Studying thousands of years of world system history is an ambitious undertaking. Attempting to explain what happened over the long term in a non-descriptive fashion is even more ambitious. For that matter, 'merely' describing what took place in recorded history is no simple feat. Thus, it is fortunate that a number of social scientists have begun to tackle such projects. The undertaking will no doubt require a small army with each platoon chipping away at their version of history's reality. The armies are necessary because the task is imposing. It is also important. For, fundamentally, what each perspective on world system history shares is a commitment to the idea that contemporary structures and processes are embedded in a long-term, historically contingent context. To make sense of these contemporary structures and processes, it is necessary to appreciate how and why they have assumed their present form. In many cases, the present forms may not be much different from older forms. Indeed, a central question is to what extent have major structures and processes been characterized by continuity — and how far back in time does that continuity extend? Social science students of world system history are betting that the continuity extends back much farther than most people realize. In the 1970s, world system analysis was provocative and revolutionary because the argument was that we needed to encompass the last five hundred years in our models of socioeconomic and political behavior, and to do so from a systemic perspective. In the 1990s, the new assertion is that hundreds of years are no longer sufficient. Now, it is thousands of years that must be accommodated in our theories and analyses of systemic change.

But how should we best go about accommodating thousands of years? How many thousands of years are really necessary? How much difference is there in the different approaches that are currently available? The first question (how should we interpret world system history) ultimately requires a subjective answer, although it is certainly not beyond the reach of empirical testing. The second question (how many years) hinges on the answer to the first question. The appropriate temporal span of one's inquiry depends on what is thought to be important. The third question (how much difference is there among approaches) requires neither a subjective response nor is it a derivative of a set of assumptions. Moreover, it is important to pause from time to time in order to evaluate the distances separating approaches to similar phenomena. Are we converging or

[288]

diverging on our answers to what makes the world tick? To what extent may this convergence or divergence be attributed to a priori assumptions and analytical preferences?

A basic divide currently separates analytical approaches to world system history into two basic camps: either 1500 is a cardinal watershed or it is not. To make the comparison easier, I will focus here only on the schools of thought for whom 1500 is not a or the basic watershed. While more approaches are sure to develop or are being developed, there currently appear to be four major paths to explaining the very long-term development of the world system. To avoid overly personalizing the work going on in the various schools of thought, I assign relatively neutral labels based on some distinctive aspect of what they emphasize. The four approaches are: the continuing world system (Frank and Gills), the comparative study of world systems (Chase-Dunn and Hall, the engulfing world system (Wilkinson), and the evolving world system (Modelski and Thompson).' Each approach deserves a brief overview prior to any attempt to compare the four. To facilitate comparison, the overviews focus selectively on ideas about systemic origins, division of labor, hegemony/leadership, long-term economic foci, and affinities for evidence and data analysis.

System origins The world system originated in the overlapping intersection of the spheres of influence of Sumer, Egypt, and northern India (2700-2400 BC). None of the early agrarian states were self sufficient. In order to acquire resources that were not available locally, a combination of division of labor subordination and participation in long-distance trade was necessary. The expansion of several states simultaneously increased their tendencies toward competition and conflict over control of the sources of supply, and the routes used to deliver desired commodities. The basic political dynamic thus was one of perpetual rivalry among competitive units interested in safeguarding their individual abilities to accumulate surplus.

Division of labor Center-periphery-hinterland - the center extracts surplus from a subordinated periphery and a less subordinated hinterland.

Hegemony/leadership Hegemony involves the concentration of privilege in one or more center predicated on the ability to accumulate surplus more effectively and at the expense of other zones of the world system. A condition of super-hegemony prevails to the extent that one center establishes itself at the apex of an overarching hierarchy of smaller-scale hierarchies. Possible candidates include India, China and the Abbasids before 1500. Post-1500 candidates are 'Iberia,' the Netherlands, Britain, the United States and perhaps Japan (in the twenty-first century). However, the idea of intersecting hegemonies (as in the sixteenth century case of Portugal, the Habsburgs, Ottomans, Mughals, and Ming China) seems to be

[289]

preferred to the concept of a sequence of singular hegemons. Similarly, the idea of hegemonic transition is seen as occurring in several places at about the same time.

Long-term economic foci Capital accumulation has been a constant concern of world system actors. However, accumulation activities are characterized by A (expansion) and B (contraction) phases lasting roughly two hundred or so years at least prior to the modern period. Evidently, these phases gradually became shorter and took the more familiar form of Kondratieff waves. Hegemonies form in A phases and decline in B phases. B phases are also characterized by decreased trade, increased conflict and war, and intensified intra-elite and class struggle.

Evidence advanced The continuing school has yet to emphasize the need to advance explicit evidence for their assertions however they have encouraged others to explore the accuracy of their generalizations.

System origins The basic driving forces behind the development of world-systems are multiple and interactive. Population growth leads to environmental degradation and increased demand for food and raw materials. One response is migration until or unless this becomes too costly due to resistance from environmental and human barriers (circumscription}. Two alternatives to migration are conflict over scarce resources and/or intensified production based on technological development, that may be accompanied by a more hierarchical organizational structure in order to reduce internal conflict. However, since technological development tends to create new types of scarcities that can be rectified by trade and/or conquest, the second alternative is unlikely to be conflict-free. Technological development may also encourage greater rates of population growth and consequent pressures for greater resources.

World-systems come in four types - kin-based, tributary, capitalist, and socialist or some other future organizational principle. Each type of system operates on a different accumulation logic but each generates four types of interaction networks through which the accumulation logics are pursued: bulk goods, prestige goods, political-military, and information. Whether a world-system can be said to exist depends on the extent to which its key reproduction and transformation processes are relatively autonomous. Kin-based world systems can be traced back 10-12,000 years but have gradually been absorbed by tributary and capitalist systems. While bulk-goods networks (BGN) have tended to be 'down-the-line' interactions, an Afro-Eurasian prestige-goods network (PGN) or super-system first came into existence with the interactions between the Roman and Han Chinese tributary systems (circa 200 BC). Since that point in time, the central super-system has expanded and contracted (a pulsation-decoupling process). It has also gradually changed, especially after the seventeenth century, from a tributary mode type of

[290]

system in which state coercion was dominant in surplus extraction to a capitalist mode system in which private entrepreneurs manage surplus extraction, assisted by their control of 'weak' capitalist city and national states.

Division of labor Core-semiperiphery-periphery - The core-periphery differentiation is a function of interaction between groups characterized by greater complexity and population density and lesser complexity/density. Whether the differentiation also develops into a hierarchical relationship in which the core dominates the periphery is an open question. The semiperiphery is an intermediate zone in terms of location and characteristics, combining features of the core and the periphery without assuming all of their liabilities. As a consequence, the semiperiphery has an unusually good structural position to exploit opportunities for upward mobility opened up by core decline and uneven development.

Hegemony/leadership Any hierarchical system will have dominant powers (hege-mons) by definition that rise and fall without exception. However, the manner in which hegemons rise and fall apparently varies with the type of world-system and its predominant accumulation mode. The Netherlands, Britain, and the United States are considered successive hegemons in the central super-system after the seventeenth century.

Long-term economic foci Different accumulation modes characterize different eras at the super-system level. Within any given era, multiple world-systems coexisting more or less autonomously may exhibit a variety of accumulation modes. Capitalistic practices have been around for a long time but only became prevalent in the seventeenth century. Emphasis is placed on the transition between modes. The spread of capitalism is attributed in part to the weakness of the tributary mode in Europe and the vigorous existence there ofsemiperipheral, capitalist city-states.

Evidence advanced The comparative school has so far emphasized the advancement of framework construction over the analysis of evidence although some examination of city size data has been conducted.

System origins Mesopotamia and Egypt formed a 'central civilization' around 1500 BC. A civilization or world system is an urban network that is militarily, politically, and geotechnically isolated from significant outsiders (who invac.;, conquer, or make alliances). The historical dynamic is basically one of a central civilization located where Asia and Africa come together gradually engulfing twelve other autonomous societies. However, coexisting with the engulfment process is a multicivilizational macroeconomy (the Old World oikumene) that is predicated on long-distance trade linkages. The origins of this world economy predate the creation of the central civilization by two millennia. Mesopotamia

[291]

and Egypt were linked by trade by the fourth millennia BC. India was linked after 200 BC. East Asia was linked after the early seventh century AD. Throughout most of recorded history, the scope of the world economy remains more spatially extensive than the boundaries of the central civilization until the twentieth century AD. In fact, an important dynamic is civilizational expansion in pursuit of greater control over the world economy.

Division of labor Core, semiperiphery, and periphery - the core of a civilization is its most powerful and wealthy center. The semiperiphery is poorer and weaker than the core but strongly connected to it. The periphery is only weakly connected to the core through trade. Core domination and exploitation of the semiperiphery is more likely to be attributable to politico-military coercion than to an economic division of labor.

Hegemony/leadership While the location of military, political, economic and cultural domination, and the identity of dominant powers may shift from one zone in the central civilization to another (or within civilizations), parahegemony has been virtually absent. Parahegemony is a situation in which one state within the world economy derives extraordinary economic benefits and privileges as a function of its leadership in invention, investment, or entrepreneurship. It must also be able to defend its advantageous position and/or outside the striking range of its rivals. Only nineteenth-century Britain and the United States for a brief period of time after World War II satisfy the criteria.

Long-term economic foci Economic organization fluctuates toward and away from capitalism/statism according to whether the system is characterized by many states that are relatively weak and small (toward capitalism) or by few states that are strong and large (toward statism). The world economy is always characterized by a mix of production modes.

Evidence advanced Arguments about assertions in the engulfing approach have so far relied heavily on city size data, measurements of polarity, and taxonomies of categories.

System origin The world system is conceptualized as a set of four, nested structures or networks of city-based interactions: economic, political, social, and cultural. Each network experiences structural evolution but not necessarily at the same time. Yet because the networks are nested, the pace of change is synchronized. In general, the world system has moved through three historical phases. An initial Bronze/Middle Eastern phase (3500 BC to 1000 BC) was followed by an Iron/Eurasian phase (1000 BC to AD 1000), which, in turn was superseded by a western/global phase (AD 1000 to 3000). A fourth, 'post-modern' phase is predicted to begin around AD 3000.

[292]

Division of labor Center-hinterland - centers are created by their lead in, and monopoly of, innovation. Hinterlands are alternatively passive and assertive and can conceivably reverse their variable dependency on the center. To what extent the center-hinterland differentiation leads to a dependency relationship must remain an open question. Note, however, that this differentiation need not be restricted to a material distinction. The conceptualization also emphasizes a dynamic of center concentration and deconcentration brought about by the periodic leveling activities of populations resident in the hinterland.

Hegemony/leadership Economic leadership is predicated on constituting the primary source of innovation in the system (the active zone). A continuous transition of economic active zones began after 1000 AD with sequential locations in Sung China, Genoa, Venice, Portugal, the Netherlands, Britain, and the United States. The evidence examined to date suggests that earlier economic active zones existed but not in a continuous sequence. Economic leadership, especially after 1000 AD, leads to leadership in naval power in order to protect maritime routes for long distance trade. To what extent economic and naval leadership leads to hierarchy and subordination outside the world economy remains an open question.

Long-term economic foci The Eurasian world economy has been characterized by long pulses of concentration and dispersal. Innovation, urbanization, and economic growth are initially highly concentrated activities that are diffused. Diffusion, in turn, facilitates the growth of trade. After the tenth century AD, successive Kondratieff waves (viewed as radical innovations in commerce and production) drive the evolution of the world system. Kondratieffs come in paired surges with a period of systemic crisis separating the first and second wave. The first wave tends to be located in a new zone and signals new economic leadership. The consequent conflict of a destabilized status quo in the systemic crisis period increases the probability of a second wave in the same active zone although not necessarily associated with the same leading sectors.

Evidence advanced Leading sectors have been measured more or less from the tenth century on. Naval concentration data have been generated for the post-1494 period. Population and city size data have been examined over a four to five thousand year range. Migration data have been used to periodize the center/ hinterland process.

The four schools of thought converge on the need to examine structures and processes prior to 1500. They agree more or less on specifying the origins of the central world system around Mesopotamian - Egyptian activity in the fourth and third millennia BC. They also agree on the general significance and early emergence of an Afro-Eurasian trading network, specializing initially in the exchange of 'luxury' commodities over long distances. If pressed, analysts associated with

[293]

most if not all of the perspectives would probably accept the idea of nested political, economic, military, and cultural networks. All four perspectives do emphasize the idea of urban networks as providing armatures for the world system's structure(s). Once we move beyond these basic appreciations, however, disagreements become more noticeable.

The division of labor question All four schools use similar vocabulary but the similar-sounding concepts do not always mean the same thing. The center/core distinctions are relatively convergent as long as one does not pursue very far the source of the center—periphery-hinterland differentiation. Yet, because the differentiation processes are not the same, it is unlikely that each school of thought would place the same parts of the world into the same categories at the same time. And even if considerable overlap did occur, it would not be entirely meaningful because the categories do not have the same meanings. Once one moves beyond the center/ core, the potential for categorical confusion increases exponentially. For instance, the continuing school's hinterland seems similar to the engulfing school's periphery while the comparative school's periphery seems to approximate the engulfing school's semiperiphery. The evolutionary school's hinterland presumably encompasses everybody else's semiperiphery, periphery, and hinterland, in addition to some of their core/center.

Perhaps even more important than the disarray of the labeling is the disagreement on whether the differentiation implies unidirectional exploitation and subordination. The continuing school assumes that this is the case. The comparative

Table 15.1 Schools of thought

[294]

and evolutionary schools regard it as an open question. However, one gathers that the comparative school would expect differentiation and hierarchy in most cases involving actors of unequal capability. Innovation in the evolutionary school's center is likely to create technological and commercial gradients that imply some type of dependency arrangement. Contrastingly, the engulfing school views center — semiperiphery subordination from a military coercion angle. Yet, to date, little work has been attempted on this question of variable subordination, dependency, and dependency reversal. One exception to this generalization is the agreement suggested by arguments associated with the comparative and evolutionary perspectives on the rise of Europe to centrality within the world economy - one of the more spectacular cases of dependency reversal. Both perspectives stress the significance of autonomous merchant republics surviving and thriving in circumstances that would have been difficult to replicate outside Europe.

The hegemony/leadership question Fundamental disagreements about the nature of politico-military, economic, and cultural power concentration and its implications have long plagued the analysis of world systems. All analysts converge on the analytical centrality of something like power concentration but there is far less convergence on how best to conceptualize it. The continuing school makes little distinction between land and maritime-based dominant powers. Thus Safavid Persia and Portugal can be hegemonic simultaneously but with different (yet overlapping?) domains. Concurrently, this school's ultimate position vis-a-vis 'super-hegemony' is ambiguous suggesting that a consistent preference for 'intersecting hegemonies' versus super-hegemons has yet to be worked out.

In contrast, the evolutionary school makes clearcut distinctions between land and seapowers but not solely because of their varying strategic orientations. Sea-power is both a key ingredient and byproduct of active zone dynamics and the concentration of innovation after the tenth century AD. As a consequence, the Sung dynasties were not fully comparable to the Ming dynasty in this sense, any more than the Ming dynasty was comparable to Venice, Portugal, or the Netherlands. Venice, Portugal, and the Netherlands were at the center of the active zones of their times. The Ming dynasty was not. On the other hand, the comparative school stresses the transition of capitalist logics over tributary logics in the seventeenth century. Thus each stage is characterized by a different set and type of dominant powers. Dominant powers in the tributary stage do not behave exactly as dominant powers in the capitalist stage.

The engulfing school's conceptualization of 'parahegemons' overlaps conceptually with the comparative school's emphasis on capitalist hegemons and the evolutionary school's stress on innovation leaders. The problem is that only two parahegemons can be found while three capitalist hegemons and eight or nine innovation leaders (depending on whether one counts Britain twice) are advanced.

The resulting problem is that any effort at generalization about hegemony/ leadership in the world system must be confined to one school of thought. Each

[295]

school of thought identifies a different population, summarized in Table 15.2, because each school of thought operates with a different conceptualization of what constitutes dominance/hegemony/leadership. There is overlap but probably not enough for cumulative, cross-school understandings.

The long-term economic foci question Disagreement on economic foci is particularly marked. In fact, several questions can be subsumed under this heading. One has to do with the question of capitalism. The continuing school of thought has it as continual. The comparative school of thought stresses that capitalistic practices only became predominant about 300 years ago. The other two approaches do not view this dispute as one of major interest.

The evolutionary school instead stresses the AD 1000 break-point which marked a shift toward the continuous sequence of innovation surge-systemic crisis-innovation surge. The continuing school sees no meaningful breakpoints. Behavior before AD 1000 and after are thought to be inherently similar. Breakpoints in the comparative school hinge on transitions to new production modes. The functional equivalent of breakpoints in the engulfing perspective may revolve around the appropriate dates for incorporating previously autonomous civilizations into the central civilization. For instance, the Far East is not seen as becoming significantly linked to the Afro-Eurasian world economy until the early seventh century AD while the other three perspectives argue for much earlier incorporation dates.

A third area of dispute concerns the centrality of economic expansion/ stagnation rhythms. They are part of the continuing school's conceptual triad

Table 15.2 Dominant powers/hegemons/leader foci

[296]

(hegemony/rivalry, core-periphery, and A/B phases) and are seen to extend far back in antiquity. Economic fluctuations are critical to the evolutionary school too but with the emphasis placed primarily on the nineteen K-waves of the past millennium. The continuing school's emphasis appears to be more focused on fluctuations in prices and trade while the evolutionary school's focus is on commercial and technological innovation. Economic rhythms of this sort are not particularly prominent in the work of the comparative and engulfing approaches.

Evidence advanced There can be no question that the types of questions that are being raised by very long-term foci on world system history do not lend themselves readily to operationalization. The data are simply hard to find. Nevertheless, the willingess to develop explicit empirical evidence for the assertions that are being made is highly variable. That is unfortunate inasmuch as the assertions that are made are often controversial. They also concern phenomena with which many social scientists are unfamiliar. How many readers are equally familiar with such topics as the development of language in die Bronze Age, the diffusion of metallurgical technology in the early Iron Age, the differences between Han, T'ang, and Sung trade routes, and the strategic preferences of the Portuguese, Ottomans, and Ming? Generalizing about these topics is an uphill battle. Not only are there extraordinary start-up costs in developing familiarity with these activities, there is very little about diem that it is safe to assume that readers are already familiar. Therefore, an important part of the process of persuading audiences of the accuracy of one's generalizations must be placed on generating convincing evidence mat can be summarized and presented in compressed form. The presentation of evidence alone will not suffice to convince skeptical audiences but without evidence, it is only too easy to dismiss generalizations about long-term history as esoteric figments of our collective imaginations. It should go without saying that we also need data to test our theories, not merely to persuade skeptics, but also to be able to assess to what degree the various theoretical arguments have explanatory utility.

In general, then, there are some important similarities among the four social science approaches to the history of die world system. Yet die deeper one probes, the more superficial die similarities begin to appear. When one looks at die processes we choose to emphasize or ignore, the dissimilarities begin to appear more impressive than die elements on which some convergence is registered.

To some considerable extent the divergences reflect me authors' path dependencies in analytical preferences. Two of the schools - die continuing and comparative approaches — 'betray' their roots in die world-systems movement. Some of the questions emphasized are identical as are some of me answers. The engulfing perspective emerged from a comparative civilizations background and may, for mat reason, find it difficult to move beyond charting me incorporation process. Nor is it a coincidence that the evolutionary perspective 'evolved' from a long-

[297]

cycle interpretation of die past 500 years to a perspective mat highlighted long cycles over the past 1,000 years, and, most recently, to a five thousand year plus perspective. As in the case of the world system, we are what we once were (and not mat long ago) to some extent, with some room permitted for progressive development of our individual research programs.

What would it take to bring about greater convergence? Unfortunately, greater convergence may only be possible through a process of analytical conversion. By conversion, I do not mean the wholesale adoption of a new set of assumptions and an abandonment of current commitments. It is unlikely that analysts associated with one school of thought will abandon their own research program for another. Rather, the best that we may be able to hope for would be die adoption/cooptation of elements of one research program into others. An example is die strong reliance on urbanization data in die engulfing school. Not only have the urbanization data been employed to examine propositions from the continuing school, the utilization of urbanization data has been emulated to different extents by me comparative and evolutionary schools.

Alternatively, we may find increasing elements of overlap through a form of serendipity. The strong emphasis on the early emergence of an Afro-Eurasian trading network is an important example because it constitutes die fundamental context within which world system development has taken place. Continued participation in die trading network imparts synchronicities to other important activities just as a common emphasis on the influence of this central trading network should work toward synchronizing some of the outcomes of our research into how die processes work, and at what pace they work. For instance, once one accepts the premise of an old, central trading network, it is difficult to escape an improved appreciation for the historical and pivotal role(s) of Central Asia. All four schools share this appreciation and an interest in mapping the historical fluctuations in the interdependence of Afro-Eurasia.

Another type of convergence may come through partial overlaps between two schools of thought. For instance, the comparative and engulfing schools share an interest in how multiple world systems move toward one central system that is not equally prominent in the other two schools that prefer to stress the unity of die world system. As a consequence, comparative and engulfing generalizations and conclusions about the process of incorporation have some greater probability of converging.

But partial overlaps need not be made more likely only through die agency of similar starting assumptions. A case in point are the similarities in the comparative and evolutionary school's explanation of the rise of the West. The explanations are not identical but they are strikingly compatible. Both accounts attribute die 'missing link' to me relatively novel emergence of the European merchant republics, me regional environment which facilitated the survival of mese capitalist niches, and the connections of die trading city-states to the larger, Afro-Eurasian trading system.

Nonemeless, what amounts to a moderately pessimistic evaluation of the chances for increased convergence through die explicit cooptation of some

[298]

findings from other schools, through serendipity, or through coincidence needs to be tempered by the fact that we have not been doing this type of research for very long. We do not yet have a number of detailed examinations to review. We do not yet have a number of well-specified theories to evaluate. We do not yet have a critical mass of scholars engaged in analyzing the history of the development of the world system from social scientific points of view. In other words, we have only just begun to crawl out of die primordial analytical mud. The prospect of considerable evolution to come awaits us.

More specifically, the continuing school's arguments about continuity have not yet been subjected to close scrutiny. The comparative school's arguments about the value of comparing different types of world systems have not yet been demonstrated by compelling analytical demonstrations. The engulfing school's fixation on the incorporation process has yet to move much beyond its taxonomical interests in, and descriptive observations about, multiple civilizations. The evolutionary school's long attachment to the post-1494 era, and more recendy die last millennia, leaves a great deal to be done in integrating its 'modern' findings with 'premodern' history.

Perhaps, then, it would be naive to expect much convergence so early in the game. Premature convergence may also prove to be undesirable. It may not take a hundred blooming flowers to bring some order to die complexities of world system development but, evolutionally speaking, variety is more desirable than uniformity. All four of die main schools of thought currently available appear to be undergoing evolutionary processes of their own as they engage in a trial-and-error confrontation with die enormity of world system developments. Other perspectives are sure to emerge and evolve as well. Selection processes, ideally, will ensure die survival of die optimal explanations until somebody comes up with somediing better.

1 The choice of labels is not meant to imply that any one approach monopolizes an evolutionary stance. The comparative study of world systems school claims to be studying systemic evolution as well. Equally, that school has no monopoly on a comparative stance. The choice of labels hinged on explicit emphases. It should also be noted that these schools of thought or approaches are hardly static monoliths. Changes in assumptions and arguments should be anticipated. My comparative comments are based on a reading of the following works: for the continuing school, see Frank 1993b; Frank and Gills 1993; Sanderson 1995b; for the comparative school, see Chase-Dunn and Hall, 1991, 1997; for die evolutionary school, see Modelsld 1994, this volume; Modelski and Thompson 1988, 1996a, 1996b; Rasler and Thompson 1994; Thompson 1988, 1995; for the engulfing school, see Wilkinson 1992a, 1993a; Frank and Gills 1993, and Sanderson 1995b.