1.0 Setting the Scenery

"Cognitive Science is en vogue."

At least, that is what I wrote a few years ago at the very beginning of my Licentiate's dissertation. Since that time, Cognitive Science has lost some of its attractive liveliness to me. Partly because it has inched its way a bit further to becoming an established field of research, with all the attendant trappings of conferences, journals, and university programs, but also because it has seen the end of the discussion between defenders of 'classical', 'symbolic' and Good Old-Fashioned Artificial Intelligence (Haugeland 1985) and the new breed of connectionist researchers. The occasional acrimony of their previous discussion notwithstanding (see Van Everbroeck 1994 for an overview), the two parties now seem to have reached the stage where peaceful cohabitation has become possible: both symbolicists and connectionists feel that they can go about minding their own businesses without having to slam the other as well. What's more, the hybrid models which are currently popping up in abundance attest that many think there is mileage to be had from combining elements from both sides (see e.g. Sun 1996).

In this M.Sc. dissertation, however, I will not be dealing with the historical or philosophical issues underlying recent developments in Cognitive Science. Instead, I will try to address a piece of criticism that I received from some readers of my Licentiate's dissertation: what they felt was missing, was an original model illustrating the relevance of connectionist simulations to theoretical linguistic research. Although I still feel that the omission of such a model was justified at the time -- if only for reasons of time, support, and focus -- I also admit that the essence of the criticism rang true. Hence, CLASPnet.