Atelier Théorique AT1

Connexions and symbols

Oct 15 and Oct 22, 2012

Where: Piaget Non B (615) - Rue des Saint Pères, 75006 Paris; 9h00-12h00.

Séance 1 (15 Oct): setting up the debate

Connectionism attracted a lot of attention in the Cognitive Science communities in the 80s through the publication of two landmark volumes (Rumelhart & McClelland, 1986; McClelland & Rumelhard, 1986). Paralleled Distributed Processing was described as a new theory of cognition challenging the idea of symbolic computation that was at the center of debate in theoretical discussions about the mind. This ignited a debate (see the special cognition issue Connections and Symbols, Pinker & Mehler, 1988), but also a wealth of empirical and theoretical results that have profoundly modified the field of Cognitive Science.

In this course, we will review the central ideas of Parallel Distributed Processing and discuss some of the critiques that were voiced, in particular by Fodor and Pylyshyn (1988).

The students should read the following paper; a small group will present the central argument of the paper, and the rest of the studentw will prepare questions.

To read and to prepare:

PDF of session 1

Reference books

Séance 2 (22 Oct): The devil is in the details

In this course, we will review two technical contributions: the recurrent networks (Elman, 1990), and the tensor products (Smolensky, 1990). We will discuss the extent to which these network architectures enable to respond or not to the fundamental criticism of Fodor and Pylyshyn.

To read and to prepare:

Download PDF of session 2

Further readings