94 ROYAL SOCIETY OF CANADA 



natural to an English person, the two forms are foreign to each other-in 

 the Pacitic islands. Take again the word for egg : it is munttro in Liang, 

 mantirhui in Morella, munteloa in Batumerah, momatiro in Lariki, but in 

 all these cases the tirst syllable stands for memo, a bird. All of these 

 dialects, therefore, are Melanesian and not Malay. It is very common 

 among uncivilized people to call the lingers the children of the hand. 

 Were this the case in Malay- Polynesian, the word children would come 

 tirst as in the Tongan cow-nima, the company of the hand, in which cow 

 is company and nima hand. But in ïeor fingers are limintayin ; in 

 Larika limahato ; in Cajeli limam-kokon ; in Liang rima-kuhatu ; in 

 Ainblaw lemnati-kokoli. In these cases the first word is pure Mala}', lima 

 or rima, the hand, but the construction is that of a people who had not 

 submitted to Malay syntax, 



it may be objected that this Melanesian syntax is found not only in 

 Ilaidah, but also in a very large number of American aboriginal lan- 

 guages. This is true. It is the order in Iroquois and Dacotah, Cherokee 

 and Choctaw, Shoshonese and Zuni, Aztec, Peruvian and Chileno. It is 

 also the Turanian order in Europe and Asia, counting out the Chinese 

 and their monosjdlabic associates. But these other American, and the 

 European and Asiatic postponers, have not, like the Haidahs, a Melane- 

 sian-Mala}' vocabular}-. Their words are. with a few exceptions that 

 tend to show the unity of all speech, quite different from those of the 

 Queen Charlotte Islanders. The Haidahs have articles, definite and in- 

 definite. The Turanians proper of Europe, Asia and America, have none. 

 But the Algonquins have, and the Malay-Polynesians and the Caffres of 

 Africa, and also the Melanesians. Whether the latter borrowed them 

 from the Malays or not who can tell ? The Haidah articles are ninuj and 

 Ith, and the Australian are unni and (jali. The Haidah thus ])resents a 

 peculiar philological study as a purely Turanian language, in syntactical 

 order, that has borrowed extensively from the Malay vocabulary, and 

 that, probably from the same source, has differentiated itself from other 

 Turanian languages by the appropriation of a spurious article, its post- 

 positional particles are not without analogy to the Japanese and cognate 

 tongues, but their affinities ai'e all with those of the Melanesian area, and 

 in particular with those of far distant Australia. In Australian speech 

 we probably have the Melanesian at its purest and, unfortunately, at its 

 scantiest. 



Commerce has carried the Malay numerals all over the Pacific into 

 almost every Melanesian habitat except Australia. The original Mela- 

 nesian type, of which the Ilaidah is a rescript, is lost ; even Australia, 

 which only counts as far as four, does not know it. It has to be picked 

 uji in fragments scattered over the whole insulai- area. A reference to 

 the appended vocabulary will show that the chief affinities of the Ilaidah 

 numerals are with those of Timbora, or Tambora, and Sumbawa, con- 



