Review by Choice Review
This volume by Danks (Carnegie Mellon Univ.) makes two main claims. First, many types of cognitive processes involve representations that take the form of graphical models. Second, this provides a framework for "unifying the mind" as a cognitive architecture typified by a single form of representation (graphical models) that can be accessed by different types of processes (such as category learning and causal inference). Danks provide both intuitive and more mathematically precise characterizations of graphical models, but this is definitely a book principally for readers already familiar with a range of theories in cognitive psychology and their treatments in mathematical and computational terms. This volume's chief importance lies in its arguments for an account of cognitive architecture at a fairly abstract level, rather than at a computational or implementational level. Danks stakes out a position that should be of interest to researchers in cognitive psychology or knowledge representation in AI, and philosophers whose work deals with those fields. Along the way the author provides useful critiques of David Marr's three levels of description and of modularity and advocating a pragmatic account of explanation. Summing Up: Recommended. Graduate students and researchers/faculty. --Steven W. Horst, Wesleyan University
Copyright American Library Association, used with permission.
Review by Choice Review
