Friday, August 04, 2006

Mailbag: Externalism and Dialectical Skepticism

A while back, I wrote about why epistemological externalism is a legitimate response to skepticism. I recently received an interesting email from Rex Hubbard revisiting this debate (quoted with permission):

Many months ago I posted an entry on my blog, the thesis of which is that Nozick begs the question against the skeptic in his well-known defense of the tracking account of knowledge. You rebutted my post, and I in turn responded to you. This resulted in a series of exchanges. Several months ago I deleted the post because it was difficult for the reader to follow, and I was too lazy to edit it. But I have not stopped thinking about the exchange that we had, and I recall enough of it to allow for an accurate reproduction. The intuitions on which I was relying during our exchange are still very strong, and so I feel that you might want to hear my response to your thoughtful comments after this period of reflection. Furthermore, I have recently come across an essay that expresses many of the points I tried to express. This will assist the clarification of my position.

My argument was simple: Nozick seems to assume throughout his argument, especially with his comments such as "If I were not in Jerusalem (or wherever) I would be in the U.S" that he is actually not a BIV. I argued that he does not argue for this assumption. Until he does argue for his assumption, then he will have no response to the skeptic who is himself giving a well-reasoned argument that nobody can show that he is not a BIV.

You responded in a now familiar manner: Nozick is arguing for an externalist theory, not an internalist one. Therefore, he does not need to give any argument to show that he is not a BIV: that is an internalist obligation, not an externalist one. Nozick merely has to show that it is POSSIBLE that we have knowledge. He does that, because, assuming that we are not BIV's, THEN we have knowledge.

I responded in the following way: Nozick has no grounds for assuming that we are not BIV's, and that this assumption stacks the deck against the skeptic, since the total absence of evidence for this assumption is what the skeptic's argument depends on.

You responded that Nozick has the right to make this assumption, because a person can assume anything that he wants FOR THE SAKE OF ARGUMENT.

This is as much of the exchange I recall. So now let me give you my latest thoughts.

My first objection to your first response: There are actually two assumptions that must be made. The first assumption, like we have already seen, is that we are not BIV's. But a second assumption must also be made: that Nozick's analysis of knowledge is correct. I do not think that this assumption is justified because I think that Nozick's analysis, like all externalist analyses, are VERY inaccurate. I think that knowledge is very much related to epistemic norms such as the right to be certain, etc, and I think that it can only be all right to be certain if you can justify your certainty with an argument. But since any disagreements on this issue end up being fruitless intuition battles most of the time, I will say no more about it. I would just like to make the point that there is a second assumption here, not just one.

My second objection: I do not think that your characterization of the skeptic's position is the best one. You characterized the skeptical thesis as the claim that knowledge is impossible. I would like to offer a different characterization of the skeptic's position that I think better captures the point he is trying to make. Here I borrow from the essay I mentioned above, "Skepticism and Naturalized Epistemology" by David Shatz. He claims that there is the following sort of skepticism that most externalists unjustifably ignore:

Dialectical skepticism: There is no dialectically effective argument for our first-level belief commonsense beliefs or our second-level commonsense beliefs.

Clearly Nozick is helpless against this form of skepticism, and it is this form that I was assuming in my critique. I think that Nozick realizes this, though since I recall (it has been months since I read the essay) that he concedes that this form of skepticism cannot be refuted.

Now I would like to borrow more from Shatz; these thoughts should show that Nozick's point is, if any sort of victory at all, a very small one.

Consider a man who thinks that the BIV scenario is true. He uses this hypothesis to account for the irritations of his sensory receptors in the same way that we use the hypothesis of commonsense realism to account for ours. When this fellow has table-like experiences, he hypothesizes that his vat caretakers are now stimulating his nervous system to have table-like experiences. Surely we would want to object that this fellow has no warrant for his hypothesis. But he can respond to our challenge in the same cowardly way that externalists respond to the skeptic: "That belief MAY be reliable. If I really am a BIV, then (to parody Nozick's account of knowledge) I know that I am now being stimulated by the caretakers to have these table-like experiences provided that, if these caretakers WEREN'T stimulating me, I WOULDN'T have such experiences." (Shatz 130). Now clearly this man has no argument that his belief is reliable in this way, just as the externalist has no argument whatsoever for the claim that his beliefs ARE reliable. I think that this thought experiment shows that this assumption that we are not BIV's, whether for the sake of argument or what have you, is just laughably without warrant in the same way that the fellow's assumption that he is a BIV is without warrant. This creates a serious problem for any externalist who thinks that he has even come close to refuting skepticism of the (most) important sort I have described.

I hope that these thoughts have at least been stimulating. Feel free to respond.



Shatz, David. "Skepticism and Naturalized Epistemology". Naturalism: A Critical Appraisal. Notre Dame. 1993.

I agree that we cannot rationally compel the committed skeptic to agree that we actually have knowledge. But that alone doesn't prove that our beliefs lack warrant. That's why I previously focused on the stronger skeptical claim that the justification required for knowledge is impossible. That's a more threatening claim, but one that - as we've seen - the externalist can rebuke.

If I understand him correctly, Rex wants us to address the more moderate skeptical position which holds that we contingently lack knowledge. Rex's skeptic grants the externalist claim that IF our commonsense realism is true THEN our beliefs amount to knowledge. But he denies the antecedent.

What grounds does he have for this denial? It sounds awfully dogmatic -- surely a true skeptic would instead remain agnostic about the ultimate nature of external reality, rather than bravely asserting our commonsense beliefs to be false. That is, he should allow that the antecedent might well be true for all he knows. Hence, granting externalism, the consequent might just as well be true. That is: the rest of us might well have externally justified knowledge, for all the skeptic knows. The careful skeptic will not go so far as to deny us knowledge. He will merely refrain from granting it.

We cannot compel the skeptic to accept our position -- but nor can he compel us to his. We have a dialectical stalemate. This won't worry the externalist. He can still claim to have knowledge. Maybe he's mistaken, but again: maybe not. (Note that it would be question-begging to ask externalists to refrain from making claims that they cannot show are justified. That would be precisely to insist on internalism!)

To be perfectly clear, here is the situation according to externalism:
(1) If realism is true, then we have knowledge.
(2) If realism is false, then we lack knowledge.
(3) It cannot be shown whether realism is true or false.

Realist externalists will further claim:
(4) Realism is true.

Our moderate skeptic grants 1-3, and is agnostic on 4. None of 1-4 has been shown false. The realist externalist emerges unscathed.

Rex brings up the vat-theorist, and suggests that "we would want to object that this fellow has no warrant for his hypothesis." Realist externalists will agree: the vat-theorist's beliefs are objectively false, unreliable, or what have you. That doesn't imply that our beliefs are in the same sorry state. Whether they are or not is a contingent matter, and depends on the external world. (The vat-theorist can defend himself in exactly the same way, and there's nothing wrong from his perspective in doing so. Realist externalists should grant that his view, like the skeptic's, is rationally tenable. They think that what makes him wrong is the external world.)

We may grant that "the externalist has no argument whatsoever for the claim that his beliefs ARE reliable." (Only an internalist would care.) To suggest that this shows the externalist's beliefs to be "laughably without warrant" is completely question-begging. Remember, the externalist holds "warrant" to consist in objective reliability and such, and he hasn't been shown to lack this.

So really, I don't think Rex has two objections here at all. The second one is shot through with internalist assumptions from start to finish. Dialectical skepticism, concerned as it is with what can be shown rather than what is objectively the case, won't much interest the externalist. This just leaves us with the first objection: that externalism is a poor theory of justification.

(In a future post, I'll suggest some internalist reasons for preferring common-sense realism to vat-theory.)



  1. Bottom line is the skeptic should be skeptical about skepticism (thus resulting in the agnostic conclusion), otherwise there is inconsistency.
    (And yet - this is a different tack to your approach to utilitarianism where you differentiate between the attitude to the theory and the attitude to the day to day decisions)

    As for the externalist, I’m not sure if “I don’t have to justify things because I’m externalist" is much of a defense. Externalism would then be a form of insanity. But I presume an externalist will have a different way in which options are whittled down and structured within reality (just as a skeptic must be grounding their beliefs in something in order to even be engaging in the debate). And that method might or might not be a good or consistent one.

  2. I've started to be very doubtful too of whether externalism really offers the resources to give a satisfying response to scepticism, for the reasons Crispin Wright has offered. (Some of the relevant material is published in 'Warrant for Nothing', though much of it is in unpublished papers on Dogmaticism and to McDowell's disjunctivism. I'll do my best to only say things that can be found in the published stuff).

    As Wright sees things, whether or not we have knowledge or justification really isn't the main target of scepticism, properly understood. The real target is our right to *claim* knowledge or justification - the suggestion is that scepticism properly understood is second-order from the get-go. Vindicating the possibility of knowledge or justification just doesn't even make it up to the start line as an adequate response to the sceptic, from this perspective.

    I'm finding myself persuaded that this is the right way to think about these issues, and that the externalism response is inadequate.

  3. An extreme externalist might insist on an externalist account of the "right to claim knowledge" too. He could then assert that he does as a matter of fact have this right, though he cannot show it.

    I think all this shows that there's something a bit unsatisfactory about externalism itself. At least, there is surely a much-needed place for internalist justification, since we need internally accessible guidelines when deliberating about what to do or believe. But, as suggested in the post, this isn't really anything to do with skepticism per se. The objection is to externalism itself as an inadequate account of justification (or "right to claim" stuff, or whatever...)

  4. I would disagree that it is trivial as this kind of uncertainty is not necesssarily akin to everyday uncertainty - where there is normally some form of evidence for beliefs and maybe some semi-reliable way of gauging probabilities. I also think it's misleading to characterise my post as stating that "we can't be absolutely certain of most things". I was saying that we cannot have ANY degree of certainty about almost anything. If I were to accept your account of knowledge and thus accept that it is possible that we have knowledge, it would not follow that we have any knowledge whatsoever. Possibility does not establish actuality. And although it would be very difficult to assess the probability that a given belief is true given the possibility that all our experience is unreliable, i would still tend to think that it might be very small, due to the large number of competing possibilities. That we might in fact have no knowledge at all, and that it might even be LIKELY that we have no knowledge (taking a sceptical stance that our attempts at knowledge are merely guesses) is, i think, far from trivial.

  5. [Continuation of discussion from here]

    Hi Sam, what is a "degree of certainty"? On the standard view, we have degrees of belief, between 0 - 1, with each endpoint representing certainty (that the proposition is false or true, respectively), and everything in between being some non-certain credence. (E.g. I might just be 99.9% sure that I truly have hands. Nothing you've said implies that this credence is unreasonable.)

    The mere fact that we "might", possibly, lack knowledge is indeed trivial. Of course it's logically possible that our beliefs are false -- who would ever have thought otherwise? But it doesn't follow that this possibility is remotely likely. For all that's been established, our beliefs might be very safe indeed: true in all close possible worlds.

    Truth is easy enough to come by (at least in principle, there's nothing barring us from having true beliefs). That's why the skeptic traditionally claimed that knowledge was much harder (even impossible) to come by: even if our belief happened to be true, they still wouldn't qualify as knowledge -- or so argued the Cartesian skeptic. But once you retreat from the strong skeptical view, you end up having to concede that there's nothing barring us from having knowledge, either. So there goes the radical skeptical threat. All that's left is our ordinary fallibility.

  6. In saying that we cannot have any degree of certainty about almost anything I was saying that according to the sceptic our knowledge claims are akin to guesses out of a wide range of possibilities. I'm not going to enter the realm of scepticism in regards to the veracity of reason itself, but the beliefs that reason manipulates are, I think, in doubt. Most beliefs are not isolated premises, rather they are conclusions built up from other, more fundamental, knowledge claims. To argue that a belief that is the result of an enormously long reasoning process based on many dubious beliefs is likely to be true does not make sense to me. I think there is something (other than a lack of indubitably reliable information) barring us from having true beliefs, and that is the probability of randomly (i say randomly here because of the unreliability of the premises/evidence in the reasoning process) choosing a true belief. For every true belief, say Barack Obama was elected president, there are many correspondingly false ones, say John McCain or Hillary Clinton was elected president. The doubt about the truth of our more fundamental knowledge claims, as well as every successive claim built from these, effectively makes claims that are much further down the argument chain random guesses against huge odds. This is far different from the ordinary fallibility that would hold in a world immune to sceptical doubt.

  7. *should have ready "john mccain or hillary clinton or... etc."

  8. I see. But what is the argument for thinking that "our knowledge claims are akin to guesses"? Guesses are paradigmatic unsafe beliefs -- beliefs that could very easily be mistaken. But the realist externalist does not think this is true of our everyday beliefs (if my beliefs are true at all, then things would have to be radically different for me to have been a BIV instead!), and you've offered no argument to the contrary.

    You claim that our foundational beliefs are "unreliable", but again, what's your basis for this claim? They could be as safe as safe can be, short of being absolutely certain. They are "dubious" only in the weak sense that they can be doubted - they are possibly in error - but that's compatible with their being 99.999...% sure.

    I suspect you are implicitly making some internalist assumptions, to the effect that if we can't show our beliefs to be justified then they are thereby unjustified (we have no reason to assign them greater credence than any competitor). But that seems question-begging in this context, since you are simply assuming the falsity of externalism from the start, rather than offering any independent argument against the view.

  9. Sorry Richard, my use of the word "unreliable" might have been a bit confusing there. In calling beliefs unreliable i did not mean to imply that they are false beliefs, rather only that we have no method that we can be certain justifies our belief in them. As Aidan mentioned, the possibility that you have knowledge does not defeat the (properly defined) sceptic, at most it offers an alternative position. I'm not quite sure whether many people would try and defend the strong sceptic position that knowledge is impossible, as that claim surely constitutes an item of knowledge. As to which position, sceptic or realist, is more rational to hold, i would tend to favour the sceptic as his position is potentially simpler. Occam's Razor states that one should not needlessly multiply entities, and the whole universe of entities required by the realist is surely more complex than the simplest sceptic scenario.


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.)