SUBANU PHONETICS AND COMPOSITION MEMBERS. 67 
when the attributive passes into verb sense. Next, inversion of the 
former syllable in the resultant gapo gives us agpo; itis very simple and 
general phonetics to find the sonant g attracting its neighbor p from 
surd to sonant in its own series; therefore agpo becomes agbo, and lack 
of vowel fixity is so characteristic of this rude speech that ugba is quite 
explicable. 
We are next to examine a phonetic usage which is not properly to 
be dealt with as a case of consonant mutation, for it affects certain con- 
sonants, the two palatal mutes, positionally; that is to say, only when 
they are used as the initial consonant of the vocable. It will be seen in 
the vocabulary that many vocables which begin with the syllable ca are 
duplicated by forms which lack the c and that to a lesser extent this 
double form is true of vocables with ga initial. Even where the double 
form does not appear in the Subanu vocabulary, a reference to the 
Visayan affiliates will show that the uncertainty exists. There are three 
seeming exceptions to this principle of uncertainty as restricted to the 
beginning of words, cotooto, gibusibus, and gonauna. It will be seen 
that while the loss appears to have taken place in a position inner with 
respect of the vocable, it is initial with respect of the stem duplicated. 
To a certain extent it has been possible to associate these variant 
forms with the several sources of the vocabulary material. Yet after 
all that leads nowhere, for there is no uniformity; the source which 
affords us the abraded form in one vocable may yield the full form in 
another and precisely similar vocable, and each in turn applies or 
neglects the initial palatal mute in the case of vocables for which we 
have a Visayan or even Spanish original. Thus from the Spanish 
caballo the Subanu borrows the transliteration cabayo and parallels it 
with an abraded form abayo. It will not be difficult in scanning the 
vocabulary under these initials to find a sufficiency of instances to show 
that the Subanu abrades the mute initial in words which clearly pos- 
sessed it in the source of the loan. On the other hand there are quite as 
many instances to show that Subanu, through some principle in its own 
phonetics, assumes C or g as initial to words which in the Visayan are 
devoid thereof; for instance gama is Visayan amahan. In our later 
examination of the exterior relations of both Subanu and Visayan we 
shall observe this word in its proper class and shall discover that the g 
is really a Subanu assumption upon a stem which in its genesis began 
with a consonant wholly distinct in series and in the speech-organ em- 
ployed. We are warranted in the statement that the Subanu assumed 
an initial consonant and that this assumed consonant tends to disappear. 
I have had a sense that this matter of the assumed initial palatal 
mute represented a senior and a junior stage of the language; that it 
was an ancient Subanu character to assume the mute, and that in the 
more recent stage it was being dropped in avoidance of dialectic rude- 
ness, as intercourse became more free with more advanced Visayan 
