The doctrine clarifies even astonishing statistical miracles. We know, e.g., that every year there come together in a certain region a large number of suicides, fractures of arms and legs, assaults, unaddressed letters, etc. When, now, we discover that the number of suicides in a certain semester is significantly less than the number in the same semester of another year, we will postulate that in the next half-year a comparatively larger number of suicides will take place so that the number for the whole year will become approximately equal. Suppose we say: “There were in the months of January, February, March, April, May and June an average of x cases. Because we have observed the average to happen six times, we conclude that it will not happen in the other months but that instead, x+y cases will occur in those months, since otherwise the average annual count will not be attained.” This would be a mistaken abstraction of the principle of equal distribution from the general Humian law, for the Humian law applied to this case indicates: “For a long series of years we have observed that in this region there occur annually so and so many suicides; we conclude therefore that in this year also there will occur a similar number of suicides.”
The principle of equal distribution presents itself therefore as a subordinate rule which must not be separated from the principal law. It is, indeed, valid for the simplest events. When I resolve to walk in x street, which I know well, and when I recall whether to-day is Sunday or a week day, what time it is and what the weather is like, I know quite accurately how the street will look with regard to the people that may be met there, although a large number of these people have chosen the time accidentally and might as well have passed through another street. If, for once, there were more people in the street, I should immediately ask myself what unusual event had taken place.
One of my cousins who had a good deal of free time to dispose of, spent it for several months, with the assistance of his comrade, in counting the number of horses that passed daily, in the course of two hours, by a café they frequented. The conscientious and controlled count indicated that every day there came one bay horse to every four. If then, on any given day, an incommensurably large number of brown, black, and tawny horses came in the course of the first hour, the counters were forced to infer that in the next 60 minutes horses of a different color must come and that a greater number of bays must appear in order to restore the disturbed equilibrium. Such an inference is not contradictory to the Humian proposition. At the end of a series of examinations the counters were compelled to say, “Through so many days we have counted one bay to every four horses; we must therefore suppose that a similar relationship will be maintained the next day.”
