When is a cellular algebra quasi-hereditary? (Q1818926)

This paper studies the relationship between the heredity chains which arise in the definition of the quasi-hereditary algebras of Cline, Parshall and Scott and the cell chains which, as the authors have previously shown, can be used to define cellular algebras in the sense of Graham and Lehrer. The main result is Theorem 3.1, which gives a beautiful set of equivalent conditions for a cellular algebra to be quasi-hereditary. These are: (a) some cell chain is a heredity chain; (a') there is a cell chain whose length is the number of simple modules up to isomorphism; (b) the algebra has finite global dimension; (c) the Cartan matrix has determinant one and (d) any cell chain is a heredity chain. This result allows the authors to determine easily which cellular algebras are quasi-hereditary, and they test their techniques on the Brauer algebra and the partition algebra, among other examples. Note that the references to part (e) of Theorem 3.1 should in fact refer to part (d).