Davidson, Skepticism and the Pragmatics of Justification



This paper is concerned with Davidsonís argument that very general properties of the theory of interpretation make the skeptical claim that most of our beliefs could turn out to be false insupportable. Conceived as a 'straight' answer to the global skeptic (that is, as a proof that we do, in fact, have true believes about the world around us) Davidson's argument is not especially convincing.[1] In particular, Davidsonís answer to the skeptic presupposes a framework that allows for a new and seemingly more radical skepticism according to which we might not even have beliefs at all.[2] Nevertheless, there is a sense in which Davidson's account of content remaps the conceptual terrain in a fashion that absolves us of the need to rule out the scenarios the global skeptic describes. The following will both present the problems Davidsonís position has as a ëstraightí solution to skepticism, and discuss the way in which his ëinterpretationalí externalism does weaken the strength of the skeptical challenge.

The details of Davidson's position (particularly his claims that interpretation presupposes massive agreement, that talk of an omniscient interpreter is unproblematic, and that a 'Swampman' would have no thoughts) are not uncontroversial, but I will accept them for the sake of argument here. The purpose of this paper is to investigate the status of skepticism within a Davidsonian framework, not to evaluate the merits of the framework itself.



Davidson claims that a proper understanding of the workings of the interpretation process rules out the possibility of radical skepticism because interpretation presupposes massive agreement between the interpreter and the interpretee.


We can make sense of differences all right, but only against a background of shared belief. What is shared does not in general call for comment; it is too dull, trite, or familiar to stand notice. But without a vast common ground, there is no place for disputants to have their quarrel .... The basic claim is that much community of belief is needed to provide a basis for communication and understanding.[3]


One might think that the mere fact that our interpreters must share most of our beliefs provides no assurance that these beliefs will be true, since it seems that the agreement could be produced by a large common ground of mistakes about the world. Davidson, however, claims that agreement insures general correctness because even an omniscient interpreter would interpret us as having beliefs which were mostly true by his own lights.[4] Beliefs which are correct by the omniscient interpreterís lights must be correct by objective standards, so most of our beliefs must be true.[5]

However, even if Davidson has managed to show that "massive error about the world is simply unintelligible,"[6] it is not at all clear that he has really addressed the skepticís worries. The conclusion of Davidsonís argument is not that we must have mostly true beliefs, but rather that we cannot have mostly false ones, and these need not amount to the same thing. While a clear way not to have mostly false beliefs is to have mainly true beliefs, another way is to fail to have beliefs (be interpretable) at all. (A rock, for instance, cannot be "massively mistaken.") Davidson presupposes that the omniscient interpreter will be able to come up with an interpretation of us, and then argues that most of the beliefs assigned to us on such an interpretation must be true, but it is precisely Davidsonís presupposition that the skeptic may deny. How can we be sure, the skeptic will ask, that the omniscient interpreter will be able to assign any beliefs to us at all? To be interpretable, our utterances must be reliably correlated with various objects in the world around us, but the presence of such a reliable correlation is precisely what the skeptic questions.[7] It certainly may ëseemí to us that we have beliefs, but this alone is not enough, at least within Davidsonís framework, to assure us that we do.

Indeed, Davidson himself makes this extremely clear in his notorious discussion of "Swampman":

Suppose lightning strikes a dead tree in a swamp; I am standing nearby. My body is reduced to its elements, while entirely by coincidence (and out of different molecules) the tree is turned into my physical replica. My replica, the Swampman, moves exactly as I did; according to its nature it departs the swamp, encounters and seems to recognize my friends, and appears to return their greetings in English. It moves to my house and seems to write articles on radical interpretation. No one can tell the difference.

But there is a difference. My replica canít recognize my friends; it canít recognize anything, since it never cognized anything in the first place. It canít know my friendsí names (though of course it seems to), it canít remember my house. It canít mean what I do by the word ëhouseí, for example, since the sound ëhouseí it makes was not learned in a context that would give it the right meaning -- or any meaning at all. Indeed, I donít see how my replica can be said to mean anything by the sounds it makes, nor to have any thoughts.[8]

How, the skeptic may ask, can we be sure that we are not like the Swampman just described, stuck with apparent thoughts and memories which lack any real content? Davidsonís argument may show that, to the extent that we have contentful beliefs, most of those beliefs will be true, but the framework needed to establish this conclusion undercuts or assurance that have any beliefs at all. The general problem of the connection between our mind and the world remains. In Descartesí case, it expressed itself in the worry that all of our beliefs might be false. The same problem expresses itself in Davidsonís case in the form of the worry that we might fail to have beliefs at all. Rather than answering the skeptic, then, Davidsonís externalism seems to lead to a type of skepticism at least as, if not more, disturbing.[9]



However, externalism strengthens skepticism only in the sense that, if both externalism and its skeptical hypothesis were true, we would be worse off than we would be if internalism and its skeptical hypothesis were true (assuming that having false beliefs is better than having no beliefs at all). Skepticism is not, however, strengthened in the sense that the skepticís challenge needs to be taken more seriously. Indeed, it will be argued below that quite the opposite is the case.

It may initially seem as if externalism merely moves the skeptical bump in the rug to an even less appealing location, but the new location actually has many advantages over the previous one. Even if the externalist canít prove that we arenít thoughtless swampmen etc., this does not mean that nothing has been gained. One is still left with a picture in which our beliefs are generally veridical. The point is not to provide an argument that shows that the skepticís scenario couldnít happen (or that it is unintelligible etc.), but rather to understand our semantic and epistemic concepts in a way that the possibility of its happening does not undermine our claims to knowledge in other cases. The relevance of content externalism is not just a matter of shifting the bump in the rug. Rather, it amounts to a shift to a type of conceptual bookkeeping in which things like knowledge can find a more secure place.

Doing so involves cutting the philosophically relevant joints in a way that separates cases like the Swampman (who purportedly knows nothing about the world around him) from those in which we do know about the world around us. Davidsonís account cuts the joints in such a way that the cases of genuine knowledge are conceptually isolated from the skeptical scenarios, (the knower has beliefs with a certain content, the Swampman lacks beliefs entirely) while internalist theories do not (the knower and the person deceived by an evil demon have beliefs with the same content).

There is, after all, no practical imperative to answer the skeptic. We will continue to act the way we do whether we have an answer to the skeptic or not. The felt pressure to answer the skeptic is of a different sort. We feel more of a rational imperative to answer the skepticís challenge to our beliefs. The real interest of skepticism is the possibility that the rational grounds for all our (empirical) beliefs could be investigated and turn out to be unjustified. Treating something as a belief at all presupposes that it is subject to rational evaluation, criticism and justification.[10] Many of the beliefs we have (even true beliefs) can turn out to be found wanting when subject to such evaluation. Such beliefs, even if true, cannot (according to a familiar line of thought) count as knowledge, since knowledge is, at least, justified true belief.[11] The skeptic takes this process, which is presupposed by the notions of belief and knowledge themselves, and argues that, in fact, our beliefs (about a certain domain -- e.g. the external world) never pass muster (and thus that we have no knowledge in that domain). The skeptical scenario serves as a general recipe for showing that no such beliefs can be justified.

f the internalist is right in thinking that our thoughts are constitutively independent from our environment, then the rationality requirements would be binding upon us even if we were swampmen, etc., so these skeptical scenarios would have to be understood as ones in which the rationality constraints were not satisfied.[12] On the other hand, if thoughts are not autonomous (as the externalist suggests), then the rationality requirements would not be binding on swampmen, since the requirements are only binding upon things that actually have beliefs. The skeptic is engaged in the process of rationally evaluating our thoughts, and this process presupposes that there are thoughts to be evaluated. If the externalist is right, and the skepticís scenario is one where there are no thoughts to be evaluated, then skeptical doubt involves itself in a type of ëpragmatic contradiction.í The skepticís practice presupposes the very phenomena he attempts cast into doubt.[13]

Externalism pushes the comparatively moderate skeptical claim that most of our beliefs could be false to the more radical claim that we do not have any beliefs at all. However, by ëstrengtheningí skepticism in this fashion, externalism easier to ignore.

Of course, in our practical life, skepticism is always easy to ignore. So perhaps one should say that externalism makes it easier to legitimate this everyday practice. Of course, if one were not practically disposed to ignore skepticism and were plagued by the ëvertigoí such skeptical scenarios can induce in some, an externalist move will not help one. Such vertigo calls for the type of ëstraightí solution which may never be forthcoming.

The pragmatic contradiction involved does not, of course, presuppose that the skeptic and his interlocutor are separate people. An actual skeptic could be accused by his interlocutor of assuming that there is an external world, other minds, etc., simply in virtue of the fact that he takes himself to be engaged in a debate with another person. The dialog itself presupposes that there exists someone other than the skeptic himself. This familiar kind of pragmatic contradiction does not arise when one, as Descartes did, ëinternalizesí the skeptics position when one is evaluating oneís own beliefs. The position defended above, suggests, however, that even in such first-personal meditations, one is commited to the general veridicality of oneís beliefs.

Davidsonís position can be understood, then, as entailing that the general veridicality of our beliefs about the world must be presupposed by our practice of giving and asking for reasons, and so cannot be taken to be at stake in such a practice.[14] Skeptical doubts are ëidleí not just in the sense that we can never act upon them, but also in the sense that their rejection is presupposed by the very normative practices that they grow out of.[15]




1. Which is perhaps a good thing, since many would consider its ability to provide such a quick refutation of skepticism to be a reductio of any theory of meaning. (See Falvey and Owens, "Externalism, Self-Knowledge, and Skepticism," Phil Rev 1994, and McGinn, "Radical Interpretation and Epistemology" in LePore (ed.) Truth and Interpretation, for claims of this sort)


2. For a criticism of Davidson of this sort, see P. Kleinís "Radical Interpretation and Skepticism" in LePore (ed.) Truth and Interpretation, and Stich "Might Man be an Irrational Animal" in Kornblith (ed.) Naturalizing Epistemology, (Cambridge: MIT 1994) p. 356.


3. "The Method of Truth in Metaphysics", p 200. See also "Just as too much attributed error risks depriving the subject of his subject matter, so too much actual error robs a person of things to go wrong about." (ibid.)


4. While Davidson appeals to the possibility of an omniscient interpreter, omniscience is probably more than what is needed on the interpreterís part (and it may lead to other issues of whether an omniscient being would feel any need to interpret others the way we do). The (mere) requirement that the interpreter have, as a matter of fact, only true beliefs should be enough for Davidsonís argument.


5. See "A Coherence Theory of Truth and Knowledge" p. 317; "The Method of Truth in Metaphysics", p.201.


6. "The Method of Truth in Metaphysics", p.201.


7. This point is stressed by Klein.


8. "Knowing oneís own Mind", pp.443-444.


9 The further possibility that Davidsonís position leaves us with the possibility of their being a skepticism about the content of our beliefs will not be dealt with here, though Davidson himself argues that such wories about our self-knowledge are unfounded in his "Knowing oneís own Mind" and "First Person Authority".


10 See Sellars Empiricism and the Philosophy of Mind, Brandom Making it Explicit, McDowell Mind and World.


11. At least when one presupposes an ëinternalistí conception of epistemic justification. If one presupposes an externalist one, where justification is understood in terms of the reliability of the relevant belief-forming processes, such skeptical scenarios do not threaten justification.


12. Once again, assuming that one is working with an internalist sense of justification.


13. One can find oneself in a similar position vis a vis the problem of free will. We do have a practice of inquiring whether certain agents are (should be held) responsible for their actions. Agents who are responsible for their actions should be punished or rewarded, those who are not, should not. Just as the skeptic tries to extend our practice of rational criticism in such a way that none of our beliefs turn out to be rationally justified, the incompatibilist extends our practice of evaluating the actions of others in a way which requires that no one is responsible for anything. This is often taken to suggest that we are never justified in punishing others since no one is responsible for their actions. Unfortunately, the entire process of evaluating whether others should be held responsible for their actions presupposes at least some cognitive or ethical responsibility on our part (the concern is with what we should do). If no one were responsible for their actions, then neither would we, in which case the evaluations would have no point. Questioning whether our reactive attitudes are justified presupposes that we can be held responsible for them, so any argument attempting to show that no one is responsible for anything will involve a type of pragmatic contradiction (if it is meant to show that our reactive attitudes are not (either ethically or cognitively) justified).


14. Such radical skepticism is not, then, a 'relevant alternative' that the thinker must rule out to legitimate his claim to knowledge. If we assume that to be entitled to make a claim to knowledge, the claimer must rule out all relevant alternatives, the skeptic can be understood as trying to come up a type of ëuniversalí alternative that rules out all claims to knowledge. Davidsonís positions suggests, however, that such universal alternatives cannot be treated as relevant, since they are already ruled out by the procedure of asking for justification itself. Swampman is not a relevant alternative since it is ruled out by the presuppositions of the discourse.


15 "The point is not that we must accept this conclusion in order to avoid skepticism, but that we must accept it in order to explain the existence of the conceptual scheme in terms of which the skeptical problem is stated. But once the conclusion is stated, the skeptical problem does not arise." (Strawson, Individuals, 106.)

Back to Henry Jackman's Home Page