4.4 The End

So, where does this road lead to? In an article entitled 'It's About Time' an answer to that question has recently been given by Van Gelder & Port (1995). Discussing the merits of 'dynamic' and 'computationalist' approaches to cognition, they write (1995: 34):

From the perspective of a genuinely dynamical conception of cognition, classic PDP-style connectionism [...] is little more than an ill-fated attempt to find a halfway house between the two world views. This diagnosis is borne out by recent developments. Since its heyday in the mid- to late 1980s, this style of connectionist work has been gradually disappearing, either collapsing back in the computational direction (hybrid networks, and straightforward implementations of computational mechanisms), or becoming increasingly dynamic (e.g., the shift to recurrent networks analyzed with dynamical systems techniques). Connectionist researchers who take the latter path are, of course, welcome participants in the dynamical approach.

It's manifesto time again. I'd better get moving. (Note 2)

Note 2
Articles describing dynamical connectionist approaches to language are Elman (1995), Petitot (1994, 1995) and Dorffner (1992). Unsurprisingly, all of them also refer to work in Cognitive Linguistics.