—i.e. reason—in the lower, whereas the reverse is the case; for in the highest animals—e.g. man—reason predominates, and instinct is at a minimum, while in lower animals—e.g. fish—instinct predominates, and reason is at a minimum. Very plainly, therefore, in higher animals there has been retrogression, not evolution, as regards instinct. Again, instinct and reason do not merge at any point, as they would were one derived from the other. The fact that one is inborn and the other acquired separates them sharply, and if, in the case of any given action, which is the outcome both of instinct and reason, as, for instance, a man's choice of a wife, we are unable to say how much of it is instinctive, and how much rational, our difficulty is due to our ignorance, not to there being no dividing line, but a border-space.
According to Mr. Spencer, "rational action arises out of instinctive action when the latter grows too complex to be perfectly automatic " (Principles of Psychology, vol. i. p. 458). His hypothesis, like Mr. Lewes', necessarily involves the supposition that there is a border-space where the instinctive and the rational merge. We have seen that he defines instinct as compound reflex action. In his view, therefore, rational action is reflex action still more highly compounded. Concerning instinct he says—