Tag. -q & -h

[potetjp]

    Laurie Reid's remark that in Cordilleran and Manobo languages no _h_ is
inserted between vowel-final stems and suffixes is both capital and
illuminating.
    What happens is such languages is the normal phenomenon. After all, we
expect a glottal stop in such a position, not a glottal fricative. Let's
compare for instance Ilok. _sim[ba] + -an >  sim[ba]qan_  with Tag.
_sim[ba]_ + -an > sim[ba]han_ "church".  Why then does Tagalog have an _h_
in this derivation ? Probably because there already is a phonemic final
glottal fricative in Tagalog that does not exist in Ilokano, hence the
necessity to contrast it with another phoneme. But why _h_? Why not any of
the other Tagalog sounds that can fill the final position? Why don't we have
_*simbakan, *simbalan, *simbanan, *simbasan_ etc.? Is it because  _h_ is the
only other glottal available? Probably. But can it spring from thin air?
Doubtfully. So where is the phoneme that generated this [h] phone?
    I have tried to address Daniel Kaufman's valuable suggestion that
CVC-CVC words tend to have their last C filled by the glottal stop.
    I used my corpus of Tagalog di- and trisyllasic stems and examined those
beginning with B and S, taken randomly from the alphabet. I found 436 stems
ending in a vowel in typical, unaccented Tagalog spelling. 212 (48.5%) of
these stems - percentages are rounded - have the final glottal stop, and 224
(51.5%) the final nihil phoneme [Even if you won't buy it, just see it as a
convenient fiction.].
    Among the former, 148 (34%) are of the CV-CVq type, 48 (11%)  are of the
CVC-CVq type, 5 (1.15%) of the CVC-hVq type, and 1 (0.25%) of the CV-hVq
type. Without being conclusive, these amateurish statistics tend to stultify
the hypothesis that there might be a relationship between the CVC-CVC type
and the final glottal stop. Nor do Spanish loanwords support this hypothesis
either. Too bad.
    Best regards
Jean-Paul  G. Potet