Vous êtes sur la page 1sur 16

Imprint

Philosophers

volume8,no.6 july2008

1. Introduction Since the publication of G. E. Moores Principia Ethica it has become commonplace for philosophers to distinguish between questions in metaethics and those in normative ethics.1 A sympathetic characterizationofthecenturyofself-consciouslymetaethicalresearchthatfollowedwouldemphasizetheextraordinarydevelopmentbothinour understandingofthecentralmetaethicalproblemsandinthesophisticationofthetheorieselaboratedtomeetthem.However,someare notsosympathetic.Inthispaper,Iexamineonesourceofdistrustin metaethicalresearch:itsapparenttensionwiththenotionthatmoralityisautonomous. Tobegin,IbrieflysketchhowIamthinkingofmetaethics,ofthe autonomyofmorality,andofthetensionthatcanappeartoexistbetweenthem.Onetraditionalconceptionofmetaethicstakesittoconcern only the analysis of moral language.2 However, contemporary philosophers typically use the term more expansively.3 Here, I use the term to pick out elements common to these contemporary discussions.Thiscommoncoreencompassesmoralontologyandmoral psychologyaswellasmoralsemantics.Bycontrast,normativeethics (sometimesalsocalledsubstantiveethics)concernsthestructureand contentofthecorrectmoralevaluationofagents,statesofaffairs,and actions. Normative ethical theories typically offer accounts of moral valueandmoralreasons,ofvirtuouscharactertraits,ofrightness,and oftherelationshipsbetweenthese.
1. The word metaethics came into regular philosophical usage much later. Moores discussion emphasized the distinction between two kinds of questionsaboutmorality,whichinturnencouragedthedevelopmentofindependentresearchprogramsinwhatwouldlaterbecalledmetaethicsandnormativeethics. 2. ObviousworriesaboutthisconceptionarisefromQuineanconcernsabout analyticity.SuchworriesledZimmermantoargueforanontologicalturnin metaethics(1980,653). 3. VariousbroaderconceptionsofmetaethicsaresuggestedinDarwall,Gibbard, andRailton(1997,7),Miller(2003,2),andSinnott-Armstrong(2006,6).

Metaethics & the Autonomy of Morality


Tristram McPherson
University of Minnesota Duluth

 2008 TristramMcPherson <www.philosophersimprint.org/008006/>

tristrammcpherson

Metaethics and the Autonomy of Morality


The apparent tension between metaethics and the autonomy of morality derives from the plausibility of taking metaethical theorizingtohavetwofeatures.First,metaethicscanseeminglybearonthe justificationofnormativeethicaltheories.ShellyKaganmakesanespeciallystrongversionofthisclaimthatmetaethicsisrelevanttothe justificationofnormativetheory:[I]nthecourseofdefendingagiven theoryaboutthefoundationsofnormativeethics,whenwetrytoexplainwhyitisthatthevariousfeaturesofthattheoryshouldseemattractiveandplausible,inevitablytheclaimswemakewillthemselves simplybemetaethicalclaims(1998,6).Kagansclaimiscontroversial. However,evenifitdoesnotseeminevitablethatmetaethicsplaysthis sortoffoundationalroleinnormativetheorizing,theideathatourbest accountsofmoralpropertiesormoralconceptscouldntcountforor againstthejustifiabilityofanormativetheorymightseempuzzling. Becausemetaethicalclaimsarethemselvesclaimsaboutmorality, this point does not by itself conflict with the autonomy thesis. The apparentproblemarisesfromconjoiningthispointwithreflectionon thetaskofmetaethicaltheory.Onaplausiblegloss,thecentraltaskof metaethicsistoexplainhowmoralityfitswiththeotherelementsof ourbroaderconceptionoftheworld.(Notethatthisdescriptionofthe task of metaethics is consistent both with conceiving of these other elementsinexclusivelynaturalisticterms,andwiththinkingthatour bestoverallconceptionoftheworldmustaugmentortranscendthe scientificworldview.)Thissuggeststhatcoherencewithwell-justified generalaccountsofontologyandsemanticscountsinfavorofametaethicaltheory.Hence,itseemsnaturalformetaethicaltheorizingto involveappealtonon-moralpremises. Together, these two apparent features of metaethical theorizing appeartoconflictwiththeautonomyofmorality.Thus,ifnon-moral premisesarerelevanttothejustificationofaccountsofmoralsemantics or ontology, and these accounts in turn play a role in justifying foundational normative theories, then non-moral premises can play an indirect role in justifying normative ethical theories. This conclusionappearstobeinconsistentwiththeautonomythesis. Ifoneissympathetictotheautonomythesis,thisapparenttension

AdomainofinquiryisautonomousinthesenseIintendifresultsin otherdomainsdonotcontributetothejustificationoftheoriesinthat domain.4Theautonomy of moralityisthustheclaimthat: autonomy: Non-moral theses are irrelevant to the justificationofnormativeethicaltheories. Iwillsometimesrefertothisastheautonomy thesis.5Theideathatnormativeethicaltheorizingisautonomousmayappearintuitivelyattractive.Thus,itmayseemonreflectionthattheonlylegitimategrounds for abandoning a basic moral principle will involve tension with othermoralcommitments.6Further,theautonomythesispromisesto explain the plausibility of Humes dictum that one cannot derive an oughtfromanis:ifmoralityisautonomous,itwillalwaysbeanerror toderiveamoralconclusionfromstrictlynon-moralpremises.
4. My use of the term autonomous here is inspired by Harmans distinction between naturalism and autonomous ethics in moral theory (2000, 79). However, I use it to make a slightly different distinction. Harman suggests thatanyapproachtoethicsnotdominatedbyaconcerntoplacevaluesin the natural world discovered by science is autonomous (2000, 79). On my gloss, science is only one domain with respect to which morality might be autonomous.Forexample,Ialsotakedivinecommandtheorytobeincompatiblewiththeautonomyofmorality,asitwouldrenderevidenceofGods willprobativefornormativeethicaltheory. 5. Therestrictionofthescopeoftheautonomythesistotheories,understood as consisting of fundamental principles, is crucial to the plausibility of the autonomythesis,becauseitisimplausiblethatevidencefromotherdomains isnotrelevanttothejustificationofderivativeprinciplesorofparticularmoraljudgments.Forexample,considerthemoralprincipleItiswrongtofire loadedgunsatinnocentpeopleandtheparticularmoralclaimLeesfiring hisgunatJohniswrong.Anaturalwaytojustifysuchclaimswouldappeal inparttomorebasicnormativeprinciples,andinparttothetypicalorpredictablecausalconsequencesoffiringloadedgunsatpeople.Theautonomy thesis should thus be understood as compatible with a role for non-moral thesesinexplainingwhyanormativemoralprincipleappliestoaparticular case,orsupportsaderivativeprinciple. 6. CompareHarmansdescriptionoftheautonomousethicistsreasoningabout acase:[W]eshouldcontinuetobelieve[thatacriminalhassufficientreason not to harm his or her victims] unless such continued belief conflicts with generalizationsorothertheoreticalprinciplesinternaltoethicsthatwefind attractivebecausetheydoabetterjobatmakingsenseofmostofthethings weoriginallybelieve(2000,87).

philosophersimprint

2

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


implications. Thus, if metaethics was necessarily neutral between normativetheories,thesensitivityofmetaethicaltheorytonon-moral considerations(thesecondapparentfeatureofmetaethicaltheories) couldnotimpugntheautonomyofnormativeethicaltheorizing. Onecouldthinkoftheneutralitythesisineitheroftwoways.First, one could think of it as a hypothesis about metaethical theories, to beevaluatedbyexaminingsuchtheories.However,itiseasytofind counterexamplestothisneutralityhypothesis,intheformofmetaethicaltheorieswithdirectnormativeimplications.Considerasimpleexample:theanalyticutilitarianismthatMoore(1903,1718)attributes toJeremyBentham.Thissemantictheorystatesthatthewordright meansconducivetogeneralhappiness.Analyticalutilitarianismisan accountofthesemanticsofacentralpieceofmoralvocabulary,and hence a paradigmatic metaethical theory. However, it also transparentlyhasimplicationsforthecontentofthecorrectnormativetheory. The neutrality thesis thus seems implausible if read as a hypothesis concerningthepossiblecontentoftheoriesthatweintuitivelycount asmetaethical. Onecouldinsteadthinkoftheneutralitythesisasaconstrainton a theorys counting as metaethical. Thinking of the neutrality thesis asahypothesisrunsintotroublewhereseeminglyparadigmaticmetaethicaltheorieshavenormativeimplications.Bycontrast,aneutrality constraint would entail that the same data showed such theories notreallytobemetaethical.Thus,acceptingsuchaconstraintmight seemtopromisetoresolvethetensionbetweenmetaethicsandthe autonomyofmorality.Intheremainderofthissection,Iexaminethe implicationsofacceptingthisneutralityconstraint,focusingonissues highlighted byDworkinspaperObjectivityandTruth:YoudBetter BelieveIt.Thisdiscussionwillshowthatacceptingtheneutralityconstraint is objectionable in virtue of obscuring philosophically importantdistinctionsbetweenmetaethicaltheories. Dworkins main target is what he calls the neutral Archimedean skeptic: a philosopher who combines first-order normative ethical commitment with a deflationary metaethics such as noncognitivism

mightsuggestthatthereissomeerrorimplicitinmetaethicaltheorizing.Somephilosophershaverecentlyoffereddiagnosesofthisalleged error. Thus Christine Korsgaard suggests that the dialectic between metaethical realists and antirealists rests on a mistaken conception ofthefunctionofmoralconcepts.Sheclaimsthatwhenourthinking aboutthesubjectiscorrected,wewillnotbeinclinedtothinkthat there is a difference between doing meta-ethics and doing normativeorpracticalethics(2003,121n.44).RonaldDworkinalsofinds errorinmuchapparentlymetaethicaltheorizing,arguingthatitrests onmistakingsubstantiveethicalclaimstobemetaphysicalorphilosophical(1996,100). This paper examines three ways to address the apparent tension betweenmetaethicsandtheautonomyofmorality.Insection2,Iexaminetheattempttoaddressthistensionbyrejectingtheclaimthat metaethics can be relevant to the justification of normative ethical theories. I evaluate this strategy in part by examining the reasoning thatleadsDworkintothediagnosismentionedabove.Insection3,I examinevariantsontheproposaltovindicatetheautonomyofmoralitybyappealingtoadeepcontrastbetweenpracticalandtheoretical reason. Korsgaards diagnosis is best understood as following from such a commitment. Finally, in section 4 I suggest what I take to be themostpromisingwaytoresolvetheapparenttensionbetweenmetaethicsandtheautonomyofmorality. 2. NeutralityandQuietism Onewaytoresolvetheapparenttensionbetweenmetaethicsandthe autonomy of morality would be to endorse what I call the neutrality thesis: neutrality: Nometaethicaltheoryisrelevanttothejustificationofnormativeethicaltheories. Theneutralitythesisresolvesthetensionbetweenmetaethicsandthe autonomy of morality by denying the first of the apparent features of metaethical theses mentioned above: their possible normative

philosophersimprint

3

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


wordsubjectivelymightbeusedtoclaimthatthevalueofsocceris onlyagent-relative,therelevantagentsbeingthosewhoshareataste forthegame.Onlysuchsoccer-enjoyershavereasonstobringitabout thattheywatchsoccer.Ontheotherhand,ifeveryonehadreasonto bring it about that people watch soccer, then the value of watching soccer would be agent-neutral. Agent-relativity is a normative-level featurebecauseitdescribesadistinctivestructureintherelationbetweenvalueandreasons,twonormativenotions.Dworkinsexemplaryuseofsubjectiveisthusbestunderstoodasmakinganormative claim,ashesuggests. Nonetheless,thereappearstobeaproblemwithDworkinsuseof thistest:someusesofthewordsubjectiveindicatetheevaluator-relativityofamoralterm.Evaluator-relativityisanapparentlymetaethical notion.Thusthesemanticsofrightwouldbeevaluator-relativeifthe truth of statements of the form -ng is right depended upon who wasmakingorevaluatingthestatement.Consideranexample:people sometimessayvaluesaresubjectiveasawayofendorsingasimple subjectivistmetaethics.Accordingtosimplesubjectivism,utterances oftheform-ingisrightaretrueinapersonsmouthjustincaseand becausesheapprovesof-ing.Suchusesofsubjectivethusappear tofailDworkinsfirsttest. Other vocabulary also exhibits the possibility of both first-order and meta-level uses. Consider another of Dworkins examples, the wordreally.IfIsayThatbookisntreallyred,Imightbeattempting toconveyoneof(atleast)twothings.First,Imightbeattemptingto getyoutoseethatyouvegottenthecolorofthebookwrong:itsmore ofanorangey-brown.Second,Imightbeattemptingtoconvinceyou onmetaphysicalgroundsthatcolorsarenotrealpropertiesofphysical objectslikebooks.Thepossibilityofsuchvocabularysbeingusedto expressbothmetaethicalandnormativejudgmentsthusappearsunsurprising,beinganinstanceofageneralphenomenon. Dworkinsfirsttestthusappearslessthanfullysuccessful.Hehas
objective reasons (1970, Ch. X). Here I abstract away from the significant complexitiesthatafullyadequateaccountofagent-relativitywouldface.

orerrortheory(1996,92ff.).However,partofDworkinscentralargumentative strategy seemingly involves deploying the neutrality constraint.HereIfocusonthisuseoftheconstraintanditsimplications formetaethics.7Dworkinappliesatwo-stagetesttoanarrayofcentral metaethical proposals in his paper. First, test whether the seemingly metaethical claim can be interpreted as having normative import. Second,testwhetheritcanbeinterpretedashavingdistinctivelymetaethical import (Dworkin 1996, 97).8 If the result of the first test is positiveandtheresultofthesecondisnegative,thentheseemingly metaethicalclaimwillbeshowntobeadisguisednormativeclaim.I considerthesetestsinturn,andexplainhowtheneutralityconstraint isindispensabletoDworkinsapplicationofthesecondtest.  Dworkins first test examines whether we can interpret pieces of seemingly metaethical discourse as expressing normative ethical claims. He suggests that people typically use seemingly metaethical adverbs, such as subjectively and really, to clarify their normative ethicalopinions.Hethusnotesthatsomeonemightusethewordsubjectivetosignalthathisclaimthatsoccerwasbadmeantthathehad areasontoavoidsoccerthatisnotsharedbythosewhohaveataste forthegame(1996,98). Inordertoproperlyperformthefirsttest,weneedtodistinguish agent-relativityfromevaluator-relativity.Agent-relativityisafeatureof valueorreasonsinanormativetheory.Veryroughly,thinkofastate ofaffairsashavingagent-relativevalueifnotallagentshaveareason tobringitabout.9Dworkinsglossonhisexamplepointsoutthatthe
7. Myconcernhereistoteaseouttheimplicationsofacceptingtheneutrality constraint.IthusfocusonelementsofDworkinsdiscussionthatappearto pursue this strategy. Dworkins argument in (1996) is complex and not alwaysclear,andIdonotclaimtoofferafullyadequatereadingofitinthis b  riefdiscussion. 8. Hereandbelow,thisglossmustbereadextensionally.WhileDworkinsleadingtargetsinthispaperareclearlymetaethicaltheories,hegenerallyprefers to talk of collapsing the distinction between first-order evaluative claims andphilosophicalclaimsaboutmorality(e. g.1996,100). 9. ThelocusclassicusforthisdistinctionisNagelsdiscussionofsubjectiveand

philosophersimprint

4

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


itisexhausted,inthesameway,bythatsubstantivethesis.Thereisnodifferenceinwhattwopeoplethinkifone thinksthattheonlythingthatcanmakeanactrightisits maximizing power, so that it makes no sense to evaluaterightnessinanyotherway,andtheotherthinksthat thepropertyofrightnessandthepropertyofmaximizing power are the very same property. The second opinion usesthejargonofmetaphysics,butitcannotaddanygenuineideatothefirst,orsubtractanyfromit.Itsoundsmore philosophicalbutitisnolessevaluative.[1996,1001] Ireadthisseriesofclaimsthattheontologicalclaimisexhausted bythenormativethesis,thatthereisnodifferenceinwhattwopeople thinkifonethinksthenormativethesisistrueandtheotherthinks the ontological thesis is true, and that the ontological thesis merely uses the jargon of metaphysics to sound more philosophicalto beforcefulwaysofclaimingthat,whilethetwoclaimsmightappear distinct,theyareinfactidenticalincontent. Dworkinarguesthatanextremelybroadrangeofapparentlymetaethical theories have significant normative consequences (1996, 98ff.).10 This in turn has striking implications for our understanding of metaethics. Discussing a series of seemingly metaethical theses thathecallsfurtherclaims,hesays,[W]earetryingtodecide,not whetherthefurtherclaimscanbetranslatedtomakethemseemmore philosophicalormetaphysicalbutwhetherwecanunderstandthose philosophicaltranslationsasthemselvesanythingbutfirst-orderevaluativeclaims(1996,100). Dworkin intends his second test to show that the answer to this question is no. This suggests that, just as with metaphysical utilitarianism, a surprising range of seemingly metaethical claims can
10. ItisnotentirelyclearwhetherDworkinintendshisconclusiontobeperfectly general,orwhetherhewouldbepreparedtocountenancetheexistenceof some genuinely metaethical claims. Dworkins discussion prompted Dreier (2002)toofferacarefularticulationofametaethicaltheorythatisclaimedto havenonormativeimplications.

shown that a variety of vocabulary seemingly used to express metaethicalclaimscanbeusedtoexpressnormativeclaims.However,he hasntshownthistobetrueforsuchvocabularyasitisusedincontexts where we would intuitively take it to have metaethical import. ThissuggeststhatDworkinssecondtestisthecrucialone.Thegoal ofDworkinssecondtestistoshowthat,forarangeofputativelymetaethical vocabulary, no distinctively metaethical interpretation is in factavailable.Werethistestsuccessful,itmightbepossibletoaccommodatetheapparentcounterexamplestothefirsttestjustdiscussed asinstancesofdiversenormative-levelusesofsubjectivelyandreally. Dworkinsleadingstrategyinapplyinghissecondtesttoapparently metaethicalclaimsistoarguethatthoseclaimsentailnormativecommitments (1996, 99112, cf. especially 100). The example of analytic utilitarianism, mentioned above, suggests that at least some apparentlymetaethicaltheorieshavesuchimplications.However,Dworkin repeatedly moves from the premise that an apparently metaethical theoryhasnormativeimplicationstotheconclusionthatitistherefore actuallyanormativetheory.Thisinferenceseemstopresupposethat nogenuinelymetaethicaltheorymayhavenormativeimplications.In otherwords,itpresupposestheneutralityconstraintoncountingasa metaethicaltheory. ConsiderDworkinstreatmentofanexampleofanapparentlymetaethical thesis with obvious normative implications: the thesis that the property of rightness is identical to the property of maximizing happiness.Callthisthesismetaphysical utilitarianism.Thisappearsto beametaethicalthesisaboutmoralontologythatentailsanormative theory. This normative theory, simple act utilitarianism, states that an actisrightjustincaseandbecause,amongtheoptionsavailableto theagentoftheaction,itwouldbringaboutthemosthappinessinits consequences.However,Dworkinrejectstheclaimthatmetaphysical utilitarianismandsimpleactutilitarianismaredistincttheories: Theidentityclaimaboutrightnessistheupshot,inaparallelway,ofasubstantivemoralthesisutilitarianismand

philosophersimprint

5

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


theclaimthattheabovebiconditionalistruebecauseGodsapproval makesactionsright.Thenaturallessontodrawfromthisexampleis thatdistinctmetaethicaltheoriescanentailthesamenormativeethic, becausecoextensionevennecessarycoextensionisnotsufficient tosettlemetaethicaltheoryidentity.Semantic,determination,andexplanatoryrelationsarealsoofparamountimportance. The defender of the neutrality constraint is thus faced with a dilemma.Ontheonehand,Dworkinsstrategyofidentifyingapparently metaethical theories like metaphysical utilitarianism with intuitively normativetheorieslikesimpleact-utilitarianismfails.Thisisbecause, aswehavejustseen,simpleactutilitarianism(Dworkinsexemplary normative theory) is compatible with metaethics that are incompatiblewithmetaphysicalutilitarianism.Indeed,itisevenentailedbymetaethicaltheoriesthatdenytheidentityofthepropertiesofrightness andhappinessmaximization.Forexample,itisentailedbysomeversions ofthe theorythatidentifiesthepropertyofrightnesswith the propertyofbeingendorsablebyanideallyinformedimpartiallysympatheticobserver. Ontheotherhand,itwillnothelptoinsistthat,invirtueoftheir implications, these apparently metaethical theories are nonetheless normativetheories.Consideragaintheimaginedtheoristscanvassed thus far: the analytic utilitarian, the metaphysical utilitarian, the divine-command utilitarian, and the ideal-observer utilitarian. These theoristsdisagreeaboutmorality.Thefirsttheoristdisagreeswiththe restaboutasemanticthesis:whethermorallyrightactandactthat maximizes happiness are synonymous. The rest disagree amongst themselves concerning a metaphysical issue: namely, whether the propertyofbeingrightisidenticaltothepropertyofmaximizinghappiness,tothepropertyofbeingapprovedofbyGod,ortotheproperty of being approved of by an ideal observer. Because these disagreements concern moral semantics and metaphysics, they are naturally understoodasmetaethicaldisagreements. Simplyinsistingthatthesedisagreementsshouldbedescribedas normativeethicaldisagreementswouldnotavoidtheapparenttension

be revealed to be normative ethical claims because they fail his second test. Dworkin thus appears to defend a kind of quietism about metaethics.11 Isuggestedatthebeginningofthissectionthatacceptingtheneutralityconstraintseeminglypromisestoresolvetheapparenttension betweenmetaethicsandtheautonomyofmorality.Thediscussionof Dworkin shows that accepting the constraint would also entail that manyphilosopherswhotakethemselvestobeaddressingdistinctively metaethicalquestionsaredeeplyconfused.Thisisbecause,according totheviewsketched,manyofthesephilosophersareactuallyaddressingfirst-ordernormativequestionsratherthanmetaethicalquestions. WhileDworkinsargumentclarifiestheconsequencesofapplying theneutralityconstraint,hisownexampleshelptoillustratewhythere isgoodreasontorejectit.ConsideragainDworkinsidentificationof metaphysicalandnormativeutilitarianism,anditsimplicationthatthe apparentlymetaphysicalcharacteroftheformertheoryisillusory.The problemwiththisidentificationcanbetracedtooneofthecentralmotivationsfordistinguishingmetaethicalfromnormativetheories.This isthatitispossibletoagreeaboutthecorrectnormativetheorywhile disagreeingaboutthecorrectmetaethicaltheory.12 Thedeepmotivationforthisclaimcanbebroughtoutespecially clearlybyconsideringthedisagreementbetweentwotheistswhoagree about which acts are right. Suppose, for example, that they both acceptsimpleactutilitarianism.Supposefurtherthattheyalsoagreethat, necessarily,anactisrightjustincaseGodapprovesofit.Thesetheists canstilldisagreeaboutthetruthofthedivinecommandmetaethics:
11. Inareplytocritics,Dworkinrejectsthislabel,saying:Idontliketheterm quietist for the reason I mentioned: it suggests that some more robust sense of objectivity makes sense but is wrong (1997). I intend my use of quietism topick outexactly thisdenialoftheintelligibilityofdistinctively m  etaethicalcontent. 12. This point has been suggested across the history of discussion of metaethicsandnormativeethics.SeeforexampleFrankena(1951,45),Zimmerman (1980, 659), and Kagan (1998, 5), who mentions but does not explicitly endorsethispossibility.

philosophersimprint

6

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


fact. In this section, I consider variants on an important strategy for explainingwhymoralityisautonomous. Thisstrategybeginsbynotingthatnormativeethicalprinciplesare paradigmaticallypractical.Thus,determiningthatanactionwouldbe wrongplaysadistinctiveandatleastoftendecisiveroleindeliberation regarding whether to perform it. This point is common ground: the distinctivemotivationalandnormativeaspectsofmoralityarecentral explanandaincurrentmetaethics.However,onedistinctiveandinfluentiallineofthoughtincontemporaryethicsinsiststhatunliketheoreticalreasoning,moraltheorizingmustproceedfromtheperspective offirst-personpracticaldeliberation.Thisdeliberative requirementmight seemtomakethedividebetweentheoreticalandpracticalreasoning deepenoughtoprecludetheoreticalreasoningfrombeingrelevantto thejustificationofmoraltheories.Accountsthatdevelopthiscontrast thuspromisetovindicatetheautonomyofmorality.13 I begin this section by sketching an apparent counterexample to the autonomy of morality, and the basic form of a response to this counterexample that appeals to the deliberative requirement. I then considerthreewaysthatthisresponsecouldbedeveloped.Thefirst twoalternativesarevariantsoftheideathatpracticalandtheoretical reasondeliverdistinctiveevaluationsofagivenproposition.Thefinal alternative satisfies the deliberative requirement by characterizing a distinctivenotionofpracticalcontent.Finally,Isummarizetheconsequencesformetaethicsoftheseapproaches.Astheissuessurrounding theseproposalsarebothdeepandcomplex,Idonotevaluateoreven fullydevelopthem.Instead,Isketchjustenoughstructuretoascertain therelevantconsequencesofsuchviews. DerekParfitscentralargumentinPartThreeofReasons and Persons suggests an apparent counterexample to the autonomy of morality. Parfit argues that the metaphysical conception of personal identity
13. Darwall,Gibbard,andRailtonnotesomethinglikethislinkintheircharacterizationofwhattheycallpracticalreasoningtheories:Itisethicsintrinsicallypracticalcharacter,itsholdonusas agentsthatexplainstheopenquestionand,theysay,marksethicsofffromscience(1997,9).

betweenmetaethicsandtheautonomyofmorality.Recallthatthisapparent tension is generated in part by the pressure to reconcile our accountsofmoralsemanticsandontologywithourgeneralsemantic andontologicaltheories.Nomatterwhatwecallthem,thesemantic andmetaphysicalviewsidentifiedherefacesuchpressure. Dworkinsuseoftheneutralityconstraintonmetaethicaltheories seemingly promised to undercut the apparent tension between metaethicsandtheautonomyofmorality,byshowingthattheapparent metaphysicalcharacterofmanymetaethicaltheoriesisanillusion.If sound,thisuseoftheneutralityconstraintwouldvindicatethecurrent ofdistrustinmetaethicalresearchmentionedatthebeginningofthis paper.Thisisbecausethisconclusionsuggeststhatmostmetaethicists arebadlyconfused,eitheraddressingpseudo-questionsordoingnormativeethicswithoutrealizingit. Inthissection,Ihavearguedagainsttheattempttousetheneutralityconstrainttoresolvetheapparenttensionbetweenmetaethicsand theautonomyofmorality.Ifirstarguedthatametaethicaltheorythat entailssomenormativethesisshouldnotbeconflatedwiththatnormativethesis(oranyother).Thisisbecausemetaethicaltheorieswith normative implications also have further distinctive content. I then arguedthattheappealtotheneutralityconstrainteitherisobjectionableinvirtueofobscuringthiscontent,orfailstoresolvetheapparent tensionbetweenmetaethicsandtheautonomyofmorality.Thissuggeststhatweshouldlookelsewheretoresolvetheapparenttension betweenmetaethicsandtheautonomyofmorality.Inthenextsection, Iexamineaclusterofstrategiesthatseektovindicatetheautonomyof moralitybyappealingtoacontrastbetweenpracticalandtheoretical reason. 3. ThePracticalityofMorality IsuggestedintheIntroductionthattheautonomyofmoralitycanappearintuitivelyattractive.However,evenifoneissympathetictothe autonomythesis,itwouldbemoresatisfyingtohaveaplausibleexplanationofwhymoralitymustbeautonomousthantotakeitasabrute

philosophersimprint

7

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


nondoxasticcommitmenttowardsit.Iwillcallthesethetwonorms andtwoattitudesproposals,respectively. Theminimalformofthetwonormsproposalsuggeststhatthepracticalandtheoreticalperspectivesdeliverdistinctivenormsforbeliefin agivenproposition.Thisclaimmustbeaugmentedintwowaysifitis tovindicatetheautonomyofmorality.First,theautonomyofmorality issupposedtomarkacontrastbetweenthejustificationofnormative ethical principles and the justification of ordinary beliefs. However, circumstances can probably make it useful (for example) to have arbitrarybeliefsonanytopic.14Thespecialroleforpracticalnormsthus needstobeconstrainedinitsscope.Isetthisproblemasidehere. Second,theautonomyofmoralityrequiresthatrelevantpractical normsforbeliefmustalwaystrumptheoreticalnormsinRawlsian argot,practicalnormsmusthavelexicalpriority.Brieflyconsidertwo alternatives.Ontheonehand,epistemicandpracticalnormscannot be(evenroughly)commensurable.Thisisbecausesuchanalternative wouldallowthepossibilitythatpracticalreasonsinfavorofbelieving a normative principle could be outweighed by countervailing theoreticalconsiderations,therebyviolatingtheautonomythesis.Onthe other,thepossibilitythatitcouldbefullyrationaltobelieveapropositioninvirtueofpracticalconsiderations,whilealsofullyrationalto disbelieveitforepistemicreasons,isalsounappealing.Besidesitsintrinsicpeculiarity,suchaviewwouldfailtovindicatetheautonomy ofmoralitybecauseitwouldsuggestthatitcouldbefullyrationalto believeamoralpropositionontheoreticalgrounds.15 Suppose then that there are distinctive practical and theoreticalnormsthatapplytotheevaluationofbeliefswithmoralcontent, andthatthepracticalnormshavelexicalpriorityoverthetheoretical
14. Thinkhereofiterationsofasciencefictioncaseinwhichabillionairewitha brainscannerwillrewardyouifyoubelievesuch-and-such. 15. Nelkin(2000)considersavariantofthisapproachthatsuggeststhatwecan rationallybelieveobviouslycontradictorypropositions,providedthatwedo sofromdifferentstandpoints.Nelkinpersuasivelyemphasizesthedifficultiesthatfollowfromsuchaproposal.Isetthisviewaside,notingonlythatthe upshotofsuchaviewformetaethicswouldbebroadlysimilartothatofthe twonormsviewdiscussedinthetext.

thatmattersformoralityisverythin.Hethenusesthismetaphysical premisetoarguefortheparityofmoralsignificancebetweenintrapersonalintertemporalconsiderations,andinterpersonalconsiderations (1984,199350). Korsgaard claims that Parfits argument misses its mark, because the conception of personal identity relevant to deliberation is one required by practical reason, not one identifiable by metaphysical inquiry (1989b, 112). She argues further that we all have sufficient practicalreasonstoviewourselvesasunifiedpersons.Thesereasons are twofold. First, there is the raw necessity of eliminating conflict. Korsgaardclaimsthatunlessonespartsbothonessubpersonalprocessesandonestemporalslicescoordinateinsuchawayastoallow one to act as a unified agent, anything resembling action would becomeimpossible,andpracticalparalysiswouldresult.Second,thedeliberativestandpointrequiresthatadeliberatorseehimselfassomethingbeyondamerecollectionofdesiresorothersubpersonalstates. Thus,Korsgaardobservesthatonecannotdeliberateaboutwhattodo merelybywaitingtoseewhichofonesdesireswinsout.Onecanonly choosefromapracticallyunifiedstandpoint(1989b,1101). Suppose that Korsgaard is correct, and that the conception of identity that we have practical reason to accept is distinct from the conceptionthatParfitdefends.Thiswouldsuggestatestcaseforthe threewaysofdevelopingthedeliberativerequirementthatIwillnow sketch:Parfitsmetaphysicalargumentssuggestonethingaboutpersonalidentityandnormativeethics,whilepracticalreasoningpresupposesanother.Iftheaccountstobesketchedaretovindicatetheautonomyofmorality,theymustexplainwhyParfitsargumentsarenot relevanttonormativetheorizing. Consider first the idea that one can have divergent practical and theoreticalreasonsconcerningtheacceptanceofagivenproposition. Thisdistinctionsuggestsapossiblecontrastbetweenthemetaethics forwhichwepossessthebestevidence,andthemetaethicsthatwe havedecisivepracticalreasonstoaccept.Howonedevelopsthisproposaldependsuponwhetheronetakespracticalreasoningtorequire agentstobelieveinametaethics,ormerelytohavesomespeciesof

philosophersimprint

8

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


theoriesdiscoverablebytheoreticalreason,andhenceagaindoesnot vindicate the autonomy thesis. However, it would vindicate the autonomyofourmoralcommitmentsfromtheoreticalreasoning.Thisis becauseitmakestheoreticalreasoningthewrongkindofinquiryto determinethecontentofthemoralitythatonehaspracticalgrounds toaccept.Ifordinarymoraltalkexpressesourpracticalcommitments, thenthisproposalsuggestsaspeciesoffictionalismaboutourmoral discourse,coexistinguneasilywithunexpressedmoralbeliefs.Forexample,itsuggeststhatonemightreasonablybelieveanactiontobe permissible,whilebeingpracticallycommittedtoitsbeingwrong. Despite vindicating a kind of autonomy for moral discourse, this account again does not impugn metaethics. Like the two norms account,itisconsistentwiththeoreticalinquiryintoethics.Further,developing this sort of view would in part involve metaethical inquiry. Considerthreecentralelementsofanadequatedevelopmentofsuch aview.First,suchdevelopmentwouldrequireaproperlydeveloped defenseoftheallegeddeepcontrastbetweenthepracticalandtheoretical perspectives. Second, it would require a plausible account of thepsychologicalstateofpracticalcommitment.Andfinally,itwould require a fictionalist semantics that explained how we express such astate,ratherthanourbeliefsaboutmorality,inourordinarymoral utterances.Thesecondandthirdofthesetasksarerecognizablymetaethicalprojects. Wehaveseenthat,inasmuchasweareseekingadeepvindication of the autonomy of morality, the two norms and two attitudes proposalsarevulnerabletoanimportantobjection.Whileeachofthese accountsmakesspaceforadomainofautonomousmoralreasoning, neither proposal fully vindicates the autonomy of morality. This is becauseeachiscompatiblewiththerebeingatruenormativeethical theorydiscoverablebytheoreticalreasoning.17
17. Theoreticalinquiryinto ethics would,however,lose a standard motivation ontheseaccounts:toguideaction.Thisisbecauseitwouldbesilentconcerningthenormsthatwearepracticallyrequiredtoaccept.Wemightevenhesitatetocallinvestigationshornofthispracticalpointethicalinquiry,despite itsbeingaimedattheidentificationoftruemoralpropositions.

norms. This account can grant that Parfits arguments might give us goodepistemic reasonstobelieveinathinaccountofpersonalidentity andintheallegedconsequentialistnormativeimplicationsofthisaccount.However,thepracticalnecessityofacceptingastrongeraccount ofidentitywouldneverthelessrequireustobelievethatthestronger account is correct. This proposal would not vindicate the autonomy thesis:forallitsays,Parfitsargumentsepistemicallyjustifynormative ethicaltheses.However,itdoesentailsomethingrecognizablysimilar totheautonomythesis:thatitisnevercorrecttoformamoralbeliefon theoreticalgrounds,wherepracticalconsiderationscountagainstit. This proposal has interesting implications for normative and metaethical thinking. On the one hand, it is clearly consistent with the existence of true metaethical and normative theories, each discoverablebytheoreticalreasoning.Ontheotherhand,itraisestheperhaps disturbing possibility that we might be practically required to disbelieve the true moral theory, or to refrain from theoretical inquiry i  ntomorality.16  The two attitudes proposal differs from the two norms account invirtueofthepsychologicalstatethatitclaimstoberationalizedby practicalnorms.Accordingtothisaccount,whileepistemicnormsrationalize belief, practical norms rationalize a distinct state of practicalcommitmenttowardsaproposition.Thisideaagaincriesoutfor moredetailedtreatment.However,thissketchsufficestosuggestthe implicationsrelevanttothispaper.Likethetwonormsaccount,the twoattitudesaccountcangrantthetheoreticalforceofargumentslike Parfits.However,itclaimsthat,invirtueofconflictingwithpractically necessarysuppositions,theseargumentstellusnothingaboutthedistinctivepracticalcommitmentsthatweoughttoaccept. This proposal is thus consistent with the existence of true moral
16. The two norms account will also seem objectionable to some in virtue of presupposingacontroversialviewconcerningtheethicsofbelief:thatthere could be practical requirements not just on the actions that lead to beliefs butonbeliefsthemselves.Thishasseemedwrongtosome,invirtueofthe thoughtthatbelievingisinvoluntary,andthatpracticalrequirementscanapplyonlytovoluntaryaction.

philosophersimprint

9

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


willedtobeauniversallaw)canbeextractedfromtheseconstraints. Thelessonisthat,sincetheseconstraintsaresimplydescriptionsof the details of the relevant practical problem, we are in a position to extractnormativeprinciplesfromthestatementofapracticalproblem alone(2003,1135).Korsgaardconcludes:Thecategoricalimperative is a principle of the logic of practical deliberation, a principle that is constitutiveofdeliberation,notatheoreticalpremisethatisappliedin practicalthought(2003,115). Korsgaard suggests that this sort of practical logic constitutes the correct moral methodology: in moral inquiry we attempt to discover moral principles, exactly by understanding these inescapable practicalproblemsandthestructureoftheirsolutions(2003,1156).20 Sheclaimsthatthisaccountofpracticalconceptsalsoprovidestruthconditions for propositions with practical content. On this account, normative principles, understood as solutions to practical problems, functionasguidestothecorrectapplicationoftheconceptsthatthey name.Korsgaardthussuggeststhatapropositionwithpracticalcontentistrueif(inter alia)itconformstotheapplicationconditionsofthe relevantpracticalconcept(2003,117). Thisaccountofpracticalconceptsfleshesoutthetwocontentsaccount by suggesting what distinctively practical conceptual content couldbe.Thepracticallogicmethodologythatfollowsfromitalsoexplainshowthetwocontentsaccountwouldvindicatetheautonomy of morality. This is because the practical logic conception of moral methodologyimpliesthattheoreticalpremisesaresimplythewrong sortsofthingstoenterintonormativeethicaltheorizing.21Thus,ifthe
20.This conception of practical methodology is, of course, controversial. One centralquestionishowmuchcanbeextractedfromthelogicofourpractical concepts.Gibbard(1999),forexample,ispessimistic. 21. Thisrequiresthatwethinkoftheoutputofmoraltheorizingintermsofgeneralprinciplesormaxims,justasmyglossoftheautonomythesisrequires. However,paceKorsgaardsclaimmentionedabove,itdoesseemthatthese maximswouldconstituteakindofmoralknowledgethatwewouldneedto applyinactioninlightofourbeliefsabouttherelevantfeaturesofouroptions(e. g.thebeliefthat-ingwouldcauseinjury,orthat-ingwouldbethe tellingofalie).

The last way to develop the contrast between practical and theoreticalreasonthatIconsideravoidsthisobjection.Thisaccountclaims that practical and theoretical reasoning involve distinctive types of content.Ithuscallthisthetwocontentsaccount.Suchanaccount suggests that nonderivative theoretical reasoning about morality wouldinvolveacategorymistake.Thisinturnensuresthattherecannotbeatruemoraltheorydiscoverablebypurelytheoreticalinquiry. The two contents account thus promises to more fully vindicate the autonomyofmoralitythantheotherproposalssketched.Becauseof this, I consider the two contents account to be the most interesting developmentofthecontrastbetweenpracticalandtheoreticalreason thatgroundsthedeliberativerequirement.Consequently,Iconsiderit ingreaterdetailthantheothers. Themostpressingquestionfacingsuchanaccountiswhatexactly the notion of distinctively practical content could come to. Here, I pursue a helpful suggestion offered by Korsgaard.18 She claims that practical concepts do not function to describe reality (2003, 105). Rather,theessentialfunctionofapracticalconceptistorefertothe solutiontoa[practical]problem(2003,115).19 Korsgaard offers as an example of this model a reading of Kants derivationoftheUniversalLawformulationofthecategoricalimperative.Thepracticalprobleminthiscaseisthatthewillmustadopta principleforitselfonpainofheteronomy.Further,thatprinciplemust bealaw,because(accordingtoKant)forthewilltobeacause,itmust conform to a law. Korsgaard claims that the Universal Law formulationofthecategoricalimperative(Actonlyonamaximthatcanbe
18. It is not entirely clear whether Korsgaard consistently endorses what I call thetwocontentsaccount.Forexample,NelkinsuggestsreadingKorsgaard asadoptingthetwostandpointsproposalmentionedinthepreviousnote (2000,5678).However,NelkinexpressesuncertaintyregardingKorsgaards ultimate view. I intend here to apply Korsgaards account of practical content, while making no claim to have identified her considered opinion on t  hesetopics. 19. Compare also her claim that morality requires a different conceptual organization of the world from that required by the theoretical perspective (  1989a,37).

philosophersimprint

10

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


toconstructaresponse-dependentmetaethicsbyappealingtoakind ofsemanticdeference.Justasonecandefersemanticallytoaperson oragroupofexperts,Iproposethatatheoreticalreasonerwouldbe abletodefertothepracticalperspective.This,Isuggest,wouldallow such a reasoner to construct a response-dependent metaethical accountofmoralproperties.Thus,takethetermsagentandagencyto beuseddeferentially,topickoutthemeaningsofthesetermsgivenby thelogicinternaltothepracticalperspective.Thesetermscanbeused from within the theoretical perspective to construct an instance of theresponse-dependenceschemaforpermissibility,practical responsedependence:  prd Optionsarepermissiblejustincaseandbecause agents would judge them to conform to norms that they accept, when they are informed about, and responsive to, the conditions necessary to fullymaintaintheiragency.

twocontentsaccountwerecorrect,thedefenderofthepracticalityof morality would be in a position to say something much stronger in responsetoParfitthanwaspermittedbytheotheraccounts.Onthis accountshecouldconvictParfitofmakingafundamentalmistake,becausemoralconceptsandconceptsofpersonalidentityarepractical, andhencenotappropriateobjectsforthesortoftheoreticalreasoning thatheattemptstoapplytounderstandthem. Therelationbetweenthetwocontentsaccountandmetaethicsis lessobviousthanthatofthepreviousaccounts.Onemightthinkthat the proposed distinctively practical character of moral concepts was incompatiblewithmetaethics.Theworryisthat,sincemetaethicsisa theoreticalproject,themetaethicistinevitablymakesacategorymistakeonthisaccount,likeachessplayerdecidingthatmakinghimself acoffeewouldbethebestmove.However,thisdiagnosiswouldfail toappreciatethescopeandflexibilityoftheoreticalreasoning.Iwill arguethatKorsgaardsaccountofpracticalconceptscanratherbebest understoodasentailingaspeciesofresponse-dependentmetaethical realism. Tobegin,considerthefollowingschemaforaresponse-dependent accountofapropertyF:  rd XsareFjustincaseandbecauseYsrespondR-ly toXsinconditionsC.22

This is a schema for response-dependent accounts of a property, because it is some response of Ys that metaphysically determines that anXisF.Forexample,acruderesponse-dependenttheoryofhumor wouldclaimthatforajoketohavethepropertyofbeingfunnyjustis forpeopletotendtorespondwithamusementuponhearingthejoke inordinaryconditions. ItispossibletousetheKorsgaardianaccountofpracticalconcepts
22. Johnstonintroducedthetermresponse-dependenceinhis(1989).Isetaside avarietyofimportantissueshere.Mostimportantly,Iassumetheontological respectabilityofresponse-dependentproperties.Readersattractedtoamore austereontologyshouldtranslatewhatfollowsintoadefenseoftheclaimthat thetwocontentsaccountpermits(derivative)theoreticalmoralconcepts.

This account grants that moral concepts can be nonderivatively graspedonlyfromwithinthepracticalperspective. According to practical response-dependence, facts derived from the practical perspective are necessary and sufficient to fix whether an option is permissible for an agent. However, practical responsedependenceisnonethelessanidentifiablymetaethicalview.23Thisis becauseitoffersanon-triviallyexplanatoryaccountofthepropertyof permissibility,graspablebyapurelytheoreticalreasoner. If this reading of the metaethical upshot of the Korsgaardian accountofpracticalconceptsiscorrect,thenthetwocontentsproposal, likethetwoattitudesproposal,turnsouttoentailmetaethicalcommitments.24However,unlikethetwoattitudesproposal,thisaccountdoes
23. Practicalresponse-dependenceisalsoobviouslyincomplete;afullresponsedependentmetaphysicsofmoralitywouldhavetoaccountforthepluralityof normativeproperties.However,thiscomplicationisorthogonaltowhether themostpromisingversionofthetwocontentsaccountentailsaresponsedependentmetaethics. 24. Tosaythatthetwocontentsaccountentailsaresponse-dependentmetaethics

philosophersimprint

11

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


referenceofanethicalconceptwillpointfairlydirectlyto practicalramifications.[2003,121n.44] Thepracticallogicaccountofmethodologythatshesuggestsdoesentailthatacorrectunderstandingofpracticalconceptswillentaildeterminate normative implications. However, this does not threaten the distinction between metaethics and normative ethics, as Korsgaard may appear to suggest. The lesson of the discussion in section 2 of metaphysicalutilitarianismandsimpleactutilitarianismwasprecisely the need to distinguish metaethical and normative theories, even if thereareentailmentsbetweenthem.Justasinthatcase,itisconceivablethatametaethicsthatmadenomentionofdistinctivelypractical conceptscouldentailanaccountoftherelationbetweenactions,motives,andrightnessstructurallyidenticaltotherelationsthatKorsgaard claimstobeentailedbythepracticallogic.Wewouldwanttosaythat suchametaethicswasdistinctfromtheKorsgaardianaccount,butentailedthesamenormativetheory.25 Inthissection,Ihaveexploredthreewaysofdevelopingthedeliberative requirementthe claim that moral theorizing must take place from within the perspective of first-person practical deliberationwiththeideathatsuchstrategiesmightvindicatetheautonomy
25. IhaveemphasizedthatIamusingsomeofKorsgaardshelpfulsuggestionsto elucidatethetwocontentsaccount,withoutclaimingittobeherconsidered view.Oneofherclaimsaboutmetaethicsisbothinconsistentwiththetwo conceptsaccountandindependentlypuzzling.   Sheclaimsthatbothmetaethicalexpressivismandrealismaretrue,albeit in a way that makes [them] boring (2003, 118; 122 n. 49). In elucidating thispoint,sheagainappealstothecontrastbetweenthepracticalandtheoreticalperspectives.Expressivismistruefromthetheoreticalperspective,she suggests,whilerealismistruefromwithinthepracticalperspective.Perhaps this is best read as a rhetorical overstatement. Note first that this appears inconsistentwithKorsgaardsowndiscussionofpracticalconcepts.Forif,as sherepeatedlyinsists,practicalconceptsdonotfunctiontodescribereality (2003, 105), then (as I have suggested above) correct theoretical reasoning aboutthepracticalcannotconflictwiththepracticalperspective.Notesecond that, as truth is arguably essentially perspective-independent, Korsgaards claim is extremely unappealing, because it makes truth itself perspectiverelative.HelpfulfurtherdiscussionofthisclaimcanbefoundinHussainand S  hah(unpublishedmanuscript,33ff.).

not permit conflict between the evaluations suggested by practical andtheoreticalreason.Thisisbecauseofthedeferentialnatureofthe theoreticalreasoningaboutmoralitypermittedbythisaccount. Inthissection,IhavedrawnonKorsgaardsdescriptionofthebasic tension between Parfits argument and the deliberative requirement, andIhaveusedsomeofherdiscussionofpracticalconceptstodevelopwhatItaketobethemostpromisingversionofthatrequirement. However,intheIntroductionInotedthatKorsgaardtookacorrectunderstandingofmoralitytoentailthaterrorsinfectstandardmetaethicalpractice.Whateverthedetailsofherview,Korsgaardisclearlysympathetic to some variation of the view that there is a deep contrast betweentheoreticalandpracticalreason.Havingexploredthreeways ofdevelopingthiscontrast,Inowexaminetwostrikingimplications thatshetakesittohaveformetaethics. First,Korsgaardclaimsthatmetaethicaltheorizingisdistortedby thepresuppositionthatallauthenticallycognitiveconceptsfunction todescribetheworld(2003,105).Inotherwords,mostmetaethicists do not see the possibility of the sort of practical concepts that she suggests. This is arguably true. However, the practical response-dependenceaccountthatIsuggestshowsthatthisaccountofpractical concepts is not incompatible with metaethics, but rather suggests a determinate metaethics. So understood, Korsgaards first criticism is bestunderstoodasa(quitedeep)methodologicalcriticismwithinmetaethics,ratherthanadiagnosisofanerrorinevitablyimplicitinthe domainasawhole. Second,Korsgaardsuggeststhatonceheraccountisunderstood, [W]ewillnotbeinclinedtothinkthatthereisadifference between doing meta-ethics and doing normative or practicalethics.Theattempttospecifythemeaningand
isnottoimplythatindevelopingtheformer,onewouldnecessarilybeaimingtosolvemetaethicalproblems.Rather,thepointissimplythatsuchan accountpermitsakindofderivativetheoreticalreasoningaboutmorality.By engaginginsuchreasoning,Iclaim,onecoulddeterminethatthepractical response-dependencemetaethicsfollowsfromthetwocontentsaccount.

philosophersimprint

12

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


theoutcomeofidealizedautonomousnormativeethicalinquiry.Such a view could be developed into a determinate metaethical theory.26 However, by making moral facts parasitic on what is by stipulation autonomousmoralinquiry,thisaccountwouldsecuretheautonomy thesis.Second,whilethediscussionofsection2suggeststhatitisa mistake to make neutrality between normative theories a constraint on counting as a metaethical theory, a metaethics could secure the autonomy of morality by being neutral between normative theories. Some contemporary sensibility theories and expressivisms are perhapsbestunderstoodasbeingneutralinthisway.27 Itmightbeobjectedthatametaethicscannotvindicatetheautonomyofmorality,becausethejustificationofsuchametaethicswillitself involveappeal tonon-moralconsiderations.As I emphasizedinthe Introduction,someofthecentraldesiderataonmetaethicaltheories appeartoconcernthefitbetweenmoralityandtheotherelementsof ourbestaccountsoftheworld.Theobjectorinsiststhat,farfromvindicatingtheautonomyofmorality,themetaethicaltheoriessuggested wouldthusbeinconsistentwithit.28 The plausibility of the apparent tension between metaethics and theautonomyofmoralityrestsonexactlytheconfusionthatmotivates thisobjection.Ananalogywillhelptoillustratewhytheobjectionis misguided.Supposethatthereisacorrectepistemicruleofinference Double Negation:
26. CompareSayre-McCord,whosuggeststhatmoralkindsarethosegroupings deemedsignificantbythebestmoraltheory(1997a,315;cf.also1997b). 27. Thus the sensibility theory sketched in Dreier (2002) is constructed to be scrupulouslyneutralbetweennormativetheories.Gibbardsnorm-expressivismalsoplausiblyvindicatestheautonomythesis.Inhisdiscussionofmoral inquiry,Gibbardidentifiescharacteristicpsychologicalpressuresthataselfconsciousnorm-expressivistwillbeunderinhernormativethinking(1990, 274ff). On Gibbards account, the appropriate response to those pressures couldonlybesettledautonomously,bywhetheronetookthemtobenormativelysalient. 28. This line of objection suggests that the autonomy of morality can only be defendedasabasiccommitment.SuchaviewmightexplainDworkinsotherwisepuzzlingdisinterestinjustifyingtheneutralityconstraint.

ofmorality,andsuggestaresolutiontotheapparenttensionbetween theautonomythesisandmetaethics.Iarguedthateachofthethree accountsthatIexamineddoesvindicatesomeapproximationofthe autonomy of morality. However, I suggested that only the two contentsaccountisfullysatisfyinginthisrespect:theotheraccountswere consistentwiththetruemoraltheorysbeingdiscoverablebytheoreticalreasoning.Ialsoarguedthateachofthesetheorieswascompatible with metaethical theorizing, and suggested that this discussion permittedaclarifyingresponsetoKorsgaardscriticismsofmetaethics. 4. Resolvingthetension IntheIntroduction,Iexplainedanapparenttensionbetweenmetaethics, as it is ordinarily conceived, and the notion that morality is autonomous.Insections2and3,Ihaveexaminedtworelativelyradical strategiesforvindicatingtheautonomyofmorality.Inthissection,I suggestadifferentstrategyforresolvingthistension.Recallthattwo of the accounts examined in section 3 involved or entailed particular metaethical theories. This suggests a general insight that can be appliedtoresolvetheapparenttensionbetweenmetaethicsandthe autonomyofmorality:ratherthanbeingnecessarilyinconsistentwith metaethicalresearch,theautonomyofmoralitycouldbevindicatedin partbyaparticularmetaethicaltheory. The fictionalist and practical response-dependence metaethics introduced in section 3 can be understood as instances of this phenomenon. Thus, the two attitudes and two contents proposals were introducedasmeansofvindicatingtheautonomythesis.However,I suggestedthattheseproposalseachinvolvedcommitmenttoviewsin metaethics.Somemetaethicswouldvindicatetheautonomyofmoralitymoredirectly.Ametaethicscouldvindicatetheautonomyofmoralitybycharacterizingthesemanticsorontologyofmoralityinsucha way asto secure the irrelevance ofnon-moral theses tothejustification of moral theories. Consider two examples of how a metaethics mightachievethis. First, consider the thesis that the moral facts are determined by

philosophersimprint

13

vol.8,no.6(july2008)


 dn

tristrammcpherson
If you believe that not-not-P then you ought to believethatP.29

Metaethics and the Autonomy of Morality


is plausible that the autonomy of morality depends for its truth on metaethicalfact. Itisimportanttodistinguishthisclaimaboutexplanatorydependence from the epistemological relationship between the autonomy thesis and metaethical theorizing. Thus, suppose that one finds the autonomy of morality intuitivelyplausible. On some epistemologies, suchplausibilitywouldentailthatcompatibilitywiththeautonomyof moralitywouldcontributetojustifyingtheacceptanceofametaethicaltheory.Vindicatingtheautonomythesismightthusserveasadesideratumononesmetaethicaltheorizing.30 Insection3,Iexaminedonestrategyforvindicatingtheautonomy thesis:developinganaccountofthedeliberativerequirementonnormativetheories. Thissectionsuggestsanotherstrategy:defendinga metaethics that directly vindicates the autonomy thesis. There is an importantcontrastbetweenthesestrategies.WhileIhavearguedin thissectionthattheautonomyofmoralityisnotintensionwithmetaethicsperse,itisintensionwithavarietyofmetaethicaltheories. Thissuggeststhatunlessanaccountofthedeliberativerequirement suchasthoseexaminedinsection3canbesuccessfullydefended,the defensibilityoftheautonomythesisrestsonthemetaethicalcasethat canbemadeforautonomy-vindicatingmetaethics,suchasthosementionedabove. Conclusion Ibeganthispaperbeganbybringingoutanapparenttensionbetween metaethicsandtheautonomyofmorality.Ithenconsideredtworadicalstrategiesforresolvingthistensionandsecuringtheautonomyof morality.Isuggestedthatthesestrategiesentailedavarietyofsignificant consequences for metaethics. Thus, some ways of securing the autonomy of morality would, if correct, suggest that contemporary
30.Onepossibilityisthatsuchadesideratummightexplainanimportantpartof what is worth salvaging from Moores notorious open question argument. ThispossibilitywassuggestedtomebyananonymousreaderforPhilosophers Imprint.

DoubleNegationentailsthatwhenonebelievesthatnot-not-P,one shouldbelievethatP.However,otherwisereasonablebutveryunlucky inquirersmightacquiremisleadingevidencethatDoubleNegationis notacorrectruleofinference.Andthatmightleadthemastray,when theythinkaboutwhattoconcludegivennot-not-P.However,ifDouble Negationisacorrectrule,itjustsettleswhattheyshouldconclude;the justificationofDoubleNegationisitselfaseparatematter. Thiscontrastbetweenthecontentofaruleanditsjustificationalso holdsfortheautonomyofmorality.Recallthattheautonomyofmoralityisthethesisthatnon-moralpremisesareirrelevanttothejustificationofnormativeethicaltheories.Now,supposethatametaethicaltheoryMistrue,andthatMentailsthatnon-moralpremisesare irrelevanttothejustificationofnormativeethicaltheories.Thenthe truthofMjustsettlesthatmoralityisautonomous.Theobjectorbalks becausethejustificationofMinvolvesnon-moralpremises.Thiscan onlyappeartobeanobjectionifonetakestheautonomythesistoapplynotjusttothejustificationofnormativeethicaltheories,butalso toits ownjustification.However,suchself-applicationisnopartofthe contentoftheautonomythesis. Nor could a plausible autonomy thesis ensure that it is always reasonable to believe that morality is autonomous. Compare again DoubleNegation:itmightbeunreasonableforourimaginedunlucky inquirers to accept or apply Double Negation, given their misleadinghigher-orderevidence.However,thisdoesnotentailthatDouble Negationisnotacorrectepistemicnorm.Anautonomythesisthatattemptedtoprecludethepossibilityofsuchmisleadingevidencewould simplybeobviouslyfalse.Theobjectionthusfails,andwithittheprimary motivation for resisting the simple resolution of the apparent tension between metaethics and the autonomy of morality. Thus, it
29. Thissuppositionisimplausiblysimple.Atminimum,adoublenegationeliminationruleofinferencelikelyneedstoread:Ifyoubelievethatnot-not-p thenyououghteithertobelievethatporceasetobelievethatnot-not-p.

philosophersimprint

14

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality


WorksCited Darwall, Stephen, Allan Gibbard, and Peter Railton. Towards Fin de SicleEthics:SomeTrends.Moral Discourse and Practice.Ed.Stephen Darwall, Allan Gibbard, and Peter Railton. New York: Oxford UniversityPress,1997.347. Dreier,James.MetaethicsandNormativeCommitment.Philosophical Issues 12(2002):Realism and Relativism.Ed.ErnestSosaandEnrique Villanueva.24163. Dworkin, Ronald. Objectivity and Truth: Youd Better Believe It. Philosophy & Public Affairs25.2(1996):87139. Dworkin,Ronald.RepliestoCritics:Zangwill.Brown Electronic Article Review Service. Ed.JamieDreierandDavidEstlund.Posted4/9/1997. <http://www.brown.edu/Departments/Philosophy/bears/homepage.html> Frankena, William K. Main Trends in Recent Philosophy: Moral Philosophy at Mid-Century. Philosophical Review 69.1 (Jan. 1951): 4455. Gibbard,Allan.MoralityasConsistencyinLiving:KorsgaardsKantian Lectures.Ethics110(October1999):14064. Gibbard, Allan. Wise Choices, Apt Feelings. Cambridge, MA: Harvard UniversityPress,1990. Harman, Gilbert. Explaining Value. Oxford: Oxford University Press, 2000. Hussain,Nadeem,andNishiShah.MetaethicsandItsDiscontents:A CaseStudyinKorsgaard,unpublishedmanuscript. Johnston, Mark. Dispositional Theories of Value. Proceedings of the Aristotelian SocietySupp.Vol.LXIII(1989):13974. Kagan,Shelly.Normative Ethics.Boulder,CO:WestviewPress,1998. Korsgaard,Christine.MoralityasFreedom.Kants Practical Philosophy Reconsidered.Ed.YirmiyahuYovel.Boston:Kluwer,1989a.2348. Korsgaard, Christine. Personal Identity and the Unity of Agency: A KantianResponsetoParfit.Philosophy and Public Affairs18.2(Spring 1989b):10132. Korsgaard, Christine. Realism and Constructivism in Twentieth-

metaethical theorizing typically involves a fundamental error. I also arguedthatsomeofthesestrategieswereindependentlyunattractive. In the final section, I suggested that the apparent tension between metaethicsandtheautonomyofmoralitycanbedissolved:farfrom puttingmetaethicsintoquestion,theautonomyofmoralitymaymost simply be secured by defending a metaethical theory that explains whynon-moralthesesareirrelevanttonormativeethicaltheorizing. While the approaches discussed in sections 2 and 3 at least initially suggestedadiminishedroleformetaethicaltheorizing,theresolution proposed here suggests that metaethical theorizing has an indispensibleroletoplayintheevaluationoftheautonomythesis.31

31. I am grateful to many people for helpful comments and discussion, especiallyMichaelSmith,TomKelly,GilHarman,PhilipPettit,NadeemHussain, SarahBuss,MarkSchroeder,BrendanRitchie,AmyShuster,JohnMaier,Nick Stang,RochelleEdinburg,RyanDavis,SamArnold,andSariKisilevsky.Iam especiallyindebtedtotwoanonymousreadersforPhilosophers Imprint,who each made many extremely useful suggestions. I would also like to thank PrincetonsUniversityCenterforHumanValuesforafellowshipthathelped tosupportmyworkonthispaper.

philosophersimprint

15

vol.8,no.6(july2008)

tristrammcpherson

Metaethics and the Autonomy of Morality

CenturyMoralPhilosophy.Philosophy in America at the Turn of the Century. APA Centennial Supp. to Journal of Philosophical Research (2003):99122. Miller,Alexander.Introduction to Contemporary Metaethics.Cambridge: Polity,2003. Moore,G.E.Principia Ethica.Cambridge:CambridgeUniversityPress, 1903. Nagel, Thomas. The Possibility of Altruism. Princeton: Princeton UniversityPress,1970. Nelkin,Dana.TwoStandpointsandtheBeliefinFreedom.Journal of Philosophy97.10(2000):56476. Parfit,Derek.Reasons and Persons.Oxford:ClarendonPress,1984. Sayre-McCord, Geoff. Different Kinds of Kind Terms. Philosophical Issues8(1997a):31323. Sayre-McCord, Geoff. Good on Twin Earth. Philosophical Issues 8 (1997b):26792. Sinnott-Armstrong, Walter. Moral Skepticisms. New York: Oxford UniversityPress,2006. Zimmerman, David. Meta-Ethics Naturalized. Canadian Journal of Philosophy10.4(December1980):63762.

philosophersimprint

16

vol.8,no.6(july2008)