| |
| | SUNY Geneseo CSci 341 Logic Programming Part 2 Spring 2000 |
 | | This is the heart of resolution as a theorem-proving technique: If a series of resolutions simplifies a system of Horn clauses to nothing, it means that the system is inconsistent (i.e., to satisfy it, at least one term in an empty clause must be true). |
 | | So to "prove" a statement, phrase it's negation as a Horn clause, add it to a Horn clause system of "axioms", and run resolution until either it can no longer simplify the system (the theorem isn't provable), or it produces an empty clause (the theorem is true). |
 | | Typically the theorem to be proven is a single proposition, e.g., A, its negation is thus (not A), i.e., a headless Horn clause. |
| www.cs.geneseo.edu /~baldwin/csci341/spring2000/lp0424.html (418 words) |
|