1 92 Heredity. 



that will be committed in each country during the year. If we 

 look into the French criminal reports and compare several years, 

 we shall be surprised to find that various crimes and offences, 

 classed under a score of heads, oscillate within very restricted 

 limits. The number of suicides, too, is much the same for each 

 year ; in five years it varied in London between two hundred and 

 thirteen and two hundred and sixty-six. Nay, even occurrences 

 which might appear to be governed entirely by chance, and to 

 result from pure stupidity, are not without regularity. It has been 

 shown that in London and in Paris about the same number of 

 letters without an address are posted every year. 



I have no wish to discuss here whether or no we are free agents, 

 nor whether that problem can be resolved by the present method. 

 My object is only to inquire whether it can lead to quantitative 

 determination that is, to absolute certitude. It is plain that it 

 cannot do so. When we are told that the statistical method 

 enables us to predict the number of murders, larcenies, suicides, 

 marriages, etc., the meaning is that they are foreseen in the gross 

 and approximatively ; but in true quantitative knowledge nothing 

 is determined in the gross or approximatively. Given a great man 

 in a family, does any one imagine that by means of Galton's 

 averages we can determine how many illustrious brothers, sons, or 

 nephews he will have, with as much certainty as we can calculate 

 the day and the hour of an eclipse ? 



It is, therefore, a mistake to fancy that because mathematical 

 processes are employed we can arrive at mathematical certainty. 

 The real service rendered by figures is this : there is a multitude 

 of scattered facts, which have no visible connection, and appear 

 to be perfectly fortuitous. The statistician compares these to- 

 gether, and discovers in them uniformities, or, in other words, 

 laws. And as from uniformity of effects we may infer uniformity 

 of causes ; as from moral and social facts we can ascend to the 

 psychological states from which they result, the consequence is 

 that statistics can be of service in the study of morals and even of 

 psychology. By grouping together certain phenomena of social 

 life it gives us a means by which we can verify and check our 

 conclusions ; it gives to the purely subjective views of the mind 

 the means of acquiring an objective value, and so of passing from 



