RE: SUO: OpenCyc Motion Open for Discussions
<snip>
> 3. There has also been a great deal of ferment in
> linguistics about the inadequacy of any fixed number
> of word senses to capture the full range of meaning
> of any word. Alan Cruse, for example, has coined
> the term "microsense" to mean the particular sense
> of any word within a given context. Cruse maintains
> that there is a continuum of meanings of any word and
> that any particular list of "word senses" will always
> be incomplete. That is just one more way of repeating
> what Peirce, Wittgenstein, and Waismann were saying.
>
> 4. As an approach to accommodating such an open-ended
> range of meanings, I proposed the lattice of all possible
> theories in Ch. 7 of my KR book. I also have an article
> on the web that proposes that lattice as the conclusion:
<snip>
[ELP]
Hi again John;
Perhaps I shouldn't get concerned when folks bring words into a
discussion of ontologies. But my head keeps much clearer when I divide
the world distinctly into "language" and "meaning". Words, for me, go
in a lexicon, axioms encoding meaning go in an ontology, and semantic
lexicographers create bi-directional mappings between the two - probably
using lexical axioms to further constrain or compose ontological
definitions to construct lexical definitions. Perhaps your lexical
definition of "Cat" considers tails mandatory (ignoring dismemberment
etc.). You simply build your lexical definition for "Cat" out of the
ontology's DomesticCat and add a lexical axiom for requiring the tail.
We probably don't disagree here.
But isn't the proliferation of word senses and micro-senses an artifact
of the lexicon and somewhat orthogonal to the structure of the
ontologies in a lattice?
I'm not yet seeing why we need to beef up or structure the ontology so
as to directly include all this proliferation of word senses and
micro-senses. Human dictionaries, of course have the disadvantage of
being defined in terms of their own entries. But, I claim, that we
ontologist and semantic lexicographers are breaking out of that
boot-strapping, cyclicly defined world. Our lexical denotational
semantics are expressed in the language of the ontology. So aren't
denotational semantics usually expressed in a simpler language than the
one they are helping to define? Can't we lean toward a kind of
Jackendoffian simplicity in having a relatively simple set of
ontological building blocks from which we build lexical definitions of
all these micro-senses?
The micro-sense of "run" in the "horse" context doesn't need to define
any new ontological axioms that aren't already there in a good
"equestrian" microtheory. I claim that if we model horses well once in
the ontology, those axioms will suffice for all word micro-senses that
can be remotely applied to horses.
I think that we would agree that packages in Common Lisp or Java are
arbitrary bags of definitions and are used for software engineering
convenience. I would make a similar claim about lattices of
ontologies/microtheories. They are conveniences that need not strictly
be there. They make our axioms a little simpler and prettier by
allowing us not to have context arguments in our predicates. They group
related things. But they, of course, can also give us serious
headaches. I certainly find them useful.
But if Miriam Websters has survived all these years without
micro-senses, which suggests to me that an axiom-based lexicon and its
supporting ontology might do well without them for most purposes for a
while. And an axiom-based lexicon can craft its lexical definitions
using general ontological definitions where appropriate. These general
definitions can then be leveraged over all subsumed definitions.
Actually, if such general definitions don't subsume all micro-senses, do
those micro-senses even have the right to exist? Dictionary senses, I
claim, must be general enough to properly subsume all their
micro-senses.
For what it's worth,
-Eric Peterson