A Virtual Laboratory for the Study of History and Cultural Dynamics

©CopyrightJASSS Juan-LuisSuárezandFernandoSancho(2011) AVirtualLaboratoryfortheStudyofHistoryandCulturalDynamics JournalofArtificialSocietiesandSocialSimulation 14(4)19 <http://jasss.soc.surrey.ac.uk/14/4/19.html> Received:28-Jul-2010Accepted:28-Aug-2011Published:31-Oct-2011 Abstract ThisarticlepresentsaVirtualLaboratorythatenablestheresearchertotryhypothesisandconfirmdataanalysisaboutdifferenthistoricalprocessesandculturaldynamics.ThisVirtualCulturalLaboratory(VCL) isdevelopedusingagent-basedmodelingtechnology.Individuals'tendenciesandpreferencesaswellasthebehaviorofculturalobjectsinthetransformationofculturalinformationaretakenintoconsideration.In addition,theeffectoflocalinteractionsatdifferentscalesovertimeandspaceisvisualizedthroughtheVCLinterface.Informationrepositories,culturalitems,borders,populationsize,individual'tendenciesand otherfeaturesaredeterminedbytheuser.Finally,theresearchercanalsoisolatespecificfactorswhoseeffectontheglobalsystemmightbeofinteresttotheresearcher.Allthecodecanbefoundat http://projects.cultureplex.ca/ Keywords: CulturalDynamics,CulturalComplexity,Multi-AgentBasedSimulation,Netlogo,VirtualLaboratory ModelingCulturalDynamics 1.1 Thisworkderivesfromthenecessitytoundertakestudiesofculturaldynamicsthatincludeculturalactionsatalocallevelaswellasanysuccessivetransformationsthatanyinformationalunitorculturalproduct undergoesinacommunicativechain. 1.2 Wehavecreateda VirtualCulturalLaboratory (VCL)usingagent-basedcomputationalmodelingthathelpsstudyhowhumancultureistransformedandtransmittedthroughactsoflearning,imitation,andthe creationofculturalobjectsastheymightbeexperiencedbyanyhumanbeingfrombirthtodeath,independentlyofthespecificcommunitytowhichtheculturalobjectorindividualbelongs.Equallyasimportantis thebroadercontextinwhichthelocalspaceissubsumedbyprocessesbetterexplainedatapopulationlevel,bynumerousculturalgroups,orthatinvolvethemovementofpeopleorculturalobjectsbetween differentecosystems[1].Furthermore,alltheseprocesses—asmuchtheinitialproductionasthefinaluse—willrepeattimeandtimeagainthroughtheyears,allowingfor,ineachrepetition,thealmostinevitable possibilityofatransformation,howeverminute,oftheculturalinformationbeingcopiedandmoving. 1.3 Abetterunderstandingofthelawsthatgovernthesetransformationsofculturalinformationatthelocallevelisascrucialasunderstandingwhathappenswhenbothelements,thelocalandthemobile,activatein casessuchasmigrations,colonisations,ormassivedisplacementsofsymboliccontentfromoneculturetoanother.Furthermore,andfromatheoreticalpointofview,amodelofculturaldynamicsmustmanagethe conceptsandtoolsthatenabletheresearchertoconnectwhathappensculturallywhenahumanbeinginteractswithabook,asonata,apainting,oralibrary,butalsothatwhathappensatalargertemporalor geographicalscale. 1.4 HistoriansandculturalresearcherscanapplythisVirtualLaboratorytodifferentscenarios,inmostofcaseswithouttheneedtoimplementchangesintheVCLEngine.Inourcase,theneedtocreateatheoretical modelofculturaldynamicsinwhichtheindividuals'behaviourplaysastrongrolederivesfromthemanyculturalcommunitiesinvolvedandthedifferentscalesusedtostudythespreadofthebaroqueculturefrom EuropetoAmericaatthebeginningoftheEarlyModernPeriod.Aswehadtodescribebaroquepatternsinordertoexplaintheirorigins,evolution,transmission,andeffectiveness,werequiredthedevelopmentofa modelthatrationalizedtheculturalandsymbolicmovementsbetweenEuropeandLatinAmerica,aswellasthetransformationsandmutationsthatculturalobjectsundergointheirsuccessiveinteractionswiththe varietyofecosystemsandgroupsthroughwhichtheypassontheirjourneys. 1.5 Itisimportanttonotethat,ingeneralterms,theVCLaddressesthreedifferentissuesthatarerelevantforthehistorianandtheculturalresearcher.First,theVCLoffersaplatformtocheckontheeffectsofhistorical eventsandprocessesaboutwhichtheresearcherhascomprehensivesetsofdata.Havingthedataletstheresearchertorefinethemodelheisusingtoexplainthegivenculturalandhistoricprocesses,asboth dataandmodelhavetoshowamutuallycoherentbehaviour.Second,whentheresearcherdoesnothavegooddataaboutthephenomenaheisstudying,theVCLhelpstestthehypothesesandtheassumptions usedbythehistorian,anddouble-checktheresultsofthesimulationwiththelogicofthosehypotheses.Third,theresearchercantakeadvantageoftheVCLbyrehearsingdifferentwhat-ifscenariosthatheknows didnothappen,butwhoseresultswouldbeimportanttoshedlightintothecontextinwhichactualeventstookplace. 1.6 Ingeneraltermsandforthisresearch,weconsidercultureasinformative,representational,andevolutionary.Therefore,weadopttheapproachofRichersonandBoyd,forwhomcultureis"informationcapableof affectingindividuals'behaviourthattheyacquirefromothermembersoftotheirspeciesthroughteaching,imitation,andotherformsofsocialtransmission"(Richerson&Boyd2006,p.5),whichalsoallowsusto incorporatethetransitivecharacterofinformationasdescribedbyJanKahre("informationisalwaysinformation about something"(Kahre2002,p.3)),sincetransitivityengenderstheintentionalandattentional dimensionsexclusivetohumanculture,asstudiedbyTomasello[2].Aculturalsystemis"definedaccordingtoitscapacityandtheformsusedtorepresenttheinformationitneedstoprocessinordertoadapt" (Suárez2007,pp.38-39),thatis,itisinitsrepresentationalabilitiesandproductsattheindividuallevelthatthemainpreferencesandtendenciesaremanifest.Inaddition,thedynamicsofcommunicationwill requirethataculturalsystemdevelopsadequatebehavioursforthecontroloftheinformationalcontentflowinginthecommunitysothatitsveryidentityasagroupispreserved.Behavioursconnectingthe representationanddominationofsystemicinformationtakeplaceattheindividuals'level,butitseffectswillarticulatebothandwillemergeatthepopulationlevel,thatis,whenweseethesystemasacommunity. 1.7 Inordertodevelopthismodelofculturalanalysis,wehaveoptedforagent-basedcomputationalmodeling,sincethesetypesoftoolsallowustobuildabstractmodelsthatcanadapttodifferent"real"situations, takingintoaccountissuessuchasscalabilityandincreasinglycomplexworldsandbehaviours.Agent-basedcomputationalmodelinghasbeenusedwithinthesocialsciencesstartingwiththeworksofEpsteinand Axtell(1996)on"sugarscape"andmorerecentlyinwhatsciencehasdenominated"generativesocialscience",suchasEpstein'smodellingoftheAnasaziculturecomplex(Epstein2006,pp.7-8).Epstein'swork defendstheuseofagent-basedcomputationalmodelsbecausespatiallytheyarevery"powerfulinrepresentingdistributedsystemsofheterogeneousautonomousactorswithboundedinformationandcomputing capacitywhointeractlocally"(Epstein2006,p.6).Inaddition,forEpstein,modelingrepresentsavalidinstrumentforempiricalresearchaswellasforrespondingtocertaininterdisciplinaryquestions,or determiningrulesofindividualbehaviourtomacroscopicregularities,ortoprovingthesensitivityoftheories(Epstein2006,pp.38-39). 1.8 Withinthehumanities,agent-basedcomputationalmodelingisinfrequentatbest,giventhereticenceofasignificantportionofthehumanisticcommunitytotechnology,andthedoubts[ofmany]astowhetherthe complexbehaviourofhumansisevenopentomodellingthroughthereductionofourbehaviourtoafewfundamentalelements,thatistosay,thosethatdefineusashuman[3].Inlightofthisobservation,wehasten toaddthatmodelsdonotsupplantrealityandthatinallmodelling,aswithallabstractions,areductioninthelevelofgranularityinbenefitofagreaterextensibilityofpostulatesandtheirconsequencesalwaystakes place.Consequently,thetaskoftheinvestigatoristoevaluate,atgreatlength,theintellectualbenefitsofamodeldesignedforaspecificcase,thatis,tobalancethereductionofcomplexitytothatofreallife. 1.9 Inourcase,theinterestresides,indeed,inreachingthatbalancebetweenmodelingindividualbehaviourswithinthescopeofcommunicationandstoredinformation,andtherelationsbetweentheseindividual behavioursandthecommunitiesorculturestowhichtheybelong.Inthissense,computationalmodelingofculturalprocessesisparticularlypromisingforthestudyofculturalfusions,militaryconquestsbyforeign powers,migratoryprocesses,enterpriseexpansions,andfusionsinaglobalspaceand,ingeneral,theimpactoflargescaleeventsonculturalevolution.InthespecificcaseofTheHispanicBaroqueProject,the coreprojectdealswiththeculturalmergers,transmissionandadaptationofbaroquepatternsofculture[4]intheHispanicworldoftheEarlyModernAge.Modelingwithcomputer-basedagentsisparticularlyuseful inunderstandingthemajorclashinModernHistory(Diamond1997,p.68),theencounterbetweenAmericaandEurope,becauseoftheamountofinformationavailableabouttheeventsinvolvedinthisprocess, theextremediversityofculturesimpacted,andthefactthataprocessof long-durée overseveralcenturiesisinvolved[5].AllthesefactorsmaketheHispanicBaroqueparticularlywell-suitedtothistypeofmodelling; however,asmentionedabove,theVCLaimsatservingawideraudienceofhistoriansandhumanistsinterestedinuncoveringtherelationsbetweenthecreativebehaviourofculturalagentsandthecultural communitiesandtraditionstheyinhabitandchangeintheirdailyinteractions. 1.10 Twoelementsarerelevantforthisapproach.First,theconceptualizationofaculturalarea[6]intermsof"adynamicsystemofculture[…]thatinculturalterms,developsaseriesofcommonrepresentativecodes, theso-calledbaroquestyleinart,music,andliterature,withvariationsinitscontentanddiachronicappearancesaroundthemiddleofthe16thCenturyandtheendofthe18 thCentury.Theculturalsystemofthe HispanicBaroquewould,therefore,betheresultoftheencounterbetweentheculturalformationsthatpreviouslyexistedinAmericaandSpain,anditwouldconstitutethefirstcomplexandtranscontinentalsystem ofmodernity,asystemthatwouldalsodevelopthefirstmodelofEuropeancolonization"(Suárez2007,p.38).Secondly,thedefinitionofaculturefromitsdynamicstructurewillleadustoitscharacterizationasa system,inthiscase,asacomplexsystem,"thosethatareformedofagreatnumberofpartsthathavemultipleinteractionsbetweenthem"(Simon1996,pp.183-184),aresystemsofadaptationand,assuch,have manylevelsoforganizationinwhichthereareagentsthatserveasthefoundationfortheagentsinhigherlevels(Holland1988,pp.117-118). 1.11 Takingthebaroqueasaculturalsystemenablesustoobserveindividualworksandtheirinteractionswiththehumanbeingswhocreate,contemplateandusethem,toexaminetheemergenceof"cultural"patterns fromtheseinteractions,andtodeterminethediversestatesoftheresultingculture.Theanalysisofthesedynamicsovertimeandinacontrolledspacebymeansofacomputationalagent-basedmodelwould revealtheadaptivecapacityaswellasbifurcationsofthatculture,itsreliefs,and,possiblythelimitsofitselasticityandbreakthroughlines(Suárez2008,pp.61-62).Inshort,thestudyofaculturalsystemat multiplescalesisthefundamentalbenefitofthismethodology.Althoughbaroquecultureisthemaintargetoftheresearchprogramweareworkingon,inthisarticlewedonotintroduceanyspecificapplicationof themethodologyandengineoftheVCLtothatscenario.Here,wefocusonthemethodologyandthetool,andinfuturepublicationswewillshowindetailedtermshowtheycanbeusedinspecificcases. http://jasss.soc.surrey.ac.uk/14/4/19.html 1 12/10/2015 1.12 Inthebalanceofthisarticlewedescribetheelementsandbehavioursoftheagentsinourmodel.Thisisfollowedbyanexplanationofthestepsaresearchermustfollowinordertocreateavirtualculturalmodel (includinglocationoftheagents,descriptionofthephysicalworldinwhichtheycoexist,levelofpreferences,existenceandplaceoftheborders,etc.,thecharacteristicsoftheagents,andincludingthevisualization oftheinformationpanelthatexecutestheinformation).Finally,wepresentthreesamplecasesthathighlightthepotentialofthistypeofsimulationfortheunderstandingofdynamicvirtualenvironments. OurModel:Agents,Messages,andBehaviours 2.1 Ourmodelattemptstoreflectthetransformationsthatinformationundergoeswhenexchangedamongindividuals.Forthispurpose,wedefineinformationbythemessagesthatagentscarryintheircommunicative andlearningprocesses.Wearealsointerestedintheeffectsoftheseexchangesatthepopulationlevel.Inthissectionwedefinewhatamessageisinourmodel,whattypesofagentswehaveincorporatedinthe model,andhowtheagentsbehave.Foramoredetaileddescriptiononthedefinitionsandhowtheyhavebeenmodelled,seetheappendixattheendofthepaper. Messages 2.2 Themodelstudiesthebehaviourintheevolutionofinformationasitistransmittedandtransformedbydiversetypesofagentsinthecommunicativeprocess.Inourcase,informationtakestheformofmessages thatcanbecommunicatedamongagents,whereeachmessageisaunitofinformationthatcanbestoredinthememoryofanagentand/orexchangedwithotheragents. Agents 2.3 Anagentisanautonomousentitywhichobservesandactsuponanenvironmentaccordingtotheinformationitcarriesandreceives.Thismodelincludesseveralsetsofagentswhointeractwitheachotherand havethecapacitytostore,transform,andcommunicatemessages.Herewedefinetheessentialcharacteristicsofeachagentset: 1. Individuals :Thisagentsetisdesignedtorepresentmembersofsocialgroups.Theyaremobile,meaningtheymaymovethroughouttheworld[7],carryasetofmessagesthatcanbetransmittedtoother agentstheyinteractwith,haveaspecificemissionfrequency[8],andusepreferencesandavariablerangeofbehaviourstomodifythemessagestheytransmitandtoaccept,ornot,themessagesfrom otheragents.Thesepreferencescanadaptdynamicallytothemessagesreceived(providingatypeoflearningthatwillmodifyhisfuturebehaviour).Itmustbenotedthatthememoryofthistypeofagents islimited,suchthatifanindividualhasafullmemory,whenitreceivesanewmessage,theoldestmessageislostordiscarded[9]. 2. Repositories :Thisagentsetisdifferentfromtheprevioussetinthatitsmembersarefixedorimmobile.Repositoriescancommunicatewith,andstoremessagesfrom,individualsthatinteractwiththem. Theyarealsoprovidedwithpreferencesandatolerancefeaturethatguidetheirbehaviourwithrespecttotheacceptanceofamessage,butwithnolearningprocessmodifyingthepreferences. Repositoriespossesalimitedstoragecapacity.Tosomeextent,theyactaslibraries:ifamessagearrives,therepositoryconsultsitspreferencesandeitherstoresorrejectsthemessageaccordingly. 3. CulturalItems :Thesearethesimplestagentsetwithinthemodel,butperhapsoneofthemostinfluentialintheglobalbehaviourofthesystem.Inadditiontobeinganothertypeofimmobileagents,cultural itemsdonothavereceptioncapabilitiesandpossessjustonemessageintheirmemory;thereforeculturalitemsalwaysemitthesameinformationandcannotmodifytheirmessagethroughouttheexecution ofthemodel.Theysharewithindividualagentsanemissionfrequencyfeaturesuchthatintheirencounterswithotheragents,theymayormaynotcommunicateamessage.Tosomeextent,theseagents representtheculturalobjectswhoseinformativecontent,inprinciple,doesnotchangethroughouttime,butmaybesusceptibletodifferentinterpretationsaccordingtothehistoricalmomentorthesocial groupthatinteractswiththem(exampleofculturalitemswouldincludeabook,aworkofart,aruinofoldcivilizations,amountain,etc.). Behaviours:reception,memory,learningandemission 2.4 Communicationbetweenallagentsetstakesplacewhentwoofthemoccupythesamevicinity( patch ,intheterminologyofNetLogo,thetechnologythatisbehindthistool).Sinceindividualsaretheonlymobile agents,themodeldoesnotallowforthecommunicationofmessagesbetweenrepositories,culturalitems,orbetweenmembersoftheseagentsets;communicationalwaystakesplaceusinganindividualasthe intermediary(andhisinterpretationasbasicdynamics).Notwithstanding,giventheirrespectivecharacteristics,eachtypeofagentsethasarangeofpossiblebehavioursthataffecthowtheymodifythemessages theytransmit. 1. CulturalItems :Aswenoteabove,culturalitemsdonotreceiveinformationandalwaysemitthesamemessage.Thuswedonotexpectachangeintheirbehaviourbeyondthatnecessarytocompletetheir assignedtasks:torecognizewhenanindividualisinitsproximity,andbasedonitsemissionfrequency,totransmititsmessagetotheindividual. 2. Repositories :Therepositoriespossesspreferencesthatallowthemtodecideaboutmessagestheyreceive.Onceanindividualsendsamessagetoarepository,thelatterfollowscriteriathatdependas muchonthemessageasonthepreferencesandtoleranceoftherepositorytostoreorrejectthemessage.Ifthemessagedoesnotsatisfythecriteriaimposedbythepreferencesoftherepository,itis rejected.Thisprocessensuresthatarepository'sbehavioursimulatesspecializationinaspecifictypeofinformation(suchaspaintings,books,music,etc.);theydonottrytoretaininformationfromall domainsacrossthecultureinwhichtheyexist. Thememoryoftherepositoriesissortedbytime,whatmeansthatthepositioneverymessageoccupiesonthismemorydependsonthetimeatwhichthemessagearrivestoit,establishingakindof relativeantiquityofthemessageintherepository. Thestoragecapacityoftherepositoriescouldnotexceedaprefixedpercentageoftheglobalamountofmessagesintheworld.Ifthisnumbersurpassestheamountallowedthen,basedonthe characteristicsoftherepository,oneofitsmessagesisselectedforelimination.Byusingwhatwecall oblivioncurves ,theusercandefinewhichpartofthememorymightbemorepronetooblivion, dependingonthemomentinwhichtheirmessageswerestored.Inthenextsectionweshowhowtodefinethiscurveanddifferentwaystouseit.Thistechniqueallowsustoconstructourmodelsimulating thenormaloperatingprocedureinlibrariesandmuseums[10]. 3. Individuals :Thisagentsetisassumedtoexhibitamorecomplexbehaviourwithrespecttothetreatmentofinformation,andthereforetheywillbecometheheartofanyprocessofculturaldynamic.Inorder toenrichtheirinteractions,threepossiblebehaviourshavebeendefinedforindividualsasbothemittersandreceiversofinformation:Creative,DominantandPassive.Creativeagentsareattentionally engagedwiththeinformationtheyreceiveandtheydosoinacreativewaythatresultsinsignificantchangesintheinformationcontentandintheindividuals.Dominantagentsonlyexercisetheminimum attentionpossibletobeabletocontroltheinformationtheyreceivebyadaptingittotheirowninformationalpreferences.Finally,Passiveagentsdonotsufficientlyengagewiththeinformationthey encounterintheirecosystems. 2.5 Thesebehavioursrepresentthree"frozen"statesinthepossiblerangeofbehavioursthatagentscanactuallyhave.Itisnotthatagentsareexclusivelypassiveordominantorcreative,butthatthewaytheybehave foreachcommunicationexchangeresultsfromacombinationoftheinformationbeingreceivedandtheattentionthesubjectisabletodevotetothatinformationaccordingtocognitiveability,thetypeofinformation andstrengthofthesender,andotherfactorsdifficulttoquantify.Worksthathaveanalyzedtheinformationsocietyhaveinsistedontheideathatdigitalmedia,thetechnologicalrevolution,andthesocialand economicimplicationsoftheseprocesseshaveshownhowthis"economyofattention"[11]resultsfromcommunicationexchanges.Whatweproposehereisthattheintentionalityofthesubjectwithrespectto culture—previouslydefinedasinformationthataffectsindividual'sbehaviours—canbedescribedasaresultoftheconfluenceoftheinformationreceivedandtheattentionlenttoitbytheagent.Thus,our classificationofagents'behavioursas passive —whoabsorbtheinformationtheyareexposedto— dominant —whorejectthatinformationasforeign—and creative —whodealwithitaccordingtotheirown interestsandpreferences—intendstoreflectthewaysinwhichthatconfluenceofinformationandattentionoccurs.Inthisrespect,Schmajuketal.haveproposedthattheattentional-associativemodelof conditioningcandescribecertainaspectsofthecreativeprocesssinceitisassumedthat"theattentiontonovelstimuliisincreasedincreativepeople"(92)[12]. 2.6 Thesethreebehavioursreflecttheintentionalityoftheindividualwithrespecttoinformationaccordingtotheirownattentionalstateand,byextension,theyalsoreflecttheirintentionalitywithrespecttotheother agentwithwhichitinteractswhenreceivingoremittingamessage.Hence,thesebehaviourswilloccurbothatthemomentofmodifyingandsendingamessagetoanotheragentandwhendecidingtoacceptornot amessagesentbyanotheragent. 2.7 Emission:Theemittingbehaviourofanindividualagenttakesoneofthefollowingforms: Inthecaseof Dominant behaviours,ifthemessageanindividualisgoingtotransmitsatisfiesitspreferences,thenthemessageisemittedwithoutchanges;alternatively,ifthemessagedoesnotsatisfyits preferences,theagentmodifiesthemessage"atthemomentoftransmission,"sothatthemessageadapts,sufficiently,toitsownpreferences; Inthecaseof Creative behaviours,andindependentlyofthecharacteristicsofthemessage,themessageismodifiedfollowingthepreferencesoftheindividual,insuchawaythattheindividualaddspartof itspreferencestothecontentofthemessage. Inthecaseof Passive behaviours,themessageisemittedwithoutmodification,independentofthepreferencesoftheowner. 2.8 Inordertoensurethatanindividual'sbehaviourdoesnotsimplyrepeatitself,butcanexpressdifferentbehaviours,wehaveincorporatedatendencypropertyforeachindividualthatgovernstheintensityofeach behaviour.Thisensuresthatanindividualwhoexhibitsahighlydominantetendencymayinsomeinstancesbehaveasacreativeorpassiveindividual. 2.9 InFigure1,weshowtwoexamplesofindividuals'tendencybehaviours,wherethevalueofeachpossiblebehaviourisproportionaltotheproximitytothecorrespondingvertex. Figure1.Twoexamplesofindividuals'tendencybehaviours 2.10 ReceptionandStorage:Intheprevioussectionweshowhowtendencieshavearoleinthemodificationofthemessageemittedbyanindividual,buttheyalsoinfluencethereceptionofamessagebythereceiving individual.ThismeansthatifthereceivingindividualisinaPassivebehaviour,itwilladmitthemessagewithoutconditions,butifitactsasDominantorCreative,theacceptanceofthismessagewilldependon whetherornotthemessageiscloseenoughtothepreferencesofthereceivingagent.Additionally,duetothelimitedstoragecapacityofanindividual'smemory,onceareceivingagenthasreceivedandaccepted anewmessageitmusteliminatetheoldestmessagetoensurethatthetotalnumberofmessagesinitsmemorydoesnotexceedcapacity. http://jasss.soc.surrey.ac.uk/14/4/19.html 2 12/10/2015 2.11 Learning:Ourmodelalsointroducesalearningfactorforindividualssuchthattheinformationeachindividualreceivesintheformofamessageaffectshimandresultsinsomeformofadaptationtotheinformative surroundings.This"learning"isdefinedasatransformationintheindividual'spreferencesasaresultofitsinteractionwiththeinformationencodedinthereceivedmessage.Theindividualwillhave"learned"bythe timeitmusttransmitthenextmessage,whichwillbetransformedaccordingtothe"new"preferenceintheagent.Thisfactorimpliesthatanindividualwhodoesnotacceptmessageswillnotbeabletolearn regardlessofitsbehaviouraltendency.Differencesinlearningratesbetweenindividualsareanticipatedtobeanextremelyusefulaspectofthismodelwhenstudyingtheperformanceofindividualsfromdifferent worldsorcultures.Theywillalsobeimportantwhendetermininghowahigherlearningratemightofferadvantagesintermsofretainingaspecificculturalgroup'sowninformationinasituationofcontactwithother groups. 2.12 Itisimportanttounderscorethattheintentionalityofindividualsalwaysoccursatthelocallevel,sincethisiswherethehumanbeingpronouncesitselfascultural;thatistosay,thisisthespaceinwhichthehuman beingimaginestheintentionality—theabilityofanindividualtorepresenttheintentionsofotherhumanbeingspresentinitssurroundingsaccordingtotheinformationitreceivesandtheattentioncapturedbythe information—ofotherhumanbeings.Thisdoesnotmeanthatculturalexchangestakeplaceexclusivelyornecessarilywithneighbouringindividualsatalocallevel,ratherthattheexchangebetweenneighboursas communicativepartnersbecomessymbolic,cultural,andthereforehuman.Theextentofthespacenecessarytomaintainthecharacteristicsofaculture'sscopeissomethingthatwehopetolearnbymeansofthe extensiveuseofnewinformationtechnologiesinculturalpractices.Indeed,technologiesofculture—thatis,thespecificationsaculturalgroupsusesinordertoreproduceandtransmititsowncollectiveknowledge —playadecisiveroleinhowindividualsestablishtheirinteractionwithsymbolicobjectsdistantinspaceandtime.Asculturalinformationisstoredbothinhumanbrainsand"encodedinsomematerialobject" (Richerson2006,pp.61-62).Wehopetobetterunderstandtheevolutionaryimportanceofobjectsandculturalsymbolsinthecontextofhumansocietiesbyfindingputhowtheyshapeupthecontentofthe informationexchangedthroughthemand,consequently,thebehavioursofagentsusingthem.Takingintoaccounttheseideas,thebehavioursspecifiedinourmodelcanbeappliedasmuchtocommunitiesin geographicproximityastothevirtualcommunitiescreatedwithinatechnologicalinfrastructuresuchassocialnetworks. 2.13 Whereasthepreferencesanddegreesofintentionalityarefundamentaltounderstandingtheinteractionbetweenindividualsandmessagesatthelocallevel—theytakepartdirectlyasmuchinthemodificationof themessagesas,indirectly,inthemodificationsthattheagentsundergo—thisfeatureismitigatedwhenweconsiderthebehaviourthatemergesfromthecirculationoftheinformationatthepopulationlevel.This doesnotmeanthatpreferencesandintentionalityarenotimportant;ratherthattheyservetopayattentiontotheindividual'sfreedomtotransformandactonthereceivedinformationaswellastotheeffectsof intentionalityinthedistributionoftheinformationonagreaterscale,communitarian,orintercultural.Thus,ourresearchquestionscanbetransformedintothefollowingforms:Howelasticisaculture?Howefficient isasystem'sculturalcommunication?Doesthesymbolicinfrastructureofagivencommunityallowfortheself-organizationofdiverseandnumeroushumangroupsoveralargerspaceandlongertimeperiod?How effectiveisone'sowncultureinhelpingonerelatetothosedifferentand/ordistantfromone'sself? 2.14 Aswehaveseen,theindividual'slearningprocessisdescribedasafunctionoftheagents'tendencyandpreferencestowardthedifferentcharacteristicsofthemessageswithwhichitinteracts.Notwithstanding, thebroadcastofinformationbytheagentsalwaysimpliesachangeinthemessageatthemomentofemission.Here,weareassumingthatourunitsofinformation,themessages,donotbehavelikememes,but thattheybehavelikementalrepresentations.Therefore,theyarenon-discreteandcontributetoaprocessofculturaltransmissionthatis"highlyinaccurate"(Heinrich2002),butnotlesseffectiveinitsabilityto makecommunitiesemergeasaresultofconstantcommunicativeexchanges. TheVirtualCulturalLaboratory:EditingWorlds 3.1 OurVirtualCulturalLaboratory(VCL)iscreatedinNetLogo( Wilensky1999),asimpleprogramminglanguage(basedonLogo)well-suitedtothemodelingandsimulationofphenomenainwhichmanyindividuals interact,asinphenomenathatoccurinnature,socialcommunities,ormanyareasofmathematicsandphysics.NetLogoincorporatesthetoolsnecessarytomodelcomplexsystemsastheyevolveovertime;itis capableofmodelinghundredsorthousandsofindividuals(people,bacteria,insects,organizations,nodesofagraph,etc.)thatinteractwitheachotherandwiththeirenvironments.Inshort,NetLogoletsusto explorelocalinteractionsbetweenindividualsandalsomacroscopicpatternsthatemergeasaresultofthesemicroscopicinteractions.Theprogramisintuitiveandallowsprogrammerstocreatemodelswith relativeeaseinordertoverifyhypothesesondecentralizedsystems(Gilbert2008:49). 3.2 TheVCLcombinesthreetoolsthatworkinacooperativemanner:aworldeditorthatletstheuserdefinetheglobalfeaturesoftheenvironment;anenginethatexecutesandvisualizesinrealtimethecultural dynamicsofadefinedworld;andatoolthatallowsbothtoreproduceandtoanalyzethedynamicsstoredbytheengine.Theprotocolfollowedbyauserofourvirtuallaboratoryconsistsofthreesteps.Inthefirst step,theresearcher/useremploysaneditortodefinetheworldshewishestosimulate(thisincludescharacteristicssuchasgeographicareasandagentsetsandtheirproperties).Inthesecond,afterthese definitionsareloadedintotheengine,theuserrunsthesimulation,generatingdiversevisualizationsthatrepresentthedynamicsofthevirtualworldasdefinedbytheinitialsettings;inthisstep,thesystemwillalso allowtostorethedatafromthisdynamics.Lastly,theusercananalyzeandcomparetheseveralshotsstoredintheformerstep,reproducingthekeyinstantsintheevolutionofthesimulation.Insection3,wewill detailthespecificsoftheeditorofworlds,andinsection4wewillexplainhowtousetheengineandthevisualizationtools. 3.3 Figure2showstheprimaryeditorinterfacewhereallthecontrolsarelocatedsuchthattheyareeasilyaccessiblefromthesamescreen.Thecontrolsare:ViewsSelector;ElementsTools;GroupsDefinitions; RepositoriesandCulturalItemsTools;AreaDefinitionTools;FileControls;Summary. Figure2.EditorInterface 3.4 Thereareseveralcontrolsinthisinterfacetodefinesomestatisticalpropertiesofthemodel.Thismeansthatthesepropertiesarenotdefinedforeachindividual,butratherasstatisticaldistributionsforanagent set.Forexample,thelearningfactorofagroupofindividualsisnotacharacteristicdefinedforeveryindividual,butitcorrespondstoagroupfunction,approximatedbyaGaussiandistributionofthediversevalues ofthelearningfactorfortheindividualsinthegroup.Thestatisticalpropertiesofadistributionmaybespecifiedusingapairofsliderslabelled'mean'and'variance'.Onceset,theViewbuttondisplaysthespecified curvesintheDistributionplot. 3.5 Figure3illustrateshowtheseparametersaffectthedistributionsgeneratedwithpairsofvalues(mean=0,1;variance=0,08)and(mean=0,66,variance=0,16). Figure3.TwoexamplesofDistributioncurves 3.6 Following,thezoneoffunctionalitywithintheVCLaredetailed: 1. FileControls (Figure4):Thisisaseriesofthreebuttonsusedtoeithercreateanewworld,readaworldpreviouslycreated,orsaveanewworldonceithasbeencreated. http://jasss.soc.surrey.ac.uk/14/4/19.html 3 12/10/2015 Figure4.FileControlszon 2. ViewsSelector (Figure5):Thiszoneisusedtovisualizetheelementsoftheworldunderconstruction:theAreasbuttonhelpsrepresentthegeographicareasofthesimulatedworldanddisplaysthe positionsoftheagents(andgroupsofindividuals)withintheworld;theMessagesbuttondetailstheinitialsettingforthemessagesassociatedwithIndividualsandCulturalItems;andthePreferencesbutton showstheinitialpreferencesofIndividualsandRepositories. Figure5.ViewsSelectorzone Astheelementsoftheworldaredefined,theyaregeneratedinthe AreasView ingraphicform:theGroupsofIndividualsareplacedintheirselectedareasandassignedtheuser-definedcolor;all Repositoriesarerepresentedasbrownbuildings;andCulturalItemsaredepictedastowers. 3. AreaDefinitionTools (Figure6):Thesetoolsallowtheusertoassigncolorstodifferentareasofthephysicalworld.Thestepsneededtodefineanareaare:(1)selectaPenSizeandAreaColorandpress theDrawAreabutton;(2)whilethisbuttonisactivatedtheusercandrawdirectlyoverasectionoftheworldtomarkitwiththeselectedcolor. Therearetwocolors—blackandgrey,respectively—reservedinthemodelandassociatedwithspecificareasandthatallowtheresearchertoinsertaWallorBorder.Whenanareaoftheworldispainted black,symbolizingaWall,Individualswillbepreventedfrommovingacrossthearea.IfanareaisdesignatedasaBorder,accessbyindividualstothosezonesiscontrolleddirectlybytheengine,meaning theusercanopenorclosethosezonesatwill. Figure6.AreaToolszone 4. ElementTools (Figure7):Throughthesecontrolstheusercancreate,remove,select/unselect,moveandupdatethedifferentagentsintheworld(Groupsofindividuals,RepositoriesandCulturalItems). http://jasss.soc.surrey.ac.uk/14/4/19.html 4 12/10/2015 Figure7.ElementsToolszone Tocreateanewelement,theuserjustneedstoclickon"Create"andfollowthedirectionsprovidedbythesystem:inallcases,theusermustenteranIDfortheelementtobecreated,selectitsgeographic location,andthenprovidethemessages/preferencesassociatedtoit.Todeleteoneoftheelements,theuserjustpresseson"Remove"and,whilethisbuttonisstillpressed,clicksontheelement(s)tobe deleted.Iftheuserwantstoselectanelementandloaditspropertiesagain,s/hewillpresson"Select"andthenclickonthechosenelement.Oncetheelementhasbeenselected,theusercanmodifyany ofitspropertiesandthenpresson"Update"tostorethechangesintheelementproperties.The"Unselect"buttonallowstodeselectthepreviouslyselectedelement.Whilepressingon"Move"theusercan changethelocationofanyoftheelementsactuallyintheworld. 5. RepositoriesandCulturalItemsDefinitionZone (Figure8):ThepropertiesofrepositoriesandculturalItemscanbesetbyusingthesecontrolsaftercreatingtheagentwiththeElementTools.Theonly propertyforCulturalItemstheusercanmodifyisthefrequencyofemission,avaluethatindicatestheprobabilitytoemitamessageineverytickofthemodel(inadditiontolocationandthemessagethat willconstantlyissue).InthecaseofaRepository,theusercandefineitsToleranceandthedistributionparametersassociatedtoits"OblivionCurve"(asweindicateabove,thisisoneofthepropertiestobe definedbya( mean,variance )pair).Inaddition,asliderallowstheusertodecidethemaximumpercentageofmessagesthatthecurrentRepositorywillbeabletostoreduringtheexecutionofthemodel. Figure8.RepositoriesandCulturalItemszone Rememberthatitisnecessarytoselecttheagentbeforewhenmodifyingitspropertiesandaftertoupdatethesepropertiesinordertostorethemodification. 6. GroupsDefinitionZone (Figure9):TheeditorlabelsasaGroupanysetofindividualsthatsharecommoncharacteristics( memorysize,receptiontendency,learningfactor,behaviour,initialarea,initial messages ,etc.).Thesecharacteristicsmayservetodefineindividualswithdifferentbehaviourswithinasingleculturalgrouportodefinedifferentculturesinthephysicalworld.Thegeneralpropertiesofa group(notindividuals)are: population (controlledbymeansofaslider), ID-Group (asmeanstoidentifyalltheindividualsinthegroup,andprovidedduringthecreationprocess), Color,Max-Memory-Size, InitialArea (toprovideanareawhereindividualsfromthegroupareoriginallydistributed,andidentifiedbythecolorofthearea)and ForbiddenArea (anareathattheindividualsfromthegroupmaynot passthrough,itisdefinedasanyotherarea,butunliketheWallsorBorders,theseareasdependexclusivelyonthegroup). Figure9.Groupsdefinitionzone The Max-memory-size determinesthemaximummemorycapacityofeachIndividualwithinagroup,buttheexactmemoryvalueforeachindividualisderivedfromthenormaldistributionprovidedbythe pair( Mean-memory,variance-memory ).Theotherindividuals'propertiesthataredefinedbyanormaldistributionare ReceptionTendency (whichalongwiththepreferencesandtendencyofeach individualwilldeterminetheindividual'sbehaviour—Passive,DominantorCreative—whenreceivingamessage),the FrequencyActivity (thisvaluedeterminestheprobabilitythattheindividualemitsa messagetoanotheragentinthesamepatchateverytick),and LearningFactor . Aspreviouslystipulated,individualsbehaveaccordingtothreepredeterminedpatterns(Passive,DominantorCreative).Thetendencyoftheindividualdeterminestheproportionoftimesthatanindividual willbehaveasoneofthesethreepatterns.Thistendencyisdefinedprobabilisticallybymeansofsliderssimilartothatusedtospecifythenormaldistributioncurves,butinthiscaseitisdepictedasa trianglewithtwo-dimensionaldistribution(seeFigure1).Theproximityofthepointtoeachoftheverticesofthetriangleindicatesthepropensitytoexhibitaspecificbehaviour(theclosertoavertex,the morepronetothatbehaviourtype).Twoslidersallowtheusertopositionthecenterofthedistribution(bi-dimensionalmean)asapointinthetriangleandathirdslidersetsthevariance.Figure10showsan exampleofadistributionofbehavioursthatindicatesagreaterindividualpropensitytobehaveinaDominantorPassiveway,andasmallerpropensitytobehaveinaDominantmanner. Figure10.Exampleofdistributionofbehaviours Toaddtheinitialdistributionofmessages/preferencesforoneofthegroups,theusermustfirstselectthegoalgroupintheAreaView,thenchangetheviewtoseeMessages/Preferences,andfinallypress theAddbutton.Whilethisisactivetheusermaydrawinthemessages/preferencesassociatedwiththatgroup(seeFigure11). http://jasss.soc.surrey.ac.uk/14/4/19.html 5 12/10/2015 Figure11.MessagesandPreferencesview 7. DistributionGraphics (Figure12):Asweindicatedpreviously,thiszoneshowsapre-visualizationofdistributioncurvesrelatedtothoseagents'propertiesassociatedwithGaussiancurves. Figure12.CAPTIONGOESHERE 8. Summary (Figure13):Thiszoneshowsdataaboutthepresentstateoftheworldasitisbeendefined(asummaryofthediverseagentgroupsthathavebeencreatedtothemomentalongwithsomeof theirbasicproperties).EachgroupofIndividualscreated,eachRepository,andeachCulturalItemisassignedaconsecutiveidentifyingindexnumber(Idx). Figure13.Summaryzone VCL:TheEngineandVisualization 4.1 Oncetheworldhasbeendefined,thesecondtool,theVCLengine,isusedtosimulatetheresultingdynamicsoftheinteractionsamongthediversedefinedagents. 4.2 Again,thepurposehereistooffertheresearcheraneasy-to-usetoolthatdisplaysonasinglescreenalltheinformationrelatedtotheevolutionofthemodel.Figure14showsascreenshotofthecontrolpanel, wherewecanfindsixfunctionalzones: http://jasss.soc.surrey.ac.uk/14/4/19.html 6 12/10/2015 Figure14.EngineInterface 1. FileandRunbuttons (Figure15):Thesebuttonsallowtheusertoloadtheworldcreatedintheeditor,start/stoptheexecutionofthesimulationandresettotheinitialconditionsoftheloadedworld.Itaddsa fourthcontrolthatallowstoexporttheshotstakenthroughouttheexecution. Figure15.FileandRunbuttons 2. WorldReport (Figure16):Thisreportisanoutputframewherethebasiccharacteristicsoftheworldareshown(numberofculturalgroups,repositories,andculturalitems).Italsoshowsthestatusofcertain parametersthatmaybemonitoredduringexecutionofthemodel,andtheactionsfromtheagendathathavebeenexecuted. Figure16.WorldReport 3. ViewZone (Figure17):Thisframesurroundsthegraphicalrepresentationofthemodelandprovidessomedifferentviewsbypressingthebuttonsintheupperpartoftheframe. http://jasss.soc.surrey.ac.uk/14/4/19.html 7 12/10/2015 Figure17.Viewzone TheWorldViewisthedefaultsetting,itshowsthegeographiczonesandsetsofexistingagentsintheworld,aswellastheirmovementandinteractionswhilethesimulationisrunning(Figure18,wherewe showtwodifferentsampleworlds,showingbordersandpathsconnectingdifferentareas). Figure18.Twodifferentworlds TheagentpreferencesarefoundbyselectingthePreferencesVieworPreferences(Stat)Viewontheselector.Thepreferencesarerepresentedaspointsina2Dspace(seeappendixforanexplanation) withacolorthatcorrespondstotheholderofthispreference(Figure19).Inthefirstoptioneverypreferenceisshownasaseparatepointinthe2Dspace,inthesecondoptiononlyacircleexpressingthe meanandvarianceofthepreferencesofthegroupisshown(alsothepreferencesoftherepositoriesareshowninthespace). Figure19.Preferencesview TheMessagesViewandMessages(Stat)Viewprovideanupdatedviewinrealtimeoftheinformationspace(themessagesthatareactuallyintheworld).Theyarecolouredaccordingtoitsownerand,as preferences,representedaspointsina2Dspace.ThemessagesassociatedwiththeRepositoriesareshowninbrown(Figure20). Figure20.Messagesview TheCultureViewprovidesthemostcomplexvisualizationofalltheviews.Becausethemessageschangewitheachinteraction,wefoundincompletetoonlyreportindividualmessagesintheinformative contentofindividualagents;consequently,wedecidedthatanimportantinformationalaspectofthemodelrequiresthehighlightingofthezonesoftheinformativespacethatareoccupiedmostfrequently byasetofmessages.Inordertorenderthisvisualizationwehaveimplementedthefollowingprocedure:theinformativespaceisasquaredlatticeformedbycellsofequalsize,andateachunitoftimethe messagesdepositintheirlocationaconcentrationofculturalcontent(similartoachemicaldeposit)thataccumulatesinthiscell.Toensurethattheeffectofamessageonagivencellisnotpermanent,the accumulationofculturalcontentdecreasesovertime,whichcausestheamountofculturalinformationalsotodiminishovertime.Thismechanismpermitstheresearchertomeasuretheamountof informationineachcell,andtherebyidentifieswhichzonesoftheinformativespaceareoccupiedmostfrequentlybythemessages.Hence,frequentlyoccupiedzonesrepresenttheculturalinformationthat staysinthepopulationwhiletimepassesby. http://jasss.soc.surrey.ac.uk/14/4/19.html 8 12/10/2015 Furthermore,throughthe"Color-messages-in-Culture-View"buttonfromthe Actions&Parameters ,theenginepermitstheresearchertogroupmessagesaccordingtodifferentcriteria:a)"ByGroup", accordingtothesetofagentsthemessagebelongsto;b)"ByOriginalArea",accordingtheoriginalareaofitscarriers;c)"Byownercolor",inthecasethattwoormoregroupsexistwiththesameIDand theycan'tbedifferentiate"ByGroup",butitcanbedonebycolors(itisawaytocreatesub-groupsinsideamainbiggergroup).Itisimportanttorememberthatallthreegroupingsareactingonthesame setofmessages,thereforethesetofaffectedcellsisalsothesame;thismeansthatachangeinthegroupingcriterionwillbemainlyreflectedasachangeofcellcolordistributionontheshape,butnoton theformoftheshapeitself(Figure21). Figure21.ThreeviewsofthesameCulture(bydifferentgroupingcriteria) Thesliderlabelledas Bright ,locatedundertheViewwindow,allowstheusertoregulatethebrightnessofthecellsaffectedintheCultureView.Thischangedoesnotaffecttheaccumulatedvaluesof culturalcontent;itssolepurposeistoaidinthevisualizationoftheCultureview.Additionally,iftheresearcherwishestopreventthemessagesontherepositoriesfromaffectingtheculturalrepresentation oftheindividuals,themodelprovidesaswitchcontrolthathidesthesemessagesinthisvisualizationin Actions&Parameters (seebelow). 4. Actions&Parameters (Figure22):Thiszonecontainssomecontrolsthatcanbeusedwhilethesimulationisbeingexecutedandthataffecttheworldbehavioraswellashowtheinformationisbeen presented.Withthe"Open"and"Close"buttonstheusercanopen/closetheglobalborders(delimitedbythegreycolor).Ifanagentstandsinaborderzoneduringtheclosingofborders,itwillbemovedto anear-by,butpermitted,area.Withthe"Movement"controltheuserwilldecidewhattheagents'movementwillbelike.Ifthe"Follow-paths"optionisactive,theagentswillberandomlymovedfollowing thosepositionsinwhichtheydonotaltertheirorientation;bydoingthisthelikelihoodoftheagentsgettingstraightlinesishigherandthesystemwillbeabletobetterworkwithpathwaysthatconnect separatedareas(asshowninFig.18);otherwise,themovementwillbecompletelyrandom(faster,butmoreinefficientinthecaseofnarrowpassageways).The"MakeSnapshot"buttonallowsgettingan instantshotoftheworldthat,alongwithothershotstakenthroughoutthesimulation,canbeexportedtotheanalysistoolandbereproducedasthekeyinstantsintheevolutionofthedynamicsofthe simulation.Theeffectof"Show-Repositories-in-Graph?"parameterisexplainedinthenextparagraph. Figure22.ActionsandParameterszone 5. Graphs (Figure23):Thiszonehostsdifferentgraphsshowingpossiblerepresentationsofrealtimeaspectsfromthesimulation.Thethreefirstoptions("CulturalDynamicsbyArea","CulturalDynamicsby Group"and"CulturalDynamicsbyOwnerColor")showthedynamicsofthetotalamountofmessagesintheworldgroupedbytheselectedcriteria.Theoptions"PreferencesDistances"and"Messages Distances"showtheevolutionofrelativedistancesbetweenpairsofmeansofgroups'preferences/messagesandrepositories.Theoptions"PreferencesVariance"and"MessagesVariance"showthe evolutionofvariancevaluesofpreferences/messagesforthesameelements.Bychangingtheparameter"Show-Repositories-in-Graph?"wecandecideiftheinformationfromrepositoriesisincludedor notinthegraphs. Figure23.Graphszone Nextfigureshowstheresultyoucanobtainfromthedifferentoptionsingraphs(amountofmessagesgroupedbyoriginalareaoftheowner,amountofmessagesgroupedbyowner,distancesbetween meansofmessagesofelementsandvarianceofmessagesofgroups): http://jasss.soc.surrey.ac.uk/14/4/19.html 9 12/10/2015