Re: (1) loi; (2) le v. la

[Logical Language Group <lojbab>]

la .and. cusku di'e

> (1) It is fairly clear that I misunderstood loi.
> I think it is better if we don't have masses automatically
> inheriting all the properties of their constituents. That
> way we if my mother weighs 1 bunda, and my father weighs 2, then:
>   ci lo bunda loi mi rirni cu junta
> But "re lo bunda" would be false.

Beware of constructs like "ci lo bunda"; they mean "three things which
weigh a pound each".  The right way is to use "bunda" as the main selbri, thus:

	lo junta be piro loi mi rirni cu bunda li ci
	The weight of the-whole-of the-mass-of my parents is-in-pounds the-number 3

Non-pedantically, we can use the object itself instead of its weight:

	piro loi mi rirni cu bunda li ci
	The-whole-of the-mass-of my parents is-in-pounds the-number 3.

This isn't relevant to your point, but it's important to understand how
to manipulate the Lojban measurement words, as they are very non-English.

> I wonder if a lohe-like cmavo is warranted for the purpose I
> had thought "loi" served: that is the view that every instance
> of some category is the same instance. This really would have
> every property of every instance of the category. So both
>   ci lo bunda *XOI* mi rirni cu junta
> and
>   re lo bunda *XOI* mi rirni cu junta
>   pa lo bunda *XOI* mi rirni cu junta
> would be true.

I think that "loi" will work here.  Note that I have added "piro" to your
examples above; if we don't assume property inheritance, then it is 
{piro loi} (the whole of the mass) that weighs 3 pounds.  But some (-specific)
part of the mass, which is {pisu'o loi} or just {loi} can weigh 2 pounds or
1 pound or anything in between, since the -specific reference means simply
"There exists a portion of the mass of my parents which weighs N pounds."

> (2) John has pointed out that "la" is +specific. In this case,
> "la gerku" means "a certain entity that I'm calling 'gerku'".
> This seems exactly like "le gerku", which means "a certain
> entity that I'm calling 'gerku'".

Well, yes.  However, the LA cmavo indicate that the association between the
name and the thing is purely conventional.  There is a difference between
naming your child "Happy Monarch" and describing your child as a happy
monarch, although in neither case is it required that your child
actually be a happy monarch.

> Does "la" allow for more than one non-cmevla following it? That
> is, does it behave like a normal gadri, terminable in "ku", etc.?

Yes.

> If so, "LA" seems to subsume "LE".

In the sense that LA can participate in the same grammatical constructions
as LE, yes.  LA can also participate in a grammatical construction peculiar
to it (and DOI), viz. being followed by one or more cmevla without a separating
pause.  We keep to the minimum number the words which can do this, to minimize
restrictions on cmevla.  (It's hard enough to respect the restrictions
we've got:  Mark Shoulson is the Official Bad-Cmene Catcher of Lojban List,
and he catches plenty.)

In the sense that every referent of "la broda" is a referent of "le broda"
also, definitely not.  Greg Bear is or might be "la cribe", but to call him
"le cribe", though not downright incorrect, is at least peculiar.

> I think it is good that "la" is +specific, so if LA does indeed
> subsume LE, I suggest changing this, by making LE +veridical
> (LA is of course -veridical - "la gerku" needn't refer to a dog).

Would almost never refer to a dog, except for those who like to call their
dogs "Dog" and their cats "Cat", a nonzero but small class.

LA with following brivla is primarily for name matching with cultures where
meaning counts for more than sound, e.g. American Indians.  Crazy Horse
would be "la fenki xirma".

> "Le gerku" would therefore refer to a certain thing that really
> is a dog. At present it could refer to a banana.

Semantically, yes.  Pragmatically, not likely.

> The resulting paradigm would be:
>    lo +veridical -specific
>    le +veridical +specific
>    la -veridical +specific

I still think the current layout is superior.  As Jorge and others have
said, +specific really removes the need for +veridical.  If the speaker is the
standard of reference, why appeal to a separate standard of truth?

-- 
John Cowan		sharing account <lojbab@access.digex.net> for now
		e'osai ko sarji la lojban.