is incomplete, if the reference of a passage is not certain, he treats it as if it did not exist.  He forgets the old story of the faggots, which, weak singly, become strong when combined.  His scales will not admit of any evidence short of the highest.  Fractional quantities find no place in his reckoning.  If there is any flaw, if there is any possible loophole for escape, he does not make the due deduction and accept the evidence with that deduction, but he ignores it entirely, and goes on to the next item just as if he were leaving nothing behind him.

This is really part and parcel of what was pointed out at the outset as the fundamental mistake of his method.  It is much too forensic.  It takes as its model, not the proper canons of historical enquiry, but the procedure of English law.  Yet the inappropriateness of such a method is seen as soon as we consider its object and origin.  The rules of evidence current in our law courts were constructed specially with a view to the protection of the accused, and upon the assumption that it is better nine guilty persons should escape, than that one innocent person should be condemned.  Clearly such rules will be inapplicable to the historical question which of two hypotheses is most likely to be true.  The author forgets that the negative hypothesis is just as much a hypothesis as the positive, and needs to be defended in precisely the same manner.  Either the Gospels were used, or they were not used.  In order to prove the second side of this alternative, it is necessary to show not merely that it is possible that they were not used, but that the theory is the more probable of the two, and accounts better for the facts.  But the author of ‘Supernatural Religion’ hardly professes or attempts to do this.  If he comes across a quotation apparently taken from our Gospels he is at once ready with his reply, ’But it may be taken from a lost Gospel.’  Granted; it may.  But the extant Gospel is there, and the quotation referable to it; the lost Gospel is an unknown entity which may contain anything or nothing.  If we admit that the possibility of quotation from a lost Gospel impairs the certainty of the reference to an extant Gospel, it is still quite another thing to argue that it is the more probable explanation and an explanation that the critic ought to accept.  In very few cases, I believe, has the author so much as attempted to do this.

We might then take a stand here, and on the strength of what can be satisfactorily proved, as well as of what can be probably inferred, claim to have sufficiently established the use and antiquity of the Gospels.  This is, I think, quite a necessary conclusion from the data hitherto collected.