Page:Bergson - Laughter (1911).djvu/183

 To penetrate too far into the personality, to couple the outer effect with causes that are too deep-seated, would mean to endanger and in the end to sacrifice all that was laughable in the effect. In order that we may be tempted to laugh at it, we must localise its cause in some intermediate region of the soul. Consequently, the effect must appear to us as an average effect, as expressing an average of mankind. And, like all averages, this one is obtained by bringing together scattered data, by comparing analogous cases and extracting their essence, in short by a process of abstraction and generalisation similar to that which the physicist brings to bear upon facts with the object of grouping them under laws. In a word, method and object are here of the same nature as in the inductive sciences, in that observation is always external and the result always general.

And so we come back, by a roundabout way, to the double conclusion we reached in the course of our investigations. On the one hand, a person is never ridiculous except through some mental attribute resembling absentmindedness, through something that lives upon