REVIEW: AN INVITATION TO COGNITIVE SCIENCE.

G. William Moore, MD, PhD.
http://www.medparse.com/rvcognis.htm


Send comments and correspondence to: George.Moore4@va.gov
See also: http://www.medparse.com/gwmcv.htm ............. http://www.medparse.com/rvgodell.htm ............. http://www.medparse.com/rvneuroc.htm .............
http://www.netautopsy.org/rvflatte.htm

United States Government Work, uncopyrighted, public-domain. This document does not necessarily represent the views or policies of any United States Government agency. This document is provided "as is", without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and noninfringement. In no event shall the authors be liable for any claim, damages or other liability, whether in an action of contract, tort or otherwise, arising from, out of or in connection with the document or the use or other dealings made with the document. Published in: Neurocomputing. 2001 Jan;42(1):.



Scarborough D, Sternberg S.
An Invitation to Cognitive Science. Second Edition. Methods, Models, and Conceptual Issues. Volume 4.
Cambridge, MA: A Bradford Book. The MIT Press. 1998.
ISBN 0-262-65046-0, 950 pages.

      This book is an overview of cognitive science, accessible to the advanced undergraduate, that covers all the major areas of this important field. The book is written by multiple authors, but is unified by an editor's introduction to each chapter, and by a consistency of organization and style. The book is perhaps best regarded as a companion to an ordinary textbook. In many cases, the author presents his/her own opinion about an area, and defers alternate viewpoints and technical details to the reference section.

      Each chapter is intended to be read as a self-contained unit, if necessary, in isolation from other chapters. This means that the fundamental intellectual assumptions of an area are summarized at the beginning of each chapter, and each chapter achieves some level of closure at the end. For such a long volume (over 900 pages), this is an important constraint. Unlike Sheherezade's captor, who had to wait anxiously each night for the conclusion of last night's episode, the reader can set this book aside for a few nights without losing track of the story as a whole.

      Some chapters are highly informative, and attempt to provide an overview of an entire area of investigation or methodology. Prof. Wickens' chapter on statistical methods covers all the major topics that one would expect in an introductory course: a sample experiment; the role of statistics in evaluating research results; the problem of variability; population versus samples; hypothesis testing; decision theory; analysis of variance; contingency tables; correlation; and multivariate analysis. Detailed formulas and calculations obviously cannot be covered in such a short space. However, the reader emerges from this chapter with a good sense of what statistics is all about: why, how-to, and how-far you can reach with statistical reasoning.

      Other chapters are polemical rather than instructional. Prof. Lewontin's chapter assumes a single point of view on the evolution of cognition: it cannot be proved, period. The author repeats this conviction in the form of an intellectual spat with the editor's reviewer as the final section in the chapter. In building his case, Prof. Lewontin lays out many of the cardinal arguments of evolutionary theory, and explains how they might (but do not) apply to the evolution of cognition. These arguments include: analogy/homology arguments, such as the panda's thumb and the whale's tail; the evolution of moth coloring during the English industrial revolution; and linguistic studies in primates and cetaceans. Whether or not the reader accepts the author's conclusion, the argument is impressive, and the intellectual tour alone makes the chapter worth reading.

      Another chapter that is richly laden with low-hanging intellectual fruit is Prof. Anderson's chapter on learning arithmetic with a neural network. Although clearly comfortable with convoluted scientific ideas, Prof. Anderson has a peppy writing style, and a remarkable ability to reduce complex concepts to their essentials. The answers to many of those nagging questions about science and mathematics that you were embarrassed to ask in class, are dashed off in a paragraph or two in this chapter. For example: what is the difference beween a biological neural net (as we currently understand it) and the mathematical formalism? What is "associative inference" in humans and in computers? How might one envision a mathematical proof, such as the infinite-prime-number theorem of number theory? How did Einstein envision special relativity?

      Readers of this book who are most interested in neural nets as computing devices will be disappointed. The book discusses some computer algorithms, but is aimed primarily at understanding biological processes. We learn about how humans distinguish, say, a lower-case-c from a lower-case-e, but not how optical character recognition software solves the same problem; how a human finds a face in a crowd, but not how computer software addresses this problem.

      The short biographies at the end of each chapter are fun to read. Some of the authors seem to have led fairly humdrum academic lives; but others' lives have zigzagged along interesting pathways. One author read science fiction while attending elementary school in a "friendly, relaxed backwater"; squandered his predoctoral years seeking intelligence among garden slugs; and then spent two exciting postdocs studying human cognition.

      This book will appeal to the intelligent but discontented undergraduate or graduate student in fields akin to cognitive science, including psychology, philosophy, mathematics, and biophysics. There is a range of opinions, presented aggressively by their champions in academia; current controversies are bared and explored; yet behind it all, the scientific underpinnings are solid. I highly recommend the book.

Last updated: 1/6/2006, by G. William Moore, MD, PhD.