Moral Realism and Epistemology

Over at Pea Soup, Jason Kawall considers Russ Shaffer-Landau’s argument tying realism in morality with what Russ takes to be an appropriate view in epistemology:

I think that intrinsic normativity is ineliminable. To see this, consider the parallels between conditions of epistemic and moral assessment. We say that agents, if they have reason to believe anything at all, have reasons to believe the truth, and to conform their reasoning to truth-preserving schemas, even if believing the truth is not conducive to the goals they set themselves (205).

I believe that there is intrinsic reason to think that two and two are four – the fact itself provides one with reason to believe it. […] The basic idea here is that certain things can be intrinsically normative – reason-giving independently of the value actually attached to them by agents (206).

I favor positions that tie the normativity of ethics and epistemology together, but I doubt this account will work.

First, I don’t think it is true that if we have reason to believe anything at all we have reasons to believe the truth, since often our only reasons to believe are reasons to believe something false. The starkest example of this point is found in the new evil demon problem, but one need not accept that argument to see the point. I also don’t think we automatically have reasons to conform our reasoning to truth-preserving schemas, though we surely have reason to do so if we know that a pattern of reasoning is truth-preserving. The last claim in the first paragraph is the only one I agree with: epistemic reasons to believe are not dependent on agents actually having the goal of getting to the truth and avoiding error. But that point alone won’t get us very far.

The second paragraph fares no better. My reasons for thinking that 2+2=4 are multifarious, but the fact itself doesn’t justify anything. The fact itself is not evidence for the claim in question, and even if it were, evidence needs to be possessed in order to function to justify a belief. To possess the evidence in this case would be to be aware of the fact in question, but it is untoward to say that my awareness of the fact that 2+2=4 justifies my belief that 2+2=4.

This leads to a worry about what moral realism is supposed to be committed to, and I guess that should be addressed by those more competent in ethics. But the talk of intrinsically normative facts, ones that are reason giving independent of any value attached to them by agents, doesn’t transpose well into the epistemic context. The interesting question in epistemology is, I think, the question of whether a relation between e and p can be reason-giving independent of the subjective features of the individual in question. Subjectivists such as Lehrer and Foley think that the right relation has to be 3-place, so that e is reason-giving relative to the basic epistemic commitments or standards of the person in question. Other, such as Ginet and Chisholm, hold that the reason-giving relation is agent-centered as well because it depends not only on the truth of certain epistemic principles but on the a priori knowability of those principles. To avoid agent-centeredness, one has to hold that the relation is a 2-place one that depends only on the truth of the epistemic principles, and not at all on whether one believes those principles, would believe them on reflection, or even if those principles are in principle knowable. If we divide epistemological theories in this way, though, how does the divide provide any useful analogy for those interested in tying the normativity in epistemology to that in ethics? It would be strange to hold that agent-centered views in epistemology make plausible some kind of noncognitivism in ethics, since they don’t incline in that direction in epistemology.


Moral Realism and Epistemology — 27 Comments

  1. Jon, I’m having trouble parsing this passage:

    Subjectivists such as Lehrer and Foley think that the right relation has to be 3-place, so that e is reason-giving relative to the basic epistemic commitments or standards of the person in question. Other, such as Ginet and Chisholm, hold that the reason-giving relation is agent-centered as well because it depends not only on the truth of certain epistemic principles but on the a priori knowability of those principles.

    and I think it’s because I’m unsure what “as well” modifies. Is it that Ginet and Chisholm, like Lehrer and Foley, think that the reason-giving relation is 3-place?

    And (moving on to substance) is that because it depends on whether the relevant epistemic principles are a priori knowable to the agents? So that, if agent A is constituted so as to be able to know principle P a priori, and agent B isn’t, e might give A a reason to believe q but e might not give B a reason to believe q?

  2. Matt–no, the idea is that Ginet and Chisholm have agent-centered restrictions on which epistemic principles can be true. Those that can be true are all knowable a priori. I don’t think they imagine the possibility that the principles are knowable in this way only to some agents, and they might dismiss the idea as impossible. I’m not sure. But if it were possible, they’d have to agree with your point.

  3. Hi Jon,
    A quick question (though it’s not really on the main focus of your post): Would you judge a true belief, even if formed without justification/evidence (or warrant, for that matter) as still in some sense a success or correct? As desirable?

    I’m not sure what to think about the claim that we have reason to believe the truth (as such). I’m still tempted to say that true beliefs are successful to that very extent, and as such we have some (overridable) reason to form such beliefs, even if our behaviour would unreasonable insofar as it lacks evidence or justification.
    Would you hold that while true beliefs are desirable, this doesn’t yet give us any reason to form such beliefs in the absence of evidence/justification?

  4. Hi Jon,
    A quick question (though it’s not really on the main focus of your post): Would you judge a true belief, even if formed without justification/evidence (or warrant, for that matter) as still in some sense a success or correct? As desirable (at least given that you’re forming a belief on a given issue)?

    I’m not sure what to think about the claim that we have reason to believe the truth (as such). I’m still tempted to say that true beliefs are successful to that very extent, and as such we have some (overridable) reason to form such beliefs, even if our behaviour would unreasonable insofar as it lacks evidence or justification.
    Would you hold that while true beliefs are desirable, this doesn’t yet give us any reason to form such beliefs in the absence of evidence/justification?

  5. Hi, Jason, thanks for posting the passage from S-F. I think the answer to your question is “no”, for reasons related to your post. A reason to believe that p is, ipso facto, a reason to believe p is true, so a good theory of reasons to believe simply addresses the question of what reasons there are to believe which claims, and though the goal is to get to the truth and avoid error (perhaps), the substantive principles of the theory will not include the goal itself but will rather formulate principles which constitute means to that goal. After completing that task, it will turn out that some of the things we have reason to believe are false and some are true. So, take the true ones: do we have another reason to believe them, beyond what our theory says? No, since believing the truth is not an appropriate means to the goal of truth; it is simply the goal itself.

    That’s a response to the second question. To the first question, yes I think true belief is valuable, but there’s a difference between what’s valuable from a cognitive point of view and what we have a reason to believe. It’s a good thing to believe the truth; it’s a different thing to have a reason to believe what is true. To use “reason to believe” to cover both of these points is simply to equivocate, or so it seems to me…

  6. Matt Weiner’s comment that apparently I just deleted!

    So is the upshot that Chisholm and Ginet are 2-placers; their restrictions are agent-centered, but the idea is that they depend on whether the principles are knowable a priori (by the agent) and that that is true for all agents if it’s true for any?

    Sorry for being dense but I’m having trouble understanding exactly what you mean by agent-centeredness here; I usually think of “agent-centered” as introducing a place for the agent. (I think I’m importing metaethical uses of “agent-centered.”)

  7. Matt–I don’t think you’re being dense; I’m not quite sure how to think about the Chisholm/Ginet claims about a priori knowability. Maybe because I don’t think I understand a priority, and part of my confusion relates to your question, the question of whether a claim can be a priori for some cognitive beings but not others. The answer is yes for trivial reasons: some being can’t conceive of the claim in question, and others lack the epistemic resources to acquire the knowledge in question. But if you restrict the quantifiers to range over beings with the right conceptual resources who are not skeptically challenged, I think a priority is supposed to involve an all or none claim, that it’s a priori for all or for none.

    If that’s right, then a priority is more a property of the proposition in question than some relation between a mind and a proposition. If it were the latter, then I think the right thing to say about the C/G view would be to posit some third relata. But given the former account, then I think they are 2-placers, with strong restrictions on what epistemic principles can be true.

    But I’m not clear on this either. Do you think this makes any progress?

  8. Jon,
    I think you’ve located an interesting line of reasoning, for moral theorists anyway, if it should turn out that we should have a seamless theory of reasons, across epistemology and ethical theory.
    One point about the quote: I wonder whether anything hangs on that pesky distinction between having a reason and there being a reason. It seems less jarring to say that if reasons exist for me to believe anything, they exist for me to believe the truth, than to say that if I possess reasons reasons to believe anything, I possess them for believing the truth. In both quotes, there is a movement back and forth between these. So, for instance, in the second quote, ‘the fact itself provides me with reasons’. Well, that seems false. But what if what he meant was just ‘the fact itself IS a reason–whether provided to me or not’?
    Just a thought to help rescue some of the position Russ is stating…

  9. Robert, that does help some, since in terms of having reasons the position is not plausible at all. But I still think there is equivocation going on to say that there is a reason to be believe the truth. Believing the truth is valuable, so if you believe the truth, you’ve succeeded in raising the total goodness score in the universe. But to say that there is a reason to believe a claim amounts, for evidentialists like me, to the claim that there is evidence in favor of that claim (whether one has that evidence or not). The two uses of “reason” are not the same, and only in the second use are we getting a view on the normativity of reasons for belief.

  10. I think I see.
    Regarding the general question regarding the reason-givingness of fact in ethical theory: I take it Russ is trying to rescustate the now unpopular Moorean nonnaturalistic realism. Moral facts are sui generis in having built-in ‘tobedoneness’. Your incredulity is widely shared, I think, but moral realism of the colder, more northern (though just barely) variety found in Cornell is still available. Still, I should be surprised if it could be done in so quickly by pointing out that facts about which claims are made aren’t evidence for those claims.
    The parallel in epistemology would be something like facts that have ‘tobebelievedness’ built right into them. Are there such? Self-evident truths seem like this. Just taking the phenomenology, the fact that 2+2=4 does seem to have that quality, doesn’t? Say, unlike the fact that there are more beetles than worms.

  11. Robert, I think if the issue is whether there are self-evident propositions, not many will side with the moral realist here. At least, not if self-evidence is supposed to be a property of a proposition. The problem for appeals to self-evidence is that it is claimed to be a property of both p and ~p by different folk. Now, of course, one may be wrong and the other right, so that, say, p really is self-evident even though you think it’s not (because you think ~p is self-evident). But self-evidence, by the very term, ought to be transparent to all cognitive beings capable of conceiving of the proposition, I would think–otherwise, I don’t know what “self-evident” is supposed to mean.

    Since the transparency claim is suspect, let’s assume there aren’t any self-evident propositions, and so no propositions with the property of “tobebelievedness”. Would that somehow count in favor of the view that there are no epistemic facts? That it is not an objective matter whether or not a person’s belief is justified? I doubt many find such an inference plausible in epistemology (and then there is the matter of the inference itself, but I’ll let that be for now).

    So, if this is what moral realism is about, it strikes me that we can’t get a unified picture of normativity in ethics and epistemology. But it also strikes me that it shouldn’t be this easy to pull them apart.

  12. Robert,

    … moral realism of the colder, more northern (though just barely) variety found in Cornell is still available.

    Ithaca 42° 27′ N

    Madison 43° 8′ N

    Cambridge 52° 20′

    Right around 43° the air thins out enough for simple non-natural properties to become, uh, visible.

    (I am hoping the entity symbols for the degree symbol work — they look fine in the Preview.)

  13. Jon, I think that helps–though I clearly need to look at the C/G stuff myself! I would guess that most folks see a priori knowability as a property of a proposition; that property determines whether we can know it a priori (rather than facts about us determining a-priori-knowability).

  14. Well, Jon, ok, I give up on trying to make sense of Russ’s position as stated. Try a different view, then. Take ‘Being respectful is obligatory.’
    1. The proposition states that a certain trait or disposition possess a property.
    2. The fact that it has this property, obligatoriness, seems to lay down a norm for us regarding the trait.
    Set aside whether this normativity is an ‘intrinsic’ property, or is ‘intrinsic to’ the property of obligatoriness. I see the pull toward thinking that it is its possessing the property of obligatoriness itself that requires somethig of us.
    Now suppose ‘(pâ��q) & p’ is true. I do have the intuition that the bare fact that this possesses truth requires something of us as well, namely, accepting ‘q’. Call it what you want, but there’s something in the neighborhood that looks parallel to the ethical case, and it appears as if it’s ‘in’ the facts, as it were.

  15. Robert, Here’s what I’d say about the two examples. In the first case, it’s the possession of a property by a trait. In the second, it’s the possession of a property by a proposition or sentence. In the first case, the property is itself a normative one; in the second case, not so. In the second case, if normativity is to be found, it will have to come from the normativity involved when you belief p implies q and p–in that case, you ought to believe q (ignoring the issue about whether there are doxastic obligations…). The mere truth of the conjunction doesn’t appear to me to require accepting q, though it does imply the truth of q. I’m repeating myself here, but only accepting the conjunction could requiring accepting q, and that’s different from the moral example you give. No?

  16. It is my imperfect way of putting this I’m sure, and I realize you’re just repeating yourself (and me myself). But you do accept my first point, right? It is the fact that the trait possesses the property of obligatoriness that seeems normative. If you do accept that, then you agree that these moral facts, in this sense at least, seem to the untutored view normative. If you accept it in that case, then you agree, that is, with the Mackie point (or what I take to be his point.) This is perhaps a sui generis relation between moral fact and action. What’s left is whether there is a parallel relationship of nonmoral fact to belief. My thought was that facts about validity would be something like this. So let me back off of saying that it is the fact that p arrow q and p that requires *accepting* q; it is the fact that ‘p arrow q and p, therefore q’ that seems to require something of me. It seems to law down a rule for the direction of my mind.
    Perhaps you will say, ‘No, it is *accepting* that this hold that requires me to direct my mind in some way. Its being the case that p arrow q and p therefore q doesn’t require anything of me.’ But don’t you see the pull to see it my way too?

  17. I guess I think that if it is a fact that a certain character trait is obligatory, there is something normative here: namely, the property in question. And suppose it is a fact that p implies q and that p is true. I don’t see any normativity here at all. People say logic is normative, and some say truth itself is, but I don’t think they mean by such claims what I think of when I think of normativity.

    I would have thought that the analogy ought to be with epistemic facts in any case. Something like “believing what you are told by those you trust is rational”. Here there is normativity, and it is plausible to think that such facts give us rules for belief formation.

    So I think there will be a good analogy here; I just don’t see what it has to do with realism/antirealism, cognitivism/noncognitivism, attitudinalism-expressivism, etc.

  18. But Jon, if you think that such facts give us rules for belief formation, then you just are saying that certain facts are of themselves normative. That is what the Moorean nonnaturalist thinks in ethics too. I was merely using the example of the fact that (p arrow q, p therefore q), to be a fact no different (though more formal) from ‘Believing whom you trust is rational’ or whatever other facts about rational belief formation exist. It seems that this does have centrally to do with realism debates.

  19. Robert, do you agree with the following? “The facts are normative” should be understood as claiming that (i) there are some true propositions that (ii) have normative properties as constituents. If we accept that account and think it’s a distinctive commitment of realism in either the moral or epistemic realm, then the opposing camp has to hold that there are no truths with normative properties as constituents. Since the property of having a rational belief is a normative one, that would imply that only realists think that there are any rational beliefs.

  20. Yes, I that is roughly what the realist line is. I hope that doesn’t make me look more foolish than I already look. I take it moral realists frequently conclude exactly what you do in ethics: “You anti-realists don’t think that there are any right actions…anything goes on your view!” I would expect them to say the same in epistemology: “You anti-realists think we can believe anything we want! No facts about what it’s rational to believe? Really?!” Anti-realists would have to deny your (i) or (ii). In the end, talking plain talk, there are no rational beliefs because there is no such property as ‘rational’ for non-realists. (No flying reindeer either.) Maybe they’d say, with Gibbard, that our concept is of something intrinsically normative here, though the facts aren’t.
    It seems to me that there is something very intuitive about seeing certain facts as being, of themselves, directive. I don’t believe it. But it sure is mighty tempting, in both ethics and epistemology.

  21. I guess the first thing I’d say is that the anti-realist shouldn’t have to say this. But I’d also want to say that if this is the way the analogy goes, then one can be a moral anti-realist only be endorsing that there are no rational beliefs either. It shouldn’t be that easy to undermine the view, should it?

    If I understand the Gibbard reference correctly, such a position will have to hold that there are intrinsically normative concepts but no intrinsically normative properties? That sounds, to use a technical term, metaphysically messed up…

    Oh, one more thing. It shouldn’t turn out that only realists get to talk about truth, should it? Maybe if truth is substantive in some important way, but maybe it’s not.

  22. Well, I agree that the anti-realist shouldn’t have to say it. But it is what realists say they should have to say. And some anti-realists embrace it. I suppose some might want to hold a non-normative view of rationality (some readings of Hume suggest this was his view, but I wouldn’t want to set off a war on that). The antirealist views I am familiar with wouldn’t balk at being required to endorse what you say, as long as it is properly understood. (True, beliefs cannot be RATIONAL or IRRATIONAL. But they can be rational or irrational. That sort of thing. You won’t like that.)
    I’m going on memory, but I think in the Gibbard view, there are only natural properties, but he divides properties from concepts and says that in some cases we have both natural and nonnatural concepts for these proeprties. Whether its messed up or not, well, not someting I’m qualified to venture a view on I think.

  23. Robert, you’re right–my likes and dislikes often outrun my arguments, too. On the Gibbard view, I grant the need for a triadic conception of intentional attitudes to be able to handle the problems involved in Frege’s puzzle, as well as the Kripke problem. I should think about this more, but properties and propositions are relevant to truth-bearers, so whatever work is left for concepts will not play a role in ascertaining truth-values. And then the realist wins on how to characterize antirealism (and going to lower case won’t help because that presupposes that we can sneak something into the account of truth besides properties and propositions).

    This is ridiculously programmatic…

  24. Robert, Jon,

    For what its worth, here is (my version) of the anti-realist view about normativity as it pertains to epistemology: beliefs can be true, false, effective, and ineffective. The most basic classification is between effective/ineffective. Sometimes, effective is coextensive with true and ineffective with false, sometimes the opposite. With respect to the issue at hand, the anti-realist claims that there is simply no determinate content to the ascriptions ‘rational’ and ‘irrational’, if these are meant to convey something other than belief-fixing processes (including inferences) that result/preserve truth and those that don’t. Furthermore, such scepticism extends to the notion of knowledge, another concept the anti-realist finds without determinate content.

    Hume, indeed, comes close to espousing this view, and arguably is committed to it. Nietzsche rejects normativity as well. More recently, this is very close to Quine. Notice, with Quine, how he speaks of ‘normative epistemology’ as being a technological endeavor. What he is saying is that normative principles are simply those that function as true descriptions of processes that lead to truth. To speak of rationality, normativity, correctness, or what have you, is just a fancy way of making the same point.

    Similar points apply in ethics: there is no correct view about what is good or bad. But to accept value without normativity is not espouse an ‘anything goes’ view. Non-normativity is not relativism. What is good or bad is what one endorses or rejects, but one can self-consciously endorse and reject things without the accompanying thought that things ought and ought not be a particular way, where this thought is understood to be distinct and determinate.

    My suspicion is that anti-realism in ethics and epistemology is so quicly dismissed because of the failure (on the part of defenders as well as critics) to distinguish the rejection of normativity from relativism. Clearly, some beliefs are better than others, and some belief-fixing processes are better than others, i.e. more effective for both cognitive and global aims. But better and worse do not imply correct and incorrect. If the anti-realist can make that distinction a plausible one, then I think the view itself gains in plausibility.

    Sorry this was so long!

  25. Joe, I don’t think I understand the remark about effective belief being sometimes ‘coextensive with truth’ and ‘sometimes coextensive with falsehood’. Do you mean simply that sometimes effective beliefs are true and sometimes false?

    Also, what kind of argument can you give for claim that ‘rational’ has no determinate content outside the context you specify?

    I wonder about the coherence of the following conjunction: there is no correct view about what is good or bad & what is good or bad is what one endorses or rejects.

  26. Jon, to your questions. Firstly, your supposition is correct, I do only mean to say that sometimes an effective belief is a true one, sometimes not.

    Secondly, I cannot give a sound argument (not even sure about a ‘mereley’ valid one) for restricting the terms ‘rational’ and ‘irrational’ to the contexts I urge. But I am convinced (!) that there can be no such argument that would establish a determinate content for those terms in any other context. This needs much more elaboration than is appropriate here, but I think that we are concerned with establishing ‘rational’ or ‘justified’ methods of belief fixation for the sole purpose of increasing their reliability. We are looking for methods that are more likely to result in true beliefs. If there is some other purpose that a ‘rational’ inference of belief-fixing method would serve, then my claim would be contradicted. But I am unware of a determinate account of any such alternative purpose.

    Thirdly, if one assumes that value judgments arise from a perspective of ‘reflective endorsement’, then it seems that one is committed to the view that what one a) can coherently reflect upon and b) can either endorse or reject, is wholly a function of one’s individual nature. Put another way, one cannot at any given time endorse/reject anything other than what one endorses/rejects (one can, of course, endorse/reject something different at some other time). Now this strikes me as a decidedly ‘anti-realist’ account of value, and hence the notion of ‘truth’ has no place. But, and this is my basic point about normativity, the notion of ‘correctness,’ assuming this means something over and above ‘truth,’ has no place either.

    What lurks behind all this is, I think, the following: our concepts of ‘truth’ and ‘correctness’ (what I am understanding as the fundamental normative notion, so if we disagree about this then this is what needs clarification) are not coherently distinguishable. Since I don’t think there is truth in ethics, I don’t think ethics is normative. Since I don’t think that its true that beliefs need to be true in order to be effective, I don’t think epistemology is normative either.

    Not sure if that makes it anymore plausible.

Leave a Reply

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