Quantum state systems that count perfect matchings
Scott Baldridge, Ben McCarty
TL;DR
The paper advances a categorification program for Penrose-inspired colorings by constructing bigraded $n$-color vertex homology and a filtered variant via a spectral sequence, with the graded Euler characteristic recovering the $n$-color vertex polynomial $\llangle \Gamma\rrangle_{n}$. It establishes a deep link between quantum-state–style constructions and combinatorial face colorings on ribbon graphs, enabling, in the $n=2$ case, a homological count of perfect matchings and a broader interpretation in terms of partial $n$-face colorings. The vertex polynomial $V(\Gamma,n)$ and the total matching polynomial $TM(\Gamma,n)$ emerge as convenient encodings of these homological invariants, and the theory extends to $4$-regular graphs with analogous structures. Overall, the work provides a robust, algebraic framework for understanding colorings of ribbon graphs, connecting Penrose-type formulas to topological quantum field theory and suggesting avenues toward new graph and manifold invariants. The constructions yield computable invariants and open pathways to relate colorings to harmonic/color-based decompositions and potential spin-network–style interpretations in topology.
Abstract
In this paper we show how to categorify the $n$-color vertex polynomial, which is based upon one of Roger Penrose's formulas for counting the number of $3$-edge colorings of a planar trivalent graph. Using topological quantum field theory (TQFT), we introduce a quantum state system to build a new bigraded theory called the bigraded $n$-color vertex homology. The graded Euler characteristic of this homology is the $n$-color vertex polynomial. We then produce a spectral sequence whose $E_\infty$-page is a filtered theory called filtered $n$-color vertex homology and show that it is generated by certain types of face colorings of ribbon graphs. For $n=2$, we show that the filtered $n$-color vertex homology is generated by face colorings that correspond to perfect matchings. Finally, we introduce and give meaning to what the vertex polynomial counts when $n \geq 2$. This polynomial is a new abstract graph invariant that can be inferred from certain formulas of Penrose.
