Computationalism After Dynamicism and Enactivism: A Mechanistic Response

Abstract

This paper examines two influential objections to computationalism: the dynamical objection, according to which computation fails to capture the temporality of cognition, and the enactivist objection, according to which computational systems are heteronomous and therefore incompatible with the autonomy proper to cognitive systems. The aim is to show that these objections are not sufficient to rule out a computational characterization of the nervous system if the notion of computation is formulated in mechanistic terms. On this view, a computational system is understood as a mechanism whose function is to manipulate medium-independent vehicles according to a rule. Against this background, the paper argues that a mechanistic computational characterization of the nervous system is compatible with a characterization of it as a dynamic and autonomous system. This is because its mechanisms operate on neural vehicles whose functional role depends on temporally structured features, such as spike rate and spike timing, so that temporality becomes part of the individuation of the computational process itself. In turn, this characterization does not conflict with the autonomy of the nervous system, because the mechanistic notion of computation does not require heteronomy, a linear input-output organization, or externally imposed goals. The conclusion is that objections based on temporality and autonomy are not, by themselves, sufficient to exclude a mechanistic computational characterization of the nervous system.