Documentation

Deontic.canon3

We verify that the Maple model from 1996 is actually a canon₃_II model. Thus the definition of canon_II was foreshadowed in 1996; In this example canon_II can be characterized as the model obtain by a certain greedy algorithm for A5-C5 and automatically satisfies E5.

def canon₃ {α : Type u_1} [Fintype α] [DecidableEq α] (A B C : Finset α) :
Finset α → Finset (Finset α)
Equations
Instances For
    def canon₃'' {α : Type u_1} [Fintype α] [DecidableEq α] (A B C : Finset α) :
    Finset α → Finset (Finset α)
    Equations
    Instances For
      def canon₃''' {α : Type u_1} [Fintype α] [DecidableEq α] (A B C : Finset α) :
      Finset α → Finset (Finset α)
      Equations
      • One or more equations did not get rendered due to their size.
      Instances For
        def Ought {α : Type u_1} [Fintype α] [DecidableEq α] (ob : Finset α → Finset (Finset α)) (Y Z : Finset α) :

        Ought (Y | Z) as in 1996 Maple code.

        Equations
        Instances For
          theorem not_empty_inter {α : Type u_1} [Fintype α] [DecidableEq α] {A B : Finset α} {i : α} (hA : i ∈ A) (hB : i ∈ B) :
          A ∩ B ≠ ∅
          theorem paradox_in_CJ2022_general'' {n : ℕ} {ob : Finset (Fin n) → Finset (Finset (Fin n))} (b5 : B5 ob) (e5 : E5 ob) (f5 : F5 ob) {A B C D : Finset (Fin n)} (hC : C ≠ ∅) (hab : A ∩ B = ∅) (hbc : B ∩ C = ∅) (hAD : A ∩ D = ∅) (hf₁₀ : A ∈ ob (A ∪ B)) (hf₀₀ : C ∈ ob (C ∪ D)) :
          C ∈ ob (B ∪ C ∪ D)

          Limited Principle of Explosion: If C is at least somewhat desirable and the most desirable (A) is violated then C is obligatory.

          theorem limited_explosion {n : ℕ} {ob : Finset (Fin n) → Finset (Finset (Fin n))} (b5 : B5 ob) (e5 : E5 ob) (f5 : F5 ob) {A C U V : Finset (Fin n)} (h₀ : A ∈ ob U) (h₁ : C ∈ ob V) (h₂ : C ≠ ∅) (h₃ : C ∩ U = ∅) (h₄ : A ∩ V = ∅) (h₅ : A ⊆ U) (h₆ : C ⊆ V) :
          C ∈ ob (U \ A ∪ V)
          theorem paradox_in_CJ2022_general' {n : ℕ} {ob : Finset (Fin n) → Finset (Finset (Fin n))} (b5 : B5 ob) (e5 : E5 ob) (f5 : F5 ob) {A B C : Finset (Fin n)} (hC : C ≠ ∅) (hab : A ∩ B = ∅) (hbc : B ∩ C = ∅) (hf₁₀ : A ∈ ob (A ∪ B)) (hf₀₀ : C ∈ ob C) :
          C ∈ ob (B ∪ C)
          theorem paradox_in_CJ2022_general {n : ℕ} {ob : Finset (Fin n) → Finset (Finset (Fin n))} (b5 : B5 ob) (e5 : E5 ob) (f5 : F5 ob) {A B C T : Finset (Fin n)} (hA : A ≠ ∅) (hC : C ≠ ∅) (hac : A ∩ C = ∅) (hab : A ∩ B = ∅) (hbc : B ∩ C = ∅) (hAT : A ⊆ T) (hBT : B ⊆ T) (hCT : C ⊆ T) (ought₀ : Ought ob A T) (ought₂ : Ought ob C (T ∩ (A ∪ B)ᶜ)) :
          C ∈ ob (B ∪ C)
          theorem paradox_in_CJ2022_generalest {n : ℕ} {ob : Finset (Fin n) → Finset (Finset (Fin n))} (b5 : B5 ob) (e5 : E5 ob) (f5 : F5 ob) {A B C : Finset (Fin n)} (hA : A ≠ ∅) (hC : C ≠ ∅) (hac : A ∩ C = ∅) (hab : A ∩ B = ∅) (hbc : B ∩ C = ∅) (ought₀ : Ought ob A Finset.univ) (ought₂ : Ought ob C (A ∪ B)ᶜ) :
          C ∈ ob (B ∪ C)
          theorem paradox_in_CJ2022 {n : ℕ} {ob : Finset (Fin n) → Finset (Finset (Fin n))} (b5 : B5 ob) (e5 : E5 ob) (f5 : F5 ob) (a b c : Fin n) (hab : a ≠ b) (hbc : b ≠ c) (hac : a ≠ c) (ought₀ : Ought ob {a} Finset.univ) (ought₂ : Ought ob {c} {a, b}ᶜ) :
          {c} ∈ ob {b, c}

          This is paradoxical if a is the best world, then b, then c.

          theorem characterize_canon₂_II₀ {α : Type u_1} [Fintype α] [DecidableEq α] (A B : Finset α) (hAB : A ⊆ B) :
          theorem characterize_canon₂₀ {α : Type u_1} [Fintype α] [DecidableEq α] (A B : Finset α) (hAB : A ⊆ B) :
          inductive Ob_chisholm_b {n : ℕ} (A B : Finset (Fin n)) :
          Instances For
            inductive Ob_chisholm_bdf {n : ℕ} (A B : Finset (Fin n)) :
            Instances For
              noncomputable def ob_chisholm_b {n : ℕ} (A B : Finset (Fin n)) :
              Finset (Fin n) → Finset (Finset (Fin n))
              Equations
              Instances For
                noncomputable def ob_chisholm_bdf {n : ℕ} (A B : Finset (Fin n)) :
                Finset (Fin n) → Finset (Finset (Fin n))
                Equations
                Instances For
                  theorem characterize_canon₂_II {α : Type u_1} [Fintype α] [DecidableEq α] (A B : Finset α) (ob : Finset α → Finset (Finset α)) (b5 : B5 ob) (ought : Ought ob A Finset.univ ∧ Ought ob B Aᶜ) (X : Finset α) :
                  canon₂_II A B X ⊆ ob X

                  Theorem 2 in the first version of the paper.

                  def least_model {α : Type u_1} [Fintype α] [DecidableEq α] (ob : Finset α → Finset (Finset α)) (axioms : (Finset α → Finset (Finset α)) → Prop) :

                  Note that if ob depends on some parameters A, B, ... then so can the axioms.

                  Equations
                  Instances For
                    theorem ought_complement {α : Type u_1} [Fintype α] [DecidableEq α] (A : Finset α) (ob : Finset α → Finset (Finset α)) :
                    Ought ob A Aᶜ

                    Since canon₂_II also satisfies A B C E G we can furthermore characterize it as the least model of those axioms and B5 and these Oughts. It is remarkable that A, C, E, G do not add anything to this least model.

                            least model of  also of
                    

                    canon₂_II A B C E G B canon₂ A B C D B D F

                    What are the least models of the paradoxical ABCDEFG and ABCEFG?

                    theorem least_model_canon₂_II {α : Type u_1} [Fintype α] [DecidableEq α] (A B : Finset α) (h : A ⊆ B) :
                    least_model (canon₂_II A B) fun (ob : Finset α → Finset (Finset α)) => B5 ob ∧ Ought ob A Finset.univ ∧ Ought ob B Aᶜ
                    theorem canon₂_II_eq_ob_chisholm_b {n : ℕ} (A B : Finset (Fin n)) (h : A ⊆ B) :

                    We need to assume A ⊆ B, perhaps relevant to referee's question, although it was already in least_model_canon₂_II. The contribution of this theorem is to obtain canon₂_II as equal to a certain closure, not just satisfying a "least model" predicate. January 5, 2026.

                    theorem characterize_canon₂_case {α : Type u_1} [Fintype α] [DecidableEq α] {A B X Y : Finset α} {ob : Finset α → Finset (Finset α)} (b5 : B5 ob) (d5 : D5 ob) (f5 : F5 ob) (H₀ : ¬X ∩ B = ∅) (H₁ : X ∩ A = ∅) (ought : Ought ob B Aᶜ) (h : X ∩ B ⊆ Y) :
                    Y ∈ ob X
                    theorem BD5_of_B5_of_D5 {α : Type u_1} [Fintype α] [DecidableEq α] {ob : Finset α → Finset (Finset α)} (b5 : B5 ob) (d5 : D5 ob) :
                    BD5 ob
                    theorem II_2_2_finset {α : Type u_1} [Fintype α] [DecidableEq α] {ob : Finset α → Finset (Finset α)} (a5 : A5 ob) (b5 : B5 ob) (c5 : C5 ob) (d5 : D5 ob) :
                    F5 ob
                    theorem characterize_canon₂ {α : Type u_1} [Fintype α] [DecidableEq α] {A B : Finset α} {ob : Finset α → Finset (Finset α)} (b5 : B5 ob) (d5 : D5 ob) (f5 : F5 ob) (ought : Ought ob A Finset.univ ∧ Ought ob B Aᶜ) (X : Finset α) :
                    canon₂ A B X ⊆ ob X

                    canon₂ is the minimal model satisfying certain Ought's and B5, D5, F5. Actually F5 follows from B5, C5, D5, A5 (Lemma II-2-2) although here we do not need to assume C5.

                    June 6, 2025

                    theorem least_model_canon₂ {α : Type u_1} [Fintype α] [DecidableEq α] (A B : Finset α) (h : A ⊆ B) :
                    least_model (canon₂ A B) fun (ob : Finset α → Finset (Finset α)) => B5 ob ∧ D5 ob ∧ F5 ob ∧ Ought ob A Finset.univ ∧ Ought ob B Aᶜ
                    structure Rule (α : Type u_1) :
                    Type (max (max u_1 (u_2 + 1)) (u_3 + 1))
                    Instances For
                      inductive Closure {α : Type u_1} (𝒜 : Set (Set α)) {R : Type u_2} (rule : R → Rule α) :
                      Set α → Prop
                      Instances For

                        General setup for CJ rules #

                        structure sets_instance (α : Type u_1) (n : ℕ) (P : (Fin n → Finset α) → Prop) :
                        Type u_1
                        Instances For
                          structure closure_rule (α : Type u_1) :
                          Type u_1
                          Instances For
                            inductive closure_under {α : Type u_1} (𝒜 : Finset (Finset α × Finset α)) (rule : Finset (closure_rule α)) :
                            Instances For
                              theorem closure_under_sub {α : Type u_1} (𝒜 : Finset (Finset α × Finset α)) (rules₁ rules₂ : Finset (closure_rule α)) (h : rules₁ ⊆ rules₂) :
                              closure_under 𝒜 rules₁ ⊆ closure_under 𝒜 rules₂

                              If we add more rules then the closure becomes larger.

                              def B5rule (n : ℕ) :
                              Equations
                              • One or more equations did not get rendered due to their size.
                              Instances For
                                def ObArule {n : ℕ} (A : Finset (Fin n)) :
                                Equations
                                • One or more equations did not get rendered due to their size.
                                Instances For
                                  def ObBArule {n : ℕ} (A B : Finset (Fin n)) :
                                  Equations
                                  • One or more equations did not get rendered due to their size.
                                  Instances For

                                    This shows, together with earlier results, how to express canon₂_II as a closure under a set of operators.

                                    theorem close_under_eq_canon₂_II {n : ℕ} (A B Y : Finset (Fin n)) (h : A ⊆ B) :

                                    The closure of a set under some operations can be defined in a very noneffective way as the intersection of all closed sets. Second, it can be defined in a recursively enumerable way in terms of being generated by some operations. Sometimes, it can be characterized even more concretely in a quantifier-free way, like here. For example, in a vector space over F_2 the subspace generated by v is just {v,0}.

                                    structure DisjointUnionParams (α : Type u_1) :
                                    Type u_1
                                    Instances For
                                      def disjointUnionRule (α : Type u_1) :
                                      Rule α
                                      Equations
                                      Instances For
                                        theorem canon₂_eq_ob_chisholm_bdf {n : ℕ} (A B : Finset (Fin n)) (h : A ⊆ B) :
                                        theorem least_model_canon₂' {α : Type u_1} [Fintype α] [DecidableEq α] (A B : Finset α) (h : A ⊆ B) :
                                        least_model (canon₂ A B) fun (ob : Finset α → Finset (Finset α)) => A5 ob ∧ B5 ob ∧ C5 ob ∧ D5 ob ∧ Ought ob A Finset.univ ∧ Ought ob B Aᶜ

                                        A version of least_model_canon₂ that mentions only older axioms than F5.

                                        def canon₃_II {α : Type u_1} [Fintype α] [DecidableEq α] (A B C : Finset α) :
                                        Finset α → Finset (Finset α)
                                        Equations
                                        Instances For
                                          theorem maple_eq_canon₃_II (ob : Finset (Fin 4) → Finset (Finset (Fin 4))) (a b c d : Fin 4) (h : ob = canon₃_II {c} {c, d} {b, c, d}) (h₀ : a = 0) (h₁ : b = 1) (h₂ : c = 2) (h₃ : d = 3) :
                                          ob ∅ = ∅ ∧ ob {a} = ∅ ∧ ob {b} = {{b}, {a, b}, {b, c}, {b, d}, {a, b, c}, {a, b, d}, {b, c, d}, {a, b, c, d}} ∧ ob {c} = {{c}, {a, c}, {b, c}, {c, d}, {a, b, c}, {a, c, d}, {b, c, d}, {a, b, c, d}} ∧ ob {d} = {{d}, {a, d}, {b, d}, {c, d}, {a, b, d}, {a, c, d}, {b, c, d}, {a, b, c, d}} ∧ ob {a, b} = {{b}, {b, c}, {b, d}, {b, c, d}} ∧ ob {a, c} = {{c}, {b, c}, {c, d}, {b, c, d}} ∧ ob {a, d} = {{d}, {b, d}, {c, d}, {b, c, d}} ∧ ob {b, c} = {{c}, {a, c}, {c, d}, {a, c, d}} ∧ ob {b, d} = {{d}, {a, d}, {c, d}, {a, c, d}} ∧ ob {c, d} = {{c}, {a, c}, {b, c}, {a, b, c}} ∧ ob {a, b, c} = {{c}, {c, d}} ∧ ob {a, b, d} = {{d}, {c, d}} ∧ ob {a, c, d} = {{c}, {b, c}} ∧ ob {b, c, d} = {{c}, {a, c}} ∧ ob {a, b, c, d} = {{c}}