obvious that no explanation of B has been given, and that neither A nor B is understood. A is potentially B,' if it means anything, must mean that in some way A already is B, and that B is needed to explain A. The late G. H. Lewes was not prejudiced in favour of old philosophies, but he most fully recognises the fact that we can only understand the lower from the point of view of the higher: "We can only understand the Amoeba and the Polype by a light reflected from the study of Man". So that even within the sciences it is not really possible to begin at the beginning'. The attempt to do so will generally mean that some dimly accepted view about the "end" is influencing the observations of the beginning; for, as Lewes reminds us, closest observation is interpretation". Even for the study of origins an examination of the end or most complete state as it exists is not superfluous, and such an examination, apart from historical methods, must be analytic, or, in Kant's phrase, critical. Before we proceed to ask what history tells us, it may be worth while to ask what history can tell us. By knowing what something was, we do not always know what it is, sometimes only what it (now) is not.

our

To discover the à priori element in knowledge, i.e., that element which, though known to us only in connexion with sense-experience, cannot be dependent upon sense-experience for its validity, is the business of a philosophical theory of knowledge. And if we call that a part of Metaphysics, it is a Metaphysics with which we cannot dispense. Suppose that Self-consciousness,' 'Identity,' 'Substance,' 'Cause,' 'Time,'' Space,' be amongst the 'Categories' so discovered, to arrange these categories in a system, to see their relations to one another and to the world of nature and of human action, will be the business of Philosophy or Metaphysics in a wider sense. 'Speculative Metaphysics,' as distinct from Critical, we might call it, because the method it must adopt can never have the logical precision and certainty of the Critical Method. The only test of the validity of a system of Speculative Metaphysics must be its adequacy to the explanation and arrangement of the whole Universe as it becomes known to us. Thus this Metaphysics can never be complete, but must always be attempted anew by each thinker. The Critical examination of the nature of knowledge may logically precede any or all of the special sciences, although it is only the advance of science that has suggested the need of such an examination; but the Metaphysician in this second sense can never be independent of any of the

1Study of Psychology, p. 122.

sciences or of any branch of human knowledge or effort. They are his material.

To make knowledge possible there must (in Green's phrase) be "a comparing and distinguishing self"; but since Time, though relatively a form, is yet also one of the contents of knowledge, this self must in some way be independent of Time. I know I am a series of experiences in Time. Therefore, in some way, I am not in Time-but an Eternal (i.e., Time-less) Self-consciousness. But the Critical Philosophy can tell us nothing further, can tell us nothing as to what this Eternal Self-consciousness is or how it is related to our individual selves, which are the subject-matter of Psychology. The attempt to find some expression for this relation, ie, to show how an Eternal Self-consciousness reveals itself in Time and in Space is the business of Speculative Philosophy or Metaphysics. That there is an Eternal Self-consciousness we are logically compelled to believe, and that it is in some way present in our individual selves; but in what way is a matter of speculation: and it is still quite competent to any one who accepts the main result of the critical examination of knowledge to maintain that this latter problem is altogether insoluble; although it is a problem (or rather series of problems) which we cannot leave alone, because we are met by it at every step in our ordinary experience, if we only begin to reflect on the meaning and mutual relations of the conceptions we are obliged to use.

It is not our present concern to give an exhaustive list of the à priori conceptions and principles which are involved in ordinary knowledge and in the procedure of scientific investigation and proof. An Intuitionist Philosophy, which professes to get at these principles by a simple introspection into the contents of consciousness, may fairly be met with the challenge to produce its list of intuitive principles. But if the term à priori be understood in the way which has been explained above, no such challenge can be justly made. It is only as experience progresses that we can become fully aware of and can formulate the conceptions and principles which that experience logically involves. Only if knowledge were completed, could we know all that knowledge implied: and it is only as knowledge approximates to that apparently ever-receding goal that we can enlarge our view of what has been there implicitly from the first. Thus, in the very simplest acts of thought the principle of Identity and the principle of Contradiction (A is A; A is not not-Ă) are involved; and yet it was late in the history of mankind when the science of Logic was first enabled to discover and formulate these principles. Nevertheless they are à priori in the

sense that without them all knowledge would be impossible. So it is with the axioms of the science of quantity. That "things which are equal to the same thing are equal to one another" is implied in all the experience which Mill thought went to prove the principle. Every carpenter who uses a foot-rule, every barmaid who draws off half-a-pint implies the principle and acts on it, though totally ignorant of the elements of Geometry. Similarly, the rudest ideas about Nature imply the conception of a Cosmos, of an order of nature, though that order may include gods, demons, fairies, and goblins, of whom the modern scientific man takes no account, and may exclude gravitation, electricity, and other forces which he has come to recognise. The principle that every event has a cause, i.e., is related to some other event (or events) without which it would not happen and with which it must happen, the two clauses of this definition of cause are sometimes mistakenly separated as the principles of Causation and Uniformity of Nature respectively, is involved in the mental action of the savage who hears the thunder and looks round for an explanation, though he may be quite wrong in his explanation, and though it may be late in time before any human being comes to reflect on the processes of experience and to formulate its principles. But the history of how men came to recognise Uniformity of Nature and how their conceptions of Cause and of Nature have varied is one thing: the logical character of the presupposition of all inductive inference is another. The former is a question of historical psychology; the latter of philosophical criticism. The proposition, "Every event must have a cause is not à priori because it convinces every person the moment he understands it, but because no knowledge of natural events is possible without a connexion of them with other events as belonging to one system of nature. That nature is a system is the assumption underlying the earliest mythologies: to fill up this conception is the aim of the latest science. capacity for discovering true causes may be capable of development as the race advances; so may be a capacity for philosophical analysis; but the presupposition of all investigation of causes cannot itself be derived from the experience either of the individual or of the race.

The question for the logician is not: 'How have I (or mankind generally) come to believe this?' That is a question for the psychologist and sociologist. The logical question is: Why am I or any one else justified in believing this?' A confusion between these two questions underlies Mill's famous attack on the Syllogism. The essential and permanently significant portion of the Aristotelian doctrine

of the Syllogism is the recognition that all inference (and συλλογισμός just means 'inference') implies a Universal. As a psychological fact there may (though even this may be questioned) be in our minds a particular proposition and then immediately afterwards another particular proposition suggested by it. But, if the one can be described as an inference from the other, we must be able to answer the question why we get the one from the other. And the answer to the question must, if we formulate it, take the form of a universal proposition, of which, till we have to face the question, we may be perfectly unconscious, and will constitute the major premiss of the Aristotelian Syllogism (Barbara or Darii being taken as typical), the middle term being, in the scientific inference, the cause or ground (sufficient reason) of the conclusion. Thus the death of some one I know may suggest to me my own mortality; but the reason of the inference is our common possession of the attributes of human and so of animal life. It is always with a question of validity that the logician as such has to deal: Are we justified in inferring that?'-not with the psychological process through which any particular person or persons have gone in arriving at their beliefs. Psychological introspection can, therefore, never solve logical difficulties. The formula of the Syllogism (major premiss, minor premiss, conclusion) is not an exposition of what actually takes place in any one's mind, but a logical exposition of that to which any actual inference must conform in order to be correct. It would not even be accurate to say it is the form according to which the normal reasoner actually reasons; because a man may reason quite correctly and be the normal reasoner while quite unconscious of logical analysis. The reasonings of the normal reasoner are those which will conform best to the strict syllogistic form, when they are so analysed by the logician. The incorrectness of an apparent inference becomes clear, when the reasoner is compelled to formulate the universal according to which he is reasoning, though without being aware of it. "If he were fully aware of it, he could not commit fallacies. If we were fully aware of everything that every proposition implies, we could not assert false propositions.

Take another logical illustration, a minor matter. Mill says that proper names have no connotation. It may be true enough that the name John Smith' suggests nothing to me or to you; but, if I am a philological ethnologist, it may suggest a good deal; if I have a friend of that name, it may suggest a good deal more. These are matters of psychological interest, and no definite answer independent of time,

place, circumstances and persons can be given. But the name of an individual, not as a mere word, but as the name of an individual, must logically have an infinite connotation. That we can say quite definitely, and that is the reason why the proper name cannot be defined. Any given person may be unable to say anything about any given proper name; whether he can or not is a matter of fact. But logic has to do with the ideal possibilities of definition. And we can answer quite certainly: We never can exhaust the signification of the individual.

The controversy whether mathematical judgments are analytic or synthetic is of a similar kind. As a psychological question it is a matter of degree, and, in the case of arithmetic, will depend solely on the extent to which a person has learned the multiplication table, &c. This is one of the merely psychological distinctions that intrude themselves into Kant's theory of knowledge. Whether any proposition conveys any new information to a person is always a question which cannot be answered irrespective of time, place, &c. In one sense nothing we ever can learn is new, else we could not learn it: it would be quite irrelevant to our already existing knowledge. (This is the truth in the old Sophistic paradox.) In this way all reasoning is reasoning in a circle; but it is a circle so large-as large as the Universe-that we need be under no immediate fear of completing it. To omniscience all propositions must be analytic (identical). That is the ideal of knowledge, and it is the standard by which all statements and all professed inferences are ultimately judged. This amounts to saying, in other words, that the inconceivability of the opposite is the ultimate test of all truth. Only it is a test that we cannot safely apply in practice, except where we can be perfectly sure that we have eliminated all risks of ambiguity and have fully realised all the conditions under which we are making an assertion. Thus we can only safely apply it in very abstract sciences, such as geometry. We know exactly what we mean and what others will understand by a "straight line," and by "enclosing a space"; and therefore we can quite certainly say "Two straight lines cannot enclose a space"; because to suppose that they do involves us in contradiction, and would make us assert that the straight line was also not a straight line. But, if any one at the beginning of this century had said 'It is inconceivable that a message should be sent from London to New York in a few minutes,' his statement would only be correct if he were to insert the qualification: 'the modes of transmitting messages being such as I know of'; for then it would be