Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: ONT RE: Ontology case study




On 5/29/02 15:03, "Chris Partridge" <chris_partridge@csi.com> wrote:

> This means we need techniques to identify which inferences preserve enough
> accuracy to be workable. The way that this is done in operational systems is
> two-fold. One the requirements are clear the data is made sufficiently
> accurate for the specified process (and their inferences). That is why
> database people are always a bit wary of new processes ­ one of the checks
> they apply is around data quality. In a system with unrestricted inference ­
> the data has to be unrestrictedly accurate ­ i.e. exact.

Actually, I don't think this is right, Chris.  Let's say for example I have
a database with an "ontological" constraint:

  No physical object may simultaneously occupy two
  non-overlapping locations

One could argue that a database (and I have built such a database) with this
constraint demands too much precision.  I think this is confused for several
reasons:

  1) How could this constraint be violated?  Certainly not by any state of
     affairs in the world.  But it could easily be violated by some error of
     belief - an epistemological error.

  2) Taking the case of measurement, as in the example from Mike Uschold
     that you cite - these errors have to do with the act of measurement
     and the (measured) accumulated error of some dimension composed of
     parts with like measurement errors.  Thus, there is no cause to toss
     out the ontological baby with the bathwater just because the ontology
     isn't complete enough to take measurement into account.

  3) Pragmatically, if you can't detect conflict, then you live in ignorant
     bliss.  Again no reason not to try to detect the conflict in the first
     place.

> I also note, in passing, that FOL as it stands cannot do what Aristotle called
> Œpractical reasoning¹ ­ no amount of logical/inferential processing will
> result in an action. This is not a problem in database systems.

Huh?  I saw Chris Menzel's response to this, but really - how is a database
system supposed to be capable of "action" in a way that say, an FOL theorem
prover cannot be?

> This leads onto another problem with what this discussion has labelled
> Œontologies¹. The strong roots in predicate logic particularly FOL. As is well
> known (see e.g. p. 48 of Lowe¹s latest book) predicate logic was developed for
> mathematic applications and so is not well crafted for more mundane uses. For
> example, if you believe in a distinction between exemplification and
> attribution, this is not well marked ­ and, at the very least, the temporality
> of predication needs some explaining. Note that database systems typically
> have this distinction built into them ­ but in such a simplistic way that it
> cannot be practically used to reliably mark the distinction.

This is just false, Chris.  Databases are simply machines that implement the
relational algebra, which, when recursion/iteration is added, is equivalent
in expressive power to FOL.  What you can do in one you can do in the other.

 .bill