@@ -38,7 +38,7 @@ to the idea of a reality whose structure is perfectly captured by these refined,
3838mathematical models, unencumbered by the computational and cognitive limitations that
3939typically obscure such understanding.
4040
41- ## Definitions
41+ ## Coda
4242
4343### 1.1 Syntactic Canonicity
4444
@@ -71,38 +71,36 @@ behave coherently up to homotopy.
7171
7272Formally, in HoTT: ` Π (t: ℕ), Σ (n: ℕ), Path ℕ t n `
7373
74- ## Canonicity Across Type Theories
74+ ### 1.4 Canonicity Across Type Theories
7575
7676| Type Theory| Syntactical| Propositional| Homotopy |
7777| -----------| -----------| -------------| -----------------------------------|
7878| MLTT | Yes| Yes | Yes |
7979| HoTT | No| Yes | Yes (Bocquet, Kapulkin, Sattler) |
8080| CCHM | Yes| Yes | Yes (Coquand, Huber, Sattler) |
8181
82- ## Proof Sketches of Canonicity Results
83-
8482Different type-theoretic frameworks impose different levels of canonicity.
8583While MLTT has full syntactic, propositional, and homotopy canonicity, HoTT
8684lacks syntactic canonicity but retains homotopy canonicity. Cubical HoTT
8785restores full canonicity using its explicit cubical structure. Understanding
8886these distinctions is crucial for developing computational and proof-theoretic
8987applications of type theory.
9088
91- ### Failure of Syntactic Canonicity in HoTT
89+ #### Failure of Syntactic Canonicity in HoTT
9290
9391In Homotopy Type Theory, function extensionality and univalence introduce
9492higher-inductive types, making reduction ambiguous for closed terms.
9593Specifically, closed terms of Nat may contain elements that do not
9694normalize to a numeral but are still provably equal to one in homotopy.
9795
98- ### Proof Idea for Propositional Canonicity in HoTT
96+ #### Proof Idea for Propositional Canonicity in HoTT
9997
10098Bocquet and Kapulkin-Sattler established that every term of Nat is
10199propositionally equal to a numeral. The idea is to use a strict Rezk
102100completion of the syntactic model to construct a fibrant replacement
103101where each closed term can be shown to be propositionally equal to a numeral.
104102
105- ### Proof Idea for Homotopy Canonicity in Cubical Type Theory
103+ #### Proof Idea for Homotopy Canonicity in Cubical Type Theory
106104
107105Coquand, Huber, and Sattler proved homotopy canonicity using cubical models,
108106where paths (identity types) are explicitly represented as maps over the
@@ -118,7 +116,7 @@ Table 2: Mechanisms Ensuring Canonicity in Different Type Theories
118116| HoTT | Homotopical fibrant replacement (propositional & homotopy canonicity) |
119117| CCHM | Cubical paths + hcomp enforcing structured identity types |
120118
121- ## Example of Violating Syntactic Canonicity
119+ ### 1.5 Example of Violating Syntactic Canonicity
122120
123121` ℕ ` defined in CCHM through ` W ` , ` 0 ` , ` 1 ` , ` 2 ` doesn't compute numerals expressions to same terms,
124122however they are propotionally canonical in CCHM though ` hcomp ` .
@@ -142,7 +140,7 @@ def ℕ-ind (C : ℕ → U) (z : C zero) (s : Π (n : ℕ), C n → C (succ n))
142140 (λ (f : 𝟏 → ℕ) (g : Π (x : 𝟏), C (f x)), 𝟏⟶ℕ C f (s (f ★) (g ★))))
143141```
144142
145- ### The Code
143+ #### The Code
146144
147145* ` ℕ-ctor ` is defined as a two-point inductive type,
148146 which is essentially the structure of natural numbers,
@@ -158,7 +156,7 @@ def ℕ-ind (C : ℕ → U) (z : C zero) (s : Π (n : ℕ), C n → C (succ n))
158156* The terms ` 𝟎⟶ℕ ` and ` 𝟏⟶ℕ ` define the transport functions for zero and successor cases,
159157 respectively, using transposition (transp).
160158
161- ### Syntactic Canonicity
159+ #### Syntactic Canonicity
162160
163161In the case of natural numbers through ` W ` , ` 0 ` , ` 1 ` , ` 2 ` , this would mean that terms involving
164162natural numbers reduce to either 0 or succ n for some n. In this case,
@@ -176,7 +174,7 @@ they involve higher inductive types and path spaces.
176174 terms due to the nature of the recursion and the transport between
177175 different levels of the inductive structure.
178176
179- ### Failures in Canonicity
177+ #### Failures in Canonicity
180178
181179* Non-normalizing terms: Because of the presence of path-dependent
182180 types ` PathP ` and recursive definitions involving higher inductive
@@ -188,7 +186,7 @@ they involve higher inductive types and path spaces.
188186 to their normal form, especially if the path spaces themselves
189187 are complicated or not trivially reducible.
190188
191- ### Reformulating Canonicity for Natural Numbers
189+ #### Reformulating Canonicity for Natural Numbers
192190
193191To reformulate canonicity for natural numbers built using this approach, consider the following:
194192
0 commit comments