Time and Free Will · Henri Bergson
FREE WILL
By going deeper into this twofold argument, we shall find, at its very root, the two fundamental illusions of the reflective consciousness. The first consists in regarding: intensity as a mathematical property of psychic states and not, as we said at the beginning of this essay, as a special quality, as a particular shade of these various states. The second consists in substituting for the concrete reality or dynamic progress, which consciousness perceives, the material symbol of this progress when it has already reached its end, that is to say, of the act already accomplished together with the series of its antecedents. Certainly, once the final act is completed, I can ascribe to all the antecedents their proper value, and picture the interplay of these various elements as a conflict or a composition of forces. But to ask whether, the antecedents being known as well as their value, one could foretell the final act, is to beg the question; it is to forget that we cannot know the value of the antecedents without knowing the final act, which is the very thing that is not yet known; it is to suppose wrongly that the symbolical diagram which we draw in our own way for representing the action _when completed_ has been drawn by the action itself _whilst progressing,_ and drawn by it in an automatic manner.
So you then alter your hypothesis; you realize that time does not require to be seen, but to be lived; and hence you conclude that, if your knowledge of the line M O was not a sufficient datum, the reason must have been that you looked at it from the outside instead of identifying yourself with the point M, which describes not only M O but also the whole curve, and thus making its movement your own. Therefore, you persuade Paul to come and coincide with Peter; and naturally, then, it is the line M O X Y which Paul traces out in space, since, by hypothesis, Peter describes this line. But in no wise do you prove thus that Paul foresaw Peter's action; you only show that Peter acted in the way he did, since Paul became Peter. It is true that you then come back, unwittingly, to your former hypothesis, because you continually confuse the line M O X Y in its tracing with the line M O X Y already traced, that is to say, time with space. After causing Paul to come down and identify himself with Peter as long as was required, you let him go up again and resume his former post of observation. No wonder if he then perceives the line M O X Y complete: he himself has just been completing it.
In order to put our finger on this vital difference, let us assume for a moment that some mischievous illustration genius, more powerful still than the mischievous genius conjured up by Descartes decreed that all the movements of the universe should go twice as fast. There would be no change in astronomical phenomena, or at any rate in the equations which enable us to foresee them, for in these equations the symbol _t_ does not stand for a duration, but for a relation between two durations, for a certain number of units of time, in short, for a certain number of _simultaneities:_ these simultaneities, these coincidences would still take place in equal number: only the intervals which separate them would have diminished, but these intervals never make their appearance in our calculations. Now these intervals are just duration _lived,_ duration which our consciousness perceives, and our consciousness would soon inform us of a shortening of the day if we had not experienced the usual amount of duration between sunrise and sunset. No doubt it would not measure this shortening, and perhaps it would not even perceive it immediately as a change of quantity; but it would realize in some way or other a decline in the usual storing up of experience, a change in the progress usually accomplished between sunrise and sunset.
Indeed it is even conceivable that this consciousness could live so slow and lazy a life as to take in the whole path of the heavenly body in a single perception, just as we do when we perceive the successive positions of a shooting star as one line of fire. Such a consciousness would find itself really in the same conditions in which the astronomer places himself ideally; it would see in the present what the astronomer perceives in the future. In truth, if the latter foresees a future phenomenon, it is only on condition of making it to a certain extent a present phenomenon, or at least of enormously reducing the interval which separates us from it. In short, the time of which we speak in astronomy is a number, and the nature of the units of this number cannot be specified in our calculations; we may therefore assume them to be as small as we please, provided that the same hypothesis is extended to the whole series of operations, and that the successive relations of position in space are thus preserved. We shall then be present in imagination at the phenomenon we wish to foretell; we shall know exactly at what point in space and after how many units of time this phenomenon takes place; if we then restore to these units their psychical nature, we shall thrust the event again into the future and say that we have foreseen it, when in reality we have seen it.
But we forget that states of consciousness are processes, and not things; that if we denote them each by a single word, it is for the convenience of language; that they are alive and therefore constantly changing; that, in consequence, it is impossible to cut off a moment from them without making them poorer by the loss of some impression, and thus altering their quality. I quite understand that the orbit of a planet might be perceived all at once or in a very short time, because its successive positions or the _results_ of its movement are the only things that matter, and not the duration of the equal intervals which separate them. But when we have to do with a feeling, it has no precise result except its having been felt; and, to estimate this result adequately, it would be necessary to have gone through all the phases of the feeling itself and to have taken up the same duration. Even if this feeling has finally issued in some definite action, which might be compared to the definite position of a planet in space, the knowledge of this act will hardly enable us to estimate the influence of the feeling on the whole of a life-story, and it is this very influence which we want to know.
All foreseeing is in reality seeing, and this seeing takes place when we can reduce as much as we please an interval of future time while preserving the relation of its parts to one another, as happens in the case of astronomical predictions. But what does reducing an interval of time mean, except emptying or impoverishing the conscious states which fill it? And does not the very possibility of seeing an astronomical period in miniature thus imply the impossibility of modifying a psychological series in the same way, since it is only by taking this psychological series as an invariable basis that we shall be able to make an astronomical period vary arbitrarily as regards the unit of duration? [Sidenote: Difference between past and future duration in this respect.]
Thus, when we ask whether a future action could have been foreseen, we unwittingly identify that time with which we have to do in the exact sciences, and which is reducible to a number, with real duration, whose so-called quantity is really a quality, and which we cannot curtail by an instant without altering the nature of the facts which fill it. No doubt the identification is made easier by the fact that in a large number of cases we are justified in dealing with real duration as with astronomical time. Thus, when we call to mind the past, i.e. a series of deeds done, we always shorten it, without however distorting the nature of the event which interests us. The reason is that we know it already; for the psychic state, when it reaches the end of the _progress_ which constitutes its very existence, becomes a _thing_ which one can picture to oneself all at once. Here we find ourselves in the same position as the astronomer, when he takes in at a glance the orbit which a planet will need several years to traverse. In fact, astronomical prediction should be compared with the recollection of the past state of consciousness, not with the anticipation of the future one. But when we have to determine a future state of consciousness, however superficial it may be, we can no longer view the antecedents in a static condition as things; we must view them in a dynamic condition as processes, since we are concerned with their influence alone. Now their duration is this very influence. Therefore it will no longer do to shorten future duration in order to picture its parts beforehand; one is bound to _live_ this duration whilst it is unfolding. As far as deep-seated psychic states are concerned, there is no perceptible difference between foreseeing, seeing, and acting.
This last form of the determinist argument differs less than might be thought from all the others which have been examined above. To say that the same inner causes will reproduce the same effects is to assume that the same cause can appear a second time on the stage of consciousness. Now, if duration is what we say, deep-seated psychic states are radically heterogeneous to each other, and it is impossible that any two of them should be quite alike, since they are two different moments of a life-story. While the external object does not bear the mark of the time that has elapsed and thus, in spite of the difference of time, the physicist can again encounter identical elementary conditions, duration is something real for the consciousness which preserves the trace of it, and we cannot here speak of identical conditions, because the same moment does not occur twice. It is no use arguing that, even if there are no two deep-seated psychic states which are altogether alike, yet analysis would resolve these different states into more general and homogeneous elements which might be compared with each other.
This would be to forget that even the simplest psychic elements possess a personality and a life of their own, however superficial they may be; they are in a constant state of becoming, and the same feeling, by the mere fact of being repeated, is a new feeling. Indeed, we have no reason for calling it by its former name save that it corresponds to the same external cause or projects itself outwardly into similar attitudes: hence it would simply be begging the question to deduce from the so-called likeness of two conscious states that the same cause produces the same effect. In short, if the causal relation still holds good in the realm of inner states, it cannot resemble in any way what we call causality in nature. For the physicist, the same cause always produces the same effect: for a psychologist who does not let himself be misled by merely apparent analogies, a deep-seated inner cause produces its effect once for all and will never reproduce it. And if it is now asserted that this effect was inseparably bound up with this particular cause, such an assertion will mean one of two things: either that, the antecedents being given, the future action might have been foreseen; or that, the action having once been performed, any other actionals seen, under the given conditions, to have been impossible. Now we saw that both these assertions were equally meaningless, and that they also involved a false conception of duration.
To assert the regular succession of two phenomena is, indeed, to recognize that, the first being given, we already catch sight of the second. But this wholly subjective connexion between two ideas is not enough for common sense. It seems to common sense that, if the idea of the second phenomenon is already implied in that of the first, the second phenomenon itself must exist objectively, in some way or other, within the first phenomenon. And common sense was bound to come to this conclusion, because to distinguish exactly between an objective connexion of phenomena and a subjective association between their ideas presupposes a fairly high degree of philosophical culture. We thus pass imperceptibly from the first meaning to the second, and we picture the causal relation as a kind of prefiguring of the future phenomenon in its present conditions. Now this prefiguring can be understood in two very different ways, and it is just here that the ambiguity begins. [Sidenote: Causality, as the prefiguring of the future phenomenon in its present conditions, in one form destroys concrete phenomena.]
In the first place, mathematics furnishes us with _one_ type of this kind of prefiguring. The very movement by which we draw the circumference of a circle on a sheet of paper generates all the mathematical properties of this figure: in this sense an unlimited number of theorems can be said to pre-exist within the definition, although they will be spread out in duration for the mathematician who deduces them. It is true that we are here in the realm of pure quantity and that, as geometrical properties can be expressed in the form of equations, it is easy to understand how the original equation, expressing the fundamental property of the figure, is transformed into an unlimited number of new ones, all virtually contained in the first. On the contrary, physical phenomena, which succeed one another and are perceived by our senses, are distinguished by quality not less than by quantity, so that there would be some difficulty in at once declaring them equivalent to one another. But, just because they are perceived through our sense-organs, we seem justified in ascribing their qualitative differences to the impression which they make on us and in assuming, behind the heterogeneity of our sensations, a homogeneous physical universe.
Thus, we shall strip matter of the concrete qualities with which our senses clothe it, colour, heat, resistance, even weight, and we shall finally find ourselves confronted with homogeneous extensity, space without body. The only step then remaining will be to describe figures in space, to make them move according to mathematically formulated laws, and to explain the apparent qualities of matter by the shape, position, and motion of these geometrical figures. Now, position is given by a system of fixed magnitudes and motion is expressed by a law, i.e. by a constant relation between variable magnitudes; but shape is a mental image, and, however tenuous, however transparent we assume it to be, it still constitutes, in so far as our imagination has, so to speak, the visual perception of it, a concrete and therefore irreducible quality of matter. It will therefore be necessary to make a clean sweep of this image itself and replace it by the abstract formula of the movement which gives rise to the figure.
Picture then algebraical relations getting entangled in one another, becoming objective by this very entanglement, and producing, by the mere effect of their complexity, concrete, visible, and tangible reality,--you will be merely drawing the consequences of the principle of causality, understood in the sense of an actual prefiguring of the future in the present. The scientists of our time do not seem, indeed, to have carried abstraction so far, except perhaps Lord Kelvin. This acute and profound physicist assumed that space is filled with a homogeneous and incompressible fluid in which vortices move, thus producing the properties of matter: these vortices are the constituent elements of bodies; the atom thus becomes a movement, and physical phenomena are reduced to regular movements taking place within an incompressible fluid. But, if you will notice that this fluid is perfectly homogeneous, that between its parts there is neither an empty interval which separates them nor any difference whatever by which they can be distinguished, you will see that all movement taking place within this fluid is really equivalent to absolute immobility, since before, during, and after the movement nothing changes and nothing has changed in the whole.
The movement which is here spoken of is thus not a movement which actually takes place, but only a movement which is pictured mentally: it is a relation between relations. It is implicitly supposed, though perhaps not actually realized, that motion has something to do with consciousness, that in space there are only simultaneities, and that the business of the physicist is to provide us with the means of calculating these relations of simultaneity for any moment of our duration. Nowhere has mechanism been carried further than in this system, since the very shape of the ultimate elements of matter is here reduced to a movement. But the Cartesian physics already anticipated this interpretation; for if matter is nothing, as Descartes claimed, but homogeneous extensity, the movements of the parts of this extensity can be conceived through the abstract law which governs them or through an algebraical equation between variable magnitudes, but cannot be represented under the concrete form of an image. And it would not be difficult to prove that the more the progress of mechanical explanations enables us to develop this conception of causality and therefore to relieve the atom of the weight of its sensible qualities, the more the concrete existence of the phenomena of nature tends to vanish into algebraical smoke. [Sidenote: It thus leads to Descartes' physics and Spinoza's metaphysics, but cannot bind future to present without neglecting duration.]
Thus understood, the relation of causality is a necessary relation in the sense that it will indefinitely approach the relation of identity, as a curve approaches its asymptote. The Principle of identity is the absolute law of our consciousness: it asserts that what is thought is thought at the moment when we think it: and what gives this principle its absolute necessity is that it does not bind the future to the present, but only the present to the present: it expresses the unshakable confidence that consciousness feels in itself, so long as, faithful to its duty, it confines itself to declaring the apparent present state of the mind. But the principle of causality, in so far as it is supposed to bind the future to the present, could never take the form of a necessary principle; for the successive moments of real time are not bound up with one another, and no effort of logic will succeed in proving that what has been will be or will continue to be, that the same antecedents will always give rise to identical consequents.
Descartes understood this so well that he attributed the regularity of the physical world and the continuation of the same effects to the constantly renewed grace of Providence; he built up, as it were, an instantaneous physics, intended for a universe the whole duration of which might as well be confined to the present moment. And Spinoza maintained that the indefinite series of phenomena, which takes for us the form of a succession in time, was equivalent, in the absolute, to the divine unity: he thus assumed, on the one hand, that the relation of apparent causality between phenomena melted away into a relation of identity in the absolute, and, on the other, that the indefinite duration of things was all contained in a single moment, which is eternity. In short, whether we study Cartesian physics, Spinozistic metaphysics, or the scientific theories of our own time, we shall find everywhere the same anxiety to establish a relation of logical necessity between cause and effect, and we shall see that this anxiety shows itself in a tendency to transform relations of succession into relations of inherence, to do away with active duration, and to substitute for apparent causality a fundamental identity. [Sidenote: The necessary determination of phenomena implies non-duration; but we _endure_ and are therefore free.]
Now, if the development of the notion of causality, understood in the sense of necessary connexion, leads to the Spinozistic or Cartesian conception of nature, inversely, all relation of necessary determination established between successive phenomena may be supposed to arise from our perceiving, in a confused form, some mathematical mechanism behind their heterogeneity. We do not claim that common sense has any intuition of the kinetic theories of matter, still less perhaps of a Spinozistic mechanism; but it will be seen that the more the effect seems necessarily bound up with the cause, the more we tend to put it in the cause itself, as a mathematical consequence in its principle, and thus to cancel the effect of duration. That under the influence of the same external conditions I do not behave to-day as I behaved yesterday is not at all surprising, because I _change,_ because I _endure._ But things considered apart from our perception do not seem to endure; and the more thoroughly we examine this idea, the more absurd it seems to us to suppose that the same cause should not produce to-day the effect which it produced yesterday.
We certainly feel, it is true, that although things do not endure as we do ourselves, nevertheless there must be some reason why phenomena are seen to _succeed_ one another instead of being set out all at once. And this is why the notion of causality, although it gets indefinitely near that of identity, will never seem to us to coincide with it, unless we conceive clearly the idea of a mathematical mechanism or unless some subtle metaphysics removes our very legitimate scruples on the point. It is no less obvious that our belief in the necessary determination of phenomena by one another becomes stronger in proportion as we are more inclined to regard duration as a subjective form of our consciousness. In other words, the more we tend to set up the causal relation as a relation of necessary determination, the more we assert thereby that things do not _endure_ like ourselves. This amounts to saying that the more we strengthen the principle of causality, the more we emphasize the difference between a physical series and a psychical one. Whence, finally, it would result (however paradoxical the opinion may seem) that the assumption of a relation of mathematical inherence between external phenomena ought to bring with it, as a natural or at least as a plausible consequence, the belief in human free will. But this last consequence will not concern us for the moment: we are merely trying here to trace out the first meaning of the word causality, and we think we have shown that the prefiguring of the future in the present is easily conceived under a mathematical form, thanks to a certain conception of duration which, without seeming to be so, is fairly familiar to common sense.
If, then, we decide to picture the causal relation in this second form, we can assert _a priori_ that there will no longer be a relation of necessary determination between the cause and the effect, for the effect will no longer be given in the cause. It will be there only in the state of pure possibility and as a vague idea which perhaps will not be followed by the corresponding action. But we shall not be surprised that this approximation is enough for common sense if we think of the readiness with which children and primitive people accept the idea of a whimsical Nature, in which caprice plays a part no less important than necessity. Nay, this way of conceiving causality will be more easily understood by the general run of people, since it does not demand any effort of abstraction and only implies a certain analogy between the outer and the inner world, between the succession of objective phenomena and that of our subjective states. [Sidenote: This second conception of causality leads to Leibniz as the first led to Spinoza.]
Ideas discussed in this chapter
Text from Project Gutenberg, public domain in the United States.