Skip to content
Documentation out of dateLearn more

Eta-equivalence

Eta-equivalence (η-equivalence) is a notion of proof equivalence in natural deduction logics with introduction and elimination forms. Roughly, it says that any proof of a proposition is equivalent to one which introduces the proposition’s principal connective. It is a form of extensional equivalence.

Consider the simply-typed lambda-calculus with arrow types.

A::=a∣A1→A2A ::= a \mid A_1 \rightarrow A_2

e::=x∣λx:A. e∣e1 e2e ::= x \mid \lambda x{:}A.\, e \mid e_1\ e_2

Eta-equivalence for terms e:A→B\texttt{}e : A \rightarrow B is the least congruence relation e1=ηe2\texttt{}e_1 =_\eta e_2 closed under the η\texttt{}\eta axiom:

(x∉FV(e))λx:A. e x=ηeη{(x \not\in \mathit{FV}(e)) \over \lambda x{:}A.\, e\ x =_\eta e} \eta

In logics and typed calculi, eta-equivalence is usually oriented to the left yielding a notion of eta-expansion. For example:

e:A→B⟹ηλx:A. e xe : A \rightarrow B \Longrightarrow_\eta \lambda x{:}A.\, e\ x

Eta-expansion transforms an arbitrary proof of a proposition into a proof that introduces the proposition’s principal connective. A term with no sub-terms that can be eta-expanded without introducing beta-redexes is said to be eta-long. Being eta-long is one aspect of being canonical.