Forums before death by AOL, social media and spammers... "We can't have nice things"
|    sci.logic    |    Logic -- math, philosophy & computationa    |    262,912 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 262,431 of 262,912    |
|    olcott to Mikko    |
|    Re: Prolog formally resolves the Liar Pa    |
|    09 Jan 26 09:53:50    |
      XPost: comp.theory, sci.math, comp.lang.prolog       From: polcott333@gmail.com              On 1/9/2026 4:03 AM, Mikko wrote:       > On 09/01/2026 01:28, olcott wrote:       >       >> Non-programmers and non-Prolog programmers only       >> understand Occurs‑check failure as “Prolog doesn’t like it”.       >       > I don't know about non-programmers but everyone who knows enough about       > programming to be able to read the definition of the predicate       > unify_with_occurs_check/2 can understand that its failure means that       > the programmer does not like a cyclic structure at that point.       >              That is so stupidly wrong that it must be dishonest.              Conclusion: LP = not(true(LP)) is formally ill‑founded       in Prolog’s term model. Its structural expansion is       infinite, it violates the well‑foundedness requirement for       terms, and unify_with_occurs_check/2 correctly detects       this. Because the term cannot be well‑founded, it cannot       be assigned a truth value in any well‑founded interpretation.              --       Copyright 2026 Olcott |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca