Zambezia (1996), XXIII (i).PEASANTS, CHIEFS AND KINGS: A MODEL OF THEDEVELOPMENT OF CULTURAL COMPLEXITY INNORTHERN ZIMBABWEG. PWIT1Department of History, University of Zimbabwe'Every archaeologist who works on the transition from egalitarian societyto ranked society is dealing with a change in ideology.'(Flannery, K. and Marcus, J., 1993, 263)AbstractThis article examines the nature and causes of socio<ultural changes thattook place amongst prehistoric farming communities in northern Zimbabwe.Farming was established in northern Zimbabwe by the fifth century AD as aresult of human population movements from further north. For the greaterpart of the first millennium AD, the early farmers were organised as non-stratified village communities. Early in the second millennium AD, complexforms of socio-political organisation developed in northern Zimbabwe. It isargued in this article that rather than migration, the development of complexitywas initially the result of changes in economic practices, ideology andpopulation increase. The development of chiefdoms is associated withpopulations of the Musengezi tradition who in the 15th century, becamesubjects of the Mutapa state. This was a secondary state, resulting from thenorthward expansion of the Great Zimbabwe tradition.INTRODUCTIONONE OF THE questions that have dominated archaeological researchthroughout the world has been the question of why and how ancientsocieties transformed from relatively simple socio-political organisationto more complex forms. Archaeologists working in Southern Africa havefor many years been pre-occupied with the transformations that tookplace amongst early farming communities in the region around thebeginning of the second millennium AD, transformations that ultimatelysaw the emergence of state societies. This article examines the issue withspecific reference to an area in northern Zimbabwe defined by the Manyameriver to the west and south, the Mazowe river to the east and the Zambezito the north (see Figure 1). This research area has been the focus of a longterm research project initiated in the late 1980s (Soper and Pwiti, 1988).I trace developments in this part of the country from the stage atwhich farming communities were established in the early part of the first3132CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEFigure 1LOCATION OF RESEARCH AREA(.. Cabora BassaZAMBIAmillennium AD, through to the second millennium AD, when we seeevidence of the emergence of hierarchically organised societies andultimately the establishment of the Mutapa state. This overall pattern isbased on evidence from previous archaeological research as well as thenew evidence coming out of research currently in progress (Soper andPwiti, 1992; Pikirayi, 1993; Pwiti, 1994). I then look at the evidence fordifferent theoretical frameworks to explain the development of complexity.EARLY FARMING COMMUNITIES IN NORTHERN ZIMBABWEThere is now general agreement that the establishment of settled farmingcommunities in Southern Africa dates from around the second century ADand was a result of the movement of populations into the region fromfurther north (Huffman, 1970; Maggs, 1984a; Phillipson, 1985, 172; Hall,1987, 31; Pwiti, 1991). While this is generally accepted, there continues tobe some debate on the nature of the movements, as well as the area oforigin. The old view of the movements being characterised by waves ofG. PWITI 33migrations, popularly referred to as the 'Bantu migrations', is no longeraccepted as a realistic interpretation of the movements. The radiocarbonchronology considered in relation to the distances involved in themovements, as well as studies of the variations in ceramic style, support amodel according to which populations have spread spasmodically, andwithout continuity (Collett, 1982; Hall, 1987, 31). There is also generalagreement that apart from introducing farming and settled village life inthe region, the communities brought with them knowledge of iron workingas well as cattle as part of the domestic animal economy.Previously, early farming communities in the research area were datedfrom around the eighth to the tenth centuries AD and associated withpottery of the Coronation and Maxton phases of the Gokomere/Ziwatradition (Huffman, 1971). This suggested that farming was established ata relatively late stage here compared to other parts of the country. Somearchaeologists argued for population movements from the south centralparts of the country to the north during the later part of the secondmillennium AD (Sinclair, 1987,141). Such models were proposed against abackground of very limited research and therefore a general lack of data.For parts of the current research area, such as the Zambezi Valley,earlier archaeologists (for example Summers, 1960; Robinson, 1965) hadthought it unlikely that the hot, dry environment, ridden by tsetse fly,would have attracted any prehistoric populations, particularly farmingcommunities. The area was thought to offer very little by way ofenvironmental resources. These were assumptions largely based on modernand Western assessments of economic potential and preferred areas ofsettlement. Such assessments failed to take into account the fact thatenvironments which we may see as unfavourable according to our modernperceptions may not have been similarly perceived by past societies.New research on the northern edge of the Zimbabwe plateau (Pikirayi,1993) and the mid-Zambezi Valley (Pwiti, 1994) has resulted in revisions ofour understanding of the chronostratigraphic framework of early farmingcommunities and their distribution in the area. Archaeological survey aswell as excavation in both localities have documented the presence ofearly farming communities dating from the fifth century AD. Despite earlierassumptions about its economic potential and suitability for occupation,the area was settled by farming communities at about the same time asother parts of Zimbabwe.In the Zambezi Valley, these early communities are represented bythe characteristic comb stamped pottery with thickened rims. Whilesimilarities have been noted with the previously known pottery of theZiwa tradition, differences are evident to the extent that a new ceramicunit known as Kadzi (named after the site where the pottery was firstexcavated and dated Š Pwiti, 1994) has been defined. This work has34 CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEresulted in a revision of models of population movements on theZimbabwean plateau towards the end of the first millennium AD. Theevidence from the survey and excavation work shows that like other earlyfarming communities in the region, those in the mid-Zambezi Valley grewcrops, kept domestic animals, lived in fairly large villages and used ironfor a variety of purposes.SOCIAL ORGANISATION OF EARLY FARMING COMMUNITIESIt has for sometime now been accepted that the early farming communitiesof Southern Africa were organised as non-stratified village communities(Maggs, 1984a; Hall, 1987), the 'tribal' societies of Friedman and Rowlands(1977, 206). The evidence of settlement structure and patterning fromnorthern Zimbabwe, as well as elsewhere in Southern Africa, shows thatthe villages were self-sufficient in economic and socio-political terms (seefor example Maggs, 1984b). Maggs has suggested from his surveys in theTugela Basin in Natal that the spatial separation of the villages and thestructure of the sites show that each village was almost a model of thelocational pattern as a whole. This is a picture comparable to that obtainedin the Zambezi Valley. Each village appears to have exploited theagricultural land in its vicinity, produced its own iron and other productsand made its own pottery and other things. Essentially therefore, thesewere egalitarian peasant agricultural people. There is no evidence ofelaborate political structures or evidence of social ranking, except possiblyalong age and sex lines. This form of organisation remains evident untilsometime towards the end of the first millennium AD.How did such a socio-economic system maintain and reproduce itselfthrough time? This is a question which archaeologists in Southern Africahave only begun to address (eg Garlake, 1982; Hall, 1987). In the past, itwas a matter which was apparently taken for granted.The best approach to this question, as with the question of theemergence of social inequality to which it is naturally related, shouldground itself within a framework which considers among other variables,the relationship between ideology and economic organisation, althoughnot in a deterministic manner. Egalitarian societies possess an ideologywhich emphasizes economic and social equality (Flannery and Marcus,1993), and social rules are in place to make them work. Within our earlyfarming communities, such an ideology would have existed to regulateagricultural production, for example, in a manner which made it difficultfor a household to produce too much of a surplus. This could have beenachieved by ideological sanctions controlling access to quality and quantityof agricultural land. While members of the community observe theserules, the system can reproduce itself again and again, but only as long asG. PWITI 35the economy remains basically the same and individuals do not start toredefine the ideology and negotiate new rules.In Southern Africa, such egalitarian communities reproducedthemselves for most of the first millennium AD. The basic agriculturaleconomy remained more or less unchanged, and livestock numbers,particularly cattle, remained comparatively small.Some time towards the end of the millennium, a number of changestook place, changes that heralded the development of what I refer to aslater farming communities. The archaeological evidence shows that cattleherds increased in number in different parts of the region. In northernZimbabwe, this has been documented at a number of sites in the currentresearch (eg Guta, 1988; Mawoko, 1995). Exotic goods from distant sources,such as glass beads, started to land on the Indian Ocean coast and to findtheir way into the interior (Sinclair, 1987, 150). Such exotic goods havebeen recovered from excavations in the current research area (see Pikirayi,1993; Pwiti, 1994). These developments represented major changes in theeconomic organisation of the early farming communities, and they set thestage for changes in social organisation, which I now turn to in a discussionof later farming communities in northern Zimbabwe.THE DEVELOPMENT OF COMPLEXITY: LATER FARMING COMMUNITIESIN NORTHERN ZIMBABWEOur earliest evidence for later farming communities in northern Zimbabweis represented by the appearance of pottery of the Musengezi traditioncurrently dated from around the 12th century AD. Partly contemporary isthe Great Zimbabwe tradition, evident in northern Zimbabwe from the15th century AD. Finally, there is the Refuge tradition, which hastraditionally been associated with unsettled conditions on the Zimbabweplateau during the 19th century (see Huffman, 1971). This article isconcerned with the two earlier traditions. I start with a discussion of thecommunities identified with pottery of the Musengezi tradition.Origins of the Musengezi traditionHuffman (1978) argued that the makers of Musengezi pottery arrived innorthern Zimbabwe as part of what he has called the Kutama tradition,consisting of a number of ceramic groupings in different parts of Zimbabwe.He argued that this tradition was the archaeological manifestation of thesettlement of ancestors of the Shona people in Zimbabwe from across theLimpopo, introducing the culture of the later farming communities andreplacing the previous culture. The argument was essentially migrationistin its persuasion and was largely based on ceramic typology.Many archaeologists, however, have been skeptical of this as anexplanation for the changes observable in the archaeological record around36 CULTURAL COMPLEXITY IN NORTHERN ZIMBABWE1 000 AD (eg Garlake, 1982, 12; Maggs, 1984a). The argument has been seenas based on rather dubious ceramic links with groups across the Limpopo.More important, it failed to account for what happened to the earliercommunities.More recently, Huffman (1989) has proposed an alternative explanationfor the origins of the Musengezi tradition, this time seeing it as originatingfrom across the Zambezi to the north. This is again based on ceramictypology and cultural practices like burial systems. As with the earlierproposal, this new argument remains grounded within a migrationistparadigm and seems to face the same problems. No explanation is offeredfor the fate of the earlier communities, and the ceramic links claimedbetween Musengezi and the parent cultures in Zambia have not beenclearly demonstrated. It would thus appear that while migrationist thinkingworked very well in accounting for the initial establishment of early farmersin the region, it has not been successful for the later period.In a refreshing departure from migrationist thinking, Garlake (1982)has proposed that the appearance of cultures attributed to the Kutamatradition in Zimbabwe could be explained in terms of changes in socialand economic organisation among the local early farming communities.He argued that the changes that we see in the archaeological recordduring the period could be a reflection of culture change occasioned byeconomic growth and fundamental reorganisation of society. This line ofthinking appears more promising, taking account of the relationshipbetween ideology, economy and social organisation. Garlake recogniseseconomic change via the evidence for increased numbers of cattle amonglater farming communities as a major development that led to socialtransformation. In addition, he sees these changes as related to change insocial organisation from matrilineal to patrilineal society. These economicand social changes are used to explain the changes in material culture,such as style and quality of pottery.Ceramics of early farming communities show better manufacturingtechniques and were more elaborately decorated. Garlake explains this interms of specialist male producers in a matrilineal society. Ceramics oflater farming communities are poorly finished and less elaboratelydecorated and, in Garlake's argument, were made by non-specialist femalepotters.Garlake's model is interesting for its attempt to examine culturalchange in local terms, but I do not think it fits very well with the evidenceon changes in social organisation. Recent work in South Africa has shownthat the southern Bantu cattle pattern of settlement can now be tracedback to the first millennium AD (Huffman, 1993). This is a settlementpattern that is associated with southern Bantu cattle-keepers and is rootedin an ideology which emphasizes male dominance of society. If the earlyG. PWITI 37farming communities in our region shared a similar ideology, as seems tobe the case, then it becomes difficult to accept the changes within theframework proposed by Garlake. I shall now explore the changes in socialorganisation in northern Zimbabwe in local terms, but in a different waythat does not call for a shift from matrilineal to patrilineal society.Archaeologists now recognise that major changes in socio-politicalorganisation are occasioned, or accompanied by, change in twocomponents of a system Š economy and ideology (Price, 1977; O'Shea,1982; Kipp and Schortman, 1989; Paynter, 1989; Flannery and Marcus,1993).Others have given prominence to environmental and demographicchange (Cogwill, 1975). While some recent work has argued for severalepisodes of climatic change in Southern Africa (Tyson and Lindesay,1992), including the often cited 'Little Ice Age' during the second millenniumAD, most of the evidence used in the reconstructions has been derivedfrom South African sites. There is no reason to assume that such climaticchanges affected our area. For the moment therefore, in the absence ofclear evidence for climatic and environmental change in northernZimbabwe, it is difficult to consider this variable.Demographic change remains a possibility, and its contribution to thedevelopment of complexity in northern Zimbabwe cannot be dismissed.There is in northern Zimbabwe a clear increase in the number of settlementsites during the second millennium AD which is a reflection of populationincrease. Such population increase is likely to have contributed towardsthe evolution of a new form of socio-political organisation.The main argument presented here is that the fundamental changeswitnessed among the early farming communities in northern Zimbabwe,which are manifested by the appearance in the later farming of theMusengezi tradition community and its attributes, were a result of economicchange accompanied by a shift in ideology. The ideological shift wascharacterised by a change in attitudes towards accumulation of wealth,which itself lies at the root of social and economic inequality.As already noted, major economic changes took place in northernZimbabwe at the turn of the first millennium AD. Cattle herds increasedand exotic goods found their way into the interior from the Indian Ocean.These two developments in themselves, however, should not necessarilyresult in changes in social organisation. They only provide conditions forchange. Such developments can be prevented from resulting in structuraltransformations in a system which is based on an ideology of equality. It ispossible that potential changes arising, the new modes of production maybe negated by appropriate mechanisms controlling access to wealth orrestrictions of such access. Those who begin to possess large herds ofanimals, which are a potential source of social and economic power, may38 CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEbe called on to slaughter them for communal rituals or feasting. In thisway, the potential to accumulate wealth is checked and the ideology ofequality is maintained and perpetuated. What is being argued for here is asystemic approach to cultural stability where the self-regulatorymechanisms of a system come into play to maintain the basic position.For our Southern African farming communities, however, the systemwas not maintained. What seems to have happened is that change in theeconomic sub-system was accompanied by change in the ideological sub-system. Individuals who realised the advantages offered by the new anddeveloping elements in the economy redefined the rules governing orrestricting the accumulation of wealth. As inequalities developed, a newideology developed to make these inequalities seem natural and acceptable.Within such a context, the egalitarian early farming communities graduallytransformed, and society became more complex.This model is offered as one which accounts for change in socialorganisation in northern Zimbabwe, without involving large scalepopulation movements. It is also offered as an alternative to Garlake'shypothesis. To support this argument, an interesting ethnographic parallelcan be cited which is relevant to the relationship between economy,ideology and socio-cultural change, even though it comes from a somewhatdifferent social formation.In a study of the integration of herding of small stock into hunting andgathering economies of the !Kung San of north-west Botswana, Yellen(1984) shows how ideology can be manipulated to accommodate neweconomic conditions that encourage accumulation and thus inequalityamong a people who previously emphasized equality and an ethic ofsharing. The change in the economy and social structure among thesepeople is also physically reflected in the changes in intra-site spatialorganisation. The new pattern gives increasing prominence to goats in thesettlement structure, the economy and presumably the social standing ofpeople (Yellen, 1984). The system adjusted to a number of changes, whichinclude ideological change to accommodate something new. Thisethnographic example is instructive for archaeologists dealing with culturechange for two reasons. It is a living example of how economy, ideologyand social organisation are intimately related. Secondly, it gives us anexample of a social system in transition as a result of changing economicconditions, and how social organisation and ideology will also change ifthe scale of the changes is such that the system cannot absorb them.The overall model that I have proposed is one which accounts fortransformation in social organisation, but it has not made reference tochanges in material culture evident in northern Zimbabwe and elsewherein the region around 1 000 AD, particularly change in ceramic style.Ceramic style is one of the major items of material culture by which theG. PWITI 39changes have been archaeologically recognised and consequently deservesexamination and comment. I once again appeal to theory in the attempt toexplain this phenomenon in Zimbabwe in general and northern Zimbabwein particular.It is now generally agreed that material culture is socially and politicallyactive (Collett, 1987; Hodder, 1982; Ndoro, 1991), and that it carriesmeanings and messages in the contexts in which it is manufactured andused. In some instances, it makes social statements about the user. Thismeans that changes in ideology and changes in material culture should insome way be related, and change in the latter will be a reflection of changein the former. In this sense, it may become possible to offer an explanationfor the change in ceramic style evident not only in northern Zimbabwe,but throughout the region around the beginning of the second millenniumAD. The heterogeneity of the pottery of later farming communities, asopposed to homogeneity in earlier periods can be taken as signallingideological change, from an ideology which emphasized culturalrelatedness, to one emphasizing separateness. The result is that we havethe more diverse ceramic styles in Zimbabwe, with the different culturestaking different directions.At the micro level, for example, among people of the Musengezi culture,the meanings and messages contained in this class of material culturehave also obviously changed. Techniques of manufacture changed, aspart of the new technology, economy and ideology. My argument here isthat rather than new populations coming into northern Zimbabwerepresenting Musengezi pottery makers, what we have is the gradualdevelopment in situ of a new socio-economic package and culture. TheMusengezi tradition is thus presented as a local development.This model has limitations which should be the subject of futureempirical research and theoretical inquiry. At the empirical level, we needto find transitional sites: it is unlikely that the gradual changes proposedwould leave no traces. Theoretically, such research should be accompaniedby the construction of models which focus on the relationship betweeninternally generated changes in social culture and material culture.Musengezi culture social organisationThe social organisation of the populations identified with the Musengezitradition has not been explored, perhaps because it has been taken forgranted or because of the limitations of the evidence available. The changeappears to have been small compared to that introduced by the GreatZimbabwe tradition. In this section, I use the archaeological evidence fromthe current research area and from other parts of northern Zimbabwe tosee what can be said about the society and its organisation.Unfortunately, the attempt to reconstruct the nature of Musengezisociety must cope with the problems of limited evidence. Although a40 CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEcomparatively large number of sites of the tradition have been long knownand more have been recovered from surveys in the current research, onlya few have been excavated. The two settlement sites that have beenexcavated, Ruanga (Garlake, 1973b) and Wazi Hill (Soper and Pwiti, 1988),do not provide enough data on settlement structure and layout to informus on the nature of society. No area excavations have been completed.Excavations at Wazi Hill in the Centenary district of northern Zimbabweattempted to expose settlement patterns, but only recovered clear houseremains on the hilltop part of the settlement site. These house remainshowever, present a rather complex picture, which has been brieflycommented on by Soper and Pwiti (1988) and Soper (1992), and which isfurther examined here within the context of relations between theMusengezi and the Great Zimbabwe traditions.Fortunately, another class of sites of this tradition has been widelyknown and more have been found in the current survey work. These arethe cave and cleft burial sites common in northern Zimbabwe. I shall usethis class of data together with the observed distribution of sites toattempt a reconstruction of the nature of Musengezi society.The majority of known burials in the Musengezi tradition are in rockshelters, crevices or caves. In most cases they are multiple burials, withthree or more individuals in association with a number of pots and bowlsand other items of material culture.The largest of the burial sites, Monk's Kop, was partially excavated byCrawford (1967) and recently interpreted by Mahachi (1986) and Pwiti andMahachi (1991) using the archaeological data and Shona ethnography.Monk's Kop is a large burial cave located on a hill and dated to the 13thcentury AD. A minimum of 70 individuals were found during excavation,together with a hundred complete or almost complete pottery vessels.The individuals were adorned with iron bangles on the arms and ankles,as well as bead necklaces (Crawford, 1967). Of importance in the contextof this interpretation of the evidence, was the presence of an unspecifiednumber of conus shell discs, known in Shona as ndoro, associated with theburials.The burial is unusual in a number of respects. It differs from otherknown Musengezi burials in northern Zimbabwe because of the largenumber of individuals. Other sites usually contain three or four individualstogether with a limited range of pottery vessels. This fact, and the associatedmaterial culture, as well as its location, have persuaded us that it is not anordinary burial site but a site for the burial of important individuals insociety.This archaeological evidence pointing to the existence of rankingwithin Musengezi society is supported by Shona ethnographic data. Wefirstly note that the practice of multiple burials among some Shona groupsG. PWITI 41of the recent past is associated with chiefs and not ordinary people(Mahachi, 1986). In such cases, successive chiefs and members of theirfamilies within a polity are buried in a cave reserved for them as royalty.Secondly, we see the location of Monk's Kop as significant in the contextof the association between royalty and hills in Shona belief systems(Mahachi, 1986; Pwiti and Mahachi, 1991). Lastly, the conus shell discsfound associated with the people buried at Monk's Kop are part of chieflyregalia among the historic Shona. In all, this large Musengezi burial issuggestive of the existence of the chiefdom level of social organisationduring Musengezi tradition times. The use of Shona ethnography in thiscase is justified because we can trace historical and cultural continuityhorn the past to some modern Shona people in the northern part ofZimbabwe.The evidence from inter-site settlement patterning would be consistentwith the interpretation offered above, if site size can be accepted as acriterion for ranking. Musengezi sites in the current research area seem tofall mainly into two sizes, large sites covering at least two hectares andsmall sites usually of less than half a hectare. Large sites are comparativelyfew: so far, only four large sites have been recorded, compared to over 55small sites (see Figure 2). This is a settlement pattern consistent withchiefdoms (Renfrew, 1977). Unfortunately we do not have information onthe living sites around Monk's Kop, which would have lent more weight tothe model.The picture that becomes possible to reconstruct is one in which thepeople identified with Musengezi pottery had established small chiefdomsover much of the northern plateau and the Zambezi Valley by the 12thcentury AD. This process is here argued to have been the result of economicchange accompanied by ideological change. The chiefdoms were supportedby agricultural production and livestock herding, and chiefly power wasenhanced by an element of external trade. This was achieved in part bythe redistribution of exotic goods in society. There is evidence that a classof craft specialists had also emerged, working with copper and iron andmanufacturing cloth. Sometime during the 15th century, with the extensionof the Great Zimbabwe tradition northwards, the relations of powerchanged, even though the Musengezi people maintained their culturalidentity and may possibly have retained a measure of political autonomy.THE GREAT ZIMBABWE TRADITION IN NORTHERN ZIMBABWE: THEESTABLISHMENT OF THE MUTAPA STATE AND ITS ARCHAEOLOGICALIDENTIFICATIONIn this section, I look at the Great Zimbabwe tradition in northern Zimbabweand try to relate this to the development and history of the Mutapa state.42CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEFigure 2MUSENGEZI TRADITION SITES IN NORTHERN ZIMBABWE© Large Musengezi Tradition Sites >1HaŁ Small Musengezi Tradition Sites <1 HaG. PWITI 43I also look at its relationship with the Musengezi people, using data derivedfrom previous research as well as from the results of the current work.The various attributes of the Great Zimbabwe tradition have beendescribed in detail elsewhere (Garlake, 1970). The most well known anddistinctive are the stone wall enclosures, enclosing housing units of smallruling elites as symbols of their power and prestige. Those currentlyknown in northern Zimbabwe (Figure 3) show a tendency to cluster infairly tight groups, with groups or single sites rather evenly spaced atintervals of 60 to 100 kilometres (Soper, 1990).The distribution of sites, and their place in the local context, has beencommented on by a number of people in various ways (Garlake, 1973a,1973b; Beach, 1980, 82; Soper, 1990). At the local level, Garlake has seenthe sites and their associated occupants, as representing the superiorculture of a small dominant elite lording it over the locals. He argued thatthe number of people occupying these sites was so small that they had todepend on the local subjects for food, building and other necessities.Essentially, they were 'small ruling groups sustained by the labour andproduction of a much wider population' (Garlake, 1973a, 134). Thisconception of enclosure occupants living in isolation from locals seemsinconsistent with some of the evidence. Beach (1980, 82) rejects it asinconsistent with how social or political importance, especially of adominant group, are measured in traditional Shona society.The yardstick of power and influence was the number of wives andrelatives living around you. This necessarily means that settlements ofimportant people had to be large (musha mukuru or muzinda in Shona). Inaddition, Portuguese documents of the 16th century refer to Mutapa capitalsas large settlements which included stone walled structures (see Pikirayi,1993, Chapter 5). The picture that is now emerging from current surveyand excavation work, and even Garlake's own work at Ruanga, shows thatGreat Zimbabwe populations were living with the local communities.At a higher level, considering the settlement pattern, Garlake (1973b)regarded the zimbabwes as representing provincial centres or courts ofthe Zimbabwe state, reflecting the extension of political and economiccontrol from the centre at Great Zimbabwe. In a later article consideringthe pastoral economy of the Great Zimbabwe tradition, the distributionpattern of the sites was seen as a strategy for supporting livestocktranshumance (Garlake, 1978). Here he revised his earlier view andconcluded that it was unlikely that the sites were part of the Zimbabwestate in the sense of a unitary state. The use of Theissen polygons suggestedto him that they represented several semi-autonomous or autonomouspolitical units.Beach (1980, 83) on the other hand saw some of the sites as reflectingthe gradual movement of Karanga dynasties northwards, to introduce the44CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEFigure 3GREAT ZIMBABWE TRADITION SITES IN NORTHERN ZIMBABWE3DIDCOTFM ŁMaronderaG. PWITI 45Zimbabwe culture here, but not necessarily as part of an expandingZimbabwe state. From one of these dynasties the Mutapa state emerged.Pikirayi (1993) sees the establishment of the state as having been a resultof one dynasty subduing the rest. This kind of analytical frame seemsmore consistent with the evidence now available. The dates now at ourdisposal from the Zvongombe cluster in the Centenary district of northernZimbabwe, extending from the early part of the 15th century, seem to fit inwithin this interpretation. If the Zimbabwe state collapsed around themiddle of the 15th century, then quite clearly the Mutapa state was alreadyin the making. Dates from other zimbabwes in northern Zimbabwe rangefrom the 14th to the 16th century.Following this line of argument, it becomes possible to interpret theoccurrence of our different stone-walled sites in northern Zimbabwe intwo ways. The earlier sites represent the initial expansion of the culturenorthwards, at some stage in the early part of the 15th century or evenearlier, resulting in the establishment of the Mutapa state. The later sites,like Zvongombe and the mid-Zambezi Valley cluster, dating from the 15thcentury onwards, can be seen in the context of shifting capitals of anestablished Mutapa state, as suggested by Beach (1980, 83). Portuguesedocuments clearly show us that by the end of the 15th century, the statewas already established and Great Zimbabwe was no longer a majorcentre. This suggestion is supported by historical evidence both from oraltraditions and written records. These sources particularly associate thelater mid-Zambezi Valley cluster of stone structures with the Mutapa state(Abraham, 1959). To this must be added the results of Pikirayi's recentarchaeological work in the Mount Darwin area at Baranda, a major 16thcentury trading site of the Mutapa period (Pikirayi, 1993).It remains to explore two further questions relating to the GreatZimbabwe tradition and the Mutapa state. One concerns the process bywhich dominance was established over the local people by the newpopulations of the Zimbabwe culture. The other concerns the economicbasis behind the development and survival of the state.The archaeological record has unfortunately not been informative onhow Musengezi people, constituting the basic populations of our area,came under the control of the Mutapa state. What the archaeologicalrecord has permitted however, are possibilities for us to suggest what didnot happen, and to offer some suggestions of the possible relationshipsbetween the two populations.It is highly unlikely that the Mutapa state was established by militarysubjugation of the locals. There is no archaeological evidence for the useof force to establish or maintain power. The plans and other attributes ofthe stone enclosures themselves do not support possibilities of a defensivefunction. Very few finds of weapons have been made from the sites. The46 CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEfew arrowheads found were probably for hunting rather than combat.Thus, while we do not discount the existence of armies in the Mutapastate, as Portuguese records testify to their existence for the later periods,military force remains an improbable explanation for the founding of thestate in northern Zimbabwe.It has been suggested by some that religion had something to do withthe process in question (Garlake, 1973a). In as far as religion is an importantcomponent of ideology, this suggestion is plausible. It is generally acceptedthat religion is part of statecraft and plays an important role in reinforcingand maintaining political power. Its role however, should not be consideredin isolation and it is difficult to conceptualise how it can successfully beused as an explanatory factor in the present context.If the establishment of the state was not by military conquest and if itwas not a result of the possession of supernatural powers, how was itachieved? Soper (1990) quite rightly concludes that we do not yet haveadequate evidence on how this process took place. However, somepossibilities can be explored.Presumably, if the founders of the Mutapa state were expanding fromGreat Zimbabwe, they were familiar with the potential of two of the state'sseveral branches of production and how they could be used as sources ofpower. These are external trade and large-scale cattle herding. If they hadpossessed large herds of cattle during their expansion, or alternativelyhad built up herds in the north, then it may have been possible for them touse these as a useful power base among the locals. The use of cattle as asource of social or political power among the Shona in Zimbabwe in thedistant and recent past is now well known (Mudenge, 1974, 1988). Indeedfor the Mutapa state, the Portuguese refer to their importance in thisregard.The hypothesis presented here is that cattle rich immigrantcommunities settled among a people who were not so rich, but who werevery keen to use cattle products or own more cattle herds. The immigrantscould easily have used cattle gradually to build up economic power,prestige and social dominance. At some stage they would have translatedthis into political power, without involving the use of force.To this scenario should be added the use of the trade goods from adistant source as a way of further developing and building up a strongpower base. This would be through the redistribution of such goodsamong a people who may not have participated in long distance trade ona large scale. This fits in well with the contention that 'luxury goods froma distant source are often distributed to reproduce a system of rank statusor offices within a polity' (Kipp and Schortman, 1989). If one of the reasonsfor the collapse of Great Zimbabwe was indeed the shift in the focus oftrade to the north, then it becomes logical to credit the early rulers of theG. PWITI 47Mutapa state with the introduction of large scale external trade in northernZimbabwe. These suggestions are offered as useful working hypotheseswhich form a basis for the discussion of Mutapa-Musengezi relationswhich I now wish to explore.Our spatial data from site distributions, and the chronological datafrom excavations, are beginning to show that the relationship betweenGreat Zimbabwe and Musengezi communities was one of interaction atvarious levels in a ruler-ruled context. At Ruanga and Wazi Hill, the evidencesuggests that the two communities lived together on the same sites. Thissuggests a very close relationship indeed, although at Wazi Hill the presenceof the Great Zimbabwe tradition seems to have been limited to a fewindividuals on the hilltop part of the settlement.This has some interesting implications for Huffman's model of adichotomy between settlements on the hilltop and those below as asymbolic expression of social organisation. The possibility of theapplication of the model at Wazi Hill is strengthened by the existence ofthe Zimbabwe culture type of house on the hilltop. Although the settlementhistory at Ruanga is not very clear on this, it appears that a similarsettlement structure may also be present at this site.In other situations across northern Zimbabwe, the different populationslived in the same area but not necessarily at the same site.The two traditions do not seem to have interacted much in ceramicterms. A few graphite burnished potsherds of the Great Zimbabwe traditionoccur both on the hilltop and the lower part of the settlement at Wazi.They have also been found at other Musengezi sites in the research area.Similarly, Musengezi sherds have been found on sites in the Great Zimbabwetradition.This paucity of ceramic interaction between the two communitiesmay seem a little odd. How do people interact so closely without influencingeach other's material culture? The answer may lie in an examination of themeaning of material culture to different people. Material culture, in thiscase pottery, may be decorated in a particular style for a variety ofreasons, including the fact that it is socially active and may be a vehicle bywhich cultural messages are transmitted. When two cultures live togetherand interact, at least one of two things can be expected to happen dependingon the relationship between them. If there is ideological conflict betweenthem in terms of the meanings of material like ceramic style, then it couldbe expected that the dominant group will seek to change and bring suchmaterial culture in line with what is ideologically acceptable. On the otherhand, in the absence of conflict, it is possible for two different groups tomaintain their styles, and to continue to make their pots for consumptionwithin their respective communities. This way, they may remain culturallydistinct, at least in ceramic terms, for long periods of time despiteinteraction in other areas.48 CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEI suggest that this is one way in which we can account for the continuedco-existence of Musengezi and Great Zimbabwe pottery. As long asMusengezi ceramic style, and whatever cultural significance it had, didnot threaten the relations of power, then it continued to be acceptable tothe Mutapa rulers. On their part, the Musengezi people may have seen itas a way of maintaining their cultural identity, rather than adopting thepottery of the newcomers. This in some ways echoes Hodder's (1982, 35)'negative reciprocity' theory based on his study of material culturesimilarities and differences among ethnic groups living in the same area innorthern Kenya.If the model proposed for the establishment of the Mutapa state isacceptable, then it may also be possible to extend this line of thought tothe political realm. It has been proposed earlier that Musengezi societywas organised into chiefdoms. In this context I suggest that rather thantotally destroying this system, the emerging Mutapas used it by makingallies of Musengezi chiefs and letting them maintain a measure of politicalautonomy. This not only ensured good government, but also made politicalsense for a state of this size. It has been known in the Shona past that bigrulers made use of this strategy (Beach, 1980, 113). In return, Musengezichiefs were assured of continued political power, and access to cattle andtrade goods.ECONOMIC ORGANISATION OF THE MUTAPA STATEThe economic organisation of the Mutapa state has been the subject of afair amount of historical research (Randies, 1979; Beach, 1980; Mudenge,1988) and recently archaeological research (Pikirayi, 1993). One of themajor advantages we have when studying this state, particularly itseconomy, is that it has been fairly well documented by the Portuguese.The Portuguese were in Southern Africa for economic motives. Althoughthey did not always get things right, they did record different aspects ofthe economy of the state in some detail, particularly those branches ofproduction which they saw as directly relevant to their commercialinterests, such as mining. In this section, therefore, only a summary of thedifferent branches of production is presented. Focus is more on how theywere articulated in the development and life of the state. This is an areathat neither historians nor archaeologists have covered adequately,although Mudenge's work is an exception.Our archaeological evidence from the excavated sites shows thatthere are several branches of production represented, namely livestockherding, mining, participation in long distance trade and agriculture.Livestock herding is clearly reflected in the faunal remains, while thepresence of finished metal products as well as waste from processingG. PWITI 49shows the mining activities. Imported goods (mainly glass beads andceramics) testify to the participation in long distance trade. Directarchaeological evidence for agriculture is very rare, but this economicpursuit can be inferred from the archaeological record through thesettlement and some of the artefactual evidence.One of the main weaknesses of the examination of state developmentin Southern Africa has been the tendency to view external trade as theprime mover. That the Mutapa state was a major trading power is not indoubt. Apart from documentary data, the site of Baranda in the Mt Darwindistrict has yielded abundant evidence of this (Pikirayi, 1993). However, tosee this as the only key factor in the economy is to ignore the manner inwhich this factor is related to everything else. Trading activities requirethat what is sold on the external market be procured in the first place. Forthe Mutapa state, this is where agricultural production and livestockherding play an important role. We are told in the Portuguese documentsthat when the Mutapa needed gold for trade, he would give his subjectscattle, and they in return would mine and supply the gold (Randies, 1979,86). Thus, a look at the relations of production shows that there was aninteresting two-way exchange.There is no doubt that the rulers appropriated a fair amount of goodsand services to support their external trading and other activities. But thegoods and services were rewarded, and power relations were maintained.There seems to be a clear potential for a systemic analysis here to avoidnarrowness in accounting for state development and survival. Livestockproduction supports trade, which in turn supports the political system.To this we should add agriculture in which, besides producing forthemselves, subject peoples produced for the state through the paymentof tribute to the rulers either by way of actual agricultural produce orthrough the provision of agricultural labour. We are told that one day outof each month, different parts of the state offered labour to the royalfields, the zunde (Mudenge, 1988, 164). Besides agricultural labour orproduce, subject peoples could also pay tribute with hunting productslike animal skins. These were used by the rulers themselves or traded.What seems clear is a situation in which many different branches ofproduction related in a complex web, rather than one factor of the economyemerging as the prime mover.CONCLUSIONI have tried to account for the development of socio-cultural complexity innorthern Zimbabwe from the formative years at the end of the firstmillennium AD to the second millennium AD. I have argued that thesedevelopments were a result of economic and ideological changes which50 CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEtook place among the early farming communities rather than of newpopulation movements.At a later stage, we see the intrusion of the Zimbabwe culture into theregion. The people of the Zimbabwe culture arrived as dynasties. Throughtime, these became united as a single political entity and as a result, one o fSouthern Africa's biggest political structures emerged. I have tried toshow that its early development and its growth through time was supportedby the successful integration of different segments of the economy, ratherthan by a single element. My approach therefore emphasizes multicausalityrather than monocausality.ReferencesABRAHAM D. P. (1959) 'The Monomotapa dynasty', Rhodesia Native AffairsAnnual, XXXVI, 59-84.BEACH, D. N. (1980) The Shona and Zimbabwe, 900-1850: An Outline ofShona History (London, Heinemann).COGWILL, G. (1975) 'On causes and consequences of ancient and modernpopulation changes' American Anthropologist, LXXVII, 505-25.COLLETT, D. (1982) 'Models of the spread of the early iron age', in C. Ehretand M. Posnansky (eds.), The Archaeological and LinguisticReconstruction of African History (Berkely and Los Angeles, Universityof California Press), 182-98.Š (1987) 'Contribution to the study of migrations in the archaeologicalrecord: The Ngoni and Kololo migrations as a case study', in I. Hodder(ed.), Archaeology as Long Term History (Cambridge, CambridgeUniversity Press), 105-116.CRAWFORD, J. R. (1967) 'The Monk's Kop Osssuary', Journal of AfricanHistory, VIII, (iii), 373-82.FLANNERY, K. AND MARCUS, J. (1993) 'Cognitive archaeology', CambridgeArchaeological Journal, III, (ii), 260-70.GARLAKE, P. S. (1970) 'Rhodesian ruins: A preliminary assessment of theirstyles and chronology', Journal of African History, II, (iv), 495-513.Š (1973a) Great Zimbabwe (London, Thames and Huddon).Š (1973b) 'Excavations at Nhunguza and Ruanga in Northern Mashonaland',South African Archaeological Bulletin, VII, (cvii, cviii), 107-53.Š (1978) 'Pastoralism and Zimbabwe', Journal of African History, XK, (iv),479-93.Š (1982) Great Zimbabwe Described and Explained (Gvferu.Mambo Press).Š (1983) 'Prehistory and ideology in Zimbabwe', in J. D. Peel and T. O.Ranger (eds.), Past and Present in Zimbabwe (Manchester, ManchesterUniversity Press), 1-19.GUTA, T. (1988) 'A Comparison of Two Faunal Assemblages From Iron AgeSites in Northern Zimbabwe' (Harare, BA (Hons) dissertation, HistoryDepartment, University of Zimbabwe).G. PWITI 51HALL, M. (1987) The Changing Past: Farmers, Kings and Traders in SouthernAfrica AD 200-1860 (Cape Town, David Phillip).HODDER, I. (1982) Symbols in Action (Cambridge, Cambridge UniversityPress).HUFFMAN, T. N. (1970) 'The early iron age and the spread of the Bantu',South African Archaeological Bulletin, XXV, (xcvii), 3-21.Š (1971) A Guide to the Iron Age of Northern Mashonaland (Salisbury,Occasional Paper, No. 4, The National Museums of Rhodesia), 3-21.Š (1978) 'The origins of Leopard's Kopje: An 11th century difaqane',Arnoldia, VIII, (xxiii), 1-12.Š (1989) 'Ceramics, settlement and late iron age migrations', AfricanArchaeological Review, VII, 155-82.Š (1993) 'Broederstroom and the Central Bantu cattle pattern', SouthAfrican Journal of Science, LXXXIX, 220-26.KIPP, R. AND SCHORTMAN, E. (1989) 'The political impact of trade in chiefdoms',American Anthropologist, XCI, 370-85.MAGGS, T. O. (1984a) 'The iron age south of the Zambezi', in R. Klein (ed.)Southern African Prehistory and Palaeoenvironments (Rotterdam, A. A.Balkema), 329-60.Š (1984b) 'Iron age settlement and subsistence in the Tugela Basin', in M.Hall, G. Avery, D. M. Avery, M. Wilson, and A. J. Humphreys (eds.),Frontiers: Southern African Archaeology Today (Oxford, OxfordUniversity Press), 194-206.MAHACHI, G. (1986) 'Some Zimbabwean Iron Age Burials Interpreted in theLight of Recent Shona Mortuary Practices' (Cambridge, University ofCambridge, Mphil. thesis).MAWOKO, P. (1995) 'An Analysis of Faunal Remains From Kasekete, A GreatZimbabwe Tradition Site in the Mid-Zambezi Valley, NorthernZimbabwe' (Harare, History Department, University of Zimbabwe, BA(Hons) dissertation).MUDENGE, S. I. G. (1974) 'The role of foreign trade in the Rozvi empire: Areappraisal', Journal of African History, XV, 373-91.Š (1988) A Political History of Munhumutapa C. 1400-1902 (Harare,Zimbabwe Publishing House).NDORO, W. (1991) 'Why decorate her?', Zimbabwea, III, 60-65.O'SHEA, J. (1982) 'A friend in need is a friend indeed: Social storage and theemergence of social ranking', in Renfrew, C. and Shennan, S. (eds.),Resource, Ranking and Exchange (Cambridge, Cambridge UniversityPress), 54-67.PAYNTER, R. (1989) 'The archaeology of equality and inequality', AnnualReview of Anthropology, XVIII, 130-42.PHILLIPSON, D. W. (1985) African Archaeology (Cambridge, CambridgeUniversity Pres).PIKIRAYI, I. (1993) The Archaeological Identity of the Mutapa State: Towardsan Historical Archaeology of Northern Zimbabwe (Uppsala, SocietasArchaeologica Uppsaliensis).52 CULTURAL COMPLEXITY IN NORTHERN ZIMBABWEPRICE, B. (1977) 'Shifts in production and organisation: A cluster interactionmodel', Current Anthmpology, XVIII, 240-65.PWITI, G. (1991). 'Trade and economies in Southern Africa: Thearchaeological evidence', Zambezia, XVIII, 119-29.Š (1994) 'Prehistoric farming communities of the Mid-Zambezi Valley,Northern Zimbabwe', Zimbabwean Prehistory, XXI, 7-14.Pwrn, G. AND MAHACHI, G. (1991) 'Shona ethnography and the interpretationof iron age Zimbabwe burials: The significance of burial location',Zimbabwea, I, 57-9.RANDLES, W. G. (1979) The Empire of Monomotapa From The Fifteenth to theNineteenth Century (Gweru, Mambo Press).RENFREW, C. (1977) 'Space, time and polity', in J. Friedman and M. J.Rowlands (eds.), The Evolution of Social Systems (London, Duckworth),89-112.ROBINSON, K. R. (1965) 'A note on iron age sites in the Zambezi valley andon the escarpment in the Sipolilo district, Southern Rdodesia' Arnoldia,I, 1-2.SINCLAIR, P. J. J. (1987) Space, Time and Social Formation (Uppsala, SocietesArchaeologica Uppsaliensis).SOPER, R. (1990) 'Great Zimbabwe tradition in local context', in P. Sinclairand G. Pwiti (eds.), Urban Origins in Eastern Africa: Proceedings of the1990 Workshop, Harare, and Great Zimbabwe (Stockholm, Central Boardof National Antiquities), 67-75.SOPER, R. AND PWITI, G. (1988) 'Preliminary report on excavations at WaziHill, Centenary', Zimbabwean Prehistory, XX, 16-20.Š (1992) 'Excavations at Zvongombe, Centenary district, NorthernZimbabwe', in P. Sinclair and A. Juma (eds.), Urban Origins in EasternAfrica: Proceedings of the 1991 Zanzibar Workshop (Stockholm, CentralBoard of National Antiquities), 146-61.SUMMERS, R. (1960) 'Environment and culture in Southern Rhodesia'Proceedings of the American Philosophical Society, CIV, (iii), 266-92.TYSON, P. AND LJNDESAY, J. A. (1992) 'The climate of the last 2 000 years inSouthern Africa' Holocene, II, (iii), 271-78.YELLEN, J. (1984) 'The integration of herding into prehistoric hunting andgathering economies', in M. Hall, et. ai, Frontiers: Southern AfricanArchaeology Today, 53-54.