ledge or supposition that a relation or connection exists. In prediction, for instance, of the action of substances according to chemical laws and in reliance on the working of gravitation, we may be unaware of what the connection is, but suppose that there must exist some connection, so that like effects to those experienced in the past are to be expected.

Ordinary scientific induction does not reveal what the connection is, but only that there is a connection which involves effects of a certain character. We seem, however, to have this much knowledge of connection in general-namely, that no existence comes into being of its own accord, and that a particular kind of existence in coming into being involves a particular kind of connection to other existence-the so-called law of causation and the law of uniformity of Nature. In any case, the experienced uniformity of natural forces renders, according to the laws of chance, the supposition of absolute independence of existences highly improbable.

In our own consciousness we seem to be capable of a more or less clear perception of what the connections upon which inference is based really are. For instance, we seem not only to suppose that there must be some connection between pain and the avoidance of the source of the pain, but also to understand what that connection is; and in virtue of this understanding we hold ourselves capable of predicting that painful objects will, as a general rule, be avoided.

It is not here contended that, whenever on perception of such relation or connection we infer some existence not experienced, the inference is always quite certain. For it may be that our perception of the connection is not clear, or we may overlook the possibility of influences which would prevent the normal result. But in so far as we do perceive the connections in which various existences, just because they are what they are, stand to one another, so far can we safely

infer what lies outside our experience.

And if we are aware that there are such connections, even though we do not know what they are, then we may with some probability argue and draw conclusions. We do not always, or even usually, arrive at general propositions by an elaborate induction, but rather postulate these propositions and then test them by experience. But the only ground for accepting them as true when they do fulfil the test applied is that this indicates a more or less necessary relation, on account of which a like result is commonly to be depended on.

interconnection of

The knowledge of necessary existences being the ground of inference, can we make inferences with certainty, or any degree of probability, about those existences which constitute the subject-matter of theology? It will be convenient to examine the manner in which Kant limited the sphere of legitimate inference, with the view to answering the question whether and to what extent theology is possible.

Kant, in his "Critique of Pure Reason," claimed to state the true via media between speculative philosophy on the one hand, with its supposed conclusions about realities beyond human ken, and radical empiricism on the other hand, which involved complete scepticism about all except actual experience. The human mind, he maintained, can validly reason about matters of possible human experience. This he held to be due to the fact that the mind imposes certain forms on the matter of experience; whence it can know that it will always discover these forms in experience. We can be sure of certain propositions about any possible experience-for example, geometrical and causal relations-simply because these relations are imposed by our own minds.

But this theory of knowledge is untenable, though it may point to truth. For, in the first place, know

ledge of the a priori faculties of the mind whereby it forms and orders experience must be attained by inference, as they are not part of the content of experience. For surely it would come as a surprise to most people to be told that space and time and the regularity of causation are due to powers of their own mind. These faculties are not, generally at least, either things of experience or of possible experience. But if so, how can we be certain that they operate with strict uniformity, so that the laws of space and time and causation will always hold good of our experience? How indeed on Kantian principles could we know at all about these faculties?

Again, the implication in Kant's doctrine of causality seems to be that the causal nexus and uniformity in the abstract are imposed by the mind, whereas the particular content of causal connections is given by thingsin-themselves. Perhaps there is a little fluctuation as regards this, the things-in-themselves being occasionally conceived as giving mere sensation in general, or even just the impulse to mental construction. But the former description of their function seems more in accord with Kant's general theory. For instance, in discovering the properties of a hitherto unknown chemical substance, it is to be supposed that the mind determines that the chemical shall act and react regularly, but the particular manner of its action and reaction is due to the thing-in-itself. This general causal regularity is all that the mind can be certain of a priori, and, presumably, therefore all that springs from the mind. It would seem, on Kant's theory, that an event might be causally connected with almost any other event as far as the mind is concerned, but that if once thingsin-themselves have so acted on a human mind that the mind causally connects two events, things-inthemselves are under the necessity of continuing to act in this definite way on all human minds. But this would imply that the human mind controls things

in-themselves in a way far beyond the influence of ordinary human volition; as if things-in-themselves had a choice of many modes of combining events, but having chosen one were obliged to keep to it-an incredible idea, and one quite contrary to the spirit of the Kantian philosophy. Or else things-in-themselves act regularly of themselves, in which case the uniformity of causation belongs to them, and is rather recognized than made by the mind.

Thirdly, Kant's principle that knowledge can only be of matters of possible experience, if taken strictly, involves hypothetical solipsism. Since we cannot, under ordinary human conditions at least, experience the consciousness of other men, we cannot, on the Kantian principle, know about them. Or, to put it differently, the sensations which suggest to us other minds and their ideas are equally with the sensations of material objects not the work of our minds, but affections of our sensibility caused from without. But the objects of knowledge in both cases are the result of synthesis by means of mental relations which do not, so far as we know, correspond to anything in the source or sources of the sensations. Why then, on Kant's theory, should we believe that our ideas of other minds have independent existences corresponding to them, if there are not independent existences corresponding to our ideas of inanimate things? On the other hand, if we can so far overstep the bounds of our experience as to be certain of both the existence and some of the thoughts and feelings of other minds, on what ground are we to be debarred from reasoning about other objects which we cannot experience?

Kant was prompted to his theory of knowledge by distrust of the facile and mutually conflicting cosmologies of speculative philosophy. But in avoidance of this he did not wish to be pushed into radical empiricism, Hume's doctrine practically, that we can know only what we experience. Between his Scylla and

1

his Charybdis, the claim to know all and the claim to know only what one is immediately conscious of, he traced a middle way-the claim to know only things of possible experience. But we should ask, "In what sense possible?" Except through clairvoyance, it is not possible for human beings to see the other side of the Moon. It is not possible for people in Europe to see what will take place in an hour's time in China. It is not possible to see an event which happened yesterday. Yet all these are matters of possible experience in Kant's sense.

It seems that there are two conditions for our knowing and reasoning about existence which we do not experience. The existence must be of like nature to existence of which we have experience. And it must be connected with existence of which we have experience through existence of like nature to existence of which we have experience. Two observations are called for. First, memory enables us in a manner to revive experience, so that inferences may be drawn by means of past experience. Secondly, material processes we know as forms of causation in general, causation being experienced in our own psychical processes; certain laws to which substances conform are deduced from the conceptions of space and time and are verified by facts; and different kinds of substances are distinguished by differences in their sensible effects, and are found, so far as we can test them, to be fairly constant in their mutual relations.

Is then valid theology possible? Let us first remark that there is a tendency in man to speculative philosophy and theology, which gives, prima facie, a presumption in favour of the possibility of valid theology. No doubt much of the reasoning of theologians has been inconclusive, and has even led to false results. But that shows that theology is difficult, not that it is impossible. Well then, are there the conditions of valid reasoning about non-experienced existence in the subject-matter of theology?