Cohomology is a mathematical tool that assigns algebraic data, like groups or rings, to geometric spaces in order to capture their shape and structure. When a space has a symmetry group acting on it, one can refine this tool to produce "equivariant cohomology," which keeps track of both the shape of the space and how the symmetry acts. A natural expectation, built from experience with many well-behaved spaces, is that equivariant cohomology should only be nonzero in even degrees. Odd-degree equivariant cohomology is considered exotic and surprising, and while a few isolated examples have been known, they have mostly been used just to confirm that odd-degree phenomena can exist at all, not to illuminate any broader pattern.
The paper constructs an explicit infinite family of geometric spaces, called toric varieties, that all exhibit nontrivial odd-degree equivariant cohomology in many different odd degrees simultaneously, not just degree one. These spaces arise in a natural and structured way: they are the orbits traced out by a symmetry group called a torus acting on certain points inside a Grassmannian, which is a classical and well-studied geometric object that parametrizes flat subspaces of a given dimension inside a higher-dimensional space. The authors also show that for small cases, they can compute this odd equivariant cohomology explicitly when working with rational numbers as coefficients, giving concrete and calculable examples of this unusual behavior.
This matters because it moves the phenomenon of odd equivariant cohomology from a curiosity to a systematic feature of a recognizable class of spaces. Having a whole family of examples, rather than one-off constructions, suggests there is genuine geometric structure behind the oddness, and it opens the door to understanding why and when such behavior occurs. It also raises new questions about how well-studied objects like Grassmannians can harbor unexpected algebraic complexity in the spaces associated with them.