Re: Linguistic Terminology

Eric Christopherson wrote:
> True. But I was wondering, are [S] and [s] allophones of /s/ in
> Japanese, or are they separate phonemes? I would think phonemes, since
> [s] can occur before /a/, /u/, /e/, or /o/ and [S] can occur before
> /a/, /i/, /u/, or /o/, thus the domains of each intersect.
Right, they are seperate phonemes.  When in doubt, make a list of where
each can occur.  So, for [S] and [s]
[S] i, u, o, a
[s] e, u, o, a

Thus, either can be used before u, o, and a, and therefore they are
seperate phonemes.  If, however, they were in complementary
distrubution, such as [t] and [t_h] in English:
[t]: syll-initial, after /s/
     preceding an approximate (which is devoiced, /pjr=/ = [pj_0r=])
     syll-final (depends on dialect)
[t_h]: syll-initial, not after /s/ or followed by an approximate

Thus, they are complimentary, there is no overlap, and thus they are
allophones of one phoneme.

Allophones may also be in *free variation*, where they can occur in the
same position, but not consistently.  I can't think of any examples off
the top of my head, but essentially this would be a case where one could
have, say, either [sale] or [Sale] for the same word in a given
language.  These are not complimentary, but they are not consistent, in
other words, there are no words where you would only hear, say, [sako]
but never [Sako], or vice versa.  This isn't common, and is rarely, if
ever, random (social factors often play a role - different social
classes might use one phone more than the other, for instance)