資源描述:
《modelling the moral dimension of decisions決策的道德維度建?!酚蓵?huì)員上傳分享,免費(fèi)在線(xiàn)閱讀,更多相關(guān)內(nèi)容在教育資源-天天文庫(kù)。
1、ModellingtTheMoralDimensionoOfDecisionsKatieSteele,UniversityofQueenslandInthispaperweexploretheconnectionsbetweenethicsanddecisiontheory.Inparticular,weconsiderthequestionofwhetherdecisiontheorycarrieswithitabiastowardsconsequentialistethicaltheories.We
2、arguethatalthoughthereisasenseinwhichotherethicaltheoriescanbeaccommodatedbystandarddecisiontheory,theseethicaltheoriesarenotnaturallysuitedtotheprojectofmaximisingexpectedutility.Inparticular,wearguethatthemotivationfortheotherethicaltheoriesandthepsych
3、ologicalprocessesoftheagentswhosubscribetothoseethicaltheoriesarelostorpoorlyrepresentedintheresultingmodels.1.INTRODUCTIONDecisiontheoryhastwocomponents:probabilitiesandutilities.Fromtheformalpointofview,thesetwocomponentsplaysymmetricalrolesindecisiont
4、heory.Foreachact–statepair(oroutcome)weassignaprobabilityandautility,thenwemultiplythesetogetherandsumtheproductsacrosseachact.Theresultingsumwecalltheexpectedutility.Standarddecisiontheorythentellsustochoosetheactwiththegreatestexpectedutility(ifthereis
5、suchanact).Thisisallveryfamiliar.Whatwedrawattentiontoisthat,intermsoftheformaldecisioncalculus,probabilitiesandutilitiesaretreatedsimilarly—theyarejusttworealnumberstobemultipliedandthenadded.Thesimilaritiesbetweenprobabilitiesandutilitiesrunevendeeper.
6、Bothareconstrainedbyaxiomatictheories:probabilitiesbytheKolmogorov(1956)axiomsandutilitiesbyvonNeumann-Morgenstern(1944)utilitytheory.Theseaxiomatictheoriesplaceminimalstructuralconstraintsonprobabilitiesandutilities.NotesOnemightargueaboutwhethertheresp
7、ectiveaxiomsareinfactminimalconstraints.Wewon’tpursuesuchissueshere.Ourinterestliesinwhatotherconstraintsweneedinadditiontotheaxioms.(Otherrepresentationtheorems,suchasSavage’s(1954)expectedutilitytheory,explicitlybindtogetherbeliefandpreference,andthust
8、heresultingprobabilityandutilityfunctions.)Whicheverwaywejustifytheirconnectiontobelieforpreference,tTheaxiomsensurethattherearenoviolationsofconsistency.Forexample,theKolmogorovaxiomsruleoutprobabilitiesthatdon’tsumto1,so