Monthly Archive for January, 2006

Nathan, if you want to hold up Parallel Distributed Processing as a “replacement” for the Information Processing model, you should probably call it something like “Radical PDP”. I think you’ll find that most people in psychology consider the network approach to be compatible with the information processing approach, not a competitor to it. Networks very easily fit within the Information Processing picture as parallel processing modules within a serialized pipeline.

Of course, if you are really subscribing to Radical PDP, then that makes for an interesting debate. That would probably mean that every receptor in your body is connected anonymously to a neural soup, which then performs some sort of parallel processing devoid of architecture. It’s a black box approach that reminds me a lot of behaviorism.

Or maybe you’ll argue that some sort of architecture emerges out of the neural soup. Well, now we’re back at the information processing model, which makes no claim as to how the architecture is implemented, only that a symbolic processing architecture exists.

Throwing the baby out with the bathwater is a common tendency in science, and it’s particularly bad in psychology. People say “this old theory is wrong because of X,Y, and Z, and this new theory handles X, Y, Z, and Q like a champ!” In science we can only critique theories by disproving parts of them, so there is little incentive to say “you know what, this old theory did A, B, and C really well–and it still does!”

But at the end of the day, all models are imperfect–that’s why we call them models. We extract value from a model by using it to tell us something about the world. And both the information processing model and the network model can tell us a lot about cognition.

If you want to see an interesting challenge to the information processing approach see J.J. Gibson’s 1986 The ecological approach to visual perception. That’s an approach that raises some genuine questions about information processing.