Defendiendo La Responsabilidad Epistemica

8/12/2019 Defendiendo La Responsabilidad Epistemica

1/28

Defending Epistemic Responsibility

by COREY CUSIMANO*

Brown University

Abstract

Te claim that we have epistemic responsibilities, and that we can beblameworthy or ailing to meet those responsibilities, has recentlybeen challenged by a number o theorists on the grounds o non-volunteerism: the claim that we do not have control over our belies.I argue that previous responses to this challenge ail to adequatelyaddress the problem and, urthermore, provide ancillary reasons todesire a proper deense o epistemic responsibility. I then argue that,

despite earlier challenges, we can be responsible and blameworthyor our belies, and that this responsibility is properly considered to

be epistemic in nature.

Introduction

M

y concern in this paper is to argue or the claim that weare epistemically responsible or at least some o our belies.As a consequence o this, I will also be arguing that we

are appropriately blameworthy or, at least, some o our belies andthat such blame is properly epistemic blame much in the way that,or certain actions, I am morallyblameworthy. Such a view has mostrecently been argued against by William Alston (2005)1. Since then,various epistemologists have attempted to deend some o our deonticconcepts (like the use o oughts) by sidestepping the importantchallenge o doxastic non-volunteerism. I will argue that these theories

either ignore or ail to accommodate the concept o responsibility intheir deontic ramework.

Tere are at least two reasons, I believe, to desire a deense o

* [email protected]. Received 1/2011, revised December 2011. the author.Arch Undergraduate Journal of Philosophy, Volume V, Issue 1: Winter 2012. pp. 32-59


2/28


3/28

judging each other as moral agents. I take it to be a very good excuse,when someone has done something wrong, to reply that they didnt

believe it was wrong and thereore should not be punished/rebuked/orso on. O course, ofen times will retort that one should havebelievedthat doing x was wrong. However, we only appropriately blamepeople and punish or chastise others when they are responsible orthe negative outcomes o their behavior. I we are never responsible(or cannot be responsible) then our claims that people should havebelieved certain things, especially in the moral domain, lose much otheir potency.

Some philosophers, including Alston3, argue that there is a sensein which we can be responsible or some o our belies but that it isnot epistemic responsibility. Is there a reason to think that, given theconcern above, we should preer to conceptualize our responsibilityas epistemic? raditionally, philosophers associate the goal o truth asbeing the dening eature o the realm o the epistemic. Furthermore,it seems that when a person is orming belies about what is right or

wrong, their primary goal is discovering and believing the truth (asopposed to ullling some other set o goals or duties that could alsoapply to belie ormation). Our duties concerning belie ormationmay be derivative o broader moral duties or responsibilities, but itseems to me that they are nevertheless epistemic.

Te Challenge of Doxastic Non-Volunteerism

In the ace o these two motivations, the claim that we areresponsible or our belies suffers rom an apparent critical problem:we do not seem to have any voluntary control over what we believe.Generic responsibility, that is as it may apply morally, prudentially,or epistemically, seems to require that the agent in question havesome control over the outcome or their actions in any scenario beorethey can be held responsible or blameworthy or what happens. Tischallenge is related to, but importantly different rom, the long-

standing principle o ought-implies-can.For the range o deontological concepts we employ, we can

distinguish between those in which something good or bad happened,

34 ARCH V:1 WINTER 2012


4/28

and those or which we are responsible or blameworthy or. Concerningthe ormer, we may claim something like he should have gone to an

all-boys prep school. Tis seems true only i there is a way in whichthey actually can (or cannot). Tat is, it makes sense to make thesekinds o claims because the boy could have easily been enrolled in adifferent school by his parents or that the salt actually does dissolve.raditionally, it does not seem true, or make sense to say things likethe boy ought to have been a Martian or ought to have been born todifferent parents. Tis is true o moral oughts as well. It seems trueto say that she should be nice to her new classmate but not o anyaction that the girl could not possibly perorm.

For a subset o these deontic ideals or claims, the agent has effectivecontrol over whether the ideal realized. Only or these do we claimthat a person was responsible or that outcome (or, in the case o a badoutcome, at least, blameworthy). For instance, we certainly would notblameor hold the boy responsible or not going to a different school he had no control over which school he went to. However, the girl had

control over how she acted toward her new classmate and, thereore,can be blamed or being mean. Te rst kind o oughts can be true andapplicable even i they do not have voluntary control over whetherwe ulll them, as long as they are something we can ulll, whereasjudgments o responsibility and blame necessitate voluntariness.

Given this background, we can understand Alstons challenge (oDoxastic Non-Volunteerism) as ollows4:

(1) Generic responsibility or blameworthiness or an action or staterequires an agent have voluntary control over that action or state.

(2) People do not have voluntary control over their judgments(actions) or belies (states).

(C) Tereore, people are not responsible or blameworthy or theirjudgments or belies.

Te rst premise ollows rom our discussion o ought-implies-can

and voluntary action. Tis premise will be challenged later by severalphilosophers but let us grant it or now. Premise (2) is the main parto Alstons attack and the one that concerns us now. What we will see

DEFENDING EPISTEMIC RESPONSIBILITY 35


5/28

is that Alston, and a number o other philosophers (like JonathanBennett, 1990), claim that we do not have voluntary control over our

belies because they are not properly responsive to our will.Alston invites us to introspect about our experiences o perorming

certain kinds o voluntary actions and contrast them to our experienceso orming and holding belies. Alston notes that we have varyingdegrees o voluntary control. We have the kinds o actions that wecan enact and complete just as soon as we can will ourselves to dothem; turning ones head or voluntarily blinking are good exampleso these kinds o actions. Ten there are more complicated actions,like opening a door or uttering a sentence, where it ar more likelythat we will somehow be interrupted and prevented rom completingour action. It even seems like we have voluntary control over certainlong-term goals and activities. For instance, I show a kind o long-range control over how clean my house is. I can sweep the oors, dothe dishes and laundry, and so on, but it likely that I will have to stopin between these activities. I may get tired or need to eat or take a

nap, but it is possible or me to return to my activity and eventuallycomplete it.

Forming belies, or ailing to orm a certainbelie, does not seembe under our control the way these simple activities are. Unlikeblinking or closing a door, we cannot simply will to believe something.Similarly, judging a certain proposition to be true is not something wecan accomplish simply by going about ullling certain sub-goals (likecleaning the dishes). In these examples the only necessary prerequisitesto accomplishing something were willing it and a relatively stableenvironment (i.e. one that would not interere with the activities set inmotion by our will). Judgments o the truth or propositions, or stateso belie and disbelie, are ormed. ake, or instance, what I considerto be an absurd proposition:

(a) My television has the power to elongate its cords and stranglepeople while they are asleep.

It is obvious to me that I could never get mysel to believe thisproposition5. It is also obvious to anyone who attempts to orm thatbelie that it is impossible to do. Tere are many other kinds o attitudes



6/28

I may take toward this proposition. I may hold (a) as a hypothesis oractor it into my decision making or even hope, wish, or pretend that

it is true, but the attitude o beliefis impossible to willully create.Not only do we seem unable to willully to orm specic belies,

but the belies that we do orm are also outside our control. Alstonwrites:

When I see a car coming down the street, I am not capable obelieving or disbelieving this at will . . . when I look out my windowand see rain alling, water dripping off the leaves o trees . . . I ormthe belie that rain is alling willy-nilly. Tere is no way I can inhibit

this belie.6

From this introspective exercise we conclude that belies are neitherthe kinds o things we can willully induce in ourselves nor the kindso things we can inhibit when they occur.

Alston concludes that, i we have any kind o control over ourbelies, it is indirect. Although my belies are not responsive to mywill or agency, I can come up with a long term goal involving selective

exposure to certain kinds o evidence and so to achieve a specicbelie. Some long-term projects, i they were reliable, would entailsome kind o effective control over our belies. However, the chanceso success o such a plan, according to Alston, are likely very slim.ake the proposition (a), above. I may try to watch horror lms abouttelevisions, draw angry eyes above the screen and tell my riends tomove our V to my bedside while I sleep so that, when I wake up,

I am (hopeully) startled. But even afer all o this effort, my successis not guaranteed. Although I can attempt to inuence my belies inthose ways, I am not guaranteed to achieve my goal and, in act, mostlikely will not succeed as a result o my willul actions. Tat is, I donot have an effective choiceas to whether I believe something or not7.Effective choice o the outcome o a judgment is, or Alston, what ittakes to exhibit voluntary control and, thereore, to be responsible orblameworthy.

Some philosophers have appealed to a dichotomy between practicaland epistemic reasons as an explanation or why we lack voluntarycontrol over our belies. For instance, Jonathan Bennett argues that



7/28

our will, the source o voluntary action or the idea o actions/statesoriginating rom the agent, is the capacity to be responsive topractical

reasons8

. Alternatively, belies are responsive to epistemic reasons,where epistemic reasons are those based solely on the truth or alsityo a proposition. Te reason why the proposition that my televisioncan strangle peopleis so strange is because there is so little evidence orit. Tese two kinds o reasons correspond to two different capacities,willing and believing, and normally they do not interact.

Tere are different ways in which our practical reasons, desires, orinterests do inuence belies. For instance, i I am highly interestedin baseball, I will orm belies about the game, various teams andplayers, and so on. In this sense, my belies are sensitive or responsiveto my interests. Te claim made by Bennett and Alston, however, isthat even i we can inuence generally what we orm belies about,we still cannot control what we believe about those things. Tat is, wecan inuence the range o propositions we orm judgments about, butnot whether each individual proposition is true (or alse)9. So, despite

the act that I can choose to orm a series o belies (o undeterminedcontent) about the inside o a room (by deciding to walk through it), Istill have not exhibited the relevant voluntary control over my beliesbecause I cannot decide to believe a specic proposition. Tere areplenty o times when the involuntariness o belies does limit what ourbelies are about (i.e. belies about rain or whether a car is speedingtoward you), but the relevant limitation is on the speciccontent o thebelie (whether it israining or not raining, and so on).

Finally, there is one other way in which the dichotomy betweenwill and belie seems to break down. Wishul thinking, andother orms o sel-serving or delusional belies, are apparentcounterexamples to the claim that our practical reasons do notinuence our belies. Consider the case o a person who is in love withsomeone else and desperately desires that person to love them back.Upon being rebuked, that person might, instead o orming the belie

that that person did not love them, believe that they really did lovethem and that the supposed rebuking was nothing other than a test.Such scenarios seem,prima facie, to be cases o our desires or practicalinterests inuencing what we believe.



8/28

However the processes o rationalization and evidential appraisaloperate, they nevertheless seem to be unresponsive to our will (even i

they are things we do want). Let us suppose that our best explanationor the situation above is such that our agent treats the others behavioras evidence of his or her love much the same way we treat seeing a caras evidence o a car to us. In this way, we can see how the belie is stillinvoluntary. Although our practical interests or desires may inuenceour belies in this indirect way, we can still see a division betweenthe will and our practical reasons, and belies and evidential reasons.Our voluntary actions are still constrained by practical reasons andwe are unable to will any indirect or direct inuence on our belies(or how we appraise evidence or certain propositions) with any sorto effectiveness. From these limitations, and the premise that we musthave effective choice/voluntary control over actions or states in orderto be responsible or them, we maintain our conclusion that we cannotbe responsible or our belies or judgments.

Various theorists, when aced with the challenge o doxastic non-

volunteerism, have responded in one o two ways. Tey have claimedeither that deontological judgments do not require voluntary controlto be true, or they deny premise two, and claim that we actually dohave voluntary control over our belies. Te problem with the rstkind o response is that, even i when it succeeds, it jettisons conceptso blame and responsibility, or at least declaws them. Te problem withthe second response is that it ails to address the actual challengedissued by doxastic non-volunteerism above.

Rejecting Ought-Implies-Can

Richard Feldman (2004)10, Hilary Kornblith (2001)11, and MatthewChrisman (2008)12provide accounts o epistemic deontologism on thebasis o rejecting the rst premise o Alstons challenge as explicatedabove. Tey each neatly build on each others theories and I willpresent them in turn. Alston assumes that deontic terms like should

and ought are necessarily intimately tied to their effectiveness asinjunctions and our responsibility or responding to them. Feldman,Kornblith, and Chrisman (FKC) have responded by arguing that we



9/28

can still use plenty o our deontic terms without relying on the premiseo voluntary control over belie.

o motivate rejecting the ought-implies-can principle, RichardFeldman shows us that, or plenty o our deontic locutions, we alreadydo so. For instance, we requently say things like eachers ought toexplain things clearly. Parents ought to take care o their kids. Cyclistsought to move in various ways13. Tat is, we apply standards to peopleall the time even when they are not able to do any better. Feldman reersto these standards as role oughts14.Importantly, these prescriptionsseem true and appropriate even when the agents reerred to wouldnever be able to ulll those demands even i they did the best theycould. A teacher may be incompetent, a parent incompetent, and acyclist uncoordinated but, in virtue o their respective roles, there arestandards that apply to them. Feldman grants that we do not hold themto standards at superhuman levels and that our prescriptions are, insome way, limited by their capacities. However, these standards are notlimited to the point o having voluntary control over ones behavior or

outcome and may still outstrip the capacities o a particular individual.Feldman argues that we should think o our epistemic oughts

as role-oughts. Tese oughts apply simply because, as cognitiveagents, we requently take on the role o being believers15. In act, weinvoluntarily take on the role o a being a believer but, according toFeldman, this does not undermine the appropriateness o the epistemicprescriptions. We are also involuntarily thrust into the role o beingeaters and breathers or which we can rightly say there are ways oneought to eat and breathe. For any activity that we take on, voluntarilyor involuntarily, there is a standard o perormance (consistent withthe goal o the role) that applies.

Feldman purposeully sets his sights short. He admits that someterms, especially those associated with praise and blame, are to bereserved or voluntary behavior16. His argument is limited to our useo requirements and permissions (that is, our use o the word ought).

He rightly points out that Alstons actual attack is against the notiono being ree rom blame or believing17 but dismisses the ideabecause he does not think that this more narrowly dened notion odeontologism is so natural or common18.



10/28

Whether Feldman has shown that we normally break the ought-implies-can principle, Feldman runs into trouble because role oughts

only seem to apply in virtue o taking on a role. Kornblith and Chrismanobject on the grounds that doxastic oughts seem to be categorical;thatis, they seem to be appropriate and applicable even when we are notactively taking on a role (or no matter which role we are taking on).Tis may be true or certain moral belies: It seems true that I oughtto orm a belie about whether what I am doing is wrong even i I amnot engaging with whether or not it is. We can also appeal to negativeprescriptions. It seems true that, no matter what I am doing, I shouldnever orm the belie that I am an alien, a robot, a princess, or so on.Kornblith and Chrisman also point out that it would be true roma role-ought perspective that, should a kleptomaniac involuntarilytake up the role o thie, it would still be true that she should steal19.However, categorically, the opposite is true: No matter whether sheadopts a certain role (or not), it seems true that she should not steal20.

Kornblith offers an alternative approach. He argues that Feldman

is right that some oughts are derived rom an evaluation o goodperormance (and that they can reject the ought-implies-can principle)but, instead o evaluation o roles and their goals, they derive romhuman ideals. He agrees with Feldman that there is a large middleground in between prescriptions beyond anyones capacities and thosebeyond merely a ew individuals. However, he claims, because idealsapply to individuals regardless o whatever role they are taking on (asa thie, slave, etc.), they maintain a normative orce that role-oughtslack, namely, categoricity. Similar to Feldman, Kornbliths treatmento epistemic oughts is orthogonal to the question o voluntarycontrol. For instance, in developing his view, he considers some oour epistemic oughts deriving rom our natural mental capacities.Tat is, he envisions certain oughts about what kind o inormationwe should believe (such as i one sees rain, one ought to believe it israining) based on how effectively we can make valid inerences (e.g.

the reliability o our visual-belie system).However, even i we grantthat we should be sensitive to the reliability or capacity o our mentalsystems, it still doesnt have any bearing about whether exercising thosemental capacities is considered voluntary. Te debate about ought-



11/28

implies-can, and arguing or oughts/ideals placed in some middleground, does not yet engage with our original problem o epistemic

responsibility or blame. Tis problem is more clearly seen in anaccount o epistemic deontologism provided by Matthew Chrisman21.

Kornblith and Feldman have developed accounts based on treatingour epistemic ideals as applying to our actions. Chrismans argues thatideals, insoar as they apply to belies, operate a certain way based onthe act that belies are states. Chrismans argues that ideals, when theyapply directly to actions, should be action guiding. Tis action guidingrequirement, argues Chrisman, is especially prominent in our moralideals: moral ideals that ail to lead to action or engage with how webehave seem to be bad ideals. Chrisman worries that because beliesare involuntary, any epistemic ideals, the way Kornblith understandsthem, will ail to engage our behavior. Instead, he points to manykinds o ideals that do not apply to actions but to states. For instance,our doctor may tell us that we ought to have a certain blood pressureor a mechanic might say that a bike ought to have a certain center o

gravity and so on. Tese oughts can be true, like the kinds o oughtsproposed by Feldman and Kornblith, even when our blood pressure orthe balance o a bike are not within our (or someone elses) voluntarycontrol that is, nobody can choose to do anything to change it.

Chrisman, in developing his account, ollows a distinction usedby Wilred Sellars between ought-to-dos and ought-to-bes22. Tatis, oughts that govern actions are ought-to-dos, or rules of action,and oughts that govern statesor, ways of being are ought-to-bes, orrules of criticism. Rules o action presuppose the ought-implies-canprinciple whereas rules o criticism do not. Rules o criticism do nothave to directly engage with our behavior. Because o this, Chrismanbelieves, we can maintain the categoricity o epistemic evaluationsdesired by Kornblith but avoid the worry o non-volunteerismbecause they apply to states instead o actions. However, according toChrisman and Sellars, rules o criticism do indirectly engage with our

behavior: they imply rules o action.Chrisman highlights a ew ways that rules o criticism imply rules o

action: conditionally, universally, and existentially. ake the ollowingrule o criticism: X ought to be . According to the conditional view,



12/28

the rule o action that is derived is:

I someone is responsible or Xs being , then that person ought to

do what he/she can (ceteris paribus) to bring it about that X is .

According to the universal view:

Everyone ought to do what he/she can (ceteris paribus) to bring itabout that X is .

And, nally, according the existential view:

Someone ought to do what he/she can (ceteris paribus) to bring itabout that X is .

Ultimately, Chrisman does not endorse one o these methods over anyother and leaves it an open question whether some or all o them areappropriate methods or deriving rules o action23.

Te problem discussed above in Feldman and Kornblith (and nowChrisman) becomes prominent in Chrismans analysis: Chrismanrelies on the notion o responsibility in explicating this view. Tisis signicant or a number o reasons. Firstly, it highlights the act

that the kinds o analyses being offered by Feldman, Kornblith, andChrisman are inappropriate (or missing the goal o) developing a viewo how we can, i we can, be responsible or our belies. Secondly, ithighlights a need or an account o responsibility because, whetheror not people can be responsible or their own belies or the belieso others will determine how effectively we can derive rules-o-actionrom rules o criticism. I Chrisman ails to be able to derive rules o

actions rom rules o criticism then he loses the supposed advantagehe had over Kornblith concerning an analysis o human ideals.

From our critique o volunteerism above, we have reason to besuspicious both o whether we are responsible or own belies (whichis the original charge) and also o whether we can effectively modiyother peoples belies as well. Te reasons we lack control over our ownbelies seem to apply to any scenario in which we were attempting tocontrol or change someone elses belies. Tat is, while we can do our

best to provide evidential reasons or a certain belie, we cannot controlor effectively inuence how they appraise or respond to that evidence(just like we cannot control how we respond/appraise evidence we



13/28

are evaluating). Given this, it seems unlikely that the universal orexistentialmethods o deriving rules-o-action would adhere either.

All three o these theories ail to address Alstons primary challengeagainst epistemic responsibility and blame. Some, like Feldmans,purposeully sidestep it or, as we saw, dismiss it as being unnaturalor uncommon. Kornblith and Chrisman do not engage with the issueor, as we saw with Chrismans account, even attempt to build a theorypresupposing some aspect o responsibility.

Tis ailure mirrors an important question about moral oughtsand responsibility. We might think that moral oughts are derivedrom human ideals and yet wonder i people are ever responsible orailing or succeeding their mandates. For example, beore we reproacha teacher or ailing to be an ideal teacher, we do not wonder merelyi she had any sort o contractual obligation and whether or not shemet the requirements. We inquire into how much effort she exerted,whether she knew she was doing a poor job and whether she had othercourses o actions available to her (like nding another job). Once

we have some understanding we begin to orm a judgment about thecharacter o the teacher. I we nd out that she is a terrible teacherbecause she just doesnt care, we will orm negative judgments abouther above and beyond the initial judgment that she is a bad teacher.

Beore attempting to develop an account o epistemic blame andresponsibility, it will helpul to review Alstons challenge against it.

Responsibility without Effective Voluntary Control

One o the specic claims Alston was arguing against was theclaim that we had control over our belies (and, thereore, epistemicresponsibilities) because we had the ability to deliberate when orminga judgment. One o his opponents, Roderick Chisholm aptly describesthe intuition motivating this claim:

When a man deliberates and comes nally to a conclusion, hisdecision is as much within his control as is any other deed we attribute

to him. I his conclusion was unreasonable, a conclusion he shouldnot have accepted, we may plead with him: But you neednt havesupposed that so-and-so was true. Why didnt you take account o



14/28

these other acts? We assume that his decision is one he could haveavoided and that, had he only chosen to do so, he could have made

a more reasonable inerence. Or, i his conclusion is not the resulto a deliberate inerence, we may say, But i you had only stoppedto think, implying that, had he chosen, he could have stopped tothink. We suppose, as we do whenever we apply our ethical or moralpredicates, that there was something else the agent could have doneinstead.24

Alstons response to Chisholm is to show that, when we attempt to ormspecic proposals about how voluntary control gures in the practice

o deliberation, they all turn out alse. For instance, Chisholm could besuggesting that i we deliberate, we are able to gather new inormationor uncover new evidential reasons rom which we may be able to willajudgment essentially, exhibit immediate voluntary control over ourbelies. However, obtaining reasons, be they evidential or practical,does not allow us any control over which o those reasons will applyin the outcome o the belie. Our belies will still be outside our will,

no matter how many new reasons we obtain to believe something.We can imagine that I may deliberate and call to mind many practicalreasons why I should believe a certain proposition without gainingany ability to act on them.

Instead o relying on an implicit thesis o immediate voluntarycontrol, Chisholm could be suggesting that such deliberation offersus long range control over our belies. Tis long range control beanalogous to indirect voluntary control, discussed near the beginning

o the essay. Tis is a more plausible, but ultimately misguided, thesis.Tis is because deliberation, as weve already discussed, is limited tothe activity o mobilizing or discovering certain reasons (evidentialor otherwise). While we can choose to engage a certain question, andengage in deliberation, all we are merelydoing voluntarily is decidingto inquire into some proposition. Alston writes:

Claims like [Chisholms] . . . ignore the difference between doing A

in order to bring about E, or some denite E, and doing A so thatsome effect within a certain range will ensue . . . At most it showsthat I have long-range voluntary control over whether I take up somepropositional attitude toward some proposition (Alston 2005, p. 71)



15/28

According to Alston, what we need in order to display effectivevoluntary control and, thereore, to be responsible or our belies/

judgments is the ability to take up a specicattitude toward a specicproposition. Which, as discussed earlier in the essay, is something wecannot do.

Despite Alstons critique, the intuition that we somehowhave responsibility over (some) o our belies because o some actabout deliberation (or our ability to deliberate) still seems right tome. Tis is or the specic reason that deliberation does not merelyentail the ability to take up some attitude toward a proposition but,rather, the ability to take up some propositional attitude of a certainqualitytoward some proposition. It would be meaningless i the onlyquality that could be applied to our belies afer deliberation was themerely quality o having been extensively deliberated about. However,deliberation can give our belies many kinds o qualities, such as beingevidentially rich, one-sided, narrow-minded, and so on, dependingon how we deliberate.

Our deliberation (or lack thereo) will (at least some o the time)play a determining or necessary role25in the eventual judgmentwe orm. When we do have voluntary control over something thatplays a determining role in a state then we can, sometimes, be heldresponsible (or blameworthy) or that state even when we do not havedirect voluntary control over it. Alston introduces this option as anoption or the epistemic deontologist:

Consider the general point that we can be blamed or a state o affairsF, provided something we voluntarily did (didnt do) and should havenot done (done) was a necessary condition (in the circumstances)o the realization o F. Tat is, F would not have obtained had wedone (not done) something we should have done (not done). I mycholesterol buildup would have been prevented had I regulated mydiet in the way I should have done, but didnt, I can be blamed orthat buildup, whether or not I have direct effective voluntary controlo my cholesterol level. (Alston 2005, p. 74)

Tat is, occasionally the only reason a belie we have ormed is deective(that is, not true) is specically due to our activity o deliberation. Tebelie was only deective because we did not deliberate and, had we



16/28

deliberated, would not have been deective. Returning to the grocerystore example rom the beginning o the paper, I could have engaged

with the question o whether buying non-organic ood was appropriateand searched my memory or inormation or evidence. I I had donethat I would have swifly remembered the conversation I had with mypartner and ormed the true belie that buying non-organic ood isinappropriate. My ailure to deliberate was the only thing preventingmy belie having a negative epistemic quality, namely, the quality obeing alse. Just as I would be blameworthy or my bad cholesterollevel i my diet was really the only thing preventing its eventualdeterioration, I would be blameworthy or my alse belie.

Te number o things we can voluntarily do that inuence beliesextends beyond deliberation. We can search or evidence, talk to ourriends, consult experts, and so on. Alston is aware that our beliesare affected in these ways; however, he claims these activities asconstituting intellectual, instead o epistemic, obligations. He claimsthat:

S is intellectually to blame or believing that p iffi S had ullled allher intellectual obligations, then Ss access to relevant considerations,or Ss belie-orming habits or tendencies, would have changed insuch a way that S would not have believed that p. (Alston 2005, p. 77)

One o my intellectual obligations might be to search my memory orpossible evidence o a propositions alsity. I doing that would havegiven my belie-orming mechanisms or habits access to inormation

that would have resulted in a different belie, then I am blameworthy.Although Alston grants that we have intellectual duties, hemaintains that they are not properly considered epistemic. His claimthat they are not epistemic is strange since, as he himsel claims, ourintellectual responsibilities are rooted in an obligation to seek thetrue and avoid the alse in belie (Alston 2005, p. 76) which strikesme as eminently epistemic in value. He is claiming that we havea responsibility to seek the truth and avoid alsity but it is not an

epistemic responsibility. Given that epistemologists dene the domaino epistemology as that governing truth, excluding our intellectualobligations rom epistemic value seems inappropriate.



17/28

How, then, does Alston reconcile this oddity? Alston claims thatwe actually do not every have epistemic responsibilities because any

apparent ones (such as the intellectual ones) do not actually yieldepistemic value.

Alston and Epistemic Value

Like many epistemologists, Alston considers the domain o theepistemic, both in normative and descriptive theorizing, to governprocesses or goals essentially related to truth (as opposed to the good

as in the domain o morality, or instance). He writes,[EpistemicDesiderata] are those eatures o belies or bodies thereo are valuablerom the epistemic point o view, dened in terms o the aim atacquiring true rather than alse belies about matters that are ointerest or importance to us (Alston 2005, p. 47). BonJour provides asimilar sentiment:

I our standards o epistemic justication are appropriately chosen,bringing it about that our belies are epistemically justied will also

tend to bring it about that they are true. I epistemic justicationwere not conducive to truth in this way then epistemic justicationwould be irrelevant to our main cognitive goal and o dubious worth.(2005, p. 7-8)

Te problem with the proposed model o intellectual obligations isthat it does not sufficiently connect us (or, our belies/judgments) withthe truth:

It is prima acie conceivable that being ormed in a way that doesnot depend on violations o intellectual obligations should be away o rendering a belie probably true. Nevertheless . . . there arevery many sorts o cases in which one does as much as could bereasonably expected o one in the way o voluntary acts leading up toa given belie without the belie s thereby acquiring any considerablelikelihood o truth. (Alston 2005, p. 78)

Te analogous case or a potential moral obligation is establishing

whether ullling that obligation yields something good. His argumentis as ollows:

(1) Te ulllment o our intellectual obligations be considered



18/28

epistemically desirable iff it is the case that it entails a considerablelikelihood o truth.

(2) Te ulllment o our intellectual obligations does not entail aconsiderable likelihood o truth

(C) the ulllment o our intellectual obligations is not epistemicallydesirable.

Beore critiquing this argument, it is worthwhile to explore someother criteria that are epistemically desirable to understand why ourintellectual obligations are restricted to truth-conduciveness. Besidestruth-conduciveness, Alston acknowledges that something can alsobe epistemically desirable i it is avorable to the discrimination andormation o true belies (Alston, 2005, p. 43). Epistemically desirableeatures include:

(A) Having some high-grade cognitive access to the evidence, and soon, or a belie (B).

(B) Having higher-level knowledge, or a well-grounded belie, that Bhas a certain positive epistemic status, or

(C) Tat one can carry out a successul deense o the probability otruth or B.

As opposed to rendering a belie likely to be true, these qualitiesalready presuppose the truth o a belie. Tese constitute higher-levelknowledge that indirectly leads to more true belies. Tat is, i I know

that certain kinds o belies, like visual belies, are generally truth-conducive (or that a belie is true because it was a visual belie) I willbe in a position later to orm true belies about what I see. However,ullling our intellectual obligations does not yield these kinds ohigher-level epistemic resources.

Te last group o Alstons epistemic desiderata includes eatureso systems o belies that we typically hold as goals o cognition.Tese include having belies that are coherent, provide reinorcingexplanations, result in an agent acquiring some kind o understanding,and so on. Importantly, these desiderata are not valuable becausethey produce true belies. Tey are valuable only when they rely on



19/28

true belies. Alston claims that unless truth can be assumed, theseeatures o belie systems would ail to exhibit the intrinsic cognitive

desirability that would otherwise attach to them (p. 46). Alstonsargument or including them on the list o epistemically desirablequalities is their otherwise intrinsic cognitive value and their relationto a system o true-belies. Our intellectual obligations ail to qualiyor this category, according to Alston, because even i we establishedsome intrinsic value independent o truth, that value is in no waydependent on being associated with a preponderance o true belies(Alston 2005, p. 78).

Returning to Alstons primary argument, he admits that, primafacie, it seems like premise 2, the claim that ullling our intellectualobligations actually does yield a high-probability o true-belies is true(see the quote above). However, Alston points to scenarios in whichit clear that not only have we ullled our intellectual obligations, butthat belie still is not likely to be true. For instance, it is a commonoccurrence that people are too unintelligent to orm the correct

belie. No amount o deliberation will help a rst-year student graspthe ner points o Descartes i they simply cannot understand it. Incontrast, something like a sufficiently reliable belie-orming processis, by stipulation, truth-conducive so long as it is ullled. Alstonalso points to areas o our lives where it would be unreasonable tohold any such obligations even though our belie orming habits arenotoriously unreliable. For instance, he claims accepting inormationrom authorities is unreliable and, at the same time, unreasonable tokeep people obligated to constantly check sources, ollow-up on theinormation, and so on. Alston also points to requent occurrences owhen people are overcome by certain belies or certain ideas, like God.No deliberation (or other voluntary action) will make them attune toevidence or reasons that will yield a different (true) belie. I will returnto this part o Alstons argument later. For now I will concern myselprimarily with the notion o epistemic value.

Te rst problem with Alstons argument is that the requiremento a considerable likelihood o truth or epistemic value proves toomuch. Tere are many things I would take to qualiy as an epistemicideal (and on the basis o some conduciveness criterion) that would



20/28

not qualiy under Alstons thesis. ake, or instance, the ability toreliably orm accurate analyses o the validity o syllogisms. Tose who

can reliability do it are more ideal epistemic agents that those who areunreliable. Te problem is that such ability does not guarantee any sorto reliability in the belie about the conclusion o the syllogism. Tisis or the simple reason that, as valid or invalid as many syllogismsare, so many o them are simply not sound. Te premises rest on alseconclusions; thereore, the vast majority o syllogisms are alse, and oursupposed ideal is rendered epistemically impotent. We might extendthis to ideals about Bayesian belie updating. Tat it, it seems plausibleto me that having our belies undergo perect Bayesian updating (thatis, the total incorporation o prior, likelihoods, and so on into ourposteriors) is epistemically ideal. However, so many o our belies arealse in so many domains that, even with perect belie updating inlight o evidence, our belies are not guaranteed likely to be true.

Tese two examples highlight an important class o epistemicactivities not captured by Alstons taxonomy. Many o the constraints

o rationality (with Bayesian inerence being a controversial epitome)are, at least to, intuitively epistemically valuable. Tis is not becausethey guarantee a certain likelihood o truth but because they arethe most sure-re way o getting at truth. As such, they are heavilyimplicated in the aim o truth even i they do not guarantee itssuccess (or a certain threshold likeliness o success). Teir value then,is not an objective threshold (like the truth-conduciveness qualityAlston deends) but a conditional truth-conduciveness. Tis leadsto a strange picture. Te ulllment o these different rational orevidence-seeking obligations makes us ideal conduits or truth. Giventhe right environment (this includes aspects o our cognition outsideour will) our willmaximizes the likelihood o true belie. Alston isright to point out that many other parts o us all prey to problemslike sheer unintelligence or deunct belie-orming mechanisms. Tisleads the surprising disconnect between our otherwise great epistemic

characters and our lousy belies. Tat is, we can do the very best wecan to ollow the aim o truth without thereby rendering it likely thatour belies are true.

I will not argue or a conception o epistemology that is not tied



21/28

to the acquisition o true belies26. Instead, I propose a more lenientcriterion o when epistemic obligations can be induced. Instead o

limiting epistemic desirability to only belies (or processes whichguarantee a signicant likelihood o truth), I propose that epistemicvalue is added to a process (or activity, unction, etc.) whenever itmakes the likelihood o truth in the resulting belie more likely. Now,instead o only having intellectual obligations which always guaranteea signicant likelihood o belie (o which there are none), we haveintellectual obligations whenever ullling them leaves our resultingbelie more likely to be true than had we not.

Tis more lenient criterion has the benet o applying tomany o qualities o rationality that intuitively seemed to provideepistemic value. Additionally, many o the activities we envisionedas constituting our intellectual responsibilities qualiy. Tings likesearching or more evidence, spending more time on an argument,and so on, make our resulting belies at least a little bit more reliable(even though there are times, as Alston has apparently pointed out,

when they are not very effective). By accepting a more lenient versiono the conduciveness criterion we are most likely limiting the role thatepistemic blamelessness can play in other normative matters. BonJourand Alstons emphasis on a robust connection or guarantee o truthis a popular notion and has played an important role in developingaccounts o justication and knowledge. Given this, the epistemicdeontologist may have to accept that being epistemically responsibleor blameless does not entail being epistemically justied or havingknowledge. I I am right about the importance that epistemicresponsibility and blamelessness play at least in our social lives, thenthis act is not disconcerting.

Te account we have developed here posits the ollowing: Wehave the epistemic goal o acquiring a true belie. From this epistemicgoal, we have epistemic obligations to do what we can to increase thelikelihood o acquiring a true belie. Tese obligations supervene only

on behaviors we have voluntary control over, severely limiting theextent to which we actually do have responsibility (or can be blamed)or a belie. However, when we ail to ulll those obligations we areepistemically irresponsible (and blameworthy) or our belie being less



22/28

likely to be true than it could be). When our irresponsible epistemicbehavior is (at least signicantly) implicated in the production o a bad

belie (alse, ill-grounded, etc.), we are blameworthy or that belie.We can also understand how, when other kinds o obligations

in other domains o value invoke the goal o truth they, in act, areimplicating epistemic value. When we have a moral duty to orm truemoral belies and we act epistemically irresponsible we are rightlythought o as both immoraland a bad epistemic agent. We are a personwho not only ails to appropriately care about our ethical dutiesbut also a person who does not care about truth. Te ailing o ourintellectual responsibilities (which we now understand as epistemicresponsibilities) can yield blame in more than one domain o value.

Whether or not our appeal to the more lenient value criterion issuccessul, it is important to engage with Alstons second premise.Tere is at least one reason we should hope it is not true: mainly, iit is, that then those kinds o intellectual obligations would rarely, iever, operate. I, even afer all o our hard work, we are still incredibly

unreliable, then those responsibilities are not appropriate or the goalo achieving true or alse belies. Recall that such responsibilities arerequently co-opted by moral and practical obligations on the basiso acquiring true beliefs being relevant to those domains. It seems asthough one crucial reason we have a moral obligation to discoverthe correct moral principles (or orm the correct moral appraisals oactions and so on) is so that they increase our ability and tendency todo good. I it turns out that actually, all o the different ways we couldhave (indirectly) inuenced our belies never helped achieve thatgoal then the goal o those moral obligations is lost and we lose thosederived obligations (epistemic or not). Te same problem applies inother domains o value.

For reasons I will get to in a moment, I do not nd Alstonsevidence or Premise 2 to be very compelling. However, work inrecent evolutionary and social psychology has threatened the idea

that our deliberation and active reasoning play even a positive role inour goal or true belies. Psychologists, responding to years o researchshowing how common and robust certain ailures in reasoning arein human beings have proposed and deended a model that claims



23/28

that our tendency or rational error is built in to the very unctiono explicit reasoning (see, or instance, Mercier & Sperber, 2011).

Tat is, our explicit deliberation and reasoning unction to reinorcepreviously held belies and reject evidence against them. Instead oabiding by principles o rationality that increase the likelihood ogetting our belies right, deliberation and explicit reasoning actuallymake us more dogmatic and closed-minded27.

We can conceptualize the challenge as ollows. I we leave ourmental obligations too lenient, then they are easy to accomplish butthey are susceptible to the biases we mentioned above. Alternatively,we can attempt to limit the effect o the biases by making our mentalobligations specic (and balanced against these biases). For instance,we might say that we have an obligation not merely to seek outevidence but, rather, to specically seek out evidence contrary to ourbelies. However, these are more difficult to ulll, even when a personis trying to do so. I a person ails even they were attempting to ullltheir obligation they are not blameworthy. On the one hand, we have

a people succeeding their obligations but those obligations themselvesnot increasing the likelihood o truth. On the other, our obligationsare well-ormed but people are blamelessly ailing to accomplish them.Either way, these mental obligations are ill-suited or epistemic value.

I ocus on deliberation and explicit reasoning or two reasons.First, many o the obligations we were supposing actually appliedinvolved things like searching our memory or evidence, analyzingarguments, and so on. I it turns out that these are subject to the biasmentioned above and that engaging in them does not actually help usget to the truth then many o our supposed obligations disappear. Tesignicance o this point is related to the second reason. Recall romChisholms long quotation above that many o the intuitions aboutthe role we play in our epistemic lives is connected to our mentalbehavior. One o the supposed advantages o establishing a theory oepistemic deontology was its ability to capture many o these kinds

o intuitions. So, although we could probably nd other kinds oepistemic obligations (that deal more with non-mental behavior), iwe lose the obligations over our mental lives, the victory is bittersweet.

Te solution to this problem is just that we have not done a good



24/28

job speciying what our mental obligations are and, importantly, inwhat kinds o contexts they apply. While it is true that we cannot have

obligations such as update your belies according to Bayesian idealsbecause people are generally not cognitively capable o that, we canhighlight, in certain contexts, what kinds o mental activities tend tobe the most conducive. So, when orming belies about subjects weare not amiliar with, our epistemic goal is to deer to authority. WhenAlston gives the example o the student who is unable to understandDescartes, we can simple alter the obligation. No longer does thatstudent have an epistemic obligation to deliberate to personally mullover the philosophical arguments but, rather, to ask his proessor or.A. or so on. Even i he still could not orm the right belies, at leastapproaching authority in this case has made a true belie more likelyto result.

We can respond to the challenge in part by admitting thatsometimes we should not rely on our mental prowess alone. However,there are still plenty o times when what our obligation consists in

is remembering certain pieces o relevant inormation. Rememberingspecic acts is, in plenty o relevant scenarios, not limited by theargumentative biases noted above. Recall the aorementioned exampleabout orming the wrong belie because o my ailure to rememberwhat my partner said to me mere minutes earlier. Such obligations assearching or memory can greatly increase the likelihood o orminga true belie in these kinds o scenarios. Finally, we might considerthat some people have ar more intellectual obligations than others.A brilliant scientist is going to be held ar more responsible or theirbelies than someone vastly undereducated. Te epistemic deontologistcan happily accept this development.

Alstons challenge about the reliable o our voluntary behaviors,and the similar arguments developed rom that, are serious worries.I have attempted to show that perhaps there is hope or nding theright kinds o rules and obligations that will tip the scale o truth-

conduciveness in its avor. However, the challenge is undamentally anempirical one and is dependent on the deontologist actually creatingthese systems o responsibilities, which I do not have the space to dohere.



25/28

Conclusion

As we saw, Feldman, Kornblith, and Chrisman were primarily

interested in the legitimacy o claims such as you ought not believex. We came to avor an account based on epistemic ideals, with thevarious obligations alling in a large middle ground what peoplewere capable o and what they were not. Te middle ground is abalance between rejecting the ought-implies-can principle and thenotion o engaging with, and guiding, our behavior. For instance,i a person is presented with a syllogism and they orm the wrong

belie about the validity o that syllogism, we might rightly say thatthey ought not have done that because it is within the human ideal tocorrectly interpret a syllogism, even i it is just beyond that particularindividuals capabilities.

Where all o these theories ell short, however, was in providingan explanation about when people were responsible or those ailuresand when they were not. It seemed that, because we have no will tobelieve and, thereore, no voluntary control over our belies, that

we are never responsible or our belies. However, we ound that wecan be blamed or our belies when we ail any number o epistemicresponsibilities we hold over our voluntary actions and those aresignicantly responsible or a deect in our belie. One challenge wedealt with was that o understanding how ullling those obligationswas properly understood as epistemic in value.

As long as those responsibilities are epistemic, we have the tools

to connect the projects o Feldman, Chrisman, and Kornblith, with anaccount o epistemic responsibility. Consider another case o someonebeing presented with a syllogism. We maintain our epistemic oughtthat they should believe that the syllogism is alse. Tis person,however, orms the belie that it is in act a valid syllogism but they doso because they didnt actually read the entire thing (merely the rstpremise, say). In this case, they violated our epistemic ought and theyare responsible and blameworthy or it.

Although we have established that we have epistemicresponsibilities, and that we can be blamed or our bad belies, wemust still understand that any kind o responsibility and blame we



26/28

have or our belies is derivative o our responsibility or the parts oour epistemic lives we actually have control over. However, because

we do have control over certain epistemic parts o our lives, we can bebetter or worse epistemic agents, much in the way we can be moral orimmoral. Peoples expectations o the kinds o belies have are inormedby the way we lead our epistemic lives, that is, the voluntary parts wehave control over, and the epistemic situations we nd ourselves in.

Te author writes: I am grateful for the helpful comments suggestedby the editors of Arch. I have especially beneted from the manystimulating conversations with and comments from revor Brothers,Geoffrey Sayre-McCord, and Maegan Fairchild.

Notes

1 William Alston, Beyond Justication: Dimensions of Epistemic Evaluation

(Ithaca: Cornell University Press 2005), 58-73.

2 We ofen make a distinction between when a person deserves blame or is

responsible foran action or state and when it will serve some instrumentalpurpose or unction to punish them or change them. For instance, the

rst time a child draws on the dining room walls with crayons we may

be upset and punish them. Most likely, we will do this without making

any sort o evaluation about what kind o child they are. However, i, the

very next day, they go back and draw on the wall again, then we engage

in a different kind o judgment. Although they did something badin

both circumstances, only in the latter would we say they are culpable or

blameworthy or that bad behavior. When we judge in terms o blame ordesert, the ocus o our disapprobation moves rom external eatures (like

an individuals situation) to internal eatures (like a persons character).

3 Alston, Beyond Justication: Dimensions of Epistemic Evaluation, 73-75.

4 Notice that this syllogism does not mention at all epistemic responsibility.

I we can show that we do have somekind o responsibility over our belies

then we still need to show that it is properly considered epistemic. Tis task

will be taken up later in the essay.

5 O course, I can imagine a crazy world or an unlikely set o experiences (inthis world) which, i they occurred, I may orm a belie this belie.

6 Alston, William. Te Deontological Conception o Epistemic

Justication. Philosophical Perspectives2 (1989): 129.



27/28

7 As noted by Feldman (2004), there are very rare circumstances in which

we can control what we believe in virtue o controlling our environment.

For instance, i I want to believe that the light is on I can go over to theswitch and turn it on. However, as noted by Feldman, these examples are to

disparate among the vast array o our belies to constitute an argument or

belie volunteerism.

8 Bennett, Jonathan. Why Is Belie Involuntary?Analysis 50, no. 2 (1990):

87-107.

9 Feldman (2004) gives the example o controlling ones belie about whether

the light is on by simply controlling whether the light is on. He dismisses

the example on the grounds that such inuence over our belies is too rare

to be relevant or arguing that we are responsible or our belies.

10 Feldman, Te Ethics o Belie, 667-695.

11 Kornblith, Hilary. Epistemic Obligation and the Possibility o

Internalism. In Virtue Epistemology: Essays on Epistemic Virtue and

Responsibility, by eds. A. Fairweather and L. Zagzebski, 231-248. Oxord

University Press, 2001.

12 Chrisman, Matthew. Ought to believe.Journal of Philosophy105, no. 7

(2008): 346-370.

13 Feldman, Te Ethics o Belie, 676.14 Using should/ought locutions when a person could not have done

differently (or better) is not restricted to the domain o role oughts.

15 We might also consider the relevant role to be a person trying to discover

the truth a person attempting to deduce a conclusion and so on. I think

that these are wrapped up in the concept o a believer but I do not think

that, should they be different, it matters to the dialectic o this essay.

16 Feldman, Te Ethics o Belie, 676

17 Specically, epistemic justication as reedom rom blame rom believingbut my target is not whether blame is connected at all with justication,

but rather i it is properly applied at all.

18 Feldman, Te Ethics o Belie, 677.

19 Tis example is taken directly rom Chrisman (2008).

20 Feldman may have recourse against some o these objections. First, it is

not clear that we are ever not in the role o an epistemic agent. One might

claim that evaluating orming judgments about our environment (and a

whole host o other topics) is an automatic unction o our psychology. Itcannot be the case that we are actually uninterested in the truth because

being so is a necessary unction o our psychology and thereore always

subject to such a perormance analysis. Because roles are the kinds o



28/28

things that can be ever present, Feldman may also respond to the case o

the unortunate kleptomaniac arguing that she is still in the role o being a

moral citizen (or something akin to that).21 Chrisman, Ought to Believe, 346-370.

22 Sellars, Wilred. Language as Tought and as Communication.

Philosophy and Phenomological Research29, no. 4 (1969): 506-527.

23 See Chrisman (2008), p. 12-14 or a uller description o deriving rules o

action rom rules o criticism.

24 Chisholm, Roderick. Lewis Ethics o Belie. In Te Philosophy of C.I.

Lewis, by ed. P. Schilpp. Salle, IL: Open Court, 1968, 244.

25 We might also weaken it to simply playing a major role. Tis way we can

nd someone partially responsible, partially blameworthy, and so on.

26 For such an attempt, see Foley, Richard. Te Teory of Epistemic

Rationality.Cambridge, MA: Harvard University Press, 1987.

27 Te mere act o the requency o these ailures o cognition is enough o

a challenge. Te added act that this may stem romproperunction o

cognition does not, on its own, mean we should or should not engage in

deliberation.


Defendiendo La Responsabilidad Epistemica

Documents

Transcript of Defendiendo La Responsabilidad Epistemica