the period 1836–47 were "suppressed" and that there was a conspiracy to hide those letters, which when examined were seen entirely to disprove the "legend"; that the real reasons for the fabrication of the "legend" were an attempt by Atkinson, Eells and Spalding to get possession of the Mission land at the Dalles, valuable for a townsite, and later, on the part of Eells especially, to create a basis for an appeal for contributions to Whitman College. Such is substantially the line of argument.
Let us now consider the most important part of this whole matter, the application of those two fundamental historical postulates to the evidence, written and memory, in the Whitman case.
In connection with these two historical canons we must consider a third equally vital. This is, that the testimony of the witnesses to an event takes precedence over all other testimony, other things being equal. Now we come to the vitals of the whole subject—the interrelations of these three canons and the qualifications and limitations of each. We not only admit, but we insist upon the general validity of each. But truth can be arrived at only by remembering that each has its necessary limitations and exceptions.
Let us first consider then the proposition that memory testimony cannot be accepted unless supported by contemporary writing. As a general proposition this is entirely valid. Common observation shows of course that memory and imagination become interlocked, that with the passage of time clouds obscure the clearness of vision, and that statements made after events must be subjected to the test of comparison with any existing records of those events. But now note the vitally important matter of qualifications to this general rule of historical evidence. First, it makes all the difference in the world whether the memory testimony be directly contradicted by the written record, or whether the written record merely fails to mention certain things contained in the testimony of memory. If the written record declares positively that a given thing did not take place, which given thing is claimed in the subsequent recollections, we must perforce, other things being equal, decide in favor of the written record. If on the other hand the written record merely omits the mention of certain things later embodied in recollections, those recollections would not necessarily have to be rejected at all. Their acceptability would depend entirely upon the circumstances, and here at once we come to another necessary qualification of that canon of evidence, the second essential qualification. It is this: In order to give the written record that paramount authority claimed for it, the conditions under which it is written must have covered all the subject matter of the subsequent
