Type theory MOC

Judgemental equality

Judgemental equality is equality defined as a judgement, in contrast to propositional or typal equality. The judgements have the form1

Ξ“βŠ’π΄=𝐡,Ξ“βŠ’π‘Ž=𝑏:𝐴

for β€œπ΄ and 𝐡 are judgementally equal (well-typed) types in context Γ” and β€œπ‘Ž and 𝑏 are judgementally equal (well-typed) terms of (well typed) type 𝐴 in context Γ” respectively. Informally, judgemental equality operates at a metatheoretic level: It is proof irrelevant, and cannot be used within the theory e.g. as the antecedent of a conditional. The actual meaning of judgemental equality depends on the type theory used

Usual inference rules

Both sides of a type equality judgement are themselves types:

Ξ“βŠ’π΄=π΅Ξ“βŠ’π΄Ξ“βŠ’π΄=π΅Ξ“βŠ’π΅

Similarly, both sides of a term equality judgement are themselves terms:

Ξ“βŠ’π‘Ž=𝑏:π΄Ξ“βŠ’π‘Ž:π΄Ξ“βŠ’π‘Ž=𝑏:π΄Ξ“βŠ’π‘:𝐴

Judgemental equality of types is an equivalence relation

Ξ“βŠ’π΄Ξ“βŠ’π΄=π΄Ξ“βŠ’π΄=π΅Ξ“βŠ’π΅=π΄Ξ“βŠ’π΄=π΅Ξ“βŠ’π΅=πΆΞ“βŠ’π΄=𝐢

and likewise for terms

Ξ“βŠ’π‘Ž:π΄Ξ“βŠ’π‘Ž=π‘Ž:π΄Ξ“βŠ’π‘Ž=𝑏:π΄Ξ“βŠ’π‘=π‘Ž:π΄Ξ“βŠ’π‘Ž=𝑏:π΄Ξ“βŠ’π‘=𝑐:π΄Ξ“βŠ’π‘Ž=𝑐:𝐴.

We have the variable conversion rule, for a generic judgement thesis J

Ξ“βŠ’π΄=𝐴′Γ,𝐴,Ξ”βŠ’JΞ“,𝐴′,Ξ”βŠ’J(V)

which guarantees Indiscernibility of identicals for types. Finally, judgemental equality is a β€œcongruence” with respect to substitution:

Ξ“βŠ’π‘Ž=π‘Žβ€²:𝐴Γ,π‘₯:𝐴,Ξ”βŠ’π΅Ξ“,Ξ”[π‘Ž/π‘₯]⊒𝐡[π‘Ž/π‘₯]=𝐡[π‘Žβ€²/π‘₯](TV=)Ξ“βŠ’π‘Ž=π‘Žβ€²:𝐴Γ,π‘₯:𝐴,Ξ”βŠ’π‘‘:𝐡Γ,Ξ”[π‘Ž/π‘₯]βŠ’π‘‘[π‘Ž/π‘₯]=𝑑[π‘Žβ€²/π‘₯]:𝐡[π‘Ž/π‘₯](tV=)

Further inference rules on judgemental equality are given for individual types in the type theory, and are often classified under Ξ±-conversion, 𝛽-computation, and Ξ·-reduction.2 #m/def/type


#state/tidy | #lang/en | #SemBr

Footnotes

  1. We depart from the usual convention: Type-theoretic literature will usually reserve ( =) for typal equality, using either ( ≑) or ( Λ™=) for judgemental equality. We take the opposite approach, which is more akin to the notation used in proof assistants. This is also used in Principles of dependent type theory. ↩

  2. 2025. Introduction to Homotopy Type Theory, Β§1, pp. 11–19 ↩