Limitations on X-Phi?

So here’s an X-Phi investigation I and others interested in logic and its philosophy needn’t care about: use of the term ‘contradictory’. (This thought occurs to me when listening to reporters and commentators on the news, proclaiming that some people’s testimony about various events is “contradictory” or “inconsistent,” by which I taken them to mean not much more than that successive querying by authorities or reporters generates different answers on different occasions.) That is, the study and development of logic has only the remotest connection to ordinary language and thought involving the use of the term ‘contradictory’ and ‘inconsistent,’ remote enough that it would be a matter more of individual psychology than required methodology to guide theory by such X-Phi investigations.

I expect widespread agreement on this point, and it raises an interesting question about when and where X-Phi investigations should matter more. Some, for example, think they matter more in epistemology, especially that part of epistemology constituting the theory of knowledge (shame, shame, shame on those or you who have thought that’s the entirety of the discipline! 🙂 ). So what’s the difference between logic and epistemology on this point?


Comments

Limitations on X-Phi? — 8 Comments

  1. Jon, I suppose that by “logic” you mean (roughly) the science of logical consequence in artificial formal languages. If that’s right, then one important difference between logic and epistemology is that logic studies something formal and artificial that is only remotely connected to ordinary social cognition, whereas epistemology studies something natural that is very important in ordinary social cognition. This difference in no way diminishes the value or importance of logic, thus understood. But it does matter to the methodological question you posed.

    Some logicians might view the relation between logic and ordinary language differently from you, though. Barwise and Etchemendy write, “One of our major concerns in this book is to examine [the] notion of logical consequence as it applies specifically to the language FOL. But in so doing, we will also learn a great deal about the relation of logical consequence in natural languages,” _Language, Proof and Logic_, p. 5. (The commonality here, I take it, is at the level of a certain relation between claims, rather than the usage of any particular word.) And I’ve heard more than one logician defend their approach by appeal to what we mean by “truth” or “proof” — though, consistent with the sentiment you express in the OP, not “contradiction” or “inconsistent”.

  2. I’d like to answer Jon’s question by boldly claiming that the difference is nil, because intuitions, whether generated in the armchair or gathered in the field, play no role for epistemology or logic.

    That’s, of course, an over-statement. But it gets me your attention. And now that I have the latter, I’d like to use the opportunity for some shameless advertising. The first pages of this paper sketch how we can do philosophy with less reliance on intuitions:
    http://huber.blogs.chass.utoronto.ca/wp-content/uploads/2014/06/New-Foundations-for-Counterfactuals.pdf

  3. Franz Huber, I am quite confused by something you say in the paper. You write,

    “One reaction might be to go experimental (Knobe and Nichols 2008) and to see which intuitions, or subjective evaluations of the quality of explanations, are more widespread. On my view this would not help much, because truth is not a matter of democracy. What we need, or at least what we should aim at, is a principled account of the semantics of counterfactuals.”

    I certainly agree that polling subjective evaluations of the quality of explanations would be not truth-conducive. However, this is simply not what experimental philosophers do. By and large, they do not poll. By and large, they do not ask for subjective evaluations of *the quality of explanations*. Instead, they ask for “intuitions” — if we want to call it that — regarding particular cases.

    In the counterfactual literature, such judgments would probably be about the admissibility of certain controversial counterfactuals into the explananda. For example, Paolo Santorio has begun to investigate the nature of backtracking counterfactuals.

    Another debate with a similar shape is on epistemic modals, where semanticists are using experimental methods to gather data regarding, e.g. retraction and disagreement. Joshua Knobe and Seth Yalcin has a paper forthcoming on this topic and Justin Khoo has a paper in progress.

    So, to return to the topic at hand, even if we all agree that subjective evaluations of the quality of explanations plays no role in philosophical debates, do you think experimental methods can be useful in other ways?

  4. Shen-yi Liao, thank you very much for your comment!

    My claim is that it would not help much to see which intuitions are more widespread, because truth is not a matter of democracy.

    In personal communication Sarah Moss suggested that she and Thony Gillies do not have differing intuitions. Rather, they disagree about what the best explanation for a shared intuition is. In order to take this into account I have included “the subjective evaluations of the quality of explanations”. While I think that, as a matter of fact, subjective evaluations of the quality of explanations do play a role in philosophical debates, let us bracket them and focus on intuitions only.

    My claim then is that it would not help much to see which of these intuitions are more widespread, because truth is not a matter of democracy.

    Now you say that experimental philosophers, by and large, do not poll. And you go on and say that, instead, they “ask for intuitions […] regarding particular cases”. We do not have to call it polling (a term you introduced). My claim is:

    the results of asking subjects for intuitions regarding particular cases would not help much, because truth is not a matter of democracy.

  5. I guess I’m still not sure what the slogan “truth is not a matter of democracy” amounts to. Suppose you want to give a theory for a piece of natural language, say, counterfactuals. (Note that I am not saying that theorizing about natural language is the only thing that philosophers have been trying to do with counterfactuals.)

    Now suppose you think a counterfactual sentence S is felicitous to assert in some particular context, but you also find out that most natural language speakers do not find S to be felicitous to assert in that particular context. Is the thought that the judgments of this majority of natural language speakers are irrelevant because “truth is not a matter of democracy”?

    I introduced the term “polling” because the term “democracy” naturally conjures up images of voting. But that’s not, by and large, what experimental philosophers ask participants to do. Observing and manipulating participants’ behaviors — including their linguistic behaviors — is like a democracy in what sense?

    I guess my lack of understanding of the slogan “truth is not a matter of democracy” really comes down to not understanding which truths are being covered and in what sense is the practice of experimental philosophy like democracy.

  6. “Is the thought that the judgments of this majority of natural language speakers are irrelevant because “truth is not a matter of democracy”?”

    By and large, yes! (As I said initially, my aim is to do philosophy with less reliance on intuitions, as it may not be possible to do philosophy with no reliance on intuitions.)

    The behavior you observe and manipulate may tell you something about how these participants behave in certain contexts — say, how they use the words ‘knowledge’, ‘factive’, ‘earth’, and ‘flat’ in certain contexts.

    However, even if all subjects use the words ‘knowledge’, ‘factive’, ‘earth’, and ‘flat’ in exactly the same way in a given context, that does not mean that the earth really is flat.

  7. I agree. We don’t know for a fact .. we don’t know it is True, that counterfactuals are required to explain causation, another human conceptual reality. But counterfactuals have become a very popular intuition, a large subjective consensus which empowers and underlies the aptness of all other analogies also. So we don’t know if we use true premises in some argument, but that doesn’t seem to deter authors in “Causation and Causality” which you cited. How does your paper differ in the quality of truth invoked from those authors in C&C unless it’s more plausible? I guess some people forget that the only truth available to us is democratic, at least in its origin.

  8. I agree with Sam that “truth is not a matter of democracy” is not a useful slogan here. P need not be determined by our cognition, for x-phi to be useful — all that is needed is that x-phi results speak to whatever sort of considerations they are that we appeal to in our debates about P. When these are intuitions, then x-phi research about the distribution of, variation in, and nature of intuitions is going to be (at least in principle) relevant. Ditto for, say, introspection.

    Compare, e.g., to the relevance of work like Gigerenzer’s to improving medical practice. “Medical truth is not a matter of democracy” is surely true as well — but a better understand of human psychology is highly relevant to coming to do medicine better.

    Btw this (very!) old blog post of mine seems somewhat relevant:
    http://philosophycommons.typepad.com/xphi/2005/03/two-metaphilosophies-on-intuitions-the-folk-and-matters-empirical.html

    I agree with John on the logic question, fwiw. One might also compare that usefully to, say, why it wouldn’t be relevant for research physicists to worry about folk judgments using the term “mass”. If someone wanted to use “knowledge” as a technical term — as, say, “warrant” frequently is, in epistemology — then x-phi would likely be of limited use there as well.

    Is this really true, btw? “Some, for example, think they matter more in epistemology, especially that part of epistemology constituting the theory of knowledge….” I guess, for the reasons just stated, I would defend the idea that x-phi is more relevant within epistemology to knowledge or, say, understanding or wisdom than it is to, say, justification or warrant or positive epistemic status. But I’m not aware of anyone suggesting that x-phi is more relevant to epistemology than it is to research in ethics or agency or semantics. No?

Leave a Reply

Your email address will not be published. Required fields are marked *