Davies and Egan (2013) have recently offered a doxastic and Bayesian account of delusions from the perspective of fragmentationism, the view according to which our belief system is fragmented, in opposition to it having a single structure or unified web. In the present paper I'll assess the explanatory potential of the account in light of an objection frequently raised to doxastic accounts of delusion: the Behavioural Objection. According to this objection, these accounts leave unexplained the fact that delusional subjects do not act following the behavioural profile that would be expected if they really believed the delusion. I shall conclude that, because of its commitments to Bayesian laws of belief formation and revision, Davies and Egan account leaves unexplained rather 1) the persistence of the delusional belief once adopted or 2) the inconsistent behaviour exhibited by the subject. My aim will then be to model delusions in another fragmented (Mandelbaum and Bendana forthcoming), non-Bayesian model (Mandelbaum 2019), and conclude that it can better account for the behavioural profile of delusions and its doxastic nature, since it takes into account the relation between the subject beliefs and motivational factors concerning the self.
Presented at the Fifth PLM Workshop on Delusion in Language and Mind, in Amsterdam, October 23---24, 2020