H.— ANTHROPOLOGY. 147 
was directing his attention to the method which I am here speaking of as 
that of Comparative Sociology. 
In the change of point of view that he made in 1911 Rivers was there- 
fore representative of a general tendency. There had been a growing 
dissatisfaction with the theories of social anthropology. From the point 
of view of a desire for historical explanation that dissatisfaction is, I think, 
justified. A historical study ‘ explains’ by revealing particular relations 
between particular phenomena or events. History does not generalise or 
cannot legitimately doso. It shows us that at a given moment a particular 
event occurred, and as a result of this something else happened. Thus, a 
cause in historical explanation is something which happened once and was 
followed by certain results. It is not similar to what is called a cause 
in natural science, which is an event that recurs or may recur repeatedly 
and always produces the same effect. Historical explanation is always 
concerned with particulars, normally with showing a chronological relation 
between two or more particulars. The value of historical explanation is 
therefore directly proportional to the amount of certain and detailed 
_ knowledge that we have of the events with which we are concerned. 
| It may be said in one sense that the ethnologist explains the existing 
similarities and differences between peoples by means of his historical 
hypotheses. Actually, however, he is not interested, at any rate primarily, 
in explanation. Where he attempts a reconstruction of history it is 
because he wishes to discover something about a past of which we have 
no records in written documents. He is interested in a knowledge about 
_ the past, as far as it is attainable, for its own sake. Or if the ethnologist 
believes himself to be following some other aim, then he is pursuing the 
wrong method. All that his hypothesis can give him will be a certain 
number of more or less probable statements about the past. And his 
results will only be valuable or valid if he avoids basing them on assump- 
tions as to general principles of historical change which have not been 
demonstrated by sociology, for it is the specific task of sociology to discover 
such principles. 
The methodological difficulty in ethnology is, and always will be, the 
demonstration of its hypotheses. I do not suppose that anyone has 
ever accepted, or ever will accept, Rivers’s elaborate reconstruction of the 
history of Melanesia. The theories of culture cycles, that are held so firmly 
by some ethnologists that they speak of them as though they .were 
demonstrated beyond any possibility of doubt, are totally rejected by other 
_ competent and open-minded students. The Egyptian theory of the origin 
of culture has its special devotees, but so has the Atlantis theory. 
It is certain that the ethnological method carefully used may give us 
“avery limited number of highly probable, if not quite certain, conclusions. 
Thus there is no doubt that the language of Madagascar and a good deal of 
is culture are derived either from Indonesia or from some region from which 
1¢ Indonesian languages and culture were also derived. In such an 
istance we are dealing with a great number of resemblances between the 
vo regions which cannot be otherwise explained, and the matter of the 
languages is conclusive. Similarly it might be possible to demonstrate 
some sort of general relationship between Australia and South India, or 
between Indonesia and Melanesia. But it seems to me highly doubtful 
L2 
= 
