Integral table algebras, affine diagrams, and the analysis of degree two (Q1906653)

Integral table algebras are commutative algebras over the complex numbers with a distinguished basis which satisfies a few rather primitive axioms. They possess many common features of character rings and centers of group algebras of finite groups, and of combinatorial structures such as association schemes. In particular, there is a homomorphism from the algebra to the complex numbers which assigns to each basis element a positive integer called its degree. We show in this paper that the assumption of the existence of a faithful basis element of degree 2 yields conclusions on the structure of such algebras which go a long way toward a complete classification. Theorem 1 adopts the rather natural further hypothesis that there are no linear basis elements of degree \(2^m\) for any \(m\geq 0\), and explicitly determines all such integral table algebras. As corollaries, extensive information is obtained even when the assumption on linear elements does not necessarily hold. In Theorem 2 we assume instead that the given element of degree 2 is real (an abstraction of the real-valued character of a self-dual representation of a group). We proceed in this situation by assigning to each algebra a labeled graph which turns out to be the underlying graph of a generalized Cartan matrix with an additive function. These structures are defined, and a classification is given, in [\textit{D. Happel, U. Preiser, C. M. Ringel}, Manuscr. Math. 31, 317-329 (1980; Zbl 0436.20005)]. The resulting 20 graphs are called the generalized Euclidean diagrams, and the 16 of these where no edge is a loop are known as the affine diagrams. (We give the definitions and display all the diagrams in Section 5 and 6.) We show that exactly 13 of the diagrams yield a total of 14 distinct integral table algebras and thereby achieve a complete determination. The last section of the paper examines a slightly more general method of assigning labeled graphs to integral table algebras which are themselves slightly more general than those in Theorem 2. The conclusion, stated in Theorem 3, is that 19 of the 20 diagrams are obtainable in this way. This result, with its proof, encompasses the realization of the affine diagrams from finite subgroups of \(\text{SL}(2,\mathbb{C})\). So the framework of integral table algebras broadens (and, we hope, illuminates) the context of the link between affine diagrams and finite linear groups of degree two.