[Paper Review] ARI, GARI, Zig and Zag: An introduction to Ecalle's theory of multiple zeta values

Leila Schneps|arXiv (Cornell University)|Jul 6, 2015

Advanced Mathematical Identities12 citations

TL;DR

This paper provides a comprehensive, accessible introduction to Jean Ecalle's mould-theoretic framework for multiple zeta values (MZVs), translating his advanced formalism into the context of classical MZV theory. It establishes rigorous proofs of key identities—particularly those linking the double shuffle relations to the ARI (anti-rhizomatic) and GARI (generalized ARI) groups—offering a systematic bridge between Ecalle's mould language and the standard non-commutative algebra of MZVs.

ABSTRACT

This text has two goals. The first is to give an introduction to Ecalle's work on mould theory, multiple zeta values and double shuffle theory and relate this work explicitly to the classical theory of multiple zeta values and double shuffle expressed in the usual terms of non-commutative variables. The second is to provide complete proofs of those of his main results and identities which are strictly useful in the context of (non-colored) multiple zeta values. Many of these proofs are difficult, laborious and not enlightening and have been relegated to appendices. The emphasis in the text is to provide an easily approachable introduction to Ecalle's language while placing it almost from the start in the context of multiple zeta value theory. Disclaimer: This text is not final and is not submitted for publication. The intention is to continue to add to and complete it over time.

Motivation & Objective

• To make Jean Ecalle's advanced mould-theoretic framework for multiple zeta values accessible to researchers unfamiliar with his formalism.
• To explicitly connect Ecalle's ARI and GARI groups to the classical theory of multiple zeta values and double shuffle relations.
• To provide complete, rigorous proofs of core identities in Ecalle's theory that are essential for non-colored multiple zeta values.
• To place Ecalle's language in the context of standard non-commutative algebra over the alphabet {x,y}, enabling direct application to MZV research.
• To clarify the role of shuffle and stuffle regularizations in extending zeta values to all words in the non-commutative algebra.

Proposed method

• Uses the shuffle and stuffle products in the non-commutative algebra Q⟨x,y⟩ to define and regularize multiple zeta values via the Drinfel'd associator Φ.
• Introduces the ARI and GARI groups as algebraic structures that encode the double shuffle relations through mould-theoretic operations.
• Applies the mould pair (pal, pil) to analyze symmetries and derive identities in the context of multiple zeta values.
• Employs recursive definitions of shuffle and stuffle products via surjective maps and weighted sums over shuffles, formalizing the regularization process.
• Uses projection operators π and π_y to isolate convergent words and define regularized zeta values ζ*(v) via the corrected associator Φ*.
• Relies on advanced combinatorial tools such as Bernoulli numbers, binomial coefficients, and matrix identities to verify key identities in the appendix.

Experimental results

Research questions

• RQ1 How can Ecalle's mould-theoretic language be systematically translated into the standard framework of multiple zeta values over non-commutative variables?
• RQ2 What is the precise algebraic relationship between the double shuffle relations and the ARI/GARI groups in the context of multiple zeta values?
• RQ3 How do the shuffle and stuffle regularizations of multiple zeta values relate to each other and to the Drinfel'd associator?
• RQ4 What is the role of the mould pair (pal, pil) in encoding symmetries and identities in the double shuffle Lie algebra?
• RQ5 How can the deep identities in Ecalle's theory—particularly those involving Bernoulli numbers and binomial coefficients—be rigorously proven in the context of non-colored MZVs?

Key findings

• The shuffle regularization of multiple zeta values satisfies ζ(sh(u,v)) = ζ(u)ζ(v) for all words u,v ∈ Q⟨x,y⟩, establishing the shuffle algebra structure.
• The stuffle regularization extends the product relation ζ*(st(u,v)) = ζ*(u)ζ*(v) to all words in the additive alphabet {y_i}, with ζ*(v) defined via the corrected associator Φ*.
• For convergent words v (not starting with y₁), the regularized value ζ*(v) coincides with the classical multiple zeta value ζ(v).
• The stuffle-regularized value ζ*(1,1) is equal to −1/2, a polynomial in single zeta values, illustrating the non-trivial nature of the regularization.
• The identity der(dupal) − dur(dapal) = 0 is proven via a detailed combinatorial analysis involving Bernoulli numbers and binomial coefficients, with the equality verified in depth d when d is odd.
• The coefficient of u_j in the linear factor of the depth-d part of der(dupal) − dur(dapal) matches the expression derived from the Bernoulli number identity, confirming the identity in the case of odd d.