Sunday, March 30, 2014

The Argument from Intelligibility for Moral Realism

I've previously suggested that the non-contingency of moral realism can help to undermine parsimony-based objections to the view.  I'm now wondering whether it can further help us to provide a positive argument in favour of the view.  Consider:

P1. Moral realism is intelligible -- even Mackie grants that moral claims are “not meaningless but false.”
P2. To be truly meaningful, there must be some (metaphysically) possible property (whether or not it is actually instantiated) that moral claims are about.
C1: So, moral properties are metaphysically possible (and some fundamental moral claims are possibly true).
P3: Fundamental moral claims are non-contingent: they are necessarily true if actually true, and necessarily false if actually false.
C2: So some moral claims are actually true. There are actually instantiated moral properties.

I'm guessing that P2 will be the most controversial premise here.  But it seems at least prima facie plausible to me.  Can you suggest any obvious counterexamples?  Or other considerations that might count either for or against this key premise?

Also, has an argument along these lines been proposed before?  References welcome!


  1. Contradictions are usually classified as false rather than meaningless, and I don't think it's a crazy interpretation of Mackie to say he thinks there's a contradiction between being an objective property of the world and having normative force.

    1. Hmm, so you're thinking Mackie would grant that there really is a possible property of having [categorical] normative force, just not one that can be co-instantiated with certain other properties? But can it be instantiated at all? Presumably not, for an error theorist, but then it seems they're committed to the property being incoherent in itself, rather than just inconsistent with certain other things...

    2. I would take Mackie to argue that "objectively prescriptive: is, itself, like "is a round square".

    3. I am skeptical of this way of interpreting Mackie, since it seems to wreak havoc on most of his main lines of argument, which regularly presuppose truths about counterfactual situations involving moral facts -- i.e., much of his argumentation, if this line of interpretation were correct, would end up depending on sensitive and controversial matters of per impossibile reasoning and conceptual analysis in impossible domains rather than, as Mackie surely intended, with factual analysis. Is there any passage in which Mackie actually argues that moral facts are conceptually contradictory rather than simply factually false?

    4. Brandon: On reconsidering, I think you might be right that I was attributing too strong a position to Mackie. I was thinking the metaphysical argument from queerness had an implication like this, but that is clearly stronger than what Mackie had in mind.

  2. This seems like a similar argument to one put forward by your colleague, Christian Coons, in “How to Prove That Some Acts Are Wrong (Without Using Substantive Moral Premises)."

    Anyway, it seems like a pretty weak argument to me. Isn't it relevantly similar to Plantinga's ontological argument? So, something like:

    P1. The claim that there is a maximally great (i.e., 3-O) being is intelligible.
    P2. To be truly meaningful, there must be some (metaphysically) possible property (whether or not it is actually instantiated) that the claim is about.
    C1: So, the property of being maximally great is metaphysically possible.
    P3: A maximally great being, if it exists, is non-contingent. So, the claim that there is a maximally great being, if true, is necessarily true.
    C2: The claim that there is a maximally great being is actually true.

    1. Ah, thanks -- very silly of me not to have remembered Christian's paper!

      I've two thoughts on how to respond to the "proves too much" objection:

      (1) The principle in P2 should be restricted to simple properties, considered singly, and the individual properties attributable to God (omnipotence, etc.) are indeed individually possible.

      * Problem: is the property of being "maximally great" composite, as this response requires, or is it arguably simple?

      (2) I'm thinking that P3 is false in the God case. Any concrete particular could fail to exist, so the property of being maximally great (if this means: the most great possible) cannot -- contrary to the claims of many theologians -- include non-contingency, or the property of being a necessary existent.

      Could these responses be adapted to undermine the moral argument? It's not clear to me that they can. Helen suggested to me that naturalists might deny that moral properties are simple. But at least the concepts are clearly basic, whereas the concept of "maximal greatness" seems more susceptible to analysis...

    2. We don’t need “maximal greatness” to make the parallel. Moral perfection suffices, so if the concept of moral perfection is basic, then it seems plausible that moral perfection is metaphysically possible, but not instantiated.

      So, an objection to the argument from intelligibility might hold that it might be that moral goodness, immorality, etc., are all possible, but not instantiated, and further that no human being or similar entity can possibly instantiate them, even if other beings possibly do.

      Of course, a defender of the intelligibility argument might hold that moral goodness, moral obligation, etc., are basic concepts, but moral perfection is not.

    3. I think that moral perfection is metaphysically possible, but not instantiated. I'm not seeing the objection here, because I don't see any plausible linking principle (to play the role of P3) to suggest that a morally perfect being, if possible, must actually exist.

    4. Yes, I agree with all of that (i.e., I think it's possible, but I see no good reason to think it's instantiated).
      I misunderstood what you were arguing for, so I thought the instantiation was a problem.

  3. A few responses:

    1. I agree with the "proves too much" objection. The ontological argument is a particularly close cousin, but more generally, conceivability to possibility arguments are controversial (rightly, I think). I've been reading Maudlin's "The Metaphysics Within Physics" recently, and the final chapter there is a nice (I think) attack on the use of conceivability arguments in metaphysics (e.g., it's conceivable therefore possible that there be homogenous spinning disks, so Humeanism is false, or it's conceivable therefore possible that there be gunk) to argue for claims that aren't just about our concepts.

    2. Setting aside worries along the above lines, if this is supposed to be an argument for the sort of realism that's incompatible with expressivism, then expressivists and naturalist reductionists (e.g., Schroeder) will deny 2. And I think rightly so. I find claims like: "it would be wrong to torture people, even if nobody objected to torture" clearly intelligible in a pre-theoretic sense, but I don't find it remotely obvious that metaphysically loaded realist interpretations of those claims--glosses that I'm assured are supposed to be inconsistent with expressivist or realist reductionist interpretations of what's being said--are intelligible. That is, I don't find it at all obvious that (a) such claims can be read in a metaphysically lightweight (e.g., expressivist) or a metaphysically heavyweight (realist) sense, and (b) given that distinction, the heavyweight sense is intelligible.

    I wonder what you think about other cases where robust realism is generally thought to be less plausible. E.g., is it at least intelligible that some clothes are objectively fashionable, in the sense in which moral realists think some acts are objectively right? Or that some food is objectively tasty, regardless of how it tastes to any particular perceiver? Maybe you'll say you just don't find it intelligible that there could be some non-natural property that is identical with objective tastiness or fashionability, though things are different in the case of objective rightness. For my own part, I find these claims equally hard to make sense of/unintelligible when they're supposed to be about non-natural properties, though much easier to make sense of when they're understood as the expressions of various attitudes pro and con.

    1. Interesting! I should check out the Maudlin, as I'm generally sympathetic to conceivability arguments (when done right, at least).

      Expressivism isn't really on the radar here. I'm more imagining a dialectical context where the choice is between non-naturalism and error theory.

      Your challenge from "objective tastiness" is an interesting one! I think that my fittingness view can help here. Tastes are not judgment-sensitive, so it really isn't intelligible that there be reasons for tastiness (or fitting tastes), and if all normativity is to be understood in terms of fitting reasons, then it makes no sense to look for normativity within this domain.

  4. I was going to raise an “egold” objection (i.e., a thing is made [exclusively] of egold (by definition) if it's made [exclusively] of gold and it’s made [exclusively] something with atomic number 80. The property of egoldness is metaphysically impossible, but the claim that there is egoldness is coherent).

    The “simple property” restriction blocks that objection, though. How about the following one?

    1. If by “fundamental moral claims” you mean claims about whether a certain entity with a certain sort of mind described in non-moral terms behaved immorally, or has a moral obligation, or is morally good, etc., or something along those lines, then even if some claims like “For all agents A, if A has such and such mental properties, then A has a moral obligation to X” are true, it does not follow that any actual agents A meet the criteria.
    On the other hand, if you mean something else by “fundamental moral claims”, then I don’t know (it depends on what you mean), but there is the following variant:

    2. If, from the hypothesis that some fundamental moral claims are not contingent and some are necessarily true – and hence, some are actually true -, it followed that some moral properties are actually instantiated, it seems that similarly, it would follow that some moral properties are necessarily instantiated, since the actual world does not appear to play any special role in this argument.
    But it seems plausible that there are possible worlds in which no moral properties are instantiated.
    For example, a world in which there are no minds seems to be such a world. Of course, here I’m accepting that there is no metaphysically necessary morally good or morally perfect being.

  5. Contrary to some of the commenters above, I do not think ontological arguments are close cousins, unless we are using the term to include realism arguments in phil. mathematics, which seem to be much closer kin. e.g.:

    P1. The position that arithmetical claims are true is intelligible.
    P2. To be truly meaningful, there must be some (metaphysically) possible property (whether or not it is actually instantiated) that arithmetical claims are about.
    C1: So, arithmetical properties are metaphysically possible (and some fundamental arithmetical claims are possibly true).
    P3: Fundamental arithmetical claims are non-contingent: they are necessarily true if actually true, and necessarily false if actually false.
    C2: So some arithmetical claims are actually true. There are actually instantiated arithmetical properties.

    1. There is a difficulty, as usual, with regard to what counts as “realism”.

      But that aside (I leave it aside only because I do not know what you define the term), I see a relevant disanalogy between the two arguments unless your math argument is for actual instantiation of some mathematical property, in some being “out there” so to speak (sorry for the non-philosophical term, but I’m not sure how else to put it), beyond an abstract scenario.

      In other words, if your argument is to the conclusion that actually, there are 2 balls, or a prime number of galaxies, at least one instantiation of that sort, then I would see it as analogous. But I do not think that the conclusion follows, because just because necessarily, 2 + 2 = 4, it does not follow that there are 2 or 4 beings, or anything like that (moreover, if it did, one could just give the same argument for all mathematical properties that are possible, so there would be (for example) at least n actual objects for any natural number n, or even for any cardinal number n.).

      On the other hand, I don’t think your argument is an argument in support of that, but if I’m right about the interpretation of your argument, then I would say it’s disanalogous, since the actual instantiation of moral properties involves (as far as I can tell; I may have misunderstood Richard’s argument) that there actually is a morally good agent, or an agent with some moral obligation, etc. Incidentally, if the argument for the instantiation of moral properties were successful, it seems it would work for all possible moral properties, and all possible worlds (since the argument does not seem to rely on one specific property, nor on any actual feature of the world), so there would be at least one agent with moral obligations in every possible world, etc.

      Then again, if I misinterpreted Richard’s argument and he only meant to conclude that some universal moral claims (like “If an agent A has such and such properties, then A has a moral obligation to X”) are true, without making any implication that there are actual agents that meet the conditions (and so, that do have moral obligations), then I find the two arguments more similar, though I’m not entirely sure that the conclusion supports moral realism. I guess if that interpretation of Richard’s argument is correct, here I would have to ask what is meant by “moral realism” in this context.

    2. It's not 'my' math argument; the point was to give a simple example of my point that the argument was plausibly seen as kin to arguments of that sort.

      However, your argument for the disanalogy seems to me to be based on a very strange reading. Where in Richard's argument is any statement of "a morally good agent, or an agent with some moral obligation"? You seem to be importing a set of controversial assumptions about moral claims that the argument does not require.

      I'm not sure at all what you are taking moral realism to be; the phrase has a standard meaning that seems directly relevant here, namely that position which denies (and is denied by) both moral noncognitivism and moral error theory. C1 cuts against moral noncognitivism and C2 against moral error theory, without any appeal to the existence of moral agents, so if the premises are acceptable, we seem comfortably in the realm of moral realism.

    3. By “your” math argument I did not mean anything beyond identifying the argument you gave as an analogy.
      That aside, I don’t see my reading of Richard’s argument as strange, but I guess it might be mistaken. How do you read it?
      As for me, that part of my reading of Richard’s argument is based on the second sentence in the conclusion C2, namely the claim that there are actually instantiated moral properties. As I see it, moral properties are properties like moral goodness, or the property of having a moral obligation, or something along those lines.
      How would you understand the idea that a moral property is instantiated, without involving actual agents that have the property in question (like moral goodness, or having a moral obligation, etc.)?

      As for the meaning of “moral realism”, I do not think there is a single standard meaning (see, for example, the SEP entry on moral antirealism for some of the discussion).

      Also, for a specific example, is David Copp's moral theory a form of moral realism?
      He says it is (so do I, and so does the criterion you mentioned, since he denies both noncognitivism and error theories), but it seems to me Sharon Street would not, and that seems to be due to a difference in how they're using the terms (they do disagree on other issues, but aside from that).

    4. Brandon's right about what I intended: it suffices, for my purposes, for there to be true moral conditionals, regardless of whether the non-moral antecedents are satisfied. Even in a (physically) empty world, the fundamental moral truths remain: pain is bad, whether or not there's any actually around. But I take your point that it's odd to speak of moral properties being instantiated in empty worlds, so I probably need to be more careful in the wording of the argument!

    5. Okay, thanks for the clarification, and sorry I got your argument wrong. So, that objection does not work.

      I would like to ask what you have in mind when you say "moral realism", given the terminological differences (at least roughly, or by means of examples).

  6. Two more potential objections:

    1. P2 does not work for properties involving complex terms, and I do not see why one should accept that it does for properties involving by simple terms.

    2. Another potential objection – but that depends on what you mean by “moral realism” - is that moral claims, while intelligible, may not have the meaning that moral realism holds they have, so P1 would be inapplicable.

    For example, Mackie argues for a moral error theory on the basis of a conceptual analysis of moral terms that I reckon is mistaken – if I thought Mackie’s analysis is correct, I would be an error theorist.

    If moral realism, as you use the expression, does not involve anything like what Mackie means by “objective values”, then your argument escapes the objection below.

    On the other hand, if moral realism, as you use the expression, involves something like what Mackie means by “objective values” (he’s not very precise, but it seems he gives enough information to tell there would be some sort of action-guiding feature that any rational being would have to at least take into consideration, which is I think not consistent, even if meaningful), then I would object to P1 since I would reject the idea that moral claims have the meaning realism – in that sense – holds that they have.

    Also, regardless of what the correct analysis of Mackie’s view is, I would not accept the claim that moral realism so understood is coherent (i.e., both meaningful and not contradictory), even if it’s meaningful. In fact, I tend to think it’s contradictory, since conceptual analysis of the concept of rationality allows in my view for entities that can be rational and far more intelligent than humans but who could not care less about morality, even if they have of course their own set of values they care about.

    That said, if the conclusion about possible entities who are rational but couldn’t care less about morality required something more than conceptual analysis, then perhaps moral realism in that sense would be coherent. However, even then, I would see no reason to think that the concepts involved in moral realism so understood (i.e., including Mackie’s semantic claims about moral terms, or something like it) are simple, so even if I were to accept P2, I would not see why it would apply to moral realism so understood.

    Of course, if you’re not using “moral realism” in that manner, the second objection is inapplicable.


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.)