Re: Ontology and Physics
> Rich,
>
> Those are fine things to read. They give some ideas
> of the latest theories about neurophysiology and how
> the brain works.
>
> But as any neurophysiologist will be the first to admit,
> the amount we know is enormously greater than the amount
> that is unknown. And for every question that is answered,
> hundreds of new questions arise.
>
> As I said to Avril, there is enough work to keep scientists
> busy for at least another few centuries.
>
> Bottom line: If you're looking for a solid foundation for
> ontology, don't hold your breath. The best thing to do is
> to keep all options open -- and that's exactly what the
> infinite lattice does: it provides room for the finitely
> many known theories while keeping open infinitely many options.
>
> John
Every infinite lattice starts with a very small lattice; possibly
the upper and lower bounds would be an example of an inductive
approach to the lattice.
Animals, including humans, have some built in functionality
that is very basic at birth. So their lattices must be very small
at that time, even if not mathematically minimal. As they age,
their lattices seem to get bigger due to experience.
As I remember Tom Mitchell's algorithms for learning, he used
each positive and negative example to expand the lattice he used
for distinguishing among situations. Though Mitchell's algorithms
were very pure mathematical implementations, something a bit
messier happens when animals, including humans, incrementally
expand their subjective lattices.
That's the part I find interesting. Rob mentioned that a dynamically
changing lattice, i.e. a learning process, is one way to view a more
natural ontology. One difficulty is that no two people actually
share the identical ontology as we've seen demonstrated repeatedly
on the SUO list. So a dynamically changing lattice seems like
one way to resolve the effect that one person's ideas have on
another person - but they each still have their own unique and
dynamically changing lattices.
In an earlier discussion, John, you mentioned that an object could
be viewed as a proces, and that there are certain researchers who
like the idea of every object being represented by a process. So
why not an ontology object? Thinking of an ontology as a process,
and formulating ways and means for growing lattices or resolving
negative experiences in lattices could be a very useful way to
think about them. Borrowing from Mitchell's early work might
be a good way to stimulate some new material in that area.
JMHO,
Rich