Re: CHAT: THEORY/CHAT: Re: Jackendoff's "Semantic (?) Structures"

Thank you to everybody having replied about
Jackendoff. This is really something you can discuss
about ! (and for a long time), and I'm glad to have
gathered other possible author and book references
(but for the moment, I'll try to concentrate on
Jackendoff).

I wondered about the exact difference between
semantics and cognition. I always thought that
semantics meant something like "the science of
meaning". You suggest that "for many linguists,
'semantics' means not our conceptual inventory but
only what is linguistically encodable. My question is:
what is encodable, and what is not ? Do you mean
"something that might be encodable, in theory", or
"something that is indeed encoded in at least on
natlang, and is attested" ? Where is the border
between encodable and not-encodable ?

Also, it seems to me (I have to read it again) that
Jackendoff clearly says that part of the meaning is to
be found in the lexicon (the lexical entries can be
decomposed semantically). So shall we call this
"encoding" ? There is no mark on the word "cat", for
example, indicating that this is a living thing, or an
animal, or whatever (not like adding an -s to "cat" to
indicate that there are several of them).

I think that, with some experience of the world, one
can try to build some theory of it. It being true, or
false, or approximative, is not really the point. The
point is: what kinds of things ("things" in the
meaning of concepts) could I possibly have to express
when talking with a human interlocutor ? Only then
comes the secund point: how shall I do it ? Clearly,
if some language is already given (and shared by my
interlocutor and I), then I have to squeeze my
concepts into the mould of syntax and lexicon, one or
another way, and I will more or less succeed (= be
understood the right way). This is very similar to
programming: you have some result to produce, but you
can use a hundred different tools (=languages), and
even using the same tool as your neighbour, you can
write your program a hundred different ways. The
result will be more or less correct, accurate,
complete, clean, elegant, tricky, maintainable, etc,
but probably never will two programers write exactly
the same program to answer the same problem (insofar
it's not a trivial one).

But the problem is different if you want to build your
own language, for ex. In that case you can decide that
there will be a special class for "things similar to
chewing-gum" for ex (in case you feel the need of it),
and even if no language on Earth ever did so. Or you
can decide that a word like "gwra" will mean "the way
a cat cleans its neck with its licked paw". The thing
is to find out what are the useful concepts (for your
purpose) and how to relate them together a consistent
way.

So, again: what is the difference between semantics
and cognition ? As I understand it, semantics refers
to one particular language (just like lexicon and
syntax), and cognition to the way every human's brain
works ? In that case, clearly my concern is cognition,
and not semantics. But I had not the feeling that it
is so when reading Wierzbicka's "Semantics - Primes
and Universals".

As to "smearing" and "loading", my opinion is that you
just can't compare them right if you stick to their
syntax, you have to understand deeply what each of
those both verbal concepts mean. Then you can ask
yourself "is it relevant to look here for a goal, or
for a patient, or whatever", and above all "is it
relevant to decide that the goal, or the patient, can
be retrieved by the same algorithm in both cases".
Syntax here is just blinding us. If you consider the
example on J. page 125 for ex:

a. Sue hit Fred.
b. The car hit the tree.
c. Pete hit the ball into the field.

first of all, if I translate them into French, I will
probably say:
a. Sue a frappe Fred.
b. La voiture a heurte l'arbre.
c. Pete a projete le ballon dans le champ.
so it's not the same verb any more ! What happened to
"the verb 'hit'" ? There are three of them !

then I disagree with reformulations like "What
happened to the tree was the car hit it": I don't care
so much about the tree, but rather about the car !
(especially if it's mine). The same with: "What
happened to the ball was Pete hit it into the field":
I won't consider just the fate of the ball, but the
result (ball in field: both terms are important; plus
'hit').

The same for "The sodium emitted electrons",
reformulated: "What the sodium did was emit
electrons". The sodium did not "do" anything at all,
there just was a natural phenomenon of emitting
electrons coming from a sodium source...

OK, that's enough for now, I guess. There will be much
more to discuss later on.

--- And Rosta <a.rosta@...> wrote:
[...]

=====
Philippe Caquant

"High thoughts must have high language." (Aristophanes, Frogs)


	
		
__________________________________
Do you Yahoo!?
Yahoo! Photos: High-quality 4x6 digital prints for 25¢
http://photos.yahoo.com/ph/print_splash

Reply