Re: GLI Re: Indirect questions

[And Rosta <a.rosta@UCLAN.AC.UK>]

Jorge:
> >But my main objection is that the "inherent meaning" is determinate
> >and the "noninherent meaning" is indeterminate.
>
> But what's indeterminate about the meaning under discussion?
> Given a context in which the utterance is produced, it will either
> be the case that she is sad, or it will be the case that she isn't.
> (Or some other value in the middle.)

The determinate meaning can be established independently of
context. Except for deictics.

The step you then want to take involves establishing which world
the proposition is asserted to be true of. That's where the
indeterminacy comes in.

> Now, given that context,
> it is perfectly determinate what is {le sedu'u xukau ko'a badri}.
> Any {lu ... li'u} that serves to report the fact will do.

First, *once we have overcome the indeterminacy of which world
the proposition is claimed to be true of*, it is indeed determinate
that if {ko`a badri} then {djuno lo du`u xu kau ko`a badri}
entails {djuno lo du`u ko`a badri}.

Second, my reservations about {le se du`u xu kau} remain. Though
see below.

> >In English one often gets exchanges like:
> >
> >  A: What are you complaining about? You *asked* me to open the
> >     window.
> >  B: No I didn't. I just said it was warm in here.
> >
> >  A: I've known women be better at maths than men.
> >  B: I strongly disagree that women are better at maths than men.
> >
> >- based on misunderstandings, which in turn arise from an erroneous
> >inference of noninherent meaning from inherent meaning.
>
> All right, that's not what I had in mind. I meant unequivocal meanings.
> Noninherent only in as much as they depend on the context. Somewhat
> like deictics.

Not like deictics. It is possible to specify the meaning of
deictics exactly, but involving variables that are to be
bound by properties of the utterance. For nondeictic terms
the meaning can be specified completely, without any
context-dependency.

Your case is less like deixis and more like "She painted the
house his favourite colour". If this is true of world W, and
in world W his favourite colour is blue, then it is entailed
that she painted the house blue.

> >Lojban's
> >cultural literalism is well-advised, and I think the distinction
> >should be carried over to the meaning of {la`e}.
>
> I agree. I'm not suggesting a deviation from literalism here.

OK. But I think your position entails that la`e should include
all entailments, given that it is established which world the
utterance is talking about. So I think you are in effect
arguing that {la`e lu she painted the house his favourite
colour} can be the proposition "She painted the house blue".
I'm unhappy with that.

> >> To those worried about the horribly arcane nature of this discussion,
> >> we are trying to decide whether {le sedu'u xukau ko'a badri} makes
> >> sense, as in {mi cusku le sedu'u xukau ko'a badri}, which to me
> >> means "I say whether she is sad".
> >
> >My position is that I don't see how it makes sense if we simply
> >extrapolate from known cases. But that does not rule out declaring
> >this a valid usage.
>
> I understand {mi cusku le sedu'u ...} to mean what {mi cusku le du'u ...}
> would mean if cusku meant  "x1 says proposition x2". Am I committing
> some blunder that I don't see here?

Hardly a blunder, even if I am right and you are wrong.

I think it has been established that the effect of Q-kau on the
logicosemantic structure varies according to the context of
Q-kau within the sentence. That is, the meaning varies according
to whether Q-kau occurs within a sumti of djuno, of preti, of
cucli, of "discuss" (I forget the gismu - it starts with a C
I think), or of frica. (I would LOVE to be proved wrong on this,
mind you.)

There are certain contexts where Q-kau just doesn't make any sense
(e.g. if it occured within a sumti of the majority of gismu).

{cusku le se du`u xu kau} is one of those contexts where I
hadn't been able to get it to make any sense, though I can accept
that by stipulated convention it should mean "say whether".

However, I think I now find myself able to rationalize such a
convention. {cusku le se du`u xu kau Y} would mean "utter a
piece of text that says whether Y is true". Logically, that would
be:

 utter a piece of text, t, such that for every x, a seljetlai of Y,
   t expresses (a truth-conditional equivalent of) the proposition
   that x is seljetlai of Y.

whereas {cusku le se du`u Y} would be

 utter a piece of text, t, such that t expresses (a
   truth-conditional equivalent of) Y.

So, in conclusion, I think that {cusku le se du`u xu kau Y} can
coherently be ruled to mean "say whether", but that without
further investigation, it does not seem as if there is any
regularity to the mapping from syntax to semantics. (I hope
that a discovery of the regularity is waiting over the horizon.)

> ...
> >The price of having the unexpanded forms is that they don't
> >all expand in the same way. If we required all expansions to
> >be automatic (i.e. insensitive to lexical semantics) then
> >either {djuno} or {kucli} could not have their current meaning.
>
> Yes, it would seem that you're right.
>
> co'o mi'e xorxes

--And