Feldman - The Ethics of Belief (2000)

8/22/2019 Feldman - The Ethics of Belief (2000)

1/29

Philosophy and Phenomenological ResearchVol. LX,No. 3, May2000

TheEthicsof BeliefRICHARDFELDMANUniversity of Rochester

In this paperI will addressa few of the many questions that fall under thegeneral heading of "the ethics of belief." In section I I will discuss theadequacyof what has come to be knownas the "deontologicalconception ofepistemicjustification" n the light of our apparentack of voluntarycontrolover what we believe. In section II I'll defend an evidentialist view aboutwhat we ought to believe. And in section III I will briefly discuss apparentconflictsbetweenepistemicconsiderations nd moral or otherconsiderations.

I. Epistemic Deontologism'A. The ProblemOur talk about epistemic mattersparallels our talk about ethical matters innoteworthyways. In the case of ethics, we say that a person oughtto performa certain action, that someone should not do a certainthing,thatpeople haveobligations to act in some ways, thatthey have rights andduties, and thatthey deservepraise or blame for what they have done. We,make seeminglyanalogous epistemic judgments about beliefs. We say that a typical well-informed contemporaryAmerican ought to believe that the Earth revolvesaroundthe Sun and should not believe that the Earthis flat. We say that aperson has a right to believe one thing and perhapsa duty to refrain frombelieving somethingelse. We sometimespraisethose who believe the thingsthey should and we criticize those who fail in their believings. We candescribe all these judgments as deontological judgments about beliefs.

In "The Deontological Conception of Epistemic Justification,"WilliamAlston says thatthe most naturalway to understand pistemic justificationisdeontological.2By this he means thatit is to be understood n terms of epis-

1 This section is a shortenedversion of my paper,"Doxastic Voluntarismand the Ethics ofBelief," forthcoming in a collection of papers on epistemic responsibility edited byMatthiasSteup.

2 William Alston, "The Deontological Conceptionof Epistemic Justification,"PhilosophicalPerspectives 2 (1988): 257-99. See especially p. 258. Reprinted in William Alston,Epistemic Justification(New York: CornellUniversity Press, 1989).

SYMPOSIUM 667


2/29

temic "obligation,permission,requirement,blame,andthe like."3He regards"requirement, rohibition, and permission as the basic deontological terms"and the other terms as derivativeones.4Alston eventually arguesthatdeonto-logical conceptionsarein the end ill-suited to epistemic purposes.5Recently,Alvin Plantinga and Alvin Goldman have independently argued that theviability of a deontological conceptionof epistemic justification is crucial tothe debate between internalistsandexternalistsaboutepistemic justification.6Goldman thinks that a central,but mistaken, line of supportfor internalisttheories begins with the assumption of a deontological account ofjustification. Plantinga also argues that internalism derives much of itssupport roma deontologicalview of justification.7The meritsof the deonto-logical view of epistemicjustificationthus is of considerableepistemologicalsignificanceto contemporary pistemology.A centralproblemthat bothPlantingaand Alston find with deontologicaljudgments about beliefs is that they presuppose that we have voluntarycontrol over what we believe. Yet, reflection on our mental lives suggeststhat we have no such control. Alston writes of one particulardeontologicalanalysis:...this conceptionof epistemic justificationis viable only if beliefs aresufficiently under volun-tary control to rendersuch concepts as requirement, permission, obligation, reproach, andblame applicableto them. By the time honoredprinciplethat"Ought mplies can",one can beobliged to do A only if one has an effective choice as to whetherto do A.8

He goes on to arguethatwe don't have an effective choice over what webelieve. In the process of objectingto Chisholm's views aboutjustification,Plantinga says of a particular proposition that "...whether or not I accept it issimply not up to me; butthen acceptingthis propositioncannot be a way inwhich I can fulfill my obligationto the truth,or, indeed, any obligation..."'Thus, according to Plantinga,our lack of control over beliefs implies that3 "The Deontological Conceptionof EpistemicJustification", . 257.4 Ibid.5 The most prominentadvocate of this sort of analysis is RoderickChisholm. He defends

this view as recentlyas in his in 1991 paper,"Firthand the Ethics of Belief," Philosophyand Phenornenological Research 50 (1991): 119-28. For a thorough list of Chisholm'spublications on this topic, see Susan Haack, "'The Ethics of Belief' Reconsidered,"inLewis E. Hahn, (ed.) ThePhilosophy of RoderickChisholin(LaSalle: Open Court, 1997),pp. 129-44. See especially footnote 2. See also Hilary Kornblith,"JustifiedBelief andEpistemically Responsible Action," Philosophical Review 92 (1983): 33-48 for aninfluential discussion of the idea that a belief is epistemically justified provided it isformed in an epistemicallyresponsible way.

6 See Alvin Plantinga,Warrant:The CurrentDebate (New York: OxfordUniversity Press,1993) and Alvin Goldman "InternalismExposed,"the Journal of Philosophy 96 (1999):271-93.Warrant:The CurrentDebate, Chapter1.

8 "The Deontological Conceptionof EpistemicJustification," . 259.Warrant:The CurrentDebate, p. 38.

668 RICHARD FELDMAN


3/29

they are not the sort of thing that can be a matterof obligation, and thisundermines Chisholm's deontological conception of epistemic justification.MatthiasSteup presents a similar argument, houghhe goes on to defend thedeontological conception on the grounds hatbelief is voluntary.10

For the purposes of the discussion that follows it will be helpful todistinguish two elements of the arguments ust presented.Their target is adeontological conception of justification, a conception according to whichepistemic justification s to be understoodor analyzed n terms of the deonto-logical concepts of obligation, requirement,and the like. This conception is"viable", n Alston's terms, only if belief is sufficiently under our voluntarycontrol. Presumably,the deontological conceptionis viable only if it can betrue that we are requiredto believe things, that we ought not believe otherthings, and so on. That is, the deontological conception of epistemicjustification s viable only if deontological udgmentsabout beliefs aresome-times true.My goal in what follows is to arguethatdeontologicaljudgmentsare sometimes trueandthatgood sense can be made of the idea that we canhave epistemic requirements,obligations,and the like. I will not be arguingfor the furtherclaim thatepistemicjustificationshouldbe analyzed n deonto-logical terms.

We can formulate the key partof the argumentat issue in the followingway:

The VoluntarismArgument1. If deontological judgments aboutbeliefs are true, then people have

voluntarycontrolover their beliefs.2. People do nothave voluntarycontrol over their beliefs.3. Deontologicaljudgmentsaboutbeliefs arenot true.

Epistemologistshave threekindsof responseto this argumentopen to them:i) they can arguethat we do havetherequisitesort of control overourbeliefs,therebyrejecting premise (2); ii) they can arguethatdeontologicaljudgmentsdo not have voluntarist implications, thereby rejecting premise (1); or iii)they can acceptthe argumentand admitthat the familiardeontologicaltermsof epistemic appraisalreally are inapplicable."This in itself is a surprisingconclusion, whetheror not a deontological analysis of justificationis accept-able.10 Matthias Steup, "The Concept of Epistemic Justification,"Chapter4 of Contemporary

Epistemology (Upper Saddle River: Prentice-Hall, 1996).1 This last alternative implies that either epistemic justification is not to be analyzed indeontological terms or else that epistemic justification is itself inapplicableto our beliefs,just like the otherdeontological epistemicterms.

SYMPOSIUM 669


4/29

I will discuss response (i) in section B below andresponse (ii) in sectionC. I will not discuss response (iii) except in passing andby implication.B. Doxastic VoluntarismIn this section I will examine some arguments or doxastic voluntarismandsome argumentsagainst it. I will eventually defend a version of doxasticvoluntarismand I will also argue that this fact is of absolutely no epistemo-logical significance and thatit does nothingto help resolve the voluntarismpuzzle. Although some philosophers have argued that it is a conceptualimpossibility that anyone form a belief voluntarily, I will not discuss thatview."2 nstead,I will examine the claim that it is a contingentmatterof factthatwe lack thatability.B]. TheArgument or InvoluntarismAlston has given the most thorough defense of the contingent inabilitythesis, the thesis that as a contingent matter of fact people are not able toacquirebeliefs voluntarily.Alston's paperincludes an excellent survey of avarietyof notionsof voluntarycontrol.For each type, he arguesthat we lackthat sort of control over beliefs. (Alston does admit that there is one veryweak notion of control that does apply to belief. But he contends that thissort of control does not providean adequatebasis for a good response to theVoluntarismArgument.)

Alston begins by discussing basic voluntary control."3We have basicvoluntarycontrol over those actions that we can "justdo." Simple bodilymotions are the primeexamples.I canjust raisemy hand,close my eyes, andbend my knee. Some people, but not I, can wriggle their ears and curl theirtongue. Alston correctly says thatforminga belief is not like that.We can'tjust do it at will.

Alston turnsnext to nonbasic immediatevoluntarycontrol.14 One has thissort of control over the things one can do right away by doing somethingelse, typically somethingover which one has basic voluntarycontrol. Stan-dard examples are opening doors and turningon lights. We can, in typicalcircumstances,do these things simply by moving our bodies in the appropri-ate ways. There'svaguenesshereconcerningwhat counts as "rightaway"but12 For defenses of the conceptual impossibility claim, see BernardWilliams, "Deciding toBelieve," Problems of the Self (New York: CambridgeUniversity Press, 1973), JonathanBennett, "Why Is Belief Involuntary?,"Analysis 50 (1990) and Dion Scott-Kakures,"OnBelief and the Captivity of the Will," Philosophy and Phenomenological Research 53(1993). For discussion of Williams, see BarbaraWinters, "Believing at Will," Journal of

Philosophy 76 (1979). For a discussion of Scott-Kakures' article, see Dana Radcliffe,"Scott-Kakures n Believing at Will,"Philosophy and Phenomenological Research 57, 1(1997).

13 "TheDeontologicalConceptionof EpistemicJustification,"Section III, pp. 263-68.14 See Section IV, pp. 268-74.670 RICHARDFELDMAN


5/29

that vagueness is in no way problematic. This is because the boundarybetween nonbasic immediatevoluntarycontroland the next weakerkind ofcontrol, long range voluntarycontrol, is acceptablyimprecise.'5The sortsofthings over which we have long range voluntary control are the sorts ofthings we can do over time by doing other things. Perhaps painting myhouse is an example.Or,moreprecisely,I have long range voluntarycontrolover what color my house is because I can do things like paint it. Finally,there is indirect voluntary influence.16This is the kind of control we havewhenwe canundertake courseof actionthatmay affect some conditionoverthe long term. Perhapsa person has indirect voluntaryinfluence over theconditionof herheart,since diet andexercise (coursesof actionwe can moredirectlycontrol)canaffectit.

Alston claims that we have only this weakest sort of control over ourbeliefs. Onecan engage in a courseof actionfor thepurposeof inculcatingorlosing a particularbelief. Consider my belief that the Earth is not flat.There's nothingI can do to get myself to stop believing that rightaway. It'snot like shutting the door or turning out the lights. And there's nothingmuch I can do long rangeto controlit either.I might enroll in the Flat EarthSociety, read conspiracy literature asserting that satellite photos are allphony, and so on. Perhapsthis will help rid me of my belief. Alston wouldagree that it might. But this gets us, at most, indirect voluntaryinfluence,and this is not the sort of effective voluntarycontrolrequiredto refuteTheVoluntarismArgument."7

I believe that, for most part,what Alston says is right. However, I willarguethat we have considerablymore controlover some of ourbeliefs thanAlston acknowledges. Still, this control does not underminethe basic ideabehindThe VoluntarismArgument,although t may show thatthe argumentneedsreformulation.B2. AnArgumentor Doxastic VoluntarismMy argument or doxasticvoluntarismbegins with the assumptionthatthereare states of the world over which people have nonbasic voluntarycontrol.For example, I have nonbasic voluntarycontrol over whether the lights inmy office are on. All I have to do is move in a certainway to get the lightson or off. And I can do this. The next step of the argumentnotes that mybelief about whether the lights are on tracks their actual state almostperfectly. As a result, I have a similar amount of control over whether Ibelieve thatthe lights areon. All I have to do is move a certainway andthen

15 See Section V, pp. 274-77.16 See Section VI, pp. 277-83.17 Perhaps this point suggests that both (2) and the consequent of (1) should be abouteffective voluntarycontrol.

SYMPOSIUM 671


6/29

I'll have the relevant belief. Moregenerally,when I have control over a stateof the world andmy beliefs about that state track thatstate, then I have justas much control over my belief aboutthe state as I have over the state itself.Thus, we have nonbasic immediate voluntarycontrol over our beliefs aboutstates of the world over which we have control, provided our beliefs areresponsive to those states. Furthermore,f we know that we will respond insome mistakenway to some state of the world over which we have control,we also have controlover theresulting(erroneous)belief. So, doxastic volun-tarism s true after all. Premise (2) of The VoluntarismArgument s false.

I believe that the existence of nonbasicvoluntarycontrol over beliefs canhave prudential nd moral significance.If thedepartment hair announces hatshe'll give a raise to all andonly those membersof the departmentwho in 30seconds believe that the lights in their office are on, I'll head for the lightswitch and turn on the lights to make sure I have the desired belief. If thechair perversely announces that the graduatestudents will be mercilesslytortured-say, by being forced to take additional prelims-unless in 30seconds I believe thatmy lights areon, thenI'd better make sure that I havethat belief. I'm in control.

Thus, we do have control over many of our beliefs. Of course, we don'toften exercise this control.Thatis, we don't often believe things voluntarily.And this leaves this defenseof doxasticvoluntarismwithouta greatdeal epis-temological significance.The central reason for this is that we make deonto-logical epistemic judgments about beliefs that we can't control and thesejudgments are as routineand commonplaceas arejudgments about beliefsthatwe can control.Thus, our ability to controlwhat we believe in the waydescribedhere is epistemically nsignificant.We can takethis into accountbyreformulatingTheVoluntarismArgument:

The Voluntarism Argument (Revised)1. If deontological judgments about beliefs concerning states of the

world people can't control are true, then people have (effective)voluntarycontroloverthose beliefs.

2. Peopledo not have (effective) voluntarycontrol over those beliefs.3. Deontologicaljudgmentsabout those beliefs concerningarenot true.

This argumentis as troublingas the original. It implies that an enormousnumber of the deontological epistemic judgmentswe routinelymake cannotbe true.Furthermore,he argumentmightbe extended to beliefs that we cancontrol but have not formed in this voluntary way. Thus, I don't think that

672 RICHARDFELDMAN


7/29

the fact thatwe have nonbasic mmediatevoluntarycontrol over some beliefsprovidesthe basis for an effective defense of epistemic deontologism.18B3. A More Robust From of Doxastic VoluntarismCompatibilists n the free will debate contendthat we voluntarilyperformanaction when the action has the right sort of cause. A defender of doxasticvoluntarism can argue that, analogously, we believe voluntarily when webelieve as a result of the right sort of causal process. Roughly, the idea isthat when unconstrained deliberation about evidence leads to belief, webelieve voluntarily.1 This is an interesting ine of thought,and it is difficultto refute because we lack a fully adequate understandingof what counts as"the rightsort of causal process." Still, I thinkthat therearegood groundstorejectthe idea thatthe deliberativeprocessthattypically leads to belief is thekindof processthatmakes its producta voluntaryaction.

A key fact about the clearest cases of voluntaryactionis that the action iscaused by an intention to performthat action. I turnon the lights because Iintendto turnon the lights;I type a certainword becauseI intend to type thatword. No doubt there are puzzles aboutthe details of these matters,such aspuzzles resulting from the fact that our intentionsare not as precise as theresultingactions-I need not have intendedto turnon the light in exactly theway I did.But however we deal with thesedetails,the case of belief is dramat-ically different.We simply don't, in the typical case, forma belief as a resultof an intention to form that belief. It may be that in some sense we docontrol what we believe, in much the way we control other involuntaryprocesses in us. However, epistemic deliberationdoes not result in effectiveintentions to believe. Except in rare cases, we don't form intentions tobelieve. But such intentions are essential to voluntarycontrol.20The cases ofvoluntarybelief formationthatI describedearliermake it easy to see what itreally would be like to believe voluntarily.In those cases, I do believe as aresult of an intention to believe. But those are, for this very reason, quiteunlikeordinary ases.

Furthermore,f this defense of doxastic voluntarism were sound, then itseems that many otherplainly involuntarybehaviorswould also be volun-tary. Deliberating about something can result in states other than belief. It18 If this revised argument s strong, then it would also be reasonableto reject deontological

analysesof epistemicjustificationas well.19 For a defense of this view, see Matthias Steup, "Doxastic Voluntarism and EpistemicDeontology," forthcoming. For a similar view, see James Montmarquet, "TheVoluntariness of Belief," Analysis 46 (1986). Bruce Russell endorses Steup's view in"Epistemicand Moral Duty," forthcoming.

20 Perhapsthe process in which by which we formulate ntentions and act on their basis canbecome automatedin certain ways, so that voluntaryactions need not involve explicitconscious formulations of intentions.Belief formation is not like that either. It does nottypically involve the formationof intentionsto believe at all.

SYMPOSIUM673


8/29

can result in desires or in panic. If we say thatbelief is voluntarybecause itis the outcome of deliberationaboutevidence, then it is hard to see why weshouldn't say that these other outcomes of deliberation are also voluntary.Yet it is clear that they are not. Again, what seems clearly missing is theright sort of intention.Thus, I reject this more robust form of doxastic voluntarism. There isgood reason to say that the deterministic processes that typically lead toaction render those actions voluntary.The processes lead to effective inten-tions to act. Deliberatingaboutevidence does lead to belief, but not via suchintentions. Such deliberation does not make belief voluntary. We need abetterresponseto The VoluntarismArgument.C. Oughtand CanThe second main sort of responseto The VoluntarismArgumentdenies thatthe deontologicaljudgmentsabout beliefs imply that those beliefs arevolun-tarily adopted.It denies premise (1) of the argument. t may be thattherearedifferences among the various assertions thatI've describedas "deontolog-ical." I'll focus firstonjudgmentsaboutwhat one is obligatedto believe andwhat one ought to believe and then discuss other judgments later in thesection.One way to defend deontological epistemic judgments in the light ofdoxastic involuntarismis to argue that we can have epistemic obligationseven thoughwe can't fulfill them (or even if we can't help but fulfill them).This is to deny that"epistemicallyobligatedto" or "epistemically ought to"implies "can".And one way to make this denialplausibleto show thatthereare other kinds of oughtstatements hatdon't imply voluntarycontrol.I willconsiderseveral candidates or this otherkind of ought.C1. ContractualObligationsYou can have an obligationto pay your mortgageeven if you don't have themoney to do so. Perhapsstudents in a class have an obligation to do thecourse work even if they are incapableof doing it. Otherexamples of thissort are rathereasy to construct.Perhapsepistemic obligationsareanalogousto these financial and academic obligations. Though I once defended thisview,21 t now strikesme as an implausiblemodel for epistemic obligations.The obligation to pay one's mortgageand the obligationto do one's coursework are contractualobligations, although n the lattercase the contract s insome sense implicit. It's difficult to see any basis for saying that we all havesome sort of contractualobligationto believe things. Surelyno such contractis explicit, and nothing analogous to enrolling in a course establishes animplicit contract.21 "EpistemicObligations,"Philosophical Perspectives2 (1988), pp. 240-43.674 RICHARDFELDMAN


9/29

C2.ParadigmObligationsIn a recentpaper,Nicholas Wolterstorff ays thatthere are two kinds of obli-gations,paradigm obligationsand responsibility obligations.22Only obliga-tions of the latter sort are associated with voluntariness. As examples ofparadigmobligations,he presents:

1. "Youought to be walking on it in two weeks"-said by a physicianas he finishes bindingup a person'ssprainedankle.

2. "That'sstrange;you ought to be seeing double"- said by a psychol-ogist to his subject while conductingan experiment n perception.

Wolterstorffsuggests thatepistemicobligationsare similar to the obligationsdescribedby these sentences.They lack anyimplicationof voluntarycontrol.No friend of epistemicdeontologismshould be comfortedby the idea that

epistemic obligations are like the obligations describedby (1)-(2). This isbecause there are no obligationsdescribedby (1)-(2). Sentences (1)-(2) are"ought" entences; they arenot obligationsentencesandthey cannot be para-phrased n any straightforwardway into obligation sentences.Your ankle hasno obligationof any sort to heal; you have no "perceptualobligation"to seedouble. So, if there are epistemic obligations, they are not like the obliga-tionsdescribedhere,since thereareno obligationsdescribedhere.

Furthermore,he ought sentencesWolterstorffdescribesarenot relevantlylike epistemicoughts. (1) and(2) describenormal,or paradigmatic,behavior.Thus,forexample, baringunforeseendevelopments,youranklewill heal in 2weeks. But I don't thinkthatthis carries over to the epistemic case. That is,epistemic oughts don't describe paradigmaticor normal function. Someresearchersreportthat people typically make various unjustified nferencesandpredictably ormunreasonable rerroneousbeliefs.23 Whatever heproperinterpretationof this researchactually is, it is at least possible that peoplenormallymake epistemic errors. It may be that we epistemically ought notdo whatwe normallydo. Epistemicoughtsare not paradigmoughts.24

22 Nicholas Wolterstorff, "Obligations of Belief: Two Concepts," in The Philosophy ofRoderickChisholm,Hahn, Lewis E (LaSalle,Illinois: OpenCourt, 1997).23 For a discussion of many of the examples allegedly showing thatpeople are irrationaland

an examination of their philosophical implications, see Edward Stein, Without GoodReason: TheRationalityDebate in Philosophyand CognitiveScience (Oxford:ClarendonPress, 1996).24 Thus, a researchermight say, while awaiting the subject's reply to a questionthatnearlyeveryone misses, "He ought to make the wrong inference here." The use of "ought"here is Wolterstorff's paradigmought. But to describe the inference as "wrong" s to saythat it is not the case that the person epistemically ought to do what the researcher

expects him to do.SYMPOSIUM 675


10/29

C3. Role OughtsThere are oughts that result from one's playing a certain role or having acertainposition. Teachersought to explain things clearly. Parentsought totake care of their kids. Cyclists oughtto move in variousways. Incompetentteachers, incapable parents,anduntrainedcyclists may be unable to do whatthey ought to do. Similarly,I'd say, formingbeliefs is something people do.Thatis, we form beliefs in responseto ourexperiencesin the world.Anyoneengaged in this activity ought to do it right.In my view, what they ought todo is to follow their evidence(rather han wishes orfears).I suggestthatepis-temic oughts are of this sort-they describe the right way to play a certainrole. Unlike Wolterstorff'sparadigm oughts, these oughts are not based onwhat's normal or expected. They are based on what's good performance.Furthermore,t is plausible to say that the role of a believer is not one thatwe have any real choice abouttakingon. It is ourplight to be believers. Weought to do it right.It doesn't matterthat in some cases we are unable to doso. Thus, I reject the first premise of the revised VoluntarismArgument.Even in cases in which a believer has no control at all, it makes sense tospeakof what he oughtto believe andoughtnot believe.

I think that most, possibly all, of the deontological terms we ordinarilyuse to evaluate beliefs can be explained in similar terms, althoughthe casefor some may be weaker thanthe case for others.Thus, if the standardsofgood believing allow eitherof two attitudes n a given situation,then it willbe true that the believer has a rightto either of those attitudesandthateitherof them is permitted.25Other attitudesmay be prohibited. It seems to mereasonable to say that when only one attitude is permitted,then one has anepistemic obligation to have that attitude.(I will add a qualificationto thisclaim later in the paper.)It may be that some terms, especially those associ-ated withpraise andblame, are to be reservedfor voluntarybehavior. Evenhere,thecase is less thanperfectly clear,since we do praiseand blamepeoplefor attributes, such as beauty, that they are unable to control. Thus, Iconclude thatdeontological epistemicjudgmentscan be true even if doxasticvoluntarism s false.C4. Is ThisDeontology?I want to contrast the view I'm defendingfrom a different view aboutepis-temic deontologism. Possibly some criticsobject primarily o the idea that tobe epistemically justified in a belief is to do the best that one can do withrespect to that belief. Their target really does imply voluntarism.I haven'tattempted o defend thatview, andI don't think thatit is true. I've takenonthemuchmoremodesttaskof defendingthe legitimacyof the widespreaduseof deontological language aboutbelief. My contention is that we can have25 I'll return o some related issues in section II, partA.676 RICHARDELDMAN


11/29

epistemic requirements,permissions, and the like even if voluntarism isfalse.

In his influential paper on this topic, Alston begins by saying that it is"natural" o understandepistemological terms in a "'deontological' way, ashaving to do with obligation, permission, requirement, blame, and thelike."26As an example of the view he wants to question, he mentions CarlGinet's view that explains justification in terms of whether one is asconfident as one ought to be.27But by the end of his paperAlston seems tobe directinghis attackmorenarrowly. n his concluding paragraph e charac-terizes epistemic deontologism as the view that analyzes "epistemicjustification in terms of freedom from blame for believing."28I haven'tattemptedto defend this sort of deontologism here. Moreover,I don't thinkthat this more narrowly defined sort of deontologism is so natural orcommon. My defense is only of the possible truthof the deontological epis-temic claims with which Alston began his paper, claims to the effect thatpeople can have epistemic rights, duties, permissions, etc. They can, nomatterwhat the truthabout doxasticvoluntarisms.29

Given thatdeontologicalepistemic judgments can be true,I turnnext to aview abouttheconditions underwhichthey aretrue.

II. Epistemic OughtsA. EvidentialismIn this section I will defend a variationon William K. Clifford's frequentlyquotedclaim that "It s wrong,always,everywhere,and for anyone to believeanything upon insufficientevidence."311 t is not my purposehere to explainor defend the thesis Clifford assertedwith these words.My view may differfrom Clifford's in any or all of the following ways. First,he may have beenmakinga moralevaluationof believing on insufficientevidence.31 The claim26 "TheDeontological Conceptionof EpistemicJustification," . 257.27 Ibid., p. 259.28 Ibid., p. 294.29 It is possible that some criticswill contend that sentenceslike "Smithought to believe that

the earth is not flat" are ambiguous between deontologicaland non-deontological senses.They might then object that I've merely argued for the possible truth of the non-deontological interpretation f the sentence. I don't see any reasonto admit that there aremultiple senses of this sort. Thedeontologicalsense of the sentence just mentioned seemsto be nothing other than the conjunction of the non-deontological sense with theproposition that it is up to Smith whether Smith does believe the earth is not flat.Obviously, so interpreted, t does imply voluntarism. But the question with which webegan was whether the original claim was compatible with voluntarism.I don't thinkthatthis compatibilitycan be ruled outby fiat.

30 W. K. Clifford, "The Ethics of Belief," reprinted in The Theory of Knowledge 2ndEdition,editedby Louis P. Pojman, Belmont,California:Wadsworth,1999), pp. 551-54.31 In "The Ethics of Belief Reconsidered" Susan Haack argues that Clifford did notdistinguishepistemicand moralsenses of the key terms.

SYMPOSIUM677


12/29

I want to make concerns epistemic ratherthan moral evaluation. Second,Clifford evaluates believing on insufficientevidence as wrong whereastheclaim I will defend is that one ought not believe on insufficient evidence.Whetherthereis a differencebetween a thing being wrongandbeing some-thingone oughtnot do is unclearto me. If thereis a difference,it is the latterclaim thatI wantto defend.Third, n thequotedpassageClifford only objectsto believing on insufficient evidence. He does not say that one ought tobelieve when one does have sufficientevidence.However,I do wantto makethis additionalclaim. A succinct way of statingmy thesis is thatone alwaysought to follow one's evidence. This evidentialist thesis is what I willdefend.32

The following principleprovidesa preliminary tatementof the evidential-ist idea:

01. For any propositionp, time t, and person S, S epistemically oughtto have at t the attitude owardp thatis supportedby S's evidence att.

The attitude n questioncan be belief, disbelief, or suspensionof judgment.33It's importantto realize that (01) does not imply that a person ought tobelieve a propositionsimply because some of the person'sevidence supportsthatproposition.Belief is the attitude o have only if the evidence on balancesupportsp.

Some philosopherswho are generally sympatheticto evidentialism maythinkthatmerely having one's evidence on balance supporta propositionisinsufficient for it to be true that one epistemically ought to believe theproposition.That evidential supportmight be exceedingly weak, and somewill think that strongerevidence is required.Those who insist on such aconditioncan interpret 01) in a way thatfits with this idea. On this' nterpre-tation, believing p is supported by one's evidence only if the evidencesupportsp sufficientlywell.

There is, however, reason to doubt (01). Suppose that a person hasevidence that conclusively establishessome proposition,q. Thereare then ahuge number,perhapsan infinitenumber,of obvious logical consequencesofq thatarealso supportedby thisevidence.Forexample,theevidence supportsq disjoinedwith every otherproposition,or at'leastevery otherpropositionSunderstands. 01) seems to imply that S ought to believe all these disjunc-tions. Many of these disjunctionswill be trivial;many will be of no practical32 In "Evidentialism,"Philosophical Studies48 (1985): 15-34, Earl Conee and I defend the

evidentialist thesis that believing a proposition is justified if and only if believing theproposition "fits"one's evidence.33 Those who think that beliefs come in degrees may wish to replace this principleby onesaying thatone oughtto believe propositions o the degreethatthey aresupportedby theirevidence.

678 RICHARDFELDMAN


13/29

or theoretical interest. It may be that in some sense believing these proposi-tions is justified-the person has good supportfor them-but it is not truethat he ought to believe them. To do so would constitute an exceedinglyfoolish use of S's cognitive resources.34

Whetherthis objection is a good one dependsin part upon what exactlygoes into believing a proposition. The more dispositional belief is, the lessthat (01) demands of us. If the person with conclusive evidence for q cansatisfy its demandsby being disposed to assent to each of the disjunctions ifasked about them, then (01) doesn't reallyrequire hat he expend any cogni-tive resources.He doesn'thave to do anything o meet its standards. f, on theother hand, it takes more than that to believe a proposition, then perhaps(01) is excessive. I'm inclined to thinkthat (01) does ask too much of us,but I won't arguefor that here. Instead,I propose the following refinementthatavoids the problem:

02. For any person S, time t, and propositionp, if S has any doxasticattitude at all towardp at t and S's evidence at t supportsp, then Sepistemically ought to have the attitude towardp supportedby S'sevidence att.

(02) in effect conjoins three principles: if a person is going to adopt anyattitude toward a proposition, then that person ought to believe it if hiscurrentevidence supports t, disbelieve it if his currentevidence is against it,and suspendjudgment aboutit if his evidence is neutral or close to neutral).A person might never even considera propositionand not be aboutto adoptany attitude toward it. In that case, (02) does not imply that there is anyattitudethe person ought to have toward the proposition. Thus, (02) avoidsthe problem concerningthe use of cognitive resourcesthat(01) faced. (02)merely stateswhich attitude o take towardpropositionsaboutwhich you aregoing to have one attitude or another.It does not instruct you to believelogical consequences that you don't entertain or to otherwise squanderpreciousresources.B. Requirements nd Permissions(02) helps clearup a confusingissue about whetherepistemic principlesstatewhatyou arerequired ought)to believe or whatyou arepermitted o believe.Suppose some unimportant nd unconsideredproposition obviously followsfromyour evidence. For the reasons ust stated,this fact conjoinedwith (02)doesn't imply that you ought to believe it. But (02) does imply that if youtake any attitudetoward the proposition, it ought to be belief. So, it's not

3 Alvin Goldman makes this point in "Epistemics: The Regulative Theory of Cognition,"The Journal of Philosophy 75 (1978): 509-23.

SYMPOSIUM679


14/29

just that believing it is permitted.That's the only permissible attitude. It'sthe one you ought to have if you have any attitude. Thus, no attitude isepistemicallyrequired,butonly one is epistemicallypermitted.

(02) thus rules out the possibility that each of two attitudes toward aproposition is permitted at a time. Some critics might regard this as amistake. To see whether t is, let's begin by considering a moral analogue ofthis possibility. On standard iews, there are cases of ties in morality:each oftwo actions is morally permissibleand neither is obligatory.It's permissiblefor me to give my extramoney to the CancerSociety and it's permissible togive it to the Heart Association. It's not permissibleto spend it frivolously.It follows then, thatit is permissible to give my money to the Heart Associ-ation and also permissible to refrain from giving my money to the HeartAssociation. Each of two incompatibleactions is thus permissible. This is aconsequence of the existence of moral ties. If the epistemic case is similar,then there are situations in which each of two attitudestoward one proposi-tion is permissible. But I think that in the epistemic world, things aren'tquitethe same. There are no ties, in the relevantsense.

Potential cases of epistemicties areof three sorts.The first,andeasiest todispose of, are cases in which the evidence for and against a propositionisequally weighty. In thatcase, evidentialismimplies thatthe sole acceptableattitude is suspending judgment. Neither belief nor disbelief is permitted.That strikes me as the rightresult.

The second kindof case consists in those in which two competing propo-sitions are equally well supported by the evidence. For example, there isequally good evidencesupportingheguiltof each of two suspectsfor a crimethatcould have been committedby only one person.Casual reflectionmightsuggest thatyou could believe thateither is guilty, but obviously, not both.In thatcase, it is permissibleto believe that suspect A is guilty and suspectB is not, and it is permissible to believe that suspect B is guilty and A isnot. In fact, however, I thinkthat in the situation describedyou should notbelieve that either suspect is guilty. Again, if you have an attitude towardthese propositions,you shouldsuspend udgmentabouteach, though perhapsnot about theirdisjunction.

There arepossible cases like the one just describedin which some actionis forced on you and it seems that you ought to form a belief about whichaction to take. Suppose thatyou know that one of two boxes in front of youhas a bomb that is about to explode. You have time to open only one boxand disarmthe bomb if it's there. You have no reason to preferone box tothe other. Surely you ought not suspendjudgmentaboutthe location of thebombanddo nothing.That evaluation s correct,but it's correctbecauseyouought not do nothing. Failing to act would be imprudent,perhapsimmoral.But you needn't have a belief about the location of the bomb in order to act.You can simply choose to open one box. And if you are the sort of person680 RICHARD FELDMAN


15/29

who can only pick a box in this case if you have the relevant belief, thentheremay well be a sense in which you ought to form a belief aboutwherethe bomb is (if you can). But the ought here is a prudentialor moral ought,not an epistemic ought.35You epistemically oughtto suspend udgment.

The third and most difficult sort of case to think about is one in whichyou have some modest amount of evidence supportinga proposition. It'stemptingto thinkthat it's permissibleto believe the propositionandit's alsopermissible to suspend udgmentaboutit, on the groundsthatbelief requiresmore thanjust this modest amount of evidence. In supportof this line ofthought is the idea that believing on only modest amounts of evidenceinvolves taking some epistemicrisk. Thereis no uniqueamount of risk thatis rightor rational.Rather,people simply have varyingattitudestowardthisrisk. My own inclination, which I won't defend here, is that you shouldbelieve when your evidence is supportiverather than neutral,even if theevidence is not at all decisive. Those who say that one's evidence supportsbelieving a propositiononly when that evidence is sufficiently strongwill inall probabilitybe faced with borderlinecases. In a case in which this modestamount of evidence is slightly in favor of a proposition, they might thinkthatbelieving it is permissible andthatsuspending udgmentis permissible.Of course, they are not forced to this conclusion, since they might hold thatit is simply indeterminatewhich attitudeone is permitted o have and that itis not simply true thateitherattitude s permitted.36

The evidentialist account of what we ought to believe relies crucially onthe notion of evidentialsupport.Analyzing this notion in a fully satisfactoryway is no easy task. Among the problems to be worked out is that ofdeterminingwhich logical consequencesof a body of evidence aresupportedby thatevidence. There arepossible cases in which apersonhas evidence thatimplies some proposition,but the connection between thatevidence andthatconsequence is distant and difficult to see. It may be well beyond theintellectual talentsof theperson.I believe that in such cases the personoughtnot believe the consequence. Given his failure to see that it is a consequence,to believe it (barringother reasons to believe it) would be rash.Furthermore,as I understand 02), it has exactly the rightresult in this sort of case. Thefact that a person's evidence implies some proposition is not sufficient forthe evidence to provideevidentialsupport or the proposition.Roughly, onlythose propositionswhose connection to the evidence the personapprehends35 Perhapsthe reasonit seems right to say that "Youought to form the belief if you can" isthat the "ought" n this judgment is one that does imply "can". This is furtherevidencethat the intended udgmenthereis not anepistemicjudgment.36 HudHudson helped me to see this point. Notice thatphilosopherswho thinkthat ties arepossible can still be evidentialists. That is, they can say that you always follow yourevidence and never believe contraryto your evidence. They just think that there arecases in which the evidence leaves open two possibilities and there's nothing else that

could eliminate one.SYMPOSIUM681


16/29

are actually supported by his evidence. And I think ascertaining thisconnection is itself an element of the person's evidence. These issues arecomplex, however,andI will notpursuethem furtherhere.

This completesmy descriptionof the evidentialist dea that one epistemi-cally ought always to follow one's evidence. I turnnow to a defense of thisidea.C. Evidentialismand Epistemic ValueOne way to explain why we ought to do something is to show that it is ameans to some goal that we have. A similar,but slightly different, way is toshow that it is a means to some valuable result. Both alternativesexplainwhat we ought to do instrumentally.To use the formeralternativeto showthatwe oughtto follow our evidence wouldrequireshowing thatpeople havesome goal-true belief or knowledge, perhaps-and that following theirevidence is the propermeans towardthat goal. If the oughts in questionaresupposed to be means to goals thatpeople actually have, then it seems thatonly people who do have the epistemic goals just mentioned would besubjectto the relevantepistemicrequirements.However,(02) is notrestrictedin that way. It says that all people epistemically ought to follow theirevidence, not just those who have adopted some specifically epistemicgoals.37

One mightarguethat all people naturallyhave epistemicgoals or perhapsthat the goal of true belief or knowledge is in some sense an "ideal"goal. Iwon't pursuethoseclaimshere.Instead,I wantto defend(02) by arguing hatfollowing one's evidence is theproperway to achievesomethingof epistemicvalue. My approachdoes not depend upon the assumption that epistemicvalue is a kind of instrumental alue. Of course,the view I'll defenddoes notimply thatbeliefs havingepistemicvalue lack any sort of instrumentalvalue.C]. TrueBelief and EpistemicValueThe dominantsource in the literature or discussions of epistemic goals andepistemicvalue is WilliamJames' famouspassage:There are two ways of looking at our duty in the matter of opinion,-ways entirely different,yet ways about whose difference the theory of knowledge seems hitherto to have shown littleconcern. We must know the truth;and we nust avoid error,-these are our first and greatcommandmentsas would-beknowers...38

37 One could take epistemic oughtsto be conditional.To say thatone epistemically oughttobelieve somethingis to say that if one hasepistemicgoals, then one ought to believe it.38 William James, "The Will to Believe," reprinted in The Theory of Knowledge, 2nd

Edition,pp. 555-62.682 RICHARD FELDMAN


17/29

One idea commonly extracted from this is that our epistemic goal is tobelieve all and only truths.39 ut if this implies that all people actually havethe goal of believing all truths,then I doubt that it is true. People's goals area varied lot, and I doubt thatall people have this one. Moreover,believing alltruths is obviously an unattainable goal. We simply can't believe all thetruths. Furthermore,attaining it is not desirable. As anyone who livedthroughPresident Clinton's impeachment rialknows, there aremany truthsone would prefernot believing (or even considering). So, it is doubtful thatbelieving all truths s anactual, attainable,ordesirable goal.

Whatever our goals, it makes sense to suppose that believing truths(orsome relatedstate) has epistemic value, thatit is a good thing from an epis-temic pointof view. I turnnow to the idea thathavingtrue beliefs and avoid-ing false beliefs has epistemic value and we should follow our evidencebecause it is the best way to have valuable beliefs. Consider, then, thefollowing principle:

VI. Each personmaximizes epistemic value by makingit the case thatfor every propositionp, he or she believes p if p is true anddoes notbelieve p if p is false.

One might then argue that people can best maximize epistemic value byfollowing theirevidence.There are a few problemswith a defenseof (02) that s based on (V1). For

one thing, (VI) would seem better suited to a defense of (01), since (VI)assigns value to believing all truths,not just to the ones about which onehas some attitudeor other.Furthermore,he allegedconnection between(VI)and(02) is basedon the assumption hatfollowing yourevidence is the bestway to get at the truth. And I just don't see why this is true. If you are inunfortunate ircumstancesn which the informationyou have will lead you tofalsehoods, following your evidence is not the best way to the truth.Thatis,it is not the most effective way to get at the truth.In these circumstances,ignoring yourevidence is a betterway to believe (some) truths.

A possible reply to at least the second problem just mentioned for theargumentfrom (VI) to (02) relies on the claim that while following one'sevidence might not be an infallible way to achieve the goal statedin (VI) oreven always the most effective way, it is nevertheless the best or mostreasonableway to tryto have true beliefs. This defense of (02) is quite simi-lar to the one I will present n section (C3) below and I will discuss it there.

3 The idea that this is our epistemic goal is widespread.It is the startingpoint for numerousdiscussions of a variety of epistemological issues. See, for example, Alston's "TheDeontological Conception of Justification,"p. 258 and Richard Foley's The Theory ofEpistemic Rationality (Cambridge,Massachusetts:HarvardUniversity Press, 1987), p. 8.

SYMPOSIUM 683


18/29

C2. Knowledgeand Epistemic ValueWhat James actually said in his famous passage wasn't that you shouldbelieve truthsor avoid believing falsehoods. Nor did he say that you shouldhave truebelief or the avoidance of false belief as a goal. What he actuallysaid is that our commandment,as would be knowers, s to know the truthandto avoid error. One might extract from this the idea that what has epistemicvalue is not mere true belief but ratherknowledge. Here's one way to spellout this idea:

V2. Each personmaximizes epistemic value by making it the case thatfor every propositionp, he or she knows p if p is true and knows -pif p is false.

(V2), like (VI), seems better suited to support something like (01) than(02), but I will ignore this point for now. One argument rom (V2) to (02)goes as follows. Given that one can have knowledge only if one has goodevidence for what one believes, one can get beliefs with epistemicvalue onlyby following one's evidence. In other words, following one's evidence is anecessary condition for getting what (V2) says has epistemic value. If weepistemically ought to do whatever is necessary to obtain this epistemicvalue, then we epistemically oughtto follow our evidence.

The idea thatknowledgeis what has epistemicvalue has a certainplausi-bility. Knowledge is valuable,andit makes sense to think thatknowledge isa particularlyepistemic kind of value. However, the argument rom (V2) to(02) is unsound.The centralproblemwith the argumentturns on cases inwhich a person has strong evidence for a false proposition, . According tothe evidentialistposition expressedin (02), sucha personshould believe thatfalsehood.And the argument or (02) based on (V2) relies on the assumptionthat the personshould adoptthe attitudethat is necessaryfor him to have inorder to have knowledge. But since f is false, there's an equally goodargument or the conclusionthat the person oughtto believe -f, since knowl-edge also requirestruth.If adoptingan attitudethat will yield knowledge iswhat's valuable,then in such a situationadoptingone attitude-believingf-will satisfy one necessaryconditionfor knowledge and adoptinga differentattitude-believing -f-will satisfya differentnecessarycondition for knowl-edge. (V2) provides no reason to think that one ought to believe f in thiscase. Thus, it fails to support(02).

A related point concerns cases in which one has strong evidence for aproposition, but evidence insufficiently strong to yield knowledge. (02)implies that one ought to believe in such a case. But it can't be that oneought to believe in such a case because that is a means to knowledge. Thesituationprecludes having knowledge, yet (02) has implications regardingwhat one oughtto do. Again, (V2) fails to provide support or (02).684 RICHARDFELDMAN


19/29

C3.ReasonableBelief and EpistemicValueWhile true beliefs may have considerableinstrumentalvalue, a person whoirrationally believes a lot of truths is not doing well epistemically. Incontrast,a personwho forms a lot of rational but false beliefs is doing wellepistemically. While knowledge also has a kind of value, seeing it as theonly thing of epistemic value fails to explain what is valuableabout formingbeliefs that fall shortof knowledge. We avoid the problemsassociated withidentifyingepistemic value with truebelief or with knowledge if insteadwesay that what has epistemic value are rational beliefs. To do well as abeliever, to achieve a kind of epistemic excellence, one must form onlyrational beliefs. I will discuss a defense of (02) based on this account ofepistemic value in this section.

One way to understand his evidentialist perspectiveon epistemic valueandepistemic oughts is as follows. Considera personwho is contemplatinga particularproposition. To carry out the role of being a believer in anepistemicallygood way, in a way thatmaximizesepistemicvalue, the personmust adopta rationalattitude oward this proposition.Thereare other valuesthatbeliefs might yield. Some beliefs might have prudentialor moral value.They mightmakepeople feel good or providecomfort for others. They mighthelp one to undertakerisky but beneficial behavior. They might give oneself-confidencethat can be help to advance one's career.But beliefs that arebeneficial in these ways can neverthelessfail to be rational.They can lackepistemic value. To achieve epistemic value one must, in each case, followone's evidence.

Here, then,is a principleaboutepistemicvalue thatsupports 02):V3. When adopting (or maintaining)an attitude owarda proposition,p,

a personmaximizes epistemic value by adopting(or maintaining)arationalattitude owardp.Given that (V3) specifies what maximizes epistemic value, a believer

epistemicallyought to form attitudesas directedby (02): he ought to followhis evidence.This defenseof (02) dependson the substantive pistemologicalthesis that rationality consists in making one's beliefs conform to one'sevidence. This thesis is apparently denied by some, for example somereliabilists. If reliabilismhas substantive mplicationsthatdiffer from thoseof evidentialism, it must be because in some cases it implies that it isrationalto formbeliefs in the absence of evidentialsupport.I will not under-takea defense of this evidentialistclaim here,norwill I discuss thereliabilistalternative.

Anyone who is about to adopt an attitudetoward a proposition, p, andwho adheresto the dictatesof (02) will get knowledge(of p or of -p) when-ever such knowledge is attainable. f the available evidence is strongenough

SYMPOSIUM685


20/29

to yield a belief well-enough ustifiedfor knowledge,and the otherconditionsfor knowledge are satisfied, then the person will have knowledge. Whereknowledge is not attainable, he personwill have reasonable belief. He'll bedoing as well as he can, epistemically speaking. By seeking rational belief,then, one will get knowledge where one can. And, of course, unless one is inunfortunatecircumstances n which one's evidence frequently eads to falsebeliefs, anyonewho follows (02) will mostly have true beliefs.

(V3) and(02) concern themselves with how to maximizeepistemic valuewhen adopting an attitude toward a proposition. There are many ways inwhich one might behave over the long term that will help one to gainknowledge of important acts. The relevant behaviormight include enrollingin suitable courses of study, maintaininga healthy life-style to keep one'smind sharp,and cultivatingsound inferentialhabits.Evidentialism is silentabout such practices. It focuses on the epistemic value to be obtainedimmediatelyfrom the adoptionof an attitudetowarda proposition.The wayto do that,in every case, is to follow the evidence one has.

At the end of section (CI) I mentioned he idea thatone might defend (02)on the grounds that having true beliefs is epistemically valuable and thatfollowing one's evidence is the most reasonableway to tryto get beliefs thathave epistemicvalue. The currentview is thatreasonablebeliefs areepistem-ically valuable and thatfollowing one's evidence is a perfectmeans to gettingvaluable beliefs. There are then two ways to support (02). The differencesbetween these defenses of (02) may not be great, but there are two pointsworthy of note. They can be broughtout by considering a person whoseevidence supportsa great manyfalsehoods.First, supposethe personfollowshis evidence. Accordingto both views, this personis believing as he ought.But accordingto theearlierview his beliefs, being false, lack epistemic valuewhereas according to the current view they are epistemically valuable.Second, suppose that the person does not follow his evidence. In thatcase,both views imply thathe is not believing as he ought, but the earlier viewimplies that he is, by luck, achieving epistemicallyvaluablebeliefs. In bothcases, I find the implicationsof currentview moreappealing.I don't see any-thing epistemically good aboutthe personwho irrationallygets truebeliefs. Idon't think thatit would be correctto say of him thathe's achievedepistemicexcellence, even thoughhe's done it in an irrationalway or merely by luck.Rather,I thinkhe's failed epistemically,not only because he isn't believingas he ought but because he does not have rational beliefs. Of course, theremay be some instrumental value in those true beliefs. They may help theperson negotiatethe world in a betterway. But that s a differentmatter.

686 RICHARD FELDMAN


21/29

D. Three Objectionsto EvidentialismDL. Evidence One ShouldHave HadConsider a person who is negligent about collecting evidence. Suppose that Ihave the firm belief that

G. Taking ginkgo supplements s a safe and effective way to improvemy memory.

I have a modest amountof evidence supporting G). I then see on the coverofa reputable magazine that it contains a major article about the merits ofgingko. Rather hanreadthe article,I avoid it for fearthatit will undermineabelief I preferto keep. Thus, I am negligent in collecting evidence; there isevidence I don't have butshould have.Furthermore, upposethat if I readthearticle,I would haveacquiredstrongevidenceagainst (G). Thus,while (G) issupportedby the evidenceI do have,I shouldhave had additional videnceandI shouldn't believe (G). So, (02) has the wrongresult in this sort of case. Itimplies that I ought to follow the limited evidence I do have ratherthanthelargerbodyof evidenceI havenegligentlyavoided.4"

The details of examples such as this one are quite important.There arepossible examples in which a person has some good reasons to believe some-thing,but knows thatthereare available strong considerations o the contraryalthough he chooses not to make himself familiar with that evidence.Suppose,that the magazineheadline in ourexampleis "GingkoShown to beIneffective." Upon seeing this headline, I immediately stop reading since Idon'twant evidence like that. I continue to believe (G) on the basis of my oldreasons. Surely my conductin this case is reprehensible.Butjust as surely,Ihave acquired evidence against (G) andmy belief losses some considerablesupportthe moment I see the title of the article. It gives me good reason tothinkthat there are strong objectionsto (G), even if I'm not yet in a positionto say in any detail what they are. That significantly alters the evidentialstatus of the propositionfor me. Given the credibilityof the source and thenatureof the articletitle, most likely my overall evidence no longer supportsmy belief. I no longer oughtto believe (G).

Variationson the examplearepossible. The articlemightbe in a publica-tion whose reliabilityis entirelyunknownto me or in a magazineI know tobe thoroughly disreputable.Other variations on the example concern thewording of the title. It could be something neutral such as "Some NewInformation on Gingko." Again I ignore the article. Again, assume that it40 Keith DeRose presentedan objection along these lines when he commented on a version

of this paper presentedat the Rutgers Epistemology Conference.SYMPOSIUM687


22/29

contains the same devastatingobjections.In this case, the mere awareness ofthe existence of the article has much less evidentialforce. Nevertheless, youmight plausibly think that in these cases, or at least some of them, I amterribly negligent in ignoringthe article. I shouldn'tdo that.Had I not doneit, I wouldn't have maintainedmy belief in (G). So, I ought not have thatbelief.

The statementof the objectionseems to me to state the heart of the reply.The name and source of the message do not provideme with much, if any,reason to stop believing (G).4'Until I've read the article,it would be bizarrefor me to stop believing what's supportedby my old evidence. I don't haveany reason to. Suppose that instead of negligently ignoring the article, I'mbusy and I simply set it aside until I can give it properattention. There'snothing negligent about this behavior. What should I believe in the mean-time? ShouldI think, "Well, somebodyhas writtensomethingaboutgingko,so I'd better stop believing that it is effective"? I think that the answer isobvious: I shouldn'tchange for that reason. And this fact doesn't changewhether my motives for settingaside the article are laudatoryor reprehensi-ble. In every case, the right answer depends upon what evidence I alreadyhave, includingthe evidence about thepossibleexistence of theseobjections.

No matterwhat the answersareto questionsabouthow I oughtto conductmy inquiry, where I ought to look for evidence, and so on, there alwaysremain the questions, "Whatshould I believe in the meantime?" "Whatshould I believe until I have a chance (or the courage) to look at that newevidence?" It's that natural and central question to which evidentialismprovidesa good answer.42

One might think that when one should look at additionalevidence, oneshouldalways suspend udgmentin the meantime.But that is a clearmistake,as the exampleabove illustrates.You should follow yourcurrentevidence. Ifyou do get moreevidence,thenyou should follow thecombined evidenceyouhave at that time.D2. Dutyto GatherEvidenceEven if it is true that in the examplesjust discussedI am believing as I oughtgiven the limited evidence I have, it is plausibleto thinkthat there are otherepistemic requirements hatI oughtto fulfill. In particular,n some versionsof these cases, I ought to obtainthe additional evidence that is available tome. More generally, in a wide variety of cases, it might be thought, one41 One might argue that we all know thatmagazines usually runarticleslike this only when

they are negative. If that's the case, then seeing the seemingly neutral title does provideme with some evidence against (G). The case is then similarto the first version in whichthe title is explicitly negative.

42 Richard Foley takes a similarapproachto epistemic questions. See especially Chapter 1of WorkingWithouta Net (Oxford:Oxford University Press, 1993).

688 RICHARD FELDMAN


23/29

oughtto gatheradditionalevidence about the propositionsone considers. It'snot enough simply to follow the evidence one actually has. Thus, evidential-ism overlooks important pistemic oughts. In a recent paper, "The EpistemicDuty to Seek MoreEvidence," RichardJ. Hall and CharlesR. Johnson arguethatyou have an epistemic dutyto seek more evidence about every proposi-tion about which you are not certain.43While many epistemologists wouldreject the demanding condition Hall and Johnson describe, it may seemobvious that in many cases you ought to seek evidence regarding proposi-tions about which you will form beliefs. However, the evidentialist viewdefended here apparentlyconflicts with this seemingly obvious truth. If thefundamentalepistemic goal is just to have reasonablebeliefs, then nothingabout evidence gathering techniques or the like follows as a means to thatgoal.44This goal just has implications concerningwhat attitudeswe oughttotakegiven the evidence we have.

By seeking out new evidence concerningsome importantpropositionandthen believing what that evidence supports,I don't do a better ob of achiev-ing the goal of believing reasonably.I achieve that goal at any moment bybelieving what is then supportedby my evidence. It's surely true that thereare times when one would be best off findingnew evidence. But this alwaysturnson whatoptionsone has, whatone caresabout,and othernon-epistemicfactors. As I see it, these are prudential or moral matters, not strictlyepistemicmatters.

A familiardistinctionmay help makethe evidentialist accountseem moreattractive.This is the distinctionbetween short-termandlong-term goals andthe related distinction between synchronic and diachronicrationality.Theformerconcerns questions of rationalityat a given moment while the latterconcerns rationalityover time. Evidentialismis best seen as a theory aboutsynchronic rationality.It holds that the epistemically rationalthing to do atany moment is to follow the evidence you have at that moment. It doesn'taddressquestions of how to conduct inquiryover periods of time. Thus, itdoes not addressquestionsabout how to gatherevidence, when one oughttoseek additionalevidence,and so on. Inmy view these diachronicquestionsaremoral or prudential questions rather than epistemic questions. You shouldgathermoreevidenceconcerninga propositiononly when havinga true beliefabout the subjectmatter of the propositionmakes a moral or prudentialdif-ference and gatheringmore evidence is likely to improve your chances of

43 See Richard J. Hall and Charles R. Johnson, "The Epistemic Duty to Seek MoreEvidence," American Philosophical Quarterly 35 (1998): 129-40. Hall and Johnsonactually defend the thesis that this remarkableduty follows from the assumptionthat ourgoal is to believe all and only true propositions.They do not defend the claim that this isour goal.44 Similarly, if the goal is just to now have true beliefs aboutthe propositionsabout whichwe are about to form attitudes,then there's no need to gather new evidence.

SYMPOSIUM 689


24/29

getting it right. Of course, whetheryou ought to gathersuch evidence alsodepends upon what otherthings you could do with your time. Epistemolog-ical considerations implydon'tresolve such matters.

There are cases in which one can spend one's time gathering evidenceabout propositions concerning inconsequential and trivial propositions oraboutmoreweighty matters.Evidentialismprovidesno guidance about whatto do. As I see it, this is not a weakness of evidentialism,since such choicesare not to be made on epistemic grounds.What topics you ought to investi-gate depend upon what topics are of interest to you, whatinvestigationscanhelp you to make your own life or the lives of othersbetter, and other suchmatters. Evidentialism is silent on those moral and prudential ssues, and Idon't see why it should address hem.D3. Being RationalBy AvoidingEvidenceThe evidentialistprinciplesstatedhere imply that one can get oneself into ahighly rational state by riddingoneself of as much evidence as one can andthensuspending udgmentaboutvirtuallyeverythingthatcomes to mind. If apersonfindsa drugor a machine that can erase memoriesfromhis brain andarranges o be immersedin a sensory depravation ank,he'll have very littleevidence regardinganything. By believing very little, he'll then be highlyrational according to evidentialist standards.Yet, this may seem just theoppositeof being rational.45

Evidentialists are committed to the view that the person just describedwould have little evidence regarding anything and would be rational tosuspend judgmentaboutnearly everything.But that seems to be exactly therightconclusion. Once the personhas lost his evidence, he has no reasons tobelieve much, and he'd be unreasonable f he did believe things thatwouldhave been well justifiedfor him hadhe been in more normalcircumstances.But he's not in normalcircumstancesand evidentialism concerns itself withassessing whathe shouldbelieve in the circumstanceshe is in, not with whathe shouldbelieve if he werein differentcircumstances.

Criticsmight thinkthathe shouldn'tget himself in a conditionof eviden-tial deprivation in the first place. Evidentialists can agree, though not onevidentialistgrounds.By puttinghimself in those conditions,he's made him-self unable to act on behalf of others who may rely on him. Perhaps hemorally ought to avoid renderinghimself useless in this way. Furthermore,by putting himself in a condition of evidential deprivation, he depriveshimself of all the pleasurethat comes from experiencingthe world. Unlesshis priorcircumstancesand futureprospectsareextremely grim,it is likely tobe imprudentto put himself in that situation. Evidentialists can agree withthese criticalevaluationsof this behavior,but these are not epistemicevalua-45 This objectionwas suggestedto me by TimothyWilliamson.690 RICHARD ELDMAN


25/29

tions. They are moral or prudentialevaluations of behavior related to theformationof beliefs.

I conclude that objections such as the ones considered here do not under-mine the evidentialist view of what we epistemically ought to do. Evidential-ism says that when adopting a doxastic attitude toward a proposition, aperson epistemically ought to adoptthat attitude hat is supportedby the evi-dence the person has at that time. How the person came to have that evidence,whether by conscientious inquiry or by avoiding potentially troublesomeinformation, is irrelevant to this epistemic fact. Similarly, how the personought to proceed in future inquiry is also irrelevant. Evaluations of thisbehaviorcanbe made, of course, but these evaluationsare of a differentnaturethan those made on evidentialist grounds. The evidentialist evaluationassesses whetherthe attitude ormedin the circumstancesone is actuallyin isthe attitudeone ought to have formedin those circumstances.

III. Epistemology and EthicsThere are several questions about the relation of epistemic duties (or oughts)to ethicalandother duties. Forexample,thereis in the literature onsiderablediscussion of the idea that there's some importantconnectionbetween whatone is epistemicallyjustifiedin believing andwhat one is morallyjustifiedinbelieving. On one side, there is the line of thought suggested by Clifford,according to which believing without epistemic justification is alwaysmorally wrong.To defendthis thesis, you'd have to argue thatsuchbelievingis always voluntaryor else thatinvoluntarybehavior can be morally wrong.I'm not inclined to arguefor either of these theses. Otherquestionsconcernwhetherepistemicduties are a "specialcase" of moraldutiesandwhether herecan be cases in which there is a conflict between these different kinds ofduties.46I won't take up these topics here. I will discuss one puzzling issuein this area.

The question I do want to address can best be raised by attendingto aremarkHall and Johnsonmakein defenseof their claim that we have anepis-temic dutyto seek more evidence for all uncertainpropositions.They say thatthis duty "is probablya prettyweak duty;most moral andprudentialdutieswould trump t in cases of conflict."47They quote Chisholm as holding thatmoral duties always trump epistemic duties. What I want to discuss iswhetherany sense can be made of the claim thatmoralduties trump,or donot trump,epistemic duties.

There is no problem with the idea that duties of the same kind can beweighed in importance.For example, a person might have a moral duty to46 For discussion of the latter question, see Eugene Mills, "The Unity of Justification,"

Philosophy and Phenomenological Research 58 (1998): 27-50. For discussion of theformer question, see Susan Haack, "TheEthics of Belief Reconsidered."

47 "The Epistemic Duty to Seek More Evidence,"p. 136.SYMPOSIUM691


26/29

his studentsto show up for class and a moral duty to his sick child to stayhome andcare for her.And one might say that one of these duties outweighs,or trumps,the other. The idea here, I take it, is that there is some scale ofmoral value and that fulfilling one duty contributesmore to that value thandoes fulfillingthe other.What is far less clear is that there is such a thing asjust plain ought, asopposedto the variouskinds of oughts philosophershave succeededin distin-guishing.48To understand he issue, it's importantto distinguish two verydifferent points. Suppose you accept the view that it is morally wrong tobelieve on the basis of insufficientevidence. (Set aside worriesaboutvolun-tarism.)I take it that at most you would thinkthat the fact that a particularactionis a case of believing on insufficientevidence counts towards ts beingwrong. As some would put it, it is primafacie wrong. This leaves open thepossibility that this case of believing on insufficient evidence has otherfeaturesthat countin its favor,morallyspeaking.It might be a case of beingsupportiveof one's family or disregarding ome excusable flaws in others. Itmight have much to be said for it. I have no trouble at all understandingheidea thatwe can in principlecomparethe negative moral value of failing tofollow one's evidence to the positive moral value of the other features of thecase. But herewe arecomparing wo moral factorsto reachan overallmoralassessment.

I take it that when people say things such as "Moraloughts trump epis-temic oughts" they are not saying that the moralweight of epistemic oughtsis less than the moral weight of other moral considerations. I believe thatwhatthey aresaying is thatthere is some sort of genericoughtthat somehowencompasses moral considerations, epistemic considerations, and perhapsothers,and thenweighs themagainstone another o come up with an overallassessment. This is not any particularkind of ought. It is just plain ought. Itake Hall and Johnson to be suggestingthat when you epistemicallyoughttogathermore evidence andyou morallyoughtto do somethingelse, the moralought "wins" and you just plain ought to do that otherthing. It's this that Ijust don't understand.Of course, by this I mean to suggest that no one elseunderstandst either. It makesno sense.

I know of no way to establish thatthe notion of just plain ought makesno sense. But I can give some partialdefense to my view by noting that itwould be a mistaketo assume that there must be such a thing as "justplainought"simply because there are various kinds of oughts. There needn't beanything meaningful about their combination. To see why, consider ananalogy. Suppose a child has two dolls. One is short andsquat.The other istall andthin.The child asks you which doll is "bigger".You arefaced with a48 I learned this way of formulating this question from an unpublished paper by Owen

McLeod.692 RICHARDFELDMAN


27/29

problem. One doll is taller. The other is wider and, let's assume, more volu-minous. The child could be asking which is taller, which is wider, or whichis more voluminous. Each of those measurescould be expressed by "bigger".We needn't assume thatthere'sany meaningful measurethat combines thesethree, yielding "biggerall things considered."Or, more precisely, we couldarbitrarily ombine the variousmeasures n any way we like. But it would bea mistaketo pick any suchmeasureas what really counts as "overallbigger."It would make no sense to ask whether height trumps width or volumetrumps height in assessing overall size. We've disambiguated the word"bigger". There's no puttingthe senses back together, at least no way thatisn't simply arbitrary.There's no such thingas "justplain bigger."Therearesimply the variousbigger-than elations.

Consider anotherexample. Suppose we are considering the wealth andstrengthof two people, A and B. A is stronger,B is wealthier. We oughtnotassume thatthere's some properway of combiningthese two measures intoone, which we mightcall "strealth".We can of course combine measuresofstrengthand wealth in a limitless number of ways. It might be that somecombinationwould in fact best suit ourpurposes.But the merefact that thereare two measures,wealth and strength, does not guaranteethat there is anyproper way to combine them to form strealth. The existence of numericalmeasuresof wealth and strengthought not deceive us here. We couldmeasureone's wealth in rubles and one's strength in terms of the number of gramsone can bench press. We could then add those numberstogether to get aproposedmeasureof strealth.Butwe also could add one's wealthin dollarstothe numberof poundsone can benchpress.That will give a differentnumber.And there are a limitless numberof other such sums. Each will yield poten-tial strealthrankings.The differentpossibilitieswill have the effect of count-ing one measuremore heavily thanthe other.If someone were to object toone of these strealth measures on the groundsthat it conflicts with the factthat wealth trumps strengthin measures of strealth,I think we should saythat we have no idea what thatpersonis talkingabout. We have no indepen-dent concept of strealthfor which there are correct weightings of the twocomponents.

I think that the same is true of the attemptto comparevarious kinds ofoughts to determinewhatonejust plain ought to do. Supposethat one beliefis prudentfor me-it will maximize my well being or what I care about orwhat is important o me-but it is not a belief I epistemically ought to havesince I lack evidence for it. We can understandthe idea that forming anunjustifiedbelief might be imprudent,since it might foster a bad habit. So,we can imagine weighing the short-termprudentialgain againstthe long termprudential ost. But I can see no values to which we could be appealingwhenwe ask whetherthe overallprudentialbenefittrumps he epistemiccost.

SYMPOSIUM693


28/29

For each "ought" here is an associatedvalue. We ought, in the relevantsense, to do the thing that maximizes thatvalue, or perhaps something thatdoes well enough in achieving thatkind of value. For example we morallyought to do what maximizes, or producesenough, moralvalue. We pruden-tially ought to do what maximizes, or producesenough, prudentialvalue. Ifthere is such a thing as "just plain ought"then there is a value associatedwith it. The thing we just plain ought to do is the thing that comes outhighest, or high enough, accordingto that measure. It's far from clear whatthat value would be-it isn't to be identifiedwith any of the more determi-nate kinds of value and there seems to be no uniquely correct(or range ofcorrect)ways to combinemoral,practical,epistemic,andothervalues. We'vedisambiguated"ought"and we can't put the various senses back togetheragain. There is no meaningful question about whether epistemic oughts"trump" r aretrumpedby otheroughts.

Stewart Cohen has presented the following objection to this view.49Suppose a child comes to her parentswith a problem.She is contemplatingsome actionand she's figuredout that it is the prudent hingto do butnot themoralthing to do. In otherwords, she realizes that she prudentiallyoughttodo it, but she morally ought not do it. She asks, "Whatought I do?" She isclearly looking for a way to resolve the apparentconflict. The view I'vedefendedrequires he parents o say that there is no meaningfulquestion here.Moreprecisely,when she uses the word"ought"n herquestion,she is eitherusing it in one of the disambiguated enses, in which case she alreadyknowsthe answer to her question,or else she is using it in the sense of "just plainought",in which case she's not asking a meaningful question. In the lattercase, the questionis like thatof child who wants to know which doll is reallybigger."'

It is difficult to resist the inclination to want to give the child adviceabout how to conductherself. And there is advice that one might give. Thequestion invites furtherreflection on whetherthe moral evaluation is reallycorrect, given thatthe act is harmful o her. The questionalso invites furtherreflection on whether the prudentialcalculation is really correct, given thatimmoral behavior is also often imprudent.And in the particularcase of achild askinga parentsuch a question,there is the matterof whetherthe childwill have parent's approvalwhichever way she acts. There is, finally, thequestion of what the parent predictshe or she would do in similar circum-stances. Thus, there are several meaningfulquestionsthat one might ask inthe sort of case Cohenenvisions. However,as I see it, none of themamountsto the questionof what onejust plain oughtto do.

49 In conversation.50 Another possibility is that she is using "ought" n yet anothermore determinate sense.This optionseems irrelevant o the present example.

694 RICHARD ELDMAN


29/29

IV. ConclusionI've argued in this paperthatdeontologicaljudgments about beliefs do notimply doxastic voluntarism. I've also argued that we do have a form ofvoluntarycontrol over a substantialnumberof beliefs, but that this fact is ofno greatsignificancefor epistemology.Epistemicevaluationsdo not dependuponwhether we have or exercise this voluntarycontrol. I've also argued, orasserted,that whatwe epistemicallyoughtto do is follow ourevidence:whenwe have a doxastic attitude towarda proposition,it ought to be the attitudethat is supportedby the evidence we have. I haven't attemptedto say any-thinghereabout what evidentialsupport s or what it is to have something asevidence. Those are topics for anotheroccasion. I have claimed thatthere isnothing more that a person epistemically ought to do than to follow herevidence in this way. In particular,activities such as gathering additionalevidence for propositionsabout which one is uncertainare not among thethings one epistemically ought to do, even if they are on other groundshighly desirable.Finally,I've expressedskepticism about the meaningfulnessof questions about whether or not epistemological considerationsare out-weighed by moral or prudential onsiderations n figuringout whatwe oughtto do all thingsconsidered.

The topics I've addressed by no means exhaust the issues that couldproperlybe raisedundertheheading"theethics of belief."But I hope to havemade some small contributionto our thinking about the specific topics Ihaveaddressed.51

51 In working on this paper I've benefitted greatly from numerous discussions with EarlConee, JohnBennett,StewartCohen,JonathanVogel, andall the students n my Fall 1998epistemology seminar. I'm also grateful to John Greco and MatthiasSteup for helpfulcomments. Earlier versions of the paper have been presented at the University ofRochester and at RutgersUniversity.I'm gratefulto the audiences on both occasions andto KeithDeRose for his provocativecommentson the latteroccasion.

Feldman - The Ethics of Belief (2000)

Documents

Transcript of Feldman - The Ethics of Belief (2000)