Oh dear oh dear. Repeat this saying ten times before you go to bed
every night:
The meaning of a word is its use. (L. Wittgenstein)
This is where arguments about the meaning of meaning finally end up.
I think he was right, but no wonder the guy killed himself.
| Meaning/semantics is surely given by schemata that associate element
| types, etc with ontologies.
I agree in principle, but those ontologies at this point are, as far
as I know, completely dependent on natural languages for their own
meaning. It seems to me that the understanding of meaning by machines
can come only after we have agreed on a universal taxonomy of ideas
that transcends all the different assumptions that underly natural
language -- something like Roget's original system, but consisting of
language-independent primitives. I'm not holding my breath until we
see that. And in the absence of this semantic substrate, I think that
the only meaning we get is behavior and appearance, which the machines
can handle, and good old prose, which they can't.
Jon