not break down. I believe that this inkstand is hard, because this belief works.'"
I conclude therefore that all knowledge of the kind we are now describing is based on belief (viz. the belief that what has been will be) tested by experience. I think it must also be admitted that Imagination contributed to the result: for the child not only remembers his two past consecutive sensations but gradually images in his mind a kind of bond between them, which memory pure and simple could not have contributed. Memory reproduces "Inkstand and then hardness;" Imagination paints, or begins to paint, a new idea, "Inkstand and therefore hardness." Again, Memory reproduces vaguely numerous instances, "The inkstand was hard ten, eleven, twenty, many times;" then comes Imagination and at a leap sets before the mind an entirely new notion, and invents for it the word "always."
Concerning other and more complex kinds of knowledge what need is there to say a word? For if such simple propositions as "a stone is hard," are shown to depend upon Imagination for suggesting, and Faith for retaining, a conviction of the uniformity of Nature, much more must these influences be presupposed if the child is to attain knowledge about matters avowedly future, e.g. "the sun will rise to-morrow." In reality all knowledge of any practical value has to do with a future, immediate or remote; and therefore I do not think I shall be exaggerating in saying that for all knowledge about things outside us we depend largely upon Imagination and Faith.
But I pass now to consider a child's knowledge about himself. Take for example such a proposition as this, "I like sugar." Is Faith or Imagination required to enable a child to arrive at the knowledge of this proposition about himself? I think so. The very use of the word "I," if used intelligently, appears to need some imagin-
