the semiperiphery

Chapter 5 of C. Chase-Dunn and Thomas D. Hall, Rise and Demise: Comparing World-Systems, Boulder, CO: Westview Press.

The Semiperiphery:

Seedbed of Change

In this chapter we propose that core/periphery hierarchies are important structural elements in the reproduction of world-systems and that semiperipheral locations within core/periphery structures are important loci of forces that transform world-systems. The comparative study of core/periphery hierarchies began with several older works on "frontiers" (e.g., Lattimore 1940; McNeill 1964; Adams 1977), but the task of sorting out the different types of regional dominance/dependence relations and examining the importance of these for the dynamics of social development is yet in its infancy.

Here we discuss our general conceptualization of the semiperiphery. Then we consider four different, but related, general formulations of a theory of social change. Finally we examine several types of social change that may be understood as instances of semiperipheral development. We employ the conceptualizations of core/periphery differentiation and core/periphery hierarchy already considered in Chapter 2.

The Semiperiphery Concept

The idea of the semiperiphery was first applied to the modern world-system by Immanuel Wallerstein (1974a 1974b, 1979b) and it has been developed and further explored by Giovanni Arrighi (1985). In various parts of Wallerstein's historical and more theoretical analyses of the Europe-centered world-system he suggests several meanings of the concept of the semiperiphery. In Chapter 2 we conceptualized the semiperiphery to include:

1. regions that mix both core and peripheral forms of organization.

2. regions spatially located between core and peripheral regions.

3. regions spatially located between two or more competing core regions.

4. regions in which mediating activities linking core and peripheral areas take place;

5. regions in which institutional features are in intermediate in form between those forms found in adjacent core and peripheral areas.

For all of these we expect that semiperipheral regions will be dominated by the core, but at the same time will dominate peripheral areas.[1] These five alternative forms of semiperipherality are dependent on the nature of the core/periphery relations that exist in any particular world-system. This is because core, periphery, and semiperiphery are relational concepts that are context-dependent. Nevertheless, we can abstract from particular systems to consider whether or not there may be patterned regularities that correspond to these analytical distinctions.

Four Approaches to Semiperipheral Development

Generally stated, we contend that semiperipheral areas are likely to generate new institutional forms that transform system structures and modes of accumulation. These changes often lead to the upward mobility of these same semiperipheral actors in the core/periphery hierarchy. We will see that the semiperiphery is fertile ground for social, organizational, and technical innovation and has an advantageous location for the establishment of new centers of power. That is why the structural position of the semiperiphery has such evolutionary significance.

Before we turn to a consideration of several types of semiperipheral development that illustrate this general principle we will summarize four earlier theoretical approaches that utilized different language, but that overlap significantly with the theory of semiperipheral development that we propose. The purpose of this review is to provide several possible related approaches that may help us to more clearly specify our general formulation and to fine-tune its application to different types of circumstances. The four approaches we will consider are:

* Leon Trotsky's "laws of uneven and combined development,"

* Alexander Gerschenkron's "advantages of backwardness,"

* Elman Service's "evolutionary potential,"and

* Carroll Quigley's "institutionalization of an instrument of expansion."

Once we summarize the relevant parts of these approaches we will show how they relate to our theory of semiperipheral development.

Trotsky

Leon Trotsky (1932), a professional revolutionary, wrote a history of the Russian revolution that began with a discussion of the "historic laws" of "uneven and combined development." He used this formulation to explain the "peculiar" form of capitalist development in Russia and the stage-skipping evident (from a Marxist view) in the sequence of Russian revolutionary transformations compared to the English and French revolutions. We will see that both Service and Gerschenkron bear an intellectual debt to Trotsky.

Trotsky's most general formulations are as follows:

Unevenness, the most general law of the historic process, reveals itself most sharply and complexly in the destiny of the backward countries. Under the whip of external necessity their backward culture is compelled to make leaps. From the universal law of unevenness thus derives another law which, for lack of a better name, we may call the law of combined development--by which we mean a drawing together of the different stages of the journey, a combining of separate steps, an amalgam of archaic with more contemporary forms (emphasis in the original, Trotsky 1932:5 ff.).

And elsewhere Trotsky (1932:4) says:

A backward country assimilates the material and intellectual conquests of the advanced countries. But this does not mean that it follows them slavishly... Although compelled to follow after the advanced countries, a backward country does not take things in the same order. The privilege of historic backwardness--and such a privilege exists--permits, or rather compels, the adoption of whatever is ready in advance of any specified date, skipping a whole series of intermediate stages. Savages throw away their bows and arrows for rifles all at once, without traveling the road which lay between those two weapons in the past. ... The development of historically backward nations leads necessarily to a peculiar combination of different stages in the historic process (Trotsky 1932:4-5).

Trotsky formulates the idea of the advantages that "backward" societies have in being able to easily adopt or innovate new forms, their ability to leapfrog "stages of development," and the beneficial effects of being able to recombine different elements.

Gerschenkron

Alexander Gerschenkron (1966), an economic historian, further developed the idea of "the advantages of backwardness" to explain the rapid industrialization of certain countries that followed Britain's lead in the late nineteenth and early twentieth centuries. According to Gerschenkron countries that have certain natural and organizational features favorable to industrialization can rapidly industrialize by importing production technology, although the mechanisms they employ and the sectoral path of economic change they follow will differ significantly from those found in the original industrializer. Gerschenkron's discussion focussed primarily on Germany but he also considered France and Russia.

Gerschenkron noted that rapid secondary industrialization has aspects that are similar to social movements in that economic change occurs across many sectors together and the whole process involves a spirit that combines entrepreneurship with collective enthusiasm. He emphasized the importance of innovations in banking and state sponsorship of economic development in these cases of "catching up" with the leader. According to Gerschenkron, the most important "natural condition" that facilitates economic growth is the availability of raw materials. Organizationally it is important to have a unified state, as opposed to a collection of baronies (e.g., Germany before 1870). Serfdom is mentioned as a social organizational barrier to industrialization, but Gerschenkron argued that the very lack of an industrial labor force might facilitate rapid industrialization by encouraging the adoption of the most up-to-date labor saving production technology.

Though "backwardness" is generally argued to be a boon for rapid change, under some circumstances Gerschenkron suggests that there is such a thing as too much backwardness, as in his contrast between Russia and Germany. This is felicitous for our reinterpretation of his approach in terms of a distinction between periphery and semiperiphery.

Service

The anthropologist Elman Service first published his discussion of "evolutionary potential" in a volume co-edited with Marshal Sahlins (Sahlins and Service 1960; Service 1971). Service argues that development is usually discontinuous in space, with older localities losing dominance to new centers. He posits an interaction between adaptation and adaptivity that produces this uneven development in space. Adaptation refers to an institutional adjustment to social and environmental pressures for change. In contrast adaptivity is the ability to make adaptive changes in response to social or environmental changes. A highly adapted society typically has very low adaptivity (Service 1971). An innovation in technology, social organization, or cultural institutions may allow a society to become well-adapted to a natural and socio-economic environment, but the very investment in that innovation, or constellation of innovations, will, at a later point in time, inhibit the ability of the society to adapt to new circumstances. Adaptation usually involves specialization and the investment of resources, while adaptivity is the quality of being able to make new changes.

Service argues that a general process of social evolution -- the interaction between adaptation and adaptivity -- accounts for the rise and fall of civilizations, the spatial discontinuity of innovations in cultural techniques such as writing, and the rise and fall of contemporary nation-states. He addresses the issue of the benefits and costs of specialization raised in the literature on biological evolution, but his formulation is clearly directed toward an explanation of uneven social development. Older regions become entrenched in earlier adaptations, while newly developing areas can "take the best and leave the rest." Service acknowledged his debt to Trotsky by quoting Trotsky on the rapid, stage-skipping nature of Russian industrialization.

Quigley

Carroll Quigley (1979), a historian, reformulated Arnold Toynbee's approach to the rise and fall of civilizations. According to Quigley civilizations rise because semiperipheral regions innovate instruments of expansion. These become institutionalized and eventually the organizations stultify and decay, and then the process repeats itself. Quigley (1979:366) actually used the term "semiperiphery." His theory of rise and fall is formulated as a series of seven stages: mixture, gestation, expansion, age of conflict, universal empire, decay, and invasion.

Quigley's discussions of the stages of mixture and gestation are suggestive of semiperipheral development. He says,

Every civilization, indeed every society, begins with a mixture of two or more cultures. ... But such casual cultural mixture is of little significance unless there comes in to existence in the zone of mixture a new culture, arising from the mixture but different from its constituent parts. .... The contributing societies may be civilizations or merely producing societies (agricultural or pastoral) or merely parasitic societies (with hunting or fishing).... Since cultural mixture occurs on the borders of societies, civilizations rarely succeed one another in the same geographical area, but undergo a displacement in space.... But on the borders of societies there is a considerable mutual interpenetration of social customs, and there arise, accordingly, alternative ways of satisfying human needs.... civilizations have generally arisen on the periphery of earlier civilizations. Canaanite, Hittite, and Minoan civilizations arose on the edges of Mesopotamian civilization (Quigley 1979:147-8).

Missing from Gerschenkron and Trotsky is any consideration of the

mechanisms that bring about organizational inertia and resistance to new adaptations in older core regions. They discuss why "backwardness" is an advantage, but not why being first is sometimes a disadvantage. Service and Quigley do consider both sides of these processes, but only in the most general terms. Quigley wrote quite a lot about ossification and efforts to reform institutions in old core regions.

One problem with all these approaches is that they are formulated in terms of levels or stages of development and therefore they largely ignore the hierarchical and structural aspects of relations between societies within a system. Backward or semiperipheral areas catch up, but there is little attention to the importance of this as upward mobility within a larger socially structured regional hierarchy.

All these approaches however are compatible with the notion that a semiperipheral location is a fruitful locus of transformational changes. But how might world-system mechanisms of intersocietal domination and exploitation, and processes of the development of underdevelopment fit in to this approach?

A theory of semiperipheral development is not a claim that catching up or becoming a new center of domination is possible for all regions. As we have said, the extent to which core/periphery hierarchies and uneven development are reproduced needs to be determined for each world-system. There have been world-systems in which the spread effects of development are much stronger than the backwash effects of underdevelopment. All hierarchical world-systems seem to experience uneven development and cycles of political centralization/decentralization. We would like to determine those aspects of semiperipheral development that are common to different types of world-systems and those aspects that differ across types.

We also need to point out that upward mobility and transformational action are not necessarily the same thing. It is possible to succeed within a system without transforming the rules of that system. And it is possible to change the logic of social action without moving toward a more powerful or central location in a system. But these two analytically separable features often do occur together. We now turn to a consideration of cases in order to concretize this discussion.

Types of Semiperipheral Development

Once again we present a heuristic typology. This, we promise, is the last. This time we want to designate different kinds of semiperipheral development. Some of these types overlap, and in one category there is only one empirical case. Nevertheless in order to see construct the general notion it is necessary to designate important differences between the following types:

1. conquest by semiperipheral marcher chiefdoms

2. conquest by semiperipheral marcher states

3. extensive and intensive commodification by semiperipheral capitalist city-states

4. the rise of European hegemony -- an upwardly mobile and transformational region,

5. the rise and fall of hegemonic core states within the Europe-centered world-system, and

6. the emergence of revolutionary challenges to capitalism in the semiperiphery of the contemporary world-system.

Semiperipheral Chiefdoms

The phenomenon of the rise and fall of chiefdoms has been studied in several different contexts by anthropologists.[2] Chiefly polities interact with and compete with one another in sets that we refer to as "interchiefdom systems." The process of rise and fall involves competition among actual and potential paramount chiefs. The territories and peoples under the control of a single paramount grow as he succeeds in expanding his domain, usually by conquest. The strains of maintaining an expanded domain often eventually lead to fragmentation back into smaller independent polities. Marshall Sahlins (1972:141-8) describes such a process for the case of precontact Hawai'i. Kristian Kristiansen (1991) sees an analytically similar process in the cyclical changes that occurred among polities in Bronze Age Scandinavia and Northern Europe. Along with the rise and fall phenomenon, which is cyclical around some equilibrium of polity size, rises sometimes occur that dramatically expand the scale of political organization by forming a single polity that is much larger than any previous domain in the region.

Our model of semiperipheral chiefdom conquest is taken from Patrick Kirch's (1984:199-202) analysis of the evolution of Polynesian chiefdoms. Kirch presents a model of population growth and lineage stratification for a single hypothetical Pacific island, though he is thinking about the actual case of the big island of Hawai'i. The principle of successive primogeniture is organized in Hawai'i (and in Polynesia generally) on the basis of conical clans.[3] Genealogical lines are ranked in terms of closeness to the original ancestor, and each person theoretically has a unique and intransitive position in this system of ranked seniority. In practice this system operated as a theoretical ideal. Competing claims were settled by force and genealogies were reconstructed according to the outcomes of struggles among contending chiefs. Nevertheless, the principle of genealogical seniority was an important one in legitimating rule. In Hawai'i this principle evolved to include a radical separation between chiefs and commoners, and the practice of brother-sister marriages among sacred chiefs and chiefesses produced the highest possible seniority for offspring.

Kirch's model for colonization and population growth works as follows. A colonizing party of Polynesians lands on an island and occupies the most favorable location, usually on the windward side of the island. The windward side receives the most rainfall and is generally the best for agriculture. Another locational desideratum is offshore reefs that make for good fishing locations. When the best location becomes fully occupied because of population growth this community "hives off" -- splits in two -- to occupy the second most favorable location on the windward side. The second community is headed by a chief who has less seniority than the line that rules the original founding community. This process continues until the windward side is occupied, at which point locations on the leeward side of the island will be occupied in the order from best to least desirable. This produces an island in which the ecological desirability of locations maps perfectly with the kinship hierarchy.

Kirch goes on to propose that after the island is filled continuing population growth will generate pressures for either further migration to other islands or will result in intensification of conflict and, eventually, pressures for island-wide hierarchy formation. He suggests that successful conquest efforts are likely to come from regions that are ecologically less favored such as the leeward side of the island of Hawai'i, where large-scale irrigation was not feasible (Kirch 1984:204). The junior line from a moderately marginal ecological location had both the motive and the means to challenge the old hierarchy.

This model may be interpreted in world-system terms as follows. An intra-island core/periphery structure emerged in which ecological factors and kinship hierarchy corresponded. Continuing population growth created pressures for intensification of production, intensification of conflict, and hierarchy formation. These pressures eventually resulted in the formation of island-wide chiefdoms by means of conquest, and the successful conqueror chief was likely to come from one of the junior lines on the less ecologically-favored side of the island. This is fascinating because of its analytic similarity with a process known to occur in state-based systems -- the formation of empires by semiperipheral marcher states.

Semiperipheral Marcher States

Secondary state-formation on the marches[4] has frequently been recognized as a phenomenon that is related to the rise and fall of empires and the shift of hegemony within interstate systems. Of interest to a theory of semiperipheral development are the processes that facilitate new and adaptive organizational forms in marcher states and which inhibit or obstruct effective responses in older core regions. Both of these have been discussed in general terms by civilizationists, historians, and historical sociologists. Rather than reviewing this rather extensive literature we illustrate with a particular case, the conquest of Sumer by Akkad.

There have been many clear cases of semiperipheral marcher state conquest of older core regions. A list of such cases known to most world historians would include: the Akkadian empire, the Kassites, Assur, Upper Egypt, the Medes, Achaemenids, Hittites, Hyksos, Macedonia, Rome, Normans, Maurya, Shang, Chou, Manchus, Toltecs, Aztecs, Huari, and Inka. In all these cases the named semiperipheral state conquered an older core region and set up a new, larger core state. We mean to exclude conquerors that were either peripheral or who simply decimated the old center without setting up a new regime. We also exclude core-wide conquests by states that have long been in the core, or who are restoring their former glory by making a comeback. The original position of a conqueror within a core/periphery hierarchy is important, and the position of the conquered region is important. Conquests of regions that do not have core status do not count. Thus, this conceptualization contains two components: the strength of the emerging semiperipheral marcher and the strength of the old core. Some cores fall to marauders largely due to their own internal processes of disorganization, while others are not rapidly decaying but yet are not able to stave off powerful challengers. The intent of constructing the marcher-state category is not to produce a pure ideal type, but rather to help specify a set of processes of semiperipheral development that are very different from those that operate in the other general categories we consider later.

Akkadian Empire

The Sumerian interstate system and world-economy lasted for seven hundred years before becoming transformed into an empire for the first time by Sargon of Agade. These city-states interacted within the context of a regional economic network that included a core/periphery division of labor with surrounding pastoralists, rain-watered horticulturalists, and specialized quarrying and manufacturing villages. In the Sumerian core irrigated agriculture was the basis of the first cities. These cities were politically autonomous although they shared a pantheon and the first written language. The early states were theocracies governed by temple priests and an assembly of lineage heads. As development of this form of organization spread up and down the alluvial plain of the fertile crescent and as land appropriate to this form of production became more scarce, the city states began to transform into political organization based on an elected war leader, a Lugal. The power of the Lugal gradually increased as warfare among city-states became more frequent and more important as a means of controlling trade routes and access to raw materials. While recent evidence reveals the existence of some forms of commodification of land and money economy, most production, distribution, and trade was carried out by reciprocal kin networks linked together by the political apparati of temple and palace.

Both spread and backwash effects apparently operated in the core/periphery hierarchy although the extent, nature and forms of domination and exploitation are uncertain. The emphasis has been on spread effects because it is obvious that irrigated agriculture diffused within the Tigris-Euphrates flood plain and leap-frogged to other areas. It is also clear that surrounding pastoralists and horticultural communities became specialized in production for exchange with the core region. The core region exported grain, but also textiles and other manufactured goods. We also know, however, that manufacturing emerged also in remote villages near steatite (soapstone) deposits, as we have seen due to transportation costs (see Chapter 2). Similarly, throughout the Bronze and Iron ages, metal-working was often associated with mountain societies.

Friedman and Rowlands (1977) contend that the main dynamic of exchange among cities within the core area was based on a prestige goods economy. This, they argue, made it hard for older core cities to monopolize power resources. Prestige goods economies are vulnerable to copying or redefinition of the symbolic goods that signify high status. The pantheon of shared gods within the Sumerian core indicates a struggle over the claims of different city deities to superior position within the regional pantheon.

Igor Diakonoff (1991) argued that exploitation within the Sumerian region was primarily based on "internal" extraction of surplus product from subjugated classes rather than "external" exploitation based on political-military subjugation. He contrasts this early period with a later period of "warrior empires" in which conquest and exploitation became much more important. From the surviving mythical literature of Sumer we know that kings took armies into peripheral areas to obtain scarce objects such as wood for construction, etc. However, Diakonoff claims that such dominance relations were unstable. The multicentric nature of the Sumerian core may have allowed peripheral groups to play competing cities against one another. We know that the Sumerian cities attempted to control peripheral resources by establishing settler colonies in peripheral regions (Algaze 1993). Yet, Michael Mann (1986) argues that even these colonies were difficult to control under the circumstances of high overland transport costs and severe logistic problems.

The question of the importance of core/periphery hierarchy in the Sumerian or any other world-system and the relative importance of spread vs. backwash effects can only be determined by surveying the whole system. Kohl (1988) compares two peripheral regions, Transcauscasia and Western Turkestan, and concludes that these two areas developed rather autonomous and successful new centers in interaction with the Sumerian core in the Bronze age. But a study that focussed on the nineteenth century United States or twentieth century Japan (or Korea or Taiwan) might conclude that upward mobility or rapid spread effects are the usual case in the modern world-system. Any decision about what is typical must be based on a survey of the whole system and must take account of relative rates of development. Peripheries are not areas that do not change at all. They are areas that develop core-like features more slowly than the relevant core. Peripheralization often involves the development in the periphery of social structural features that impede further development toward the core. Any comparative study of core/periphery hierarchies should take these matters into account.

Sargon, the eventual king of the city of Agade, served as a young cup-bearer to the king of Kish, one of the old cities of the Sumerian core in Southern Mesopotamia. Sargon was originally from Agade, perhaps a recently established city up river from the old core region. Agade may have been populated by recently settled nomadic pastoralists who had settled on the fringes of the Sumerian core (McNeill 1963). In any case the residents of Agade were speakers of Akkadian, a Semitic (i.e. non-Sumerian) language.[5]

The Sumerian interstate system had already begun to exhibit the features of a balance of power mechanism and the rise and fall of hegemons, but no city had managed to conquer the whole core region. Sargon led the Akkadians on a military campaign that defeated all the other cities as well as much of the peripheral hinterland and erected the first empire-state on Earth composed of formerly sovereign city-states.

There is great disagreement about the relative importance of different factors that led to the Akkadian conquest. Because of the scarcity of evidence it is impossible to know for sure which interpretation is the best, but it is nevertheless interesting to elaborate and distinguish between two models, both of which are compatible with a general theory of semiperipheral development. First, Mann's (1986) developmental history of power techniques stresses the importance of military technology and organization in the expansion of empires. In his chapter on the "first empires of domination" Mann lays out an admirable model of the contradictory forces of centralization and decentralization that operated in the ancient empires. Mann (1986:130) also suggests, "What had been hitherto semiperipheral areas became, in a sense, the new core of civilization. Marcher lords were the pioneers of hegemonic empire."

Speaking explicitly of the Akkadian conquest, Mann claims that the recently settled pastoralists combined peripheral with core-type military techniques in a way that gave them an advantage over the older Sumerian core. The Sumerians used heavy infantry phalanxes and cumbersome chariots drawn by equids, "perhaps onager and ass hybrids" (Mann 1986:132). True horses for riding or more mobile and speedy chariots had not yet been developed. The core infantry were "suited for slow, methodical campaigns whereby small densely

settled areas could be conquered and defended. They arose from the necessity to defend the early city-state and perhaps to conquer its immediate neighbors" (Mann 1986:132 ff).

According to Mann the Akkadians combined heavy infantry force with the use of a newly developed composite bow. Mann's contention that Sargon combined peripheral and core-type military techniques rests on his claim that this type of bow was a peripheral product. He says, "But archery was apparently developing rapidly from hunting practices, and the use of the bow seems to have given a comparative advantage to the marchers if combined with infantry force" (1986:133).

Mann also asserts that the old core was in some ways "ripe for the picking." The reliance on trade routes became increasingly dependent on military protection as core development created new needs and abilities in peripheral peoples. Though the old core still had a comparative advantage in production, reliance on long distance trade exposed merchants to raiding and tribute extraction. Competition among core city-states often spurred the development of semiperipheral marchers directly as core states made alliances with marchers in order to win struggles within the core.

Mann's (1986:141) discussion of "piecemeal treachery" suggests that cleavages within the core state societies were developing. In the context of a discussion of the difficult supply problems of an ancient conquering army, Mann points out that campaigns approached one city-state at a time, and success was dependent on a quick victory. Both superior force and "coercive negotiations" were important. Mann says, "The defenders were not being offered much of a choice. If they resisted, they might be killed or enslaved; if they surrendered, their entire visible surplus might be pillaged and their walls knocked down. But a discontented cousin or younger son and his faction could be promised more, and the city delivered up by them. This faction would be added to the army or left in charge of the city" (1986:141).

A second, rather different, yet complementary, picture is presented by the Soviet Assyriologist Igor Diakonoff (1991). Diakonoff focussed more on class relations in his explanation of how the Akkadians conquered Sumer, although he also mentions the importance of Sargon's use of the bow. Diakonoff (1991:85) says, "It is entirely possible that Sargon had access to yew-tree (or hazel-tree) groves in the foothills of Iran and Asia Minor, or that a composite bow, glued together from horn, wood and sinews, had already been invented at that time.[6]

Diakonoff claims that the conquest by Akkad was based on both ethnic and class factors. He says:

Legends of much later times describe Sargon the Ancient as a man of very humble origin, and there is no reason to doubt the credibility of this tradition. It was said that he was a gardener, the adopted son of a water-bearer, and that he became a cupbearer of the lugal of Kish.... The fact that Sargon had no roots in the traditional nomes and did not depend on the nobility allowed him to draw his support from the common people, forming a militia which might have been more or less voluntary (1991:84).

Another contributing factor to the superior solidarity and motivation of the Akkadian army may have been based on their status as Semitic speakers who had only recently settled down from pastoralism (but see note 5 above). Pastoralists generally have a kinship structure that promotes solidarity among male lineage heads, and their experience with domesticated animals is easily transformed into the husbanding of peoples (McNeill 1963; Hall 1991a). A kin-based reciprocal society can more easily mobilize collective energies than a more stratified urbanized society.

In contrast, in city-states private property and the differentiation between temple and palace increased competition within the urban ruling class, and increasing stratification between classes made mobilization of the commoners for warfare more dependent on material incentives and less susceptible to calls for sacrifice in the name of the society. The "piecemeal treachery" that Mann mentions may well have been a function of increasing competition and conflict within the ruling classes of the old Sumerian core. As to increasing interclass polarization Diakonoff (1991:85) says:

In Lagash the events which led to the coup of Uruinimgina attest to the accumulation of many grievances against the prevailing order. Sargon could encounter support everywhere. The poorest community members may have been interested in curbing the inordinate growth of the nome aristocracy's power; service in Sargon's army offered them hope for social and material betterment... But also within the temple and the state economies, the personnel was stratified to such a degree that it was always easy to find here people who were willing to help destroy the nome order.

The disagreements over matters of fact regarding the Akkadian conquest stem from problems of the availability of evidence for this most ancient case of semiperipheral marcher conquest. This makes it difficult to draw firm conclusions about which particular mechanisms were important and about the general implications of this example for cases of semiperipheral marcher conquest. The Akkadian case is important because it was first, and because it transformed the basic logic of a world-system more completely than did later instances. Ironically, as Diakonoff (1991:87) points out, "Thus, the popular masses, who had supported Sargon, gained little from his victory and eventually lost considerably, because a despotic and bureaucratic form of government became more established in Mesopotamia and lasted for millennia." The semiperipheral region that combined elements of a peripheral kin-based mode of accumulation with those of the core tributary mode succeeded in eliminating many of the vestiges of the kin-based mode that had remained in the old core and in establishing a more centralized, more exploitative, purer form of the tributary mode than had ever existed before.

In a related, but somewhat different approach to the phenomenon of semiperipheral marcher states Randall Collins (1978) contends that the advantages of states in the marchlands is primarily geomilitary. Because they are near the edge of the core "heartland" they do not need to defend several borders at once and so they can concentrate their resources on a strategy of conquest that adds territory sequentially without threats from the rear. The disadvantage of older core powers is that they must defend themselves from many sides and so their resources are spread thinly. This explanation may account for some of the semiperipheral marcher state phenomenon but we doubt that the purely geopolitical advantages of location on the marches is the most important factor. The exact combination of elements that allowed semiperipheral marchers to conquer older cores undoubtedly varied from instance to instance, and these different combinations also varied in their degree of fit with a general theory of semiperipheral development. Only a comprehensive study of a large and representative number of instances of semiperipheral conquest and failed attempts can sort out the general from the specific aspects of this phenomenon.

Autonomous Capitalist City-states

The historical process of commercialization within the states and empires dominated by the tributary modes of accumulation has not received sufficient scholarly attention. It has been ignored because of the vociferous debate between those who focus on the differences between the ancient and the modern world and those others who argue that the instinct to truck and barter is a universal feature of human nature, and thus all societies can be analyzed in terms of the model of "economic man." We side with Polanyi and the substantivists with respect to the distinction between normative, political, and market-based forms of social integration. We do not think that market rationality is "natural." Rather, as Marx claimed, the market and the commodification of aspects of life are socially constructed institutions.

The debate between the substantivists and the formalists has clouded the analysis of the historical development of commodified forms within the context of normative and tributary modes of production. The substantivists tend to argue that either market forces did not exist in the ancient world, or that they were so encumbered as to be unimportant. The primitivists, such as Weber (1981) and Finley (1973) argue that ancient capitalism was fundamentally different from modern capitalism and that the dynamics of ancient society were not importantly affected by market forces. This debate about similarities and differences has obscured the study of the actual processes of commodification of land, labor, wealth, and goods and the causes and consequences of commodification for the dynamics of development within the tributary modes of accumulation. The fact that market relations are not dominant does not by itself prove that their existence in certain spheres is unimportant, and in any case, since we know that capitalist production eventually became the dominant mode of accumulation, and since we want to know how modes become transformed, it makes sense to study the actual processes of commercialization and their agents.

Our third type of semiperipheral development focusses on capitalist city-states in the semiperipheral interstices of empires dominated by the tributary mode of accumulation. What do we mean by capitalism in this setting? We reject the dichotomous distinction between ancient and modern capitalism in favor of the notion of degrees of commodification. Even within modern capitalism perfectly operating price-setting markets are a rarity, especially for "problematic" commodities such as labor. The extent to which competitive bidding by buyers and sellers sets exchange ratios (prices) is a variable that is influenced by the operation of both normative and political (coercive) regulation. Capitalism has become the dominant mode of accumulation in a socio-economic system when market forces have greater weight in the determination of the dynamics of growth, reproduction, and decline (and greater weight in the competitive struggle determining the distribution of social resources) than normative or political-coercive forms of regulation. In the ancient world money (commodified wealth), markets, interest, commodified labor, and the production of commodities were more limited and less purely developed than these institutions have become in the modern world. Commodification was a process that was uneven. But there were increasingly significant degrees of commodification, and some actors are helpfully understood as having been merchant capitalists and production capitalists in the ancient world-systems.

The question of the control of states by "capitalists" is also problematic. Control is also a variable, and is always shared or conditioned to some extent. Control is most easily seen in its effects. State policy is bent toward the provision of protection rent in Frederick Lane's (1973, 1979) sense.[7] Lane defines protection rent as the differential returns received by merchants whose trading efforts are supported by a cost-efficient and protection-providing state. This assumes that the point of state policy is to maintain (and extend) the conditions for profitable trade at minimal cost. This definition can be extended easily to cover the foreign and domestic policies that are beneficial to capitalist accumulation through production, as well as trade. No state measures up to the pure type because all states must make compromises to some extent with other groups besides capitalists, but some of the autonomous capitalist-dominated city states of antiquity approach the pure case closely.

The mix between those emphasizing merchant capitalism (accumulation through exploiting price-differentials across different regions) and production of commodities for sale varies across cities and, over time, within cities. Some cities combine these forms of capitalism with a more typical (in the context of tributary mode empires) accumulation through taxation and tribute.

There are also important differences depending upon the sectors in which capitalist production is conducted. The Greeks combined production of wheat, olives, and wine for use and for sale, a partially-commercialized economy that required control over agricultural land (Rostovtzeff 1941). This was a more commodified mode of accumulation than that found within the large territorial empires, but it was less commodified than the approach employed by the Phoenicians. The Phoenician city-states concentrated on merchant capitalism and the production of manufactures for export. This meant that they did not need to control large tracts of land. Thus, their cities were built on promontories that could be protected from armies by naval force. They did utilize nearby land for truck gardens to supplement staple food imports, but this did not weigh them down with the necessity to control large land areas. This advantage for capitalism is emphasized by Fernand Braudel (1984) in his study of merchant cities in Europe.

Another complication in the delineation of a class of autonomous capitalist city-states is their degree of autonomy. One reason why the autonomous city-states of antiquity were semiperipheral is that they were on the edges of, or the boundaries between, large territorial empires. These empires were the dominant core polities of the ancient world-systems. Spatially the capitalist city-states were often located such that they could easily mediate trade between the core empires and peripheral regions. But the autonomy of these cities varied. Often they acted as allies to one or another empire, and they were sometimes swallowed up by imperial expansion (Frankenstein 1979).

Some cities within empires were allowed partial autonomy for merchant and production capitalism, and this was increasingly the case as tributary mode conquerors became more sophisticated about how to extract surplus. As empires themselves became more commercialized, kings and emperors learned to tax merchants rather than expropriating them outright, and conquered cities were granted some autonomy to pursue production and trade. The point here is that autonomy is another variable. We focus primarily on those cities that had formal sovereignty, but we recognize that merchant cities within tributary empires sometimes played an important role in the expansion of commodification (e.g., Tlateloco, Canton, Babylon, Osaka).

Some cases of sovereign semiperipheral capitalist city states are: Dilmun, Byblos, Tyre, Sidon, Carthage, Malacca, Venice, Florence, Genoa, Antwerp, and the cities of the Hanseatic League. Another list might be constructed of semi-autonomous towns that carried on extensive trade as a result of being located on overland routes, such as those on the Silk Road that made a living by linking the distant cores of Eurasia. Some of these, like the specialized maritime cities, obtained much of their basic foodstuffs through trade, and served as agents of commodification.

Karl Polanyi's students and colleagues (Polanyi, et al. 1957) utilized the concept of the "port of trade" to characterize long distance state-administered exchange in the context of the early empires. The idea is that real price-setting markets are not operating because exchange is regulated by political deals among states, and merchants operate primarily as state agents. Subsequent research has revealed new evidence that some of the cases studied by Polanyi and his students did indeed have independent merchants trading on their own account and extensive monetization (e.g., Curtin 1984). Polanyi's overall implication is that ports of trade are simply reflections of the dominant tributary mode of production. They supposedly do not constitute important agents of social change toward a more commercialized type of economy. It may be the case that some cities in neutral zones mediating trade between empires truly were inert ports of trade in the Polanyian sense. But Polanyi did not choose to analyze most of the maritime city-states we have listed above. These were agents of market forces within the interstices of tributary empires and their activities were an important stimulus to the further commodification of large regions, including peripheral areas and the tributary empires themselves.

Marx's (1967:Chap. 20) analysis of merchant capitalism discusses how, under some circumstances, the action of merchants in buying cheap and selling dear not only equalizes prices across different regions but also encourages production for exchange and specialization. For Marx labor only becomes a commodity under the wage system in which workers sell their labor power (time) to capitalists. Market forces subject producers to the constraint of average "socially necessary labor time."[8] This means that market forces will eliminate forms of production that are inefficient in terms of the utilization of labor, i.e. new forms of labor-saving technology will be implemented and drive older forms out. Marx considered wage labor to be the only form of labor mobilization appropriate to capitalism. While we agree that this is the most flexible form (allowing market forces to most easily restructure the production process), we note that various other forms of partly commodified labor (e.g., chattel slavery) and even uncommodified labor (e.g., serfdom) have been used for commodity production in both the ancient modern world-systems. Though this kind of capitalism was not as extensively commodified as the modern kind is, it should be studied rather than swept under the rug.

The processes by which regions became integrated into a larger market economy began with the commodification of goods through the carrying trade of merchants. Merchants move things from places where prices are low to areas where prices are high. The ability to do this depends on the existence of forms of transportation that are economic in the sense that the transport costs do not consume the profit. A human carrier cannot economically carry corn further than the distance over which he will eat his pack full of corn (Drennan 1984). And this is true regardless of the mode of accumulation. Thus transport costs are the main determinants of the concentric rings within which bulk goods and lighter prestige goods move, and this is also an important reason why trade-based cities are usually built near waterways.

Another reason why semiperipheral capitalist city-states tend to rely on water transportation and naval power has to do with their interaction with conquest-based tributary empires. Merchant cities located inland are much more susceptible to conquest than those located on coasts, promontories or islands that can be supplied and defended by naval forces. These maritime powers usually combine merchant-shipping with naval supremacy, and are often considered "pirates" by their competitors.

But it is mistaken to characterize many of the cities on our list simply in terms of merchant capitalism. In many cases they specialized in manufacturing particular products for export in conjunction with their carrying trade, and this aspect of their operation also functioned to expand and deepen market demand. The Phoenician cities of Sidon and Tyre borrowed glass-making technology from the Egyptians and manufactured relatively cheap glass vases which they exported to the entire Mediterranean littoral. The Carthaginians reproduced Greek-style pottery and statuary for export. Besides reminding us of Taiwan, these examples show how the Phoenicians were agents of technological change through the operation of socially necessary labor time within an ancient world-system.

Semiperipheral capitalist city-states performed the role of trade diasporas (see Chapter 1; Curtin 1984). They not only encouraged commodity production and exchanges by providing demand for local surpluses, but they also helped to forge the intercultural bases of regularized cross-cultural trade by familiarizing people with the products and ideas of distant lands. This was transformational action, but not upward mobility as long as the tributary modes of accumulation remained dominant. It was only with the emergence of a dense and very commercialized regional system in Europe that capitalist states moved from the semiperiphery to the core.

"The Rise of the West"

Although we deal at greater length with the "rise of the West" later (in Part III), we also need to consider it briefly here as a type of semiperipheral development. The rise of the West combined two different types of semiperipheral development -- marcher states and capitalist city states -- to produce a new form that might be labeled a "semiperipheral marcher world-system." But because Europe was never really a separate system, it should rather be called a semiperipheral marcher region.

The confusion about Europe's peripherality can be resolved by analysis at four spatial scales: developments within Europe; changes that occurred in Europe's peripheral and then semiperipheral relationship with the Eastern Mediterranean and West Asian core region; Europe's links (sometimes interrupted) with the other core regions in India and China; and changes that were occurring in the Afroeurasian PGN as a whole. The three separate cores of the Afroeurasian system constituted separate political/military networks (PMNs) that had long been linked in a larger prestige goods and information network (see Chapter 8).

At first Europe was a periphery, then a semiperiphery of the Near Eastern core. Then it began to form its own internal core region, and to dominate its own periphery. Finally it came to dominate the older cores of the Near East, India and China. These shifts are considered in detail in Chapters 8,10 and 11.

The Rise and Fall of Hegemonic Core States

According to Wallerstein, hegemony is a situation in which one core state has an unusually large share of world economic and military power over other core states. Under this definition, there have been three hegemonies in the Europe-centered system since the seventeenth century CE: the Dutch, the British, and the United States. In the modern world-system, we argue, the context is a larger world-economy in which capitalism has become the dominant mode of accumulation. Thus the nature of competition and the types of institutional innovations employed by successful rising semiperipheral powers are somewhat different.

All of the three states that became hegemonic core states were semiperipheral before they achieved core status. The Netherlands was a wetland where Protestant rebels from Antwerp retreated from the Spanish Duke of Alva. The strongest European state at the time was Spain. Protestantism was a religion of the semiperiphery, an ideology that democratized access to the deity and challenged the authority of the old core. The Dutch Republic uniquely combined features of earlier capitalist city-states with a federalist nation-state policy dominated by the merchants and production-capitalists of Amsterdam (Taylor 1994). The Dutch had the first capitalist state with core status.

England was a relatively strong medieval state, but its role in the international economy was as an exporter of raw materials, mainly wool, to manufacturing cities on the continent. Efforts to gain control of trade (e.g., the expulsion of Jews, formation of the Merchant Adventurers) and to support import substitution began in the time of Queen Elizabeth, as did colonialism and the raiding of Spanish galleons by state-sanctioned privateers. The eventual success of these policies in expanding trade and manufacturing for the home and export market led England to core status in the eighteenth century, and then to hegemony in the nineteenth.

The United States was peripheral during the eighteenth century, although the "triangle trades" shipping, ship-building, and some manufacturing had already developed in New England before the Revolutionary War. The South remained a classical producer of peripheral raw materials employing coerced (slave) labor until the Civil War, but the North and the West developed core capitalism. Aspiring core capitalists and statesmen struggled politically against those who had a vested interest in the export of raw materials to the European core from 1816 until the Civil War. The north and the south struggled over control of the federal state and such issues as tariffs and land-distribution policy. The western farmers shifted their sympathies in synchrony with the world market price of wheat. When the price was high they supported free trade. When the price was low they supported the Henry Clay's "American System" -- tariff protection for industry and government-subsidized transportation links (Chase-Dunn 1980). The U.S. was semiperipheral in the sense that it contained within it a mixture of core and peripheral activities, and U.S. merchants mediated trade between the European core and the Caribbean and Latin American periphery. The U.S. reached core status in the 1880s, and economic hegemony after World War I. Only after World War II was political/military hegemony embraced.

The hegemonic sequence, stages of hegemony, and the causes of both rise and decline have been analyzed using comparisons of the three hegemons with each other and with other core and peripheral states within the modern world-system (Chase-Dunn 1989:Ch. 9). Though these cases fit the general model of semiperipheral development, a detailed comparison of the mechanisms of rise and fall with those operating in earlier semiperipheral marcher states reveals important differences as well as general similarities. In the capitalist mode of accumulation, the paths to success of these rising hegemons relied much more on comparative advantage in the production of commodities and the use of military power to protect trade routes and access to raw material inputs to commodity production. Military power was used to facilitate capitalist accumulation. In earlier semiperipheral marcher states coercive power was itself the direct mechanism of accumulation. This is simply another way of saying that the logic of competition shifted away from the extraction of taxes and tribute by military force to the accumulation of profit through commodity production.

The most successful states in the modern capitalist world-system have been capitalist states. Of course, there have been other "paths of development" within the modern world-system. Prussia and Japan attained early success through state-building and military strength that was much more reminiscent of the tributary strategy, only later converting to commodity production. The case of Russia/USSR was also very different. The dependence of capitalism on the interstate system (see Chase-Dunn 1989:Ch. 7) has reproduced the logic of geopolitics even while commodity production became the dominant mode of accumulation.

Another important difference between uneven development within the modern world-system and in earlier systems is in the degree of disruption caused by the fall of old core centers and the rise of new ones. Since the capitalist world-economy is politically organized as a permanently multicentric interstate system, with multiple competing states in the core, the process of rise and fall of hegemons appears to be more routinized and less disruptive for the system as a whole. When tribute and taxation are the main forms of accumulation, the failure of an empire is quite disruptive of all social relations. But in a system of generalized commodity production the center of power, based now more on competitive advantage in production, can move from Holland to England to the United States with relatively less disruptive consequences for the operation of the system as a whole. For example, the fall of the western portion of the Roman Empire was not fatal for the European region, but recovery took a long time, at least compared to the consequences of the decline of a modern hegemon. We contend that this difference is largely a consequence of the transformation of the mode of accumulation.

The most problematic instances of semiperipheral-based change are actual or potential challenges to capitalism. Trotsky's approach can be extended to consideration of the socialist experiment in the Soviet Union. Chase-Dunn has argued elsewhere (1982, 1992c) that the Soviet Union never successfully established an autonomous and self-reproducing socialist mode of production. Nevertheless, it posed the most powerful challenge to capitalism in the modern world-system. China may pose a similar challenge. If so, both would be cases of challenge from semiperipheral regions. To the degree that this is so, they support our argument that fundamentally new organizational forms, activities with very different logics of operation, are likely to emerge first in semiperipheral areas where both core and peripheral forms are combined and development is subjected to very contradictory forces. We explore this theme further in the following section and in Part IV.

Conclusions

We now return to the discussion of theoretical problems in the light of our consideration of the different types of semiperipheral development. The different vocabularies used by Trotsky, Gerschenkron, Quigley, and Service are only semantic problems. Whether we talk of stages, phases, or levels of development, or evolution is not really crucial. A more serious problem confronting the effort to formulate a coherent theory of semiperipheral development is the potential for confusion and circularity in the definitions of structural positions within a core/periphery hierarchy. It could be true by definition that new cores are previous semiperipheries, but this tautology would not explain anything.

When we are considering the question of upward mobility--the moving of a semiperipheral polity into core status (or hegemony or empire-formation)--we need to distinguish between position in the core/periphery hierarchy and changes in that position. In the modern world-system this is accomplished by examining relative indicators of world-system position comparatively at one point in time and rates of change over time in those indicators relative to the population of competing actors. For example, if we can consider, as some do (e.g., Arrighi and Drangel 1986), GNP per capita to be a good indicator of a country's position in the contemporary core/periphery hierarchy, then the growth rate of GNP per capita can be used as an indicator of change or stability in world-system position when it is compared to the growth rates of all the other countries in the system. A semiperipheral country will have a GNP per capita that is roughly intermediate in the distribution of crossnational comparisons, while an upwardly mobile semiperipheral country will have a relatively high economic growth rate (e.g., South Korea, Taiwan).

Of course an entire army of researchers and bureaucrats have worked on methods of national economic accounting and crossnational comparisons for fifty years to produce the concepts, operations, and actual data that make these quantitative measurements possible, and even so, the data are skimpy before 1950. But the existence of such a system for the global political economy demonstrates that the distinction between position in a core/periphery hierarchy and change in that position can, in principle, be meaningful for earlier intersocietal hierarchies. The solution is to develop concrete understandings of the forms that earlier core/periphery hierarchies have taken, and to conceive of mobility within these hierarchies as relative to the development of the existing intersocietal network as a whole.

Our discussion of semiperipheral marcher states intentionally excluded both core and peripheral conquerors, but a real test of our hypothesis must include these, as well as conquerors who emerged from regions previously unconnected with the relevant world-system. Our proposition is that conquerors from the semiperiphery are most likely to perform empire-formation and the expansion of empires, while peripheral or external invaders are more likely to be destroyers or to form very short-lived states. Transformational action by semiperipheral marcher states involves the expansion and further institutionalization of the tributary mode of production.

The relationship between nomadic pastoralists and states supports these contentions (for more details see Chapter 8 and Hall 1989, 1991a). Most often nomadic pastoralists are employed as border guards of empires (e.g., Comanches, Cossacks). Those that succeed in conquering core areas are rarely successful in forming long-lasting empires. It is rather recently-settled pastoralists who have already undergone some state-formation of their own who are the prime candidates to be empire-builders. In the act of settling down on the edges of a core region these nomadic societies become semiperipheral (e.g., Akkad). A comprehensive test of this thesis would need to survey the universe of empire-formations and classify the core/periphery positions of all the empire builders.

David Wilkinson's (1991:124-145) examination of cores and core shifts in thirteen political-military networks (PMNs) is one way to solve the circularity problem mentioned above. There are several logical alternatives involved in core shifts. The newly risen core power may be a formerly semiperipheral city or state, but it may also be: a resurgent older core state, a peripheral people that conquer and destroy a core region but do not establish a new core state, or a peripheral people that conquer an old core and do establish a new core state. These logical alternatives eliminate the tautology. Wilkinson's study provides support our hypothesis, but differences in our respective definitions of semiperiphery cloud that support.[9]

Wilkinson's survey of "core shifts" in thirteen PMNs reveals that the most frequent type of core shift is one in which a semiperipheral state conquers an old core and creates a new hegemonic power. But other types of core shifts also occur frequently. Wilkinson notes a pattern that he calls a "shuttle" in which hegemony moves back and forth between two regions. And there are many cases of semiperipheral ascent, the rise to core status of semiperipheral regions that do not conquer the old core.

The theory of semiperipheral development also claims that semiperipheral areas are disproportionately the locus of agents of major social transformations. In order to evaluate this broad claim we would need to define and operationalize the differences between activities that simply reproduce the dominant mode of accumulation from activities that expand and develop the dominant mode or that expand the logic of a new mode.

Do both success within the modern world-system and transformative influences on the mode of accumulation come disproportionately from semiperipheral countries? Obviously we are not arguing that all semiperipheral areas are transformative or upwardly mobile. Rather, our contention is that those countries that display the greatest successes at capitalism have been formerly semiperipheral, and also that the most significant challenges to the logic of capitalism have emerged from the semiperiphery. We submit that the first claim, that the most successful capitalist countries (i.e. the hegemons) were indeed previously semiperipheral. This does not establish that semiperipherality as a prerequisite to success. What we have not addressed is why some challengers succeed and others do not. This is a much more complicated matter.

The transformation to a socialist mode of accumulation is even more problematic. We contend that the large Communist states -- China and the Soviet Union-- underwent the most transformative structural changes toward socialism, though they did not succeed in establishing a self-reproducing socialist mode of accumulation. Russia/USSR was obviously semiperipheral. But what about China? By most economic measures (e.g., GNP per capita) China is still a peripheral country. But we contend that China was never completely peripheralized as a result of incorporation in to the expanding Europe-centered world-system, though regions within China were. The cultural and political strength of China enabled it to resist colonization by the West and to maintain political unity and some military strength. These features contributed to the strong establishment of institutional features of socialism that have proven difficult to reverse despite more than a decade of effort. These features also contribute to China's upward mobility within the capitalist world-economy. Aspects of socialism have also emerged within both core and peripheral countries but we submit that these have been less transformative than the changes that occurred in China and the Soviet Union.

We are not arguing that socialism is impossible in the core or the periphery, but rather that it is more likely to emerge strongly in the semiperiphery. This is because the modern core/periphery hierarchy stimulates class struggle in the semiperiphery, while it cross-cuts and dampens it in both the core and the periphery (Chase-Dunn 1989:Ch. 10). The current (1990s) phase of "free market reforms" in both the former Soviet Union and China do not vitiate our argument. It may well be that of these regions may invent a new mode of accumulation that is neither socialism nor capitalism. While that remains to be seen, it is clear that semiperipheral regions will continue to challenge the capitalist world-system in various ways.

The notion of semiperipheral development is only part of our theoretical explanation of the historical evolution of world-systems. In the next chapter we consider a sequence of interactions among processes of population growth, environmental degradation, population pressure, conflict, hierarchy formation, and the intensification of production. Iterations of these interactions, along with the transformations of modes of accumulation, account for the main patterns of social change over the last ten thousand years in which myriad small-scale egalitarian world-systems have become incorporated in to a single global capitalist system.

Notes