The Greeks were not addicted to moderation, either in their theories or in their practice. Heraclitus maintained that everything changes: Parmenides retorted that nothing changes.
Parmenides was a native of Elea, in the south of Italy, and flourished in the first half of the fifth century B.C. According to Plato, Socrates in his youth (say about the year 450 B.C.) had an interview with Parmenides, then an old man, and learnt much from him. Whether or not this interview is historical, we may at least infer, what is otherwise evident, that Plato himself was influenced by the doctrines of Parmenides. The south Italian and Sicilian philosophers were more inclined to mysticism and religion than those of Ionia, who were on the whole scientific and sceptical in their tendencies. But mathematics, under the influence of Pythagoras, flourished more in Magna Graecia than in Ionia; mathematics at that time, however, was entangled with mysticism. Parmenides was influenced by Pythagoras, but the extent of this influence is conjectural. What makes Parmenides historically important is that he invented a form of metaphysical argument that, in one form or another, is to be found in most subsequent metaphysicians down to and including Hegel. He is often said to have invented logic, but what he really invented was metaphysics based on logic.
The doctrine of Parmenides was set forth in a poem On Nature. He considered the senses deceptive, and condemned the multitude of sensible things as mere illusion. The only true being is 'the One', which is infinite and indivisible. It is not, as in Heraclitus, a union of opposites, since there are no opposites. He apparently thought, for instance, that 'cold' means only 'not hot', and 'dark' means only 'not light'. 'The One' is not conceived by Parmenides as we conceive God; he seems to think of it as material and extended, for he speaks of it as a sphere. But it cannot be divided, because the whole of it is present everywhere.
Parmenides divides his teaching into two parts, called respectively 'the way of truth' and 'the way of opinion'. We need not concern ourselves with the latter. What he says about the way of truth, so far as it has survived, is, in its essential points, as follows:
'Thou canst not know what is not—that is impossible—nor utter it; for it is the same thing that can be thought and that can be.'
'How, then, can what is be going to be in the future? Or how could it come into being? If it came into being, it is not; nor is it if it is going to be in the future. Thus is becoming extinguished and passing away not to be heard of.
'The thing that can be thought and that for the sake of which the thought exists is the same; for you cannot find thought without something that is, as to which it is uttered.'1
The essence of this argument is: When you think, you think of something; when you use a name, it must be the name of something. Therefore both thought and language require objects outside themselves. And since you can think of a thing or speak of it at one time as well as at another, whatever can be thought of or spoken of must exist at all times. Consequently there can be no change, since change consists in things coming into being or ceasing to be.
This is the first example in philosophy of an argument from thought and language to the world at large. It cannot of course be accepted as valid, but it is worth while to see what element of truth it contains.
We can put the argument in this way: if language is not just nonsense, words must mean something, and in general they must not mean just other words, but something that is there whether we talk of it or not. Suppose, for example, that you talk of George Washington. Unless there were a historical person who had that name, the name (it would seem) would be meaningless, and sentences containing the name would be nonsense. Parmenides maintains that not only must George Washington have existed in the past, but in some sense he must still exist, since we can still use his name significantly. This seems obviously untrue, but how are we to get round the argument?
Let us take an imaginary person, say Hamlet. Consider the statement 'Hamlet was Prince of Denmark.' In some sense this is true, but not in the plain historical sense. The true statement is 'Shakespeare says that Hamlet was Prince of Denmark,' or, more explicitly, 'Shakespeare says there was a Prince of Denmark called "Hamlet".' Here there is no longer anything imaginary.
Shakespeare and Denmark and the noise 'Hamlet' are all real, but the noise 'Hamlet' is not really a name, since nobody is really called 'Hamlet'. If you say '"Hamlet" is the name of an imaginary person,' that is not strictly correct; you ought to say 'It is imagined that "Hamlet" is the name of a real person.'
Hamlet is an imagined individual; unicorns are an imagined species. Some sentences in which the word 'unicorn' occurs are true, and some are false, but in each case not directly. Consider 'a unicorn has one horn' and 'a cow has two horns'. To prove the latter you have to look at a cow; it is not enough to say that in some book cows are said to have two horns. But the evidence that unicorns have one horn is only to be found in books, and in fact the correct statement is: 'Certain books assert that there are animals with one horn called "unicorns".' All statements about unicorns are really about the word 'unicorn', just as all statements about Hamlet are really about the word 'Hamlet'.
But it is obvious that, in most cases, we are not speaking of words, but of what the words mean. And this brings us back to the argument of Parmenides, that if a word can be used significantly it must mean something, not nothing, and therefore what the word means must in some sense exist.
What, then, are we to say about George Washington? It seems we have only two alternatives: one is to say that he still exists; the other is to say that, when we use the words 'George Washington', we are not really speaking of the man who bore that name. Either seems a paradox, but the latter is less of a paradox, and I shall try to show a sense in which it is true.
Parmenides assumes that words have a constant meaning; this is really the basis of his argument, which he supposes unquestionable. But although the dictionary or the encyclopaedia gives what may be called the official and socially sanctioned meaning of a word, no two people who use the same word have just the same thought in their minds.
George Washington himself could use his name and the word 'I' as synonyms. He could perceive his own thoughts and the movements of his body, and could therefore use his name with a fuller meaning than was possible for any one else. His friends, when in his presence, could perceive the movements of his body, and could divine his thoughts; to them, the name 'George Washington' still denoted something concrete in their own experience. After his death they had to substitute memories for perceptions, which involved a change in the mental processes taking place when they used his name. For us, who never knew him, the mental processes are again different. We may think of his picture, and say to ourselves 'yes, that man'. We may think 'the first President of the United States'. If we are very ignorant, he may be to us merely 'The man who was called "George Washington".' Whatever the name suggests to us, it must be not the man himself, since we never knew him, but something now present to sense or memory or thought. This shows the fallacy of the argument of Parmenides.
This perpetual change in the meanings of words is concealed by the fact that, in general, the change makes no difference to the truth or falsehood of the propositions in which the words occur. If you take any true sentence in which the name 'George Washington' occurs, it will, as a rule, remain true if you substitute the phrase 'the first President of the United States'. There are exceptions to this rule. Before Washington's election, a man might say 'I hope George Washington will be the first President of the United States,' but he would not say 'I hope the first President of the United States will be the first President of the United States' unless he had an unusual passion for the law of identity. But it is easy to make a rule for excluding these exceptional cases, and in those that remain you may substitute for 'George Washington' any descriptive phrase that applies to him alone. And it is only by means of such phrases that we know what we know about him.
Parmenides contends that, since we can now know what is commonly regarded as past, it cannot really be past, but must, in some sense, exist now. Hence he infers that there is no such thing as change. What we have been saying about George Washington meets this argument. It may be said, in a sense, that we have no knowledge of the past. When you recollect, the recollection occurs now, and is not identical with the event recollected. But the recollection affords a description of the past event, and for most practical purposes it is unnecessary to distinguish between the description and what it describes.
This whole argument shows how easy it is to draw metaphysical conclusions from language, and how the only way to avoid fallacious arguments of this kind is to push the logical and psychological study of language further than has been done by most metaphysicians.
I think, however, that, if Parmenides could return from the dead and read what I have been saying, he would regard it as very superficial. 'How do you know,' he would ask, 'that your statements about George Washington refer to a past time? By your own account, the direct reference is to things now present; your recollections, for instance, happen now, not at the time that you think you recollect. If memory is to be accepted as a source of knowledge, the past must be before the mind now, and must therefore in some sense still exist.'
I will not attempt to meet this argument now; it requires a discussion of memory, which is a difficult subject. I have put the argument here to remind the reader that philosophical theories, if they are important, can generally be revived in a new form after being refuted as originally stated. Refutations are seldom final; in most cases, they are only a prelude to further refinements.
What subsequent philosophy, down to quite modern times, accepted from Parmenides, was not the impossibility of all change, which was too violent a paradox, but the indestructibility of substance. The word 'substance' did not occur in his immediate successors, but the concept is already present in their speculations. A substance was supposed to be the persistent subject of varying predicates. As such it became, and remained for more than two thousand years, one of the fundamental concepts of philosophy, psychology, physics, and theology. I shall have much to say about it at a later stage. For the present, I am merely concerned to note that it was introduced as a way of doing justice to the arguments of Parmenides without denying obvious facts.