This is a very critical review of a recent book by Paul Churchland. I can't say if it is fair or not, although to me it read as if the reviewers were rightly annoyed at the lack of attention to opposing views ("While his book has many virtues, it is unfortunate that he repeatedly fails to do justice to his opponents' views."). But I was struck by the last paragraph:
A final striking fact about Churchland's book is that it seems almost wholly divorced from empirical psychology. Remarkably, indeed, in a book that advances a theory of the mind that is supposed to be empirically supported, Churchland provides only around thirty scientific references, just a third of which date from the twenty-first century, and many of which are computational rather than experimental in nature. One would like to think that he chose to provide only a judicious selection so as not to overwhelm his audience with references. But since he ignores a great many results that appear inconsistent with his main theses, we fear that the paucity of references requires a different explanation. Indeed, Churchland ignores almost entirely the extensive work in developmental and experimental psychology, in neuroscience, and in studies of comparative cognition that have been conducted by cognitive scientists, especially over the last twenty years. And it is precisely once we examine the theories supported by empirical phenomena of these psychological sorts that past and present arguments for nativism and for LOT (appropriately understood) begin to emerge.
Now I doubt I have a firm grip on the computational vs experimental distinction here. (Is the former a simulation with a Bayesian networks and the latter work on real humans and other animals?) Either way, this paragraph made me wonder: does a book entitled How the Physical Brain Captures a Landscape of Abstract Universals, have to cite experimental literature?
Recent Comments