150 THE AMERICAN JOURNAL OF SOCIOLOGY
which he finds, unless they conflict with other accounts of the same events or appear to be inherently improbable.
"To take an illustration of Nietzsche's, the vague feeling, as we lie in bed, that the soles of our feet are free from the ,usual pressure to which we are accustomed in our waking hours demands an explanation. Our dream explanation is that we must be flying. Not satisfied to leave its work half done, dream logic fabricates a room or landscape in which we make our aerial experiments. Moreover just as we are going to sleep or awaking we can often actually observe how a flash of light, such as sometimes appears on the retina of our closed eyes, will be involuntarily interpreted as a vision of some human figure or other object, clear as a stereopticon slide. Now anyone can demonstrate to himself that neither dream logic nor the 'mind's- eye faculty,' as it has been called, desert us when we are awake. Indeed they may well be, as Nietzsche suspects, a portion of the inheritance bequeathed to us, along with some other inconveni- ences, by our brutish forebears. At any rate they are forms of aberration against which the historian, with his literary tradi- tions, needs specially to be on his guard. There are rumors that even the student of natural science sometimes keeps his mind's eye too wide open, but he is by no means so likely as the historian to be misled by dream logic. This is not to be ascribed neces- sarily to the superior self-restraint of the scientist but rather to the greater simplicity of his task and the palpableness of much of his knowledge. The historian can almost never have any direct personal experience of the phenomena with which he deals. He only knows the facts of the past by the traces they have left. Now these traces are usually only the reports of someone who commonly did not himself have any direct experience of the facts and who did not even take the trouble to tell us where he got his alleged information. This is true of almost all the ancient and mediaeval historians and annalists. So it comes about that 'the immense majority of the sources of information which furnish the historian with starting points for his reasoning are nothing else than traces of psychological operations,' rather than direct traces of the facts.
