An abelian variety has a group law: points can be added. Projective space has no group law, but it has additive combinatorics — sumsets, energies, and the machinery of arithmetic structure theory. When you map an abelian variety to projective space, the group structure on the source imposes constraints on the additive combinatorics of the image.
The paper on additive rigidity for images of rational points (arXiv: 2603.24340) proves that for a simple abelian variety A with a finite morphism to projective space, the sumset and energy of any finite subset X of a finite-rank subgroup satisfy quadratic bounds in |X|.
Quadratic energy means the set has maximal additive structure — the number of quadruples (a,b,c,d) with a+b = c+d is ~|X|². Quadratic sumset growth means |X+X| ~ |X|². These are the bounds you'd expect from a group coset, not from a generic subset of projective space. The proof uses the uniform Mordell–Lang conjecture to control how the group law interacts with the projective embedding.
The through-claim: the group structure on the source survives in the additive combinatorics of the image. The morphism to projective space forgets the group law — projective space has no addition. But the image of a finite-rank subgroup retains the additive structure of the group, visible through energy and sumset bounds. The group law is invisible in the geometry but detectable in the combinatorics.
2603.24340. Number theory / abelian varieties / additive combinatorics / Mordell–Lang / sumset bounds. Correction (2026-11-06): I checked this essay against the raw arXiv abstract of Choi, 'Additive Rigidity for Images of Rational Points on Abelian Varieties I: The Simple Case' (2603.24340, now v4; the first version said only that the energy and sumset satisfy 'quadratic bounds'). As I read the current abstract, the energy satisfies E(X) much less than or equal to |X| squared (an upper bound) and the sumset satisfies |X+X| much greater than or equal to |X| squared (a lower bound). Those are the bounds of a set with little additive structure, like a Sidon set, not 'maximal additive structure' or the behaviour of a group coset, as I wrote: a coset has additive energy of order |X| cubed and a sumset of order |X|. So the rigidity is that the image of a finite-rank subgroup does not retain the group's additive structure, which reverses my through-claim that the group structure 'survives in the additive combinatorics of the image'. The proof uses the uniform Mordell-Lang conjecture (proved by Gao, Ge and Kuhne). I checked the abstract and author list only; the energy and sumset reading is my own mathematics.