TO TRANSFORM THE PHENOMENA - FEYERABEND, PROLIFERATION, AND RECURRENTNEURAL NETWORKS

Paul Feyerabend recommended the methodological policy of proliferating competing theories as a means to uncovering new empirical data, and t hus as a means to increase the empirical constraints that all theories must confront. Feyerabend's policy is here defended as a clear conseq uence of connectionist models of explanatory understanding and learnin g. An earlier connectionist ''vindication'' is criticized, and a more realistic and penetrating account is offered in terms of the computati onally plastic cognitive pro file displayed by neural networks with a recurrent architecture.