World Systems Theory, Core Periphery Interactions and Elite Economic Exchange in Mississippian Societies

World Systems Theory has been one approach used to explain the rise of the Mississippian social and political phenomenon. In this paper it is argued that a hierarchical model of core- periphery interaction docs not explain the Cahokian phenomenon, because several crucial clements of such a model cannot be demonstrated to have existed within the Mississippian system. It is suggested that looking at Mississippian society as a differential core-peripheral system may have utility as a framework for including concepts such as gateway communities and interacti on spheres previously used to describe the economic interactions between Cahokia and its neighbors.


Introduction
Archaeologists have long sought an explanation for the rise of Mississippian society in the major river valleys of the American Midwest and Southeast between A.D. 1000 and 1500 (Smith 1978). Over the years, our explanations have changed with the changing fashions of then-current theories of cultural evolution. From diffusion to cultural ecology to economic models ofredistributivc exchange, we have attempted to put a finger on the causal variables involved in the production of certain ceramic wares, the construction of earthen platform mounds, and the large and highly organized residential and ritual sites of these people.

Journa I of World-Systems Research
None of these models have proven completely satisfactory, failing at one level or another to account for the complexity of Mississippian intergroup int eractions revealed by the archaeological record. With the emergence of World Systems Theory (WST) (Wallcrstcin 1974) within the anthropological community, it is only natural that archaeologists should attempt to use it a~ a way to gain some insight into this longstanding archaeological problem (cf. Peregrine 1991). In this study, I will attempt a critical examination ofWST a~ it might pertain to Mississippian society. A brief review of what we know about Mississippian social and political economy is necessary before we explore how well WST works to explain what we sec in the Midwest and Southca~t United States between AD 1000 and 1500.

Traditional View of Mississippian Social-Political Economy
Middle Mississippian is a term used to describe an archaeological culture that flourished in the major river systems of the midwcst and southca~t United States between A.D. 1000 and A.D. 1500 ( Figure 1 ). Middle Mississippian society is traditionally viewed a~ a ranked level society (Phillips and Brown 1978). The social system is seen a~ pyramidal, with ruling elites at the top, a mid-level grouping of semi-elites and a larger population of non-elites. Data for this demographic make-up arc provided by studies ( e.g., Peebles and Kus [1977]) that have demonstrated a correlation of grave goods and health indicators with spatial location in cemeteries. In addition, the distribution of exotic or non-local artifacts in graves is usually skewed towards males, and there may be a correlation of materials within genetically related clusters of individuals. 9 Second·l ine site @ ~i rd/founh~l inc Sile figure 1. American Bottom Core Arca. \ l11e subsistence economy i~ generally depicted as a maiz.e ( Zea lllays) and sqnash ( CiJl1J rbita /l<JW)) agriculture snpp lementcd by vvild plant gathering and hunting. Plants such as M:aygrass ( P1wk1rl~ camlir. f(IIW ). sunflower {,flelkmthus ar.r.u,, ). Goosefoot ( C11er.1)1wxliulll sp.). erect kno~veed (P()l)'g()r.Ulll eredulll), little barley(fi()rdeum pusill1m1) and berries (e.g .. Vi,,"·ir.ium sp.) were all part oftl1e fvfississippian diet. Beans (Phaseolus sp.) arc a later introduction in some areas of the Mississippian world, but arc not found in the American Bottom. Tobacco (Nico tania mstica) was grown for ritual or recreational purposes (Parker 1987). While irrigation was not used, at least some fields were improved through the use of a raised field technique to aid in drainage and frost protection (Riley 1993). Aside from the dog ( Canisfamiliaris), no animals arc known to have been domesticated.
There is biological and archaeological evidence for warfare. In addition to skeletons bearing evidence of violent death (Milner ct al. 1991), some Mississippian sites arc stockaded (Goldstein and Richards 1991), and there arc symbolic representations of pottery and shell engravings, suggesting that a warrior class or at least some form of warrior veneration existed (Phillips and Brown 1978). The archaeological evidence is supported by cthnohistoric data from southeastern groups such as the Natchez, who were following this basic pattern at the time of European intrusion ( Stcponaitis 1978). The cthnohistoric data do give the impression that high-level elite s at larger sites exerted influence, if not control, on individuals at other locations. However, none of the historically known groups approach a state-level of social and political complexity. In particular, complex bureaucracies with the power to coerce taxation and draft an army were not features of the social and political structures of these groups. It is still a matt er of debate whether the Mississippian world of 200-400 years earlier contained an incipi ent state-level sociopolitical system.

[Page 3] Journal of World-Systems Research
World Systems Theory and Mississippian A" defined by Wallcrstcin, the world system<; perspective cmpha<;i zcs the a<; ymm ctrical political and economic exchange between a highly developed core and a lesser developed periphery. The core is highly developed both economically and politically, with centralized authorities supportin g an exchange system that encourages the accumulation and investment of surplus (Stein 1993). These elites control the flow of goods betwe en the core and periphery through colonial administration or control of local elites, who arc dependent upon core elites for their own power. The periphery provides a flow of staple good<; and raw mater ials to the core in exchange for value -added or finished commodities.
A number of scholars (e.g., Cha<;c-Dunn and Hall 1991; Schneider 1991) have argued that Wallcrs tcin's initial formul ation of WST , designed to explain a European capitalist environment, is probab ly inadequate for non-state societies. For one thing, Wallcrstcin's view a<;sumcs that there is inequality inher ent in the core-periphery exchange, but Cha<; c -Dunn and Hall would like to sec WST account for those situations where exchange inequities arc not readily apparent. In order to provide for a more flexible approach, Cha<;c-Dunn and Hall offer a typology of possible world systems forms covering sociopolitical-cconomic situations from band-level, kin-ba<;cd lineage systems to fully industrialized state-level capitalist systems. To denote the more generalized approach, they favor the use of the term core-periphery or even more generically, intcrsoci ctal interactions, rather than Wallcrstcin's world systems. Their eclectic approach to world systems allows one to operationalize expectations for what a world system would look like on a ca<;c by ca<;c ba<;is. Again, in order to be flexible enough to cncompa<;s a<; much variation a<; possible under the rubric of core-periphery, they break core-periphery interactions into two larger groups: core-periphery differentiation, where a large group interacts with a smaller group; and core-periphery hierarchy, where the core demonstrably dominates the smaller group economically, militaril y, politically, or ideologically.
Based on the previous discussion of what we think we know of Mississippian society, our ca<;c study here falls under Cha<;c-Dunn and Hall's taxon of chicfdom--a non-stat e, but stratified society. There is probably little controversy on this point, but we need to determine whether the interactions between Cahokia and surrounding smaller politics wa<; In a hi erarchical system, the core creates the periphery by pulling it into the exchan ge system a<; a politically and economically depend ent area (Stein 1993). This creation of a dependent area impli es some form of coercive power over the peripheral area, either through military threat or some entic ement so powerful that the population in the targ eted area accepts their subservient role in the exchange system . In effect, the core must extend to the periphery an offer that they cannot refuse.
Stein suggests that a hierarchical system requires three ba<;ic assumptions about the relation ships between the core and periphery. First, there must be a fundamental power a<;ymmctry such that the core can dominate the periphery. Second, the model a<;sumcs that a<; a result of this power asymmetry, the core can control an exchange system crucial to its ex istence. Third, this exchan ge system must structure all other a<;pccts of th e economy in the periph eral society. In the particular ca<;c presented here, we envision a chiefdom level socie ty, with a small cadre of ruling elites ba<;cd in Cahokia who have political, economic, and marriage alliances with each other a<; well a<; with elites in outlying area <; of the American Bottom and beyond. For a hi erarchical mod el of corcpcriphcry relations to work, we need to show that the core elites had the capabi lity to coerce other elites, a<; well a<; non-elit es, into an unequal system of economic exchange. Such coercive power could conceivably consis t of mili tary, economi c, or ideologica l control of access to desired and/or necessary resources. We can examine Mississippian social relations at Cahokia to sec if our example meets these three criteria for a hierarchical core-periphery relationship.
World Systems Theory and Mississippian: Does it Fit'?
When placing Mississippian culture into the taxon of core-p eriph ery int eraction, the problem of area bounding ( i.e., the geographical extent of th e system) is imm ediat ely apparent. The problem of bounding core /periphery hierarchie s has been di scuss cd by Cha..,c-Dunn and Hall (l99l) and the probl em is a significant one in our particular ca..,c study. Cahokia and it.., immediate environs arc clearly uniqu e and represent a core area. But the extent of the core, and the further extent of the periph ery arc much less definit e.
The distribution of Middle Mississippian sites in the Midwest shows that, out side of

[Page 5] Journa l of World-Sys tems Research
Cahokia and the American Bottom, the lower Mi ssissippi , the centr al Illin ois, the lower Wabash, and the Ohio rivers all contain major habi tations, suggesting a large scale, farflun g geographic extension of Mississipp ian political and economic heg emony (Figure l). It is not entirel y clear, how ever, that the entir e Am erican Bottom would qu alify as a core. It may be that Cahokia and its imm ediate suburb s were a core, with the Am eri can Bottom reg ion a semi-periphe ry and the sites in the Illinois, W abash, and Ohio river valleys the periph ery. However, it may also be that the entir e American Bottom is the core. If Cahokia and the American Botto m is the core, where arc the semi-p eriph ery and periphery? Th e semi-p eriphery and periphery mi ght well includ e those sites in the river valleys ju st mentioned. Maj or sites in this pe riphery or semi-p eriphery includ e the Angel Site in lndian a (Black 1967), and Dickson Mounds in the Central alllin ois riv er va lley (Harn 1980). T here is also the Caddoan area of Oklahoma and Arkan sas, in parti cular the Spiro site (Brown ct al. l 978) . Caddoan sites share many Mi ssissippian traits, yet retain a rcg ionali zation that suggests they mi ght be part of a Middl e Mississippian periph ery, or again, at least semi-p eriphery. As a furt her wrinkl e, there is a much larger exte nsion of Mississippian into the north , cast , and west , which archaeologist s have termed Upp er Mississippian . Sometimes referred to as Cahokia' s Hint erland s ( cf. Em erson and Lewis 1978; Mehrer and Collin s 1995), it is quite po ssible that group s in these areas made up the true periphery of the Mississippian world . T hese arc group s whose pottery shares man y of the same motifs as Middl e Mi ssissippian, but who live at sites which arc more simil ar to Late Woodland villages th an to the hierarchical Middle Mi ssissippian hamlet/town/c ity pattern , and whi ch lack pyramid al mound s . Oneota and Lan gford arc variants of th is Uppe r Mississippian cultural ph enomenon found in northern Illin ois, northweste rn ln diana, Wisconsin, Iowa and Minn esota . These groups were generally less dependent upon maize and more involved in a mixed economy of hunting-gathering and maize horticulture (Brown 1982(Brown , 1990Jeske 1990;Pollack and Henderson 1992). Fort Ancient wa<; a southern variant of Upper Mississippian found in central Indiana and Ohio. The Fort Ancient subsistence regime may have been somewhat different from that of Oneota and Langford, including a significantly greater reliance on maize agriculture, and later inclusion of beans, in the diet (Pollack and Henderson 1992; Watson 1988).
Of special interest is the site of Aztalan in southea<;tern Wisconsin (Goldstein and Richards 1991). This site is a seemingly Middle Mississippian settlement separated from the core by 500 km and surrounded by several different variants of Upper Mississippian and Late Woodland cultures. The site contains a platform mound and exhibits evidence for hostilities with its immediate neighbors. Aztalan ha<; sometimes been considered either a colony of

Journa I of World-Systems Research
Cahokia or a trading center between Cahokia and its northern periphery (Barrett 1933; Gibbon 1974;Griffin 1960). Others (Fowler and Hall 1978) have suggested that Aztalan is not representative of direct Cahokian contact, but is a "hybrid resulting from interaction between Middle Mississippian and [local Woodland cultures]" (Hurley 1975, cited in Goldstein and Richards 1991). In a recent study, Goldst ein and Richards (1991 :206) a<;sert that the site is an example of direct Cahokian contact, although the rca<;ons for the contact arc unclear. The site is located on the Crawfish River, which they argue places it in geographic context for movement of trade goods. Howev er, there is no evidence for trade goods flowing through the site, and there are demonstrably many more strategic area<; in the north if the Cahokians wished to place a settlement for the control of trade goods from the Great Lakes region to the American Bottoms.
In fact, it appears that we may have a nested core/periphery phenom enon (Cha<;e -Dunn and Hall 1991), with Middle Mississippian American Bottom sites, Middle Mississippian Central Illinois-Waba<;h-Ohio Rivers sites, Caddoan sites in Arkansa<;, and Oneota -Fort Ancient-Langford sites oflowa , Minnesota, Wisconsin, north ern Illinois, central Indiana , and Ohio all displaying varying degrees of periphcralization from the central place of Cahokia. Finally , there are hundr eds of Mississippian sites on major waterways of the southea<;tern United States, many of them large regional centers containing platform mounds, plaza<; and organized residential centers (Smith 1978). How ever, it is rea<; onable to a<;sume that Mississippian wa<; a multi central phenomenon ( Cha<;e -Dunn and Hall 1991), and for the purposes of this study, I will treat these southea<;tern sites a<; independen t and separate socio-political unit<; that had their own core-periphery interactions outside of any connections with Cahokia and the Midwestern Middle Mississippian sites. We do know that this is not entirely true ; sea shells from the Gulf Coa<; t are found at Midw estern Mississippian sites. There wa<; some int eraction between Cahokia and at least some groups to the south, although direct contact between Cahokia and Mississippian sites outside of the Mississippi Valley proper is not definite.
Even ifwc ignore those "other Mississippians", taking a world systems view of Midwestern Mississippian society is problematic, but teasingly plausibl e. Looking at Stein's first a<;scrtion about hierarchical systems a<;sumptions, we must posit that the American Bottom area possessed a power a<;ymmctry with its periphery. Judging by the size and complexity of Cahokia and its a<;sociatcd suburbs, a strong argument can be made that it certainly had a huge population relative to other inhabited area<; in the American Bottom.

Journa I of World-Systems Research
The population numbers suggest that the Cahokia elite commanded significant potential power, a<;suming that institutions capable of organizing that population into active coordinated labor existed. The building of Monk's Mound, containing 22 million cubic feet of earth, suggests that Cahokia's elite did possess the ability to organize large scale labor intensive activities. It is not unreasonable to suggest that Cahokia possessed th e potential for an a<;ymmctrical power relationship with its a<;sumcd periphery and scmipcriphcry. Cahokia itself could plausibly have controlled the American Bottom a<; part of its core.
But we also have to contend with the notion of power distance decay (Stein 1993). That is, how far can a core polity extend coercive power over peripheral area<;? The short answer for Cahokia may be: 'Not Far". Ross Hassig (1992) has shown that the ability of the state-level Aztecs to throw their military weight around wa<; bounded fairly tightly. The Aztecs had difficulties with groups a<; close a<; the Chichimccs, and did not have the ability to take on ca<;ily the Maya city states of the Yucatan. To expect the Cahokians, who did not have nearly the population or social integration of the Aztecs, to defeat groups militarily in Minnesota, Ohio, Iowa, and Wisconsin stretches credulit y.
Evidence for the degree and evolution of social integration at Cahokia and its environs is provided by Mehrer and Collins (1995). Ba<;cd on excavations at the ICT -II tract at Cahokia, th ey show that during the Lohman Phase (AD 1100-1150) the resid ential community plan wa<; highly structured and oriented on a central grid system, indicating a high level of community control by a central power. Centralized authority is perhaps best symbolized by the dramatic burials within Mound 72, where an aged male wa<; laid out on a cloak of shell beads; several individual<; buried alongside him arc interpr eted a<; sacrificed attendants. Four young males between the ages of 18 and 25, who were buried minus their heads and hands, with arms interlinked , arc interpret ed as an 'honor guard'. Also nearby, an ossuary contained the remains of 33 young wom en, int erpreted also a<; sacrifices. It is conceivable that the aged individual buried with such ceremony, or someone like him, was a central authority capable of organizing the growth of th e large ceremonial and residential site of Cahokia along a well-defined grid system. At this time, population wa.., expanding rapidly, and the "mound and town " centers characterizing the Mississippian settlement pattern in the American Bottom first appear. In the hinterlands (for Mehrer and Collins, the American Bottom excluding Cahokia 's immediate environs), the Late Woodland settlement pattern oflargc villages is replac ed by the hierarchical system with individual households or farmsteads a.., the ba..,c unit. Political and

[Page 8]
Journa I of World-Systems Research social control became centralized at this time a.., individual.., left villages and moved into larger towns or smaller hamlet.., (Mehrer and Collins 1995:43). By the Stirling Pha..,c (AD 1150-1200), however, with Cahokia's population at its maximum, the socio-political system already seemed to be segmenting. The same ICT-II tract at Cahokia shows that structures were no longer oriented on a central grid, and the resid ential area wa.., separated from the central plaza and mound area of the site by stockade walls. Mor e importantly, local resid ences were structur ed around their own mound and plaza complexes. Out..,idc of Cahokia proper, differentiation of individual structur es within farmsteads indicates local stratifica tion. In sum, it appears that local communities were oriented to local elites rather than to a centralized authority (Mehrer and Collins 1995:4 7). As Cahokia and the American Bottom population declined through the following Moorehead and Sand Prairie Pha..,es, it appears that local elite control continued at the household level, rather than a.., a centrally ba..,cd system (Mehrer and Collins:57) .
Even ifw c discount the direct authoritarian control by a singl e or small group of elit es over the entire population, it is conceivable that Cahokia, because of its sheer size and magnitude relative to neighboring Mississippian groups , mi ght have been perceived a.., a credible milita ry threat to far off group s. Core elites may well have been able to enlist or coerce significant military supp ort through political/ economic /marriag e tics with local elites. Manpower necessary for significant milita ry activity is readil y available to elites in ethnographically known chiefdom-level societics--espccially at the level of raiding (Keeley 1996). Warfare a.., persist ent raiding is a pattern seen among non-stat e societies in the historic record . Historic Iroquois extended terroristic raiding parties far into the Illinois Country, for example, destroying the Grand Village of the Ka..,ka..,kia in the Upper Illinoi s River Valley in 1680 (Brown 1961). The Miami, ba..,cd in north ern Indiana, maintain ed a long-standing blood feud with the Chicka..,aw of Georgia durin g the 18th century (Call ender 1978).
Archaeological data suggest that Cahokians could conceivabl y have used intimidation through intermitten t raiding a.., a coerci ve tactic. Data from mort ality profil es and skeletal pathologies at a peripheral Oneota site in the Central Illinois River Valley (Norris Farms #36) dating to the 13th and 14th centuries suggest that long-term, int ermittent , smallscale raiding resulted in a high rate of homicides among adults (Milner ct al. 1991) . Th e site is contemporaneous with both Middle Mississippian sites and other Oneota sites in the Central Illinois Valley . Evidence for significant violence is also pres ent on skeletons from the Fisher site, on the Kankakee River near Joliet, Illinois (Langford 1927). The Fisher site is a Langford and Oneota site, also occupied during the 13th and 14th centuries. Decapitated

[Page 9]
Journa I of World-Systems Research skulls and other evidence for violence were also recentl y recovered from the Tremaine site, a 14th century Oneota occupation in the Mississippi River valley in west-central Wisconsin (Eric Hollinger, personal communication). While it is not known who their antagonists were, there is no doubt that raiding and warfare wa.., a significant a..,pcct of life in the Cahokia periphery and semi-periphery. Such violenc e can be seen a.., evidence of inter group coercion.
Stein's second a ... sumption, that because of this power a..,ymmctry, th e core elite were able to control trade, is much more difficult to a..,ccrtain. There is little evidence that peripheral area.., were trading anything of econo mic importanc e to Cahokia. Th ere is no direct evidence that American Bottom populations needed anything from the north and northw est.
The organization of settlements in the American Bottom strongly suggests that th e population within the region itsclf wa.., deployed in a manner that insured adequate agricultural production if we posit institutionali zed redistributi on of resources (Kelley 1978) . Sites arc located on the floodplain, terrac e, and uplands in a way that minimi zes risk from either flooding or drought. Potential failure in one portion of the region could be made up by redistributing surpluses from another. Control of this redi stribution ha.., been argued by some to account for the rise of Cahokia and its elite (Fowler 1969). Th e need for redist ribution ofresourccs within the American Bottom, however, docs open up the possibility that at lea..,t some agricultural produce may have been sent from th e scmipcriphcry to the core area and rcchannelcd. It is also possible that dried bison meat may have been sent to Cahokia from periphera l sites. While such meat would be archaeologically invisible at Cahokia, bison kills indicative of mor e than local-scale consumption at peripheral sites would not be. A systematic investigation of Mississippian era bison kill sites on the plains with this hypothesis in mind might be in order.
1n addition, faunal remains at Oneota residen tial sites in the Midwest suggest that it is possible that meat may have been traded to Cahokia. Kuznar (1994), in an cthnoarchaco logical study of Andean herding communities, ha.., demonstrated that faunal a..,sc mbl agcs from sites of pa..,toralists who trade meat to agriculturalist.., show a "charqui effect". Charqui is dried meat sent to agriculturalists by pastoral producers. Th e production of charqui at pastoral sites yields a faunal assemblage that is heavily biased towards heads and lower limb bones. Animal portions associated with high m cat:bon c ratio arc shipped; the low mcat:bonc ratio parts arc retained at the pastoral sites for local consumption. A brief survey of the faunal remains reported from Upper Mississippian sites suggests that the charqui effect may have operated at some of these sites. Although Upper Mississippian groups arc [Page 10] Journal of World-Systems Research characterized as having a general subsistence economy (Brown 1982;Jeske 1990;Michalik 1982), and several data sets support this assertion (Brcitburg 1992;Y crkcs 1985), several site reports from Upper Mississippian sites in Wisconsin and north ern Illinois indicate a bias towards heads and lower limb portions in faunal assemblages (Ncusius 1990;Styles and White 1993). However, the situation is far from clear, and a larger, more formal review of this evidence is suggested for futur e research.
Moreover, there is no evidence that Cahokians needed or desired dried meat. Th e fact that Cahokians did not grow beans, an important supplier of amino acids in a prot ein-poor maize-base diet, suggests that they had other protein sources, such as meat. Unfortunately, wheth er the protein source was locally hunted game or was import ed as dried meat is not known, but it seems likely that a protein-stressed population would supplem ent a meat-poor diet with as many alternatives as possible.
An alternative economic commodity that may have been sent from the periph ery to the core is slaves . Like dri ed meat, slave trading would be largely invisible in the archaeological record. Unlike bison huntin g, there would be no local production areas present in the periph ery. Although it would be very difficult to sec slave tradin g archaeologically, there is also very little cthnohistoric evidence that indicates clearly that slaves wer e an integral part of the Mississippian and historic Native Americ an economy. War captives often became slaves, but active, large scale trade in slaves is not recorded for Midw estern or southeastern groups (Ritzcnthalcr and Ritzcnthalcr 1991; Snow 1995;Swanton 1946;Trigger 1990). It is not likely that slaving was an important, widespread, or frequent economic exchange betwe en periph ery and core.
Althou gh Wall crstcin was insistent that economically important goods arc prime actors in core periphery interact ions, Schneider (199 1) and others have since argued that exchange in elite-controlled ritual parapherna lia may be more important than staples, especially in pre-capitalist economics. Peregrine (1991) has argued that trade in exotic or prestige goods was crucial for the evolution of Mis sissippian social complexity. He contends that contro l of exot ic and prestige goods necessary for 'social reproducti on' by elit e m ales provided an impetus for competition and eventual elaboration of the social system. Pcrcgrin c's argument is that the distribution of exotic goods docs not reflect pure economic exchange or competition for material resources, but is actually a reflection of individual males' desire for greater prestige and acceptance into an elite hierarchy. Although he refers to his model as a

[Page 11]
Journa I of World-Systems Research world systems approach, Peregrine actually harkens back to older models of kinshipbased exchange as a basis for social complexity (Malinowski 1922;Rosman and Rubel 1971;Strathcrn 1971 ). Although Peregrine docs not broach the topic, th e power-pr estige approach may also be used in a sociobiological context, in which males who garner status through economically non-optimal but risky activities have increased access to women, enhancing their ability to reproduce them..,elvcs biologically (Chagnon 1988;Hawkes 1994).
In an attempt to show that Mississippian exchange was not centered around economic necessity, but was structured by elite power-prestige tics, Peregrine (1991) put.., forth three major hypotheses: that core sites should have more exotic and prestige goods than peripheral sites, that adult males should control exotic and prestige goods and be buried with more of these ite1rn than others, and that goods should be dir ected by adult males to particular people or places. He tests his hypotheses by examining the distribution of shell bead artifacts, "exotic goods" (shell beads plus selected artifacts found in burials), and "prestige goods" (selected artifacts minus shell beads), in burials at four sites in the American Bottom and two sites in the Little Tennessee River area (Peregrine 199 1:73) . Unfortunately, the data at his disposal arc inadequate to answer hypotheses one and three, so he uses his data from the six sites to test them indirectly. Despite Percgrinc's best efforts, most of his tests arc statistically insignificant, leavin g his hypot hese s unsupported by the archaeological record.
If the data do not allow one to demonstrate that elites controlled trade using Peregrine's example, is it possible at least to discuss the distribution of trade items in the Midwest? In fact, there is evidenc e that certain classes of exotic or ritual-a..,sociatcd artifacts were exported from Cahokia to the north and northwest. In particular, Ramey Incised Pottery (Figure 3), long-nosed god ma..,quctt cs, and marine shell ornaments seem to be a..,sociatcd strongl y with a Cahokian influence or presence on sites in the semi-periphery and periphery (Hall 1991). Trade in these exotic or ritual items can be seen a.., elites from Cahokia cementing relationships with local elites in the periphery, or alternatively, thes e items refl ect colonial administrators in the periphery. In addition, there is some c"idencc to ,nggcst that the core:iicriphcral trade W!l~ asymmetrical in that some ritual items seem to be sent to the periphery, with little return to the o.;rc. John Kelley (l99 lJ notes that when we examine the di~tril:mtion of exotic artifacts in Cahokia and iu; northern periphery, we sec that trade items made from materials to the soclth of Cahokia arc found at Cahokia and iu; northern periphery. These artifacts inctc14lc marine she lb and hoes m114lc from a particular, highly localizc(l material called !I.fill Creek Chert. Howc-.,;:r, northern artifacu; arc not frmml to the soclth of Cahokia. To Kelley, this pattern ,nggcsts that the Amcricim Bottom !lJJPCars to be a o.;ndnit, or gateway commcmity for the mo'1cmcnt of soctthcrn gomt~ ,etch as marine shell north"w"llnl. Kelley ,nggcsu; that the Cahokia (lc"clopmcnt begins as an outpost on the northern edge of a soctthcrn core area, an(l c-.,;:ntually is clabomtcd throngh its ftmction as a gateway for the mo"cmcnts of these ritual items. Howc"cr, his '1icw beg,; the qncstion, what i~ being rctnrncd to the soclthcrn core? If most c"crything i~ .flowing in one direction, there i~ no tm(lc. ; there is distribcttion withoctt co.;nomic context. Some northern tnulc items, howc"cr, may have gone soclth to Cahokia iu;clf. 'for example, long nose god masqncttcs arc m!Ulc from shell or copper. Sc-.,;:ml o.;ppcr examples come from the Amcricim Bottom, while masqncttcs m!Ulc from shell arc focmd in the ~riphcry. It is ·cL~uallyas,nmcd that copper came from mines arO(md the northern Great Lakcs--cithcr the material or the masqncttcs thcmsch,;:s were tn1\lcd from the periphery to the core. If it "w1l~ the material, then the "alnc iv.klcd items--thc mR~qctcttcs thcmscl"cs--rcmaincd in the core and were not sent back octt to the periphery !l~ we wonld expect if there W!l~ an !l~}'Tl'lmctrical power relationship controllingtm(lc. In the case of the mR~qctcttcs, the di~tribcttional datit ,nggcst that the mo"cmcnt ofmR~qctcttcs WR~ conditioned by factors other than direct control by a core elite. As an !Uklitional problem, nati"c copper nnggcu; arc ooc!l~ionally frrcmd in !1-fi~si~sippi Ri"cr -.1tllcy gra"cL~. James A. Bn,..,,11 (personal mmmnnication.J hR~ suggested that all of the kno"'11 copper artifacts found at Cahokia could ea.:;ily have been produced from one large nugget obtained locally and opportunistically. Perhaps of greater importance, copper is not found at Cahokia until after the elaboration of Mississippian social complexity and the major period of growth at the site. If this is the ca.:;e, it wa.:; clearly not the accumulation of copper from the periphery that caused the elaboration, but rather the elaboration of social complexity that allowed access to materials and goods from far-flung localiti es. It ha.:;, in fact, been suggested by Hall (1991) and others that the ma.:;qucttcs were symbol.:; that operated similarly to the historic Calumet pipes used in the Calumet ceremon y among Plains Indians. The ceremony wa.:; an adoption rite that created fictive kinship relationships between unrelated groups. It can be seen a.:; a way to provide a stable relationship between potentially competitive politics and a.:; a means to provide safe conduct for priest.:; and/or traders who moved between and among politics. In this view, the ma.:;quettes were not commodities to be controlled by an elite, but badges of office that facilitated trade in other goods and/or scrviccs--quitc likely connected with ideologically shared ritual performances.

[Page l3] Journal of World-Systems Research
Cahokian Ramey Incised pottery, strongly correlated with the presence of exotic shell and copper ornaments, is found at sites throughout the main river drainages of th e Midwest, and might be seen a.:; signifying the presence of the economic giant in the core to those in the periphery. Ramey Incised pottery contains decorative clements that historicall y were powerful symbols of the continuity oflifc, a.:; well a.:; warrior status (Hall 1991). The falcon or thunderbird motif is present on Cahokia pottery and in stylized form on several peripheral wares such a.:; Oneota and Langford materials throughout the core, semi-periph ery, and periphery. Elite ownership and display ofCahokian pottery (or imitations thereof), may be seen a.:; a way for local elites to bolster their own power usin g support and symbols from the center. A wooden baton, carved into the likeness of a falcon, found in an elite burial at Aztalan may have function ed in this way.
Herc again, however, there is a catch. Hall (1991) points out that the distribution of peripheral Oneota ceramics bearing the thunderbird and related motifa match es the distribution of Ramey Incis ed pottery (Figure 3). If Ramey Incised pottery is an indication that peripheral elites were signalling to their local populations that they were backed by core elites, it is equally possible that the Oneota pottery found at core sites suggests that core elites desired peripheral pottery for similar rca.:;ons. The isomorphic distribu tion of Oneota and Ram ey Incised pottery with their rituall y important decorativ e motifa clearly suggests that these items were moving across the Mississippian world via a mechanism free of core elite control.
Moreover, Ramey Incised pottery is often recovered from general habitation middens, and not necessarily in elite graves or other places of social or ritual significance (Hall 1991). If elites controlled the distribution of these pot<;, we would expect them to be found clustered in elite house structures and/or graves. The non-clustered depositional context strongly suggests that ownership or trade of Ramey Incised pottery was not under the control of a ruling elite at all. These examples do not support the notion that core elites controlled trade with the peripheral populations. However, it is clear that direct and indirect contact between Cahokia and sites in the far-flung hinterlands wa<; fca<;ible.
We can now look at Stein's third proposition, that a hierarchical systems view requires us to a<;slllnc that the trade economy transforms the local economy into a dependent supplier of goods to the core. Herc we have the most problematic a<;pcct of Mississippian hierarchical core/periphery interaction. As the earlier discussion about meat trade and faunal remains indicated, there is little evidence that Mississippian sites in the Central Illinois, Waba<;h, or Ohio Rivers were organized to produce any specialized commodities such a<; bison or deer

[Page 14]
Journa I of World-Systems Research meat (Black 1967;Harn 1975Harn , 1978Harn , 1980. The situation is even worse for Caddoan, Oneota, Fort Ancient, or other Upper Mississippian populations (Brown 1982(Brown , 1990Jeske 1990;Michalik 1982;Rossen 1992). ln fact, the opposite appears to be the ca<; c. The subsistence economics of Upper Mississippian and Fort Ancient sites demonstrate a very generalized economy with little evidence for multigroup interactions other than raiding.
Where docs that leave an investigation into Mississippian core -periphery interactions ? It is clear that some exotics such a<; marine shell did move from the south, through Cahokia to the semi-periphery and periphery. In addition, small amounts of copper and smaller amount<; of other exotic items or raw materials possibly moved from the periphery to the core. Moreover, it appears that the movement of these exotics wa<; accompani ed by, perhaps aided by, ceramic pots and long-nosed god ma<;qucttcs symbolizing an ideology revolving around warriors, the continuity of life and the thundcrbird. In sum, shell moved from core to periphery, copper from periphery to core (maybe), and symbol-laden ceramics moved in both directions.
If specific 'value-added' items such a<; shell artifacts were sent from the core to the periphery, but few economic goods or appreciable amounts of preciosities were not sent from periphery to core, then what else might possibly have been returned? Perhaps loyalty and subservi ence of the peripheral popula tions. Perhaps local elites wer e coerced into controlling their populations for the aggrandizement of American Bottom elites. Th e payoff for the local elites wa<; access to important symbols and political alliances that enhanced their own status within their local group. Unfortunatel y, such loyalty is archaeologically invisible, as Pcrcgrinc's (1994) research has demonstrated.
By AD. 1300, Cahokia itself went into decline while the periph eral Oneota populations expanded. Although the pattern of core decline and pcriphc ral ascendence is expected in WST, it is difficult to argue that these later Oneota populations ever approximated a true core in terms of social and political complexity that overshadowed surrounding populations. The Oneota world of the 15th through 17th centuries did not contain sites with large scale architecture nor did it demonstrate the movement of exotic goods that we use to infer the existence of a ruling elite for Middle Mississippians.

Summary and Conclusions
What then can we conclude about taking a world systems approach to Mississippian society? First and foremost, Mississippian society docs not seem a compelling case for a hierarchical model of core-periphery interaction. Th e problems with such a m odel revolve around four primary areas: l) The difficul ty of boundedness of the core/periphery int eract ions.
2) The lack of evidence for a highly integrat ed core for mor e than a short period wit hin the entire time span of Cahokia's rise and fall.
3) The data necessary to demonstrate elite control of economic resources arc not present in the archaeological record . 4) Th ere arc insufficient data to suggest that elites controlled access to exotic artifacts and materia ls in a power-prestige hierarchy that functioned entirely to allow elites to reproduce themselves socially.
However, the notion of differential core-periphery interact ion is a framework in which the concepts of gateway communities and prestige -goo ds exchange can be examin ed in relationship to each other to find a comprehensive approach to Mississippian society. Th e task now is to operationaliz e our expectat ions for what such a system would look lik e archaeologically, and to devise testable hypotheses that would differentiate a world systems approaeh fro m other concepts such as interaction spheres (Strucvcr and Houart 1972) or other models of reciprocal and redis tributive trade netwo rks . While thi s paper has not explored these hypotheses in any detail, future work to refine our expectat ions for the archaeological record and to test these expectations will be forthcomin g.