The relative $h$-principle for closed $\mathrm{SL}(3;\mathbb{R})^2$ 3-forms
Laurence H. Mayther
TL;DR
This work proves the relative $h$-principle for closed $\mathrm{SL}(3;\mathbb{R})^2$ 3-forms on oriented 6-manifolds by employing convex integration with avoidance. By reducing the problem to a fibred differential relation $\mathscr{R}_+(a)$, the authors navigate the non-ampleness obstacle through a novel generic-hyperplane analysis and the concept of macilence, establishing ampleness in a generic setting. Consequently, any manifold that admits an $\mathrm{SL}(3;\mathbb{R})^2$ 3-form represents every $3$-class with such a form, and the Hitchin functional on each class is unbounded above. These results extend the scope of stable forms known to satisfy the relative $h$-principle and illuminate the interaction between rank-3 distributions and hyperplane data in 6-manifolds, with implications for calibrated geometry and special holonomy.
Abstract
This paper uses convex integration with avoidance and transversality arguments to prove the relative $h$-principle for closed $\mathrm{SL}(3;\mathbb{R})^2$ 3-forms on oriented 6-manifolds. As corollaries, it is proven that if an oriented 6-manifold $\mathrm{M}$ admits any $\mathrm{SL}(3;\mathbb{R})^2$ 3-form, then every degree 3 cohomology class on $\mathrm{M}$ can be represented by an $\mathrm{SL}(3;\mathbb{R})^2$ 3-form and, moreover, that the corresponding Hitchin functional on $\mathrm{SL}(3;\mathbb{R})^2$ 3-forms representing this class is necessarily unbounded above. Essential to the proof of the $h$-principle is a careful analysis of the rank 3 distributions induced by an $\mathrm{SL}(3;\mathbb{R})^2$ 3-form and their interaction with generic pairs of hyperplanes. The proof also introduces a new property of sets in affine space, termed macilence, as a method of verifying ampleness.
