The Evolution of Morality. 239 



data are so scanty, the lacunce so numerous, that 

 almost any hypothesis, it would seem, might es 

 tablish some claim to verification. Our informa 

 tion is made up of a collection of scattered 

 observations on the marriage customs of a small 

 part of the human family. Moved by the scien 

 tific impulse, we attempt to discover their origin 

 and causes. But if even in physical investiga- 

 gations, where complicating conditions may be 

 eliminated, we are always liable to error from the 

 possibility of a plurality of causes, how much 

 more so in dealing with social phenomena which 

 are inextricably entangled and intertwined. The 

 ignoring of this limitation is the weak point in 

 the argument of Professor Robertson Smith, 

 whose &quot; Kinship and Marriage in Early Arabia &quot; 

 is otherwise (if I may say so) a model of philo- 

 logico-historical research. &quot;When Professor Smith 

 lays down (p. 132) that &quot; the very object of hy 

 pothesis is to inquire whether a real cause (vera 

 causa) has not had a wider operation than there 

 is any direct evidence for,&quot; his position may not 

 be disputed ; but when he adds &quot; the necessary 

 and sufficient proof that this is so is the wide 

 prevalence of effects which the cause is adequate 

 to produce,&quot; he overlooks altogether the possi 

 bility, and, indeed, in human affairs the proba- 



