Goodbye, Descartes: The End of Logic and the Search for a New Cosmology of the Mind, Keith Devlin

Reading takes you many surprising places. The dis-oriented author recently visited one of those place while reading Keith Devlin's Goodbye Descartes. When I picked this book of my shelf, I could not remember why I had bought it. I could not remember what the book was about and the title was not much help. In fact in Devlin's introduction, he notes that librarians will find this book difficult to classify.

Well, after reading the book, I know why I initially bought it and put it on my to read shelf. This book combines mathematics, linguistics, computer science, philosophy, psychology and the nature of thought. Exactly the kind of book I love.

When I was in college, in addition to my major classes in mathematics and computer science, I took forty hours of foreign languages — a combination of French, German, Spanish, Russian and Koiné Greek. Devlin's book takes a look at artificial intelligence, linguistics and the nature of human intelligence.

The Descartes of the title is of course French philosopher and mathematician René Descartes. Descartes is one of the pioneers of the scientific method. In the mid-twentieth century, developments in logic, computer science and mathematics led to the development of mathematical linguistics.

I recall writing a paper for my linguistics class on mathematical linguistics and the work of Noam Chomsky.  As a student of mathematics, I found Chomsky's work intriguing. Chomsky attempted to come up with an axiomatic, logical framework for languages. Not just a language but languages in general. If such an axiomatic treatment were possible, then perhaps it would be a step towards the dream of strong AI — true machine intelligence.

Devlin discusses some of the work of Alan Turing whose Turing Test plays a pivotal part in the book. The Turing Test was Turing suggestion for how to determine whether a machine were truly intelligent. In essence, Turing suggested that if a human asked questions (perhaps through a terminal) and the questions were answered by a human and a computer, if the human questioner could not tell the difference then the machine would be said to be intelligent.

There are many nuances to the Turing test but I believe that in 1980 I was a guinea pig in just such a test. That summer my father had taken a job as a visiting scientist in a NASA laboratory. I went to work in the lab as a student intern along with a number of engineering students, all male. My job involved writing mainframe code in FORTRAN to access a variety of plotters and CRT displays to draw maps of satellite data.

Those were the early days of networking and our lab was connected via DARPA-NET (the early form of the Internet)  to mission control at Johnson Space Center (JSC) in Texas. At the time we had a program that worked like AIM and let us chat with the team at JSC. We soon discovered that almost all of the interns at JSC were women. We spent many an hour chatting and flirting back and forth in a precursor to today's IM programs.

Near the end of that summer one of our guys took a trip to Kentucky to visit one of the JSC girls. After Louis came back we were full of questions but Louis never said a word... Hmm.... Maybe there were no girls at all at JSC and this was a NASA Artificial Intelligence program.

In the end Devlin rejects the notion that it is possible to codify language and understanding into a formal mathematical framework. He even makes the suggestion that  there might be soft mathematical techniques that come into play.

The title reference indicates the author's letting go of formal scientific and mathematical methods in the search for understanding of language and intelligence.

Goodbye, Descartes gets 4 of 5 dis-oriented smileys

Purchase Goodbye, Descartes from Amazon.com.