points of difference found at one moment would cease to exist in the next. In the first case any expectation would be fulfilled by one thing as well as by any other; in this latter case no expectation would be fulfilled by anything. Thus while the possibility of distinguishing between things implies that they differ from each other, it also implies that each of the things distinguished has some characteristic uniformity of appearance or way of acting which marks or constitutes its own individuality or nature.

An oxygen atom acts in one way in the presence of hydrogen and in another in the presence of nitrogen, but it never ceases to act like oxygen and begins to act like iron or chlorine. A formula which told the various ways in which it does act under all possible circumstances would be a definition of its nature.

We not only assume that every thing has a nature of its own, but we assume also that there are absolute similarities between them, so that they can be divided into various 'kinds', all the things of a given kind having so much the same nature' that one might be substituted for another without producing any perceptible change in the result.

Along with the uniformities characteristic of certain individuals and classes of things we come to have also some idea of the uniformities characteristic of things in general, particularly of material things. A material thing, as contrasted with a mere phantasm, is expected to be tangible, to resist pressure more or less, to have a continuous existence in time and space in the sense of not passing from one point to another without passing through all the intermediate points; and perhaps also to obey the laws of inertia and gravitation. Finally, since things are only given as aspects of the sum total of experience called the world as a whole, by reference to which aspects the essential uniformity of the whole can be conceived, it is a part of the very nature of a thing to be bound up with others in a whole world, to act upon them and to be acted upon by them, and to observe the same

general laws. To speak of a thing out of all relations to everything else is to speak of an aspect apart from that whose aspect it is, and is therefore absurd.

By the Circumstances of a thing as contrasted with its nature we mean, not the general rule of its action, but the particular conditions to which the rule must be adapted at any given instant. These conditions include the present or vanishing state or activity of the thing itself. An iron, for example, that has just been heated will not act precisely like one that has not. They include also the position and nature, and present or vanishing states or acts, of other things. Oxygen does not act in the presence of hydrogen precisely as it acts in a vacuum, and it does not act in the presence of hot hydrogen precisely as it does in the presence of cold.

When I say that the conditions which help to determine an act include the position and state of other things besides the agent, I mean, not merely that sometimes an act is determined in part by something outside of the agent, but that it is always. The moth in its chrysalis seems to develop wholly from within, shut off from all the rest of the world; but take away the warmth of the sun outside and how long would the development last? Even within the chrysalis we have not one simple thing moving by a wholly inward law from one state to another, as we may be inclined to assume at first, but rather a whole system of cells acting and reacting upon each other; and each one of these cells again is composed of atoms, all acting with reference to what lies beyond them. In short, all explanations in natural science come down finally to atoms; and no explanation assumes that an atom ever acts wholly from within, regardless of the rest of the world. Therefore every ultimate explanation is made on the assumption that every act or state of a thing is determined partly from without. In other words, causation always involves interaction.*

* This is quite as true in psychology as it is in physical science. When we 'explain' a person's thought we either regard it as due to the

Circumstances, like things, can be divided into various 'kinds'; but in their case the division is always rather loose. Two things can be absolutely alike in their nature, for there is no reason why they cannot both act according to precisely the same general law. But two circumstances can never be precisely alike. If one exists here at this time the other must exist somewhere else or at some other time, and consequently they must always differ in time or place if in nothing else.* We often speak of two circumstances being exactly alike, but when we do so we only mean that they are alike in every respect that is worth considering for the purpose in hand. Circumstances of essentially the same kind, like things of the same kind, can be substituted for each other without any essential change in the resulting phenomena. If they could not, the division of experience into the two aspects of things and circumstances would not help us in our conception of its uniformity. Wind blowing upon the green leaves of an aspen-tree here to-day makes them move here and to-day, if every other circumstance is the same, just as wind blowing upon them there and yesterday made them. move there and yesterday.

play of different feelings and ideas or else we turn to physiology and regard it as due to various currents' in the cells and fibres of the brain. In either case the explanation involves something beyond what is explained. Of course this causal analysis does not prevent us from believing in the real unity of the mind, and a similar analysis in natural science does not prevent us from believing in the unity of the world. The fact that explanations take account of aspects of reality does not turn these aspects into something independent and self-existent.

*I do not refer to their absolute position in time and space, --to mere position apart altogether from the relation of events to other events that precede or coexist with them and of things to other things that surround them. Such absolute position does not enter into our explanations at all, and is therefore not a 'circumstance' in the sense in which the word is used here. When we say that the result of an act was different in two different cases because the act occurred at two different times it is not the bare time as such that we have in mind, but the preceding or concurrent events; and the same is true, mutatis mutandis, of differences in place.

Having divided our world into these two aspects of thing and circumstance we are not satisfied to say merely that under precisely similar circumstances precisely Law. similar things will produce precisely similar

results. We go further, and seek for laws, which take account of the differences between things and circumstances as well as of their resemblances, and which tell not merely what takes place when the same things or precisely similar things are placed in precisely similar circumstances, but tell also how much variation in the circumstances or the things produces a given amount of variation in the result. We do not merely say that bodies attract each other; but we say that they attract each other in direct proportion to their mass and in inverse proportion to the square of their distance. This would not be possible unless we took account of the amount of difference between two situations and measured off one against another. It is this recognition of differences in amounts of difference, and their precise measurement, that enables us to introduce the conception of proportion into our formulæ and to deal with them by mathematics.

The notion of law would be impossible and our search for uniformity in the world would be doomed to failure if our ability to see different aspects of reality were limited to the general distinction between things and circumstances. But the very ability to enumerate a thing's different attributes involves the power on our part of attending to one of them at a time, and in the same way the very possibility of stating some general law such as that of gravitation involves the power of attending to some one of a thing's relations and accounting for it without saying anything about the rest. In accounting for the fall of an apple by this law, we consider its mass and the mass of the earth, its motion towards the earth, and the original distance between them. We do not say anything about its taste or color or even perhaps about its size and shape and its motion around the sun. Of course these neglected relations must be accounted for too,

and the state of affairs that produces them must be compatible with the state of affairs that produces the apple's fall. The total state of the world at every instant is the cause of its total state at the next; but we feel that we have a right in our explanations to break up this complex whole of each instant into as many aspects or different relations as we please and account for some of them at a time.

Thus the uniformity discovered by each one of our explanations is a uniformity in some one definite respect, and when we speak of the 'Cause' of an event we are almost always trying to pick out the essential elements of its nature and the relations to its own past and to other things that account for some few of its salient features. Nothing short of the whole universe would account for the event as an absolutely complete whole.

Precision in

uniformity.

The ability of which we have just spoken to attend to some one aspect of a situation involves certain dangers. For we may forget that the situation has other aspects which might also be accounted for, and therefore rest with the feeling that our explanation of the situation as a whole is complete when it is really very incomplete. Worse still, the explanation which we give of the aspect of the situation that we happened to notice may be quite inconsistent with any reasonable explanation of the other aspects that we did not happen to notice, and therefore wrong, without our detecting the fact, as we should have done if we had realized how much there was to explain. We may say that the light on the wall comes through a certain window, and be perfectly satisfied with the explanation so long as we fail to notice that the glass in the window is blue and the light on the wall is not.

Nay, even if we notice such a discrepancy, we may deliberately disregard it, with the feeling that in some way or other it can be detached from the rest of the phenomenon, and therefore makes no difference, and that in any case it must not be allowed to interfere with the conclusion already