Theorem 10, proof (converse step) — quasisupermodularity of g + p·x for every p forces supermodularity of g
ProvedMonotoneCompStatics.LinearPerturb.theorem10_converse_stepLet and give the componentwise order, so that and are the componentwise maximum and minimum. Let , and for a price vector write . If for every the function
is quasisupermodular on , then is supermodular on :
This is the step of the proof of Theorem 10 that recovers supermodularity in (for a fixed value of the parameter , which the paper suppresses) from quasisupermodularity of all linear perturbations. It shows that quasisupermodularity, although an ordinal property, pins down the cardinal property supermodularity once it is required robustly against all linear price terms.
Formalization Note is Fin n → ℝ with Mathlib's pointwise order and lattice operations, and is p ⬝ᵥ x (dotProduct). The hypothesis quantifies over all , not over some .
import Mathlib import Definitions.Def_Supermodularity_Monotonicity_SupermodularOn import Definitions.Def_MonotoneCompStatics_Monotonicity_QuasiSupermodularOn
namespace MonotoneCompStatics.LinearPerturb
/-- Milgrom and Shannon (1994), p. 166, proof of Theorem 10 (the step for `x`, with `t`
suppressed): if `g + p · x` is quasisupermodular on `ℝⁿ` for every price vector `p ∈ ℝⁿ`, then
`g` is supermodular on `ℝⁿ` (componentwise order, meet and join). -/
theorem theorem10_converse_step {n : ℕ} (g : (Fin n → ℝ) → ℝ)
(h : ∀ p : Fin n → ℝ, MonotoneCompStatics.Monotonicity.QuasiSupermodularOn (fun x => g x + p ⬝ᵥ x) Set.univ) :
Supermodularity.Monotonicity.SupermodularOn g Set.univ := by sorry
end MonotoneCompStatics.LinearPerturb
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.