Here there would seem, at first sight, to be no room for error, since this movement of retrospective imagination may be said to involve a direct measurement of the distance, just as a sweep of the eye over the ground between a spectator and an object affords a direct measurement of the intervening space.
Modern science, however, tells us that this mode of measurement is by no means the simple and accurate process which it at first seems to be.

In point of fact, there is something like a constant error in all such retrospective measurement.

Vierordt has proved experimentally, by making a person try to reproduce the varying time-intervals between the strokes of the pendulum of a metronome, that when the interval is a very small one, we uniformly tend to exaggerate it in retrospection; when a large one, to regard it, on the contrary, as less than it actually was.[117] A mere act of reflection will convince any one that when he tries to conceive a very small interval, say a quarter of a second, he is likely to make it too great.

On the other hand, when we try to conceive a year, we do not fully grasp the whole extent of the duration.

This is proved by the fact that merely by spending more time over the attempt, and so recalling a larger number of the details of the period, we very considerably enlarge our first estimate of the duration.