Abstract
Functions, Representations, and Zombies: Information Theoretic Perspectives on Three Standard Confusions in Cognitive Science
After a brief introduction to the language of Chaitin's version of algorithmic information theory, this talk summarises solutions to some standard problems of cognitive science and philosophy of mind. Less compressed discussions appear in the book Mind Out of Matter.
Problem 1: Functional systems. Arguing that usual formulations of functionalism, including Chalmers's definition of a combinatorial state automaton, are mathematically trivial, I introduce a novel measure of process complexity and apply it to ground an objective and precise account of functional decomposition.
Problem 2: Representation. Making sense of the notion of representation has long challenged philosophy of mind and cognitive science. I argue that representation itself may be understood in terms of mutual information content and show how the different question of whether an item is used as a representation by a cognizer may be understood in terms of representation within a functional system.
Problem 3: Supervenience and physical facts vs. consciousness facts. Comparing the relationship of supervenience to the relationship of provability, I argue that typical philosophical discourse about the relevance of 'absent qualia' zombies for physicalism erroneously assumes a logic which is both complete and consistent. An interesting analogy comes from considering the tantalisingly noncomputable number Omega and the formal specification of the corresponding Universal Turing Machine.
|
