558 THE ROYAL SOCIETY OF CANADA 



table, if introduced by a user into his calculations, will cause a rapid 

 extension and increase of error under the process of multiplication and 

 division, while in history an erroneous statement adopted from a 

 record will affect the user's results but little, if any, beyond the adopted 

 statement itself. A very small percentage of gross errors, indeed, 

 would deprive a historical record of the character of trustworthiness; 

 but the gross errors to which probability as a positive historical cri- 

 terion may lead, are not the kind of error that is apt to occur under the 

 application of correct processes. Real errors, indeed, especially those 

 of minor importance, cannot be escaped entirely; and in history, as 

 in other sciences, there are often inaccuracies that are known and un- 

 avoidable, yet not of essential importance in the results as a whole. 

 Thus, in a mathematical table, where the limitations of space and the 

 convenience of the user require that the quantities shall run only to 

 a certain number of decimals, the last decimal must ordinarily have an 

 error in deficiency, or if 1 be added to it, there will then be an error in 

 excess. The editor must choose between these two inaccuracies 

 according as to which will be the greater; and neither inaccuracy will 

 affect appreciably the operations of a user. In the same way the laws 

 of perspective and limitations of space will not allow a narrator to 

 speak at the same time from more than one of, it may be, several 

 stand-points, nor to tell all the truth to the last detail even from that 

 one stand-point. The stand-points between which an historian may 

 have to choose, moreover, are sometimes equally good; and what 

 he says from the one stand-point, will seem, to a person viewing it 

 from the other, to be an error (See Origin and Treatment of Dis- 

 crepancy, pp. 167-168). As result in part of real errors of a minor 

 nature and in part of the above unavoidable restrictions, especially 

 in general works of history, a special investigator of any particular 

 period, or series of events, or single event, finds, where a general work 

 touches on his particular field, quite a percentage of points or state- 

 ments which he feels should have been put differently, or would be 

 the better for a little explanation. An average of 1 of these real minor 

 errors and unavoidable inaccuracies (for want of explanations which 

 space and perspective forbid) together for every 100 statements would 

 seem low, and 1 in 50 moderate (making the required average of essen- 

 tial correctness 99/100 and 49/50 respectively), but even 1 in 25 would 

 not be out of the question, for a narrative that was correct in major 

 points and 24/25 right in minor would still have historical value; 

 and might be worth more than another narrative with a greater pur- 

 suit of accuracy in detail and a correspondingly poorer perspective. 

 The value of an historical narrative by no means depends entirely on 

 minute accuracy. Green's History of the English People, e.g., has a 



