- From: Christopher Welty <welty@us.ibm.com>
- Date: Mon, 8 Mar 2004 07:53:26 -0500
- To: Pat Hayes <phayes@ihmc.us>
- Cc: public-swbp-wg@w3.org, public-swbp-wg-request@w3.org, schreiber@cs.vu.nl
- Message-ID: <OFA58E509A.21DEA86C-ON85256E4E.0062F32C-85256E51.0046CFF5@us.ibm.com>
[ Many people may not find this interesting. ] Pat, How many times do I have to win this argument? <pat> The idea that mereology is fundamental, as opposed to simply being a useful theory of parthood, is one example of a truly bad piece of ontological doctrine. (c.f , from http://esw.w3.org/topic/PartWhole :" The partOf relation is one of the basic structuring primitives of the universe" Er...nonsense. The relation of PartOf cannot be used to "organize the universe", which is why mereology never made it as a serious rival to set theory, in spite of Nelson Goodman's strenuous efforts; and probably why it plays no role in any of the sciences (Is the magnetism part of the magnet?). It also is, arguably, not even a very good model of human common-sense intuition, eg people are still arguing about some of Plato's examples). </pat> <chris> I am perfectly happy to tone down the dogmatism in the PartWhole pattern description. After all, *I* certainly do not find mereology to be a serious rival to set theory, I merely (heh) wanted to point out that some people do . Still, it is a basic and domain-independent relation, and characterizing how to represent it (especially the common variants and the most frequent pitfalls) I consider to be a very widely useful exercise. But I would appreciate hearing a more constructive re-wording of the background section. My goal was to point out that many people have thought about it for a long time, there are many carefully crafted axiomatizations (I find Achille's quite good), and then describe it in general. </chris> <pat> Another is the pernicious idea that Clear Thinkers *must* make some kind of basic ontological division of the universe into two disjoint categories of enduring things and dynamic processes (cf ontology.buffalo.edu/smith/articles/SNAP_SPAN.pdf ), </pat> <chris> Whoa! Hold on thar, Tex! Although I personally find this basic division useful, I certainly would not and do not espouse preaching from on high (i.e. putting it on a "best practices" web page) that clear thinkers must or must not do anything. The purpose of the Ontology patterns I would like to construct (like the PartWhole one) and make available through this WG are to help get people up to speed on how to use OWL to represent some fairly common things. They are advisory, though their association with "best practices" does imply they should be fairly good. Use them if they help. Provide your own, Pat, if you have rivals. </chris> <Pat> and another is the slightly barmy idea that modal reasoning is somehow connected with keeping your databases up-to-date. </pat> <chris> I feel pretty strongly that modal logic should have no part in any of the documents we produce. However, I see no problem in referencing a paper as background material that does, as long as it is relevant to the issue being described (or if I am one of the authors). That said, however, and between you and me I suppose, I will continue to maintain that modal axioms (not modal reasoning, you keep missing the point) are necessary to characterize database change. I thought at the last AAAI Spring Symposium you had finally understood the point, but you appear to have reverted to a previous state. But again, we can take this discussion off the SWBP group. Finally, I hope you will participate in the Ontology Patterns and Engineering task force of this working group. </chris> -Chris Dr. Christopher A. Welty, Knowledge Structures Group IBM Watson Research Center, 19 Skyline Dr., Hawthorne, NY 10532 USA Voice: +1 914.784.7055, IBM T/L: 863.7055, Fax: +1 914.784.7455 Email: welty@watson.ibm.com, Web: http://www.research.ibm.com/people/w/welty/ Pat Hayes <phayes@ihmc.us> Sent by: public-swbp-wg-request@w3.org 03/04/2004 01:37 PM To: schreiber@cs.vu.nl, public-swbp-wg@w3.org cc: Subject: Re: Tech Plenary: agenda Best Practices Sorry if this is too late for the Cannes discussion. My 'top 3' would be: 1. Tell people how to put RDF/RDFS/OWL onto (or attach it to) a web page so that it has some relevance to what is on their web page already (which is almost certainly largely HTML). Or at least give them an inkling of an idea how to do that and why it might be worth doing. In other words, take on the task of the public-rdf-in-xhtml-tf effort and get it done. 2. Get across the idea that the SW will work best when people use one another's concepts rather than invent their own, and tell people how to do that. The point of having a topic in a Web ontology is to make communication easier for agents. We should be thinking of ways to make this easier to do: right now there is very little support. Obviously there should be ways to find existing concepts and check them out (to find out if they are close you the intended meaning one has in mind, or can be tweaked so as to be) but we need also to deal with trust issues: like, to what extent am I making my ontology hostage to J's ontology if I use J's person-concept? Maybe (??) we need to think about a notion of 'meaning stability' analogous to the best-practices rules for keeping URIs stable. 3. (Hendler's #2 - explain the mess) ----------- 4. Finally, this is a negative suggestion, but I would oppose any attempt to tell the world how best to write ontologies; or if we cannot avoid doing that, then let the advice be severely pragmatic and free from philosophical punditry. There is a lingering (festering?) tendency among some folk to want to give instruction from on high to the great unwashed on how they should best think and express themselves. Unfortunately this advice is similar to most religious doctrine: most of the energy is spent in endless debates between rival doctrines, you can find some of it somewhere to justify almost any action you want to take anyway, and when the rubber meets the road most of it isn't really directly applicable in any case without an expert there to interpret it for you. The idea that mereology is fundamental, as opposed to simply being a useful theory of parthood, is one example of a truly bad piece of ontological doctrine. (c.f , from http://esw.w3.org/topic/PartWhole :" The partOf relation is one of the basic structuring primitives of the universe" Er...nonsense. The relation of PartOf cannot be used to "organize the universe", which is why mereology never made it as a serious rival to set theory, in spite of Nelson Goodman's strenuous efforts; and probably why it plays no role in any of the sciences (Is the magnetism part of the magnet?). It also is, arguably, not even a very good model of human common-sense intuition, eg people are still arguing about some of Plato's examples). Another is the pernicious idea that Clear Thinkers *must* make some kind of basic ontological division of the universe into two disjoint categories of enduring things and dynamic processes (cf ontology.buffalo.edu/smith/articles/SNAP_SPAN.pdf ), and another is the slightly barmy idea that modal reasoning is somehow connected with keeping your databases up-to-date. (By the way, it may be of interest to note that the first two of these both have their intellectual roots in the same strand of Polish philosophy from the late 1800s in Warsaw, for some reason. It is salutary to try reading what the founder, Brentano, actually said. But just because Brentano ( http://ontology.buffalo.edu/smith/articles/brentano/ http://www3.baylor.edu/~Scott_Moore/Contemp_Philosophy/Brentano.html http://grimpeur.tamu.edu/~colin/Phil251/lect2-brentano.html) was confused, there is no reason why the rest of us need to be, a century later) Most of our philosophical ontological ideas have never been seriously tested in the real world, and there are almost certainly real, hard-to-solve problems out there that we have never thought of before. If anything, now that we are asking the planet to do ontology, it might behoove us to listen and learn, rather than have the hubris to think we can instruct. Pat -- --------------------------------------------------------------------- IHMC (850)434 8903 or (650)494 3973 home 40 South Alcaniz St. (850)202 4416 office Pensacola (850)202 4440 fax FL 32501 (850)291 0667 cell phayes@ihmc.us http://www.ihmc.us/users/phayes
Received on Monday, 8 March 2004 07:54:10 UTC