“-— 
POLYNESIAN AND MALAYAN. 111 
vavy. In the utter absence of 1 we may not be positive in associating 
these forms with fine; yet it is possible that they may be anomalous 
devolution products from the Celebes fafine type, and this possibility 
is made more probable by the occurrence of fa/ay1 in the Igorot. It has 
seemed to me that fa of fafine may be associable with the next ensuing 
item, for we find in fa and faka a sense of resemblance in addition to 
the more frequent causative employment. 
13. fa, faka, formative; Subanu po, poc, pocoid. P. W. 270. 
faka Futuna, Tonga, Paumotu, Uvea,Ro- | pag Tagalog, Bicol, Visayan. 
tuma. poco Subanu. 
fak Rotuma. pog Subanu. 
vaka Viti. poc Subanu. 
whaka Maori. maka Sulu. 
haka Marquesas, Paumotu, Nukuoro, | mak Sulu. 
Tongarewa, Rapanui. mag ‘Tagalog, Bicol. 
hanga Rapanui, Paumotu. 
hoko Moriori. fa Samoa, Uvea, Paumotu. 
aka Rarotonga, Mangareva, Bukabuka. | va Viti. 
anga Mangareva. wha Maori. 
fa‘a Samoa. ha Tonga, Rapanui, Marquesas, Hawaii. 
faa Tonga, Uvea, Tahiti. ho Hawaii. 
haa Tonga, Nukuoro, Tahiti, Marque- | a Rotuma, Rapanui, Paumotu, Ta- 
sas, Hawaii. hiti. 
hoo Hawaii. 
—> pa Visayan, Bontoc Igorot. 
paga Visayan. po Subanu. 
Mr. Tregear has frequently called upon me to suggest some expla- 
nation for the hoko and ho forms of this most largely utilized of all the 
composition members in Polynesian. While I do not incline to regard 
the vowel mutability here in the Philippines as of much diagnostic 
value, I think that our Subanu wiil answer his query, which up to this 
discovery has wanted a satisfactory reply. It will be noted that the 
Subanu is a secluded speech within the area of the Malayan archetype ; 
the Moriori is an equally remote and ancient form of the Polynesian; 
and the Hawaiian ho, without this knowledge of its source, I have 
already employed in proof of the early settlement of those islands by 
Proto-Samoans long before the era of the Tongafiti migrations. Pre- 
bendary Codrington writes upon this formative prefix (Melanesian 
Languages, page 184): 
The causative is almost universally va, alone or with a second syllable 
ka, ga. The form va, fa, pa undoubtedly appears to be the original particle, to 
which ka, ga, ha has been attached. ‘This may perhaps be the verbal particle 
ka, ga, which is used in several languages. 
I am by no means convinced of the justice of this determination. 
The Polynesian exhibits a complete devolution system faka—fak—fa, and 
in the Philippines we now see similar systems, paga—pag—pa and poco- 
poc-po. In the biological study of the upbuilding of the Polynesian 
I shall give due weight to Codrington’s suggestion, but merely as a 
matter of the etymology of the languages in their present phase it is 
quite clear that we pass by abrasion from paga to pag, from poco to poc. 
