Back to Contents

Part IV The world system: 500 years or 5,000? Discussing the theoretical, historical, and political issues


David Wilkinson

Interested in wars and their causes, I found that to study them with any hope of success it was important to study the whole system that generated them. This in turn led me to the study of "civilization" and civilizations, which seemed the best name for such macrosocial systems; but this study in turn compelled the study of their subsystems - e.g. cores; of their nonmilitary aspects - e.g. world economies; and of their social "containers" - e.g. oikumenes. And at that point my work intersected the work of Christopher Chase-Dunn, Barry Gills, Andre Gunder Frank, and others, who had arrived at the same subject from their different provenances. The purpose of this piece, then, is to provide a set of definitions and theses which summarize the way I now organize my work on civilizations, cores, world-economies, and oikumenes, so as to present it at once as a counterpart and a contribution to their undertaking.


1 In the time and space-bounded "ocean" of human sociocultural phenomena there exists a kind of vast social entity, a collection of interacting cities, a civilization, which functions in varying degrees as a real unity or "atomism," and as a field. A "civilization" is not a "culture," a "state," or a "nation." Ordinarily the boundaries of this social entity transcend the geographical boundaries of national, state, economic, linguistic, cultural, or religious groups.'

2 Due to the interdependence of the whole civilization as a system/field and its parts, these vast civilizational social networks tangibly condition most of the surface ripplings of the sociocultural ocean, including the historical events and life-processes of smaller sociocultural systems and the actions of individuals and groups living in a given civilization.

3 Without an adequate knowledge of the civilization we can hardly understand the structural and dynamic properties of its important parts of all its subsystems, regions, and components - just as without a sufficient knowledge of a primate troop as a whole, of its gross structure and. gross functioning, we cannot understand the nature and behavior of its member individuals.

4 Screening a list of some 70 candidates yielded a list of 14 entities which appeared to be societies at a civilized level (criterion: cities; further evidence: record-keeping, economic surplus, nonproducing classes, etc.) which were also connected world systems - militarily closed, geotechnologi-cally isolated social-transactional networks with an autonomous political history during which they did not take or need not have taken much account of the possibility of conquest invasion, attack (or alliance and cooperation) from any outsiders, although the members of each such system did recurrently conquer, invade, attack, ally with, command, rule, legislate, cooperate with, and conflict significantly and effectively with, and only with, one another. own 1961 revision of his earlier list, mainly by combining members of the prior rosters. The current civilizations list is different from Toynbee's and Quigley's, but still more different from Spengler's (1926-8) or Danilevsky's (1920).

6 Vis-a-vis Toynbee: of Toynbee's revised 1961 list I recognize Aegean under that name, Egyptic as "Egyptian," Middle American as "Mexican," Andean as "Peruvian," Sumero-Akkadian as "Mesopotamian;" combine Indus and Indie as a single "Indie;" combine Sinic and the Toynbeean "satellites" of Sinic - Korean, Vietnamian, and Tibetan - as "Far Eastern;" promote the Toynbeean satellites Mississippian, North Andean (as "Chib-chan"), Japanese, and South-East Asian (as "Indonesian") and a combination of Toynbee's "abortive" Far Western Christian and Scandinavian (as "Irish") to full civilizational status. Of Toynbee's full civilizations, five are not on my list: Syriac, Hellenic, Orthodox Christian, Western, and Islamic are regions or phases of a single continuing civilized society which I call (see Table 7.1 and Figure 7.1) "Central civilization." The same treatment is meted out to some of Toynbee's satellite civilizations -"Elamite," Hittite, "Urartian," Iranian, "Italic," and Russian - and to several of his abortive civilizations - Nestorian, Monophysite, and the Medieval Western City-State Cosmos. Toynbee's abortive First Syriac civilization I have treated as a shared semiperihery of Mesopotamian and Egyptian civilizations.

7 Vis-a-vis Quigley: my list includes Egyptian, Mesopotamian, and Japanese civilizations, and contains reasonable matches to Quigley's Cretan, Mesoamerican, and Andean. I do not accept the separateness of Quigley's "Indie" and "Hindu" civilizations, nor of his "Sinic" and "Chinese." And Quigley's Hittite, Canaanite, Classical, Islamic, Orthodox, and Western civilizations seem to me to constitute cultural regions and epochs within the polyculture of a larger civilization, that which I have called Central civilization.

8 Vis-a-vis Spengler: while my list mentions Egyptian and Mexican civilizations, and contains reasonable matches to Spengler's Babylonian, Indian, and Chinese cases, I do not recognize the separateness of his Classical/Apollinian, Arabian/Magian, Western/Faustian or (suppressed) Russian civilizations; these are, rather, conflicting cultures within a single civilization, namely Central civilization.

9 Vis-a-vis Danilevsky: while my list contains Egyptian, Mexican, and Peruvian entities, and reasonable matches to Danilevsky's Ancient Semitic, Chinese, and Hindu-Indian, I do not recognize the separateness of Iranian, Hebrew, Greek, Roman, Arabian, European, or Slavic civilizations, all of which are (to me) conflicting cultures within the polycultural compost of a single larger society: Central civilization.

10 My differences with the four lists cited reflect my application of a social criterion, while Danilevsky and Spengler employed cultural criteria and Toynbee and Quigley used mixed sociocultural criteria. The similarities between lists reflect this coincidence: where e.g. Spengler or Danilevsky found cultural coherence in Egypt, Mesopotamia, China, Mexico, Peru, and India, I found a period of geosocial isolation and historical autonomy.


11 The various civilizations are not necessarily based upon any major premise, nor do they necessarily articulate, develop, and realize such, nor are they necessarily logically or aesthetically consistent or complementary. Each civilization is a causal system; it may or may not be a "meaningful" one, or evolve toward or away from "meaningfulness." Since civilizations are not assumed to be "meaningful" unities, they need not possess any major premise, prime symbol, ultimate principle, or fundamental value that is articulated by their cultural phenomena. But they might in fact do so.

12 Do they in fact do so? I would guess that they do not, but, rather, that each will be found to articulate a different evolution of a different dialectic, i.e. a different struggle among a different set of conflicting premises, symbols, etc. Artists, philosophers, charismatics, and prophets within civilizations frequently seek or seem to create or discover meanings, premises, prime symbols, ultimate values, and Utopian reorderings in and for their civilizations. Instead they ordinarily create dialectical controversies.

13 Since we need not assume that the cultural field of any civilization is completely unified, nor that it is meaningfully consistent, the question of whether, when, and how cultural unity, consistency, or interaction exist becomes hypothetical, to be explored empirically rather than by definition or axiom.

14 In such exploration, I would begin with the guess that over many generations the culture of any civilization will tend toward greater second-order integration - mutual agreement on what its areas of discord are -with continuing first-order inconsistency (continued discord). Its causal 'unification will likely be dialectical, organized as a continuing struggle of changing opposit(though withoutany final synthesis ever terminating the dialectic).

15 Indeed, the likelihood that we will find all civilizations actually highly and evolvingly contradictory, conflicted, dialectical, is so strong that we might reasonably study civilizations on the assumption that each, far from being an organic cultural unity, is in fact "a cultural field where a multitude of vast and small cultural systems and congeries - partly mutually harmonious, partly neutral, partly contradictory - coexist" (Sorokin 1950: 213).

16 On this assumption, one would research a civilization's cultural individuality precisely by identifying, not a prime symbol, major premise, fundamental value or ultimate principle, but the collection of such symbols, premises etc., that coexisted, conflicted, and coevolved within it, their mutual relations of dominance and displacement, challenge and response, fusion and fission.

17 Systematic cross-civilizational cultural research would explore such questions as: is there usually or always a dominant core culture in a civilization? How long does such dominance persist? How is it displaced and by what? When civilizations collide, how is the evolution of cultural dominance affected? Does second-order integration emerge, and at what time scales?

18 I would not want to assume that civilizations necessarily contain a dominant cultural system - the question of dominance is once again properly empirical - but would regard it as an empirical fact that most civilizations, most of the time, indeed contain dominant cultural cores, which have geographic locations and are frequently "dominant" in more ways than one: i.e. militarily, technologically, economically, and demographic-ally, as well as culturally.

19 The most striking effect of the new definition on accustomed lists of civilizations, as has been shown above, is that such familiar entities as classical/Hellenic/Graeco-Roman civilization, Hittite civilization, Arabian/ Magian/ Syriac/ Iranic/ Islamic civilization(s), Orthodox Christian civilization, Russian civilization, and even our own familiar Western civilization, must be reclassified either as episodes/of or as regions within a previously unrecognized social-network entity, by my definition both a civilized society and a world system, hence a single civilization. This civilization I have labeled Central civilization.2

20 Thus today there exists on the earth only one civilization, a single global civilization. As recently as the nineteenth century several independent civilizations still existed (i.e. those centered on China, Japan, and the West); now there remains but one. Central civilization.

21 The single global civilization is the lineal descendant of, or rather I should say the current manifestation of, a civilization that emerged about 1500 bc in the Near East when Egyptian and Mesopotamian civilizations collided and fused. This new fusional entity has since then expanded over the entire planet and absorbed, on unequal terms, all other previously) independent civilizations.

22 Central civilization was created in the Middle East during the second millennium bc by an atypical encounter between two pre-existing civilizations. Civilizations may coexist, collide, break apart or fuse; when they have fused, they have typically done so by an asymmetric, inegalitarian engulfment of one by the other. But the linking of the previously separate Egyptian and Mesopotamian civilizations through Syria was an atypical, relatively symmetric, and egalitarian "coupling" which created a new joint ***** network-entity rather than annexing one network as a part of the other entrained to its process time.

23 The Central city-network, in unbroken existence and process since its inception, has been atypical in another way: it has expanded, slowly by the reckoning of national and state turnover times, but quite rapidly by comparison to other civilizations, and in that expansion has engulfed all the other civilizational networks with which it once coexisted and later


24 "Central" is a historical and positional nomenclature which deliberately avoids any specific geographic or cultural references, thereby indicating that this society is not to be characterized by references to a single river basin, and that its development has not been determined by that of a single culture, nation or people. It would be too parochial to label that civilization by the nomenclatures of any of the nations that have successively populated, of the states that have successively dominated, or of the regions that have successively centered it. At this moment, in this place, and in this culture, it seems not mistaken, and not too parochial, to call it Central civilization.

25 Central civilization is of course positionally "central" only in retrospect, by reason of its omnidirectional expansion: this network, originally Afro-Asiatic in being located where Asia and Africa meet, spread over time in all directions, encompassing the civilized networks of Europe, west Africa, and the Americas by moving west and those of south and east Asia by moving east, and thereby rendering itself historically "Central" as well.

26 The subsumption of a variety of putative civilizations under the single rubric of Central civilization is illustrated by Figure 7.1, which shows two such candidates, "Graeco-Roman" and "Western," as epochs of regional dominance within Central civilization; these dominant regions in fact constituted long-lived, but impermanent, cores of Central civilization. The Near Eastern, medieval, and global phases of Central civilization also possessed cores; but they were larger and less culturally homogeneous than the Graeco-Roman and Western cores. Another way of comprehending the subsumption is that what has in the past appeared to be the end of one civilization/world system and the beginning of another is easier to comprehend as a core shift within a single continuing civilization - a shift of military, political, economic and cultural domination from one region to another.

27 The taxonomic principles that yield Central civilization as a recognizable entity are three in number. First: any two "civilizations" that were always adjacent and vigorously politico-militarily interacting were ipso facto parts of a single civilization. In the medieval period of the northwest Old World (i.e. Europe, south-west Asia, north Africa) there were Western cities. Orthodox cities, Muslim cities; there was no Western civilization, no Orthodox, no Islamic civilization. There were civilized peoples and territories in the north-west Old World; they were members of a single civilization.

28 Second: any two historically autonomous civilizations which become adjacent and vigorously, continuously politicomilitarily interactive (through expansions or shifts) thereby become a single civilization. Either a new entity emerges (if they unite on relatively equal terms); or one of the old civilizations absorbs the other.

29 Third: any two alleged "civilizations" adjacent in time are but periods in one single civilization unless the earlier civilization's cities are entirely depopulated and abandoned (like those of Mississippian civilization). Unless a civilization's urban centers vanish, it does not fall. It may terminate by fission into two separate and more or less equal autonomous entities which cease to interact dynamically; it may terminate in fusion with some other civilization. Without fission, fusion, or fall there is no end to the civilization's system and process. If there is no end, there can be no succession.

30 Inasmuch as Central civilization combines from 4 to 14 of the civilizations discerned by more pluralist civilizationists using cultural criteria, it is to be expected that, and it is in fact the case that, Central civilization is not a language group, or a religious group, or a state group. Yet it is bonded, bonded oppositionally: for continuing warfare is a social bond, and continuing hostility is a cultural bond. Central civilization is a strongly bonded entity, even though it be a cultural potpourri. Central civilization is a conglomeration ofsociocultural phenomena, adjacent inspace and time, that is integrated by causal ties - including collision, warfare, and coevolution - and by quasi-meaningful ties of mutual consciousness, awareness of differences, and hostility.

31 Our time is unique in that only one civilization now exists on earth, of global Scope, without a periphery into which to expand further. Central civilization seems never yet to have been a "meaningful" but always a "causal" unity; but now that it has reached the limits of its oikumene, after having absorbed the whole human species and all other civilizations, there is a good chance that it will in the future evolve toward a recognizable "meaningful" unity.

32 Central civilization does however have a presently dominant culture within its polycultural mix. The dominant culture is what Sorokin labeled "Sensate" - and also theoretic, secular, Promethean, scientific, technological. I would additionally label it cosmopolitan, bourgeois, capitalist, liberal, democratic, and above all "modern." Sensate modernity's culture continues to expand against resistance while simultaneously generating internal schisms and coopting and incorporating external resistances in a manner which maintains both its variety and its dynamism. Being only dominant but not yet all-pervasive, Sensate culture, whose dynamic expansion is called "modernization," has not yet reached its full attainable social limits, and (consequently?) continues to expand savagely against savage resistance, as it has done within Central civilization for the past seven or eight centuries, even while Central civilization itself has been expanding to global dimension.

33 Sensate culture may well be dominant in the now global cultural field, but it is a near thing: there are enormous masses yet being culture-colonized against active or passive resistance. Such masses are to be found, e.g., among Africans and the Indians of the Americas; among non-Protestant Christians, and nonmainstream Protestants; among Muslims, Hindus, Buddhists; among still Marxist-ruled but non-Marxist populations (Chinese especially); among tribal peoples, peasants, and genuine proletarians (Servian, not Marxian - i.e. the urban homeless).

34 That Sensate culture has not expanded to its conceivable demographic and social limit, and that it continues to recruit and expand toward those limits, does not mean that it will get there. There are also signs both of Sensate disintegration and of the beginnings of many coumertrends. But the latter are neither integrated nor expansive, and their resistance thus far seems more like diehard reaction than like the genesis of a new counterculture.


35 An ideal-type civilization/world system/macrosociety, because its characteristics are unequally distributed over space; and because they are distributed centrically; and because their unequal distributions overlap; and because the inequalities are connected intrinsically to its past history of expansion (for civilizations tend strongly to expand. Central civilization being an extreme rather than an exceptional case), characteristically pos-

(a) a core (central, older, advanced, wealthy, powerful);

(b) a semlpenphery strongly connected to the core (younger, fringeward, remote, more recently attached, weaker, poorer, more backward); and

(c) a weakly connected periphery (nomads; peasant subsistence producers not yet attached to a city; and other civilizations that trade but do not habitually fight or ally with the subject civilization).3

36 Civilizations usually begin in a geographically restricted area composed of cities and the hinterlands their fighters can aspire to control; this is surrounded by an area to which the new cities are politically irrelevant. We may call these zones the (initial) urban core, controlled or disputed semiperiphery, and uncontrolled periphery of the civilization.

37 Civilizations usually expand over time by raiding, invading, and conquering adjacent areas, by sending out colony-cities and military settlements and trading forts, by fascinating and addicting previously indifferent peripheral people to their products (gods, drugs, laws, weapons, music, ornaments, etc.). The territories affected by this civic expansion -whether the expansion be colonialist, imperialist, cultic, developmental -may be considered to have been incorporated by the civilization when their occupants - settlers or settlees - undergo urbanization and begin to interact politically on a regular basis - as subjects, allies, tributaries, enemies - with the civilizational core. This area of later expansion and control is the (enlarged) semiperiphery of the civilization.

38 Once a semiperiphery exists, and it comes to exist quickly, it also persists. Thus one of the main continuing patterns that reveals itself in the history of civilizations and world systems is that they-tend - not by definition, but empirically - to be markedly geographically tripartite. In the core, military force, political power, economic wealth, technological progress, cultural prestige, and theogony are concentrated. The periphery is far from the core in all senses, containing peoples and territories known but scarcely noted. The semiperiphery, more or less recently penetrated or engulfed, is a zone characterized by military subjection, powerlessness, relative poverty, technological backwardness, low cultural prestige.

39 But while the tripartition of a civilization is very durable, no area has permanent tenure in any role, and tenure of coredom is rather precarious. Cores are not eternal; civilizations can outlast their original cores. A history of cores must therefore be kinematic, describing their rises, shifts, and falls; a theory of cores must ultimately be dynamic, accounting for their motion and change.

40 A civilization's core may have any of several political forms. It may be a single state, as in Mesopotamian civilization, perhaps, during the rise and fall of Assyria, or as in Central civilization during the rise of Media and Persia and Rome, and the era of Justinian. The core may be the metropolitan region of a universal state, as in Central civilization during the Assyrian, Persian-Macedonian, and Roman empires. Or the civilization core may be a functionally divided set of areas in a universal state, as in Far Eastern civilization during the Qin-Former Han and in Japanese civilization during the Kamakura period. The core may contain several states, successively hegemonic: in Mesopotamian civilization, the Sumerian core c. 2500-2360 bc (Ur, Lagash, Umma). It may constitute several states simultaneously balanced, as in Central civilization between its universal empires, and for most of the time since the Roman empire's fracturing. The most frequent core forms are: the single dominant or hegemonic state; several competing states; and the universal-empire metropole.

41 Cores pulsate. Core areas enlarge and contract. Central civilization's core shifts - westward in the Graeco-Roman phase, eastward in the medieval phase, westward again in the Western phase - involved expansion at one edge synchronic with contraction at the other; the global phase saw core expansion east and west. Contractions are naturally enough associated with hegemonic struggles and universal-state periods, expansions with all-core epochs; but not perfectly.

42 Cores shift. Cores may move in a single general direction, or oscillate. The Central core half moved west, then east, then drifted west and north and west again. No signficant patterns are evident.

43 Cores decline and rise. Does past experience as a core preclude or assure return to core status? Apparently neither. Setting aside all the apparent civilizational-startup first-time cores (e.g. Central civilization's Penile Crescent + Nile valley), there are many cases in which a semiperipheral area, never before a core, rose to core status. In Central civilization, such first-time core entrants included Assyria, Persia, Greece (previously an Aegean core), Macedonia, Rome, Byzantium, western Europe, America, Russia. But there are several other cases in which a fallen core area hasreturned from semiperipheral status, or has regaia solitude it had lost to upstart sharers. In Central civilization: Abbasid Mesopotamia, and the classic case of Renaissance Italy. In the transition from semiperiphery to core, history seems somewhat more favorable to renaissance than to renaissance, but renaissances do happen.

44 Different areas may serve as military-political, economic, and cultural-religious cores, and core shifts may occur in these features at different times. The most notable discrepancies between Central civilization's economic-technical and politicomilitary cores are attested by being corrected: the shift from Rome to Constantinople, the Renaissance-ending invasions of Italy, the revolt of the Netherlands, the involvement of British finance and fleets in Continental wars, and the American entry into the world wars of the twentieth century. There is thus some tendency for geographically separated functions to be pulled together: the politicomilitary core may conquer the others (the post-Renaissance invasions of Italy), migrate to them (by a movement of the capital, e.g. to Constantinople or Lo-yang), or usurp them (by taxation and subsidy, e.g. Tokugawa Edo);

or economic cores may invest in politicomilitary potency (Dutch, British, American).

45 Are semiperipheries necessary? Apparently not, since civilizations are often all-core, i.e. lack a semiperiphery. Central civilization has always had a significant semiperipheral area. Semiperipheries exist more often than not, particularly in universal-empire periods when the metropole is especially favored, but they do not seem necessary features of a civilization: power, wealth, creativity, can all be rather widely dispersed, though dispersal usually alternates with concentration.

46 Tenure in the semiperiphery is more secure than core tenure (cores decline) or peripheral tenure (peripheries are devoured). But there is some upward mobility. A semiperipheral area remains semiperipheral as long as it is politically annexed to, urbanologically subordinate to, militarily dominated by, culturally provincialized by, economically outaccumulated by, technologically outcompeted by, cultically devoted to, the old core. When and where the semiperiphery acquires states as influential, forces as dangerous, cities as populous and wealthy, culture as attractive, technique as progressive, gods as efficacious as those of the core, that part of the semiperiphery becomes core; the core area expands to encompass it. And if the old core should peak and decline, be overtaken and passed in its military and political, demographic and economic, cultural and technical and theological development by its semiperiphery (or a pan of it), so that the old core becomes a historic backwater, becomes marginal to the affairs of the civilization, while the former semiperiphery becomes the new core, we may properly say that the core of the civilization has shifted. And cores do shift: witness Karnak, witness Babylon.

47 There is some relationship between the transition of a state from semiperiphery to core and its later ability to impose hegemony and universal empire over the states system. Recent arrivals to core status have some advantages in competitions to destroy states systems, but they are not overwhelming, nor entirely self-evident.

48 A theory of peripheries must largely account for their secular decline. Civilization as such - the sum of the territories and peoples of the various civilizations - has expanded continually since its origins, despite some regional setbacks and a single holocivilizational collapse (that of Mississippian civilization), by conquering and colonizing and assimilating its non-civilized peripheral peoples and territories. This contradicts the idea that civilizations rise and fall, rise and fall: they almost never fall. It also contradicts the image of peaceful, sedentary civilized peoples always threatened and occasionally overwhelmed by neighboring barbarians: most of the "overwhelming" has been inflicted by the civilized societies on their peripheral neighbors. When noncivilized peripheral peoples - usually nomads - attack and conquer civilized territory, the result has ordinarily been that they settle down, take over, enjoy ruling the civilization, and continue expanding it; on the whole, peripheral peoples have not developed a sense of peripheral identity and pride sufficient to impel them to destroy the civilizations they have sporadically conquered. Civilizations, on the contrary, strongly tend to destroy their peripheries, through incorporation.

49 "Coreness" and "semipherality" are multidimensional phenomena, but certainly have politicomilitary, economic, technological, demographic, religious, and cultural components. Politicomilitary driving variables seem more obvious and accessible to analysis than others, but are unlikely to function alone. Forces need to be posited to explain both the motions and changes of cores - formations, expansions, pulsations, shuttles, drifts, evaporations - and core persistence and stability.

50 Interesting speculative questions about core and periphery include: can an all-core global society evolve? Would it require a states system? Does the end of the periphery increase the chances for an all-core society? - or a freezing of current core-semiperiphery boundaries? - or a speedup in core shift? - or a narrowing of the core to a single hegemonic state or imperial metropole?


51 I have elsewhere (Wilkinson 1987: 48-53) provided some impressions of the economic "facts" about Central civilization which comparative theory needs to accommodate and explain. The civilizationist most noted for attention to economic issues is Carroll Quigley (1961); the analyst of world-systems most so noted is Immanuel Wallerstein (e.g. 1974, 1975, 1979a, b, 1980, 1982, 1983, 1984). To what extent can Quigley's and Wallerstein's ideas be deployed for such an explanatory purpose?

52 Carroll Quigley's economically driven model of the evolution of a civilization is elegant, lucid, consistent, and tight. There are serious problems in its delimitation of the units of macrosocial analysis, and in its dependence upon a relatively homogeneous structure and process to explain fluctuations in relatively heterogeneous social systems. It is not at all clear that such systems have "stages" rather than "phases." Nevertheless Quigley's concept of an instrument of expansion is more generally useful than the alternative "mode of production," which suffers from the same defects while not directly addressing the crucial issue of the general phenomenon of macrosocial expansion. Similarly, Quigley's ideas about core and periphery relationships, and about expansion/stagnation cycles, are of great value in broadening later views of the same topics.

53 The world-systems school of Immanuel Wallerstein and his colleagues has produced a large body of provocative work with great internal complexity. It too delimits the units of macrosocial analysis in ways that seem to call for revision, though in different ways from Quigley's work. It would be Useful for world-systems analysts to consider Quigley's work as a potential contributor to their own.

54 For the study of Central civilization, Quigley and Wallerstein are resources despite the fact that Quigley would deny that such an entity ever existed, while Wallerstein would accept it only for the past two centuries or so. Nonetheless this entity displays core-periphery phenomena, and probably buffers "globally" the effects of "local" expansion/ stagnation cycles which its world wars probably also "locally" entrain. Even if one does not accept the tight policy-economy linkages implied in the Quigleyan civilizational and Wallersteinian world-systems schemata, one cannot come away from reading Quigley and Wallerstein without accepting that there must be some such linkages: if not quite that posited by either, then perhaps mixtures of their pure types, and perhaps softer, more delayed, sometimes even inverted versions of their harder couplings. No two writers seem better sources for hypotheconcerning the political economy of Central civilization.

Extracivilizational as well as intracivilizational trade characterized Central civilization's Egyptian and Mesopotamian predecessors, and Central civilization itself from its inception until its incorporation of the globe. Wallerstein's propositions about the "rich trades" help to account for the existence, distance, and relatively low impact of such external trade. Still, if it is highly rational to trade what is "worthless" for what is "precious," one must expect traders (and tribute-seekers, and predators) to flock toward preciosity. Such a tendency may help explain the marked inclination of civilizations to couple with or engulf onc* another, on the assumption that uneven distribution of resources and uneven development of technology tend, while civilizations are separated, to create what will be viewed as "preciosities" as soon as they begin to communicate. This proposition, and all those hereafter asserted for Central civilization, may well be true of other civilizations, and certainly should be treated as comparative hypotheses or heuristics.

56 There existed an Old World oikumene, an ecumenical macroeconomy, a multicivilizational structure which apparently provided the highest-level largest-scale economic order until the global reach of Central civilization, as the evolving context of the world economies of the various Eurasian civilizations linked by the silk, spice, slave, gold, and ivory trades. This economy was larger than any polity (universal empire or states system) it contained, encompassing Central civilization, Indie, Far Eastern, and others. It may require theoretical treatment as a whole; its theory is likely to be quite special, precisely because of the absence of a polity.

57 Local economies and short-range trade probably account for most economic activity most of the time, with the extraction of food from each city's hinterland and its distribution to the city population of primary importance.

58 World-economic commodities in Central civilization have tended strongly to be elite goods - luxury food, clothing, shelter, and display items - along with the trade tools of elite-supporting soldiers and bureaucrats (weapon-metal; paper for record-keeping). Elites, classes, and the associated inequalities must not be treated as recent phenomena.

59 Early Central trade in precious metals may, and coinage does, imply the development of mobile free persons, merchant classes, and economic (vs. politicomilitary) elites, characterized by private property in portable

wealth. These elements of capitalism similarly must not be treated as of recent vintage.

60 The entry into the Central world-economy of fish, wheat, oil, and wine suggests mass consumption driven either by political redistribution (to hire loyalty of armed men, clients, voters, etc.) or by markets, probably varyingly by both. Luxury goods may also have spread more widely through the social structure.

61 The general trend over time is clearly toward a continuing increase in the number and variety of commodities traded in the world economy of Central civilization. Within this trend there are temporary and permanent commodity dropouts, shifts in regional contributions, epochs of faster and slower commodity increase; but the trend remains. Commodities and commodification too precede modernity, and must be attributed to some early cause, perhaps simply to civilization's division of labor, increased scale, and increased population. Commodities and markets are not intrinsically related: granite appears to be a state commodity for Egyptian monument-builders; granite-hewers worked not (primarily) to build for their own tombs, but to satisfy the monumental egos of the state elite. The increase in the number and variety of commodities over time is one piece of evidence for a secular trend to expansion in Central civilization over the past 5,000 years.

62 There is as yet discernible no clear increase in the per capita wealth or living standards of the median individual during the premodern periods of Central civilization. It appears that increased production is mostly utilized to increase total population and total urban population. The aggregate wealth of the wealthiest strata (typically politically rather than economically defined) must have increased, but it is not clear that the per capita wealth of these strata also increased.

Modernization seems another story. But if Wallerstein is right regarding "absolute immiseration," it is an even less cheerful story. One wonders, for instance, if the forward days of contemporary world food reserves are more or fewer than in the first food-storing cities. At best, there is room for doubt, and for inquiry.

63 There is no clear evidence of an endogenously economic general crisis or collapse ever having occurred in the Central world economy, although there have been city-level and state-level disasters, and system wide periods of setback and stagnation, usually deriving from politicomilitary events. This has been argued elsewhere at some length (Wilkinson 1987: 39-48). A very long-term expansive trend appears to underlie various cycles of expansion and stagnation. If Quigley is right, this implies very frequent reforms and circumventions. If an economy is very mixed, with strong regional differentiation, regional failure by institutionalization may lead to the semiperipheralization of the failing region and the destruction of the failed institutions by intruders from another region of the same civilization - a combination of Wallerstein's core-shifting and Quigley's semiperipheral-success ideas.

64 The basic expansive process in Central civilization appears to be circularly causal, dependent upon the presence of an unpopulated or underpopulated geoeconomic periphery and a Malthusian pressure: population expands; more and larger and more dispersed cities with more populous hinterlands extend and intensify settlement; there is greater division of labor and specialization; sufficient demand arises to mobilize new products or longer routes to more distant sources; total production rises; increased production mainly serves to support an enlarged population; etc. While seas, seabeds, poles, deserts, mountains, forests, tundra, atmosphere, and space remain in many ways peripheries and frontiers of expansion, they are also barriers. Whether the ultimate bounds of human expansion are those of landmasses or of the universe is not clear. Can a civilization avoid taking out all its economic expansion extensively, by a corresponding population growth? Perhaps not.

65 The borders and cities of Central civilization expanded preferentially toward commodity sources, but not always quickly, effectively, or uniformly. Quigley and Wallerstein employ circumferential rather than radial images of expansion; otherwise their theories meld well with this observation. We may add that a preferred direction of expansion could well be precisely toward "rich trades." Otherwise, areas likely to be brought into the semi-periphery sooner would include likely population outlets and tribute sources.

66 Whatever may be true for state and local economies, it is not correct at any time to describe the world economy of Central civilization as fundamentally feudal, nor slave, nor hydraulic, nor free-peasant, nor communal, nor corporate, nor hierocratic; nor is it fundamentally, in the Wallersteinian sense, either a "world-economy" (capitalist) or a "world-empire" (tributary). In the fifth century bc, to take an apparently extreme example of variety, but a binding one, what was the Athenian economy? A slave economy (there was a very large slave population)? A peasant economy (most citizens were country-born and bred, and landowners, producing sheep, cattle, grapes, olives, grain)? A merchant capitalist economy (exporting wine and oil, providing coinage and a carrying trade)? An industrial economy (based on the silver, lead, zinc, and iron mines, importing grain)? With an industrial proletariat (slaves included skilled workers; free workers' wages hovered at subsistence)? A world-empire (Athens imposed trion other states)? A welfare state (much of the population was on the pupayroll via the mass-jury system)? A socialist state (massive expenditures on public works - harbors, fortifications, temples, naval expeditions)? Clearly something of all: a very mixed economy. And all this in a tiny fraction of the total area of Central civilization!

67 It is an interesting fact, and one worth reflecting on, not just a given, that Central civilization has never yet been completely penetrated by any particular "instrument of expansion" (in the Quigleyan sense) or "mode of production" (in the Marxian sense). A possible hypothesis is that there are a limited number of possible modes of production (Wallerstein); that all have inbuilt self-destructive propensities (Quigley); and that the only available choices at times of reform or circumvention are the items from the same old menu.

68 A possible reason why the world economy of Central civilization has never been fully statist is that the universal states of Central civilization have been either short-lived, with their extraction capabilities confined to the civilizational core, or tolerant of private property and merchant classes. Since the same could be said of universal empires in Indie, Far Eastern, Japanese, and Mexican civilizations, we might want to look at the Inca empire, also brief but apparently ultra-statist, to question its extremism, explain its divergence, and thereby explain the norm. Similar questions might be usefully put to statist national economies, e.g. the Soviet and Chinese, within states systems.

69 A possible reason why the world economy of Central civilization has never been fully capitalist (private-propertarian, individualist, marketive) is the unbroken prominence of the political state, based on force, and of political-military-religious elites based on ground rents, taxes, and extraction by force. Why can these elements apparently not be expunged? How far can they be suppressed, and kept suppressed? These are questions of interest at least to libertarians, and to those socialists who are in touch with the anarchist rather than statist tendencies of that movement. Wallerstein's idea of the marketer's mixed motives and the consequent need of capitalists for states is very much in point here.

70 For whatever reason, the Central economy is at all times a mixed, political economy, embodying trade and war, coercion and bargaining, the one-few-and-many. The balance shifts with time, scale, region, commodity. And possibly other variables. The determinants of the mix need study, The coexistence, with regional and temporal variations, is so marked as to suggest a theory of the mixed economy as historic norm, and the idea of capitalism and socialism as ideal-typical extremes needs developing.

71 The balance shifts more toward "capitalism" (without ever coming close) as states are small, weak, and numerous, more toward "statism" as they are few, strong and large.

72 One useful indicator of the statist/capitalist balance in the civilization might be the balance between cities of the same size that are state capitals (i.e. power-maintained) and that are commercial centers (i.e. trade-maintained).

73 The core/semiperiphery distinction is not that of a straightforward division of labor between political coercion and economic supply, nor between primary and higher-tech products; but both divisions are notably present. The element of time-delayed expansion over space, of institutional aging, of destructive core wars, of unequal "materialism" and exogenous technical development will also all doubtless prove factors in determining and shifting cores.

74 It is the politicomilitary predominance of the core, not any purely economic differentiation or "unequal exchange" tradition, that mainly accounts for the tendency for the core to drain the semiperiphery: loot, tribute, taxes, price controls, confiscations, trade route closures, and enforced monopolies are primarily political ventures. A significant fraction of primary products come from within the core, from the hinterlands of core cities. This becomes less true in the nineteenth century with the development of railroads; the British .policy of agricultural free trade may mar the shift. However, it remains true that...

76 Citification, and eventually core status, tends to move toward major semiperipheral supply sources.

77 Wherever it is possible to map the distribution of wealth in Central civilization, inequality prominently appears: by city, by region, by political power, by inheritance, in law, by age and family status, by gender. The several inequalities do not appear to be reducible to, any one fundamental root inequality.

78 There is abundant scope here for theory and observation, dialectic and eristic, in the contemplation of the world economies of Central and other civilizations. A world system will certainly have a world economy associated with it, and it is worthwhile trying to describe such an economy, and seek theoretical assimilation of the description. Terminology adequate to describe holistically the economic structure of a civilization does not yet exist. It cannot be produced simply by adapting and generalizing "macroeconomic" terminology suited to describe the economy of a state or the economic institutions of a culture, since a civilization is neither a state nor a culture. World economies do not appear to be characterized by sufficiently homogeneous class systems, property systems, production relations, divisions of labor, or instruments of expansion, to make holistic Marxian, Wallersteinian, or Quigleyan characterizations very revealing. There are coexisting and contradictory property types rather than a prevailing property type, coexisting and inconsistent class structures rather than a prevailing class structure, heterogeneous divisions of labor, and several competing instruments of expansion. We have clearly only just begun the theoretically salient description of Central civilization's world economy.

79 It is the case that there have been economic structures of larger scale than the political structures they contain. I would refer to the theoretical problem of establishing the interconnections between economy and polity when the two are not coextensive as the problem of the oikumene.4

80 An "oikumene" is here defined as a trading area, a domain internally knit by a network of trade routes, in which there is enough internal trade so that the whole trading area evolves to a significant degree as a system, while trade outside the area, though perhaps important both to the oikumene and to other oikumenes with which it trades, is not sufficiently dense and significant to cause system-level development to encompass these external systems.

81 Oikumenes may contain no civilization (pre-urban trade networks); may contain and be coextensive with one civilization (the present global economy); may contain but outlie one civilization; may contain and outlie more than one civilization.

82 In an empirical examination (Wilkinson 1992-3), it proved possible on the whole to correlate "civilizations" (politicomilitarily linked urban' networks) with "oikumenes" (economically linked urban networks) in which they nested. What similarities and differences exist in the nature and development of oikumenes, as trading areas, and civilizations, as systems of states and empires?

83 A world economy, lacking a coextensive world polity, but containing world polities of smaller area than its own, existed from (at least) the fourth millennium BC (when it linked the world polities of Egyptian and Mesopotamian civilizations) to the nineteenth century ad (when a world polity became global, and coexistensive with the world economy that had theretofore contained it). Other such "oikumenes," trade-linked but not politicomilkarily bonded, probably connected Chibchan with Peruvian civilization, and may have linked Mexican with Mississippian and/or Mexican with Peruvian civilization. But it is particularly noteworthy that Central civilization, from c. 1500 bc to c. 1900 ad, formed a politically coherent social system smaller than, nested wit, expanding in pace with and into the space pioneered by, an economically coherentbut politically unlinked oikumene. Because that oikumene seems to have been the globe's oldest "world economy," it is designated herein the Old oikumene. The Old oikumene is not only the eldest of the several members of its species (there have been Indie, Far Eastern, and Japanese oikumenes at least, in addition to those of the New World); in its expansion it, like Central civilization, engrossed all others, and today, grown to global scope and (for the first time) coextensive with a polity, is the sole survivor of its species.

84 Oikumenes contain civilizations, but not the reverse. Oikumenes organize larger areas more weakly. Why should this be? Perhaps because politicomilitary ties (rule, attack, threat, alliance) are more costly for actors to maintain than economic ties; or because they impose a net economic loss on the whole system that maintains them, while trade ties produce a net gain. Politics (or political economy) may be a negative-sum game, economics a positive-sum game. Western neoclassical economists would be happy to think so; redistributionists would not.

85 Oikumenes tend to expand. Despite occasional setbacks (reflected by losses of urban population), there have been underlying upward trends in numbers of megacities and in their sizes. Oikumenes tend to expand in area as well as in human and urban numbers: the Old oikumene expanded from the Middle East to global scope, in the process colliding with and absorbing the other oikumenes.

86 There is a parallelism between the tendency of oikumenes to expand, collide, and merge and that of civilizations to do the same. But there is also a major difference: namely, the apparent absence of the distinction between (he inegalitarian "engulfment" and egalitarian "coupling" relationships in oikumenical fusion. In particular, during the interval between the fusions of the Old oikumene with Indie and Far Eastern oikumenes, and the later fusion of Central civilization with Indie and Far Eastern civilizations, i.e. between about 326 bc and 1000-1600 ad in the Indie case, and between about 622 ad and 1900 ad in the Far Eastern case, it is hard,; to argue for any kind of extreme inequality in the transactions between the formerly separate oikumenes. Intense complaints and resistance seem to appear as a result not of economic penetration, but of politicomilitary penetration, not of oikumenical fusion but of civilizational fusion, in which politicomilitary predominance also alters the terms of economic redistribution in the direction of the penetrating powers.

87 Civilizations follow oikumenes, and "the flag follows trade," and not the reverse. There appears to be a powerful economic incentive, once trading areas have expanded beyond the politicomilitary reach of the powers in a civilization's political system, for those powers to extend the reach of their rule, violence, threat, and power-bargaining. No doubt there is a reciprocal incentive for traders and colonists to get outside civilizations' polities, and then to reach back for economic ties. Economy flees polity, which pursues.

88 Oikumenes do not allocate their benefits equally and impartially, except in the Malthusian sense that populations '"granted" a surplus tend to use it to become numerous and poor rather than few and rich (though elites within such populations seem to do the opposite). On the assumption that a notable growth (or shift) in megalopolitan population implies, and results from, a notable growth (or relative shift), of "wealth," the question of which world city was the largest when becomes of theoretical interest.

89 No clear system-level processes exist that give or remove primacy of wealth and population to or from chief cities of the civilizations in polycivilizational oikumenes; urban primacy at the oikumene-level appears to be mostly an epiphenomenon of asynchronous imperial unions and collapses at the civilization level.

90 It would seem consistent to expect that in a monocivilizational oikumene (like the current one), economic inequality is likelier to be the result of politicomilitary than of purely economic processes.

91 There are some apparent, though not in principle unresolvable, discrepancies between this argument and the recent and current findings of other workers, notably Barry K. Gills and Andre Gunder Frank (chapter 3 above; cf. Frank 1990: esp. 228-33). On the one hand, their argument that the world system developed from its origins in Mesopotamia, Egypt, and Indus, into the "Asio-Afro-European ecumene" and incorporated the Western hemisphere after ad 1500 (p. 81-2 above) is virtually identical to the interpretation I would put on Tertius Chandler's (1987) city data, though I prefer the term "Old oikumene" to both the "world system" and "the 'Afro-Eurasian ecumene.'"

92 Furthermore, I fully concur with their defense (p. 85-7 above) of Central Asia's very important and unduly neglected role in the development of "the world system" (for me, of Central civilization and of the Old oikumene).

93 On the other hand, I feel compelled to use a substantially later dating of the incorporation of several key areas into "an over-arching system of inter-penetrating and competitive super-accumulation" (p. 81-2 above) than is implied in their work, which brings the Indus zone into the "world system" by about 2700 bc (p. 81-2 above) and China apparently by 500 bc. To the extent that I am constrained by Chandler's data, I see the Indus as inside the Old oikumene in Chandler's "snapshot" for 1800 bc, but Indie civilization as thereafter outside the Old oikumene in 1200, 1000, 800, 650, and even 430 bc, and not back until 200 bc. To that same extent, I see Far Eastern civilization as outside the Old oikumene up to the 500 ad "snapshot" and inside it only in and after the 622 AD "snapshot."

94 The reasons for our differences are two, and the same in these two cases. One reason is approachable by theory, one by research. The theoretical reason is that I am unwilling to accept that the connection of two oikumenes has produced a single system until the trade routes that connect the two have been studded with entrepot cities whose population and polity are pretty clearly sustained by brokering (and guarding, warehousing, servicing, repackaging, rerouting, and parasitizing) the trade. Thus the rise of Rayy, Baikh, Broach, and Taxila are to me important and necessary indicators of the reincorporation of Indie civilization's private oikumene into the Old oikumene by 200 bc; the rise of Samarkand and Kashgar serves similarly as indicators of the incorporation of the Far Eastern oikumene into the Old oikumene by 622 ad.

95 The researchable reason might, however, reduce or even resolve our chronological disagreement without requiring changes in theory on either side. Chandler's 1987 data takes the threshold of city size down only to 30,000 in 430 bc, and to 40,000 in 500 ad. Were data to be collected down to the threshold of 10,000 which I prefer, it may be taken as certain that each of Chandler's tables of cities would be greatly expanded, perhaps in the case of some of the later tables expanded by one or two orders of magnitude. Inspection of the Chandler tables suggests very strongly that city sizes form a near-Zipfian distribution - the larger the fewer; the smaller the size the more cities at that size. In the process of such expansion, it is highly probable that many cities which, like Samarkand and Kashgar, crossed a 40,000 threshold by 622 ad, would have crossed a 10,000 threshold by 500 AD and not impossible that they did so much earlier, or that other cities on the same route crossed the lower threshold long before those crossed the higher. It is therefore quite conceivable that further research will fully resolve our chronological disagreements, with or without a resolution of our theoretical differences.

96 A second difference between the argument developed here and that of Frank and Gills has to do with the system-level phenomenology of my "Old oikumene" and their "world system." I have not located prior to the nineteencentury the phenomenon they characterize as "super-hegemony": a "privileged position... in whicone 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" (p. 103 above).

97 I prefer (to "superhegemony") the term "parahegemony," based on the multiple connotations of the prefix "para-": related to; almost; closely resembling the true form; abnormal; beyond. "Parahegemony" is a position in an oikumene in which the parahegemon derives economic benefits similar to those which a true hegemon is able to extract by the use or threat of force. But the parahegemon does so without the need to spend on force, because it has the economic advantage of being a highly privileged fore-reacher (a center of invention, and/or saving and investment, and/or entre-preneurship) and/or a rentier (monopolizing a scarce resource, a trade-route intersection or choke point, an enormous market, etc.); and because it has the politicomilitary advantage of being strong enough to defend its centers and monopolies, or of being outside the politicomilitary striking range of its rivals and/or victims.

98 The terminological difference is not crucial. "Parahegemony" could not unreasonably be called "superhegemony," even though it involves less relative power than "hegemony," because it may be more secure, less assailable, cheaper to maintain than genuine politicomilitary hegemony.

99 There have, I believe, been recognizable parahegemons on the world-systems. Britain, often mistakenly styled "hegemonic" in the nineteenth century, was a parahegemon - able to defend itself from anyone though not to conquer or control any of its great-power rivals; advantaged by being first or fastest in industrial development and then in finance.

100 So, after the Second World War, was the United States parahegemonic rather than hegemonic? The USA was incapable of compelling positive compliance by Russia (Stalin's violation of Yalta), China (failure of the Marshall mission; failure of the 1950 Acheson initiative), France (general intractability of General De Gaulle), India (defection from 1950 Korean war support coalition; foundation of nonaligned movement), even North Korea (1950-3) or North Vietnam (1954 ff.). It was, however, fully capable of defending itself, all its trade routes and major trading partners, and it possessed relative superiority in agricultural and industrial capacity and in innovative capacity and achievement. By contrast, the position of the USA in 1991 was far closer to hegemony than to parahegemony: it was better able to coerce, and less able to compete.

101 But were there pre-nineteenth-century parahegemons? I have not found their trace in the Chandler data. The historical traces of oikumenical parahegemony ought to include cosmopolitan accumulation of wealth; and, if we accept that a wealthy cosmopolis will contain a luxuriating patriciate and a proliferant and/or immigrative plebs, remarkable growth in population ought to be as usable a sign of parahegemony as would be the accumulation of palaces and temples, pleasances and theaters, monuments and brothels, warehouses and ministries, harems and hippodromes.

102 The largest city in an oikumene is, then, perhaps also the sign of the oikumenical parahegemony of the state within which it lies. But there are other possible explanations for cosmopolitan size. A city might be largest by reason of direct hegemony (not parahegemony) over the oikumene as a whole. Or it might be largest for reasons accidental to the oikumene but well-grounded for some region within the oikumene, e.g. because its state was locally hegemonic (or parahegemonic) to the most populous or wealthiest region within the oikumene.

103 In reviewing Chandler's list of "Cities that can have been the largest" (Wilkinson 1992-3: Table 30) most such seem to have their status plausibly explained on grounds that relate to their regional rather than their oikumenical role. Most commonly they rose in population as their state acquired hegemony, empire or universal empire, not within the whole of the Old oikumene but within a civilization that was a politicomilitarily linked region within the economically bound system of the oikumene, and they fell in size in proportion as the scope of the regional domination of their state shrank.

104 On the whole, therefore, the achievement of oikumenical parahegemony seems to be a relatively recent phenomenon. Why? The answer is no doubt partly to be found by closer examination of the rise of nineteenth-century London and twentieth-century New York; but also in the failure to reach parahegemony of earlier plausible candidates. These would be those cities that acquired large populations without acquiring empires large enough to account for those populations, and which accordingly probably prospered mainlymrough success in trade, but which never rose to demographic primacy: perhaps this list should include Kerma (Nubia), Hazor, Ugarit, Saba (Yemen), Hastinapura, Miletus, Broach, and Canton; surely it would include Tyre, Athens, Carthage, and Venice. If the experience of the latter quartet is characteristic, then the usual pattern of the failure on the road to parahegemony is dual: one becomes a target for the attacks of dominant powers on their way to hegemony or universal empire, and is thereby distracted from wealth-seeking to defense, or destroyed, or taken over and drained; and/or one turns from the road to economic parahegemony to the parallel but different road to politicomilitary hegemony, and finds oneself unfitted to be a hegemon by just those social characteristics that made one a fit candidate for parahegemony, e.g. (perhaps) an open, fluid, volatile, mercantile social order.

105 Whether the USA has acquired the attributes needed by a hegemon, and in the process lost those required of a parahegemon, is a question that might be raised in this connection; but not in this paper.

106 Since Gills and Frank do not as yet ascribe "superhegemony" to any particular pre-nineteenth-century state, it cannot be said that we are as yet in substantive disagreement. But I am now pessimistic about the likelihood that empirical research will in future locate such an entity, while I believe they remain rather more hopeful. To the extent that their "superhegemon" and my "parahegemon" mean the same thing theoretically - the overlap is not complete, but substantial - this difference of expectations is also resolvable by research rather than otherwise. I have tried in this essay to begin where my own independent research began, and to end by coming to grips with the challenging, stimulating, and important theses about the world system and superhegemony lately articulated by Gills and Frank. Certain issues divide us still, and these are always to me the more interesting. I regard the question of hegemony, parahegemony, and superhegemony as the premier issue for the next phase of the debate, which this collection opens.


1 This proposition, like most of those in the next three sections, is derived by way of a critical encounter with the work of Pitirim Sorokin. See Wilkinson forthcoming.

2 For an expanded treatment of this issue, see Wilkinson 1987.

3 These issues are treated at greater length in Wilkinson 1991.

4 This problem is investigated further in Wilkinson 1992-3.


Chandler, Tertius (1987) Four Thousand Years of Urban Growth: An Historical

Census, Lewiston/Queenston: St David's University Press. Danilevsky, Nikolai la. (1920) Russland und Europa, trans. Karl Notzel, Stuttgart: Deutsche Verlags-anstalt. Frank, Andre Gunder (1990) "A theoretical introduction to 5,000 years of world

system history," Review (Fernand Braudel Center) 13 (2) (spring): 155-248. Quigley, Carroll (1961) The Evolution of Civilizations, New York: Macmillan. Sorokin, Pitirim A. (1950) Social Philosophies in an Age of Crisis, Boston: Beacon.

覧 (1956) Fads and Foibles in Modern Sociology and Related Sciences, Chicago: Henry Regnery.

覧 (1964) Sociocultural Causality, Space and Time, New York: Russell & Russell.

覧 (1966) Sociological Theories of Today, New York: Harper & Row.

Spengler, Oswald (1926-8) The Decline of the , trans. Charles Francis Atkin-son, New York: Knopf.

Toynbee, Arnold J. (1961) Reconstderations. A Study of History, vol. 12, Oxford: Oxford University Press.

Wallerstein, Immanuel (1974) The Modern World-System, vol. 1: Capitalist Agriculture and the Origins of the European World-Economy in the Sixteenth Century, New York: Academic Press.

覧 (1975) "The present state of the debate on world inequality," in World Inequality: Origins and Perspectives on the World-System, Immanuel Wallerstein, Montreal: Black Rose Books, 9-28.

覧 (1979a) The Capitalist World-Economy, Cambridge: Cambridge University Press.

覧 (1979b) "Underdevelopment and phase-B: effect of the seventeenth-century stagnation on core and periphery of the European world-economy," in The World System of Capitalism: Past and Present, ed. Walter L. Goldfrank, Beverly Hills: Sage, 73-83.

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

(1982) "World-systems analysis: theoretical and interpretative issues," in World-Systems Analysis: Theory and Methodology, ed. Terence K. Hopkins, Immanuel Wallerstein, et al., Beverly Hills: Sage, 91-103. (1983) Historical Capitalism, London: Verso. (1984) The Politics of the World-Economy, Cambridge: Cambridge University

Press. Wilkinson, David (1985) "General war," Dialectics and Humanism 12 (3-4); 45-57.

覧 (1986) "Kinematics of world systems," Dialectics and Humanism 13 (I): 21-35.

覧 (1987) "Central civilization," Comparative Civilizations Review (fall): 31-59.

覧 (1988) "Universal empires: pathos and engineering," Comparative Civilizations Review (spring): 22-44.

覧 (1991) "Cores, peripheries and civilizations," in Core/Periphery Relations in Precapitalist Worlds, ed. Christopher Chase-Dunn and Thomas D. Hall, Boulder, CO: Westview Press, 113-66.

覧 (1992-3) "Cities, civilizations and oikumenes," Comparative Civilizations Review (fall 1992); 51-87 and (spring 1993): forthcoming.

(forthcoming) "Sorokin vs. Toynbee on congeries and civilizations: a critical reconstruction," in Sorokin and Civilization: A Centennial Assessment, ed. Michel Richard, Palmer Talbutt, and Joseph B. Ford, Rutgers, NJ: Transaction Books.



Samir Amin

The modern world has produced a general image of universal history founded on the proposition that (European) capitalism is the first social system to unify the world. The least that can be said in that respect is that this statement seriously distorts reality and - I submit - is basically an expression of the dominant Eurocentric ideology. In fact, societies prior to the sixteenth century were in no way isolated\from one another but were competitive partners within at least regional systems (and perhaps even a world system). Overlooking their interaction, on*can hardly understand the dynamics of their evolution.

Simultaneously I maintain that capitalism is a qualitatively new age in universal history which started around 1500. Therefore I insist upon distinguishing the modern capitalist overall structure from protocapitalist elements which indeed appeared in anterior societies, sometimes since quite ancient times; I also insist upon the specificity of the capitalist center/ periphery dichotomy vis-a-vis previous forms of polarization.


The theoretical contribution of the Marxist concept of the capitalist mode of production is crucial to this discussion. Its eventual dilution (fashionable nowadays of course) does not help clarify the issues. The capitalist mode of production entails private ownership of the means of production which are themselves the product of labor, namely, machinery. This in turn presumes a higher level of development of the forces of production (compared to the artisans and their instruments) and, on this basis, the division of society into two fundamental classes. Correspondingly, socially necessary labor takes the form of free wage labor. The generalized capitalist market thus constitutes the framework in which economic laws ("competition") operate as forces independent of subjective will. Economistic alienation and the dominance of economics are its expression.

No society prior to modern times was based on such principles. All advanced societies from 300 bc to 1500 ad are, from one end of the period to the other, of a profoundly similar nature, which I call tributary in order to show this essential qualitative fact; namely, that the surplus is directly tapped from peasant activity through some transparent devices associated with the organization of the power hierarchy (power is the source of wealth, while in capitalism the opposite is the rule). The reproduction of the system therefore requires the dominance of an ideology - a state religion - which renders opaque the power organization and legitimizes it (in contrast to the economistic ideology of capitalism which makes economic exploitation opaque and justifies it through this means, counterbalancing the relative openness of political relations, itself a condition for the emergence of modern democracy).

Having taken a stand on some of the debates of historical materialism, I believe it helpful to recall my essential conclusions. They affect my suggestions on the nature of the one (or more) premodern system(s). I have rejected the supposedly Marxist version of "five stages." More precisely I refuse: 1) to regard slavery as a necessary stage through which all the societies that are more "advanced" have passed; 2) to regard feudalism as the necessary stage succeeding slavery. I have also rejected the supposedly Marxist version of the "two roads." More precisely, I refuse to consider that only the "European" road (slavery-to-feudalism) would pave the way to the invention of capitalism, while the "Asiatic" road (the supposed Asiatic mode of production) would constitute an impasse, incapable of evolving by itself. I have described these two interpretations of historical materialism as products of Eurocentrism. I refer to my alternative suggestions in Class and Nation. I suggested the necessary succession of two "families of modes of production": the communal family and the tributary family. This suggestion comes from highlighting two qualitative breaks ** the general evolution: 1) later in date: the qualitative break from thej dominance of the political and ideological instance (state plus metaphysical' ideology) in the tributary phase into the dominance of the economic instance (generalized market and economistic ideology) in the capitalist phase; 2) previously: the qualitative break from the absence of a state and the dominance of the ideology in the communal phase into the crystallization of social power in the statist-ideological-metaphysical form in tributary phase. This proposition entailed identifying various forms of the two phases and, more particularly, defining the "central/periphe forms of the tributary phase, with precisely the description of feuda as a peripheral tributary form.

To some, the forms I call "tributary" would not constitute "a" me of production in the sense that they believe Marxism attaches to the coc



of the mode of production. I shall not indulge in this kind of Marxology. If it is a "nuisance" I am ready to replace the term "tributary mode of production" with the broader expression "tributary society." Of course my suggestions remain within/a framework dominated by the search for "general laws." Include in this, on the basis of these conceptualizations I have suggested, their "transition" toward capitalism, marked by the development of the "protocapitalist" elements which appeared earlier in history. There is of course a strong current nowadays rejecting any search for general laws and insisting on the "irreducible" specificity of various evolutionary paths. I take this epistemological orientation to be a product of a Eurocentrism concerned above alwith legitimatizing the "superiority" of the West.


The first question tdebate on this subject encounters concerns the character of worldwide capitalist expansion. For my pan, along with others (including A.G. Frank), I hold that the processes governing the system as a whole determine the framework in which local "adjustments" operate. In other words, this systemic approach makes the distinction between external factors and internal factors relative, since all the factors are internal at the level of the world-system. Is there any need to stress that this methodological approach is distinct from prevailing (bourgeois and even current Marxist) approaches? According to the latter, internal factors are decisive in the sense that the specificities of each ("developed" or "undeveloped") national formation are mainly due to "internal" factors, whether "favorable" or "unfavorable," to capitalist development.

My analysis remains broadly based on a qualitative distinction (decisive ', in my view) between the societies of capitalism, dominated by economics s (the law of value), and previous societies, dominated by the political and i ideological. There is, as I see it, a fundamental difference between the | contemporary (capitalist) world-system and all the preceding (regional and tributary) systems. This calls for comment on the "law of value" governing I capitalism.

On that ground I have expressed my point of view in terms of what I jhave called "the worldwide expansion of the capitalist law of value." Generally speaking, the law of value supposes an integrated market for the iroducts of social labor (that then become commodities), capital and labor. Cithin its area of operation it brings a tendency to uniformity in the price f identical commodities and returns on capital and labor (in the form if wages or returns to the petty commodity producer). This is a close pproximation to the empirical reality in central capitalist formations. But the scale of the world capitalist system, the worldwide law of value ates on the basis of a truncated market that integrates trade in goods and the movement of capital but excludes the labor force. The worldwide law of value tends to make the cost of commodities uniform but not the rewards for labor. The discrepancies in world pay rates are considerably broader than in productivities. It follows from this thesis that the polarizing effect of the worldwide law of value has nothing in common in terms of its quality, quantity, and planetary scope with the limited tendencies to polarization within the former (regional) tributary systems.

In this context the qualitative break represented by capitalism remains totally valid; it manifests itself in a fundamental reversal: the dominance of the economic replaces that of the political and ideological. That is why the world capitalist system is qualitatively different from all previous systems. The latter were of necessity regional, no matter how intensive the relations they were able to maintain among each other. Until the reversal has occurred it is impossible to speak of anything but protocapitalist elements, where they exist, subject to the prevailing tributary logic. That is why I am not convinced of the usefulness of a theoretical view that suppresses this qualitative break and sees a supposedly eternal "world system" in a continuum whose origin is lost in the distant past of history.

The significance of the qualitative break of capitalism cannot, therefore, be underestimated. But an acknowledgement of it reveals its limited historical application, as it is stripped of the sacred vestments in which bourgeois ideology has dressed it. The simple and reassuring equations can no longer be written, such as capitalism (nowadays "market") equals freedom and democracy, etc. For my part, along with Karl Polanyi, I give a central place to the Marxist theory of economic alienation. With Polanyi, I draw the conclusion that capitalism is by its nature synonymous not with freedom, but with oppression. The socialist ideal of bringing freedom from alienation is thus reinvested with all the force of which some sought to deprive it.

The critique of Eurocentrism in no way implies refusal to recognize the qualitative break capitalism represents and, to use a word no longer fashionable, the progress (albeit relative and historically limited progress) it ushers in. Nor does it propose an "act of contrition" by which westerners renounce describing this invention as European. The critique is of another kind and centered on the contradictions the capitalist era opens up. The system conquers the world but does not make it homogenous. Quite the reverse, it effects the most phenomenal polarization possible. If the requirement of universalism the system ushers in is renounced, the system cannot be superseded. To sum up in a phrase the critique I suggested in Eurocentrism: the truncated universalism of capitalist economism, necessarily Eurocentric, must be replaced by the authentic universalism of a necessary and possible socialism. In other words, the critique of Eurocentrism must not be backward-looking, making "a virtue of the difference," as the saying goes.


The world-system is not reducible to the relatively recent form of capitalism dating back only to the final third of the nineteenth century, with the onset of "imperialism" (in the sense that Lenin attached to this term) and ths- accompanying colonial division of the world. On the contrary, we say that this world dimension of capitalism found expression right from the outset and remained a constant of the system through the successive phases of its development. The recognition that the essential elements of capitalism crystallized in Europe during the Renaissance suggests 1492 - the beginning of the conquest of America - as the date of the simultaneous birth of both capitalism and the world capitalist system, the two phenomena being inseparable.

How should we qualify the nature of the "transition" from 1500 to 1800? Various qualifications have been suggested, based on the political norms prevailing at the time (ancient time or "the age of absolute monarchy") or on the character of its economy (mercantilism). Indeed, the old mercantilist societies of Europe and the Atlantic and their extension toward central and eastern Europe are problematic. Let us simply note that these societies witnessed the conjunction of certain key preliminary elements of the crystallization of the capitalist mode of production. These key elements are a marked extension of the field of commodity exchanges affecting a high proportion of agricultural production; an affirmation of modern forms of private ownership and the protection of these forms by the law; a marked extension of free wage labor (in agriculture and craftsmanship). However, the economy of these societies was more mercantile (dominated by "trade" and "exchange") than capitalist by virtue of the fact that the development of the forces of production had not yet imposed the "factory" as the principal form of production.

As this is a fairly obvious ** of a "transitional" form, I shall make two further comments on this "conclusion." First, the elements in question - that some have called "protocapitalist" (and why not?) - did not miraculously and suddenly emerge in 1492. They can be found long before in the "region," in the Mediterranean precinct particularly, in the Italian cities, and across the sea in the Arab-Muslim world. They had also existed for a very long time in other regions: in India, China, etc. Why then begin the "transition to capitalism" in 1492 and not in 1350, or in 900, or even earlier? Why speak of "transition to capitalism" only for Europe and not also describe as societies in transition toward capitalism the Arab-Islamic or Chinese societies in which these elements of "protocapitalism" can be found? Indeed, why not abandon the notion of "transition" altogether, in favor of a "constant evolution of a system in existence for a long while, in which the elements of protocapitalismhave been present since very ancient times"? My second comment explains in part my hesitation in following the suggestions made ab. The colonization of America accelerated to an exceptional extent the expansion of the protocapitalist elements indicated above. For three centuries the social systems that participated in the colonization were dominated by such elements. This had not been the case elsewhere or before. On the contrary, the protocapitalist segments of society had remained cloistered in a world dominated by tributary social relations (feudal in medieval Europe). So let us now clarify what we mean here by the domination of tributary relations.

One question we might ask is whether the dense network of Italian cities did or did not constitute a "protocapitalist system." Undoubtedly protocapitalist forms were present at the level of the social and political organization of these dominant cities. But can the Italian cities (and even others, in south Germany, the Hanseatic cities, etc.) really be separated from the wider body of medieval Christendom? That wider body remained dominated by feudal rural life, with its ramifications at the political and ideological levels: customary law, the fragmentation of powers, cultural monopoly of the Church, and so on. In this spirit it seems to me essential to give due weight to the evolution of the political system of "protocapitalist" Europe from the sixteenth to the eighteenth centuries. The evolution that led from the feudal fragmentation of medieval power to the centralization of the absolute monarchy kept pace precisely with the acceleration of protocapitalist developments. This European "specificity" is remarkable, since elsewhere - in China or in the Arab-Islamic world for example -there is no known equivalent of "feudal fragmentation"; the (centralized) state precedes "protocapitalism." I have attributed this European specificity to the "peripheral" character of the feudal society - the product of a grafting of the Mediterranean tributary formation onto a body still largely at the backward communal stage (the Europe of the barbarians).

The (belated) crystallization of the state, in the form of absolute monarchy, implied, at the outset, relations between the state and the various components of the society that differed abstractly from those that were the case for the central tributary state. The central tributary state merged with the tributary dominant class, which had no existence outside it. The state of the European absolute monarchies was, on the contrary, built on the ruins of the power of the tributary class of the peripheral modality and relied strongly in its state-building on the protocapitalist urban elements (the nascent bourgeoisie) and rural elements (peasantry evolving toward the market). Absolutism resulted from this balance between the new and rising protocapitalist forces and the vestiges of feudal exploitation.

An echo of this "specificity" can be found in the ideology accompanying the formation of the state of the ancien regime, from the Renaissance to the Enlightenment of the eighteenth century. I stress the "specificity" -and in my opinion advanced character - of this ideology, which broke with the tributary ideology. In the latter scheme, the predominance of a metaphysical view of the world is based on the dominance of the political instance over the economic base. To avoidany misunderstanding, I stress that metaphysics is not synonymous with "irrationality" (as the radical currents of the Enlightenment have painted it), but seeks to reconcile Reason and Faith (see my discussion of this theme in Eurocentrism). The ideological revolution from the Renaissance to the Enlightenment did not suppress metaphysics (metaphysical needs), but freed the sciences from their subjection to it and thereby paved the way to the constitution of a. new scientific field, that of the social sciences. At the same time of course (far from accidental) concomitance between the practices of the new state (of the ancien regime) and developments in the field of ideology stimulated protocapitalist expansion. The European societies began to move rapidly toward the "bourgeois revolution" (1688 in England, 1776 in New England, 1789 in France). They challenged the absolutist system that had provided a platform for protocapitalist advances. New concepts of power legitimized by democracy (however qualified) were introduced. It is also from there on that the Europeans developed a new "awareness" of their specificity. Before the Renaissance the Europeans (of medieval Christendom) knew they were not "superior" (in power potential) to the advanced societies of the Orient, even if they regarded their religion as "superior," just as the others did! From the Renaissance on, they knew they had acquired at least potential superiority over all the other societies and could henceforth conquer the entire globe, which they proceeded to do.


Everybody knows that the Arab-Islamic Mediterranean and Middle East region enjoyed a brilliant civilization even before the Italian cities. But did the Arab-Islamic world constitute a protocapitalist system? The protocapitalist forms are present and, at certain times and places, inspired a glorious civilization. The views I have put forward on this subject (see The Arab Nation, Eurocentrism) tie in with Mansour Fawzy's book (1990) on the historical roots of the impasse of the Arab world, and, in some regards, with the works of the late Ahmad Sadek Saad. Beyond possible divergences - or shades of meaning - we are of the common opinion that the Arab-Islamic political system was not dominated by protocapitalist (mercantilist) forces but, on the contrary, that the protocapitalist elements remained subject to the logic of the dominant tributary system power. In fact, I consider the Arab-Islamic world as part of a larger regional system which I call the Mediterranean system.

I have suggested (in Eurocentrism) that we can date the birth of this "Mediterranean system" from the conquests of Alexander the Great (third century Be) and conceptualize a single long historic period running from this date to the Renaissance, encompassing at first the "Ancient Orient"] (around the eastern basin of the Mediterranean), then the Mediterranean I as a whole and its Arab-Islamic and European extensions.

I have in this regard put forward the thesis that we are dealing with a I single tributary system from 300 bc (unification of the Orient by Alexander the Great) to 1492. I refer to a single "cultural area" whose unity is manifested in a common metaphysical formulation (the tributary ideology of the region), beyond the successive expressions of this metaphysics (Hellenistic, eastern Christian, Islamic, western Christian). In this tributary area I find it useful to distinguish between its central regions (the Mediterranean Orient) and its peripheral regions (the European West). Within this entity exchanges of every kind have (nearly always) been highly intensive and the associated protocapitalist forms highly advanced, particularly evident in the central regions (in the period of the first flowering of Islam from j the eighth to the twelfth centuries and in Italy for the succeeding centuries). These exchanges have been the means of a significant redistribution of surplus. However, the eventual "centralization" of surplus was essentially I tied to the centralization of political power. From that point of view the I cultural area as a whole never constituted a single "unified imperial state" I (except for the two brief periods of the Alexandrine empire and the Roman empire occupying all the central regions of the system). Generally speaking, the peripheral region of the European West remained extremely fragmented under the feudal form (and this is the very expression of its peripheral character). The central region was divided between the Christian Byzantine Orient and the Arab-Islamic empires (the Umayyad, then the Abbasid dynasties). It was first subject to internal centrifugal forces, then belatedly unified in the Ottoman empire, whose coincided with the ... end of the period and the overall peripheralization of the eastern region to the benefit of a shift of the center towthe previously peripheral I region of Europe and the Atlantic.

Could this "system" be described as protocapitalist? In support of the thesis is the presence of undeniable protocapitalist elements (private ownership, commodity enterprise, wage labor) throughout the period, expanding in certain places and times (especially in the Islamic area and in Italy), declining in others (especially in barbarian Europe of the first millennium). But in my view the presence of these elements does not suffice to characterize the system. On the contrary, I would argue that, at the crucial level of ideology, what began in the Hellenistic phase of this period (from 300 bc to the first centuries ad), and then flourished in the (eastern then western) Christian and Islamic forms, is purely and simply the tributary ideology, with its major fundamental characteristic: the predominance of metaphysical concerns.

What we are talking about is indeed a "system," but not a "protocapitalist system," that is, a stage in the rapid transition from tributary society to capitalist society. On the other hand, we are dealing with a "tributary system," not a mere juxtaposition of autonomous tributary societies (in the plural), which just happened to share some common elements, such as religion, for example, or integration - albeit of limited duration - in an imperial state, such as that of Rome, Byzantium, or the Umayyad or Abbasid dynasty.

The distinction implies in my view a certain degree of centralization of surplus, which took the form of tribute and not, as in capitalism, that of profit from capital. The normal method of centralization of this tributary surplus was political centralization, operating to the advantage of imperial capitals (Rome, Byzantium, Damascus, Baghdad). Of course this centralization remained weak, as did the authority of the centers concerned. Byzantium, Damascus, and Baghdad could not prevent their staging-posts (Alexandria, Cairo, Fez, Kairouan, Genoa, Venice, Pisa, and so on) from frequently achieving their own autonomy. The entirety of barbarian Christendom (the first millennium in the West) escaped such centralization. In parallel, the logic of the centralization of authority stimulated protocapitalist relations to the point that mercantile handling of part of the surplus never disappeared from the region, and took on great significance in some areas and epochs, notably during the glorious centuries of Islam, and the emergence of the Italian cities following the Crusades. On this basis I have described the social formations of the Arab world as tributary-mercantile formations. All this leads me to conclude that capitalism "might have been" born in the Arab world. This takes me back to other discussions on this issue with which I have been associated. I have argued that once capitalism had appeared in Europe and the Atlantic, the process of evolution toward capitalism was brutally halted in its development elsewhere, The reason why the evolution toward capitalism accelerated in the Atlantic West (shifting the center of gravity of the system from the banks of the Mediterranean to the shores of the" Atlantic Ocean), it seems to me, is mainly due to the colonization (of America, then of the entire globe) and contingently to the peripheral character of western feudalism.


My methodological hypothesis leads me to regard the other "cultural areas" as further autonomous tributary systems. In particular, it seems to me that the Confucian-Chinese tributary system constituted a world on its own and of its own. It had its own center (China), characterized by a strong political centralization (even if the latter under the pressure of internal centrifugal forces exploded from time to time. But it was always reconstituted), and its peripheries (Japan especially) had a relationship with China very similar to that of medieval Europe with the civilized Orient.

I leave a dotted line after the question whether the Hindu cultural area constituted a (single) tributary system.

This having been said, the question is: was the Mediterranean system "isolated" or in close relation with the other Asiatic and African systems? Can the existence of a "permanent" world-system, in constant evolution, be argued beyond the Mediterranean area and prior to its constitution? A positive response to this question has been suggested to some (notably Frank) by the intensity of exchange relations between the protocapitalist Mediterranean, the Chinese and Indian Orient, and sub-Saharan Africa, and perhaps even the significance of the exchanges in earlier times between these various regions of the ancient world. For my part, I do not believe that it is possible to answer the question, given the current state of knowledge. It is, however, useful to raise it in order to provoke a systematic exchange of views on what can be deduced from our knowledge, the hypotheses it may inspire, and the directions of research indicated for verification of these hypotheses.

I do not intend to substitute my own "intuitive views" for the eventual results of these debates. I advance them here only provisionally, to open the discussion. I should therefore suggest the following (provisional) theses.

First, humankind is one since its origins. The itinerary of the earth's population begins from the nucleus of hominids appearing in East Africa, going down the Nile and populating Africa, crossing the Mediterranean and the Isthmus of Suez to conquer Europe and Asia, passing the Bering Straits and perhaps crossing the Pacific to install themselves (in the most recent epoch) in the Americas. These successive conquests of the planet's territory are beginning to be dated. The following may be the pertinent question: has the dispersal brought a "diversification" of the lines of evolution of the various human groups, installed in geographical environments of extreme diversity and hence exposed to challenges of differing kinds? Or dpes the existence of parallel lines of evolution suggest the conclusion that humankind as a whole has remained governed by "laws" of evolution of universal application? And as a complement to this question it might be asked what effect have relations between the scattered human populations had on the fate, intensity, and rapidity of the transfer of knowledge, experience, and ideas?

Intuitively it might be imagined that some human groups have found themselves fairly isolated in particularly difficult circumstances and have responded to the challenge by particular adaptations unlikely to evolve of themselves. These groups would then be located in "impasses," constrained to reproduce their own organization without the latter showing signs of its own supersession. Perhaps included here would be the (still highly fragmented) societies of hunters/fishers/gatherers of the Arctic, the equatorial forest, small islands, and some coasts.

But other groups have found themselves in less arduous circumstances that have enabled them to progress simultaneously in mastery of nature (passage to settled agriculture, invention of more efficient tools, and so on) and in tighter social organization. In regard to the latter the question arises of "possible laws of social evolution of universal application" and the role of external relations in this evolution.

Secondly, in regard to societies that have clearly "advanced," can one detect similar phasing followed by all, albeit at faster or slower rates? Our entire social science is based on this seemingly necessary "hypothesis." For the satisfaction of the spirit? As legitimation of a universalist value system? Various formulations of this "necessary evolution" succeeded one another up to and during the nineteenth century. They were based either on the succession of modes of exploitation of the soil and instruments utilized (Old Stone Age, New Stone Age, Iron Age), or on the succession of social forms of organization (the of Savagery, Barbarism, Civilization). Various evolutions in these "particular" domains were regrafted on to what we regarded as fundamental general tendencies. For example, the "matriarchal-patriarchal" s, the succession of the ages of philosophical thought (primitive, animist, metaphysical, Auguste Comte-style positivist), and so on. I shall not spend time here discussing these "theories," which are almost always more or less overridden by subsequent research. I merely point to their existence as evidence of the persistence of the need to "generalize," beyond the evident diversity that is the property of the scientific approach.

It seems to me that the most sophisticated formulation of all the theories of general evolution was that proposed by Marxism and based on the synthetic notions of "modes of production." The latter comes from a conceptualization of the basic elements of the construction (forces of production, relations of production, infrastructure and superstructure, etc.). They are then "enriched" by the grafting on of particular theories articulated to those of "modes of production" (such as theory of the family, of the state, etc.). Here again I shall not discuss whether these Marxist constructs are indeed those of Marx himself, or the product of later interpretations that may or may not be consonant with the spirit of the Marxism of Marx. Nor shall I discuss the validity of these theories in the light of our present-day greater knowledge of the societies of the past. Once again I merely point to the formulations as the expression of this same need to "Understand," which implies the possibility of "generalizing."

Thirdly, on the basis of the conceptualization proposed, it is not difficult to identify several tributary societies at more or less the same level of maturity of general development: production techniques, instruments, range of goods, forms of organization of power, systems of knowledge and ideas, and so on. Noteworthy too is a fairly dense web of exchanges of all kinds between these societies: exchange of goods, knowledge, techniques, and ideas. Does this density of exchange justify speaking of a single world-system (albeit described as tributary) - in the singular? Frank and Gills provide an explicit criterion: an integrated system arises when reciprocal influences are "decisive" (A would not be what it is without the relation it has with B). So be it. But the overall question remains: were these relations "decisive" or not?

However, the universality of the laws of social evolution in no way implies the concept of a single system. Two distinct concepts are involved. The first refers to the fact that distinct societies - separated in geographical distance or time - have been able to evolve in a parallel manner for the same underlying reasons. The second implies that these societies are not distinct from one another but ingredients of the same world society. In the evolution of the latter - necessarily global - the laws in question are inseparable from the effects of the interaction between the various components of the world society.

I would in this context make two prefatory comments. 1) Economic exchanges are not necessarily a "decorative" element, making no lasting impression on the "mode of production" and hence on the level of development. Exchanges may be a significant means of distribution of surplus, decisive for some segments of the interrelated societies. The question is not one of principle but of fact. Were they? Where and when? I discount any hasty generalization that they were always (or generally) so or that they were never (or with rare exceptions) so. In the case of the Arab-Islamic region, for example, I have said that the exchanges were significant. They were enough to mark the formation of a "tributary-mercantile" character essential to an understanding of its involuted history of succession from a "glorious" phase to one of "degeneration," and of shifts of the centers of gravity of wealth and power in the region. I have also said that the "protocapitalist" formation of mercantilist Europe (seventeenth-eighteenth centuries) rapidly climbed the step toward capitalism thanks to these exchanges it dominated. But whether the exchanges had a matching role in China, India, the Roman empire, etc., I personally am in no position to say. 2) The exchanges in question must not be limited only to the economic field. Far from it. The writing of the history of the precapitalist epochs puts greater emphasis on cultural exchanges (especially the spread of religions) and military and political exchanges (rise and fall of empires, "barbarian" invasions, etc.), whereas the accent is on the economic aspect of relations within the modern world-system. Was this distinction wrong?

I do not think so. I believe, on the contrary, that the historians -*albeit intuitively - have grasped the reversal of dominance, from the political and ideological to the economic, which is the central core of my own thesis. At this level is it possible to speak of a single tributary political and ideological world-system? I do not believe so. I have therefore preferred to speak of distinct tributary "cultural areas" founded precisely on broad systems of particular reference - most often the religious: Confucianism, Hinduism, Islam, Christianity. Of course there is a certain relationship between these various metaphysics since they express the fundamental requirement of the same type of (tributary) society. The relationship in turn facilitates mutual borrowings.

To approach an answer to the question (of one or more systems), it is necessary to combine three elements: the density of economic exchanges and transfers of surplus distributed through this channel; the degree of centralization of political power; the relative diversity/specificity and hence autonomy of the ideological systems.

Autonomy of the various tributary systems does not preclude economic relations and other exchanges among them, nor even that such exchanges could be significant. It would be impossible to understand many historical facts and evolutions without reference to these exchanges: the transfer of technology of all kinds (the compass, gunpowder, paper, silk that gave its name to the roads in question, printing, Chinese noodles becoming Italian pasta, etc.); the spread of religious beliefs (Buddhism crossing from India to China and Japan, Islam traveling as far as Indonesia and China, Christianity as far as Ethiopia, south India, and Central Asia), etc.

There is certainly no centralization of surplus at the level of a world-system comparable to that characterizing the modern world in the exchanges that led here and there to lively protocapitalist links (from China and India to the Islamic world, the African Sahel, and medieval Europe) and transfers of surplus - perhaps even decisive at key points of the network of exchanges. The explanation is that centralization of surplus at the time operated mainly in association with centralization of power, and there was no kind of "world-empire" or even a "world power" comparable to what British hegemony would constitute in the nineteenth century or US hegemony in the twentieth.

The ancient (tributary) epochs had nothing comparable to the "polarization" on a global scale of the modern capitalist world. The earlier systems, despite significant levels of exchange, were not polarizing on a world scale, even if they were on a regional scale to the benefit of the centers of the regional systems (for example, Rome, Constantinople, Baghdad, the Italian cities, China, India). By contrast, the capitalist system is truly polarizing on a global scale and is therefore the only one deservedly described as a world-system.

This methodology for the analysis of the interactions between the tributary systems may call for a reassessment of the "traditional" findings in the history of the notorious "barbarians" who occupied the interstices of the great tributary cultural areas. Was the role of these "barbarians" really as it has been made out, a purely negative and "destructive" role? Or did their active role in Intel-trexchanges give them a certain vocation to take decisive initiatives? The latter would explain their success (not only military) in "unifying" immense territories (Genghis Khan's empire), their capacity to situate themselveat the heart of ideological initiatives (Islam born in Arabia, the "barbarian" crossroads of Mediterranean-Indian-African exchanges), their capacity to hoist themselves rapidly to central positions in a tributary system (the glorious example of the Khwarizm area in the first centuries of Islam), etc.

A final reservation concerning the systematization of the hypothesis of the existence of a single world-system throughout history: is it possible to speak of tributary systems and significant exchange networks among them before the fifth to third centuries bc? I do not think so for the following three reasons at least: 1) because the social systems of the greater part of humankind were still backward at the stage I have described as communal; 2) because the islets of civilization at the stage where the state was the recognized form of the expression of power had not yet found complete tributary ideological expression (see the argument on the ideology of the ancient world in Eurocentrism); 3) because the density of the exchange relations between these islets remained weak (this did not prelude some exchange relations; for example, technological borrowings that were able to travel unexpected distances).


The theory according to which all human societies have been forever integrated in a single world-system, in continuous evolution (capitalism not representing therefore any kind of qualitative break in this respect), arises from a philosophy of history which is in the end based on the notion of competition. Certainly it is based on a realistic observation of facts, namely, that all societies on earth, in all eras, are to some extent in "competition" with one another. It would not matter whether the relations they did or did not entertain showed their awareness of it. We know that the strongest must carry the day. At this level of abstraction there is indeed a single world, because there is a single humankind. It might perhaps be added that most "open" societies with intensive relations with the others have a greater chance of measuring up to this competition and facing up to it more effectively. It is otherwise for those who shy away from competition and seek to perpetuate their way of life; they risk being overtaken by the progress made elsewhere and later being marginalized.

This discourse is not wrong, but merely at such a high level of abstraction that it begs the real issue, namely, how this competition is manifested. Two bourgeois historians - themselves philosophers of history - deliberately placed themselves at this most general level of abstraction (in order to refute Marx). Arnold Toynbee in this regard suggests an operative model reduced to two terms: the "challenge" and the "response to the challenge." I suggest that, as a model valid for all times and all places, it teaches us nothing that is not already obvious. Toynbee suggests no law to explain why the challenge is taken up or not. He is satisfied with a case-by-case treatment. There is an almost natural parallel with the contradiction between the axioms of neoclassical bourgeois economics defined in terms claiming to be valid for all times ("scarcity," "utility," etc.) and the historical concept of qualitatively differing successive modes of production, determining specific institutional frameworks in which the "eternal rationality of human beings" is expressed. Jacques Pirenne, far superior to Toynbee in my opinion, suggests a refinement of constant contradiction between (sea-going) "open" societies and (land-based) "closed" societies and does not hesitate to describe the former as "capitalist" (Sumer, Phoenicia, Greece, Islam in the first centuries, the Italian cities, the modern West) and the latter as "feudal" (from ancient Persia to the European Middle Ages). He never hesitated to attribute to ^hat I call "protocapitalist elements" the decisive place in the progress of the "open" societies making them the driving force of development of the forces of production. He likewise never concealed that his thesis was intended to discount the "closed" experiences of the Soviet Union and salute the dynamism of the Atlantic world. Hence Pirenne managed - certainly with skill - to replace class struggle with a constant struggle between the capitalist tendency and the feudal tendency within human societies.

I still believe that Marx's method is superior, precisely because it situates the abstraction at the appropriate level. The concept of modes of production gives back to history its explicit real dimension. At that level the significance and character of the capitalist break can be detected. The break is such that I do not think that competition between societies of earlier times and within the modern world-system can be treated in the same way. First because the competition of earlier times rarely crossed the threshold of consciousness and each society saw, or believed, itself "superior" in its own way, "protected by its deities," even when a looming danger imposed a greater consciousness (as between Muslims and Crusaders). Moreover, the discrepancy between the great tributary precapitalist societies is not such that the superiority of one over another is obvious; it is always conjunctural and relative. There is nothing comparable to the subsequent overwhelming superiority of capitalist societies over the rest. That is why I see the seizing of consciousness of this superiority as crucially important and therefore date the beginnings of capitalism to 1492. From then on the Europeans knew that they could conquer the world and went on to do so (see my arguments on this point in Eurocentrism'). We know a posteriori - but the actors of the time were unaware - that the "strongest" is the one who has advanced to a qualitatively superior mode of production - capitalism. I would add that in the competition of earlier times geographical distance had a blunting effect. However intensive exchanges between Rome and China, I find it difficult to believe that the "external" factor could have a similar impact to that of the discrepancies in productivity of our own times. I believe that this distancing gave strictly internal factors a considerably more decisive relative weight. It also explains why those concerned had difficulty in assessing the real balance of forces. Quite different, it seems to me, is competition within the modern world-system, where consciousness is so acute that it is a plaintive chorus in the daily discourse of the authorities.


The diagram opposite (Figure 8.1) illustrates my concept of the "ancient world-system" (reduced to societies of the so-called eastern hemisphere: Eurasia-Africa) for the periods covering the eighteen centuries between the establishment of the Hellenistic system in the Middle East (300 bc), the establishment of the Han state in China (200 bc), the Kushana and Maurya states in Central Asia and India (200 ac), and the European Renaissance, that is, from 300 bc to 1500 ad. I wish to summarize its characteristics as follows.

First, as I have already said, all societies of the system in question are, from one end of the period to the other, of a tributary nature. Nevertheless, it is possible to distinguish among all these societies those which I would call "central tributaries" from those which are "peripheral tributaries." The former are characterized by a surplus centralization at the relatively high state level, with its redistribution placed under its control; while in peripheral formations, the embryonic character of the state (and even its virtual nonexistence) leads to a complete disintegration of surplus distribution monopolized by local feudal systems. The centers/peripheries antithesis is not, in this case, analogous to that which characterizes the (modern) capitalist world. In the latter, the relationship in question is an economicdomination relationship in which the centers override the peripheries (and this is associated with economic dominance). This is not so in the ancient relationship. Dominated by the ideological authority, the tributary structures are either or peripheral depending on the degree of the completion of the power-centralization process and its expression through a state religion. In the central formations, the latter takes the form of a state religion or a religious-oriented state philosophy with a universal vocation which breaks with the specific local religions of the former periods which I called "communal formations" (see Class and Nation). There is a striking relationship between the establishment of big tributary societies in their completed form and the emergence of great religious and philosophical trends which were to dominate civilizations over the ensuing two thousand years: Hellenism (300 ac), Oriental Christianity, Islam (600 ad), Zoroaster, Buddha, and Confucius (all three 500 bc). This relationship -which in no way excluded the reciprocal concessions provided by the relations that all tributary civilizations maintained among themselves - is not, in my view, an accident, but rather one of the consistent bases of my thesis on the dominant "tributary mode."

The establishment of great philosophical and religious movements associated with the formation of tributary systems represents the first wave of revolutions related to universal history, which is expressed by a universalist-oriented vocation transcending the horizons of the local -almost parochial - line of thinking in the ancient periods. This revolution sets up the tributary system as a general system at the entire level of mankind - or almost does so - for 2,000 to 2,500 years. The second wave of universal-oriented revolutions, which opens up capitalist modernity and its possible socialist overtaking, is marked by the Renaissance (and the revolution in Christianity with which it is associated) and, subsequently, by the three great modern revolutions, the French, Russian, and Chinese revolutions (see Eurocentrism).

The "model" par excellence of this tributary mode is, in my view, provided by China, which, without, it seems, a long incubation period (there is only one millennium between the Shang and the Zhou and the establishment of the Han dynasty), crystallizes in a form which undergoes no fundamental change, either with regard to the organization of productive forces and production relationships or ideology (the Confucianism-Taoism tandem replaced for only a brief moment by Buddhism), or with regard to power concepts during the 2,000 years between the Han dynasty and the 1911 revolution. Here, surplus centralization is at its height, at the level of an enormous society, not only during the brilliant periods where political unity was entirely or almost entirely achieved in this continent-country by great successive dynasties (Han, Tang, Song, Yuan, Ming, and Qing), but even during the periods of imerdynastic disturbances when the country was divided into several kingdoms whose size was nonetheless considerable for the period. At the borders of China, Korea and Vietnam also turned, during the course of the first millennium of our era, into similar tributary systems which, in spite of their political independence with regard to China, borrowed its model of organization and Confucian ideology.

In the Middle East, the tributary system derived its completed form from the conquest of Alexander the Great. I have recommended in this connection (see Eurocentrism) this reading of the successive philosophical and religious orientations of Hellenism, Oriental Christianity, and Islam, However, in this region, the incubation period lasted for as long as thirty centuries for Egypt and Mesopotamia, ten centuries for Persia, Phoenicia, etc., and five centuries for Greece. Hellenism, Christianity, and Islam were, moreover, to produce a synopsis which borrowed some elements crucial to each of these ancient components and even from Persia and India as well. Here, too, surplus centralization for the ensuing 2,000 years is remarkable.

Doubtless, the region was split after the precarious political unification in the Alexander era; but it was split into large kingdoms for the period. Hence, divided between even bigger empires - those of Byzantium (300 to 1400 ad) and the Sassanids (200 to 600 ad) - and subsequently reunified gradually through the expansion of the Muslim caliphate, formed in the seventh century ad, which conquered Constantinople at the end of our period (in 1453), the spaces of surplus centralization were still either vast (during the first three centuries of the caliphate), or at the very least, considerable, after the break-up of the caliphate from the year 1000 to the advantage of Arabo-Berber dynasties in north Africa and Turco-Persians in the Mashreq and western part of Central Asia. The western Roman empire finds its place in this reading of history as an expression of an expansion of the tributary model to the banks of the western Mediterranean. Of secondary importance in universal history, the Roman empire owes its place to the fact that it has transmitted tributary ideology - in the form of western Christianity - to the "European" periphery.

A Eurocentric reading of history (see my critical appraisal in Euro-centrism) has, in this regard, distorted the achievements which, beyond the Italian peninsula, failed to resist barbaric feudalization (that is, the disintegration of the tributary system).

A third completed tributary center was established on the Indian continent in 200 bc from the Maurya period, followed by the Kushana state (which overlaps the western part of Central Asia) and Gupta after the long incubation period which began with the Indus civilizations (Mohenjodaro and Harappa - 2500 bc). The Muslim conquest from the eleventh century on which followed after a "pulverization" period (of the seventh and ninth centuries) re-established together with the Ghazhavids, the sultanates of Delhi (1200-1500 ad), and subsequently the Mughal empire (1500-1800 ad), a tributary centralization on a large scale, while'the Hindu states of Dekkan, also tributaries, equally represented considerable kingdoms for the period.

Three zones appear in Figure 8,1 whose peripheral character is striking during the entire or almost entire period under consideration (from 300 bc to 1500 ad). Europe (beyond the Byzantine region and Italy, that is, "barbaric" Europe) was the product of a tributary graft (transmitted by the ideal of the Roman empire and Christian universalism) on a social -body still organized, to a large extent, on deteriorated community bases. Here, I wish to refer to the analysis I made (see Class and Nation) which simultaneously gives an account of the disintegration in the control of surpluses, and which defines feudalism as an uncompleted peripheral form of the tributary system, although the collapse of the state system was partially offset by the Church. Europe was slowly moving toward the tributary form, as testified by the establishment of absolute monarchies (in Spain and Portugal after the reconquista, and in England and France after the Hundred Years War). This belatedness constitutes, in my view, the crucial advantage which facilitated the early qualitative strides made by the Renaissance and capitalism (see Class and Nation).

Japan constituted, at the other end of the Euro-Asian continent, a peripheral tributary mode whose resemblance to Europe had struck me even before Mishio Morishima came to confirm my thesis. The degraded form of Japanese Confucianism, the feudal disintegration which preceded the belated formation of a monarchical centralization from the Tokugawa state (1600 ad), bear testimony to this peripheral character (see Eurocentrism), which, here, too, explains the remarkable ease with which Japan switched over to capitalism in the ninteenth century.

Sub-Saharan Africa constituted the third periphery. It was still lingering at the communal stage developing toward tributary forms. At this stage the tributary surplus centralizations still opeonly on societies with limited size. Disintegration therefore remained the rule.

The status of Southeast Asia was ambivalent. It seems to me that here it is possible to recognize some central type of tributary formations - even if they only cover spaces than those of other great Asian systems - and peripheral zones (denned by surplus disintegration). To the first type belongs the Khmer empire, followed by its Thai, Burmese, and Cambodian successors from the fifth century and, perhaps, in Indonesia, the Majapahit kingdom from the thirteenth century. On the other hand, the organized societies of Malaysia and Indonesia which crystallized into states under the influence of Hinduism (from the fifth century) and subsequently Islam, seem, in my view, to belong to the peripheral family, crumbled by the scattering of the surplus, collected in very small and relatively numerous and fragile states.

The status of the Central Asian region was special. The region itself is less defined in its borders than the others. Some large states were established in this region at an early period - such as the Kushana empire - which directly linked up the Hellenistic Middle East and the Sassanids and then the Islamic Middle East to India and China. The region itself became the center of gravity of an immense empire at the time of Genghis Khan (1300 ad). Before and after this final crystallization, it had entered the Islamic orbit. Its modes of organization were tributary-oriented, at one time advanced (where the expression of centralized power on a large scale makes it possible), at another time relapsing into "feudal" disintegration. But the major feature of the region was that, by virtue of its very geographical position, it was the indispensable transit zone for East-West trade China, India, the Middle East, and beyond to as far as the peripheries of the system). Having been in competition with the sea route from time immemorial, the continental route lost its importance only belatedly in the sixteenth century.

As for the second characteristic of the ancient world-system: during the entire eighteenth-century period under consideration, all the societies represented in Figure 8.1 not only existed together, but still maintained trade links of all types (trade and war, technological and cultural transfers) which were much more intense than was generally thought. In this very general sense, one can talk of the "general system" without, of course, mistaking its nature for that of the modern (capitalist) world-system. In Figure 8.1, I represent these links by eleven arrows. Of course, the intensity of flows that each of these arrows represents varied considerably with time and space. But above all - and I wish to emphasize this point - their connection with the internal dynamics peculiar to the different tributary systems they link up is not only fundamentally different from that which characterizes the "international links" within the modern world-system, but has also operated differently from one tributary formation to another. To clarify things, I want to distinguish four sets of links:

1 The links mutually maintained between the three major centers (A -Rome and Byzantium, the Sassanid empire, the caliphate; B - China; C - India) are marked by arrows 1 (Middle East-China through Central and northern Asia), 2 (Middle East-India across western Central Asia), and 3 (Middle East-India by sea route). These links were undoubtedly the most intense of all, merely in view of the wealth and relative power of the centers in question, at least in the glorious years of their history.

2 The links maintained by the Arabo-Persian Islamic center with the three peripheries (Europe, Africa, Southeast Asia) are shown by arrows 4 (Middle East-Malaysia, Indonesia sea route), 5 (north Africa-African Sahel trans-Saharan route), 6 (Middle East-Swahili eastern coast sea route), and 7 (caliphate and Byzantium-Europe). The trade in question was less intense than that of the previous group (due to the relative poverty of the peripheries), and especially important is the fact that it was asymmetrical (a concept that I clearly distinguish from the specific inequality of the centers/peripheries relationships of the modem world) in the sense that they were perhaps neutral in their effects on the center, but crucial for the development of the peripheries. These relationships considerably accelerated the establishment of states in the African Sahel and East Africa (see Class and Nation) as well as in Malaysia and Indonesia and thus opened the way for the Islamization of these regions (Islam then replacing the ancient local religions in line with the needs of the tributary world). They also contributed immensely to the emergence of Italian trading cities, and, through these cities, of infiltration throughout the whole of feudal Europe.

3 The links maintained by the Chinese center with the Japanese periphery (arrow 8) and the Southeast Asian periphery (arrow 9) are of the same nature as those in the second group. Here, I wish to refer to arrow 11, which indicates a direct communication establishment between China and Europe, using of course the routes of Central Asia but without passing through the canal in the heart of the Islamic caliphate. This direct relation existed only for a relatively short period, within the framework of the Mongol Pax (the Genghis Khan empire in the thirteenth century). But it was crucial for subsequent events of history because it made it possible for Europe to resort to China's vast technological accomplishments (gunpowder, printing, the compass, etc.); Europe was mature enough to do this and take the qualitative leap from a peripheral tributary (feudal) system to capitalism. Furthermore, shortly thereafter, Europe substituted the sea route it dominated for all ancient forms of long-haul transport, thus establishing direct links between itself and each of the other regions of the world (Africa, India, Southeast Asia), "discovering" and then "conquering" America at the same time. 4 The links maintained by the Indian center (Buddhist and Hindu) with its Southeast Asian peripheries (arrow 10) are similar to the China-Japan links.

It obviously appears that the relative intensity of "external" flows, as compared with the different masses constituted by the regional formations under consideration, varies considerably from one region to another. The three key central regions, A, B, and C (Middle East, China, India), represented, in terms of economic weight, a multiple of what constituted each of the other regions. If, therefore, the volume of the surplus identified in each of these key central regions is measured by index 1,000, it could hardly have exceeded index 100 for each of the other regions (Europe, Africa, Japan, Central Asia, and Southeast Asia). Moreover, only a part and probably a relatively minor part (10-20 per cent perhaps) of this surplus could involve long-distance trade.

The four arrows which concern China (major 1, minors 8 and 9, and transitory 11) could, for instance, represent an index "value" of about 100 (10 per cent of the surplus produced in China). The three arrows which concern India (majors 2 and 3 and minor 10) probably hardly exceeded index 50 or 70. All historians have observed that the "external" trade of these two continental masses were marginal as compared with their volume of production.

On the other hand, the weight of external trade seems more pronounced for region A, which is the only region in direct relationship with all the others. To major arrows 1, 2, and 3 representing A's trade with B and C (total index value: 115 in our assumption) is added the region's trade with the peripheries of Europe (arrow 7), Africa (arrows 5 and 6), and Southeast Asia (arrow 4), making a total index value of about 25. In sum then, external trade, in this case, would have represented an index value of 140 (almost 20 per cent of the surplus?).

For each of the peripheries too the contribution of external trade would appear relatively considerable: index 20 for Europe, 10 for Africa, 20 for Southeast Asia, and 20 for Japan, that is, 20-30per cent of the surplus generated in these regions. Similarly, transit flows through Central Asia (arrows 1, 2, and 11) on the order of index 100, might have accounted for a volume even greater than that of the locally produced surplus.

The index values assignto both the surplus volumes produced in each region and the trade volumes indicated by each of the arrows are, of course, mere fabrications on my part created with a view to suggesting some relative orders of magnitude. It is for historians to improve upon them. Failing this (and we have not found any figures in this regard) the figures I have used constitute some orders of magnitude which seem plausible to me and which can be summarized in Table 8.1 below:

Table 8.1 Locally generated external flows

Surplus (1)


% (2/1)

Middle East
























Southeast Asia




Central Asia




Geography has assigned to key central region A an exceptional role without any possible competitor until modern times, when Europe, through its control over the seas, overcame the constraints. Indeed, this region is directly linked to all the others (China, India, Europe, Africa) and is the only one as such. For two millennia, it was an indispensable transit route to Europe, China, India, or Africa. Besides, the region does not reflect a relative homogeneity similar to that of China or India, either at the geographical level (stretching from the Moroccan shores of the Atlantic to the Aral Sea, Pamir and to the Oman Sea, it does not have the features of a continental block as in the case of China and India), or at -the level of its peoples, who themselves are products of the early proliferation of the most ancient civilizations (Egypt, Sumer, Assyria, Mesopotamia, Iran, Hittites, Phoenicians, and Greeks) and speak languages from various families (Semitic, Hamitic, Indo-European). The conquest of Alexander the Great and the triumph of the Hellenistic synthesis triggered a collective awareness which was subsequently strengthened by Oriental Christianity (limited by the Sassanid border) and subsequently and, above all, by Islam.

One of the keys to the success of Islam relates, in my view, to this reality. The region was finally firmly established within the short period covering the first three centuries of the Hegira. It was thus composed of their three superimposed strata of Islamized peoples, namely, the Arabs from the Atlantic to the Gulf, the Persians beyond Zagros to Pakistan, the Turks in Anatolia and in the entire Turkestan from the Caspian Sea to China proper. Thus, Islam did not only unify the peoples of the so-called classical "East" but annexed, at the same time, Central Asia, the indispensable transit route to China and northern India. I think that this success should be attributed to the fact that, in spite of all the conflicts witnessed by history internal to this region, it created a certain solidarity and strengthened the sense of a particular identity with regard to the "others"; that is, specifically, the Chinese, Indians, Europeans, and Africans that the Muslim Umma borders on along each of its frontiers. In Central Asia, the success of Islam created regional unity, which, until then, was absent. For the civilization in this region, in which trade flows represent larger volumes than the surplus produced locally, depended on the capacity to capture, in passing, a part of these transit flows.

The magnitude of the links with the others for the entire key central region A and its Central Asia annex bestows on its social system a special character which I venture, for this reason, to call "mercantile-tributary," thus indicating even the magnitude of protocapitalist forms (commercial links, wage labor, private property, or estate) in the tributary societies of Islam. Moreover, beyond the original boundaries of Islam, the gradual conquest of African and Southeast Asian peripheries is also worth putting into close relationship with its mercantile dynamism of region A (see The Arab Nation and Class and Nation).

Thirdly, the world-system described above for the eighteen-century period preceding the Renaissance is not analogous to the modern system that follows it (in time). To talk about the ancient system in its spatial and time universality or even in its Arab-Islamic component as the "ancestor" of the modern system would be misleading. For this is only a platitude -succession in time and nothing more; or it implies that there was no qualitative break but only quantitative development and a "shift" of the system's center of gravity from the southern shore of the Mediterranean to its northern shore (Italian cities) and then to the Atlantic shores, and this boils down to eliminating the essential, that is, the qualitative change in the nature of the system: the law of value which governs the dynamics of the modern system but not those of the tributary system. This universalization of the law of value is exclusively responsible for the establishment of one single antinomy which operates worldwide (a center composed of historically established national centers as such and peripheries all economically dependent on this center), thus creating an ever increasing differentiation from one period to another between the center and the peripheries, over the entire five-century history of capitalism and for the entirely visible or imaginable horizon within the framework of its immanent laws. In this connection, there is nothing comparable to the lasting relative balance (for 2,000 years!) between the key central regions of the tributary period. This qualitative difference forbids talking about "interdependence" - unequal, as it were - of the different components of the ancient system in terms similar to those that govern the modern world. Key regions A, B, and C are certainly in "relation" with one another (and with the other regions); it remains to be demonstrated that this "interdependence" would have been essential. The parallelism in their trend is no evidence of the crucial nature of their "relations"; it only reflects the general character of the laws governing the social development of all humankind (thus defining the status of the "specificities"). The possible concomitance of the "rise" and the "fall" of states of the past is far from obvious.

A cursory glance at Table 8.3, which describes the parallel history of the three key centers and the other regions, shows that this concomitance is merely a matter of pure chance.

Pirenne had already observed - a view taken up by Andre Gunder Frank and Barry K. Gills - the concomitance of the fall of the Roman empire and that of the Han dynasty. But the Roman fall was followed by the rise of Byzantium, the Sassanid, and the Kushana state, while the decline of the Hans was followed, right from the year 600 (the height of barbarianism in the West) by the rise of the Tang, and, three centuries earlier, by that of the Guptas, whose fall coincided (also by chance) with the rise of Islam. There are no clues to the identification of the "general" cycles of the rise and fall. The very term "fall" is, even in this context, misleading; it is the fall of a form of state organization in a given region, but, in most cases, as regards the development of productive forces, there is no parallel fall. I am struck rather by the opposite phenomenon, that is, the continuity of these long parallel historical events: from Rome-Byzantium-Sassanids-Islam to the Ottomans and the Safavids, from the Maurya dynasty to that of the Mughal state, from the Han dynasty to those of Ming and Qing, there were only a few qualitative changes but a great quantitative progress on the same organizational (tributary) bases. This does not exclude the fact that, in examining local developments, it is possible to explain any particular political rise (or fall) - which may still be relative - by a special link in which "external relations" have occasionally played a role. Once again, there is nothing similar to the "cycles" of the capitalist economy, whose scope ireally global as a result of the universalization of the law of value, the basis of the modern capitalist economy.

The crystallization of a new modernity in Europe which was achieved within a short time (from the rise of Italian cities to the Renaissance; three to four centuries) is not t"repetition" of a "general" phenomenon under which would be subsumed all together the "birth" of civilizations (Egypt, Sumer, Harappa, Shang) and the "establishment of empires" (Achemenid, Alexander, Rome, Byzantium, Sassanid, Umayyad, Abbasid, Ottoman, Safavid, Maurya, Gupta, the Mughal state, Han, Tang, Song, Ming, Qing, and the Genghis Khan empire).

I proposed an explanation of this fact (see Class and Nation) that the qualitative break is first made within a tributary periphery (Europe) and not in one of its centers (A, B, or C) and is then repeated in another periphery (Japan). I based my explanation on the contrast between the flexibility of the peripheries and the rigidity of the centers, that is, while keeping to the logical context of the general nature of the laws of the evolution of societies (the "uneven development" which is the general form of an identical overall evolution). I consider this explanation more satisfactory than those proposed by the different characteristically Eurocentric conceptions (see Eurocentrism). I also think it is more satisfactory than Pirenne's theory, which I have referred to as being based on the permanent contrast between "capitalism" (the synonym of "openness," especially in "maritime" terms) and "feudalism" (the synonym of "closure," especially in "landlocked" terms). Like Andre Gunder Frank's and Barry K. Gills's (which is close to the extreme), Pirenne's theory is a transformation of the Eurocentric deformation: it "attributes" the European miracle to the maritime openness of the region, since each of the theories is based on the negation of the specific nature of the capitalist modernity.

Of course the crystallization of capitalism in Europe has a history (it is not done by magic, in 1493 for instance) and entails specific consequences for the subsequent evolution of the other regions. The rapid development of Italian cities, which of course accounted for such crystallization, is in turn a result of the tributary mercantile expansion of the Arabo-Islamic region. However, it is because it operated within an outlying zone (feudal Europe) that this Italian expansion set fire to the grassland and accelerated the rate of evolution to the extent of creating in Europe a system that was qualitatively superior to that of the formerly more advanced societies. I have given (in Class and Nation) a detailed explanation of this conjuncture which establishes a link between the state's weakness and the establishment of an area of autonomy for a veritable new class - the middle class - to appear, then the state's alliance with the latter in order to go beyond the breaking up of the feudal system by creating a new absolutist and mercantilist state, and so on. The general consequence of the new crystallization of Europe (capitalist and no longer feudal) is obvious: it blocked the evolution of the other societies of the world, which were gradually marginalized in the new global system. Moreover, the capitalist crystallization of Europe brought about a specific hostility toward the Arabo-Islamic region. We recall at this juncture the observation I made earlier about the specific position of the Islamic world in the old system. In order to establish direct links with the rest of the world to its advantage, Europe had to break the indispensable monopolistic and intermediary position enjoyed by the Islamic world. Ever since the early attempt of the Crusades, which was followed immediately by the establishment of the link between Europe and China that was opened by the Mongolian peace during the era of Genghis Khan, this hostility has been pursued to date and has found expression in a particularly neurotic attitude toward Muslims and generated in turn a similar response from the opposite direction. It is finally to break up this inevitable intermediate zone that Europeans set off on the seas. Contrary to Pirenne's thesis, such a choice was not the result of some geographical determinism.

Fourthly, the remarks made concerning these 2,000 years are not valid for the previous periods: on the one hand, the civilized societies known during previous periods - a fortiori the barbarians - were sometimes organized in a manner that was different from those of the subsequent tributary period; on the other hand, the network of relations that they ^1 engaged in among themselves was also different from the one illustrated ' with Figure 8.1 and Table 8.3.

Certainly, our scientific knowledge of the past becomes even less as we recede further in time. Nevertheless, it seems to me that two lines of thought relating to the "pretributary" eras can be distinguished (two philosophies of history). Pirenne's theory - which on this basic point is similar to the points of view defended by Andre Gunder Frank and Barry K. Gills - does not recognize any qualitative break around 300 bc, either around the Christian era or from the end of the Roman empire (the end of Antiquity, according to contemporary textbooks), just as it does not recognize any qualitative break separating "modem times" from "ancient times." Indeed, as I already mentioned, according to Pirenne, all periods of human history are marked by the same contrast between open, maritime, and "capitalistic" societies and closed, landlocked, and "feudal" societies. Moreover, like Frank and Gills, Pirenne emphasizes the exchange relations that existed among the societies at all times, irrespective of the distance separating them (for example, on the exchanges among Sumer, the Indus civilization, Egypt, Crete, Phoenicia, and Greece). Like Frank and Gills, Pirenne's theory is based on a philosophy of linear history: the progress is quantitative and continuous, without any qualitative change; in the words of Gills and Frank, it is the "cumulation of accumulation." On the other hand, the commonly accepted theory of Marxism distinguishes three stages of civilization that are different in terms of quality: slavery, feudalism, and capitalism. I do not enter into this field of Marxology, to resolve the question of knowing whether this theory is really that of Marx (and of Engels) - and to what extent - or whether it is only that of the subsequent Marxian common understanding. In any case, this theory states that all the societies listed in Table 8.3 are "feudal" societies: for Europe, from the end of the Roman empire; for the Byzantine and Islamic Middle East, right from their constitutions; for India, since the installation of the Maurya dynasty; and for China, since the Han era. Previously, on the other hand, according to this theory, they must have passed through a phase of "slavery" whose obvious and indisputable existence would be exemplified by Greece and Rome. In my opinion people put forward by analogy a state of slavery in China (from the Shang to the Han), in India (the Indus and Aryan civilizations), in the Middle East (in Mesopotamia). The existence of slavery located elsewhere and later on in certain regions of Africa, produced by the disintegration of earlier forms of communal formations, proves - according to this theory - that the passage through slavery constitutes a general requirement.

I do not share this point of view (see Class and Nation) and have offered instead a theory according to which: 1) the general form of class society that succeeded the previous communal formations is that of the tributary society; 2) the feudal form is not the general rule but only the peripheral form of the tributary type; 3) various conditions determine the specific form of each tributary society (castes, estates of the feudal era in the European sense - Stande; peasant communities subjected to a state bureaucracy, etc.); 4) slavery is not a general requirement - it is absent from most of the landmarks of history (Egypt, India, China); it hardly undergoes any important development unless it is linkedto a commercial economy and is therefore found within ages that are very different from the point of view of the development of productive forces (Graeco-Roman slavery and slavery in America up to the nineteenth century). Are the periods before the "break of tributary societies" which is marked iTable 8.3 not then to be distinguished from the rest of the precapitalist history? For instance, Egypt in particular offers the example of a tributary society having practically nothing to do with slavery whose history begins 3,000 years before the crystallization of the Hellenistic era. Assyria, Babylon, Iran of the Achemenids and probably pre-Mauryan India and pre-Han China sometimes practiced slavery but this practice did not constitute the main form of exploitation of productive labor. Finally, according to my theory, a tributary society is not crystallized into its complete form until it produced a universal ideology - a religion based on universal values that go beyond the ideologies of kinship and country religions peculiar to the previous community stage. In this perspective, Zoroaster, Buddha, and Confucius announce the crystallization of the tributary society. Until then, I prefer to talk about "incubation" or even the "long transition from communal forms to the tributary form." This transition, which is perhaps relatively simple and rapid in China, is made more complicated in India as a result of the Aryan invasion that destroyed the Indus civilization. In the Middle East the diversity of the peoples and trajectories, as well as the mutual influence of one people by the other, compels us to consider the region as a "system." I place within this context the early maturing of Egypt into a tributary society, the distinctive mercantile nature of slavery in Greece, and therefore I give particular importance to the Hellenistic synthesis, the prelude to the Christian and Islamic revolutions which were to take over the unification of the region.

Does the intensity of the exchange relations among the societies of these distant eras make it possible to talk about a "system"? I doubt it, considering that the civilized societies, that is, those advanced in the transition to the tributary form, still remain islets in the ocean of worlds of communities. Even when they are parallel, the trajectories do not prove that the societies in question do constitute a system but establish only the validity of the general laws of evolution.


This chapter first appeared in 1991 as "The ancient world-systems versus the modern capitalist world-system," in Review 14 (3) (summer): 349-85.


Ahmad, Sadek Saad (1985) Tarikh Misr al IjtimM, Cairo. Amin, Samir (1978) The Arab Nation, London: Zed.

覧 (1980) Class and Nation, Historically and in the Current Crisis, New York: Monthly Review Press.

覧 (1989) Eurocentrism, New York: Monthly Review Press.

Favzy, Mansour (1990) L'Impasse du monde arahe, les racines historiques, Paris: L'Hannattan.

Frank, Andre Gunder (1990) "A theoretical introduction to 5,000 years of world system history," Review 13(2) (spring): 155-248.

Pirenne, Jacques (1948) Les Grand Courants de I'histoire universelle, 4 vols, Louvain.

Polanyi, Karl (1987) La liberta in una societa complesse, Milan: Boringheri.

Toynbee, Arnold (1962) A Study of History, 12 vols, Oxford: Oxford University



One world system or a succession of systems?

Janet Abu-Lughod

I've recently published a book on the world system in the thirteenth century, entitled Before European Hegemony. It was intended in pan as a corrective to Immanuel Wallerstein's work on the sixteenth-century et seq. world-system.2 My criticism was that Wallerstein, while creatively extending the work of other historians and correcting for some of their biases, had still accepted the main line of western historical scholarship: namely, that the "story" becomes interesting only with the "Rise of the West" after 1450.3

This, I contend, is much too late. Because his account begins essentially with the sixteenth century, Wallerstein tends to overemphasize the discontinuity between the new Eurocemered capitalist world economy that began to come into being then and the system of world-empires and world-economies that had preceded it. And what is less defensible, he refuses to "dignify" any pre-sixteenth-century patterns of global trade by applying the term "world-system" to them. Indeed, he defends reserving that term only for the modem world-system, with its capitalist structure.

In contradistinction, my position is that a very advanced world-system already existed by the second half of the thirteenth century, one that included almost all regions (only the "New World" was missing) that would be reintegrated in the sixteenth century. Indeed, nascent capitalism was present in various parts of that system, without actually succeeding in dominating all parts.4 However, it was a world-system that Europe had only recently joined and in which it played only a peripheral role. Furthermore, this earlier world-system was organized in a very different way from the one over which Europe would ultimately exercise hegemony. The major metatheoretical dilemma in my work was (a) to see elements of continuity and discontinuity between what I conceptualized as successive but linked world-system stages; and (b) to account for how and why the transition occurred when it did.5

Andre Gunder Frank and I are now having a friendly debate conducted by long-distance mails in which, I confess, he has been writing more regularly and voluminously than I have been answering. Like the earlier one I had by mail with Wallerstein when I was writing my book, this disagreement also has no resolution.6

Both debates have been over some "simple" (but ultimately unanswerable) questions:

1 Has there been only one world-system, the one that began with the sixteenth century?

2 Have there been several successive world-systems, each with a changing structure and its own set of hegemons?

3 Or has there been only a single world-system that has continued to evolve over the past 5,000 years

Wallerstein espouses the first position, I have taken the second, and Frank and Gills contend the third. The present volume is part of this debate.

Is this a real controversy or is it merely a frivolous debate of the kind in which academicians sometimes engage for the sake of selling more books? I hope not the latter. What I would like to explore here is not whether one answer is right and the others wrong; cleanly, there is. no right answer. Rather, what I want to do is challenge us to think about what can be gained, intellectually and in terms of a research agenda, from a strategy that emphasizes continuities, versus one that emphasizes discontinuities.

It might be useful, however, to distinguish two levels of the argument: one on the regional level, the other on the international. On a regional (or what I have called a subsystem) level, one can argue not only for continuity but even development and expansion of economic and cultural linkages, without having to assume that the international system itself exhibited such continuities. To put it another way, one might find that local patterns persist and even prosper, while, at the same time, acknowledging that the role of the local region in a wider system has undergone a real transformation. Such an approach might help to explain long-term consequences in a more fruitful way.

First, then, I would like to support the argument about persistence. This might best be illustrated by reference to the series of maps (based on McEvedy) that were prepared for and appear in Before European Hegemony (pp. 138-40, reproduced here as Figures 9.1, 9.2, and 9.3). If one examines only the right halves of the maps, one is struck by the remarkable continuity in trade routes and volume in *** discussion. But one can make an equally good case for systemic transformation. The left halves of the maps illustrate that the context within which this continuity existed was undergoing a radical expansion and restructuring, which made its meaning in the larger system more problematic.

The ediof this volume have asked me to set the historic stage by describing what was going on in the world in the thirteenth century. I shall do this, but I cannot resist moving beyond that time period to describe how the containing system drastically altered after the middle of the fourteenth century and especiallafter the early fifteenth century. It would be totally transformed by the sixteenth.


As can be seen from Figure 9.4, I have conceptualized the thirteenth-century world-system as one that stretched between north-western Europe and China at its geographic extremes and have hypothesized that it was internally organized into eight overlapping circuits of trade that connected three (or possibly four)7 core regions that were politically and culturally distinctive. While each of these core regions had one or more hegemons, no single subsystem exercised hegemony over the entire system. Rather, a rough and somewhat stable balance existed - not necessarily because of detente, but because, given the technological level of transport, as well as siynfiunt cultural-religious barriers, there were real limits ro span of control that fell far short of the entire system's scale.8

My book traces the processes whereby these subsystems were formed and gradually linked to adjacent ones in the centuries between, roughly, the eleventh through the opening decades of the fourteenth, when the peak of commercial integration was reached. By that time, high levels of surplus were being produced throughout the system, as evidenced, inter alia, by a cultural and artistic efflorescence that was remarkable for its level and extensiveness. My thesis is that this level was not only a symptom but a product of the connections that had been forged and were stimulating local economies throughout the system.

Some time after the opening decades of the fourteenth century, however, signs of decline were already evident - although it is hard to make a case that all of them were related to what was happening in the world-system. By the mid-fourteenth century, however, the case becomes clearer. Along the pathways that connected the various subregions a pandemic outbreak of bubonic plague occurred. It spread widely (see Figure 9.5), decimating populations along its path, shaking dynastic power bases, creating fissures and breaks within and between subsystems, and disturbing the modus vivendi that had, at its height, permitted almost frictionless trade and exchange.

One century later, one could observe these discontinuities very clearly. At the eastern extreme, where the plague evidently had originated, a process was set in motion that led first to the Ming Rebellion (1368), which overthrew the Yuan dynasty that for two centuries had unified China with Central Asia and facilitated trans-Eurasian land trade, and eventually (after 1430) to the withdrawal of the Chinese fleet (and its subsequent port rot) which had previously played so important a role in maintaining the eastern sea connections. Once again. Central Asia was poised in opposition to China - as it had traditionally been9 - and presented a barrier rather than (what I have called) a "frictionless medium" through which trade and exchange moved relatively freely. Even after the closing of the Central Asian frontier, however, the sea trade from the Red Sea to the east persisted, as we shall see. It was only after a later Ming policy shift that its final destinations - the ports of south-east China - were closed, which further reduced volume and viability.

Within the Middle East, the effects were no less dramatic. The Egyptian-based Mamluk system underwent a similar cycle.10 The apogee of that subsystem was achieved under the Ayyubids and their successors, the Bahri Mamluks, during which period both Crusader incursions were thrust out and Mongol threats repelled, albeit not without the loss of Baghdad. Especially under the long, if discontinuous, rule of Sultan an-Nasr Mohammed between 1293 and 1341, prosperity was great, thanks to the operations of the so-called-Karimi merchants who sustained and mediated the Mediterranean sea trade of the Italians, conducted the eastern sea trade via port entrepots along the Islamicized west coast of India, and finally reached a working arrangement with the newly Islamicized Ghazanids in Iraq, along the overland route of reopened trade.

This prosperous period peaked roughly in the first few decades of the fourteenth century, but it was short-lived. The plague hit the Middle Eastern region with particular virulence, and the eventual transformation of the Mamluk system under the Burji Mamluks (after 1381) may be seen as parallel to the Chinese changes. In Egypt the fifteenth century was a period of increasing inflation during which the currency was debased, and one in which the Mamluk state expanded its active role in controlling trade, monopolizing export crops, squeezing local producers, and intimidating external traders.

Finally, the conquest of Constantinople by the Ottoman Turks in 1453 must be viewed in the context of a shift in the European subsystem, which had also been set in motion by events of the preceding century. The major European actors in the sea trade that linked northwestern Europe to the Middle East were the Genoese and Venetians, whose rivalry for sea control in the Mediterranean constituted a continuing plot-line in how the post-twelfth-century world-system was organized.

Between 1204 and 1261, Venice had been the major power in Constantinople and thus the guardian (and major beneficiary) of the northern gateway to Central Asian trade. The fall of the Latin kingdom in the latter year, or rather the restoration of Byzantine rule with the assistance of Venice's arch rival, Genoa, led to a redirection of Venetian trade via Egypt. It was a partnership that would benefit both for decades to come.

But Genoa's relative exclusion from the system was not without its eventual effect. By the last decade of the thirteenth century it .had already turned its attention to the Atlantic, effecting a sea link with Flanders that bypassed central France and making forays down the western coast of Africa along a path that would eventually be opened by Portugal, Genoa's new rival in the Atlantic.

In the mid-fourteenth century, however, both Venice and Genoa were hard hit by plague mortalities. Thanks to their sea connections with the Black Sea, from which the plague spread from Mongols to Europeans, the two port city states suffered proportionately greater mortalities than any other parts of Europe that were more peripherally situated. This led to a mid-century depression in both- cities, from which Venice eventually recovered, but Genoa did not." Indeed, the period 1378-84 marks the substantial defeat of Genoa in the Mediterranean rivalry between the two powers, although the final coup de greice was not administered until the end of the fourteenth century.

The relevance for Middle Eastern developments, and particularly for the subsequent rise of Ottoman power, should now be clear. In the course of the thirteenth and early" fourteenth centuries, a division of labor, or a modus vivendi, had been worked out. Genoa gained pribrity over the northern land route via its preferred status in Constantinople and its dominance over the Black Sea and the trading ports on it. (Genoese traders also benefitted from their role as providers of new recruits to the Mamluk dynasty centered on Egypt.) The middle route, which went overland from Palestine to the Persian Gulf, underwent a severe decline as an attractive alternative, after the last Crusader kingdom was eliminated towards the end of the thirteenth century (the so-called "fall" of Acre in 1291). This left, as the major rival to the Black Sea-overland route to China, the southern route, in which, thanks to the increasingly close symbiosis between the Mamluks and the Venetian traders, the Venetians were becoming more important in the sea trade with the farther east. The Egyptians still dominated that trade via their monopoly over the Red Sea route.

With the reduction in trade over the northern land route, by the plague and then the break-up of the Mongol empire that had unified Central Asia with China, the Genoese were no longer in a strong position vis-a-vis their rival, Venice. Indeed, there are no records of any Genoese traders in China after 1340. Thus, Genoa was weakened economically and, consequently, militarily. Texpansion of the Ottoman Turks into (newly named) Istanbul was in part the fruit of Genoa's final defeat in the Mediterranean.

This left the southern, mostly sea, route as the only one to which Europeans (largely through Venetian traders) had access. While throughout the fifteenth century this route continued to prosper, the peculiar alliance between the Mamluks, who controled access to eastern markets and suppliers, and the Venetians, who transshipped most of the eastern goods to European markets, was controling a larger share of a declining amount of world trade. This monopoly was finally broken by the Portuguese, beginning in the early sixteenth century.


This is the final argument in my book. In it, I contend that the entire Indian Ocean arena lay open to foreign "conquest" for two reasons. First, patterns of nonhegemonic trade had prevailed for many centuries in that arena.13 Multiple naval powers not only shared the trade, but even carried each others' goods and merchants. This meant that the powers involved in the Indian Ocean trading system had absolutely no preparation to resist the Portuguese incursion into their waters in the early decades of the sixteenth century. The second reason, of course, was the prior Chinese withdrawal from the sea and the rotting in port of its former navy.14 Since the Chinese fleet was the only force that (earlier) could have marshaled sufficient strength to offer resistance to the Portuguese, the latter's ability to "skim" the surplus from the continuing sea trade could not be prevented.15

The successful conquest of the Mamluk empire by the Ottoman Turks, which was roughly contemporaneous with Portuguese expansion into the Indian Ocean, must in part be attributed to these changes in the larger system. When Egypt came under Ottoman rule in 4516, the ease with which the former was defeated was not unrelated to the setbacks it had earlier experienced in the eastern trade.

The rest is, as they say, history. By the latter part of the sixteenth century, not only had the Ottoman fleet been defeated by the Venetians (in the battle of Lepanto in 1571) but they ceased any pretensions to remaining a sea power. After that, the Mediterranean was not a Muslim "lake." Furthermore, the major arena of the world system had begun to shift to the Atlantic. Braudel documents the eclipse of the Mediterranean in his two-volume work on the Mediterranean in the age of Philip II,16 while there is voluminous work on the growing importance of the so-called "New World." There is no need to document this.

However, these naval defeats did not mean that Ottoman power declined, nor did they mean a break in the continuity of the land system that connected Anatolia with south-eastern Europe (notably the Balkans) or with northern India and beyond. But they did mean that thereafter the Ottoman strength was to be over land. The importance of Turks and Arabs in the sea trade of both the Mediterranean and the Indian Ocean was at an end.


None of this global analysis implies that regional subsystems disappeared, or even declined, if measured in absolute terms. In this sense, Gunder Frank is correct to speak of one long march, rather than a set of equal cycles. My own metaphor is one of a very long up-cycle with fluctuations that at times are so extreme that it is analytically useful-to speak of "breaks" and restructuring. There is, then, no necessary contradiction between seeing the persistence and even improvement in economic activities over time within a given region and seeing that this region was falling increasingly below the average change for the system or the exponentially increasing shift in a region that, due to restructuring, was far outdistancing the subregion in question.

It is not enough to fight the stereotype of decay in the Ottoman-north-India region, because, despite its prominence in historical discussions, it was just not true. At least, my readings of the serious work that has been done by scholars of the seventeenth century on that region is a sufficient refutation of the stereotype.

I would suggest, however, that we need to pay more attention to the changed role of the region in question in terms of the larger system. In the final analysis, although the Venice-Cairo axis continued to operate down through Ottoman times, and the rumors of the demise of the spice trade were exaggerated, the fact is that the context of this trade, had undeniably altered. The Venetian fleet did vanquish the Ottomans. The world-system did eventually restructure away from the Mediterranean and the sea powers that controled it. The real arena did move outward to the Atlantic and the Atlantic rim nations of Portugal and Spain, before shifting to north-western Europe. The fact is that the axis of Central Asia, Anatolia, northern India, and the Levant-Egypt - an axis of central importance in earlier times which was scarcely destroyed by the seventeenth century -never again occupied the center stage of the world-system. I urge study of not only the continuities at the subsystem level, but also the discontinuities most evident at the large scale.


This chapter first appeared as a manuscript in 1990 as "Discontinuities and persistence: one world system or a succession^of systems?" from the New School of Social Research in New York.

1 See Before European Hegemony: The World System a.d. 1250-1350, New York:

Oxford University Press, 1989.

2 Particularly Immanuel Wallerstein, The Modem World-System, vol. 1, New York; Academic Press, 1974.

3 The term is, of course, drawn from the title of William McNeill's famous book, The Rise of the West: A History of the Human Community, Chicago: University of Chicago Press, 1963, even though the lengthy text of his book stresses the earlier and continuing importance of the East.

4 Contention over the meaning of "capitalism," and over when it began (whether in the thirteenth century or the sixteenth), has been going on for a very long time, and I have no wish to enter that debate in this article. It is, however, amply covered in my book.

5 See my "Restructuring the premodern world system," in Review 8 (spring 1990): 273-85.

6 See his review essay of Before European Hegemony in Journal of World History 1 (2) (fall 1990): 249-56, in which he sets forth his own views. He has been sending me materials and book outlines for the past year.

7 At the minimum, the cores were north-western Europe, centered on the triangular axis of Flanders, central Prance, and northern Italy, the Middle East, radiating from Baghdad and then Cairo, and China, along the axis of the Grand Canal that connected the Yellow River with the Yangtze. The intervening zone - including the Indian subcontinent and the East Indian archipelago - is harder to conceive of as a "core" since its mixed character and shifting limits often fragmented it into parts that were within the orbits of separate cores. I would not argue strongly against viewing this central region as a core, so long as political factors were excluded.

8 In the thirteenth century these limits had expanded considerably over those that prevailed in the two preceding world-system formations: the first which centered on the Mesopotamian-Indus Valley connection several millennia before Christ, the second which connected the western Mediterranean with the west coast of the Indian subcontinent in the centuries just before and after the start of the Christian era. The limits of the second era were somewhat widened during the early centuries of Islam, eventually expanding to encompass those in force in the thirteenth century.

9 Barfield's book is significant for our purposes because itindicates that the thirteenth century Yuan period was the only exception to the long history in which nomads from Central Asia were pitted against the settled population of China. See Thomas J. Barfield, The Perilous Frontier: Nomadic Empire and China, Cambridge, MA: Basil Blackwell, 1989.

10 These vicissitudes can best be traced through stuthe capital city of that empire, Cairo, whose growth and decline sensitively reflected Mamluk fortunes. See my Cairo: 1001 Years of the City Victorious, Princeton; Princeton University Press, 1971.

11 See, among others, B.Z. Kedar, Merchants in Crisis: Genoese and Venetian Men of Affairs and the Fourteenth-Century Depression, New Haven: Yale University Press, 1976.

12 See my "Did the West rise or did the East fall? Some reflections on the thirteenth century," presented to the annual meetings of the American Sociological Association in Chicaco in summer 1987 and subsequently printed as Working Paper No. 50, New York: New School for Social Research, Center for Studies of Social Change, 1987.

13 In this evaluation I depend heavily on (he prior work by K.N. Chaudhuri, Trade and Civilisation in the Indian Ocean: An Economic History from the Rise of Islam to 1750, Cambridge: Cambridge University Press, 1985.

14 The Ming withdrawal from the sea is a subject of considerable debate in the field. While scholars disagree on why, they do agree on when. The scuttling of *****, the Chinese fleet did not occur as soon as the Yuan dynasty was overthrown. In fact. Admiral Cheng-Ho's "treasure ships" continued their voyages up through the early 1430s in a remarkable show of force that carried them to the Persian Gulf, as well as to all important intermediary ports. Nevertheless, the Chinese navy was left to rot in ports after that, which meant that they could offer no resistance to Portuguese men-of-war when they arrived some seventy years later.

15 A small fleet mounted jointly by the Mamluks and the Indian Muslim rulers to protect their control over the Arabian Sea between them was easily defeated near Diu in the first decade of the sixteenth century by far superior Portuguese armed ships.

16 See Femand Braudel, The Mediterranean and the Mediterranean World m the Age of Philip II, 2 vols. New York: Harper & Row, 1972.



A critique

Immanuel Wallerstein

I wrote my article "The West, capitalism and the modern world-system" ('Wallerstein 1992) before I read Andre Gunder Frank's and Barry K. Gills's various recent writings in which they insist that there was no historic transition from anything to capitalism (anywhere, and specifically not in sixteenth-century Europe) because whatever happened in Europe in the sixteenth century was simply a (cyclical?) shift within the framework of an already existing "world system," which has existed (for Frank and Gills) for several thousand years. Frank and Gills refer primarily to a geographic zone, called by some the oikumene, which goes from eastern Asia to western Europe and southward to include at least south Asia, south-west Asia, and northern Africa.

This is an interesting and important thesis, but its argument is directed at me only to the degree that it is directed at anyone and everyone who does not wish to "abandon [the sacrosanct belief in] capitalism as a distinct mode of production and separate system" - apparently so large a group that it includes (dixit various acknowledements) even the "friends" whom they have asked to make "reflective comments" on their papers.

My paper was written not at all contra Andre Gunder Frank and Barry K. Gills but rather contra, all those - from Maurice Dobb to E.L. Jones to W.W. Rostow - who believe two things simultaneously: (a) something distinctive occurred in (western) Europe which was radically new somewhere in early modern times; (b) this "something" was a highly positive or "progressive" happening in world history. My position is that (a) was true but that (b) was distinctly not true.

I shall not repeat the detailed argument of my previous paper. But permit me to spell out the logic of my presentation there. Basically, the paper has two parts. First, I sought to establish that most of the traditional ways of distinguishing capitalism from other previous historical systems used weak distinctions in that they did not hold up under the light of empirical investigations. These traditional differentiae spedficae included extensive commodity production, profit-seeking enterprises, wage labor, and a high level of technology. I called all these elements "protocapitalism" since, without them as a part of the whole, one couldn't have capitalism. But I argue their presence was not enough to call a historical system a capitalist system.

They were not enough because, I argued, each time the agents who used these elements seemed as if they might be able to go further and create a true capitalist system, they were repressed or destroyed in one way or another. And what then distinguishes a self-sustaining long-lived capitalist system, I asked? To which my answer was that the differentiae specificae was, and was only, that the system was based on a structural priority given and sustained for the ceaseless accumulation of capital. Not, I insist, merely for the accumulation of capital, but for the ceaseless accumulation of capital.

It is my view that such a system was created, initially in Europe in the sixteenth century, and then expanded to cover the entire world. It is my view also that no historical system that ever existed before can be plausibly seen as operating on the principle of structural priority to the ceaseless accumulation of capital.

I made this argument not (I remind readers) in order to counter Frank and Gills but in order to counter all those who regarded such a transformation as a progressive "miracle." That is what brought me to the second half of my article - the attempt to account for the peculiar weakness(es) of western Europe that it permitted such a disaster to occur. I found the weakness in the implausible contemporaneity of four collapses - those of the seigniors, the states, the Church, and the Mongols.

Let me speak to the Mongols issue once again, since Frank reopens it in chapter 6 above. The importance of the Mongols is negative. My argument was that the three other "collapses" were not enough since one might have expected that they would have led, by occurring jointly, to the conquest of western Europe by an external power, which would have ended the possibility of the descent into capitalism. However, since the Mongols "collapsed," this led (through several intervening steps) to the momentary collapse of the world trading system of which Frank speaks, the weakening of its component sectors, and hence the impossibility for anyone to conquer western Europe at that particular moment in time. For one moment in historical space-time, the protective anticapitalist gates were opened up and capitalism "snuck in," to the loss of all of us.

Having restated my position on the "contra-miraculous" nature of the origins of capitalism as a historical system, let me briefly address Frank's own views. In his article (1990), he makes a case for the growth over thousands of years of an interrelated trade network that he calls the "world system." I believe in fact his account is a fairly acceptable initial and partial outline of what had been happening in the world between 8000 bc (or so) up to 1500 ad. I agree that there were many major nodes of political-economic activity, which I prefer to call "world-empires," and that these world-empires entered into long-distance trade (often? regularly? this is still to be demonstrated) with each other. I agree too that these world-empires included in the trading network of the oikumene various zones that were not organized as "world-empires." I even agree that, as a consequence, there may have been some common economic rhythms between them.

However, I do not believe that this trading network at any point of time was based on an axial division of labor involving integrated production processes. And therefore, for me, by axiom they did not form a singhistorical system, since I use that term to mean precisely something based on an axial division of labor involving integrated production processes. Of course, we may all define terms as we wish. This is the definition I have found useful, since it is the only one that accounts for the lives of limited duration of all these various systems, and for ways in which they have functioned historically during their lives.

I do not believe that trade alone makes a system. I have tried on at least four occasions (Wallerstein 1973, 1976, 1989: ch. 3; Hopkins and Wall-erstein.1987) to spell out the distinction between trade in "luxury" goods and trade in "bulk" goods or "necessities," and to indicate the consequences of the distinction. Even if it is difficult on occasion to draw the line empirically between the two kinds of trade, I continue to believe the distinction to be key analytically. It permits us to distinguish trade within a historical system (primarily in "necessities") and trade between separate systems (primarily in "luxuries"). Because of the technology of transport before modern times and hence because of its high cost, "long-distance" trade had necessarily to be in low-bulk, high-profit goods, and these had to be "luxuries."

Note a detail in word usage that distinguishes Frank and Gills from me. They speak of a "world system." I speak of "world-systems." I use a hyphen; they do not. I use the plural; they do not. They use the singular because, for them, there is and has only been one world system through all of historical time and space. For me there have been very many world-systems. For example, I do not consider that what many historians call China or the Chinese empire has been one system. There have been a number of successive systems in the geographic zone called China. The Han rose and fell. The Tang or the Ming is not the same historical system, even if the geographic location, the outward form (a "world-empire"), and some cultural features were the same. The "modem world-system" (or the "capitalist world-economy") is merely one system among many. Its peculiar feature is that it has shown itself strong enough to destroy all others contemporaneous to it.

This brings us to the hyphen. My "world-system" is not a system "in the world" or "of the world." It is a system "that is a world." Hence the hyphen, since "world" is not an attribute of the system. Rather the two words together constitute a single concept. Frank and Gills's system is a world system in an attributive sense, in that it has been tending over time to cover the whole world. They cannot conceive of multiple "world-systems" coexisting on the planet. Yet until the nineteenth century, or so I contend, this has always been the case.

Far from being Eurocentric, my analysis ''. 封封oricizes" Europe. Europe is historically aberrant. In some ways this was a historical accident, not entirely Europe's fault. But, in any case, it is nothing about which Europe should boast. Perhaps Europe and the world will one day be cured of this terrible malady with which Europe (and through Europe the world) has been afflicted.

This brings us to the future. For that we have to return to a schematic view of the past. Thus far, I believe, we have had three historical eras on the planet earth. There was the period before 8-10,000 bc about which we still know very little. The world was probably composed of a large number of scattered minisystems.

Then, there was the period from 8-J.O,000 bc to circa 1500 ad. There were in this period multiple instances of coexisting historical systems (of the three main varieties; world-empires, world-economies, minisystems). None of them was "capitalist" in that none of them was based on the structural pressure for the ceaseless accumulation of capital. Gloria Deo! As I said, I do not disagree that, among many of the major "world-empires," there was a growing network of long-distance trade. And perhaps this "crowding together" accounts in part for the outbreak of the malady that is capitalism. I say perhaps, because I do not like the ideological implications of this. I prefer my explanation of a fortuitous simultaneity of events. The two modes of explanation are not necessarily incompatible one with the other.

The third period began circa 1500 ad. The aberrant system, our capitalist world-economy, proved aggressive, expansive, and efficacious. Within a few centuries it encompassed the globe. This is where we are today. I do not think it can last too much longer (for my arguments, see Wallerstein 1982). When its contradictions make it no longer able to function, there will be a bifurcation, whose outcome it is not possible to predict. This outcome, however, will be radically affected by small input, hence by our input. The world is neither continuing to inch forward to a perfect oikumene, as some might suggest, nor remaining in a relatively stable state of social imperfection. Just because our inadequate analyses based on nineteenth-century social science are now proving to have badly misled us does not mean we have to fall into a variant of eighteenth-century triumph of universal reason. Just because it is useful to probe more intelligently into the patterns1 of the pre-1500 era does not mean we may ignore the unpleasant and dramatic caesura that the creation of a capitalist world-economy imposed on the world. Only if we keep the caesura in mind will we remember that this historical system, like all historical systems, not only had a beginning (or genesis), but that it will have an end. And only then can we concentrate our attention on which kind of successor system we wish to construct.


This chapter first appeared in 1991 as "World system versus world-systems: a critique," in Critique of Anthropology 11 (2).


Frank, A.G. (1990) "A theoretical introduction to 5,000 years of world system history," Review 13 (2) (spring): 155-248. Hopkins, T.K. and Wallerstein, I. (1987) "Capitalism and the incorporation of new zones into the world-economy," Review 10 (5/6) (supplement): 763-79,

Wallerstein, I. (1973) "Africa in a capitalist world," Issue: A Journal of Africanist Opinion 3 (3) (fall): 1-11.

覧 (1976) "The three stages of African involvement in the world-economy," in P.C.W. Gutkind and I. Wallerstein (eds) Political Economy of Contemporary Africa, Beverly Hills, CA: Sage, 35-63.

覧 (1982) "Crisis as transition," in S. Amin, G. Arrighi, A.G. Frank, I. Wallerstein, Dynamics of Global Crisis, New York: Monthly Review Press; London: Macmillan, 11-54.

覧 (1989) The Modem World-System, vol. 3: The Second Great Expansion of the Capitalist World-Economy, 1730-1840s, San Diego, CA: Academic Press.

覧 (1992) "The West, capitalism, and the modem world-system," Review 15 (4) (fall): 561-619. Prepared as a chapter in J, Needham (ed.) Science and Civilization m China, vol. 7: The Social Background, pan 2, section 48: "Social and economic considerations" (forthcoming).



Andre Gunder frank and Barry K. Gills

In this book about history in the "old" Afro-Eurasian world system oikumene, the central debate is really about how to write a world (system) history. We editors/contributors view the world system primarily as a "world economy" or better as a "world political-economic system." It has been the "same" system by virtue of its real historical continuity and the persistence of its structural patterns and processes for at least 5,000 years. However, these patterns themselves promote change, especially through the constant competition among the participants. In our view therefore, we are dealing with the same world system over 5,000 years, even though it is not always the same.

The debate among the contributors has been primarily about continuity versus discontinuity in world history. There are two main positions in this "debate. One position is that political/ideological determination of the mode of production or social formation in world history before about 1500 ad and of ceaseless capital accumulation and economic determination (through the "law of value") at least in the modem capitalist world-system thereafter 'makes a sharp break or discontinuity between the pre-1500 and post-1500 periods. This first position is taken here by Amin and Wallerstein, who at least therein represent the nearly universally accepted received wisdom on this matter. The other position is that the capital accumulation did not begin or become 'ceaseless" only after 1500 ad, but has been the motorforce of the historical process throughout world system history. Therefore, there was no such sharp break between different "systems" around 1500. This second, still very small-minority, position is taken by us editors/contributors and by Ekholm and Friedman. Though Wilkinson does not emphasize it, his long-term analysis also seems closer to this second position. Abu-Lughod seems closer to the first and still dominant position, despite her reference to a "thirteenth-century world-system." In his contribution here, Wallerstein clarifies this difference between many historical world-systems (with a hyphen) and one world system (without a hyphen), while Abu-Lughod tries to take an intermediary position.

Wallerstein and other world-system theorists stress the material transcivilizational exchanges in the world system via the "division of labor," which in turn determines their approach to the issue of "incorporation" into the world(-)system. We and others (e.g. sociologists Chase-Dunn and Hall, political scientist Wilkinson, and Kohl and other archaeologists) prefer a framework based on center-periphery complexes and a larger hierarchy of such complexes. Whether center-periphery relationships are "developmental" or "underdevelopmental" (Kohl takes a position against underdevelopmental relations in the ancient world) is an important but in our view not a crucial debate.

The real debate/disagreement revolves around the question of what structures constitute a "system" or a "world(-)system" in particular. We contend that a hierarchy of center-periphery (and hinterland) complexes within the world system, in which surplus is being transferred between zones of the hierarchy, necessarily implies the existence of some form of an ("international" (though this is not the best term) division of labor. In our i view, Amin and Wallerstein continue in the footsteps of Polanyi and Finley land underestimate the importance of capital accumulation via trade and the market in the ancient world system. Therefore, they do not see participation in the system in the same way we do and look for the "incorporation" of peoples and their societies and economies into the world-system at a point when we see them as having long been part and parcel of the historical development of the world system.

The real dispute then is over the character of the "international" or world system division of labor - not over its very existence. It may be true that as time passes the world system division of labor becomes ever more integrated, and time and space become ever more "shortened" (in the long run at least - allowing for temporary historical "setbacks" in the process - particularly in crisis periods). Wallerstein stresses what in our view is only a particular modern phase in the development of this world system division of labor at a higher level of integration than may have generally prevailed earlier. In his contribution in chapter 7 above, Wilkinson is prepared to accept our world system framework, albeit as manifested through more connections than economic trade and political conflict. However, the dating of the formation of such a "Central world system" (as Chase-Dunn and Hall would have us term it to satisfy all of us) is later and its geographical extent is consistently smaller through the ages for Wilkinson (e.g. as displayed in the date maps that accompany some of his writings) than it is for us editors/contributors. The reason is that Wilkinson asks for more stringent, empirically verifiable criteria of system inclusion, in particular demonstrable strings of connecting entrepot cities between hegemonic centers. Yet Wilkinson also recognizes above, and even more so in his still unpublished "test" of our chapter-5 cycle datings, that his de facto operational criteria may be too stringent. For instance, he relies on Chandler's data on city sizes with a lower cutoff at 40,000 people. This restriction probably makes many cities escape the net, which would be there if we had readily available data for cities of say 10,000 population.

We have no objections to and indeed welcome such "tests" and refinements of our definition and identification of the world system. We see no reason why such procedures should contradict our own premises about the existence and importance of a world system division of labor much earlier and perhaps (though this remains to be empirically established) at a lower "intensity" of integration. To take a more contemporary example, no one can any longer reasonably deny that China or India are pan of the modern world system, though reference to China's "socialist system" might give this (false) impression. Their "internal" division of labor may be less intensely integrated into the world system wide division of labor than that of Singapore or South Korea, but China and India are not "out of it." We disagree with those, among whom Charles Tilly is prominent, who suggest that we set arbitrary (trade and/or other) percentage levels of integration or systemicity in our definition of world system. We still believe this would be neither necessary nor useful and could be very misleading. (Until the 1970s, foreign trade accounted for no more than 5 per cent of American exports or imports; but not for that was the United States "out" of the system!) Such arbitrary limits/requirements on the level of system integration could constitute a barrier to posing questions about world system developments/interactions/correlations/exchanges, which we regard as so important. We do not accept criteria that may amount to a projection of the prevailing conditions of the present onto the past. We reject this as a form of "now centrism." The evidence suggests to us that there has been a world system wide division of labor even in the distant past some 5,000 years ago. Its form does not necessarily have to be identical with the modern form. Why should it? The labor of the ancient lapis-lazuli miners of Afghanistan and the textile workers in urban Sumeria was surely interlinked in a "world" economic/system division of labor even in the fourth or third millennium BC. They were both in the same world economy and the productive labor of one was connected, though perhaps indirectly, to the labor of the other in one overall exchange nexus.

In her contribution in chapter 9 on the other hand, Abu-Lughod recognizes long continuities on a regional level, such as Anatolia to India, which she calls "subsystem levels." However, she declines to "assume that the international system itself exhibited such continuities." So what she so magisterially analyzed in her book Before European Hegemony: The World System a.d. 1250-1350 and the earlier Islamic "system" to which she refers in passing were each only one-shot deals for her. We editors/contributors stand by our argument with her that the system-wide "disorganization" and "failure" to which she refers should be recognized as being rather recurrent phases of a continuing "international" world system, whose continuity is far longer than any of her "regional subsystems." Indeed, is there not a contradiction in her own reference to a long-lasting "subsystem level" in the absence of a continuous world system?

We also find a contradiction in Abu-Lughod's qualification of her thirteenth-century world system first as one of booming economic expansion and then as economic crisis and decline. Both could be admissible if the boom appelation were only to the first pan of her 100-year period; but it was not confined to that. Or the decline and fall could have started only after 1300. However, by our reckoning, and also by Wallerstein's and others' at least for Europe, the declining B phase of a long cycle began around 1250, that is her starting date, and lasted until 1450. By that reckoning,Abu-Lughod's entire world-system and period were part of a long B phase of decline between two periods of A-phase expansion in the eleventh-twelfth centuries and again in the "long sixteenth century" from 1450 to 1600, both of which are also recognized by Wallerstein and others. In that case however, as we have already argued, especially in our chapters 5 and 6 above, hois it possible to deny world systemic and long-cyclical continuity over this period, if not a still longer one? This area is certainly open to future investigation and debate. Hopefully such further explorations of this issue will help clarify our respective positions and perhaps break new ground.

Another area for future refining of positions over similarities and differences in the world system is the form of hegemony. We editors retain the concept used by economic (Wallersteinian) and political (Modelskian) world-systems and international relations theorists (Keohane), historians (Kennedy), and others. However, we insist again that hegemonic dominance over the world system, not to say the entire globe, has only rarely been achieved, if ever. More common, in the modern world system as well as in the world system before 1500, have been a series of simultaneous regional but unstable, temporary hegemonic powers. We see these regional hegemons both as forming a "system" of hegemony and of being constituent parts of the "world system," whose own structure and dynamic is also expressed by these hegemonic powers as well as their rise, fall, and mutual rivalries and alliances. Their hegemony, however, is not only "political." It is also "economic" in the sense that they centralize and use to promote their own "development" the economic "surplus," which they derive from their at least in that sense "dependent" peripheries and even hinterlands. That, and not only "power for its own sake," is why they seek to expand, maintain, and defend their hegemony as far and as long as they can. In chapter 3 especially, we editors/contributors broached the idea of a "super-hegemon" who "super-accumulates" on a world system wide scale. However, Wilkinson says he can find no example of such super-accumulating super-hegemony in the world (system) before 1500; and he is doubtful even about the only two recognized instances of hegemony since then, the British in the nineteenth century and the American in the twentieth century. The Mongols in the thirteenth century, whom we put forward as a possibility in chapters 3 and 5, may well be a doubtful case. However, we contend per contra Wallerstein and Modelski, that the case for super-accumulating super-hegemony on a world system scale was greater for the thirteenth-century Mongols than for Portugal in the sixteenth century and Holland in the seventeenth century.

These issues in our debates about system continuity also intercede in some older, indeed "classical," debates. In the classical Marxist view, capitalism was progressive. Capitalism developed the means of production and opened up new areas of the world to the concomitant process of capital accumulation in the "world capitalist system." Bill Warren (1980) among others also defended this position against writers who had begun to question it explicitly or implicitly. Baran, Sweezy, Frank, Amin, and others had begun to argue that capitalism as a world-historical process also generated the "development of underdevelopment" in the periphery and therefore was not progressive, at least for most of its people. Amin and Wallerstein have argued that most people in the modern world(-system) are now worse off than their forefathers and mothers were in premodern times. Wallerstein and other world-system theorists also argued that the semiperiphery represented an avenue of mobility within the world system hierarchy that also provided for the rise and decline of hegemonic powers.

On these issues, the present contributors largely do agree. None of us can support Marx's (or rather Stalin's) unilinear view of capitalist agency in world history. However, we can retain and seek to refine the basic perspective of historical materialism, which of course was neither original nor exclusive to Marx. However, the contributors differ in the role they assign to the capitalist mode of production or the supposed transitions between them in world history. On this issue, Amin and Wallerstein remain in the classical Marxist tradition; and we editors, Ekholm and Friedman, and Wilkinson do not. Why? Because we find too many big patterns in world history that seem to transcend or persist despite all apparent changes among modes of production and supposed transitions between them. The evidence available to us (notwithstanding our ignorance in comparison to specialists) suggests to us that there has been a profound misunderstanding of the character of modes of production. In particular there has been widespread underappreciation or underestimation of the role of capital accumulation, markets, the profit motive, "entrepreneurial elements," and of long-distance trade for most of world history.

Therefore, also, as contributors to this debate, we editors seek to extend the insights of dependency and world system theory much further back. We argue that the world system is rather like a giant and never-ending game of musical chairs. This "game" is not child's play but is based on incessant "rat-race" and "devil-take-the-hindmost" competition among the "players." The driving force behind this game is competitive capital accumulation, whether by state or by private elites, and usually by both. In the course of this competitive "game," particular "players" and areas change position within the world system from one time to another. This change is particularly accelerated, if not generated, during periods of system-wide economic crisis and the accompanying hegemonial shifts when "the music stops" or at least slows down. Peripheral (regional and political-economic) positions within" the world system act as a mechanism of exploitation via transfer of surplus to the distant center(s) and the subordination of the periphery's "development" to the requirements of the center's "development." Some "development" may take place even in the periphery; but this is not the real point, at least over the long pull. The important issue is the position in the world system, and how, when, why, and where that position does or does not change and permit or deny -and literally in any case unevenly distribute - the benefits (and costs) of development in and of the world system as a whole.

The outcome of "development" in any particular part of the system is pan and parcel of the prevailing conditions of "development" (especially capital accumulation) of the whole world system. This seems to be a general law of all world system history. This game of musical chairs is literally as old as the hills and as old as the world system in any event, and it continues right up to the present. The recent collapse of the Soviet Union and the regimes in eastern Europe and their subordination to the International Monetary Fund and the Group of 7 ( de facto G3) is a case in point. The attempt to understand these events as a transition between modes of production, let alone as a change of "system," can only confound much more than it clarifies. Instead, these dramatic events should be seen as the outcome of these regions' and regimes' inability to compete adequately in the accentuated rivalry during and because of the present period of world-economic crisis within the world system as a whole. Reaganomic "star-wars" bankrupted both "super-powers," although the USA has been able to paper over the "twin deficits" in its budget and balance of payments through capital shots in the arm by its allies (but even more its competitors) in Europe and Japan, which have bailed the United States out (so far but for how long?). We are not witnessing a "reincorporation" of the "socialist East" into the world economy run along the canons of the "free market/capitalist" West. On the contrary, the East - no less than the South, of which most of it had been and now again will be a part - had in a disadvantaged position in this same world system all along; and that is the principal reason why it failed.

We editors/contributors view this change in position in the world system as part and parcel of change in the overall development of the world system itself and as the result of its governing motor-force process of competitive capital accumulation and hegemony-rivalry. Tour position reaffirms the earlier break by Frank, Wallerstein, and others with the then dominant state-centric framework and unilinear modernization theory. We reaffirm the correctness of the shift to the whole world system as the essential unit of analysis. However, we seek to carry it further in space, time, and analysis than Wallerstein and other "world-system" theorists. This is not to say that the world system simply "determines" everything that can or cannot happen or the "chances of mobility" of any regional, social, or individual "player" within the system. However, the whole is more than the sum of its pans, and no pan can ever be understood apart from the whole system, which helps determine if not its fate, at least its choices or lack of them.

Therefore, we argue that it is not the mode of production which determines the overall developmental patterns and outcomes of this game - but the nature of the game itself, of which the various modes are (only) an element. The search for any supposed "transitions" between such "modes" only obscures the essential continuity of participation in the same one and only world system. Soviet "socialism" was an element in the game after 1917, but it never determined the nature or pattern of the game. Similarly, the whole period of world history since 1500 has been less "defined" by "capitalism" than it was generally defined and characterized by shifts in the routes of trade, centers of accumulation, and the location of hegemonic power from "East" to "West." The now emerging period of world history is again, or rather still, characterized by competition between centers of accumulation, which may possibly even be accompanied by a continuous westward shift of the center of accumulation back to Asia in "the East," if there is not a (temporary) "breakdown" of the world economy into rival political-economic zones.

We are arguing that the nature and rules of the game do not change so much as the players change position. Techniques of competition change, but it is also true that many basic techniques have been around for millennia. However, the world system process of competitive capital accumulation goes on and remains the ultimately determining process in world system development - and in the (temporary) costs and benefits, which panicular peoples and regions derive therefrom, mostly on the basis of their (also temporary) under/privileged, competitive, or monopoly/oligopoly position in that same world system. This is what we mean by the continuity of and in the world system, by comparison with which the discontinuities are only minor from a longer world-historical (system) perspective.

The point on which we cannot agree with either Amin or Wallerstein concerning discontinuity is the role of capital accumulation in world history. We affirm Ekholm and Friedman's challenge to the Polanyi and Finley ortodoxy on capital in the ancient economies. The evidence suggests to us (overwhelmingly) that there was no sharp break pre- and post-1500 in the predominance or even in the supposed nodes of capital accumulation in the world system. What there was was a dramatic and important shift in what McNeill refers to in his foreword as the communications network of the world system. That is, pre- and post-1500 there was a historic shift from trade centering on the three corridors in the east Mediterranean, west and Central Asia to transatlantic linkages, which however did not really replace the former until the eighteenth century (not the sixteenth!). Within this shift there was the "windfall" profit to European centers of accumulation of the conquest and incorporation of the Americas as a periphery to the rising European core, thus bolstering its new world system position. The surplus and especially the bullion extracted from the Americas (a la Blaut) was injected into global exchange circuits in Europe's favor while also stimulating within Europe what is usually called the "rise of capitalism."

The key in these changes is a shift in locus of accumulation in the world system accompanied by a shift in hegemonic power and the global reorganization of centers and peripheries. These shifts were part of the underlying changes in competitive position in the world accumulation process, but they are much older than "modern" "capitalism" in the world system. Polanyi et al. were mistaken about the supposed absence of capital accumulation in the ancient economy. We believe that Amin and Wall-erstein are still mistaken about the in/'significance they accord to capital accumulation in pre-1500 world history, even though both admit the existence of capital accumulation before 1500.

This question of continuity/similarity versus discontinuity/differences in the debate is related to other patterns and issues about which the contributions to this book contend as well. Among these issues are the following:

1 Long cycles of expansion and contraction in the world economy.

2 Hegemonic world system cycles of simultaneous rise and decline.

3 Large-scale periodic migrations and invasions during crisis periods.

4 Ideological confusion and fragmentation during crisis periods.

5 The increase of repression, heightened class conflict and greater incidence of war and general destruction during crisis periods.

6 Shifts and the rise of new centers of accumulation during crisis periods.

7 The economic, social, and political retrogression of some regions during crisis periods (so-called "dark ages"), which involve less well-developed connections to the centers of world system economy and "reversion" to locally more self-reliant economies.

8 The eventual re-establishment of a renewed exchange and communications network, renewed economic growth, and new hegemonic structures.

9 Shifts in centers of production and accumulation, related to shifts in trade routes and urban and demographic settlement patterns in the world system in periods of renewed expansion.

10 New urban growth and larger demographic growth and the territorial expansion of the world system as a whole in periods of expansion.

11 Periods of economic expansion, accompanied by growing trade at long distances over both land and maritime routes, higher frequency of diplomatic contact, cosmopolitan exchange of ideas and technologies, consolidation of political systems and hegemonies, all of which contribute to higher rates of capital accumulation.

12 Eventual overconcentration of capital, which leads to "overaccumulation" and "underconsumption," growing gaps between rich and poor, overextension of the state apparatus, falling rates of investment and profit, slowdown of expansion, and contraction of economic growth.

All of these patterns may be associated with "capitalism" since 1800 or the "modern world-system" since 1500. However, inspection of the evidence shows that they were equally present and significant over at least 5,000 years of world system history.

Albert Bergesen (1992) suggests that neither the "pre-" nor the "post-" 1500 camp in this debate has yet made a breakthrough to a new theory of world system history. Perhaps not. However, we suggest that the extension of post-1500 insights to the long pre-1500 period may indeed constitute the kernel of a new theory of world system history and of its center-periphery structure and governing processes of capital accumulation and hegemony-rivalry. This one world system process develops, underdevelops, brings "progress," peace, prosperity and also war, destruction and depression in its wake. This is the stuff of all world (system) history. The challenge for world system theorists is not only to rewrite all world history in this light, but also to help guide at least some of the more destructive processes of world system history into socially mobenign if not beneficial channels.

This new departure in world history writing leads us "full circle" back to the issue of ecology. The world system's origins lie in ecology, as does its entire developmental history and of course its future. We agree with McNeill, writing in the foreword of this book, that one of the next steps in the new rewriting of world history must be a new understanding of hworld system development both altered and was in turn altered by the natural environment. The natural environment places limits on and establishes parameters for world system development. Ecological crisis and perhaps even ecological cycles have accompanied world system development in the past. The now looming ecological crisis reminds us of the urgency of this emphasis on ecology in world history.

Thus, this is not an arcane debate about ancient history. The disagreement goes to the nub of a larger theoretical argument over how best to study the modem world (system). This debate is at the fulcrum of historical-materialist, including Marxist, analysis of world history. The questions we debate above must be answered satisfactorily in order better to be able to address pressing political questions about "what is to be done." Both "camps" in this debate (perhaps surprisingly to some) share the same good political intentions. Namely, how to change the world for the better in the interests of its people and not just its present or future rulers. These political implications may explain why some of the disagreements debated in this book sometimes arouse such strong feelings: they are related to the question "what is to be done?" Three of the present contributors, Amin, Frank, and Wallerstein, along with Giovanni Arrighi and Marta Fuentes, recently addressed this question jointly under the title Transforming the Revolution: Social Movements and the World-System (1990).

Thus McNeill may be right (though excessively generous in his foreword) that our debate may be a potential turning point in the writing of world history. McNeill says that whether or not the "5,000-year world system" framework (which he suggests is a "confluence of Marxism with more inchoate liberal ideas about world history") will constitute a "landmark" or not depends partly on the future of Marxism and partly on the future of world-history writing as a professional undertaking. The key to this, and again we agree with McNeill, is how material processes are treated in the writing of world history. We are very adamant that capital accumulation, as well as its long cycles of world system wide economic expansion and contraction, merits much greater recognition in world historical analysis and writing.

If this point is recognized, as McNeill says, the debate about the unit of analysis takes on a very different and less central aspect. Traditionally, the unit has been framed as "nation-state" "societies" or "civilizations" and more recently "world-systems." Following McNeill, material transcivilizational processes in the entire Afro-Eurasian oikumene (or world system) would become much more central to the concern of world-historians and express the "unity in diversity" of humanity in world history. McNeill proposes that we emphasize the material transcivilizational exchanges in the "communications network," which act as the skeleton of the world system body. In that case, the "awkwardness of terminology" recedes in importance.

The "central question" is whether such material exchanges have really been the key determining process in world (system) history. We agree with McNeill about these material processes in the world system and that their recognition does not require or imply the exclusion of civilizational or ideational aspects of human experience. Rather, we suggest that due recognition of these long-term "material" structures and processes can provide a framework through which to analyze these "ideal" aspects in an organized way. We also accept that the world system develops in and through a dynamic communications network across which capital, surplus, commodities, peoples, armies, ideas, technologies, diseases, and more travel. Thus, McNeill's suggestions do not contradict and are entirely compatible with emerging world system history framework, premises, or hypotheses - and vice versa. His suggestions can help refine, develop, and elaborate our framework for the writing of world history as well. We can only look forward to this new joint enterprise and invite our supporters and critics alike to continue the debate over the ideas developed in this book. The editors' perspective on world system history extends and deepens me rejection of Eurocentrism that was made by earlier world-system theory. However, all the contributors in this book agree on arid participate in this important task. We affirm that in the future all world-history writing must be humanocentric and as objectively as possible assess the overall unity of human history while encompassing the diversity of its cultural expressions. No centrism based on the temporary historical "glory" of any nation or region should any longer be allowed to distort our universal human understanding of our one world history.


Amin, S., Arrighi, G., Frank, A. G., and Wallerstein, I. (1990) Transforming the Revolution: Social Movements and the World-System, New York: Monthly Review Press.

Bergeson, (1992) Political Economy of the World System (PEWS) Newsletter.

Warren, B. (1980) Imperialism: Pioneer of Capitalism, London: New Left Books and Verso.

Back to Contents