Return to
Representation Day Page
Recently, several alternatives have been proposed: as emergent patterns of non-linear dynamical processes, or appropriate behavior in embedding environments. Because they reject the traditional view, these alternatives are often said to be non- (or even anti-) representational.
I believe this characterisation is too narrow. Instead, I will argue, we must reclaim the word 'representation' from the traditionalists, broaden its meaning, and use it to characterize the rich and multiply-textured middle ground among these limiting ideals. Towards this end, I will present a generalised theory of representation: as appropriately stabilized patterns of disconnected coordination. On this view onotology becomes an achievement, not a given.
I will conclude by asking my fellow speakers a question: Is this new view powerful enough to analyze emergent systems and interactive robots?