Lately I have been reading some books on cognitive architecture, and browsing Amazon to see what else is available. There is as yet no single book that covers all of the subject, but there are some amazing works that handle parts of cognitive architecture in depth, and that are well suited for a computer implementation.

These cornerstones of cognitive architecture are:

1. Artifical Minds (1995) - Stan Franklin [Amazon]
   Great introductory work touching on many aspects of the architecture. It introduces some of the works below or describes similar techniques.
2. A Cognitive theory of consciousness (1988) - Bernard J. Baars [Amazon]
   Describes the central function of the Global Workspace of the mind: working memory, working contexts and unconscious processes. I think this provides the basic framework in which the other parts should be fit.
3. The Emotions (1986) - Nico H. Frijda [Amazon]
   This is not an easy read, but it provides a full functional account of the role of emotions. Emotions are reactions to affected concerns. Concerns are the basic drives of an organism or agent, so this book describes why we do things.
4. Scripts, Plans, Goals, and Understanding: An Inquiry Into Human Knowledge Structures (1977) - Roger C. Schank, Robert P. Abelson [Amazon]
   I have not read it yet, but is referred to again and again as the source of how our knowledge is structured. It explains how we are able to understand and act in complex situations.
5. Analogy-Making as perception (1993) - Melanie Mitchell [Amazon]
   This book describes in detail how creative intelligence is built, how we solve problems by using previous similar experiences as analogy. The creative brain behind this work is Douglas R. Hofstadter and his Fluid Analogies Research Group. For a context of this work see [Fluid Concepts...] and Hofstadter's new work [The essence of thought].
6. The atomic components of thought (1998) - John R. Anderson, Christian Lebiere [Amazon]
   I am reading this at the moment. You need to look past the claim of the authors that this is the architecture, but once you have done that, it contains very useful information about the symbolic and subsymbolic structure of production rules and declarative knowledge. Moreover, these structures have been used successfully by many scientists in very diverse experiments.