Books

Short works

Books : reviews

Marvin L. Minsky, Seymour A. Papert.
Perceptrons: extended edn.
MIT press. 1988

Marvin L. Minsky.
The Society of Mind.
Simon & Schuster. 1985

rating : 1.5 : unmissable
review : 22 August 2004

This seminal work on how a brain might make a mind is just as relevant today as when it was first published twenty years ago. Minsky explains how lots of very simple agents might come together as agencies and result in intelligent behaviour. The structure of the book reflects this architecture -- each individual page is a short "chapter" explaining some nugget of the theory, and the cumulative effect is the totality of the argument. Some of the ideas are quite specific, and may be wrong in detail (for example, I remain unconvinced by the sections on humour), but the overall argument remains sound.

(One thing that has changed dramatically since the book was first published: the fact that I've finally managed to get hold of a copy of this long out-of-print work is a testament to Amazon's second-hand sales webpages.)

Marvin L. Minsky.
Computation: Finite and Infinite Machines.
Prentice-Hall. 1967

Marvin L. Minsky.
The Emotion Machine: commonsense thinking, artificial intelligence, and the future of the human mind.
Simon & Schuster. 2006

rating : 2.5 : great stuff
review : 19 March 2007

This is a direct sequel to The Society of Mind, with some more details and explanations worked out, and the same style of diagrams. This time we get a discursive wander around possible structures of various cognitive processes, with various interjections from a range of critics.

Despite the title, there is actually very little on emotions as such. Minsky has an interesting take on them: he regards them as being a consequence of thinking, not as a causal factor in the process:

pp226. [Patient Elliot's inability to decide, and emotionlessness] led Damasio to suggest that "reduced emotion and feeling might play a role in Elliot's decision-making failures." However, I'm inclined to turn that around to suggest that it was Elliot's new inability to make such decisions that reduced his range of emotions and feelings.

pp233-4. The expressions of rage ... could have served in primordial times to help to repel or intimidate the person or creature that one is angry with; indeed, any external expression of one's mental state can affect how someone else will think. This suggests an idea about what we mean when we use our most common emotion-words; they refer to classes of mental conditions that produce external signs that make our behaviors more predictable to the persons with whom we are dealing. Thus, for our ancestors, those bodily signs served as useful ways to communicate such so-called "primary" emotions as Anger, Fear, Sadness, Disgust, Surprise, Curiosity, and Joy.
...
The body and face could also serve as a simple sort of memory ... your external expressions of anger may serve not only to frighten your enemies, but to also ensure that you will stay frightened for long enough to carry out some actions that might save your life.

If emotions are to be thought of as a consequence of cognition, I wonder how this fits with animal emotions? Minsky says only a little about evolutionary development, but his architecture does admit increasing levels and layers of sophistication in the processing (which is one of the regions where it is more developed than in The Society of Mind). Maybe Minsky would say that animals have much simpler cognitive systems, but that their emotions are nevertheless consequences of these simpler systems?

Despite not doing exactly what it says on the tin, however, this is a fascinating read. There seems to be more structure in the underlying model this time around, not just a pandemonium of agents competing to be heard over the din. Some questions are still begged: how do certain of the Recognisers manage to do their jobs? But it certainly feels like a not implausible model of what's going on inside our heads.