> > Missing from what?  If the sounds you mentioned are the only ones not
> > your conlang then that means you must have a really awful consonantal
> > system, full of retroflex nasal clicks and voiced uvular ejectives
> > other such monsters.  <<shudder>>
> Are there even such things as voiced uvular ejectives?  Voiced
> are, IIRC, very very rare (Mam has one, I think) -- though voiced
> implosives are not.

That actually was my point.  I don't think voiced ejectives exist, and
I'm pretty sure that they're physically impossible.  In an ejective the
glottis is closed to force air upwards and create the pressure for the
ejective, which means that it cannot be open to create voicing.  When I
try, I find myself either making a voiced implosive or a creaky-voiced
stop.  In any case, I was being sarcastic and pointing out that if you
only exclude a few phonemes, you imply that what remains is everything
else that's even conceivable.

> > I'm confused now.  It sounds like you're talking about changing one
> > language into another, saying that "[v] was reassigned to [f]" and
> > "Affricates were done away with."  Yet you never directly mention
> > nor daughter language.
> I interpreted Steve's comments as saying that he started out with some
> set -- not necessarily an ideal one -- and then changed it.  That could
> one or more of at least two things:
> (1) He was envisioning an earlier stage of the same language, and what
> happened to it after those changes;
> (2) He just didn't like his earlier set, for esthetic reasons, and so
changed it
> In either case, he seems to have done so with regular *rules*, though.

This is possible.  I await Steve's response.

> > My guess is that you are (subconsciously?) starting from your idea of
> > "ideal" phoneme set, and then deciding which phonemes are added to
> > set, and which are lost.  The problem is that I don't know what your
> > ideal phoneme set is, and in fact, there *is* no universal phoneme
> True, but there are phonemes that are typologically more common than
> others -- the stops [p t k] are more common than their glottalic
> [p' t' k'], and the same voiceless stops are more common than their
> equivalents [b d g].

True, true, but the set of sounds that are nearly universal is quite
small.  I think 95% of all languages have at least /p t k s n/ and /a i
u/, though Hawaiian leaps to mind as an exception for the consonants, and
Japanese doesn't have [u].  Which direction they go for more is very
uncertain.  They might add aspiration as a feature, or voice, or add more
fricatives, or affricates, etc.  Most languages have at least one
rhotic/lateral, too, but that can be almost anything phonetically.

Jesse S. Bangs     Pelíran
"There is enough light for those that desire only to see, and enough
darkness for those of a contrary disposition."  --Blaise Pascal