Whiskey Creek logo

Fluid Concepts and Creative Analogies:

Computer Models of the Fundamental Mechanisms of Thought

by Douglas Hofstadter and the Fluid Analogies Research Group

Fluid Concepts cover From Goedel Escher Bach to Metamagical Themas, I've been following Douglas Hofstadter's struggle with what he calls the "alphabet problem." Representatives from varying scientific disciplines have been trying to develop computers to model human thinking for 15 to 20 years now. One of the great strengths of the human mind is pattern recognition as opposed, for example, to number crunching. Unlike computers, human beings seem to be able to recognize an A in almost any deformed shape. A computer really issn't any good at recognizing "near A's." Goedel, Escher, Bach cover

Hofstadter entered the early artificial intelligence frays when people were arguing whether to construct computer architectures from the top down or the bottom up. The top down supporters felt to build intelligent machines, designers simply needed to ask experts what they did and cram the computer as full of as much information as they could pick from expert brains. But like the alphabet problem, researchers have discovered that a large measure of what makes an expert an expert is the ability to see pattern or similarity in situations, even if that pattern may have been seen only once before. Metamagical Formulas cover Of course, Roger C. Shank describes in Tell Me a Story, how those who were working to develop computers that could generate stories were discovering a somewhat parallel problem. Human beings have a large stock of assumed knowledge in the form of various kinds of scripts, knowledge that is understood and needs no explanation in story telling. These kinds of problems indicated there is more to modeling the workings of the mind, including expert systems, than just stuffing a computer full of as much data as possible.

Hofstadter's Fluid Concepts outlines the history of the computer modeling efforts to overcome the alphabet effect. He says he used to consider himself a researcher in artificial intelligence, but he feels this group has become too tied up in the commercial development of expert systems and has too little respect for cognitive science and philosophy.

This book is a collection of essays, with Hofstadter's introduction providing a history of the research on the alphabet problem. He notes the book is mainly about computer modeling efforts, with a critique of artificial intelligence methodology. Anyone following the research to discover the secrets of how the human mind works will find a good summary of the ground the research has covered as well as analysis of the direction Hofstadter thinks the research is going.

Hofstadter's group is still struggling with the alphabet problem, and notes that in attempting to model creativity, figuring out how to get a machine to "imbue all 26 lowercase letters of the roman alphabet with the same abstract essence -- visual style, that is -- is one of the most fascinating and deeply challenging tasks that we can imagine." (p. 8)

Back to Brain-Mind Connection Books Reviews

Books Reviews About Design, Writing, and Editing

Back to Whiskey Creek Document Design's Home Page

© Copyright 1996-2003 Whiskey Creek Document Design. All rights reserved