Justification in Action

It is officially Spring Break for us at SMU and while my students party on islands so exclusive their names are unknown to me, I’ll be heading to Oklahoma for the Epistemic Goodness Conference. I’m not jealous, by the way. Riggs and Pritchard have put together what promises to be a really exciting conference and I won’t have to worry about sunburn. I thought I’d do a quick post on something that I’ll be discussing next Saturday.

In old paper of Richard Feldman’s (‘Subjective and Objective Justification in Ethics and Epistemology’), he stakes out a view concerning the relationship between the justification of belief and action that I’d like to discuss here. In criticizing some recent work on ‘actionable intelligence’ (e.g., Neta’s forthcoming proposal about when it’s permissible to treat something as a reason for action and a similar view defended by Fantl and McGrath in their forthcoming book), I’ve worked from the assumption that considerations that bear on the justification of action bear on the justification of beliefs about the justification of that action. If because of certain facts, someone cannot justifiably A, I’ve claimed that these very facts prevent someone from justifiably believing that they must A. (And, of course, whatever it is that ensures that your belief that you must A is justified will thereby provide an adequate justification for the intention to act in accordance with this judgment as well as the action itself.) Feldman thinks this isn’t right.

Concerning the justification of action, he writes:

With regard to ethical justification, it is clear that there are cases in which what is moderately subjectively justified differs from what is objectively justified. Such cases occur when a person has reasonable but incorrect beliefs about what action is morally best. One may have reasonable but incorrect beliefs about what the consequences of an action will be or about the values of correctly identified consequences. In such cases, one may have a good reason to believe that an action is best when it actually is not best. Thus, there can be actions that are subjectively justified, but are not objectively justified (1988: 415).

Concerning the justification of belief, however, he insists, “all subjectively justified beliefs are also objectively justified.” This isn’t because he has an unforgiving account of subjectively justified belief on which someone can fail to have subjectively justified beliefs even if they are unaware of that which prevents their beliefs from being justified. Instead, he thinks that whatever it is that ensures that our beliefs are subjectively justified will thereby ensure that our beliefs are objectively justified. Here’s his argument:

Whenever one is [moderately] subjectively justified in believing p, then one is objectively justified in believing that one’s reasons for believing p are good ones. But then the evidence for this [second-order] belief together with the reasons for thinking that those are good reasons constitute an objectively good reason for believing p. Hence … moderate subjective justification implies objective justification (1988: 416).

There’s a bit of jargon here that I should explain. Feldman distinguishes between two kinds of subjective justification. If you believe that you are objectively justified in A-ing, this is a kind of subjective justification but not one that is all that interesting. If you have evidence or good reasons to believe this, this is a more interesting kind of subjective justification and we can say that when you have this your belief is moderately subjectively justified.

If Feldman’s view is correct, I can’t say that because S’s A-ing is beyond justification, S’s believing that she must A is also beyond justification. That’s too bad because I like to say this. Here’s what I don’t get. The argument I’ve sketched here is an argument that there’s no gap between moderate subjective justification and objective justification for belief. If successful, why shouldn’t a parallel argument erase the (alleged) distinction between subjective and objective justification of action?

If you say, as Feldman does, that there’s a gap between objective and subjective justification for action but not belief, then it seems that if you think of justification in deontological terms (i.e., it’s never permissible to act/believe without justification and permissible always to act/believe with justification) you have to deny (at least) one of these two principles:

Detox: If you intend to A and it’s not the case that you shouldn’t so intend, it’s not the case that you shouldn’t A.
Krasia: If you believe you should A and it’s not the case that you shouldn’t so believe, it’s not the case that you shouldn’t intend to act in accordance with this belief.

I like Krasia and Detox. Deny Krasia or Detox and it seems you’ll have to deny one of these three claims:
Pro: If some reason, R, counts in favor of a belief, judgment, intention, or intention, R counts in favor of the beliefs, judgments, intentions, or actions that it rationalizes.
Con: If some reason, R, counts against some action, intention, judgment, or belief, it counts against any intention, judgment, or belief that rationalizes that action, intention, judgment, or belief.
Deon: If S’s A-ing is justified, there is no unmatched or undefeated reason not to A.

I think you shouldn’t deny Krasia or Detox because you shouldn’t deny Pro, Con, or Deon. So, bracketing any qualms I have about Feldman’s argument for identifying objective justification of belief with moderate subjective justification, I’d say that this better be right for action and belief if it is right for either. For what it’s worth, I think there’s a serious problem with Feldman’s argument for the claim that moderate subjective justification entails objective justification. There’s a hole in it. It seems that the passage above contains an argument for the claim that reasons to believe that there are reasons to believe just are among the reasons there are to believe. Suppose that’s right. That might tell us that whenever there is a reason to believe there are reasons to believe there must be some reason to believe, but that’s perfectly consistent with the presence of reasons not to believe. And, if permissibility is a function of both the reasons for and the reasons against (i.e., the pros and the cons), there’s nothing in the passage above that addresses the crucial point. If you think there’s a difference between objective justification for belief and a more subjective notion, that’s probably because you think there can be reasons to refrain from believing that aren’t provided by the subject’s evidence. I see nothing in the passage above that addresses that possibility. That’s the hole.

Now, maybe you think that this doesn’t need to be addressed because you think that (alleged) reasons against aren’t really going to prevent something from being justified unless it is provided by the subject’s evidence, available, accessible, within the agent’s perspective, etc… Okay, that’s a sensible thing to say on Feldman’s behalf. That’s what I’d say about action if I was sympathetic to the view that took the justification of action to be an internalist notion. If I thought that stuff of which the agent is non-culpably ignorant can’t prevent a belief from being justified but only show that there are sometimes unfortunate consequences that follow in the wake of forming a justified belief I’d say the same for action. If I unwittingly bring about some bad set of consequences, that’s bad and unfortunate but not wrong. If it’s not wrongful, it doesn’t threaten the justification of the action. For some reason, Feldman doesn’t think this shows that moderate subjective justification for action suffices for objective justification for action. He rejects that view and so asserts that it is possible for actions to fail to be justified because of faultless wrongs. (He must if he’s to say that moderate justification doesn’t entail objective justification.) Whatever his reason is for dismissing the argument for internalism about the justification of action, that’s the reason I’d steal and offer for saying that there is a gap on the epistemic side as well.

So, are there ways of filling in the hole in Feldman’s argument so we can say that subjective justification of belief is just justification while saying that there’s a real difference between subjective and objective justification in ethics?


Comments

Justification in Action — 11 Comments

  1. I read your JESP paper, or maybe misread it. In spite of pointing out the distinction between moral and epistemic “permissibility”, it seems to me that you’re still confusing the two.

    The only alternative is to say that circumstances can arise in which it is permissible to deliberate from the belief that you ought to Φ even though it is impermissible to Φ. That possibility is difficult to make sense of.

    To reprise your example, suppose our goal is to avoid trouble, and staying produces more trouble than going. However, our Nameless Protagonist happens to justifiably believe the opposite.

    In this case I’d say it is epistemically permissible for NP to deliberate from the belief that they ought to stay (since the belief is justified by hypothesis). It’s morally impermissible* for NP to stay, since that will cause more trouble. But asking whether it’s epistemically permissible for NP to stay strikes me a a nonsensical question, a category error. Unlike “believing” or “deliberating”, “staying” simply is not an epistemic act.

    *(This assumes a definition of “moral permissibility” which says that making well-intentioned, non-negligent mistakes is morally impermissible. If you want to define the term in a way that makes such mistakes morally permissible, then NP’s mistake may be morally permissible. But that’s a tangential issue.)

  2. Jacob,

    This has to be quick, I’m running to the airport in a few minutes. You’re assuming a conception of moral impermissibility it seems we have an argument against if we also assume what you are assuming about epistemic permissibility. If we work from the assumption that you both believe you must A and that it’s not the case that you oughtn’t believe this, by Krasia it seems that you are permitted to intend to A. By Detox, you are permitted to A in accordance with your belief. Now, if you don’t like the conclusion because you think it’s obvious that there can be faultless failures to do what you really should do, my first question is why are you so certain that there couldn’t be similar epistemic failures? Surely it doesn’t follow from the concept of the impermissible that there couldn’t be such failures. You’ve just stated the conditions under which you think such things could occurr–when you are non-culpably ignorant of the considerations in light of which you oughtn’t A.

    Now, you might think that arguments for a view that combines Krasia and Detox invariably involve a confusion between epistemic and some other sort of permissibility such as moral permissibility. Perhaps. Let me note a few things. I see no reason to deny that the fact or consideration that bears on whether to A wouldn’t also bear on whether to judge that you should A. I see no reason to deny that this one consideration couldn’t serve as both an epistemic and moral reason. Suppose you knew all of the relevant facts and knew that if you go there will be trouble but if you stay there will be double. I don’t see what’s wrong with saying that under these conditions the very same facts in light of which you morally ought to leave are the facts in light of which you epistemically oughtn’t believe that you ought to stay. Now, suppose you didn’t know all of the relevant facts. Why prefer a view that says under _these_ conditions but not the previous conditions the facts or reasons that bear on what to believe and what to do are different things?

  3. Hi Clayton.

    Interesting paper & post. Rich and Earl argue that not every internalist theory of justification is a deontological theory of justification, and point to evidentialism as an example. So, if memory is serving me correctly, Rich would deny the antecedent in your argument, namely, “if you think of justification in deontological terms…”. I can imagine him answering, with a wry smile, “…then don’t think in those terms.”

  4. Hey Greg,

    I’ve had people say this on C&F’s behalf, but my response would be twofold. I’d be interested in knowing whether the response is apt because I can’t see what’s wrong with it and refs apparently can’t see what’s right with it. First, I’d say that justification just is a deontological notion. The notion of justified action is clearly a deontological notion and I think it would be weird if the notion of justified belief was not a deontological notion in the sense introduced above. Above, all I said about the deontological character of justification was this:
    (D1) If S’s belief that p is justified, S is permitted to believe p.
    (D2) If S’s belief that p is not justified, S shouldn’t believe p.

    If that’s all it takes for justificaton to be a deontological notion, for C&F to deny this they’d have to point to cases where someone has a justified belief but is epistemically obliged to give up that belief or is permitted to believe without justification. I don’t know where C&F has said that either of these is a genuine possibility but I doubt that they could show that either of these claims could be wrong.

    I’ve had people say that their view is not a deontological view, but I’ve never found anything in the texts that would justify attributing to them the claim that justification is not a deontological notion in the sense intended above. On pp. 61 they wrote, “According to deontological conceptions of epistemic justification, one has a justified belief in a proposition when one deserves praise (or does not deserve blame) for having the belief or when it is one’s duty or obligation to believe that proposition (or believing it violates no duty or obligation)”.

    They reject most of this, rightly. They “deny that internalism depends on a deontological conception of justification” (62) but this is of course consistent with the claim that justification is deontological in the sense introduced above. They go on to offer independent support for internalism that doesn’t rest on deontological assumptions. However, the resultant view looks for all the world like a view on which (D1) and (D2) turn out to be true. Concerning epistemic justification, they endorse the following claim:
    (EJ) Doxastic attitude D toward proposition p is epistemically justified at iff having D towards p fits the evidence S has at t.

    They then say:
    “(EJ) sets an epistemic standard for evaluating doxastic conduct. In any case of a standard of conduct … it is appropriate to speak of ‘requirements’ or ‘obligations’ that the standard imposes. The person who has overwhelming perceptual evidence for the proposition that the lights are on epistemically ought to believe the proposition … We hold the general view that one epistemically ought to have the doxastic attitudes that fit one’s evidence. We think that being epistemically obligatory is equivalent to being epistemically justified (87).” Similar remarks in Feldman’s “Ethics of Belief” suggest that while he’ll try to derive evidentialism from assumptions that don’t include the assumption that justification is a deontological notion he nevertheless ends up with a view on which the justified belief is the permissible belief.

    So, I’ve had referees say that C&F’s view is not a deontological view in ref reports, but I’ve never seen any explanation of the remarks like those on pp. 87 that justifies saying that C&F reject the idea that justification is a deontological notion in the weak way I’m thinking of deontological notions.

  5. Try: Richard Feldman, (1988). “Epistemic Obligations,” Philosophical Perspectives 2, Epistemology, edited by James Tomberlin, Ridgeview Publishing Co., pp. 235-256.

    D1 is pretty weak and you might have some room to hang something like that on Conee and Feldman. However, neither man will go for D2. Both think that there are non-epistemic norms for belief and that sometimes non-epistemic norms should govern. (I agree.) So an agent may have an *epistemic* obligation to believe p in so far as he has good evidence for p, but there may well be other obligations on that arise from other norms (practical, moral, legal) that the agent’s belief should (?!) be governed by.

    I remember a classroom example of Rich’s, the gist of which was to imagine that you have evidence that your wife has betrayed you, but that you also have a moral reason to believe she hasn’t (assuming you do not want to shred your marriage). You need to hang more bells and whistles on the example to avoid dubious claims of voluntary control of belief, avoid creating a defense for blind-faith dogmatism, and show that evidentialism is not drained of its normative juice with this move. But you can get the idea: your belief that your wife is faithful is epistemically unjustified on this new evidence; nevertheless, it is not necessarily the case that you shouldn’t (tout court) believe her to be chaste.

    An example like this one *might* be in Feldman’s Epistemology textbook, 2003.

  6. Hey Greg,

    I should have been clearer about (D2) as I had intended to convey the idea that the permission to believe would be _epistemic_.

    There are passages in Evidentialism where I take it that they deny some things some deontologists have said (e.g., that there’s an interesting connection between J and blamelessness) and I’m wondering if some of the unclarity can be removed in this way. Someone like G.E. Moore thinks that justification is deontological in one sense. He says in Principia Ethica that the only possible way an action can be justified is if it has the thing that he identifies as the thing that makes an action right. However, Moore is also quite clearly not defending a deontological approach to ethics since his account of rightness is characterized in consequentialist terms. When I said that justification is a deontological notion, I only meant the weak thing that someone like Moore could accept and to deny the thing that I think someone like Alston might believe, which is that statements about what you (epistemically) ought/oughtn’t believe could be dispensed with and we could still apply the concept of justification to beliefs.

  7. I think I’m with you, Clayton. Here’s the original argument:

    1. Whenever one is [moderately] subjectively justified in believing p, then one is objectively justified in believing that one’s reasons for believing p are good ones.

    2. But then the evidence for this [second-order] belief together with the reasons for thinking that those are good reasons constitute an objectively good reason for believing p.

    3. Hence … moderate subjective justification [of belief] implies objective justification [of belief].

    The hole in the argument—-the hole that I think you are pointing out—-is in the inference from (2) to (3). As you say, the presence of an objectively good reason for believing p seems to be compatible with the presence of an objectively better reason to not believe p, and hence is entirely compatible with not being objectively justified in believing p.

    The gap is most obviously (but perhaps not plausibly) closed by just adding the premise:

    2.5. The presence of an objectively good reason to believe p is incompatible with the presence of an objectively better reason to not believe p.

    So the analog argument for action would be:

    1*. Whenever one is [moderately] subjectively justified in doing A, then one is objectively justified in believing that one’s reasons for doing A are good ones.

    2*. But then the evidence for this [first-order] belief together with the reasons for thinking that those are good reasons constitute an objectively good reason for doing A.

    2.5*. The presence of an objectively good reason to do A is incompatible with the presence of an objectively better reason to not do A.

    3*. Hence … moderate subjective justification [of action] implies objective justification [of action].

    Which premise would Feldman reject? (2.5*) maybe? Maybe he could say something like: when it comes to belief, one has at most one reason for/against a given belief—namely, an epistemic reason; but when it comes to action, one often has two reasons for/against a given action—namely, one epistemic reason (constituted by what you believe and your reasons for believing it), and one non-epistemic reason (say, a moral reason). He could thus hold that in the case of action, the epistemic reasons often gets “trumped” by a stronger non-epistemic reason, and hence having an objectively good reason to do A is not sufficient for being justified in doing A.

    The key to this kind of move is going to be treating cases in which some people say that there are ‘multiple epistemic reasons’ as actually cases in which there is just a single epistemic reason (which is determined by one’s entire epistemic state), and similarly for cases in which people say that there are ‘multiple non-epistemic reasons’. I have no idea how plausible such a move would be.

  8. Dustin,

    That’s the sort of worry I had in mind. I suppose that your 2.5 could plug the hole. The other idea I had was to say that moderate subjective justification requires an absence of reasons to refrain from believing p/reasons to disbelieve p that aren’t themselves defeated that are provided by the evidence. Then, they could add a premise to the effect that there can only be reasons against, as it were, that are provided by a subject’s evidence. The problem with this additional premise is that looks like using evidentialism to motivate evidentialism and it raises the obvious question as to why someone would hold that in the epistemic case a reason against must be provided by the evidence when this isn’t true of reasons against acting. I couldn’t see any good answer to that question.

    I think that Feldman would be right to reject the practical version of 2.5 in your argument. Interestingly, I think Bernard Williams came close to rejecting the epistemic version of 2.5 when he noted that in the practical case reasons can line up as pros and cons and we don’t have to say that the defeated reasons were spurious. However, if we knowingly judge that we should believe p we should regard the evidence against p as misleading.

    I’m not sure that we’d get the asymmetry, however, even if we thought that there was just ‘one big’ epistemic reason and many potentially conflicting practical or moral reasons. Suppose, for example, you thought that belief was governed by some fundamental norm (e.g., truth, knowledge, evidencet, etc…) but that when it came to action Rossian pluralism is the right view. It could be that when, say, the duty of beneficence defeated the duty of fidelity there was a decisive reason to refrain from A-ing that you have because of the pro tanto reasons there were that spoke in favor/against A-ing but there was nevertheless one ‘big’ epistemic reason not to believe that one should A in the circumstances. That epistemic reason might be that the belief didn’t constitute knowledge, it wasn’t true, it wasn’t the belief supported by the available evidence, etc… So, I’m not even sure that the ‘one big’ epistemic reason view will help. But, I could be mistaken about that as it’s hard to imagine the details of such a view.

  9. Hi, Clayton. I’ve just taken a closer look at Feldman’s paper and thought I’d share my reconstruction of the argument with you. I’d love to hear what you think. Here goes:

    Feldman first argues for the following accounts of the notions in bold.

    S is objectively ethically justified in doing A iff A is best.

    S is subjectively ethically justified in doing A iff S had good reasons to believe that S is objectively ethically justified in doing A.

    S is objectively epistemically justified in believing p iff S has good reasons to believe p.

    S is subjectively epistemically justified in believing p iff S has good reasons to believe that S is objectively epistemically justified in believing p.

    Note that Felman’s accounts of the subjective notions each employs the corresponding objective notion. Hence, we can substitute his accounts of the latter into his accounts of the former:

    S is objectively ethically justified in doing A iff A is best.

    S is subjectively ethically justified in doing A iff S had good reasons to believe that A is best.

    S is objectively epistemically justified in believing p iff S has good reasons to believe p.

    S is subjectively epistemically justified in believing p iff S has good reasons to believe that S has good reasons to believe that p.

    Given the substitutions, his argument is quite straightforward:

    It is possible to have good reasons to believe that A is best without A’s being best. Hence, subjective ethical justification does not imply objective ethical justification.

    However, if you have good reasons to believe that you have good reasons to believe that p, then you have good reasons to believe that p. Hence, subjective epistemic justification implies objective epistemic justification.

    Given his accounts of the notions in bold, these argument seem pretty good to me. I think the real issue is whether he has the right accounts. As you have previously pointed out, given the possibility of having better reasons to, e.g., not believe that P, these might not be the best accounts.

  10. Hey Dustin,

    I still think the argument suffers from the same problem, but you’re right that the problem stems from his account of one of these notions (I’d say it’s with his account of objective epistemic justification). According to Feldman, if S is objectively ethically justified in A-ing, there can only be reasons not to A if there are better reasons to A. But, for reasons he never explains, you can’t draw the same sort of inference when we say that S is objectively epistemically justified in believing. (Well, not on his account of that notion.) So, I guess my beef is with his definition of objective epistemic justification. It should follow from S’s A-ing/believing is (objectively) justified that if there were reasons to refrain there were reasons that defeated the reasons to refrain. That inference holds on his account for the ethical but not the epistemic. If I know it’s not wrong for me to A, I think I can infer that there aren’t better reasons not to A. If I know it’s not wrong for me to believe, I think I can infer that there aren’t better reasons not to believe.

    I’m guessing his response will be that there aren’t good reasons unless the reasons are undefeated and he’ll say that the defeating reasons supervene on your evidence, but why someone would think this is obvious for the epistemic case and not worth considering for the ethical case is beyond me.

Leave a Reply

Your email address will not be published. Required fields are marked *