Lemma 1 — Bound-variable standardization
OpenBergerSchwichtenberg1991V2.lemma_1At every prefix the standardization of two -equivalent raw terms is identical. For any term with no free index extending , standardization at gives an -equivalent term.
import Definitions.Def_BergerSchwichtenberg1991V2_Syntax
namespace BergerSchwichtenberg1991V2
/-- Lemma 1 (p. 204). Standardization depends only on the α-class, and a prefix
excluded by the free indices gives an α-variant. These statements apply also to raw
terms without a typing derivation. -/
theorem lemma_1 (k : Index) :
(∀ r s : Term, AlphaEq r s → alphaRename k r = alphaRename k s) ∧
(∀ r : Term, (∀ m ∈ freeIndices r, ¬ IsPrefixOf k m) → AlphaEq (alphaRename k r) r) := by sorry
end BergerSchwichtenberg1991V2Read-back
What the Lean code literally says, in plain math · OpenAI Codex (GPT-6)
For every index , a finite list of Boolean bits that may be empty, rename the binders of a raw term by giving each binder the indexed name concatenated with its path from the root. Paths append into an abstraction body or the left child of an application, and into the right child of an application. A variable is its name together with its type. Each bound occurrence takes the new name of its nearest matching binder and each free occurrence keeps its name and type. Two raw terms are -equivalent if the trees obtained by recording only the types of binders and the type and binder path of each bound occurrence, while recording each free variable in full, are equal. For all raw terms , if they are -equivalent, their renamings at are literally equal raw terms. Also, for every raw term , if for every index of a free indexed variable of there is no bit list with , then the renaming of at is -equivalent to . Free variables with natural-number names do not enter this prefix condition. At , every index has prefix , so the condition holds exactly when has no free indexed variables. These assertions quantify over open and ill-typed raw terms as well.