computational architecture

*The structure and organization of a given computing device (the way in which it handles memory, the organization of date, the set of primitive instructions it executes, and the ordering of instruction application or execution) define a device's computational architecture. Computational architecture involves only the structure and organization relevant to computation; implementational details to not constitute computational architecture. See also computation, functionalism.*

**Details:**

In actual practice, the notion of computational architecture inherits some ambiguity/vagueness from the ambiguity/vagueness in one's notion of computation. For example, computation might refer exclusively to forms of effective computation (Turing-equivalent computation) or it might also include weaker computational devices (e.g. finite state automata and push-down automata) or analog computational devices.

On the hypothesis that cognition is computation, one research goal is the determination of the computational architecture that constitutes the structure and organization of cognition. In other words, if the brain is a type of computing device, one wishes to know exactly what type of computing device. Even on the hypothesis that cognition is a form of effective computation (Turing-equivalent computation), one still wishes to know exactly which of the computationally equivalent forms of computation constitutes the hypothetical computational architecture of cognition. Even if diverse forms of computation are equivalent in the sense of computing the same functions, one wishes to know exactly what mechanisms the brain uses to compute those functions.

Ken Aizawa