- From: pat hayes <phayes@ai.uwf.edu>
- Date: Tue, 26 Sep 2000 12:27:58 -0500
- To: www-rdf-logic@w3.org, "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
>In general I have a tough time with proposals that require the use of >mechanisms such as the ones that Pat is proposing. I wasnt aware that I had proposed any mechanisms. (I wish I was able to.) Seems to me that I was only pointing out some problems which are going to arise and making some tentative suggestions for how to approach them. We will have to find SOME way to deal with them (or legislate them out of existence? Unlikely.). Tim B-L has made some similar comments, I've since discovered (see http://www.w3.org/DesignIssues/Inconsistent.html) >I have a hard enough >time trying to figure out monotonic, certain representations without adding >certainty factors, probabilities, etc., etc., even for small examples. >For large examples, with many objects, assertions, classes, knowers, etc., >the situation becomes, to me, hopelessly complex. If the web seems "hopelessly complex" to you, Peter, then you maybe should be working on something else. But in any case, I wasn't advocating the use of certainty factors or probabilities; I was just giving an exposition of a basic idea about nomonotonicity. >Of course, it is possible that someone will present a logic that makes it >easy (easier?) to perform this kind of representation. But without such a >logic, and I mean a full logic, including semantics, deduction, and >algorithms, I don't see how one could in good faith propose these >mechanisms as a vehicle for representing information that will be acted on >by agents that do not have the level of ``common sense'' that we ascribe to >human beings. Im not sure that this kind of debate is useful, but let me just register some optimism in response to your pessimism. I think we are all agreed that we can't expect agents to just 'have' common (or any other kind of) sense, which is why we want a logic which can embody some of that stuff so that agents can transmit it and utilize it to a limited extent. Being a good oldfashioned logical type, I'm all for good oldfashioned logics (as 'full' as we can manage), but I think that we will need to modify our old ideas about semantics to accomodate to the web's messiness. In particular, we have to come to terms with the fact that logical names arent just logical constants any more (ie essentially skolemised existentials) but can cross-refer to other places over which we have no control (and may even get used in 'nonlogical' ways, eg as words in NL text.). We will have to be able to deal with the fact that inconsistencies will arise involving assertions from disparate sources, and the processes of resolving them may need to take into account the nature of these sources and some kind of notion of their warrantability. These are just facts about the web and the way it is already used in B2B and P2P applications which we need to come to terms with. So instead of complaining about how loose and ill-defined this stuff is, I think that people with a logical training have some new fields to conquer here, and I'd like to encourage us to get on with it rather than complain about a 'lack of logic'. Its our job to try to help get a suitable logic invented, right? Now, got any good ideas? 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
Received on Tuesday, 26 September 2000 13:25:35 UTC