Tuesday, April 14, 2009

Intuitions about Cases in Metaphysics

I think it’s safe to say that intuitions about cases tend to be taken less seriously in material-object metaphysics than they are in (e.g.) epistemology, philosophy of language, philosophy of mind, and ethics. Does anyone know of any explicit discussion or defense of this differential treatment in the literature? In particular, is there any discussion (even in passing) of either of the following two claims:

(i) That we should be more skeptical of particular-case intuitions about material-object metaphysics (or metaphysics generally) than we are of particular-case intuitions about other matters (e.g., in epistemology, phil language, ethics).

(ii) That we should be more skeptical of particular-case intuitions about material-object metaphysics (or metaphysics generally) than we are of general-principle intuitions about material-object metaphysics (e.g., anti-colocation intuitions).

The can only think of two discussions. The first -- bearing on question (i) -- is in Rodriguez-Pereyra’s Resemblance Nominalism (p.217) where he contrasts intuitions about metaphysics with intuitions in philosophy of language, and he suggests that the latter are reliable only because the range of facts intuited (e.g., about meanings) are themselves determined by our conceptual activities. The other -- bearing on question (ii) -- is the last couple sentences of Ted Sider’s paper “Parthood,” where he suggests that general-principle intuitions are more trustworthy because judgments about cases tend to be “infused with irrelevant linguistic intuitions.” I’ve also encountered various responses in conversation, e.g., that metaphysics is about what exists, or that it’s misguided to rely on conceptual analysis in this domain. But I’ve never seen any proposal worked out in any detail, and I have my doubts that any of them can draw the line in the right place between (on the one hand) cases and principles and (on the other hand) metaphysics and other areas.

I’d be grateful for any references, as well as any thoughts on how (i) or (ii) should be defended.


  1. Two things spring to mind. Firstly, despite the fact that some philosophers are monists, the anti-collocation intuition is not robust. Most people have this intuition before doing philosophy and *most* drop it upon learning about the statue and the lump.

    Secondly, 'object' is often used in a technical sense. Sometimes to mean Spelke object, but more often to mean Fregean object. Intuitions about material objects are guided by something like the notion of a Spelke Object, whereas lots of philosophical discussion is guided by the notion of a Fregean object.

  2. Hi Dan,

    What is the difference between a particular case intuition and a general case intuition? Is this distinction supposed to track the universal/non-universal distinction?

    If so, it seems to me that philosophers of language and linguists take ordinary judgments about universal sentences just as seriously as they take atomic or existential sentences. For example, they seem to take the ordinary judgments about the truth of the sentence "all of the beer is in the fridge" just as seriously as they take the judgment about the falsity of "there is beer" when there's no contextually salient beer. (This is not to say that philosophers of language or linguists always agree with folk judgments, but they seem to take them equally seriously.)

    Does your distinction correspond to the distinctions between quantified/quantifier-free sentences. It seems to me that Sider is completely right if he means that quantified claims - whether universal or existential - can do work to suppress irrelevant features of the judgment which are present in non-quantified sentences. Judgments containing singular terms or definite descriptions can trigger irrelevant presuppositions which can be avoided by appeal to quantified claims. Philosophers of language can be very sensitive to these considerations. So, I think it would be a mistake to think of them as more prone to prefer judgments about non-quantified sentences to those concerning quantified sentences.

  3. Lee,

    I'm not sure I follow your suggestions. Your first point is intended as a point *against* claim (ii), right? I take it the second point is meant to support (i). Is the suggestion that, even when philosophers try to consult their intuitions about whether there's *any* (Fregean) object composed of my nose and the Eiffel tower, they end up having the intuition that there's no Spelke object composed of my nose and the Eiffel Tower, and then mistakenly think they had the intuition that there's no Fregean object composed of my nose and the Eiffel Tower? If so, I wonder why this inscrutability of intuitional content wouldn't infect other areas (e.g., maybe we're only having the intuition Smith doesn't know *with certainty* that the man who gets the job has ten coins in his pocket).


    I didn't mean to suggest that general claims aren't taken seriously in other areas; perhaps (as you’re always reminding me) it’s misleading to put this in terms of cases vs. general claims. Here's what I have in mind. Sometimes the following sort of thing happens: there's a theoretical claim F (e.g., that no proposition can be both necessary and a posteriori); G is an instance of F (e.g., that it is not both necessary and a posteriori that water is H2O); and, intuitively G is false. What I am suggesting is that, in situations like this, it’s far less common in material-object metaphysics than it is in other areas to abandon F on the basis of G. This is so with your example as well. If there were some theoretical claim that entailed that ‘all of the beer is in the fridge’ is infelicitous, the claim would be rejected on that basis.

  4. Hi Dan,

    I'd take a look at Brian Weatherson's "What Good Are Counterexamples?" -- he says some stuff relevant to both (i) and (ii) there, although his case study is the post-Gettier epistemology literature.

    I think lying in the background of a lot of people who endorse (i) (and probably (ii), though it takes more work to see how the connection goes) is the picture that philosophy proceeds in two parts: the METAPHYSICS part (where we figure out how the world is `in itself') and the ANALYSIS part (where we figure out which of our descriptions are appropriate for which ding-an-sich states). The thought is that how the METAPHYSICS part is has got nothing to do with us or how we describe things, so our intuitions (which only reveal to us how we're disposed to describe things) can't tell us anything about that. But on the ANALYSIS side, since which descriptions are proper depends on (even if it is not wholly constituted by) our dispositions to describe, finding out about these dispositions can have some evidential weight. This weight should be defeasible, of course, to the extent that we're externalists about our descriptions --- this is what Weatherson points out in his paper, which could be taken to suggest that we as philosophers are over-reliant on counterexamples because we haven't fully appreciated the import of the externalist revolution.

    Second, although I don't have any citations to hand, I suspect the literature surrounding our knowledge of moral properties (if there are any) will be relevant here, too. There you have arguments like: If there were mind-independent moral properties, then what is in fact right and wrong might diverge largely from our (deeply held) intuitions about it. These sorts of arguments seem to me to very much rely on the sort of "our intuitions get no grip on METAPHYSICS, but only on our descriptions of the world" picture that seems to lend support to (i).

  5. Hi Dan,

    A couple of points wrt (i):

    First, I suspect that the lack of trust in metaphysical intuitions and especially our intuitions about material objects has its source in the (at least nominally) naturalistic bent of much contemporary philosophy. QM and SR have substantially weakened our confidence in the reliability of our intuitions about how the world is (cf. Eddington's two tables) and evolutionary considerations seem to suggest that the way we conceptualize the world may be noything more than a relatively blunt tool apt at most for our survival in certain specific environments. (I think you can probably find this kind of arguments against our metaphysical intuitions in Ladyman and Ross's Everything Must Go: Metaphysics Naturalized) I don't know how seriously analytic metaphysicians take such considerations but, if analytic metaphysics could develop after the logical positivist ban was partly because analytic metaphysicians displayed a certain degree of "humility" about common-sense intuitions and deference to science in order to be taken seriously.

    Second, I think Jason's comment is descriptively accurate, but I'm not sure we should think of things this way. In particular, whereas metaphysics is not merely conceptual analysis, conceptual analysis seems often to be a precondition to asking the more substantial metaphysical questions. Before asking questions such as 'Do ordinary objects have temporal parts?' 'Are there any laws of nature?', all sides would seem to have to agree as to what temporal parts or laws of nature are if there are any, which is often where a lot of the disagreement takes place.

  6. Jason,

    Yeah, Weatherson’s paper has played a big role in my thinking about this stuff. But I worry about how relevant his reference-magnet-based worries are to justifying any kind of differential treatment, either of type (i) or type (ii). Presumably, *if* there are reference magnets, all of the central objects of inquiry in philosophy are going to be referents magnets (parthood, knowledge, permissibility, causation, consciousness, meaning). As you say, he directs his skepticism at Gettier intuitions, which I was taking to be the paradigm case of intuitions that typical parties to the material-object debates do take seriously. Plus, his worries should apply equally to general-principle intuitions (e.g., that knowledge requires justification, that parthood is transitive, that material coincidence is impossible).

    On the two-phase (analysis and metaphysics) picture: I think there’s something right about this. Phase one: First we figure out what it would take for there to be Fs. Phase-two: we figure out whether anything satisfies the conditions for being F. Gabriele suggests (and I suspect you’ll agree) that at least some metaphysical questions fall into phase one. But it seems to me that even the areas of metaphysics where particular-case intuitions are meant to be less-reliable-than-elsewhere fall cleanly into phase one. Think about composition: We figure out what relational property would have to hold among some things in order for them to compose something, and then we figure out whether there actually exist any things that together instantiate that property. Phase-two should be the (largely nonphilosophical) investigation into the distribution of matter in spacetime, or something like that. So it’s not clear how this two-phase picture is going to support (i).

    On moral intuitions: I’ll definitely look into that literature to see what I can find. I do wonder how our moderate skeptics (who are skeptical about intuitions in metaphysics but are generally happy to engage in intuition-driven philosophy) would feel about their skepticism extending to particular-case intuitions in ethics. (E.g., that we should treat the intuition that it’s permissible to unplug from the violinist the way we treat the intuition that my nose and the Eiffel Tower can exist without composing anything.)


    I don’t see why the skepticism generated by QM and SR or the skepticism generated by evolutionary considerations justify (i). If indeed our intuitions about (say) space, time, and objective indeterminacy are called into question by QM and SR, how does that impugn intuitions about composition, but not intuitions in epistemology, phil mind, phil language, etc.? And if intuition is indeed “a relatively blunt tool apt at most for our survival in certain environments”, why doesn’t that simply support a skepticism towards all intuitions about matters that are more or less irrelevant to our survival (which will include twin earth, Gettier man, Godel-Schmidt, Thomson’s violinist, etc.)?

  7. Hi Dan,

    A few thoughts. First, to clarify: The two stages you're thinking of aren't the two I had in mind, exactly. In fact, what I had in mind doesn't map on to two temporally distinct stages in a single project, but rather two distinct projects --- which perhaps cannot be undertaken but in tandem.

    Let's put it this way: I think the metaphysical picture a lot of practitioners have in mind is the 'Relativity' picture. The spacetime structure is what it is independent of our intuitions (they think); figuring out what the spacetime structure is, well, that's a job for METAPHYSICS. (No, its science!, you say --- but the people I have in mind insist that these two things aren't going to be mutually exclusive.) And intuitions just have got nothing to do with what spacetime structure is like; the fact that the untutored will insist that of *course* two events can happen at the same time, full stop, just has no evidential weight.

    Of course, given that spacetime has thus-and-so a structure, there's a further question: why should things separated in this way could as "time-like" and in that way as "space-like", etc. In other words: why does our concept of time fit (even if somewhat unhappily) with this aspect of the structure instead of that? This further job is a job for ANALYSIS. And so an investigation of our intuitions will be useful there --- not for figuring out that the world has thus-and-so spatiotemporal structure, but for figuring out that thus-and-so portion of the world's structure is the bit that deserves to be called "spatiotemporal".

    The idea these metaphysicians have, I'm guessing, is that questions of ontology (or of "fundamental" metaphysics, whatever that turns out to be) are all like questions about what spacetime structure is like, and so intuitions don't have very much sway there. But wherever they think that they're arguing about something less than fundamental, they think intuitions are more important --- for the intuitions will dictate how the non-fundamental concept will relate to the METAPHYSICAL structure. (You get a feel for this in Lewis comparing him on causation and mereology; also in his treatment of the Humphrey objection, where the idea is that, so long as the modal sentences that come out true under the analysis get the intuitions right, it just doesn't matter whether the analysans jibe with our intuitions or not.)

    This isn't to say that intuitions have nothing to do with the project --- the job of figuring out what METAPHYSICS (the ultimate structure, whatever you want to call it) is like will be constrained by the intuitions in the following manner: whatever story we tell about the ultimate structure must show how it can "account for" (provide enough analyzing material, I guess) to make the various analysandums (analysadi?) true.

    OK, all that being said, my second thought: you've got me thinking, and I'm now not sure that your claims (i) and (ii) really are all so safe to say after all. Particular cases are certainly used in material-object metaphysics --- think of van Inwagen's people shaking hands (a new thing doesn't pop into existence!), or Merrick's reliance on the intuition that he has seen his wife (leading him to conclude that people are distinct from their brains, since he has never seen her brain), or what have you.

    There are two features of the material-object debate that make appeal to intuitions there seem not quite as straightforward as in (say) the Gettier or Twin Earth cases. First feature: most people working in this area think we simply have prima facie inconsistent intuitions (think of how the Leibniz' Law arguments are generally presented), and so something has to give. So simply pointing out that some theory runs awry of some intuition won't settle the issue, because every theory does that. But this doesn't mean intuitions aren't important, or even that it's ok to simply say "well, nuts to that intution, I like these better"; most practitioners then go out of their way to do as much to salvage/explain away/etc. the damaged intuition.

    The second is more a speculation than a an assertion, but here goes: it's not at all clear that the material-object intuitions in question are as robust as many of the classical examples (Twin Earth, violinist, etc.) It takes a lot less to talk most people into accepting that there is something composed of my nose and the Eiffel tower than it does to talk most people into accepting that, say, Smith knows that Jones owns a Ford or Brown is in Barcelona. At least, that's how things look to me --- maybe I'm wrong about this? (Xphi? Anyone?)

    But recognition of either of these points isn't the same as saying we should be more skeptical about these intuitions --- just that we can't do them as much justice as we can in other areas. And there are plenty of other debates that are rather similar; debates in epistemology over the Lottery paradoxes have quite a bit of this flavor, for instance. We can't accomodate all our intuitions about knowledge, so something will have to go; what can we get rid of with the least damage to the rest? But that's not the same as being skeptical about them; just realist about what our theoretical options are.

  8. Hi Dan,

    It seems to me that you are seriously overestimating the extent to which ethicists and philosophers language, mind and logic are willing to respect pre-theoretic judgments when compared to metaphysics.

    In philosophy of language, there is still open debate about whether almost all utterance of the sentence 'There is no beer' are false, despite our pre-theoretic judgments. Direct reference theorists - still an open position last time I checked - often say we speak falsely when we report beliefs with proper names. Radical Contextualists often have to "explain away" our strictly false pre-theoretic judgments that speakers say the same thing by two utterances. Let's not even get started on the relativists.

    In philosophy of mind, one still hears functional role people and others saying that strictly speaking no two people have the same beliefs. I don't take this position to have been refuted despite its obvious tension with ordinary beliefs (I do think it's false though).

    In Ethics, act utilitarianism seems to conflict with many pre-theoretic judgments. Nonetheless, I don't think of it as a refuted position. Also, pre-theoretically, I firmly believed that there is nothing wrong with buying meat from a grocery store. I still don't think that the action is wrong, but I don't think the pre-theoretic judgment counts very much against the ethicist who argues for vegetarianism because it follows from various ethical principles.

    About epistemology, I'm not so sure. Is skepticism considered to be merely false or completely refuted? If it's not refuted, we'd have at least one position which seems to conflict with some pre-theoretic judgments. I'm pretty sure there must be others. I've heard a few people suggest that justification entails truth, which I wouldn't have expected. Is this a conflict with pre-theoretic belief? Is this position refuted?

    Anyhow, it seems to me that one will think metaphysicians are more willing reject folk believes only insofar as one views these as closed positions. Even though I disagree with many of these positions, I think that they are open possibilities in their respective fields. As a result, I don't see philosophers in these areas as any more prone to respect ordinary judgments than metaphysicians.

    One final point. I agree with Jason's suggestion above that many metaphysicians view our pre-theoretic judgments about the statue and clay may very well be inconsistent. This would give us a very good reason to reject some of these claims.

  9. Hi Dan,
    I'll have to admit that I haven't read all the comments, so I'll write what I write with the hope that it is helpful.

    For the most part, I like Michael Huemer's principle: if it seems to S as if p, then p is prima facie justified for S. The thing is that prima facie justified beliefs can be defeated. Perhaps the reason people are less trusting of intuitions about certain cases in metaphysics is because they have defeaters.

    Here's an example. It seems to me that there is a chair, and that the chair is composed of parts. This gives me prima facie justification for my belief that there is a chair. But Merricks, as you know, argues that we have an undercutting defeater for the belief. He points out that how we are perceptually appeared to would be the same even if there were no chair and there only atoms arranged chairwise. I would add that since what seems to be true (that there is a chair) is dependent on how we are appeared to, and how we are appeared to, Merricks points out, does not give sufficient evidence for believing there is a chair, it follows that we have a defeater for our belief that there is a chair.

    (So I am distinguishing the belief, the perceptual appearance, and the intuition/seemings. Merricks' reasoning gives us reason to think that both the perceptual appearance and the intuition/seemings is truth-indicative.)

    To use another of Merricks' examples, if somebody pointed out to me that I am wearing green tinted glasses, I would gain a defeater for my belief that the objects around me are green. This is in spite of the fact that I am appeared to greenly, and it seems to me (intuitively) that the objects around me are green. (We talked about this when we first met at MEW1 the night before the conference started, I don't know if you remember.)

    In other areas having to do with material constitution, I think that the defeater arises when we see that our intuitions lead us to conflict with other intuitions. This can also provide one with a defeater. (I think that Michael Rea lays this out really well in the introduction of his material constitution anthology.) This leads us to have a general distrust of all of our intuitions about material constitution.

    Could the same things about defeat apply to phil. language and other areas? Maybe... but then I think that that would be reason to distrust intuition in those cases too. I guess you have to go on a case-by-case basis.

  10. sorry, in my fourth paragraph, I should have said "not truth-indicative."

  11. Hi Dan,

    I think the most important point is the one Pickel made at length above: intuitive counterexamples always matter, they're just not decisive, and if the history of philosophy has taught us anything, it's that some of our pre-theoretic judgments are going to turn out false.

    As for this particular case, I think part of the explanation for the difference between the composition literature and the Gettier literature is that almost everyone is convinced that there isn't an answer to the special composition question that will accord with all our pre-theoretic intuitions. It's either brutalism or giving up pre-theoretic intuitions, and brutalism is bad philosophy. (Plus, brutalism conflicts with our pre-theoretic intuitions as well, just not about specific cases.) Hence our willingness to give up pre-theoretic intuitions.