Re: More about scopes

[ucleaar <ucleaar@UCL.AC.UK>]

Jorge:
> > > > (4)     le ci nanmu cu bevri pa tanxe goi ko'a
> > > >         Each of the three men carries it, one box.
> > In this case, I don't think (4) shd  entail they all carried the same
> > box. Subsequent uses of koha will remain within the scope of le ci
> > nanmu, and it would not be a problem for there to be three boxes.
> Hmm... I'm not happy with that mainly because {pa tanxe goi ko'a} is
> assigning up to three referents to ko'a, which looks odd. The {goi}
> assignment becomes dependent on the whole context, rather than on the
> single sumti to which it attaches. In general, assigning more than
> one referent to ko'a, I think is to ask for trouble.

Obviously I see your qualms, but (a) the logic of the current system
supports my interpretation, and (b) occasionally one might actually
want to have a koha with reference varying in this way. Your qualms
I think can be generalized: users are liable to make many "errors"
as far as scope is concerned. Even in English, people do that (e.g.
using "Everyone didn't go" to mean "Not everyone went"). I think
we should accept that any grammar that can unambiguously encode
scope is going to sometimes make excessive demands of its users.

> A related question:
>        le ci nanmu cu prami ri
> Does that mean "each of the three men loves each of the three men", or
> "each of the three men loves himself"? What about with {vo'a} instead
> of {ri}?

This is an excellent question. I see no basis for {ri} and {voha}
behaving differently.
I note with satisfaction that Livagian uses different anaphors for the
two meanings. In Livagian they have the following logical form:

  [1] Ea, a is a set, 3 is cardinality of a; Ab, if b is a member of
  a then b is a man; **Ac, if c is a member of a then** b loves c.
      "Each of the three men loves each of the same three men"

  [2] Ea, a is a set, 3 is cardinality of a; Ab, if b is a member of
  a then b is a man, and b loves **b**.
      "Each of the three men loves himself"

The bit of logical form provided by each of the contrasting Livagian
anaphors is shown flanked by **. (For expository purposes I've
ignored the specificity of {le} in your example.)
[I mention this not to advertise Livagian, but to - I hope - clarify
the nature of the problem.]

I think it desirable to have both types of anaphor.

---
And