Summary Foundations of Logic 2 (Modal logic) (c) 2001-06-25 Fabian M. Suchanek http://suchanek.name/texts/summaries/fol2.txt This is a summary of the Modal Logic lecture held by Prof. Wolfgang Lenzen in the WS 2000 at the University of Osnabrueck. By reading the following text, you accept that the author does not accept any responsibility for the correctness or completeness of this text. If you have any corrections or remarks, please send me a mail. This is the only way to make the publication of this summary useful for me, too. My e-mail address is f.m.suchanek@zweb.de, but the letter 'z' has to be removed from the address. Types of truth: * ordinary truth (contingent facts) * necessary truth (true in all possible worlds) * logical truth (tautologies) * analytic truth ("Bachelors are unmarried") * laws of nature Necessity: N(p) := p is necessarily true N(p) = ~M(~p) = I(~p) = (p->q) & (~q->p) It is always N(p) & N(q) <=> N(p & q) Possibility: M(p) := ~N(~p) M(p) = ~N(~p) = ~I(p) It is always: M(p & q) -> M(p) & M(q), M(p | q) <=> M(p) | M(q) Impossibility: I(p) := ~M(p) I(p) = N(~p) = ~M(p) = (p->~p) Contingency: C(p) := M(p) & M(~p) Incompatability: IC(p,q) := I(p & q) Modal operators according to Leibniz: Necessity: p is true in all possible worlds Contingent truth: True in our world, but not in every possible world Possible truth: Not true in our world,but in at least 1 possible world Types of Implication: * material implication: p > q (boolean expression) * logical implication: p -> q (necessary inference) * strict implication: (p => q) := N(p > q) reflexive, transitive, not symmetric Types of Equivalence: * material equivalence: p = q (boolean expression) * strict equivalence: (p <=> q) := N(p = q) reflexive, transitive, symmetric Types of Proofs: * syntactic proofs: * Shifting: Convert the formula to a known one * Constructive: Assemble the formula by rules * Indirect: Assume that the formula is false, lead to contradiction * Falsify: Find a counter-example * semantic proofs: * Constructive: Substitute the modalities by their meaning, proove the resulting expression by propositional calculus * Indirect: Assume that the formula is false, substitute the modalities by their meaning and find a contradiction * Falsify: Find a diagram of worlds as a counter-example Rules: MP: Modus ponens: p, p > q |- q RN: Rule of necessitation: p |- N(p) N1: N(a > b) > (N(a) > N(b)) N2: N(a) > a N3: N(a) > N(N(a)) N4: ~N(a) > N(~N(a)) Models: A model is a triple , where * V is a Valuation function, evaluating a boolean expression V(world, expression) maps to {true,false} * W is a non-empty set of worlds * R is a Relation function, returns accessability R(world1,world2) = (world1 can access world 2) Models: 1. K-Model Properties of R: none Additional rules: Propositional Calculus, MP: p,p>q |- q RN: p |- N(p) N1: N(a > b) > (N(a) > N(b)) Modalities: Infinite 2. T-Model Properties of R: reflexive Additional rules: N2: N(p) > p Modalities: Inifite 3. S4-Model Properties of R: reflexive, transitive Additional rules: N3: N(p) > N(N(p)) Modalities: N(p) -> N(M(N(p))) -> N(N(p)) -> M(N(M(p))) -> M(p) \----------------bare-truth-------------------^ 4. S5-Model Properties of R. reflexive, transitive, symmetric Additional rules: N4: M(p) > N(M(p)) Modalities: {M|N}M(p)=M(p) {M|N}N(p)=N(p) Provability: "|- a in X" means a is provable in model X Adequacy of a model: 1. The model must be self-consistent, i.e. if |- a in model X then a is semantically valid (true in every possible world of every possible model with a relation function R satisfying the conditions of model X) 2. The model must be complete Proof of semantic validity of a model X: 1. Every axiom of X is valid in X 2. If a is valid in X, then also N(a) must be valid (this is obvious) 3. Proofs of N(i) i=1..5, using the properties of R in X Proof of semantic validity of the S5-model: 1. (numerous proofs) 2. (obvious) 3. Proof of all rules: (FMS version) N1: N(a > b) > (N(a) > N(b)) <=> N(a > b) & N(a) > N(b) <=> N( (a > b) & a) > N(b) <=> N(a & b) > N(b) <=> N(a) & N(b) > N(b) (obvious) N2: N(a) > a Assumption Ex V: Ex w: V(w,N2)=f <=> V(w,N(a))=t & V(w,a)=f <=> (All w': R(w,w') > V(w',a)=t) & V(w,a)=f => V(w,a)=t & V(w,a)=f N3: N(a) > N(N(a)) Assumption Ex V: Ex w: V(w,N3)=f <=> V(w,N(a))=t & V(w,N(N(a)))=f <=> (All w': R(w,w') > V(w,a)=t) & (Ex w'': R(w,w'') & V(w'',N(a))=f) <=> (All w': R(w,w') > V(w,a)=t) & (Ex w'': R(w,w'') & Ex w''': R(w'',w''') > V(w''',a)=f) Contradictory, since R(w,w'') & R(w',w''') => R(w,w''') N4: ~N(a) > N(~N(a)) Assumption: Ex V. Ex w: V(w, N4) = f <=> V(w,~N(a))=t & V(w,N(~N(a)))=f <=> (Ex w': R(w,w') & V(w',a)=f) (Ex w'': R(w,w'') & V(w'',~N(a))=f) <=> (Ex w': R(w,w') & V(w',a)=f) (Ex w'': R(w,w'') & All w''': R(w'',w''') > V(w''',a)=t) Contradictory, since R(w,w'') & R(w',w''') > R(w,w''')