In arguing his thesis, Foucault makes much of the modifications which can be proved to take place if any one is asked to repeat a dream at intervals of months. Under the influence of time and repetition a dream becomes more coherent and more conformed to reality. In illustration Foucault presents two versions of an insignificant dream in which a lady imagines that she is out with her husband for a drive, and in the course of it experiences a natural need which she seeks an opportunity to satisfy; the details of the first version were highly improbable; some months later they had become much more like what might have occurred in real life. Such a process, Foucault thinks, is taking place from the first in the making of dreams as we know them awake. There are, undoubtedly, facts which may seem to support Foucault's argument that the logic of the dream, as we know it, is not in the original dream, but is introduced afterwards. Thus I once dreamed in the morning that I asked my wife if she had been into a certain room, and that she replied, 'Can't get in.' I immediately awoke and realised that my wife had actually spoken these words, not to me, but to an approaching servant, in anticipation of a message about entering a neighbouring room of which the door was locked. It is thus evident that although it seemed to me in my dream that the question came first and the answer followed in the ordinary course, in reality the answer came first. The question was a theory, supplied automatically by sleeping intelligence and prefixed to the answer, in which order they both appeared to sleeping consciousness, that is to say, in the only way in which sleeping consciousness can ever be known, as translated into waking consciousness.[7] It must be borne in mind that such a dream as I have recorded--in which an actual sensory experience is introduced, untransformed, as a foreign body into sleeping consciousness--is not a typical dream. Dreams are, however, without doubt of various kinds, and we may well admit that there is a class of dreams formed in this way. That supposition will, indeed, be helpful in explaining several dreams I shall have to record. The process is much the same as when a nervous person receives a telegram, and at once assumes that some dreaded accident has occurred, and that the telegram is the announcement of it. The craving for reasons is instinctive, and the dreamer's sense of logic even dominates his sense of time.