Sifting out in this way unworthy formulations, she can leave a residuum of conceptions that at least are possible. With these she can deal as hypotheses , testing them in all the manners, whether negative or positive, by which hypotheses are ever tested. She can reduce their number, as some are found more open to objection. She can perhaps become the champion of one which she picks out as being the most closely verified or verifiable. She can refine upon the definition of this hypothesis, distinguishing between what is innocent over-belief and symbolism in the expression of it, and what is to be literally taken. As a result, she can offer mediation between different believers, and help to bring about consensus of opinion. She can do this the more successfully, the better she discriminates the common and essential from the individual and local elements of the religious beliefs which she compares.
I do not see why a critical Science of Religions of this sort might not eventually command as general a public adhesion as is commanded by a physical science. Even the personally non-religious might accept its conclusions on trust, much as blind persons now accept the facts of optics—it might appear as foolish to refuse them. Yet as the science of