Re: Troponym Structures?
Hmm. I think I was too telegraphic with that question. More specifically:
>> Do you consider a class of [verb + object-type] (generalized verb
>> phrase) to
>> be a subclass of the verb?
>
>
> No, how can it be a subclass of verb? That is a category error.
> 'Verb' is a part of speech.
Yes, "Verb" is a part of speech, but when I wrote "the verb" I was referring
to some specific verb, e.g. "to cook", specifying some particular class of
actions. So, if the class "Cook" in the ontology labels a class of actions in
which food is prepared by a heating process, then the question is, is the phrase:
"Cook Meat"
. . . a subclass of those actions, distinguishable from another subclass:
"Cook Vegetables"?
Better examples of diverging semantics might be cases like "Play a game"
versus "Play a sonata".
I would distinguish the grammatical categories (verb, verb phrase) from the
specific classes of concepts ("Cook Meat") that could be labeled in language
by some combination of words that conform grammatically to that syntactic
category. I think that was your point, and I agree with that. I should have
elaborated.
What I am curious about is your comment:
> Moving to clauses can give an important boost to getting to
> something deeper, semantically if not ontologically. This
> lets you start getting to grips with the general structures
> that you might need, rather than those that happen to turn
> up as verbs in some language.
I agree that meanings that are logically full assertions associate with
phrases rather than individual words. I also think that the meanings of
individual nouns or noun phrases can be coherently specified in a logical
(ontological) structure. What I am curious about is how one would logically
specify the meanings of phrases (such as verb phrases) that are not just noun
phrases and not full sentences. I believe this is an issue that you have
studied for along time, and was wondering if your comment was related to that
question.
>
>> Or are the semantics (meaning in the one particular
>> sense) of a verb already implicitly determined by the class of things
>> it can
>> take as an object?
>
>
> The semantics should be an interpretation of the clause,
> possibly of a verbphrase. Given a particular semantic
> intent, one can select the linguistic units over which
> that semantic content will be distributed. If one has
> an activity type of meaning, then one of the possible
> ways of expressing this is across a clause, including
> selections across verbs, participants and additional
> bits and pieces as the grammar of a language supplies.
>
> A verb therefore may well have a partial/underspecified
> meaning contribution that already constrains possible
> participants,
That is what I has in mind.
But -- Does the comment above mean that you don't think that isolated verbs
can have a semantics of their own? I have thought of verbs as each being labels
for some class of actions, the whole class including all of the possible
realizations of the case fillers. That, is, the class (of processes, actions,
events) labeled by the verb alone would still be implicitly phrasal, having at
least some implied case filler. That is why I was wondering if specifying one
of the case fillers ("Cook Meat") would then constitute a subclass (which in
this case would still have the subject case-filler unspecified).
> I would not say that it is "determined
> by the class of things it can take as an object" as
> this seems to assume a priority of a particular kind
> of meaning that is unwarranted.
>
I am not sure what "priority" in this context this means, but it does open
paths to some interesting lines of thinking.
Thanks for the reply. If you have a pointer to some specific on-line
resource where these particular ideas are elaborated, please share it with us.
Pat
=============================================
Patrick Cassidy
MICRA, Inc. || (908) 561-3416
735 Belvidere Ave. || (908) 668-5252 (if no answer above)
Plainfield, NJ 07062-2054
internet: cassidy@micra.com
=============================================