lemmaroutine

Hessian cofactor is column-divergence-free

Shows Hessian cofactor C(H(u))C(H(u)) has divergence-free columns
enables divergence-form integration by parts of the determinant
Lemma·lem:cofactor-divergence·§3.3 Column divergence and the boundary term

Statement (paper)

For k=3k=3 or 44 and a smooth real uu, the Hessian cofactor satisfies ∑a,αeˉα ∂aαC(H(u))ab=0(1≤b≤m). \sum_{a,\alpha}\bar e_\alpha\,\partial_{a\alpha} C(H(u))_{ab}=0 \quad(1\le b\le m).

In plain English

For k=3k=3 or 44 and smooth real uu, the cofactor matrix C(H(u))C(H(u)) of the quaternionic Hessian is divergence-free in each column: ∑a,αeˉα ∂aαC(H(u))ab=0\sum_{a,\alpha}\bar e_\alpha\,\partial_{a\alpha} C(H(u))_{ab}=0 for 1≤b≤m1\le b\le m.

Why it's here

This is the quaternionic analogue of the classical fact that the cofactor matrix of a real Hessian is divergence-free. It lets det⁡H\det H be written in divergence form after integrating by parts against C(H(u))C(H(u)), which is what the energy balance and the determinant-energy computations on fibers (the Green identity with M=C(H(⋅))M=C(H(\cdot))) require. The proof rests on the cofactor being a polynomial derivative of the Moore determinant (the quaternionic matrix lemma) and on the Hermitian structure of H(u)H(u) (the Hessian calculus lemma).
Classified routine:
Quaternionic analogue of the classical divergence-free property of the cofactor matrix of a real (or complex) Hessian, as used for real and complex Monge–Ampère operators. It follows from the Moore determinant cofactor properties and the Hermitian structure of the quaternionic Hessian established in the earlier lemmas.

Sources

PaperOpen
Section
§3.3 Column divergence and the boundary term
Page
—
Label
lem:cofactor-divergence
Line
sections/calculus.tex:188
Leannot aligned

Phase 2 aligns this node to a Lean declaration.