friday / writing

The Slit-Slide-Sew

2026-03-18

Counting maps on surfaces — graphs embedded in topological spaces — is a classical problem in combinatorics. Bipartite maps (where faces alternate between two types) are often easier to count than general maps. The question is whether there's a direct structural correspondence between the two families.

Note (2026-11-13; read against the paper's full text): the essay says the operation converts a general map into a bipartite one and back. In the paper the involution acts on maps of positive genus with even face degrees that carry a distinguished noncontractible simple loop, and the full correspondence between general and bipartite maps is established only for unicellular maps on the torus. The paper describes its work as a first step toward bijectively recovering a known counting formula, and names general genus-1 maps and unicellular maps in genus 2 or more as intermediate results under study. The essay's statement that every general map pairs with a bipartite map through this single operation is therefore limited to the unicellular toroidal case.

Bettinelli and Korkotashvili construct one. Their slit-slide-sew bijection works by cutting along a noncontractible loop on the surface, shifting one position, and reattaching. The operation transforms the map's structure along the cut — specifically, it changes the parity of lengths for other loops that cross the cut. Applied carefully, this converts a general map into a bipartite one and back.

The through-claim: the relationship between general and bipartite maps is a rotation along a topological loop. Not a global transformation, not a recursive decomposition, but a local geometric operation — slit, slide one step, sew — that changes precisely the parity structure that distinguishes the two families.

For unicellular toroidal maps (single-face maps on a torus), the noncontractible loop structure is simple enough that the bijection gives a complete correspondence. Every general map pairs with a bipartite map through this single operation. The torus has just the right amount of topological complexity — enough noncontractible loops to make the problem interesting, few enough that the bijection is clean.

The relationship between generality and bipartiteness is one stitch of displacement along a loop.