The Connoisseur’s Guide to the Mind

Roger Schank directs the Institute for Learning Science at Northwestern University; his vita also includes stays at Yale (where he directed the Artificial Intelligence Project) and Stanford. Along with his impeccable academic credentials, Schank has established himself as a prolific popularizer of AI with books such as THE COGNITIVE COMPUTER: ON LANGUAGE, LEARNING, AND ARTIFICIAL INTELLIGENCE and TELL ME A STORY: A NEW LOOK AT REAL AND ARTIFICIAL MEMORY.

As a popularizer, Schank has a profitable scam going: He keeps writing the same book, repackaging it slightly each time. In THE CONNOISSEUR’S GUIDE TO THE MIND, he uses a number of scenarios relating to eating and drinking as vehicles for “thinking about thinking.” This device depends for its effectiveness on the provocative incongruity of its two principal themes: eating and thinking. Lest the reader not appreciate the novelty of this juxtaposition, Schank himself frequently comments on it.

Think of Schank’s rapid shifts between themes as the equivalent of the patter and misdirection that accompany the classic shellgame. What’s being concealed in THE CONNOISSEUR’S GUIDE TO THE MIND is the staggering banality of Schank’s insights. The pleasingly whimsical chapter titles—“Searching Memory in Minnesota”; “Trying to Believe Champagne and Armagnac”—perform a similar function. Flip to the end of the Minnesota chapter and you’ll discover this nugget: “Finding things in memory depends upon having been interested in them at the time that they occurred.”

Here’s another insight, from the chapter “Finding and Following Scripts in Atlanta”: “The story Jean-Francois ordered caviar but he spilled it on his tux does not seem at all odd, while Fred ordered caviar but it fell in the hay does. The reason is simply one of what was expected and what wasn’t.”

Is there a more general truth to be discerned here, about AI as currently practiced? Are we seeing a naked emperor, or a naked army?