We must restrict to mere probability not only statements of comparatively great uncertainty, like predictions about the weather, where we would cautiously ...