The intermediate-derivation character ψt and Sideward detection (MCB Cor. 3.5.4) #
[MCB25] §3.5.2.2, eq. (3.5.7)
The character ϕt of Prop. 3.5.3 only weights the complete construction L(T) → T, whose grading
is always ≥ 0 (MCB Lemma 3.5.5) — so ϕt is nonpolar and cannot detect Sideward Merge. Cor. 3.5.4
introduces
ψt(T) = Σ_{(F,F') ∈ FT} (F → F') · t^{δ(F→F')},
summing over all intermediate derivation pairs (F, F') in the construction of T (forests with
L(F) = L(F') = L(T) lying on some chain L(T) → ⋯ → F → ⋯ → F' → ⋯ → T). Since δ depends only on
the endpoints (δb₀(F→F') = b₀ F − b₀ F', etc.), each segment contributes a single grading monomial.
This file is the lean t-grading model of ψt, parameterized by the multiset of intermediate
segments S (the full FT-enumeration engine — derivation reachability for a target — is deferred;
here S is the input). The essential content of Cor. 3.5.4 — and the reason ψt succeeds where ϕt
fails — is captured by polarHahn_psiSeries: the polar (divergent) part of ψt is exactly the sum
over its Sideward-divergent segments (δb₀ < 0, where b₀ increases — MCB's Sideward types
3(a)/3(b)). A single Sideward segment already makes ψt polar (polarHahn_psiSeries_sideward_3a),
which the Birkhoff factorization then renormalizes away (Core/Algebra/RootedTree/BirkhoffLaurent).
Main definitions #
Minimalist.Merge.psiSeries:ψtas a graded sum over a given segment multiset.
References #
[MCB25] (Cor. 3.5.4, eq. 3.5.7; Lemma 3.5.5; Prop. 3.5.6)
A workspace transformation (intermediate derivation segment) F → F'.
Equations
- Minimalist.Merge.Segment α β = (RootedTree.Forest (RootedTree.Nonplanar (α ⊕ β)) × RootedTree.Forest (RootedTree.Nonplanar (α ⊕ β)))
Instances For
ψt in the lean t-grading model (MCB eq. 3.5.7, δ = δb₀): a graded sum Σ tᵟ over the
multiset S of intermediate derivation segments of a target.
Equations
- Minimalist.Merge.psiSeries S = (Multiset.map (fun (p : Minimalist.Merge.Segment α β) => Minimalist.Merge.gradeMonomial (Minimalist.Merge.δb₀ p.1 p.2)) S).sum
Instances For
The polar part of ψt is exactly the divergent (Sideward) segments #
The polar (divergent) part of ψt (MCB §3.5.2.2): R·ψt is the sum of grading monomials
over exactly the Sideward-divergent segments (δb₀ < 0, i.e. b₀ increases). The polar part
detects Sideward Merge — the property ϕt lacks (Lemma 3.5.5).
If every intermediate segment respects no-divergence (δb₀ ≥ 0), then ψt is nonpolar —
the ϕt-like case (a construction using only External/Internal Merge).
The nonpolar projection (1 − R)·ψt keeps exactly the non-divergent segments (δb₀ ≥ 0 —
the External/Internal/Minimal-Yield-respecting derivations). At the level of a single target's
series this projection already separates the good derivations from the Sideward-divergent ones;
the subtlety MCB Prop. 3.5.6 addresses is that 1 − R fails to do this on the character level
(products of series), where the full Birkhoff factorization is needed because R is not an
algebra homomorphism.
A single Sideward 3(a) segment already makes ψt polar (δb₀ = −1): its whole value is the
divergent monomial t⁻¹. This is the contrast with ϕt (always nonpolar) — ψt sees the
Sideward Merge, which the Birkhoff factorization then removes.