Paradigms of Artificial Intelligence; A Methodological and Computational Analysis Book

Eric Baum points out that until recently, most scientists thought biological life could never be rationally explained but today few people doubt that a scientific explanation of life is possible. In the same way, we should now believe that the mind can be explained in terms of physics, biology, mathematics, and other scientific principles. This book claims to provide that explanation.

Baum chooses the theory of evolution and principles of computer science to sculpt a remarkable image of the mind as computer program - not a mind like a computer, but literally a computing program. What does it compute? Understanding of the world, leading to behavior for genetic propagation. Who was the programmer? Evolution by natural selection.

One of the most important things DNA does, he says, is to create a brain as an algorithm, a recipe for dealing with the world. The recipe does not contain facts about the world, but instructions for getting around in the world. Instructions are much simpler than facts, just as the recipe for a cake can fit on one 3x5 card but a specification of its chemical composition would take up many pages.

What about the mind though? The brain is the mind, in Baum's narrative. The terms are used interchangeably, without qualification, throughout the book. Since science can only deal with physical things, the mind must be the same as the brain for Baum's thesis to work.

One standard objection to that idea is that thoughts don't seem like brain processes. Baum says thought occurs when we are aware of the brain-computer working, but that doesn't seem right. When I think about the world, I am not aware of my brain at all. Baum has no serious answer for that and seems not to grasp the objection.

Another well-known objection points out that a computer program is a syntactical system, a kind of grammar. Semantics, the meaning of the words, is independent of the syntax. ("'Twas brillig, and the slithey toves…" is syntactically correct but meaningless). Grammar does not reveal what words mean and likewise, a recipe doesn't know what the cake tastes like. Even if the brain was a computer program, as a syntax, it would have no capacity for meaning.

Baum addresses that objection by re-defining semantics to be appreciation of perceptual invariants. An invariant is something that does not change when things around it change. Invariants are essentially, Platonic essences. The idea that the brain deals in invariants goes back decades. J.J. Gibson proposed a theory of perception based on detection of the invariant structures of the world. In physics and mathematics, invariant theory is well developed.

DNA was imprinted, through evolution, with the world's invariant contingencies and that causal relationship gives DNA semantic capability (as semantics is now redefined). The DNA program in turn generates the brain program, so the brain/mind also refers the world by being susceptible to its invariants.

It's true, there are neurons in a frog's brain that fire when a moving dot (a "fly") moves across the visual field, and neurons in a cat's brain fire differentially to horizontal and vertical stripes, but that seems a far cry from explaining why I might prefer democracy over fascism.

Baum calls the process of invariant recognition "exploitation," which is no more helpful than Gibson's term for the process, "resonance." So if meaning is exploitation of invariants, we don't know how it works.

Baum declares nevertheless, "…I am offering a theory that is, in my view and I hope by now in the reader's as well, capable of explaining everything" (p. 435). Well, yes. If my only tool is a hammer, all the world is made of nails.

Why should anyone take seriously a theory of mind that includes the principle that theories have no meaning? Words are simply labels for environmental invariants, according to Baum, and that is their only meaning (and we must disregard questions about how language fits into this formulation). By his own rights then, the contents of this book are the traces of somebody's muscle spasms as that person confronted his world. Why should anyone else care about that?

The book could profitably be used in an introductory interdisciplinary course cognitive science, psychology and philosophy. As a persuasive argument, it does not succeed, but the journey is interesting.