Idealization and the Mechanistic Theory of Computation
The MIT Press eBooks · 2025
Overview: This chapter focuses on Piccinini's (2015, 2020) influential mechanistic theory of computation.A key idea behind the mechanistic theory of computation is that a computational explanation of a system takes the form of mechanistic explanation.Another central idea that motivates the mechanistic view of computation is that computation involves three ele ments:(1) rule, (2) medium-independence, and (3) teleological function.Hence, the mechanistic view of computation can be applied in a defense of the CTC, given the assumption that cognitive pro cesses involve manipulation of medium-independent vehicles, whose (teleological) function it is to carry repre sen ta tional content in accordance with rules mapping across inputs, internal states, and outputs.The central point that I defend in this chapter is: the claim that the nervous system and cognition are computational is false if taken literally.Maley (2023) argues that computational pro cesses are not executed over medium-independent vehicles.I will not address this issue here.My target is the specific condition that computational pro cesses in the nervous system are done in accordance with a rule.I shall show that the "rulegoverning" condition on computation is not reflected in the target system.Rather, it sits entirely on the side of the scientific model (or, the computational model); not the nervous system.Again, we will have reasons to think that this can be made consistent with a version of scientific realism.However, once the rule-governing condition cannot be met, realism about the brain and cognition being computational cannot be defended. IntroductionAn advantage of the mechanistic account of computation is that it does not rely on question-begging notions from logic and computer science such Idealization and the Mechanistic Theory of Computation