POLYNESIAN AND MALAYAN. 127 
It will tend toward the simplification of this nexus, which seems the 
more complicated as we extend the view, if we dissect out the negatives 
which are found singly or in combination in these Polynesian languages 
before we advance upon other allied tongues. From the variety of the 
Samoan /é particular and sé indefinite negatives, corresponding in the 
functional value of the consonantal modulant to the weak demonstra- 
tives (article value) le particular and se indefinite, we infer a primal neg- 
ative é@. ‘This we find in an unmodulated condition in Rapanui, and in 
the Rotuma composite eaki corresponding to Samoan Jea‘s it again 
appears. For the l and the Ss modulants I can discern no value other 
than that of indicating precision. In composition with other elements 
they recur in the following forms, postponing consideration of the value 
of the composition members: ohe, aohe, ole, aole, ore (0e), aore (aoe), hore, 
ahore (kahore), kore (koe), kakore (kakoe), kare. ‘The same primal nega- 
tive receives the consonantal modulant t in certain languages of the 
Tongafiti stock, #é in Maori, Rapanui, Mangareva, and the Marquesas. 
This modulant may be regarded as the definite modulant, such asin the 
same migration group we find in the article te; at the same time we may 
find reason to assign to the t in this composition a negative value of its 
own. ‘Toward the latter interpretation operates the fact that in the 
range of Polynesian #é is considered so strongly negative that it requires 
no bolstering with other negative particles, which we have just observed 
to be so extensive in the case of sé and in an even more highly marked 
degree of /@. ‘The second stage away from the primal negative é shows 
the preface of ko to sé and to /é, thus producing a typical kosé which is 
inferential from Hawaiian ohé, and kore. In Hawaii and Tahiti, which 
lack k, we find this stage in ohe, ole, ore; and in the Marquesas, which 
drops the liquid also, we find no more than oe. ‘This preface syllable 
is itself a negative, as we may see from Rapanui ko; the composite 
is, therefore, a determinant compound in which two stems carrying 
inter alia one signification in common are compacted in order to set the 
meaning beyond doubt. With kore we shall probably associate the 
hore which gives us a series of three membersin Maori. Iam unable to 
discover ho elsewhere in Polynesia in a negation use, and the k—h muta- 
tion, while it is phonetically possible, I can not find in Polynesian use. 
A variant kare in Rarotonga and Mangaia has peculiar interest because 
we meet the same form far to the westward in Maewo kare, dehortative 
“do not.’’ In the discussion of Melanesian negatives I hope to be able 
to show that ka in itself carries negative value. In this place we shall 
assume this to be fact and shall estimate kare not as a vocalic mutant of 
kore, for vocalic mutation is almost unknown in Polynesian, but as a 
/é compound with the negative preface ka. ‘This same ka gives us the 
third stage of the Polynesian negative, kose and kole prefaced by ka, 
doubly a determinant compound, “‘no—no-no,”’ which ought surely to be 
beyond all chance of miscomprehension. In this third type we have 
from se the Hawaiian aohe rising in the loss of k twice, and from lé we 
