On Dec 25, 2010, at 4:03 PM, Patrick Durusau wrote:
> John Sowa captures an unspoken part of the disagreement when he says:
>> But if you aren't using logic, you don't have anything that
>> can be combined with an inference engine.
> But the need to use an inferencing engine is surely a requirement for a
> particular context/project.
Sure -- a big part of the vision of the Semantic Web.
> I think part of the
problem is an admixture of marketing claims with
> statements about formal logic (in the sense ChrisM used the phrase),
> about which there is little disagreement.
> For example, I don't think there is much disagreement that there are any
> number of things called "ontologies" (by people other than ChrisM) that
> ChrisM and others would say violate the precepts of formal logic.
Let me remind you of what I actually said. I explicitly acknowledged that there are probably lots of ontologies written in languages that lack clear syntactic definition and a rigorous semantics. What I argued is that they will have limited use vis-á-vis the central goals of the Semantic Web. In particular, absent a clear logical foundation, ontologies cannot easily (if at all) be shared, integrated with other ontologies, or reasoned upon. If those are not your goals, then an ontology written in an ill-defined language that
lacks a clear semantics might still be useful. How useful they might be seems like a pretty interesting matter for empirical study and I'll not pretend to know the answer. My intuition is that they could at best be useful inside a fixed context in which it is possible to clarify intended meanings with face-to-face conversation and in which there is no need for any sort of robust automated reasoning support.
> The SOA Ontology for example.
Let's be clear that the problem I spoke of in that case wasn't the ontology's lack of clear logical foundations -- even if such foundations weren't explicitly given, there are plenty of frameworks to draw upon to flesh out the basic logical notions of class, subclass, an instance that the SOA ontology avails itself of. The problem there was an explicit, elementary modeling error -- the confusion of instance and subclass.
> Now, what ChrisM and others would like to say is that
> advantages if your ontology does not violate the precepts of formal
Well, I never spoke imperiously of "violations of the precepts of formal logic" (which seems to me to have been composed with the subtle sneer I was talking about). I have only ever spoken of the importance of using an ontology language with a clearly defined syntax and semantics -- the model for which is formal logic.
Be that as it may, imperiously delivered or not, bear in mind that it would be a logical howler to mistake this claim for its converse, namely, that there are no advantages to ontologies that lack clear syntactic and semantic foundations. Indeed, once again, I have explicitly acknowledged that such ontologies might be useful, albeit not in ways envisioned by the Semantic Web.
> But I hear (and perhaps they actually say it this way, would have to
> ask) them as saying: The only things that qualify
as ontologies are
> things that do not violate the precepts of formal logic.
> Patrick Durusau
> Chair, V1 - US TAG to JTC 1/SC 34
> Convener, JTC 1/SC 34/WG 3 (Topic Maps)
> Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300
> Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps)
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx