Aline, Gloria and I met together to discuss John Campbell’s 2001 paper ‘Rationality, meaning and the analysis of delusion’. As a result I think I have a clearer picture of why I disagree with it than before. The following are thus not my own thoughts (though neither do I want to saddle - my expression of - them onto either of my colleagues).
Campbell’s paper divides between two phases of argument. In the first, he deploys a Davidsonian link between meaning and rationality to press problems with the interpretation of characteristic expressions of the Capgras delusion. The characteristic type of utterance – “That woman is not my wife!” - is insufficient to specify an interpretation since that type of utterance, freed of context, might be used to flag a mere discovery of illegality in the wedding ceremony. (I worried that by the end of the paper, Campbell had lost the right to this refutation but Aline pointed out that the proof of the pudding was in the lack of acceptance by the Capgras subject of this interpretation.) But the most plausible interpretation – “This [demonstrated] woman is not that [remembered] woman” - fails because the subject fails to deploy paradigmatic or canonical forms of checking. They do not do what they ought to do to check such a thought. Given the meaning-rationality link, this apparent failure of rationality undermines such an interpretation.
The positive phase aims to respond with a suggestion about the shape, at least, of the thought-content involved. The very fact of the failure to adopt paradigmatic checking strategies would be rational if the delusion had the status of a (third period) Wittgensteinian hinge proposition. So the form of the thought can be identified even though not the content (in G’s helpful terminological analogy though it will conflict with my philosophy of content use of ‘content’, sorry).
It thus seems as though the negative phase presses a failure of rationality and hence a failure of interpretation in accord with the meaning-rationality link. The positive phase suggests a ‘sort of’ rationality and hence a ‘sort of’ interpretation that reaches only as far as the form, not the content, of the thought.
I think this summary of what is going on enables me to state my earlier worry in slightly different terms. If one takes the negative phase seriously, the positive phase is unavailable. It, the positive phase, says of we-know-not-what content that it has the form of a hinge proposition. But surely the form-content distinction is an abstraction from the motley of thoughts subjects have rather than something independently understandable. (To be so would require prior commitment to something like the representational theory of mind in a strong sense of it being independently characterisable and understandable as the a priori engineering of minds rather than merely a post facto explanation of how they could be possible.) Given the meaning-rationality link (ie the constitutive ideal of rationality underpinning interpretation through the principle of charity) and given a plausible additional claim that rationality is not codifiable, drawing a distinction of form and content would require first articulating thoughts as a piece of radical interpretation and then abstracting forms and contents (from thought-contents, I want to say, with a danger of ambiguity). So if that could be done, one would then be in a position to say that delusions have the form of hinge propositions. But the first phase of Campbell’s paper is an argument against the possibility of that necessary preliminary work.
To assume access to the form of delusions without their contents is to assume something about the shape of their intentional content (in the usual philosophy of content sense of that word). Aside from abstraction from the output of radical interpretation, the only other route I can see to that would be to start with a picture of the internal vehicles of content and describe their functional roles. This approach faces a dilemma. If the functional roles presuppose the structure of rationality then the approach cannot sidestep the meaning-rationality link since rationality governs thought-contents. But if they are just any dispositional causal connections between inner vehicles of content (whatever the content turns out to be) then this threatens the connection between what is being described and our ways of making sense of one another since only some connections would be (so much as, as they say in Oxford) intelligible. There would need to be some reason to think that both structures kept in step. In the absence of that, such an approach seems a non-starter and hence the assumption that delusions are some sort of shaping of thought-content in the form of a hinge proposition despite being inaccessible to rational interpretation looks illicit.
Campbell, J. (2001) ‘Rationality, meaning, and the analysis of delusion’ in Philosophy Psychiatry and Psychology 8: 89-100