An orthodoxy in cognitive science is that cognition is best explained in terms of mental representations. Anti-representationalist views reject this orthodoxy, but they in turn face what’s known as the "scaling-up problem” – they seem unable to explain higher-order cognition such as language use and abstract thought. This paper assesses a recent solution to the scaling-up problem proposed by Matej Kohár. Kohár’s solution appeals to a theory of mechanistic explanation that purports to explain hig…
Read moreAn orthodoxy in cognitive science is that cognition is best explained in terms of mental representations. Anti-representationalist views reject this orthodoxy, but they in turn face what’s known as the "scaling-up problem” – they seem unable to explain higher-order cognition such as language use and abstract thought. This paper assesses a recent solution to the scaling-up problem proposed by Matej Kohár. Kohár’s solution appeals to a theory of mechanistic explanation that purports to explain higher-order cognition by decomposing it into the activity of neural mechanisms. However, I argue that Kohár’s solution fails because the neural mechanisms he cites as explanans just are representational mechanisms in virtue of their functional profile. I anticipate and respond to objections to my argument on two fronts – one concerning the supposed mysteriousness of the representational relation, and another concerning the explanatory utility of representational content. The upshot of my discussion is twofold. First, I illustrate that mechanistic explanations of cognition are insufficient to overcome the scaling-up problem. Second, in defending the notion of representational mechanisms, I offer a novel account of the explanatory utility of content which states that content non-causally explains the success of representational mechanisms.
