Fuzzy Cognitive Maps Of Local People Impacted By Dam Construction: Their Demands Regarding Resettlement

0.27 MB·English

by  Uygar Ozesmi

Fuzzy Cognitive Maps Of Local People Impacted By Dam
Construction: Their Demands Regarding Resettlement
6
0
0
2 UygarO¨zesmi

n
a
J

1 DepartmentofEnvironmentalEngineering,ErciyesUniversity,38039Kayseri,Turkey,[email protected]
2

]
C Abstract
N
o. Fuzzycognitivemappingwasusedtounderstandthewantsanddesiresoflocalpeoplebeforeresettlement.Vari-
i
b ablesthattheaffectedpeoplethinkwillincreasetheirwelfareduringandafterdamconstructionweredetermined.
-
q Simulationswere done withtheir cumulative social cognitive map to determine which policy options would most
[
increasetheirwelfare. Theconstruction ofroads, jobopportunities, advancepayment ofcondemnation value, and

1
schoolsarecentralvariablesthathadthemosteffectonincreasingpeople’sincomeandwelfare. Thesynergisticef-
v
2 fectsofvariablesdemonstratedthattheimplementationofdifferentpoliciesnotonlyaddcumulativelytothepeople’s
3
0 welfarebutalsohaveanincreasedeffect.
1
0
6
0
INTRODUCTION
/
o
i
b
-
q Halfofthe192damprojects,whichwerefinancedbytheWorldBankbefore1994,hadnoresettlementplan(WCD,
:
v 2000). In 70 percent of the ones that had a plan there was only financial compensation. Of these 75 percent did
i
X
notcomplywiththecriteriasetbytheWorldBankitself. Thelossofproductivecapacitywascompensatedinnone.
r
a Accordingto anassessmentin allof these projectsthesocialand environmentalproblemscouldbe avertedwithout
affectingtheirfinancialviability(WCD,2000).
When resettlement programs for large-scale dam projects do exist, they usually focus on physical settlements
ratherthandirectandindirecteffectsonsocialandeconomicdevelopment. Howeverthesocio-economicconditions
toprovideforthewelfareofthelocalsisatleastasormaybemoreimportantthanphysicalsettlement. Resettlement
severs people’s geographical, social and economic ties and functions. The land, the house, the livelihood, and the
commonsof the localsaffectedby the dam are endangeredor lost. Thisreducesthe socio-culturalresilience of the
people.Theendangermentofeconomicpowerandcomplexsystemsoflivelihoodpreventspeoplefromfulfillingtheir
needs(WCD,2000).
1
Usually the amountof people affected by a dam has always been underestimatedand the condemnationmoney
payments have been low. In most projects only condemnation values or single-payment monetary compensations
weregiven. Allpaymentshaveusuallybeenlate andthe amounthasbeeninsufficient. Forexample,inTurkeyina
studycommissionedbytheWorldCommissiononDamsaboutAslantasDam,thepaymentsmadeforcondemnation
were not sufficient to buy the same amount of land. The same situation was valid in Tarbela dam in Pakistan and
KiambereDaminKenya,too(WCD,2000).Unfortunatelypeoplebenefitingfromdamsarenotthelocalpeoplewho
arenegativelyimpactedbythedam. Especiallythemostimpactedlocalpeoplearethe oneswhodonothavemuch
capital, are poor,andare marginalized. Localpeopleare theoneswhoareaskedlast whethertheywantthe damor
not. The importanceof resettlement policies and their implementation is emphasized, as this is a major factor that
determineswhetherlocalpeoplewillriskresistanceorriskresettlement(Dwivedi,1999).
ThisstudyresultedfromamoreprogressiveapproachthatwastakenbytheTurkishStateHydraulicWorkstoask
thelocalpeopleabouttheirdemandsbeforedamconstruction.Ihopethatthisapproachwillbecomethenorminsuch
largestateprojectsandwillleadtopoliciesthatimplementthesedemandsattheend. A540MWdamistobebuilt
ontheCoruhRiverinnortheasternTurkey. Thedamistobe270mhighandthereservoirareawillbe33km2. The
reservoirwillcompletelyfloodthevillageofYusufeliandpartiallyortotallyflood17surroundingvillages,andaffect
approximately 16,000 people. The average annual household income of these people is approximately4500 USD.
Overhalfofthemobtaintheirincomefromagriculture,whichconsistsofintensivemarket-orientedgardenagriculture
onsmallplots. Alloftheagriculturallandwillbeleftunderwaterasitislocatedinthevalleys(SaharaMuhendislik,
2001). This study deals with the actions needed so that the local people are not negatively impacted or driven to
povertyafterthedamconstruction.Iusedfuzzycognitivemappingtodeterminethedemandsanddesiresofthelocal
peoplewholiveinsettlementsthatwillbedirectlyaffectedbytheYusufeliDam.
WHAT IS FUZZY COGNITIVE MAPPING?
Fuzzycognitivemapping(FCM)isaprocesswheretheparticipant(s)listthefactorsthattheyperceiveasimportantto
aparticularissueandthendrawthecausalrelationshipsamongthesefactors.Insteadofadichotomousyes-no(affects
or doesn’taffect)the fuzzyaspectallowspeopleto weightthecausalrelationships, whichusually cannotbe known
precisely,butaredescribedasoccurringalittle,somewhat,alot,etc. Anyonecanfreelydrawcausalpictures,orcog-
nitivemaps,ofproblemsorsystemsofanykind. FCMisapplicableforanalyzingcomplexproblemsthatinvolvethe
interrelationshipsbetweenvariousaspectsofsocietysuchaseconomics,environmentandpolitics(Nakamuraetal.,
1982). FCMisespeciallyappropriateinsoftknowledgedomains(e.g. politicalscience,militaryscience,history,in-
ternationalrelations,organizationtheory)whereboththesystemconcepts/relationshipsandthemeta-systemlanguage
are fundamentally fuzzy (Kosko, 1986). For example, FCMs have been used to model organization behavior and
jobsatisfaction(Craigeretal.,1996)andtosuggestwaystoimprovedecisionmakinginsport-fisheriesmanagement
(RadomskiandGoeman,1996).
2
A briefhistory ofFCM
Cognitive mapping takes its roots from graph theory put forward by Euler in 1736. Harary et al. (1965) put this
theory into use in quantitative anthropology to make a structural analysis of observations. The structural analyses
weredoneonmapscalled”digraphs”. LaterthepoliticalscientistRobertAxelrod(1976)changedthesegraphsfrom
thesubjectiveinterpretationsofanthropologiststotheobservationsofindividuals.Thesegraphs,whichAxelrodcalled
’cognitivemaps’, were codedfromtextor werefirst-handexperiencesverballystated by informants. Kosko (1986)
openedthewaytoamorerigorousanalysisofcognitivemapsthroughtheuseofneuralnetworkanalysis. Mapscan
be simulated under different scenarios via signal processing which then becomes an auto-associative memory. He
coinedthetermfuzzycognitivemaps(FCM) forthesemapswith weightedconnections. Carley(1988)putforward
someofthestatisticaltechniquessothatstakeholdersanddifferentgroupsofpeoplecouldbecomparedandcontrasted
throughtheircognitivemaps. Cognitivemappinghasbeenusedtoexaminedecisionmakingprocessesandcomplex
socialsystems(Axelrod,1976;Bauer,1975;Bougonetal.,1977;Brown,1992;CarleyandPalmquist,1992;Cossette
and Audet, 1992; Hart1977; Malone, 1975; MontazemiandConrath, 1986; Nakamuraet al., 1982). These studies
openedthewayforusingcognitivemapstorepresentlocalknowledgesystemsastoldbyinformants(O¨zesmi,1999).
Comparisonwithother methods
In this study I use fuzzy cognitive mapping to approach the question of what people impacted from the dam want
andneedduringandafterdamconstruction. Fuzzycognitivemappingcanofferadvantagesoverothersocialscience
research methods such as questionnaire surveys, unstructured interviews, or rapid assessment processes (RAP). A
disadvantage of questionnaire surveys is that the researcher asks questions that they consider important and also
typicallygiveachoiceofanswers. IntheFCMmethodologytheintervieweedetermineswhatisimportantandwhat
shouldbeincluded. Questionnairesurveyshavetheadvantagethatlargenumbersofpeoplecanbesurveyed. Infact,
questionnaire surveys (n = 2522) were employed as one research method in the Yusufeli Dam resettlement plan
(SaharaMuhendislik,2001). Howeverbecauseofananticipatedlowparticipationinwrittensurveys,thesurveyshad
tobecompletedinperson,andthustookaconsiderableamountoftimeonthepartofresearchers.AlthoughtheFCM
technique was done with less people than the questionnaire surveys in this study, it provided more insight into the
desires and needs of the local people (Sahara Muhendislik, 2001). However, I am not advocating that only a FCM
methodshouldbeusedindamresettlementprojects. Detailedinformationneedstobeobtainedfromtheindividuals
that will be affected about their losses so that they can be adequately compensated, and a FCM method would not
be used to obtainthis informationfrom allaffectedindividuals. Rather the FCM becomesa way to obtainpeople’s
opinionaboutthesocio-economicconditionsthatwouldbemosthelpfulforincreasingtheirwelfareduringandafter
resettlement.
Open-endedinterviews could be used to obtain this information, but the researchers must extract the important
3
conceptsandrelationshipsfromtheinterviewtranscripts,whichistime-consumingandmayleadtobiasesandmis-
judgments.IntheFCMtechniquetheimportantconceptsandrelationshipsaredrawnonthemapbytheinterviewees,
thusremovingpotentialresearcherbiasaswellasreducingtheamountoftimespentonanalysisaftertheinterviews.
FCMscanbequicklycodedfromthedrawnmaps. Inadditionresultsfromopen-endedinterviewsarehighlyqualita-
tiveanddonotpermitstatisticalanalysis. WithFCM,mapsofdifferentstakeholdergroupscanbecomparedandthe
results,whilequalitative,doyieldrelativeinformationastowhatismostimportantorwhatthingshavethemosteffect
onasystem. FCMissimilartoRAPinthatitisasystemsapproach.InthecaseofFCM,theintervieweesthemselves
determinethethingsaffectingthesystemandhowtheyrelatetoeachother.Itcanalsobeaniterativeprocess,asmaps
canbeobtainedfromanynumberofintervieweesandthesemapscanbesuperimposedtoformasocialcognitivemap
orthemapscanbesuperimposedaccordingtostakeholdergrouptocomparedifferencesandsimilaritiesamonggroups
ofpeople.ThustheFCMmethodalsohasadvantagesforenablingstakeholderparticipationandanalyzingstakeholder
groupmapsforparticipatorymethods(O¨ andO¨zesmi,2001)andalsoincasesofconflictamongstakeholdergroups
(O¨zesmi,1999).CertainlyFCMcouldbeincorporatedasoneofthesuitabletechniquesforuseinRAPbecauseFCM
is especially appropriatefor informationgatheringwhere the researcherwants to obtainfrom the localpeople what
they think is important. In FCM, the question can be and usually is very general, such as, ”what are the important
factors/variablesaffectingthesystemandhowdothesefactorsaffecteachother?”
METHODS
A standardizedcognitivemappingmethodologywas used in this research to bringto light the wants and desires of
thepopulationimpactedfromthedam. Theanalysisofdatahasbeendonethroughgraphtheoryandfuzzycognitive
mapping,aformofsignalprocessinganalysisusingneuralnetworksimulations(O¨zesmi,1999).
FieldResearch
Interviewsweremadeinsettlementswithinthefourgeographicareasthatwillbefloodedbythedamlake.Theelected
villageheadsandtheirspouseswerecalledbytelephoneoneortwodaysaheadoftheinterview.Anappointmentwas
madetomeetwiththevillageheadandtheirspousesfor1to1.5hourstomakeagraphicalexerciseaboutthedamto
bebuilt(permissionforanappointmentwasnotrefusedbyinformantsexceptforanNGOattheYusufelitownship).A
totalof13cognitivemapsweredrawnwithvillageheadmenandtheirspouses,andanNGOheadandhisspouse.The
totalnumberofpeoplewhoparticipatedinthisresearchwas26people. Theexercisewasdonewithmarriedcouples,
so that participation of both genders could be secured. Women in general made different contributions to maps at
about30 per cent, and have assured usthat they agreewith the contributionsof their husbands. Men have likewise
agreedon the contributionsof women. ThusI believeI have representedthe concernsof the women sufficientlyin
thesemaps.
4
BeforestartingthecognitivemappingexerciseIexplainedthestateofthebuildingofthedamandwhythisresearch
wasbeingmade. ThenIexplainedwhatkindofgraphweweregoingtodrawusingacompletelyunrelatedmapthat
summarizedtheoccurrenceoftrafficaccidents. Onthecornerofalargepieceofpaper(50X30cm)thelocation,date,
time and duration,andthe namesofthe participants, their gender,age andoccupationwere noted(Figure1). Then
thetwoquestionsrelatingtothelistingofvariablesandthedrawingofthegraphwereaskedandwrittenontopofthe
paper.Thesequestionswere:”Whatarethevariables,factors,causesandthingsthataregoingtoincreaseordecrease
thewelfareofthepeopleafterthedamconstructionduringresettlement?”and”Howdothesevariables,factors,causes
andthingsaffecteachother?”Astheintervieweesstatedthesevariables,theywerelistedontheleftofthepaper. We
drewthevariable’peoples’welfare”inthemiddlewiththerestofthevariablesdrawnaroundit. Thenwedrewthe
causalrelationshipsamongthevariables.Thecausalrelationshipsweregivenbytheparticipantsaseitherpositive(+)
ornegative(-)andwereweightedbythemasstrong(1),medium(0.5)orweak(0.25)(Figure1).
Data Analysis
Variableswerelistedandpooledcumulativevariablerichnessplotsweredrawntofindoutwhethervariable(sampling)
saturation was reached or not. For this purpose Monte-Carlo simulations were done using random numbers from
Minitab10.5statisticalpackagetodeterminetheconfidenceintervalofobservedsaturation.
Thecognitivemapsdrawnbyinformantswerecodedasneighborhoodmatrices. Indicescharacterizingthestruc-
ture and type of variables were calculated by cell programming. A macro program written in the spreadsheet was
usedtomakeneuralnetworksimulations. Neuralnetworksimulationisdoneusingmatrixalgebra. Thestatevector
of variables is multiplied with the neighborhoodmatrix and then the results are squashed into the interval of [1,0]
bya logistic function. The resultis thenew vector-state. Themultiplicationis continueduntilthesystem stabilizes
into one variable, falls into an oscillation or goes into chaos. All the simulations ended in a stable state. Policies
canbesimulatedbyclampingthevector-stateofavariabletoadesiredvaluewhiledoingthemultiplicationprocess.
Thesocialcognitivemapisobtainedbymatrixaugmentationinwhicheachparticipantmapisgivenanequalweight.
DetailswithregardstotheabovementionedmethodologycanbefoundinO¨zesmi(1999)oratthiswebsiteaddress:
http://env.erciyes.edu.tr/Kizilirmak/UODissertation.html.
RESULTS
Saturation ofCollectedData
Cognitivemappingexercisesintheareatobeaffectedbythedamtookanaverageof56±15SDminutes(min35,
max 95 min.). The increase in new variables considerablydeclined at the end of 13 cognitive maps and additional
5
newvariablesdecreasedto3pernewmap(Figure2).Sincethesenewadditionswereverymuchpersonalandspecial
issuestothepeopleinterviewedIbelievetheresultsindicatethewantsanddesiresofthemajorityofthepeopleinthe
regionaffectedbythedam.
The Structure andCharacteristic ofCognitiveMaps Obtained
Individualcognitivemaps
The results obtained with the maps from the settlements affected by the dam are comparableto previousstudies in
Turkey(Table1)andotherplaces.Inastudyofasmallbusinessenvironment,57conceptsand87linkswerementioned
bytheowner,whichresultedinaconnectiontovariableratioof1.53andadensityof0.027. Astudywithamodal
averageof75minutesofinterviewingof116informantselicited32conceptsintherangeof14-59. Edenetal. found
typical1.15-1.20ratiosoflinkstonodes. Ifound1.64±0.95SDratioofconnectionstovariablesintheKizilirmak
Delta,whichisalmostthesameasYusufeliwith1.76±0.51SD(Table1).
ThemanytransmittervariablesmentionedbyYusufelipeopleshowthattheyseemanyfactorsbeyondtheircontrol
affectingthem. Theveryfewreceivervariablesotherthan’thewelfareof people’meansthattheydo notsee much
utilityintheexistingsystemsandthattheprospectsarelimited.Theirmapsarenothierarchical.Variablesareaffecting
eachothersometimesinloops. Thenon-hierarchicalmapsarealsoaresultofthehighconnectivityofthemaps. The
highconnectivityisanindicationoftheinterdependentcharacterofthementionedvariables(Table2).
Socialcognitivemaps
Inthisstudythenumberofvariablesandconnectionswereagaincomparabletopreviousstudiessimilartotheindivid-
ualcognitivemaps. CarleyandPalmquistreportedthat29undergraduatesmentionedupto244conceptsonresearch
writing, and 45 studentsproduced217conceptson tutor selection. One individualcouldonly consider between 30
and 40 concepts in a session lasting between 20-40 minutes because of combinatorial explosion. Nakamura et al.
obtained152conceptsand265connectionsfrom5documentsontrafficproblemsinJapan. O¨zesmi(1999)obtained
136variablesand616connectionsfrom31cognitivemapsthatincluded3distinctlydifferentstakeholdergroupsin
perceptionsofawetlandecosysteminTurkey. InthisstudyIfound97variablesand360connectionsin13mapsof
26people(Table1).
VariablesMentioned andTheir Importance
Therearethreedifferentwaysofatlookingatdifferentvariablesmentionedinacognitivemap. Variablescanbeor-
deredfirstbyhowmanytimestheywerementionedineachmap,secondbytheircentrality(includingthecomponents
6
ofhowmuchinputtheyreceive,indergree,andhowmuchoutputtheygive,outdegree),andthirdbythetypeofthese
centralvariables(transmitter,ordinary,receiver). Thevariabletyperevealshowpeoplethinkaboutthevariables.
Theimportanceofvariablestopeoplewhoaregoingtobeaffectedbythedamcanberepresentedbyhowmany
timestheyhavebeenmentionedincognitivemaps(Table3). Schoolwasmentionedinall13cognitivemaps. Con-
structionofroads,jobopportunities,andadvancepaymentofcondemnationvaluewereallmentionedin12outof13
maps. Orderingvariablesbythenumberoftimestheyhavebeenmentionedgivesanorderofimportanceintheeyes
ofthepeoplewhoaregoingtobeaffectedbythedam.Theyindicatewhattheywouldliketoseehappenafterorwhile
thedamisconstructed.
Moreimportantthantherepetitionofvariablesistofindouthowthesevariablesareinrelationtoothers.Centrality
andotherindicescanbeusedtoevaluatethestructureofthecognitivemap. Thecentralityofa variableshowshow
connected it is to other variables and the cumulative strength of these variables. The centrality is composed of the
indegree and of the outdegree. If the indegree is zero and there is a non-zero outdegree the variable is called a
transmitter, if in the opposite it has a non-zero indegree and a zero outdegree then it is a receiver variable. If the
variablehasbothanon-zeroindegreeandoutdegreethanitiscalledanordinaryvariable.Ofcourseordinaryvariables
basedontheratiooftheirindegreesandoutdegreescanbemoreorlessreceiverortransmittervariable.
The centrality of the variable is therefore not only a frequencyof expression but also how importantthat given
variable is given the whole structure of the cognitive map. The type of variables, their in and outdegreesand their
centrality for the social cognitive map of the people affected by the dam, are given in Table 4. Variables that have
a centrality of 0.5 and lower are not included in Table 4 for brevity. After people’s welfare, the most central vari-
ableswerejobopportunities,constructionofroads, schools, governmentsupport,andindustrialfacilities, allwith a
centrality¿2.0.
CognitiveMap Simulations
The real advantage of cognitive mapping is to be able to evaluate differentpolicies and judge their outcome to the
people’s welfare based entirely on the people’s cognitive map without any interference and judgement imposed by
authoritiesorscientists. Thepolicies,theoutcomevariables,andthecausalrelationshipbetweenvariablesareentirely
that of the informants. What remains afterwardsis the implementationof the wants and desires of the people after
damconstruction.Themostimportantvariable,whichneedstobeincreasedafterandduringdamconstruction,isthe
welfareofthepeople. Theeffectofthemostmentionedorcentralsinglepoliciesandthesynergeticeffectsofthese
differentpoliciesweresimulatedtomaximizethewelfareofpeople.
Arelativeincreaseoverthesteadystateofvariableswhennopoliciesareimplementedisusedtocomparedifferent
policiesmentionedbyinformants(Figure3and4). Thesegraphsaretheresultofsimulationsasdescribedinthedata
7
analysissection. Thevariables(policiesimplemented)inthegraphsareorderedfromthemosteffectivetotheleast.
Thesynergisticeffectsof variablesthataremostcentralare simulatedinpairsandintriplestodemonstratethatthe
implementationofdifferentpoliciesdonotonlyaddcumulativelytothewelfareofthepeoplebuthaveanincreased
effect(Figure5and6).
Asallpoliciesareimplementedasaturationofeffectwillbeobserved(Figure7). Themostcentralvariableswill
addaconsiderableamounttothewelfare. Asotherpoliciesareaddedthetotaleffectofpoliciestowelfarewillslow
down,butisneverthelessimportant.ItisnotasurprisegivenFigure3,5and6thatthemostcentralvariableswilladd
moretothewelfareofpeopleandareimportanttoaddress,sincethesearevariablesthatpeoplemostconcuron.
These simulationshelp to pinpointthe most importantissues to be addressed by governmentagenciesafter and
duringdamconstruction.
DISCUSSION AND CONCLUSION
Lookingatthementionedvariables,educationisthemostmentionedvariablethatprovidesforeconomicindependence
andwelfare(Table3). Alltheintervieweeshadeducationintheirmaps. Thereconstructionofschools,whichareleft
underwater,willhelplocaleconomyandtheconstructionofnewschoolsisgoingtoincreasethewealthofthearea.
Informantsthink that students coming from other places in the country, such as the ones in state boardingschools,
livenuplocaleconomiesandcreateaneducationeconomy. Schoolsarealsothethirdmostcentralvariable(Table4,
excludingpeople’swelfare,whichwearetryingtomaximize)andoneofthemostimportantcontributorstowelfare
(Figure3and7).
The second most mentioned variable, construction of roads (Table 3), was a very prominentconcern that came
upineverydiscussion. Constructionofroadswasalsothesecondmostcentralvariable(Table4)andhadastronger
outdegreemakingitclosertoatransmittervariable,henceaforcingfunction. Itwasalsothesecondmostimportant
variable contributing to the people’s welfare (Figure 3 and 7). It is one of the variables with highest synergistic
effectonwelfare(Figure5and6). Thevillageheadmendemandthattheroadsbeconstructedinparalleltothedam
constructionandbefinishedcontemporaneouslywiththedam.
ThepeopleofYusufeliareveryafraidthattheyaregoingtoloosetheirjobsandwanttosecureafuturefortheir
children,thereforejobopportunitiesisalsoafrequentlymentionedvariable(Table3)anditisthemostcentralvariable.
Accordingtoinformantsalotofothervariablesdependonhavingajob. Inaddition,ifothervariablesareaddressed
thenmorejobswillbecreated,becausetheindegreeofjobopportunitiesishigherthanits outdegree(Table4). Job
opportunitieswillnotonlyhaveasignificantimpactonwelfarebutitisalsooneofthefewvariablesthatdoinfluence
theincomeofthepeopledirectly(Figure3and4). Othervariablesthatimpacttheincomeofthepeopleinorderof
contributionareadvancepaymentofcondemnationvalues,revenuesfromthedam,alivelycitycenterandgovernment
8
support(Figure4). Theadvancepaymentofcondemnationvalueswasanotherissuethateverybodyraisedduringthe
investigator’s visit to the area. It is not only considered to have the largest contribution to income but also to the
welfareofthepeople. Thepaymentforlossduetocondemnationmustbejustandtimely,preferablymorethanthe
actualworthoftheland. Implementingsuchapolicywillhelppeopleimpactedbythedamtogeneratemoreincome
andincreasetheirwelfare.
Theconstructionoftheroads,jobopportunitiesandtheadvancepaymentofthecondemnationvalueareissuesthat
everydamprojecthastotakeseriouslyintoaccount. Thesethreeissueshavealsocomeupinalmostallthecognitive
mapsas some of the most importantvariablesto affectthe welfare of the people. The inhabitantsdo notthinkthat
theseissuesaregoingtoberesolvedeasily. Infrastructureandgovernmentsupporttorecreatethepreviouslyexisting
facilities is again among the most mentioned and most central values (Table 3 and 4), they also contribute directly
(Table5and9)andhavesynergisticeffectsinincreasingthewelfareofpeoplewithothervariables(Table7and8).In
countriessuchasChina,Thailand,andMalaysiatheroads,electricity,waterandschoolfacilitieslosttothedamwas
neverreplaced,oritwasconstructedsolatethatpeopledispersedslowlyleavingnotracebehind(WCD,2000).
In China the government reports on dams mention ”the seven difficulties” of resettlement projects. These are
electricity,drinkingwater,schools,foodsecurity,healthservices,telecommunicationandtransportation(WCD,2000).
The people of Yusufeli have seen all these difficulties without any knowledge of previousdams and have put them
into their cognitive maps frequently and with high centralities (Tables 3 and 4). These variables have important
consequencesfortheirwelfare(Figures3-7).
Againaccordingtothecognitivemaps,informantsstatethatthepeopleofYusufeliwanttobeabletoworkinthe
damduringconstructionbutalsoafterwardsasmaintenanceandrunningstaff.
Evenif all these policies are implementedthe misfortune of the people who lost their livelihood, their property
andmemoriesunderwatercannotprevented.Therearetwoimportantvariablesmentionedinthemapsthathavebeen
addressedintheworldandcanbeaddressedinthecaseofYusufeli. Thefirstisrevenuesfromthedam,whichisalso
thesecondmostimportantvariabletogenerateincomeaccordingtosimulations(Figure4).Thesecondisinexpensive
or freeelectricity. Thereare manyexistingexamplesfromthe worldandthere isno reasonwhythese couldnotbe
implementedintheYusufelidamresettlementproject.
InNorwaythereisanexamplewhereeitheraspecialpriceforelectricityorfreeelectricityisgiventodisplaced
people. TherearerevenuesharingprogramsinUrra1DaminColombiaandinthecross-boundaryItaipuprojectin
both Brazil and Paraguay. There is also revenuesharing in Minastuck projectas well as Hydro-Quebecprojectsin
Canada(WCD,2000).Intheworldenergysectoritiscommonpracticethatacertainpercentageofproceedsisgiven
backtothelocalcommunitiesormunicipalities.
The main goalof anyresettlementprojectmust be to increase of the welfare of the people who are goingto be
undoubtedlynegativelyimpactedbythedamafterandduringitsconstruction. A dam,beingdecidedsomewherein
9
thestateapparatus,isanimpositiononthelocalpeople. Therefore,atleasttodecidehowtohelpthesamepeople,it
is wiser to ask themfirst, ratherthan decidingtopdownonwhatthe bestway toincrease theirwelfare is. Through
cognitivemappingonecanbestapproachthewantsanddesiresofthepeoplethroughtheirownwayofthinkingina
bottom-upway.
With fuzzy cognitive mapping the people themselves decide what variables are important and how much they
wouldaffecttheirwelfare. Themapsareeasyforpeopletodraw. Itissimpletocodethemapsandtoaugmentthem
into one social cognitivemap. Any numberof maps can be combined; the maps can be of differentsizes and with
differentvariablesandcausalconnections.Bylookingatthemostmentionedvariablesandmostcentralvariablesitis
clearwhichvariablesaremostimportantforthepeopleasagroup.Finallysimulationscanbedonequicklyandeasily
toseetheeffectofdifferentpolicyoptions.
REFERENCES
Axelrod,R.(1976)StructureofDecision,TheCognitiveMapsofPoliticalElites. Princeton,NJ:PrincetonUniversity
Press.
Bauer,V.(1975)Simulation,EvaluationandConflictAnalysisinUrbanPlanning,in: M.M.Baldwin(ed.) Portraits
of Complexity: Applications of Systems Methodologies to Societal Problems, p. 119-126. Columbus, OH: Batelle
Institute.
Bougon, M., Weick, K., and D. Binkhorst(1977)Cognitionin Organizations: An Analysis of the UtrechtJazz Or-
chestra,AdministrativeScienceQuarterly22:606-639.
Brown, S. M. (1992)CognitiveMappingandRepertoryGrids for QualitativeSurveyResearch: Some Comparative
Observations,JournalofManagementStudies29:287-307.
Carley,K.(1988)FormalizingSocialExpert’sKnowledge,SociologicalMethodsandResearch17:165-232.
Carley,K.,andM.Palmquist(1992)Extracting,Representing,andAnalyzingMentalModels,SocialForces70:601-
636.
Cossette,P.,andM.Audet(1992)MappingofanIdiosyncraticSchema,JournalofManagementStudies29:325-347.
Craiger, J.P., Weiss, R.J., Goodman, D.F., and A.A. Butler (1996) Simulating OrganizationalBehavior with Fuzzy
CognitiveMaps,InternationalJournalofComputationalIntelligenceandOrganizations1:120-133.
Dwivedi,R.(1999)Displacement,RisksandResistance: LocalPerceptionsandActionsintheSardarSarovar,Devel-
opmentandChange30:43-78.
10