RE: gadri

[cbmvax!uunet!oasis.icl.co.uk!I.Alexander.bra0122]

lo terspuda be tu'a la djan. kau,n.

> > {le cukta} means
> > "the-thing-which-I-am-describing-as-a book", but with the
> > rider that I don't feel the need to be more specific, because
> > I expect you to know from the context which book I am talking
> > about.  This is an alternative way of referring to previously-
> > mentioned sumti, without always assigning a KOhA.
>
> Not necessarily.  When I talk of "le cukta", the context for figuring
> out which book may be extra-linguistic.  There is no reason to assume that
> I have necessarily mentioned this book before.  In fact, we recently
 introduced
> the particle "bi'u" to distinguish between old and new information:  "lebi'u
> cukta" is a newly mentioned book.

I intended the word "context" to be understood in the broadest
sense, including all the extra-linguistic factors.  Obviously
I didn't make this clear.

> > Note that this makes the specific/definite descriptions
> > ambiguous.  When I use {le}, I _am_ referring to something
> > specific, *but I'm not specifying it now*.  It is something
> > which has been specified earlier.
>
> Or not.  "le vi cukta" may be just "this book here in my hand" even if I
> have not >mentioned< the book before -- I still expect you to figure out
> from the >total situation< which book is meant.

OK, it is something which is specified by means other than
the words I am using _now_.

> > When I use {lo}, I am
> > almost certainly immediately going to start telling you
> > enough about it so that it becomes specific.
>
> Again, perhaps not.  I may simply not care about the specifics:
> > "lo remna cu xekri" means "some humans are black", without any
> intent to specify which.

OK, perhaps a put it a _little_ bit too strongly.  Still, I think
such statements are rare in isolation.  And you've already become
more specific than it was to start with.  If you followed it with
{le remna}, this would probably be understood as {ro lo remna poi xekri},
in other words "_these_ humans".  This leads us in to Bob's comments.

tu'a la lojbab.

> "lo ratcu" is not necessarily non-specific; it IS veridical - it claims
> that whatever is described REALLY IS a rat, and is not, say, merely being
> described as one for convention or convenience (which might be the case
> with "le").  It is the implicit quantifier on "lo" that makes it indefinite
> AND non-specific - the outer "su'o" means that ANY thing meetin the
> description will do.  If the outer quantifier is "ro", the result is quite
> definite: you are claiming about every single thing meeting the (possibly
> restricted by a relative clause) description.

Initially I was quite worried about this tie-up between specificity
(ugh - {ka satci}?) and quantifiers, but it's gradually beginning
to make more sense.  I _think_ you're saying that if you want to
_be_ (veridically) specific, then you have to talk about _all_ things
which satisfy the given conditions, which it is then up to you to make
as restrictive as necessary.  This is distinct from _claiming_
specificity, which you can do with {le}, at the expense of necessarily
relying on the audience's good will to interpret your incomplete
description.

>  Similarly, if you use a
> relative clause with "voi", then you remain veridical on the main
> description, but the restriction is to a definite/specific subset.

I'm not quite sure what point you're making here.  A {poi} restriction
would also be to a definite/specific subset.  But {lo remna voi ratcu}
allows you to talk about some (real) people, who you're _describing_
as "rats".

co'omi'e .i,n.