§16 Case A Branch A: the g₃ sub-case assemblies for Mix (Pi, Lambda). #
This is the PL-specific m = 1, b₁ = d₁ sub-case of §16 Case 2. Here g₁ = g(1) is the
minimal (odd-rank) nonpolarized gene, g₂ = g⁺(2n'+2) is the minimal gene of X - g₁, and
X - g₁ - g₂ contains a negative or nonpolarized gene g₃ of minimal rank t. The
mutation g₂ + g₃ → g(2n'+1) + g(t+1) (t even, type7) or
g₂ + g₃ → g(2n'+1) + g⁺(t+1) (t odd, type6) gives Z ≤ Y, using the level-1-anchored
a-propagation branchA_g3_aprop (odd j) and half_le_sigma_diff_at_r (even j).
prime^[j] Y ≠ 0 for j below the rank of any gene of X (via dominance).
§16 g₃ existence: when g₁ = g(1) and g₂ = g⁺(2n'+2) are the two minimal genes and
a₁ < c₁ (ha), the remainder X - g₁ - g₂ contains a negative or nonpolarized gene.
Otherwise every non-positive gene of X is g₁ (rank 1), and the a-propagation
branchA_g3_aprop forces 1 ≤ 0 at a level beyond maxRank.
Mirror of one_le_signature_fst_of_contains_positive_mix for the b-component:
a negative gene g⁻(r) forces 1 ≤ b-signature at level r - 1.
§16 g₃ top boundary (type7, Mix (Pi, Lambda)). With g₃ = g⁻(2nn+2) a negative
gene of X (disjoint from Y) and t = 2nn+2 even, prime^[t] Y ≠ 0: otherwise the level
t-1 survivors of Y would all sit at rank 1 and, being neither g⁻(t) (disjoint) nor
NP(t) (even rank), contribute 0 to the b-component, contradicting the b-bound from
g₃.
§16 g₃ assembly, t odd (type6): g₂ + g₃ → g(2n'+1) + g⁺(t+1), with g₂ = g⁺(2n'+2)
and g₃ = NP(2nn+3). Odd levels absorb the (1,0) boost via the a-propagation; even levels
absorb the (1/2,1/2) via half_le_sigma_diff_at_r.
§16 g₃ assembly, t even (type7): g₂ + g₃ → g(2n'+1) + g(t+1), with g₂ = g⁺(2n'+2)
and g₃ = g⁻(2nn+2). Odd levels absorb the (1,0) via the a-propagation; even levels
(including the top j = t, gated by hYwin) absorb (1/2,1/2) via half_le_sigma_diff_at_r.