Explaining delusional beliefs: a hybrid model

ABSTRACT

Introduction: In this paper we present and defend a hybrid theory of the development of delusions that incorporates the central ideas of two influential (yet sometimes bitterly opposing) theoretical approaches to delusions—the two-factor theory and the prediction error theory.

Method: After introducing the central ideas of the two-factor theory and the prediction error theory, we describe the motivations for our conciliatory project, explain the theoretical details of the hybrid theory we propose, and answer potential objections to our proposal.

Results: According to the hybrid theory we advance, the first factor of a delusion is physically grounded in an abnormal prediction error, and the second factor is physically grounded in the overestimation of the precision of the abnormal prediction error. Against anticipated objections, we argue that the hybrid theory is internally coherent, and that it constitutes a genuine hybrid between the two-factor theory and the prediction error theory.

Conclusion: A rapprochement between the two-factor theory and the prediction error theory is both possible and desirable. In particular, our hybrid theory provides a parsimonious and unified account of delusions, whether monothematic or polythematic, across a wide variety of medical conditions.

KEYWORDS:

• Delusions
• belief formation
• two-factor theory
• prediction error theory
• cognitive neuropsychiatry

Acknowledgements

We received helpful feedback at the seminars and conferences in which early versions of this paper were presented, including the PERFECT Reading Group at the University of Birmingham (March 2017), the Philosophy, Psychology, and Informatics Group at the University of Edinburgh (March 2017), The 31st International Congress of Psychology at PACIFICO Yokohama (July 2016), and the Delusions Lunchtime Seminar at the University of Birmingham (June 2015). We also thank two anonymous referees for their insightful comments and suggestions. Some parts in this paper come from revising Section 4.4 of Delusions and Beliefs: A Philosophical Inquiry. Routledge: Abingdon and New York. © 2019, Kengo Miyazono. Kengo Miyazono acknowledges the support of JSPS KAKENHI (grant number 16H06998 & 18H00605). Ryan McKay acknowledges the support of the Cogito Foundation (grant number R10917).

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 As Coltheart and colleagues have often characterised the first factor in their scheme as a prediction error (e.g., Coltheart et al., 2011), contrasting the two-factor account with the “prediction error” theory may be misleading. Nevertheless, we use this terminology to be consistent with much commentary about the two theories.

2 Following Coltheart et al. (2010), we use the phrase “abnormal datum” rather than “abnormal experience” because the former is more neutral as to whether the data are consciously accessible or not. We assume that the abnormal data can be, but need not be, consciously accessible. Conscious or not, we take “abnormal datum” to be a psychological notion that belongs to the psychological (rather than neurophysiological) level of explanation. This will be crucial later in the discussion of our hybrid theory, which connects the two-factor theoretic explanation, which belongs to the psychological level, with the prediction error theoretic explanation, which belongs to the neurophysiological level.

3 The hybrid theory we advance is not the only way of combining the two-factor theory and the prediction error theory. Other hybrid theories are possible (e.g., see endnote 4; see also Gadsby and Hohwy (forthcoming) for a related discussion).

4 According to another proposal, the second factor is a bias of discounting or neglecting evidence against adopted delusional hypotheses (Coltheart et al., 2010). This proposal can also be combined with a (top-down) prediction error story to form another hybrid theory. For instance, the bias of discounting or neglecting counterevidence might be physically grounded in the underestimation of the precision of prediction errors; i.e., the failed predictions from delusional hypotheses.

Additional information

Funding

This work was supported by Cogito Foundation [grant number R10917]; Japan Society for the Promotion of Science [grant number 16H06998, 18H00605].