We have said before, and we cannot repeat too often, that the science of matter proceeds like ordinary knowledge. It perfects this knowledge, increases its precision and its scope, but it works in the same direction and puts the same mechanism into play. If, therefore, ordinary knowledge, by reason of the cinematographical mechanism to which it is subjected, forbears to follow becoming in so far as becoming is moving, the science of matter renounces it equally. No doubt, it distinguishes as great a number of moments as we wish in the interval of time it considers. However small the intervals may be at which it stops, it authorizes us to divide them again if necessary. In contrast with ancient science, which stopped at certain so-called essential moments, it is occupied indifferently with any moment whatever. But it always considers moments, always virtual stopping-places, always, in short, immobilities. Which amounts to saying that real time, regarded as a flux, or, in other words, as the very mobility of being, escapes the hold of scientific knowledge. We have already tried to establish this point in a former work. We alluded to it again in the first chapter of this book. But it is necessary to revert to it once more, in order to clear up misunderstandings. When positive science speaks of time, what it refers to is the movement of a certain mobile T on its trajectory. This movement has been chosen by it as representative of time, and it is, by definition, uniform. Let us call T_{1}, T_{2}, T_{3}, ... etc., points which divide the trajectory of the mobile into equal parts from its origin T_0. We shall say that 1, 2, 3, ... units of time have flowed past, when the mobile is at the points T_{1}, T_{2}, T_{3}, ... of the line it traverses. Accordingly, to consider the state of the universe at the end of a certain time _t_, is to examine where it will be when T is at the point T_t of its course. But of the _flux_ itself of time, still less of its effect on consciousness, there is here no question; for there enter into the calculation only the points T_{1}, T_{2}, T_{3}, ... taken on the flux, never the flux itself. We may narrow the time considered as much as we will, that is, break up at will the interval between two consecutive divisions T_{n} and T_{n-|-1}; but it is always with points, and with points only, that we are dealing. What we retain of the movement of the mobile T are positions taken on its trajectory. What we retain of all the other points of the universe are their positions on their respective trajectories. To each _virtual stop_ of the moving body T at the points of division T_{1}, T_{2}, T_{3}, ... we make correspond a _virtual stop_ of all the other mobiles at the points where they are passing. And when we say that a movement or any other change has occupied a time _t_, we mean by it that we have noted a number _t_ of correspondences of this kind. We have therefore counted