Criminal Psychology: A Manual for Judges, Practitioners,...

There are two means of correction. Either have the witness determine the time in terms of some familiar form, i.e., a paternoster, etc., or give him the watch and let him observe the second hand. In the latter case he will assert that his ten, or his five, or his twenty minutes were, at most, no more than a half or a whole minute.

The problem of time is still more difficult when the examination has to be made with regard to the estimation of still longer periods—weeks, months, or years. There is no means of making any test. The only thing that experience definitely shows is, that the certainty of such estimates depends on their being fixed by distinct events. If anybody says that event A occurred four or five days before event B, we may believe him if, e.g., he adds, “For when A occurred we began to cut corn, and when B occurred we harvested it. And between these two events there were four or five days.” If he can not adduce similar judgments, we must never depend upon him, for things may have occurred which have so influenced his conception of time that he judges altogether falsely.

It often happens in such cases that defective estimates, made in the course of lengthy explanations, suddenly become points of reference, and then, if wrong, are the cause of mistakes. Suppose that a witness once said that an event occurred four years ago. Much later an estimation of the time is undertaken which shows that the hasty statement sets the event in 1893. And then all the most important conclusions are merely argued from that. It is best, as is customary in such cases, to test the uncertainty and incorrectness of these estimates of time on oneself. It may be assumed that the witness, in the case in question, is likely to have made a better estimate, but it may equally be assumed that he has not done so. In short, the conception of periods of time can not be dealt with too cautiously.