To the statement that the mind cannot abstract _out of_ human experiences what was not already in them, Mr. Spencer could make, I think, but one answer, to wit: that while the operations of the mind are generally reliable, and while there has been an element in human experience which seemed to warrant conclusions derived from them, nevertheless, mankind has egregiously erred in thinking that it had the power to build up a valid content to Religion, since the very nature of Religion is such, that the mental operations which are reliable in the realm of Science cannot be so in the realm of Religion. To answer this, we must consider the argument for conceivability as the touchstone which is to separate the "Knowable" from the "Unknowable." Corresponding to small objects, a piece of rock for example, where the sides, top, and bottom can be considered as practically all present in consciousness at once, and large ones, like the earth, where they cannot, our author divides conceptions into complete and symbolic. Great magnitudes and classes of objects also produce symbolic conceptions which, while indispensable to reasoning, often lead us into error. "We habitually mistake our symbolic conceptions for real ones." The former "are legitimate, provided that by some cumulative or indirect process of thought, or by the fulfilment of predictions based upon them, we can assure ourselves that they stand for actualities," otherwise "they are altogether vicious and illusive" and "illegitimate" and here belong religious ideas.