RE: SUO: RE: More KIF-ified Ontology Content
>Hi Pat,
>
>I seem to have struck a raw nerve.
Nah, I just expressed myself a little too colorfully.
>I have just got back to the raft of messages my original email seems to have
>provoked.
>I have looked through them and I think that there may be some
>misunderstandings, which I will try to clarify.
>
>You wrote:
>There is an alternative conclusion, which is that alternative ways of
>thinking about a topic might all be equally valid and coherent, even
>though they differ from one another. The proposal to find a single
>coherent upper-level ontology then amounts to an insistence that all
>but one of these alternative ways of thinking are wrong. This is a
>kind of intellectual fascism which has never succeeded in the past
>several thousand years, and is unlikely to make progress now either.
>Maybe the SUO should focus on ways of allowing alternative
>conceptions of the world to co-exist, rather than trying to legislate
>which of them is 'right'. That approach would at least have the merit
>of providing a standard that more than a small fraction of the user
>base could use without discomfort.
>
>I do not see how 'The proposal to find a single coherent upper-level
>ontology then amounts to an insistence that all but one of these alternative
>ways of thinking are wrong.' Any more than arguing that a law that says we
>drive on the right side of the road, says that driving on the left side is
>inherently wrong.
"wrong" was too simplistic. I didnt mean inherently wrong, I meant
more like 'one of these sides will be chosen by fiat, and the others
thereby forbidden.' Which wouldnt matter if the choice were
arbitrary, but it isn't. So a better analogy might be a law which
insisted that one drive only at a precise speed while holding the
wheel with ones right hand, or some such.
>It seems to me that there are a number of possible different enterprises
>here, let me characterise them as:
>- philosophical,
>- scientific/engineering, and
>- everyday.
>
>It is true that the goal of some attempts in philosophy is to arrive at a
>'right' answer - and that they have failed. Indeed some philosophers think
>in terms of the current best effort (e.g. David Lewis justifying his views
>on possible worlds.
>
>The history of science sets us a different example. It seems that what
>happens (according to Kuhn et al) that the development of science requires
>the choice of a framework (what he calls a paradigm). An example might be
>Newton's choice of absolute space. These frameworks are essential for
>progress - if the lessons of the last few thousand years - and particularly
>the last five hundred years - are anything to go by.
>
>As these historians and philosophers of science have pointed out, choosing
>the framework is as much a political as a rational process, and involves
>persuading the community that the framework is the 'right' one. Typically it
>involves a choice between several competing theories. Furthermore this is
>only ever successful if the framework is sufficiently good (where this is
>characterised in a number of ways). I'm not sure how this qualifies as
>intellectual fascism.
That doesnt, but (1) scientific theories are not chosen by standards
organizations (fortunately), and if they were then it could be so
described, so your metaphor here is misleading; but more to the point
(2) there in fact isnt a single *ontological* framework in science,
and it couldnt possibly be done if there were. Scientists re-think
their ontologies constantly, without changing their basic Kuhnian
paradigm. Interestingly enough, the RKF effort is working through
some genuine sceince in the form of a biology textbook, and one finds
alternative (and conflicting) ontological presumptions all the time,
sometimes in the same sentence. We all know what DNA actually is; but
should we think of (formalize) it as a tube, a spiral, a molecule, a
linked structure of atoms, a series of pairs, a pair of sequences, or
as a genetic 'sentence' in base-pair code? The text dances between
these alternative views without even noticing the ontological shifts
involved, since human beings can make those ontological shifts
without effort; but ontologies can't. (Well, not with current
technology; but to make this possible would be a significant advance
in AI, not just a standard ontology. The RKF effort might finish up
using explicit metaphorical mappings, for example, to allow several
logically incompatible ontologies to co-exist.)
>If we look at everyday uses of the concepts dealt with by philosophy and
>science, then here there is a plethora of theories - and probably always
>will be. So adopting a 'standard' in one field of work does not eliminate
>variety in others.
>
>You wrote (in a subsequent email):
>'However, I honestly cannot reconcile what you say with what Chris says.
>What is "the ontological paradigm" if it is not the One True Ontology?'
>
>An ontological paradigm would be the one accepted ontology - but like all
>science subject to both evolution and revolution.
>
>My concern is that the work involved in producing a single coherent general
>theory is enormous. I suspect that this is way beyond the resources of the
>SUO (hence my vote at set up was ABSTAIN). If no real effort is made to
>produce a high level then the work is surely doomed.
>
>My guess is that science's method of working (with a single framework) is
>because the cost of working with a variety of frameworks is astronomic. As
>philosophy has shown the high level 'concepts' are inter-related and working
>out these for each of the choices on each of the topics is a really
>substantial piece of work - especially given that there are not really any
>coherent theories covering the full range.
This sounds to me like you are thinking of the alternatives as all
fitting into some kind of super-heirarchy. As I tried to argue in a
previous message, they don't really do so ('really' because they
could be forced into one, but its upper levels would then just be a
kind of philosophical catalog of alternative metaphysical stances,
rather than any kind of unified upper ontology in the sense I think
y'all have in mind.)
>My concern is very different from yours - and it focusses on a different
>domain/purpose. Enterprises are finding that they need to link together
>systems, and the biggest problem they are having is at the semantic level
>(Mike Uschold pointed this out in an earlier email). They are going to
>develop something to do this - unless someone introduces them to the work
>done by philosophers et al. (and the fact that there are a range of
>choices - and how these choices interlink) we are going to end up with a
>standard that is unaware of the substantial issues and alternatives.
I understand your pragmatic point, and even sympathise with it. What
worries me is that some of the pragmatic enthusiasts of
standardization have a very oversimplified vision of what the
problems are, and that they therefore are being led (or allowed) to
expect a lot more from the SUO than it can possibly deliver to them.
If someone thinks of this as being like the choice of a standard
book-cataloging scheme, or a standard math notation, or a standard
paper size, then they really havnt understood what ontologies *are*.
Pat Hayes
---------------------------------------------------------------------
IHMC (850)434 8903 home
40 South Alcaniz St. (850)202 4416 office
Pensacola, FL 32501 (850)202 4440 fax
phayes@ai.uwf.edu
http://www.coginst.uwf.edu/~phayes