Re: plural

[jorge@PHYAST.PITT.EDU]

la lojbab cusku di'e

> Well, I don't like your trnaslations to start with - try
> >        le re prenu cu citka lo plise
>          Each of the two persons eats some apple(s).
>
> >        lo plise cu se citka le re prenu
>          Some apple(s) are eaten by each of the two persons.
>
> I had always presumed that in the above sentences "some apples" was a fixed
> meaning reference - in the above case, pragmatics dictates that the two people
> eat different apples (Cowan points out proivately though that they could have
> eaten different pieces of the same apple), but that the order did not change
> which apple(s) are eaten by whom.

What you are saying is that the {ro} of {le re prenu} has scope over the
{su'o} of {lo plise}. What is the rule? Universal quantification always
first, unless different order is set in the prenex?

> The "da poi" equivalence changes this, since
> your second sentence means that the two people have to be eating the same
> apples, whereas the first sentence is not specific.  So we are forced by
> logical convention to have the conversion mean something that people are
> unlikely to want to say.

Why unlikely? Consider

        lo cukta cu se tcidu le ro prenu
        Some books were read by all the people.

Why would the meaning "All the people read books" be more likely to be the
one that is wanted?

> And using "loi" or "lei" doesn't help, apparently,
> because they also export to the prenex in some order dependent manner, since
> they are not quantified with "pa" or "ro".

Using {lei} would be very easy with default quantifier {piro}. Then order
becomes irrelevant for it.

(Outside quantification with pa does not give a singular term, in general.
What you need for a singular term is that outside quantification ro = pa)

> Cowan also told me privately in phone con tonight that "le" also is subject
> to thius.  Since the inside quantifier is not necessarily one, we must
> presume that distribution COULD take place  in any logical analysis, and thus
> an analysy that did not knwo the in-mind specific referent of le broda would
> have to assume that it is plural to avoid a possible logic error - in which
> case order affects the "le" sumti as well.  (Cowan will of course correct me
> if I misunderstood this.)

The difference is that to understand what the sentence is telling you, you
have to know what is the referent of {le broda}. Once you know that it is
a single referent, the order is not important.

If you are making a logical analysis then under some changes {ro le broda}
goes to {su'o le broda}. If it happens to be a single broda, then "each of
the single broda" says the same as "at least one of the single broda", and
there is no change in what the sentence says, even if superficially the
quantifier did change.

> So we end up with a situation where probably every sumti value is at least
> potentially order dependent, and for purpose of loguical manipulation must not
> be converted or otherwise rearranged.

Not indiscriminately, that's right. It can be rearranged by making the
appropriate changes in quantification.

> I see this as a major philosophical impact on the langauge design, though
> I might be hard pressed to find examples where it really makes a difference,
> especially when the writing was done by an English or other speaker used to
> the English order conventions.  I believe And Rosta and Colin Fine are most
> likely to have used FA in unconventional way that could be rendered incorrect.
> "se conversion" becomes more like a true passive.  AFterthought addition of
> sumti becomes probably invalid.

I don't think you'll find many incorrect uses, because overwhelmingly {le broda}
is a singular term, and it makes little difference where you use it. If the
sentence has one or more {lo broda}, and no {le su'ore broda}, order
is irrelevant.

Order only matters when you have universal quantification {le su'ore broda}
mixed with existential quantification {lo broda}. This doesn't happen very
often.

> Cowan mentioned tonight that one option might be some marker that says to
> consider all sumti in the sentence being spoken as prenexing in some
> arbitrary order (say the standard x1/x2/x3) even if they do not appear in
> that order).  He suggested an allolex of CU, but I think this is really
> metalinguistic discursive if anything is.

I think that would be really confusing, and I don't see the point of it.

> > I only said that singular terms (ie the only ones that
> >are simultaneously existentially and universally quantified) are much more
> >easy to manipulate. Of course the other ones are useful, too.
>
> But no Lojban sumti has these as the default quantification.

If {piro} was the quantifier of {lei}, it would.

And {le} isn't a singular term in principle, but yes for most practical
purposes.

> I'm not saying this change is impossible - I am just saying that it is a
> major change from the way >I< at least thought the language would work.

I don't understand why you say it's a change. The scopes of ro and su'o
are either in the order they appear (which is what I always understood) or
there is some other rule. If there is some other rule, what is it?

Jorge