First-hand information for everyone
Researchers have mapped the way certain mathematical objects called surface‑rigid algebras break into connected components. The key finding is that the support τ‑tilting graph of a surface‑rigid algebra is disconnected only when the algebra is weakly symmetric, has at least two simple modules, and its Cartan matrix has rank one. In that special case the graph consists of exactly two connected portions.
The study examined algebras that arise from partial tagged triangulations of punctured surfaces with a boundary. By analysing the algebraic data, the authors proved that if the algebra is connected but meets the three conditions above, the graph splits into two pieces; otherwise the graph stays whole.
For any non‑zero τ‑rigid pair u over a connected surface‑rigid algebra, the completion graph F A(u) is always connected. The number of components is twice the number of blocks of the algebra on which both parts of u vanish. This result gives a clear rule for how smaller building blocks of the algebra influence the overall connectivity.
The team also linked τ‑tilting finiteness—a property indicating that only finitely many τ‑tilting modules exist—to the geometry of the signed endpoint graph of the triangulation. They showed that τ‑tilting finiteness is equivalent to the Cartan matrix being positive definite, which happens precisely when every component of the signed endpoint graph is a tree or a single unbalanced cycle. When τ‑tilting finiteness holds, the g‑fan, a geometric object encoding mutation relations, is complete.
However, the paper cautions that when a weakly symmetric algebra with Cartan rank one and at least two simple modules appears, its support τ‑tilting graph will have exactly two connected components and no further fragmentation occurs. Thus, under these conditions, the algebra’s mutation behaviour is fully captured by two distinct classes.
These results deepen the understanding of how combinatorial surfaces govern representation theory and clarify the precise circumstances under which mutation graphs split.