Hi  Greg 

Re: the voice change and resulting clause markers around user defined
dictionary pronunciations:

I played with it some with voice locking on and off,
and personally, I think the pause it introduces is just right.

It might be that we are using the feature in very different contexts; you're
reading literature whereas I'm reading software manuals
and program source code

but for instance take the following sentence from the OLE manual;
it's a partial line of C code.

HRESULT ( __ stdcall __ RPC_ FAR *QueryInterface )(

without appropriate dictionary definitions the above line is pretty hard to
deal with unless you painfully have HRESULT and stdcall spelt out.
After defining pronunciations as
HRESULT == H result
stdcall == std call

with my dictionaries active and split caps on
(which results in QueryInterface being spoken correctly) I hear

(I'll indicate pauses with commas)

,H RESULT,  ( __ ,std call, __ RPC_ FAR *QueryInterface )(

Notice that the pauses around the phrase " h result "
and the pauses around the phrase " std call "
help you realize the grouping; otherwise you would have a hard time parsing
the utterance  in that it would not be obvious that "std" and "call" are part
of the same phrase --a phrase that results from the dictionary substitution.
When I listen to the same with voice locking turned off but dictionaries
active  I  find that one hears "call __rpc" as one phrase --whereas the "call"
is actually part of phrase "std call"

For the present,  I suggest you turn off voice locking if you really find the
clause separation disturbing when reading non-technical material;
however, having played with different intonational structures on the Dectalk
in the past,
I'm fairly convinced that what emacspeak is currently producing is in fact an
optimal choice in terms of being expressive as well as succinctly unambiguous.
Even when reading literary texts, when forced to define pronunciations that
have a space in them, ie are made up of a group of I suspect you'll run into
the same "ambiguous grouping" problem
if it weren't for the voice change and resulting pause.

At the risk of rambling, let me add that AsTeR used this same inflection/pause
trick to enable the ear to logically group related parts of fractions

Note: if others on the list have opinions/comments, feel free to voice them
here --which is why I am copying the list.


Best Regards,

      Adobe Systems                 Tel: 1 (408) 536 3945   (W14-129)
      Advanced Technology Group     Fax: 1 (408) 537 4042 
      (W14 129) 345 Park Avenue     Email: raman@adobe.com 
      San Jose , CA 95110 -2704     Email:  raman@cs.cornell.edu
      http://labrador.corp.adobe.com/~raman/raman.html (Adobe  Internal)
      http://www.cs.cornell.edu/Info/People/raman/raman.html  (Cornell)
    Disclaimer: The opinions expressed are my own and in no way should be taken
as representative of my employer, Adobe Systems Inc.