ภาษาไทย
Best Method to Determine If a First Order Formula Is Logically Valid?

First order logic is undecidable, so there is (as you note) no mechanical method for determining first order validity. So what's the next best to a mechanical method for trying to establish (in)validity?The trouble with looking for natural deduction proofs is that the introduction rules allow you to deduce more and more complex wffs, and thus don't stop you spinning off down blind alleys (even when there is a proof to be discovered).Better, then, to use a method without introduction rules, where as you grow a proof the length of wffs at least doesn't grow and ideally decreases. Two options spring to mind. One is downward-branching tableau proofs as in Richard Jeffrey's lovely Formal Logic: Its Scope and Limits, or my Jeffrey-for-Dummies, officially titled An Introduction to Formal Logic. The other option is certain sequent systems. The tableau system in particular is very student-friendly, and with a bit of practice, on "nice" examples you'll usually find a closed tableau if there is one to be found, and find an open tableau and be able to read off a valuation which falsifies the wff being tested if the wff is invalid. This is perhaps your best buy!For example, let's use a tableau to test \$(forall x P(x) rightarrow forall x Q(x)) rightarrow forall x (P(x) rightarrow Q(x))\$. We start by assuming that's false (i.e. its negation is true), which comes to assuming\$\$forall x P(x) rightarrow forall x Q(x)\$\$

\$\$negforall x (P(x) rightarrow Q(x))\$\$The latter is equivalent to\$\$exists xneg (P(x) rightarrow Q(x))\$\$So, if that's true there must be an object in the domain which satisfies the condition, which we can dub \$a\$, so\$\$neg (P(a) rightarrow Q(a))\$\$whence \$\$P(a)\$\$

Copyright © 2020 - Guangdong Pinzheng Furniture Manufacturer Co.,Ltd. | Sitemap