CHAPTER 4 THE EARLY WORLD-EMPIRES In this section we shall discuss the primary world-empires, those areas in which an overarching political apparatus first emerged through conquest to incorporate several primary states. These areas include most of the same ones which we considered in the section on primary states. In the Indus river valley the early states were destroyed by invasions from the north, not conquered and unified into a larger world-system. In Egypt the early states achieved a "precocious centralization," unifying the whole valley of the Nile under a single administration around 3000 B.C. Because of the geographical isolation of the Nile, the Egyptian dynasties did not evolve an interstate system of competing sovereign city-states. In fact there were no large cities in ancient Egypt. Rather the Egyptian civilization long remained a theocracy, a "temple economy writ large" (McNeill, 1963). The empires which emerged in Mesopotamia, Mesoamerica, Peru and China all fit the category of relatively autochthonous developments of primary empires. Along with these we shall discuss a few later empires which seem to have similar characteristics. The evidence we have about most of these world-systems is considerably improved over the early world-economies because, in addition to archeological remains, we have greater numbers of written documents and, in the Meso-American and Peruvian cases, we have the reports of the Spanish conquerors. It is easier to know the boundaries of empires than the boundaries of world-economies, and Rein Taagepera (1978b) has studied the changes in territorial size of the Old World early empires. These changes are depicted in Figure 1. Taagepera includes in an empire all the territories which are paying even symbolic tribute. Of course there may be economic relations with areas outside the political boundaries, both with peripheral raw material producers and with other empires. In general, the economic networks were probably larger than the political boundaries of empires. (Figure 1 about here) Semi-peripheral conquerors The first conquest empire was that of Akkad (Agade) led by Sargon, a former cup-bearer at the Sumerian city of Kish. Akkad was upriver from Sumer, an area in which Semitic former pastoralists had recently taken up irrigated agriculture. Akkadian state-formation was secondary, using elements of organization and productive technology diffused from Sumer. But it did not follow the Sumerian path of development. The stage of theocracy was eliminated. Elements of the pastoral society, especially the great authority of a central leader to command a following among lineage heads, led to the creation of a centralized military state from the beginning. According to McNeill (1963), the relatively egalitarian pastoral society was not well-adapted to the division between a ruling class and a class of farmers and canal-diggers. Rather the Akkadians, recently in charge of herds of sheep, preferred to husband a conquered people. Sargon's son extended direct imperial rule over the Sumerian cities, eliminating the old local aristocrats and replacing them with governors from his own family. Lattimore (1940: 305-306) describes a somewhat similar process in the shift from Shang to Chou hegemony in China: Other groups began to follow, though more slowly, the same line of evolution as the Shang, some of them becoming tributaries while others remained independent and hostile though imitating as far as they could the characteristics that had given strength to the Shang. Some of these groups, which had not been sufficiently advanced in general culture to develop for themselves the technique of working in bronze as the Shang had developed it, were able to take over this technique once it had been developed. This narrowed the gap in military efficiency between what had now become backward or "less-Chinese" groups and the progressive Shang or "more Chinese" group.... As the backward groups became more formidable in war the most advanced and cultured group became more vulnerable. War, for the Shang people, came to mean not only the raiding of others but the protection of their own slaves, granaries, wealth, and lands. The difference between slave cultivators and ruling warriors meant that interest in war was not equal throughout the society. More backward groups with fewer slaves and more free warriors, even though they were less "noble" than the Shang aristocracy, became able to hold their own in war against the Shang.... When this equalizing process had gone far enough it was only a matter of time until the Shang would be overthrown by a cruder people who were not necessarily invaders from beyond the horizon of what was now definitely the early Chinese culture, but a people already belonging to the Chinese culture, although not to its most refined and advanced form. Peripheral peoples sometimes overran civilizations which were in decline, despoiled them, and left only a scattered remnant of the former hierarchical society. This happened in the Indus river valley. In Mesoamerica the cities of Teotihuacan, Tula and Monte Alban may have suffered similar fates. As we shall see, later world-empires were conquered by semi-peripheral peoples, "high barbarians" McNeill ( 1983) calls them, who succeeded in establishing themselves as the new ruling class. The pattern for the early empires suggests that it was semi- peripheral states that succeeded in unifying the early world-economies and their interstate systems. In the archaic terms often employed by historians, barbarians, after all, are not savages. At this point we need to discuss the meaning of the term "semi- periphery." This term has been used by world-system analysts in a number of different ways. I have counted seven different meanings in the definitions and usages in recent literature.9 In this discussion of historical world-systems I would like to employ it in a way which helps us understand the interaction between societies at different levels of development. Part of our definition of core and periphery (p. 41 above) includes the distinc- tion between societies at different levels of socio-cultural complexity. The actual content of that distinction varies greatly as we compare different types of world-systems. By semi-periphery I mean something similar to Trotsky's (19) "combined and uneven development." Some societies that are changing continue to have features of their earlier level of development while simultaneously adopting features from other societies that have previously attained a greater level of socio-cultural complexity. Of course it matters exactly what these combined features are as to how they affect the semi-peripheral society's nature and interaction with other societies. We have mentioned the cases of Akkad in Mesopotamia and Chou in China. The Chinese example continues with a shifting back and forth between "marcher states" nearer the frontier and lowland centers which led "nativist" reactions to the upland conquerors. The Chou dynasty, attacked by new barbarians in the West, moved its capital further east toward the Great Plain and then declined. It was followed by the Ch'in, another semi-peripheral marcher state from the West that succeeded in founding the first unified Chinese empire. The Ch'in conquered the Great Plains but never completely subdued the Yangtze valley state of Ch'u. Drawing on resistance from the Huai and Yangtze peoples, the Han dynasty (firmly based on the Great Plans) was founded in B.C. 206 (Lattimore, 1940: 311).108 The Peruvian case provides another example. Irrigation, a highly productive fishing economy, and state-formation first occurred in the coastal river valleys, but the states which created early empires were the highland Huari and Tiahuanaco. Service (1975: 197) describes these as "cultural outposts" of the developed coastal region, and ascribes their motivation for conquest as based on the need to control trade in the context of a coarse-grained ecology that linked the highlands with the coast. McNeill (1963) describes a similar force driving hill peoples near the Mesopotamian heartland. As they became more specialized as pastoral and agricultural producers for exchange with the valley agriculturalists they came to depend on imports for their livelihood. When political disorder disrupted the exchange networks, these people were motivated to conquer the lowlands in order to control the exchange network of which they had become a part. Murra's (1980) study of the Inca economy and polity (a later and larger highland empire in Peru) supposes that the Incan marketless economy had been imposed to regulate unstable market exchange among the ecologically differentiated areas. If Service (1975) is correct, the Egyptian case fits the pattern of primary empires being created by semi-peripheral states. He claims that both the marcher state based in Thisis which first unified Egypt under the Old Kingdom, and the lord of Thebes which reunified Egypt to create the Middle Kingdom were from the south, (Upper Egypt) which he characterizes as a "relatively backward area" (Service, 1975: 235). McNeill (1963: 81 Note 22) is responsible for this idea, and he further claims: Moreover, if tradition is correct in identifing the first king of Upper and Lower Egypt as a pastoralist conqueror from the south, then the truly extraordinary prerogatives of the later Pharaohs can be seen as resulting from the application to conquered agricultural populations of the necessarily broad powers assigned by any successful pastoral community to its leader.11 The Mesoamerican case may also fit this pattern. The early states of Teotihuacan and Tula fell in part because of invasions of peripheral peoples (Chichimecs) from the northern deserts. After the fall of Tula the valley of Mexico lost population and no central state emerged to integrate the already advanced horticultural economy for around 250 years. One of the Chichimec groups that moved into the valley was the Mexica, originally a group of tribal hunter-gatherers. These people are known to us as the mighty Aztecs, and as Blanton, et. al. (1981: 152-153) put it: They rapidly adopted the behavior and technology of the valley's more urbanized dwellers, however, and soon rose to political prominence. The story of one of these immigrant groups--the Mexica--is one of the greatest tales of epic progress from rags to riches ever recorded. The Mexica were originally considered to be so undesirable that they were forced to live on a series of low, swampy, bug-ridden islands in the middle of brackish Lake Texcoco. They survived by exploiting the meagre lacustrine resources, and by selling themselves as mercenary soldiers. Through cunning and generally shabby behavior they eventually established themselves and their island city of Tenochtitlan as a force to be reckoned with. Within a century after its founding Tenochtitlan had become an independent city, and from then until the Spanish conquest it carried on an aggressive policy of military expansion and alliance which created a large empire extending far beyond the boundaries of the valley of Mexico. Although the Mexica could not be described as semi-peripheral at the time of their arrival in the valley, after 100 years of adopting "the behavior and technology of the valley's more urbanized dwellers" they might easily be so considered. And though not pastoralists, their egalitarian kin mode of integration may have given them the type of military advantage over their more stratified neighbors which Lattimore ascribes to the Chou. Before we consider explanations for this pattern of semi-peripheral conquest let us describe some of the other features of the early empires. Political structures in the early empires Most of the early empires were more centralized than the smaller states of the early world-economies had been. Although they often adopted many of the religious institutions of the civilizations they conquered, these were generally under the firm control of the political-military apparatus. Diakonoff (1969) reports that, during Sargon's reign, the temple estates of Sumer fell into the king's possession. Sometimes the ruler was both commander-in-chief of the armed forces and the head priest, or as in Egypt, god himself. The apparatus for controlling conquered states sometimes took the form of indirect rule, granting of imperial titles to the indigenous rulers (Incan), or grants of land to loyal local leaders (Chou). In other cases direct rule was created by the elimination of indigenous oligarchs and their replacement by members of the emperor's family or close associates (Akkad). The Incas declared that all the land in the empire belonged to the Inca, and then parcels were "given" to local lineage heads who were loyal (Murra, 1980). The democratic forms which originally characterized the relationship between a semi-peripheral king and his own cadre usually evolved into an absolutism as the empire was consolidated. The Mexica exhibited a radical restructuring of formerly equal lineages into a class society of nobles and commoners as their success in war led to empire (Padden, 1967). Some of these societies developed written legal codes to help administer a centralized law across different regions brought together by conquest. The Babylonian code of Hammarabi is the most famous. The Akkadian empire, probably the most centralized of the early empires, was short-lived, while some of the least centralized lasted a very long time. Chou granted "feudal" titles to indigenous local leaders, and yet its hegemony lasted for hundreds of years. Although historians refer to the Aztec empire, the Aztecs always ruled in alliance with other city-states in the valley of Mexico. And, while Tenochtitlan was the political-military center, Tlateloco, a neighboring city, was the market center of the Mexican world-system. The relatively weak and somewhat illegitimate character of Aztec rule may partly account for the politics of terror by which the Aztecs attempted to maintain the subservience of their own commoners and conquered states.12 The unbelievable orgies of human sacrifice surely must have stemmed from the inability of the Aztec center to control a burgeoning political economy. This weakness is further indicated by the amazingly rapid conquest of this great empire by a handful of audacious Spaniards (Padden, 1967). Economic institutions in the early empires Here again we face the substantivist/formalist controversy. Polanyi (1957) claims in the strongest terms that there were no price-setting markets in Babylonian times or before. His discussion of marketless trade carried on by merchants in the context of state-administered agreements about exchange rates is illustrated by the case of an Assyrian trading post. He finds merchants who are paid a commission on exchanges that they under- take at set prices. They gain by increasing their volume, but not as a result of speculating on price changes. Thus they engage in trade which is risk-free with respect to price changes. Strong support for the existence of non-market centralized economies is given in studies of the Inca economy (Murra, 1980; La Lone, 1982). The Incan state erected large storehouses for the accumulation of food and cloth produced by the labor services of community members on state and church lands, as well as by a staff of state-workers. A specialized group of women weavers directly employed by the church produced fine types of cloth used in ritual ceremonies. All those donating labor services to the state or church were fed and clothed during the period of their labor out of rations from state or church stores. The immense magazines, overflowing with stored goods, found by the Spanish conquerors throughout the Incan empire, were utilized as a method of state finance for public works and military efforts. They did not play a role in the provisioning of inter- regional trade for the consumption of commoners. The subsistence of commoners was provided by community redistributional mechanisms which had predated the Incan state. As has been mentioned above, lineages in one ecological zone had early colonized other zones where lineage members produced the products which could only be obtained within each of the several ecological zones of the "vertical archepelago." These goods were then reciprocally exchanged within each lineage. The Incan state adapted a version of this form of production and exchange to its own uses, sending colonists to intensify production in newly conquered territories. Murra (1980) contends that those who have understood the Incan political economy as a socialist welfare state have mistakenly attributed the redistributional communities which continued to function within the empire to Incan invention. The capital city of Cuzco was inhabited only by members of the royal family, and weekly distributions of tributary goods to the royal lineage in the main square were mistakenly identified as markets by the Spanish conquerors. LaLone (1982) points to the importance of reciprocity as an ideology at the highest level of the Incan state. The Inca himself, a priest king, son of the Sun, participated in ritual labor with the planting of the first maize. Godelier (1977) contends that religious ideology was an "internal element in the relations of production" in the Incan political economy. As has been mentioned above, the Incan political system was a form of indirect rule in which the kings and chiefs of local states were given officials titles and were "given" land which had formerly been theirs but which, under the Incas, was defined a belonging to the state. These local kings and leaders brought tribute to Cuzco and in return were given prestige goods, such as the clothes of the Inca and other items of high status. Murra (1980) contends that the Incas required primarily labor services rather than tribute in kind. This principle is reflected in the way in which state farms were worked. A broad field in one of the mountain valleys was divided into strips and each strip was, in theory, the responsibility of a different lineage. Lineage heads and village leaders were responsible for sending a crew of laborers to meet their corvee obligations. This system, called the mita, was utilized in modified form by the Spaniards to provide labor for the mines at Potosi', where over eight million "Indians" met their deaths digging and transporting silver for the European conquerors (Frank, 1978). But this runs ahead of our story. Another non-market economy in an early empire is the Egyptian case, described by McNeill (1963) as a "temple economy writ large." Because of the geography of the Nile valley, with the main transport being the Nile itself, control of exchange among regions became an important basis of support for the state. As McNeill (1963: 71, 74) says: By controlling shipping, the king automatically and easily regulated all major movements of goods and people, and therewith possessed the means for effective rule over Egypt.... Trade and all large-scale economic enterprises were managed and controlled by officials of the divine household. At intervals the Pharoahs dispatched expeditions of a semi- military character to Syria for timber, to Sinai for copper, to Nubia for gold, to Punt for myrrh; while within the country similar enterprises brought granite and other special stone from Upper Egypt and gathered taxes in kind into the royal treasuries. So far as thee goods were not utilized directly by the god-king for the maintenance of his court or for the construction of his tomb and other buildings, they were distributed among the courtiers of the royal household. The students of Polanyi who explored ports of trade and administered trade (Chapman, 1957) among the states of Mesoamerica have been supplanted by a new wave of formalism using evidence not available before. Blanton, et. al. (1981) strongly emphasize the ebb and flow of market mechanisms versus state control in their studies of the valleys of Oaxaca and Mexico and the lowland Mayan area (see also Hassig, 1985: 67-84). They ask whether or not a profit motive existed in prehispanic Mesoamerica (Blanton, et. al. 1981: 38-39)13 and their answer is in the affirmative. As evidence they put forth the following: Why for instance, would cacao-bean money have been counterfeited? Judges in the markets punished those who committed fraud in transactions. "Good" merchants were distinguished from "bad" ones. The latter profitted excessively by selling fake stones as precious stones, or bad-tasting, old tamales as fresh ones, and so on. This sort of thing sounds very much like "business as usual," and certainly provides overwhelming evidence for a profit motive. Curtin (1984: 70) weighs in with a further critique of Polanyi's characterization of marketless trade in the Assyrian trading post of Kanesh: Karl Polanyi took the trading post at Kanesh as a prime example of marketless trade, a cross-cultural "port of trade" where prices were not allowed to fluctuate but were fixed by treaty. Unfortunately, he had to work from the early and incomplete studies of the Kultepe tablets. Even more, he was led astray by the dominance of the now-displaced hypothesis of a Mesopotamian temple economy. All later authorities who have studied the Kultepe tablets portray an economic order where merchants were extremely sensitive to small changes in price, where prices fluctuated rapidly, and where political authorities interfered in trade as much as they could to suit their particular interests--but never to the extent of successfully suppressing the underlying play of supply and demand, thought here as elsewhere in the history of commerce, protection costs were an inescapable cost of doing business. One possible interpretation of the above flows from Murra's analysis of the Inca state. When early empires were strong, as in Peru and China, they controlled the internal economy and foreign trade. When they were weak, as in Mexico and the Middle East, or rather when the density of economic inter- action was larger and more multicentric than the state apparatus, market systems emerged to facilitate inter-regional trade. Local markets, where petty trading of locally produced goods occurred, and international market trade of the kind described by Curtin, may have predated the emergence of price-setting markets in the intra-regional (or "national") scale. Blanton (et. al., 1981) suggest that states are weakened by the existence of strong market networks, which create the possibility for accumulation outside of their control. This idea is also expressed by Ekholm and Friedman in their discussion of Mesopotamian differentiation. Private accumulation beyond the purview of the state is thought to threaten the ability of the state to obtain needed resources. This kind of analysis is supported by Diakonoff (1982) who notes that commodity production became more prevalent in Old Babylonia. This led to the replacement of communal lineage-based forms of organization by more individualistic types of property. "Territorial communities became no longer a collective proprietor, but rather a collective of proprietors" (Diakonoff, 1982: 34). But when political order later became more problematic communal forms of organization reemerged. Rich land owners began to seek foreign labor--captives and refugees, who enlarged the class of slaves and dependent workers. The workers on royal lands were paid subsistence rations or alloted only small subsistenance plots. In Old Babylon, Diakonoff (1982: 79) surmises that "a certain inner commodity market for household objects and products must have existed on a small scale." During periods of rising commodity production slave labor increased, along with private enterprise in "international" and wholesale trade. Farber (1978) shows evidence of cyclical fluctuations of prices in Old Babylonia. But in the absence of a generalized monetary economy credit was the main economic form which was stimulated, and usury flourished despite the attempt by Hammarabi to control it. According to Diakonoff, usury stifled economic growth, which led to contraction, political instability and the reemergence of family-communal type of organization, as well as a scaling down of international trade. Core/periphery relations in the early world-empires A discussion of empires and imperialism necessarily implies core/ periphery relations. As we have seen above it was, in several cases, semi-peripheral states which succeeded in unifying the early world- economies by conquering the old core areas and extracting tribute. We have discussed the various ways in which control and exchange relations operated within the boundaries of these empires. The institutional nature of core/periphery relations is also varied in the early world-empires. We have mentioned the colonizing efforts of the Incas, who sent new people to develop intensive agriculture in peripheral areas. Scarce raw materials were obtained by the method of colonization. Military incursions into peripheral lands also were utilized to obtain scarce raw materials. The Egyptian case has been mentioned above. During the Old Kingdom a colony was established just below the second cataract of the Nile in Nubia to smelt copper. Nubians were apparently employed to transport copper ore to the smelter, but the smelting was done by Egyptians and mineral prospecting was carried out by official agents of the Pharaoh (Adams, 1977: 170-174). Adams claims that Egyptian imperialism in Nubia during the Old Kingdom was "characterized at first by sporadic and uncoordinated raiding and trading expeditions into the southern lands. ...no effort was made to extend Egyptian political control or to establish permanent relations with the Nubian people, except perhaps for some frontier chiefs in the immediate vicinity of Aswan" (Adams, 1977: 165). It was under the Middle Kingdom that there was: a period of armed trade monopoly, operating through one or more established trading posts in the interior. Its main concern was not the subjugation of territory or of the native population, and production (except in the case of minerals) was left in Nubian hands. Animal and forest products, which were obtained through subsidization of native suppliers, meaning in all probability local rulers. There was no significant movement of Egyptian settlers into the southern lands. However, an enormous military effort was devoted to the protection of the trade routes to the south, and the assurance of a complete Egyptian monopoly of the trade along them.... Finally, the New Kingdom saw the extension of imperialism from the economic to the political sphere. Outright Egyptian control was extended over the Nubian territory and people, displacing or subordinating the native rulers with whom the Egyptians had formerly been content to deal. Control of raw material production, and probably also of agriculture, passed directly into Egyptian hands, and the Nubians in their turn became fellaheen (Adams, 1977: 165-166). As Service (1975) and McNeill (1963) have pointed out, the main problem for the civilizations of the river valleys was not how to exploit peripheral peoples, but how to fend them off. The accumulation of wealth was greater than ever before, and this constituted a license to steal for peripheral nomads. This led to a racist ideology of "subhuman barbarism." As Cooper (1983) points out, the Mesopotamians did not generally vilify different ethnic groups, but: There are two important exceptions to this absence of ethnic stereotypes: the Gutians and the Amorites, who represent the nomadic and seminomadic hordes of the mountains to the north and east and the desert to the north and west, respectively. Both are described as subhuman, but in a somewhat different terms. The Gutians are characterized as savage, beastlike imbeciles, whereas the Amorites are usually curious primitives, less horrible, if every bit as threatening militarily, than the Guti. Even though the Gutians and Amorites did conquer, temporarily, in the lowlands their lack of familiarity with civilized culture made their rule short. Of course the priests interpreted this in terms of their failure to understand how to worship the gods, much as the failures of more civilized dynasts were attributed to impiety. The maintenance of power required political savvy as well as military prowess, and in this the nomads were lacking. The above discussion of early Egyptian relations with the Nubian periphery supports Diakonoff's (1969: 28) assertion that the early empires relied more on trade and colonization to obtain raw materials than direct political subjugation and tribute-taking. According to Diakonoff it was the later "great warrior empires" that came to depend on unequal exchange supported by conquest. The early empires did not fill the larger economic networks of which they were a part, and their relations with raw material producers were primarily ones of equal exchange. Of course there were some conquests, slave capturing and booty-taking. But these were not yet fundamental to the mode of production. Ross Hassig's (1985: Chapter 6) description of the valley of Mexico under the Aztec empire appears to be an exception. He describes how the Aztecs used political coercion to extract tribute and to promote a structure of market exchange which concentrated "secondary" manufacturing at Tenochtitlan and "primary" production of less processed goods in secondary centers. Hassig implies that the Aztec policy successfully deindustrialized secondary centers, concentrated goods in the core such that prices there were low for many goods, and facilitated the further specialization of manufacturing production in the core (Tenochtitlan). As Hassig (1985: 130) puts it: Tenochtitlan altered this pattern by encouraging primary production throughout the Valley of Mexico while simulta- neously discouraging secondary production. It did so by controlling the flow of tribute and by exercising mono- polistic control over much pochtecah trade, rendering the valley cities incapable of effectively competing with the secondary production of Tenochtitlan for several reasons. First, overall craft specialization in these cities was weakened by the influx of competing tribute goods from the outside. ...These goods consisted primarily of finished goods that could be rechanneled into the valley economy through the markets. The cost of tribute goods to the recipients was nil. There were, of course, both military and administrative costs involved in tribute, but the costs weighed more heavily on the commoners than on the nobility while benefiting the latter disproportionately. These tribute goods could be sold at or below the cost of production of similar wares produced in the Valley of Mexico. It is impossible to say how important these forms of core/periphery exploitation were for the Aztec political economy, but the interaction between tribute from conquered hinterlands, and the creation of a more hierarchical division of labor in the valley of Mexico is a fascinating instance of the use of both imperial and market mechanisms to restructure a core/periphery division of labor. Is this an exception to the general- ization that core/periphery exploitation was not very important in primary empires? Perhaps. But maybe the Aztec empire is better placed in the next category, that of secondary empires. Both Teotihuacan and Tula had been the centers of previous empires in the Valley of Mexico. The interstices between empires contained neutral ports of trade, and autonomous states that became engaged in trade with different empires. In some areas early states specializing in maritime trade emerged, as the Minoans on Crete. What can we make of the frequency of semi-peripheral conquerors who created the primary empires? As we shall see, the later warrior empires were most often conquered by peoples who had adopted certain military techniques such as the chariot, or developed new techniques, such as the production of iron weapons. To describe the above conquerors or the later ones as semi-peripheral instead of peripheral is certainly problematic. This may be a false problem which results from our inability to be precise about boundaries between zones within a world-system. Even in the modern world-system no one has yet done much of a job distinguishing empirically between the periphery and the semi-periphery. If my impressions are correct that early conquerors were more semi-peripheral than peripheral, this may partly be due to the fact that, as much of the evidence above has indicated, the core/periphery hierarchy within the primary world-economies was not well institutionalized. Friedman and Rowlands (1977) note that a prestige good system is difficult to monopolize and thus encourages competing centers that grow at the expense of the original center. And the lack of firm imperialism, with the predominant reliance on colonization and equal exchange with peripheral producers, may have also produced only a weak operation of the processes of the "development of underdevelopment." Nevertheless, something like a process of combined and uneven development seems to characterize the pattern by which marcher states formed the primary empires. The use of the term "semi-periphery" in this context means a kind of uneven and combined development in which late developers going through state-formation were able to utilize less complex forms of solidarity among lineage heads and their clan followers to forge a strong state capable of unifying a number of "more developed" (more complex, differentiated and stratified) primary states. While we are on this subject we may as well point to an important difference between the kind of semi-peripheral upward mobility we see in the early empires and that which occurs in our own global political economy. In our system, at least so far, conquest of the core by upwardly mobile semi-peripheries has not occurred, although it has been tried. It has been rather the powers which combine military prowess with competitive commodity production for the world market which became hegemonic in what then still remains a multicentric world-economy. The reasons for this difference will be discussed below.