90 THE SUBANU. 
to light the fact that its vocabulary preserves many vocables which in 
the general Visayan are but rarely encountered. 
It is wide of the purpose of this work to point out in the vocabulary 
the Malay affiliates. We do not need them for proof that the language 
is Malayan, and it would do no more than duplicate work already 
familiar. The only exception which I have made is in the case of a few 
vocables for which I have noted affiliations in several languages in the 
Celebes waters. ‘These were collated in Dr. Elbert’s Sunda Expedi- 
tion, while the study of Subanu was in progress, and are therefore quite 
new and undoubtedly welcome additions to the stock of Malay com- 
parative material. 
There remain for our consideration a double-handful of examples of 
a mutation which is certainly anomalous, for it does not fall within the 
ordinarily established categories of phonetic variability. Absolutely 
these examples are few, yet it is clear from their repetition for so many 
as are here shown that they represent a speech principle. Therefore 
they are worthy of our examination in order that we may discover the 
principle which is operative. So far, when it has been necessary to refer 
to any of these vocables in passing, I have classed them as inversion; 
now we Shall see of what nature they really are. 
For readier vision, let us first see what would be the result if the 
principle were a part of our own speech equipment, since it is much 
easier to study it when applied to vocables more familiar than any 
which may be drawn from the distant Subanu. From the Latin ruga 
we derive the adjective rugose; if this principle were at work in our 
linguistics we should have gurose or urgose; from the German beten we 
should have tebe or edbe in place of bede as in bederoll. ‘These illustra- 
tions in familiar material, so far as they have true applicability to the 
matter, will show us that there are two possible forms; that there is an 
interchange of consonants between the second and first syllables; that 
the syllable itself is inverted and its initial consonant is made final. In 
our Subanu material we must discover which of these takes place. 
Omitting here all consideration of the usual phonetic variation 
which may or may not affect the form of the compared vocables, we 
present the examples in similar groups: 
gosod 
sogot 
casit 
saquit 
Subanu 
Visayan 
socog 
cosog 
In these the change has affected the initial syllable; we find the 
same number of instances in which it has been applied to the final syl- 
lable. : 
tungdong 
tongod 
tondong 
tongod 
timod 
tigom 
Subanu 
Visayan 
