RE: [lojban] "knowledge as to who saw who" readings

["And Rosta" <a.rosta@dtn.ntl.com>]

pc:
> a.rosta@dtn.ntl.com writes:
> It seems to me that {ce'u} ought to be analysable as an unevaluated
> {ma kau}, btw.
>
> I think there will be scope problems here. What does "unevaluated"
> mean in this context? As bound variables, they are all inherently
> unevaluated.

Under Jorge's mooted reinterpretation of the formal meaning of qkau
constructions, {du'u qkau} is the category of propositions in which
the qkau is replaced by every possible value. (I hope that's not
phrased too clumsily to be understood.) I then say that the difference
between {ma kau} and {ce'u} is that the latter is not replaced by
any value.

I was making an incidental observation, but am conscious of certain
implications, in that just as {ma kau} has an unreplaced counterpart,
{ce'u}, so other qkaus, e.g. {xo kau} should have an unreplaced
counterpart. E.g.

du'u <unreplaced-xo-kau-counterpart> prenu cu tavla
"the property of being the number of people talking"

du'u <unreplaced-xo-kau-counterpart> prenu cu viska ce'u
"the relation between x and y such that x is the quantity of
people that see y"

This would then raise the issue of How To Say It In Lojban, and
one particularly obvious way would be to have a counterpart of
{kau} that indicated nonreplacement -- call it {kau'u}. This
would then make {ce'u} merely an abbreviation for {ma kau'u},
and the above examples would be:

du'u xo kau'u prenu cu tavla

du'u xo kau'u prenu cu viska ma kau'u

> <I'm not sure what you mean by "adequate"; certainly we can't do
> without having a way to represent intensional forms of beliefs,
> but at the same time I think we can't do without having a way
> to represent extensional forms of beliefs, and I don't readily
> see the snag: what's wrong with saying "the truth conditions of
> p are blahblahblah and John believes p"?>
>
> OK, and that will be helpful for deciding whether John's belief is
> true or not, but does nothing to help the problem -- which I thouhgt
> was the one you were on -- of connecting what John believes with some
> other propositions that you can work on more easily. For this task,
> no amount of truth-condition information will complete the task, even
> if, ala xorxes, you know that both propositions are answers to the
> question.

I, perhaps-obtusely, truly don't see why it doesn't help the problem of
connecting what John believes with some other propositions that you can
work on more easily. "the truth conditions of p are blahblahblah and John
believes p" doesn't tell us the intensional form of what John believes
(i.e., p) but it does tell us the truthconditions of what John believes
(i.e. p). And this is exactly the outcome that (I think) I want.

> > But your approach still does not get over the extension-intension
> > gap. You just know extensional equivalence and that says nought
> > about intensional anything (well, if they are not extensionally
> > equivalent, they are not intensionally either).
>
> I agree (I think -- I can't ever be sure we understand one
> another right) that my approach says nought about intensional
> anything. But I don't see that as a problem.>
>
> Then I guess I don't understand what you are trying to do. Most of
> what I have seen from you looked to be trying to rewrite what John
> believes in terms of the extension of some property and John's
> beliefs about that extension, based on the extensional eqquivalence
> between the proposition which John believes and a certain proposition
> about the extension of the property. But those rewrites are not
> generally legitimate and in the cases given clearly are not.

I think you accurately summarize the situation. I want to be able to
say "John believes something that is extensionally equivalent to q"
(spelling q out as a bridi). I don't see where the supposed illegitimacy
is.

> <I don't understand everything you say, but I had taken it as one of
> the strengths of the xorxesian set-of-answers approach that it isn't
> intensional (or so I understood).>
>
> I don't know quite what "xorxes' system is not intensional" means.
> He has a set of answers better, a property "is an answer to..." and a
> number of propositions that meet that property), within that set the
> answers can be pretty much reduced to the model answers (just like
> the question with the kau words replaced) on extensional grounds, but
> the various answers, even all the true model ones, are still
> intensionally distinct.

I don't want to labour this, because under Jorge's latest revision of
his proposals I do think that the intensional form of each answer is
determinate.

But at some earlier stage of the analysis, lo du'u ma kau broda was
*anything* that was an answer to {ma broda}, regardless of the intensional form
of the answer. Crucially, if John knows every true answer, he
knows every extensionally-distinct answer but not necessarily every
intensionally-distinct answer.

But, as I say, this has changed, as Jorge's thought has shifted from a
relatively intuitive notion of 'answer' to a more explicit and formalizable
analysis.

> <So, to reply to what you say, I think the "and nobody else" is
> implied by an answer that is understood to be exhaustive.>
>
> That claim is vague enough to avoid any objection I could think of.
> On the other hand, taken in its simplest sense, it is often false.
> The eharer may take the speaker's stopping for a exhaustive
> completion, the speaker may only intend an exhausted run out of
> patient or memory. I would like a "and nobody else" explicit before
> I bet even the chicken coop.

I, on the other hand, think that there is a cultural (and perhaps
logical and linguistic) distinction between 'partial' answers and
'exhaustive' answers, and that interlocutors are usually griceanly
aware which sort of answer is relevant. In talking in Lojban about, say,
"knowledge as to who saw who", I would like to be able to distinguish
explicitly between knowing a partial answer and knowing the/an exhausative
answer.

--And.