing a historical shift in the locus of accumulation from "East" to "West."

A definition of hegemony that emphasizes the integral nature of the hierarchy of power and the hierarchy of centers of accumulation shares much in common with the emerging Gramscian approach to international relations (Cox 1981, 1983, 1987; Gill 1990, 1991; Gills 1993). What they share is a move away from the single-power model of hegemonic succession and toward a more complex multilayered international political economy of hegemonic transition.

The key significance of the emerging Gramscian perspective on international hegemony is that it encourages us to examine not only the military ; and productive capabilities of states as the motor of hegemonic transition, but also, and perhaps most importantly, to examine how class coalitions are constructed and how ideology and culture are employed both to construct and to legitimate a hegemonic order. It is very rare in history that any elite dares to rule by force alone. It must seek consent, if not consensus, for its leadership or even domination. In my view this very "political" perspective actually enriches a materialist analysis of hegemonic power. It also reinforces the shift toward a more fluid and flexible concept of hegemonic power, one in which power is much more diffuse than any model of single-state hegemonic succession implies.

Previously, the term "transition" was usually reserved for change on a very broad sociohistorical scale. For historical materialists in particular, the most important transitions that shaped the course of world history were those between modes of production. I will argue that "hegemonic transition" is as useful a concept, if not more so, as transition between modes of production, or as "hegemonic succession" for understanding the patterns of change of accumulation, power, and world order throughout world history.

Taking this general point somewhat further, I would argue that it is possible to view all of international or world history as a series of hegemonic transitions entailing recurrent shifts in the locus of accumulation in the world economy. It follows from this hypothesis that these hegemonic transitions, or alternatively "center shifts," are a central form of historical change, i.e. as much a fundamental change as transitions between "modes of production," "historical social systems," or "civilizations" were previously presumed to be. Perhaps hegemonic transition and center shift are more real than transitions between the above, which exist primarily as analytical constructs, and therefore their boundaries are more easily identified.

If hegemonic transition is the central concept this implies a fundamental rethinking of the agencies of change in world history. It poses anew the problem of the relationship between "internal" "external" factors as explanations of historical change. This debate was already begun between "productionists" and "circulationists" some time ago. However, if the nature of the "international" or "external" arena is reconceptualized as a hierarchy of centers of accumulation in which the hierarchy of power is embedded, the debate could enter a new phase. It also re-poses the problem of the relationship between change "from above" and change "from below" in the social hierarchy.

My intention is to build upon the insights of Janet Abji-Lughod (1989), who argues that a world system is not always dominated by a single hegemon, but may be characterized by a number of coexisting core powers (or in my terms interlinked hegemonic powers) that via both conflictual and cooperative relations become increasingly integrated. To Abu-Lughod, therefore, "hegemonic transition" would not be best understood as a process of absolute rise and fall by states. Rather, she emphasizes relative position in a complex multilayered hierarchy. Over the course of world history some nations, or groups of nations, gain relative power vis-a-vis others. Thus they occasionally succeed in "setting the terms of their interactions with subordinates." This is a "rise." Conversely, the loss of such a (temporary) advantageous position is referred to by Abu-Lughod as a "decline" (Abu-Lughod 1989).

Within Abu-Lughod's formulation there is an implicit conception of movement up and down a complex, multilayered hierarchy of economic and political power in the world system. The hierarchy of political power is embedded or "nested" in a hierarchy of economic power embedded in the world economy. If we accept that there has been a world economy at Eurasian scale for far longer than the past five hundred years (Chase-Dunn 1989; Gills and Frank, chapters 3 and 5) then several other points follow. First it becomes possible to view hegemonic rivalries as a continuous process accompanying the development of the world economy. Secondly, it becomes necessary to distinguish between purely regional hegemony (the "empires") and world hegemony. I would argue that hegemony on the scale of the world economy, unlike the regional form of hegemonic power (and perhaps not even that), has never been held exclusively by a single power or its ruling/propertied classes. Rather, especially global or world hegemony is always shared hegemony, exercised through a complex network composed of class coalitions, and also alliances and other forms of association between states, including competitive ones.

The world system as a whole is certainly never simply dominated by one great hegemonic power, but rather is characterized by interlinking hegemonic powers - which are typified in their mutual relations by both competitive and cooperative interactions, i.e. "independence," "interdependence," and "dependence". Changes in the configuration of relations between these hegemonic actors can have a truly profound impact on the course of history and social development. This impact may be equal to or even greater than the impact of class struggle between the exploiter and the exploited classes (i.e. the accumulating and the producing classes). In fact, the outcome of class struggle may often depend ultimately on the outcome of these hegemonic struggles, at least as much, if not more, than the other way around.

This idea flows from a conception of the world system as an interlinked hierarchy of centers of accumulation, as opposed to a simple hierarchy of states and their power. For example, the Pax Americana is probably better understood as a complex coalition of classes and states in a shared global hegemony than as the overwhelming power of a single state (Gill 1990; Van der Fiji 1984). The consolidation of US hegemony after 1945 is accompanied by west European and Japanese economic power, of course, but also by European and Japanese political influence, operating largely in subordinated harmony with US power. This coalition operated in a context of global rivalry with the Soviet Union and other communist powers for hegemonic position. Likewise, British global hegemony in the nineteenth century cannot be properly assessed in isolation from the coexisting (global) imperia of other contemporary great powers and the specific relations established among the great powers within Europe after the Congress of Vienna. To venture much farther back in world history for a moment, our western view of the sole dominance of the Roman empire in the ancient world is fundamentally flawed by the prevailing Eurocentrism. In reality, the regional Roman imperium coexisted with other very powerful and wealthy hegemonic actors, such as the Parthians in Mesopotamia, followed by the Sassanid Persian empires, all of which were embedded in the same Eurasian-wide economic relations, which included Indian, Central Asian, and Chinese states and empires as well.

As an alternative conception to the single-hegemon-succession model, it can be argued that the world system as a whole goes through a cycle composed of periods when several hegemonic powers rise and coexist together, and periods when several hegemonic powers decline together or when hegemonic power is in disarray and competition and conflict increase, i.e. a period of general world political and economic crisis. These hegemonic power cycles seem to be correlated with long cycles of economic expansion and contraction (or at least slower growth or some form of dislocation). Gills and Frank (in chapter 5) trace the occurrence of these cycles back at least two thousand years. These are not simply parallel developments, but are synchronized. That is to say, there is a common causal link between them. My preferred hypothesis is that this link comes from their mutual participation in the world economy and in its single hierarchy of accumulation. However, though some hegemonic powers decline there are always ascending powers, even in periods of general crisis. Even the worst economic crisis, though it certainly brings about much political, social, and economic restructuring and a change of the geopolitical landscape, does not mean the disappearance of hegemonic power altogether. The world economy as a whole never ***, rather the ways in which it is constituted and the linkages through which it operates are I changed. This process favours some at a particular time while discriminating against others, and so on through time me world historical process to which I refer above is not merely a 1 rearrangement of players through time and space, but entails the restructuring of all the players as well as of the world system itself. It could be more broadly understood, as I have argued elsewhere, as a perpetual politico-economic process of mutual societal penetrations and transformations ... [in which] coexisting classes and states interlock in competitive/cooperative relationships of accumulation and rivalry. These relationships not only determine shifts in the "balance of power" or configuration of international hierarchy over time, but equally, if not more importantly, they constantly force restructuring on all of the classes, states, and societies imer-locked into these competitive/cooperative relationships. This constant process of societal restructuring should be recognized as the real subject matter of the discipline of international relations. (Gills 1993; see also Gills and Palan 1993)

From this (new) perspective, hegemonic transitions in the world system may be viewed as an unbroken series entailing cumulative development: but composed of both secular and cyclical change. Over the tongue duree, the long passage of sociohistorical time in which fundamental social structures are embedded, the world system expands spatially, for instance (a secular trend), while simultaneously undergoing internal restructuring (often of a cyclical character) or "deepening." The hegemonic transition is therefore not simply a repetitive cycle. At the beginning of each new historical period certain conditions will have changed that make it different from the preceding period. In particular, as the pace of technologicchange increases the difference between one hegemonic period and another may be considerable, despite other continuities.

For example, underpinning all hegemonic transitions is a secular developmental and underdevelopmental process which restructures the hierarchy of center-periphery relations, and center-center relations. This constant process of restructuring occurs locally, regionally, and now globally. There is an underlying process of capital accumulation on a world scale, which itself demands that certain types of restructuring occur in order for world accumulation to continue and expand. Therefore, secular developments in technology and the organization of the production system intertwine with cyclical rhythms of capital formation, and both with social and political developments. Mandel (1980) has examined in a very sophisticated manner the developmental logic of such interacting secular and cyclical patterns for the period of modern history since the 1780s. The long-term relationship of consumption to production, the rates of profit, investment, and exploitation, the technological cycles of innovation, the Kondratieff waves, and the form of social regulation, all appear in Mandel's examination of the development of modem capitalism. However, Mandel did not fully integrate the notion of hegemonic-power transitions ***** analysis. The locus of accumulation, and with it the locus of hegemonic power, shifts in response t6 all of these world historical forces above, operating in conjunction with one another.

I argue that the accumulation process is the ultimate driving force of hegemonic transition and thus of world order. This materialist analysis of the primacy of economic processes in the evolution of world order is not a "return" to past positions, but is even more relevant today than in the past. It stands in contrast to the explanations of a reinvigorated and redeployed idealist analysis of world order which explains macrohisiorical change as the working out of some great historical idea, such as "freedom," or more topically "democracy." If world history has any real "end" it is most likely the (capital) accumulation process itself, in whatever specific historical form. Hegemonic power is a means to that end. As the forms of accumulation change so do the forms of hegemonic power and thus the form of world order. I believe that the historical evidence shows that the sequence is ultimately in that order and not the other way around.

The second set of insights I wish to expand upon are those of Gilpin (1981) concerning the cycle of hegemonic rise and decline and the role of economic surplus. I hope the reader will pardon the exceptionally long quotation which follows, but it is necessary to do justice to the full range of Gilpin's formulation in order that I may later relate these points to the arguments above and those which follow. According to Gilpin (1981):

The territorial, political and economic expansion of a state increases the availablility of economic surplus required to exercise dominion over the system (Rader, 1971, p. 46.). The rise and decline of dominant states and empires are largely functions of the general and then the eventual dissipation of this economic surplus [p. 106]....

The type of social formation is extremely important because it determines how the economic surplus is generated, its magnitude, and the mechanism of its transfer from one group of society to another (Amin, 1976, p. 18); it influences the distribution of wealth and power within societies as well as the mechanism for the distribution of wealth and power among societies [p. 108]. ...

The distinguishing features of premodern and modern international relations are in large measure due to significant differences in characteristic social formations. The displacement of empires and imperial-command economies by nation-states and a world market as the principal forms of political id economic organization can be understood only as a development associated with the change from an agricultural formation to industrial formation [p. 110].

... the predominant form of political organization before the modern era was the empire ... the history of interstate relations was largely that of successive great empires. The pattern of international political change during the millennia of the premodern era has been described as an imperial cycle (Rader, 1971, pp. 38-68; Rostow, 1971, pp. 28-9). World politics was characterized by the rise and decline of powerful empires, each of which in turn unified and ordered its respective international system. The recurrent pattern in every civilization of which we have knowledge was for one state to unify the system under its imperial domination. This propensity toward universal empire was the principal feature of premodern politics.... The principal determinant of this cycle of empires was the underlying agriculture-based social formation... the size of the economic surplus from agriculture and imperial tribute was principally a function of the extent of territorial control. Therefore, other things being equal, the greater the territorial extent of an empire and of its political control, the greater the taxable surplus and the greater the power of the empire....

Although the generation of an economic surplus during the imperial era was dependent on agriculture, its distribution was frequently influenced by commerce and international trade ... the control of trade routes has been an objective of states and a source of great wealth and power. The great and enduring empires frequently have arisen at the crossroads of trade, and struggles over control of the principal arteries of commerce have been constant sources of interstate conflict. Changes in the control of these trade routes and changes in the locations of the routes themselves have played decisive roles in the rise and decline of empires and civilizations....

The cycle of empires was broken in the modern world by three significant interrelated developments: the triumph of the nation-state as the principal actor in international relations; the advent of sustained economic growth based on modern science and technology; the emergence of a world market economy. These developments reinforced one another and in turn led to displacement of the cycle of empires by the European balance-of-power system and, later, a succession of hegemonies in the nineteenth and twentieth centuries [pp. 110-16].

Since it is my firm contention, following Ekholm and Friedman chapter 2 above and Silver (1985) that "capital" existed in the ancient economy in much the same form as later in world history, and that capital accumulation is the driving force of world-historical development (Gills and Frank, Chapter 3 above), it follows that the history of capital accumulation and e history of hegemonic-power transitions are inextricably linked not only r the modern world but throughout most of world history. Chase-Dunn (1989) already argued that "both political-military power and the propriation of surplus value through production and sale on the world key play an integrated role" in hegemonic-power cycles. But how far k in world history could this be said to hold true? Perhaps much farther k. than we are normally led to think is the case. Chase-Dunn explains *** the "low overhead strategy" of Venice, which was later emulated by Holland, Britain, and the USA, relies for its success on a decentralized political apparatus of domination which reduces the cost of administration pot empire, while surplus extraction is accomplished by trade. By contrast, high-overhead imperia which rely on a direct and centralized political apparatus of control and extraction of surplus via coercion/tribute are less ^: successful when in competitive relations with the low-overhead types.

We have been taught that ancient economy and empire were basically; about coercion, bureaucratic centralization, and tribute. I believe there is |much evidence to the contrary. Gilpin's statement suggests that the regional power dynamic of empire cannot readily be separated from the trade dynamic transcending regional territorial boun. Many important and the long-lasting hegemonic or imperial powers in world history depended not * only on the agricultural surplus or on direct extraction of the same, but crucially upon exchange of products via market relations conducted over long distances. That is, they were embedded in a world economy and their power position was interrelated with their economic position within it. States, even ancient ones, pursue wealth through the pursuit of sources of surplus, of which trade has always been a key, if not decisive, element.

Even the earliest cities ** Sumer prospered via long-distance trade, though they engaged in imperial rivalry and expansionism in order to protect or expand their vital trade. The example of the Minoan thalossoc-racy comes to mind as another ancient example of a centre of accumulation prospering not primarily through military imperium or territorial expansion but through long-distance trade. Even more so, the early trade cities of the Levant and later the Phoenician cities provide an example of ancient capital accumulation on the "Venetian" model that was very successful for many centuries. Many of the principal classical Greek cities also rose to economic prominence in a similar manner, and eventually in competition with the Phoenicians, both of the Levant and of Carthage. The Byzantine empire's strength probably persisted for so long due to the important role the metropole of Constantinople played in the world economy, because of the strength of its gold currency and its pivoal geopolitical location.

The lift could easily be expanded.

Perhaps it would be more correct to hypodtMBie that tribute has never been the sole, or the most effective or competitive, means of accumulation, but rather that it has always been trade and commerce which constitute the most significant means of accumulation. This significance lies in the key role that transfer of surplus via trade has in determining change in the hierarchy of accumulation and power, and also in stimulating social change (Denemark and Thomas 1988; Denemark 1990). Therefore, our both crude and incorrect inherited notions of ancient "command-economy" (Gilpin 1981: 112) and of "tributary" modes of production (Amin 1976, 1989), determined by coercion or political means of extraction (Anderson 1974), require significant reformulation.

Gilpin is perfectly right to argue that the rise and decline of empires, great states, hegemons, etc., is a function of the generation and dissipation of economic surplus. He is also right to imply that this is a principle which applies to all world history and not only to the modem world. The quotation above from Gilpin illustrates, if only implicitly, that both the "domestic" character of surplus extraction and the "international" arena of competition over control of the flows of trade are probably of equal historical and analytical importance when attempting to understand the patterns of hegemonic transition. Likewise, he is right to say that changes in the form the surplus takes and the method of its accumulation significantly influence the form of political power that dominates a historical period. Indeed, I would elevate this to a cardinal principle in the study of world history and world order.

If capital accumulation existed via trade even in the ancient world economy and this was a key element in continuous hegemonic rivalry processes, then Gilpin's sharp break between the premodem and modern forms of hegemonic power may not be quite so sharp after all. "Trading empires" are not rare in history, as we have seen, even in the premodern era. "Command economies" are not rare in modern history, indeed the twentieth century seems to have been a period of remarkable (temporary) revival of such economic systems - in direct competition with the trading nation-states which Gilpin identifies as the dominant modem form of hegemonic power. As in the past, the trading state of the twentieth century proved itself to be a superior form in competition with the command economy. Therefore, the putative historical break between the "cycle of empires" and the "succession of hegemonies" may not be quite as clear as Gilpin suggests.

The reason for the success of both premodem and modern trading states seems straightforward. Participation in world trade is participation in world accumulation. This participation greatly increases access to surplus being exchanged and thus offers the opportunity to capture more surplus than would be possible based on a purely self-reliant national economy. This has always been true. Likewise, participation in world trade is an avenue to acquiring technology and production techniques also not necessarily available to a closed, self-reliant economic system. It is likewise a stimulus to achieve superiorities in the production system which allow a state's exports to be competitive in other markets, including other core markets.

Even the so-called imperial command economies of the premodern era, and particularly their elite classes, on closer scrutiny, were most often simultaneously engaged in pursuit of wealth through trade. This was certainly true of the most ancient Mesopotamian empires beginning with the Akkadian, and of Assyria, Persia, Rome, Byzantium, the Arab caliphate and subsequent Islamic empires, Parthia, Sassanid Persia, Bactria, the Kushan empire, Tang and Song China, the Ottoman empire ... the list could go on. Therefore, the "propensity toward universal empire" should not be explained solely on the basis of the desire to expand territorially in pursuit of more tribute and tax revenue from the agricultural base. It can also be explained in many cases by a desire to control key trade routes, the source of key materials, and key cities which generate "liquid" revenue in monetary form.

So if there is an important premodern-to-modern historical break it may not be so much due to the existence or nonexistence of trade as a key element of the pursuit of power, but rather to a change in the character of that trade. Chase-Dunn argues that "the thing which distinguishes a capitalist world-economy from earlier world systems is the exent to which states in the core rely on comparative advantage in production for the world market instead of political-military power" (Chase-Dunn 1989: 111) Nevertheless, this statement would be difficult to defend even for the relatively "modern" mercantilist states of the seventeenth and eighteenth centuries, which systematically deployed naval military power on a global scale to secure their share of surplus from world trade.

Change in the character of trade seems to reside first and foremost in the production system. The modern industrial change in the production system - made possible by advances in science and technology - led to a change in the form of surplus, or at least to a drastic change in the proportions being produced via agriculture and industry, which in turn led to a change in the forms of accumulation, increasingly via commodity exchange in price-setting markets and the wage-labor form at the point of production, and then in the form of state power, and thus to a change in the form of hegemonic power and world order. We have become very accustomed to referring to this modern historical period as "capitalism," or the "capitalist world-system," i.e. characterizing the historical period by a term for its dominant mode of production.

Though Gilpin and Chase-Dunn are right to highlight the importance of a switch from agricultural surplus and territorial expansion to modem industry, the spread of market relations, commodification, and wage labor, and the appearance of the modern nation-state, it is important to note that these modern forms never entirely displaced other coexisting forms of accumulation. Following Chase-Dunn (1989), it should be accepted that nonmarket variables are important to modem "capitalist" accumulation processes. "Capital" can be defined as a social relationship in which labor transfers a surplus to appropriating classes. These appropriating (accumulating) classes are composed both of owners of means of production and of political elites. The, the "capitalist class" can be either a private or a state elite which organizes production in order to appropriate surplus from labor. "States are part of the relations of production in capitalism" and "there are many degrees and forms of the commodification of labour" so that "the subjection of labour to the logic of profit-making... is accomplished by a variety of institutional means." "Real capitalism," and the accumulation of capital in it, includes both private and state "capitalists'" accumulating via the world market, and a mix of "competitive production of commodities and political-military power." The larger arena of "capitalism" allows various forms of commodified labor, not only the wage-labor form, and includes "geopolitics," i.e. the competitive quest for accumulation and military-political hegemonic power among states. Interestingly, "peripheral capitalism does bear a greater similarity to precapitalist societies based on the tributary mode of production than does core capitalsm" (Chase-Dunn 1989: 1-43, 121).

In my reading of world history, despite many apparent changes in modes of production, or modes of accumulation, the fundamental patterning of hegemonic transitions, i.e. consolidation and deconsolidation of hegemonic power, and the concomitant concentration and deconcentration of accumulation, seems to persist and indeed to transcend change in the mixture of modes. Let us briefly consider how modes of accumulation interface with hegemonic transition. Just as hegemonic power is better understood as a multilayered hierarchy rather than a unipolar dominance, so also modes of production or accumulation are better understood as a multilayered hierarchy, i.e. always being a complex articulation of modes.

I follow Geoffrey de Ste Croix (1981) and Ekholm and Friedman (chapter 2 above) in recognizing that all the primary forms of extraction of surplus known to the modern world were already in existence even in the ancient world. As discussed in chapter 3 above, in order to retain the notion of discrete modes of production, such as the slave mode of production, de Ste Croix developed an interesting formulation. He decided to characterize a mode as that through which the ruling elite derives the main pan of its surplus. This obviates the need for one mode to be overwhelmingly common in the social formation. It is only important that it be the form through which the elite derives the main source of its wealth from other classes. That is, it characterizes the key form of the transfer of surplus; the main form of the accumulation process in that socioeconomic formation at that period of history. It can and does coexist with many other modes. Taking this Respective even further, perhaps to its logical conclusion, I would argue that the notion of transition between discrete modes of production breaks down altogether. If the reality is always a mixture of many modes in a complex articulation then what actually takes place is a change in the composition of this mixture and the hierarchy within it, not a clean transition from one mode to another. The crucial change is at the top layer of accumulation. This form changes in correspondence with a host of social, political, military, technological, demographic, and other factors. But perhaps it is not only a matter of the historical form the surplus takes, but crucially, change in the distribution of surplus between fractions of the accumulating classes whch constitutes the ultimate key to understanding what drives hegemonic transitions.

If change in the configuration of modes is only one element, change in the configuration of power among the accumulating classes is another very important and too neglected dimension of historical transitions. Despite many changes in modes the fundamental patterning of hegemonic transitions seems to persist and remain a profound influence on the course of social history. The current confusion over whether the world is witnessing a transition from "socialism" to "capitalism" or may yet experience a transition from "capitalism" to "socialism" illustrates my point that perhaps what is happening in the world today would be better understood primarily as a hegemonic transition rather than primarily or solely a transition between modes of production. The same lesson applies to the earlier "transition" between "feudalism" and "capitalism." Chase-Dunn (1989) and Gills and Frank (chapter 5) criticize Wallerstein for regarding India and the Ottoman empire as separate world systems in the sixteenth century because they were allegedly not "capitalist" while the Eurocentered world economy was "capitalist." Wallerstein has substituted mode-of-production criteria for his material-exchange criteria, but by doing so has to reinterpret the very important and expensive trade relations among Europe, India, the Ottoin.ans, and also China. Far better to recognize that "Europe was never (or only briefly) a separate world-system according to the definition of material exchange networks. Rather, there has existed for at least two millennia a multi-centric Eurasian world-system" (Chase-Dunn 1989: 45).

As the present world situation clearly illustrates, a hegemonic transition is largely set in motion by shifts in the relative position of classes and states in the hierarchy of accumulation, but has profound effects on social development and is therefore by no means merely a rearrangement of players on a chessboard. Therefore, by focusing on hegemonic transition as the key concept of change in world history, we need not abandon the problems that modes-of-production analysis sought to address - namely how struggles over accumulation affect the course of larger sociopolitical and economic development in world history. The two are inseparable. "Modes" certainly exist, but they need to be (plocated in our scheme of social change, within a framework governed by patterns of interelite rivalry and the accompanying hegemonic transition.

Two of the fundamental axes of change in the world historical process of restructuring are the spatial and the temporal. Both time and space are very real dimensions of historical change, as important as material exchange and institution building. In the discussion to follow I explore the interface of these two axes in hegemonic transition; synchronization, conjuncture, and center shift are concepts I will employ in this exploration of the effects of space and time on hegemonic transition (Gills 1989a). Though from the perspective of the world there is only one unified sociohistorlcal time, from the point of view of any locality in the world developmental process there are separate or distinctive streams of sociohisiorical time. The interaction of these distinctive sociohistorical rimes is one of the key elements of historical change.

The interaction of different local sociohistorical times (synchronization) produces specific combinations in a moment of world time (conjuncture), which in turn may also result in the spatial rearrangement of the world system (center shift) which will be an expression of the hierarchy of centers of accumulation. This process constitutes a hegemonic transition.

Of course it is not time and space themselves which interact, but rather real social formations and states. These social formations and states each have their own respective cycles of accumulation and cycles of hegemony. These local/regional cycles are in turn themselves constitutive of the one world system cycle (Gills and Frank, chapter 5). Following Gilpin, the consolidation and deconsolidarion, or rather rise and decline, of hegemonies can be understood as being part and parcel of a parallel pattern of the concentration and deconcentrarion of accumulation, both within and between states. What I seek to do is to build up a framework of how all the cycles interact. That is, how each of the local/regional patterns of change affect each other and how each of these local/regional cycles is affected by and in turn affects the patterns of the world system as a whole.

According to my own rof world history, every regional empire seems to have experienced a cyclical pattern of concentration and deconcentrarion of accumulation. They all seem to have alternating periods of more centralized accumulation and periods of more decentralized accumulation. Every region in world history also seems to have experienced some form of cyclical pattern of political power, alternating between periods of more centralized power and periods of more decentralized power.

If we accept that cycles of accumulation and cycles of hegemony are very common, possibly universal phenomena, then the key question is what historical forces propel and perpetuate these cycles?

The answer is class struggles. Historically, class struggles or class conflict, both between the elite and the exploited, and between elites, are essentially a struggle over political means ***** apportioned to the classes. The ruling/propertied classes tend to want to increase the level of surplus extraction, since this surplus is the tangible wealth they enjoy. The exploited classes tend to want to decrease the level of surplus extraction, since this is their primary means of safeguarding their standard of living (to use the modem term). In the modern capital-labor relationship capital tends to seek a higher rate of surplus transfer from labor to capital. This is done by pressure to keep wages low and to lengthen the working day. Labor, particularly in its organized trade-union form, tends to want to decrease the transfer of surplus value to capital, by increasing wages and shortening the working day. Capital can and does circumvent the "natural" limits of surplus extraction inherent in labor's reproduction costs by substituting technology for labor and thus increasing productivity. While this is not the rime or place to venture further down this well-worn path, the main point is that we should not assume that the essence of the modern capital-labor relation is quite so unique as some would have it, though the logic of development it sets in train has certain important new historical features. Capital has always sought to increase the transfer of surplus from labor to capital, whatever the method of extraction.

In short, class struggles are as old as "civilization" itself, and are essentially about the same thing. It is impossible to argue that class struggles have ever come to an end. That would indeed be the "end of history." The utopianism of early Marxist analysis lies precisely in its view that class struggles would someday end. However much this ideal may appeal to us, it is a risky notion to carry over into analysis of world history as it has so far actually happened or can be realistically expected to happen in the future. This idea of the end of class struggle was an expression of nineteenth-century notions of progress and is Implicit in the model of transition between modes of production. By contrast, a model of hegemonic transitions does not assume any necessary end of class struggles, and of their concomitant economic and political cycles, not does it assume progress in a linear or historically necessary direction.

Goldstone (1991) has recently argued that cycles of social rebellion (which one could argue are very intimately related to cycles of accumulation and cycles of hegemony) are essentially demographically driven, at least for the past few centuries. Demographic change may play an important role in cycles, but in my view it would be a mistake to explain these cycles on the grounds of demographic factors. Goldstone, of course, does not explicitly focus on cycles of accumulation and hegemony.

But there does seem to be a general historical correlation between concentration of accumulation and social rebellion. When extraction of surplus is excessive, and when it is heavily concentrated in a few hands, these conditions seem to lead time Kid time again, in many different cultures, states, and empires, to a reactive rebellion, and also to possible disintegration, war, invasion, or collapse. Indeed, Geoffrey de Ste Croix (1981) explains the "fall" of the Roman empire in the western provinces on the basis of the social effects of overconcentration of accumulation in the hands of a small oligarchic elite and the resultant overextraction of surplus by both private propertied classes and the state. Similar patterns can be found in many other historical cases.

In preindustrial socioeconomic formations this cyclical development was focused on a struggle between private vs. state elites for control over the mainly but not exclusively agrarian-derived surplus. That is, it is primarily interelite conflict or competition which governs the pattern. The struggle over control of the surplus was not primarily between the elites and the exploited masses of the population. It has to be recognized that most of the time in most places the "masses" are so subordinated to the elite-constructed institutions of social life that they are more objects than subjects of history. Even when the masses are set in motion, i.e. temporarily become self-determining subjects of history, usually via rebellion, they very seldom gain real "progress" in the end. This does not mean they should not struggle. They should and do. It means simply that the outcome of class struggle "from below" is not often what is expected. Class struggle "from above," by contrast, seems to be the real meat of hegemonic rivalry and does seem to have very profound effects on sociohistorical forms, though not very "progressive" effects from the point of view of the exploited.

Returning to Gilpin, it is true that for much of history a particular dynamic characterized the rise and demise of most empires. The centralized state contended with the landed "aristocracy," or landowning classes, for shares of the available agricultural surplus. When the demands of this struggle grew so excessive as to undermine the basis of material life for the exploited masses, they rebelled. It is thus at a particular historical point that the masses enter history as agents of change. This is usually a response to overextraction and underinvestment, often accompanied by "privatization" or "aristocratization" of the state and a resultant overconcentration of power and wealth in the hands of a few.

Private accumulation posed a constant centrifugal, disintegrative force vis-a-vis centralized state accumulation (Gills 1987: 268). Eisenstadt (1963) analyzes the dynamic of bureaucratic agrarian empires according to a model of levels of structural differentiation and the development of "free (flexible) resources," i.e. the availability of social surplus that is not constricted by the "fixed commitments of ascriptive kinship and status groups, but which could be allocated directly." The state establishes direct relations of surplus appropriation with producers and both requires and facilitates the expansion of commodified labor, upon which it becomes dependent. Thus the state's economic activities, even in the ancient world, promoted the extension of the commodity nexus and deepening of the division of labor. One example of this is the minting of coinage to pay armies and to employ artisans on contract.

Eisenstadt recognizes that "it was the combination of external and internal pressures that constituted the major foci of change in the empires." He gives five determinants of structural change in bureaucratic agrarian empires: 1) the continuous needs of the rulers for different types of resources and especially their great dependence on various flexible (free) resources; 2) the rulers' attempts to maintain their own positions of control, in terms of both traditional legitimation and of effective political control over the more flexible forces in the society; 3) the great and continuous sensitivity of the internal structure of these societies to various external pressures and to political and economic developments in the international field; 4) the consequent need of the rulers to intensify the mobilization of various resources in order to deal with problems arising out of changes in military, diplomatic, and economic international situations; and 5) the development of various autonomous orientations and goals among the major strata and their respective demands on the rulers (Eisenstadt 1963).

According to Eisenstadt "strong contradictions" between these different determinant factors developed especially when the state elite "emphasized expensive goals which exhausted the available economic and manpower resources." He identifies a pattern in which overexploitation of the resource base by the state elite depletes the flexible resources necessary to the state's existence. This overexploitation exacerbates a countenendency for resources to revert to subordination to "more conservative, aristocratic-patrimonial (or feudal) elements." This process is associated with the "aristocratization" of the -state, usually via the bureaucracy. When this occurs the state elite increasingly indulges in "parasitic exploitation," i.e. the growth of consumption by the elite relative to redistribution and investment of the surplus. This type of parasitic appropriation of surplus is most marked during periods of decline. The process creates "new ascriptive positions and groups" which eventually mount a challenge which undermines "their already overdeveloped bureaucracies." As the state succumbs to aristocratization, in cases where the center is relatively weak and where international exigencies are strong, there is a tendency for such systems to undergo final disintegration and reversion to "simpler, patrimonial, or at most feudal, units" (Eisenstadt 1963). Therefore, "feudalism" is not a mode of production that (all) societies experience at a certain stage of linear historical development, but rather a phase in a hegemonic cycle experienced .across much of historical time. The history of virtually every region and every empire or civilization seems to bear this cmt.

Following Eisenstadt, I suggest there is a pattern in the premodem era of the centralization and decentralization of accumulation. In the centralization phase of empires the state elite is dominant and subordinate private and landed elites to state ends. The power these classes derived from the concentration of wealth in private hands is reduced via state action and reform. The result is an increase in state revenues, since the state is receiving an increased share of the available surplus vis-a-vis these (rival) private elite classes. The strengthened central state is free to develop the bureaucracy, invest in infrastructure, and promote imperial expansion. Through measures to integrate the economy and standardize administration the productive base of the economy is stimulated toward expansion, and more "free" flexible resources are created by the state. Thus there is usually a period of economic and concomitant demographic growth following the success of a centralization phase. Such growth is temporary, however, and is eventually overtaken by inertia as the costs of maintaining the system grow and the private elite erodes the dominance of the state elite and increase its share of surplus appropriation.

In the resulting decentralization phase, the private elite increase its share of surplus at the expense of the state and of producers. Its surplus appropriation is more parasitic and thus the ratio of consumption grows as that of redistribution and investment declines. This process undermines the stability of both the state and the economic base of society. The state enters fiscal crisis and in order to maintain control may increase the level of state surplus appropriation through both taxation and expropriation. But this increase in state appropriation often falls upon a narrowing tax base and thus exposes it to greater exploitation which damages the base, immiserates producers, and stifles economic growth, thus eventually stimulating social rebellion. Many producers flee the jurisdiction of the state and seek a haven under the auspices of local private elites, thus transferring the surplus to them rather than to the state. In this manner the sovereignty of the state begins to be "feudalized" or parcelized. At the most acnite stage, the state may collapse altogether, leading to a period of anarchic local rivalries.

Michael Mann has posited the alternation of two historical variants of "power configuration" as evidence of a dialectic of change in world history. These two power variants are 1) "empires of domination," and 2) "multi-power-actor civilizations." The empires of domination are characterized by "combined military concentrated coercion with an attempt at state territorial centralization and geopolitical hegemony." The multi-power-actor civilizations are characterized by "decentralized power actors... [which] compete[d] with one another within an overall framework of normative regulation." Mann suggests there is a dialectical relationship between the two variants of power configuration. Empires of domination may be the culminating development of preceding multi-power-actor civilizations. Mann also makes allowance for contiguous and contemporaneous interactions between the two variants of power configuration. That is, region A may be in variant 1 and region B in variant 2 in the same period or world history for example, ***** coexistence of the multi-power-actor civilization of the classical Greek and Phoenician city states with the "near Eastern Empire," by which I presume Mann means the Persian empire. Martin Wight similarly regarded the interlinkage of Hellas and Persia as two variants of international systems coexisting in the same overarching "secondary states system" (Wight 1977). Likewise, Wight, similarly to Gilpin, suggests that most multi-power-actor civilizations, or "states-systems" have ended in a "universal empire." In fact, Wight held that most states systems experienced a succession of hegemonies, in which one power after another tried to transform the states system, or even to abolish it, by reducing it to unity (Wight 1977).

Mann's formulation of a dialectic between the two variants of power configuration can be usefully enhanced by reference to Eisenstadt's model above. According to Mann, the dialectic rests on each variant's internal capacity for innovation. Mann suggests that one type gave way to the other when further social development was possible only "when its polar opposite type arose to exploit precisely what it could not." One has to be cautious with this type of formulation. I would prefer to suggest that the outcome of transition between types is a consequence of structural forces in historical motion and largely outside the conscious control of actors themselves. For Mann the internal dynamic of multi-power-actor civilizations "seems to have led toward its opposite, greater hegemonic centralization," and vice versa. Mann suggests this dialectic is the "core of world-historical development." The empires of domination, according to Mann, "unintentionally generated more diffuse power relations of two main sorts within their own interstices: (1) decentralized property-owning landlord, merchants and artisans; and (2) ideological movements." Mann's hypothesis is that "If these diffuse power relations continue to grow interstitially, a decentralized multi-power-actor civilization may result, either from the collapse of the empire or from its gradual metamorphosis" (Mann 1986).

In modern industrialized socioeconomic formations, the cycles are focused on a competitive struggle for the realisation of surplus value in the commodity circuit among separate bodies of capital within and between "national economies", mediated on the one hand by elaborate mechanisms of state intervention and regulation and on the other hand by the use of force in the arena of inter-state competition for the apportionment of the surplus value available for capture in the world market. (Gills 1987: 268)

Wallerstein maintains that in the modern world-system a state's strength correlates with the economic role of the owner-producers of that state in the world economy ... the modern history of the state... [is] one long quest to create structures sufficently strong to defend the interests of one set of owner-producers in tworld-economy against other sets of owner-producers as well as, of course, against workers. (Wallerstein 1980: 113-14)

So what is really changing is the configuration of relative power, measured in shares of the capture of surplus in the world market, between sets of owner-producers, backed by their respective state elites, which in turn affects the relative power positions of these state elites one to another.

However, unlike the traditional empires, direct competition between state elites and private elites in the modem "capitalist" state is not the main problem. Rather, it is intensification of the competition between private elites, and thus possibly also of their associated states, in the world market that drives hegemonic cycles in the modern era. WaUerstein assumes that capitalist states use both mercantilist and military techniques to assist their owner-producers to compete, but modern states do so most effectively if they can keep the costs of such assistance low enough not to "eat up the profits" (public finance). Each capitalist state's economic activity is more competitive if its political rule reflects "a balance of interests among owner-producers, such that a 'hegemonic bloc' (to use a Gramscian expression) forms the stable underpinnings of such a state" (Wallerstein 1980: 113-14). Much work on the political economy of neomercantilist or "capitalist developmental" states in east Asia emphasizes the importance of a symbiotic relationship between the state elite (including both political parties and the bureaucracy) and the private owner-producers with an aim of achieving competitiveness in the world market (Johnson 1982; Nester 1990).

In the premodern period of world history hegemonic transition on the regional level was primary. Regions alternated between Mann's two power configurations according to their own internal dialectic, though nevertheless affected by external factors as well. Synchronization of cycles between regions was an important feature of international historical change. If the cycles of accumulation and hegemonic power in one region were parallel to those in another, this rhythm obviously had an effect on their mutual interaction. They "rose and fell" together. Therefore their economic expansion and infrastructural investment phases could be mutually reinforcing. If these states came into conflict during their expansion phase the outcome might not be decisive, since both would be in possession of considerable staying power. In their declining phase neither might be capable of taking much advantage of the situation of the other, since both are weakened. The beneficiaries of such a synchronized crisis might be other external factors, from the periphery or hinterland, or another rival empire in a centralization phase. So-called "dark ages" result from a general crisis in which the beneficiary is more backward hinterland peoples and the result of the destruction and disintegration of empires is a degree of economic retrogression.

If, on the other hand, the cycles in one region were not parallel to another contiguous or contemporaneous regional empire, that rhythm would produce potentially very different systemic outcomes. For instance, in the case where region A is in a centralization phase and region B is in a decentralization phase, region A has considerable advantages. This pattern can be identified throughout world history and even as late as the later half of the nineteenth century, when Japan was in a centralization phase and the Qing empire in China was in a decentralization phase. This particular nineteenth-century historical conjuncture brought about a center shift in east Asia toward modern industrializing imperial Japan, which in turn affected the global hegemonic configuration. One of the easiest paths to hegemonic power is a conjuncture in which the major rivals are already weakened by either their own internal dynamic or a general crisis or war. Such a conjuncture was essentially the context for the post-1945 emergence of the Pax Americana.

In the modern era synchronization and conjuncture continue to be important elements of the spatial/temporal interface of historical change. However, the dynamic of alternation between variants of power configuration does not operate in exactly the same way as in the premodern era. The shift in the axis of class conflict is the key to this difference. In the premodern era the main form of class conflict was between state and private elites internally, and between state elites externally. In the modern era, the main form of class conflict internally is between producers and appropriating classes (both state and private) and between private owner-producers of different states, while internally the state and private elites are in a very close alliance. This internal alliance is directed at the producing classes and at rival blocs of private-state elites. However, in some periods of modern history global hegemonic coalitions are built that reduce conflict between rival blocs of private-state elites. This form seems to alternate with periods of more intense direct competition between rival blocs of private-state elites over shares of surplus in the world market.

To conclude, a shift to "hegemonic transition" as the central organizing concept of world history suggests some general hypotheses on the causes of periodic crises. Above all, the primary theme in these crises is always the overall character of the struggle for control of the surplus. The key factors in the onset of crises include the following an excessive rate of the extraction of surplus (overextraction); an excessive concentration of control over capital (overconcentration); a failure of demand or expansion to stimulate growth (underconsumption); and a failure of investment in productive capacity (underinvestment) which may be reflected in the growth ot elite consumption relative to social redistribution and productive investment -but is always "parasitic appropriation" as opposed to productive investment of capital. The above are often accompanied by a fiscal crisis of the state and a crisis of political authority. The end result is economic contraction and political fragmentation or dislocation, often accompanied by social rebellion and war.

A general world system crisis occurs when the combined cumulative layers of contradictions in the world system cannot any longer be sustained by the existing social, economic, and political arrangehients of the world order - thus necessitating drastic transformation. The interaction of all the cycles in the system and the simultaneous occurrence of crises in such conjunctural moments generate a high disequilibrium within the world system, which destabilizes the whole. This disequilibrium is present in both the sphere of world trade and the sphere of political-military power. A shift in both the locus of accumulation and of hegemonic power in the world system is the result. This process is accompanied by economic, social, and political upheavals and usually by wars. A dramatic world hegemonic transition is thus both the result and the resolution of a general world crisis. It is a resolution in the sense that the old world order is eventually destroyed and conditions are laid down for the emergence of a new world order. Out of a crisis of accumulation comes the restored conditions for an expansion of accumulation and a new hegemonic order.

Most importantly, these concepts seem as relevant to today's global economic crisis and on-going world hegemonic reorganization as they are to 5,000 years of world system history. This again renders the supposed sharp historical break around 1500 ad less meaningful than some would have it.

REFERENCES

Abu-Lughod, Janet (1989) Before European Hegemony: The World System a.d. 1250-1350, New York: Oxford University Press. Amin, Samir (1976) Unequal Development: An Essay on the Social Formations of Peripheral Capitalism, New York: Monthly Review Press.

覧 (1989) Eurocentrism, London: Zed Press.

ymderson, Perry (1974) Lineages of the Absolutist State, London: New Left Books.

Braudel, Femand (1982) Civilization and Capitalism: 1th-18th Century, New York: Fontana Press.

Chase-Dunn, Christopher (1989) Global Formation: Structures of the World-Economy, Oxford: Basil Blackwell.

Cox, Robert W. (1981) "Social forces, states and world orders: beyond international relations theory," Millennium: Journal of International Studies 10 (2): 126-55.

覧 (1983) "Gramsci, hegemony and international relations: an essay in method," Millennium: Journal of International Studies 12 (2): 162-75.

覧 **********; Press.

Denemark, Robert A. (1990) "Theories of trade as a political variable," paper presented at the International Studies Association annual meetings, Washington, DC.

Denemark, Robert and Thomas, Kenneth (1988) "The Brenner-Wallerstein debate," International Studies Quarterly 32: 47-65.

de Ste Croix, G.E.M. (1981) The Class Struggle in the Ancient Greek World, London: Duckworth.

Doyle, Michael (1986) Empires, Ithaca, NY: Cornell University Press.

Eisenstadt, S.N. (1963) The Political Systems of Empires, Glencoe, IL: The Free Press.

Ekholm, Kasja and Friedman, Jonathan (1982) "Capital" imperialism and exploitation in ancient world-systems," Review 4 (1) (summer): 87-109.

Frank, Andre Gunder (1978) World Accumulation 1492-1789, New York: Monthly Review Press; London: Macmillan.

Gill, Stephen (1990) American Hegemony and the Trilateral Commission, Cambridge: Cambridge University Press.

覧 (1991) "Historical materialism, Gramsci, and international political economy," in The New International Political Economy, ed. Craig N. Murphy and Roger Tooze, Boulder: Lynne Rienner, 51-75.

Gills, B.K. (1987) "Historical materialism and international relations theory," Millennium: Journal of International Studies 16 (2) (summer): 265-72.

覧 (1989a) "Synchronisation, conjuncture, and centre-shift in east Asian international history", paper presented at the International Studies Association annual meetings, London (April).

覧 (1989b) "International relations theory and the processes of world history:

three approaches," in The Study of International Relations: The State of the Art, ed. Hugh C. Dyer and Leon Mangasarian, London: Macmillan, 103-54.

覧 (1993) "The hegemonic transition in east Asia: a historical perspective," in Gramsci and International Relations, ed. Stephen Gill, Cambridge: Cambridge University Press, 186-212.

Gills, Barry and Palan, Ronen, "Introduction" in Palan, Ronen and Gills, Barry (eds) (forthcoming) Transcending the State/Global Divide: The Neo-Structuralist Agenda in International Relations, Boulder: Lynne Rienner.

Gilpin, Robet (1981) War and Change in World Politics, Cambridge: Cambridge University Press.

Goldstone, Jack A. (1991) Revolutions and Rebellions in the Early Modem World, Berkeley: University of California Press.

Higgott, Richard (1991) "Toward a. nonhegemonic IPE: an antipodean perspective," in The New International Political Economy, ed. Craig N. Murphy and Roger Tooze, Boulder; Lynne Rienner, 97-128.

Johnson, Chalmers (1982) MITI and the Japanese Miracle: The Growth of Industrial Policy, 1925-1975, Stanford: Stanford University Press.

Kennedy, Paul (1987) The Rise and Fall of the Great Powers, New York: Random House.

Keohane, Robert 0. (198^) "The theory of hegemonic stability and changes in international economic regimes, 1967-1977," in Change in the International System, ed. Ole Holsti et al; Boulder: Westview Press, 131-62.

覧 (1984) After Hegemony: Cooperation and Discord in the World Political Economy, Princeton; Princeton University Press. Mandel, Ernest (1980) Late Capita^m, London: Verso.

Mann, Michael (1986) The Sources of Social Power, vol. 1: A History of Power from the Beginning to ad 1760, Cambridge: Cambridge University Press.

Modelski, George (1987) Long Cycles in World Politics, London: Macmillan.

Modelski, George and Thompson, William R. (1988) Seapower in Global Politics 1494-1993, Basingstoke: Macmillan. Nester. William R. (1990) Japan's Growing Power over East Asia and the World Economy, London: Macmillan.

Rader, Trout (1971) The Economics of Feudalism, New York: Gordon & Breach. Rostow, W.W. (1971) Politics and the Stages of Growth, Cambridge: Cambridge University Press.

Silver, Morris (1985) Economic Structures of the Ancient Near East, London: Croom Helm.

Toynbee, Arnold (1946) A Study of History (Somervell abridgement), Oxford: Oxford University Press.

Van der Fiji, Kees (1984) The Making of an Atlantic Ruling Class, London: Verso. Wallerstein, Immanuel (1974) The Modem World-System, vol. 1: Capitalist Agriculture and the Origins of the European World-Economy m the Sixteenth Century, New York: Academic Press.

覧 (1980) The Modem World-System, vol. 2: Mercantilism and the Consolidation of the World-Economy, 1600-1750, New York: Academic Press.

覧 (1988) The Modem World-System, vol. 3: The Second Era of Great Expansion of the Capitalist World Economy, 1730-1840, New York: Academic Press.

Wight, Martin (1977) Systems of States, ed. Hedley Bull, London: Leicester University Press in association with the London School of Economics.

覧 (1978) Power Politics, New York: Holmes & Meier.

Back to Contents