WikiDiscuss

---

> Robin Lee Powell wrote:
> > Mark E. Shoulson wrote:
> > > Jorge "Llamb??as" wrote:
> > > >(Actually, there was at some point talk about an experimental
> > > >BAhE that meant "what follows continues the utterance of the
> > > >previous speaker, even though I'm a new speaker".)
> > >
> > > Why should that be a BAhE and not a UI?
> >
> > Because UI binds to the left, and one would want to use this word
> > to start a new sentence.

Mostly it would be used to *not* start a new sentence. I think the logic
of the situation requires this marker to attach to the word that follows.
It says "what comes next, starting from this word I'm marking, should
not be considered a new utterance". It doesn't really say anything
about the last word spoken by the previous speaker.

> If we needed such a thing (and IMO we don't; that's what leaving off
> ".i" is for) it should either be a member of I or just ".i se'i
> nai".

It can't be in "I"! The parse tree would come out all wrong. I suppose
{i si} could be used as a signal for the parser with that meaning.
As I said on irc, I think it makes more sense for si-clauses to
(invisibly, but still) attach to the following word rather than the
preceding one, so in that case an {i si} would behave very much
like BAhE.

Leaving off {i} is not enough because the default assumption should
be that a new speaker starts a new utterance. Assuming that a new
speaker always continues the other speaker's utterance does not
make much sense.

mu'o mi'e xorxes

__
Do you Yahoo!?
Check out the new Yahoo! Front Page.
www.yahoo.com