The n-gon coefficient is one B-type polylogarithm
The paper gives an explicit formula for the n-gon graph contribution to the integrand of the cosmological wavefunction for conformally coupled φ³ theory in de Sitter space. The central statement is compact: the wavefunction coefficient ψ°_n of the n-gon graph, for n ≥ 1, is a single weight-n B-type quadrangular polylogarithm QLiB_n, with no additional products of lower-weight pieces or additive constants.
This is a statement about the loop-momentum integrand, not the integrated correlator. The authors verified the identification explicitly for the one-, two-, three- and four-gon graphs, corresponding to quadrangulations of the square, octagon, dodecagon and hexadecagon, and then proved it in general.
Total compatibility fixes the symbol alphabet
The result rests on the symbol of the n-gon wavefunction coefficient satisfying total compatibility with respect to the B_{2n-1} cluster algebra. Total compatibility is stronger than the cluster adjacency seen in N = 4 super-Yang-Mills amplitudes: every letter of a given word in the symbol lies in a single cluster.
The alphabet for polygon graphs consists of cluster variables of B-type cluster algebras, whereas chain graphs use A-type. This distinction is what makes the n-gon case a separate problem rather than a relabelling of the chain-graph result.
Centrally symmetric quadrangulations define the new functions
The authors introduce B-type quadrangular polylogarithms, constructed from centrally symmetric quadrangulations of a 4n-gon. The construction extends Matveiakin and Rudenko's A-type quadrangular polylogarithms by requiring every combinatorial object to be invariant under the involution a ↔ ā, with weighted symbols read off from the quotient by that involution.
A cluster algebra of type B_m can be obtained by folding one of type D_{m+1} by a Z₂ symmetry, realised combinatorially by triangulations of an auxiliary (2m+2)-gon invariant under a 180° rotation. Every quadrangulation of this auxiliary polygon that descends to the quotient has precisely one self-conjugate quadrangle. For that quadrangle the associated symbol letter is the square root of the cross-ratio, with a particular choice of sign, rather than the cross-ratio itself.
That square root is the structural reason the B-type functions are a genuinely new family and are not simply related to the A-type ones. The involution also supplies a canonical root for the dual tree, so no boundary marking is needed as it was in the chain-graph construction.
A differential recursion plus a soft limit fixes the answer
The proof relates the recursive differential equation satisfied by the n-gon wavefunction coefficients to a recursive coproduct formula for B-type quadrangular polylogarithms. The authors show that the operator K annihilates the quasi-shuffle product of all branches in the arborification, and separately that the contributions from left deletion, right deletion and weight-lowering terms cancel for each central polygon.
The differential recursion alone leaves an additive constant undetermined. That remaining constant is fixed by the soft limit, which is what upgrades the symbol-level match to a function-level identity. The paper notes that AI systems were used to draft the first version of the proof in Appendix C and to generate Mathematica files auditing the calculations, with all resulting material reviewed and verified by the authors.
The two-gon example shows the constant is right
For the 2-gon, the wavefunction coefficient can be written in terms of multiple polylogarithms, and the paper takes the kinematic limit explicitly. The four log² contributions cancel against one another, leaving a combination of Li₂ terms plus a −π²/6 term.
That surviving −π²/6 is the point of the exercise: it is a transcendental contribution the symbol cannot see, and it matches the correct expression for ψ₂. The authors state they explicitly checked the function-level relation up to n = 4.
A kinematic limit turns polygon graphs into chain graphs
For n ≥ 2, in a particular kinematic limit, the B-type quadrangular polylogarithm QLiB_n exactly degenerates to the combination of A-type quadrangular polylogarithms that compute the wavefunction coefficient ψ_n of the n-site chain graph. The same statement holds at the level of the functions, so it relates transcendental terms the symbol cannot detect.
This limit is the paper's internal bridge between its two graph families. It also clarifies why the polygon case is simpler: only the single weight-n QLiB_n function is needed, whereas the chain-graph construction required a pure weight-n linear combination of QLi functions and products thereof.
One-loop graphs, and cluster variables only for special cases
The construction covers the one-loop graphs considered here. The authors note it is natural to ask whether it extends to multi-loop diagrams or to more general graphs where n-gons appear as subgraphs of a larger graph. They also observe that while the symbol of any de Sitter wavefunction coefficient satisfies total compatibility with respect to its symbol entries, those entries are cluster variables only in the special cases of chain or polygon graphs.
Two further open problems are stated: integrating the formulas, for example ψ°_3, over the loop momentum, and studying whether the cluster-algebraic structure survives at the level of the cosmological correlators. On the mathematical side, the authors expect QLiB_n to be the lowest-weight member of an infinite sequence and leave the construction of higher-weight siblings, and the classification of weight-w functions satisfying total compatibility with respect to B_m for arbitrary w and m, to future work.
The decisive comparison is B-type versus A-type
The most useful way to read this paper is as a controlled comparison between two graph families that look similar but are governed by different cluster algebras. Chain graphs are controlled by A-type algebras and need a weight-n linear combination of A-type quadrangular polylogarithms plus products; n-gon graphs are controlled by B_{2n-1} and collapse to a single weight-n B-type function with nothing added.
The reason the two cannot be conflated is the self-conjugate quadrangle: its symbol letter is the square root of the cross-ratio, with a sign choice the authors describe as crucial. If a reader wants one threshold for judging follow-up work, it is whether a proposed extension preserves that square-root letter and the canonical root supplied by the involution, since those are exactly the features that make the B-type family new rather than a relabelling of the A-type one.
Sources used in this answer
de Sitter Wavefunction from Quadrangular Polylogarithms: Polygon Graphs
The n-gon graph contribution to the de Sitter cosmological wavefunction integrand for conformally coupled φ³ theory equals a single weight-n B-type quadrangular polylogarithm QLiB_n, whose symbol letters are B_{2n-1} cluster variables satisfying total compatibility, with no products of lower-weight pieces or additive constants.