RE: "which?" (was: RE: [lojban] centripetality: subset vs component

["Jorge Llambias" <jjllambias@hotmail.com>]

la and cusku di'e

> > It is not {mo} that makes the difference. It is the article.
>
> Maybe the answer should just be {blabi}, then?

The wanted answer is exactly the same sentence with {mo}
replaced by an informative brivla. Of course it is possible
to rephrase the answer or use ellipsis, but the information
provided should at least cover what was asked.
{do viska le mo mlatu} requires an answer of the type
{mi viska le blabi mlatu}, and {le blabi mlatu} has
to be identificatory. Other valid answers to that question
should retain this property.

> The following
> exchange doesn't seem too unreasonable:
>
> A: le mlatu cu cliva
> B: le mlatu voi mo cu cliva
> ko'a voi mo mlatu cu cliva
> A: [insert appropriate answer]

If B did not identify the cat from A's initial claim,
the best request for clarification is {le ki'a mlatu}.
A's claim assumes that {le mlatu} is enough to identify
the cat.

> where B wants the nonveridical description of the cat to be
> elaborated, for whatever reason. At any rate, I can imagine
> a context where B might already know which cat A is talking
> about.

Then he is not asking "which?". Of course he may want to and
can ask for elaboration.

> > mi pu viska lo mlatu vi le panka
> > "I saw a cat in the park."
> > i lo mo mlatu
> > "What kind of cat?"
> > i lo blabi mlatu
> > "A white cat."
>
> I still can't see why this exchange would become silly if
> {lo} were changed to {le}.

It wouldn't be silly, but it would have a different meaning.

A: mi pu viska le mlatu vi le panka
"I saw the cat in the park."
B: i le mo mlatu [pu se viska do vi le panka]
"Which cat?"
A: i le blabi mlatu
"The white cat."

B could have used {ki'a} in this case. I think {le mo mlatu}
is also valid here, but it has to be taken as starting a
different reference than the one used by A first, which failed.
A's answer, on the other hand, is the same as B's reference,
and hopefully this time it succeeds.

> > I would say that is not the most important difference.
> > In {le mlatu cu mo}, the speaker has the cat identified
> > and asks for more information about that cat. They
> > already know which cat.
>
> Not necessarily. Pace the "in mind" characterization of {le},
> I think all it does is say there's a specific referent, but
> not necessarily one that the speaker has identified (in the
> sense of being able to point to, pick out of a line-up, etc.).

Ok, given that the sentence as a whole can't be evaluated
until we have a value for {mo}. But still {le mo mlatu}
is more forceful, because in {le mlatu cu mo} the sumti
is already complete. I don't know whether the rule should
be that complete sumti should not be evaluated until the
whole sentence is ready for evaluation. It seems more
natural to allow partial evaluations.

> For example, if A says to B {le mlatu cu mo}, then A may be
> able to identify the referent only as "that which B has in
> mind".

Yes, I agree. But it is more ambiguous. B might not be
certain whether A has identified the referent or is just
making reference to his reference. Both {le ki'a mlatu}
and {le mo mlatu} are safer bets for A.

> > In {le mo mlatu} the speaker is asking for information
> > that will make that sumti an appropriate reference, i.e.
> > they are asking for an answer that will allow them to
> > identify the cat, they are asking "which cat?".
>
> This seems more an assertion than an argument. I am unpersuaded.

I'm just using the definition of a question in Lojban.
The speaker asks the listener to fill the blank so as to
make a true statement. To make a true statement with {le}
requires successful identification.

> A: A certain cat leaves.
> B: A certain cat of what kind leaves?
> A: A certain cat of white colour leaves.
>
> -- what's wrong with that?

Nothing, but they are not {le} statements.

A: lo steci mlatu cu cliva
B: lo steci ke mo mlatu cu cliva
A: lo steci ke blabi mlatu cu cliva

A is not making a specific reference there.

> > {le mo} asks the speaker to replace {mo} in such a way that
> > the sentence becomes true. For the sentence to be true, it
> > is necessary that {le broda} be identified.
>
> (a) This is true of any question containing a specific reference,
> not just ones with {le mo} in.

Yes. But if {mo} is outside the scope of {le}, then it is
at least reasonable to expect the questioner to have already
make that one identification, isn't it?

> (b) {le broda}'s referent must be identified for the truth to
> be evaluated, but it needn't be identified by the questioner.

Maybe you're right, but I don't think we have debated this
before. My feeling is that there would at least be a strong
presumption that the speaker has already identified a
{le broda} sumti in a question. Otherwise asking any kind
of question becomes a pain if the listener needs not to
worry about any identifications that the questioner wants
to make.

> > What would you use as [sumti]? In {le mlatu du ma}, the
> > speaker already has to know which cat they mean.
>
> They don't have to.
>
> A: le mlatu cliva "The cat leaves"
> B: ma du le mlatu "Which cat?" ["which is the cat you were
> referring to?"]
> or:
> ri du ma

Maybe, but I think in the end this makes things harder for
the speaker. If he should not be expected to know which cat
when asking about {le mlatu}, how does he do when he does
want to ask about a specific cat?

co'o mi'e xorxes



________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com