all, merely recording certain physiological processes, or else they indicate the existence of mentation which does not belong to any recognized human being. The first would seem to deny the doctrine of parallelism, according to which physiological processes of the degree of complexity requisite to the production of writing necessarily generate mental states, and this would lead us toward the old theory of the soul, or something like it. The second would compel the assumption either of personalities distinct from that of the subject, which is the theory of possession, or of segregated mental states. The latter is the theory which I am developing in these pages, and although I am far from satisfied with it, it is more in line with our present scientific conceptions than others, and accounts for some of the facts fairly well.
But this dilemma presents itself only when it can be shown that the subject's upper consciousness has nothing to do with the
production of the writing. I am convinced that experimenters do not pay sufficient attention to this point, and consequently much of the recorded material is to my mind of little significance. As my space is limited, I wish to lay especial stress upon this aspect of the problem.
A few years ago I had the opportunity of studying at leisure a remarkably good case of automatism. The subject, whom I shall call B, was a man of intelligence and education, with whom I had long been on terms of intimacy, and of whose good faith I can therefore speak with some confidence. The writing was at first a mere scrawl, accompanied by quite violent twisting of the arm; little by little it became intelligible, wrote "Yes"
