William James, The Will to Believe and Other Essays in Popular Philosophy… (1897)
“ The truest scientific hypothesis is that which, as we say, ‘works’ best; and it can be no otherwise with religious hypotheses. Religious history proves that one hypothesis after another has worked ill, has crumbled at contact with a widening knowledge of the world, and has lapsed from the minds of men. Some articles of faith, however, have maintained themselves through every vicissitude, and possess even more vitality to-day than ever before: it is for the ‘science of religions’ to tell us just which hypotheses these are. ”
