A noncommutative manifold whose diffeomorphism group is the Connes–Kreimer character group, and a reconstruction formula for its elements

The Connes–Kreimer Hopf algebra $\mathcal{H}_R$ of rooted trees encodes the forest combinatorics of subdivergences in perturbative renormalization and is the algebra of coordinates on a nilpotent formal group $\mathcal{G}$ (the Butcher group) whose Lie algebra $\mathcal{L}_1$ extends to a Lie algebra $\mathcal{L}$ that maps onto the Lie algebra of formal vector fields on the line; renormalization is inversion in $\mathcal{G}$. The 1998 paper that established this closed by asking for its geometric home: find a noncommutative manifold $X$ whose diffeomorphism group is $\mathcal{G}$, and find an explicit formula reconstructing elements of $\mathcal{G}$ from coordinates, analogous to the classical formula $\psi(x)=\int_0^x\exp(-\sum_{n\ge 1}\delta_n u^n/n!)\,du$ for formal diffeomorphisms, with the Feynman integral $\int\exp(-L_0+\sum_\Gamma L_\Gamma)$ as the tantalizing analogue. Of the three tasks announced there, the detailed treatment of overlapping divergences was published by Kreimer in 1999 and is settled. Later work gave $\mathcal{G}$ other geometric meanings — a group of diffeographisms acting on the space of coupling constants (with a scattering formula reconstructing counterterms from the residue) and a motivic Galois group — but no construction of $X$ and no general reconstruction formula for $\mathcal{G}$ from rooted-tree coordinates is known. This question asks for such a construction and/or formula, which would give the renormalization group of characters a genuine differential-geometric realization.

Background

Perturbative renormalization in quantum field theory removes ultraviolet divergences from Feynman amplitudes by adding local counterterms, one for each divergent graph, organized by the nested or disjoint structure of its divergent subgraphs (subdivergences). Connes and Kreimer (1998) showed that this combinatorics is encoded in a commutative Hopf algebra $\mathcal{H}_R$ of rooted trees. A Hopf algebra is here a commutative algebra $A$ equipped with an algebra homomorphism $\Delta:A\to A\otimes A$ (the coproduct, describing how an element decomposes into parts), a counit, and an antipode $S$ (implementing inversion). A rooted tree is a finite connected graph without cycles with a distinguished vertex, the root; $\mathcal{H}_R$ is the polynomial algebra on isomorphism classes of finite rooted trees, with coproduct defined by admissible cuts (sets of edges with at most one cut edge on each root-to-leaf path):

$$\Delta(\delta_T)=\delta_T\otimes 1+1\otimes\delta_T+\sum_c\Big(\prod_i\delta_{T_i}\Big)\otimes\delta_{R_c(T)},$$

where the $T_i$ are the cut branches and $R_c(T)$ the remaining trunk. Admissible cuts record the possible divergent subgraphs of a Feynman graph, and the antipode reproduces the classical forest formula for counterterms (the recursive subtraction of subdivergences).

Two structural results of that paper matter here. First (their Theorem 3), $\mathcal{H}_R$ is the graded dual of the universal enveloping algebra of a graded Lie algebra $\mathcal{L}_1$ spanned by elements $Z_T$ indexed by rooted trees, with bracket built from grafting of trees (the pre-Lie structure of rooted trees: a product satisfying $(a\cdot b)\cdot c-a\cdot(b\cdot c)=(b\cdot a)\cdot c-b\cdot(a\cdot c)$, whose antisymmetrization is a Lie bracket). Consequently $\mathcal{H}_R$ is the algebra of coordinates on a nilpotent formal group $\mathcal{G}$ — a group whose elements are described by formal power series in the tree coordinates, with the group law terminating degree by degree — whose points over a field $K$ are the characters (algebra homomorphisms) $\mathcal{H}_R\to K$. This group is the Butcher group, introduced in the numerical analysis of one-step integrators (Runge–Kutta methods) for ordinary differential equations, whose elements are formal series indexed by rooted trees (B-series). In the renormalization picture, the unrenormalized Feynman rules of a regularized theory (e.g. by dimensional regularization, which shifts the spacetime dimension to $D=4-2\epsilon$ so that amplitudes become Laurent series in $\epsilon$) define such a character, and renormalization is inversion in $\mathcal{G}$.

Second, the paper relates this renormalization side to noncommutative geometry (NCG), the theory that studies 'quantized' spaces through noncommutative algebras of coordinates, typically via spectral triples — the NCG replacement for a Riemannian spin manifold, given by an algebra represented on a Hilbert space together with an unbounded Dirac-type operator. The Hopf algebra $\mathcal{H}_1$ of Connes and Moscovici, which governs the transverse geometry of codimension-one foliations (the Hopf-theoretic symmetry of the leaf space, arising in their transverse index theorem), is identified (Theorem 8 of the paper, due to Connes–Moscovici) with the Hopf algebra of coordinates on the group $G_2$ of formal diffeomorphisms of $\mathbb{R}$ tangent to the identity, $\psi(x)=x+o(x)$: the generators $\delta_n$ correspond to the functionals $\gamma_n(\psi^{-1})=(\partial_x^n\log\psi')(0)$, the Taylor coefficients of the logarithmic derivative. Inside $\mathcal{H}_R$, the iterates $\delta_k=N^k(e)$ of the natural-growth operator $N$ (which grafts a new vertex onto a rooted tree in all possible ways, starting from the one-vertex tree $e$) close under the coproduct and realize these affine coordinates. Extending the Lie algebra $\mathcal{L}_1$ by two generators (one implementing $N$) yields a Lie algebra $\mathcal{L}$ that surjects onto the Lie algebra of formal vector fields on the line, so that $\mathcal{G}$ is an extension of the group of formal diffeomorphisms.

The paper ends with the geometric interpretation of this state of affairs. For $G_2$ itself one has the classical reconstruction formula: a formal diffeomorphism tangent to the identity is reconstructed from the coordinates $\delta_n=-(\log\psi')^{(n)}$ by

$$\psi(x)=\int_0^x\exp\Big(-\sum_{n\ge 1}\frac{\delta_n}{n!}u^n\Big)\,du,$$

which makes composition and inversion transparent. The authors write that "it would be desirable to find a non-commutative manifold $X$, whose diffeomorphism group is $\mathcal{G}$", and, of the formula, that "we would love to have a similar formula for the group $\mathcal{G}$ and it is tantalizing to consider the Feynman integral $\int\exp(-L_0+\sum_\Gamma L_\Gamma)$ as a direct analogue of the above expression", where $L_0$ is the original Lagrangian and the $L_\Gamma$ are the counterterm Lagrangians labeled by Feynman graphs $\Gamma$. The same paper's appendix sketched how overlapping divergences (divergent subgraphs that are neither nested nor disjoint) reduce to sums of rooted trees, with the detailed treatment announced as future work.

Problem Statement

Let $\mathcal{G}$ be the group of characters of the Connes–Kreimer Hopf algebra $\mathcal{H}_R$ of rooted trees (equivalently, the Butcher group), with graded Lie algebra $\mathcal{L}_1$ extended by two generators (one of them implementing the natural-growth operator $N$) to the Lie algebra $\mathcal{L}$, and let $L_0$ and $\{L_\Gamma\}_\Gamma$ denote the original and counterterm Lagrangians of a renormalizable perturbative quantum field theory. The open question has two linked components, serving one objective — giving $\mathcal{G}$ the geometric meaning of a diffeomorphism group together with its explicit coordinate calculus:

1. (Geometric realization) Construct a noncommutative manifold $X$ — a specified algebra or spectral triple in the sense of noncommutative geometry — together with a definition of its group of diffeomorphisms that is intrinsic to the chosen algebra or spectral triple, reduces to the classical diffeomorphism group when $X$ is a smooth manifold, and does not presuppose $\mathcal{G}$ or its points, such that $\mathrm{Diff}(X)$ is $\mathcal{G}$, thereby giving the extended Lie algebra $\mathcal{L}$ and the natural-growth operator $N$ a geometric meaning as an extension of the formal-diffeomorphism calculus on the line.

2. (Reconstruction formula) Provide an explicit closed-form formula reconstructing an element $g\in\mathcal{G}$ from tree-indexed coordinates, analogous to the classical formula

$$\psi(x)=\int_0^x\exp\Big(-\sum_{n\ge 1}\frac{\delta_n}{n!}u^n\Big)\,du$$

for formal diffeomorphisms tangent to the identity, including — if possible — a precise statement of the sense in which the functional integral $\int\exp(-L_0+\sum_\Gamma L_\Gamma)$ defines an element, or coordinates, of $\mathcal{G}$.

Each component is a question in its own right, independent of the other.

References

1. Alain Connes and Dirk Kreimer, Hopf Algebras, Renormalization and Noncommutative Geometry, Commun. Math. Phys. 199 (1998) 203–242, DOI 10.1007/s002200050499, https://doi.org/10.1007/s002200050499 (arXiv:hep-th/9808042, https://arxiv.org/abs/hep-th/9808042)
2. Dirk Kreimer, On Overlapping Divergences, Commun. Math. Phys. 204 (1999) 669–689, DOI 10.1007/s002200050661, https://doi.org/10.1007/s002200050661 (arXiv:hep-th/9810022, https://arxiv.org/abs/hep-th/9810022)
3. Alain Connes and Dirk Kreimer, Renormalization in quantum field theory and the Riemann–Hilbert problem I: the Hopf algebra structure of graphs and the main theorem, Commun. Math. Phys. 210 (2000) 249–273, DOI 10.1007/s002200050779, https://doi.org/10.1007/s002200050779 (arXiv:hep-th/9912092, https://arxiv.org/abs/hep-th/9912092)
4. Alain Connes and Dirk Kreimer, Renormalization in quantum field theory and the Riemann–Hilbert problem II: the β-function, diffeomorphisms and the renormalization group, Commun. Math. Phys. 216 (2001) 215–241, DOI 10.1007/PL00005547, https://doi.org/10.1007/PL00005547 (arXiv:hep-th/0003188, https://arxiv.org/abs/hep-th/0003188)
5. Alain Connes and Matilde Marcolli, Renormalization and motivic Galois theory, 2004, arXiv:math/0409306, https://arxiv.org/abs/math/0409306 (published as: Renormalization, the Riemann–Hilbert Correspondence, and Motivic Galois Theory, in Frontiers in Number Theory, Physics, and Geometry II, Springer, 2007, pp. 617–713, DOI 10.1007/978-3-540-30308-4_13, https://doi.org/10.1007/978-3-540-30308-4_13)
6. Spencer Bloch, Hélène Esnault and Dirk Kreimer, On Motives Associated to Graph Polynomials, Commun. Math. Phys. 267 (2006) 181–225, DOI 10.1007/s00220-006-0040-2, https://doi.org/10.1007/s00220-006-0040-2 (arXiv:math/0510011, https://arxiv.org/abs/math/0510011)
7. Jean-Louis Loday and Nikolay M. Nikolov, Operadic construction of the renormalization group, 2012, arXiv:1202.1206, https://arxiv.org/abs/1202.1206
8. Frédéric Fauvet and Frédéric Menous, Écalle's arborification-coarborification transforms and Connes–Kreimer Hopf algebra, Ann. Sci. Éc. Norm. Supér. (4) 50 (2017), fasc. 1, 39–83, DOI 10.24033/asens.2315, https://doi.org/10.24033/asens.2315 (open-access full text at https://hal.science/hal-00767373)
9. William G. Faris, Rooted tree graphs and the Butcher group: combinatorics of elementary perturbation theory, 2021, arXiv:2101.09364, https://arxiv.org/abs/2101.09364 (published as a chapter in Sojourns in Probability Theory and Statistical Physics II, Springer Proceedings in Mathematics & Statistics 299, Springer, Singapore, 2019, pp. 135–166, DOI 10.1007/978-981-15-0298-9_6, https://doi.org/10.1007/978-981-15-0298-9_6)
10. Dirk Kreimer, Chen's Iterated Integral represents the Operator Product Expansion, 1999, arXiv:hep-th/9901099, https://arxiv.org/abs/hep-th/9901099
11. Alain Connes and Henri Moscovici, Hopf Algebras, Cyclic Cohomology and the Transverse Index Theorem, Commun. Math. Phys. 198 (1998) 199–246, DOI 10.1007/s002200050477, https://doi.org/10.1007/s002200050477 (arXiv:math/9806109, https://arxiv.org/abs/math/9806109)