The Freudian unconscious and the Bayesian brain

This chapter presents a philosophical synthesis of a range of work relating psychoanalysis and neuroscience. The overall argument is (1) material in these and related fields can usefully be integrated via the notion of representation; (2) the appropriate notion of representation is a biological one common both to psychoanalysis and the Holmholtz/Bayes tradition in neuroscience; and consequently (3) the Freudian unconscious may be understood as realized in what is now described as the Bayesian brain. Psychoanalysis began as an extension of the intuitive commonsense psychology in whose terms we make sense of one another in everyday life. In this we understand one another as persons who have subjective experience of an objective world, and whose actions are informed by mental states and processes such as those of emotion, desire, belief, intention, and will. Such states and processes have intentionality: that is, they are about things and situations in the world, and they are capable of both truth (or accuracy) or falsehood (or inaccuracy). This ‘aboutness’ has historically been regarded as the defining feature of mental phenomena. Recent work in philosophy, however, has enabled us to see that the intentionality of the mental can be understood in terms of the notion of representation; and Ruth Garrett Millikan has argued that representation itself can be seen as a biological phenomenon which has evolved for the selective advantages it confers. So on this account the intentionality of mental phenomena can be understood as realized by the biological functioning of the brain. Such representational functioning is here explicated via the current Helmholtz/Bayes approach to neuroscience developed from work by Geoffrey Hinton and his colleagues. This approach has recently been advanced by Karl Friston in such a way as to unify a number of basic theories in neuroscience and to integrate them with evolution. This framework is particularly relevant for psychoanalysis, since Freud was trained in the school of Helmholtz and framed his early theories in terms of free energy; and this has enabled Friston and Richard Carhart-Harris to relate Freud’s claims to data from neuropsychology, neuroimaging, and psychopharmacology. The same framework can also be used to link Freud’s theories with the findings of affective neuroscience, including the homeostatic regulatory emotional and motivational systems described by Jaak Panksepp and Antonio Damasio, and the interoceptive system delineated by A. D. (Bud) Craig, and in conjunction with research in attachment and development. In this way we can start to describe the wishfulfilling Freudian unconscious as part of the predictive and error-minimizing working of the computational Bayesian brain.