Exact rank of the full nonempty-block incidence projection
ProvedFormalCapacity.Finite.rank_blockIncidencefinitecombinatoricsformalcapacitylinearalgebrapartition-incidence
Let be a nonempty finite label set with . Let be its nonempty subsets and its genuine set partitions. Over , define by . Then
The formula measures how many independent additive block statistics remain after all partition identities are taken into account. It is a theorem for every nonempty finite label set, not only for four labels. The map is the transpose of the block-incidence projection from signed partition laws to block coordinates. The formal right-hand side uses natural-number subtraction; no claim is made for .
Preamble
import Mathlib
import Definitions.Def_capacityFinitePartitions
import Definitions.Def_capacityBlockIncidence
set_option autoImplicit false
/-!
# Rank of first-order block incidence
This file upgrades the row-relation characterization in `Partition.lean` to
the finite-dimensional rank formula `2^n - n`.
-/
open FormalCapacity.Finite
open scoped BigOperators
open Finset Module
variable {α : Type*} [Fintype α] [DecidableEq α]Formal statement
theorem FormalCapacity.Finite.rank_blockIncidence [Nonempty α] :
finrank ℚ (blockIncidenceTranspose (α := α)).range =
2 ^ Fintype.card α - Fintype.card α := by
sorry
Source
The four-label correlation threshold, Theorem 2.1, rank assertion in (2.2). Unpublished research note (2026), N4_BLOCK_CORRELATION_NOTE.md, SHA-256 a9b67c132ca0e0af807fb917ca2322833b84e162c45a4da7329ee88fbe4f1c83. Exact formal source: formal_capacity/FormalCapacity/Finite/BlockRank.lean, lines 166–185, declaration FormalCapacity.Finite.rank_blockIncidence, source-file SHA-256 0adb1c853d3a750450a8b0f6899884f5e03206466fc1f165f86e21e0849a29bc. Local source archive; no public repository URL or commit is asserted. Upload checked with Lean 4.33.1 and Mathlib 0df444a360eaa60ab8c11dca51a86af692955474.