“Moore’s Argument and Varieties of Transmission Failure”

Hamid Vahid sent this paper and has agreed to have it placed here for discussion. Dr. Vahid is with the Analytic Philosophy Faculty at the Institute for Studies in Theoretical Physics and Mathematics (IPM) in Tehran, and given the growing interest in recent years concerning Moore’s proof and what value it has epistemically, it is a very timely paper. A word version can be downloaded here.


“Moore’s Argument and Varieties of Transmission Failure” — 4 Comments

  1. Hamid,

    I have a couple of comments for you. One is a quibble on your treatment of Wright, which may end up being quite important, while the others are directed at your analysis of the I-II-III arguments.

    I could only find you giving one criticism of Wright’s view on entitlement, since you (rightly) dismissed your own “negative dependence” notion as a red herring — you accuse Wright’s notion of entitlement as being non-epistemic and thereby objectionable. Now, Wright’s notion of entitlement is in fact non-evidential, but it only follows that it is non-epistemic if you implicitly identify “epistemic” and “evidential.” You don’t seem to give any further argument, and since this is precisely the point in question (Wright is challenging the normal identification of “epistemic” with “evidential”), I don’t think you’ve made your case in rejecting Wright’s view.

    You divide up the I-II-III arguments into two groups, the “easy knowledge” types (Zebra, Wall, BIV) and the Moorean types (Moore, Election, Soccer), and assert that the first group fails to transmit warrant while the second is unsatisfactory for another reason. I have some problems with your arguments in both cases.

    I’ll start with the “easy knowledge” group of arguments. Here is your view as I see it: you said that proposition I justifies proposition II only in a context of background assumptions that include the assumption of something close to proposition III. Since an argument for a conclusion cannot assume that conclusion, the argumentative context when the argument is given from II to III requires that the relevant background assumption be removed. In this case, proposition I can no longer justify proposition III (because it cannot justify proposition II). (You distinguish this from Wright’s view by making it an issue of argumentative context rather than justificatory structures.) So, you conclude, the argument fails to transmit warrant. There is a problem with this, though — the question is whether proposition I transmits warrant to proposition III given that it justifies proposition II. To answer this question, you can’t remove the background assumption you want, since that is the necessary condition for proposition I to justify proposition II. As a result, your view either collapses into Wright’s view or the arguments are not in fact transmission-failures (in which case, you still need to say what is wrong with them, since you are right to say that they are intuitively defective).

    Moving on, to Moore’s proof. You think the proof transmits warrant but is defective in another way: propositions II and III stand “so close” that any evidence that justifies II “ipso facto” (your quote) justifies III. (You just use the “closeness” sentences to illustrate this.) I confess that your meaning escapes me here. You have, it seems to be, two alternatives. On the one hand, you could mean that I justifies II and also justifies III, but it justifies III directly, without needing to infer III from II. If this is true, then the I-II-III argument is actually a transmission-failure, since warrant does not transmit over the II-III inference. Instead, it transmits directly over a I-III inference. If this is so, then you’ve just sided with Wright in calling it a transmission-failure. Also, I don’t see how a direct I-III inference is possible, since I can see no way to directly infer a type III proposition from a type I without proceeding through a type II. On the other hand, you could just mean that II and III are “close” in that II entails III. In this case, though, the I-II-III argument is just a normal argument, and you are left without an account of why it seems defective.

    I hope these are helpful, and I of course welcome any responses you may have. I’m especially interested if you could clarify the last point, your treatment of Moore’s proof.

  2. Don,
    Thank you for your helpful comments. Here are a few points.

    Re: Wright

    The argument from, what I called, negative epistemic dependence to positive epistemic dependence is Wright’s not mine. This argument was criticized in the paper. In addition, I made two further remarks about Wright. First, his thesis of disjunctive template helps itself with an underdetermination principle that would render the question of warrant transmission across the entailment in the I-II-III inferences redundant, and secondly that negative epistemic dependence is not what marks such inferences.

    By epistemic justification, I mean ‘justification’ in a truth-conducive sense. Wright appeals to a kind of warrant that, as you say, is not a matter of evidential support. He seems to think that we enjoy some sort of a default justification with respect to certain beliefs though he does not elaborate on where this comes from. Neither does he say much about the epistemic standing of his notion of entitlement. If it is not evidential, is it deontological? If so, in what sense? Does he take deontological justification to be truth-conducive? We are eventually told that this is not an evidential warrant to believe a proposition p but something like a warrant to act on the assumption that p, take for granted that p, or trust that p. Given such standing issues, and also considering the examples he gives to substantiate his notion of ‘entitlement’, not to mention his concluding remark that his skeptical solution concedes the basic point of the skeptical arguments, namely that we do indeed have no claim to know (justifiably believe) the truth of ordinary claims, I argued that one could recast Wright’s point about entitlement or rational acceptance in terms of the distinction between pragmatic reasons for believing and epistemic reasons for believing (see my discussion of Bratman).

    Re: Moore

    There is certainly a I-II-III inference in the case of Moore. This is in accord with my account of epistemic closeness of M-II and, its consequence, M-III. One can view Moore as a ‘normal’ argument where, by being derived from the justified premise M-II, the conclusion M-III is also justified. But M-II and M-III are not just any two propositions with one being a consequence of the other. Rather, M-II and M-III are also epistemically close in the sense that one cannot believe M-II without believing M-III so that when some evidence e makes M-II more likely true (justified), it will also justify M-III. Accordingly, given the epistemic closeness of M-II and M-III, the argument’s job is already done by the time we get to its second premise. So while Moore can be seen as a ‘normal’, sound argument, it is only so in a trivial sense, for, unlike substantial arguments, its dialectical function has already been discharged by the time we get to M-II although we may proceed to derive and thereby ‘establish’ its conclusion in a trivial sense pretty much like establishing ‘there is an election in process’ by deriving it from the proposition that John has just voted where our belief in the latter happens to be justified. The idea of epistemic closeness was intended to explain why Moore-type arguments strike us a defective.

  3. Hamid,

    Thanks for clarifying — your comments are helpful for understanding your position. I have a few responses.

    Re: Wright

    I think there is a lot that can be said about Wright’s proposal, and I cannot hope to exhaust that here. Your suggestion that we recast the non-evidential entitlement/evidence distinction into a pragmatic/epistemic reasons distinction is interesting, and I think that it gets at something important. My first concern is that in calling entitlement a merely “pragmatic” reason for believing, you devalue it — in fact, I take it that this is your criticism of Wright. I don’t think this criticism is fatal, but I do think that Wright needs to give some account of why the entitlements we have are truth-conducive (I actually think that a theistic framework lends itself to this task). This is an uncompleted task as of now. Again, though, this doesn’t yet amount to a very strong criticism of the view.

    Re: Moore

    Your view of Moore’s proof depends on saying that any evidence e (in this case, M-I) that justifies M-II also justifies M-III without the need of an extra inference from M-II to M-III. The problem is that I just don’t see how M-I directly justifies M-III. I would think that “I see a hand, therefore external things exist” is even more objectionable that “I see a hand, therefore there is a hand, therefore external things exist.” Either the first is more objectionable than the second, or both are equally objectionable; in either case, you haven’t given an account of why the proof is objectionable.

    Any thoughts?



  4. Hamid,

    Sorry, one more comment on Moore. In addition to the fact that I cannot see how M-I can justify M-III directly, my original point still stands: since warrant transmits directly from M-I to M-III, then it doesn’t transmit across the II-III inference (since you are already justified in believing III by believing I). Therefore, the I-II-III inference is actually a transmission-failure, because the II-III inference is. (A direct I-III inference may not be a transmission-failure, assuming of course that it is a valid deduction, which I have challenged.) But you wanted to say that the argument was not a transmission-failure and give another explanation of why it is unsatisfactory.

Leave a Reply

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