I am grateful to John Schwenkler for giving me the opportunity to present my first book Cognitive Structural Realism, which aims to consolidate the ties between the philosophy of science and cognitive science. There already is some connection between these fields, given that the philosophy of cognitive science is a branch of philosophy of special sciences. My book though does not offer a philosophical treatment of some specific problems in cognitive science. Rather it draws on resources of the cognitive science and neuroscience to find a handle on some persisting problems of the general philosophy of science (i.e., the problem of scientific representation). This makes the book an inheritor of (and a debtor) to the Cognitive Models of Science Approach (CMSA) which was originally developed by Ronald Giere and colleagues in the 1990s. However, my theory—i.e., Cognitive Structural Realism (CSR)—is developed in the context of the Structural Realist philosophy of science (SR). Also, I rely on some edge-of-the-art scientific theories that were not considered by (or available to) Giere et al. I am referring to scientific theories of Karl Friston and colleagues, theories such as Free Energy principle and predictive coding. This makes the book’s solution to the problem of scientific representation radically naturalistic.
Scientific realism is a popular theory of the philosophy of science. It needs to be supplied by a realistic account of scientific representation. The problem of scientific representation then consists of how to provide a viable account of the veridical inferential relation between theories and the world. An antirealist (e.g., Larry Laudan) may summon counterevidence from scientific practice or history of science (which are naturalistic grounds) to challenge the assumption of the existence of such a veridical relationship. Under the circumstances, structural realists (such as John Worrall) appreciate the force of counterevidence from the history of science but argue that the realistic account of the veridical relationship between scientific theories and the world can be retained at the level of structure or form instead of content. To unfold their structuralist account of scientific representation, some notable structural realists such as Steven French use quasi-set theory and some innovative forms model theory (with partial structures), because as James Ladyman has observed, the set/model theory wears its structuralist commitments on its sleeve. The book aims to make a great improvement on the orthodox forms of SR and their formal account of scientific representation.
The main critical insight of the book is that, although formal tools that are used by orthodox versions of SR can regiment the structure of scientific theories precisely, they fail to account for a veridical relationship between the underlying structure of theories and the causal structure of the world. Please note that epistemological and ontological commitments of scientific realism are inseparable from the assumption concerning the veracity of scientific representations (regimented structurally according to SR). However, my insight is that, owing to their abstract nature, the semantic formal tools (e.g., set/model-theoretic tools) that are used by French (as well as some information-theoretic tools used by Ladyman) fail to explicate how it is that the theories contain significant veridical information about the world. I surmise that it might appear foolhardy to expect from the formal framework of theories to explicate the issue of scientific representation.
It might be contended that we should rely on philosophical arguments (such as No Miracle Argument or NMA) so as to defend the realist credentials of the theory. If so, expecting from semantic (structural) relations to substantiate the realist component of SR would be foolhardy. However, the general insight of the book is that if we use the representational capacity of the structures that embed the theories to substantiate the thesis of realism, we will accomplish a stronger and down-to-earth version of SR. I argue that replacing formal set/model-theoretic structures with cognitive structures contributes to achieving this goal. Scientific theories are the by-products of cognitive systems of human beings (as agents who interact with their ecosystems). Accordingly, I submit that it would be best to account for the theories-world relationship on the basis of the brain-world relationship. To do so, we may replace the formal model-theoretic structures with embodied informational structures. The latter kinds of structures could regiment the structure of scientific theories with enough formal precision. More interestingly, they could also be incorporated into a viable evolutionary story of how cognitive agents garner information from their environment and how pieces of information coalesce into sophisticated scientific theories. CSR is committed to the existence of embodied informational structures. The book relies on the recent theories of computational neuroscience, such as predictive coding and the Free Energy Principle, to offer a radically naturalised account of scientific practice and scientific representation. I shall flesh out further details in the next posts.