Proving Invalidity in Predicate Logic – The Finite Universe Method In general, to show that an argument is invalid, we show that the premises can be true when the conclusion is false, that is, we show that the truth of the premises does not guarantee the truth of the conclusion. In predicate logic we do not have simple truth values for statements, so we do not have truth tables. So the standard truth table test, or indirect truth table test, from propositional logic, is unavailable. However, we can talk about the truth or falsity of a statement in predicate logic in regard to a universe. If that universe is THE UNIVERSE (1 in Boolean systems), then we assume that it contains an infinite number of things and, of course, we cannot list them. On the other hand, we can talk about the truth or falsity of particular statements relative to finite universes and we can use those to demonstrate invalidity. (Note: we will assume all universes have at least one member; the empty universe is of no interest here. Furthermore we will display universes as sets of names.) Handy definitions: Substitution instance: A substitution instance is the result of eliminating the quantifier from a quantificational proposition and substituting a name for every occurrence of the variable of quantification. Examples: (let a and b be names) 1. ~ [ Sa Wa ] is a substitution instance of (x) ~ [ Sx Wx ] 2. Wb • Bx] Bb is a substitution instance of 3. Note, however, that (x) Wx • substitution instances because this quantificational proposition; it is conjuncts (x) Wx and (x) [Bx • instances, however. (x) [ Wx • (x) [Bx • Tx] has no is not a a conjunction. Both Tx] have substitution Quantificational Equivalence in a Finite Universe: Given any predicate logic proposition, we can create its quantificational equivalent in any finite universe. The -2procedure is to form all of the substitution instances for every name in the universe of a quantificational proposition and link these substitution instances by conjunction, if the quantifier is universal, or disjunction, if the quantifier is existential. Other operators, in this procedure, stay the same. If there are names in the original, those names must be part of the finite universe. The result has no quantifiers and therefore can be treated like a proposition in propositional logic. Examples: Let the universe be {a, b} 1. The quantificational equivalent of (x) ~ [ Sx Wx ] is ~ [ Sa Wa ] • ~ [ Sb Wb ] 2. . The quantificational equivalent of (x) Wx • (x) [Bx • Tx] is: (Wa V Wb) • [(Ba • Ta) V (Bb • Tb)] 3. The quantificational equivalent of ~(Wa V Wb) ~(x) Wx is: 4. The quantificational equivalent of Ha ~(Wa V Wb). Ha ~(x) Wx is: Quantificational Validity: An argument is quantificationally valid if in every universe, when the quantificational equivalents of the premises are true, the quantificational equivalent of the conclusion is true as well. In a given universe, if an argument is such that when the quantificational equivalents of the premises are true, the quantificational equivalent of the conclusion is true as well, we can say that this argument is valid in that universe. So another way of defining quantificational validity is to say that an argument that is valid in all universes is quantificationally valid. FACT: if an argument is valid in a universe of size n, then it is valid in all universes smaller than n. Quantificational Invalidity: An argument is quantificationally invalid if there is a universe such that the quantificational equivalents of the premises can be true when the quantificational equivalent of the conclusion is false. Or, an argument is quantificationally invalid if the argument is invalid in some universe. 2 -3- FACT: If an argument is invalid in a universe of size n, then it is invalid in all universes larger than n. Proving Invalidity: The method consists in choosing the smallest universe that will make the premises true and the conclusion false, form the quantificational equivalents of the premises and conclusion, and then apply the indirect truth table test. The proof of invalidity requires a universe and a set of values. If an argument turns out to be valid in the universe you initially choose, then go to the next larger universe. Finally, it should be noted that if an argument contains names, these have to be in the universe and if there are existential quantifiers you usually need a universe with more than one name. But not always, the argument (x) Sx V (x) Hx (x) Sx can be shown invalid in a universe of one thing. Here is an example of the method: (x) Sx (x) [Sx Mx] Ma Let the universe = {a,b} The quantification equivalents of the premises and conclusion would be: Sa V Sb (Sa Ma) •(Sb Mb) Ma Here are values that make the premises true and the conclusion false: Sa F Sb T Ma F Mb T 3 -4Note: Since Ma, the conclusion, must be false, Sa must be false as well, or the second premise could not be true. But at least one substitution instance of Sx must be true in order to make the first premise true. Note that we need at least two things in the universe to make this work and one of them -- b -- is such that both Sb and Mb are true. This set of values makes the premises true and the conclusion false, and hence is a proof of invalidity. 4