176 THE POLYNESIAN WANDERINGS. 



Beyond this we may not venture. The effort has been made to 

 tabulate for each of these languages a scheme of consonant mutation, 

 and the effort has proved vain. The trouble begins even earlier than 

 the tabulation of variant forms. 



There are cases in which there is a resemblance of form between 

 some Semitic vocable and some vocable in Efate and Melanesia and 

 Polynesia. If to the form resemblance were added correspondence 

 in signification, then we should not truly in any one instance have 

 proof that though widely separated in space they are homogenetic. 

 A single instance will suffice as well as a thousand to show the utter 

 lack of evidential value in such a case. The Hawaiian like means 

 to resemble, to be like, yet no one has yet arisen to predicate upon 

 the double identity of form and sense a theory that Hawaiian and 

 English derive from the same source, except in so far as Fornander's 

 project might be susceptible of extension to such an absurdity. 



A long catena of such dual resemblances of the common element 

 in Melanesian and Polynesian with a Semitic parent speech would 

 add confirmation with each new link as welded to the chain. Yet 

 the weakness of any chain is in its weakest link ; that weakness meas- 

 ures its utmost strength. Many of these links have been vitiated 

 by the falsity in definition hitherto animadverted upon. We have 

 pointed out instance upon instance in which through ignorance, some 

 through manifest design, the definition of Efate vocables has been 

 distorted for the clear purpose of establishing a sense resemblance 

 where a form resemblance has already been noted, either by eyesight 

 or in a rather fertile imagination. Losing our faith in this link and 

 in that we can place no trust at all in the chain. So much for that. 



We find Semitic triliterals proposed as the parents of vocables in 

 our Pacific islands in which the consonantal skeleton consists of but 

 a single consonant, or of two. We find island vocables having a 

 skeleton of three consonants proposed as in direct descent from 

 Semitic stems in which it is impossible to discover a trace of more 

 than two consonants, sometimes not even two. We find some Sem- 

 itic triliterals identified with triconsonantal stems in island languages 

 in which the order of the consonants is deranged. We find some 

 Semitic triliterals identified with triconsonantal island stems in 

 which the first Semitic element has vanished, the second and third 

 are respectively first and second in the island stem, to which is added 

 a third which corresponds to nothing in the Semitic; expressed dia- 

 grammatically it is sought to establish ABC as equal to BCD. 



In vain we strive to unravel Dr. Macdonald's attempts at con- 

 sistent explanation. These things remain a tissue of irreconcilables, 

 and the explanation but serves the more to confound them. 



Pretermitting as hopeless the task of comprehending these anoma- 

 lies so gravely proposed and so argued with what must be regarded 



