Understanding And Engaging Local Knowledge And Practice: Practical Approaches To Natural Resources Research And Development

.

I

|

\ iC / ^ fI zLi nr LA! !h C , /y t

I

I

L

,

a

l'1

-

' 1a

+

UNDERSTAI{DING and ENGAGING LOCAL KNOWLEDGE and PRACTICE Practical Approachesto Natural ResourcesResearchand Development

illustratedby projectexperiences in Zanzibar

by

Martin T. WalshandSharonP. Harvey

1997

NaturalResources Institute Universityof Greenwich

CONTENTS

Preface

Chapterl:

INTRODUCTION

Chapter2:

REDEF'INING LOCAL KNOWLEDGE ANI) PRACTICE

Chapter3:

UNDERSTANDINGAND ENGAGING LOCAL KNOWLEDGE AND PRACTICE

Chapter4:

T]NDERSTANDING AND ENGAGING PROJECT PRACTICE

Chapter5:

CONCLUSION

Bibliography

x

x

BOXES ACCOMPAI\TYINGTHE TEXT

Box l. I

T,anzibr,r: historical background

Box 2.I

The cuftigensand cultivarsof Zanzibar

Box2.2

Leopardsand witchcreft on Unguje

Box2.3

The declineofclove productionon Pemba

Box2.4

Agricultural devclopmentin casternPembr

Box 2.5

Agricultural innovationin northern Unguja

Box 3.1

The evolution of PRAs in ZCCFSP

Box 3.2

Ranking experiencein ZCCFSP

Box 3.3

On-farm trials in ZCCFSP

Box 3.4

Flrmer reselrch groups end networks in Zlnzibrl.

Box3.5

Understandingrnd engaginginnovationwith cese studies

Box 3.6

Engaging traditional agroforestry practicc for sust*inebleland use

Box 4.I

ZCCFSPas a local institution

Box4.2

Disciplinarity,interdisciplinarityand participation within ZCCFSP

PREFACE

This book, as its title suggests,is aboutunderstandngand engagSng local knowledge and practicein the context of natural resourcesresearchand development. It has its immediate origins in our recently-completedwork for the ODA-funded and NRl-managedZanzibarCashCropsFarmingSystemsProject(ZCCFSP). In particular it grew out of our joint preparationof a workshop on the subjectof 'Methodologies for ParticipatoryResearchand Extension'for the staff of the Ministry of Agriculture, Livestock andNatural Resources,our host institutionin Zanzibar. In writing the book we have rangedmore widely than we did in our workshop presentations,examining local knowledge and practice in their wider context rather than focusing upon 'participatorymethods'per se. Our approachto local knowledgeand practicereflects our diferent and overlapping backgrounds. One ofus (N[fW) is a social anthropologistwith an active interestin ethnobiologyand communitywildlife management;while the other (SPII) has wide experienceof agroforestryand farmerparticipatoryresearc[ with a specialinterestin the socialaspectsof local ecologicalknowledgein community-based natural resource management.We do not claim to haveanythingradicallynew to say,only perhapsa different way of expressingit, As active 'developmentpractitioners' we have had preciouslittle time in which to write this book, but havehad to work arounda variety of other commitmentswhile basedoverseas.To makelife more difficult, neitherofus had accessto ow own collectionsof the relevantliterature (MTW's library being in storage,and SPH's having been conzumedby fire, along rJeithher house and other worldly goods). If we succeedin enrichinganyone'sknowledgeor enliveningtheir practice,then it is duein no smallmeasureto the zupportandencouragement ofthe manyinstitutionsand individualswho havehelpedus. Our primaryinstitutionaldebt is to NRI, in particular to the Head and other staff of the Social SciencesDepartment,the staff of the Publishing and Publicity Group, and the staff of the library. We also received invaluableassistance and advicefrom researchers and other staff of the IIED Resource Centre in London; the School of DevelopmantStudiesin the University of East Anglia; and the Schoolof Agricultural and Forest Sciencesin the University of Wales at Bangor. For their direct andindirectcontributionsto this book, andmuchmore,we would also like to thank all of our past and presentcolleaguesin the Ministry of Agriculture, Livestock andNatural Resources(MALNR) in Zanzibar;especiallythose who worked for ZCCFSP,as well as other staff of the Commissionfor Researchand Extension and the Commissionfor Natural Resources,including membersof the Jozani-ChwakaBay ConservationProject (JCBCP). Last, but not least,specialthanks are due to our familiesand friendsin MombasaandZanzibarfor toleratingour absence from their lives while writing this book, and for performingmore than their ordinary shareof domesticlabour,child-careincluded.

In so far aswe fail to enrichanyone'sknowledgeor enliventheir practice,then noneof the aboveshouldbe blamed. The responsibilityis entirelyours.

CHAPTERONE

INTRODUCTION

"Reversingtraditionalattitudes to developmentresearchthereforemearuuniting researchand practice, understandingand action, researcherand researched,into a single unitary process. And this in tum impliesthat researchers must acceptto be changedby the resultsof their researc[ mustbe accountable to the subjectsof their researc[ and must be preparedto seethe valueof their work judged according to its relevancein improvingthe livesofthe peopleconcemed.This doesnol meanthat all researchthat is relevantalso hasto be 'directly participatory'. Researchwhich analysessimilaritiesand differences over time and spacecan be extremely'relevant', but the usefulnessof such'secondary'researchwill be a function of its effectivenessin changingattitudes among the powerirl in a direction which will ultimatelyenablethelesspowerfulto thinkandact for themselves." (Edwards1994:281)

The abovequotationis takenfrom the author'sown paraphraseofan earlier pap€ron 'The irrelevanceof development studies'(1989). We havequotedit herebecauseit sumsup our own perspective, and toucheson someof the principalthemesof this book. Like Edwards,we areboth 'developmentpractitioners',and duringour work 'in the field' over a numberof yearswe have often wonderedjust how relevant 'developmentstudies'are to our everydaypractice. Unlike manyofour colleagues, however,we havefoundtime to readsomeof the recentdevelopment literature,and in the chapterswhich follow havetried to combinesomeof its best argumentswith our owrr insightsgainedfrom field experience. We hope that the result will be 'useful' in 'secondaxy' the way thatEdwardssuggests researchcanbe, andthat it will helpto changethe attitudesofour fellow practitioners to the ultimatebenefitofthose they practiseupon(but shouldpractisewith). Our main concernis how to understand andengagepeople'sknowledgeandpractice in a developmentcontext,and more specificallyin the contextof naturalresources researchanddevelopment.Understanding andengagingarecriticalactivitiesin these contexts;they specifuthe development practitioners'task much more preciselythan the increasinglynebulousexhortationto 'participate' does. We have not used 'participation' as the binding thread of this book, but have asked instead 'participationwith whom?', and 'participationhow?'. Unless the participatory imperativeis definedin this way, it eitherlosesits meaningor remainsa one-sided recommendation for oneparticularsetof interactionsin the development arena.This is not to saythat the emphasisuponparticipationhasitself becomeirrelevant,or is without value for developmentpractice. It has undoubtedly served an important purposein changingthe attitudesof at leastsomepractitionerstowardsthe people who meantto be the primarybeneficiaries oftheir projects.It is now time, however, to building on the (theoretical)gainsof the pastquartercenturyand recognizethat significant(practical)deficitsstill remain.

A brief outlile

The structureof the book can be summarisedas follows. In chaptertwo we will examine what local knowledge and practice are and why they are important. Conventionalunderstandingsof knowledgeand practice are strongly influenced by the 'indigenousknowledgedebate'and the way in which it has developed. This debatecan be characterized in terms of the (incomplete)shift from a 'naive' to a 'sophisticated'conception indigenous of knowledge'(IK). The naiveconceptionof lK is overwhelminglytechnical,while the sophisticated conceptionemphasizes its (political, social economic,etc.) dimensions.The latter is an improvementon the former, but it still focusesnarrowly on 'knowledge'and has led to a number of impracticalproposalsfor practrce. It is more satisfactoryto treat knowledgeas just one aspectof practice, and to put more effort into understanding differentmanifestations of local practice. Different ways of doing this are suggestedby recent social developmenttheorists, and we examinethe relativeusefulness of these,with particularreferenceto 'structuralist' and'actor-oriented'approaches. Havingestablished the needto analysehow practice changesovertime,we thentum our attentionto the meaningof 'local'. Conventional approaches to IK (andpractice)focusalmostexclusivelyon one set of actorsin the developmentarena,membersof the (usuallyrural) 'community' By contrast,we expandthe definitionof local to includethe knowledgeand practiceof otheragents, including govenrmentand non-govemmentorganisations(NGOs), and even pro.iects themselves. Chapter three looks at how local knowledge and practice can be engaged at community level. Different methodsand approachesare examined,starting with the 'participatory' packages standard and including ParticipatoryRural Appraisal (PRA). Theseare found wantingin a numberof respects, especiallywhenthey are used(as they usually are) in an inflexible and uncreativeway. They often provide no more than a superficial understandingof local knowledge and practice, and an insufficientlydevelopedmeansof engagingthem. The secondpart of the chapter considersaltemativeapproaches which addressthesedeficiencies.The needto use different methods and approachesstrategically, and against a background of a coherentunderstanding of localpractice,is emphasized. In chapterfour, the questionof engagingprojectpracticeis considered.We discuss what we call the 'extemal interventiondebate',returningagain to some earlier theoreticalthemesand arguingthat interventionhasto be constructivelyengagedand not naively rejected. Different proposalsfor incorporating local knowledge into project practice are then considered,focusing on local knowledge as it is conventionallyunderstood.In the secondhalf of the chapter,projectsare examined as local institutionsand the implicationsof this approachare exploredin terms of differentinstitutionaloptionsandthe waysin whichthe skills of projectactorsmight be improvedto enablethem to understandand engagethe practiceof other actors(as well as their own) more effectively. Someof the wder implicationsfor policies relatingto naturalresourcesresearchand developmentare alsooutlined. The final chaptersummarizesour argument.

Illustrative material Someof the main points in the text are illustratedwith boxesusing casematerial from our working experiencein Zanzibar,in particularwith the ZanzlbarCashCrops FarmingSystems Project(ZCCFSP)andthe JozaniChwakaBay Conservation Project (JCBCP). It shouldbe emphasized that theseare not intendedto illustrateall the points of the argument,althoughsomeboxesdo relateto points madeat different placesin the text, in somecasesin differentchapters(to this extentthe locationof someboxesin the text is fairly arbitrary).It shouldalsobe notedthat theseboxesare not intendedto providea comprehensive reviewof all tlre activitiesof the projects concerned, andthat we havebeendeliberatelyselective.

l.l Zanzibar: historicalbackground Zanzibarcomprisestwo islands(anda numberof smallerislets)off the coastof East Africa. The largestisland,Unguja,has a surfacearea of about 1600 km2: The capital,Zanzibartown, is locatedon Unguja. The smallerisland,Pemba,hasan area ofabout 1014kmr. It lies some50 km to the northof Ungujaandis separated from it by a deepseachannel. Thetopographies ofUnguja andPembaaresuperhciallysimilar,thoughtherearealso importantdifferences.Both islandscanbe roughlydividedalonga north-southa.xis; the westemside of each island is more elevatedand has deepersoils, while the easternlandsare flatterand litteredwith outcropsof limestone,generallyreferredto as 'coral rag'. The coralragof Ungujais extensive,dominatingthe southernhalf, the eastand the far north of the island. The west of Pembais hillier and more deeply dissected, andthecoralragis largelyrestrictedto its eastemfringe. Thetropicalclimatemeansthat it is hot andhumidfor mostof the year. The westem areashavea high rainfall (over2000mm/yearin someinlandlocations),which falls in two mainseasons:masika(abouMarch-May)andvali (aboutOctober-December). The coral rag is somewhatdrier, and the vzrli rains in theseareasare notoriously unreliable. The naturalvegetationof the islands(of which little survives)reflects their topographyand micro-climate;moist forestonce dominatedthe westemhills andvalleys,anddry forestdominatedthecoralrag. The islandswere first settledby Swahili-speaking fishers,farmersand livestockkeepersaroundthe middle of the first millenniumAD. Different Swahili groups continuedto migate to Unguja and Pemba in the centurieswhich followed. Meanwhile,the annualmonsoonwindsbroughttradersfrom acrossthe IndianOcean, especiallyfrom the PersianGulf andthe southemArabianPeninsula.Someof these tradersandtheir followerssettledpermanently, inter-marryingwith the local Swahili ion and adontinstheir The middleof the millenniumsaw

the beginningsofa protractedstrugglefor politicaldominationoverthe coastaltowns betweenOmani Arabs and the newly arrived Portugese. The latter were finally ousted at the end of the 17th century and henceforth Zarzibar was ruled by an independentline of Omani Arab sultans(who had earlier broken away from their cousinsin Muscat). The economy of the sultanaterevolved around ffade in slaves,ivory, and natural productsobtainedfrom the mainland. h Zavlbu itself (and the parts of the coastal strip which were under Zara;ibari control), Omani Arab landownersdeveloped a systemof plantationagriculturebasedon the labourof importedslaves. In the 19th century,most of the forestsof westernUnguja and Pembawere clearedto make way for theseplantations(theseare still referredto asthe 'plantation areas',in contrastto the lessproductivelandsof the coral rag). A largenumberof exoticcultigenswere introducedduringthis period,especiallyfruits andspices.The mostimportantof the spiceswascloves,first introducedin the early 19thcentury.Realizingtheir potential, in 1834SultanSeyyidSaid orderedhis fellow Arab landownersto plant cloves;in doing this, he unwittingly laid the foundation for an era of agricultural prosperity whichwasto lastfor the next 150years. Increasing European involvement in the affairs of Zatvibar culminated in the establishment of a British Protectoratein 1890. The British governedthe islands throughsuccessive sultanswho therebylost much of their independence as well as their mainlandpossessions.The new colonial authoritiesabolishedslaveryand ensuredthat cloveproductionwasmaintainedby substitutingslaveswith paid labour. In subsequentdecades,they intervenedon more than one occasionto counterthreats to the clove economy,guaranteeing the laboursupplyand preventingthe transferof plantationsto Indian moneylenders. They also made a number of (largely unsuccessful)attempts to foster agricultural diversification. However, cloves continued to reign supremethrough to, and beyond, the granting of Zanzlbar's independence in 1963. In January1964,the govemmentof the Sultanwas overthrownin a bloody coup, known as the ZatuibarRevolution,organizedaroundoppositionto continuingArab dominance. In April 1964, the leader of the new RevolutionaryGovemment, PresidentKarume, agreedto the unification of Zanzibarwith mainland Tanganyika, creatingthe United Republicof Tanzania. Karume,while remainingPresidentof Zanzibar,alsobecameVice-Presidentof Tanzania,with Nyerereas its first President. Zanzibu retained its own elected Council of Representativesand various non-strategicministries, including those responsiblefor agriculture and the environmentof the islandsin general.Karumeand his successors followedsocialist policies for the next two decades,an era characterizedby excessivestateintervention in the economyandmanyotherwalksof life. The mid-1980susheredin a gradualshift towardseconomicliberalization. This changein policy wasalsocarriedacrossto mainlandTanzaniawhereits impactwas subsequently evenmore evidentimpact. The 1990shavebeencharacterized by a combinationof both economicand political cri a drastic decline in

world clovepricesandthe introductionof multi-partypolitics. The negativeimpacts of botl of thesedevelopments havebeenfelt mostseverelyon Pembawheremostof the clovesare produced,and where the vast majority of the populationvoted for the oppositionCUF (Civic United Front) party dwing the first national multi-party electionsin 1995. Many observerswereunhappywith the way theseelectionswere conducted,as well as with the ruling CCM (Chama cha Mapinduzi, Revolutionary Party)govemment'ssubsequent teatment of Pembans.In early 1996,mostwestem donorsfroze their aid to Zarulbar, and a numberof bilateral projectsendedabruptly without any promise of renewed funding. So far, however, the Revolutionary Govemmentof Zanzibarhas weatheredthis storm, while intemational trade through the port, and the growing popularity of Zaraibar as a tourist destination, have continuedto stimulateeconomicsrowthin Zanzibartown.

CHAPTERTWO

REDEFTNING LOCAL KNOWLEDGEAND PRACTICE

Introduction Whatarelocalknowledgeandpractice?More to thepoint,whatdo we meanby local knowledgeand practice? The simple answeris 'what people say and do in a particularcontext',whetherthis contextis specifiedlocationally,institutionally,or otherwise. We do not necessarilyequate 'local' with 'rural', 'p€asant', or 'indigenous', asmanywritersdo. In our usageit rnayreferto anygoup of actorsin a specifieddevelopment arena,includingthe staffof projects,aid agencies, andoutside comm€ntators(such as ourselves). The phrase 'stakeholders'knowledge and practice'is perhapstechnicallymore accurate.However,we havepreferred'local knowledgeand practice', largely becauseit is less cumbersomeand reflectsour starting-point in the 'indigenousknowledgedebale'. We have reformulated'knowledge'as 'what people say' for two main reasonsFirstly, we are dissatisfiedwith the way in which the term (or rather construct) 'knowledge'has beenusedin recentdevelopment discourse.In this context,the word 'knowledge'drawsattentionaway from systematicconsideration of people'sideas andopinionsby nanowlyfocusingon oneparticularaspectofthem. The complexity of actors' mental and verbal actions is thereby reducedto a single construct ('knowledge'),packagedin a form whichis convenientlycompatiblewith the notions andprejudicesof thepackagers themselves. Secondly,it is evidentthat our knowledgeof other people'sknowledgeand other aspectsoftheir mentalworld is primarilybasedon interpretation ofwhat theysayand do. 'Kaowledge'is not a physicalentitywhich canbe handledor observeddirectly, althoughit maybe temptingto think of it this way. Eliciting indigenousor any other kind of knowledgeis a complexenterprise,and a combinationof strategiesmay be used to reconstructthe knowledgeof any given individual or group_ Direct questioningis one strategyand observationof relevantpracticesis another. This process is mediated by language,and is particularly evident in studies of ethnotaxonomy(indigenousclassification)when researchersdistinguishbetween explicit andimplicit categories, signallingthe fact that the laftercan only be infened from informants'statements andbehaviour. What peoplesay and what they do are not necessarilycorrelated. Our own approach is to give priority to practiceand treat people'sverbalpronouncements (including thoseelementswhich we useto reconstruct their 'knowledge')asjust one aspectof their practice. We believethat expressionof ideasis an integralpart of practice, while their reformulationas 'knowledge'may reflectthe observers'own practiceas much as that of their subjects. We recognize,however,that the understanding of theseissuesby manydevelopment practitioners'hasbeeninfluencedby the growing

demandthat they shouldtake accountof indigenousknowledge'when formulating their own practice. The title we have given to this book acknowledges this fact, although our text suggeststhe need for adopting a broader and more flexible approach.

Starting with knowledge T'heimportanceof indigenous knowledge' Over the pastdecadeor so, the conceptof indigenousknowledge'has becomean integral part of developmentorthodoxy. In the introduction to a recent review ('Moving the IndigenousKnowledgeDebateForward?'),Thompsonprovidedthe followingsuccinctaccountof its riseto prominence: "D*ailed study of indigenous knowledge(IK) date3backto the late 1970swhentwo seminalcollected works drew togetherresearchthat examinedthe capacities,skills and rationaleof peasantfarmersand pastoralists. Thefirst,a specialissueofthe 1DS8ll//etn (HowesandChainbers, 1979),hadwidespread influencedespitea limitedcirculation.The secondcollection@rokensha et al., 1980),published a year later,provedto be evenmoreofa landmarktext. Thereafter, a numberof otherinfluentialbooks(e.g. Cemea,1986;Richards,1985;Chambers, l9E3) arguedforcefullyfor the involvementof local people and the incorporation of their knowledge into processesof technologicaldevelopment. From the mid-1980sresearchinto IK expandedrapidly. SeveralintemationalIK resourcecentresare currentlyin operation,with yariousintemationally co-ordinated research programmes, andfollowingAGENDA2l a globalnetworkof indigenous people'sorganisations hasbeenestablished to promotepoliciesprotecting indigenous property rights to genetic resourcesand supporting the conservation of indigenous knowledgeofbiodiversitywithinthecontextin whichit hasdeveloped." (1996:105)

The need to record, understandand incorporate'indigenousknowledge' into developmentpracticein someconstructiveway is well establishedin professional circles. As the blurb accompanying on€ recentcollectionof essaysputs it: "The mainconclusions from this impressiveanay of expertiseis that local peopledo know a greatdealabouttheir environment, in which they haveoftenlived for generations; andthis knowledgemustbe takeninto accountin the planningandimplementation of development, if this is to be acceptable to the people,and effective"(Wanen et al. 1995:back cover). Accordingto the proponentsof indigenousknowledge',its importancefor developmentis no longeran issue(note herethe assumptionthat it once was). The challenge,rather,is to translatethe recognitionof this fact into practice. 'Naive'snd 'sophisticated' conceptiowof indigenous krutwledge' This is easiersaidthandone,not leastbecause'indigenousknowledge'hastumedout to be a much more complexaffair than its originalproponentsthought. Reviewing the (relativelyshort)historyof the notionof indigenousknowledge',it is possibleto distinguishbetweentwo differentconceptions, one which might be characterised as 'naive' and the other 'sophisticated'. as This is no doubt an over-simplification; differentcommentators draw the dividing line betweenthe naiveconception(s) they criticise and the sophisticated one(s)they advancein different ways, giving them

differentlabels(for example,compareScoones andThompson,1994,with Blaikie er al., 1996). Despitetheir differences,however,the theoristsagreethat a paradig'rn shift of somekind hasoccurred(or shouldoccurif their prescriptions are followed). We concurwith this view, and in the following sectionsprovideour own perspective on the 'naive' and 'sophisticated'conceptions of indigenousknowledge',indicating wherewe think the debateis leadingto (or shouldlead to if orr prescriptionsour followed). Indigenousknowledgeas technicalkttowledge The 'naive' conceptionof indigenousknowledge'(IK) stressesits technicaland instrumentalaspects. This is capturedperfectly in the once favoured formulation 'indigenoustechnicalknowledge'(ITK), which is still usedby manypractitionersin the field. A similar orientationis implied in pluaseslike 'indigenousagricultural knowledge'(IAK) and'indigenousecologicalknowledge'(IEK), which specig'the particulartechnicalfield underinvestigation. Theseexpressionsreveal the basic assumptionunderlyingthe naiveconceptionof IK, that it can,and shouldbe, treated as the mirror image(or seriesof local images)of formal scientificknowledge,the knowledgewhich researchers themselves havegainedthroughtheir training in, and applicationof, the universalprocedures of scientificenquiry. Thosepossessing IK areassumed to havearrivedat (at leastsome)valid conclusions aboutthe environmentin whichtheylive, andhowto manageit, tkough a cumulative processof trial anderror. This processis conceivedasbeingparallelin somewaysto that ofscientific experiment, andthoughgenerallylesssystematic, is believedto have producedresultswhich arenot only comparable with, but in manycasessuperiorto, those reachedby scientists,especiallyin contextswhere the latter have limited experience.With this scenario- or somemore cautiousversionof it - in mind, the practitionersis "first to empowerlocal peoplevis-i-vis research taskof development scientistsand developmentplannerswho, it is assumed,will be convincedof indigenous'wisdom' only throughscience;and second,to 'blend' IK with formal Westernsciencein theresearch process"(Thompson,I 996: 106). anddevelopment This,the originalconceptionof IK, evolvedin the wakeof the increasingadvocation of 'participatory'approaches to development.If sustainabledevelopmentwas to dependon people'sparticipationin the development process,it madesenseto take a 'bottom-up' look at their own ideas and practicesand act upon these. Some commentators havethereforelabelledthis the 'populist'or 'neo-populist'approachto IK (Blaikieet e ., 1996;Thompson,1996). It is the approachwhich dominatesmany of the contributionsto Farmer Frrsr (Chamberset al., 1989),including the IDS Workshoppaperon 'Farmer'sKnowledge,Innovations,and Relationto Science'. Blaikie er. al. (1996)suggestthat the contributionsto BeyondFalmer Fir.st(Scoones and Thompson,1994),the first pafi of which is devotedto 'TheoreticalReflections on Knowledge,Power and Practice',belongto the same'neo-populist'paradigm. According to our own characterization (caricature?)of naive and sophisticated conceptionsof IK, they do not, although we would a$ee that the influence of the naiveapproachcontinuesto loomlargein certainrespects.

It is no accidentthatthe naiveconceptionof IK evolved,andhashadmostinfluence, in the contextof agriculturaldevelopment.In this field, the confiastbetweenformal scientific researchand indigenouspractice has alwaysbeen sharperthan most. An importantrole has long been assignedto the expertiseof agronomistsand other specialists,since well before the Green Revolutionand beyond into the era of FarmingSystemsResearch(FSR). Given their backgound and training it is not surprising that enlightenedagricultural experts should have constructedIK in the imageof their own profession,once it becameobviousthat they could no longer ignoreit. Ironically,the sameimpetuscan still be seenin the standardmethodologyof Farmer ParticipatoryResearch(FPR),despitethe claim that this approachhas been freed from the top-downorientationof its predecessors.Farmersare typically encouraged to developtheir own on-farmexperiments,completewith control plots and evaluation procedures, oftenwith the objectof testingintroducedgermplasm.To a largeextent farmers are viewed and treated as proto-researchers, working in the image of their professionalcollaborators. This is undoubtedlypreferableto ignoring them completelyin the processof researchand development;but unfortunatelyit can sometimesalso provide an excusefor ignoringtheir practicesoutside,and in the absence of, projectinterventions. In some fields, the technicalapproachto IK has been developedto extremesof scientificprecision. Ethnobotanicalresearchprojectsoften tend in this direction, especiallywhenthey havea universalscientificgoal (for examplethe discoveryof globallymarketablemedicines,or the preservation of areaswith a high biodiversity value)andonly a subsidiaryinterestin the concemsof local people.The work of the EcologicalKnowledgeResearch Groupin the Universityof Walesat Bangor,andthe, inputs of this group and othersto the ODA ForestryResearchProgrammeProject, provide examplesof highly technical approachesin the field of agroforestry (Southern,1994;Thap,, 1994;Walker, 1994;Kendonet al., 1995;Sinclairer a/., 1995;Thapaet al.,1995: WalkerandSinclair,1995;Walkeret al., 1995). Studiesof this kind have the virtue of indicatinghow complexIK can be, even when it is definedin narrowtechnicaltermsandthe socialandeconomiccontextsin which it is producedand reproducedare largelyexcludedfrom consideration. Whenresearchon IK is undertaken by naturalscientists,it is perhapsinevitablethat certain aspectsare emphasizedat the expenseof others. Eliciting IK, even in specifiedtechnicalfields,is not the simpleactivitythatit is sometimes assumed to be, but requiresa rangeof skills. Theseincludethe skills possessed (in theory)by social scientists,as well as a degreeof linguisticcompetence (at leastsomeknowledgeof the languagein which the IK is expressed).This last requirementis all too often neglected,with unfornrnateresults. For example,botanists and others frequently recordthe vemacularnamesof plants(and evenspeculateon their meanings)with no more than a crude understandingof the language(s)concemed. Their orthographies are often inadequate,especiallyin the caseof unwritten languages,and it may even be unclearwhich languages or dialectsare being recorded.This is a problemwith

manyethnobiological databases andnotjust ethnobotanical collections(seeBox 2.1). The simplesolutionis to employand/orseekthe adviceof peoplewho possess the relevantskills,includingnativespeakers (ideallypeople andcompetentprofessionals who areboth).

2.1 The cultigensand cultivars of Zanzibar An incrediblevarietyof cropsis grownin Zanzibar,morethanin mostotherpartsof Africa. The great diversityof cultigensreflectsthe geographicalposition of the islandsand the many and varied historical influencesto which they have been exposedover the past2000years. ZCCFSPresearchers begantheir examinationof the viability andpotentialof differentincome-eaming cropsin Zanzibarwith a list of 42 'candidatecashcrops',mostof which werefruits and spices.Thesecompriseda merefractionof the known cultigenson the islands,which includea varietyof root andgraincropstraditionallyclassifiedby theMinistryof Agricultureas 'food crops'. A furtherlevel of complexityis addedby the fact that farmersrecognizeand namea largenumberof differentvarietiesof sornecrops,especiallythosewhich havebeen the subjectof multipleintroductionsand/orwhichhavebeenimportantin the islands' economyfor manyyears(bananas andrice aregoodexamples).AlthoughZanzibaris relativelysmalland mostfarmersare nativespeakers of a singlelanguage(Swahili), thereis also considerable local variationin the namesof recognizedcrop varieties. This reflectsthe islands'complexculturalandlinguistichistory,which hasfostereda high degreeof terminologicalheterogeneity, even within the main dialects(more precisely,dialectcontinua)which linguistsrecognise.As a result,the Swahilinames of crop varietiesmay differ not only from one areato another,but in someinstances from villageto village,andevenfrom onelocal speakerto another. Although Zanzibar has hosted a number of donor-fundedagricultural projects in recent yea$, relatively few attemptshave been made to record the local namesof cropvarieties,let alonedescribeandidentifythem,assess their statusascultivars,and examinetheir useby farmers.In trueGreenRevolutionstyle,researchers havetended to focusonly on thecropswhichinterestthem,includingthe importedcultivarswhich theybelieve(not alwayswith justification)to be superiorto local varieties.ZCCFSP startedto recordinformationon farmers'knowledgeof differentcrop varieties,and beganto investigatesomeof them (mostnotablymangocultivars)in greaterdetail. However, this work was never completed,partly becausesome researcherswere scepticalof its valueandthoughtit wouldconsumetoo muchtime. A full inventoryof Zanzibar'scultigensandcultivarsandtheir Swahilinameshasyet to be compiled. The best written sourcesremaina coionial handbook(Williams, 1949) and the informationscatteredin unpublishedreports,some of which was compiled,albeit uncritically,by Koenders(1992a)in his suwey of agricultureon Pemba. In manyrespects,Zarzibar'sfarmersare much better informedaboutthe thanthe researchers andextensionists who to advisethem.

The compilationof such a list (or lists) shouldnot be seenas an end in itself ("farmers'knowledgeof cropsandcrop varieties"),but as a meansof understanding farmers'practiceandengagingwith it moreprofitably. The diversityof cultigensand (presumed)cultivars on Zanzlbw sayssomethingimportant about farmers' past and present strategies(including diversificationas a strategyto maximize limited resources andminimizerisk).andtheseshouldnot be ienored.

Practitionersoperatingwithin the boundsof the naive conceptionof IK can be justifiably accusedoffailing to carry their own programmethrough,and of havingtoo narrowa view of their own practice,the scientificmethodologywhich they are using to elicit and evaluateIK. This is reflectedin the use of the term 'knowledge'(as enshrinedin the phrase'indigenousknowledge'and all of its derivatives),as well as by the commoncall for the needto 'validate'this 'knowledge'. Anyoneschooledin Popperianprescriptions for scientificmethodwill immediatelyrecogaisetheseusages as being symptomaticof an outmoded'inductivist' or dogmatically'positivist' epistemology.A follower of Popper'sopen-ended approachto scientific enquiry might reasonablywonder why the object of our enterpriseis called 'indigenous knowledge'.Would it not be moreaccurateto describeit as 'indigenoustheory',or betterstill ' theeries'? Sucha revisionofour conceptionof IK (nowIT or ITsl) wouldcarrywith it a number of advantages.It would serveas a constantreminderthat neitherour theoriesnor those espousedby others are necessarilyprivileged, and that even those which are favouredmust alwaysbe treatedas provisional,and never accordedthe certainty of 'knowledge'. It would encourageus to be more careful in assessing the relation between,and relativevalue of diflerent theories. Over-enthusiasticproponentsof IK sometimesfind it embarrassing to admit that IK may be wrong,an embarrassment which stemsfrom the claim that recognitionof the validity of IK is what separates it from earlier approaches.However,if IK is recastas IT, then we shouldbe well preparedto find that some ITs are less adequatethan others, including our own theories. The challengeis to explainhow and why sometheoriesare superiorto others,whatevertheir source. It is widely believedin sub-Saharan Africa that physicalcontactwith chameleonsand their salivacausespoisoningin humansandlivestock. Zoologistshave(as far as we are aware)found no evidenceto supportthis proposition,and haveno hesitationin describingit as mistaken,consigningit to the categoryof 'belief. However,a good IT researcher wouldwantto know how thetheoryof chameleon toxicity arose,how it spread,andwhy it remainsso prevalent.Only thencouldthe practicalconsequences of this IT (or setofITs) be addressed, assumingthattherewasa perceivedneedto do so (for example,if a particular group of camel herderswas avoiding good browsing becauseof the presenceof chameleons, or killing a rare speciesin order to avert sicknessin theirherds).

This semi-hypothetical exampleshowsthat it is not sufficientsimplyto considerone IT as 'wrong' and anotheras 'right', but that it is importantto explorethe relation betweenthem and the consequences of this for practice. When it matters,the incommensurability of theorieshas to be translatedinto the commensurabilityof practice(for a non-hypothetical exampleof this see Box 2.2). Sometimes,the problemmay be lessacute,for examplewhen an IT can be readily incorporatedinto a moregeneraltheorywhichexplainsthesameeventsandmorebesides.In eithercase, however,the real work beginswhen the practical consequences of theory (whether 'ours' or 'theirs')are addressed.

2.2 Leopardsand witchcraft on Unguja The ZanzibarLeopardis unique. It lives solelyon the islandof Unguja,whereit is the largestcamivore and only wild felid. It has been separatedfrom its mainland relativesfor morethan 10000years,andsomeauthoritiesconsiderit to be a distinct Pantherapardus adersi. As settlementand agriculturehaveexpandedon subspecies, Unguja, the ZanzibarLeopardhas sufferedincreasinglyfrom habitat destructionand predationby humans.lt is not knownhow manyleopardsremainon the island,but recentresearchsuggeststhat the populationmay havereacheda critically low level. Most rural inhabitants believethat althoughsomeleopardsarewild, othershavebeen groups bred and fed by of witches (wachawi)who usethem to harassand intimidate their fellow villagers. This belief is elaboratedin manywaysand includesdetailsof the witches'practice(for example,their useof charmsto sendthe leopardsfrom one joint-owner to anotherin a different village) which are recountedwith considerable consistency from one end of the island to the other. These accounts of leopard-keeping are so convincingthat most people who hear them, including educated townspeople, acceptthemwithoutquestion-A numberof outsideobservers, amongthem wildlife researchers, havelikewisebeenpersuaded that theremustbe a core of truth in thesereports,and more than one has beentaken on a 'wild goose chase'to seea keptleopard. Thebelief in leopard-keeping andfearofthe witcheswho keepthemhasa very direct impact upon people'sattitudesto leopardsand their conservation.Many villagers would simply like to get rid of them, and throughoutrecordedhistory local hrmters have trapped,spearedand shot leopards,with addedhelp from their own magical charms,in order to reducethe threat. Shortly after the 1964ZarulbarRevolution this culminatedin an island-widecampaign,encouragedby the govemment,to eradicate leopardsand neutralizetheir allegedkeepersonce and for all- By the mid-1970s, more than 100 leopardshad been killed in this campaign,which was led by a witchfinder known as Mzee Kitanzi. The killing has continuedto the presentday, with the now illegaltradein leopardskinsprovidingan additionalincentiveto local hunters. Leopard numbers have declined to such an extent that some villagers, particular the younger generation, are now prepared to consider some way of ine them. The commonest is to keeo them in a zoo or a more

extensiveenclosureand chargetourists to see them. Many Forestry officials (includingthoseresponsible for wildlife conservation) agreewith this approach,and suggestthatthe simplestcoursewouldbe to persuade leopard-keepers to displaytheir kept leopards. At first sight,the widespreadtalesof leopard-keeping are completelyat odds with scientific knowledge. However, closer examinationshows that this belief is constructedaroundan attemptto explain the very real problem of conflict between human and leopard populations. Villagers designateleopards as wild or kept accordingto their behaviour. A leopardwhich is seendeepin the bushand flees from the observeris generallyassumedto be wild, but a leopardwhich doesnot run away, approachesfarmlandand humanhabitation,and/orwhich posesa threatto the lives of peopleandtheir livestock(therearemanyrecordsofpredation,includingfatal attacks on children),is assumedto be kept and acting on the ordersof its owne(s). The belief in leopard-keeping alsodrawson commonideasaboutthe natue and role of witches in local society,and functionsas a multi-purposetheory which, amongother things,explainswhatmightotherwisebe thoughtofas normalbehaviourfor leopards living in closeproximityto humans. Equipped with this knowledge (a more general theory which encompassesand explainsthe 'indigenous'theory),it is easierto approachthe problemof conserving the ZanzibarLeopard. It is evidentthat local concernsaboutthe 'antisocial' activities of leopardswill haveto be addressed if their long-termsurvivalis to be assured.The Jozani-ChwakaBay ConservationProject(which is fundedby CARE Austria) and the ForestrySectionof the Commissionfor NaturalResources arecurrentlyworkingon a seriesofproposalsdesignedto balancethe interestsof differentstakeholders involved in this issue,includingthoseofthe nationalandlocal hunterswho haveactuallybeen doing the killing. Whetheror not the divergentviews of the different 'experts' can be reconciled,remainsto be seen;it is more important,perhaps,that some form of agreement is reachedoverthe coursesofaction whichcanandshouldbe taken.

Indigenousknowledgeas indigenouspractice Sincethe beginningofthe decade,the naiveconceptionof IK hascomeunderattack (Bebbington,1991;Fairhead,1991;Long and Long, 1992;Scoonesand Thompson, 1994;Thompsonand Scoones,1994;Thompson,1996). Thesecritics all makethe point that 'knowledge', including IK, is always socially (culturally, politically, economicallyetc.)grounded,andthat its articulationin development practicecannot be separatedfrom questionsof power and the relationsbetweenthe different actors andagenciesinvolved.ThompsonandScoones state: "The attempt

to "blend" or "incorporate" local knowledgeinto eisting Westem scientificprocedures assumesthat rural people's knowledge(RPK) representsan easily definable"body'' or "stock" of knowledgereadyfor extractionand incorporation. The critics point out, however,that rural people's knowledge,like Westernscientificknowledge,is alwaysfragmentary,partial, and provisionalin nature. It is neverfully unified or integratedin termsof an underlyingcultural logic or systemof classification.

Moreover, knowledgeis embeddedin and emergesout of a multidimensionaluniversein which diverse cultural, economic,environmental,aad sociopoliticalfactors intersectand influenceone another. The proc€sstskes place on the basisof existingconceptualframeworksand processesand is atrectedby various social contingencies,such as the capacities,experiences,interests,resources,and pettems of social interaction characteristicof the particular social group or groups of individuals. Finally, knowledge,whether"indigenous"or "scientific", is inclusivein the sensethat it is the result of a great manydecisionsandselectiveassimilationsofprevious beliefs,values,ideas,andimages,but at the same time exclusiveof other possibleframesof conceptualiz:tionand understanding. Hence, it is not an accumulationof "facts" but involyesways of comprehendingthe world: knowledgeis always in the making."(1994:59)

The naive conceptionof lK, with its over-emphasison the technical and its over-simplisticview of the easewith which IK can be appropriated,lacks this deeper sociologicalperspective.Accordingto its advocates, the sophisticated conceptionof IK makesup for this failing, and requiresdevelopmentpractitionersto deal with the issuesit raisesheadon. The essaysin BeyondFarme,,First (ScoonesandThompson, 1994)providesomeindicationof how theymight approachthis tash especiallythose workingin agriculturaldevelopment.The basicmessage (at leastof the editors'own contribution)is that theyshouldplungethemselves into the complexitiesof practice, reflectingcritically upon their own as well as others'; acting not just as smiling facilitators but also as canny catalysts and crafty negotiators in the fields and battlefields of knowledgebefore them. However, Beyond Farmer Firut promises morethanit deliversandsuffersfrom someofthe sameblind spotsas its predecessor (Chamberset al., 1989). Farringtondrawsattentionto someof these,the first of which is particularlyrelevanthere: "[One shortcoming]is

an inadequateconceptualis8tion of the interactionsamongknowledge,the forms (i.e. technologies) in which it might be applied,and the processes (i.e. innovation)of applyingit. Knowledge, on which the book focuses,is only one componsnlof innovation: others include the complementaryinputs - whetherland,labour or capital- necessaryto put new ideasinto practice. The vast literature documentinglocallevel conflict over physicalaccessto resourcesis ignorcd- If it hod beentakeninto account,therewould havebeenlessoptimismaboutthe willingnessof"the community" to wotk togetherin prioritisingneeds,about the possibilitiesof arriving at an agreedagenda,about the apparentpermanence of "unanimous"decisionsoncethe team of outsidershasdeparted,and about the capacityof participatory methodsto resolveconflict."(1995:5)

The conceptionof IK espousedin BeyondFarmer L'irst and other recent writings does provide a more sophisticatedperspectiveon the same subject, (indigenous) knowledge. Although it promotesa view of knowledgegroundedin practice,it preservesthe naive conception's prioritization of knowledge over practice, and thereforecontinuesto pay little attentionto those aspectsof practice cited by Fanington(and others). A more recentreview of different approachesto IK therefore combinesthe naive and sophisticated conceptionsunder the label 'neo-populist', describingthe latter as merely a more radical versionof the former and equally deficientin its proposalsfor practice(Blail
by Fanington.Instead,sociologicalconcemsappearto havebeensimplytackedon to the originaltechnicalapproach.Althoughthe resulthassomevalue,it falls shortof the obviousconclusionthatindigenous knowledgeis an aspectof indigenotspractice, practiceis muchmorethanthat. On this vien the theoristsof IK andthat indigenous havetravelled an extremelyroundaboutroute to get halfway to a goal which should havebeentheir startingpoint. Theoristsof the sophisticatedconceptionof IK claim an impressiveintellectual pedigree.The list of authoritiescitedby Scoones and Thompson(1994)readslike a Llho's Who of social theoristsand philosophersof sciencespanningthe past few decades.The commonthemewhich connectsthe work of thesedifferentauthorities, including Kuhn, Feyerabend, Habermas,Bourdieu,Foucaultand Denida, is their rejection (explicit or otherwise)of 'positivism', and in particular of scientific epistemologyas an adequateaccountof how knowledgeis constructedand scienceis really carriedout. As Blaikie er al. (1996)suggest,the altemativeswhich they and the sophisticatedtheoristsof IK sketch out lead inexorablyin the direction of epistemological andculturalrelativism(this is mostexplicit in Feyerabend's lgairsl Method, and should be apparentto anyonervho has followed the developmentof post-structuralisttheory). This is somewhatunsatisfactorywhen taken to its inevitableconclusionand translatedinto proposalsfor practice. Some of these proposals(Blaikie et al. refer to those of the 'extreme' FPR lobby) are clearly impractical,and are in dangerof leadingto an attitude of 'anl.thinggoes', and ultimatelyto theabandonment ofpractice. We all recognizethat knowledgeis socially constructedin the sensethat it is producedandreproduced in particularsocialcontextswhichmay shapeit in different ways. However, not all of us are preparedto accept the relativists' implicit conclusionthat everyone'sknowledgeis equalin the sensethat no one person'sor group'sknowledgeis moreextensive,accurateor usefulthananother's(we avoid use of the term'valid' herebecauseit canhavedifferentmeaningsin this context). This doesnot meanthatwe candecidein advance whoseknowledgeis 'better',or discount anyone'sknowledgeasbeinginelevant. In a development context(anddevelopment parlance)the knowledgeof all the stakeholders is relevant,and the challengeis to searchsharedunderstandings which arenot ownedby any singleparty. While examiningthe naive conceptionof IK, we intimatedthat a more productive approachto the whole questionof IK might be to adopt a Popperianperspective,by treatingindigenousknowledgeas indigenoustheoryor theories(ITs). This approach providesa readyset of procedures for comparingand evaluatingtheories,regardless of who holds them, and encorragesus to theorize the theories, in other words to attemptto explainwhy differenttheoriesare held. This includestheorizationof the socialdimensionsof theoryand all thoseaspectsof ITs to which the sophisticated conceptionof IK draws attention. Epistemologicalrelativism can be jettisoned without losingthe particularinsightsgenerated by its adherents.Popper'sscientific epistemology is essentiallyaprcscriptionfor practiceratherthana descriptionof that practice,a point which seemsto havebeenmissedby the majorityof theoristswho sided.withKuhn in the wakeof their famousdebate.Readerswho wish to pursuethis l0

line of reasoningand its implicationsfor the IK debateare refened to Popper's disputewith the FrankfurtSchool(Habermasincluded)and relatedwritings. For presentpurposes, however,we havesaidenoughabouttheory,andwill retumnow to our centraltheme:practice.

The primacy of practice An anthropologicalperspective A number of contributorsto the IK debatehave acknowledgedthat before it began researchon IK was carried out primarily by social (cultural) antkopologists (Fairhead,1991;Thompson,1996). Most issuesraisedin the debatecan be tracedin the anthropologicalliterature,where many of them are examinedin considerably more detail and from a much wider varietyof angles. Anthropologistshave long grappledwith countlessfacetsof the sociologyof knowledgeamongthe peoplethey havestudied.However,muchoftheir work remainsunreadby contemporary students oflK, especiallythosetrainedin otherdisciplines.It is arguablethat ifthey hadtaken accountofthe existingliterature,thenthe IK debatewouldnot havedevelopedin the way that it has. It is alsounlikelythatIK wouldhaveremainedthe privilegedconceptthat it has. Not only would the expression'indigenousknowledge'havebeenseverelycriticized(it seemsto have been introducedas a 'soft' translationof what some American anthropologists were alreadycalling 'ethnoscience'),but also the notion that it represents.One of the anthropologist'sprimary interestsis, and always has been,to explainthe relationbetweenwhat peoplesay and what they do, regardlessof how thesestatements might be categorized('knowledge','belief, 'myth', or just plain 'exegesis').As might be expected,socialrelationsfigureprominentlyin accountsof the relationbetweeninformants'statements (presumed'thought')and their actions, although different schoolsof anthropologyinterpret their roles differently. Many anthropologists would agreethat what peoplesay(think, know, believeetc.) cannot be explainedwithout referenceto what they and othersaboul them do. Knowledge consistsof more thanjust statements or theoriesaboutthe world, with or without people. The productionandreproductionof knowledgeis somethingthat peopledo, andin this senseis an aspectoftheir practice. A practical problem The IK debatehas diverted attention away from the needto understandand engage people'spracticeand the theoristshaveonly recentlybegunto recogtrisethis need. What impactshasthis had upon developmentpractice? The majority of practitioners are now awareof the declaled importanceof IK, althougha significant proportionof them havenot progressedbeyondthe naive conceptionof IK as technical knowledge. Relativelyfew field workersknow how to elicit or use IK creativelyand many continue to treat it as somethingwhich threatensto make their work unnecessarily complicated.Their understanding of, and ability to deal with, people'spracticeis ll

evenmoreundeveloped.This is not surprising,becausethe literatue placesso little emphasison the importanceoflocal practice,andhardlyany guidanceis providedon how to understandand engageit. The standardtoolkits for participatorydevelopment only promotea superficiallevel of understanding andengagement, focusingmore on the technicalaspectsof local practice(seechapterthree). Practitionersare left to follow their own instincts,and while someland on their feet, othersdo not. This problemis perhapsmost acutein projectswith explicitly technicalgoals,but not uncommor in thosewith an institutional focus. Understandinglocal practice What do we meanby local practice? The answerto this questionhas traditionally fallenwithin the domainof anthropologists, for whomlocal practiceis everythingthat peopledo and say. Understandinglocal practice meansunderstandingwhy they do what they do and why they saywhat they say. This is not how anthropologistswould normallydescribetheirjob, but this is basicallywhat it boils downto. In the process of performing this task, anthropologistshave frequently limited their frames of reference(for example,to a particular ethnic group, or a particular type of practice) and focusedon tle minutiaeof local practiceto the exclusionof its wider context. Many anthropologistsstill operatein this way, especiallythose who are more interestedin what p€oplesaythan what they do. A significantnumber,however,have enlarged and enriched their approachto local practice by borrowing from other disciplinesandby addressing the concernsof socialdevelopment researchin general. This is the approachwhich, we suggest,can and shouldenrich developmentpractice. At 'community' level (whetherthis refersto a particularlocation or a particular population,howevernarrowly or broadlydefined)this entailsdevelopingan overview of the politicaleconomyof the community,of the principalrelationswithin it andthe principalrelationsbetweenit andothercommunities(includingthe wider community, suchas the state,of which it is a part). In this context,'relations'includethoseof productionand exchange,power, gender,social organization,land tenure and access (this is merelyan indicativelist). At individualor householdlevel the to resources task is much the same,though it shouldbe pointedout that a lot of the researcher's time andresources canbe wastedby focusingon this level- It is muchmoreeffrcient to begin with the generaland work down to the specific, given that the former often explainsthe latter. In FarmingSystems (FSR),the mistakeis oftenmadeof Research extrapolatingfrom the specific to the general(the 'farming syst€m'conceptis in any casequestionable asit is limited in bothdescriptiveandanalyticalscope).

2.3 The declineof cloveproductionon Pemba Until recently,Zanzibarwasfamousfor the productionofcloves. For morethan 100 years,its economywas heavily dependenton clove monocultureand the incomes derived by the state,traders,landownersand agricultural labourersl?om the clove The raoi ion of clove ion bv the world's main consumer t2

Indonesia, andthe consequent over-supply anddrasticfall in world marketpricesover the pastdecade,havehad a devastatingimpact on Zaruibar's clove industry and the nationaleconomyasa whole. Accordingto currentmarketpredictions,thereis little hopethatpriceswill recoverto anlthinglike their formerlevels. The futureprosp€cts for large-scaleclove productionin Zanzibarare thereforegrim, and while clove farming has not suffereda suddendeath,it is certainlyin a stateof accelerating decline. The vastmajorityof Zanzibar'sclovesare grown on Pemba. Clove farms,many of which were originally establishedas plantationsand share-plantations in the 19th century,dominatethe hilly landscapeof the westemside of the island, covering approximately two+hirdsof the land. Farmersin this areahavebeenmosthardhit by the fall in prices. Althoughfarmerselsewhereon the islandhavealso suffered,in generalit has been easierfor them to adjust to the change. There are a variety of reasonsfor this,themostobviousbeingthatclovesneverplayedsucha dominantrole in their livelihoods;relativelyfew own clove farms,and seasonal clove-pickingwas only one of a rangeof incomesourceswhich householdsexploitedto supplement their subsistence-oriented agriculture. In the west, however,the vast majority of farmersown clove treesand until recentlyrelied heavilyon their incomesfrom the cloveharvestto buy foodandmeetotherdomesticneeds. As cloveshavebeenso dominantin the local economy,farmershavefound it very difficult to adjustto the recentdemise. Most clovetreesremainon the farms. It is illegalto fell productivetreeswithoutofficial sanction,andthe formal procedures for this areslow andcumbersome. However,manyfarmersblamethe government, rather than the world market,for the low price of harvestedclovesand believethat prices will recoverif the governmentor its policieschange.As a result,thereis cunently little inclinationamongfarmersto engagein wholesalefelling and replacementof clovetrees. Instead,manyclove farmsareneglected, andthe landbetweenthe trees is revertingto secondaryforest. This providesan ideal habitatfor vervetmonkeys, whosedepredations on other food and cash crops provide further disincentiveto expandcultivation. The inherited patternsof land-holdingand labour use are further obstaclesto agriculturaldevelopment.Largeacreages of clovesare ownedby peoplewho live outsidethe localcommunity(someof themoutsidePemba),andthis limits the useof the land for other crops, especially'permanent'tree crops. Labour inputs to agricultureremaingearedto the economicsystemunderwhich they evolved. The role ofwomen is mainlyconfinedto subsistence rice cultivationandthe processing of rice and otherfood crops. Althoughmen'sinputsaremore extensive,in mostcases they could not be describedas intensive,especiallywhen comparedwith farmers elsewhere(includingwomenoutsideof the cloveplantationarea). This reflectstheir historicalrelianceon a treecrop requiringminimalmaintenance, muchof which was undertakenby labourersfrom outsidethe clove-growingareawho were paid from the proceedsof the harvest. The transitionfrom a low-inputto a higher-inputsystemis not easy,particularlyasthe financialresources to employoutsidelabourareno longer so readilvavailable. 13

The typicalresponse ofclove farmen hasbeento expandsubsistence productionin a piecemealfashionby plantingcassava plots. Many ofthese andbananasin scattered plotshavebeenopenedup in patcheswherecloveshavedied from SuddenDeath,a diseasewhich is known to attackpoorly-managedplantationsmore readily than those which are well cared for and which app€arsto be tkiving in proportion to the increasingneglect of the clove farms. For farmers,cassavaand bananashave the obviousadvantageof being familiar food crops which can help to plug the gap previouslyfilled by food purchases;any surplusescan be sold. There is also a growing market for bananas,which are being traded in increasing quantities to UngujaandespeciallyZanzibartown. Farmershavebeenslow to expandproduction of other altemativecashcrops,partly becauseof their lack of confidencein the markets,andpartly becauseof the constraintsmentionedabove. The cloveareasof Pembaareprobablythe mostpressingproblemin the agriculrural economy of Zartzibar. This problem is compoundedby the environmental risks inherentin the developmentof alternativesto cloveswithout carefulplanningand appropriateadvice from policy-makers. Throughouthistory, the westernside of the islandhasbeenperceivedas Zaruibar'slargestareaof naturallyfertile land, full of potential for agricultural development. Howeveq althoughthe potential is there (as demonstrated by morethan 100yearsof successful cloveproduction),the soilsin this areaare no more fertile thanelsewhere.In fact,the subsoilon the hills of Pembais pooranddeficientin minerals;the deepsoil pocketsin somepartsofthe coralragare muchmorefertile. The apparentfertility of thecloveplantationsis dueto a thin layer of organic mafter depositedand maintainedover thousandsof years by the forest vegetationwhich grew on it. The natual balancewas fragile, and becomeevenmore precariouswhenmostof Pemba'sindigenousforestswerefelled in the l9th andearly 20th centuries.By happyaccident,theywerereplacedby a treecrop which playeda similar, if lesseffective,biologicalrole in maintainingsoil fertility. Ironically,the periodic and cunent neglect of the clove plantationshas further helpedto maintain the balance. Withoutsomekind of treecoveror othermeasures to maintainsoil fertility, the green hills of Pembawill give way to a banen and relatively unproductivelandscape, similar to that which hasalreadydevelopedin otherdeforestedpartsof the tropics. Short-termeconomicgainshaveto be measured againstsuchmedium-and long-term consequences.There is a dangerthat theseconsequences will be ignoredor not recognizeduntil it is too late. This hasalreadyhappenedin other areasof Pemba; whereroot cropshavebeencontinuouslycultivatedfor morethan two decades,soil fertility and productivityhavedeclinedmarkedly. Given the recentchangesin the economicfortunesof the cloveplantations, this issueneedsto be tackledurgentlyto avertthe risks of unplanneddevelopmentand work towardsa sustainableagriculture.

justificationfor this approachis sketchedout in Booth'soverviewand The theoretical summing-upof RethinkingSocial Development(1994), He arguesthat social l4

developmentresearchers shoulddevelopanalysesin the 'middle ground' between top-downtheory and bofiom-upempiricism,both to accountfor the patternsof divenity which have scuttled grandtheory and to bridge the gap betweenacademic developmentstudiesand the work of practitionersin the field. To some extent, researchershave begun to do this, and the contributionsto RethinkingSocial Developmentillustratedifferentapproaches to this problem. Booth's discussionof 'neo-structuralist the relationbetweenthe political-economy' and 'actor-oriented'(or 'agency-oriented')approaches advancedby different contributorsis particularly pertinent- Although he concedesthat eachapproachhassomethingto ofler the other, he ultimately choosesto subscribeto the former. One way out of this 'minor impasse'is to think of themashavingdifferentbut complementary applications.The neo-structuralistapproachis particularly relevant to the understandingof practice, whereas an actor-orientedapproachhas a much more obvious rcle in engaging practice. The reason for this is deceptivelysimple: understandingrequires generalization,whereasengagemententails interaction between people and/or the institutionsto whichtheybelong.This distinction,however,shouldnot be carriedtoo far; understandingand engagingpractice are not just complementaryactivities, but canandshouldbe dialecticallycombinedby development practitioners.

2.4 Agricultural developmentin easternPembr For more than a centurythe agricultureof Pembahas been dominatedby clove production(seeBox 2.3). The lowlandsandcoralrag beyondthe westemplantation zone have frequently been characterized as marginal in terms of both their contribution to the clove economy and their presumedcomparativelack of agriculturalpotential. However,recenthistorysuggests that this characterization is misleading.Most of the significantagriculturaldevelopments in Pembaoverthe last 25 yearshavetakenplacein the lowlandsandthis trend continuesto the presentday. Subsistenceproductionexpandedin responseto period shortagesof basic foodstuffs, particularlyduring the 19'10-72famine,causedmainly by the government'sabrupt cessationof food imports. At the sametime, and in particularareas,specificcrops were developedas cash crops, and marketedboth within and outside Pemba. The best exampleswere sweetpotato in Makangalein the far north-west,and turmeric in Mwambe in the south-eastof the island. The recent sharpdecline in the clove economy has had a further, and undoubtedly more widespread, impact upon production for the market. Many farm householdsin the lowlands used to derive additionalincomefrom clove-picking,but the fall in clove prices,and thereforeof labourrates,hasreducedthe seasonalmigrationof labourersto a trickle. Instead, many farmerson the eastof the islandhavebegunto supplementtheir incomesby growingvegetables and other short-termcropson the coral rag. The only villages largely unaffectedby this trend are thosein the Micheweni peninsulaand on someof the smallerislandsaroundPemba,wherefishingandlong-distance tradeoffer a more securelivelihood.

l5

The villagesin theseso-calledmarginalareashavethe most traditionalsocial and economicorganization.Theywereleastaffectedby the large-scale transformation of the islandwhich accompanied the introductionofthe plantationeconomyin the l9th century,and thereforeconform more closely to the historical pattem of rural Swahili communityorganizationfound all along the East African coast. To this extent,they representthe outcomeof a long processof adaptationwhich is evident in their exploitationof the diverseresources availableto them,on land as well as in the sea. Agriculturehasa longhistoryin thesecommunities.Shiftingcultivationon the coral rag,especiallyof sorghumandmillets,hasbeenpractisedfor morethana 1000years, and the modified vegetationof Pemba'scoral rag, whereonly two dry forestsof any sizeremain(RasKiuyu andMsitu Mkuu on theMichewenipeninsula),is t}e resultof this practice. Rice growingin the shallowvalleys,and cultivationof other crops, includingbananasand root crops,on the higherland betweenthese,also hasa long history. The history of economicand agricultwaldiversificationin thesecommunitieshas madeit easierfor them to weatherrecenteconomicchangesand to adaptto new circumstancesand opportunities. However,thesechangeshavenot affectedthem all equally. Populationpressureand land availabilityhaveprobablyhad mosteffect on their differentialresponses, as shownby the historiesof cashcrop developmentin Makangale and Muwambe. These factors also help to explain why agricultural in the centralpart ofthe eastcoasthaslaggedbehindand only recently development begunto show signsof rapid change,propelledby the declinein clove incomes. Much of this development is focusedon Vitongoji and Ole, wherecomparativeease of accessto Pemba's urban markets has provided an additional stimulus for increasingcashcropproduction,especiallyofvegetables(includinggroundnuts).The impactof this development now app€arsto be spreadingnorth,following the line of the new tarmacroad. Althoughthe paceof changeis slowerin Kangaganiandothervillagesoffthis road,it is already beginning to be felt, particularly on the coral rag, where vegetable production (tomatoes in Kangagani) is developing, and efforts by the Forestry Departmentto introducetree planting aretaking off The tree planting is particularly encouragingbecauseuncontrolledexploitationof the coral rag and the fringing mangroveforestsposesa threat to the resourcebase;in someareaslarge standsof mangrovehave alreadybeen clearedto make way for the manufactureof salt. At present, coral rag land is comparatively abundant and fallows are generally maintained,thoughnot for long enoughfor naturalforestto regenerate.The coral rag is generallyperceivedto be commonland, opento exploitationby anyonefrom either within or outsidethe local community. In this respect,there are fewer controls than there were in the past, when village elders and the party branchesexercizedsome authorityoverits use. As populationpressureincreases, Pemba'scoralrag is likely to be subjectedto the kindsof intensiveexploitationwhich are alreadytaking placeon Unguja,wherelandownershipis becomingan issueandresearchindicatesthat mixed farming with livestock and tree crops provides the best prospectsfor a sustainable agriculture.

l6

Genderissuesare an importantcomponentof this situation. Whereaswomen havea relativelylimited role in agriculturein the plantationareas,they providethe bulk of the labour in the lowlands of Pembaand produce nearly all the food crops with minimal assistancefrom men. The coral rag, however,is traditionally a male preserve,and it is men who havebegunto cultivate cashcropson this land, although women'sgroupshavealsobeeninvolvedin the recentwave of tree planting. It is likely that furtherexpansionof agricultureon the coral rag will increasinglyimpact on genderrelations,and that this will not necessarily be to women'sadvantagein termsof their accessto resourcesand the proceedsfrom them. Planners,researchers and extensionists will have to considerthis issuecarefully,both in planningtheir work with farmersand in providing appropriateadvice. Gendersensitivity shouldbe an essentialcomponentin the developmentof a participatoryapproachto agricultural developmentthroughoutZawibar, but especiallyin areaslike the lowlandsof Pemba wherewomenalreadyplaya leadingrole in farming.

Theimporlanceof change An essentialcomponentof any analysisof this kind shouldbe an understanding of how practice has changedover time, how it is changingat present,and in which direction it seemsto be heading. This might seem to be an obvious point developmentis, after all, about managingchange- but it is one which is frequently neglected. One reasonfor this neglect,perhaps,is that developmentpractitioners tend to be more concemedwith the changesthat they and other stakeholderswould like to seetake placethan thosewhich are alreadytaking placeor havetaken place in the past. The participatorymethodsand tools recommendedto practitionersdo little to overcome this inherent bias. Consider, for example, the standard practice for conductingParticipatoryRural Appraisal(PRA). The overallemphasisof the PRA processis placed on the identificationof existingproblemsand possible/zlzre solutions. One of the many problemswith this approachis the generallack of att€ntion paid Io existing solutiors to past problems. The 'landscape' of problems and solutionsin any one communityor region is usuallyuneven,what remainsa problem to one group of actorsmay alreadyhavebeensolved,or is in the processof beingsolved,by anothergroupin the sameor anotherlocation. Thesesolutionsor sofutions-in-the-making may constituteopportunitiesfor those ll,ho have not yet arrivedat them, including thosewho havenot yet recognisedthe problemswhich they engage. Such opportunitiesare often missedin the processof conductingPRAs, especiallywhenthey are undertakenin a singlecommunitywithout referenceto what is happeningin others. Apart from this generalfailing, the recommendedtools and techniquesfor PRAs are inadequatefor addressingquestionsof change, especially in the hands of inexperienced researchers.While local historiesand time lines "can be extremely importantin highlighting someof the causesof certainproblemsor how changeshave t7

occurred"(Nabasaet al., 1995:27),theirelicitationoften produceslittle morethana chronologyof importantevents. Other tools are explicitly orientatedtowardsthe collection of syrchronic information. Wealth ranking, for example,providesa snapshot of perceiveddifferentiationin the economicstatusof households at the time the exerciseis undertaken.The causesof inter-household di{ferentiationare not usuallyconsidered beyondstatements ofthe obvious(suchas 'households headedby widows are in the poorestcategory'). As a result,field workersoften assumethat householdmembership in the categories they haveelicitedis moreor lesspermanent. As anthropologists havelong beenat painsto point out, this is not necessarilythe case. Households, and the largerdomesticgroupsto which they belong,invariably movethrougha'developmentalcycle'during which their economicstatuschanges. A householdassignedto one particular 'wealth category' may well fall into a different category some years later, and householdsat different stages of the developmentalcycle may appearto belong in the samecategorywhile they are actuallymovingin verydifferentdirections. The operationof factorssuchas thesecan createhavocwith unsophisticated wealth rankingsandbringinto questionany furtheranalysisandactionsbaseduponthem. In communitieswherepermanent pattemsof socialandeconomicdifferentiationarenot immediatelyapparent,oneofthe first questionsany researcher shouldask is whether they haveemerged,or to what extentthey are in the processof emerging. This is not a purely academicquestion;it is of critical importancefor developmentpractice, especiallywhen distributionaland equitability issuesare being considered. An understanding of patterns of differentiation at different levels (inter- and -group, -communityetc.) is crucial for making decisionsabout intra-household, targetinginterventions(or perhaps,'interactions')and monitoringand evaluating lmpacts. Whenexaminingchange,it is perhapshelpful to distinguishbetweenchangeat the macro-andmicrolevels. This is anotherareawhereboththe 'structure'and 'agency' approaches can be usedin a complementary way Changeat the macroJevelcan perhapsbe best understoodin terms of the 'neo-structualistpolitical-economic' approach(Booth, 1994),whereaschangeat the microJevel falls more obviously within the domainof the 'actor-oriented'approach(Long and van der Ploeg 1994). Whereasa structwal approachis the simplestroute to understandinggeneralpatterns of change,a focusuponagencycan help to elucidatethe diversityof individualand institutionalactionswhich make,and aremadeby, thesewider pattems. Innovation is one such aspectof changeat the micro-levelwhich is of particularinterestto developmentpractitioners. Innovatorscan be viewed as individual agentswho pioneerchanges in structure- structures whicharenot,however,oftheir own making, but of otheragentswhoseearlieractionshavedefinedthe pattemsof constraintand opportunityon which they act.

2.5 Agricultural innovationin northern Unguja t8

Agricultural developmenthas proceededat a much faster pace on Unguja than on Pemba,partly becauseof the proximity of Zanzibartown and its growing market for agriculturalproduce.The growthof eggplantfarmingin Gambaisjust one example among many on Unguja; others include orangefarming in Ndijani (central Unguja) andmangoproductionandexportin Muyuni(southwest Unguja). Gambalies at the northeastemfringe ofUnguja's plantationzone,borderingthe coral rag which stretches awayfrom the villageto the eastcoast. Beforethe 1980s,it was more or less marginalto the economyof the island; as in many other villages, productionof bananas,cassava,rice and other grain crops was much subsistence more importantthan the oppornrnisticsaleof theseand varioustree crops in the local markets. Somecashcrop developmenthad takenplace,focusingon turmeric and gingergrownin homegardens, but by the 1990sthe marketfor theseproductswasnot very good. Until recently, the clove economyprovided many inhabitantsof Gamba with their most reliable sourceof income. As a result of historical pattemsof migration and intermarriagebetweenthe people of northem Unguja and southem Pemba,someof the inhabitants of Gambaheldrightsin cloveplantationson the latter island, and many more sailed over to Pembaevery year to pick cloves or take advantageof other economic opportunitieswhich the clove harvest offered. The declineof the cloveeconomyfrom the mid-1980sonwardsplacedadditionalpressure on the inhabitants of Gambato developaltemativesourcesof incomeGamba'stimely responsew:rslinked to the growth of the Zanzibarurbanmarket;this, in tun! wasa functionofthe increasedtrade(andto a lesserextent,tourism)which followedeconomicliberalizationin the mid-1980s.In lessthan a decade,many of Gamba'sfarmershadreorientatedtowardsthe growing urbanmarketand transformed a significantsub-setof their local farmingsystemsby movingawayfrom subsistence productionand into vegetableproduction. The most spectacularaspectof this transformation wasthe conversionofthe fertile valleyto the eastof the village from cassavacultivation to eggplant(aubergine)production. Developmentof this kind may havebeeninevitableoncetransportlinks with Zalibar town hadimproved,but the precise form of the developmentowed much to the foresight of a single innovative farmer. He had previously worked in the town market and seen an opportunity for supplyingit with eggplantswhen they were out of seasonin Umbuji, the villagewhich hadpreviouslysourcedZanzibar'sresidents.For severalse.rsons he labouredalone,but when neighbouringfarmerssaw the good profits he was mafting, theyswitcheden zzrsseto eggplantproductionin 1993. Since then, eggplant production has continued to expand in Gamba and the surroundingvillages- The processof expansionhas been greatly assistedby the evolutionof a collectivemarketingsystem,flrst institutedin 1994. This system parallels,and in somewaysis more successful than,the systemoperatedby orange farmersin Ndijani in cenhalUnguja. It is entirelythe productof local innovationand collaboration,and is designedto ensurethat the market is never flooded and that prices therefore remain more or less constant through most of the three-month harvestingperiod. In the first year of its operation,the two villages involved took tums to markettheir andwereset for the maximumquanti 19

which they could taketo marketon any singleday. In the 1995season,the system had expandedto coverfive villages;eachtook a daily tum to markettheir produce without any limit on the amountof producethey could send. Farmersanticipatethat the systemwill be furthermodifiedin 1996,andareevenconsideringits adaptationto at leastone other vegetablecrop, tomatoes. So far, it has worked well, and farmers arepreventedfrom sellingout of tum by carefulpolicing. Gambais now Unguja's secondmost importanteggplantproducerafter Umbuji. Profits are not as high as they were for the village's first eggplantfaxmer,but good enoughto encouragecontinuingexpansion. Meanwhile,some farmershave also begun to experimentwith other vegetablecrops, such as spinach(Amaranthussp.) intercroppedwith the eggplants,and tomatoproductionhasbecomean increasingly importantfeatureofcultivation on the coral rag to the eastof the village and its fertile valley. The introduction of improved tomato cultivars for the urban market has playeda significantrole in this development.Many farmersnow combineeggplant cultivation, begunbeforethe long masikarains,with tomato cultivation, begunin the shortvrli rains,in an annualcycleof vegetableproduction.Althoughsomefarmers, mostlyof the older generation, haveresistedthis development and continueto glow 'traditional' cassavaand other crops(in somecasesderivingmost of their income from off-farm activities),the trend seemsto be set. Gambaprovidesan excellent exampleof the kind of diversificationtaking place in manydifferent partsof Unguja, especiallyin areaswherethere is land on which to expand(the coral rag uwanda and maweni)and/orno or few permanenttree cropsto preventthe conversionof the land to more productiveuse (especiallyin the rice and other valleys). The naturalend result of this diversificationwould appearto be a patchworkof complementarylocal cashcrop specializations; a patternof this kind is alreadyemerging(eggplantsin GambaandUmbuji,orangesin Ndijaniandmangoes in Muyuni). This patchworkof localized cashcrop production,much of it for the Zanzibar town market, is a far cry from tfte picture of future developmentoriginally envisagedby ZCCFSP,which beganby searchingfor cropsto substitutecloves. ln manyrespects, the farmersand tradersof Unguja have shownthemselvesto be a step aheadof the agriculturalexpertsby devisingtheir own solutionsto the declineof the plantation economyand exploringopportunitieswhich are more immediatelyrealizable. In retrospect,it makessenseto developthe local marketbeforeaiming for export, and in the caseof 'Boribo Muyuni' mangoesthis hasprovideda good springboardfor more ambitiousdevelopment.Outsideinterventions sometimes fall very wide of the mark. For example,govemmentsponsorship production rice of has persistentlyfailed to produceresultsover the past30 years,and farmershaveoften found more productive usesfor the valleysin which rice was oncegrown. On a muchsmallerscale,it can now be seenthat ZCCFSPwas wrong in promotingginger productionin Gambaat a time when the local markethad declinedand in a way which led somefarmersto believethat it hadnot. Until intensiveres€archwasundertaken in Gamba,ZCCFSP only had a vague idea of the scale of eggplantproduction and marketing and the overalltrendtowardsvegetablecultivation.This underlinesthe importanceof finding out what farmers have actually done, and are doing, before suggestingwhat they isht do in future. 20

The recordofZCCFSPandsimilarprojectsin assessing andpredictingthe impactsof agriculturaldevelopment hasbeenmixed. ZCCFSPwas instrumentalin highlighting need the to evolve sustainableagricultural practices,given the inherent fragrlity of Zanzibar'ssoils. Farmersdo not always perceivethe dangersof environmental degradationuntil the damagehas been done, and a strong ciue can be made for developingparticipatorystrategiesto tackle this issuein anticipationof future events. The increasingprivatizationof the coral rag near Gamba,and the replacementof shifting cultivation with permanent grain and vegetable production, parallels developmentswhich are taking place throughoutthe coral rag lands of Zawibar. It remainsto be seenwhetherthe 'Ndijani solution' (fruit tree productioncombined with intensivemanuring)or somethinglike it will evolveelsewherein the islands.In Gamba,new agoforcstry practiceson the coral rag are still in a tentative stageof development,but cattle-keepingand the potential for manuring appear to be in decline. The Gambacasealso highlightsthe needfor a more carefulassessment of gender issuesin cash crop development. Although turmeric and ginger productionhas providedwomenwith an additionalsourceof income,this hassometimes beenrather less than anticipated (where husbands have taken advantage of their culturally-specified monopolyon long-distance marketing).Male dominancein coral rag agriculturealsomeansthat cashcropproductionin this areais moreimmediately to their advantage.On the other hand,somewomenhave clearly benefitedfrom the developmentof eggplantproductionby securingtheir own plots in the valley near their homes. In somerespects, the expansionof cashcrop productionwould appear to increasewomen'sbargainingpowerbecause moreincomeis availablefor themto bargainover. Cashcropdiversificationat the local levelmightmakeit moredifficult for men to monopolizethe incomefrom thesecrops,especiallyif womenown the land on which they are grown and./orprovidemost of the labour. However,this is no more than conjecture and again underlines the need for further researchbefore interventionsdesignedto enhancewomen's accessto, and control of, resourcesare proposed.

Expandingthe defnition of local Sinceconcludingour discussionof the IK debatewe haveabandoned the qualifier 'indigenous'and substitutedit with 'local'. 'Indigenous'(and its derivatives)can havepatronizingconnotationsand evokeethnic particularism. 'Local' is rathermore generalandopen-ended than,for example,'rural', andis definitelylesspatronizing. So far, we havediscussedlocal practicein the conventionalsensein which'local' refersprincipallyto the inhabitantsof a particularlocale (a village, a region or a wider politico-geographical area). These'local people' are very often the primary practice,thepresumed'beneficiaries'of intervention,andthe targetsfor development possessorsof IK and all its (near-)synonyms ('ethnoscience','rural people's zl

knowledge','local knowledge','existinglocal knowledge','indigenousagricultural knowledge','indigenousecologicalknowledge'etc.). They are also the primary subjects of anthropology and social development research in its structuralist manifestations.However,the actor-oriented approachto socialdevelopment research and practiceencourages a wider conceptionof local practicewhich doesnot limit it to its traditionalor assumed definition. This is the conceDtion which we adoptin the followingdiscussion. Understandingand engaginglocal practice meansunderstandingand engagingthe practiceofall the actorsandagents('stakeholders') activein the development arena, whethertheyare 'locally'basedor not. This includesthe knowledgeand practiceof all the relevantgovemmentinstitutions,NGOsand aid agenciesinvolved in a project. It also includes the knowledge and practice of project sta{I, treating projects themselves aslocal institutions.Thepracticeofthesedifferentinstitutionsandagents is all too frequentlyoverlooked,under-emphasized, placedin a different category,or only consideredseriouslyat certainstagesof the project cycle (most often at the beginning,whenformulatingprojectproposals). Thereare different reasonsfor this neglect. Onehasbeena generallack of analytical (and practical)tools for examininginstitutionalpracticeat the wider level. When researchers haveexaminedsuchinstitutionsin greaterdetail they havetendedto do so from a critical rather than a constructiveperspective(asking, for example, why a particular project has failed, or analysingthe failure of governmentto addressthe needs of 'local' people). Developmentpractitionersare given few practical guidelineson how to understand local institutionalpractice,includingtheir own, or how to tacklethe issueswhich suchan understanding might raise. It is perhapsnot surprisingthat so many projectsare riven with intemal conJlict and working at odds with the institutionswith which their logical frameworksassumethem to be in harmony. The successand smoothrunningon a projectfrequentlydependson the 'experience'and 'commonsense'of the projectteam,and this is no substitutefor propertraining and guidancein undentandingand engaginglocal practicein all of its differentaspects. It is ironic that many of the insights of British social anthropolory (or at least the classical structural-functionalistversion of it) developedfrom a need to understand and engage the practice of local institutions as part of the wider project of colonialism.The decisionto implement'IndirectRule' produceda flurry of research on local institutionsand practiceby both administrators and govemment-employed anthropologists. Theresultsofthis researchhadmanylongJastingconsequences, and althoughit is no longerpoliticallycorrectto agreewith their goals,the methodsthey usedcontain both both positive and negativelessonsfor the present. Unfortunately, they did not beginto questiontheir own practice,or analysecriticallythe institutions which they had created and worked with, until they had lost their mandate to intervene. Participatorydevelopmenthas (we hope) very diflerent goals: this is no excuse,however,for us to approachour work with an equalmeasureof sophistication ard naivety.

22

Conclusion We are not suggestingthat every developmentpractitioner should be an anthropologist or socialdevelopmentspecialist,or that everyprojectshoulddevote itself to academic-style research.However,thereis a clearneedto developboth the methodsand skills which would enabledevelopment practitionersto understand and local knowledge practice, engage in our expandeddefinition,more effectively and than at present. Somepracticalapproaches to theseproblemswill be exploredin moredetailin the followingtwo chapters.

23

CHAPTERTHREE

UNDERSTANDING AND ENGAGINGLOCAL KNOWLEDGE AND PRACTICE

Introduction Over the pasttwo decadesdevelopmentplannersand practitionershavebegunto pay increasingattentionto what rural peoplesayand do, and in particularto what they say they want to do. Although developmentonce meant trying to 'modemize' such people,to someit now meansactingin partnership with themto solvetheir problems 'Participation' and fulfil their wishes. has becomethe key word of development rhetoricand,at leastin theory,of development practice.Devising(andin somecases, 'participatory' reviving) methodshasbecomea minor industrywith the elicitation of 'indigenousknowledge'(and the elaboration means of to mobiliseit) as one of its majorbranches. In this chapter,we will focuson some(but by no meansall) of the methodscurrently usedto understandand engagelocal lcrowledgeand practicein the context of natural resourcesresearchand development.Here, we will focus upon the conventional, narrowdefinitionof 'local', referringprimarilyto the knowledgeand practiceof the 'ordinary' people usually identified as the primary target goup or intended 'beneficiaries'of development intervention.The mostseriousweakness of the tools and methods available lies in the way that they are used, and the general 'participatory' approach(es)under which they are subsumed. In many ways 'participation'and 'participatory'have becomeempty slogans,and their repeated deploymenthasbecomea way ofavoidingthe firll complexitiesof understanding and engaginglocal knowledgeandpractice.

The Limitations of Participaiion The main problem with participatory approachesis not so much what they do, or encoumgepractitionersto do, but what they do nor do, or discouragepractitioners from doing. Participation,in the senseof interaction,is an essentialcomponentin any attempt to understandand engagethe knowledge and practice of others, but without further specificationit either lacks content,or containsonly that which the differentparticipatorymethodsprovide. If we ask 'Participationwith whom?',then we receivean answerwhich is both over- and underdeterminedat the sametime. In everydayapplication,participationis generallyrestrictedto one setof interactionsin the development arena: the interactionbetween'us', the development practitioners, and 'them', the rural peoplewith, and on whom, we practise. The agentsand actors on either sideof this equation,however,tend to be treatedashomogenousclasses.

Apart from the appllicatiionof stock formulaesuchas 'resource-poor households', 'female-headed households' etc.,relativelylittle attentionis paidto the importanceof socialandeconomicdifferentiationamongthe intended'beneficiaries'of projects,or among the various 'owners' of projectsthemselves,practitionersincluded. The failure to identiS the beneficiariesclearly is frequentlyassociatedwith a poor understanding of the socialand culturalcontextsin which they think and act. The social,political and economicdimensionsof practice,includingknowledge,tend to be weakly apprehended, if at all, by conventionalparticipatorymethodsand practice. 'sophisticated' The conceptionof IK is helpingto correctthis,but hasyet to 'trickle down' to most practitioners,many of whom continueto operatewith the 'naive' conceptionof IK astechnicalknowledge,in the contextof the equallynaivepopulism of standardparticipatoryapproaches. If we ask 'How shouldparticipationbe put into practice?',then we find an equally unsatisfactoryresponse. Natural resourcesprojectsusually fail to specify which 'mode' of participationtheyareto work in; the defaultis usuallythe mode(s)fostered by the particularmethods,participatoryor otherwise,which a projectemploys. The useof the term 'modes' in this contextderivesfrom Biggs' (1989)identificationof the following four modes of participation in agricultural research: contactual (researcherscontract with farmers to provide land or services),consultative (researchersconsult farmers about their problemsand then develop solutions), collaborative(tesearchers andfarmerscollaborateaspartnersin the researchprocess) and collegiate (researcherswork to strengthen farmers' informal research and developmentsystemsin rural areas). Thesefour modesare progressivelymore participatory,from contractualto collegiate. Blaikie et al (1996) argue that negotiationof knowledgebetweenlocal and extemalactorscan only really occur if local peopleare ableto participatein the processof researchand developmentin a collegial (collegiate)mode. In practice,Biggs' modesare usually muddled,and participatorymethodsdo little to sortthemout. The fact that Biggs' modesrefer primarily to agriculturalresearchrestrictsthe scope ofthe modelandreflectsthetechnicalorientationfrom which agriculturalresearchers are still strugglingto free themselves.This is the principalcontextin which the IK debateand participatorymethodshaveevolved,and it is possiblethat if the debate had focusedon altemativestrategiesof natural resource(for example,wildlife) management, a very differentemphasismight haveemerged.Managementusually involvespayinggreaterattentionto the interestsofall the stakeholders, notjust to one set of them; it also fostersa greaterconcemwith local institutionsand institutional arrangements. One of the difficulties of conducting participatory agricultural researchat communitylevel is that as there are often no (or very few) institutions involved over and above individual farmine households.researchersare often temptedto createthem. Natural resourcemanagement also remindsus that we neednot think of modesof participationin termsof a scalewhich slidesfrom goodto bador vice verso; instead we shouldperhapsthink of differentmodesasbeingappropriate in differentcontexts. Thereare somecontextsin which more formal t1,pesof researchare appropriate(and

may even be requestedby farmers). Similarly, situationsfrequently arise in community-basedresourcemanagementfor which the interventionand/or arbitration of a higherauthorityis required(for example,inter-villagedisputesover resources can rarely be resolved at the village level). The challengefor development practitionersis to know how differentmodesof participationcanbe deployedto the greatesteffect. In this respect,participationis merelyan unsatisfactorycoverterm for a complex sequenceof engagements,some running concurrently,others in succession, andall of them intendedto complementoneanotherin pursuitof project objectives. PMs : participation packaged For many practitioners, participatory research (and development) has become synonymouswith PRA. PRA is a family of approachesand methods with acknowledgedsourcesin activist participatoryresearch,agroecosystems research, applied anthropology,field researchon farming systems,and its immediate predecessor, Rapid Rural Appraisal(RRA). The focus has shiftedfrom extractive data collection to facilitating local people to produce and analysetheir own information(Chamben,1994a,1994b;ComwallandJewkes,1995). However,there hasbeenan increasingtendencyto standardizePRA methodsand packagethem into a singleframework. Although the results of PRA can be, and often are, a vast improvement on the outcomesof more traditional styles of investigationand engagement(including formal surveysandtopdown modesof researchand extension),they frequentlyfail to match expecations and create additional expectationswhich cannot be met. In individual cases,this kind of failure can be blamedon inadequateplanning,poor implementation,and insuffrcientfollow-up; which in tum can be blamed on inadequateunderstanding of the PttA processon the part of insufiicientlytrained participants and other poorly preparedagents(such as NGO decision-makersnot directly involved in undertakingPRAs). To a largeextent,however,the methodology itselfcan be blamedfor the failure, andthe fact that PRAsare badly conductedcan be seenasa functionof naiveover-optimism on the partoftheir proponents. The 'quick andeasy'natweof PRA,with its heavyrelianceon a'toolkit' of different methods,is at the root of many of its problems. The toolkit comprisesa random assortment(and in somerespects,a confusingassenblage)of techniques,eachof themdesignedto givea 'quick fix' on particularissues.It is assumed thattheir usein combinationwill somehowachievethe ambitiousgoals which PRA sets itself, providing information for practitionersandassistingcommuniqrparticipantsto define andsolvetheir problems.To someextentit does,but not in a very satisfactory way.

3.1 The evolutionof PRAsin ZCCFSP The main line of methodological development in ZCCFSPtook the followins couse:

(a) a seriesof PRRAswas conducted;(b) farmingsystemszoneswere classifiedon the basisof the information deriving from the PRRAs;and (c) farmer researchgroups (FRGs)werecreatedwithin zonesof particularinterest. At a later stage,a seriesof more focused'mini-PRAs'was carriedout with the FRGs. Other approaches were followed in parallel, including work with farmer networks,cashcrop and marketing casestudies,andthedevelopment ofstrategiesrelatingto these. Overthe courseof oneyear,a seriesof Participatory RapidRwal Appraisals(PRRAs) was carried out with support from all sections of the Ministry for Agriculture, (MALNR). The programmeincludedtraining, 10 Livestockand NaturalResources village PRRAs,and a one-weekworkshop. A multidisciplinaryteam (of varying composition)spentoneweekliving in eachof the l0 villages. The objectivesofthe programmewere multiple and involved the developmentof a farming systems approachwithin MALNR, analysisof researchand programmepriorities, and more specific objectivesrelated to ZCCFSP'sexport crop diversificationprogramme. Twenty-eighthouseholdfarmingsystemsrn Zuzibar were identifiedand described. This madea significantcontributionto the delineationof farmingsystemszoneson the two islands. However,the descriptionand the classificationof farmingsystems zonesdid not generatean adequateunderstandingof the complexity and diversity of livelihood strategies,nor was the classificationperceivedas relevant by many deparfrnentsin MALNR (which continued to organise its activities according to existingadministrative divisions). The PRRAs were usedto move towardsa more participatory approachin MALNR andprovidedtraining opportunitiesandan exposureto participatoryresearchmethods and tools. However,ttrey generatedlittle action in the communitiesinvolved (with the exceptionof thosewhich later becameFRG sites)and reflectedthe project's objectivesmore than thoseof the local participants-This was partly becausethe PRRAs'role in the researchprocesswasnot clear,resultingin their usefor defining cash crop constraintsand opportunitiesin the classic style of farming systems research(FSR). The use of the term ParticipatoryRapid Rural Appraisal ratherthan PRA was in itself a problem. The rapid and elicitativenatureof the researchstill predominated. The sametoolkit, which included a wide array of tools (semistructuredinterviews,time-lines,mapping,transects, calendarsand rankingactivities of various kinds) was used in each of the villages. The opportunity for further training was not adequatelyfollowed up, and the developmentof the use of these tools wasneverreally given suflicient attention. Lack of experiencein the useof PRA methodsmeantthat there was little analysisof socialandeconomicdifferencesat householdor village level and,despitethe useof time-linesandothermethods,the understanding ofthe development of agriculturein Zanzibarin its socialand historicalcontextremainedpoorly understood.A second seriesof PRAswascarriedout later in the project,partlyto gain furtherinformation on farmers'knowledgeand practice. Theseweretermed'mini-PRAs'and were,in effect, intensivestudiesof local farming systemsand livelihoods. Focusingon membersof the FRGs rather than whole villages, the mini-PRAs were actionientated:the researchthemeswhich from collective were taken

backto the FRGs,modifiedwith their suggestions, and convertedinto joint research plans. The tool-setusedwas more basicand encouraged more effectiveuseof the tools and ownershipover the process;only semi-structured interviewsand farm observationswere employed. The research process was also more integrated, involvingall the membersofthe researchteamin all the stagesfrom initial checklist brainstormingto collectiveanalysisof the results. Greateremphasiswas placedon tracing changesin the farming system,identi$'ing existing opportunitiesrather than constraints,and understandingfarmers' knowledgeand practicein their local context. (The generalaccountsof agriculturaldevelopment presentedin Boxes2.3,2.4, and 2.5 havebeenadaptedfrom the reportsof thee of thesemini-PRAs).

A significantamountof informationcan be collectedduring a PRA, althoughit generallytakes the form of series of superficial snapshotsof particular aspectsof community life rather than a comprehensiveunderstandingof structureand agency and how these have changedover time. PRAs are undoubtedlymost effective for eliciting strictly synchronictechnicaldata (for example,which crops are grown where,andwhy). Theyareleasteffectivefor answeringpolitical-economic questions and for dealing with social and institutional complexity, including issues of distribution and equitability. When such matters are addressedeffectively, it is usuallybecausepractitionerswith the relevantskills and experiencehavetaken part, andnot a resultof the applicationof the PRA toolsper se. Althoughsomeof these tools build on the experienceof anthropologistsand other social development researchers, they do not equip others(natural scientistsincluded)with the skilts necessary to carryout suchresearch themselves.Theymerelycreatean impressionof understandingandprovidepractitionerswith an excusefor not delving deeper. The ranking exercisesrecommendedin the standardPRA packageare particularly at fault in this respect.The weaknesses of wealthrankinghavealreadybeenmentioned. Other,and in somewayssimpler,methodshavebeendevisedfor rankingproblems and preferences.Although thesecan be revealing,the conversationswhich take place while the exercisesare in progressoften yield more informationthan the rankings themselves.The resultsof suchexercisesshouldneverbe takenat facevalue. As Nabasae/ al. (1995)makeclear,rankingcan be influencedby a variety of factors, including the socio-economicstatusof individual respondents.Without probing further, practitioners may fail to recognize the influence of these factors on participants'choices,or conflatewhat peoplesay theydo with what theyaetually do. Although ranking may provide a window on people'sknowtedgeand practice, other methodshaveto be employedto completethe picture.

3.2 Rankingexperiencein ZCCFSP Severaldifferentrankingexercises werecarriedout by ZCCFSP,

developmentof a candidatecash crops matrix; and (b) a method to quantify indigenous technicalknowledge(QuIK). The first methodwasnot aimedspecifically at incorporating localknowledgein its design,but the secondmethodwas. The candidatecashcropsmatrix usedinformationfrom PRAsand other sources(such as informationconcerningproposedclient groupsand potentialmarkets)to select crops for further researchand development. Criteria were listed down the side and potential crops were entered as columns to form the matrix. The criteria were selectedby the projectteam and consistedof projectobjectives,agronomic,socioeconomic,marketandpost-harvest indicators.The scoresin the cells of the matrix werealso allocatedby the projectteam. Cropswere groupedinto first, secondand third ranks,andtie outcomewas usedto prioritize cropsfor research.Apart from the issueof whosecriteriaandchoicesformedthe basisofthe scores(obviouslyfarmers were not involved),a major problemwith this methodwas that it simplified and nanowed down the selectionof crops for research,rather than building on the diversityofcrops andfarmercroppingstrategies.It alsofailedto put the choicesinto contextanddid not reflecttheir changingnaturein a historicalsense. The QuIK ('QuantifuingIndigenousKnowledge')ranking method was orientated towardsincorporatinglocal knowledgeinto the researchprogmmme. QuIK is a rapid methodof assessing crop performancewithout field trials and wasapplied in research on ginger, mango, and cinnamon. Matrix ranking involved the systematic interviewingof experiencedfarmersand the generationof numericaldata on crop performance. Data setswere compiled using severalfarmersas replicatesand were then subjectedto statisticalanalysisusingan analysisof variance(ANOVA). The advantageof QuIK is that it canprovideinformationrelevantto the farm situationat a muchlower costthan field trials. This is particularlyusefulin tree crop researchas conventionaltrials can take manyyearsto yield data. Although the data may reflect morethan a singleseason'sresults(basedon farmers'accumulated experiencewith the crop in question),the methodis strictly limited to agronomicparametersand resultsshouldthereforebe treatedcarefully. QuIK analyseslocal knowledgeusinga formal scientific procedure which cannot possibly reflect the differentiation of knowledgeof cropperformance otherthanthat betweenthe informants.The problem of 'representativeness' whichplaguesconventional trials alsoprevailshere. Also, the 'replicates' treatmentof farmersas and groupsof farmersas a 'populationof key informants'is redolentof the scientificmethodfrom which manypractitionershave soughtto distancethemselvesby adoptingmore participatorymodesof inquiry.

Even whenconductedin the contextof FarmingSystemsResearch(FSR)or Farmer ParticipatoryResearch(FPR),PRAs may still fail to providea deeperperspective. FSRhasbeencriticisedfor not payingsufficientattentionto the historical,political, economic,and institutionaldimensionsof farmers'practice(Biggs and Farrington, 1991),andthereis little evidenceto suggestthat FPRis any better. FPR hasstrong technologydevelopment, testingand informationdissemination objectives.lt departs from FSRby focusingon thedevelopment potentialof farmers'own researchprocess,

and having a strongergroup or communityorientation. FPR would thereforeappear to be more suitedto a PRA approach,and a numberof commentatorshavetaken this view. However,as Okali et al. (1994)observe,PRA tools and techniques,and the increasingly standardizedframework in which they are applied, have not made a significant contributionto the understandingof group dynamicsand the processesof informallocal experimentation and informationdissemination in which FPR is most interested. PRA fares little better in assistinglocal peopleto define their problemsand the possiblesolutionsto these. The problem orientationof PRA has alreadybeen discussed in Chaptertwo. Askingpeoplewhattheir problemsare,andwhat solutions they can envisage, is not necessarily the best way of identifliing needs and opportunitiesandthe structuralfactorswhich underliethem; it shouldcertainly not be the only way. PRA often resultsin a 'shoppinglist' of local wishes;and it is not unusualfor the agencyconductingit to selectjust one or two itemsfrom this list for furtheraction. Researchand development activitiesinevitablyraiseexpectations in the communities in which they are undertaken;PRAs do this to an even greater degreethan normal,but frequentlyfail to deliver. To exacerbatematters,PRAsprovide few guidelinesfor further engagement,which is one reasonwhy they are usually undertakenat the start of the project processand rarely later. Also, and despitethe rhetoric of participation,PRA tools are obviously 'owned' by the outsiderswho introduce them and are widely perceived as suchPRAs are still essentiallypackagesfor the extraction of information and a tool for helpingoutsidersmakedecisionsabouttfteir intewentions.As PRAsarepresented as packages,practitionerstend to use them without further tliought, either becausethey lack the skills and/or confidenceto adapt and experimentwith methods,or because they are constrainedby time (field practitionersare often askedto conduct PRAs at relativelyshortnotice). Thereare good reasonsfor abandoningthe PRA packagewhile retaining someof its methods.Individualtools canbe usedseparately provided andfor specificpurposes, that their limitations are recognizedand/or they are combinedwith other methods. The tendencyamong practitionersto conductcomprehensive PRAs using all the recommended methodsshouldbe discouraged; topicalPRAsusingfewertechniques but targetingparticularissuesareoftenpreferable.Similarly,it may be more useful to undertaketopical investigationsin a larger sampleof communities,enablinga comparisonof results,than to conducta full PRA in a singlevillage;this would be intrinsicallymore sensitiveto inter-villageheterogeneity in whateverform (social, economic,institutionaletc.). Ii for example,oneofthe objectsofan actionresearch exerciseis to identify institutional opportunitiesfor natural resourcemanagement, suchopportunitiesaremuchmore likely to appearin a wide sampleof villagesthan in a single village study. It may also only be possibleto uncoveragricultural innovationsby examiningfarming practicesin severallocations. Homogeneityshould never be assumedin advance;it can be demonstratedonly by wideningthe n€t of investigation. The standardPRA packagedoeslittle to encourage

this, especiallyif it is linkedto othergeneralizing (suchas thosepromoted strategies by FSR). In spite of claims to the contrary,full PRAs can be both costly and time-consuming,and projectscan rarely afford to conductmore than a few. Unless they arecombinedwith otherkindsof investigation, a project'sunderstanding of the communitiesin its targetareamaybe correspondingly restricted. PRAsoften compareunfavourablywith what can be achievedby a singleexperienced fieldworkerin a fractionof the time andat a fractionofthe cost(a goodvillage study can be producedin a month,with additionaltime for writing-up). The full costsof PRAs are usually buried deepin prqject accounts,and it can be arguedthat thorough cost-benefitanalyseswouldshowthat the more 'traditional'methodsof investigation and interactionare consistentlymore useful and cost effective than PRAs. However, 'traditional'methodshave rarelybeendeployedin this context;the increasinguseof anthropologistsand other trained field workers as development consultants and projectstaff has largelycoincidedwith the growingemphasison the PRA package, and it is extremelyunusualfor theseworkersto be sentinto the field for more than shortperiods(or on grandtours). At the sametime, the apparent'success'of many PRAs can probablybe attributedto the presenceof such experiencedfield workers, ratherthan to the PRA processper se. PRAsare rarelysubjectedto critical evaluation.It is not in the interestsof project organizers to declaretheir shortcomings, andexternalprojectevaluatorsmay not wish to cast aspersionson the current fad (which is what PRA has become). On the positive side, the emphasisPRA placeson understanding (and to a lesserextent, engaging)local knowledgeand practiceis a distinct advanceon the more formal and technicaltypesof investigation. The activeparticipationof the targetpopulation,and govemment and NGO personnel, in the PRA process is also laudable. It is particularly important for providing training and experienceto members of the researchteam itself, especiallyif they are working largely as technicalextension agentsand/or have insufficient resourcesto undertakeintensiveinvestigationsin the field (asis thecasein manygovemmentdepartments). However, using the PRA packageis not necessarilythe best way to achieve participationor train field staff, especiallyif the outcomesof an exerciseare superficial or inaccurate;participation is pointless if it leads to unfulfilled expectationsand misguidedproposalsfor action. Similarly, there is no point in trainingfieldworkersto usetechniques whichdo not necessarily producedthe desired resultsandwhich may discourage themfrom thinkingfor themselves.A muchbetter approachis to 'unpack' PRAsand usevariousPRA tools in a more selectiveway; recognisingwhateachcanandcannotachieve,andcombiningthem,wherenecessary, with other(in somecasesmore'traditional')methods. Ironically,the mostusefultechnique,'semi-structured interviewing',is a 'traditional' technique drafted into the PRA package under a new and technical name. Open-endedinterviewing hasbeenaroundfor much longer than PRA and is usedby anthropologistsand others as a standardmeans of eliciting and exchanging informationin the field. Asking peoplequestionsand respondingto their queriesis

an everydaysocial activity, not just the principalelementin so-called'participant observation'.Trainingfield staffto do this well, andto recordthe outcomesof their interactions,is relativelyeasy,as most peoplepossessbasic discursiveskills. In additionto participatingdirectlyin otherpeople'sactivities,talking and listeningto them is one of the most powerlulmeansavailablefor establishingand maintaining 'participation').As we will seebelow,effective (or engendering socialrelationships communicationis essentialto understanding and engaginglocal knowledgeand practicein any context.

Undertakingparticipatory researchwith farmers The emphasison participatorymethodshasdevelopedits own specialmomentumin agriculturalresearchand developmentand relatedfields (including agroforestry). The PRA processis usuallytreatedas being most appropriatelyemployedin the initial stagesof problemidentification(diagnosis)and the provisionalproposalof solutions. Thereafter,different waysofworking with farmersare recommended,such as the formation (or transformation)of farmer researchgroups(FRGs). This kind of institutionalinterventionis partly necessitated by the fact that householdsor other domesticgroupsarethe primary,and often only, unitsof farm organization, and the sheernumbersof thesein any singleproject area precludesintensiveinteractions without some form of selection. The formationof farmer networksrepresentsa different,and sometimesmore effective,solutionto the sameproblem. Relatively little attention has been paid to the implicationsof creating such institutions; researchers havetreatedthem largelyas researchtools (without describingthem as In this such). section,we will examinesomeof the prosandconsof doingthis. We will begin,however,by examiningthe practiceof conductingparticipatoryon-farm trials, which exposesmost clearlythe continuingprevalenceof traditionalresearch agendas, albeitmodified$eaterfarmerparticipationin the researchprocess. On-farmtrials: ure they alwaysnecessary? Much effort hasbeenexpendedon the developmentof participatoryon-farm research methodologies,including powerful and sophisticatedmeans for analysing the complexanddisorderlydatawhichoftenemergefrom trials. It is clearlybetterto test cropsandtreatmentsin a field thanin a greenhouse, and it is betterstill to test them in the intendedbeneficiaries'own fields. Despitethe emphasison farmers'participation,however,the underlyingagendafor this type of researchis still generallyresearcher-driven. Researchers often set the objectives,designthe trials,choosethe treatments, analysethe results,andwrite them up and disseminate them. Even when farmersare given a more active role, it is difficult to escapethe conclusionthat they are essentiallybeing treated as 'proto-researchers', that and their practiceis beinggentlymanipulatedinto the mould of formal research.Somefarmersmay be happyto be treatedthis way, and accorded sucha novel degreeof respect,but it is hardto acceptthat the conditionsof on-farm

trials,evenwith the bestandmostparticipatoryof designs,replicatethoseofthe 'real world' withoutresearchers andthe expectations generates. whichtheir presence Participatoryagricultural researchdemandsthat agronomistsconduct on-farm trials, but in certain contextsthere may be no needfor them to do so. By striving to solve agricultural problems in order to meet the requirementsof domestic survival and subsistence, all farmersare conductingagriculturalresearch,year in and year out, althoughthey might not call it 'research'and formal researchersmight not recognize it as such. The number of experimentswhich agronomistscan undertake,with or without farmers' participation,palesinto insignificancebesidesthis, especiallyin situationswhere agriculturaldiversificationis alreadytaking place. Under these circumstances,it might be more useful for researchersto examinefarmers' practice as it is, determinewhat 'experiments'they havealreadycarried ouVarecarrying out, and note what the outcomeshave beer/arelikely to be. Furthermore,on-farm tnals may well divertresources awayfrom morerelevantinvestigations, andthesemay,for example,revealthat interventionis mostappropriateat a non-agronomic levels,such as in the marketingchain,ratherthan in farmers' fields.

3.3 On-farm trials in ZCCFSP The on-farmtrials in ZCCFSP'sprogrammewereoriginallyconceivedwith a view to converting farmers into researchers,modelling farmer practice on formal research. All the trials in the researchframework,from exploratorytkough to stationand then to pilot trials, were either designedby the project or largely researcherJed. Participationwasthereforecontrolledunderthe researchers' terms. Diflerent approachesto on-farm trials were tried, someof which were subsequently modifiedor rejected.For example,pilot trials basedon the commercialdevelopment of a single crop were dropped. Within a given area,farmerswere identified (opento any farmer), potential traders were found, and planting materials were distributed. The mainproblemswith were: the difliculty of conectlyidentifringa crop with real scope for expansionto commercial scale; the usual problems of farmers' expectations;and the difficulty of providing sufficient good quality planting material at the right time. However,the trials carriedout by the ZCCFSPagroforestry networkhad the explicit objectiveof buildingon farmers'existingsystems,usingfarmer-designed trials, from the outset of the programme. The mandatewas simply to observehow network farmers experimentedwith and manageddifferent tree and crop combinations (indigenousand introducedspecies),and assess their preferences for plantingniches and arrangementsand the ways in which theserelated to their householdand farm characteristics. Localknowledgeof indigenousspecies,andlocal practicein termsof tree/crop combinations, and planting sequencesand anangements in both the traditional bush fallow system(in sparselypopulatedareas)and permanentsystems in more settledareas).were evaluated.This enabled and analvsisof l0

changingpractice,and identificationof potentiallysustainable management options. A casestudy approachwas adopted,and differencesin practiceaccordingto gender and land tenurearrangementswere identified. This helpedin the targetingof species accordingto householdandgenderneeds. The first approach(researcher led) focusedmoreon adoptionand the second(farmer designed)on adaptation. It is doubtful, however,if any approachcould match the unaided work of the farmers themselves. Rural Zanzibar comprisesmore than 100000farm households,most with two or more membersinvolved (to varying degrees)in farming and (to a lesserextent)the saleof farm produce. Every year,they makeinnumerable decisionsaboutthe cultivationandharvestingofa wide varietyof farm plots,includingmajorchoicesaboutwhatto plant,whatto sell, andhow to sell it. This hasresulted,especiallyon Unguja,in an impressivedegreeof home-grown agriculturaldiversification(seeBox 2.5). In this context,on-farmtrials, whether farmer designedor not, are unlikely to have a wide impact. Although formal researchersmay feel that an inordinateproportionofthe farmers' 'experiments'have uninterestingdesignsor results,the cumulativeeffect can be very impressiveand have far reachingconsequences.In short,a lot can be leamedfrom looking at what farmers(and traders)are alreadvdoine.

Groupsand networks: institutionol intenentions The pros and cons of working with FRGs,farmer networksand various intermediate forms of institution,havebeenextensivelydiscussedin the literature(for example, Drinkwater, 1994). Participatoryresearchwith FRGsundoubtedlyhas the potential for buildingon local knowledgeandpractice;ideally,researchprioritiesshouldarise from the problemsand opportunitiesfacedby the farmers within the group. If it is well chosenand representative, new ideasor technologiesdevelopedby the group will be relevantto a wide sectionof the farmingcommunitywithin the area. Froman pointof view, it is easierto work with a groupof 15 farmersthanwith organizational individuals. If farmersdrop out, otherscanjoin; therewill be a groupmemoryof what hasbeendone,and thereis more chancethat activitieswill continueif project assistanceends. The leaming processwill be quicker as farmerscan exchange information and leam from each others' experiences. Other potential benefits of group co-op€rationinclude sharingresources,collectivemarketing,and attracting attentionfiom other organizations.FRGscan providea mediumfor trainingboth researchand extensionstaff, and the farmersthemselves. Therearealsomanydisadvantages of workingwith FRGs,especiallygroupscreated by researchers. Theformationof a groupis an institutionalintervention,andthis may havea numberof consequences. It is extremelydifficult to ensurethat FRGsare representative eitherof the local farmingcommunity,or of the wider areaor zonein which they are located. It is difficult to avoid treatingthem as research'tools', and easyto rely too heavily on interactionswith 'progressive'group memberswhile neglectingthe knowledgeand practiceof resource-poor farmers. It may also be

difiicult to show farmershow they might benefit from group membership,especially when their expectationsdiffer from those held by the researchersand project stafl Most of all, it is difficult to enswethe sustainabilityof FRGsafter projectshave withdrawntheir support. In orderto makeFRGs'work', significantinputsof time, expertise,materialsandotherresources aretypicallyrequired. Meanwhile,tiere are problems many as associated with the dissemination of researchresultsand group experiencesastherearefor on-farmtrials conductedoutsidea groupcontext. The formationof farmernetworkscanovercomesomeof theseproblems.Networks may comprisea mixture of individuals(and their households) and existinggroups (FRGscanalsobe formedon the basisof existingcommunitygroups).Theymay be geogaphicallydispersedand/orclusteredin particularcommunities;this helps to enswe a greaterdegreeof representativeness.In many respects,networksare more flexible than FRGsand allow a wide rangeof working methods,often at much less cost. However,the formation of a network is also an institutional intervention,and as such may be just as difficult to sustainbeyondthe lifetime of a project or other externalfunding(sustaininga local groupmaybe slightlyeasier"thoughonly if it has developedits own rationalefor existing). In view of their inherent 'ftagility', considerablethought should be given as to whether researcher-created groups and networksneedto be setup in the first place.

3.4 Farmer researchgroupsand networksin Zanzibar The developmentof farmer researchgroups(FRGs)and networks in Zanzibar took placegradually.Towardsthe end ofthe project,the tkee FRGson eachislandwere seen as key for enhancingfarmer participatory research. A network of farmers involvedin agroforestryresearchdevelopedin parallelwith the FRGs. The initial work with FRGs was candidatecrop-orientated,with researchfocusing on simple trials for measuringthe efilectsof different treatmentswith relatively few variables. However,the crops/varietiesintroducedoften proved unsuitableand failed to match farmers' expectations.A new sfiategywas thereforeevolved wherebyfarmerswere providedwith subsidizedseedlingswhich they could selectthemselvesand plant where and how they wished. However, this strategy was clearly inadequatefor dealingwith complexissuessuchas soil fertility or marketingproblems,although farmerswerehappywith the improvedsupplyof plantingmaterial. Otheractivitieswith FRGsincludedtrainingcoursesand workshops,and helpingto organizeor raisewider issueswith the relevantauthoritiesby lobbying. The main focus of the FRG approachwas on technology development. Problems with the approachcentred on the dependencyit generated. Groups were essentially researcher-created and farmers had high expectations. Other questionscould be raisedabout their cost effectiveness,the time taken by staff with individual groups, FRG's representativeness in termsof the wider targetgroup,andtheir effectivenessin disseminating the resultsofresearchandtraining.

The agroforestry farmers' research network provided an altemative focus for interaction between farmers and researchers(as well as between farmers and farmers). It comprised70 farmers,including women'sgroups,individual womenand individualmen, from 10 differentlocationson the coral rag of Unguja. Its main objectiveswere: (a) to assessthe importanceof different agroforestry(AI) species and systemsfor different locationsand socio-economicgroups;and (b) to expandand strenglhenthe network in eachvillage, building on farmer-to-farmerextension(using farmen as a sourceof ideas/plantingmaterial). The following activities were carried out: farmerdesignedon-farm fials linking researchand extension; study tours to exchange knowledge and experience about different A-F species and systems (farmer-to-farmer extension);and farmers'workshopsto €valuateresearch,identif the strengthsand weaknessesin the designof different systems,set priorities (using farmers' criteria), andplan future research. Some problemswere common to both the FRGs and the AF network becausethe network was also researcher-created and neededresourcesto bring its members together. However,the individuals and groups comprisingthe network were more 'natural' in a way that the FRGswere not; the problems of representativeness could be overcome becausetheir geographicaldispersal covered much more variation, including socio-economicvariation, and enabledmore effective targeting. Networks alsohelpto dissolvethe research andextensiondichotomy.

The focus of agriculturalresearchand developmentshould be farmers' existing practice and efforts to build upon that. The imposition of new 'methods' and 'institutions'(temporary or otherwise)shouldtake secondplace to this pragmatic emphasis. We agreewith Okali and Sumberg's(1988) view that a potentially powerful altemativeto the more classicmodesof on-farm researchis to begin with the 'farming system'itself; rather than developingnew systemsor identifuing 'technologicalbreaklhroughs', the approachshouldbe to 'work from and build upon the existingproductionsystem'. The objectof researchanddevetopment shouldnot be to introducepermanentinnovations,but to involve peoplein a processof'selfgeneratedinnovation' with the smallestnumber of technologiesnecessaryfor achievingsignificantsuccess(BunchandLopez,1995). In somecases,farmersare capableof achievingmuch the same,and more, without no extemal intervention at all.

Understendinglocal knowledgeand practice In the following sections,we discussaltemativeapproachesto natural resources researchand development. Methods should be shapedless by their participatory packaging and more by coherent approacheswhich focus on understandingand engaging both social and agroecologicaldiversity. This requires a deeper understandingof the theoretical underpinningsof different approachesin social developmenttheory, researchand practice,on the part of all practitionersin natural

resourcesdevelopment.A well-informedapproach,which acknowledges the theory that underliesit, will of coursebe subjectto changein the projectprocessand in the context of theoreticaldevelopments.However,the continualdevelopmentof an approachshould be seenas an important part of project practice and an essential 'tool' for understanding andengaginglocal knowledgeand practicein termsof both the nanowerandbroaderdefinitionswe haveprovided. Understandingagencyo.ndstructure: knowledgeandprcctice in context Local knowledge and practice shape, and are shaped by, patterns of social differentiationandthe socio-cultural, historical,institutionalandpolitical contextsin which they emergeandchangeover time. An understanding of the diversityof local knowledgeand practiceimpliesa needto understand socialdifferencesin the local context(suchasthe socialorganization ofthe family,kin, andcommunity;and social relationsof power,gender,and land tenure)andhow theserelateto wider processes of social,economicandpolitical change.This reflectsboth the 'actor-oriented'and 'neo-structuralistpolitical-economy' approachesalready outlined and their importancein understanding local knowledgeand practicein the contextof social changeat the microandmacrolevel. The actor-orientedapproachis useful for understandingsocial processesat the microJevel (including the knowledge processesembeddedin them) and the emergenceof different patterns of social organization, which "result from the interactions,negotiations,and social strugglesthat take place betweenthe several kinds of actor" (Long, 1992:21). By adoptinga broaderperspective, Booth draws "insights processes attentionto social in their wider contextandto reconciling about indigenousalternatives"(whichcanbe interpretedasour own understanding of local knowledge and practice) "with the kinds of urderstandingsof larger structwes without which they will lack realism"(1994a:17). This bringsthe two approaches togetherand stressesthe different 'kinds of understanding'requiredin our own pfactice. Understandinglocal knowledgeand practicein contextrequiresa historical approach which emphasizes the importanceof changein socialand ecologicalprocesses.By conffast,an actor-orientatedapproachstresses"the importantextentto which changes in the well-beingof rural peopleare the result of complex interactionsbetween individualsand groupsendowedwith differentmd changingamountsof knowledge and power" (Booth 1994a:11, emphasisadded). A broader,structuralapproach which viewsthe changingnatureof local knowledgein the contextof wider change, "the knowledgeof local groupsis not is equallyimportanthere. As Thruppsuggests, a static body of wisdom, but instead,usually consistsof dynamic insights and techniqueswhich arechangedover time throughexperimentation and adaptations to environmentaland socioeconomic changes"(1989: 15,author'semphasis).On this we can agree;it shouldbe noted,however,that adaptations of techniquesmay also refer to innovationsin institutionalpractice, which we addressbelow, and not j ust the adaptationin farming practicesto which Thrupp refers.

t4

Much of the recentliteratureon local knowledgesystemshighlightsthe importanceof institutionalchangeat both the micro-andmacrolevel. At the microJevel,the need to work with and throughexisting organisationsis an areaof specialconcem(Pretty, 1995;Scooneset al., 1994;Warrenet al., 1995) However,as Bebbingtonstates, "despite the importance accorded it, to the theme of local organisationsremains underdeveloped"Q99a:212). The participatoryapproach,and the use of rapid PRA methods such as Venn diagramming techniques to describe different Opes of institution,doeslittle to overcomethis problem. An explicit institutionalapproachis neededto providea thoroughanalysisof local institutionsand their complexand sometimesconflictingpractices. An understanding of institutionaladaptationand innovation(and the negotiationprocessesfrom which they emerge)asa form of local practice is a useful entry point to this kind of analysis. Further to this is an understandingof the changingconditionswhich give rise to institutional innovation. Bebbingtonnotesthat if organizationsareto be the agentsof a strategywhich is based on local knowledge,it is necessaryto "understandthe conditionsthat will structure the possibilitiesandtendencies of that agency"(1994:212). This pointsagainto the importanceof both structureand agencyapproachesto our understandingof local knowledgeandpractice,and institutionalpracticein particular. The political dimensionsof local knowledgeand practiceform an importantbut neglectedarea in nafural resourcesresearch. While practitioners have begun to recognizethe importanceof the social differentiation of knowledgethis is rarely extendedto includethe political differencesthat shapepeople'spractice. Political relationsshouldbe the subjectof moredetailedinquirythanthe useofrapid methods allows. For practical purposes,political di{ferentiationcan be viewed at the micro-levelin muchthe sameway asgender,classandethnicdifferences,especially when it is concernedwith access to resources(including lnowledge) and decision-making. These differencesmay have indirect consequences, if certain groupsareexcludedfrom the membership of local committeesfor example,or direct and negativeimpactsif the managementof naturalresourcesbreaksdown completely dueto partypoliticalconflict. The forcesof politicalchangearefelt on di{ferentlevels(local,regionalor national), and the relationshipbetweenthesedifferent levels needsto be understoodin its historical perspectiveand in the context of the broaderstructuralapproachwe have outlined.It is importantto recognize, however,the impactthattheseforcesmay have at the micro level,especiallyon local organizations, which,asBebbington(1994)has shown,are as much vehiclesfor political expressionas for the administrationof agriculturalprojects. This has implicationsfor our own practiceand the role that projectsmayplay in eitherfurtheringor hinderingthe claimsof particulargroups.Of key importance, therefore, is an understandingof the various forms of political expressionand the ways in which different positionsare negotiated. Peoples' manoeuvringsin this respect are an important part of local practice with which practitioners shouldbe mostconcemed. Understandinginnovation: the casestudyapproach

15

Observationis a simpleyet effectivetool for understanding local practice.Whether informalor systematic, the needfor researchers to usetheir own observations as well asthoseofthe local people,is againbasedon the premisethat what peoplesaythey do often differs from what they actually do. However, the role of observationis frequentlyunderestimated by practitionersengagedin participatoryresearch. This may be partlybecausethey associate it with the moreconventional, time-consuming and extractivemethodof 'participantobservation'usedin anthropological research. The way forward may therefore involve a synthesisof the two (i.e. participatory researchand participant observation)which recognizesthat both have somethingto teachthe other (Nelsonand Wright, 1995).The practicalimplicationsof this are clearlyspeltout by Richards,whoseinterpretation of 'participantobservation'in its mostactivesenseis particularlyuseful: "Where time and resourcespermit there is little doubt that 'participant observation'(i.e. taking part directly in the farm work, preferablyacrossa full farming season)is the best of thesediagnostictools (Johnny,1979; Richards,1985). Some apparentlyfamiliar problemstake on an sltogethernew significance perspective"(1985:l5l). whenseenfrom a participant's

Richards further advocatesthat successfulparticipatory researchin its wider sense will dependon regular and continuouscontactbetweenresearchersand user groups and a "willingress of researchers to live and work for considerableperiodsunder village conditions"(1985: 154). It would seem,however,that the participatory researchhe originally envisagedwas a little different from what PRA has now become,with its emphasison speed. The case study approachtakes a different starting point and begins with local knowledgeand practicein context,ratherthan problemdiagnosis. It assumesthat farmers and other actors are already conducting research, finding solutions to problems,and sometimessharingthem with others(whetherintentionally by word of mouth, or unintentionallyby demonstratingtheir efficacy). Casestudiesconcentrate on understanding the processof innovation(technicalczd institutional),the meansof its spread,and the existingpracticesor opportunitieswhich can be most effectively engaged.In termsof the wider influenceswhich impedeor give rise to innovation, the casestudy is most concernedwith the ways in which theseinfluencesare locally perceivedand contextualized.The tool set for the casestudy is more basicthan for a PRA and relies simply on semi-structured interviewsand farm observations.The time frame dependson the resourcesavailable and the amount of researchalready undertakenin the area. Furthercasestudiescarried out at regular intervalsand with differentobjectiveswill maintainthe regularandcontinuedcontactto whichRichards refers. Although distinct from the PRA approach,casestudiesneedbe no less participatory, especiallyif they involve 'participantobservation'in the active sensedescribed above. Although the type of ethnographicinquiry with which the casestudy is most often associateddoes not necessarilyentail direct action in the short term, the understandingit can generatecanl leadto more effective interventions,strategiesand actionsin the mediumterm. A detailedethnography of the kind describedby Long (1992), in which he focuseson a single good (beer) to explorethe relationship IO

betweenknowledge,practiceandsocialdifferenliation,is likely to revealmoreabout socialgroupingsand conflicts,'participants',and the scopefor action,than a PRA evercould. The casestudyalsohasa potentiallyusefulrole in assessing the impact of interventions and the innovationswhich may resultfrom them. Casestudiescan thereforeprovidea detailedunderstanding of the processand impactof innovation, andan approachfor more effectiveengagementof local practice.

3.5 Understandingand engaginginnovationwith casestudies ZCCFSPundertooka numberof casestudiesaimed at understandingand building on farmers' and traders' existing knowledgeand practice,designingmore eflective padicipatory researchand extensionstrategies,and targeting resourceswhere they were most needed. They were also used to help the government formulate appropriatepolicies and to inform the wider developmentcommunity in Zanzibar. The hrst studieswere carriedout with the six FRGswhich ZCCFSPhad begunto work with. What werethen termed'mini-PRAs' (with a casestudyorientation)led to the formulation of detailed plans for further participatory researchand extension activities with the groups. They also provided a key input to policy and planning initiatives,includingcollaborationwith otherdepartments in the Ministry to develop a zonalapproachto Zaruibar's farming systems. A furtherset of casestudieswas carriedout on cashcrop historiesin order to gain insightsinto the factorswhich stimulatecashcrop development.Studieson five relatively 'successful'crops looked at how their productionand marketinghad evolved over time, often without any direct input from governmentresearchersand exlensionists, andin somecases,in spiteof it. The cashcropschosenincludedsome which until then had been widely perceivedby the Ministry, and the project, to be eitherfood cropsor'local' cashcropsunworthyof researchers' attention. The cash crophistorieswerenot ony usedto identif,rexistingopportunities, but alsoto helpthe projectdeterminewhat kindsof researchandextensionstrategies wererequired(and which oneswereto be avoided)to promotesimilar developmentsin the future. The casestudieswere particularly useful for revealingimportant links betweencash crop developmentand patternsof social and economic differentiation in different areasof the islands. The studyon the developmentof orangesas a cashcrop in Ndijani revealedthe importanceof groupswhich had generallybeen excludedfrom previousanalyses; it highlightedthecentralrole in the local economyof squatters and agricultural labourersof mainland origin, and provided an understandingof social stratificationandthe differential accessto resourcesin the community. The emphasis on the historical developmentof agriculture in different areas,and the comparative natureof t}restudies,providedusefullessonsfor identifuingcashcrop potential,and possible impacts, in areaswhere permanentpattems of differentiation had not yet emersed. The case

was also usedto i l7

for cashcropson both islands. Productionand marketingstudiescarriedout for a wide rangeof cropsled to the formulationof cashcrop strategiesin collaboration with farmersand traders. The developmentof a SpiceStrategyfor Pembaand the ZaruibarMango Strategyare prominentexamples.All the casestudiesfocusedon understanding local knowledgeand practicein the contextof developingcashcrop productionand marketing,with an emphasison identifyingand further developing e-ristingopportunities.Basedon the premisethat farmersand tradersare already carryingout researchandfindingsolutionsto their problems,oneofthe attractionsof this kind of approachis that it canbe usedto maximizescarceresources, especiallyin Zanzibarwherefundsfor researchareextremelylimited. contemporary

Engaginglocal krowledgeand practice The following discussionis basedon the premisethat a comprehensive and coherent understanding is essentialif the processof engaginglocal knowledgeand practiceis to be effective. An importantelementof this approachlies in understanding and engagingthe diversityof local knowledgeand practicewithin agriculturalsystems and ecosystems.Existingpractice(whether'traditional'or'new') needsto be built on, as well as existinginstitutionsand local networksresponsiblefor the spreadof innovation. The importanceof engagingsuccessin order to generatesuccessis highlighted,althoughit is notedthat this shouldnot be at the expenseof practices oflenperceived to be of minorimportance. Building on local pructice and repertoire Effective engagement of local knowledgeand practicemeansbuilding on existing practicein orderto identify sustainable land useand resourcemanagement options. This approachis not new. Richards(1985) saw the traditionalsystemof shifting cultivationasa'compendiumof skills', andstressed that the mosteffectiveapproach to agriculturalchangemight be to recombinethe skills and methodswhich already exist within the shiftingcultivator'srepertoire,ratherthan designnew systemsfrom scratch. As an exampleof understanding and engagingdiversitywe can take the caseof ethnobiologicalresearch.The casestudyapproachis aimedlargelyat providingan ethnographicunderstandingof local knowledgeand practice in terms of social diversityand the pattemsof socialdifferentiationemergingand changingover time. The approachwe describehere has a slightly different focus and is aimed at generatingan ethnobiological understanding of local knowledgeand practicein the contextof agroecological diversity,as well as of the socialdiversityfrom which it emergesandto which it givesrise. The ethnobiological approachhasits entrypoint at the level of local practice,linking local plant and animalresources with the users andmanagers ofthoseresources.Theexamplediscussed belowrelatesspecificallyto ethnobotanical studies(people/plant interactions).

l8

An ethnobotanicalapproachto understandinglocal knowledgeand practice is concemedwith the plant-human intenelationships at the interfaceof peoplewith their "ethnobotany environments.In the wordsof Alcom, is the studyof contextualised plantuse"(1995:24). Traditionally,ethnobotanical researchhasbeenframedby the technicalapproachto IK andthe socialusesand management of planl specieshave frequentlybeenoverlooked.Ethnobotanical studieshavealsotendedto be extractive, for examplein their emphasisupon identifldngmedicinalplants for commercial exploitation. More recently,however,someworkershavewidenedtheir conceptof ethnobotanyand emphasizedthe needto concentrateon the ecologicalcontextsand dynamic aspectsof plant-humaninteractions. This approachcan be particularly fruitful if associated with effortsto develop(or reinstitute)an ecologicallysustainable agticulture. If ethnobotanicalsurveysare undertakenas part of a wider effort to understandland tenure and other relationships(including genderdifferencesin resource use and managementpractices), shifting the focus of interaction from researchto actionand from understanding to engaging(seeBox 3.6) is only a small step.

3,6 Engagingtraditional agroforestrypracticefor sustainableland use The agroforestryon-farm researchprogrammein the coral rag areasof Unguja was a collaborativeeffort, linking the Sub-commissions for Forestryand Research,and ZCCFSPand the FINNIDA-fundedZanzibarForestryDevelopmentProject(ZFDP). The traditional bush-fallowor agroforestrysystem,which is the main farming system on the coral rag, is becomingincreasinglyunsustainable as demandsintensi$ for forestproductsand cultivableland. Collaborativeefforts to solve the problemsof decreasingfallow periods,decliningsoil fertility, and wood and fodder shortages, focusedinitially on participatoryresearchinto agroforestry systemswhich incorporate trees,food and cashcrops. This researchwas intendedto supportboth the Village Forestry extensionprogramme,which had already started to promote agroforestry with exoticmulti-purpose trees,andthe objectivesof ZCCFSPwhich wereto testthe candidatecashcropsidentifiedfor the coralrag(seeBox 3.2). However,the emphasisof the researchchangedover time as it becameobviousthat elementsof the traditionalsystem,and the role of indigenoustree speciesin the fallow andin farmers'fields,werecriticalboth in termsof sustaininglivelihoodsand maintaininglocal biodiversity. Many of the indigenoustreesand shrubsprovide fruits, fodder, fuelwood,building materials,and medicines. Thesewere far more 'multi-purpose' and better adaptedto the harshconditions of the coral rag than the exotic specieswhich had been screenedand testedin on-stationtrials, and which were susceptible to the widespreadproblemsof drought,livestockdamageand fire damageassociated with coral rag agriculture.The role of indigenoustrees,and the ways in which they were managedin the bush fallow system over time, became importantfeaturesof the research. Oneof the most

research

ities.aimedat t9

usefuli

multipurpose trees and shrubs for agroforestry,involved an ethnobotanical assessment of the usesand management of tree,shruband herb speciesfound in the forestsand agricultural lands of the coral rag. Field researchwas carried out by a teamcomprisingfarmers,foresters,agronomists, agroforesters, local botanistsand an anthropologist. An important startingpoint was understandinghow land tenure and genderrelationsinfluencedlocalpeople'saccessto, andconhol,of coralrag landand how this in tum influencedtheir use and management of the indigenousspeciesit supported.The informationcollectedduringthis exercisewasanalysedby the whole researchteam, and then taken back to the farmers involved in order to generatea discussion of agroforestry opportunities andfutureresearchplans. Further information on indigenousspeciesand their managementwas derived from farmer-designed trials, workshopsheld with membersof farmernetworks,and field daysduring which farmerssharedtheir knowledgeof indigenoustreesand related practiceswith other farmers. One important output of this work was the preparation of a guideto indigenoustreesby a memberof the researchteam(Kombo, 1996),and a Swahilitranslationof a paperon agroforestry systemson the coral rag (Kitwanaet al., 1996). Thesepaperswerewidelydistributedandhelpeddisseminate the research findingsto farmersandotherinterestedpractitioners.Meanwhile,the discoverythat the agroforestryoptionsexercisedby farmers,and planting choicesin particular,were influencedby differencesin livelihoodstrategies (degreeof dependence on off-farm income),accessto labour(for land clearingand groupwork), and genderand land tenurerelationships,formedan importantinput to ZCCFSP'sown approachto the development of sustainable farmingsystems on the coralrag.

This is one exampleof the potentialfor building on local knowledgeand practice onceit hasbeenthoroughlyunderstood.Thereare no shortcutsto success.Prance arguesthat those ethnobotanicalstudies which have yielded useful ecological informationhavebeenlong-termand multidisciplinary:"Such work is successfully canied out only when somedeeperunderstanding of culture,not just tlat of plants, hasbeenachieved;oftenlong-termand,at times,tediousresearchis involved"(1995: 64). This point can be generalisedto the whole processof understandingand engaginglocal knowledgeandpractice,andit castsa shadowoverour own cultureof rapidappraisalandthe 'quick fix' whichthis so oftenpromisesbut frequentlyfails to deliver. Changingpattemsof differentiationat the level of local agencyboth influenceand are shapedby wider structuralchange. This is also true of human influenceon natural systemsand the way in which the changingnatureof the environmentshapes the livelihoodsof thosewho interactwith it. An agroecosystem is, after all, the interfacebetweensocial systemsand nafural ecosystems.The study of what goeson at the interface,and the natureof the interactionswhich form the basisof social and ecologicalrelationships, shouldbe of mostinterestto thoseworkingtowardssocially andecologicallyacceptable change.

20

Engagingsuccess Engagingsuccessinvolvessupportingthe processof local innovationand seeking opportunitiesfor empoweringknowledgeand practice. Empowermenthereis viewed as helpingpeopleto developconfidencein their own knowledge,ideas,insightsand capabilities,thusbuildingup their self esteemandsenseof power;this canhelp them to "selectivelyincorporate,adaptand take advantageof extemaltechnologiesand ideas,ifthey want them" (Thrupp"1989:20). Thruppnotesthat peoplecan become empoweredin this rvay if they are enabledto demonstratethe validity of their knowledgeto other farmerstkough farmer-to-farmereKensionand groupworkshops. This approachis echoedby BunchandLopez(1995)who recogaizethat an important goal in supportinginnovationis for farmersto be confident,leam, and become motivated to continue developingtheir own agriculture. Farmerswill incorprate traditionalcomponents into extemaltechnologyandthey shouldnot be madeto feel that theyhavefailedifthey do. In community-based conservation in particular,successfulexamplesof participatory projectsarehardto find, andthe participatorymelhodsseenas 'recipes'for success havehada disappointing impact(Wells,1995).However,success is a mostimportant recipe for just that - success- and can help establishthe legitimacy('effective engagement'or practice)of local knowledgeand justi! calls for more funds and resourcesto expandactivitieswhich seekto build on local knowledgeand practice (Thrupp, 1989). Also, 'sustaininginnovation'after interventionmeansachieving (BunchandLopez,1995). However,it is importantto evaluate recognizable success successaccordingto local people'sneedsand aspirations. Local knowledgeand practice, and tle successesto which their understandingand engagementmay give rise,shouldbe perceivedin localpeople'sterms. All too often, developmentpractitionersfocus on the problems identified by themselvesand their clients,and the PRA processoften encourages this, with its emphasisuponproblem-ranking and relatedmethods- They usuallypay ratherless attentionto the identificationof opportunitiesand examplesof success. Listing problemsis relatively easy,althoughit may require more thought to provide a structuralanalysisof them. Proposingsolutionsis a little more difficult, especially solutionswhich areappropriate to the wider contextof the problemsin questionand which can be realisticallyachieved.Identiffing opportunitiesis harderstill, and the methodsat our disposalfor doing this are generallyinadequate.Only by paying closerattentionto farmers'andothers'practicecanwe can recognisesuccess andthe potentialfor it; andonly by creativelyengagingtheir practicecanwe hopeto build on success. Conclusion ln this chapter,we havelookedcloselyat someof the differentmethodsemployedto understandand engage local knowledge and practice in the context of natural resourcesresearchand development. We have contrastedsome well-known 'participatory'approaches with more consideredapproaches to the samequestion. )l

Many more examplescould havebeenprovidedand somepointsmentionedbriefly couldhavebeendevelopedfurther. However,it is hopedthat sufficientdirectionhas beengiven to enablereadersto pursuethis task for themselves.So far, we have focusedlargely on local knowledgeand practicein its narrow definition, where 'local' refersto the world of the primary targetsof development intervention.In the next chapter,we will expandour argumentto a wider arena,and considerthe knowledgeand practiceof projectsand the differentactorsand institutionsdirectly associated with themat a moreinclusivelevel.

22

CHAPTER FOUR

UNDERSTANDING AND f,NGAGING PROJf,CT PRACTICE

Introduction In the last chapterwe looked at different aspectsof the social (institutionaletc.) differentiationof local knowledgeand practice,and how an understanding of these shouldshapeapproaches andinfluencethe choiceandapplicationof methodsusedin naturalresourcesresearchand development.We focusedon local knowledgeand practiceas conventionallyunderstood,at the level of the 'community' or other primary targetgroup. In the presentchapterwe will return to our expandeddefinition of'local', and look more closelyat the wider institutionalcontextof development practice. We will focus in particularon projectpractice(includingthe knowledge and practiceof their stall), treatingprojectsas 'local' institutionswhoseknowledge and practicemust also be understoodand engagedby developmentpmctitioners, whethertheyareworkingwithin theseprojectsor not. The simplepoint we makeis that we should understandand engageourselvesas well as other actors in the developmentprocess. This is more than a question of engagingin critical self-reflection. In this context engagingproclice unequivocally meanschanging practice,andwe will suggestsomeofthe practicalwaysin whichthis mightbe done. Theexternal interventiondebate Beforewe tum to look at projectsaslocalinstitutions,we shouldsaysomethingabout whatmight be calledthe 'externalinterventiondebate',which in somewaysparallels the 'indigenousknowledgedebate',alreadydiscussed.In recentyears,a growing numberof academics, includinganthropologists, havescrutinizeddifferentstagesof process,from policy formulationthroughto impactevaluation,and the development 'deconstructed' the conceptswhich underlieits standardprescriptions.Long andvan der Ploeg (1989) provide perhapsthe most authoritativeanalysisof intervention practice,having first exposedthe theoreticalassumptionsunderlyingthe policy modelswhich areadvancedin supportof it. Accordingto this analysis,intervention is "an ongoing,socially-constructed andnegotiatedprocess,not simplythe execution plan of action with expectedoutcomes"(1989: 22S) We of an already-specified agreewholeheartedlywith this view and are happyto be identified with what might be calledthe 'sophisticated' conceptionof extemalintervention. However,aswith the sophisticated conceptionof IK, this perspective on development practicecan easilyslideinto a naiveanarchism.Longandvan der Ploegsuggestthat their approach"can be usednot only to understand,but also to transformthe practice of intervention"(1989: 242). We have written this book as our own modest contributionto sucha programme,althoughwe havenot embracedan actor-oriented approachto theexclusionofa structuralist perspective.At the andpolitical-economic sametime, we do not believethat the transformation of interventionshould.or will.

result in its abandonment, as someresearchers imply, and we hold this view for a numberof reasons. Firstly, there is no evidenceto suggestthat the different international and national agencieswhich cunentlyhavestakesin development interventionare aboutto give themup. Secondly,we believethat we shouldbe strugglingto changethe practiceof these agenciesbecausewe are stakeholdersin many of them (both as potential employeesof them and as citizens of a nation which has its own agency and contributesto others).Thirdly,evenif theydid tum their backson intervention,other govemmentand non-govemment agencieswould probablycontinueto intervene. Plannedinterventionis not the preserveof intemationaland national agenciesbased in developedcountries,but is the practiceof many agentsat diflerent levelsin the developmentarena. None of these agenciesor agentsshould be automatically privileged,althoughthis is the positionadopted,explicitly or implicitly, by some adherents of actor-oriented research andradicalpopulism. Our own pragmaticapproachis to acceptthe reality of extemalinterventionwhile activelytrying to changeits theoryand practiceso that it morecloselymatchesthat reality. To someextentwe agreewith the positionadvancedby Hulme (1995)in his discussionof altemativeapproaches to projectidentificationandplanning.He argues "there that is no optimalmodelfor planningagriculturalor rural projectsbut, rather,a set of alternativesthat should be considereddependingon the specific context and objectivesof an intervention" (1995: 211). The conclusionto his paper is reproducingin full below. "This paperhasattempted to chart the wide rangeof responsesopen to those seekingto make project planning and identification more effective, in terms of achievingthe stated objectivesof rural and agriculturaldevelopment,It haspointedto the inadequacies of the onhodoxmodelupon which the training of project personnelis largelybasedandfrom which most project planningmethodologieshave beenderived. The conflictingimageofprojects aspolitical arenasin which powerful groupsconflict and bargainin their attemptsto set and manipulatethe agendafor public action hasbeenpresented. This appearsto havesignificantexplanatorypower in termsof actual project processes,but its rejection of technicalanalysesas mereguisesfor self-interestis unsatisfactory.Thosewho continueto pusue the narrow confinesof the orthodox model are likely to define 'what shouldbe' without relating this to 'what is'. Those who remain within the boundariesof the political model may be academically comfortable,but may conderm themselvesto being permanentlymarginalcritics who are not able to contributeto changesin pfoject practice. It is the ground betweenthese two positions, termed the hybrid model in this paper, that offers opportunitiesfor the developmentof project methodologiesthat are both desirableand feasible. This searchfor improvedmethodologiesis not for one optimal project identificationprocedure,but for the production of a numberof altemativemethodologiesthat can be recognisedas being more or less appropriatein certaincircumstances.It is not simply a questionof blueprintversusprocess(Sweet & Weisel,1979),but a questionofwhich form ofblueprintor process, in whichcircumstances, andevenof what meansmay be used to irftegrateblueprint and processapproaches. For rural and agricultural initiatives in developing countries, where uncertainty is high, knowledge is limited and intended beneficiariesare commonlyperipheralto centresof local and national decision-makingpower, then processapproacheswhich make a seriousattempt at beneficiaryparticipationand informal institution buildingare likely to be mostrelevant. The dominantirnageof projectsas a technocraticexercisemust be replacedwith a revisedimagethat recognisesprojectsas arenasfor conflict, bargainingand trade-offs, and in which data and technical

tools havethe potentialto clarii/ likely outcomesand shapearguments.This is an altogetherlesscosy imageofwhat projectsandprojectidentification areabout,but in realitythetextbooknotionsofproject planninghavenevertranscended theircoverc."(1991:229-230)

In the following sections,we will applya similar perspective to differentaspectsof projectpractice,andnotjust to projectidentificationandplanning.

Incorporating local knowledgeinto project practice Local knowledge(with or without practice)is generallyunderstoodto mean the knowledgeof people in a project's target 'community' (usually rural dwellers). Incorporatinglocal knowledgeinto projectpracticeis thereforetreatedas part of the problemof makingprojectsmore 'participatory'by drawingthesepeopleinto the projectprocessfrom startto finish andbeyond. Oneway of tacklingthis problemis to examineeachstageof the projectprocess,lookingat waysin which this might be achieved. Project designand evaluation Reg ( l99l ) arguesthat "If we want to give indigenousSWC [soil and water conservation]and other forms of local environmentalknowledgea real chance,then projectdesignshouldbe thoroughlychanged": conventional "Many donor

agenciesfield a numberof missionsfor projert identification,preparationand appraisal. Thesemissionsoften lake 3-4 weeksin the field followedby a similarperiodfor report writing at headquarters,they involve severalconsultantsand gaps of severalmonths betweeneach missionare common.Thisdesignchainis highlyinadequate. Projectidentification missions tendto spendhalftheir time in the field talking to public administrators,staff of various ministries(agriculture,environment) andto somerepresentatives ofthe targetgroup- oftenvillageelite. The restofthe time is spentin the capitalon data collection(censusdata,pricedata,etc.) and on discussing with ministriesand donor agencies.This type of identification missionis usuallynot in the positionto identi8/and analyselocal perceptions,prioritiesandenvironmentalknowledge.Assumingsomecontinuitybetweenwhat hasbeen identifiedandwhat hasbeenappraised,it is importantthat identificationmissionsget it right. Therefore the emphasisduring designshould shift from appraisalto identification. A solution is to field small (2 or 3 consultants identification missions insteadof6 or 7), who know the regionwell andareprepared to stay3 monthsin thefield."(Reij 1991:15).

Reij's characterization of identificationmissionsshould,perhaps,be qualified by noting that not all projectsare'new' in the sensethat they are not all building on nothing. It is importantthat identificationmissionsshouldtalk to publicofficialsand membersof otherinstitutions(includingresearchinstitutions),not only becausethey may haveimportantinformationon, and insightsinto, local knowledgeand practice, but alsobecausetheyare importantstakeholders in their own right. Otherwise,it is not unusualfor projectsto be proposedanddesignedaroundthe experienceof others which have precededthem in the same location (see Box 4.1). In this case, evaluationsof former project(s)may provide much of the requiredinformation, especiallyif they have been conductedin a 'participatory'way. This does not eliminatethe need for stakeholderanalysisand stakeholderworkshops,although again,theseshouldnot be confinedto just the donor agencyand membersof the 1

target 'community'. If project identificationis startingfrom scratch,then Reij's recommendations havemoreforce. A morewide-ranging assessment ofthe implicationsfor projectdesignis providedby Blaikie et al. (1996), basedupon their more 'sophisticated'conceptionof local knowledge. Farringtonprovidesa concisesummaryof their main argumentsas follows. "B&riers to fuller incorporation of LK flocal knowledge] into researchand developmentprojects include the (often unacknowledged)social and political agendasof outsideagencies,the professional and cultural backgroundof their sta$ and the often contradictoryagendasamongstlocal people and outside agencies. Structural and behaviouralfactors often lead outside agenciesto design projects without allowing time for local participation. In additioL outsidersare often not adequatelytrained to recognise and deal with unacknowledgedprofessionaland cultural agendaswhich may underlie interactionsbetweenoutsidersandinsiders. Progresstowardsremovingsomeofthese barrierscanbe madeby introducingnew skills into the design and implementationof projects, including facilitation, conflict management,and negotiation,and by supportingexistinglocal networksandpathways,andsuccessfulinsider- outsiderinstitutions. In terms of projea management,somepreliminaryfunding may be required to ensureearly dialogue betweenscientists,developmentworkersandintendedbeneficiaries,so that a clearsharedunderstanding of the designand implementationprocesscan be gained. Furthemore, policiesat the nationallevel in developingcountries should be examinedto see in what ways they encourageor discouragethe resilience,vitality andadaptabilityofLK. Theymayhaveparticularrelevanceto, for instance,the extent to which LK incorporatedin plant selectionis protectedby legislationon intellectualproperty rights." (r9e6.2)

It is interestingto notethat althoughBlaikieer al. (1996)examineda largenumberof ODA (DFID) prqectsfor evidenceof a concemwith local knowledge,theyhavelittle to sayaboutprojectevaluationas a specificactivity,apartfrom recommending that projects action researchon on-going involving local knowledge(in their restricted definition) should be undertaken. Their silence may be related to the general weaknessof DFID and other agenciesin project evaluation,especiallywhen the projectshavebeencompleted.This is becausethe shortcomings of pastprojectsare unlikelyto be acknowledged duringthe processof self-evaluation by projectstaff and otherindividualaid organization stakeholders. Opportunitiesfor institutional leaming are therefore lost, though individual practitionersmay gain somethingfrom their own experience.This problemexistsin all sectors,not just in natural resourcesresearchand development. Time and resourcesfor detailedpost-projectevaluationsare generallyscarceand certainly not availablefor evaluatingeverypastproject. In this respect,the recommendation put forwardin Blaikie et al. is quitereasonable, especiallyif it canbe ensuredthat action rs taken on the recommendations of suchaction research.Otherwise,more could certainly be done to record and leam from practitioners' and other stakeholders' opinionswhenprojectscometo an end (at the very leastby askingthem to fill in a moreelaborateversionofthe typeof form usedto evaluateworkshops,or by holding workshopsthemselves). Some platform from rvhich they can air their opinions shouldbe orovided.

Theneedfor time Blaikie et al. (1996) make the importantpoint that project designseldomallows sulficienttime for localparticipation.A similarpointis madeby Reij, who notesthat a major constraintupon incorporatinglocal knowledgeinto projectsis the fact that "govemment,donoragenciesand the presswant quick and tangibleresults"(1991: 15). "It may take3-5 yearsbeforethe bestandmostacceptable technicalpackageis identified, hencetangible resultscan rarely be obtainedbefore 5-10 years have elapsed.It is essentialthat donoragencies andgovemments acceptthesetime frames for projects"(1991:16). Althoughnot everyoneis aiming for'technical packages', this point is valid. Oneof the supposed advantages of'participatorypackages',which was mostexplicit in ,lR.A,is the relativespeedwith which they can be carriedout; this characteristic('quick, easy and cheap') has helped to make them widely and engaginglocal knowledge(and practice) acceptable.However,understanding canbe a lengthyprocess,andprojectcyclesoftenprovideinsufficienttime. (Reij, andBlaikie el a/., recommend This is not only neededfor outsideresearchers that more formal researchon local knowledgeshould be undertaken),but also for other projectstakeholders.Researchby an experiencedresearcher, especiallyif it focusesupon particulartopicsor locations,can often be undertakenin a relatively shortperiodof time. For example,an antkopologistwith relevantexperienceshould be able to producea detailedanalysisof a village communityin a month or less, dependingon how much researchhas alreadybeendone in the generalarea(PRAs, which typically result in poorer analysis,may take much longer in terms of person-days).In orderto engagelocal knowledgeand practice,however,theremay be no alternativebut to follow the paceof the peopleand institutionsbeingengaged (this appliesequallyto the posfPRA process).Projectmilestonesand calendarsare generallymeaningless exceptin projectoffices. Naturalresourcesprojectsmay also demandmoretime by their very nature. Many tree species,for example,whetherwild or cultivated,do not reachmaturity during the short life time of forestry,agoforestryand cashcrop developmentprojects. This problem is particularlyacute in conservationprojects and those which support initiatives. The development is communitymanagement of communitymanagement a longtermprocessandunlessprojectsaredesignedsothattheir key activitieswill be (whichtheyrarelyall canbe),the long-termprognosismaybe grim. sustainable Time constraintscan sometimesbe overcomeby reproducingprojects;extending them into anotherphaseor initiatinga furtherprojectbasedon the experienceof its predecessor. ad hoc Althoughthis maybe effectivein somecases,it is a dangerously approach.Decisionsaboutsecond(third etc.)phasesandprojectsareusuallymadeat a very late stagein the cycle, and funding is often not secureduntil after a project has ended. Meanwhile, uncertaintyprevails; project staff (and sometimesother may not know whetherthey shouldbe preparingto concludetheir work stakeholders) it is difficult to planfor sustainability;it is or continueit. Underthesecircumstances, probably safestto assumethat the project will end and that attemptsshouldtherefore be made*o rneetits originalobjectivesin thetime available.

Altemativeapproaches to projecttime framesareclearlyrequired. 'Processprojects' arenot enough;a greaterflexibility hasto be built into the projectprocessitself, even if this meansabandoningthe conventionalnotion of what a project is and what it shouldlook like. This will obviouslybe diffrcult for mostaid agencies,in which the paceof institutionalchangerarelymatchesthat which they demandof their clients overseas.However,this issuemust be tackled if our capaci$ to understandand engagelocal knowledgeandpracticeis to be developed. One option may be to provide funds for open-endedsequencesof intervention' which placea greaterrelianceon on-off consultancy inputsthan permanentpostings in the field. Progresscan be evaluatedduring and after eachdiscreteinput, and inputs canthenbe clearlytargetedandtheir institutionalimpactscarefullyconsidered.Such a processwould providethe motivationfor promotingsustainabilityand a meansof ensuringits development, especiallyin the periodsbetweenextemalinputs. It would not be appropriatein everycontext,but it is evidentthat the traditionalprojectcycle is often inappropriate.Under the circumstances, it would seemquite rational to experimentwith altematives Project objectives Blaikie et a/. found that relatively few ODA natural resourcesprojectsmadeexplicit reference to localknowledgein their titlesor statedobjectives,andconcludedthat the data"suggestthatLK hasbeena minimallyidentifiedcomponentin ODA projectsto date" (1996:24). The increasingemphasison 'participatory'approaches, however, impliesthat this situationis changing.Blaikieet al. alsonotethatthe (ratheroblique) recognitionof the importanceof theseapproachesin ODA's revisedRenewable NaturalResources ResearchStrategtfor 1995-2005(ODA, 1994a)offerssomehope for the future, athoughthere is a dangerthat a gap betweenrhetoric (the rhetoric of 'participation'andthe 'incorporation oflocal knowledge')andpracticewill develop. Thereis no intrinsicreasonwhy local knowledgeshouldfeatureexplicitly in project objectives. It is more importantthat projectsshould have clear and achievable objectives,that differenttargetgoups or stakeholders areclearlyidentified,and that (whether their own objectives theseare construedas 'knowledge'or not) form an input to project objectives. This might be achievedin a numberof ways. The organizationof stakeholderworkshopsas part of the projectplanningprocesshas much to recommendit, especiallyif theseworkshopsare facilitatedby consultants who are relativelyindependent of the principalstakeholders.Processes of this kind should ensurethat local knowledgeand ptactice are engagedfrom the very outset, whatever specific project objectives emerge from negotiation between the practice,and stakeholders. This kind ofengagement shouldalsopermeatesubsequent to this extent local knowledgeand practiceshouldbe an integral componentof every projectwhetherstatedin its objectivesor not. This is alreadythe casein termsof our expandeddefinitionof local knowledgeand practice,althoughit may not be recognizedas such. All projectsoperatewithin the

arena of local knowledgeand practice and engagethem, knowingly or not, in differentways. The problemis that existingmodesof understanding andengagement are often deficient,especiallywhen they are not critically and creativelyengaged themselves.The meta-objective of all projectsshouldbe to understand and engage localpracticein its widestsense,includingtheir own practice.

Project practiceas local practice Theproject as a locdl institution Projectsshouldbe consideredas local institutions,not only becausethey may be based 'locally', but also becausethey directly engageother local institutions, whereverthe projectitself is located.Projectsoftenform part of otherinstitutionsor are linkedto themin someway. Theymay be strictlytemporaryinstitutionslasting only for the durationof the 'project cycle', or they (or partsof them) may have a longerlife, eitheras(semi-)autonomous institutionsor asintegralpartsof others(see Box 4.1). Whenprojectsarebeingplannedit is importantquestionsto askwhat kind of institutionsthey shouldbe. Thereis no singleanswerto this questionbecauseit has to be approachedon a case-by-case basis,taking into accountboth project objectivesandthedifferentinstitutionaloptionsavailable.

4.1 ZCCFSPas a local institution ZCCFSPwas both an institutionin its own right and a part of other institutionsto varying degrees. It was fundedby ODA who also directly employedsomeof its expatriate TCOs (Technical Co-operationOflicers) and, through a different arrangement,the APOs (AssocateProfessionalOfficers) attachedto the project. OtherTCOswereemployedtkough NRI (which,until it was privatizedin the final monthsof the project,was the 'scientific arm' of ODA) who managedZCCFSP. Apartfrom supplyingfundingandstaff,ODA's mainrole wasto provideadvisoryand relatedinputsvla the British DevelopmentDivision in EastemAlrica (BDDEA), its regionaloffice in Nairobi, Kenya. The most important inputs were providedthrough annualreviewmissions,includinga majormid-termrevieworganizedby BDDEA in late1993. Day-to-day managementwas left to NRI. In the early years, the designatedNRI ProjectManagerin the U.K. playeda key role in settingmilestonesandensuringthat they weremet. Later on, a new ProjectManagerwho gavemuchmore autonomyto ZCCFSP'sFieldManager(an NRI employee).NRI headquarters wasthenconcemed mainlywith providingadministrative supportandmoregeneraladvice. Up to a point, this systemworked well, although some TCOs regrettedthat BDDEA and./orother extemaladvisorswerenot availableto providemoreactiveguidancealongthe lines of the mid-termreview.

ODA andNRI werenot the only 'extemal'stakeholders in ZCCFSP.TheForeignand CommonwealthOffice (FCO) providedTechnicalCooperationFunds (to support TCOsin post,includingtheir families)throughthe British High Commission(BHC) in Dar es Salaam,Tanzania.As the ultimateoverseerof British aid to Tanzania,the FCO,tsHCalsohada vestedinterestin the project'sprogress.For example,towards the end of the projectthe Field Managerwas askedto provideBHC with a list of ZCCFSP's'tangibleoutputs'. Followingdissatisfaction with political developments during and after Zanzibar'sfirst multi-partv elections in November 1995, the FCO,tsHCintervenedand suspended British aid to Zarzibar. This put an abruptend to plansfor a newprojectdesignedto build on ZCCFSP'swork to date. Planningfor this was well advanced,had involved the active participationof the Zanzibar govemment,and was actively supportedby ODA.tsDDEA. Ironically, the new projectwasto havehada moreexplicitlyinstitutionalfocusthanZCCFSP. ZCCFSP'sprincipal institutional locus in Zanzlbar lay within the Ministry of Agriculture, Livestock and Natural Resources(MALNR). At the start of the project, it wasattachedto theparastatal CashCropsandFruit Commission(CCFC),but it was later relocatedto the Sub-commission for Researchin the Commissionfor Research and Extension (the organizationof the MALNR into this and three other Commissions, Planningand Administration,Agricultureand Livestock,and Natural Resources,was relativelynew). ZCCFSPwas one of four donor-fundedprojects dominatingthe work of this Sub-commission, andwasgenerallyreferredto as 'Cash Crops' to distinguishit from 'Food Crops' (under the IFAD-funded Zanzibar SmallholderSupportProject),'Coconuts'(the GTZllDA-fundedNational Coconut DevelopmentProgramme), and 'PlantProtection'(the Dutch-fundedprojectentitled 'Strengtheningthe Plant ProtectionDivision of Zanzibar'). All these projects developedextensionactivities,in manycasesindependently of the Sub-commission for Extension(supportedby a UNDP-fundedproject and implementedby FAO). Althoughthe differentprojectscollaboratedon specificprogrammes, their different approaches to researchand extensionremaineda constantsourceof friction, both statedand unstated. ZCCFSP'sinstitutionalpresencewas further complicatedby the attachmentof an APO agroforesterand two employeesof the Sub-commission for Forestry(in the Commissionfor Natural Resources),and the appointmentof a TCO marketing economiston the understanding thathe wouldalsowork within the Ministry of Trade, Industryand Marketing. ZCCFSP'srvideninginterpretationof its objectives,and especiallyits drive to institutionalizea Farming Systems/Farmer Participatory ResearchapproachthroughoutMALNR, conflicted with its primary institutional location(not to mentionthe widely-heldperceptionthat its main job was still to developcashcrops,and in particularaltemativesto cloves). As one projectwithin a singleSub-commission, ZCCFSPdid not havethe institutionalstrengthto carry its expandedobjectivesthrough. This is onereasonwhy the proposedfollow-onproject wasto be basedin the Commissionfor Planningand Administrationwhereit would potentially have a much wider impact on MALNR and be able to tackle the recognized weaknesses of MALNR asan institution.

However, relatively little thought was given to the sustainabilityof ZCCFSP's activitiesas the 'CashCrops' sectionin the Sub-commission for Research.It could be arguedthat if ZCCFSPhadstuckmorecloselyto originalobjectives,especiallyas thesewere perceivedwithin MALNR, more attentionmight havebeenpaid to this issue. One TCO even suggested that ZCCFSPshouldhave remainedwithin the CCFC, and thereforeoutsidethe line structureof MALNR altogether. Planning for the new project,and the generaloptimismthat it would go ahead,delayedserious consideration of the 'CashCrop' section'sfutureuntil it was too late. Negotiations secureda promise of support(for one year) from IFAD firnds, but past experience suggested that this promisemay not translateinto practice. When ZCCFSPwas discontinuedin June 1996,capitalequipmentwas handedover to the 'CashCrops' section(who continuedto usethe projecttitle). Within a monthof the handover,the project's successorhad run into diffrculties, partly becauseMALNR could not afford to pay its runningcosts(ZCCFSPhad subsidizedsomeof thesecosts,althoughthe original project agreementhad specifiedthat this was the responsibilityof the Zamibar government).The local institutionswhich ZCCFSPhad created,including its FRGs.havehadsimilardifficulties.

Ihstitutionaloplions There are severalobviousways in which projectscan be strengthened.As local institutions,they shouldbe accountable to their stakeholders.The problemis that somestakeholderinterestsare frequentlyignoredat the expenseof others. Many projectsthereforebecomesites of conflict betweendifferent interests,and are riven with disputes over purpose,the distribution of resources(including training resources),and the ownershipof their tangibleand intangibleproducts(including researchresults). Some degreeof conflict is to be expectedin any institution, especiallyin one with the diverseanay of individualand collectivestakeholders so projects. planners, tlpical of Project however,are often unableor unwilling to consider all the options which might reduce the potential for conflict; project practitionersmay alsobe ill-equippedto dealwith conflictswhenthey arisesbecause of a similarblindnessto altemativesandlackof management skills. Someform of stakeholder engagement shouldbe the startingpoint for everyproject. Plannersand practitionersmay be more familiar with stakeholderanalysis(SA). Grimbleel al. (1995)suggest the followingdefinitionof SA: "An approachfor understandinga systemby identifyingthe key actors or stakeholdersin the system, andassessing their respectiv€interestsin that system.Stakeholdersincludeall thosewho affect,and/ or are affectedby, the policies,decisions,and actionsof the system;they can be individuals,communities, social groups or institutions of any size, aggregationor level in society. The term thus includes policy-makers, planners and administratorsin govemment and other organisations,as well as comrnercial andsubsistence usergroups."(1995:3-4: originalpassage in italics)

Grimbleet al. alsoidentifuwo key objectivesof SA:

"(i)

to improve the ef/ecti',vnessof policies and projects on the ground, by explicitly considering stakeholders'interestsand the challengesthey may present,identi!,ing and dealingwith (before they arise) conflicts over natural resourcesbetweenstakeholdergroups, and consideringth€ potential fcr andcompromise. cooperation (ii) to better addressthedistributional and rocial impactsofpolicies andprojectsby brealing down the analysisto assessseparatelythe interestsof, and impactsof intervention o4 diferent stakeholders. Considerationis also gjvento trade-ofls betweendiferent policy objectivesand priorities (in particular (1995:4: italicsin the original) economicandequityconsiderations)." betweenenvironmental,

In these quotations,SA is being describedin the context of natural resource management. Grimbel et al. argte convincingly that SA can be both a powerful to activitiesdesignedto empowerlesspowerful analytictool azd a stepping-stone stakeholders. SA can be applied at different levels and at different stagesof the projectprocessfor a varietyof purposes.Althoughthey do not specificallymention this, SA canbe appliedequallywell to the analysisof projectsas institutions,to help determinewhat institutional form they should take and how they might be adapted overtime to thechanginginterestsoftheir differentstakeholders. The major limitationof SA in this contextis that it is primarilydesignedas a tool to extemalprojectplannersand practitioners. be usedby just one set of stakeholders, primary As its name suggests,the emphasisis on analysis. To the extent that prerequisite is a necessary for practice,this is satisfactory.However, understanding on its own, SA doesnot ensurestakeholderengagement,so the meansfor achieving workshopshasalready this shouldbe activelyexplored. The holdingof stakeholder beenmentiioned.Thesemay be particularlyeffectivein the early stagesof project design if all Ihe major stakeholdersare represented,and if the workshops are local consultants.This facilitatedby relativelyneutralagents,suchas independent just as SA can,and phase alone; should not be limited to the design kind of approach should, be undertakenat different stagesof the project cycle, SE (stakeholder engagement) shouldalsopermeateprojectpractice.Both formaland informalmeans involvementshouldbe employedthroughout. of ensuringstakeholder Ownership and sustainability are perhaps the two most important issues to be consideredwhen determiningthe institutional form of projects; they are key considerationsfor institutional interventionat any level and are inextricably intertwined.To a largeextent,the natureofan institution'sownershipwill determine its capacityto persistover time, so it would be a mistaketo plan for sustainability without payingcarefulattentionto ownership.'Ownership'is not a simplevariable in this contextasmanydifferentinstitutionsmay havea stakein the project. Unless thereare goodreasonsfor treatingparticularprojectsas one-offin-out interventions, the emphasisshouldalwaysbe on giving local institutionsas large a sharein the ownershipof projectsas possible;this will meangiving them control over project resowces,not just 'paper' shares. Potentialownersor part-ownersmight include government institutions, NGOs and community-basedorganizations. In some contexts,it may be appropriateto designprojectsin sucha way that they subsequently 'melt into' an existinginstitutionalframework,disappearing as local institutionsin local institutionsonce their own right. Altematively,theymayremainasindependent replaced. expaftiatestaffhavebeen l0

Much the samecan be saidaboutthe institutionswhich projectsthemselvescreate. Mosseet al. (1995)discussthe needto identirythe socialconditionsfor sustained participationwithin rural development projects.Theysuggesttwo linesof attack: to build into projectsthe ability to analyseand interpretproblems,needsand priorities as social constructs;and to identifu appropriatesocial contexts(local groups)for planningandsustainability.SA andSE havethe potentialto fulfil muchofthis work with a more explicitly institutional focus. The refinement of SA, and the developmentof methodsfor ensuringSE, also have potentiallykey roles in the understanding and engagingof local knowledgeand practicein the wider sense. Although SA and SE are, in effect, elementsof an actor-orientedapproachto understandingandengaging,they arenot suflicient for all our purposes.SA is not the sameas structual analysis,thoughthey sharethe sameacronjryn. Institutiondl actors If projectsareto functioneffectivelyaslocal institutions,local stakeholders areto be enabledto exercisetheir different stakesin the project, and projects are to aim for sustainability in the medium-to long-term,carefulattentionmustbe paid to the roles, skills,andtrainingofproject practitioners andotherindividualactors. The institutional culture of donor-funded projects typically revolves around dichotomiesof 'us andthem', insidersandoutsiders,operatingat differentlevels. At one level, 'we' are the expatriatetechnicalstaff and 'they' are our local (usually govemment-employed)counterparts;at anotherlevel, 'we' are all the practitioners directlyattachedto the projectand 'they' areotherministryemployees, includingthe superiorsof our counterparts;at yet another level, 'we' are all the donor-funded projectsand 'they' arethe hostgovernment.Perceptions of this kind, while in some ways unavoidable,can also be highly counter-productive, especiallywhen they are translatedinto project practice and affect the ways in which different institutional actorsare engaged.Similar dichotomiesmay operatewithin the wider institutional culture,for example,within the ministryor department to whicha projectis attached, but this is no reasonfor reproducing them. As an aspectof local practice,including our own, perceptions suchastheseshouldfirst be understood andthen engaged(and possible. wherenecessary, countered)in asconstructive way as a

4.2 Disciplinarity, interdisciplinarityand participationwithin ZCCFSP In its last year (1995-96),ZCCFSPhad an unusuallylarge expatriateteam of four TCOsand two APOs. Theseincludedthe project'sField Manager(an economist), (Unguja),a three agronomists(two on Ungujaand one on Pemba),an agroforester (Pemba).To someextent, marketingeconomistGlnguja)anda socialanthropologist partly becausethey were basedon differentislandsand, in they workedseparately, two cases,becausethey had close links with other institutions (the agroforester workedwith staffin the Sub-Commission for F while themarketingeconomist ll

was attachedto the Ministry of Trade,Industryand Marketing). In spite of this, there was a remarkabledegreeof collaboration,and many researchexerciseswere undertakenjointly (the social anthropologist, in particular,commutedbetweenthe two islands,asdid his colleagues, althoughlessoften). Most of the expatriateteam'sZanzibaricounterparts wereMALNR employees.The majorityhadreceivedbasictrainingin agriculturalscience(with someexceptions, to certificateand diplomalevel). On both islands,the expatriateand nationalproject staff were allocatedto threesections:agronomy,post-harvest, and socio-economics (a fourth section,agroforestry,was confinedto Unguja). This division was often counter-productive and madeit difficult to co-ordinatethe work. On Pemba,the three sectionswere effectively integratedinto one for the purposesof most research exercises.On Unguja,however,attemptsto achievea similardegreeof integrationin everydayactivitiesfailed,partlydueto resistance andantipathyamongstsomeofthe personnelinvolved. For a long time, decision-making was dominatedby the expatriateteam meeting aloneon Unguja. Major decisionswerethen presented to Zanzibai counterparts as fait accompli. However,in the last yearsof the project,somepositivechangeswere made. Monthly projectmeetingswereheld and importantissueswere discussedby all the technicalstaff, includingdecisionsaboutthe allocationof funds for training activities of diflerent kinds (including study tours). The meetingsfrequently heateddebateand werea distinctimprovementon the previouspractice. engendered Towardsthe endof the project,a ZanzibariHeadof Projectwaschosen.Prior to this appointment,the expatriate Field Manager had no effective counterpart and the Zanzibai staff had no effective representative.The Field Managerhad dealt directly with individualstaff at one level. and the AssistantCommissionerfor Researchat another. When ODA funding endedand the expatriatestaff left, the new Head of Proiectremainedin controlof whatwasleft ofZCCFSP.

Projectresources and includepersonnel,skills,and opportunitiesfor trainingas well as financialand materialresources, and thesehaveto be allocatedand controlled. Financialcontrolis alwaysa sensitiveissue,especiallyifhost institutionsareshortof funds and/orhavea reputationfor managingthem poorly (and at worst, comrptly). If the personnelofa hostinstitutioncannotbe trusted,it is arguablethattheyshouldnot (unlessthe projectis designedto promote be chosenasrecipientsof projectassistance greateraccountabilityand transparencywithin the institution). In more regular circumstances, thereareseveralof waysof haadlingthis issueeffectively. Insteadof handingoverall financial control to a singleexpatriateprojectmanager(who may also be financially incompetentor comtpt), a systemof checksand balancescan be introduced;a boardof'directors' or otherpersonnelcouldbe createdto reviewmajor decisionson the allocationof projectresources.Iffinancial andotherresponsibilities canbe safelyhandedto the hostinstitution,it maybe appropriate to appointa project managerfrom within the institution and leave the expatriate staff with primarily technicalandadvisoryroles. t2

Whenemployingeitherexpatriateor nationalpersonnel, carefulconsideration should given be to the skills requiredby the project. Althoughmultidisciplinaryteamsare often preferred,they may not be entirely satisfactoryif the expatriatescome from a rangeof disciplinesbut their nationalcounterpartsare mainly mono-disciplinary. Multidisciplinarityin itself doesnot necessarily promoteinder-disciplinarity; thereis oftentoo muchspecialization within teamsand an over-emphasis on technicalrather than institutional skills. Technical skills are sometimesbest acquiredthrough consultancy; whenplanningprojects,seriousthoughtshouldbe givento creatingan effectivebalancebetweenlong-termappointmentsand consultancies.Either role may be given to nationals,and in somecontextsthe emplolnnentof local staff and (for instance,in termsof familiaritywith the consultants canbringdistinctadvantages project area, local language(s), and the institutionsinvolved). The formation of mixed teamsof expatriatesand nationals,working as equals,is an obviousway of overcomingthe 'us andthem' syndromereferredto above. Trainingstrategies shouldaim to improvethe skills availableto projectsboth in the shortand long-term. This shouldincludetacklingthe 'normal professionalism' and 'conservatism professionals' of development highlightedby Blaikie et al. (1996)by creatingmore effectiveblendsof naturalscienceand social science,technicaland institutionalskills; and this appliesto both nationaland expatriatestaff. Natural resourceprojects are frequentlyoverloadedwith expatriatesand nationalswith primarilytechnicaltraining. In orderto understand and engagelocal knowledgeand practiceat all levels,a varietyof non-technical skills is alsorequired.Theseinclude the ability to analyseinstitutions,and interactand negotiatewith other institutional actors from olfice to field level. This does not mean that every development practitionershouldbecomea socialscientist,but the importanceoftheseskills should be recognized andstepstakento ensurethattheyareexercisedmorewidely.

Implicationsfor policy Throughoutthis book,we haveemphasized the needto understand and engagelocal knowledgeand practicerathermorerigorouslythanhasnormallybeenthe case,and we havesuggested someof the ways in which this might be achieved. However, questionsof policy alsoneedto be addressed.We will beginby consideringcurrent DFID strategy,andthen discussmorespecificstrategies relatingto the allocationof resowcesfor naturalresources researchanddevelooment. A strategtfor lhefuture'l ODA,{DFID'sRenewable NaturalResources ResearchStrategy(RNRRS),which was introducedin 1990and revisedin 1993-94,is unashamedly technicaland explicitly basedon a 'productionsystemapproach'. The revisedRNRRS for 1995-2005is introducedasfollows: "The revisedStrategyaimsto generatereplicablenew

technologiesand improvedknowledgein natural, and relevantsocial, sciencesthrough a more rigorous scientific approachto problem solving. It will IJ

promote the uptake and application of researchproducts for the removal of constraintsto the sustainabledevelopmentand managementof renewablenatural resources ifl tropical developing countries. It will focusmore on demandcriteriato improvethe quality, relevance,uptakeand impactof (ODA 1994a:2; seealso1994b:vii) research."

The technical orientation of the RNRSS is evident in its definition of 'commodity/resource productionsystems'(research will focuson sevenof these)and reflected in its organizationinto 12 programmes,the majority of which are discipline-based(for example, Aquaculture Research,Fisheries Management Research, FishGeneticsResearch andFishPost-Harvest Research).TheEnvironment ResearchProgramme,which is administeredseparatelyfrom the RNRSS but is designedto complementit, is similarin its conception. The emphasison'demandcriteria'offers a little hope,but not much. Blaikie et al. cautiouslystated: "With the introduction of the revisedRNRRS it is likely that LK ocal knowledgel (together wirh 'participation') will featuremore strongly in project frameworks and in annual/ completionreports. However,thereis a dangerthat it will do so morein letter than in spirit, andthat a gap betweenrhetoric and practicewill develop. That is that researchagendasmay continueto be prirnarily set by Western scientificpractic€,with a cursorynod towardsindigenousmanagement strategiesandLK. The reasons for this haveto do with the difficulties of managingthe researchinterfacewhich, in the neo-populist paradigm,becomeconsiderablymore complex. It needsto be ernphasised that the implicationsof an explicit LK componentfor the researchprocessare considerable,and involve a profound shift in the social and political relationshipsbetweenNR lnatural resources]researchscientists,extemal agencies and their clients, and in the ways in which these structure themselves,make decisionsand relate to othersin the professional sphere."(1996:24-25)

This criticism is valid, but limited by the focus on local knowledge in its conventional, restrictedsense.Why is the RNRRSonlyaresearchstrategy?Thereis nothingwrongwith researchas such(andnothingintrinsicallywrong with 'Western scientificpractice',at leastin its lessdogmaticmanifestations), but research(either ours or theirs) must be put into practice. Natural resourcesdevelopmentand managementappearto have been overlooked. If the RNRRS actually does guide DFID practiceuntil 2005, the short-termoutlook is grim. This is a criticism, however,only of the letter (and implied spirit) of the text. DFID is a complex organizationwith manydifferentstakeholders, includinginstitutionalactorswho are well awareof theseproblemsand capableof addressing them. The proof will be in their practice. Multiple strategies The vast and growing literature in favour of participatory approachesto natural resourcesresearchand developmentpromptsthe assumptionthat they have been endorsedwith the same fervour by policy-makers,and that resourcesare freely availablefor their use. However,just as the participatorymessage hasyet to trickle down to somepractitioners,it is encounteringsimilar, if not greater,difficulty in moving upwardsto the policy-makers. While many practitionershave at least embracedthe rhetoricof participation,evidencesuggeststhat many policy-makers t4

havenot, eitherbecausethey haveactivelyresisteddoing so, or becauseare largely ignorantofthe partici patorymovement. Thereare severalreasonsfor this, someof which relateto the political economyof development.Many policy-makersare adherents(declaredor not) of neo-classical economics (the source of traditional modemization approaches)whereas many developmenttheoristsand practitionershavebeenmore stronglyinfluencedby the radical alternatives(including neo-Marxism,feminism, and the environmental movement)underlyingparticipatorypopulism. These are the ingredientsof an age-old ideological conllict which, although it rarely surfaces,exists beneath developmentpolicy and practice. It is not difficult to seewhy resourcesmight be deniedto the participatorylobbyin this context. The questionswhich this raisesarerarelyaddressed by the optimisticproponentsof participation.In his critical rcviewof BeyondFarmerFirst (ScoonesandThompson, 1994), Fanington (1995) notes that "the book is permeatedwith simplistic perceptions ofthe conditionsin whichpublicsectorresources maybe allocatedto (or deniedto) the rural poor." Referringspecificallyto less developedcountries,he arguesthat participatoryapproachesrequiremore resourcesthan are usually available in publicsectorresearch ald extensionservices: "We cannotsimply assumethat participatoryapproaches are self-evidentlydeservingof greaterresource allocations: existingpolitical pressuresare likely to favour continuedstrongrepresentationof better-off farmers. By what political and managementprocessesmight we expect some reversal of cuneni resourceallocationpatterns?Thisis a qu€stionthatthebookdoesnot adequately (1995:6) address."

Farringtonmight equally have been referring to the allocation of resourcesin developedcountries,althoughthe political pressuresagainstparticipationtake a somewhatdifferent form. He continuesby arguing that the editors' definition of 'transferof technology'nowhererises abovethe level of caricature,but is conflated with a topdown technologydevelopmentandextensionapproach: "The time is ripe for redefinitionand somerehabilitation ofthe conceptof technologytransfer: it does not necessarilyimply purely "science-driven"modesof operation(though it does admit the important possibility that ideasfor researchand for new technologiesmay come from both scienceand from farmers). Even if technologieshave been developedby profoundly participatory processes,there remainsthe needto offer themas optionsto otherswho mayfind themuseful,regardlessofwhether this "extension"or "technologltransfer".(1995:6) processis termed"sharing","dissemination",

Farringtonsuggestsrelying insteadon the use of the media (including radio) for middle and high incomefarmers,therebysavingon resourceswhich could then be allocated to lower income farmers for whom face-to-face approachesare more necessary. He also suggestsa need to explore altemativesto farming. Farm households differ widely in their capabilitiesandlevelsof commitmentto agriculture; typically,poor farmerscombinestrategies (for to meethouseholdfood requirements example,by sellingtheir labourto othersin additionto working their own farms). Farringtonconcludesby endorsingCarroll's (1992) argumentthat development plannersshouldalsoadoptmultiplestrategies by promotinga combinationof income t)

generation(farming included), the creation of casual, unskilled employment opportunities, and'safetynets'. Farringtonis arguingrealisticallythat resources shouldbe bettertargetedif we want to meetthe needsof resource-poor farmers.Theseissuesareusuallyglossedover by proponents of participatoryapproaches and in this respectwe agreewith Farrington; an enhanced understanding of practice,includingthe practiceof policy-makers, both demandsandallowsusto engagepracticesfategically.ln somecases,this may mean using methodswhich are not participatoryin the fullest sense(Biggs' 'collegiate' mode),but our ultimateobjectivesare the sameas, or very similar to, thoseof the populists. As Edwards (1994) noted, researchdoes not have to be directly participatoryto be relevant.Similarly,we areunlikelyto achieveour objectivesifwe only engagethe practiceof (or 'participatewith') the peoplewe want to help most. The knowledgeand practiceof our own policy-makers might be an obviousplaceto begin.

Conclusion There are many ways in which project desigrrand practicecan be improvedto enhancethe capacity of projects to understandand engagelocal knowledge and practice in their widest sense. There is also a considerableneed to consider alternativesto the normal 'project package', with its restrictive format and time-frame. Why the whole package? Is there scope for specific institutional interventions,longer-termbut more focused interventions,the use of rolling consultancies, closercollaborationwith NGOs? Theseare questionswhich do not apply to natural resourcesresearchand developmentalone,but must be consideredif local knowledgeandpracticeareto be engagedmoreeffectivelythanat present.For the samereasons, thereis an evidentneedto examinecriticallythe policy framework in whichnaturalresources research anddevelopment areembedded.

l6

CHAPTERFIVE

CONCLUSION

The messageof this book is simple: understanding and engaginglocal knowledge andpracticearecriticalto naturalresources researchanddevelopment.Althoughthis is not a new idea,we havegivenit a muchwider interpretation by emphasizing that understanding andengagingshouldnot be confinedto a singlesl,nchronicdomain(of primaryprojectbeneficiaries), but ratherextendedovertime andto differentlevelsof institutionalspace. More importantly,however,we haveemphasizedthe needto understandand engage knowledgeand practiceat different levelswith much greaterthoroughness than is case. The methods generally usuallythe andtools recommended for this taskarenot entirelysuitable,or appliedaswell (or ascreatively)astheymightbe; thoughtless use pRl{'package' good of the standard is a exampleof this. However,altemativesare available,includingdevelopingthe potentialof more'traditional'methodsof inquiry. Understandingis the key to effectiveengagement, andthis appliesto engagingproject practiceaswell asto planninginteractions at communitylevel. Before ending, let us recap the sequenceof our argument. Conventional understandingsof knowledgeand practice are strongly coloured by the 'indigenous knowledgedebate'and the way in which this has developed.This debatecan be characterized by the (incomplete)shift from a 'naive' to a 'sophisticated'conception of indigenousknowledge'(IK). The naiveconceptionis overwhelminglytechnical, while the sophisticated conceptionemphasizes the social (political, economicetc.) dimensionsof IK. The latteris no doubtan improvementoverthe former,althoughit still carrieswith it a narrow focus on 'knowledge'which has led to impractical proposalsfor practice. Knowledgeis more satisfactorilytreatedas an aspectof practice,and more effort should be put into understandinglocal practice in its different manifestations. Differentwaysof doingthis aresuggested by recentsocialdevelopment theorists,and we examined the relative usefulnessof these, with particular reference to 'structuralist'and'actor-oriented' After emphasizing approaches. the needto analyse how practicechangesover time, we tumed our attentionto the meaningof'local'. Conventionalapproaches to IK (andpractice)focusalmostexclusivelyon one set of actorsin the developmentarena,membersof the (usuallyrural) 'community' By contrast,we expanded the definitionoflocal to includethe knowledgeandpracticeof otheragents,includinggovernment, NGOs,andevenprojectsthemselves. We thenconsidered how local knowledgeandpracticecanbe engagedat community level. Different methodsand approacheswere examined,starting with the standard 'participatory'packages, includingPRAs. Thesewerefoundwantingin a numberof respects,especiallywhen used(as they usuallyare) in an inflexible and uncreative

way. They often provideonly a superficialunderstanding of local knowledgeand practice,and an insufficientlydevelopedmeansof engagingthese. Alternative approaches whichcanmakeup thesedeficiencieswereconsidered, includingthe case study approachand the importanceof building upon local practice (including innovationandsuccess) asa meansof engagingit. The needto usedifferentmethods andapproaches strategically andagainsta backgroundofa coherentunderstanding of local practicewasalsoemphasized. Finally, we retumedto the wider definition of local knowledgeand practice,and looked at the questionof engagingprojectpractice. We beganwith an extended discussionof the 'extemalinterventiondebate',retumingto someearliertheoretical themesandarguingthat interventionhasto be constructively engaged andnot naively rejected. Different proposalsfor incorporatinglocal knowledgeinto project practice wereconsidered, focusinguponlocal knowledgeas conventionallyunderstood.We then examinedprojectsas local institutions,and lookedat the differentimplications of this in termsof variousinstitutionaloptionsand the ways in which the skills of project actors might be improved,enablingthem to understandand engagethe practiceof otheractors(aswell astheir own) moreeffectively.We concludedwith a brief look at someof the wider implicationsof our argumentfor policiesrelatingto naturalresources research anddevelopment. To reiterate:the knowledgeand practiceof all the participants(individualactorsand institutions)in the processof natural resourcesresearchand developmentare important.Theprimarytaskof projectandotherdevelopment practitionersshouldbe and engagelocal knowledgeand practice(using 'local' in its widest to understand sense)ascomprehensively andcreativelyaspossible.Thereareno shortcutsto doing this. However,the potential advantagesof a comprehensiveunderstandingof local knowledgeand practiceare considerable;not least in fosteringa more creative approachto engagement with the promiseof more equitableand thereforeeffective i nteractions.

BIBLIOGRAPITY

The following bibliographyis divided into two parts: (1) works relatingto our main text and the generalissuesdiscussedin it, and (2) works dealing specificallywith Zavibar and the project experienceswe have used to illustrate the text. We have includedin it a numberof works which arenot cited directly in the t€xt, but which we have found useful in the course of preparingand writing it. It does not pretend, howwer, to be a comprehensive bibliography,either of the mainthemeswe discussor of the political economyandhistory of dwelopmentinterventionsn Zmzibar.

I. GENERAL (1995) 'The Scopeand Aims of Etbnobotanyin a Developing Alcom, J. B. World', in R. E. SchultesandS. von Reis(eds.)(1995). 23-29. Altieri, M. A. (1995) Agroecologt: Ihe Scienceof SustainableAgriculture (xcond edition). London: IntermediateTechnologyPublications. Appleton,H. E. and l{ll, C. L. M. (1994) 'Genderand IndigenousKnowledgein Various Organisations',IndigenousKntwledge and Developmefi Monitor, 2 (3),8-l l. Bebbingto4 A. (1991) 'Indigenous Agricultural Knowledge Systems,Human Interestsand Critical Analysis:Reflectionson FarmerOrganizationin Ecuador', Agricalture and Human Values;8, 14-24. Bebbington, A. (1993) Rural People's Knawledge,Farmer Organisations and Regiotnl Development: Implicationsfor Agricaltural Researchand Extension, ODI AgriculturalResearchandExtensionNetwork PaperNo.41. 'Theory and Relevancein Indigenous (1994) Bebbingto4 A. Agriculture: Knowledge,AgencyandOrganisation', in D. Booth(ed.)(1994). 202-225. Bebbington,A., Merrill-Sands,D. andFarrington,J. (1994) Farmer and Community Organisafionsin Agricultural Researchand Extension: Functions,Impactsand Questions,ODI AgriculturalResearchandExtensionNetwork PaperNo.47. Bentley, J. W. (1994) 'Facts, Fantasiesand Failures of Farmer Participatory Research, lgriculture andHumanValues,11, 140-150. Biggs S. (1989) Resource-poorFarmer Pafttcirytton in Research: A Sythesis of Experiencesfrom Nine Natiornl Agric,trltural fusearch Systems,ISNAR / OFCORComparative StudyPaperNo.3.

Biggs, S. D. (1990) 'A Multiple Sourceof InnovationModel of Agricultural Research andTechnology Promotion',WorldDevelopment, 18(ll\, l48l-1499. Biggs, S. D. and Fanington,J. (1991) 'Assessing the effectsof FarmingSystems Research: Time for the Reintroduction of a Potitical and Institutional Perspective',Joumal of AsianFarmingSystems Association,l, I 13-131. Biot, Y., Blaikie, P. M., Jacksoq C. and Palmer-Jones, R. (1995) Rethinking Research on land Degradation in Developing Countries, World Bank Discussion PaperNo.289.,Washington D.C. Blaikie, P., Brown, K., Stocking,M, Tang,L., Sillitoe,P. and Dixon, P. (1996) 'UnderstandingLocal Knowledge and the Dynamicsof Technical Changein DevelopingCountries', paper presentedto the Socio-economicMethodologies Workshop(ODANatural ResourcesSysternsProgramme),29-30 April. Booth, D. (1994a) 'RethinkingSocialDevelopment:An Overview',in D. Booth (ed.)(lee$. 3-34. Booth,D. (1994b) 'How far BeyondtheImpasse?A ProvisionalSumming-up', in D. Booth(ed.)(1994).298-311. Booth, D. (ed) (1994,) Rethinking Social Development: Theory,Researchand Practice. Bumt Mill, Harlow: LongrnanScientific& Technical. Brokensha,D , Warre4 D. and Wemer, O, (eds.) (1980) IndigenousKrutwledge Systemsand Development.Lardaalrl.:UniversityPressof America. Bunclq R. md Lopez, G. (1995) Soil Recuperationin Central Amerim: Sustaining Innovationafter Intervention, IIED GatekeeperSeriesNo.55. Carew-Reid, J., Prescott-Allen, R., Bass,S. andDalal-Clayon,B. (1994) Strategies for Nafional Susainable Development: A Handbookfor their Plarning and Implementation London: EarthscanPublications. Carney, J. (1991) 'Indigenous Soil and Water Managementin Senegambian Rice FarmingSystefirs', Agriculture and Human Values,8, 37-48. Canoll, T. (1992) Intermediary NGOs: The Supporting Link in Grassroots Development.WestHartford, Connecticut:KumarianPress. Carter, L (1995) Alley Farming: Have Resutrce-poorFmmers BeneJited?,ODI NaturalResourcePerspectives, No.3. Carter,J. (1996) RecentAppr
Cerne4 M. (ed.) (1985) Putting People First: Sociologiml Variables in DevelopmentProjectJ. Baltimore: The JohnHopkinsPress. Chambers,R. (1983) Rural Development: Putting the Lost First. Longmans.

London:

Chambers, R. (1993) Challengtng the Professions: Frontiers for Rural Development.London: IntermediateTechnologyPublications. Chambers, R, (l99aa) 'The Originsand Practiceof ParticipatoryRural Appraisal', lf orld Development,22, 953-969, Chambers, R. (1994b) 'Participatory Rural Appraisal (PRA): Experience',WorldDevelopment,22, 1253-1268.

Analysis of

Chambers,R. (1994c) 'ParticipatoryRural Appraisal@RA): Challenges,Potentials andParadigm',World Development,22, 1437- | 454. Chambers,R. (1995) Poverty and Livelihoods: WhoseReality CountsT, IDS Discussion PaperNo.D347. Chambers, R., Pacey,A., and Thrupp,L. A, (eds.) (1989) Farmer First: Farmer Innovation and Agricultural Research. London: IntermediateTechnology Publications. Colchester, M (1995) 'SalvagingNature: IndigenousPeoples,ProtectedAreasand BiodiversityConservation',UNNSD SocialDevelopmentNews, 12,14. Comwall, A., Guijit, I. and Welboum, A. (1993) AcknowledgtngProcess: Challengesfor Agria tural Reseqrch and Extension Methdologt, IDS Discussion PaperNo.D333. Cornwall, A. and Jewkes,R. (1995) 'What is ParticipatoryResearch?',.Socjal Scienceand Medicine, 4l , 1667-1676. Critchley,W., Reij, C. and Wilcocks,T. J. (1994) 'IndigenousSoil and Water Conservation:A Review of the Stateof KnowledgeandProspectsfor Building on Tradition', Iand Degradationand Rehabilitation,5,293-314. Davies,S. (1994) 'Introduction: Informatio4 KnowledgeandPower', IDSBulletin, 25 (2), 1-r3. Davis,E. W. (1995) 'Ethnobotany:An Old Practice,A New Discipline',in R. E. Schultes andS. von Reis(eds.)(1995). 40-51. Dei, G. J. S. (1988) 'Crisis and Adaptationin a GhanaianForest Community', AnthropologicalQumterly, 6l (2), 63-72.

Dore, R. (1994) 'Why Visiting SociologistsFatl', World Development,22 (9), 1425-1436 Drinkwater,M. (ed.) (1994) Former Participafion and Farmer ResearchGroups: Worlahop Commentary and Papers, proceedings of an ARPT National Workshopheld at Kabwe,Zambia,1-6February1993. Drinkwater, M. and Sutherland,A. (1994) 'Evaluationof FarmerParticipationin ARPT', in M. Drinkwater(ed.)(199\. 2-59 Edwards, M. (1989) 'The Irrelevance of Dwelopment Studies', Third World Quarterly,1l (l). Edwards,M. (1994) 'RethinkingSocialDevelopment:The Searchfor 'Relevance", in D. Booth(ed.)(1994). 279-297. Eyben,R. (1991)'TheProcessApproach',paperpresentedtotheNaturalResrrurces AdvisorsConference,Bangor,July 1991. Fairhead, J. (1991) Indigenous Technical Knan'ledge ond Natural Resource Managementin Sub-SalwranAfrica: A Critical Reiew. Chatham:Natural ResourcesInstitute. Farringto4 J. (1994) Public Sector Agricahural Extension: Is There Ltfe After Strilctural Adjustmezt?,ODI NaturalResourcePerspectives, No.2. Farringto4 J. (1995) Review of 'Beyond Farmer First', Tropical Agriculnre AssociatronNewslelter,15(2), 5-6. Farrington,J. (1996) Socio-economic Methds in Natural ResourcesResemch,ODI Natural ResourcePerspectives No.9. Farrington,J. andMartin, A. (1988) Farmer Participation in Agricalnral Research: A Reviewof Conceptsmtd Practices,ODI OccasionalPaperNo.9. Fernandez, M. E. (1994) 'Gender and lndigenous Knowledge', Indigenous Knantledgeand DevelopmentMonitor,2 (3), 6-7. Fernandez,P. (1995) 'Indigenous Seed Practices for SustainableAgriculture', IndigenousKnowledgeand DevelopmentMonitor,3 (1). Feyerabend,P. (1975) Against Method: Outline of an Anarchistic Theory of Krutwledge London: Verso. Fisher, R. J (1995) 'Local Knowledge of Dryland Salinity in the Hunter Valley, Australia',IndigenousKntwledge and DevelopmentMonitor, 3 (l), 13-14. Frazer,J, G. (1923-26) TheGoldenBough (third edition). London. 4

Freire, P. (1996) Pedagogt of the Oppressed(trans. M. B. Ramos). London: PenguinBooks. Gibson, C. G. and Marks, S. A. (1995) 'TransformingRural Hunters into Conservationists:An Assessmentof Community-basedWildlife Managernent Programsin Africa', WorldDevelopment, 23 (6),941-957. Grandin, B. E. (1988) Wealth Ranking in Smallholder Communities: A Field Mmtuol. London: IntermediateTechnologyPublications. Grimble,R., Chan,M-K, fulionby, J. and Quan,J. (1995) Treesand Tradenfs' A StakeholderApproach to Naturql ResourceManagement,IIED Gatekeeper Series-No.52. Hasler,R. (1995) Political Ecologiesof Scale: TheMulti-tiered Co-management of ZimbabweanWildlife Resources,IIED (in associationwith The CAMPFIRE CollaborativeGroup) Wildlife andDevelopmentSeries,No.7. Hassaneiq N. and Kloppenburg, J. (1995) 'Where the Grass Grows Again: Knowledge Exchange in the SustainableAgriculture Movenent', Rural Sociologt,60 (4), 72l -740. Hobley, M. and Shah, K. (1996) What Makes a Local Organisation Robust? Evidencefrom India and Nepal, ODI NaturalResourcePerspectives, No. I 1. Howes, M. and Chambers,R (1979) 'IndigenousTechnicalKnowledge: Analysis, Implicationsand lssrues' , IDS Bulletin, lO, 5-11. Hulme,D. (1995) 'Projects,Politicsand Professionals: AltemativeApproachesfor ProjectldentificationandProjectPlann;ng', Agricultural Systems,47,2ll-233. IDS Workshop (1989) 'Farmer'sKnowledge,InnovationsandRelationto Science', in R. Charnbers et al. (eds.)(1989),3l-38. IIED (1994) ll/hose Men? An Ovemiewo/ Comrrunity Apprmches to Wildlife (a report to ODA). London: IIED. Mcmagement ISNAR (1994) Report of a Workshop: Sttengtheningthe Role of Farmers' Organizations in Technologt Developmentand Transfer, ISNAR Briefing Paper,No.15. Jacksoq C. (1993) 'Doing What ComesNaturally? Women and Environmentin Dwelopment', lYorld Development,2l (12), 1947-1963. Kendon, G., Walker, D. H., Robertso4 D., Haegitb M., Sinclair, F. L. and Muetzelfeldt,R. L (1995) 'Supporting CustomisedReasoningin the AgroforestryDomain', TheNewReviewof ApphedExpen Systems,179-192.

Kiome, R. M. and Stocking,M (1995) 'Rationalityof FarmerPerceptionof Soil Erosion', Global EnvironmentalChange,5 (4), 281-295 Lakatos,I. andMusgrave,A. (eds.) (1970) Criticism ond the Growth of Knowledge. London: CambridgeUniversityPress. Lindskog, P. and Tengberg,A. O99q 'Land Degradatio4 Natural Resourcesand Local Knowledge in the SahelZone of Burkina Faso', GeoJoumal, 33 (4), 365-375. Long, A. (1992) 'Goods,KnowledgeandBeer: TheMethodological Significance of SituationalAnalysisand Dscourse',in N, Long and A. Long (eds.)(1992). 147-170 Long,N. andvan derPloeg,I. D. (1989) 'Demythologizing PlannedIntervention:An Actor Perspective',SociologiaRuralis, 29 (3/4), 226-249. Long, N. and van der Ploeg, J. D. (1994) 'Heterogeneity,Actor and Structure: Towards a Reconstitutionof the Concept of Structure', in D. Booth (ed,) (1994). 62-8e. Long, N. and Long, A. (eds.) (1992) Battlefieldsof Knawledge: TheInterlockng of Theoryand Practice in Social Researchand Developmenl. London and New York: Routledge. Longhurst,R. (1988) 'CashCropsandFood Security:CashCrops,HouseholdFood SecurityandNutition' , IDS Bulletin, 19(2),28-36. Mathias, E. (1995) 'Frameworkfor Enhancingthe Use of lndigenousKnowledge', IndigenousKna+,ledgeand DevelopmentMonitor, 3 (2), 17-18. Maundu,P. (1995) 'Methodologyfor Collectingand SharingIndigenousKnowledge: A Ca:c Shtdy', IndigenousKrcwledge and DevelopmentMonitor, 3 (2), 3-5. Mazur, R.E. and Titilola, T. S. (1992) 'The Social-economic Dimensionsof Local Knowledge Systemsin African SustainableDwelopment', SociologtaRuralis, 32 (2t3),264-286. Merrill-Sands,D. and Collioq M-H. (1992) 'Making the Farmers'Voice Count; Iszuesand Opportunitiesfor Promoting Farmer-Responsive Research',paper presentedto the |2th Anmml Farming SystemsSryposiazr, Michigan State University,13-18 September. Mosse,D. (1994) 'Authority, GenderandKnowledge:TheoreticalReflectionson the Practice of Participatory Rural Appraisal', Developmentand Change, 25, 497-526.

Mosse,D., with the KRIPB ProjectTeam (1995) 'People'sKnowledge'in Project Planning: The Limits and Social Conditiorrs of Porticirytion in Planning Agricultural Development,ODI Agricultural Researchand ExtensionNetwork PaperNo.58. Mosse, D., with the KRIBP Project Team (1996) Local Institutions and Farming SystemsDevelopment: T1roughts from a Project in Tribal WestemIndia, ODI AgriculturalResearchandExtensionNetworlg PaperNo,64. Nabasa,J,, Rutwara, G., Walker, F. and Were, C. (1995) Particirytory Rural Apprarsal: Practical kperiences. Chatham:NaturalResourcesInstitute. Ndufa, J. K., Ohlsson,E., Swinkels,R. and Shepherd, K. D. (1993) 'Participatory Research Methods for Agroforestry Technology Development in Western Kenya', paper presentedto the workshop on Design, Implementation and Anolysisof On-farm Trials, Arlsha,28 June- 2 July. Nelson, N. and Wright, S. (eds.) (1995) Power and Participatory Development: Theoryand Practice, London: IntermediateTechnologyPublications. Niemeijer, D. (1995) 'Indigenous Soil Classifications:Complicationsand Considerations',Indigenous Knowledge ord DevelopmentMonitor, | (l\, 20-21. ODA

(1992) The Process Approach to Projects, Technical Note No.4, Aid Economicsand SocialDepartment,ODd London.

ODA (1994a) RenewableNatural ResourcesResearchStrategt 1995-2005, final report, NaturalResourcesResearchDepartment,ODd London. ODA (1994b) A Guide to the RenewableNatural ResourcesResearchStrategt. London: ODA. Okali, C., Sumberg,J., and Faningon, J, (1994) Farmer Pmticipatory Research: Rhetoricand Realir!. London: IntermediateTechnologyPublications. Popper,K. (1959) TheLogrc of ScientificDiscovery. London: Routledge& Kegan Paul. Popper, K. (1966) The Open Society and its Enemies (fifth edition, revised). London: Routledge& KeganPaul. Prance,G. T. (1995) 'Ethnobotany Todayandin theFuture',in R. E. Schultesand S. von Reis(eds.)(1995) 60-68. Pretty, J. N. (1995) RegeneratingAgriculture: Policies and Practice for Susnirubility and Self-reliance. London: EarthscanPublications.

Pretty, J. N. and Chambers,R. (1994) 'Towards a Leaming Paradigm:New Professionalism and Institutionsfe1 dgricu!ture', in L ScoonesandJ. Thompson (eds.)(199a). 182-202. Rajasekaran, B., Warreq D. M. and Babu, S. C. (1991) 'IndigenousNatural ResourceManagernentSysternsfor SustainableAgricultural Development: A GlobalPerspective', Journalof InternationalDevelopment,3 (l), 1-15. Reij, C. (1991) IttdigenousSoil and WaterConservafionin Africa,IIED Gatekeeper SeriesNo.27. Richards,P. (1985) IttdigenousAgricultural Revolution London: Hutchinson. Rocheleau,D. (1994) 'Participatory Researchand the Race to Save the Planet: Questions, Critique and Lessons from the Field', Agriaiture and Human Values,ll,4-25. Schultes,R. E. and von Reis, S. (eds.) (1995) The Scopeand Aims of Ethrnbotany in a DevelopingWorld. Lorrdon: DioscoridesPress. Scoones,L, Melny( M. and Pretty, J. N. (1992) TheHiddm Harvest: Wld Foods and Agricultural Systems: A Literature Reviewand Anrntated Bibliography. London: IIED. Scoones,I. andThompsoqJ. (1994) 'Knowledge,Power andAgriculture - Towards a TheoreticalUnderstanding',in L Scoonesand J. Thompson(eds.) (199a). t6-32. Scoones,I. and Thompson,J. (eds.) (1994) BeyondFarmer First: Rural People's Knanledge, Agriculnral Research and Extension Pracfice. London: IntermediateTechnologyPublications. Simpsoq B. M. (1994) 'Genderandthe SocialDitrerentiationof Local Knowledge', IndigenousKnowledgeandDevelopment Monitor,2 (3),2l-23 . Sinclair,F. L., Muazelfeldt, R., Robertsoq D., Haggitb M., Walker, D. H., Kendon, G. and Randell, D. (1995) Formal Representdtionand Use of Indigenms Ecologicnl Krumledge About Agroforestry, final report of ODA Forestry ResearchProgrammeProjectR4731. Sinclair,F. L., Walker,D. H., Joshi,L., Ambrose,B. andThapa,B. (1993) Useof a Based Systems in the Imprwement ol Tree Fdder Resourceson Farmland in the fustem Hills of Nepal: Pilot Phase,report of researchundertakenin ODA Forestry and Agroforestry ResearchStrategy, AdaptiveProjectR5470.

Southem, A. J. (1994) Acquisition of IndigenousEcological Knowledge about Forest Gardens in Kandy District, Sri Innka, unpublishedM.Phil. thesis, Universityof Walesat Bangor. Sumberg,J. and Okali, C. (1988) 'Farmers,On-farmResearchandthe Development of New Technolory',Expl.Agric., 24,333-342. Sweet, C. F. and Weisel,P. F. (1979) 'ProcessversusBlueprint Models for DesigningRural DevelopmentProjects', in G. Honadle and R. Klauss (eds.) Internqtional Development Adninistration: Implementation Analysis for DevelopmentProjecls. New York: Praeger. 127-145. Thapa,B. (1994) Fmmers' Ecological Knowledgeabaut the Matngement and (/se of Fmmland Tree Resourcesin the Mid-hills of fustem lr'epal, unpublished Ph.D.thesis,Universityof Walesat Bangor, Thap4 B. (1995) 'Incorporation of IndigenousKnowledge and Perspectivesin AgroforestryDevelopmett',AgroforestrySystems,30, 249-261. Thapa,8., Sinclair,F. L. and Walker,D. H. (1995) 'Incorporationof Indigenous Knowledgeand Perspectivesin AgroforestryDevelopment(Part 2: Case-study on the knpact of Explicit Representation of Farmers'l(nowledge)', Agroforestry Systems,30,249-261. Thiollay, J. M. (1995) 'The Role of TraditionalAgroforestsin the Conservationof RainForestBird Diversityin Sumatra',Conservation Biologt,2,335-353. Thompso4 J. (1995) 'Participatory Approachesin GovernmentBureaucracies: Facilitating the Processof Institutional Change', World Development,23 (9), 1521-r554. Thompsoq J. (1996) 'Moving the IndigenousKnowledgeDebate Forward?', DevelopmentPolicy Review,14, 105-112. Thompson,J. and Scoones,I. (1994) 'Challengingthe PopulistPerspective: Rural Peoples' Knowledge, furicultural Research, and Extension Practice', Agricahure and Human Values,I l, 58-76. Thrupp, L. A. (1989) 'Legitimizing Local Knowledge: From Displacementto Empowermentfor Third World People', Agricalture and Human Values, 6, t3-24. Uphotr, N. (1992) Local institutionsand Participationfor SustaitableDevelopment, IIED Gatekeeper SeriesNo.3l. [various] (1994) Linking Farmers' Organisationsand Resemchers:Four Case Studies,ODI AgriculturalResearchandExtensionNetwork PaperNo.50.

Walker, D. H. (1994) A Knowledge-basedSystemsApprmch to Agroforestry Research ad ktension, unpublishedPh.D. thesis, University of Wales at Bangor. Walker,D. H., Sinclair,F. L. andKendo4 G. (1995) 'A Knowledge-based Systems Approach to Agroforestry Researchand Extension', AI Applicatiotts, g (3), 6t-72. Walker, D. H., Sinclair,F. L. and Thapa,B. (1995) 'Incorporationof Indigenous Knowledgeand Perspectivesin AgroforestryDevelopment@art l: Rwiew of Methodsandtheir Applications)', AgroforestrySystems,30,235-248. WarrerqD. M., Slikkerveer, L. J. andTitilol4 S. O. (1989) 'IndigenousKnowledge Systems: Implications for Agriculture and International Dwelopment', Teclmologtand Social Change,ll. Waneq D. M., Slikkerveer,L. J. and Brokensh4D. (eds.) (1995) The Cultural Dimension of Development: Indigenous Kn<mledge Systems. London: IntermediateTechnologyPublications. Webber,L. M. and Isoq R. L. (1995) 'ParticipatoryRural AppraisalDesign: ConceptualandProcesslswes' , Agricultural Systems,47, lO7-l3l . WelbourqA (1991) 'RRA andthe Analysisof Difference',RM Notes,14, 14-23. Wells, M. P. (1995) Bidiversity Conserwtion and lncal Peoples' Development Aspirations: New Priorities for the 1990s,ODI Rural DevelopmentForestry NetworkPaperNo.l8a. Young, A. (1989) International.

for Soil Consemation. Wallingford:

CAB

2. ZANZIBAR

FAO (1995) Formulation of a National Forest Policy for knzibar, report to the Govemmentof the United Republicof Tanzania.Rome: FAO. Goldmaq H. V. (1996) A ComparativeStudyof Swahili in TwoRural Communities in Pemba,Zanzibar, Taruania,unpublishedPh.D. thesis,New York University. Kitwana, M., Kombo, Y. H. and llarvey, S. (1996') Wakulina rn Kilimo Mseto hnzibm, Forestry Technical Paper No.32, Sub-commissionfor Forestry, Commissionfor Natural Resources,Ministry of Agriculture, Livestock and Natural Resources.Zanibar.

Koenders,L. feport.

(1992a)

Agriculture in Pemba: Facts and Figures, unpublished

Koenders,L, (1992b) Faurn of PembaIsland: A Checklistof AnimatsOcctrring on Pemba Island, Tanzonia (PublicationNo.l). Dar es Salaam: Wildlife ConservationSocietyof Tanzania. Koenders,L. (1992c) Flora of PembaIsland: A Checklist of Plant Species (PublicationNo.2). Dar es Salaam:Wildlife ConservationSocietyof Tanzania. Kombo, Y. H. (1996) IndigenousTreesfor AgroJorestry:High Potential Speciesfor the Coral Rag of Unguja, ForestryTechnicalPaperNo.29, Sub-commission for Forestry,Commissionfor Natural Resources,Ministry of Agriculture, Livestock andNaturalResources,Zanzibar. Krain, E. (1994) Land Terure in Zanzibar,discussionpaper,NCDP. ODA (1993) Repolt on a Visit to Zanzibar: Mid-term Reviewof Tnnzibar Cash Crops Farming SystemsProject,2g Novernber- 3 December1993. Nairobi: BDDEA. Smitb P. D. (ed ) (1992a) Agricaltural Reseorchand Developmentin hnzibar: An Anolysk of the Literature. Bangor: Centrefor Arid Zone Studies,Universityof North Wales. Smith, P. D. (ed,) (1992b) A Bibliography of Agrictrltural Resemch and Developmentin hnzibar (two volumes). Bangor: Centre for Arid Zone Studies,Universityof North Wales. de Villiers, A. K. (1996) QuanfifuingIndigenousKnowledge: A Rqid Methodfor AssessingCrop Performarrcewithout Field Trials, ODI Agricultural Research andExtensionNetwork PaperNo,66. Williams, A. and Jum4 K. M. (eds.) (1996) Local ManagementPlan Preparation Guidelines(draft), ConservationandVillage ForestrySub-sections,Commission for NaturalResources,Zanzibar. (1949) Williams, R. O. The Useful ond OrnamentalPlonts in Zanzibor and Pemba. Zanibu: GovernmentPrinter.

A full set of ZCCFSPreports is availableat NRI, as well as in the former project library in the Ministry of Agriculture, Livestock and Natural Resources,Forodhani, Zauibar.

ll

About theauthors

Martin Walsh is a social anthropologist in the Social Sciences Departmentof the Natural ResourcesInstitute in the University of Greenwich,and a Visiting ResearchFellow in the Schoolof African and Asian Studiesin the University of Sussex. He has extensivepractical and researchexperiencein rural and urban developmentin Africa, and hasworked in a variety of sectorsasa freelanceconsultant(1985-92),a ResearchFellow in the University of Sussex(1992-94), and as the social anthropologiston the ODA-fundedand NRl-managedZavibar Cash Crops Farming SystemsProject (1994-96). His most recent consultancywork has focused upon biodiversity conservation and communitywildlife management.He holds a M.A. and Ph.D. in social anthopology from the University of Cambridge,and has published widely in differentfields.

Sharon Haney is a freelanceconzultant currently working with a CARE-funded conservation project in Zaruibat. She has wide experienceof agroforestryand farmer participatoryresearch,focusing upon the social aspects of local ecological knowledge in community-based naturalresourcemanagement.Shehasworked in the Pacific(1990-92),South-eastAsia (1988-89),and East Africa (1987, 1993-96),with both NGOs and bilateral agenciesin a variety of local and nationalinstitutionalsettings. Until recentlyshewas carrying out agroforestry research for the ODA-funded Zavibar Cash Crops Farming Systems Project, working in collaboration with the FINNIDA-funded,Zanzibu ForestryDevelopmentProject. Sheholdsa B.Sc. in Agriculture from the University of Westem Sydney (Hawkesbury),Australia,and a M.Sc. in Agriculture,Environmentand Developmentfrom the Universityof East Anglia.