In a couple of places, e.g. here and here I’ve suggested that Ryle’s argument against Descartes relies (illegitimately) on general sceptical principles about induction. I now think that’s something of a mistake. (As Michael Kremer was trying to point out to me at the time.)
I still think it is true that Ryle is a little careless in how he formulates the argument against Descartes in chapter 1 of The Concept of Mind. That chapter certainly suggests that Descartes is illegitimately inferring from effects to causes, even though there are other possible explanations of the physical phenomena around us. And that would be a crazy argument; inference to the best explanation need not be inference to the only explanation. So crazy, you might think, that I’m very uncharitable to attribute it to Ryle.
Ryle does not object to this general method of argumentation, but to Descartes’ particular application of it. This becomes quite clear in the discussion of signal boxes from page 52 to 54. Ryle seems to hold the following two (quite plausible) general principles.
- To explain E in terms of C, we must know the laws, or at least the lawlike generalisations, that relate C and E
- To inductively infer from from knowledge of other causal connections that C is the cause of E, we must either have something like an exact duplicate of the (purported) C-E link, or a vast number of cases somewhat similar to it
Ryle then argues (on page 52) that many people who know nothing of how non-physical souls could cause physical action, let alone the laws regulating such interactions, know a lot about the minds of Napoleon, Euclid and assorted footballers and chess players. So their knowledge cannot be grounded in an inference to the best explanation, since they do not know how such an explanation could go.
Even if it is granted that we know how mind and body are connected in our own case (not something Ryle is disposed to grant) we only have one case, and given the diversity of minds, it probably won’t be like Napoleon, Euclid and our assorted footballers and chess players. So our knowledge of other minds can’t be based on enumerative induction or analogy.
This more detailed argument actually does seem like a good argument against the possibility of knowing about other minds given Cartesian dualism. So I should not have been so quick to accuse Ryle of general epistemic crimes.
Of course, everyone reading this probably knew this already…