826 



SCIENCE 



[N. S. Vol. XLIV. No. 1145 



people. Last come certain recent influences 

 from Micronesia and Polynesia. Polynesia, 

 moreover, is made to participate in some of 

 the other culture strata, so that a later Poly- 

 nesian culture corresponds to an earlier Mela- 

 nesian one, while the earlier Polynesian cul- 

 ture is given a share in the moulding of the 

 culture of the dual people, which, therefore, 

 also proves to be complex in character. 



The remaining sections of the volume are 

 devoted to an interpretation of the different 

 aspects of Melanesian culture in the light of 

 the cultural strata just outlined. Thus, 

 linked totemism is regarded as due to two 

 successive migrations of totemic peoples; con- 

 ventionalized art is ascribed to the influence 

 exerted by the geometrical art of one people 

 on the realistic art of another; the origin of 

 money is seen in the conditions which arise 

 when two largely independent people live side 

 by side; religion is a trait of the kava people, 

 while the dual people were addicted to magic; 

 sun and moon worship also come from the 

 kava people, while stone work is due to ideas 

 introduced by them ; the bow and arrow belong 

 to the kava as well as to the dual people, 

 although they were subsequently lost among 

 both; the plank-canoe was shared by the 

 kava and betel peoples, while the dug-out 

 originated with the dual people; the use of 

 an inclusive and exclusive plural, finally, in 

 some of the Melanesian languages points to 

 the necessity of differentiating between two 

 social strata. 



In fairness to Dr. Rivers it must be said 

 that the bare outline presented above does but 

 poor justice to the author's amazingly complex 

 argumentation. It will suffice, however, for 

 the purpose of the present examination, which 

 is not to refute the author — a task that would 

 require a volume — but to characterize and 

 expose his method. This restriction is the 

 more justifiable as the author himself regards 

 the " history " as a model of ethnological 

 method. 



In order to allow for a more deliberate 

 analysis of the second part of Vol. II., the 

 first part will be discussed very briefly. In 

 it the author applies the method of survivals 



with little regard for probabilities. "When a 

 reconstruction based on a diagnostic utiliza- 

 tion of relationship terms leads to the as- 

 sumption of an ancient state of gerontocracy 

 of a type hitherto unknown in concrete 

 ethnographic experience, and of forms of 

 marriage, such as that between individuals 

 separated by two generations (a condition 

 which, while it seems to occur, must certainly 

 be regarded as highly exceptional), one pauses 

 to think before accepting the author's conclu- 

 sion. Again, although Dr. Rivers has cer- 

 tainly made good his contention that terms of 

 relationship will reflect states of society, par- 

 ticularly of marriage — a position once held 

 as a dogma by Lewis H. Morgan — and not- 

 withstanding the new in part very striking 

 evidence which the author's book brings in 

 support of that contention, he clearly is guilty 

 of deliberately overlooking the fact that social 

 structure and function represent but two out 

 of a set of factors which may and do influence 

 relationship terms and systems. A set of 

 terms must always remain a feature of lan- 

 guage and as such it is subject to those in- 

 fluences which control linguistic changes as 

 well as to the peculiar spirit of a particular 

 language or linguistic stock. Again, a system 

 of relationship, a set of terms, are phases of 

 culture and, like other cultural features, they 

 may spread from people to people, may be in- 

 fluenced by factors extraneous to the group to 

 which they belong. While the theoretical 

 validity of these propositions seems assured, 

 one welcomes the fact that renewed interest in 

 the numerous and intricate problems presented 

 by the study of systems of relationship is 

 manifested in a series of concrete and sys- 

 tematic investigations undertaken particularly 

 by American anthropologists, investigations 

 which have already brought valuable evidence 

 in favor of a less one-sided attitude toward 

 the problems of relationship systems and 

 from which further results along similar lines 

 may ere long be expected. 1 



But Dr. Rivers's principal error consists in 



1 Cf., for instance, K. H. Lowie, ' ' Exogamy and 

 the Classificatory Systems of Relationship, ' ' Amer- 

 ican Anthropologist, Vol. 17, 1915. 



