The present book has two origins, one very remote, the other nearer and more contingent. The first goes back to the time when I initiated my career as a neurophysiologist in Pisa, a small town with the advantage of a highly stimulating atmosphere created by two famous institutions, the University and the Scuola Normale Superiore. It came quite natur ally, then, while engaged in experimental work, to start brooding over the possible analogies between neurophysiological problems and those of the physical world. This slowly induced me to become less interested in the solution of the innumerable specific problems presented by the brain, and more in the general principles on which the brain function might be based. Certainly, for several years I had no clear idea of my purposes, or of the difficulties I could encounter in the task. However, it was clear enough that there ought to be a first, indispensable step: the search for methods of quantification of nervous activity, the sole way of allowing predictions about its behavior. At first, I somehow followed the fashion of the time, experiencing the impact of information theory on neurophysiology, but soon this was revealed as unsatisfactory, since it was only one aspect of the problem, and what I was interested in was not a way of describing the flow of information, but rather, the laws of the machinery.