On Tue, Feb 22, 2011 at 10:36 AM, Ian Johnson <blindbravado@gmail.com> wrote:Indeed.
> Except that domain of discourse in general is quite explicit in logical or
> mathematical contexts. That is to say, you walk into (for example) an
> undergrad real analysis class and it is usually very explicit that that "for
> all x" means "for all x in R". When it isn't, it's always specified in one
> way or another; the original statement of the theorem you're discussing may
> be talking about a domain that is a compact interval, and then the "for all
> x" quantifies over that interval, since you're handing x to f. In more
> general contexts, such as set theory, "for all x" quantifies over literally
> everything that the theory causes to exist, such as the entire collection of
> sets.
>
> These things don't bring up Lindar's issue of "which everything are we
> talking about?". If they did, mathematicians and logicians would be doing
> something else to alleviate it.
The domain of discourse takes twists and turns together with the
> By contrast, in a general conversation, that universe of discourse is
> perhaps not some absolute one, but it is a very very large one, because a
> conversation could take twists and turns very very easily by comparison with
> a real analysis class. Thus the universe of discourse has to be an
> appropriate size to compensate for that.
discourse. Just look at any conversation in any natlang and see how
often "everybody" is meant to include the speaker, George Washington,
everybody that will be born in the year 2017, Sherlock Holmes, Santa
Claus, and anything else that qualifies as "prenu". I would say not
very often. So whatever problems anyone has with "ro da" they will
find the same type of problems with "ro prenu".
We only need to wait and see what happens in practice. You can't have
> I really think it isn't so hard, if you want to be lazy, to simply use {ro
> srana} and {da poi srana} (I forget, would {lo srana} suffice?). If you need
> a binding, {ro da poi srana} or {ro srana goi ko'a} work too.
a rule that says that the domain of discourse must always include
Sherlock Holmes as one of the values that variables may take, and you
can't have a rule that says that Sherlock Holmes is a value that
variables may never take (at least I don't think either of those rules
can work in practice) so it's almost inevitable that you will need
context to determine what the universe of discourse is.
If "ro prenu" can, in some contexts, refer to the 6 or 7 billion
people alive that make up the human population of the Earth today,
then it can just as well refer, in a different context, to the 6 or 7
people in the room now. There are, in both cases, an infinite number
of potential values being left out that may turn up in another
context.