DISSECTION OF THE THEORY. 



33 



In this table I have made use of the convenience of abbreviations, 

 M for Melanesian, V for Viti, P for Polynesian, My for Malayan, 

 S for Semitic. With each entry the name Efate is to be understood. 

 Thus the first entry in the table, M 31, signifies that Dr. Macdonald 

 notes 31 words in which he finds for Efate relation with some lan- 

 guage elsewhere in Melanesia, but without recognizing further affilia- 

 tions. Similarly the final entry signifies that he has found 29 words 

 in Efate for which he provides identification with words along the 

 whole chain from Polynesia to the Arabian Saba. 



In such a study as this the position of the Fijian speech is anom- 

 alous. It is neither wholly Melanesian nor pure Polynesian. As in 

 part it may be claimed for each household, there seemed but one 

 way of securing fair treatment wherever it chances to be involved 

 in this investigation; namely, to set it apart and in each instance 

 to sift out the Melanesian or the Polynesian affiliation. 



The total of the words for which his ingenuity has suggested 

 affiliations outside the island of Efate is 1,154, or 31.55 per cent 

 of this dictionary, which he calls complete. That is to say, for not 

 so much as one-third of the vocables in the language of Efate can 

 his liveliest fancy — how lively that may be we shall see hereafter — 

 find any sort of ground for his promise of "the Oceanic languages, 

 their material or vocabulary set forth in a complete dictionary of 

 one of them." Furthermore we find that an even smaller number, 

 in effect no more than a quarter of the whole number of entries, 

 covers all the suggestions of Semitic affiliation which he has ven- 

 tured upon. Compare his claim that no more than 29 words establish 

 the complete chain from Arabia to Samoa with his identification of 

 346 words which connect Efate with the Semitic without having 

 left a trace elsewhere in Melanesia or anywhere throughout the 



