[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
xorxes:
> la adam cusku di'e
>
> >If you are talking about in-mind sets which happen to be
> >singleton, but which in principle need not be, I think
> >that this is probably substratum influence on the way we
> >conceptualize 'le broda'. I think we should force ourselves
> >to use bare, unmarked 'le broda' even when that is plural,
> >and hopefully we will get used to evaluating it as possibly
> >plural. This would be made easier if 'le' ceased to be used
> >for inherently singleton sets (see below)
>
> The problem with that is that nonsingular {le} terms are harder
> to manipulate scopewise, so by allowing the possibility of
> plural with {le} we are actually complicating the understanding
> of the sentence quite a bit. It would be different if the default
> for singletons was either {lei broda} or {le'e broda}, because
> these two, being always singular terms, are easy to deal with for
> scope purposes even with an underlying non-singleton set. So if I
> were to abandon the "{le} is {le pa} in practice if not in theory"
> implication, I would have to switch 95% of its use to {lei} or
> {le'e}, because I want the flexibility of a singular term for the
> most common terms
Exactly. I'm delighted to be hearing this said loud and clear by
somebody other than me for a change!
> ><ultra-radical-proposal>
> >
> >What we need is a gadri for inherently singleton categories,
> >to take the burden off of 'le'. Unfortunately any cmavo
> >experimental in form would not be morphologically unmarked,
> >so that would not be a good solution for you. Therefore, I
> >(tentatively) propose that 'lau' could be used for this, since
> >no one uses it in its official meaning, and could be defined
> >thus:
> >
> >lau broda cu brode <-->
> >da zo'u
> > ge
> > ge da broda gi da brode
> > gi
> > ro de zo'u
> > go
> > de broda
> > gi
> > de du da
> >
> >i.e., Russell's iota operator or whatever it is called
> >
> ></ultra-radical-proposal>
>
> I wouldn't mind getting our hands on {lau}, {tau} and several
> others to put them to some use, but I'm not convinced this one
> would be useful. Referring to inherently singular sets with
> in-mind gadri is perfectly reasonable: if you know they are
> inherently singular, you cannot help its reference being specific
The problem is that there is a natural tendency (not in actual
Lojban use, but in language in general, and, I would say, in
open-minded Lojban use unfettered by convention) to infer the
following:
+specific = +specific, not -specific
-specific = -specific, not +specific
+distributive = +distributive, not +collective
etc.
In other words, using one value of an attribute that also
takes opposing values tends to imply that the attribute is
relevant and that the opposing value does not apply.
> The only reason to use {lau} would be when you additionally
> want to point out and emphasize this inherent uniqueness, and for
> that there is {lo pa}, {loi pa} or {lo'e pa}, any of which
> correspond to your definition for {lau}.
The case made for {tu'o} included:
1. We don't want to *claim* that lo'i broda cu pa mei. We want
to treat it as presupposition.
2. We simply don't want to have to arbitrarily choose among
redundant contrasts: e-gadri vs. o-gadri, V-gadri vs. Vi-gadri.
3. Given that one reason to mark singularity is to make the
meaning easier to process, it is counterproductive to have to
do this by adding an extra word (pa).
loi'e/lo'e satisfies (1-3).
The Adamic {lau}, which I will call {lau'au}, *is* supposed to
claim that lo'i broda cu pa mei, which is precisely what makes
it usefully distinct from loi'e/lo'e. But (2-3) still apply to
lau'au.
> If pointing out the
> singletonhood is important, it is proper to have to add the inner
> {pa}.
No, for reason 3 in particular.
> {lau} would make sense as a singular gadri in my opinion
> only if it was also +specific, so that you could also use it
> for singleton in-mind sets.
But le'e is completely satisfactory for this job, except perhaps
in its disyllabicity, which is really a symptom of a much more
pervasive problem.
> In that case it would take over the
> role {le} plays now, and {le} would be left for not necessarily
> singleton in-mind sets taken distributively, which is used
> relatively much less frequently than {le pa}, for the very good
> reason that it is much harder to process
Hear hear.
--And.