Counting in Calabi--Yau categories, with applications to Hall algebras and knot polynomials

TL;DR

This paper develops a framework for defining Hall algebras in Calabi--Yau categories, resolving longstanding problems and connecting to knot invariants through Legendrian knot theory.

Contribution

It introduces a new intrinsic Hall algebra construction for Calabi--Yau categories and proves a conjecture relating Legendrian knot invariants to augmentation categories.

Findings

Defined homotopy cardinality for evenCY categories.

Constructed intrinsic Hall algebras for oddCY categories.

Proved a conjecture linking Legendrian invariants to augmentation categories.

Abstract

We show that homotopy cardinality -- a priori ill-defined for many dg-categories, including all periodic ones -- has a reasonable definition for even-dimensional Calabi--Yau (evenCY) categories and their relative generalizations (under appropriate finiteness conditions). As a first application we solve the problem of defining an intrinsic Hall algebra for degreewise finite pre-triangulated dg-categories in the case of oddCY categories. We compare this definition with To\"en's derived Hall algebras (in case they are well-defined) and with other approaches based on extended Hall algebras and central reduction, including a construction of Hall algebras associated with Calabi--Yau triples of triangulated categories. For a category equivalent to the root category of a 1CY abelian category $\mathcal A$, the algebra is shown to be isomorphic to the Drinfeld double of the twisted Ringel--Hall algebra of $\mathcal A$, thus resolving in the Calabi--Yau case the long-standing problem of realizing the latter as a Hall algebra intrinsically defined for such a triangulated category. Our second application is the proof of a conjecture of Ng--Rutherford--Shende--Sivek, which provides an intrinsic formula for the ruling polynomial of a Legendrian knot $L$, and its generalization to Legendrian tangles, in terms of the augmentation category of $L$.