It is one of the great services of relativity theory to philosophy that it provides an answer to this question. The answer is that the entities discussed by physics are defined in terms of one another. The three hundred years of building up exact science really amounts, in the last analysis, to doing what the dictionary compiler did when he defined a violin as a small violoncello and a violoncello as a large violin. Of course, if this statement were literally true, science would give us no information about the universe at all. Nevertheless, the statement is true about the actual procedure of science, and it is in virtue of this procedure that science forms a closed system. But what is left out of this description is the scientist himself. The mysterious process which is not taken into account in this description of the scientific method is the process by which the consciousness of the scientist makes contact with the entities he is talking about. In deducting the world from “point-events”, for instance, we begin by talking about something we have no direct cognisance of, namely point-events. From point-events we deduce “potentials”――again a mere word. But from potentials we deduce “matter”, and here we are talking of something of which we have direct knowledge. Similarly, the circular definition of violin and violoncello tells us nothing as it stands. But to a man who can identify one of these entities, to a man who has ever seen a violin, it gives genuine information. We need not be surprised, therefore, that nothing from the outside ever seems to disturb the equanimity that reigns within the closed system of physics. The abstractions with which it begins are all it ever has to deal with. There are no subsequent fresh contacts with reality. If the region covered by relativity theory embraced the whole of physics it would seem that, so far as physical science is concerned, we knew all that there is to be known. But it is notorious that, of recent years, an entirely new set of phenomena has been discovered in physical science. These phenomena arise when we consider, not matter in bulk but matter in its smallest particles. These phenomena are, at present, strictly incomprehensible. The celebrated quantum theory provides us with rules for dealing with some of them, but does not make them intelligible. It seems that science has here reached its limits. Professor Eddington has even hinted that these phenomena may indicate that the universe is finally irrational, that is, that the attempt to describe nature mathematically will have to be given up. This is a possibility that Newton foresaw. But it seems more likely that our present state of bewilderment has a different cause. That cause, we shall probably find, is the insufficiency of the abstractions hitherto used in science. We have to go back to the concrete facts of experience and build up a richer, fuller set of abstractions. Physics is now paying the penalty of inadequate abstraction.