Idealization and Content in Computational Modeling
The MIT Press eBooks · 2025
Overview: In this chapter, my focus is on Egan's (2018) account of the nature and function of content in computational models.Egan's (2018) view is that semantic properties have some heuristic value only in enabling us to think about the be hav ior of cognitive systems in a useful way.Egan states: "The pragmatic ele ments and the contents they determine [read: the cognitive content] are 'quarantined' in the intentional gloss, to use Mark Sprevak's (2013) apt expression" (2018, 256).Egan labels semantic properties "cognitive content."I take Egan's treatment of cognitive content to suggest that content ascription in cognitive explanation is akin to idealization in science-as argued in the previous chapter.Egan argues that repre sen ta tional content should be seen as ineliminable parts of cognitive science; yet, Egan rejects the view that repre sen ta tional content captures anything real in cognitive systems.In this sense, Egan aligns her view on repre sen ta tional content with fictionalism-the idea that discourse about semantic content is useful but ultimately false.This is how idealization is defined under the standard view of idealization (see chapter 2).A dif fer ent way of describing Egan's position is by introducing Dennett's (1987) distinction between realism and instrumentalism.Unsurprisingly, Dennett is a realist about brains and their vari ous neurophysiological parts.Yet, he is an instrumentalist about belief-desire states that, he contends, appear as abstracta in the "intentional stance."As Dennett states: "They are reason-giving explanations, which make an ineliminable allusion to the rationality of the agent.Primarily for this reason. . . .I suggest that folk psy chol ogy might be best viewed as a rationalistic calculus of interpretation and prediction-an idealizing, abstract, instrumentalistic interpretation method" (1987, 48; emphasis added).My suggestion is that Idealization and Content in Computational Modeling