Monday, August 15, 2011

Knowing What Matters (draft)

A pdf draft of my paper, 'Knowing What Matters' (forthcoming in Peter Singer (ed.), Does Anything Really Matter? Responses to Parfit) is now online.

Parfi t's On What Matters offers a rousing defence of non-naturalist normative realism against pressing metaphysical and epistemological objections. He addresses skeptical arguments based on (i) the causal origins of our normative beliefs, and (ii) the appearance of pervasive moral disagreement. In both cases, he concedes the fi rst step to the skeptic, but draws a subsequent distinction with which he hopes to stem the skeptic's advance. I argue, however, that these distinctions cannot bear the weight that Parfi t places on them. A successful moral epistemology must take a harder line with the skeptic, insisting that moral knowledge can be had by those with the right kind of psychology -- no matter the evolutionary origin of the psychology, nor whether we can demonstrate its reliability over the alternatives.

Along the way, I also (1) argue against Street's "normative lottery" analogy; (2) argue that her constructivist view is self-defeating; (3) introduce an "internalist" version of reliabilism; and (4) offer an analysis of when it makes a difference whether some possible disagreement is actual.

I still have a couple of months to put on the finishing touches, so any suggestions would be most welcome!


  1. You might want to put in a footnote acknowledging that the existence of an a priori probability distribution over possible worlds is pretty controversial. Not everybody's OK with a priori anything, and among those who are, many would restrict a priori knowledge/justification to necessary truths, or necessary truths + special Kripke-style contingent a priori truths. For instance, Jim Pryor's dogmatism and Jonathan Vogel's explanationism can both be understood as ways of trying to avoid appeals to a priori probability distributions over possible worlds. Orthodox decision theorists will also typically reject a priori probability distributions, taking them to require something like the sort of purely logical grounding that Carnap tried and failed to provide. And it's often a big part of reliabilist/naturalist epistemology (which you cast your view as a version of) to reject a priori knowledge, (see Kitcher's "The Naturalists Return"--he describes the rejection of the a priori as a central tenet of naturalist epistemology) or at best understand it in a way that wouldn't support the use you're making of it (e.g., see Louise Antony's reliabilist characterization of the a priori).

    I'm not saying you should change any content, but the way you put things now (with reference to "the" a priori probability distribution over possible worlds) makes it seem as if you don't realize that you're helping yourself to a controversial piece of machinery.

  2. To clarify, I know you didn't cast your view as a version of naturalism, just as a version of reliabilism. I just meant that, given that many reliabilists are naturalists (and are reliabilists in part because it fits in with their naturalism), and so are hostile to the a priori (or understand it in weaker forms), it's a bit misleading to treat your view as a version of reliabilism. Those who are sympathetic to reliabilism in part because it doesn't require appeal to the a priori will not be sympathetic to your view.

  3. Fair enough. (I should clarify that I'm not particularly concerned to associate myself with the existing reliabilist programme; it just seemed an appropriate label insofar as my view appeals to the reliability of one's belief-forming mechanisms!)

  4. Paper updated, following additional (emailed) comments from Alex and Barry, besides Daniel's above -- thanks all!

    (Useful things, blogs!)

  5. Another thought--I don't know Street's view well, but I wonder if it is most charitably put as the view that all "eligible" normative judgments are true. As you point out, that view straightforwardly leads to contradiction (or maybe relativism) in light of the possibility of different agents making conflicting eligible judgments.

    Given what she says about ideally coherent Caligula cases, I wonder if the view isn't better put as something like the following. An agent S's eligible judgments to the effect that S has such-and-such reasons are true. So facts about Caligula's reasons are constituted by facts about his eligible judgments, facts about my reasons are constituted by facts about my eligible judgments, and so on.

    That view seems to me to avoid self-defeat problems, and to better fit what Street says about particular examples, whatever other flaws it may have. (Though take this with a grain of salt, since as I mentioned, I don't know her work well.)

  6. Daniel - I was actually thinking of a single agent making counter-constructivist judgments.

    Suppose I can eligibly judge the following: I always have reason to avoid pain, even in (counterfactual) circumstances in which I eligibly judge that I have no reason to avoid pain.

    The italicized judgment is about my reasons, and so if eligibly made by me, constructivism implies it is true. That is, it's true that I always have reason to avoid pain. But constructivism also implies that if I eligibly judge that I have no reason to avoid pain, then in that case I have no reason to avoid pain. Contradiction.

    That's why Street is committed (as she admits) to denying that the realist's normative judgments about counterfactual scenarios are eligible.

  7. Perhaps we could be constructivists only about eligible judgments concerning the agent's actual circumstances. But then what grounds the normative truths about what reasons we would have in counterfactual circumstances?

    I think Street needs her constructivism to be more all-encompassing than that. She needs the truth of constructivism itself to be a constructed truth that we're all committed to, for otherwise the skeptical worries are just pushed back a step: Street would face the a "striking coincidence" that of all the possible metaethical views, just one of which is mind-independently true, she happened to believe the right one!

  8. The view I had in mind would be something like Brandt's view, or Michael Smith's, but without the assumption that there would have to be convergence among ideally rational agents in order for there to be reasons.

    That is, according to the view I was thinking Street held, an agent S in circumstances C (actual or counterfactual) has (in circumstances C) whatever reasons she would take herself to have were she to go through some sort of reflective equilibrium process. As I understand Smith, this is similar to the view he defends in The Moral Problem, but he takes it that the reasons we'd take ourselves to have (well, really our desires, but I'm fudging) would all converge were we to reach reflective equilibrium, and that they'd have to converge for there to be any genuine normative reasons. The view I was suggesting for Street would differ both by denying that such convergence would occur, but also denying that such convergence would be necessary for there to be genuine normative reasons.

    You're right that this view does allow that our eligible judgments about what reasons we'd have in counterfactual circumstances might be wrong--namely, they'd be wrong if we actually eligibly judge that we ought to do one thing in counterfactual circumstances C, while were we in C, we'd eligibly judge that we ought to do something else.

    But I don't see how this leads to the problem in your second paragraph though. That's not to say I don't see how skeptical worries might be raised concerning metaethics too (though a particularly Darwinian story seems less applicable here), or how a regress might arise. I just don't see how taking the route I was suggesting for her makes those issues more or less pressing.

  9. Sorry, I included a "but" and a "though" where only one was necessary (and both was strictly overkill).

  10. The worry is a kind of tu quoque: if constructivism is mind-independently true, then it ends up in a similar epistemic position to realism (albeit at the metaethical rather than normative level) -- a position that Street considers intolerable. We might reasonably conclude that it isn't so intolerable after all, in which case we would have a (restricted) constructivist view that avoids self-defeat, but at the cost of undermining Street's argument against realism.

  11. I understand the tu quoque, what I don't understand (and I'm not saying I think it's not true, I just don't see the route) is why the version of constructivism I suggested on Street's behalf is more vulnerable to the tu quoque than the (inconsistent) version we started with.

  12. Ah, I see. Thanks for the clarification.

    Street's concern about mind-independent philosophical truths is that there seems no explanatory connection between what's true and what we end up believing. (It's not like the truths cause us to believe them, or anything.) So it looks like it's just a matter of luck whether we end up believing the right things.

    Her universal constructivism is meant to avoid this problem because it no longer claims that (any) philosophical truths are mind-independent. Her constructivism, for example, is true because it's (putatively) what we would all conclude at the end of inquiry. This creates an explanatory link between our beliefs and the truth of the matter, such that it no longer seems a matter of wild luck for constructivists to have the correct metaethical beliefs: one just needs to work out the implications of one's core commitments, and then you're guaranteed to be right.

  13. Thanks, that helps. I'm still not sure why something like the tu quoque doesn't arise for that version of constructivism, though.

    So suppose Street holds that all ethical/metaethical truths are mind-dependently true. Now the question arises whether that claim (the claim that all such truths are mind-dependently true) is mind-dependently true, or mind-independently true. And if it's mind-dependently true, what about (one level up) the claim that it's mind-dependently true?

    It seems to me like she can either say there's a mind-independent truth somewhere, or face an infinite regress of distinct mind-dependent truths, of the following sort. First-order ethical truths are mind-dependently true, and so is the claim that first-order ethical truths are mind-dependently true, and so is the claim that the claim that first-order ethical truths are mind-dependently true is true...

    Maybe it's not obviously a vicious regress, but it certainly seems odd. Analogous views in epistemology (Klein's infitism) are typically not though to provide satisfactory responses to skeptical problems in epistemology.

  14. I think Street wants to hold that it's mind-dependent truths "all the way down". I agree that there's something a little fishy about this regress, though it is at least a different problem from what she attributes to the realist.

    (It is interesting to compare it to infinitism. I feel like each earlier link in the chain is not so "dependent" on the later links, in Street's case, as it is in case of infinitism. So maybe it isn't so bad. But I'm too lazy to flesh out this intuitive thought in any more rigorous way right now...)

  15. I don't think you really get to the heart of the lottery analogy. The intuition that you shouldn't be confident about the lottery doesn't seem to depend on the lottery being fixed by a chancy process. It wouldn't be any more reasonable to think that you had the winning ticket if you thought the universe was strictly deterministic, or if the winning number was the second billion digits of pi.

    That said, I think you are right that it is not a matter of internal incoherence. The irrationality of such probability assignments doesn't arise from the very meaning of rationality, or probability, or any other concepts. Nevertheless, this doesn't commit us to normative realism. We might, say, be expressivists about epistemic norms.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)