This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
S ubtleties aside, a look at the topography of the disagree-mentdebate reveals amajordivide separatingpositions thatare generally hospitable to maintaining one’s confidence in
the face of disagreement, and positions thatwouldmandate exten-sive revision toouropinionsonmanycontroversialmatters.Letuscallpositionsofthefirstsort“Steadfast”andpositionsofthesecondsort “Conciliatory”.1The fundamental theoreticaldifferencebetweenthesetwocamps,itseemstome,liesintheirdifferingattitudestowardevaluatingtheepistemiccredentialsofopinionsvoicedbypeoplewithwhomonedisagrees.
Allpartieshold that theproperresponse to learningofanother’sdisagreementdependsonone’sepistemicevaluationof thatperson.Allpartiesholdthatone’sbeliefsabouttheotherperson’sintelligence,intellectualdiligence, acquaintancewith theevidence, and freedomfrombias, fatigue,or intoxicationarerelevant towhether (andhowmuch) that person’s disagreement should occasion one’s changingone’sbelief.Thecampsdiffer,though,onthisquestion:Inevaluatingtheepistemiccredentialsofanopinionexpressedbysomeonewhodisagreeswithmeaboutaparticularissue,mayImakeuseofmyownreasoningaboutthisveryissue?Clearly,totheextentthatImay,itwillfavorSteadfastness incertaincases.For thereasoning thatsupportsmyownviewaboutthedisputedmatterwillalsosupportthinkingthattheotherpersonhasgottenitwrong,atleastthistime,andthusthatIneednotworryaboutherdissent.
Tosimplifythediscussion,letusfocusoncaseswhereI’vearrivedatacertaindegreeofcredence inPandsubsequentlydiscover thatanotherpersonhasarrivedatadifferentdegreeofcredence.Appliedto this simple sort of case, the principle separating the two campsamountstosomethinglikethis:
Independence: In evaluating the epistemic credentialsof another’s expressedbelief about P, in order to deter-minehow(orwhether)tomodifymyownbeliefaboutP,
1. Itakethelabel“Conciliatory”fromElga(2010).
ImprintPhilosophers’
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –2– vol.11,no.6(march2011)
Conciliationism in general, aiming to show that the putting aside ofone’soriginalreasoningmandatedbyIndependenceleadstounaccept-ableconsequencesinothersortsofcases.
Below, I’ll first defend Conciliationism by arguing that Indepen-dencedoesnothavetheunappealingconsequencesthatsomehaveworriedabout.HavingmaderoomforaConciliationistaccount,I’llde-scribesomeissuesthatconfronttheprojectofdevelopingafullCon-ciliationistaccountofrationallyrespondingtodisagreement.I’llthenargue that these issuesmustbe faced evenby reasonable accountsthatrejectIndependence.Theissuesflowfromacertainfeatureofthewiderepistemologicalterritorythathasnotyetbeenwellexplored:ra-tionalaccommodationofevidencethatonehasmadecognitiveerrors.
1. Does respecting Independence amount to throwing away evidence?
ThefirstproblemI’dliketoconsiderisgivenforcefuldevelopmentbyThomasKelly(2010).KellyarguesagainstaparticularversionofCon-ciliationismhecallstheEqualWeightview:thatwhenIhavereasontothinkmyfriendisanepistemicpeer(thatis,thatsheisgenerallyequallyreliableinthedomaininquestion),andhavenoreason(inde-pendent of my own reasoning about P)tothinkherlessreliableaboutP,IshouldadjustmylevelofcredenceinPsoastosplitthedifferencewithher.Hisargumentproceedsviaaseriesofcases,twoofwhichIwilladaptheretomyterminology.
WrongandWronger:WrongandWrongeraremutuallyacknowledgedpeersconsideringwhetherP.Att0,Wrongformsa0.7credenceinP,andWrongerformsa0.9cre-denceinP.Theevidenceavailabletobothofthemactual-lysupportsa0.2credenceinP.WrongandWrongerthencomparenotesandrealizetheydisagree.Theyfollowthedictates of EqualWeight, and at t1 they compromise at0.8.3
3. TheEqualWeightviewmayofcoursebedefinedtorequirethissortofdiffer-ence-splitting(andthisisindeedanaturalandinitiallyappealingformforaConciliationistviewto take).But it’s important tosee thatConciliationismneednotbecommittedto thisgeneralpolicy. In fact, Iwouldarguethat itactuallyrunscountertothemotivatinginsightbehindConciliationism:thatwemusttakeaccountofthepossibilitythatwe’vemadecognitivemistakes,andthatthebeliefsofothersserveaschecksonourcognition.ConsideracasewhereIcometohavefairlyhighcredence—say,.92—inP,asfollows:myinitialinclinationistobeevenmorecertainofP,butIscalebackmyconfi-denceabitbecauseIknowImakesomemistakes.Ithenlearnthatmyfriend,whomI taketobemypeeronsuchmatters,hasalsoconsideredthe issueandhasbecome.91confidentinP.Isupposethatshearrivedathercredenceinmuchthesamewayas Idid.But it seemsthat learningofhercredenceshouldmakememoreconfidentthatIdidn’tmakeamistake.Ifthat’sright,I should raisemy credencebeyond .92,not lower it asdifference-splittingwould dictate. This verdict is entirely consistent with Independence. It isalsoconsistentwith(intheintuitivesense)givingmyfriend’sopinionequalweight,andevenwiththeviewadvocatedinElga(2007),whencetheterm“EqualWeightView”derives.Moreover, therearetechnicaldifficultieswiththeuniformdifference-splittingformulationofConciliationism(seeShogenji[2007]andJehleandFitelson[2009]).Butwhile I thinkthat it’s importanttonotethatneitherConciliationismingeneral,norgivingone’speer’sopin-ionequalweightinparticular,requiresuniformdifference-splitting,neitherKelly’sargument,normydiscussionofit,turnsonthispoint.TheimportantissueisaboutIndependence.
Kellynotes that theEqualWeight view counsels themboth to splitthedifference,eachendingupatt1withcredence0.5.Butthis,Kellyar-gues,iscounterintuitive.Beforetheirepistemiccompromise,RightandWrongwereinstronglyasymmetricalsituations.But,Kellysays,“ForanadvocateoftheEqualWeightview,thisseeminglyimportantasymme-trycompletelywashesout”onceRightandWrongadjusttheirbeliefs:
ItseemstomethatKellyisentirelycorrectinsayingthatweshouldnot see Right and Wrong as being in epistemically symmetricalsituationsatt1.Totheextentthatwedid,we’dbeoverlookingthebearingoftheoriginalevidenceonwhatRightandWrongshouldbelieve.AndthisisinfactthetrapKellyseestheEqualWeightviewasfallinginto:“Withrespecttoplayingaroleindeterminingwhatis reasonable for [them] to believe at t1, [the original evidence]getscompletelyswampedbypurelypsychologicalfactsaboutwhat[RightandWrong]believe”(ibid.,124).ThegeneralproblemKellylaysatthefeetoftheEqualWeightversionofConciliationism,then,is that itmakes rationalbelief indisagreementsituationsdependcompletelyonthe“psychologicalevidence”—evidenceaboutpeo-ple’sbeliefs.
Note that thisapparentproblemseems toflowdirectly fromIn-dependence. If Right could have relied on her reasoning about PinassessingWrong’sopinion,she’dhavehadareasonforthinkingWrongunlikelytobecorrectaboutP.Andinthatcase,Rightwouldnothavebeen required to compromise in theway shedid.So the
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –4– vol.11,no.6(march2011)
tellsuswhattheproperresponseistooneparticularkindofevidence.Thus the Equal Weight Conciliationist is committed to holding, inKelly’scases,thattheagentshavetakencorrectaccountofaparticu-lar bit of evidence—the evidenceprovidedby their peer’s disagree-ment.Buthavingtakencorrectaccountofonebitofevidencecannotbeequivalenttohavingbeliefsthatare(evenpropositionally)rational,allthingsconsidered.Ifonestartsoutbybotchingthingsepistemically,andthentakescorrectaccountofonebitofevidence,it’sunlikelythatonewillendupwithfullyrationalbeliefs.Anditwouldsurelybeask-ing toomuchofaprincipledescribing thecorrect response topeerdisagreement todemand that it includeacomplete recipe forundo-ingeveryepistemicmistakeonemightbemaking inone’s thinking.IfConciliationismisunderstoodintherightway,then,itisnotcom-mittedtodeemingthepost-compromisebeliefsinWrongandWrong-erautomatically rational.And inallowingus tocriticizeWrongandWronger’spost-compromisebeliefs,Conciliationismthusunderstooddoesnotentail,orevensuggest, thatWrongandWronger’soriginalevidencehasbecome irrelevant to the rationality of their post-com-promisebeliefs.4
AsimilarpointappliestotheasymmetryintheRightandWrongcase.Conciliationismdoesnot entail thatRight andWrong endupwithequallyrationalbeliefs.Nordoesitentailthattheywererational-lymandatedtomakeequallyextensiverevisionstotheiropinions.Ofcourse,itdoeshavetheconsequencethattherevisionscalled for by the
4. IshouldnotethatwhenIsaythatWrong’smovingto.5istakingcorrectac-countofonepieceofevidence,Idonotmeanthatmovingto.5iswhatWrongepistemicallyshoulddo,allthingsconsidered,uponreceivingthatevidence.ThedisagreementevidenceexertsrationalpressureonWronginthatdirec-tion,buttheoriginalevidenceexertsdifferentrationalpressure.It’snotcleartome that there is anoption forWrong that’s fully rational,holdingfixedhis initialerror.Ofcourse, there’sasense inwhich themaximally rationalresponseforWrong,andwhatheshoulddo,allthingsconsidered,istocor-rectlyassesstheoriginalevidence,andundohisinitialmistake.Butthecasesupposesthathedoesnotdothis.AskingwhatWrongshoulddo,holdingfixeda certainirrationalityonhispart,isdifferentfromsimplyaskingwhatheshoulddo.ThankstoStewCohenforpointingouttheneedtobecleareraboutthis.
Kellywrites:
On the Equal Weight view, [their] high level of confi-dencethat[P] is trueat time1 is theattitudethat is rea-sonablefor[them]totake,despitethepoorjobeachof[them]hasdoneinevaluating[their]originalevidence…However, it isdubious that rationalbelief is so easy tocomeby.(Kelly2010,126)
Again,Kelly’sintuitiveverdictonthecaseseemscorrect:weshouldnot see Wrong and Wronger’s post-compromise beliefs as ratio-nal.Again,todosowouldbetotreattheiroriginalevidenceasifitdidn’tmatter.
Thusitwould,Ithink,beverydamagingtoConciliationistviewsiftheirinsistenceonIndependenceamountedtoinsistingthatone’soriginalevidencewas irrelevant to the rationalityof thebeliefsoneendedupwithaftermakingone’sconciliatoryepistemiccompromise.Nevertheless,itseemstomethat,fortwoseparatereasons,theCon-ciliationistpositiondoesnothavethisconsequence.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –5– vol.11,no.6(march2011)
Andifeachofthemhasintheendreachedthesamecredenceonthebasisof the sameevidence,howcanwe say thatWrong’s credencefallsshortrationally,whileRight’sdoesnot?7
Tobeginthinkingaboutthispuzzle,supposeweapproachedtheexample by considering a third party confronted with E1–E3, andaskingwhat,fromaConciliationistpointofview,sheshouldbelieveaboutP. In fact, it seems tome that suchanagent confrontedwithE1–E3shouldnotendupgivingP0.5credence,asI’veclaimedthattheimpeccablyrationalRightshould.SuchanagentshouldofcoursetaketheimportofE1tobetorationalize0.2credenceinP.Butthenshe’dseethatoneotheragentagreesandonedisagrees.TheundercuttingpowerofWrong’sbeliefisdilutedbythesupportingpowerofRight’s.Sotheagentshouldendupwithcredencesomewherebelow0.5.8AndthisisinfactwhereKellythinksRightshouldendup.
ButdoesourconclusionaboutthethirdpartycarryovertoRight?Interestingly, I think, itdoesnot.And thereason for this involvesafacetoftheepistemologyofdisagreementthathasn’tbeenfullyarticu-lated:thattheevidentialforceoftheinformationexpressedinclaimslikeE2andE3dependscruciallyonwhethertheagentrespondingtotheevidenceisidenticaltooneoftheagentsmentionedinE2andE3.Inordertoseehowthisdependenceworks,letusfirstconsiderasim-plercaseinvolvingevidenceofpossiblecognitivemalfunction.
SupposeI’mparticipatinginplacebo-controlledtrialsofareason-distortingdrug.Thedrughasbeenshowntocausepeople tomakemistakes inalgebraic reasoningbut to leavemostof their cognitivefacultiesunscathed.Moreover,thoseaffectedbythedrugdonotnotice
8. This assumes that the agent has some reason for epistemically respectingRightandWrong,sosheshouldtaketheirviewsintoaccount.ItalsoassumesthattheagentdoesnotseeRightandWrongassuchexperts(relativetoher-self) that sheshouldnoteven try tofigureout the importof theevidencedirectly,butinsteadshouldjustbaseherbeliefsonRight’sandWrong’s.Onewayofavoidingbothissues:stipulatethattheagenthasexcellentevidenceofpeerhoodwithRightandWrong.
It turns out, then, that Conciliationism’s respecting of Indepen-dencedoesnotafterallrenderirrelevantthereasoningandevidenceonwhichConciliatoryagentsbasetheirinitialbeliefs.
2. A Follow-Up Objection and Agent-Specific Evidence
5. I have in mind cases where Right’s dispute-independent evaluation ofWrong’sbeliefwouldprovideextremelystrongreasontobelievethatWrongisequallywell-informed,andequally likely tohave reasoned fromtheevi-dencecorrectly,asintheMentalMathcaseabove.(SeeSections6and7formoreonthis.)
The fact that thefirst-personpsychological evidence is relativelyinertinthisrespectisexactlywhatonewouldexpect,giventhemain
9. I do not expect these judgments to be very controversial. Evenmost non-Conciliationistphilosophersconcedethatincaseslikethis,thedisagreementofafriendshouldmakemelessconfident;andtakingagreementofafriendtojustifyincreasedconfidenceisjusttheothersideofthesamecoin.
thattheiralgebraicreasoningisimpaired;infact,theyseemtothem-selvestobethinkingasclearlyanddistinctlyastheyeverdo.I’vebeenthroughseveraltrials,somewiththedrugandsomewiththeplacebo,andI’veneverseemedtomyselftohavebeenaffected;butwatchingthe tapes ofmyself in previous trials, I seemyself earnestly—evenheatedly—insistingonthepatentlymistakenconclusionsI’vedrawnontheassignedproblemswhenIgot theactivepills. It seemsclearthat,insuchasituation,ifI’mgivenapillandthenaskedtodrawaconclusionfromsomeevidencethatrequiresalgebraicinterpretation,IshouldbefarlessconfidentofmyanswerthanIordinarilywouldbe.
Nowsupposewerepresentedmyevidenceasfollows:
E1: The evidence presented as part of the experimentalproblem
E2: dchada50%chanceoftakinganactivepill
Clearly,arationalthirdpartypresentedwithE1andE2wouldnotbemuchbotheredbyE2.Infact,E2seemslikeitshouldbecompletelyevidentiallyirrelevanttothebeliefoneshouldendupwithaboutthealgebra problem—for everyone except dc. (More precisely, E2’s rel-evanceforanagentwilldependonthedegreetowhichtheagentbe-lievessheisdc.Onemightevenwanttofactorinwhethertheagent’sconfidence that she’sdc is rational.But letus leave these complica-tionsaside, and just consider caseswhereagentshave rationalandcorrectbeliefsabouttheiridentities.)Sointhissortofcase,theratio-nalimportofevidenceisagent-specific.
Now,whenI’mconfidentthatP,andfindoutthatmyfriendiscon-fident that not-P, the evidenceprovidedbydisagreement is at leastpartlyofasimilarsort.GiventhatmyfriendandIhaveaccesstothesamefirst-orderevidence,herdisagreementwithmeisalsoevidencethatI’vemisconstruedtheimportofthatfirst-orderevidence.Inthisrespect,it’sliketheunderminingevidenceinthedrugexample.ThissuggeststhattheevidenceRightandWronghaveinthecaseabovemightbedescribedmoreperspicuouslyasfollows:
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –7– vol.11,no.6(march2011)
disagreementshouldtypicallyoccasionsomechangeofbelief.Still,ina case likeRight andWrong,Kellynotes that thepsychological evi-denceisbalanced:Right’sbeliefpointstowardnot-PjustasstronglyasWrong’sbeliefpointstowardP.Suchbalancedpsychologicalevidence“tends to pushwhat it is reasonable for us to believe about the hy-pothesisinthedirectionofagnosticism”(2010,143).Thus,giventhatthenon-psychologicalevidencestronglyfavorsnot-P,it’sreasonabletoexpectthatthetotalevidenceintheexamplefavorsnot-P,thoughlessstronglythandoesthenon-psychologicalevidencealone.Ifthis“balancingargument”isright,thenevenifweadmitsomeconciliation,thecorrectcredenceinPtoadoptherewouldseemtofallwellbelowRight’s0.5,contraEqualWeightConciliationism.
11. Analoguesofthebalancingargumentclearlyfailincasesnotrelatedtodis-agreement.SupposethatJockovisitsanartmuseumtoseethenewexhibit,(unde)r(min)ed!Uponentering,JockobeholdsStudy: 4b,thefirstpaintinghecomes to. It appears tohim tobe a simple red rectangle. Jocko concludesthat(a)Study: 4bisaredrectangle,and(b)themuseum’scurrentexhibitisunlikelytoproverewardingforhim.Asheleaves,henoticesanartist’sstate-mentexplainingtheshow.Theartisthaspainted50%ofthecanvassesred,and50%white,andthenlit thewhiteoneswithdeceptivelightingsothattheylookjustliketheredones.ConsideringJocko’stotalevidenceastoStudy: 4b’sredness,wenowhave:
intuitive rationale for adjustingone’sbeliefs in the faceof disagree-mentwith anequally informed friend. Since I recognize that Imaysometimesmisconstruetheimportofevidence,Iseethatmyfriend’sreactiontothesameevidencemaywellconfirmordisconfirmmyhav-ingassessedthatevidencecorrectly.Butclearly,Icannotusemyownreactiontotheevidenceasacheckinthisway.Thusforme,psycho-logicalreportsaboutothersserveasakindofepistemicresourcethatpsychologicalreportsaboutmyselfdonot.
Thereisasense,then,inwhichRightandWronghavedifferentevi-dencetoreactto.10Ineachcase,wemaytakethefirst-personpsycho-logicalevidencetobeincapableofprovidingthesortofcheckonone’sreasoningthatthird-personevidenceprovides.Inthissense,itisrela-tivelyinert.Sotheimportantdeterminantsofwhat’srationalforRighttobelievearetheoriginalevidenceE1(whichshould,anddoes,moveher toput0.2credence inP),andWrong’sdissent(whichdoesand,accordingtotheEqualWeightConciliationist,shouldmoveherfrom0.2to0.5).Incontrast,thedeterminantsofwhatWrongshouldbelieveareE1(whichshouldmovehimtowardhaving0.2credenceinP),andRight’sbelief(whichalsoshouldmovehimtoward0.2).Lookedatthisway, it’snotsurprisingthathisarrivingat0.5rather than0.2 is lessthanfullyrational.
The upshot is this: Right’s andWrong’s evidential situations arenotsymmetrical.Uponcloserexamination,itturnsoutthattheirtwosituationsdonotrationalizethesamedegreeofconfidenceinP.AndthuswhenRightandWrongarriveatthesamedegreeofconfidenceinP,theConciliationistneednotconsidertheirdegreesofconfidenceequallyrational,orequallysupportedbytheevidence.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –8– vol.11,no.6(march2011)
through theproblem. I thencarefullycheckmyanswer,anditchecksout.Ithentakeoutmywell-testedcalculator,andredotheproblemandchecktheresultinafewdiffer-entways.AsIdoallofthisIfeelfullyclearandalert.EachtimeIdotheproblem,Igettheexactsameanswer,$43,andeachtimeIcheckthisanswer,itchecksoutcorrectly.Sincethemathproblemissoeasy,andI’vecalculatedandcheckedmyanswersocarefully inseveral independentways, I nowhave an extremely high degree of rationalconfidencethatoursharesare$43.Thensomethingverystrangehappens.Myfriendannouncesthatshegot$45!
Here,many people feel that I should not reducemy confidence in$43veryfaratall.AndthisintuitionholdsevenifwestipulatethatIcouldseemyfriendwritingnumbersonpaperandpushingcalculatorbuttons,andthatmyfriendassuresmethatshedidhercalculationsslowlyandcarefully,feltclearwhiledoingthem,andgothersamean-swerrepeatedly.ItseemsthatI’dbereasonableinthiscasetosuspectstronglythatsomethingscrewymustbegoingonwithmyfriend.12
Thisintuition—whichtoalargedegreeIshare—seemstocutdi-rectlyagainstConciliationism,andparticularlyagainstIndependence.Why,afterall,doIsuspect thatsomethingscrewywentonwithmyfriend? It’s just because she reported getting $45.And the only rea-sonthatthatwouldindicateanythingamissisthatI’mquitesurethatthe answer is not $45.Yetmy reason for being so sure that the an-swerisnot$45isjustmyownmeticulousreasoningshowingittobe$43!Thus,indescribingasimilarcase,Sosawrites:“NowIamintheMoore-likepositionofhavingtosaythat ifhisprocedurehasledtothatresult,theremustbesomethingwrongwithhisprocedure…Istill
Iwouldlikeinthenextfewsectionstoturntoexaminequiteadif-ferentsortofworryaboutConciliationism:thatitgetscertaincasesclearlywrong.Thefirstsortofhardcasesareoneswhereanagentbeginswithextremelyhighrationalconfidenceinherbelief.Invari-oussuchcases,itseemswrongtoholdthatsheshouldreviseherbe-liefmuchatall,eveniftheagent’sfrienddisagreessharply,andevenif,beforediscoveringthedisagreement,shewouldhaveconsideredthefriendherepistemicpeeronthesortof issueinquestion.Thissuggests that it is, after all, legitimate for the agent todemoteherfriend’sdissentingopiniononthebasisofherownreasoningonthematterunderdispute.Inotherwords,itsuggeststhatIndependencefailsinthesecases.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –9– vol.11,no.6(march2011)
isthatoneofushasexperiencedsomebizarrementalmalfunctionresultinginerrorsthatsomehowledtothesamewronganswerinalltheindependentwaysofdoingandcheckingtheproblem.Anotheris thatoneofus isactuallyexhausted,ordrunk,or tripping,orex-periencingaconfusingpsychoticepisode,andisreallyonlymanag-ingtogothroughtheexternalmotionsofrecalculatingandchecking,without actuallypaying clear attention.Still another is thatoneofusisjustjoking,ormessingwiththeother’sheadforfun.Anotheristhatoneofusisdeliberatelymakingfalseclaimsabouthisorheranswer,forthepurethrillofbald-facedlying,oraspartofapsycho-logicalorphilosophicalexperiment,orperhapsinanearnestattempttoproblematizethehegemonyofphallogocentricobjectivitybyanactofperformanceart.
Thisisnotanexhaustivelistofexplanationsforthedivergenceofourannouncedanswers.Butitisenough,Ithink,toshowwhyIaminapositiontobelievethattheanswermyfriendannouncedislesslikelytobecorrectthanmineis.Forexample,whileIcandefinitivelyruleoutthepossibilitythatI’vedeliberatelyannouncedanincorrectanswerforrecreational,experimental,orperformance-artisticreasons,Icannotbenearlysosureofrulingoutthesepossibilitiesformyfriend.Similarly,whileIcanbeverysurethatIwasactuallypayingattentionratherthangoingthroughthemotionsofcheckingmyanswer,Ican-notbenearlysosurethatmyfriendwas.Andwhilethereareconceiv-ablesortsofmentalmalfunctionthatwouldaffectmyreasoningwith-outmyhavinganysignoftrouble,mostreason-distortingmentalmal-functions comewith clear indicationsofpossible trouble:dizziness,seeingpatternsmovingonthewall,memoriesofrecentdrug-takingorofpsychoticepisodes.AndI’minamuchbetterpositiontoruletheseoutformyselfthanIamformyfriend.LetmeputtheinformationI’mdependingoninallthesecasesunderthecommonlabel,takenfromLackey, of “personal information”. The personal information I haveaboutmyselfinCarefulCheckingprovidesaperfectlyreasonableba-sisformycontinuingtothinkthatoursharesofthebillaremuchmore
lack independent reason to downgrademy opponent’s relevant judg-ment andhis epistemic credentials on the question that divides us.OnlybasedonourdisagreementcanInowdemotehim”(forthcom-ing,ms.18–19).AndLackeysaysthatcasesinvolvingextremelyhighjustifiedconfidence“showpreciselywhycondition(1)[aformulationofIndependencefromChristensen(2007)]shouldbeeliminatedfromChristensen’saccount”(forthcoming,ms.45,fn.33).
I think that, on closer inspection, cases involving ultra-high ini-tialrationalconfidencedonotendupunderminingIndependence.13 LetusbeginbyconsideringwhatIshouldthinkofmyinitialopinioninCarefulChecking.Beinggenerallycompetentatelementarymathproblems,havingdonethecalculationsrepeatedlyandcarefullybothon paper andwith a well-tested calculator, having checked the an-swerinmultipleindependentways,andfeelingveryclear-headedandalertthroughout,Ishouldthinkthatitwouldbeextremely unlikelyforsomeoneinmysituationtohavegotten(andverified)thesamewronganswereachtime.Thatgoeshand-in-handwiththelegitimacyofmyhavingultra-highconfidenceinmyanswer.
Butifthat’sright,here’ssomethingelsethatwouldbeextremelyunlikely:twopeople,bothgenerallycompetentatelementarymath,whoworked on the same problem, each having done the calcula-tionsrepeatedlyandcarefullybothonpaperandwithawell-testedcalculator,eachhavingcheckedtheanswerinmultipleindependentways,eachfeelingveryclear-headedandalertthroughout,andeachrepeatedlycomingupwith(andverifying)adifferentanswer.Thisisimportant,becauseitmeansthat,inthestrangescenariodescribedin Careful Checking, I have good reason to think that somethingscrewyhasgoneon.
13. MyexplanationforthisfollowsthestrategybrieflysketchedinChristensen(2007,200–203).ItalsodrawsheavilyonLackey’s(forthcomingand2010)insightful analysis of this type of example. The conclusion I’ll draw aboutIndependence,however,isoppositefromhers.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
NoticethatthisreasoningdoesnotevenrefertotheparticularanswerIgot. In fact, the reasoningcouldhavebeen formulated inadvanceofmydoinganycalculation,orevenseeingthebill!Thisshowsthatwhilethereasoningreliesoncertainfactsabout the reasoning I use, itdoesnotrelyonmyreasoningitself.Ittakesintoaccountthefactthatwedisagreed,butitdoesnotdependonthesubstanceofthedisagree-ment.Soitdoesnotbegthequestionagainstmyfriend’sbeliefinthewayIndependenceisdesignedtoprevent.
Thusmyreason formaintainingmybelief in thiscase isentirelyconsonant with the sort of positions advocated by Conciliationistwriters.Itisobviousthatinconsideringtheepistemicimportofone’sfriend’s expressed beliefs, onemust take into account certain factsaboutone’s reasoning. If Iknowthat I’vebeenreasoningwhile trip-ping,or if Iknowthat the reasoningmethod I’veused isonlymod-erately reliable, that givesme reason to accordmoreweight tomyfriend’sdisagreement(to theextent,ofcourse, that Idoubt thatherownreasoningsuffers fromthesesortsofweaknesses).Similarly, totheextentthatIknowmyownreasoningtohavebeenofaparticularlyreliable sort, that givesme reason togivemy friend’sdisagreementlessweight(totheextent,again,thatIdoubtthatherownreasoningisofthissameparticularlyreliablesort).AndImaybringthesesortsofconsiderationstobearwithoutrelyingonmyowninitialreasoningconcerningthedisputedmatter.
Itseemstomethatthissortoftreatmentappliesparticularlynicelytotheabove-mentionedextremeexamplesofferedbycriticsofInde-pendence:caseswhereone’s friendclaimstobelievethatthere’snooneelseatthetable,orthat2+2=5.Ifsuchabizarresituationwereactually tooccur, I thinkonewould reasonably take itasextremelyunlikely thatone’s friend(a)was feelingasclear-headedasoneself;(b)hadnomemoriesofrecentdrug-ingestionsorpsychoticepisodes;andmostimportantly,(c)wasbeingcompletelysincere.Thus,touse
Ifthisismybasisformaintainingmybelief,haveIviolatedInde-pendence?ItseemstomethatIhavenot.True,insupportingmysus-picionthatsomethingscrewyhasgoneonwithmyfriend,IreliedontheclaimthatIarrivedatmyanswertothemathproblembyaveryreliablemethod.Butmy reasoningdidnot relyon the resultsofmycalculationsatall.Ididnotsay,“WellI’mverysuretheansweris$43.Myfriendsaysit’s$45,sosomethingscrewymusthavegoneonwithher.”ThatsortofreasoningwouldindeedviolateIndependence.ButthereasoningIusedwasquitedifferent.Itwasmorelikethis:
Iarrivedatmyanswerbyanextremelyreliablemethod.Itisveryunlikelythattwopeopleemployingsuchmethodswouldendupsincerelyannouncingincompatiblebeliefs.Thebeliefmy friendannouncedwas incompatiblewiththeoneatwhich Iarrived.This is strongevidence thatoneofusdidnotarriveathisorherbeliefinahighlyreli-ableway,orthatoneofusisnotsincerelyannouncinghisorherbelief. Icaneliminate(viapersonal information)manyof theways that I couldhave failed touse a reli-ablemethod,aswellasthepossibilitythatmyannounce-mentwasnotsincere.ButIcannoteliminateanalogouspossibilities formyfriend.Soit’s likelythatshedidnot
Frances (2010)andFumerton(2010)alsoargue that incertaindisagree-mentswhereonebeginswithultra-highrationalconfidence,onewillreason-ablysuspectthatone’sfriendisjokingorcrazy,andthusoneneedn’treviseone’sbelief.Neithermakesthispoint inthecontextofevaluatingIndepen-dence(thoughinotherpartsoftheirpapers,Francesseemssympatheticto,andFumertonseemstodeny,somethinglikeIndependence).
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –11– vol.11,no.6(march2011)
everydaysortofslip.Unfortunately,mypersonalinformationdoesnothelpmetoeliminatethispossibility formyself.Ofcourse, therearealso theexoticpossibilities consideredabove: thatoneofus is trip-ping,psychotic,joking,lying,etc.Andmypersonalinformationdoesallowmetoeliminatevariousexoticpossibilitiesformyselfandnotformyfriend.Butsincetheseexoticscenariosaresounlikely,thefactthatIcaneliminatesomeofthemhasonlyatinyeffectontheplausibilityofexplainingthedisagreementinawaythatinvolvesthefalsityofmyfriend’sclaim.ThatiswhyIshould(incategoricalterms)suspendbe-lief,or(ingradedterms)comeclosetosplittingthedifferencewithmyfriend,inthesenseofseeingthetwoanswersasaboutequallylikelytobecorrect.15
InCarefulChecking,bycontrast,thehighdegreeofrationalconfi-denceIhaveinmyinitialbeliefiscorrelatedwithmyrationallytakingmyreasoningmethodtobeextremelyreliable.Anditistheextremereliabilityof thismethod,amethodwhicheliminates the “everydaymentalslip”explanationofourdisagreement,whichbothmakesthissortofdisagreement sounusualandmakes theexoticexplanationsvastlymoreprobable,shouldadisagreementoccur.(Thisiswhyit’sonlyinthesecasesthatI’llthinkthatsomethingscrewymustbegoingon.)Atthispoint,whenpersonalinformationallowsmetoeliminateseveralexoticpossibilities formyself,butnot formy friend, thebal-anceofprobabilityisshifteddramaticallyovertoexplanationsinvolv-ingthefalsityofmyfriend’sexpressedbelief.
Itisworthconsideringoneobviousobjectiontotheclaimthatmain-tainingbeliefintheabovecasesisconsistentwithIndependence.Theobjectionisbasedonacomparisonbetweenthecaseswe’vebeendis-cussing and cases involving significantly lower degrees of initial ra-tionalconfidence.Onemightpointoutthatthepersonalinformation,whichprovides the independentbasis formy thinkingmyselfmorelikelytoberightinCarefulChecking,isalsopresentinMentalMath,whereoneclearlyshouldreduceone’sconfidencedramatically.AndasLackeypointsout,theobviousdifferencebetweenthiscaseandCare-fulCheckingissimplythedegreeofrationalconfidenceIhaveinmyinitialopinion;soinsomeway,myhighinitialrationalconfidenceen-ablesthepersonalinformationtoplayitskeyroleinCarefulChecking.Thismightleadonetosuspect,then,thatmymaintainingmybeliefinCarefulCheckingmustafterallrely onmyreasoningconcerningthedisputedmatter.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –12– vol.11,no.6(march2011)
confident inP.But as the conversationdevelops, I findthatStrangerandIdisagreeequallysharplyaboutQ,R,S,T,andsoon—ahugelistofclaims.Andtheseclaimsarenotpartofsometightlyinterconnectedsetofclaimsthatwouldbeexpectedtostandorfalltogether:they’relargelyindependentofoneanother.DoInowhavetobe-comesignificantlylessconfidentaboutallofthem?
Seminar, Continued: In addition to believing anteced-entlythatI’mquitereliableinmeteorology,andthatthevastmajorityofothersareaboutequallyreliable, Ialsobelievethatthereareaveryfewpeople—callthemmete-orologicallyderanged—whoarehorriblyunreliable.I’m
personalinformation—muchmoreprobable.Butnoneofthisunder-mines Independence. For in adjudicating explanations for our dis-agreementinanyofthesecases,Idonotrelyonmyreasoningaboutthedisputedmatter.
Beforeleavingdiscussionofcasesinvolvingextremelyhighratio-nalconfidence,it’sworthemphasizingapointabouthowthesecasesrelatetoConciliationismingeneral.It’sobviousthatmostoftheissuesthat are subject to controversy are nothing like the issue ofwheth-erourfriendisbeforeus,orwhether2+2=4.Ahallmarkofthelattercases—whichis intimatelyrelatedtotheir involvingextremelyhighlevelsofrationalconfidence—isexactlythatbeliefsformedinthesewaysarevirtuallyneversubjecttodisagreement.Soit’sworthnotingthateveniftheConciliationistsharestheSteadfastview’sverdictoncasesinvolvingextremelyhighrationalconfidence,thereisnoreasontothinkthattherationalpermissibilityofmaintainingone’sbeliefinthesecaseswillbleedoverintothecontroversialcaseswhichgivethedisagreementissuesomeofitsurgency.Thisisthereasonthatthethe-oreticaldiagnosisof theextremecases—and inparticular, theques-tionofwhethertheyrequireviolationofIndependence—isimportant.
Seminar: I’m in a meteorology graduate seminar withStranger, another graduate student. I don’t know himvery well, but his first few comments seem quite sen-sible tome. I takemyself tobeaprettyreliable thinkerinmeteorology,thoughnotmorereliablethanmostgradstudents.Atthebreak,Idiscoverthatwe’vebothreadafairamountaboutissueP,butwhileI’mquiteconfidentthatP,Strangerexpressesequalconfidencethat~P.I,be-ingagoodConciliationist,thenbecomesignificantlyless
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
ThewayI’veexpressedtheagent’sattitudesinthecaseabovedis-tinguishesbetween“singleclaims”and largeconjunctionsofclaims.It’s worth noting that this convenient classification need not bearheavy theoreticalweight.Tobring thispointout, supposesomeoneobjectedasfollowstotheaboveanalysis:
You say that if Stranger disagrees about just one claimP, you should become agnostic on P. But suppose thatStranger’s initial claim is (~P&~Q&…&~Z),wheretheconjunctsarethenegationsofalltheparticularclaimsinvolved in the repeated disagreement you describedabove. Does Conciliationism now say you should givesignificantcredencetothisclaim?Thatwouldmeantak-ingitasreasonablylikelythatallofP,Q,R,etc.arefalse,whichwould require your becoming agnostic (at best)abouteachindividualclaim.
Toanswerthisquestion,weshouldnotethatit’snopartofConcili-ationismthatonetakesimilarattitudestoallofthepropositionsonebelieves.Imay,asstipulatedinSeminar,havefairlyhighconfidenceineachofPthroughZ.ButthereareotherclaimsIhavemuchgreaterconfidence in.Givenmy fairlyhighconfidence inP throughZ,andgiventheirrelativeindependencefromoneanother,Ioughttohaveextremelyhighconfidencethatthey’renotall false, i. e., that~(~P&~Q&…~Z).ThisisjustanotherwayofsayingthatI’mextremelycon-fidentthatI’mnothorriblyscrewedupepistemicallyinmeteorology.SoifStrangerhadassertedhisbigconjunctionrightatthebeginning
Giventheseassumptions,whenIfirstfindoutthatStrangerandIdis-agree about P,Conciliationismwould counselme to becomemuchlessconfidentinP.Butwhenwediscusstwodozenmoreclaims,andhedisagreeswithmeaboutallofthem,Ishouldnowthink:giventheextentofourdisagreements,it’sincrediblyunlikelythatStrangerandI arebothvery reliableat all.Themoredisagreementswediscover,themorelikelyitisthatoneofusisderanged.SinceI’mmoreconfi-dent(independentof thedisagreement) that I’mnotderangedthanthatStrangerisn’t,IshouldbecomemoreconfidentthatI’mbetteratevaluatingtheevidencethanStranger is.17 If this isright, Ineednotbecomemuchlessconfidentaboutall thethingswedisagreeabout.Thus,evenontheConciliationistview,Ineednotbedriventowide-spreadagnosticismbyStranger’srepeateddisagreement.Andindeed,IshouldregainmostofmyoriginalconfidenceinP.18
17. White (2009, 247–249) contains a somewhatmore formal discussion of asimilaridea.
18. Ishouldemphasizeapointabout theasymmetrythatfigures inthisexam-ple. It is crucial thatmyextreme confidence that I’mnotmeteorologicallyderangedheredoesnotjustderivefrommylowestimateoftheproportionofderangedpeople;thatwouldnotproduceanyasymmetrybetweenmeandtheStranger.TheideaisthatIbeginbybeingmoreconfidentthatI’mnotderangedthanwouldbejustifiedsimplybythelowproportionofderangedpeople,justasImightbelievethat1%ofadultsareparanoidschizophrenics,butbemuchmorethan99%surethatI’mnotoneofthem.Thankstoaref-ereeforPhilosophers’ Imprintforbringingtomyattentiontheneedtoclarifythispoint.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –14– vol.11,no.6(march2011)
inPshoulddropdramatically.Thisisbecauseinsingle-claimmatch-upsbetweenmyalmanacandarandomlyselectedone, Ishouldex-pectminetobewrongina fairproportionof the(rare)caseswhentheydisagree.However, if I lookup16more factsat random inmyalmanac,andtheStranger’salmanacdisagreesoneveryoneoftheseclaims,Ishouldnotdramaticallylowermyconfidenceinall17claims.Instead,IshouldconcludethattheStranger’salmanacisverylikelyanAcme,andinfactIshouldthengobacktotrustingwhatmyalmanacsaysabouteven thefirstof the17claims.This isbecause, inmatch-upsbetween twoalmanacs,when theydisagreeoneveryoneof 17claims lookedup, theoverwhelmingly likelysituation is thatoneofthemisanAcme(althoughAcmesaresomewhatrare,thechanceoftwohighlyreliablealmanacsdisagreeing17timesinarowisincred-iblyminuscule).Andwe’vestipulatedthatI’minitiallyextremelysurethatmine’snotanAcme.Thisinitialasymmetry—analogoustomybe-ingmuchmoreconfidentthatI’mnotmeteorologicallyderangedthanthattheStrangerisn’t—isallthat’sneededtoproducetheConciliatorystructuredescribedabove.19,20
In case the reader is suspicious about the structure of the Semi-narexample,here isamodelwith thesamestructure thatabstractsfromanycontroversiessurroundingdisagreement:Considerasitua-tion inwhicheachpersonhasanalmanac, andeachperson’s alma-nac isunique(they’renotcopiesof thesamebook).Suppose that Ihaveextremelygoodreasontobelievethatmyalmanacishighlyac-curate—that98%of thetimeit’scorrect(myreasonmightbethat Iboughtitfromaveryreliablepublisher).Moreover,Ihavegoodrea-sontobelievethatthevastmajorityofotheralmanacsareasreliableasmine.ButIalsobelievethatthereareasmallnumberofhorriblyin-accurateones(perhapstheAcmePublishingCompany—averyunre-liablepublisher—producedthemwithoutresearchinganyfacts).Thiscorresponds,inanartificiallysimplifiedway,tothesortdistributionofreliabilityamongpeoplesupposedinSeminar.SoItakethedistribu-tionofaccuracyinalmanacstobethis:
Inthissituation,ifmyalmanacsaysP,Ishouldinitiallybeverycon-fident thatP is true.But if I thenfindout thataStranger’salmanac,aboutwhichIhavenospecialinformation,disagrees,myconfidence
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
Weshould startbynoting that Independence,by itself,doesnotmandateanythingaboutbelief-change.Ittellsustoevaluatetheepis-temiccredentialsoftheotherperson’sbeliefinadispute-independentmanner.22 But it doesn’t saywhat to dowith the evaluation. To geta Conciliationist result, onemust add a principlemandating belief-revision in theevent that theevaluationyields certain results.Nowthissortofprinciplemightbedrawnupindifferentways.Letuscon-finediscussionatthispointtocaseswhereanotherpersondisagreessharplywithmyinitialbelief.
(B) Insofar as the dispute-independent evaluation gives megoodreasontobeconfidentthattheotherpersonisequal-lywell-informed,andequallylikelytohavereasonedfromtheevidencecorrectly,Imustrevisemybeliefinthedirec-tionoftheotherperson’s.23
Clearly, thefirstsortofprinciple threatens to leadstraight toskepti-cism.Forsupposethattheotherpersondisagreeswithvirtuallyallmybeliefsabouttheworld.Thatwouldleavemewithvirtuallynothing
22. By“dispute-independent”,hereandbelow,Imeannotrelyingonthereason-ingbehindmyinitialbelief.Adispute-independentevaluationcould,forex-ample, take into account the fact that theotherpersondisagreeswithmeaboutP,aswesawinthinkingabout,e. g.,CarefulChecking.
23. The contrast between (A) and (B) bears strong similarities to that be-tween Harman’s (1986) Principles of Negative Undermining and PositiveUndermining.
Insum,then,itturnsoutthatConciliationismcaneasilyaccommo-datetheintuitionthat incertaincases,repeateddisagreementswitha stranger, particularly about relatively independent claims, shouldoccasion little change inbelief,while a singledisagreement shouldsignificantlyreducemyconfidence.Anditseemstomethatthisisnotamereformalpossibility.Inmanycases,thebackgroundassumptionsneeded to underlie this phenomenon are reasonable: one believesthatoneself,andpeoplegenerally,areprettyreliableaboutacertainfield;onebelievesthatthereareafewpeoplewhoareepistemicallyvery screwed up and thus terribly unreliable; and one is extremelyconfidentthatoneisnotoneofthosepeople.
6. Does Independence Lead to Wholesale Skepticism?21
Onemightthinkthattheanalysisgivenintheprevioussectionofre-sponses toapparentlyderangedstrangerswas tooeasy,by theCon-ciliationist’s own standards. After all, Independence requiresme toabstainfromrelyingonthereasoningunderdisputeinevaluatingtheepistemiccredentialsofanother.But it’sclearlypossibleforanotheragent to dispute a verywide swath ofmy beliefs—a swath that in-cludestheclaimsIwasrelyingoninthepreviousanalysis.SupposeImeetamanonthestreetwho,afterexpressingdisagreementwithmybeliefthatP,adds,“Mostpeoplearecompletelyscrewedup.Onlya fewofusaresane.There’smassivedelusionabout.”Supposehe’swearinganaluminum foilhat, andhe tellsmehe’s just seenElvis’simageinhismorningwaffle,andthatElvistoldhimthatonlythosewithfoilhatscanthinkstraightatall.Clearly,FoilHatGuyisdenyingaclaim(thatI’mnotincrediblyscrewedup)confidenceinwhichunder-laymyresponsetorepeateddisagreementinthelastsection.Andthis,combinedwithIndependence,seemstomeanthatIcan’tdismisshiswholesalechallengetomybeliefsbyrelyingonmyconfidenceinthisclaim.Ontheotherhand,it’salsointuitivelyobviousthatIshouldn’t
ThissortofapproachmayalsohelpConciliationismmeshwithan-otherintuitivejudgmentsomehavedefended.Ithasbeensuggestedthatwhenthepropositionindisputeispartofadeepnetworkofbe-liefs,allofwhicharechallengedbytheotherperson,therationalpres-suretorevisebeliefsshouldbeless.24Elga(2007,495ff.)considersthecaseofAnnandBeth,whodisagreenotonlyoverthemoralityofabor-tionbutaboutawholenestofrelatedmoral,religious,andordinaryfactualmatters.Elgasuggeststhatinthissortofcase,therewillbenofact of thematter abouthowAnn—independentlyofdisputedmat-ters—would evaluate Beth’s likelihood of being correct about abor-tion.I’mnotsurethisisright.Afterall,theremaybequiteabitthatAnndoesknowaboutBeth:thatshe’shuman,intelligent,familiarwithcertain arguments, honest, etc. Thismightwell yield some dispute-independent assessmentofBeth’s reliability.25 But evenputting thisworryaside,thephenomenoninquestionisintuitivelyagradualone:astheareaofdisagreementbecomesgreater,thepressureforconcili-ationdiminishes.AnditishardtoseehowthetransitionfromtherebeingafactofthematteraboutAnn’sdispute-independentopinionofBethtotherebeingnofactofthematterwouldbeabletoaccountforthisgradualphenomenon.
Bycontrast,ona(B)-styleview,as theareaofdisagreementgetswider, the basis for dispute-independent assessment gets narrower,andthusthestrengthofthereasonsprovidedbythatnarrowerbasiswilloftenbereduced,andinagradualmanner.Soinsofaraswefinditplausiblethattherationalpressuretoreviseone’sbeliefsdiminishesastheextentofdisagreementwiththeotherpersonincreases,aprin-ciplealongthelinesof(B)isanattractiveoption.
on which to base a dispute-independent evaluation of the relativelikelihoodofhis reasoningcorrectly.But suchabaselessevaluationwouldclearlynotgivemeagoodreasonforconfidenceinmyhavingreasonedcorrectly;soprinciple(A)wouldrequiremassiverevisioninmybeliefs.Theproblemwith(A)isthatitineffectturnsouttorequirethatonehaveanon–question-beggingresponsetotheskeptic.
Ontheotherhand,thesecondsortofprincipledoesnotobviouslyhavethesamevulnerability.Foriftheotherpersonchallengesvirtu-ally all of my beliefs, an evaluation based on dispute-independentgrounds—i. e.,anevaluationbasedonvirtuallynothing—cannotpro-videgoodreasonformetobeconfidentabouteitherparty’slikelihoodofhavingreasonedcorrectly.
Thusthe‘insofar’inprinciple(B)shouldbeunderstoodasindicat-ing that the undermining power of disagreement by apparent epis-temicequalsisnotall-or-nothing.Whenoneevaluatestheepistemiccredentialsofanother’sbeliefsinadispute-neutralmanner,onemaygetvaryingstrengthsofreasonsforthinkingthattheotherpersonisaswell-informedandaslikelytohavereasonedcorrectlyasoneisone-self.Thestrongerone’sreasonsforthinkingequallywelloftheother’sepistemiccredentials, themoreoneshouldreviseone’sbeliefs.Andthis,Ithink,ishowitshouldbeintuitively.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –17– vol.11,no.6(march2011)
(1) One’s dispute-independent evidence supports an assess-mentofherepistemiccredentialsthatyieldshighestimatesfor:
a. thelikelihoodthatherexpresseddisagreementissincere,
b. herdegreeofwell-informedness,and
c. the likelihoodof her having reasoned correctly from theevidenceshehas;
and
(2) Thereasonsfortheseassessmentsofa–carestrong.
Incaseswhereone’sdispute-independentevidenceyieldsverystrongsupportfortheclaimthattheotherpersonissincere,aboutequallywell-informed,andaboutequallylikelytohavereasonedcorrectlyfromherevidence,oneshould cease to be much more confident that P thanthat~P.
Evenhere,however,manyquestionsremainopen:Howfar,ingeneral,shouldonereviseone’sbelief?Howdoestheprincipleextendtocaseswhere initial confidence is distributeddifferently—e. g.,wherebothparties aremore confident that P than ~P but to different degrees?Aswesawabove(fn.3),theinitiallyattractiveideaofuniformlysplit-tingthedifferenceincredencesdoesnotsitwellwiththemotivationsforConciliationism.
Sowearealongwayfromhavingaformula,orevenarecipewithquantities, describing in general how one should react to disagree-ment.Giventhecomplexityofthesubject,andtheearlystateofthediscussionofthisissue,thisshouldnotbesurprising.Andofcourse,Conciliationismisinnowayspecialhere;non-Conciliationistwritershavebeenequallyunabletocomeupwithprecisegeneraldirectionsfor accommodating disagreement evidence. But the burden of the
7. Unresolved Issues
Thus far,we’ve seen that severalworries that have been expressedabouttreatmentsofdisagreementbasedonIndependence-styleprin-ciplesmaybeassuaged.Thatis,ofcourse,alongwayfromofferingadetailedConciliationist recipe foraccommodating theevidencepro-videdbythedisagreementofothers.Inthissection,I’dliketonotetwoofthemajorissuesfacingthisConciliationistproject.
a. Formulating a principle for belief-revisionAsnoted above, Independenceneeds to be supplementedby someprincipledescribinghowthedispute-independentassessmentoftheotherperson’sepistemiccredentialsshouldinformone’sresponsetodisagreement.Andprinciple(B),asstated,isstillnothinglikeageneralprincipleofthissort.Foronething,itignoresthequestionthatwesawabovetobeimportant:whethertheotherperson’sexpresseddisagree-mentissincere.Moreimportantly,it’srestrictedtoevidencethattheotherpersonisone’sepistemicequal,andthisisobviouslyjustaspe-cialcase.Intuitively,totheextentthatthedispute-neutralassessmentgives reason to believe that the other person is one’s superior, oneshouldmoveone’sbelief farthertowardthatof theother.Andevenifthedispute-neutralassessmentgivesgoodreasonforthinkingtheotherpersononlyslightlylessqualifiedepistemically,somechangeofbeliefinthedirectionofherbeliefwilloftenbewarranted.Moreover,(B)saysnothingabouthowmuchrevisionisrequiredinanycase;asitstands,it’snotevenclearlyveryConciliatory.WemightcomeclosertoageneralConciliatoryprinciple(atleastforthemost-discussedcasesinwhichtheotherpersonexpressesanopinionsharplyopposedtoone’sown)asfollows:
Revision: In cases where one has strong confidencethat P, and another person expresses equally strongconfidencethat~P,oneshouldreviseone’sbeliefinthedirectionof theotherperson’sexpressedbelief totheextentthat
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
Onepossibleapproachissuggestedbythefollowingvariationontheexampleabove:Supposemyfriendagreeswithallofthestepsofmy reasoningup to and including that thepre-rounding amount is$42.87.Nevertheless,sheinsiststhattheansweris$45,not$43.Inthiscase,itdoesnotseemwrongformetousethestepsweagreeaboutinevaluatingherbelief.Infact,thecasethenseemsmuchliketheex-tremely-high-rational-confidencecasesconsideredabove.It’snotthatIhaveextremelyhighrationalconfidenceinP,butIdohaveextremelyhighrationalconfidenceintheconditionalofthelasttwostepsinmyreasoning,whichmy frienddenies.Thus Ihaveexcellent reason tothinkthatsomethingscrewyisgoingon.Asbefore,personalinforma-tionwillstronglyfavorthehypothesisthatthescrewinessaffectsmyfriendratherthanme.Sothecorrectresultwillbeachieved,andinanintuitivelyattractiveway.Thissuggeststhatweunderstandtheextentofthereasoningtobebracketedtobejustthatwhichisindispute(orperhapsbetter,thatwhichIhavegoodreasontobelieveisindispute).
This is, I should emphasize, a very rough pass. It doesn’t distin-guishbetweenstepswhicharedeniedbymyfriendandoneswhichsimplyaren’taffirmed.Italsoleavesopenquestionsabouthowtotreatclaimsaboutwhich Idon’t knowmy friend’s attitude: ifwehaven’tdiscussedthestepsofreasoning,Idon’tknowforsureifshegetsofftheboatatthelaststep,oratsomeearlierstep.Therearesomenaturalapproachesforfillingthesegaps,butfornow,Iwantonlytoacknowl-edgethecomplexitiesthatlieahead.
Again, it is important to note that the difficulty just describeddoesnotnecessarilycountagainstConciliationism.Thatwilldependonwhethernon-Conciliationist views can avoid encountering it, orwhethertheywillalsoneedtoaddressthesameproblem.I’llturntothatquestioninthenextsection.
argumentsabovewasnot toprovidesuchdirections. Ithasbeen toshow that there is room for a Conciliationist approach to disagree-ment; inparticular, to show that relianceonadispute-independentassessmentoftheother’sbelief’sepistemiccredentialsdoesnothavetheuntowardconsequencesitmightseemtohave.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –19– vol.11,no.6(march2011)
Thealternativewouldseemtobethat,insteadof(atleastpartially)bracketingmyoriginalreasoning,wewouldinsteadsimplycombinethereasonsforPthatsupportmyinitialbelieftogetherwiththereasonsforloweringconfidenceinPprovidedbythedisagreementofmyoth-erwisewell-qualifiedfriend.Onthiskindofview,myoriginalreasonsforbelief inParenotputasideatall indeterminingwhat I shouldbelieve.Totheextentthatbelief-changeisneeded,it’sbecausethoseoriginalreasonsarepartlyunderminedoroutweighedbythereasonsfordoubtingPprovidedbythedisagreement.
Itseemstomethatthissortofviewcannotquitebecorrect.Toseewhy,consideragaintheMentalMathcase.SupposemyfriendandIarefullyinagreementthatourtotalbillis$71.20,sotheonlyquestioniswhatoursharesofthisamountare,with20%tipandroundeduptothenearestdollar.AndsupposethatI’veinfactarrivedatthecorrectanswer, $43, by faultless appreciation of the relevantmathematicalreasons.Inthiscase,there’sasenseinwhichthereasonsformyinitialbeliefarecompletelydecisive:given theparametersof theproblem,there’ssimplynopossibilitythat$45couldberight,andnopossibilitythat$43 couldbewrong.Mathematical reasons cannotbeweighedagainstorunderminedbyadditionalevidenceinthewayordinaryin-ductivereasonscan.So insofaras I’mweighingmyoriginalreasonsin“atfullstrength”(i. e.,notputtingthemaside,atleasttosomeextent,inmycognitivedeliberations),andinsofarasIcombinealltheserea-sons togethercorrectly, it turnsout that theevidenceofmy friend’sdisagreementshould innowaydiminishmyconfidence in$43.ButthisisnottheresultthemoderatelySteadfasttheoristisafter.28
28.Onemightquestionwhethermy initial reasons in thiscasearereallydeci-sive.IfIrepresentthosereasonspurelysubjectively—e. g.,as“I,agenerallyreliable calculator, seem to remember getting $43,” or “I now seem to seethat$43isthecorrectanswer,” thenofcoursecountervailingevidencecangainsomepurchase.ButinsofarasoneacknowledgesthatmyinitialbeliefismaderationalinpartbytheactualmathematicalreasonswhichIcorrectlyappreciate,onemustacknowledgethatmybeliefissupporteddecisively.Therelationbetweenthe“data”—theamountofthebillandotherparametersoftheproblem—andthepropositionthatoursharesare$43isnotarelationthatcanbeoutweighedorunderminedbyanything.
8. Beyond Conciliationism
Themajorityofthispaperhasbeenconcernedtoargueforthedefen-sibility ofConciliatory approaches to thedisagreement problem. Indoingso,ithasexposedsomeproblemsthatafullyspelled-outCon-ciliationistaccountofdisagreementwouldhavetoface.Inthissection,Iwanttoarguethattheseproblems—withdefiningthescopeofInde-pendence, andwith incorporating the independentassessment intobelief-revision—will have to be faced by any reasonable view. Theproblems,then,shouldbeseenasindicatingthedifficultyofprovidingageneral accountof theepistemologyofdisagreement, rather thanasposinga specialobstacle toConciliationism.Moreover, I’ll arguethattheissuesraisedherearisenotonlyinthecontextofaccountingfortheepistemicsignificanceofdisagreement,butforawide,butun-derstudied,rangeofcases:thoseinvolvingevidencethatImayhavereasonedimproperly.
Ofcourse,forradicallySteadfastviews,onwhicheventhedisagree-mentofmyfriendintheoriginalMentalMathcasegivesnoreasonatallformodifyingmyconfidencein$43,thereisnoneedforworryingaboutwhatmustbeputasideinassessingmyfriend’scredentials,orwith how to take account of the resulting assessment. But such ex-tremeversionsofSteadfastnessarehighlyimplausible.27
SosupposeweconsideramoderatelySteadfastview:thatdisagree-ment,atleastincaseslikeMentalMath,shouldmakemesignificantlylessconfidentofmyoriginalbelief,butthat inmanyordinarycasesofdisagreement(say,inphilosophy),myoriginalreasoning(perhapsonly if it is in factcorrect) justifiesme in thinking thatmy friend issubstantiallymorelikelytobemistakenthanIam,andsonotmuchrevisionisrequired.Canweobtaintherightresultsonthissortofview,withoutrelyingatallonthesortofbracketingorputtingasideofmyoriginalreasoningmandatedbyIndependence?
27. CompleteSteadfastnesswouldholdthat(atleastincaseswheremyoriginalcalculationshappenedtobecorrect)I’dberationalinignoringevenhordesof friendswho all were proven calculatingwhizzes, and all agreed that Ihaderred.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
Itshouldalsobenotedthatthenecessityofputtingasideone’sorigi-nalreasoningisnotlimitedtocasesinwhichthatreasoningisdeduc-tive. The example ofmy having doubts about arithmetic calculationmakesthepointparticularlyclearly,takingadvantageoftheconclusivenatureofdeductivereasoning.Butoncetheexampleisunderstood,itslessonsclearlyapplymoregenerally.Forwhatliesbehindourverdictinthecalculationcaseisourseeingthatfailuretoputasidethereasoningunderconsiderationamountstobeggingthequestionofwhetherthatreasoningwascorrect.Andthatissomethingwemustnotdo,ifwearetotakeseriouslythepossibilityofourowncognitiveerror.30
One might worry that something in this argument must bewrong.Afterall,evenbeforelearningofmyfriend’sdisagreementinMentalMath,Ishouldn’thavebeenabsolutelyconfidentinmyanswer.How,then,couldthosereasonsbeasdecisiveastheargu-mentsuggests?
TheanswertothisworryisthatIdidhavereasontoputmymath-ematicalreasoningpartiallyaside,evenbeforelearningofmyfriend’sdisagreement. After all, amajor reason for being initially less-than-absolutelyconfidentin$43isthatIknowthatIandotherpeoplefairlycommonlymakemistakesindoingmentalmath.Butagain,ifIweresimplytocombinethereasonsforcautionprovidedbythisfactwiththefull-strengthmathematicalreasons,thefactabouthumanfallibil-itywouldhavenoeffect.Nothingabouthumanfallibilityaffectstheconnectionbetweenthetotalbill’sbeing$71.20andmyshare(with20% tip, roundedup)being$43. If I didn’t at leastpartiallyput themathematical reasonsaside, thennoamountofadditionalevidencewouldhaveanyeffect.29
Thismakes it clear that thegeneral point about thenecessityof(at leastpartially)bracketingsomeofourreasoningapplies inwideareas outside the context of disagreement. There are countless pos-sible reasons forme tobeunsureofwhether I’vemadeacognitiveerror,andthusunsureofconclusionsI’vereached.TheyrangefromknowledgeofpasterrorsthatIorothershavemade,tounexplained
29.The contrast between reasoning in one’s head and careful calculation hasbeencitedbyanti-Conciliationistsassupportfortheirview.Aswe’veseen,for example, the effect of disagreement on mental-math-based beliefs isgreater than the effect of disagreement on beliefs based on careful papercalculations.Theanti-Conciliationist’ssuggestionisthatthedifferenceistobeexplained in termsof thedifferences in thestrengthof theoriginal rea-soning,andthusthattheoriginalreasoningcan’treallyhavebeenputaside,contra Independence. But the considerations reviewed above suggest thattheexamples reallypull inexactly theoppositedirection. For the stepsofreasoningonedoesinone’sheadmaybeexactlythesameasthestepsonedoesonpaper.Insofarasthosestepscountasreasonsforone’sconclusion,theyareequally(andmaximally)strong.Whatdiffersbetweenthecasesisthesecond-orderreasonsfordoubtingthefirst-orderreasoning.Andthosesecond-orderreasonscanonlyhavepurchasetotheextentthatoneputsthefirst-orderreasonsaside.
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –21– vol.11,no.6(march2011)
self-criticism.Ontheonehand,onemustevaluateone’sbeliefs inawaythatdoesnotsimplybegthequestionof theircogency.Ontheotherhand,onemustdoso fromwithin the frameworkofone’sbe-liefs—foronehasnootherwaytothinkaboutanything.31Thereisanobvious tensionbetween these requirements,and it’s far fromclearthattherewillbeanice,cleanrecipeforresolvingthattensioninanyarena.But resolving the tensionsomehow isan inescapablepartofreactingrationallytothepossibilityofourownrationalfailures,some-thingwe’rerequiredtodoallthetime.Understandinghowthetensionisproperlyresolved,then,isessentialtounderstandinghowtocoperationallywithacentralaspectofourepistemicpredicament.32
davidchristensen Disagreement, Question-Begging, and Epistemic Self-Criticism
philosophers’imprint –22– vol.11,no.6(march2011)
Lackey, J. (forthcoming), “A Justificationist View of Disagreement’sEpistemicSignificance”,inA.Haddock,A.Millar,andD.Pritchard(eds.),Social Epistemology(Oxford:OxfordUniversityPress).
——(2010),“WhatShouldWeDowhenWeDisagree?”,inT.S.GendlerandJ.Hawthorne(eds.),Oxford Studies in EpistemologyIII(Oxford:OxfordUniversityPress),274–293.
Sosa, E. (forthcoming), “The Epistemology ofDisagreement”, in hisArmchair Philosophy(Princeton:PrincetonUniversityPress,2010).
vanInwagen,P.(1996),“ItisWrong,Everywhere,Always,andforAny-one,toBelieveAnythinguponInsufficientEvidence”,inEleonoreStumpandMichaelJ.Murray(eds.),Philosophy of Religion: the Big Questions,editedby.NewYork:Blackwell(1999).
White,R.(2009),“OnTreatingOneselfandOthersasThermometers”,Episteme: A Journal of Social Epistemology6:233–250.
References
Christensen, D. (2007), “Epistemology of Disagreement: The GoodNews”,Philosophical Review116:187–217.
——(2010), “Higher-OrderEvidence”,Philosophy and Phenomenological Research81.1:185–215.
——(2007),“ReasonableReligiousDisagreements”,inL.Antony(ed.),Philosophers without Gods: Meditations on Atheism and the Secular Life (NewYork:OxfordUniversityPress).