- From: <jos.deroo.jd@belgium.agfa.com>
- Date: Wed, 25 Jul 2001 01:41:10 +0100
- To: phayes@ai.uwf.edu
- Cc: Graham.Klyne@Baltimore.com, w3c-rdfcore-wg@w3.org
[...] > Right, though that "(in some sense)" deserves to be made more > explicit, which I will try to do before the f2f meeting. I mean, it > is an existential, strictly speaking, but queries are functionally > similar to *negated* assertions, so the machinery you need to handle > an existential in a query is the same as what you need to handle a > universal in an assertion. This is one of those things that just kind > of gets to be part of the atmosphere to anyone working in logic. In > most inference engines you can't tell asserted existentials from > queried universals, and vice versa. I'm really glad you express it so clear and I think you couldn't be more right :-) This is indeed related with Modus Tollens (MT) given P->Q and ~Q we can deduce ~P (Q like the Query and P like the Premiss an P becoming now a new given sub-query etc.) and ~forall is like forsome so indeed you can't tell asserted existentials from queried universals, and vice versa. and I still think this is kind of what also happens in the 'mini' world of anonymous node resolution (which is kind of generalized MT) ??? could it be that [ :p :o] (as one token) is an existental but that we *could* assert forall _:s such that _:s :p :o. ??? (we actually do use the former and then query with some kind of univeral/template (which can contain the 'normal' existentials)) -- Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
Received on Tuesday, 24 July 2001 19:42:15 UTC