Knowledge and conditionals: essays on the structure of inquiry robert c stalnaker - Read the ebook o by Education Libraries

https://ebookmass.com/product/knowledge-and-conditionals-

Instant digital products (PDF, ePub, MOBI) ready for you

Download now and discover formats that fit your needs...

Belief, Agency, and Knowledge: Essays on Epistemic Normativity Matthew Chrisman

https://ebookmass.com/product/belief-agency-and-knowledge-essays-onepistemic-normativity-matthew-chrisman/

ebookmass.com

Milton's complex words: essays on the conceptual structure of Paradise Lost First Edition John Milton

https://ebookmass.com/product/miltons-complex-words-essays-on-theconceptual-structure-of-paradise-lost-first-edition-john-milton/

ebookmass.com

Mastery, Dependence, and the Ethics of Authority Aaron Stalnaker

https://ebookmass.com/product/mastery-dependence-and-the-ethics-ofauthority-aaron-stalnaker/

ebookmass.com

London in the Roman World Dominic Perring

https://ebookmass.com/product/london-in-the-roman-world-dominicperring-2/

ebookmass.com

Principles and Practice of Structural Equation Modeling, Fourth Edition (Methodology in the Social Sciences) 4th Edition, (Ebook PDF)

https://ebookmass.com/product/principles-and-practice-of-structuralequation-modeling-fourth-edition-methodology-in-the-socialsciences-4th-edition-ebook-pdf/ ebookmass.com

CompTIA® A+ Core 1 (220-1101) and Core 2 (220-1102) Cert Guide Rick Mcdonald

https://ebookmass.com/product/comptia-a-core-1-220-1101-andcore-2-220-1102-cert-guide-rick-mcdonald/

ebookmass.com

A Curse of Queens Amanda Bouchet

https://ebookmass.com/product/a-curse-of-queens-amanda-bouchet-3/

ebookmass.com

Mathematical Modeling and Applied Calculus Joel Kilty

https://ebookmass.com/product/mathematical-modeling-and-appliedcalculus-joel-kilty/

ebookmass.com

Gothic War on Terror: Killing, Haunting, and PTSD in American Film, Fiction, Comics, and Video Games Danel Olson

https://ebookmass.com/product/gothic-war-on-terror-killing-hauntingand-ptsd-in-american-film-fiction-comics-and-video-games-danel-olson/

ebookmass.com

Madoff Talks: Uncovering the Untold Story Behind the Most

https://ebookmass.com/product/madoff-talks-uncovering-the-untoldstory-behind-the-most-notorious-ponzi-scheme-in-history-jim-campbell/ ebookmass.com

KnowledgeandConditionals

Knowledgeand Conditionals

EssaysontheStructureofInquiry

RobertC.Stalnaker

GreatClarendonStreet,Oxford,OX26DP, UnitedKingdom

OxfordUniversityPressisadepartmentoftheUniversityofOxford. ItfurtherstheUniversity’sobjectiveofexcellenceinresearch,scholarship, andeducationbypublishingworldwide.Oxfordisaregisteredtrademarkof OxfordUniversityPressintheUKandincertainothercountries

©RobertC.Stalnaker2019

Themoralrightsoftheauthorhavebeenasserted

FirstEditionpublishedin2019

Impression:1

Allrightsreserved.Nopartofthispublicationmaybereproduced,storedin aretrievalsystem,ortransmitted,inanyformorbyanymeans,withoutthe priorpermissioninwritingofOxfordUniversityPress,orasexpresslypermitted bylaw,bylicenceorundertermsagreedwiththeappropriatereprographics rightsorganization.Enquiriesconcerningreproductionoutsidethescopeofthe aboveshouldbesenttotheRightsDepartment,OxfordUniversityPress,atthe addressabove

Youmustnotcirculatethisworkinanyotherform andyoumustimposethissameconditiononanyacquirer

PublishedintheUnitedStatesofAmericabyOxfordUniversityPress 198MadisonAvenue,NewYork,NY10016,UnitedStatesofAmerica

BritishLibraryCataloguinginPublicationData

Dataavailable

LibraryofCongressControlNumber:2019933997

ISBN978–0–19–881034–6

Printedandboundby CPIGroup(UK)Ltd,Croydon,CR04YY

LinkstothirdpartywebsitesareprovidedbyOxfordingoodfaithand forinformationonly.Oxforddisclaimsanyresponsibilityforthematerials containedinanythirdpartywebsitereferencedinthiswork.

Acknowledgments

Nineofthepaperscollectedherehavebeenpreviouslypublished,orintwocases,are forthcominginotherpublications.Ithanktheeditorsandpublishersforpermission toreprintthemhere.Thedetailsaboutthesourcesarelistedonpageiv.

Ihavetoomanyintellectualdebtstoacknowledgethemallatthispoint,butletme mentionsomeofthepeoplethathaveaffectedmythinkingonalltheissuesIdiscuss inthechaptersofthisbook.

ManyoftheideasinthesepapersweredevelopedinseminarsIgaveatMITandat ColumbiaUniversityonconditionalsandontopicsinepistemologyoverthepast five orsixyears.Iwasfortunatetohavegroupsofverytalentedphilosophersparticipatinginthoseseminarswhosecriticalandconstructivecontributionstothediscussion helpedmetounderstandtheissues,andinfluencedmyresponsestothem.These includedJessicaCollins,NilanjanDas,KevinDorst,JeremyGoodman,DanGreco, BrianHedden,DanHoek,SophieHorowitz,JensKipper,HarveyLederman,Matt Mandelkern,DamienRochford,BernhardSalow,MiriamShoen field,Jonathan Vogel,andIanWells.

EpistemologyhasbeenalivelyareaofresearchatMIT,andoverawiderrange oftimeIhavebene ﬁtedfromthealmostconstant ﬂowofstimulatingdiscussion, informalandinreadinggroupsaswellasseminars,withgraduatestudentsand colleagues.Inadditiontothosealreadymentioned,Iwanttothankthefollowingwho havehelpedmetounderstandtheseissues,bothduringtheirtimeatMITandlater: RayBriggs,AlexByrne,AndyEgan,AdamElga,NedHall,CasparHare,JustinKhoo, SarahMoss,MiloPhilips-Brown,AgustinRayo,GingerSchultheis,JackSpencer, JasonStanley,EricSwanson,ZoltanSzabo,RogerWhite,SteveYablo,andSeth Yalcin.

MydebtstoTimothyWilliamson,DorothyEdgington,andDavidLewiswillbe evidentthroughoutthesepapers.Witheach,thereistherightmixofagreementand disagreementtomakeforfruitfuldiscussion.Eachhashadaprofoundinﬂuenceon myideas.

Thanks,asalways,toPeterMomtchiloffforhissupportandadvice,andtoDavid Balcarrasforeditorialhelp.

DetailsofFirstPublication

Ithanktheeditorsandpublishersforpermissiontoreprintthefollowingpreviously publishedpapersinthiscollection.

Chapter2 ﬁrstappearedasRobertStalnaker, “LuminosityandtheKKThesis” in Externalism,Self-knowledgeandSkepticism,editedbySanfordC.Goldberg, 17–40,CambridgeUniversityPress.Copyright©2015.Reprintedbypermissionof CambridgeUniversityPress.

Chapter4 ﬁrstappearedasRobertC.Stalnaker, “ModelingaPerspectiveonthe World” in AboutOneself:DeSeThoughtandCommunication,editedbyManuel García-CarpinteroandStephanTorre,121–37.Copyright©2016.Reprintedby permissionofOxfordUniversityPress:https://global.oup.com/academic/product/ about-oneself-9780198713265.

Chapter7alsoappearsasRobertC.Stalnaker, “ExpressivismandPropositions,” forthcomingin UnstructuredContent,editedbyDirkKindermann,AndyEgan,and PetervanElswyk,andisreprintedbypermissionofOxfordUniversityPress.

Chapter10 ﬁrstappearedasRobertC.Stalnaker, “ConditionalPropositions andConditionalAssertions,” in EpistemicModality editedbyBrianWeatherson andAndyEgan,227–48.Copyright©2011,andisreprintedbypermission ofOxfordUniversityPress:https://global.oup.com/academic/product/epistemicmodality-9780199591589.

Chapter11alsoappearsasRobertC.Stalnaker, “CounterfactualsandProbability,” forthcomingin Conditionals,ParadoxandProbability:ThemesfromthePhilosophy ofDorothyEdgington,editedbyLeeWaltersandJohnHawthorne,andisreprinted bypermissionofOxfordUniversityPress.

Chapter12 ﬁrstappearedasRobertStalnaker, “CounterfactualsandHumean Reduction, ” in ACompaniontoDavidLewis,editedbyBarryLoewerandJonathan Schaffer,411–24,Wiley-Blackwell.Copyright©2015.Reprintedbypermissionof theeditors.

Thefollowingchaptersarepublishedhereforthe ﬁrsttime:

“Reﬂection,Endorsement,Calibration ”

“RationalReﬂectionandtheNotoriousUnmarkedClock”

“ContextualismandtheLogicofKnowledge” “DispositionsandChance.”

Introduction

MorethanthirtyyearsagoIwroteabookcalled Inquiry.Thiswasagreattitlefora philosophybook,withitsallusion(orhomage)toclassicworksintheempiricist tradition,anditwasanappropriatetitlefortheaspirationswithwhichthebookwas written:itstopic,Isaidinthepreface,wastheabstractstructureofinquiry.Butitis lessclearthatthiswasanappropriatetitleforwhatwasactuallyaccomplishedinthe booksinceitdidnotgetmuchbeyondpreliminarysettingupoftheissues,andsome expositionofandmotivationfortheformalapparatusthatIplannedtousetotalk aboutthestructureofinquiry.Beforegettingtothemainissues,Ihadtoexplainand motivatemyapproachtotheproblemofintentionality,sketchandmotivatethe formalapparatususedtorepresentthatapproach(possibleworldssemantics),and respondtoproblemsthattheapproachfaced.Thattookupmostofthebook.The restofitfocusedmainlyonanotherpieceofapparatusneededtorepresentthe dynamicsofbelief(aformalsemanticsforconditionals),andIwasabletomakeonly astartonadiscussionoftheroleofthisapparatusinformingandreﬁningbothrules forrevisingbeliefs,andconceptsforgivingatheoreticaldescriptionoftheworld.

Isaidatthetime(againinthepreface)thatIhadbegunthatprojectwithanaïve hopethatIcouldgettothebottomoftheproblemsIwasconcernedwith,butthat IhadlearnedthatthebottomwasfurtherdownthanIthoughtandsowasthen preparedonlytomakeapreliminaryprogressreport.Thepresentcollectionisa furtherprogressreportonthesameproject,butIhavechangedmymindabout gettingtothebottomofthings.I’vedecidedthereisnobottom:thebestwecandoin philosophyistochipawayatbitsandpiecesoftheproblems.Wecanpaint impressionisticbigpicturesthatwehopewillgetonetoseetheissuesinanewand betterway,andwecanconstructmodelsthatachieveprecisiononlyatthecostof idealizationandsimpli ﬁcation,butthatwehopewillthrowsomelightonthe phenomena.Thatmaybeenoughtocountasprogress.

Onegainssomeperspectivefromputtingacollectiontogether,seeingconnections andrecurrentthemesthatonehadnotnoticedwhenworkingontheindividual papers.OnethingthatstoodoutformeasIselectedpapersforthiscollection,and addedtothemto ﬁllingaps,wasthecontinuitywiththeearlierbook,eventhoughall butoneofthepapersinthiscollectionwerewrittenmorethanthirtyyearsafter Inquiry waspublished.Thiscollectionalsohastwoparts,papersonknowledgeand papersonconditionals,andthesepapersdiscussthesamethemesdiscussedinthe twopartsoftheearlierbook.Thefocusofthe ﬁrstparthaschangedfrombeliefto knowledge,butIhavecometoseethattheproblemofintentionality(atleastonmy wayofapproachingit)isessentiallythesameastheproblemofcharacterizing knowledge.Knowledgewhether ϕ,accordingtoasloganIlike,isthecapacityto

makeone’sactionsdependonwhether ϕ.Knowledgeisamatterofcausalsensitivity tofactsthatarethesubjectmatterofone’sknowledge.Myearliergesturesat explainingintentionalitytookasimilarform:Itookbeliefanddesiretobethe basicintentionalstates,butarguedthatbeliefstatesgettheirintentionalcontent fromtheinformationthattheytendtobesensitiveto(undercertainnormal conditions).LookingbackfromthelaterperspectiveofTimothyWilliamson ’ s generalpictureofepistemology,Icametoappreciatethatmyaccountofintentionalityisreallyaversionofhis “knowledge ﬁrst” view:beliefiswhatwouldbe knowledgeiftherelevantnormalconditionsinfactobtained,ortoputittheother wayaround,knowledgeisfullbeliefwhenitisnon-defective.

Thepapersinthesecondhalfofthiscollectiondevelopfurthertheideasabout conditionalsthataresketchedinthelastthreechaptersof Inquiry:theirrolein epistemology,themetaphysicalstatusofthepropositionstheyexpress,andtheir relationtoprobabilisticconcepts,bothcredenceandchance.Intheearlierbook IsketchedanddefendedwhatIcalledtheprojectionstrategyforexplainingobjective modalconceptsasakindofprojectionofepistemicstatesandpoliciesontothe world,arguingthatthisstrategyhelpedtoexplaintherelationbetweenthetwokinds ofconditionals(indicativeandsubjunctive).ThestrategyhasitsrootsinHume,but IcontrasteditwiththekindofreductionistHumeanprojectthatDavidLewis developed.Onmyanti-reductionistaccount,theresultoftheprojectionisconcept formationthatreﬁnesourdescriptiveresourcesfordistinguishingbetweenthe possiblewaysthattheworldmightbe.Inthepapersinthesecondpartofthis collectionIlookinmoredetailatthesesameissues.

Iwillsketchinbroadstrokesthepictureofepistemologythatisguidingme,and thentrytoputtheindividualpapersincontextbysayinghowIseetheirrelationto thisbigpicture.

Themainproblemofepistemologyistoexplainhowwecognitivebeingsareable to findourwayaboutintheworld:howdoweacquireandusetheinformationabout ourenvironmentthatweneedtosucceedinit?Eventhesimplestanimalsacquire anduseinformation,andthey(alongwithsimpleartifacts)provideusefulmodelsof knowledge,butonethingthatdistinguishesthekindofcognitivebeingswearefrom thesesimplecasesisthatwecanreflectonourselvesascognitivebeings;partofthe informationweareabletoacquireanduseisinformationaboutourownplacein theworld informationabouthowweareabletoacquireanduseinformation.The pointisnotjustthatoneoftheinquirieswecanengageinisepistemology.Itisthat anyinquirywillinvolveatleastimplicitconsiderationofthemethodsweareusingto reachtheconclusionswereach,andwhenwearesurprised whensomethingwe tookourselvestoknowisshowntobefalse weareforcedtoreflectonwhatwent wrong:whatassumptionweweremaking,perhapsimplicitly,aboutourepistemic connectionstotheworld,andwhatchangesweneedtomakeinthoseassumptionsto recoverfromourmistakes.Theupshotisthatoneoftheimportantchannelsof informationinvolvedinouracquisitionandrevisionofknowledgeisinformation aboutourselvesandourplaceintheworld.

Aseconddistinctivefeatureofthekindofcomplexcognitivebeingsthatweareis thatwearesocialcreatureswhorelyontheknowledgeofothers.Thatis,knowers otherthanourselvesareinvolvedinthechannelsthroughwhichwereceive

information.Criticalreflectiononthesechannelsofinformationwillbereflection, fromathirdpersonperspective,onhowitisthatcognitivebeingslikeourselvesare ableto findtheirwayaround,what their sourcesofinformationare,andwhatthe worldislikefromtheirperspectives.

So,wedevelopaconceptionoftheworldaprominentpartofwhichisourselves andotherslikeus rationalagentswhoaredevelopingandreﬁningaconceptionof theworldtheyarein.Clarifyingaconceptionofthiskindwillinvolveconsidering differentperspectivesontheworld,andrelationsbetweenthoseperspectives.Ina sense,wearelookingatourselvesfromtheoutside,asagentswhoseinteractionswith natureandwithotheragentsarepartofanobjectiveworldtobedescribedand explained.Butwealsorecognizethatwearen’treallyoutside:Ourthird-personview ofourselvesisdevelopedandreﬁnedwithintheworld,fromperspectiveswithinit. Gettingclearabouttherelationshipsbetweendifferentcognitiveperspectives that ofthetheorist,thatofoneselfatthemoment,thatofoneselfatrememberedand anticipatedtimes,andthatofothers isoneoftheaimsinmanyofthepapersinthis collection.

Thepictureisanaturalisticonethatseescognitivebeingsaspartofthenatural world.TakingapagefromHume,thisnaturalisticpicturegivesnoroletopure reason,beyondtherequirementsofconsistencyandcoherence,initsaccountof inductiveknowledge.ItaketheupshotofHume’sskepticalargumentthatreason cannotjustifyinductivepractice,andhisjudgmentthatallreasoningaboutmatters offactisbasedoncauseandeffecttobesomethinglikethis:wecan’tseparatethe taskofdevelopingandjustifyingrulesfor findingoutabouttheworldfromthe substantivetaskofdevelopingaviewaboutwhattheworldislike.Weapproachboth tasksfromwithin,criticizingandrefiningthemethodsandbeliefsthatwe find ourselveswith.

Todevelopandsharpenthispicture,ithelpstohavesomeformaltools.Thebook begins,inchapter1,withareviewofaformalsemantics,inthepossible-worlds framework,ofknowledgeandbelief.Afeatureofthiswayofmodelingknowledge (pioneeredbyJaakkoHintikka)isthatitprovidesawayofrepresentingpropositions aboutwhatanagentknowsaspropositionsthatarethemselvesthecontentsof knowledge.Intheearlytheoriesofthiskind,justasingleknowerwasmodeled,but theframeworknaturallyextendstoatheorywithmultipleknowerswhohave knowledgeandbeliefsabouttheknowledgeandbeliefsofeachother,sothisisan appropriateframeworkfordevelopingthegeneralpicturesketchedabove.Inthe particularversionofamodeltheoryforknowledgeandbeliefthatIsketchinthis chapter,assumptionsaremadethatpermitbelieftobereducedknowledge,whichis appropriatetothe “knowledge ﬁrst” ideologythatisimplicitintheinformationtheoreticpictureofknowledge.Anditprovidesaframeworkforclarifyingquestions aboutfurtherconstraintsontherelationbetweenknowledgeand(full)belief(where yourfullbeliefsare,roughly,thepropositionsyourightlyorwronglytakeyourselfto know).Belief,inthissense,andknowledgewillcoincidewhenoneisrightabout everything,butwecanconsiderhowmuchwecangeneralizeaboutwhatanagent knowswhensomeofwhatshetakesherselftoknowisfalse.Thatis,whatcanone say,atthislevelofabstraction,abouttheextenttowhicherrorsinsomeofour knowledgeclaimsinfectothersofourknowledgeclaims,andtheextenttowhich

someoftheseknowledgeclaimscanbeisolatedfromothers.Iposethisquestionin themodeltheoreticframework,anddrawsomeconnectionsbetweenanswerstoit andproposalsthataroseintheverydifferentpost-Gettierprojectoftryingtoanalyze knowledgeintermsoftruebelief,plussomefurthercondition.

Thelogicofknowledgesketchedinthe ﬁrstchaptermakessometransparencyor luminosityassumptionsthatarecontroversial.While,asIhavesuggested,mypicture conformsinmanywayswithWilliamson’sexternalistepistemology,itdiverges sharplyfromhisontheissueofluminosity.ButIargueinthesecondchapter,with thehelpofsomesimplemodelsofinformation-carryingdevices,thatthese assumptions particularlytheassumptionthatonewhoknowsthat ϕ isinaposition toknowthatsheknowsit canbereconciledwithathoroughlyexternalistconceptionofknowledge.

Asecondpieceofformalapparatusthatisrelevanttothedynamicdimensionof thepictureisaformaltheoryofbeliefrevision.Themaintaskofthestandardbelief revisiontheoryistospecifyconstraintsonthewayasubjectisdisposedtochangehis overallbeliefstateasaresultofdiscoveringthatsomepriorfullbeliefisfalse.The thirdchapterisacriticaldiscussionofsomeattemptstoextendthestandardtheory togiveanaccountofthewayone’sbeliefrevisionpolicies,aswellasone’sbeliefs, shouldchangeinresponsetothediscoverythatapriorbeliefisfalse.Someelegant theoriesofiteratedbeliefrevisionhavebeenproposed,andtheyhelptoclarifythe terrain,butIarguethattheyallfacecounterexamples.Althoughthemainpoints Imakeinthischapterarenegative,thecounterexamplespointtotheimportanceof meta-information theagent’sknowledgeandbeliefsaboutherownepistemic situation inbeliefrevision.Afullysatisfactorybeliefrevisionwillinvolvean explanationofwhyrevisionwasrequired ofwhatdeviationfromnormalcondition ledonetotakeoneselftoknowsomethingthatonedidnotknow butoftenone learnsonewasmistakenwithoutlearningwhy,andthiscomplicatestheprocessof rationalbeliefrevision.

Thefourthchapterisaboutthewayself-locatingknowledgeandbeliefshouldbe represented.OnthepictureofcognitivebeingsthatIamworkingwith,allknowledge is,inasense,self-locatingsinceallofanagent’srepresentationsgettheircontent fromthatagent’srelationtothethingsthoserepresentationsareabout.Iarguethat thestandardwayofthinkingaboutself-locatingbelief,whichdistinguishessharply betweenknowingwhatpossibleworldyouareinandknowingwhereyouareinthe world,isconfused.Youcanknowwhatcountryyouareinwithoutknowingwhere youareinthecountry,but(Iargue)ignoranceofwhereyouare,orwhattimeitis,is alwaysignoranceaboutwhattheworldislike,whichistosay,aboutwhatpossible worldyouarein.Modelsthatrecognizethiscanhelpgiveaclearerviewofthewaywe thinkabouttherelationsbetweenepistemicperspectives,sincechartingthoserelationsrequirescalibratingtherelationsbetweenthecontentsoftheattitudesof differentagents,andofthesameagentatdifferenttimes.

Fullbelief,onthepictureIamdeveloping,iswhatonetakesoneselftoknow,buta cognitivestatewillalsoincludedegreesofpartialbelief,anda “knowledge ﬁrst” epistemologymustconcernitselfwiththesemore ﬁne-grainedstatesaswell.Questionsabouttherelationshipsbetweendifferentcognitiveperspectiveswillinclude questionsabouttherelationshipsbetweenthecredencefunctionsofdifferentagents,

andofthesameagentatdifferenttimes.Chapters5and6areaboutreﬂectionor deferenceprinciples:principlesthatstateconstraintsonanagent’scredencesabouta credencefunctionotherthanherownatthattime:aboutherownanticipatedor rememberedcredences,aboutthecredencesofanotheragent,oraboutthecredences thatsheideallyoughttohave.Iargueinchapter5thatreﬂectionprinciplesabout oneselfatothertimesoraboutotherscanbedefendedontheconditionthatone endorses thoseothercredencefunctions,whichmeansthatonejudgesthattheyare therightcredencesfortherelevantagenttohave.Inchapter6,Iexploreapuzzle abouttheattitudesthatanagentshouldhaveabouttherationalityofherownpresent attitudes.

Theconceptionthatourcognitiveagentisformingandreﬁninghasanormative dimension.Herinquiriesaskwhattheworldislike,whileatthesametimeasking whatrulesandproceduressheshouldadopttoformbeliefsandpartialbeliefsabout whattheworldislike.Chapter7focusesonthenormativeorpracticaldimension.It sketchesaframework,developedbyAllanGibbard,forrepresentingamixof normativeandfactualbeliefs.WhileIendorseGibbard’sexpressivistframework, Irejecthisowninterpretationofthatframework,arguingthatitblurstheline betweenarealistandanexpressivistconceptionofnorms.Thischapterismostly aboutnormsingeneral,butinthelastpartIlookatsomewaysinwhichthis frameworkhelpstoclarifymorespeciﬁcquestionsaboutepistemicnorms,andthe waystheirapplicationisconstrainedbyfacts.

Theinformation-theoreticconceptionofknowledgeisnecessarilyacontextualist conceptionforthefollowingreason:Knowledgeclaimscanbemadeonlyagainsta backgroundoffactualpresuppositionssince,onthatconception,knowledgeisbased onnaturalisticcausalrelationsbetweenaknowerandtheenvironmentthatisknown. Butthepresuppositionsrelativetowhichknowledgeisdeﬁnedcanthemselvesalways bequestioned,andaddressingthosequestionsrequiresashiftinthecontext. Chapter8developstheinformation-theoreticversionofcontextualismaboutknowledge,comparingandcontrastingitwithacontextualisttheorydevelopedbyDavid Lewisthathasaverydifferentmotivation.

Thesecondpartofthebookcontainspapersthatfocusonconditionalpropositions:theirroleinrepresentingepistemicpolicies,theircontributiontothetheoreticalresourcesfordescribingtheworld,andtheirconnectionswithotherobjective modalnotionssuchasdispositionalpropertiesandchance.

Asnotedabove,allbutoneofthepaperswerewritteninthelasttenortwelve years.Theexceptionischapter9,my ﬁrstpaperonconditionals,whichisnowabout ﬁftyyearsold.Itisincludedheresinceitisthestartingpointofaprojectthatledtoa progressreportsixteenyearslater,andtoanotheronenow.Theformallogicand semanticsforconditionalsdevelopedinthispaperweresimilartothoseinatheory beingdevelopedindependentlyataboutthesametimebyDavidLewis,butthe philosophicalideasguidingourtwotheorieswereverydifferent.Myprojectwasless ambitiousthanLewis’s,disclaiminganyattempttoprovideareductiveanalysis.The aimwasjusttoclarifytheformalstructureofaconceptandtoprovidethesemantic apparatuswithsomeintuitivemotivation.Thetheory’saimwastodoforcounterfactualconditionalswhatKripke’spossibleworldssemanticsdidfortheconceptsof necessityandpossibility,whichwasmanifestlynotareductionofmodalconceptsto

somethingelse.Theprojectionstrategywasnotexplicitinthisearlypaper,butitwas preﬁguredintheappealtoRamsey’sexplanationofindicativeconditionalsto motivateananalysisthathadcounterfactualconditionalsasitsmaintarget.Ramsey’ s suggestionwasabouthowtodecidewhethertoacceptaconditional:addthe antecedent,hypothetically,toyourstockofbeliefs,andaccepttheconditionalif andonlyiftheresultinghypothesizedstockofbeliefsimpliestheconsequent.My question(afterextendingRamsey’ssuggestiontocovercaseswheretheantecedent wasincompatiblewithyourbeliefs)wasthis:Whatshouldthetruth-conditionsfora conditionalpropositionbeifthisisagoodwayofdecidingwhethertoacceptit? Aselectionfunctionfromapossibleworldplusapropositiontoapossibleworldin whichthepropositionistruewasthoughtofasanontologicalanalogueofafunction fromastateofbeliefplusapropositiontoahypotheticalstateofbelief,andsoasa kindofprojectionofarelationbetweencognitivestatesontotheworld.

Whilemyaccountofconditionalspresupposedthattheproblemwastogivetruthconditionsforconditionalpropositions,othersataboutthistimewerearguingthat oneshouldexplainconditionalsentencesassentencesforperformingadistinctive kindofspeechact,orforrepresentingadistinctivekindofconditionalattitude. ErnestAdamsdevelopedaprobabilisticsemanticsthatbeganwiththeideathata conditionalisassertablewhentheprobabilityoftheconsequent,conditionalonthe antecedent,ishigh,andDorothyEdgington,buildingonAdams’swork,developed somepowerfulargumentsforanon-propositionalaccount.Somephilosophers suchasAllanGibbardgaveadividedaccountofconditionals,sidingwithEdgington ingivingaconditionalassertionaccountofindicativeconditionals,butwiththe propositionalistsonsubjunctiveconditionals.BothEdgingtonandIaimedfor uniﬁedaccountsofthetwokindsofconditionals,butaccountsthatallowedfor andexplainedthedifferences.Inchapters10and11Idefendanecumenical approachtothedisputebetweenpropositionalistsandthosewhowanttoexplain conditionalsintermsofconditionalspeechactsandattitudes,arguinginchapter10 thattheconditionalassertionanalysiscanbeformulatedasalimitingcaseofthe propositionalanalysis,andthatitisusefultodososinceithelpstochartthe connectionsandcontinuitiesbetweenindicativeandsubjunctiveconditionals. Chapter11focusesonEdgington ’saccountofcounterfactuals,andontherelations betweencounterfactualsandobjectiveprobability.Iargueinthischapterthatthe propositionalaccountcanallowforindeterminacy,andcanbetterexplainthe phenomenasheusestodefendheraccountoftheroleofcounterfactualsinepistemic reasoning.

Chapter12developsandcriticizesDavidLewis’sreductiveanalysisofcounterfactuals,andtheHumeansuperveniencemetaphysicsthatunderliesit.Lewis’sHumean theorycontrastswiththeusualempiricistdefenseofaHumeanmetaphysics, whichtakesthesuperveniencebasetobeobservationalorphenomenalconcepts. ForLewis,thepropertiestowhichallelseistobereducedarethefundamental propertiesofphysics.HispicturealsocontrastswiththepictureIhavebeendeveloping,whichputscausalnotionsatthecenterbothofthedescriptiveresourcesfor describingtheworld,andoftherulesforourpracticeoflearningabouttheworld. Lewisseparatesmetaphysicalquestionsconcerningwhatthereisafactofthematter

aboutfromepistemologicalquestionsabouttheproperrulesforlearningaboutthose facts,whileItrytodrawconceptualconnectionsbetweenthetwo.

The ﬁnalchapter,aboutdispositionsandchance,isthemostdetaileddiscussionof theprojectionstrategy,andthemostexplicitdevelopmentofanapplicationofit,the applicationtotheconceptofobjectivechance.Iargueintheconclusionofthis chapterthatwhilethegeneralpictureisbasedonconstitutiveconceptualconnections betweenepistemicrulesanddescriptivetheoreticalconcepts(asphenomenalistand veriﬁcationisttheorieswere),itisneverthelessathoroughlyrealistmetaphysical picture.

PARTI Knowledge

1 OntheLogicsofKnowledge andBelief

1.Introduction

Formalepistemology,oratleasttheapproachtoformalepistemologythatdevelopsa logicandformalsemanticsofknowledgeandbeliefinthepossibleworldsframework, beganwithJaakkoHintikka’sbook KnowledgeandBelief (Hintikka1962).Hintikka’ s projectsparkedsomediscussionofissuesaboutiteratedknowledge(doesknowing implyknowingthatoneknows?)andabout “knowingwho,” andquantifyinginto knowledgeattributions.Muchlater,thiskindoftheorywastakenupandappliedby theoreticalcomputerscientistsandgametheorists.¹Theformalsemanticproject gainednewinterestwhenitwasseenthatitcouldbeappliedtocontextswith multipleknowers,andusedtoclarifytherelationbetweenepistemicandother modalconcepts.

EdmundGettier’sclassicrefutationoftheJustifiedTrueBeliefanalysisofknowledge(Gettier1963)waspublishedataboutthesametimeasHintikka’sbook,andit immediatelyspawnedanepistemologicalindustry aprojectofattemptingtorevise therefutedanalysisbyaddingfurtherconditionstomeetthecounterexamples. Revisedanalysesweremetwithfurthercounterexamples,followedbyfurther re fi nements.Thiskindofproject fl ourishedforsomeyears,buteventually becameaninternallydrivengamethat wasthoughttohavelostcontactwith thefundamentalepistemologicalquestionsthatoriginallymotivatedit.Thisway ofapproachingepistemologicalquesti onsnowseemshopelesslyoutofdate,but Ithinktheremaystillbesomeinsightstobegainedbylookingback,ifnotatthe detailsoftheanalyses,atsomeofthegeneralstrategiesofanalysisthatweredeployed.

Therewaslittlecontactbetweenthesetwoverydifferentepistemologicalprojects. The firsthadlittletosayaboutsubstantivequestionsabouttherelationbetween knowledge,belief,andjustificationorepistemicentitlement,orabouttraditional epistemologicalissues,suchasskepticism.Thesecondprojectignoredquestions abouttheabstractstructureofepistemicanddoxasticstates.ButIthinksomeofthe abstractquestionsaboutthelogicofknowledgeconnectwithtraditionalquestionsin epistemology,andwiththeissuesthatmotivatedtheattemptto findadefinitionof knowledge.Theformalsemanticframeworkprovidestheresourcestoconstruct modelsthatmayhelptoclarifytheabstractrelationshipbetweentheconceptof

¹SeeFaginetal.1995andBattigalli&Bonanno1999forexcellentsurveysoftheapplicationoflogicsof knowledgeandbeliefintheoreticalcomputerscienceandgametheory.

knowledgeandsomeoftheotherconcepts(beliefandbeliefrevision,causationand counterfactuals)thatwereinvolvedinthepost-Gettierprojectofdeﬁningknowledge. Andsomeoftheexamplesthatwereoriginallyusedinthepost-Gettierliteratureto refuteaproposedanalysiscanbeusedinadifferentwayinthecontextofformal semantictheories:tobringoutcontrastingfeaturesofsomealternativeconceptions ofknowledge,conceptionsthatmaynotprovideplausibleanalysesofknowledge generally,butthatmayprovideinterestingmodelsofknowledgethatareappropriate forparticularapplications,andthatmayilluminate,inanidealizedway,oneor anotherofthedimensionsofthecomplexepistemologicalterrain.

Myaiminthischapterwillbetobringoutsomeoftheconnectionsbetweenissues thatariseinthedevelopmentandapplicationofformalsemanticsforknowledgeand beliefandmoretraditionalsubstantiveissuesinepistemology.Thechapterwillbe programmatic,pointingtosomehighlyidealizedtheoreticalmodels,somealternativeassumptionsthatmightbemadeaboutthelogicandsemanticsofknowledge, andsomeofthewaysinwhichtheymightconnectwithtraditionalissuesin epistemology,andwithapplicationsoftheconceptofknowledge.Iwillbring togetherandreviewsomeoldresults,andmakesomesuggestionsaboutpossible futuredevelopments.AfterabriefsketchofHintikka’sbasiclogicofknowledge,Iwill discuss,insection2,theS5epistemicmodelsthatweredevelopedandappliedby theoreticalcomputerscientistsandgametheorists,modelsthat,Iwillargue,conflate knowledgeandbelief.Insection3,Iwilldiscussabasictheorythatdistinguishes knowledgefrombeliefandthatremainsrelativelynoncommittalaboutsubstantive questionsaboutknowledge,butthatprovidesadefinitionofbeliefintermsof knowledge.Thistheoryvalidatesalogicofknowledge,S4.2,thatisstrongerthan S4,butweakerthanS5.Intheremainingfoursections,Iwillconsidersome alternativewaysofaddingconstraintsontherelationbetweenknowledgeandbelief thatgobeyondthebasictheory:insection4IwillconsidertheS5partitionmodelsas aspecialcaseofthebasictheory;insection5Iwilldiscusstheupperandlower boundstoanextensionofthesemanticsofbelieftoasemanticsforknowledge;in section6Iwilldiscussaversionofthedefeasibilityanalysisofknowledge,andin section7asimpli fiedversionofacausaltheory.

ThebasicideathatHintikkadeveloped,andthathassincebecomefamiliar,wasto treatknowledgeasamodeloperatorwithasemanticsthatparallelsthepossible worldssemanticsfornecessity.Justasnecessityistruthinallpossibleworlds,so knowledgeistruthinall epistemically possibleworlds.Theassumptionisthattohave knowledgeistohaveacapacitytolocatetheactualworldinlogicalspace,toexclude certainpossibilitiesfromthecandidatesforactuality.Theepistemicpossibilitiesare thosethatremainaftertheexclusion,thosethattheknowercannotdistinguishfrom actuality.Torepresentknowledgeinthiswayisofcoursenottoprovideanykindof reductiveanalysisofknowledge,sincetheabstracttheorygivesnosubstantive accountofthecriteriafordeterminingepistemicpossibility.Theepistemicpossibilitiesaredeﬁnedbyabinaryaccessibilityrelationbetweenpossibleworldsthatisa primitivecomponentofanepistemicmodel.(Where x and y arepossibleworlds,and “R” istheaccessibilityrelation, “ xRy ” saysthat y isepistemicallypossibleforthe agentinworld x.)Theideawastogiveapreciserepresentationofthestructureof anepistemicstatethatwasmoreorlessneutralaboutmoresubstantivequestions

aboutwhatconstitutesknowledge,butthatsharpenedquestionsaboutthelogicof knowledge.Thisformofrepresentationwas,however,farfrominnocent,sinceit required,fromthestart,anextremeidealization:Eveninitsmostneutralform,the frameworkrequiredtheassumptionthatknowersknowalllogicaltruthsandallof theconsequencesoftheirknowledge,sincenomatterhowtheepistemicallypossible worldsareselected,alllogicaltruthswillbetrueinallofthem,andforanysetof propositionstrueinallofthem,alloftheirlogicalconsequenceswillalsobetrueinall ofthem.Therearedifferentwaysofunderstandingthecharacterofthisidealization: ontheonehand,onemightsaythattheconceptofknowledgethatisbeingmodeled isknowledgeintheordinarysense,butthatthetheoryisintendedtoapplyonlyto idealizedknowers thosewithsuperhumanlogicalcapacities.Alternatively,one mightsaythatthetheoryisintendedtomodelanidealizedsenseofknowledge theinformationthatisimplicitinone’sknowledge thatliterallyappliestoordinary knowers.Howevertheidealizationisexplained,thereremainthequestionswhether itisfruitfultodevelopatheorythatrequiresthiskindofdeviationfromreality,andif sowhy.²ButIthinkthesequestionsarebestansweredbylookingatthedetailsofthe waysuchtheorieshavebeen,andcanbedeveloped.

Themostbasictaskindevelopingasemanticsforknowledgeinthepossible worldsframeworkistodecideonthepropertiesoftheepistemicaccessibility relation.Itisclearthattherelationshouldbereﬂexive,whichisnecessarytovalidate theprinciplethatknowledgeimpliestruth,anassumptionthatisjustabouttheonly principleofalogicofknowledgethatisuncontroversial.Hintikkaarguedthatwe shouldalsoassumethattherelationistransitive,validatingthemuchmorecontroversialprinciplethatknowingimpliesknowingthatoneknows.Knowingand knowingthatoneknowsare,Hintikkaclaimed, “virtuallyequivalent.” Hintikka’ s reasonsforthisconclusionwerenotcompletelyclear.Hedidnotwanttobaseitona capacityforintrospection:heemphasizedthathisreasonswerelogicalratherthan psychological.HisproofoftheKKprinciplerestsonthefollowingprinciple:If {Kϕ,~K~ψ}isconsistent,then{Kϕ, ψ}isconsistent,anditisclearthatifonegrants thisprinciple,theKKprincipleimmediatelyfollows.³Thereasonforacceptingthis principleseemstobesomethinglikethis:Knowledgerequiresconclusivereasonsfor belief,reasonsthatwouldnotbedefeatedbyanyinformationcompatiblewithwhatis known.So,ifoneknowsthat ϕ while ψ iscompatiblewithwhatoneknows,thenthe truthof ψ couldnotdefeatone’sclaimtoknowthat ϕ.Thisargument,andother considerationsforandagainsttheKKprincipledeservemorecarefulscrutiny.There isatangleofimportantandinterestingissuesunderlyingthequestionwhetherone shouldaccepttheKKprincipleandthecorrespondingsemantics,andsomechallengingargumentsthatneedtobeansweredifonedoes.⁴ Ithinktheprinciplecanbe

²Iexploredtheproblemoflogicalomniscienceintwopapers,Stalnaker1991and1999b.Idon’t attempttosolvetheproblemineitherpaper,butonlytoclarifyit,andtoarguethatitisagenuineproblem, andnotanartifactofaparticulartheoreticalframework.

³Substituting “~Kϕ” for ψ,andeliminatingadoublenegation,theprinciplesaysthatif{Kϕ,~KKϕ}is consistent,then{Kϕ,~Kϕ}isconsistent.

⁴ Seeespecially,Williamson2000forsomereasonstorejecttheKKprinciple.IrespondtoWilliamson’ s mainargumentinStalnaker2015,reprintedaschapter2ofthisbook.

defended(inthecontextoftheidealizationswearemaking),butIwillnotaddress thisissuehere,provisionallyfollowingHintikkainacceptingtheKKprinciple,anda semanticsthatvalidatesit.

TheS4principles(Knowledgeimpliestruth,andknowingimpliesknowingthat oneknows)wereasfarasHintikkawaswillingtogo.Heunequivocallyrejectsthe characteristicS5principlethatifonelacksknowledge,thenoneknowsthatonelacks it(“unlessyouhappentobeassagaciousasSocrates”⁵),andherehisreasonsseemto beclearanddecisive:

Theconsequencesofthisprinciple,however,areobviouslywrong.Byitsmeans(togetherwith certainintuitivelyacceptableprinciples)wecould,forexample,showthatthefollowing sentenceisselfsustaining:

Thereasonthat(13)isclearlyunacceptable,asHintikkagoesontosay,isthatit impliesthatonecouldcometoknowbyreﬂectionalone,ofanytruth,thatitwas compatiblewithone’sknowledge.Butitseemsthataconsistentknowermight believe,andbejustiﬁedinbelieving,thatsheknewsomethingthatwasinfactfalse. Thatis,itmightbe,forsomeproposition ϕ,that~ϕ,andBKϕ.Insuchacase,ifthe subject’sbeliefsareconsistent,thenshedoesnotbelieve,andsodoesnotknow,that ~ϕ iscompatiblewithherknowledge.Thatis,~K~Kϕ,alongwith~ϕ,willbetrue, falsifying(13).

2.PartitionModels

DespiteHintikka’sapparentlydecisiveargumentagainsttheS5principle,later theoristsapplyingepistemiclogicandsemantics,bothintheoriesofdistributive computersystemsandingametheoryassumedthatS5wastherightlogicfor(an idealizedconceptof)knowledge,andtheydevelopedsemanticmodelsthatseemto supportthatdecision.Butwhilesuchmodels,properlyinterpreted,havetheirplace, Iwillarguethatthetheoristsdefendingthemconflatedknowledgeandbeliefinaway thathasledtosomeconceptualconfusion,andthattheyhaveabstractedawayfrom someinterestingproblemswithintheirintendeddomainsofapplicationthatmore generalmodelsmighthelptoclarify.Butbeforegettingtothisissue,letme firsttake noteofanotherwaythatmorerecenttheoristshavemodi fied,orgeneralized, Hintikka’soriginaltheory.

Hintikka’searlymodelsweremodelsoftheknowledgeofasingleknower,but muchofthelaterinterestinformalepistemicmodelsderivesfromaconcernwith situationsinwhichtherearemultipleknowerswhomayknoworbeignorantabout theknowledgeandignoranceoftheothers.WhileHintikka’searlyworkdidnotgive explicitattentiontotheinteractionofdifferentknowers,thepotentialtodosois implicitinhistheory.Boththelogicandthesemanticsoftheknowledgeofasingle

⁵ Hintikka1962,106.

⁶ Ibid.,54.InHintikka’snotation, “Pa” isthedualoftheknowledgeoperator, “Ka” : “ ~Ka~ ”.Iwilluse “M” for~K~).

knowergeneralizeinastraightforwardwaytoamodelformultipleknowers.One needsonlyaseparateknowledgeoperatorforeachknower,andinthesemantics,a separaterelationofepistemicaccessibilityforeachknowerthatinterpretsthe operator.Onecanalsointroduce,foranygroupofknowers,anoperatorforthe common knowledgesharedbythememberofthegroup,whereagrouphascommon knowledgethat ϕ ifandonlyifallknowthat ϕ,allknowthatallknowthat ϕ,allknow thatallknowthatallknow,etc.allthewayup.Thesemanticsforthecommon knowledgeoperatorisinterpretedintermsofanaccessibilityrelationthatisdeﬁnableintermsoftheaccessibilityrelationsfortheindividualknowers:thecommonknowledgeaccessibilityrelationforagroupGisthetransitiveclosureofthesetof epistemicaccessibilityrelationsforthemembersofthatgroup.⁷ IfRG isthisrelation, thentheknowerswhoaremembersofGhavecommonknowledgethat ϕ (inpossible world x)iff ϕ istrueinallpossibleworldsthatareRG relatedtoworld x.The generalizationtomultipleknowersandtocommonknowledgeworksthesameway, whateverassumptionsonemakesabouttheaccessibilityrelation,andonecandeﬁne notionsofcommonbeliefinanexactlyanalogousway.Thepropertiesofthe accessibilityrelationsforcommonknowledgeandcommonbeliefwillderivefrom thepropertiesoftheindividualaccessibilityrelations,buttheywon’tnecessarilybe thesameasthepropertiesoftheindividualaccessibilityrelations.(Thoughifthe logicofknowledgeisS4orS5,thenthelogicofcommonknowledgewillalsobeS4or S5,respectively).

Theoreticalcomputerscientistshaveusedthelogicandsemanticsforknowledge togiveabstractdescriptionsofdistributedcomputersystems(suchasofficenetworks oremailsystems)thatrepresentthedistributionand flowofinformationamongthe componentsofthesystem.Forthepurposeofunderstandinghowsuchsystemswork andhowtodesignprotocolsthatpermitthemtoaccomplishthepurposesforwhich theyweredesigned,itisusefultothinkofthemascommunitiesofinteracting rationalagentswhousewhatinformationtheyhaveaboutthesystemasawholeto servetheirowninterests,ortoplaytheirpartinajointproject.Anditisusefulinturn forthoseinterestedinunderstandingtheepistemicstatesofrationalagentstothink ofthemasanaloguesofthekindofsimpli fiedmodelsthattheoreticalcomputer scientistshaveconstructed.

Adistributedsystemconsistsofasetofinterconnectedcomponents,eachcapable ofbeinginarangeoflocalstates.Thewaythecomponentsareconnected,andthe rulesbywhichthewholesystemworks,constraintheconﬁgurationsofstatesofthe individualcomponentsthatarepossible.Onemightspecifysuchasystembypositing asetofncomponentsandpossiblelocalstatesforeach.Onemightalsoincludea componentlabeled “nature” whoselocalstatesrepresentinformationfromoutside thesystemproper. Global stateswillben-tuplesoflocalstates,oneforeach component,andthemodelwillalsospecifythesetofglobalstatesthatare admissible Admissibleglobalstatesarethosethatarecompatiblewiththerulesgoverningthe

⁷ Moreprecisely,ifRi istheaccessibilityrelationforknower i,thenthecommon-knowledgeaccessibility relationforagroupGisdeﬁnedasfollows: xRGy iffthereisasequenceofworlds, z₁,... zn suchthat z₁ = x and zn = y andforall j between1and n–1,thereisaknower i 2 G,suchthat zjRi zj+1