ThirdWorld Quarterly, Vol. 17, No. 1, pp. 89± 108, 1996

TheTanzanian peasant and : a studyin contradictions

NANCYSPALDING

Tanzania,more than most countries, has beenshaped by the vision of oneman, thecountry’ s ®rst President,. He is reveredand respected there andabroad for his intelligence, his integrity and his devotion to his people. He was unusualamong African leaders, and personal respect has persistedsince he leftof® ce. However, has madelittle economic progress since indepen- dence,many of Nyerere’ s policieswere failures, and the country is still desperatelypoor, despite high levels of aid.It has beensuggested that `never was therea morenoble social experiment, and never was therea moremiserable failure.’ 1 Nyerere’s programmestemmed from his speci® c visionof what Tanzaniacould and should become, which was basedin part on his perception ofthenature of pre-colonialTanzanian society. Did his vision re¯ ect Tanzanian societyand historical reality? What was thepolitical , and howdid it ® twithNyerere’ s policyagenda? How much can or should political agendasand rhetoric be measured against historical reality? I willuse culture theory,as developedby Mary Douglas, 2 toaddress thesequestions. While Tanzania’s failurehas beenanalysed extensively, culture theory provides a new wayof seeing the data which can add to an understanding of the cultural bases ofpoliciesand of policy failures. Cultural analysis suggests that, while ways of lifevaried, many Tanzanian communities were quite individualist, relatively unconstrainedby prescriptions,and equality of conditionseems tohavebeen the resultof the limits imposed by technology, environment and the vagaries of subsistenceagriculture rather than a philosophicalpreference.

Thesocial formsof pre-colonial Tanzania Tanzaniais ethnicallyvaried and fractionalised, without dominant groups. Little canbe said in an overarching manner concerning the pre-colonial social and politicalforms inthe area whichcomprises present day Tanzania, except that therehas beenlittle tradition of centralised rule or authority, even in the more hierarchicalsystems. 3 Overall,the economy was predominantlysubsistence, with widevariation in rainfall and land productivity. The family was theprimary focusof loyaltyand economic activity. 4 Socialforms variedwidely, as didlocal geographicaland climatic conditions. Even the more hierarchical societies (egWahaya and Wachagga) had numerous cross-cutting cleavages, reducing leaders’ability to command obedience.

NancySpalding is attheDepartment of Political Science, EastCarolina , Greenville NC27858,USA

0143-6597/96/010089-20$6.00 1996Third World Quarterly 89 NANCY SPALDING

MaterialContext

Thesubsistence economy was basedon rain-fed agriculture, characterised by smallscale productionusing family labour and rudimentary technologyÐ which was simple,including hoes and digging sticksÐ with no draft animals. 5 Erratic rainfalland the possibility of famine led to a varietyof adaptations, based on localconditions, and different crops would be grown to ensure that some crops wouldsurvive in case therains failed. In addition to subsistence agriculture, therewere crafts (oftenwith regional specialties), providing a basis foreconomic specialisationand regional trade. Pastoral peoples who ranged across the territoryadded to theeconomic and cultural variety. Tanzanian peoples were not isolated,or in a stateof tribal war; especiallyin the well watered eastern highlands,there were lively exchanges between contrasting environments and cultures,and antagonism coexisted with lively exchanges of goods.The Swahili languageand culture also facilitated exchange. The language began as acoastal languageand became a linguafranca, lowering barriers to travel, interaction and trade,and eventually facilitating nationalist organisation. 6 Thevarieties of pre-colonial political organisation ranged from complete statelessness tochiefdoms. Tanzania was acolonisingsociety. 7 Untilrecently, landhad been freely available for the effort of clearing it. Groups and individualscould independently begin or join new communities, suggesting limitedcontrol over members, or a viable`exit’ option. In these colonising efforts,survival depended on individual effort as wellas cooperation. 8 Theeast consistedof well-watered, fertile, cool, densely populated highlands alternatingwith dryer, warmer lowlands. These communities had minimal tribal orethnic identity, in part because of the dynamic interactions, movements and colonisingactivities; people identi® ed themselves with their geographical loca- tion(eg from the hills, or thelowlands) rather than with any ethnic group. 9 There werehighly articulated political systems, fairlywell organised to resist incur- sionsfrom the plains. Varying authority systems facilitatedinteractions and arbitrateddisputes among the mix of peoples with different customs. Thewest was amoreundifferentiated plateau, with low population density andsmall political units. Chiefs were often ritual, with minimal executive or judicialpower, 10 powerbeing vested in councils. However, there was avariety ofsocial forms, including states suchas theFipa. 11 The19th century brought with it tremendous economic and social shocks, whichaccelerated social and economic change. The major changes were the rise ofcommercebetween the coast and the interior, the invasions of theNgoni from southernAfrica, and the shift from traditional, religious or ritual bases to militarybases forrule. 12 Intensesocial disruption, dynamic economic changes andgreater attempts to centralise governments resulted. Materialistexplanations spring easily from the preceding discussion; peculiar- itiesof geography, the requirements of rain-fed subsistence agriculture, dif- fusion,or migration patterns might have created a needfor certain social forms. 13 Theseexplanations are logicaland certainly important, but inadequate. If we coulddemonstrate that each rain-fed subsistence agriculture system developed similarsocial structures, then an agriculturally or climatedetermined explanation 90 TANZANIAN PEASANT AND UJAMAA wouldbe suf® cient.That is clearlynot the case. Eventhe migration does not explainvaried social forms, except insofar as theoriginal migrants who popu- latedeastern Africa brought with them their own social forms. 14 Once the broadermigrations had ended, migration and colonisation were choices made withinspeci® c waysof life and social structures. While shifting cultivation and poorsoils led to migration,they did not require it; migration was asocialchoice, nota necessityof the environment, or inevitable because of marginal land. Thoughthese material in¯ uences or variables were important, I arguethat theyinteracted with cultural factors, and that cultural factors are fundamental determinantsof the varied forms thatsocieties take, even within similar environments.These cultural factors are thefocus of this study.

Culture theory Culturetheory suggests that preferences (eg for policies, goals) are consistent, andstem froma wayof life which is acongruentcombination of social relationshipswith the values which reinforce, reproduce and explain those relationships. 15 Acoherentway of life requires that the people’ s beliefsystem andsocial relations support each other, so awayof lifeis morallycoherent and internallyconsistent, with a world-viewwhich outlines morality, the good life, andthe good society. `These distinctive ways of life have systematic frames of referencefor interpreting similar facts andmoral issues indifferent ways,’ 16 so themeaning of behaviours, values and events will vary across cultures,as will perceptionof problems, risks andthreats. Adherents believe that theirs is the correctway to see theworld, not a matterof convenienceor ideology.Therefore, cultureis noteasily malleable. 17 Culturetheory organises and explains varied social constructions of reality, makingthem comparable and comprehensible across societiesand contexts. This modelcan be used to understand societies, and learn something useful concern- ingtheir functions and the natures of their cultures quickly and comparatively. Geertz’s `thickdescription’ is valuablebut `a methodologywhich requires its users tobe brilliantly perspicacious was nota methodthat had much to recommendit to most of the rest ofus’ . 18 Collectionof ethnographic data is a long,slow meticulous process, beset by manyobstacles, and these data are often notused for any analytical purpose beyond that of the original researcher. While thisprocess providesessential data, is therea wayto use thosedata more fully? Furthermore,`can one learn something useful about cultures more quickly?’ 19 Forpolicy analysis and development planning, models which can provide readily usableand sound indications of socialrealities, based on available ethnographic andhistorical data, would be bene® cial. Culturetheory categorises ways of life along two dimensions, as shownin Figure1: `theextent to which the individual is incorporatedinto bounded units’ (or`group’ ), and`the degree to which an individual’ s lifeis circumscribedby externallyimposed prescriptions. The more binding and extensive the scope of theprescriptions, the less oflife that is opento individual negotiation’ (or `grid’ ).20 Highgroup suggests strong inclusion into groups which sustain the individual’s lifeand provide purpose and identity; low group suggests weak 91 NANCY SPALDING

Group Low High

FATALISM/ High DESPOTISM HIERARCHY Grid Low INDIVIDUALISMEGALITARIANISM

FIGURE 1.Dimensions of group and grid categories. boundariesor exclusionfrom group membership, so lifeis notde® ned by strong groupties and supports. High grid is heavilyprescribed or regulated; low grid suggestsfew external prescriptions, so the individual self-prescribes behaviour. Thereare fourviable ways of lifebased on thesedimensions: hierarchy (high grid/highgroup) supports order against social deviance; egalitarian (high group/ lowgrid) pursues equality and opposes strati® cation and authority; individualist (lowgrid/ lowgroup) ® ts intothe Rational Choice model, and pursues freedom throughcompetition and individual achievement; fatalism/ despotism(high grid/ lowgroup) which is heavilyprescribed and excludes the majority from decision making;its adherents see lifeas alottery,unpredictable and uncontrollable. Theseways of life have analogues in other social sciences; anthropology’s `acephalous’societies might fall into either the individualist or the egalitarian category,since their essence islowgrid, or fewrules and prescriptions. Weber’ s legal±rational category is re¯ected in hierarchy, which would include elitist, monarchical,caste and`authoritarian’ societies, where rules distinguish between behaviourappropriate to different members, based on position and role. It also ®ts manyde® nitions of `traditional’ society and attitudes. Generally, traditional societiesare highgroup, thus maintaining high boundaries and rejecting outside in¯uences. Of the low group ways of life, modern or market society is predominantlyindividualist. Individualism can also include `charismatic’ soci- eties,because of the development of networks of followers based on personal abilityand qualities. Fatalism/ despotismcorresponds to the society described in Ban® eld’ s MoralBasis of aBackwardSociety ,21 as wellas thosein some US innercity environments, communism in crisis, 22 andpossibly societies in collapse,such as inRussia orLiberia,where people believe (often with reason) thatthey have no control over their lives and situations, and that their leaders do notconsider their interests or preferences, so theyare notpart of a `group’. Peoplein each way of life will behave and think in such a waythat they reproducetheir way of life, and there are certainbeliefs and actions which are uniquelycompatible with each way of life. Eachway of life provides its adherents with a coherentworld view which explainsthe world for them. Each interprets the same facts andevents differ- ently,and would have a differentde® nition of justice and the good society. A hierarchicalway of life would see astrati®ed and ordered society as just,each persondoing the work and receiving the rewards appropriate to his/ her 92 TANZANIAN PEASANT AND UJAMAA station.This Platonic ideal contrasts with the egalitarian view of justice as removalof differences in status and wealth, all of which are illegitimate. Individualistswould see ajustsociety as onein whicheach individual is free to pursuehappiness (individually de® ned, usually as wealthor status)according to his/herability and ambition. The fatalist does not believe that a justsociety exists,since he/ she believesthat life is randomand arbitrary, and that rulers makedecisions without concern for their interests. These views of justice are fundamentallycontradictory, based on different values and assumptions, and peoplein one way of life will not view other systems as justor fair. Equalityis anissue fordiscussion of Tanzanian policies and African social- ism,but each way of life views it differently.The egalitarian concept re¯ ects the socialistideal, equality of outcomeor condition.Individualists would see thatas unjust,denying people the fruits of theirlabour and diminishing the incentive to work.23 Theyvalue equality of opportunity,reduction of legal/structuralbarriers topersonal enterprise. Hierarchists do not value equality, because people are not equal,and democracy leaves them wandering about without direction or purpose. Fatalistsde® ne themselves as victims,unable to control their lives, so equality is irrelevantto them. These real cultural differences affect the kind of society whichadherents of each way of life will choose and build, and they affect evaluationof institutions and policies. Asocietyis notculturally uniform or monolithic. Different ways of life coexistwithin a societyand interact with each other, de® ning themselves in oppositionto thepractices and values of theothers. 24 However,societies do have dominantcon® gurations which will explain social organisation and policies. 25 Thequalities of Tanzanian society which I havedescribed above suggest an individualistpolitical culture, mediated by hierarchical tendencies in some communities,and in family relations. Even when ethnographers use theterm egalitarianto refer tosome practices,the practices are generallyidenti® able as individualistrather than egalitarian, and even hierarchical communities have underlyingindividualist tendencies. However, ways of life and cultural con®gurations are likelyto vary across ethnicgroups, so Iwillexpect variance, notsimply in surface practices,but in ways of life. Tanzania is adynamically heterogeneousand ethnically fractionalised society, and I expectculture to re¯ect that.

Essential characteristics ofTanzanian society Generalisationsconcerning the diverse forms ofTanzaniansocial and economic organisationare uncertain.However, generally organisation and political auth- oritywas basedon personal relationships, especially through family and clans. Personalisation,together with the family focus of subsistenceagriculture, meant thatpolitical power was mediatedby cross-cutting loyalties and relationships:

Evenin the most highly organized regions the extent of speci® cally political authoritymust not be exaggerated.African states were websof relationshipswhich grew steadilyweaker with distance from thecapital until they merged into the statelessnessof peripheral peoples ¼ Menstill had many loyalties: to nuclear 93 NANCY SPALDING

family,extended family, descent group, clan, village, patron, chiefdom, perhaps eventribe. One was relevantin onesituation, another in another ¼ thesocieties of 1800were inthemidst of the dynamic, autochthonous change which came from the minglingof diverse colonists. 26 Severaloverarching characteristics of the pre-colonial Tanzanian society have beenwidely observed, which I willdiscuss inturn: 27 Non-centralizationand relativeabsence of strong authority and control; family as theunit of decision; autonomous,diverse, interacting communities. These qualities are interdepen- dentand seemed tocharacterise the many varied kinds of social and political organisationto some degree.

Non-centralisation Non-centralisationsuggests the absence of a centralauthority which makes bindingdecisions on distant groups, and enforces them, whether by force or legitimateacceptance, or delegation to other authorities (as opposedto a ritual orsymbolic government which has littlein¯ uence on local policies). The Pare andUsambara, which attempted genuinely centralised states wouldbe excep- tions.Many groups were essentially stateless, such as theIraqw, the Ndendeuli andthe Makonde, and any organisation was atvillage level. Coordination betweenand even within villages was often ad hoc.28 Thestates oftenhad internalcleavages and alternate sources oflegitimacy and organisation, es- peciallyin lineage and clan. Material factors such as distance,as inthe case of manywestern chiefdoms such as theFipa, reduced further the ability of leaders tocommand, and the willingness of people to follow. However, material explanationssuch as distance,agriculture or technology are inadequateto explainthe extent of non-centraliseddecision making, which often extended into thevillage, across levelsof organisation.While centralisation was attemptedby militarycontenders for power in the 19th century, and was imposedlater by the colonialpowers, it had uncertain effectiveness, whether for material or cultural reasons.

Family Thefamily was generallythe fundamental unit of decisionmaking, as wellas of economicproduction. Families were fairly autonomous in many decisions, and oftenwere free toleaveone village or area andmove to anotherto seek unused landand economic opportunity, or to¯ ee fromquarrels and personality con¯ icts. Whilein some groups,lives were highly prescribed, there usually seems tohave beenfreedom to exit and colonise elsewhere.

Autonomous,diverse communities Communitieswere not isolated from each other; there were high levels of interaction,from migration and colonisation, and from trade in regionally specialisedcrops and crafts. Therefore, a community’s coerciveability was 94 TANZANIAN PEASANT AND UJAMAA limited,since other options were readily available. Communities generally had internalapproaches to problems and con¯ icts, and would have freedom to determinesurvival and growth strategies. Chieftainships were relatively indepen- dent.Even where there were complex systems ofchiefs and subchiefs, their jurisdictionmight be limitedto speci® c kindsof con¯ icts (especially inter-com- munitycon¯ icts), with the expectation that the community or individuals involvedshould resolve con¯ icts peacefully. Community autonomy is also indicatedby the lack of central coercive power, or taxes and tribute to transfer wealthto thecentre. Tanzania was composedof dozensof differentgroups, each withits own form of social and economic organisation, distinct from its neighboursbut interacting. 29 Thepicture presented above suggests that pre-colonial Tanzania was com- posedof a loosecollection of interacting and intermingling (especially through theextensive and language) but autonomous communities, which containedlittle strong authority and left great scope for individual and family action,within the parameters of a subsistenceeconomy. This generalised view suggestsindividualistic cultural tendencies. I arguethat these qualities are presenteven in hierarchicallyorganised societies, as wellas inthemore informal andstateless societies,and that they help to explain post-independence peasant behaviour.If theseare indeedthe qualities which can be associated with the Tanzanianculture, there are majorimplications for Tanzanian policy and politicallife. I arguethat the Tanzanian culture was incompatiblewith Ujamaa andTanzanian socialism. Below, I willdiscuss thecultural bases ofthese policiesin President Nyerere’ s writings.

Independent Tanzania:Nyerere’ sconceptionof the peasantryand Ujamaa JuliusNyerere was acharismaticleader of the nationalist movement, and presidentof theTanganyika African National Union ( TANU).Hebecame Tanza- nia’s ®rst PrimeMinister, then President, and was themajor philosopher and architectof independent Tanzania’ s developmentpolicy. He was knownfor his personalintegrity, magnetism, popularity, and his sel¯ ess devotionto the good ofthe nation as hesaw it;stories re¯ ecting this character are widespreadand unquestioned.Early independent government programmes were based on and justi®ed by Nyerere’ s socialistcommitment. He wrote and spoke extensively of hismoral commitment to equality and community, which he basedon thenature ofthe pre-colonial indigenous Tanzanian social structure, and so claimedwas culturallyappropriate. I haveabstracted from his words to provide a tentative summaryof hisperception of the nature of Africansociety. Did Nyerere’ s view ofthe nature of African or Tanzanian society re¯ ect the cultural and historical realityof the area? 30

Traditional society Oneof the major themes in Nyerere’ s writingson African society was thatof acommunitysharing available resources withall members: 95 NANCY SPALDING

InAfrican society ¼ traditionallywe livedas families, with individuals supporting eachother and helping each other on terms of equality.We recognizedthat each of ushad a placein the community, and this place carried with it rights to whatever foodand shelter was availablein return for theuse of whatever abilities and energieswe had.The old and the sick, or thosewhose crops had been destroyed by naturaldisasters, were notleft alone in their suffering. Other people shared with them,and did so without any feeling on eitherside that this was ªcharityºfrom the better-off,or involvedany loss of humandignity for theone who was (throughno faultof his own) in need ¼ Thecommunity was aunitin which every individual was important,and among which the goods available were sharedwithout too great inequality. 31 TraditionalAfrican society, as depictedabove, is basedon three major beliefs: ®rst,an obligation to work and participate in society; second, a communal sharingof property when members werein need; and third, respect due each personfor the role they played in the communal effort, even if it was aminor role.32 Nyererefrequently commented on the centrality of hard work to traditional life,and placed a highvalue on it, suggesting that rewards should be received foreffort as wellas need.However, communal ownership of property was also central;`¼ resultsof their joint effort were divided according to well-under- stoodcustoms on the basis ofthe fact that every member of the family had to haveenough to eat before any of them had any extra’ . Furthermore,this communalownership implied a kindof socialsecurity system: `Nobodystarved, eitherof foodor of human dignity because he lackedpersonal wealth; he could dependon the wealth possessed bythe community of whichhe was amember’ evenin a situationof no surplus. Land was alsoconsidered to be communal property,and the only individual rights to it were use rights,not sale or exchange.The fundamental values suggested by these principles were equality andparticipation, with no con¯ ict between the interests of the individual and thoseof thecommunity. Individualistic preferences for competition and individ- ualadvancement to achieveprosperity were seen as alien,capitalist exploitation, inculcatedby the colonial government. 33 Nyerere’s policywas structuredin accord with these communitarian ideals; Nyerereargued that hereditary aristocracy was aforeignimport. With no hierarchicalstructures or hereditary rulers, the Tanzanian community had been governeddemocratically, as `a societyof equals and it conducted its business throughdiscussion ¼ theytalk till they agree’ . 34 Nyererebased his concept of democracyon thisassumption concerning traditional Tanzanian government, and arguedthat `our ® rst stepmust be to reeducate ourselves; to regain our former attitudeof mind’ . However,traditional equality was notenough for a modern society,however valuable, since it was `anequality of poverty, the cooperation was onsmallthings’ . Hisaim was tomodernisethe base whichtradition had laid downand use itto meet modern goals. 35 Itis controversialto assert thatTanzania’ s culturalhistory is predominantly individualistic,especially in the context of President Nyerere’ s beliefin an egalitarianpre-colonial history. Careful analysis of ethnographic studies is necessary toexplore this issue. To pursue this question, and suggest whether it 96 TANZANIAN PEASANT AND UJAMAA is supportedby more speci® c evidence,I willexamine studies of the Makonde, Ndendeuli,and Shambala groups for evidence of their ways of life. 36 These studiesexplore historical practices, which, may not be followedtoday. However Nyererebased his policies on an ª historicalºTanzania, and so Ibasedmy critiquethere as well.This is apreliminarysurvey, and these groups are not representative,but it should be suggestive.

Makonde TheMakonde peoples 37 are locatedin the southern coastal plateau area, where theymigrated from Mozambique over a periodof centuries. The plateau provideda basis forshifting agriculture and protection from enemies, by hiding inthe impenetrable bush. Migration has continuedto beade®ning characteristic ofMakonde social structure. 38 TheMakonde maintain two signi® cant group memberships, kinship and village.However, these two ties are fairlyweak in terms ofprescription of behaviourand maintenance of group boundaries. Inheritance is matrilineal,so youngmen move from their family’ s hometo their bride’ s, andoften work under herfamily’ s authority.Separations, sometimes permanent, between the residence ofa manand his family might result, weakening kinship ties. However the familydid prescribe with respect to inheritance, religion and taboos, and it providedassistance inold age and emergency, so ithad a strongerhold than the village.Villages were generally scattered settlements which exercised little authority,partly because of thepresence of multiple lineages and the prevalence of¯ightand resettlement. Other networks also formed, meeting social, economic andpolitical needs, especially through initiation rites and neighbourhoods. 39 Thesenetworks were independent of both village leader and lineage groups. Theof® ce of`leader’ in a villagewas hereditary,based on descent from the ®rst coloniser,but he hadlimited authority and few special privileges (virtually nocoercive authority). The primary duties were spiritual, involving ancestor worship.40 Anotherrole which commonly fell to leaders was mediationof disputesand negotiation of settlements,if theyhad a reputationfor wisdom and skillin mediation. Such skill was respected,since there was nouniversally acceptedcode of behaviour and no socially de® ned superior or neutral arbitra- tors.Respect based on wisdomwas earnedby theindividual, not inherited. 41 The leaderdid not rule, and countervailing power bases existed. Decisionswere made collectively by interested parties, partly because of the manycountervailing groups and loyalties within a smallcommunity. 42 The major countervailingauthority lay in the elders of other family groups, who protected theinterests of their groups. Disagreements including restitution, were negoti- ated,and advantage in negotiations often depended on whohad the greatest force orin¯ uence available. Disruptive behaviour which threatened the harmony of thecommunity was controlledthrough judicial settlements which sought compensationfor injury rather than punishment of deviance. 43 Themain economic sanction available to the leadership was assignmentof vacantland to new immigrants. However, that control was limited(after two years,new immigrants had equal rights with other residents), and there were no othereconomic sanctions. The climate and technology provided little potential 97 NANCY SPALDING foraccumulation of wealth. Taxes were minimal, and ritual tributes and gifts(including bride price) were tokens, reducing the transfer of wealth and hencethe potential for economic coercion. In general, sanctions were almost non-existent,either aimed at non-members (recent immigrants) or stemming fromnegotiation. After all, sanctions are ineffectivewhen the exit option is availableand attractive. 44 Whatwere the primary characteristics of the Makonde discernible from Liebenow’s study?Prescription was virtuallyabsent. Each community had differentpractices. There were no universally accepted codes of behaviour, and rightsor privileges were determined by negotiations which were in¯ uenced by therelative strength of the parties. 45 If localnorms were unacceptable, an individualcould leave. Leadership had little practical power, since it was not inheritableand had no sanctions. Ascribed status and roles were minimal. Groupmembership was veryloose; lineage was important,but only the most importantof severalloose ties, and the combination of marital practices and the propensityto migrate reduced the strength of the lineage group as asupport system orde® ner of social identity. The Makonde were characterised by theformation of networks of relationships, based on different criteria (line- age,location and colonisation, interest, marriage, friendship, etc) and they functionedwithin these overlapping and cross-cutting webs, 46 ratherthan as members ofascribed groups. In addition, religion did not support either group cohesionor prescription of behavioural norms, since there was nocentralised religiousauthority. Religion was basedon ancestor worship, so each person or familyunit could practise religion without outside interference or assistance. Thesequalities suggest a stronglyindividualistic way of life, characterised by personalfreedom. Thereare threemajor exceptions to this individualistic social structure, suggestinga hierarchicalsubculture. First, there was a`castelike’relationship betweenthe Makonde and more recent immigrants. This included some occu- pationalspecialisation and some contemptfor `unclean’ personal habits. How- ever,this strati® cation may not have played a majorrole in dailylife. The second hierarchicalcharacteristic was theability to `pawn’ relatives (especially female) anduse themas securityfor debts, or inplaceof labourowed as compensation. However,pawns were not permanent, had rights, were not slaves, and it is unclearhow prevalent the practice was. Even so, pawning suggests some clear statusdifferences (mostly gender, which followed in other areas offamily life). Thethird exception was thepractice of keeping domestic slaves. Slave status was essentiallypermanent, and slaves hadfewer rights than pawns. These exceptionsdo notseem tohavebeen important in daily life, but they do suggest adistinctthread of hierarchical organisation running through the culture. 47

Ndendeuli TheNdendeuli, 48 alsoof southern Tanzania, clearly de® ne themselves and their interactionsin terms ofkinshiprelations: `Rights and obligations, privileges and responsibilities¼ wereubiquitously and continually explained and morally justi®ed in terms ofthekinship relations of people.’ 49 However,this apparently 98 TANZANIAN PEASANT AND UJAMAA stronglybounded group identity was mitigatedby the highly ¯ uidand `ego centred’de® nitions of kinship. Each individual forms `complexnetworks’ for purposesof social interaction and economic cooperation. These networks (kin- sets) includethose persons who are reliableand available to assist incooperative endeavours.Each individual will de® ne his or her kin-set differently, though kin-setswill overlap. Kinship is de®ned primarily as involvingthose people with whomone has areciprocalrelationship, rather than having relationships with peoplewho are kin.This somewhat counter-intuitive system seems toorganise agriculturallabour and social interactions effectively. As withthe Makonde, the Ndendeuli are acolonisingpeople. They lived in smallcommunities, 50 andpractised shifting cultivation, growing crops in one ®eld(adjoining the home as nearlyas possible)until it was infertile.They would thenmove the ® eldby clearing unclaimed land, still remaining close to the home.After a numberof years,when no usableland was convenientlyavailable forcultivation, they would move. These moves were not coordinated, families wouldmove in different directions and at different times, and new communities createdwould not resemble the ones dissolved. A fewheads of households would® nda newlocation and settle. Other immigrants would join later, generallysponsored by a presentmember of the new community. Honour accruingto the ® rst immigrantsdid not translate into real power or authority; evenin¯ uence was minimalaside from whatever personal reputation developed. Communitiesformed clusters of households to facilitatecooperation for some of thetasks ofclearing bush and agriculture, 51 and`persisted only so longas they retainedtolerable supplies of cultivable land, and as theyadequately satis® ed men’s needsfor fairly reliable cooperation with their neighbors’ . 52 Theregular, extensive economic cooperation suggests clear forms ofsocial organisation,but based on neighbourhoods, shifting relationships and varying needs.No organising principles such as commondescent, property rights or commondefence bound communities together. 53 The`rules’ by which interac- tionswere guided were ego-centred, situational and interpreted conditionally andcontinually. Obligations were reciprocal though shifting and conditional; understandingschanged over time, and were based on practical realities. Thebasic unit of action in the community was thehousehold, headed by a marriedman who organised his household, administered it as aneconomic unit andrepresented it to the community. 54 Thehousehold would include wives, childrenand young unmarried men, and any dependents related to either the husbandor the wife (such as widowsand their children). Young married men hadlabour obligations to their fathers-in-law, and so residednear them in the earlyyears ofmarriage. In a village,any head of household was relatedto severalother heads of household,either through blood or marriage.If therewere toofew `kin neighbours’ to provide a base foreffective economic cooperation andsocial interaction, more relatives would be sponsoredto jointhe community. Anindividual would have a numberof households which would provide a logicalpool of cooperation. However, as mengrew older, by preference they wouldgradually develop `somewhat diverging sets ofactive kinship relations andthe desire for a greaterdegree of autonomy of choice in social action’ . 55 Networksare developedbased on some conceptof kinship,but it was a¯uid 99 NANCY SPALDING conceptthat depended upon convenience and personal af® nities.Each individual wouldhave many linkages to different people to whom their kin were not related;de® nition of kin neighbour was individuallynegotiated, not ascribed. Menaligned themselves with those kin-members with whom they had closer obligationsand af® nitiesand `developed their own sets ofkin-linkages that becamepartly idiosyncratic to each individual’ . Linealkinship was nota strong factor,and it seemed foolishto `impose arti® cial limits on the range of social relationships’. Practically,these networks worked out in ad hoc workgroups recruitedfor a speci®c purposefrom kin-neighbours. 56 Desirefor personal choice in relationships and household decisions is re¯ected inthe orientation towards community in¯ uence and leadership. `Household headswere considered to be equals who did not recognise any authority over themwithin the community’ . 57 Leadershipwas tentativeand based on personal prestige.If anindividual was talentedand able, he might use hiswisdom and in¯uence in resolving disputes and supporting his kin-neighbours and friends in disputes.A personcould aspire to statusas anotableor abigman, but that status was fragile,dependent on continual success. Con¯ict resolution demonstrates the network processes. Whena disputearose withina community,there was astronginterest in resolving it quickly so that thenormal patterns of cooperation could continue undisrupted. The dispute mightbe amicable, over con¯ icting work parties, or more personal, over debts, insultsor strongly disruptive or antisocial behaviour. A seriouscon¯ ict might disrupteconomic cooperation and social relationships throughout the com- munity,since networks were inextricably linked. A `moot’or meeting would thenbe called, convening an assembly of neighbours on neutral ground. The principalswould be present with their action sets andpossibly other interested partiesor notables.Individuals who did not want to supporteither side (because ofcon¯ icting obligations) might be absent, visiting distant kin. `There were fewrules of procedure; and there was noone to act as chairman,let alone as arbitratoror judge.’ The matter would be discussed, facts wouldbe agreed to,and a decisionconcerning resolution and compensation (if necessary) would bereached. Notables had the opportunity to use theirwisdom and in¯ uence toresolve con¯ icts, and would increase their prestige if effective. The mootwould usually end genially, with the drinking of beer and the restora- tionof social relations. There were no sanctions available to enforce judgments;cooperation was voluntary. 58 Economiccooperation provides more examples of the shifting, idiosyncratic networksof cooperationamong the Ndendeuli. 59 Economiccooperation involved coordinatingwork parties to share inthelabour of clearingnew ® elds,repairing houses,preparing land for planting, and the actual planting. Kin-neighbours wouldbe asked to join a workparty formed for speci® c tasks,and they would joinbased on their other obligations, their relationship with the organiser, and theirinterest in maintaining reciprocal obligations. Families participated in the workparties of those with whom they had reciprocal obligations, and those with whomthey wanted to strengthen obligations. These ties were originally formed dueto kinship bonds and over time became self-de® ning. However, each ® eld 100 TANZANIAN PEASANT AND UJAMAA was clearlythe property of theman who had claimed it and his produce was his alone. Interms ofculture theory, the Ndendeuli are clearlyvery low grid; the behaviourof heads of households is notprescribed by social conventions and mores,and individuals have tremendous autonomy in social action. Kinship is de®ned so looselythat individuals de® ne their kinsmen for purposes of cooper- ationand social interaction. There are nonegative consequences or sanctions for buildinga kin±neighbour network in aparticularway, and men participate in an action-setor networkfor a varietyof reasons. 60 Therefore,group boundaries are clearlyweak. The individualist label is supportedby culture theory’ s contention thatego-centred networks are theprimary organisational form of low group socialrelationships: `If relationshipsare organizedinto networks, the pattern you willtrace out will be unique to the individual you have chosen ¼ theyspread outall over the place, have no boundariesand are as numerousas theindividuals whobuild them.’ 61 Thisdescription clearly ® ts theNdendeuli.

Shambala TheShambala (or Shambaa) kingdom, in the Usambara mountains of northeast Tanzania,forms afar morehierarchically organised system thando theMakonde andthe Ndendeuli. It is organisedthrough lineage. Chiefs or royallineages have authorityover interlineage relations and legitimate monopoly over coercion, thoughthere are councillorswho advise and check chie¯ y power.The system’ s abilityto control all aspects oflife is extensiveand group identity is strong, thoughthe option of migrating and founding a lineageexists. Thebasic residential and economic unit is thehousehold, which is built aroundan autonomous family, ideally with multiple wives and many children. Eachwife has aseparatehouse, ® eldsand stock which are allocatedto her at her marriage,and inheritable through her by her children. This property cannot be takenfrom her without her permission. The family is nestedwithin a lineage, andmost lineages are fairlylocalised, living within a village.As withthe Makondeand Ndendeuli, the lineage of the ® rst migranthas leadershiprights, butfor the Shambala, there is morereal power associated with inherited leadership.Land tenure is permanent,rather than temporary, and land ownership andrights are inherited,though the lineage retains rights so thatland cannot be alienatedor disposed of outside the lineage. 62 Thelineage has stronggroup boundaries; any action taken by a memberof a lineageis theresponsibility of the entire lineage. When ® nes orcompensation are required,they are paidby the lineage as agroupto the offended lineage as agroup.`Kinship groups ¼ retaina considerableresponsibility for the security oftheindividual and the regulation of social relationships ¼ itis thelegitimate businessof the lineage to settle’ any disputes within the lineage. While villages are generallynot completely of one lineage, different lineages may be merged bymarriage and genealogies may be falsi® ed to increase solidarity. 63 Withinthe lineage, there are clearstatus differences which prescribe the range ofacceptable behaviour closely. Prescriptions include behaviour expected betweengenerations, generally in terms offormality and respect. Relatitve age 101 NANCY SPALDING withina generationis alsosigni® cant, since eldest sons have® rst claimon bride wealthand tend to inherit more than younger sons. Villages are ledby the acknowledgedleader of the ® rst lineageand these leaders have far greater authoritythan among the Ndendeuli, mostly thanks to theposition of allyounger menas deferentialto them, and their role as intermediarieswith non-members. Leadersalso have sanctions at their disposal, though these are weakeningover time.64 Thechiefs represent a levelof authorityabove village leaders. Lineage heads havefamilial authority, while chie¯ y authorityis `expresslypolitical in nature’ . Chiefsmonopolise the exercise of authority in interlineage relations, and some acts are de®ned as achallengeto the chief’ s authority.For example, murder is prohibited,since only the chief can allow killing of any kind: `it eats the chiefshipfor common people to do the things that only a chiefcan do’ .This pointsup adifferencebetween the Shambala and other groups; while most acts, includingmurder, were civil issues tothe Makonde and Ndendeuli, requiring compensationto the victim based on discussion and negotiation, the Shambala havea realsense that`murder is aninjury to the body politic and a challenge tothe authority of the chief rather than a privateinjury’ . 65 Thechief has theauthority to command services fromhis subjects, collect ®nes andexact tribute and taxes. The chief and the system ofwhich he is part alsode® ne the `public goods’ 66 whichare collectivelyprovided through obliga- torycommunal labour and taxes. He commandsthe allegiance of hispeople, and throughthat their property, though the burdens tend not to be onerous. Winans arguesthat the system is weakenedby the chiefs’ technological inability to controltheir distant subjects. 67 However,control of behavioris highlyeffective. Technologydoes not determine the hierarchical nature of the society and the system functionseffectively. Explicitlines of authorityand deference exist within chiefs’ courts as wellas betweendifferent level of chiefs. However, countering concentration of power upward,there is aninstitutional check to the power of chiefs. The councillors advisethe chief on many issues, announcehis decisions and determine succes- sion,based in part on agreement of commoners as tothe best chief. The legitimacyof the chiefs’ rule is basedon the support and acquiescence of their subjects,as inthe belief `that the royal chief is achiefbecause the people wish him to be’ .68 TheShambala system has clearlydelineated roles and relationships, with sanctionsfor violating norms. There are strongdeference patterns and differen- tialsin status based on ascribed criteria. Nor is thisan imposed system; there seems tobe `a beliefin the inborn ® tness ofthe ruling class torule’ . 69 These characteristicsre¯ ect hierarchy. Group boundaries are alsohigh; there is true corporateauthority, group membership is thede® ning characteristic of one’s life, andgroups act as representativesfor their members. Even though ancestors are viewedas capriciousand controlling one’ s luck,often negatively, they were believedto be controllable with the appropriate formula of sacri® ces, which suggestsa culturewhich is basedon rules and appropriate behaviour to control andstabilise the world (an extremely hierarchical view of nature or life). Evenwith this very hierarchical social system, there are individualistic 102 TANZANIAN PEASANT AND UJAMAA undercurrents.The ® rst is themore casual relations between grandparents and grandchildren,which circumvent prescriptions concerning relations between relativesand between generations. (This re¯ ects the`joking relationships’ which are foundin many African societies, whereby casual, even bizarre or violent behaviouris allowed,even expected, between people who have certain relation- ships).The second is theoption of migration, which remains viable, though probablynot the most desirable. These provide some `escape valves’for pressures buildingup within society.

ATanzanianpolitical culture? Giventhe variation in the groups discussed above, it wouldbe dif® cult to select oneorganising principle or way of life which predominates in traditional Tanzaniansociety. There are certaincommonalities, explained by culture and reinforcedby technology, but no universal or monolithicculture. These common attributesinclude the `loosely linked small communities’ , 70 thepropensity for migration,and the cross-cutting loyalties and obligations. Many authority struc- tureswere weak, but certainly not all, though the state had little ability explicitly tocontrol daily life. Among the Shambala, intrusion was unnecessary,since normsand prescriptions legitimately determined much individual behaviour and acceptedrules governed con¯ ict management. While experiences varied, many pre-colonialTanzanian peasants may have experienced some autonomyin subsistenceand lifestyle preferences. Obedience to a centralauthority was notan aspectof the Tanzanian peasant’ s life,nor was cooperationon alargescale. This is alogicaldevelopment from a subsistenceeconomy based on rain-fed agricul- turein a marginalarea withunreliable rains and low level technology, but certainlynot the only logical development. A materialistexplanation is inad- equateto explain differences, while cultural analysis helps to explain variation insocial systems builtaround different preferences.

Nyerereand Tanzanian culture Thereare anumberof problems with Nyerere’ s analysisof traditional society. Hespeaks ingeneral terms of`African society,’ or `the African family’ , with littleconsideration for the tremendous diversity of social forms foundin Tanzania,and throughout the African continent. He identi®es Africansociety as egalitarian,composed of small communities and families who owned property communallywith no status differentials or hereditary aristocracies. However, thereare hierarchiesin Tanzania; Nyerere’ s insistencethat the few Tanzanian aristocracieswere composed of immigrants, rather than indigenous peoples, avoidsthe fact that some ofthesemigrations occurred in thedistant past, and the mythsof origin, whether factually true or not, were used to legitimate the hereditaryleadership’ s rightto rule. Nyererealso suggested that small communities owned property in common. However,there is adistinctionbetween land which cannot be alienated from a familyor lineage and land which is communallyowned, with shared rights to appropriateproduce. In Mount Kilimanjaro, land was privatelyowned and meticulouslydivided. The Shambala practised inheritance of landwithin a family 103 NANCY SPALDING andthe rights were quite speci® cally divided among wives and through them to theirchildren. These property rights could not be changed without the wife’ s agreement,though the land could not be alienated outside the community. Furthermore,many communities such as theNdendeuli were composed of shift- ingcultivators, who would farm availableland until it was infertile,and then moveon to clear more unoccupied land. Unowned and unoccupied land is not communallyowned, and the concepts of property ownership, communal or individual,beyond short-term use rights(inferred by clearingbush), would have beenirrelevant. Communalland ownership implies egalitarian community relations. Nyerere suggeststhat communities shared their produce with members inneed, without obligationto repay. Generally, the basic economic unit was thefamily; families sharedtheir resources andowned them in common(though control and appropri- ationrights were divided along various criteria). Consumption and contribution ofagricultural labour were not equal within a family;in some cases, eachwife mighthave her own plot of land, from which she grewfood for herself, her childrenand husband. A moreproductive woman would feed her children better. Communalaccess tofoodand shelter did not extend to the village, clan or tribe. Tanzaniancommunities and peoples were so diversethat it isnotuseful to speak ofthem generally. However, there are some generalstatements which can be made: ®rst,outside the family unit, there was little`egalitarian’ social organis- ationas Nyerereintends it. Villages did not hold their land in common, and orseed was borrowedif shortages were found, rather than given as aright.In some communities,cooperation was usedfor agricultural chores, but the produce ofthe ® eldswas strictlythe family’ s. Itappears that Nyerere’ s pictureof a primitiveAfrican or Tanzanian socialism is notsupported by evidence.

Conclusion Tanzania’spoliticalculture Ireturnto my original question: was Nyerere’s visionof pre-colonial Tanzania accurate?Nyerere saw Tanzaniantraditional society as homogeneousand egalitarian,and he spoke of this legacy and its opposition to individualistic Europeanculture often and explicitly. The Tanzanian peasantry were natural socialistswho had been corrupted by thecolonial system. 71 Theyneeded only to relearntheir natural cooperative patterns, and be taught how to upgrade their skillsand technology so theycould take advantage of economies of scale to createmore wealth, which would make all better off. However, Tanzanian historyhas manyexamples of entrepreneurial behaviour independent of, or opposedto, colonial or independent government preferences, examples of which havecontinued in the contemporary `second economy.’ 72 Thissuggests a confusionof colonial capitalist exploitation with indigenous individualism, and afundamentalmisunderstanding of the nature of Tanzanian political culture. Nyereretook some realaspects ofTanzanianlife, generalised from them to all ofTanzania and Africa, and applied his principles beyond their appropriate- realm.While his references to`African’ traditional society may have been intendedmetaphorically, his beliefs about Tanzania were used as thebasis ofhis 104 TANZANIAN PEASANT AND UJAMAA policyprogramme, and as `a kindof mobilizing political myth which would enablethe government to explainthe ujamaa programto thepeasantry and enlist itssupport ¼ {the}historical accuracy of an ideology { is acrucialquestion} in determiningits success asamobilizinginstrument’ . 73 Nyerere’s interpretationof Tanzanianculture and the social realities which shaped Tanzania’ s historywas mistaken,and contributed to disastrous policy failures. The compromises and shiftsin his policies over the years, as socialismwas noteffectively imple- mented,and the people were no better off, may be interpreted as tacitthough reluctantacceptance of this contradiction. 74 Culturalanalysis suggests that, while ways of life varied, many Tanzanian peasantsdemonstrated individualist qualities, and could make independent choicesconcerning subsistence and lifestyle preferences which were relatively unconstrainedby complex and intrusive systems ofprescriptions. Options includedmigration, allowed exit from communities because of personalcon¯ icts orlimited opportunities. Equality of condition and limits of freedom seem to havebeen primarily the result of thelimits imposed by technology, environment andthe vagaries of subsistence agriculture. There were few indications of egalitarianculture. Though hierarchical systems existed,and were widespread, especiallyin response to the Ngoni invasions and the disorder of the 19th century,75 obedienceto a centralauthority was notthe predominant experience oftheTanzanian peasantry before colonial rule, nor was large-scalecooperation. Whilethere are dozensof other groups to consider, many of which have never beencarefully studied, my ® ndingsare suggestive.

Culturetheory and African politics Culturetheory has adistinctivecapacity to take analysis of Africanpolitics and policya stepfurther than other modes of analysis,and infer appropriate policies andstructures from the dominant ways of lifewithin society. Culture theory can explainand compare varied behaviours, in different times and places, by their functionsfor supporting a wayof life.It reminds us thatinstitutions and policies mustbe consistentwith the values and preferences of thepopulation if theyare tobe effective.It addresses fundamentalquestions of humansocial life: Who am I?Whatam Itodo? 76 Andso itfacilitates understanding of social processes. It cantake us beyondgeneralisations concerning peasant behaviour in EastAfrica, toexplain the ways in which people respond to their environment, and the varied institutionswhich they build to meet their needs. Tanzania does not have a homogeneouspeasantry, so responses cannotbe predicted by amonolithictheory. Thequestion of whether policies can be used to change ways of life is still open;Nyerere and others deeply believed that individualistic practices were unjustand would harm development. 77 Whywas theelite egalitarian and hierarchicalif the population was individualist?What differing cultural in¯uences created this division between elite and peasants? More than that, how feasibleis engineeredcultural transformation? How far canstructures be moved fromvalues without disaster? Fatton, while agreeing that pre-colonial Tanzania was notnaturally socialist, suggests that proper programmes could create a socialistculture. 78 Thisfollows the logic of egalitarian culture (and socialist 105 NANCY SPALDING philosophy),which suggests that human nature is aproductof socialintercourse; peopleare borngood and corrupted by theirinstitutions, so iftheinstitutions are changed,people can be transformed. 79 Howamenable to change are apeople? Culturecan and does change. If thematerial and political reality does not supportthe cultural world view, eventually people will shift to one which is morecongruent with realty. 80 However,natural change in responseto signi®cant contextualshifts is differentfrom engineered change, which is notoriously dif®cult, though often attempted, and frequently leads to unintended conse- quences.This remains an open question, worthy of study. Tanzania’ s failureto changeits cultures could be the result of cultural resilience, ineffectiveness of socialengineering or erratic, arbitrary and often coercive policy implemen- tation.81 Furtherresearch is necessary,though a de®nitive answer may be impossible.However, culture theory provides a potentiallypowerful tool for explainingTanzanian and other African development and political anomalies.

Notes Iam gratefulto Aaron Wildavsky for the guidance and encouragement he gaveme onan earlier draftof this paper.His untimelydeath is deeplymourned. Larry Diamond, Goran Hyden, Sheen Rajmaira andTony Waters alsogave me valuablesuggestions and it is nottheir fault that I didnot follow all ofthem. 1 JHWeaver &AKronemer, `Tanzanianand African socialism’ , WorldDevelopment ,9(9/10),1981, pp 839±849. 2 MaryDouglas is aclassically trainedBritish anthropologist, who studied ritual and religion in African societies. Fromthat she became interestedin culturally varying perceptions of risk. She developed the theory employedbelow in Purityand Danger: An Analysisof Concepts of Pollutionand Taboo ,London:Routledge &KeganPaul, 1966; NaturalSymbols: Explorations in Cosmology ,London:Barrie &Rockliff,1970; ImplicitMeanings: Essays inAnthropology ,London:Routledge & KeganPaul, 1975; and (with Aaron Wildavsky) Risk andCulture: An Essay intheSelection of Technical and Environmental Dangers , Berkeley, CA: Universityof California Press, 1982. 3 Kimambo disagrees. He suggeststhat a process ofcentralisation had been occurring in some partsof Tanzania overseveral centuries,probably accelerating inthe 19th century. Hierarchy andcentralisation are notnecessarily synonymous;hierarchy suggests a particularform of social relations,where each personhas aplace inthesystem, knowsthat place, itsduties and responsibilities, and receives rewards commensurate withthat place, as willbe discussed more fullybelow. Centralisation on theother hand suggests an authority whichdictates anddecides policyfor members, andis able toenforce thatpolicy. This is onepossible logicalextension of hierarchy, especially underconditions of coercive power,or advanced technology or communications.I Kimambo,`The interior before 1800’ , inI Kimambo &ATemu (eds), A ,Nairobi:East African PublishingHouse, 1969, pp 14± 34, esp p22. 4 J Iliffe, AModernHistory ofTanganyika ,Cambridge:Cambridge University Press, 1979,p 13. 5 Ibid, p 15. 6 Ibid,pp 15, 529. Goran Hyden pointed out to me theunique importance of theSwahili language and culture inTanzania. 7 As was muchof Africa. IKopytoff(ed), TheAfrican Frontier: The Reproduction of Traditional African Societies,Bloomington,IN: IndianaUniversity Press, 1987. 8 Iliffe,A ModernHistory; Kimambo, `The interior before 1800’ , p22. 9 Names suchas Ndendeuleand Nyasa denote`people of thebush or thevalley, the savannah, the forest, or thelakeshore’ . ROliver& GMatthews(eds), History ofEastAfrica Vol1, Oxford: Clarendon Press, 1963, p 207. 10 JFlint(ed), TheCambridge History ofAfrica ,Vol5, from c1790 to c1870, Cambridge: Cambridge UniversityPress, 1976,p 310;and Oliver & Matthews,History of East Africa, pp191±192. 11 Iliffe, AModernHistory , p 24. 12 ARoberts,`Political change in the nineteenth century’ , inKimambo &Temu, AHistory ofTanzania , pp57±84; and Flint, TheCambridge History ofAfrica , ch 8. 13 Kimambo,`The interior before 1800’ , pp22±23, 30. 14 J Vansina, Pathsin the Rainforest: Toward a History ofPoliticalTradition in Equatorial Africa , Madison, WI:University of Wisconsin Press, 1990;and Oliver & Matthews, History ofEast Africa , ch 6. 106 TANZANIAN PEASANT AND UJAMAA

15 Culturetheory has beenapplied to a varietyof policy issues, especially inthe USA andEurope. Though ithas notbeen applied systematically toAfrican andother `developing’ societies, itcan enrichour understandingof these social systems andtheir problems. 16 CLockhart& GFranzwa, `Culturaltheory and the problem of moral relativism’ ,inD Coyle& RJEllis (eds), Politics,Policy and Culture ,Boulder,CO: Westview Press, 1994,pp 175±190; p 181.This model is relativistabout fact, meaning`there is nosuchthing as objectiveknowledge of realities independentof the knower,’since thesocial environmenthas acentral rolein determining beliefs aboutwhat is andwhat shouldbe. A Flew (ed), ADictionaryof Philosophy ,New York:St Martin’ s Press, 1984,p 303. 17 Idonotargue that culture is independentof material conditions,but rather thatit interacts withthe objective environment,and shapes reactionto it. A culturalworld view will interpret and construct reality in a certain manner,and shape social responseto that socially constructed reality. If circumstances changeand the responsebecomes ineffective,people will move towards alternative world views whichcan more success- fullyexplain reality. In 19th century Tanzania, for example, some communitiesadopted more hierarchical forms basedon military discipline in response to the Ngoni raids and disorder. 18 REllis,`Why cultural theory?’ in G Grendstad& PSelle (eds), Cultureas a Wayof Life ,Oslo,Samlaget, forthcoming1995. 19 RKlitgaard,`In search ofculture: a progressreport on research onculture and development’ , unpublished manuscript,1991, p 90. 20 MThompson,R Ellis& AWildavsky, CulturalTheory ,Boulder,CO: Westview Press, 1990,p 5. 21 E C Ban® eld, TheMoral Basis of a BackwardSociety ,New York:Free Press, 1958. 22 JClark& AWildavsky,`Why communism collapses: the moral andmaterial failures ofcommand economiesare intertwined’, Journalof Public Policy 10(4), 1993, pp 361±390. 23 Fordiscussion of these twoconcepts of equality,see AFlew,`Enforced equalityÐ or Justice?’ , Journal of LibertarianStudies 8(1),1986, pp 31±41. 24 Thompson et al.,CulturalTheory, p 215. 25 AWildavsky, TheRise ofRadical Egalitarianism ,Washington,DC: American UniversityPress, 1991;R J Ellis, AmericanPolitical Cultures ,New York:Oxford University Press, 1993;D Coyle& RJEllis, Politics,Policy and Culture ,Boulder,CO: Westview Press, 1994. 26 Iliffe, AModernHistory , p 25. 27 JohnIliffe has writtenthe most extensive history of Tanzania,from pre-colonial times toindependence, and hiswork suggested these themes tome. Theyare alsoobservable in the many ethnographies which have beenwritten of different groups in Tanzania, and are suggestedin other interpretations of peasant interactionsand communities. While I may draw differentconclusions, I accept thedescriptions. Iliffe, A ModernHistory . 28 Iliffe, AModernHistory ;PHGulliver, Neighboursand Networks: TheIdiom of Kinship in Social Action Amongthe Ndendeuli of Tanzania ,Berkeley,CA: Universityof California Press, 1971;J GLiebenow, ColonialRule and Political Development in Tanzania: The Case ofthe Makonde ,Nairobi:East Africa PublishingHouse, 1971. 29 Kimambo,`The interior before 1800’ , pp26,23. 30 Thisis aproblematicquestion, since Nyerere’ sintentionwas nation-buildingand developing support for his programmes,rather thanscholarly exegesis. However, if his vision was incompatiblewith social reality, potentialsuccess ofprogrammes wouldbe affected. 31 J K Nyerere, Principlesand Development ,, InformationServices, 1966,p 9. 32 J K Nyerere, UjamaaÐEssays onSocialism ,London:Oxford University Press, 1968,p 107. 33 Ibid,pp 2±7; 106± 110; J KNyerere, Freedomand Unity: A Selectionfrom Writings and Speeches 1952± 1965,Dar es Salaam: OxfordUniversity Press, 1965,pp 105±106; After theArusha Declaration , Dar es Salaam: Ministryof Information and Tourism, 1967, p 12. 34 Nyerere, Freedomand Unity , p 103. 35 Nyerere, UjamaaÐEssays onSocialism , p 6; After theArusha Declaration , p 15. 36 These studieswere chosenunsystematically, and are notrepresentative ofTanzanian society. A more exhaustiveexploration (in process) does® ndall fourways oflife.I willaddress thisissue inthe conclusions. 37 Bantupeoples, including the Makonde, characteristically haveno central politicalauthority; their primary identityis throughcommon lineage, language and beliefs. The discussion below is drawnfrom Liebenow’ s ®eldworkand description of the Makonde. 38 Liebenow, ColonialRule and Political Development , pp 24±30. 39 Ibid, p 59. 40 Thespiritual duties of ancestor worshipprovide no scope for control, since itwould be unthinkable to individualsfrom religious observances except for heinous acts. 41 Thisdistinguishes the Makonde from hierarchical systems suchas theShambala, where theleader is consideredto have strong ritual powers, including rain-making, and the leader’ swell-beingis essential to thewell-being of the society as awhole. 42 Mostdecisions are made withoutreference tothe leader, bythe interested parties. 43 Liebenow, ColonialRule and Political Development , p 57. 107 NANCY SPALDING

44 G Hyden, BeyondUjamaa in Tanzania: Underdevelopment and an Uncaptured Peasantry ,Berkeley,CA: Universityof California Press, 1980. 45 Liebenow, ColonialRule and Political Development , p 47. 46 Ibid,pp 44± 5, 70± 1. 47 Ibid,pp 34± 5, 66. Aaron Wildavsky suggests that all fourways oflife are presentin any society, though indifferent mixes, and the speci® ccon®guration explains policy preferences. 48 Thisdiscussion is basedon Gulliver’ s discussionof the Ndendeuli. Gulliver, Neighboursand Networks . 49 Ibid, p 4. 50 Acommunitymight contain 15 or 20hamlets (small clusters ofautonomoushomes) totalling perhaps 150 to200 people in two to three squaremiles. Gulliver, Neighboursand Networks , p 54. 51 Cooperationin these tasks was believedto be necessary bythe Ndendeuli, though the actual material practical valuehas beenquestioned by agriculture specialists. Gulliver, Neighboursand Networks , pp 195± 196.Whether materially necessary ornot, cooperation facilitated social organisation. 52 Gulliver, Neighboursand Networks , p 54. 53 Ibid, p 55. 54 There were nofemale headedhouseholds among the Ndendeuli. Women needed the labour, coordination of agriculturalcooperation, and community representation which only a man couldprovide. Gulliver, Neighboursand Networks , p 56. 55 Ibid,pp 59, 61. 56 Ibid,pp 223, 61, 126, 18. 57 Ibid, p 65. 58 Ibid,pp 135± 136. 59 Ibid, ch. 6. 60 Ibid,pp 225, 227. 61 Thompson et al., CulturalTheory , pp 11, 12. 62 E V Winans, Shambala:The Constitution of a TraditionalState ,Berkeley,CA: University of California Press, 1962,pp 32±43, 49. 63 Winans, Shambala, pp 66, 104. 64 Ibid,pp 51± 54. 65 Ibid,pp 100± 105. 66 Publicgoods include maintenance ofchiefs’ housesand royal burial places, andthe construction and maintenance ofroads and irrigation canals. Winans, Shambala, p 134. 67 Ibid, p 120. 68 Ibid,pp 128, 117. 69 Ibid, p xx. 70 Liebenow, ColonialRule and Political Development , p 31. 71 See Thompson et al., CulturalTheory , pp 59±60. 72 Forpre-independence examples, see NSpalding,`Resource mobilizationin Africa: therole of local organizationsin the independence movement’ , Journalof Developing Areas ,28(1), 1993, pp89±110. The second economy in Tanzania has beenextensively documented, and provides the livelihood formany who could not support their families inthe of® cial economy.See TLMaliyamkono& MSD Bagachwa, TheSecond Economy in Tanzania ,Athens,OH: OhioUniversity Press, 1990;N Spalding,`The state±society debate in Africa: anexploration of theTanzania experience’ ,forthcoming,Polity (fall 1996). 73 MFLofchie,`Agrarian socialism inthe Third World: the Tanzanian case’ , ComparativePolitics , 8 (3), 1976,pp 479± 499, esp p488. 74 McHenryargues that shifts in policy are theresult of power struggles between ideologicalfactions within thegovernment, responding to the economic, political and social situation.D EMcHenryJr, Limited Choices:The Political Struggle for Socialism in Tanzania ,Boulder,CO: LynneRienner, 1994. Those strugglesseem tobe cultural struggles, with the individualists eventually prevailing over egalitarians, as egalitarianpolicies failed to achieve theirdesired result. 75 Kimambo,`The interior before 1800’ , pp19±29. 76 AWildavsky,`Choosing preferences byconstructing institutions: a culturaltheory of preference formation’, AmericanPolitical Science Review ,81(1), 1987, pp 3± 22. 77 Shoulda governmentwork with indigenous values and institutions because theyare culturallyappropriate, ortry to change them because theyare exploitative,unjust or inef® cient? Thisdilemma is central to developmenttheory and policy. 78 RFattonJr, `The political ideology of Julius Nyerere: thestructural limitations of ª African Socialismº’ , Studiesin Comparative International Development ,20(2), 1985, pp 3± 24. 79 Thompson et al., CulturalTheory , p 34. 80 Thompson et al.,CulturalTheory, ch 4. 81 Spalding,`The state± society debate in Africa’ ;Clark& Wildavsky,`Why communism collapses’ . 108