Skepticism and Rationality
Part three of this series dealt with the objection that knowledge of the external world is impossible, depending on a Reliabilist account of knowledge to rob the skeptic’s conjecture of its deletorious consequences. Yet I doubt the skeptic is fully satisfied. There remains another concern, that even if we have knowledge of the external world, we could never be in a position to rightfully claim that knowledge. Though we base our knowledge claims off of a conception of the world upon which our beliefs would be true and reliably formed, we are not entitled to that conception in the first place: a stubborn irrationality is at the heart of our knowledge claims. So let’s look at rationality.
It is common to think of rationality as a matter internal to a subject. In one way this is true, and in another it is not. It is true that rationality concerns the formation of beliefs, and that these are internal. It is true that beliefs are often formed on grounds which are likewise internal and accessible to the subject. Further, we allow that a subject may be rational despite drawing the false conclusions about the world, even using false beliefs to draw them, and so in this sense a lack of external correspondence does not matter. On the other hand, the principles of rationality by which we judge subjects are presumed to be objective and universally applicable, and it is this which gives them an authoritative power1. A subject may very well think they are rational, but fail to be. Yet to be objective in this way, rationality must derive from something external to subjects. Rationality, like knowledge, needs be externally defined.
There are generally understood to be two kinds of rationality: practical and theoretical rationality2. The former is concerned with the evaluation of states of affairs or objects in accordance with one’s first-order desires, a process which generates second-order desires, and thus subsumes means-ends rationality. The latter is concerned with truth, and the epistemic practices whose observance are likely to lead one to true beliefs. Both kinds of rationality have an aim – theoretical rationality aims to ensure one’s beliefs are mostly true, and practical rationality aims to ensure that one’s desires are mostly satisfied – and both kinds of rationality depend upon the external world – to yield mostly true beliefs an epistemic practice will have to be reliable3, and to correctly derive second-order desires from first-order desires the world will have to exhibit the sort of causal structure which would relate these second-order desires to the satisfaction of first-order ones. However, it is theoretical rationality that is relevant here, so I shall say no more of practical rationality.
Given that theoretical rationality is concerned with the epistemic practices which are likely to lead us to true belief, what epistemic practices would those be? Well, reasoning according to laws of classical logic would almost certainly4 count as a rational epistemic practice. Such laws are necessary, they hold in all possible worlds, therefore using them is a reliable epistemic practice in any world you like. Other practices may be more parochial – perhaps inductive reasoning is a epistemic practice which fails in some worlds, but is rational in this world, because this world contains natural kinds which make the inference from past events to future events much more secure than it otherwise might be. Still other epistemic practices are obviously world-relative, and included amongst these will be the evidentialist practice of taking experiences as prima facie evidence of how the world is. This practice fails to be reliable, and thus rational to use, in worlds where a skeptical hypothesis is true. But conversely, it is both reliable and rational at the actual world, if the actual world be one where skeptical hypotheses are false. Since it is primarily by taking our experiences in this way that we arrive at our conception of the world, it turns out that, if the world is roughly as we think it, that general conception of the world will be rational too, and with it any knowledge claims that the conception supports. The upshot is that an externalist notion of rationality can do the same job on the skeptic as our externalist account of knowledge.
True, none of this really refutes a hardened skeptic. What we’ve done instead is to frame the debate in terms which guarantee that one side must beg the question against the other, if they wish to argue for their position at all. Is something more satisfactory in the offing? I believe so. For, whereas externalism concerning knowledge does not have consequences for the skeptic’s position, as the skeptic does not claim knowledge, externalism concerning rationality does. The skeptic claims rationality, and this opens the way to rebuttal. Here is an argument to that end:
1. An epistemic practice is (theoretically) rational only if following that epistemic practice is more likely to lead to true than false beliefs.
2. A belief is rational iff it is derived according to an epistemic practice which is itself rational. It is otherwise irrational.
3. If a skeptical hypothesis is true (e.g. that I am a brain-in-a-vat), then there is no rational epistemic practice by which I could derive a belief that a skeptical hypothesis is true.
4. So, if a skeptical hypothesis is true, then a belief that a skeptical hypothesis is true is irrational.
5. If no skeptical hypothesis is true, then equally there is no rational epistemic practice by which I could derive the belief that a skeptical hypothesis is true.
6. So, if no skeptical hypothesis is true, then a belief that a skeptical hypothesis is true is irrational, too.
7. Therefore, belief in a skeptical hypothesis is irrational simpliciter.
8. Belief in a skeptical hypothesis is requisite for belief in skepticism.
9. A belief for which an irrational belief is requisite is itself irrational.
10. Therefore, a belief in skepticism is irrational.
Premise 1 states a necessary condition for theoretical rationality. If the aim of rationality is to ensure that one’s beliefs are mostly true, and if it is epistemic practices which are to be evaluated, then any epistemic practice whose observance would fail to do so is trivially irrational. 2 extends the honorific of ‘rational’ from epistemic practices to beliefs – the rationality of beliefs depends not on their content, but how we come by them, which is to say that their rationality depends upon the practices we follow which lead us to believe them. Premise 3 indulges the skeptic. Here it is agreed with the skeptic (for the sake of argument) that we cannot even say our commonsense view of the world is more probable than a skeptical alternative, meaning that we have no rational epistemic practice to favor it over the skeptical hypothesis. Of course, the reason given for this is that our evidence would be the same in either case. But this too implies that no epistemic practice could favor a skeptical hypothesis over our commonsense view. What’s sauce for the goose is sauce for the gander.
Premise 4 is a simple deductive consequence of 2 and 3. Number 5 adds that, in the case all skeptical hypotheses are false, there is no epistemic practice which would both yield beliefs more likely to be true than false, and from which I could derive the belief that such a hypothesis is true. Like 3, this premise too can be defended on the grounds that it is something a skeptic should accept – if skeptical hypotheses and commonsense hypotheses really are evidentially equiprobable, then there are no rational epistemic practices which would favor one over the other, thus there is no epistemic practice which could favor a skeptical hypothesis over the commonsense view. 6 is deduced from 2 and 5, whereas 7 states that, in any case, the belief in a skeptical hypothesis is irrational.
Premise 8 reaps the rewards of our externalist account of knowledge. Having knowledge requires that one’s belief be true, and that it be produced by means which are reliable in the circumstances. But, as we have seen, the satisfaction of these conditions isn’t something that can be assessed without some picture of what the world is like – that is, our metaphysics is logically prior to the evaluation of knowledge claims. Thus any doctrine such as skepticism which presumes to tell us the extent of our knowledge must presuppose an understanding of the world, and in the case of skepticism must presuppose some skeptical scenario on which we would lack knowledge of the external world. So belief in the doctrine of skepticism, properly understood, requires belief in a skeptical hypothesis. 9 asserts that any belief which must be founded on irrationality is irrational also. That is, if the only reason for believing in X is Y, and Y is irrational, then X is irrational, too. One irrational belief cannot be made to raise another belief into rationality. However, if one believes in X on the basis of Y, but X is rational for some reason other than Y, then Y is not requisite in the sense of 9 – X can derive its rationality from elsewhere. Lastly, 10 states our conclusion, that belief in skepticism is irrational. This amounts to a rebuttal, but not a refutation, since a refutation of skepticism would show it to be false.
Even if the above argument is sound, it is not a complete victory. Showing that skepticism is irrational does not prove that the commonsense view is rational, for both may be ultimately irrational – it may be the case that one of the skeptical hypotheses is true, and then belief in anything, save logical truths, will be irrational. Yet the argument does assure us that we have no purely philosophical reason to doubt the rationality of our common epistemic practices, and thus none to doubt a broadly commonsense conception of the world. If that’s as good as it gets, then I’m quite satisfied5.
1 Perhaps you disagree – you think that “rational” is simply an arbitrary label we give to those whose thinking is commendably like our own. I doubt that, but if it were true, the skeptic’s charge of irrationality would have no force. It would merely be his opinion.
2 Spohn, The Many Facets of Rationality.
3 This is not to say that a true rational belief counts as knowledge: the standard of reliability for the former is intuitively lower than the latter. I leave open the question of just where the line for each should be drawn.
4 Almost certainly, since there is the lurking possibility that classical logic may not apply to quantum phenomena, and that some other logic may be demanded for the preservation of truth in reasoning about the world. But this is still a matter of controversy, and even if it does apply to quantum phenomena, classical logic will lead us to likely true beliefs in the usual cases, where our reasoning concerns dependable middle-sized objects.
5This is the last post in this series.