Chapter 6 Logic in Coq


All syntactically well-formed propositions in Coq have type Prop in Coq. Propositions are not necessarily provable.

Functions that return propositions are called parameterized propositions; they define properties of their arguments.

Logical Connectives

For a conjunction, we use split to separate it and prove each of them if it appears in the goal, but use destruct if it appears in the hypothesis.

For a disjunction, we use left or right to prove only one branch if it appears in the goal, but use destruct to separate them into branches and prove the goal with each branch taken as the premise.

¬ P is a representation of to P False. Anything can follow from False; this rule is called ex falso quodlibet, or the principle of explosion. If the goal is controversial, we can use this rule by the exfalso tactic, turning the goal into False.

The “if and only if” connective is reflexive, symmetric, and transitive; it is an equivalence relation. This means we can use reflexivity, symmetry, and rewrite on “” connected propositions, like on equations.

To prove a statement containing an existential quantification, we find a value that witnesses the existential.

Programming with Propositions

We can define propositions recursively, but this method is limited by Coq’s requirements for functions, for example, “obviously terminating.” We can also define propositions inductively.

Applying Theorems to Arguments

Coq treats proofs as first-class objects. The type of a proof is the proposition which it is a proof of. We can apply theorems to statements like applying a function; they have the same inference mechanism.

Coq vs. Set Theory

In Coq, we say that a property holds for a mathematical object, not that a mathematical object belongs to some set. This is because Coq’s logical core is the Calculus of Inductive Constructions, instead of set theory (e.g. Zermelo–Fraenkel Set Theory).

Coq’s equality operator is polymorphic, we can use it to claim that two functions are equal. It is common to consider two functions the same if we observe the same behavior from them:

(∀ x, f x = g x) → f = g

This is the principle of functional extensionality, we can define it as an axiom in Coq. Sometimes adding an axiom to Coq can break consistency, that is, we may be able to prove every proposition.

Coq’s language is computational, every function it can express is computable and total. Coq doesn’t have an operation to do case analyses on propositions, since allowing such operation would allow non-computable functions.

Many computable properties are easier to express using propositions rather than booleans, since recursive function definitions in Coq are subject to significant restrictions.

If a boolean function gives true if and only if a proposition holds, we say that the boolean computation is reflected in the truth of the proposition. We could exploit boolean computation to prove propositions; this technique is known as proof by reflection.

Logics that do not assume the excluded middle are referred to as constructive logics. Every proof of existence in such logics are constructive.

Although constructive logic is useful, constructive proofs of many statements are complicated. What’s worse, some statements only have non-constructive proofs.

Comments