And Rosta wrote:
> Herman Miller, On 13/05/2009 01:56:
>> This seems like a fairly readable system, and it even leaves the
>> tones on the ends unmarked (so you can use marked tones at boundaries
>> for other purposes like marking questions or for emphasis). For the 
>> tones you could use e.g.
>> a = low level
>> A = high level
>> b = falling
>> B = rising
>> Start a phrase with high level A for emphasis, or end the phrase with 
>> rising tone B to mark questions. 
> If I understand correctly, your scheme offers a redundant a/A/b/B 
> contrast on the first word of the sentence and a redundant b/B contrast 
> on the last. Is that correct?

A always groups to the right and b to the left, so the first tone is 
always "a" and the last is always "b". Any (ab) groups at the beginning 
or end are unambiguous and don't need to be marked. So the first and 
last tones are unmarked. The second tone can be either "a" or "b", but 
it turns out that it's also unmarked. If it's a "b" it forms an (ab) 
group at the start, which is unmarked, but a marked "A" ultimately 
combines with a "B" and reduces to "B".

So you've got three unmarked words in a sentence (if you're putting one 
tone on each word). I guess you could also use "b" or "B" in place of 
the first tone and "a" or "A" for the last, since the actual tones are 
predictably "a" and "b" respectively. So you could have 8 possible 
contrasts at the beginning and 4 at the end, if you really want to get 
the most out of the unused contrasts at the boundaries.

You could even go further and make use of other unused contrasts, but 
that may be more confusing than it's worth. Since a marked "A" never 
occurs without a matching "B" later on in the phrase, you could mark an 
"a" tone for emphasis if you know the phrase isn't going to have a 
marked "B" somewhere later on. You could also mark a "b" tone if there 
are no preceding "a" tones it could combine with.

Perhaps a better use for the unused contrasts is to imply omission. If 
you run across "aaA", you know there has to be a following "B" somewhere 
: "aaA ... B", which ultimately combines with the A, and that the "a" 
preceding the marked "A" has to have a matching "b". So you could use 
"aaA" as a shorthand for "aaABb", omitting the last two words.

> (If so, then I'd use the tone on the first word to mark illocutionary 
> force, and tone on the last, rising v falling, as the very abstract 
> categories 'incomplete/open' vs 'complete/closed' associated with rising 
> and falling intonation in natural language (according to Pierrehumbert 
> and her followers).)
> --And.