Some Further Reflections

Abstract

Unfortunately, the commentaries from Covarrubias and Roca equally missed the central point of my target paper, namely, the futility of discussions about psychology’s proper subject matter. Their commentaries only circled around the approach, without commenting on it directly. In these further reflections, I reiterate and elaborate in more detail some aspects of that point. Specifically, regarding Covarrubias’s comment, I reflect on some apparent ontological confusions about properties and mental representations, clarifying that they are not the same. I also clarify that representationalism is compatible with cognitive neuroscience, the identity theory, connectionism, and even the ecological Gibsonian approach. However, I end up rejecting representationalism as conceptually obscure and, with it, all debates over representations as unintelligible. Instead, as an example of the kind of eliminative connectionism I pursue, I show a contingency learning simulation with a neural network model that does not posit representations, information processing, or even internal causal mediation. Regarding Roca’s comment, I point out his stubbornness in accusing mentalistic psychology of being dualistic, reiterating the reasons why this accusation is patently false. I also identify several fallacies in his argument against neural network models. In particular, as anticipated in my reflections on Covarrubias’ commentary, no computational model requires the postulation of mechanical causal mediation.