It is easy to point out some of the conditions upon which failure to detect actually existing individual deviations from uniformity may depend. Professor Royce has, in this connection, laid special stress upon one such condition, the limitation of what he calls the time-span of our attention. We are unable, as the student of Psychology knows, to attend to a process as a whole if its duration exceeds or falls short of certain narrow limits. Now, there seems no foundation in the nature of the attentive process for the special temporal limitations to which it is subject in our own experience, and we have no means of denying the possibility that there may be intelligent beings whose attention-span is much wider, or again, much more contracted, than our own. One can even conceive the possibility of a being with a power of varying the span of attention at will. Now, it is clear that if we could so vary our attention-span as to be able to take in as single wholes processes which are at present too rapid or too slow to be perceived by us in their individual detail, such a purely subjective change in the conditions of our own attention might reveal individuality and purpose where at present we see nothing but routine uniformity. In the same way, we can readily understand that a being with a much wider attention-span than our own might fail to see anything but purposeless routine in the course of human history. Supposing that we are placed in the midst of a universe of intelligent purposive action, it is clear that we can only hope to recognise the nature of that action in the case of beings who live, so to say, at the same rate as ourselves. A purposive adaptation to environment with consequent deviation from uniformity in reaction would necessarily escape our notice if it took place with the rapidity of the beat of a gnat’s wing, or again, if it required centuries for its establishment.[[135]]

Other similar subjective conditions which would necessarily cut us off from the recognition of purposive fresh adaptations widely different from those which occur in our own life, are the limitations of our power of attending to more than a certain number of presentations simultaneously; and again, the restriction of our sense-perception to a few types, and the impossibility of perceiving contents belonging to those types when they fall below or above the lower and upper “thresholds” of sensibility. These considerations do not, of course, positively prove that the routine uniformity of natural processes is only subjective appearance, but they are sufficient to show that there is no valid reason for taking it to be more, and in conjunction with our previous positive argument for the sentient individuality of all real existence, they suffice to bring our general interpretation of the physical order under Mr. Bradley’s canon that “What must be and can be, that is.”

§ 4. (3) What, then, are we to make of the principle of the “Uniformity of Nature”? Any principle which does actual work in science must somehow be capable of justification, and if our interpretation of the physical order really conflicts with a fundamental scientific principle, it must contain fallacy somewhere. Fortunately, there is no real conflict. In dealing with the principle of Uniformity, we must distinguish very carefully between the sense in which it is actually required for the purposes of science and the sense which has been put upon it in the set of metaphysical doctrines popularly but illogically deduced from the actual procedure of the sciences. As we have seen already, it is impossible to affirm the principle of Uniformity as an axiom of systematic thought. It is also not capable of verification as an empirical truth. Its logical character must therefore be that of a postulate, an assumption defensible on the ground of practical usefulness, but only so far as it actually succeeds.

Now, this is precisely the place which the principle fills in the actual procedure of the sciences. We have absolutely no means of showing that the concrete course of Nature is strictly uniform, as has already been seen. But also, we have no need, for our scientific ends, that it should be uniform. All that we require is that natural processes, when dealt with in the bulk, should exhibit no divergence from uniform routine except such as we may neglect for the purposes of practical calculation and control of the course of events. The actual success of the empirical sciences shows that this demand for approximate uniformity is actually fulfilled with sufficient closeness for all our practical purposes. That it would be so fulfilled we could have had no theoretical means of divining before putting it to the actual test. In this sense the principle, like that of Causality may be said to be a postulate made a priori and in advance of experience. But, once more like the principle of Causality, it could not be presumed to be trustworthy unless the subsequent results of its employment vindicated it; it cannot, therefore, be a priori in the Kantian sense of being known to be true independent of empirical verification.[[136]]

This result is confirmed by consideration of the way in which the principle of uniform law is actually applied to concrete cases. Scientific laws, as we all know, are purely general and abstract. They state not what will happen, but what would happen providing that certain specified conditions and no others were operative in determining the result. In this abstract form they are, of course, statements of exact and absolute uniformities. But in this abstract form they cannot be directly applied to the calculation of the actual course of any process. To take, for instance, an example which has been used by Professor Ward.[[137]] We learn in Mechanics that equilibrium is maintained on the lever when the moments of the weights about the fulcrum are equal and opposite. As an abstract generalisation this is a statement of a rigid uniformity. But in order that it may be universally true, we must suppose the conditions implied in the formulation of the proposition to be fulfilled. The lever itself must be absolutely rigid, and must be weightless; it must be of absolutely uniform structure, the fulcrum must be a mathematical point, in order that friction may be excluded, and so forth. Similarly, the weights must be thought of as mere masses without any further difference of quality, and thus only capable of affecting the lever through the one property of their weight; their attachments, again, must be of ideal tenuity, or fresh complications will be introduced. But when all these conditions have been taken into account, the principle has become so abstract as to amount to the tautology that what only operates by its mass and its distance from the fulcrum will not operate by any other property.

In any actual case, the course of events will be liable to be affected by all the conditions which had to be excluded from the abstract formulation of the principle. No actual lever will be weightless or incapable of being bent or broken; its construction will never be uniform. Actual loads, again, may influence the behaviour of a lever differently according to their bulk, their chemical composition, the nature of their attachments. At an actual fulcrum there will be some degree of friction between the lever-bar and its support, and so on. In actual fact, any or all of these circumstances may affect the behaviour of the lever bar when the loads are suspended from it. Consequently, it is quite impossible to apply the mechanical generalisation with certainty to determine the course of events in a concrete case.

What holds good in this instance holds good in all similar cases of the “laws” of nature. In so far as these laws are really exact they are all hypothetical, and deal only with the problem.[problem.] What would be the course of a physical sequence, assuming its complete ground to be contained in the conditions enumerated in the enunciation of the law? That is, they all, in so far as they are absolute, are different forms of the tautological proposition, that where there is nothing to make any difference between two cases, there will be no difference. But the moment we apply our laws to the calculation of the actual course of an individual process, we have to recognise that the condition for their rigid exactness is absent; in the individual process there are always aspects not comprised in the conditions for which the law was enunciated, and nothing but actual experience can inform us whether the presence of these aspects will perceptibly affect the result in which we are interested. As applied to the study of an individual process, the principle of Uniformity is thus a postulate, like the principle of Causality, which can only be justified by its actual success.

Again, like the principle of Causality, the principle of Uniformity may be successful to different degrees, according to the special nature of the processes for which it is assumed. As the causal postulate rested on the assumption that a selection from the antecedents of an event may for practical purposes be treated as equivalent to its complete ground, so the more general postulate of Uniformity rests on the assumption that individual purpose may be left out of account in assigning the ground of a process. It does not follow that these postulates will receive the same amount of empirical justification for all departments of the physical order. There may well be certain processes in which the individual purposive character is so prominent that, even for our practical purposes, we cannot safely calculate their course without taking their end or purpose into consideration. In that case the principle of Uniformity and that of Causality would, for this part of the physical order, lose their practical value. It is a popular belief that such a failure of these practical postulates actually takes place where we come to deal with the conscious volitions of human agents. The problem is one which must be kept for fuller consideration in our next Book, but we can at present make two general statements.

(1) Such a failure of the postulates of Causality and Uniformity in application to a particular sphere would not involve a breach of the fundamental logical principle of Ground and Consequence, since, as we have seen sufficiently already, both postulates impose special restrictions on that principle for which the nature of the principle itself affords no warrant. It would thus not be an unthinkable or logically untenable position to hold that no general laws of human action can be formulated.

(2) While this extreme denial of the possibility of laws of human action is logically possible, the actual success of those sciences which deal with human behaviour in the statistical way forbids us to accept it. The success of these sciences shows that human behaviour, considered in the gross, does exhibit certain approximate uniformities. But there seems to be no means of proving that all aspects of human behaviour would show such uniformity if considered in gross in the same fashion. It is at least conceivable that some social activities would fail to exhibit approximation to an average value, no matter how extended the area and period taken as the basis of investigation. We might conceivably have to admit that there are departments of social life for which no “laws” can be formulated. If we disregard this possibility in practice, the reason is a methodological one. It is our interest to discover such uniformities, and therefore, as failure may only mean a temporary check to the success of our investigations, we properly make it a rule of method to assume that it is no more than this. We treat all sequences as capable, by proper methods, of reduction to uniformity, for the same reason that we treat all offenders as possibly reclaimable. We desire that they should be so, and we cannot prove they are not so, and we therefore behave as if we knew they were so.