Monday, March 11, 2013

The Possibility of Moral Realism

I've been wondering whether the non-contingency of moral realism (it's either true in all possible worlds, or false in all possible worlds) undermines parsimony-based objections to it.

We can explain all our observations without appeal to moral facts.  So, the argument goes, that's precisely what we should do -- why multiply entities beyond necessity?  Of course, I'm inclined to start with a broader conception of the philosophical data to be explained, but parsimony is surely worth something, and it isn't difficult to get into a skeptical mindset of wondering whether the theoretical "costs" of positing moral properties shouldn't lead us to instead take anti-realism as the "default" position.

But now notice that in the standard cases (teacups orbiting Pluto, etc.), we don't take the more complex hypothesis to be impossible or anything, but merely less likely to be actual.  So, we might interpret the basic rule of parsimony as follows: given a set of possibilities, distribute your credences (all else equal) in proportion to the simplicity of each hypothesis.

The precise details don't matter here.  What's interesting about this interpretation of parsimony is that it would render the principle entirely inapplicable to the moral realism debate.  For there we are not "given" rival possibilities, either of which might (with greater or lesser likelihood) turn out to be actual.  Rather, the debate is about what's true of the possibilities themselves -- e.g., is any possible instance of gratuitous torture non-instrumentally bad?  Is fine-grained description D a case where moral property M holds?  These questions don't hang on the contingencies of what happens to be actual.  So, if parsimony only speaks to the latter, then it just doesn't have anything to say about the former questions.

Which is, perhaps, all just to say that moral anti-realism is a stronger position than it might at first appear.  It's not just to claim that there happen not to be any moral properties.  Really, the anti-realist is committed to the stronger claim that there couldn't be such properties.  But why should we think that?  Mere parsimony doesn't seem to establish impossibilities.  So it seems that the real case against Moral Realism must rest on other grounds -- arguments that attack, not its likelihood, but its very coherence.  (E.g. Humean worries about "necessary connections" and the supervenience of the normative upon the non-normative; or perhaps confused worries about how non-natural properties could be normative at all.)


  1. I think this is right, but I'm not sure that many arguments against moral realism are best construed as "parsimony" arguments. You don't mention anybody in particular, but somebody like, e.g., Mackie, I suspect is best interpreted as arguing that moral realism is incoherent/doesn't make any sense, rather than that it's a coherent hypothesis that is, because less parsimonious than its rivals, on balance unlikely to be true. Did you have any particular "parsimony" based arguments in mind?

    1. Maybe Gil Harman's arguments about the irrelevance of moral explanations (referenced, e.g., here). But I'm probably more influenced here by informal discussion than the actual literature, so you may well be right that the latter mostly contains stronger objections (as it should!).

    2. I wonder what attitude Gil Harman takes towards the modal status of anti-realism. In particular, I wonder if, given his Quineanism, he's hesitant to say anything about it at all. (I can imagine a certain sort of Quinean refusing to take a stand on the modal status of anti-realism--with maybe a whiff of suspicion that there's somehow something confused about the question--and just offering Harman-style arguments that, as a matter of fact, there are no moral properties.)

  2. That's interesting. It seems as though if that's right, the same argument might apply to parsimony based objections to the existence of God (Russell's teapot).

    1. Yeah, I think that's right insofar as God is stipulated to be a necessary being. (That view seems better opposed on the grounds that, for any proposed concrete particular, it's clearly possible for the object in question to fail to exist.)

      Parsimony objections could still apply to posited contingent deities, though.

  3. In a similar vein, Shafer-Landau argues that parsimony only comes into play for causal explanations. That seems right to me.

    This distinction would allow one to apply considerations of parsimony to God-like explanations which as Alex G pointed out, your account might not allow.

  4. Hi, Richard,

    While I don't find this particular line of arguments against moral realism (but I would actually be interested in a definition of “realism” and what kind of features a metaethical theory has to have in order to count as realist as you use the word “realism”), I don't think that the fact that if realism is true, it's necessarily so, blocks or undermines arguments from parsimony, or generally probabilistic assessments (i.e., I don't think there is anything improper per se in making probabilistic assessment about whether realism is true, or arguments to the conclusion that it's improbable; the arguments might be bad and improper for their own specific reasons, but that's another matter).

    For example, let's consider the case of gold, namely the claim that gold has atomic number 79 (or water is H2O). The claim, if true, is necessarily true, it seems (it holds even in worlds at which gold does not exist). However, there is nothing objectionable about making probabilistic assessments, based on the information available to one, about whether gold has atomic number 79. In fact, it seems to me it's precisely what one should do. The evidence at this point (in the case of gold) is very strong – beyond a reasonable doubt, I'd say -, but one may give an extremely low but not zero epistemic probability to the event that that's not the case (say, a conspiracy among scientists, completely wrong atomic theory, etc.).

    In the case of gold, the a case for the conclusion that gold has atomic number 79 (or more precisely that one should assign probability close to 1 to that conclusion, based on the available evidence), needn't take the form of a parsimony argument, but to the extent to which parsimony arguments are proper, I don't see any good reason to reject parsimony considerations (e.g., perhaps the scientists' conspiracy is less parsimonious, or something like that) just because the hypothesis, if true, it's necessarily true; at any event, even if parsimony objections were not successful for whatever reason, the more general category of probabilistic objections (i.e., arguments for the improbability of realism) wouldn't be affected by the fact that it's either necessary or impossible.


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.)

Note: only a member of this blog may post a comment.