Thus, we see that long before man had arrived at an idea of self there had been growing up an emotional predisposition to think well of self. And in this way we may understand how it is that this sentiment of self-esteem shows itself immediately and instinctively in the child's mind as soon as its unfolding consciousness is strong enough to grasp the first rough idea of personal existence. Far down, so to speak, below the surface of distinct consciousness, in the intricate formation of ganglion-cell and nerve-fibre, the connections between the idea of self and this emotion of esteem have been slowly woven through long ages of animal development. Here, then, we seem to have the key to the apparently paradoxical fact that a man, with all his superior means of studying his own feelings, commonly esteems himself, in certain respects at least, less accurately than a good external observer would be capable of doing. In forming an opinion of ourselves we are exposed to the full force of a powerful impulse of feeling. This impulse, acting as a bias, enters more or less distinctly into our single acts of introspection, into our attempts to recall our past doings, into our insights into the meaning of others' words and actions as related to ourselves (forming the natural disposition to enjoy flattery), and finally into our wild dreams as to our future achievements. It is thus the principal root of that gigantic illusion of self-conceit, which has long been recognized by practical sense as one of the greatest obstacles to social action; and by art as one of the most ludicrous manifestations of human weakness. If there are all these openings for error in the beliefs we go on entertaining respecting individual things, including ourselves, there must be a yet larger number of such openings in those still more compound beliefs which we habitually hold respecting collections or classes of things. A single illusion of perception or of memory may suffice to give rise to a wholly illusory belief in a class of objects, for example, ghosts. The superstitious beliefs of mankind abundantly illustrate this complexity of the sources of error. And in the case of our every-day beliefs respecting real classes of objects, these sources contribute a considerable quota of error. We may again see this by examining our ordinary beliefs respecting our fellow-men.