singular/plural (was: Re: veridicality in English)

[And Rosta <a.rosta@UCLAN.AC.UK>]

Lojbab:
> >I don't accept these as counterexamples. "Veridical/nonveridical"
> >do not mean "true/false". They mean "asserted (by the speaker)
> >to be true/false".
>
> Please amend that.  How about "asserted (by the speaker) to be true/
> not-necessarily-true".  A non-veridical description is not necessarily a
> false one,merely a convenient one.

Yes: I erred, as others have pointed out. They mean "asserted/
not asserted to be true" (to use Jorge's correction).

> >> So you want to reintroduce grammatical number after all the work we went to
> >> to get rid of it? Even after the baseline? Aren't there cmavo for 'single'
> >> and 'multiple' to take care of this requirement?
> >
> >The baseline is irrelevant. This is purely a matter of usage,
> >not of design. Furthermore, I don't want to reintroduce
> >grammatical number. In contexts where the contrast between
> >singular and plural referents is logically irrelevant, I
> >am all in favour of letting the distinction be blurred, and I
> >recommend lei/loi for this. However, as previously established
> >on this list, there are plenty of contexts where the difference
> >between singular & plural referents makes a *logical* difference.
> >(E.g. {le} with plural ref is scope-sensitive; with singular
> >ref it is scope-insensensitive.) Hence a logical language *should*
> >distinguish at least between singulars and distributive plurals.
>
> Well, Lojban is more than just a logical language, and it strives to
> eliminate uneccessary assumptions.  singular/plural distinction is not
> necessary, given what you say, since scope-insensitivity is not harmed if
> you treat everything as scope sensitive.  Thus, in a logical language,
> if it is not absolutely clear that only singulars are being used, then
> the speaker *should* treat sumti as scope-sensitive, whether or not this
> is truly necessary.

It looks like we're not far from agreeing.

We agree that {le pa} and {lei} are scope-insensitive.

{pa lo} and {loi} are not sensitive to scope relative to other
existential quantifiers (I think), though they are sensitive
to scope relative to other logical elements. In other words
{pa lo} and {loi} are scope-sensitive because of their
existential quantifier.

{su`o re lo} is scope-sensitive both to ex. quantifiers (because
it is plural ({su`o re lo} = {ro lo su`o re lo})) and to other
elements (because of its ex. quantifier).

{le su`o re} (= {ro le su`o re} is scope-sensitive because it is a
plural.

Maybe it would be simpler to say that plurals introduce an
extra universal quantifier. Yes: that is a simpler way of
making the point. (You could see all sumti as having an
outermost universal quantifier. This would be vacuous when
the quantification is over a singleton set, as is the case
with singulars and collectives.)

Summarizing so far:

singulars:
    the outermost universal quantifier is vacuous
    the collective/distributive distinction is vacuous
collective plurals:
    the outermost universal quantifier is vacuous
distributive plurals:
    the outermost universal quantifier is not vacuous

If you know you want to refer to a collective plural, then
clearly you use {lei/loi}.

If you know you want to refer to a distributive plural, then
clearly you use {le/lo}.

If you know you want to refer collectively to a group of unknown
cardinality, then you use {lei/loi}.

If you know you want to refer distributively to members of
a group of unknown cardinality, then clearly you use {le/lo}.

Now, if you know you want to refer to a singular, then in
principle you can choose either {le/lo} or {lei/loi} - it
makes no difference, and people generally just opt for the
shorter one. That then starts to imply that {lei/loi} is
used for collective plurals.
I, however, advocate using {lei/loi}, because it saves the
hearer the redundant and pointless mental effort in processing the
appropriate scope that using {le/lo} would entail. Unless {le/lo}
are explicitly marked as singular  ({le pa}, {pa (lo)}), the
hearer must by default assume them to be plural.

(As I have said many times before, it would, for the reasons
I have given, have been nice to have shorter versions of
{le pa}, {le su`o re} and {su`o re}. That would encourage
people to encode number distinctions when they are known, and
thereby make communication easier.)

--And