On the asymmetry between the notions of computation and of representation

In the literature of cognitive science, the notions of computation and of representation have often been closely associated. This attitude finds its more explicit and radical expression in Fodor’s slogan “no computation without representation.” But the strict connection between computations and representations has been widely accepted far beyond Fodorian orthodoxy. However, there are good reasons to maintain that a deep asymmetry between such notions exists. On the one hand, (digital, algorithmic) computation is a well-defined (mathematical) notion. On the other hand, the concept of representation is hopelessly vague, and, in any case, no concept of representation is available that is even remotely comparable to that of computation. If the Fodorian slogan “no computation without representation” was true, it would be a disaster: since we do not have any clear notion of representation, this would involve that the notion of computation itself is in danger. But since we have a fruitful and satisfactory notion of (digital, algorithmic) computation, I suspect that there is something wrong with the Fodorian slogan.