Philosophical Experiments

[Warning to the reader: the remarks that follow cite no-one and do not attempt to engage with details; but they are in part a response to the first two essays in this book.]

What is called "experimental philosophy" is diverse and does not admit of unified treatment. Some of it enlists the existing work of empirical scientists where it might be relevant to the questions of philosophy. While I may not always agree about the relevance, this seems innocent enough.

But there are more radical threads. One is a variety of "naturalism" that entails the complete rejection of a priori knowledge and non-empirically justified belief. Let the armchair blaze.

It is sometimes claimed by advocates of "naturalism" that the armchair method rests on a hopeless view of philosophy as conceptual analysis: what could one discover from the armchair, if anything, but the shape of one's conceptual space? (Can one discover even that? See below.) But whatever we make of the rest of it, the sociology of Williamson's recent book is sound when he denies that this view is orthodox. Many philosophers reject the conceptual analyst's account of a priori knowledge.

Some do not, of course, but others give other accounts, and there is a silent majority. Moreover, the pressure to countenance the a priori and to do so in a way that outstrips conceptual analysis or "epistemological analyticity" can be seen in the traditional problem of induction. According to a tempting principle,

One can be justified in believing p on the basis of evidence, q, only if one is independently justified in believing [if p then q].

This has the fairly rapid upshot that, if we are justified in believing things inductively, we must be non-empirically justified in believing contingent propositions. The argument may go wrong, but it must be faced by any honest attempt to live without the a priori or to confine it to the analysis of concepts.

This is the briefest sketch; I have not tried to say why the principle above should tempt us. But let the record state that there is an argument for a priori justification that has nothing to do with analyticity and everything to do with the threat of scepticism.

It is ironic, in this context, that the most baffling experimentalist project – the taking of surveys that elicit folk intuitions about such matters as knowledge, intentional action, and moral responsibility – would have a definite point if the content of our concepts was fixed by the corresponding dispositions, as some conceptual analysts believe. Witness the idiom, "folk concept of ____," as if one could make this theory true by stipulation. If the theory is false, we need some other incentive to care what the surveys say.

One of the reasons commonly given is that when we find that our intuitions are parochial, the beliefs that rest on them are undermined. But we should ask: what justifies that response? Perhaps the view that intuitions are evidence, akin to perceptual appearances. For if things look different to others, whose perceptual mechanisms we have no reason to question apart from the present discrepancy, that should give us pause. The problem is that we need not – should not – think of intuitions in that way. We could think of them, instead, as beliefs that are justified non-empirically, if at all, and not by the "evidence" of intuitive appearances.

Alternatively, the threat of parochiality might rest on a controversial view in the epistemology of disagreement, that we should give as much weight to the opinions of others as to our own unless we have antecedent reason to doubt their reliability. On the contrary, if some of my beliefs are justified a priori, quite apart from evidence, won't that give me reason to doubt the reliability of those who disagree, antecedent to – well, everything.

Again, these arguments may be wrong. But let the record state that inferences from surveys to the application of concepts or the justification of beliefs rely on hidden machinery: theories of concept-possession or epistemology disputed from the armchair, which stand in need of further defence. In its absence, the point of the surveys, however entertaining, is seriously opaque.

There is a final heresy, espoused by some, that is irrefutable by design: the surveys do nothing more, and need do nothing more, than map the cognitive powers by which "the folk" identify something as cause or effect, intentional action, exemption or excuse. There is no call to map them on to more familiar philosophical pursuits, as the strategies above purport to do. They stand on their own.

Still, we can ask for guidance. Why survey these particular questions? Why not study what people believe about just anything, or anything that has been a topic for philosophy: say, the meaning of life? It is no good responding that surveys are apt when philosophical problems are posed by "the basic concepts people use" or that an interest in such concepts is obviously philosophical: part of what is in dispute is how the study of folk beliefs relates to the study of concepts and their conditions of application.

Not everything is philosophy. But the boundaries of the field are controversial; and it would be wrong to refuse publication on grounds that reasonable philosophers dispute. There is, therefore, a standing risk that non-philosophy will be taught and studied as philosophy. This much follows from the proper humility of peer review and a wise refusal to police the borders. The borders should not be policed: these things must work themselves out. That doesn't mean it's philosophy, any more a heap of flesh and bones is a human being.