Systematicity and the post-connectionist era:
 Taking stock of the architecture of cognition

19-21 May 2011
San José (Andalucía, Spain)

Background and objectives

Is human thought systematic? How can we best explain it? According to Fodor & Pylyshyn’s seminal challenge, the only way for connectionist theory to account for the systematicity of thought is by implementing a classical combinatorial architecture. Connectionism, they claim, is otherwise doomed, managing at best to throw light upon details of the neural substrate. Unfortunately, by targeting the connectionism of the 80s, the discussion of how to meet the challenge of explaining systematicity got stuck in a false dichotomy. We are now in a post-connectionist era where the battle is not between classicism and connectionism, but rather between cognitivism—both classicist (symbolic) and connectionist (subsymbolic)—and a range of methodologies such as behavior-based AI, ecological psychology, embodied and distributed cognition, dynamical systems theory, and non-classical forms of connectionism. Under the banner of ‘embodied cognitive science’, these approaches are questioning not only the symbols-and-rules approach, but in some cases representationalism tout court. It is thus not clear to what extent the challenge to representational connectionism, as originally formulated and rehearsed, transfers to these other cognitive architectures. Thus, even if we granted that Fodor & Pylyshyn’s systematicity argument proved effective against classical connectionism, we could continue to ask whether it poses a compelling challenge to post-connectionist methodologies.

Our workshop seeks to take stock of the architecture of cognition and revitalize the debate by discussing these and related issues as they arise in the treatment of the systematicity challenge within embodied cognitive science. Can we identify novel lines of response that ecological psychology, embodied and distributed cognition, or neurobiologically-plausible neural network theory, for instance, may try out? Can further progress be achieved by teaming up? Would such strategies incur in the same vices as previous connectionist responses? Would an implementational embodied cognitive science still serve to inform psychology? Likewise, it is timely to rethink what we really mean when we say that thought is systematic. What’s the empirical evidence for or against the systematicity of thought? How does the systematicity of human thought relate to human and non-human systematic behaviors? What areas of research, other than language, can throw light upon the systematicity argument?