Listen to founder of the connectionism model known as parallel distributed processing Jay McClelland talking about his background in psychology and his current enthusiasm for connectionism.
Philosophy Talk, 17 February 2008. Start at 9:50 and listen until 14:06.
The discussion contains references to two non-connectionist learning theorists. One is a cognitive theorist and the other is not.
1. Who are they?
2. What objection to the cognitivist's theories does McClelland make?
3. How do connectionist theories offer a better model, in his view?
1. BF Skinner: this researcher is not mentioned by name, but the study of pigeons in psychology is attributable to his behaviorist experiments.
Chomsky: "old-fashioned AI (artificial intelligence) approach"; the structure of human language is pre-specified in our genes.
2. McLelland claims that the way we talk is influenced by our experience.
3. Computer simulations of neural networks allow psychologists to model learning as changes in connections between units (neurons). Information that was thought to have to be pre-programmed could actually be acquired.