Tuesday, January 29, 2013

Metaphysical Skepticism a la Kriegel

I'm totally biased. I admit it. But I love this new paper by Uriah Kriegel.

Taking the metaphysical question of the ontology of objects as his test case, Kriegel argues that metaphysical disputes often fail to admit of resolution. They fail to admit of resolution, Kriegel argues, not because such disputes lack substance or are merely terminological, but for the more interestingly skeptical reason that although there may be a fact of the matter which metaphysical position is correct, we have no good means of discovering that fact. I have argued for a similarly skeptical position about the "mind-body problem", that is, the question of the relationship between mind and matter. (Hence my pro-Kriegel bias.) But Kriegel develops his argument in some respects more systematically.

Consider a set of four fundamental particles joined together in a tetrahedron. How many objects are there really? A conservative ontological view might say that really there are just four objects and no more: the four fundamental particles "arranged tetrahedronwise". A liberal ontological view might say that really there are fifteen objects: each of the four particles, plus each of the six possible pairings of the particles, plus each of the four possible triplets, plus the whole tetrahedron. An intermediate ("common sense"?) view might hold that the individual particles are each real objects, and so is the tetrahedron, but not the pairs and triplets, for a total of five objects.

Now who is right? Kriegel envisions three possible approaches to determining where the truth lies: empirical testing, appeal to "intuition", and appeal to theoretical virtues like simplicity and parsimony. However, none of these approaches seems promising.

Contra empirical testing: There is, it seems, no empirical fact on which the conservative and liberal would disagree. It's not like we could bombard the tetrahedron with radiation of some sort and the conservative would predict one thing, the liberal another.

Contra appeal to intuition: "Intuition" is a problematic concept in metaphilosophy. But maybe it means something like common sense or coherence with pre-theoretical opinion. Intuition in this sense might favor the five-object answer (the four particles plus the whole), but that's not entirely clear. However, Kriegel argues, hewing to intuition means doing only what P.F. Strawson calls "descriptive metaphysics" -- metaphysics that aims merely to reveal the structure of reality implicit in people's (possibly misguided) conceptual schemes. If we're aiming to discover real metaphysical truths, and not merely what's already implicit in ordinary opinion, we are doing instead what Kriegel and Strawson call "revisionary metaphysics"; and although descriptive metaphysics is beholden to intuition, revisionary metaphysics is not.

Contra appeal to theoretical virtue: Theoretical virtues like simplicity and parsimony might be pragmatic or aesthetic virtues but, Kriegel argues, there seems to be no reason to regard them as truth-conducive in metaphysics. Is there reason to think that the world is simple, and thus that a simple metaphysical theory is more likely to be true than a complex one? Is there reason to think the world contains few entities, and thus that a parsimonious metaphysical theory that posits fewer entities than its rivals is more likely to be true? Kriegel suggests not.

As I said, I love this paper and I'm sympathetic with its conclusion. But I'm a philosopher, so I can't possibly agree entirely with another philosopher working in the same area. That's just not in our nature! So let me issue two complaints.

First: I'm not sure object ontology is the best test case for Kriegel's view. Maybe there's a real fact of the matter whether there are four, five, or fifteen objects in our tetrahedron, but it's not obviously so. It seems like a good case for reinterpretation as a terminological dispute, if any case is. If Kriegel wants to make a general case for metaphysical skepticism, he might do better to choose a dispute that's less tempting to dismiss as terminological, such as the dispute about whether there are immaterial substances or properties. (In fairness, I happen to know he is working on this now.)

Second: It seems to me that Kriegel commits to more strongly negative opinions about the epistemic value of intuition and theoretical virtue than is necessary or plausible. It sounds, in places, like Kriegel is committing to saying that there's no epistemic value, for metaphysics, in harmony with pre-theoretical intuition or in theoretical virtues like simplicity. These are strong claims! We can admit, more plausibly I think, that intuitiveness, simplicity, explanatory power, etc., have some epistemic value while still holding that the kinds of positions that philosophers regard as real metaphysical contenders tend not to admit of decisive resolution by appeal to such factors. Real metaphysical contenders will conflict with some intuitions and harmonize with others, and they will tend to have different competing sets of theoretical virtues and vices, engendering debates it's difficult to see any hope of resolving with our current tools and capacities.

Consider this: It seems very unlikely that the metaphysical truth is that there are exactly fourteen objects in our tetrahedron, to wit, all of the combinations admitted by the liberal view except for one of the four triplet combinations. Such a view seems arbitrary, asymmetric, unsimple, and intuitively bizarre, compared to the more standard options. If you agree, then you should accept, contra a strong reading of Kriegel's argument, that those sorts of theoretical considerations can take us some distance, epistemically. It's just that they aren't likely to take us all the way.


Daniel said...

Karen Bennett defends a similar position concerning certain metaphysical debates in "Composition, Colocation, and Metaontology," available on her website. Similar in that she doesn't think the debates are merely verbal or anything like that, but does think that we are not (and probably will never be) in a position to know which side is right. Moreover, she doesn't, as far as I remember, endorse strong negative (or positive) claims about the epistemic value of agreement with intuition, or simplicity, of the sort that you object to in Kriegel.

Nic M. said...

I'm not sure the simplicity consideration gets us very far. Parsimony arguments are not the same as 'intuition': the implausible, arbitrary view is lacking because it introduces complexity without explanation, and all else being equal, more complex explanations are more likely to be false, having more points of potential failure. (I'm also of the view that simplicity is not an amorphous 'virtue' but an property in good standing, but that's orthogonal for now.) So while you could make a case that the standard theoretical virtues can help the metaphysician, 'intuition' is still a lost cause on this example.

Not because intuitions are always bad. I can imagine plenty of cases where expertise and training can yield reliable intuitions. Just that in these metaphysical cases I have no standard of reliability.* I have no idea how to 'check' its deliverances, even indirectly. At best we can hope that our intuitions match up with the thin data of theoretical virtue,** but I think Kriegel is right that no empirical information speaks to many of these questions; and lacking this, his scepticism seems certainly warranted.

I'd rather an argument showing that many of these disputes are terminological, of course, but scepticism will do in a pinch: there is perhaps a difference in temperament, in Nietzsche's sense, between those who feel such questions are substantive and those who do not.

(*Sometimes friends of intuitions make analogies to perception, or to the intuitions of e.g. radiologists who, after years of experience, have reliable spontaneous judgments they find hard to explain in words. But in both cases I have a standard of correctness: when visual perception fails, as in an optical illusion, I can both discover the failure and offer an explanation of it. Same with radiology. One can easily determine whether an expert radiologist is having an 'off' day. How do I know when the expert philosophical intuiter is not getting it right? The whole process is completely private, and, like Wittgenstein, I don't put much stock in data that ain't public. Weinberg has talked about this kind of thing.)

(**And sometimes I wonder if intuition really is just our minds making a kind of judgment about theoretical virtue, which seems opaque until we unpack it: thus there are no intuitions, just under-analyzed arguments. Cappelen doesn't say exactly this but it's the same basic idea.)

Eric Schwitzgebel said...

@ Daniel: Thanks for the pointer! Yes, I know about that paper of Bennett's (and Kriegel also cites it). My memory (though maybe I should refresh myself) is that her skepticism is considerably more modest and limited than Kriegel's (insofar as he endorses the conclusions of his arguments; he does put in strong caveats near the end) or mine.

@ Nic: I don't mean here to be committing to specifics regarding whether it's intuition, simplicity, explanatory failure, or something else that makes the 14-object ontology unlikely to be true. I think I am pretty close to agreeing with you about intuition, since I don't think it counts for a whole lot. But there are better and worse cases, I suspect, and when all the sources are problematic, even things that don't count for a whole lot can't be ignored.

uriah said...

Hi Eric,

Thanks for the enthusiasm, and for your thoughts on this issue. A couple of comments back:

I don't wish to deny that there is an *intuition* that one ought to believe the 4-, 5-, and 15-object ontologies more than the 14-object ontology. To me, that's just the same intuition as that realism about the external world is more plausible than Berkeleyan idealism (even on the assumption, which I realize you contest, that the two are empirically indistinguishable). But the challenge is how to *show* that this intuition should be endorsed. What seems to tell against the 14-object ontology is its inelegance, lack of simplicity, etc. But nobody has ever shown that elegance and simplicity are truth-conducive.

This insistence on the truth-connection is at the heart of my challenge to the notion that the theoretical virtues have *any* epistemic value. The reasoning is something like this. We metaphysicians ought to believe only what's true. So a consideration C gives us reason to believe that p only if C makes it more likely that p is true. But it's very unclear how a theory's elegance, simplicity, etc. might enhance the likelihood that the world is the way the theory says it is. You could resist this reasoning, notably either by denying the truth-connection requirement or by showing how some theoretical virtues meet the requirement. I'm open to that turning out to be the case. I just don't know of anything convincing that's been written to suggest that...

Eric Schwitzgebel said...

Thanks for the thoughtful reply, Uriah!

Your essay has motivated me to look more closely into the issue of whether simplicity is indeed truth-conducive. I'm inclined to think it is, though the literature on this issue is underdeveloped. What do you make of Sober's work on the topic, e.g.: Forster & Sober 1994; Hitchcock & Sober 2004? Sober isn't mainly talking about armchairish metaphysics in those papers, of course.

John Jones said...

The issues you present are cast in a transcendentally real mode and arise only there.

But whether objects include relational properties is immaterial if we adopt a transcendental idealism where the limits of objects are not fixed ontologically through their existence.

Eric Schwitzgebel said...

Ah, the old Kant two-step! I'm not necessarily hostile to that, but I confess it's not always entirely clear to me what either Step 1 or Step 2 is.