Process and Reality — the Gifford Lectures, 1929

This astronomical contingency, and the beliefs which cluster round it, have been stated with some detail, because—as thus expressed—they illustrate the problem as it shapes itself in philosophy. Also the example of the railway time-tables illustrates another point. For it is possible momentarily, in Vermont on July 1, 1927, to forget that the unprecedented Mississippi floods happened during that May and June; so that although the estimate as to error in punctuality was justified by the evidence consciously before us, it did not in fact allow for the considerable derangement of the traffic in some states in the Union.* The point of this illustration from railway trains is that there is a conformity to matter of fact which these judgments exhibit, even if the events concerned have not happened, or will not happen. These considerations introduce the fundamental principle concerning ‘judgment.’ It is that all judgment is categorical; it concerns a proposition true or false in its application to the actual occasion which is the subject making the judgment. This doctrine is not so far from Bradley’s doctrine of judgment, as explained in his Logic. According to Bradley, the ultimate subject of every judgment is the one ultimate substance, the absolute. Also, according to him, the judging subject is a mode of the absolute, self-contradictory if taken to be independently actual. For Bradley, the judging subject has only a

* Since this sentence was written in July, 1927, a star has unexpectedly split in two, in March, 1928.

“Still less, at the time of writing this sentence, were the Vermont floods of November, 1927, foreseen.

p. 305

derivative actuality, which is the expression of its status as an affection of the absolute. Thus in Bradley’s doctrine, a judgment is an operation by which the absolute, under the limitations of one of its affections, enjoys self-consciousness of its enjoyment of affections. It will be noticed that in this bald summary of Bradley’s position, I am borrowing Spinoza’s phrase, ‘affectiones substantia.’

In the philosophy of organism, an actual occasion—as has been stated above—is the whole universe in process of attainment of a particular satisfaction. Bradley’s doctrine of actuality is simply inverted. The final actuality is the particular process with its particular attainment of satisfaction. The actuality of the universe is merely derivative from its solidarity in each actual entity. It must be held that judgment concerns the universe as objectified from the standpoint of the judging subject. It concerns the universe through that subject.

With this doctrine in mind, we pass to the discussion of the sense in which probability can be a positive fact in an actual entity; so that a proposition expressing the probability of some other proposition can in this respect agree or disagree with the constitution of the judging entity. The notion of ‘probability,’ in the widest sense of that term, presents a puzzling philosophical problem. The mathematical theory of probability is based upon certain statistical assumptions. When these assumptions hold, the meaning of probability is simple; and the only remaining difficulties are concerned with the technical mathematical development. But it is not easy to understand how the statistical theory can apply to all cases to which the notion of more or less probability is habitually applied. For example, when we consider—as we do consider—the probability of some scientific conjecture as to the internal constitution of the stars, or as to the future of human society after some unprecedented convulsion, we seem to be influenced by some analogy which it is very difficult to convert into an appeal to any definite statistical fact. We may consider that it is probablep. 306 that the judgment could be justified by some statistical appeal, if we only knew where to look. This is the belief that the statistical probability is itself probable. But here, evidently, there is an appeal to a wider meaning of probability in order to support the statistical probability applicable to the present case. It is arguable that this wider probability is itself another statistical probability as to the existence of the special statistics relevant to such types of scientific argument. But in this explanation puzzling questions are accumulating; and it is impossible to avoid the suspicion that we are being put off with one of those make-believe explanations, so useful to reasoners who are wedded to a theory. The philosophy of organism provides two distinct elements in the universe from which an intuition of probability can originate. One of them is statistical. In this and the next section, an attempt will be made to justify the statistical theory. It is therefore the more imperative to survey carefully the difficulties which have to be met.

In the first place, probability is always relevant to evidence; so, on the statistical theory, the numerical probability will mean the numerical ratio of favourable to unfavourable cases in the particular class of ‘cases’ selected as the ‘ground’ for statistical comparison. But alternative ‘grounds’ certainly exist. Accordingly we must provide a reason not based upon ‘probability,’ why one ‘ground’ is selected rather than another. We may admit such a chain of vaguer and vaguer probabilities, in which our first ground is selected as statistically probable in respect to its superiority to other ‘grounds’ of other types. We are thus driven back to a second-order ‘ground’ of probability. We may logically proceed to third-order ‘grounds,’ and so on. But if the statistical theory is to be substantiated, after a finite number of steps we must reach a ‘ground’ which is not selected for any reason of probability. It must be selected because it is the ‘ground’ presupposed in all our reasonings.

Die PR OR OS LELONS 307

Apart from some such ultimate ‘ground,’ the statistical theory, viewed as an ultimate explanation for all our uses of the notion of ‘probability,’ must inevitably fail. This failure arises by reason of the complete arbitrariness of the ultimate ‘ground’ upon which the whole estimate of probability finally rests.