[Footnote 1:  See, for example, John xix. 23-24.]

I beg those who think that I have placed an exaggerated confidence in narratives in great part legendary, to take note of the observation I have just made.  To what would the life of Alexander be reduced if it were confined to that which is materially certain?  Even partly erroneous traditions contain a portion of truth which history cannot neglect.  No one has blamed M. Sprenger for having, in writing the life of Mahomet, made much of the hadith or oral traditions concerning the prophet, and for often having attributed to his hero words which are only known through this source.  Yet the traditions respecting Mahomet are not superior in historical value to the discourses and narratives which compose the Gospels.  They were written between the year 50 and the year 140 of the Hegira.  When the history of the Jewish schools in the ages which immediately preceded and followed the birth of Christianity shall be written, no one will make any scruple of attributing to Hillel, Shammai, Gamaliel the maxims ascribed to them by the Mishnah and the Gemara, although these great compilations were written many hundreds of years after the time of the doctors in question.

As to those who believe, on the contrary, that history should consist of a simple reproduction of the documents which have come down to us, I beg to observe that such a course is not allowable.  The four principal documents are in flagrant contradiction one with another.  Josephus rectifies them sometimes.  It is necessary to make a selection.  To assert that an event cannot take place in two ways at once, or in an impossible manner, is not to impose an a priori philosophy upon history.  The historian ought not to conclude that a fact is false because he possesses several versions of it, or because credulity has mixed with them much that is fabulous.  He ought in such a case to be very cautious—­to examine the texts, and to proceed carefully by induction.  There is one class of narratives especially, to which this principle must necessarily be applied.  Such are narratives of supernatural events.  To seek to explain these, or to reduce them to legends, is not to mutilate facts in the name of theory; it is to make the observation of facts our groundwork.  None of the miracles with which the old histories are filled took place under scientific conditions.  Observation, which has never once been falsified, teaches us that miracles never happen but in times and countries in which they are believed, and before persons disposed to believe them.  No miracle ever occurred in the presence of men capable of testing its miraculous character.  Neither common people nor men of the world are able to do this.  It requires great precautions and long habits of scientific research.  In our days have we not seen almost all respectable people dupes of the grossest frauds or of puerile illusions?  Marvellous facts, attested by the whole population of small towns, have, thanks to a severer scrutiny, been exploded.[1] If it is proved that no contemporary miracle will bear inquiry, is it not probable that the miracles of the past, which have all been performed in popular gatherings, would equally present their share of illusion, if it were possible to criticise them in detail?