The Explanation Approach to Delusion

Keywords
delusion, explanation, experience, two-factor theories

We are grateful for Professor Coltheart's incisive discussion of our paper and important suggestions about how an explanation account of delusion might answer the problems we raise. To further clarify the issues, we consider the replies Coltheart offers to each of the six difficulties he discusses and indicate points of agreement and disagreement.

Before turning to these six issues, however, we rehearse an important general puzzle raised by explanation accounts of delusion. A theory of delusion must explain how delusions arise and why they are not revised. In the model of Coltheart et al., the second question is addressed by positing a deficit in belief evaluation. In the case of the Capgras delusion, the first question is addressed by saying that the delusion arises as an attempt to explain an anomalous thought¹ of the form "There's something odd about this woman." The puzzle is this: why is the patient's explanation so poor? This poverty of delusional explanation can be seen in all of the case studies mentioned by Coltheart, including the case of Capgras:

1. Mirrored-self misidentification, Case T: The patient's belief that his mirror reflection was not him is supposed to have arisen as a result of an ignorance about the function of mirrors. If so, however, it seems vastly simpler and more natural to suppose that the mirror is like a TV screen on which images (including images of himself) can be shown. This is compatible with an ignorance about mirrors and does not require positing the existence of someone qualitatively, but not numerically, identical to the patient.
2. Mirrored-self misidentification, Case F: Here there is no ignorance about mirrors but rather a face perception deficit. Given that the patient correctly identifies other people's reflections as reflections, why does he not suppose that his appearance has changed rather than that the mirror fails to reflect it? This is surely more conservative than the explanation that the mirror changes its function when he is in front of it.
3. Somatoparaphrenia: Here the violation of conservatism in explanation is clear. Inability to move a limb is compatible with any number of more or less plausible hypotheses (including the true hypothesis of illness—always in principle available in delusion). In the face of clear perceptual evidence to the contrary, the idea that the limb belongs to someone else is surely one of the least plausible.
4. Capgras delusion: Despite an initial plausibility, the idea that the Capgras delusion is a candidate explanation of an anomalous thought of the form "There's something odd about this woman" does not seem plausible. This thought is hypothesized to arise when an unconscious process detects a mismatch between a predicted event (autonomic arousal to the spouse) and the actual event (a lack of arousal). If the wife were a convincing impostor, however, there should be nothing odd about her. The hypothesis that she is an impostor, therefore, does not explain why there is something odd about her. The explanation seems to miss the mark entirely.

   There are, of course, reports of patients with Capgras who have subtle visual disorders as well (Young, Leafhead, and Szulecka 1994), and here the impostor hypothesis is more plausible. If there were something odd about someone, and they looked quite a lot—but not exactly—like one's wife, they might indeed be an impostor (though, again, other more natural hypotheses would surely come to mind). In these cases, however, it is not clearly the thought of oddness that is required to give a clue to the identity of the impostor, but the simpler fact that the impostor does not look convincingly like the original.

There are two senses in which the putative explanations of patients with delusions are bad: either they are not obviously empirically adequate—they do not, that is, really explain the anomalous thought at all—or they explain the anomalous thought in...