This is the Base.Equality.Extensionality module of the Agda Universal Algebra Library.
{-# OPTIONS --without-K --exact-split --safe #-} module Base.Equality.Extensionality where -- imports from Agda and the Agda Standard Library ------------------------------------ open import Agda.Primitive using () renaming ( Set to Type ; Setω to Typeω ) open import Data.Product using ( _,_ ) renaming ( _×_ to _∧_ ) open import Level using ( _⊔_ ; Level ) open import Relation.Binary using ( IsEquivalence ) renaming ( Rel to BinRel ) open import Relation.Unary using ( Pred ; _⊆_ ) open import Axiom.Extensionality.Propositional using () renaming ( Extensionality to funext ) open import Relation.Binary.PropositionalEquality using ( _≡_ ; refl ) -- imports from agda-algebras -------------------------------------------------------------- open import Overture using ( transport ) open import Base.Relations using ( [_] ; []-⊆ ; []-⊇ ; IsBlock ; ⟪_⟫ ) open import Base.Equality.Truncation using ( blk-uip ; to-Σ-≡ ) private variable α β γ ρ 𝓥 : Level
Previous versions of the agda-algebras library made heavy use of a global function extensionality principle asserting that function extensionality holds at all universe levels. However, we have removed all instances of global function extensionality from the current version of the library and we now limit ourselves to local applications of the principle. This has the advantage of making transparent precisely how and where the library depends on function extensionality. Eventually we hope to be able to remove these postulates altogether in favor of an alternative approach to extensionality (e.g., by working with setoids or by reimplementing the entire library in Cubical Agda).
The following definition is useful for postulating function extensionality when and where needed.
DFunExt : Typeω DFunExt = (𝓤 𝓥 : Level) → funext 𝓤 𝓥
A useful alternative for expressing dependent function extensionality, which is essentially equivalent to dfunext
, is to assert that the happly
function is actually an equivalence.
The principle of proposition extensionality asserts that logically equivalent propositions are equivalent. That is, if P
and Q
are propositions and if P ⊆ Q
and Q ⊆ P
, then P ≡ Q
. For our purposes, it will suffice to formalize this notion for general predicates, rather than for propositions (i.e., truncated predicates).
_≐_ : {α β : Level}{A : Type α}(P Q : Pred A β ) → Type _ P ≐ Q = (P ⊆ Q) ∧ (Q ⊆ P) pred-ext : (α β : Level) → Type _ pred-ext α β = ∀ {A : Type α}{P Q : Pred A β } → P ⊆ Q → Q ⊆ P → P ≡ Q
Note that pred-ext
merely defines an extensionality principle. It does not postulate that the principle holds. If we wish to postulate pred-ext
, then we do so by assuming that type is inhabited (see block-ext
below, for example).
We need an identity type for congruence classes (blocks) over sets so that two different presentations of the same block (e.g., using different representatives) may be identified. This requires two postulates: (1) predicate extensionality, manifested by the pred-ext
type; (2) equivalence class truncation or “uniqueness of block identity proofs”, manifested by the blk-uip
type defined in the [Base.Relations.Truncation][] module. We now use pred-ext
and blk-uip
to define a type called block-ext|uip
which we require for the proof of the First Homomorphism Theorem presented in Base.Homomorphisms.Noether.
module _ {A : Type α}{R : BinRel A ρ} where block-ext : pred-ext α ρ → IsEquivalence{a = α}{ℓ = ρ} R → {u v : A} → R u v → [ u ] R ≡ [ v ] R block-ext pe Req {u}{v} Ruv = pe ([]-⊆ {R = (R , Req)} u v Ruv) ([]-⊇ {R = (R , Req)} u v Ruv) private to-subtype|uip : blk-uip A R → {C D : Pred A ρ}{c : IsBlock C {R}}{d : IsBlock D {R}} → C ≡ D → (C , c) ≡ (D , d) to-subtype|uip buip {C}{D}{c}{d} CD = to-Σ-≡ (CD , buip D (transport (λ B → IsBlock B) CD c) d) block-ext|uip : pred-ext α ρ → blk-uip A R → IsEquivalence R → ∀{u}{v} → R u v → ⟪ u ⟫ ≡ ⟪ v ⟫ block-ext|uip pe buip Req Ruv = to-subtype|uip buip (block-ext pe Req Ruv)