Alternatively, several types of identity, parthood, etc. could be
well defined and well documented in the upper ontology and
different ones could be selected as appropriate in lower ontologies
and KBs. These different types could be further specified in the
appropriate microtheories.
I was taught an effective theory of identity, with two cornerstones, and one that was quite widespread among logical theorists, good enough for practidcal work in science and ontology. It is a thread that runs through many philosophers and logicians, for millennia, to the present day, I thought, as follows:
x is the same as y
is not meaningful in itself, in that it lacks the context of a qualifier T, where T is a type of thing, so that the full _expression_ is
x is the same T as y.
This is of course the theory of relative identity. I myself think the view is fatally flawed as a comprehensive account of identity (notably, it seems to presuppose absolute identity: x is the same T as y if and only if there are F and G such that F is a T and G is a T and F = G). Problems and virtues of the view aside, however, it is much too strong to say that absolute identity "is not meaningful". It may have its quirks and limitations, but meaningless it is not; we can axiomatize it, provide a clear extensional semantics for it, etc.
chris
