And we must remember that it is not merely that the same world is differently mirrored in different minds, but that they are two different worlds. If there is any truth in what I have urged in the last chapter, we _construct_ the world that we see. The sensations are, as we have seen, mental facts, in no sense resembling their causes, but representing them in mental symbolism. Percepts are the elaborated products of this mental symbolism. The question, then, is not--How does the world mirror itself in the mind of the dog? but rather--How far does the symbolic world of the dog resemble the symbolic world of man? How far is his symbolism the same as ours? Only by fully grasping the fact that the external world of objects does not exist independently of us (though something exists which we thus symbolize), shall we realize the greatness of the difficulty which stands in the path of the student of animal psychology. So long as we are content to accept John Bunyan's crude analogy of the gateways of sense, the difficulty is comparatively small. There is the outside world self-existent and independent; a knowledge of it comes into the mind through the five gateways of sense--a picture of it through the eye-gate, and so on. The dog has also five similar gateways. The world for him is, therefore, much the same as for us. But this is not a true analogy. The world we see around us is a joint product of an external existence, the independent nature of which we can never know, and the human mind. It is something we construct in mental symbolism. How far does the dog construct a similar world? The answer to this question must, as it seems to me, be largely speculative. And what help have we towards answering it? That afforded by the theory of organic evolution. If we accept that theory, and accept also the view that mental or psychical products are the inseparable concomitants of certain organic or physiological processes, then we have a basis from which to start. That basis I adopt.