Re: [engelang] intensions & extensions (Xorban)

* Thursday, 2012-09-13 at 19:15 -0300 - Jorge Llambías <jjllambias@hidden.email>:

> On Thu, Sep 13, 2012 at 7:01 PM, John E. Clifford <kali9putra@hidden.email> wrote:
> >
> > I haven't worked through the threads that talk about f very thoroughly
> > yet, so I a unclear how it functions syntactically.
>
> It's very much like Lojban NU, except there's only one of it instead
> of all the different NUs.
>
> la fa se xrme bjre snva'aka
> A/state-of-affairs(A) in which (some E/horse(E): runs(E)): dream(me,A)
> I dream that some horse runs.
>
> se xrme la fa bjre snva'aka
> some E/horse(E): A/state-of-affairs(A) in which runs(E): dream(me,A)
> For some horse, I dream that it runs.
>
> "fa <formula>" is basically the same as "NU <bridi>".
>
> There is, of course, no special distinction made between fa predicates
> and other predicates. "fa se xrme bjre" is a one-place predicate just
> like "xrma".

Having looked a teensy bit more into situation semantics, it seems there
is currently a strand of thought that you should use situation variables
to do everything Davidsonian events and tense semantics did, and some
more.

I don't know if this exactly fits f-, but the translations above make it
sound like it does.

For example, here's a sentence I saw Kratzer highlighting as one which
basically forces you to introduce situation variables when developing
semantics for natural languages:

"Whenever it snowed anywhere round here, some local person dreamed that
it snowed more than it actually did, and that the local weather channel
erroneously reported that it had snowed less, but still more than in
reality.".

Could you translate that as something like
"ra fa [snows somewhere round here] se [local person e] si fi je [snows
more than snows in a] [local weather channel erroneously reports
snows less than snows in i but more than snows in a] [e dreams i]"?

(Not clear yet how to implement "amount it snows in a", but that's
a side-issue; similarly, it would be nice to be able to indicate that
the i situations are temporally linked to the a situations, and also to
put the whole thing in the past tense (which should end up being part of
the restriction on the universal quantifier))

(whether the above semi-translation would go through in lojban with {nu}
is another interesting question)

One bonus of working situations into the semantics would be that it
gives the parthood relation ("mnake", I think?) some new power, related
to And's suggestion elsewhere in this threadtangle that there should be
a "ca'a" version of f- which doesn't let you jump worlds: if, say, u'ue
is bound to the current world(/situation?), that could be
"je fa [blah] mnaku'ue".

Martin

P.S. the slides from these talks
http://www.diffusion.ens.fr/en/index.php?res=conf&idconf=853
talk a bit about situation variables