Slides / Bullets
- TT-consequence and logical consequence, redux
- Why does TT-consequence suffice for logical consequence?
- Suppose that Q is not a logical consequence of {P1,...,Pn}. Then there is some possible situation in which P1...Pn are true and Q is false.
- But every possible situation corresponds to some line in the joint truth-table
- Informal proofs and the Boolean connectives
- Cardinal rule: in an informal proof, you’re allowed to make any inference whose validity you can legitimately assume to be obvious to your audience.
- Some very obviously valid inferences
- Conjunction introduction: from any two premises P, Q, you may infer the conclusion P ∧ Q.
- Conjunction elimination: from the single premise
P ∧ Q, you may infer either P or Q.
- Disjunction introduction: from any premise P, you may infer P ∨ Q for any Q.
- Other useful valid inferences
- Important tautological equivalences, such as
- De Morgan’s Laws:
- ¬(P ∧ Q) ⇔ ¬P ∨ ¬Q
- ¬(P ∨ Q) ⇔ ¬P ∧ ¬Q
- Idempotence:
- Distribution:
- P ∧ (Q ∨ R) ⇔ (P ∧ Q) ∨ (P ∧ R)
- P ∨ (Q ∧ R) ⇔ (P ∨ Q) ∧ (P ∨ R)
-
- Important tautologies, such as P ∨ ¬P
- These can be introduced into your informal proofs at any time, since a logical truth follows from everything.
- Indirect proof: proof by cases
- Suppose we want to prove ¬(a = b), given the premise (Cube(a) ∧ Tet(b)) ∨ (Tet(a) ∧ Dodec(b)). How do we do it?
- Proof: Suppose Cube(a) ∧ Tet(b). Then it follows that ¬(a = b), since nothing can be both a cube and a tetrahedron. Suppose on the other hand that Tet(a) ∧ Dodec(b). Then again, it follows that ¬(a = b), since nothing can be both a tetrahedron and a dodecahedron. So in either case, ¬(a = b).
-
- What we’re relying on here is the following fact: if a sentence follows from P together with certain other premises, and the same sentence follows from Q together with those premises, then it follows from P ∨ Q and those same premises.
- When we say ‘Suppose...’, we’re beginning a subproof.
- An indirect proof is a proof that uses subproofs.