ConstructingPractical
Reasons ANDREASMÜLLER
GreatClarendonStreet,Oxford,OX26DP, UnitedKingdom
OxfordUniversityPressisadepartmentoftheUniversityofOxford. ItfurtherstheUniversity’sobjectiveofexcellenceinresearch,scholarship, andeducationbypublishingworldwide.Oxfordisaregisteredtrademarkof OxfordUniversityPressintheUKandincertainothercountries ©AndreasMüller2020
Themoralrightsoftheauthorhavebeenasserted FirstEditionpublishedin2020 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:2020935466
ISBN978–0–19–875432–9
PrintedandboundinGreatBritainby ClaysLtd,ElcografS.p.A.
LinkstothirdpartywebsitesareprovidedbyOxfordingoodfaithand forinformationonly.Oxforddisclaimsanyresponsibilityforthematerials containedinanythirdpartywebsitereferencedinthiswork.
2.Reasoningandreasonjudgements33
5.Correctreasoning141
5.1Correctandincorrectreasoning142
5.2Correctnessandreasons145
5.3Correctnesswithoutreasons156
5.4Theconstitutiverulesaccountofcorrectreasoning159
5.5Developingtheaccount163
5.6Howtodeterminetherulesofreasoning172
5.7Thestatusoftherulesofreasoning185
6.Mind-dependenceandobjectivity189
6.1Thevarietiesofmind-dependence190
6.2SidingwithEuthyphro197
6.3Fallibilityanduniversality200
6.4Modalrobustness206 Concludingremarks219
Acknowledgements
Thisbookgrewoutofmydoctoraldissertation,whichIsubmittedto Humboldt-UniversitätzuBerlininAugust2013.Alotoftimehaspassed sincethen,andwhilethepositionIdevelopinthisbookhasnotchanged, someoftheargumentsandalotofthetextthatpresentsthemhave. Writingthisbookandthedissertationonwhichitisbasedwouldnot havebeenpossiblewithoutthesupportofanumberofpersonsand institutions.Firstandforemost,IamgreatlyindebtedtomyPhD supervisoratHumboldt,ThomasSchmidt,whohassupportedand encouragedmenotonlythroughoutthetimeIworkedonthedissertation,butalsoalongtimebeforeIhadeventhoughtaboutpursuinga doctorateinphilosophy.Hisinsightfulcommentsandsuggestionswere immenselyhelpfulinshapingmythinkingaboutthetopicofthisbook. DuringaresearchvisitspentatthePhilosophyDepartmentofPrinceton Universityinthespringtermof2012,MichaelSmithwasverygenerous withhistimeandhiscomments,forwhichIamverygrateful.Our conversations,inPrincetonandonvariouslateroccasions,helpedme todevelopaclearerperspectiveonmanyoftheissuesthatIdiscussin whatfollows.
Manypeoplehelpedmetoimprovethematerialthatfound itswayintothisbook.Ihavegreatlybenefitedfromconversations withandcommentsfromMaikeAlbertzart,HannahAltehenger,Max Barkhausen,PhilippBrüllmann,AnneBurkard,AndreasCassee, ChristopherCowie,SinanDogramaci,GerhardErnst,DaanEvers, ChristophFehige,SimonGaus,JanGertken,StefanGosepath,Logi Gunnarsson,TimHenning,UlfHlobil,RebekkaHufendiek,Benjamin Kiesewetter,ChristianKietzmann,FelixKoch,ErrolLord,David Löwenstein,BarryMaguire,SusanneMantel,LeoMenges,Cory Nichols,HerlindePauer-Studer,PeterSchaber,OliverSchott,Moritz Schröder,SharonStreet,ChristineTiefensee,BarbaraVetter,Jay Wallace,RalphWedgwood,andJackWoods,aswellastwoanonymous
readersforOUP.Iwouldliketoexpressmygratitudetoallofthese people,andtothoseIhaveforgottentomention.Iamalsomuchobliged toTimoJunger,whowasagreathelpinpreparingthe finalmanuscript.
Forvaluablediscussionsandcomments,Iwouldliketothankvarious audiencesinBerlin,Essen,Frankfurt,Munich,Münster,Konstanz, Leipzig,Saarbrücken,Tübingen,andVienna,whereIhadtheopportunitytopresentearlierversionsofsomeofthematerialincludedhere.Iam alsogratefultoDominikPerlerandhisLeibnitz-Preis-Projekt,tothe GraduateSchoolofPrincetonUniversity,toHumboldt-Universitätzu BerlinandtheSonderforschungsbereich644,totheStudienstiftungdes deutschenVolkes,andtotheCentreforAdvancedStudyinBioethicsin Münsterforsupportingmeandmywork,aswellastotheUniversityof Bernandmycolleagueshere,forprovidingmewiththeopportunityto continuedoingphilosophyinsuchanidealenvironment.Finally,Iowea specialdebtofgratitudetomyparents,whoseunconditionalsupporthas helpedmeoneverystepoftheway,andtoMoritzSchröder,whohas listenedtomythoughtsaboutthisbookanditscontentswithunfailing patienceandencouragement.
Chapter2containsmaterialpreviouslypublishedinA.Müller(2019). Reasoningandnormativebeliefs:nottoosophisticated. Philosophical Explorations 22(1),2–15,whichisreprintedherebypermissionofthe publisher(Taylor&FrancisLtd).
Introduction
Agentshavereasons:somethingsarereasonsforthemtoperformor refrainfromcertainactions.Thatitwillsparehergreatpaininthe future,forexample,isareasonforanagenttogotothedentistnow, thatshepromisedherfriendtohelphimmoveisareasonforhertodo so,andthatsheisalreadylateforworkisareasonforhernottoreadthe nextarticleinthemorningpaper.Whatmakeseitheroftheseconsiderationsareasonfororagainsttherespectiveaction?Whydowehavesuch reasons?Insomecases,anansweriseasytocomeby.Youmighthavea reasontobuyatrainticketsimplybecauseyouhaveareasontogetfrom BerlintoHamburg.Here,yourreasontobuytheticketcanbeexplained intermsofanotherreasonfromwhichitderives.Butnotallreasonscan bederivativereasons,sothiskindofanswerwillnotalwaysbeavailable. Canwegiveamoregeneralexplanationwhysomethingisareasonforan agent(not)toperformacertainaction,onethatappliestoallsuch reasons?
Somephilosophers,includingThomasScanlon(1998,ch.1;2014)and DerekParfit(2011a;2011b),arescepticalaboutthefeasibilityofthis explanatoryproject.Theydenythatsuchageneralaccountofour practicalreasons thatis,ourreasonsfororagainstperformingcertain actions canbeprovided.Wemightbeabletoexplain,ineachcase,why thereason-givingfactholds,suchasthefactthatgoingtothedentistnow willsparesomeonegreatpaininthefuture.Butthereasonfactitself,the factthatthisfactabouttheconsequencesofgoingtothedentistisa reasonfortheagenttogo,cannotalwaysbeexplainedinaninformative way:atleastsomereasonfactsareprimitive,fundamentalfacts,factsat whichourexplanationscometoanend.Others,likeMichaelSmith (1994;2013)andMarkSchroeder(2007),aremoreoptimisticabout theprospectsofgivingageneralaccountofpracticalreasons.They emphasizethatwhatthereisreasontodoforanagentiscloselyrelated
towhatthatagentismotivatedtodo,orwhatshewouldbemotivatedto doifcertainconditionsweremet.Thisrelation,theyclaim,suggeststhat factsaboutwhatisareasonfordoingwhatcanbeexplainedintermsof factsabouttherelevantagent’sactualorhypotheticalmotivationalstates, suchasherdesires.
Thisbookexploresanotherattempttoprovideageneralexplanation ofpracticalreasons,onethatisofferedbyaconstructivistaccountof thosereasons.Constructivismsharestheoptimismofthesecondkindof viewmentioned;it,too,purportstogiveageneralandinformative answertothequestionwhycertainreasonfactsobtain.Butitsuggests adifferentapproachtodevelopingsuchanaccount:tounderstandwhy somethingsarereasonsforcertainactions,wehavetostartbythinking abouttherelationbetweenthosereasonfactsandanagentwhoacknowledgesthembyformingthecorrespondingreasonjudgements.Thebasic ideaofconstructivismisthatreasonsarenotdiscovered,butmade.That is,therelationbetweenareasonfactandtheagent’scorresponding judgementisnotoneofdiscovery;instead,thosefactsmustbeconsideredtheresultofsomeactivityinwhichagentsareengaged.To explainwhysomethingisareasonforanagenttoperformacertain action,wehavetoexplainhowtheyaremade,whatactivitytheyresult from,andwhatrolereasonjudgementsplayinthatactivity.
AsaviewthatiscloselyassociatedwiththeworkofJohnRawlsand ChristineKorsgaard,constructivismhasattractedconsiderableattention overthelastfewdecades.Muchofthatattention,however,hasbeen critical.Theideathatconstructivismaboutpracticalreasonsmightoffer aninterestingandplausiblealternativetothemoreestablishedviewsin themeta-ethicaldebateisoftenmetwithscepticism(see,e.g.,ShaferLandau2003,ch.2;HussainandShah2006;Enoch2009;Wallace2004; 2012;andRidge2012).Ontheotherhand,thereislittleagreementasto whatexactlyaconstructivistaccountofpracticalreasonsconsistsinand whatitscentralclaimsare,overandabovethebasicideasketchedinthe previousparagraph.Hence,beforewecanproperlyassessitsmerits,we needabetterunderstandingofwhatconstructivismactuallyamountsto asanaccountofpracticalreasons.
Mymainaiminthisbookisthustoelaboratethebasicconstructivist ideaintoafully fledgedaccountofpracticalreasons,tomakeits
theoreticalcommitmentsexplicit,andtodefenditagainstsomewellknownobjections.Doingsowillrequiremetoaddressavarietyof differenttopics.¹Giventhelimitedspaceavailable,itwouldbeimpossibletodiscussallofthesetopicsasextensivelyastheydeserve,andto addressthemanypositionsthatonecouldtakewithrespecttoeachof them.Ihavethuschosentofocusonthoseaspectsthataremostrelevant forelaboratingthedetailsoftheconstructivistpositionandtoaddress onlythoseproblemsandopposingviewsthatposeanimmediatethreat totheoverallcoherenceofthatposition.Asaconsequence,myoverall argumentisprimarilyconstructive:whatIhopetoestablishisnotthat constructivism must beadoptedorthatitoffersabetteraccountof practicalreasonsthanitscompetitors,butthatit can beadoptedand thatitoffersacoherentandprimafacieattractiveaccountthatavoids manyoftheproblemsandconcernswithwhichitisconfrontedand hencedeservesfurtherdiscussion.Aboveall,thisbookismeantto facilitatesuchadiscussion.
Beforewebegin,thereareafewclarificatoryremarksthatIwouldlike tomakeinadvancesoastoavoidpotentialmisunderstandings.Firstly, constructivismaboutpracticalreasonsisameta-ethicalviewthatmust bedistinguishedfromviewsinmoralorpoliticalphilosophythatalsogo bythename ‘constructivism’,suchasJohnRawls’s(1971)positionon thefundamentalprinciplesofjusticeorThomasScanlon’s(1998,partII) positiononmoralwrongness.Thedifferencebetweenconstructivismas ameta-ethicalpositionandconstructivismasa first-ordernormative positionliesinthequestionseachpositionissupposedtoanswer.Firstordernormativeviewsanswerquestionslikethese:whenisthedistributionofascarceamountofgoodsjust?Isitsometimespermissibleto breakapromise?Ishappinessanintrinsicvalue?Whyisitwrongtolieto someone?Isthereareasontodevelopone’stalents?
Meta-ethicalpositions,ontheotherhand,areconcernedwiththe metaphysicalstatusofnormativefacts,suchasfactsaboutwhatisa reasonforwhat,andtheymightofferanexplanationforwhythethings thata first-ordernormativetheoryidentifiesasreasonsforacertainkind
¹Ipresentthemaintenetsofaconstructivistaccountofpracticalreasonsandgivean overviewofhowthisbookproceedswithelaboratingthataccountinChapter1below.
ofactionarereasonsforthataction.Theyarealsoconcernedwiththe natureofnormativejudgements,withwhetherandinwhatsensethey canbesaidtobetrueorfalse,withthemeaningoftermslike ‘just’ , ‘ wrong ’ ,or ‘ reason ’,etc.Sowhatdistinguishesaformofmeta-ethical constructivismfrom,say,Scanlon’smoralconstructivismisnotthatthe latteris ‘merely’ aboutmoralwrongness,whereastheformerisconcernedwithpracticalreasonsingeneral,fora first-ordernormative positionmightsharethiswiderscope.Thedifferenceliesinwhatquestionstheyaddressaboutwrongnessorreasonsandourjudgements abouteither.
Thatisnottosay,however,thatmeta-ethicaland first-ordernormativepositionsareentirelyindependentofeachother,orthatonecan remainentirelyneutralonanysubstantivenormativeissueswhile defendingafullyspecifiedmeta-ethicalview.Givingcertainanswersto someofthequestionswithwhichmeta-ethicalviewsareconcernedwill commityoutoacceptingordenyingcertainanswersto first-order normativequestions.Ameta-ethicalviewthatexplainsanagent’ s reasonsentirelyintermsofheractualdesires,forexample,willface significantdifficultiesinaccommodatingthesubstantivenormative claimthatsomereasonsaresharedbyallagentsregardlessofany differencesintheirmotivationaldispositions.Hence,thatmeta-ethical and first-ordernormativepositionsmustbedistinguishedbecausethey answerdifferentkindsofquestionsdoesnotmeanthatonecanbe defendedwithoutaneyetotheother.
Secondly,theconstructivistpositionthatIamconcernedwithhereis anaccountofpracticalreasons.Moreprecisely,itisanaccountoffacts aboutwhatisareasonforwhomtodowhat(andnotoftheordinary first-orderfactsthatgivetheagentareason).Itthushasawiderscope thananaccountof,forexample,moralfacts,butitdoesnotpurportto coverallnormativephenomena.Howbesttounderstandtheterm ‘normative’ hereisitselfacontroversialmatter.Someuseitonlyfor factsandjudgementsthatareabout orcloselyconnectedtofactsand judgementsabout reasons,whileothersuseitmorebroadlysoasto include,forexample,factsandjudgementsaboutcorrectness,rationality, virtue,etc.,evenifthosearenottobeunderstoodintermsofreasons (cf.Thomson2008).Butevenif ‘normative’ isusednarrowlytoapply
onlytotherealmofreasons,theconstructivistpositionunderconsiderationtargetsonlyasubsetofallnormativefacts,becauseitonlyoffersan accountofpracticalreasons,puttingotherkindsofreasons,suchas epistemicreasons reasonsforbelief tooneside.Thereisnogeneral reasontosuspectthattheconstructivistaccountofpracticalreasonsthat willemergeinthefollowingchapterscouldnotbeextended,withsome adaptions,tothecaseofepistemicreasons.Still,Ithinkthatitisnota trivialquestionwhetherepistemicandpracticalreasonsshouldreceive thesametheoreticaltreatment atleastthemerefactthatwetalkabout ‘ reasons ’ inbothcasesdoesnotsufficetoshowthatwearedealingwitha homogeneousphenomenonthatrequiresaunifiedaccount.Hence,the plausibilityofaconstructivistaccountofepistemicreasonsmustbe assessedseparatelyandonanotheroccasion.
Whatisconstructivism?
1.1Thebasicidea
Constructivism,asIunderstandithere,is firstandforemostaviewabout therelationbetweenthoughtanditsobject.Tobeaconstructivistabouta certaindomainistotakethefactsofthatdomaintobeanupshot,rather thanthemeasure,ofcorrectjudgementinthatdomain.Aconstructivist aboutpracticalreasonsthusdeniesthat,whenanagentcomestorealize thatsomethingisareasonforacertainaction,shedoesso inresponseto thecorrespondingnormativefact,thatis,afactaboutwhatisareasonfor doingwhat.Instead,suchfactsaresomehowexplicableintermsofour reasonjudgements.Thatdoesnotmeanthatthefactsarewhateverwe thinktheyare.Constructivistsdonotdenythatwecanbemistaken aboutwhatisareasonandwhatisnot.Buttheydosuggestthatnormative factsareultimatelygroundedinourmentalstatesoractivities.Inparticular,accordingtotheconstructivistaccountofpracticalreasonsthatIwill presentinthisbook,factsaboutwhatisareasonfordoingwhatobtainin virtueoftheactivityofpracticalreasoningandtherolethatthecorrespondingreasonjudgementsplayinit.Morespecifically,suchafactobtains becausethecorrespondingreasonjudgementistrue,andthejudgementis truebecausetheepisodeofreasoningthatitisapttoguideissound.
Inthis firstchapter,Iwanttostartbypresentingthemainaspectsof theconstructivistaccountIhavejustsketchedinabitmoredetail.Ido this, firstly,inordertogiveyouanoverviewandtomotivatethe remainingchaptersofthisbook,wheretheseaspectsandthequestions towhichtheygiverisewillbediscussedatmuchgreaterlength.ButIalso dothisbecausetheterm ‘constructivism’ hasbeenusedtopickouta varietyofratherdifferentpositionsinmeta-ethics,bothbyadvocatesand byopponentsoftheviewssolabelled.Hence,Iwanttobeasclearas possibleabouthowandwhyIusethatterm,toavoidaddingtothe
confusion.Letmebeclear,though,thatIdonotclaimthatmyproposed understandingofconstructivismsomehowsynthesizesorsubsumesall ofthevariousconceptionsofconstructivismthatcanbefoundinthe literature.¹TheviewthatIputforwardadoptsmanyimportantinsights fromthoseotherconceptions(asIwillpointoutalongtheway).Infact, littleofwhatIhavetosayinwhatfollowshasneverbeensuggested before.Butmyaimhereisnottoidentifyacommondenominator;Iam scepticalthatonecouldbefound.Thebasicideaofconstructivismis essentiallyametaphor:somethingsare constructed fromcertainother things;theyare made,ratherthandiscovered,asthewell-knownslogan hasit.Suchmetaphorscanbespelledoutindifferentways,noneof whichis,assuch,therightone.Nevertheless,somewaysmightbemore usefulormoreinterestingthanothers.ThepositionthatIpresentinthis bookismeanttobetruetotheideaencapsulatedinthatmetaphor,andit takesupsomeimportantthoughtsofparadigmaticconstructivistslike JohnRawlsandChristineKorsgaard,aswewillseeinthefollowing sections.²Butmostimportantly,itismeanttoofferadistinctandprima facieattractiveaccountofpracticalreasons.
1.2Constructivismandrepresentation
Letmenowelaborateonthemainaspectsoftheconstructivistposition thatIsketchedinsection1.1.Iwillstartwithitsaccountofthenatureof reasonjudgementsinthissection,andthenaddressitsunderstandingof truthaswellasthemind-dependentstatusitassignstoreasonfactsin sections1.3and1.4.
Constructivismrejectsafamiliarconceptionofthekindofattitude thatisinvolvedinthinkingaboutreasons.Considerthefollowing
¹Advocatesofconstructivistviewsinmeta-ethicsincludeJohnRawls(1980;1993),Christine Korsgaard(1996;2003;2009a;2009b),CarlaBagnoli(2002;2012;2013),AaronJames(2007; 2012),andSharonStreet(2008;2010;2012).Foracomprehensiveoverview,seeBagnoli(2017).
²ConstructivismisoftenconsideredtohaveaKantianpedigree,andbothRawlsand Korsgaard’sviewsarecertainlyinspiredbyKant.However,whetherKantisbestinterpreted asaconstructivistiscontroversialamongKantscholars,andIwillnotaddressthisissuehere. Forahelpfuldiscussionofmeta-ethicalconstructivismthatpaysspecialattentiontoitsrootsin Kant’sphilosophy,seeSchafer(2015a;2015b).
example.Laraandherfriendsareoutfordinner.Theyareabouttoorder fromthemenu,which,thisbeingaverysmallrestaurant,offersonlytwo optionsforthemaincourse:astuffedsquashandduckbreastmarinated inhoisinsauce.Laraisunsurewhichtochoose,sosheconferswithher friends,whohavebeenherebefore.Theytellherthatthesquash’ s stuffingisvegetarian,andthattheduckbreastcomeswithrice.Lara alsolearnsthathoisinsaucecontainssoyabeans,andthisshetakestobe areasonagainsthavingtheduckbreast(sheisallergictosoyabeans). Thisreasonjudgementsettlesherchoice,andsheproceedstoorderthe squash.WhatroledidLara’sjudgementthatthefactthathoisinsauce containssoyabeansisareasonagainstorderingtheduckbreastplayin theprocessthatleadtoherdecision?Whatkindofattitudedowe describeanagentashavingwhenweattributesuchareasonjudgement toher?
Thefamiliarconceptionanswersthosequestionslikethis:tobelieve thattheduckbreastcomeswithriceandthathoisinsaucecontainssoya beansistorepresentcertainfactsasobtaining,andwhenthosefacts reallydoobtain,asweshallassume,thenthoserepresentationsare accurate.Hence,byformingthosebeliefs,Laraacquiresnewinformation abouttheoptionsinachoiceproblemshecurrentlyfaces,information thatshecanapplyinmakingadecision.Thefamiliarconceptionextends thisaccountofthecognitiveroleofordinarybeliefstothecaseofreason judgements.Herreasonjudgement,³iftrue,alsoprovidesLarawith informationabouttheoptionsinherchoiceproblem,informationthat willhelphertocometoadecision.Ofcourse,thebeliefsandthereason judgementsdifferinthekindofinformationtheyprovide.Theformer providenon-normativeinformation,thatis,theyaccuratelyrepresent certainnon-normativefactsabouttheoptionsinquestion,suchasthe factthattheduckbreastcomeswithrice,orthathoisinsaucecontains soyabeans.Toformareasonjudgement,ontheotherhand,istoacquire
³Irefertotheseattitudesas ‘reasonjudgements’ ratherthan,e.g., ‘beliefsaboutreasons’ , becausethelatterisoftenassociatedwitharepresentationalistaccountofthefunctionofthe attitudesinquestion.Aswillbecomeclearshortly,constructivismrejectssuchanaccountofthe attitudes’ cognitivefunction.So,inordertoavoidprejudgingtheissue,IamusingwhatItaketo bealesscommittingterm. ‘Judgement’ isthusmeanttopickoutacertainkindofmentalstate, nottheeventofformingsuchastate.
normativeinformationaboutthechoicesituationandtorepresentsome normativefact,somefactaboutwhatisareasonforwhat,asobtaining. Butunderlyingsuchdifferencesincontent inwhatkindsoffactsthey areabout isafundamentalsimilarityintherolethosetwoattitudes playinthecognitiveeconomyofanagentwhofacesachoice.Justlike ordinarybeliefsaboutnon-normativematters,reasonjudgementsare representationalstateswhich,ifaccurate,providethesubjectwithinformationthatshecanapplyinmakingadecision.
Thedetailsofthisconceptioncanbe filledininnumerousdifferent ways,particularlywhenitcomestocharacterizingthefactsthatreason judgementspurporttorepresent.Arethey suigeneris normativefacts,or cantheysomehowbereducedtoother,non-normativefacts?Arethey causallyefficacious?Howcanwegainknowledgeofthem?Thesearejust someofthequestionsonwhichdifferentversionofthisconceptioncan disagree,andtheyarecertainlyimportantquestions.Nevertheless,we canputthemtoonesideforthemoment,becausethefundamental aspectsoftheconception,theonesthatwillbemostrelevantinwhat follows,arenotaffectedbyhowthesequestionsareanswered.These aspectsconcerntherelationbetweenthosereasonfactsontheonehand andourreasonjudgementsontheother,aswellastherolereason judgementsplayinthecognitiveprocessesofagentslikeLara. Accordingtothefamiliarconception,judgementsaboutwhatisareason forwhatrelatetocertainfactsinthesamewaythatthebeliefthathoisin saucecontainssoyabeansrelatestothefactthathoisinsaucecontainssoya beans:theypurporttorepresentthosefacts.Hence,theyprovideanagent withinformationthatisrelevanttoherchoice,iftheyrepresentaccurately.
Thisconceptionofthenatureofreasonjudgementsisquitepopularin contemporarymeta-ethics.Theideathatreasonjudgementspurportto representcertainfactsand,ifaccurate,provideagentswithinformation aboutthenormativeaspectsofhersituationseemstounderlie,inoneof itsguises,thepositionsdevelopedbyDavidEnoch(2007;2011a),Derek Parfit(2011b),ThomasScanlon(2003;2014),MarkSchroeder(2007), MichaelSmith(1994),andmanyothers.Itisoftennotsomucharticulatedexplicitlyashiddenintheassumptionthatreasonjudgementsand ordinarybeliefsaboutnon-normativemattersarethesamekindof attitudetowardsdifferentcontents.Ifyoutakethemaindifference
betweenthetwotoconsistinthefactstheyrepresent,itisunderstandablewhyyouwillfocusongivinganaccountofthefactsourreason judgementsareabout,whilesayinglittleaboutthekindofattitudeyou takethosejudgementstobe thereisnoneedto,ifitisthesameasour ordinarybeliefs.
Constructivismoffersanalternativetothisfamiliarconception.It suggeststhatthecognitivefunctionofreasonjudgementsdiffersfundamentallyfromthecognitivefunctionofordinarybeliefsaboutnonnormativematters.Accordingtotheconstructivist,reasonjudgements playadistinctroleinthepsychologyofadeliberatingagent:thepointof thosejudgementsisnottoprovideanagentwithadditionalinformation thatshecanapplyinmakingherdecision.Instead,reasonjudgements guide herinmakingadecisiononthebasisoftheinformationshehas. Touseacomputationalanalogy:whileordinarybeliefscanprovidethe inputfordecision-makingprocesses,reasonjudgementsprovidethe instructionsthatsuchprocessesfollowintranslatingthoseinputsinto outputs,thatis,intodecisions.Thus,whenLaraformsthebeliefthatthe hoisinsaucecontainssoyabeans,sheacquiresnewinformationabout oneofheroptions,butwhensheformsthejudgementthatthisisa reasonnottoordertheduckbreast,shedoesnotacquiremoreinformation.Instead,thisjudgementwillallowhertoapplytheinformation shealreadyhastothechoiceshefaces.Informationisnotenoughto cometoadecision;wealsoneedtohaveanideaofhowtobringanysuch informationtobearontheproblemofchoosingoneoftheavailable options.Thisisthedistinctcontributionreasonjudgementsmaketothe processofreasoneddecision-making.
Tobemorespecificaboutthemainpointofdisagreementbetween constructivismandthefamiliarconceptionofthenatureofreason judgementsoutlinedabove,letmeintroducethefollowingthesisabout whatkindofattitudereasonjudgementsare:
Representationalism
Reasonjudgementsarerepresentationalstates.Theircognitive functionistorepresentcertainfacts,andiftheydosoaccurately, theyprovidetheagentwithnormativeinformationbyindicating thatacertainfactconcerningwhatisareasonforwhatobtains.
ProponentsofthefamiliarconceptionacceptRepresentationalism,or oneparticularversionofit,whileconstructivistsrejectitaltogetherand takereasonjudgementstoplayafundamentallydifferentcognitiverole from,forexample,beliefsaboutthecompositionofsomecondiment. Constructivismisthusaformofanti-representationalismaboutthe natureofreasonjudgements.⁴
ConstructivistsarenottheonlyoneswhorejectRepresentationalism. Theysharethisnegativepartoftheirviewwithexpressivistssuchas SimonBlackburn(1998)andAllanGibbard(1990;2003).Thetwoviews differ,however,intheirpositiveaccountsofthekindofmentalstates theytakenormativejudgementstobe.Expressiviststypicallyconceiveof themasmotivationalordesire-likestates,thatis,statesthatarecharacterizedintermsoftheirroleintheexplanationofanagent’sactions. Constructivists,ontheotherhand,thinkthatreasonjudgementsarea distinctkindofmentalstate,thenatureofwhichcanonlybeunderstood byattendingtoitsroleinguidingtheagent’spracticalreasoning.Inother words,whereasexpressivismcharacterizesreasonjudgementsasakind ofpro-attitude,constructivismidentifiesthemintermsoftheirroleina processthatleadstotheformationofsuchattitudes.Thetwoviewsthus disagreesubstantiallyonthecognitiveroleoftheattitudesinquestion.⁵
Representationalismisalsorejectedbyglobalexpressivistssuchas RobertBrandom(1994)andHuwPrice(2011;2013).Moreover, Brandom’sclaimthatthefunctionofnormativevocabularyistomake explicit ‘anattitudethatotherwisecouldbeimplicitonlyinwhatis done namely,theendorsementofapatternofpracticalreasoning’ (1994,271)seemstoattributetoreasonjudgementsacognitiverole similartotheonedescribedinChapter2below.Theconstructivist theorypresentedinthisbookindeedbearsimportantsimilaritiestothe viewsofBrandomandPrice.Buttherearealsosignificantdifferences. Forone,constructivismaboutpracticalreasonsisa local theorythat
⁴ CarlaBagnoli(2002,125)agreesthatconstructivistsdonotconsidermoraljudgementto involvethe ‘aspirationtocorrectlyrepresentamoralreality’
⁵ SharonStreet(2008;2010;2012)alsoarguesthattheattitudesthatprovidethebasisforher constructivistaccountofpracticalreasonsmustbedistinguishedfromordinarydesires.Butshe neverthelesscharacterizesthemasattitudesthatare ‘bytheirnaturemotivating’ (2008,230), whichseemstomakeheraccountofnormativejudgementsmoredifficulttodistinguishfrom theexpressivist’sthanthereasoning-basedaccountthatIproposeinthisbook.
contraststhecognitivefunctionofreasonjudgementswiththatof,e.g., beliefsaboutone’senvironment itdoesnotrejectRepresentationalism acrosstheboard.ConstructivistsalsodonotshareBrandom’sandPrice’ s commitmenttodeflationism(cf.Chapter4).
JohnRawls,probablythemostinfluentialproponentofconstructivisminthetwentiethcenturyandtheauthorresponsibleforintroducing thatlabelintothemeta-ethicaldebate,⁶ agreesthatconstructivistsreject theideathatnormativejudgementsresembleordinarybeliefsinplaying arepresentationalroleinanagent’spsychology.Hecontrastsconstructivismwithaviewhecalls ‘rationalintuitionism’ . ⁷ Thisviewholdsthat ‘moral firstprinciplesandjudgements,whencorrect,aretruestatements aboutanindependentorderofmoralvalues’ andtheyare ‘knownby theoreticalreason’,whichmeansthattheirepistemologyisstructurally similartothatofempiricaljudgements: ‘moralknowledgeisgainedin partbyakindofperceptionandintuition’ (1993,91–92).Therational intuitionist’sconceptionoftherelationbetweenmoraljudgementsand moralfactsisthusverysimilartoRepresentationalism.Accordingto Rawls,constructivismrejectsthisconception,alongwiththe ‘ sparse conceptionofthepersonas ...a knower’ (1993,92)thatgoeswithit. Rawls’scharacterizationofconstructivismthusputstheviewinfundamentaloppositiontoRepresentationalism.
ChristineKorsgaardisthesecondmajorproponentofconstructivism inthelastdecades.Especiallyinhermorerecentwritings,sheputsthe contrastbetweenconstructivismandrealism,whichsheconsiderstobe itsmaincompetitor,intermsofthefunctionorpurposeofmoraland othernormativeconcepts:
[T]hedifferencebetweenarealistandaconstructivisttheoryrestsin thewaythetwoviewsunderstandthefunctionofconcepts ....A realist
⁶ ItisimportanttonotethatRawls’sviewsonconstructivismhavechangedovertime,orat leastshiftedinfocus.WhereashisDeweylectures(1980)aresimplyconcernedwith ‘Kantian constructvisim’,helaterdistinguishesbetween ‘politicalconstructivism’ and ‘Kant’smoral constructivism’ (1993,89–129).Onlythelatterofthesetwoiswhathecallsa ‘comprehensive moraldoctrine’,thatis,aviewthatalsoaddressesmetaphysicalandepistemologicalquestions andthusqualifiesasameta-ethicalview.Hence,IwillfocusonRawls’scharacterizationof constructivismassuchacomprehensivedoctrine.
⁷ Cf.Rawls(1980,557–64;1993,91–101).HeidentifiesSamuelClarke,RichardPrice,Henry Sidgwick,G.E.Moore,andDavidRossasparadigmaticadvocatesofrationalintuitionism.
believesthatthefunctionofconceptsistodescribetheworld,tomark outtheentitieswe findthere,whileaconstructivistbelievesthatthe functionof(atleastsome)conceptsistomarkout,inaschematicway, thesolutiontosomeproblemthatweface.(Korsgaard2008,22)
Constructivismthusrejectstheideathatthecognitivefunctionofnormativejudgementsistorepresentreality.Thatisnotwhatnormative conceptsarefor;their ‘cognitivejob’ isnotadescriptiveone(2003,105). Korsgaardthusclearlyrejectsarepresentationalistaccountofnormative thought.LikeRawls,sheunderstandsconstructivismasaviewthatis meanttoofferanalternativetoRepresentationalism.⁸
1.3Constructivismandtruth
Constructivismoffersanexplanationwhysomethingsarereasonsfor certainactions.Itexplainssuchreasonfacts,inthe firstinstance,interms ofthetruthofthecorrespondingreasonjudgement:theyobtainbecause thatjudgementistrue.⁹ Constructiviststhusdonotdenythatthereare factsaboutwhatisareasonfordoingwhat.Nordotheydenythatthose factscorrespondtotruereasonjudgements,aslongasthatclaimis understoodasasimplebiconditional:
Correspondence
Thejudgementthatpistrueifandonlyifitisafactthatp.
⁸ Heralternativeproposalforthefunctionofnormativeconceptsisadmittedlylessclear.On herpreferredwayofformulatingthatalternative,thoseconceptsdenote,ina ‘formal’ or ‘schematic’ way,thesolutiontosomeproblemwefaceinsofarasweareagents(1996,113). WhilethisdoesnotstrikemeasbeinginconflictwiththeviewthatIdevelopinthisbook,Iam notconfidentenoughinmyunderstandingofherproposaltomakethecaseforanysubstantive agreement.
⁹ Notethatwhatexplainsareasonfactaccordingtothisproposalisthe truth ofthe correspondingreasonjudgement,notthetruejudgement.Thatis,noactualjudgement understoodassomespecificpsychologicalreality isnecessaryforareasonfacttoobtain. Whatisrequired,rather,isthattheconditionsthatgroundthetruthofthatjudgementare actuallymet.
Thisclaimissymmetricalandthusleavesopenthequestionofexplanatoryprioritybetweenitsleft-andright-handsides.Itisonthislatter issuethatconstructivistsdisagreewiththeiropponents.Theymaintain thattheorderofdeterminationinCorrespondencegoesfromleftto right:theytreatreasonfactsastheexplanandumratherthantheexplanans.¹⁰ Whatmotivatesthisreversalistheconstructivist’srejectionof Representationalismasanaccountofthecognitivefunctionofreason judgements.Truth,itseems,iscloselyconnectedtosuccessfulfunctioning.So,ifthecognitivefunctionofreasonjudgementsisnottorepresent certainfacts,thentheirtruthwillnotbegroundedintheaccurate representationofsuchfactseither.Denyingthatreasonjudgementsare true,iftheyaretrue,becausetheyaccuratelyrepresentcertainreason facts,allowsconstructiviststoemployCorrespondenceaspartofan explanationofitsright-handside,ratherthanitsleft-handside.For them,areasonfactisnotthegroundofareasonjudgement’struth, butratheritsontologicalshadow.¹¹
Fortheirexplanatoryprojecttosucceed,constructivistsmustofferan alternativeaccountofwhatitisforreasonjudgementstobetrue.Here, thethreatofcircularitylooms.Toavoidit,theconstructivists’ accountof thetruthofreasonjudgementsmustnotpresupposethepriorexistence ofanyreasonfacts.Moreover,theiraccountmustunderstandtruthin termsthataresubstantiveenoughtobeartheexplanatoryweightthatthe
¹⁰ CompareBagnoli(2013,167–8),whowritesthatconstructivistsrejecttheclaim ‘that moraljudgmentsaretruth-evaluablebecausetheycorrespondtoorrepresentsomeportionof reality’ andthat,ratherthanholding ‘that[moral]factsaretruth-makers’,theyhold ‘thatsucha domain[offacts;A.M.]istheresultoftheactivityofpracticalreasoning’
¹¹Thiswayofdrawingthecontrastbetweenconstructivismanditsalternativestakescues fromKitFine’sdistinctionbetween ‘antifactualist’ and ‘factualist’ accountsofacertainpractice: Thefactualist’saccountmustinthissenseberepresentational:itmustlinkupthepractice withtheunderlyingfactsorsubjectmatter,whiletheantifactualist’saccountwillbe nonrepresentational.Intheonecase,thepracticemustbeseenasengagingwiththe possiblefactsanditmustbeunderstood atleast,inpart intermsofhowitengages withthosefacts.Intheothercase,thepracticeistakentobedisengagedfromthefacts; andratherthanunderstandingthepracticeintermsofhowitrepresentsthepossible facts,thefactsthemselvesshouldbeunderstoodintermsofhowtheyare ‘projected’ by thepractice.(Fine2001,24)
Note,however,thatdespitethesimilaritiesbetween ‘shadows’ and ‘projections’,constructivism differsimportantlyfromSimonBlackburn’sexpressivistprojectivism,whichhasprojectedfacts correspondtowhatevernormativejudgementanagentactuallyendorses;seeBlackburn(1988b) andtheintroductiontoBlackburn(1993).
constructivists’ useofCorrespondenceputsonthem.Providingsuchan accountisoneofthemainchallengesthatadvocatesofconstructivism havetomeetinordertodeliveronthepromisethattheirtheoryof practicalreasonsholds.AsIwillargueinChapters4and5,thebestway tomeetthischallengeisforconstructiviststobuildontheiraccountof thenatureofreasonjudgementsasanattitudethatguidespractical reasoningandproposethatthetruthofthosejudgementsistobe understoodintermsofthesoundnessoftheepisodesofreasoningthat theyareapttoguide.¹²
Offeringaconstructivistaccountofsomedomainthushasconsequencesforunderstandingwhatthetruthofjudgementsinthatdomain amountsto.Here,too,IaminagreementwithhowbothRawlsand Korsgaardcharacterizeconstructivism.Rawlsacknowledgesthatbecause theyrejecttherationalintuitionists’ accountofmoraljudgements,constructivistscannot ‘conceiv[e]oftruthinatraditionalwaybyviewing moraljudgementsastruewhentheyarebothaboutandaccuratetothe independentorderofmoralvalues’ (1993,92).Inotherwords,constructivistscannotunderstandtruthintermsofaccuraterepresentation.Rawls seemsunsurewhattomakeofthisconsequence.Sometimesherecommendsthatconstructivistsshouldnotthinkofmoraljudgementsor principlesasbeingtruebutratherasbeingthemostreasonableones, highlightingthatthisallowsustoavoidanycommitmentstoatheoryof truththatdeviatesfromthestandardrepresentationalistaccountthatis mostplausibleinthecaseofordinaryempiricaljudgements(1980,554). Atothertimes,however,Rawlsseemstosuggestthatconstructivismputs forwardanalternativeconceptionofthetruthofmoralprinciplesand judgements,onethatunderstandstheirtruthintermsofthenotionof reasonablenessandwhichallowsconstructiviststoaffirmthegenuine truth-aptnessofthosejudgements,despitetheirrejectionoftheintuitionist’srepresentationalistaccountofthem(1980,569;1993,126n.34).
¹²Showinghowaconstructivistaccountofpracticalreasonscanavoidcircularityisalsoa majorconcerninJames(2007).Jamesarguesthatthiscommitsconstructiviststoshowingthat thenormsofpracticalreasoningdonotderivetheirauthorityfromthefactthattheyleadto reasonjudgementsthataretrueonindependentgrounds(2007,308).Theaccountofcorrect practicalreasoningpresentedinChapter5allowsconstructiviststodischargethatcommitment.
Korsgaardislessambivalentwhenitcomestotheconstructivist’ s commitmentsinthetheoryoftruth.Herconstructivismandwhatshe callssubstantiverealismdonotdisagreeonwhetherornotmoral judgementscanbetrue,butonwhatmakesthemtrue,iftheyare.¹³ Thesubstantiverealistclaimsthattherearecorrectanswerstomoral questions ‘because therearemoralfacts ...thatthosequestionsask about’ (1996,35;heritalics).¹⁴ Inotherwords,ananswertoamoral questioniscorrectbecauseitgetsthemoralfactsright.Korsgaard’ s constructivism,bycontrast,ismeanttogiveanaccountofwhatmakes answerstomoralquestionscorrectthatdoesnotrefertoanyantecedent moralfacts.Sherejectstheideathattheprocedureswhichleadusto correctanswerstomoralquestionsareproceduresthat ‘track’ certain ‘factswhichexistindependentlyofthoseprocedures’ andtherebyprovide ‘waysof findingoutaboutacertainpartoftheworld,thenormative part’ (1996,36–37).¹⁵ Thisdoesnotprompthertobescepticalaboutthe truth-aptnessofmoraljudgements,though.Allthatisrequiredfortruth, Korsgaardclaims,isthatthereisastandardforthecorrectuseofthe relevantconcepts: ‘whenaconceptisappliedcorrectly,whatwegetis truth’ (2003,117).Suchastandardisprovidedbywhatshecalls ‘the correctconception’ oftheconceptinquestion.Butwhatmakesaconceptioncorrect,inthecaseofmoralandothernormativeconcepts,is ‘thatitsolvestheproblem,¹⁶ notthatitdescribessomepieceofexternal reality’ (2003,117).
¹³Hence,theyarebothversionsofrealisminabroadersense,whichincludesanyonewho believesthattherearecorrectanswerstomoral(andothernormative)questionsandthusthinks that ‘ethicsisn’thopeless’ (1996,34–5).Korsgaardcallsthisbroadversion ‘proceduralrealism’ , presumablybecauseshethinksthattheavailabilityofproceduresthatwillleadustocorrect answerstomoralquestionsisanecessarycondition,ifnotfortherebeingsuchanswers,thenat leastforthepursuitofthoseanswerstonotbeahopelessendeavour.
¹⁴ KorsgaardcountsSamuelClarke,RichardPrice,G.E.Moore,DavidRoss,H.A.Pritchard, PeterRailton,DavidBrink,ThomasNagel,andDerekParfit,amongothers,asproponentsof substantiverealism(1996,19;2003,101).Notethesimilaritybetweenthislistandthelistof philosophersRawlsidentifiesas ‘rationalintuitionists’ (seen.7above).
¹⁵ SeealsoKorsgaard(2003;2008,22–3).
¹⁶ Thatis,thepracticalproblemschematicallydenotedbytheconceptinquestion;seealso n.8above.
