>What I'd call a velar whistle is a relatively low-pitched whistle with a
>constriction in the back of the mouth (you can get even lower pitched
>whistles, which I guess would be uvular) and rounded lips with a very small
>opening between the lips. A labialized voiceless velar approximant? The tip
>and front of the tongue are then free to move around and change the volume
>of the resonant cavity (changing the pitch), and even to add gestures like
>tonguing, taps, and trills. With practice you can control pitch to a fairly
>good degree of accuracy and speed. *Trilled* whistles, though possible, are
>probably too difficult to use as human speech sounds, but who knows what
>non-human mouths are capable of?

Hmm... trilled whistles are cool things.  You're probably right that they,
and whistles in general, too complex to be used as human speech sounds,
though; after about the palatal "hiss", I start failing to be able to
produce them.  (What a verb form!)  It seems that these nonhuman mouths of
ours tend to be able to produce increasingly exotic speech sounds: the
famous insect clicks and whirrs, these whistles and trills, the ejective
and implosive sounds of the more exotic human tongues.  I wonder now if
there are conlangs that make distinctions more fine: I was pondering at one
point a phonemic distinction between apical and coronal /s/, but came to
the conclusion that reasonably humanoid speakers (those that use human-
speakable speech sounds, at any rate) would probably conflate those into a
series of homophonous words with double meanings.  I'm told that Malayalam
has five <n>s.  And so forth.

P.S. This post, in its ending, seems to not truly contain any information.
How stressful.

---
Shreyas