56 EASTER ISLAND. 



to the spirit of the known Polynesian. The only elements which are at 

 all to be recognized outside the Polynesian family are a very few of that 

 small group common to Polynesian and Malayan. The position of 

 this element I have discussed at great length in ' ' The Polynesian Wan- 

 derings ' ' and have established the proof of borrowing by the Malayans 

 from the earlier Polynesian peoples of Indonesia. This element, there- 

 fore, is to be held as true Polynesian, not a Malayan contamination. 

 The only other sources of such speech metamorphosis fall into two 

 classes, according as we regard the Proto-Samoan migration or the 

 Tongafiti migration as colporteurs. For the latter we can not speak; 

 not as yet can we identify its voyagings earlier than its appearance 

 in Nuclear Polynesia, except that negatively and exclusively we are 

 convinced that it did not follow the course along the Melanesian archi- 

 pelagoes. To the earliest Proto-Samoan migrants occurred the oppor- 

 tunity of acquiring Melanesian speech material. To each item in the 

 data of this work where the word is recognizable in Melanesia, despite 

 savage mutilations, I have made a note of reference to my former work; 

 from this it will readily be seen that the word in Polynesian can not be 

 due to Melanesian contamination, but that it occurs among the darker 

 race as a borrowing from the more intelligent Polynesian commorant 

 for a more or less extended sojourn in their abodes. A discussion of the 

 improbability of Melanesian contamination of Polynesian, at far greater 

 length than is here desirable, will be found in "The Polynesian Wan- 

 derings" at page 149. 



This problem is one which we shall encounter in the detailed examina- 

 tion of each Polynesian language; each will exhibit its distinctive per- 

 centage of recognized affiliates, each will have a residuum which is not 

 to be identified in any other language of the family in that modern 

 phase in which alone we may know it. The mere accident that, in other 

 languages of the family, these residual vocables have gone into disuse 

 need not rob them of their Polynesian heritage. Therefore in dealing 

 with the several sets of percentages I adopt for my denominator the 

 sum of the affiliates as being the true representative of the character 

 of the speech, the unrecognized mass being set apart as not conditioning 

 the problem. 



It is easy for a word to go into disuse in any language ; that is one of 

 the incidents of growth. Not all of us understand the English of 

 Shakespeare, a fact which is scumbled in our perception by the fact that 

 in those texts we have the keen zest in the narrative to carry us past the 

 incomprehensibilities scarcely noticed. Still less do we comprehend 

 the King James English of the Bible, a fact piously obscured in the gen- 

 eral feeling that ignorance is the handmaid of theology. If these facts 

 are undeniable in a language of written record and lexicographic exacti- 

 tude much more must such be the case in the speech of simple islanders 

 who know no letters. The Polynesian is of the earliest type of speech, 



