The Computer and the Brain [Englisch] [Gebundene Ausgabe]

This is a great book that pushed the limits of his time; his swan song, to be delivered as the Yale Silliman lecture, but never was, due to Von Neumann's tragic untimely death in his early fifties.

The "prima facie" modifier is commonly taken to mean von Neumann saw the brain as "obviously digital," or "patently digital," and that it therefore must resemble a digital computer. But as you read the rest of the book, you quickly discover that this is not what John von Neumann intended. Von Neumann uses words cautiously and precisely, and to him, "Prima facie" means exactly what it says: "on its face."

In 1956, the brain appeared digital. But von Neumann thought this impression might be superficial. He thought that deeper biological investigation might well demonstrate that the nervous system is not, in fact, digital, or not completely digital. He believed it might work in some more sophisticated way, and suggests that perhaps some intermediate signaling mechanism, a hybrid between analog and digital, might be at work in the brain. For this and other reasons he actively resisted labeling the brain as a digital computer.

In the mid 90s, evidence began to appear that von Neumann was probably right to reserve his judgment. These curious new results show that a single nerve impulse is somehow able to convey information to the brain. This signal seems distinctly un-digital. A number of theories have popped up, some attempting to explain this whopping new mystery, others attempting to explain it away. But its impact on neurophysiology, and on conventional computer models of the brain, is pretty shocking. Not to say, devastating. (See Spikes, by Rieke et al, for a readable account of this story.) When the smoke clears, it would not be surprising if people go all the way back to John von Neumann, looking for traction, fresh starting points, and for von Neumann's wonderfully broad sense of what is possible in neurobiology - a sense we have evidently lost to progress in the years since he wrote this splendid essay.

Von Neumann did not include in this book his interesting views on the nervous system of the eye. He was an early adopter of visual memory systems in digital computers, and he was evidently intrigued by the way the retinal cells of the eye are arranged to look backward, that is, toward the screen of the back wall of the eye. Possibly he thought the retinal cells saw back there a thin film diffraction pattern. You can find his interest in the nervous system of the eye remarked in his brother Nicholas Vonneumann's book, John von Neumann as seen by his Brother, and this reminiscence is also paraphrased in Poundstone's Prisoner's Dilemma. Finally, some of the worldly story of von Neumann, his digital computers, and their role in the creation of the hydrogen bomb can be found in MaCrae's biography.

The ideas contained in these lectures will come as no great surprise to most scientists today; indeed, I would expect most to simply nod in agreement at most of von Neumann's observations. For example, von Neumann notes that neurons are essentially digital in that they have an all-or-nothing activation energy. However, it is interesting to see how seriously he pursues the idea that the brain may rely upon a mixture of analog and digital encodings; he took absolutely nothing for granted, and may well have been vastly ahead of his time.

Although von Neumann's many references to vacuum tubes and differential analyzers may seem archaic today, his central points remain essentially intact. I'm certain that von Neumann would have felt somewhat vindicated by the explosive advances in semiconductor devices (in both digital and analog incarnations), as well as in machine learning and neurobiology. One can perhaps view von Neumann's lectures as the first glimmerings of what would eventually become fruitful exchanges between computer science and various biological disciplines.

If you are looking for a discussion that will give you some insight into artificial intelligence, neural networks, or brain physiology, then I'm afraid you will likely be disappointed with this book. While many of von Neumann's observations may have been controversial at the time, they have for the most part moved quietly into the collective consciousness of scientists. However, if you have interest in either the historical development of these ideas, or in seeing how one of the preeminent minds of the 20'th century approached this vexing new problem, then it will be worth your time.

What I most enjoyed about this book is von Neumann's methodical and exceedingly cautious approach, coupled with his occasional willingness to speculate. As the vast majority of von Neumann's writings are accessible only to a very small audience, such as his enormously influential treatises on quantum mechanics, geometry, and game theory, and his pioneering work in areas such as functional analysis and operator theory, this little book is perhaps unique in that it lets you in on the ground floor.