Abstract
Puzzles of Representation and Function: Using Algorithmic Information Theory
After introductory comments on Chaitin's rendition of algorithmic information theory, I discuss some traditional puzzles surrounding the notion of representation and the way we understand functional modularisation in complex systems. Applying standard algorithmic information theory to the first problem and a novel extension of it to the second, I argue that our usual ideas of representation and of function can and should be overhauled. If there's time, I'll also note how properties of the delightfully noncomputable number Omega might bear on debates about supervenience.
The talk will be aimed at a broad audience; intimate knowledge of algorithms and Turing Machines is not required.
I can easily be cajoled into discussing other topics before, after, or even during the talk; my web pages include notes on most of my current work.
|
