[bowman] fundamental PROCESSES IN HISTORICAL SCIENCE 563 



though it deals with persons and personal elements, is above personal 

 considerations. The example of the courts in civil cases, therefore, 

 does not apply here ; and the entire series of obligatory probable con- 

 clusions in practical affairs, and in practical applications of the sciences, 

 is no precedent for the acceptance of probable conclusions in scientific 

 investigation, because in such investigation no decision is ever obliga- 

 tory. The scientist has always the privilege of silence, and the his- 

 torian, wherever he regards a conclusion as not free of reasonable doubt, 

 is at liberty to omit it from his final results. 



4. Mixed phraseology concerning probability. Mixed phraseology 

 in legal works and dictionaries is another encouragement to the present 

 probable conclusions in history. Thus, a very good legal work (Phip- 

 son. Law of Evidence pp. 5, 6) uses the phrases "preponderance of 

 evidence" and "preponderance of probabilities" interchangeably, 

 thereby seeming to imply (though the inference is not specifically made, 

 and may not be intended) that probability is as good a thing as 

 evidence. In dictionaries the confusion in definitions reaches the 

 point of complete contradiction. Probability as a term by itself is 

 defined in them in the ordinary sense that the supporting evidence 

 inclines the mind to belief, but still leaves room for doubt. The same 

 dictionaries, however, under the terms "moral" and "certainty," 

 define moral certainty (which excludes reasonable doubt) as a high 

 degree of probability. Earlier editions of Webster {e.g., 1856), 

 indeed, state specifically under "probability," that this falls short 

 of moral certainty; but this statement is omitted in a more recent 

 edition (1900), which joins with the Century Dictionary (1895) and 

 the Standard Dictionary (1893, 1913) in the above contradiction. It 

 is not unusual for the same word to have different shades of kindred 

 meaning and even to show considerable variation from this kinship 

 or common starting-point in meaning. In such cases an indiscrim- 

 inate use of the word is not permissible; but an exact delimitation of 

 those differences or variations is recognized as necessary in order to 

 prevent misunderstandings, confusion of thought, and errors, which 

 would otherwise arise; and especially to prevent the basing of prem- 

 ises on one meaning and the drawing of a conclusion on another. 

 Where the difference amounts, as here, to a flat contradiction, the need 

 of an exact delimitation is so much the greater. Of the various forms 

 of probability, pure probability, regardless of its height, does not 

 exclude reasonable doubt except in averaged results. Reasoned 

 probability used as a positive basis for conclusions does not exclude 

 reasonable doubt; when used as a negative force to overthrow, if 

 possible, conclusions and inferences seemingly necessary, it can 

 assist in the exclusion of reasonable doubt. Formal probability, the 



