Page:Public Opinion (Lippmann).djvu/97

Rh thirteen more than 50%. Moreover in twenty-four accounts 10% of the details were pure inventions and this proportion was exceeded in ten accounts and diminished in six. Briefly a quarter of the accounts were false.

"It goes without saying that the whole scene had been arranged and even photographed in advance. The ten false reports may then be relegated to the category of tales and legends; twenty-four accounts are half legendary, and six have a value approximating to exact evidence."

Thus out of forty trained observers writing a responsible account of a scene that had just happened before their eyes, more than a majority saw a scene that had not taken place. What then did they see? One would suppose it was easier to tell what had occurred, than to invent something which had not occurred. They saw their stereotype of such a brawl. All of them had in the course of their lives acquired a series of images of brawls, and these images flickered before their eyes. In one man these images displaced less than 20% of the actual scene, in thirteen men more than half. In thirty-four out of the forty observers the stereotypes preempted at least one-tenth of the scene.

A distinguished art critic has said that "what with the almost numberless shapes assumed by an object What with our insensitiveness and inattention, things scarcely would have for us features and outlines so determined and clear that we could recall them at will, but for the stereotyped shapes art has lent them." The truth is even broader than that,