[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
Here I respond to Nick & then proceed to try to set the debate in
order by untangling the different issues under debate and
indicating possible resolutions to them.
Nick:
> Guys, I give up. I thought about this really hard for three hours,
> and I just don't have that many spare hours. So I'm going to write
> here what I scribbled down and did work out; if it is helpful, fine
>
> I am now also finding that my eyes are glazing over when I read
> others' contributions on this topic, so I doubt I have anything left
> to contribute anyway
Let me repeat something I've just written in another message:
# There are 3 concurrent issues:
# 1. What does {lo'ei} (= {lo'e}-as-used-by-xorxes) mean?
# 2. What should {lo'e} mean?
# 3. How do we make logically-robust generic statements (about the
# typical X)?
# These issues need to be kept separate, conceptually even if not in
# different threads.
Almost all of the recent discussion has been devoted to (1), though
the "I like chocolate thread" dipped its toe into (3).
But you are worrying about (2) and (3). We can't answer (2) until
we've answered (3) (because either (3) will give us the answer to
(2), or (3) will show that a gadri is not the appropriate device
for making generic statements). And we've so far devoted very
little energy to answering (3). Certainly not nearly enough energy
for it to be time for people to throw up their hands in resignation.
> If this is a solved problem in formal semantics, can someone *please*
> find out? Does either pc or And have any formal semanticist friends
> they could ask? If this is a problem with several schools of thought,
> could we identify them and just pick one? I find it exceedingly hard
> to believe that we here now are breaking new ground that noone has
> worked on for the past 100 years. I'll go further. I find the fact
> that no jboskeist has ever gone out and referred to papers written by
> semanticists that might discuss these recurring issues to be
> incompetence. How can I defend the jboske enterprise to others (hi
> xod :-) as benefitting from 2 millenia of logic, when I don't see
> anyone trying to benefit from the last 50 years?
In fact, over the years when we have discussed (3), my contributions
have been informed by my reading on the topic of generics. Admittedly,
I haven't gone out and tried to locate formal semantics stuff on it,
but that's partly because our own discussions of (3) haven't yet got
to a sufficiently advanced or focused stage.
> So
>
> Let us assume that we know what lo'ei (Llamban lo'e, the intensional
> article in general) is
>
> Let us call my version of lo'e (typical intension) lo'e'au
>
> lo'ei cinfo is an intension, an abstraction. It is sense without
> denotation. It is "satisfied" by any instance of {lo cinfo}. (The
> notion of 'satisfying' is one that has me worried, and I'll come back
> to it.)
>
> lo'e'au cinfo is also an intension. But inasmuch as it is 'satisfied'
> --- as it has a corresponding extension --- that extension is not all
> lions, as with lo'ei cinfo. It is a subset: all typical lions
>
> I claim:
>
> lo'e'au cinfo =
> lo'ei fadni be lo'i cinfo
Not {ro fadni}?
I've snipped the rest of your long discussion. It is constructive,
but I think it conflates (2) & (3) excessively & so makes unduly
heavy-weather of the problem.
Sticking to (3), let's get the easy case out of the way: attributing
typical properties, such as living in Africa. For this, we can say,
"Every typical lion lives in Africa", "Every typical unicorn is white".
Okay, "typical lion" needs to be firmed up, but that's part of the
definition of the relevant brivla (fadni, or whatever), and I think
we're entitled to take it for granted that such a firming up can be
provided. The appropriate brivla can be chosen to express the mean
average broda and the modal average broda.
This leaves the other class of cases, "I like chocolate", "I tame
the lion", "This depicts a snake", "I need a box", etc. To me, these
seem much more heterogeneous, but I tentatively suggest that they
fall into two classes.
Class A: {lo} within the scope of an implicit element.
* "I tame the lion" = "I am trained for a situation where there is a
lion that I tame", or, more precisely, "co'e tu'o du'u there is
a lion that I tame".
* "I need a box" = "I need for it to be the case that there is a
box such that it co'e"
Note the difference between these two:
"**co'e** tu'o du'u mi **tinbygau** lo cinfo"
"mi **nitcu** tu'o du'u **co'e** lo tanxe"
It remains to be seen whether we can define gadri (such as {lo'ei})
that allow us to say one or both of these as
mi tinbygau lo'ei cinfo
mi nitcu lo'ei tanxe
Class B. The ones that don't fall under Class A. My feeling is that
what is going on with these is that we are referring to the type
rather than the tokens. Not the set, but the exemplar. It is as
if we abstract away from the differences among the tokens to end
up with a single instance.
* "I like chocolate" -- "I like the chocolate exemplar", "If you
abstract away from all different instances of chocolate, what you
end up with is liked by me"
* "This depicts a snake" -- "This depicts the snake exemplar",
"If you abstract away from all different instances of snakes,
what you end up with is depicted by this."
This is the meaning I intend for {loi'e} and {lei'e} to have:
mi nelci loi'e cakla
ti pixra loi'e since
I have carefully avoided discussing issue (2) -- the meaning of
{lo'e}, because that is premature. Once we've got issue (3)
sorted out -- and I hope that between us we are now getting a
fair way down that road -- then issue (2) should be fairly
straightforward (-- we can assign to {lo'e} whichever of the
explicitly articulated meanings best matches the official
documents).
--And.