[bowman] fundamental PROCESSES IN HISTORICAL SCIENCE 551 



selected for the unpleasant mission by the advice and machinations of 

 William, brother of the king and archbishop of Mainz, though he had 

 never offended against that prelate^ — both of which points, so far from 

 being intimate, would in the ordinary course become common know- 

 ledge in a comparatively wide circle, especially in so imporaant an 

 affair between dignitaries of the church and state. Thus the 1st and 

 2nd groups in the above reasons are based on error and exaggeration. 

 There remains the 3rd, i.e., Adalbert's attainments and opportunities 

 for writing such a chronicle, and the coincidence between its close and 

 his appointment as archbishop of Magdeburg. Here let us apply the 

 definite test, "Could additional evidence establish another than 

 Adalbert as author, without disproving either of these two points?" 

 The answer is. Yes, because Adalbert was but one of a number with 

 like attainments and opportunities, and such a chronicle may be dis- 

 continued at any time, for any one of many reasons, or for no particular 

 reason at all. On the other hand, to a similar question concerning the 

 identification of the Annals of Rosenfeld, the answer should be. No, 

 because it is not reasonably conceivable that additional evidence 

 could show that these annals were written elsewhere than at Rosenfeld, 

 so long as the bare entry stands under the year 1130 without further 

 specification of the monastery, "Abbot Kuno died," and there is no 

 disproof that Rosenfeld was the monastery over which Kuno presided 

 until 1130. The inference identifying the annals with Rosenfeld, 

 accordingly, is necessary and scientific, while that which identifies 

 Adalbert as the author of the Continuator is unnecessary, and there- 

 fore (though his name is bracketed as such in the title of Kurze's 

 octavo edition of the Continuation, because "to-day almost no one 

 seems any longer to doubt it") is unscientific. Adalbert may, indeed, 

 have been its author. Additional evidence may yet show that 

 he was the author. But not even this would be a retrospective justi- 

 fication of present publicity unnecessarily given to an unnecessary 

 conclusion. The scientific requirement would still be to have sought 

 the evidence, not to have anticipated it. 



{b). Pure Probability (the Superior Measure). 



Probability of the superior measure, which may be called for 

 convenient identification "pure probability," produces in any consider- 

 able series of results an average of correctness corresponding to the 

 degree of probability involved. This form of probability supplies a 

 scientific basis for some rules of method applicable to constantly re- 

 curring situations in historical research. Thus, if, out of three inde- 

 pendent, trustworthy records (or witnesses), two are contradicted on 

 any point by the third, and the known circumstances of the discrep- 



