Thursday, July 25, 2013

Truth tables for quantified statements

Today we had a very interesting discussion following our attempt to do problem number 10 on page 446 in Stan Baronett's Logic. In doing the proof it looked like we might have a typo in the conclusion. To check this and see if the argument was invalid (which it looked like considering the conclusion was a conjunction of Aa and not Ha with EG applied to it, but Aa would have to be entered into the proof via Addition. Only DeMorgan's would enable us to change it. But after multiple approaches none seemed to work. So we attempted to do a truth table to check it. It was at that point I realized I have never seen a truth table applied to quantified statements before - they are not truth functional? Here is one explanation I found: (from http://williamstarr.net/teaching/2310/10.20-4up.pdf) Truth tables work by explaining the truth of a complex formula in terms of the truth of its parts Example: :P is t i P is f The problem with using truth tables for quantifi ers is that the truth of quantifi ed formulas cannot be determined from the truth of its parts Example: 8x Cube(x) is t i ??? Cube(x) is t? f? Neither! Cube(x) isn't capable of truth or falsity, it's too incomplete! So, we can't use truth tables to explain what quantifi ed sentences mean But see the link and following discussion of Tarski's World. What helped was removing the quantifiers and doing the table for instances of the statements. Doing that it became obvious that the argument was invalid by that truth table.

No comments:

Post a Comment