- From: Seth Russell <seth@robustai.net>
- Date: Wed, 29 Nov 2000 12:39:05 -0800
- To: "Sean B. Palmer" <sean@mysterylights.com>
- CC: www-rdf-logic@w3.org, timbl@w3.org
"Sean B. Palmer" wrote: > Summary of this message: I truly believe that there must be a way of > schematically asserting a 'not'! Let me make another take on this. There probably is not any way to express the essence of "not" in a purely declarative way. Were I a professor of logic I would set out to find that assertion proved in the literature, and failing that I would attempt to prove it myself. But we can easily make "not" mean something in a running system. For example, let's say that our system believes the purely declarative representation: [A, not, [s1, isLeftOf, B]]. Now that just sits there representing our fact exactly and explicitly; but doesn't mean shit. Until perhaps we write the "schema" for "not" and then instigate a running process: The schema: [not, Boolean, notBoolianProgram] [Boolean, rdf:type, method] [notBooleanProgram, rdf_1, firstAction] [notBooleanProgram, rdf_1, secondAction] .... for however many actions are necessary So now I can write the sequence of actions hanging off the notBooleanProgram such that the running process will take a [False] attitude towards any statement it encounters or implies that looks like [A, isLeftOf, B]. In other words if the active process's attention is on [A, isLeftOf, B] the process will ~feel~ false. I'm being a bit cute here, really all I'm saying is that there will be a Boolean variable in the process object and the value of that variable will be "0" when the process's attention pointer is on the proposition above. Does that make any sense ? Seth Russell
Received on Wednesday, 29 November 2000 15:36:24 UTC