Dr. Lee D. Carlson 2006-04-10
4 of 5 people found the following review helpful:
As defined by the editor of this book, cognitive psychology is a field that studies the information processing capabilities of the human mind. Having its origin in an actual college classroom, the book does give a good overview of some of the developments in cognitive psychology and can be read profitably by anyone interested in the field. Since it is targeted towards undergraduate students in cognitive psychology, articles are included that introduce the field both from an historical and a philosophical viewpoint. Those readers interested in artificial intelligence will also find many of the articles of interest, especially those that address the actual cognitive processes in the brain that are involved in problem solving or goal-setting. Philosophical argumentation on the `mind-body problem' occupies a portion of the book, but the space devoted to such speculative matters is refreshingly kept at a minimum. It is clear throughout the book that the editor wants to keep the mind-brain distinction intact.
One of the more interesting articles in this collection is the one by K.A. Ericsson and J. Smith on the empirical study of expertise. The identification and study of expertise is important not only from the standpoint of cognitive psychology but also in artificial intelligence, where it is becoming more important to identify when a machine has expertise in more than one domain. The authors have given the reader an overview of what they call the `original expertise approach'. Their goal is to first characterize expertise in a domain-independent way and then discuss chess as an example. Crucial to any study of expertise is being able to distinguish between outstanding individuals in a domain from those that are not. The authors also want to distinguish "stable" achievement from that which might be accidental or happen only once. But most importantly, they point to the need for a `control group' who, given the same opportunities to achieve, could be compared with the achievements of an individual deemed to be unusually talented. The author's `original expertise approach' tries to define conditions under which one can study critical performance and find out which of its components are responsible for making it superior performance. It is necessary in their approach to design a set of tasks in a given domain that illustrate superior performance and to be able to bring out this performance in actual laboratory conditions. This approach is therefore difficult to carry out, since the conditions in real life needed for superior performance are hard to replicate in the laboratory. Their discussion of chess performance is helpful in this light but it still leaves doubt as to why superior chess players can perform as well as they do and the exact nature of the cognitive processes that they use to attain superior performance in chess. Of course, one cannot observe directly these cognitive processes, and the authors recognize this. However, they argue that using the observations of certain tasks, one can study these processes by using an information-theoretic model of cognition.
In some research circles in artificial intelligence there is currently a great interest in creating machines that can deal with information in more than one domain without changing appreciably the "programming" or "software" that processes the information in these domains. Called `artificial general intelligence' by some, the goal of researchers in this field is to arrive at a notion of machine intelligence that models what is found in human intelligence: the ability of humans to solve problems in many different domains (sometimes concurrently). Those readers interested in these developments may find the article by J. Tooby and L. Cosmides on the functional organization of the mind and brain of some interest. The emphasis in this article is that the brain is a biological entity that has evolved to process information. The authors call the brain an `organ of computation', which means that as a physical structure it contains a collection of programs that process information, and that this physical structure exists because it contains these programs. The programs are considered to be the functional components of the brain, and they are there because they solved a particular type of problem in the past. The problems that had to be dealt with evidently had to be confronted over long time scales, in order for evolutionary adaptation to occur. In addition, in order to sophisticated structures to evolve, the brain had to confront nontrivial problems that require heavy computation. The authors emphasize, and this is very important for readers interested in artificial general intelligence, that there is no single algorithm that is able to solve every adaptive problem. The human mind is therefore composed of many different programs for solving different problems. The reverse engineering of the human brain will therefore involve the identification of those functional units and collections of computations that it uses in order to be biological successful. If these functional units are independent, then the algorithms themselves are independent, and thus there is no notion of `general intelligence' used by the human brain. Brains are built from adaptive problem-solving devices, though they may have incidental capabilities that may make them appear to be `generally intelligent'. This modular approach to the brain taken by the authors seems to be a reasonable one, and they recognize that more evidence must be accrued. And certainly an organism, human or not, will have a better chance of surviving if their neuronal systems or problem-solving abilities are independent of each other. If one module is damaged for example, the organism can still have the capability to deal with problems and issues that are confronted by the other undamaged modules.
