Re: TECH: specifity & definiteness

[Mr Andrew Rosta <ucleaar@UCL.AC.UK>]

> (This is John Cowan; snark is temporarily down.)
> la .and. cusku di'e
> > I quite agree: specificity is distinct from definiteness. Specific
> > referents may or may not be definite.
> I still don't understand this.  What is an an example of a specific
> indefinite reference (in English)?

"I bought a book". "A book" is specific, but indefinite, for I do
not presume that you can identify the referent. "I bought the
book" is specific and definite - I am presuming you can identify
the referent of _the book_.

> > Definiteness is non-truth-
> > conditional, so appropriate for a .UI cmavo,
>
> How's that again?  "I saw the man" entails "I saw a man", but not vice
> versa, so truth conditions are definitely affected.

This is +specific entailing -specific, not +definite entailing
-definite. Definiteness concerns whether the referent is identifiable
by the addressee.

At this point I ought to acknowledge the great variation of
uses of 'definite' in linguistics. I am using it in the sense
I've just given.

> > whereas specificity
> > affects truth conditions, so I would be inclined to treat le v.
> > lo as specific v. non-specific (though I do not understand what the
> > distinction between le/lo officially is).
>
> Again, with feeling:
>
> Remember that although we loosely refer to LE cmavo as articles, they are
> really more like pronoun+relatives:  "lo kanba" is not "a goat" but
> "that-which-really is-a-goat", because "kanba" is a stative verb, not a noun.

This is quite a good analysis of *English* articles! This is roughly the
way determiners work in English, too.

> le/lei/le'i descriptions need not describe accurately, and are implicitly
> restricted to those things the speaker has in mind.
>
> lo/loi/lo'i descriptions must describe accurately on pain of failing to refer,
> and are unrestricted.

*the light dawns on me* Okay: I do indeed see now that it is reasonable
to say that the _le_ series is +specific & the _lo_ series -specific.
Definiteness has nothing to do with it.

> > I don't, incidentally,
> > see that bihu/bihunai corresponds to definite/indefinite - or
> > rather, I do see that it doesn't.
>
> Correct.  It is a discourse marker.

Yes, but definiteness - "& incidentally the referent of this word
is identifiable by you" - ought also to be a discourse marker,
I feel.

> > > > Well, actually "zo'e" does well there, since "zo'e" and "le co'e" mean
>  much
> > > > the same thing.  Both of them refer to something specific-but-unspecifie
 d.
> > > > There is the difference that "le co'e" keeps the force of "le": one or
>  more
> > > > individuals, probably not a set or mass.
> > >
> > > I'm not sure I believe this.  I thought that {zo'e} was totally
> > > ambiguous, and could be specific or non-specific, universally
> > > or existentially or exact-numerically quantified, or any other
> > > (censored) thing.
> >
> > This is my understanding too.
>
> Yes, you are both right.  Nonetheless, pragmatically "zo'e" does work
> pretty well as a substitute for the non-quantificational sense of English
> "something".

You should give examples. _Something_ can be +spec or -spec, as, I suppose,
can _zo'e_. I agree that the +/-specific distinction is only relevant
in certain 'irrealis' contexts ("I want a book", "I tried to think
of something", "I will wed someone"). In realis contexts ("I wed
someone", "I had a book", "I managed to think of something") the
+spec & -spec readings are pragmatically though not truth-
conditionally equivalent.

------
mihe laho da. And da.