34 EASTER ISLAND. 



And in the present score of years London has become alarmed at the 

 remarkable spread of the a in lady to a something which we do not 

 exactly represent in type by lydy or laidy, the sudden extension of 

 a narrowly restricted Middlesex village dialect which had lain dor- 

 mant for centuries until this modern weed growth, and now baffles all 

 efforts at explanation. 



A very small area of the general vowel-change has been set apart 

 into artificial classes and designated ablaut and umlaut, active under 

 impulses which we scarcely yet begin to comprehend. In the languages 

 familiar to our use the lasting frame is the consonant, the vowel may 

 change almost in a year; but in Polynesia the skeleton of the word is 

 the vowel. The consonants are yet but few, a sign in this case of recent 

 and partial development as genetic conditions have served ; they are so 

 dotted over the buccal speech-area as to suggest that they are little more 

 than samples of what may long ages hence be needed. They are sub- 

 ject to mutation along lines which we may readily explain; they are 

 frequently subject to extinction without entailing any serious loss of 

 comprehensibility. But the vowel remains firm and unwavering; it is 

 the real skeleton of the Polynesian speech-body. 



Let us clinch the statement by a simple illustration, and in this we 

 may draw upon the Samoan as representing the central and least 

 modified type of Polynesian speech. 



We are all familiar with English types of inflection employing such 

 forms as sang, sing, song, sung. This English series has been subjected 

 to purposeful vowel change, yet the sense runs one and undivided 

 throughout; the stem has but undergone ablaut. Yet if we were to 

 attempt to subject to such vocalic mutation a similar Samoan couplet 

 of consonants, as t-ng, we should have tagi to cry, togi to peck, tugi to 

 set afire. In the Samoan series, which is not in the least a speech-series, 

 the same change of vowels gives us a new word in each case. Although 

 the consonants remain unaltered in themselves and in their relative 

 position the shift of vowel gives a complete alteration of sense. Clearly 

 the skeleton of these words is not in the consonants. Now let us 

 examine the first of these Polynesian words and notice the consonant 

 modifications it may undergo and yet carry the sense unmodified in 

 various dialects of the Polynesian family and as loan material in 

 Melanesian languages, 



tangi tani taki tai kani angi jangi hai 



Each consonant has undergone each and every of the changes which are 

 its phonetic possibility, even to extinction. In the final reduction we 

 are led to a specimen so elemental that we find no consonant other than 

 an aspirate, a mere initial breathing, scarcely more vocal than an ap- 

 pulse ; but throughout the changes the vowel a and the vowel i remain 

 unchanged in themselves and in their relative position. The life of all 



