Type trees are distance types
Lax185574.TreeEquivalence · concepts/Lax185574/TreeEquivalence.lean · lax-185574
No public endorsements yet.
Loading review…
Sign in with ORCIDNatural Language Statement
Corollary
Fix ℓ ≤ k and q, a partial graph G and ā, b̄ ∈ V(G)^ℓ. Then tree_{k,q}(ā) = tree_{k,q}(b̄) if and only if G ⊨ φ(ā) ⇔ φ(b̄) for every distFO formula φ(x̄) of distance rank (k, q) with |x̄| = ℓ.
Concept map
Evidence
Each proof establishes this claim relative to its assumptions.
In the paper
- page 43 of this submission's paper
Lean source view on GitHub
| 1 | import Lax185574.TypeTrees |
| 2 | |
| 3 | /-! |
| 4 | --- |
| 5 | title: Type trees are distance types |
| 6 | type: corollary |
| 7 | --- |
| 8 | Fix ℓ ≤ k and q, a partial graph G and ā, b̄ ∈ V(G)^ℓ. Then |
| 9 | tree_{k,q}(ā) = tree_{k,q}(b̄) if and only if G ⊨ φ(ā) ⇔ φ(b̄) for every |
| 10 | distFO formula φ(x̄) of distance rank (k, q) with |x̄| = ℓ. |
| 11 | |
| 12 | # Formalization notes |
| 13 | |
| 14 | The vertex type is finite, as everywhere in the paper. Equality of type |
| 15 | trees is equality of the objects of `Lax185574.TypeTrees`. |
| 16 | -/ |
| 17 | |
| 18 | namespace Lax185574.TreeEquivalence |
| 19 | |
| 20 | open Lax185574.PartialGraphs Lax185574.DistanceLogic Lax185574.DistanceLogic.DistFO |
| 21 | open Lax185574.TypeTrees |
| 22 | |
| 23 | /-- **Type trees are distance types.** Two tuples have the same type |
| 24 | tree of rank `(k, q)` exactly when they satisfy the same formulas of |
| 25 | distance rank `(k, q)`. -/ |
| 26 | axiom tree_eq_iff {V : Type*} [Finite V] (G : PartialGraph V) (k q : ℕ) {ℓ : ℕ} |
| 27 | (hℓ : ℓ ≤ k) (a b : Fin ℓ → V) : |
| 28 | tree G k q a = tree G k q b ↔ |
| 29 | ∀ φ : DistFO ℓ, DRank k q φ → (Sat G a φ ↔ Sat G b φ) |
| 30 | |
| 31 | end Lax185574.TreeEquivalence |
| 32 |
Formalization notes
The vertex type is finite, as everywhere in the paper. Equality of type trees is equality of the objects of .
Discussion
Ask a question or add context. Endorsements and structured flags are kept in the review panel above.
0 comments