BACKGROUND

The twentieth century has been filled with the most incredible shocks and surprises: the theory of relativity, Communist revolutions, psychoanalysis, nuclear war, television, moon walks, genetic engineering, and so on. As astounding as any of these is the advent of the computer and its development from a mere calculating device into what seems like a "thinking machine." The birth of the computer was not wholly independent of the other events of this century. The history of the computer is a fascinating story; however, it is not the subject of this course. We are concerned with the Theory of Computers, which means that we form several abstract mathematical models that will describe with varying degrees of accuracy parts of computers and types of computers and similar machines. Our models will not be used to discuss the practical engineering details of the hardware of computers, but the more abstract questions of the frontiers of capability of these mechanical devices. There are separate courses that deal with circuits and switching theory (computer logic) and with instruction sets and register arrangements (computer ar-chitecture) and with data structures and algorithms and operating systems and compiler design and artificial intelligence and so forth. All of these courses have a theoretical component, but they differ from our study in two basic ways. First, they deal only with computers that already exist; our models, on

3

4

AUTOMATA THEORY

the other hand, will encompass all computers that do exist, will exist, and that can ever be dreamed of. Second, they are interested in how best to do things; we shall not be interested in optimality at all, but rather we shall be concerned with the question of possibility-what can and what cannot be done. We shall look at this from the perspective of what language structures the machines we describe can and cannot accept as input, and what possible meaning their output may have. This description of our intent is extremely general and perhaps a little misleading, but the mathematically precise definition of our study can be understood only by those who already know the concepts introduced in this course. This is often a characteristic of scholarship---after years of study one can just begin to define the subject. We are now embarking on a typical example of such a journey. In our last chapter (Chapter 31) we shall finally be able to define a computer. The history of Computer Theory is also interesting. It was formed by fortunate coincidences, involving several seemingly unrelated branches of intellectual endeavor. A small series of contemporaneous discoveries, by very dissimilar people, separately motivated, flowed together to become our subject. Until we have established more of a foundation, we can only describe in general terms the different schools of thought that have melded into this field. The most obvious component of Computer Theory is the theory of mathematical logic. As the twentieth century started, mathematics was facing a dilemma. Georg Cantor (1845-1918) had recently invented the Theory of Sets (unions, intersections, inclusion, cardinality, etc.). But at the same time he had discovered some very uncomfortable paradoxes-he created things that looked like contradictions in what seemed to be rigorously proven mathematical theorems. Some of his unusual findings could be tolerated (such as that infinity comes in different sizes), but some could not (such as that some set is bigger than the universal set). This left a cloud over mathematics that needed to be resolved. David Hilbert (1862-1943) wanted all of mathematics put on the same sound footing as Euclidean Geometry, which is characterized by precise definitions, explicit axioms, and rigorous proofs. The format of a Euclidean proof is precisely specified. Every line is either an axiom, a previously proven theorem, or follows from the lines above it by one of a few simple rules of...