II.2 The five thousand year world system



II.2 The five thousand year world system

Our thesis is that the contemporary world system has a history of at least 5,000 years. The rise to dominance of Europe and the West in this world system is only a recent – and perhaps a passing – event. Thus, our thesis poses a more humanocentric challenge to Eurocentrism. (Andre Gunder Frank and Barry K. Gills 1993b: 3)

Analysts do not manipulate data, though many of them like to think that is what they are doing. Rather, they analyse concepts. Concepts become our friends, even our children. They take on a certain life of their own, and it is tempting to stretch their usage beyond the purpose for which they were created. (Immanuel Wallerstein 1995b: 245)

II.2.1 Circumscription and the iterative model

Chase-Dunn and Hall developed an iterative model of formation and evolution of social structures such as polities and world-systems (Chase-Dunn 1997: 45-46, 101-108; see also Bartley 1997). According to the model, population growth in a given region (e.g. in the fertile valleys of the Nile, Eufrates and Tiger) leads to increasing pressure on local resources. Emigration will follow, unless (or until) the region is circumscribed, either ecologically (e.g. in Egypt by the desert), or socially by the presence of other peoples established in the adjacent regions (e.g. in Mesopotamia with a multicentric city-state system). This will lead to competition for scarce resources, and endemic conflict and warfare.

The impasse is solved either by a large enough death rate and consequent diminished pressure on resources[4], or by the formation of hierarchical structures, that is, states, to reduce the level of conflict by enforcing peace and regulating the use of resources. The formation of hierarchies will contribute to an intensification of the production processes through technological development[5] or greater internal inequalities, and further pressure on local resources, that must be solved by expansion. “There is thus an increase in both the scale of world-systems and the scale of environmental degradation” (Bartley 1997: 374), and there is a long-term trend to increase the size of polities and decrease their number (Chase-Dunn 1995: 117-120). When societies become more complex, new paths of change become possible, and the circumscription and conflict steps can be bypassed, with pressure on resources leading directly to institutional and technological innovation.

II.2.2 Hegemony/rivalry alternation

The polities around the core states are normally in environmentally less desirable locations. Such semi‑peripheries will thus reach their ecological limits faster, and will have a stronger incentive to develop new technologies and to expand militarily against the core states. Further, circumscription of the core states can be so strong (when they create constraints that are beyond their capacity to transcend), that the “old core regions then (become) ripe for the taking by less constrained semi-peripheral actors” (Chase-Dunn 1997: 234). Ekholm and Friedman (Ekholm 1993: 63) argue that the fundamental weakness of the central state is that it must control a production/resource area much larger than itself (one could add: when it is circumscribed), and is hence unstable and prone to collapse.

Within this model, there is naturally a cyclical alternation of hegemonic with rivalry periods. When a state innovates and intensifies production, it becomes dominant (i.e. core) in the region and experiences an economic upturn (called an ‘A’ phase, marked by high levels of infrastructural investment, exchange, and expansion of the world-economy as a whole). When it no longer can overcome circumscription, the neighbouring semi‑peripheral states start to challenge it, and there is systemic conflict between competing upstart centres and the core state(s), associated with an economic downturn (called a ‘B’ phase, marked by decline, economic and political contraction, and reduced economic growth), until a new hegemony is established.

II.2.3 A/B economic phases

This ties in neatly with Frank and Gills’ five thousand year world system, of which “the trinity of core/periphery, hegemony/rivalry, and A/B-phase cycle seem to be constant or at least recurring structures and processes” (Gills 1993b: 188). An important characteristic of these cycles is that they are inter-regional and apparently world system wide, with periods of more rapid economic growth (A phases) being associated with the simultaneous rise of several regional hegemons, that then decline together, often with increased conflict levels (B phases). So, the hegemonic power cycles are correlated with the long cycles of economic expansion and contraction (or at least slower growth) (Gills 1993c: 121).

A tentative dating of the A/B cycles is given in Table I (see Gills1993b: 151-188 for a full description of the cycles, as well as Bosworth 1995: 211-220). For instance, “[b]y the first century AD, the entire world system was politically organized into an unbroken belt of inter-linking hegemonies, stretching from Rome in the Mediterranean basin, to Parthia in Mesopotamia and Persia, the Kushan in Central Asia, and the Han in China”. “From the third through the fifth centuries AD, the previous period of expansion and consolidating hegemonies was followed by a major world systemic crisis on a Pan-Eurasian scale. During this world systemic crisis the Han and the Roman, as well as the intermediary Kushan and Parthian hegemonic structures simultaneously disintegrated” (Gills 1993b: 164, 167).

II.2.4 Interaction networks

The question arises of what should be considered to constitute a world system. Chase-Dunn and Hall argue that world-systems (and we can extend their ideas to the un-hyphenated world system) have four kinds of boundaries, that rarely coincide (Chase‑Dunn 1997: 54, Hall 1996: 10). These form different types of interaction networks that interconnect different parts of the system: bulk goods network (BGN); political/military network (PMN); prestige goods network (PGN); and information network (IN). To these we will add a fifth, the environmental network (EN).

Typically, but not always, the interaction networks are nested as shown in Figure 1. It should be noted however that their boundaries are fuzzy and variable The IN, that includes a variety of elements such as ideology, religion, long distance travellers that bring information and news from distant regions, must be included as a bounding mechanism (Chase‑Dunn 1997: 52, see also Melko 1995). It is sometimes a faint network, and does not often coincide with the other ones.

Nesting of the boundaries of the five networks of exchange (adapted from Chase-Dunn 1997: 54).

The EN also must be included. For instance, global climatic change such as ice ages can induce similar pressures on otherwise unconnected regions. Localised ecological factors can also lead to widespread systemic effects, for instance, by causing local shortage of essential resources leading to migration or invasion of other regions (with possible cascade effects). Further, the environmental impact of the human activities in one region can affect far away regions (even otherwise unconnected). Gills and Frank point out that the world system has ecological origins (Gills 1993a: 82,83), and Goldstone argues that (at least in the early modern world) political and social stability (or instability) depends crucially on demographic trends, that he calls ‘ecological cycles’ (Goldstone 1991: 390-393, 459-460).

There is strong disagreement between theorists about which interactions are required to define a system. For Amin, there is not sufficient evidence to assert whether or not there was a world system in the Iron Age (Amin 1993: 256). Chase-Dunn and Hall, amongst others, argue that exchange of bulk goods, as well as political-military interaction, must be present to form a world-system. Further, these must be regularised and two-way (see Chase‑Dunn 1997: 52-55 and references therein). Ekholm and Friedman, on the other hand, consider that control of prestige goods plays a vital role in primary state formation, as “possession of extremely ‘valuable’ commodities … makes it possible (for the ruler) to accumulate a disproportionate part of the production of the larger system”, as well as “to maintain his nodal position” (Ekholm 1993: 61, 67).

Gills and Frank go further, as they argue that the PGN is as important, and sometimes even more important, than the BGN, in defining systemic relations, because it is an inter-elite exchange that embodies the social relations of production and reproduces the division of labour and the class structure (Gills 1993a: 93-94). Further, Gills considers that “regular and significant” trade is a sufficient ground to speak of a “world economy”, as it “brings in its train systemic political, social, ideological, cultural, and even religious rhythms” (Gills 1995: 145, emphasis in original). As bulk goods (wheat, fish, oil, wine) only entered world trade sometime between 825 and 375 BC (Wilkinson 1995a: 64-66), Gills is implicitly meaning trade in prestige goods before that date.

II.2.5 Some evidence on the existence of a five thousand year world system

This discussion is important to clarify the concepts at hand, however, one should examine available evidence on world systemic interconnections and try to establish a – necessarily preliminary – interpretation, which we will now do.

Bosworth used demographic data to validate Frank and Gills’s dating of A/B cycles (Bosworth 1995). He used the population of the 25 world’s largest cities, as well as the variations with time and geographic shifts, as a barometer of economic activity. Data exists (see Bosworth 1995: 209 for the source) from 2250 BC at intervals of 150 to 200 years until 1000 AD, then every 50 years, and every 25 years after 1800 AD. After a few small corrections to the dating of the cycles, Bosworth concludes that “cyclical, pendulum-like periods of expansion and contraction are indeed evident in population data” (Bosworth 1995: 226), finding moderate or strong support for 11 phases, mild support or inconclusive data for four phases, and contradictory only for one phase. The results are summarised in Table I.

Wilkinson also used city population data to test Frank and Gills’s dating of A/B cycles, using different data sources (Wilkinson 1995a: 54-63, also quoted in Bosworth 1995: 225). He tested specifically the B phases, finding the data consistent with five, ambiguous with respect to two, and contradictory with one phase only. It should be noted that his own dates of incorporation of some regions into his ‘Central Civilisation’ diverge from Frank and Gills’ dating, because he uses more stringent criteria (such as an unbroken chain of entrepôt cities) (Wilkinson 1993 : 241, Frank 1993c: 298)

Eckhardt used data from several independent sources to establish relations between civilisations, empires and wars from 3000 BC (Eckhardt 1995). He used the number of ‘geniuses’[6] to measure ‘civilisation’, the geographic area occupied to measure empires, and the number of battles to measure wars. He found (not unexpectedly) a high degree of correlation between the three parameters. More interestingly, he found that, for any given area, there is a high degree of temporal variation, described as “somewhat cyclical” (Eckhardt 1995: 92), with no clear trend visible. His results for Europe are shown in Figure 2a). This temporal variation corresponds roughly to the ‘rise and fall’ of empires. On the other hand, there is a clear global trend, with the total number of ‘geniuses’, empire area and number of battles showing the same pattern of steady increase, as shown in Figure 2b). This is an indication of global, world-wide interconnections between ‘rising’ and ‘falling’ polities. Further, there seem to be changes in the rate of increase in two different periods, the first around 1000-500 BC (which is, coincidentally or not, close to when the BGN was established), and the second around 1500 AD (which is, coincidentally or not, when the PMN, BGN, PGN and IN started to become concurrent and world-wide).

In short, there seems to be some evidence for both a five thousand year world system, and for important changes in that system. One should however be careful, inasmuch that it is the concepts used and not the data itself that determine the interpretation of the data.

Eckhardt’s empires (circles), geniuses (squares), and wars (stars). a) Europe b) Global

II.2.6 The fall of the East and the rise of the West

Abu-Lughod described in detail the Eurasian systemic connections in the period 1250-1350 AD, which she considers to be a world system on its own, separate from the previous or subsequent ones (Abu-Lughod 1989, 1993). Not wishing to enter into a detailed discussion, for our present purposes the most relevant point of Abu-Lughod’s work is her assertion that the fall of the East preceded the rise of the West after 1500 AD:

The failure to begin the story early enough has resulted … in a truncated and distorted causal explanation for the rise of the west. I hope to correct this by beginning at an earlier point when the outcome was far from determined. The time between the thirteenth and sixteenth centuries marked the transition, and geopolitical factors within the rest of the world system created an opportunity without which Europe’s rise would have been unlikely. (Janet L. Abu-Lughod 1989: 20)

In the thirteenth century, feudal Europe was a marginal participant in the world system. Hegemony was firmly in Asia, with China foremost amongst the regions of expansion (Gills 1993b: 176-178), mostly because it linked the northern overland trade routes with the Indian Ocean sea trade area (that connected to the Mediterranean via the Nile-Red Sea corridor and the Syria Mesopotamia-Persian Gulf corridor) (Abu-Lughod 1989: 347, Gills 1993a: 88). The northern Italy city states (Venice, Florence, Milan and Genoa) were at one end of the major trade routes, over which they had no control, and through which they had (some) access to the riches of the core. Europe imported manufactured Chinese goods (silk and porcelain) in exchange for bullion (Chase-Dunn 1997: 191).

With the fourteenth/early fifteenth world systemic crisis, the East was temporarily in disarray (including a “power vacuum” in the Indian Ocean) (Abu-Lughod 1989: 340-348, 361). It is in this context that Portugal, a small emerging power within Europe, could open the alternative transoceanic route and fill in the vacuum in the pre-existing system in the Indian Ocean, which “was probably the single most important logistical shift in world history and world system development” (Gills 1993a: 90).

In this context, the shift from East to West is a rather typical transition in the world system, in which a semi-peripheral region emerges and challenges the core, as described in section II.2.2 above (page 8). “It must be recognized, however, that the ‘takeover’ of that system was certainly not according to the old rules” (Abu-Lughod 1989: 361). In the next section we will study how the restructured (capitalist) world system works. We begin with the dynamics of capitalism, which we will then rejoin with the theory so far explained.



[4] Which can lead to a vicious circle of warfare and population growth (Chase-Dunn 1997: 104).

[5] Technological development will not always occur, but where it does occur strong (core) polities are formed.

[6] Whose superiority was established by the consensus of encyclopedia and textbook authors, and is likely to be rather Euro- or Westerncentric.