đź“šHome

Logic crib sheet

Propositional consequence

We say that a formula \(X\) is a consequence of a set \(S\) of formulas, denoted \(S \vDash X\) , provided that \(X\) maps to \(T\) under every valuation that maps every member of \(S\) to \(T\) .

X is a tautology if and only if \(\empty \vDash X\) , which we write as \(\vDash X\).

Boolean algebra laws

booleanAlgebraLaws

\(\alpha\) and \(\beta\) formula

alphaBetaFormula

Generalised operands

Generalised disjunction

Generalised conjunction

Normal forms

Disjunctive normal form

Start with \([ \langle X \rangle ]\), then repeatedly apply the following rules to the expression, terminating when only literals remain:

DNFrules

Conjunctive normal form

Start with \(\langle [ X ] \rangle\), then repeatedly apply the following rules to the expression, terminating when only literals remain:

CNFrules

Proof systems

Semantic tableau expansion

Start with \(\neg X\) as the root node, as it is a refutation system, and we are looking for contradictions.

In each step, select a branch and a non-literal formula on that branch, then apply the following rules (same as DNF):

DNFrules

A closed branch is when either both \(N\) and \(\neg N\), or \(\bot\) occur on a branch. If all branches of the expansion are closed, the algorithm terminates.

A tableau is called strict if each vertex in the tree only has one rule applied to it.

A tableau proof of \(X\) is a closed tableau for \(\neg X\), and we write \(\vdash_t X\) to denote this.

Resolution expansion

Start with \(\neg X\) as the first disjunction, as it is a refutation system, and we are looking for contradictions.

In each step, select a disjunction and a non-literal formula \(N\) within it, then apply the following rules to \(N\) (same as CNF):

CNFrules

There is another rule called the resolution rule.

Suppose \(D_1\) and \(D_2\) are two disjunctions, with \(N\) occurring in \(D_1\) and \(\neg N\) in \(D_2\).

Let \(D\) be the result of the following:

1) Delete all occurrences of \(N\) from \(D_1\) 2) Delete all occurrences of \(\neg N\) from \(D_2\) 3) Combine the resulting disjunctions

Finally, add \(D\) as a new line in the proof

There is a special case of when a disjunction \(D\) contains \(\bot\), in which case remove all instances of \(\bot\) from the \(D\)

A resolution is called strict if each disjunction in the proof only has one rule applied to it.

A resolution proof of \(X\) is a closed resolution for \(\neg X\), and we write \(\vdash_r X\) to denote this.

Natural deduction

A system of assumptions, with sub-ordinate proofs (lemmas), enclosed within boxes, which can then be used later in the proof. The first formula in a box is an assumption, which is taken as axiomatic, only within the scope of that.

The following rules can be applied:

naturalDeductionAxiomaticRules1

naturalDeductionAxiomaticRules2

naturalDeductionDerivedRules

Proof strategies:

We write \(\vdash_d X\) to denote a natural deduction proof of \(X\)

S-introduction

The above proof methods just prove tautologies. These allow us to prove propositional consequences by “introducing” premises from a proposition \(S\).

Semantic tableau

Any formula \(Y \in S\) can be added to the end of any tableau branch. We write \(S \vdash_t X\) if there is a closed tableau for \(\neg X\) from \(S\), allowing the S-introduction rule for tableau.

Resolution

For any formula \(Y \in S\), the line \([ Y ]\) can be added as a line to a resolution expansion. We write \(S \vdash_r X\) if there is a closed resolution expansion for \(\neg X\) from \(S\), allowing the S-introduction rule for resolution.

Natural deduction

For any formula \(Y \in S\), the line \(Y\) can be added as a line anywhere within the natural deduction. We write \(S \vdash_d X\) if there is a natural deduction derivation for \(X\) from \(S\) , allowing the S-introduction rule for resolution.