Johnathan Knibb wrote:
> Benct Jonsson wrote:
> >It claimed that the actual phonetic realization of
> >'/tSr/' where it occurs is [t`s`] -- not unreasonably
> >if the realization of /r/ is retroflex, which it
> >apparently is in most accents in both the US and
> >the UK.
>
> [t`s`] feels righter to me than [tS], partly because I
> feel a slightly different articulatory gesture towards
> the closure in /tS/ than in /tr/ - the first is laminal and
> the second apical. In fact, the gesture itself doesn't
> feel any different in /tr/ than in /t/+vowel.
(snip the rest of a very accurate and thoughtful analysis-- with which I 
agree 100%-- of the question.)

There is also the interesting phenomenon, that English has no words 
beginning with [tSr\], that is, spelled "chr..." (and don't bring up 
"Christian", that's not what I mean...), and I can think of only one that 
even comes close-- fast speech (and non-native) "cheroot" [tS@'rut]*--  
which even so is going to be 2 syllables. (Make that two exs., if you count 
"-cheree" in Ms. Poppins' "chim-chim-cheree" however it's spelled...).
----------------
*also pronounced [S@e'rut], I know; similarly no [dZr\]-- again, some come 
close; Gerard, giraffe, Jerusalem.  I don't get a [Z] transition in [dr-], 
probably because my [d] is just a tad further back than my [t], plus the 
cluster is fully voiced, which is relevant.
--------------------

A bit OT of course, but interesting too-- there are lots of s+C initials, 
but no native /sr-/; conversely there are no native sh+C except shr-. If I'm 
not mistaken, those descend from *skr- (+ front V?).  But I once saw 
(prominently displayed on a restaurant window) "srimp" for "shrimp" :-))).