Conclusion

Cognitive science is not static. The interdisciplinary project continually identifies new problems and develops solutions for solving them. At the same time, the scope of cognitive science has been expanding.

As we have noted, as new tools for studying brain processing have been developed, cognitive scientists have become increasingly concerned with how cognitive operations are implemented in the brain. How to incorporate information about neural processing poses challenges for both classical and connectionist modeling in cognitive science. Insofar as cognitive models focus on the mental activity, they must to some degree abstract from the neural detail. This frames a philosophical question about how much they can abstract from details of neural processing and still claim to provide accounts of how humans process information. A related issue involves cognitive science's traditional reliance on logical and heuristic techniques to model reasoning. Increasingly, both cognitive and neuroscience researchers are advancing probabilistic models (Chater, Tenenbaum, and Yuille 2006). For philosophers, this raises the question of whether such models should replace more traditional cognitive models, or whether the two kinds of models can be constructively related.

Another major new direction in cognitive science is the concern with the embodied, situated nature of cognition. While traditional cognitive science focused solely on what is going on in the heads of cognizers, recent theorists have argued that the non-neural body and environmental con­text are not merely inputs to the cognitive system but play a constitutive role in cognition (e.g., Clark 1997). Some of those championing more attention to the organism's body and environment have appealed to a previously untapped philosophical tradition known as phenomenology (comprising writers such as Husserl, Heidegger, and Merleau-Ponty) for insight about these issues (e.g., Wheeler 2005). When the focus is on the real-time responses of organisms to their environment, the temporal dynamics of cognitive processes is obviously important and has been emphasized by those advocating use of dynamical systems tools for understanding cognition (Port and van Gelder 1995). Many advocates of applying dynamical systems theory to cognition have, as we noted, also argued against the reliance on representations in cognitive models. Others, such as Rick Grush (2004), have tried to show how control theory, a dynamical approach, employs neural representations in accounting for motor control, and to extend this approach to other cognitive processes. These and other current debates in cognitive science provide rich opportunities for continued philosophical engagement with cognitive science.