Every balanced sequence of `{1} ⊆ ZMod n` has length at least `n`
ProvedSocialChoice.singleton_one_min_lengthaether-catalogapplications
Every balanced sequence of {1} ⊆ ZMod n has length at least n
(reproduced).
theorem SocialChoice.singleton_one_min_length{n : ℕ} (l : List (ZMod n))
(h : IsBalanced ({1} : Frame n) l) : n ≤ l.length := by sorry
Formalization Note Transplanted verbatim from the Aether Catalog source Applications/SocialChoice/IncoherenceStratification.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.
Preamble
-- Thm stub generated from Applications/SocialChoice/NonFiniteAxiomatization.lean
import Mathlib
import Definitions.Def_Applications_SocialChoice_NonFiniteAxiomatization
/-
# Non-finite-axiomatization of measurable majorities via the incoherence index
This file extends the catalog model of *standard social decision frames* from
`Applications/SocialChoice/IncoherenceIndex.lean` (the reconstructed
`MossPedersen2026` / `arXiv:2606.23853` framework: frames as finite atom sets in
`ZMod n`, *perfectly balanced sequences*, and the *incoherence index*). Because
the surrounding monorepo's build configuration does not make that module
importable in isolation, the small amount of shared infrastructure (the model
definitions and three base lemmas about the single-generator frame) is
reproduced here verbatim from the catalog file; the **main results below are new**:
* `realization_2k2` — for every `k ≥ 1` a *maximal* frame whose shortest
coherence violation has length exactly `2k+2` (every even index `≥ 4`).
* `coherence_not_finitely_axiomatizable` — no bounded finite fragment can replace
the coherence criterion (the headline non-finite-axiomatization theorem).
* `incoherenceIndex_unbounded_over_maximal` — the spectrum of incoherence indices
over maximal frames is unbounded.
-- !-- Lab Notes -- !--
HYPOTHESIS (Hypothesizer). Bold conjecture: coherence (= strict majority
representability) is *not* finitely axiomatizable. For no fixed `B` does "no
perfectly balanced sequence of length `≤ B`" imply "no perfectly balanced
sequence at all", uniformly over all finite social decision frames. The
obstruction is quantitative: incoherence indices realize the entire even tail
`2k+2`, so any finite bound is eventually overshot.
EXPERIMENT (Experimenter). Reuse the catalog model. The single-generator frame
`{1} ⊆ ZMod n` is maximal and its shortest balanced sequence is `1` repeated `n`
times (`incoherenceIndex_singleton_one`). Set `n = 2k+2` to hit every even
length `≥ 4`; set `n = B+1` to defeat the width-`B` fragment. The decisive new
lemma `singleton_one_min_length` shows every balanced sequence of `{1}` has
length `≥ n` (all atoms equal `1`, so the length is a positive multiple of `n`).
ANALYSIS (Analyst). True and provable. Structural pattern: the incoherence
index of `{1} ⊆ ZMod n` equals the additive order `n` of the unit, so the family
`{1} ⊆ ZMod (B+1)` produces, for each `B`, a frame that passes the width-`B`
test yet is genuinely incoherent — exactly a non-finite-axiomatization failure:
the bounded fragments are strictly weaker than the full criterion at every stage.
CRITIQUE (Critic). Guards: `coherence_not_finitely_axiomatizable` is proved by
`by_contra` (no `decide`/`rfl` shortcut); every witness frame is exhibited
maximal; the separating gap is quantitative (`B < l.length`). No theorem is
vacuous: each existential carries a concrete frame with a computed index, and the
separating frame is simultaneously coherent-up-to-`B` and incoherent.
SYNTHESIS (PI). Together with the catalog's `realization_even`, these results
upgrade "the spectrum is unbounded" to "the criterion admits no bounded finite
fragment" — the precise sense in which measurable majorities cannot be finitely
axiomatized. See `FUTURE_DIRECTIONS.md`.
-- !-- Lab Notes -- !--
-/
open SocialChoice
open scoped BigOperators
/-! ## Catalog model (reproduced from `IncoherenceIndex.lean`) -/
/-! ## New results: non-finite-axiomatization -/
/-
Every perfectly balanced sequence of the single-generator frame `{1}` has a
length divisible by `n`: all its atoms equal `1`, so its length is a multiple of
the additive order `n`.
-/
/-
Every perfectly balanced sequence of the single-generator frame `{1} ⊆ ZMod n`
has length at least `n` (its incoherence index).
-/Formal statement
theorem SocialChoice.singleton_one_min_length{n : ℕ} (l : List (ZMod n))
(h : IsBalanced ({1} : Frame n) l) : n ≤ l.length := by sorrySource