Alekseev–Enriquez–Torossian Construction

• Alekseev–Enriquez–Torossian construction is an explicit passage mapping Drinfeld associators to symmetric Kashiwara–Vergne solutions, encapsulating key structures in Lie algebra theory.
• It reformulates the Kashiwara–Vergne problem using tangential derivations, cyclic quotients, and Jacobian cocycles, establishing an injective map from GRT₁ to KRV.
• The construction underpins operadic and topological reformulations, linking formality morphisms, graph cocycles, and deformation quantization in a unified framework.

Searching arXiv for recent and foundational papers on the Alekseev–Enriquez–Torossian construction and its later reformulations. arXiv search query: "Alekseev Enriquez Torossian associators Kashiwara Vergne" The Alekseev–Enriquez–Torossian construction is the explicit passage from Drinfeld associators to solutions of the Kashiwara–Vergne equations. In the formulations used in later work, any solution of Drinfeld’s associator equations gives rise to a solution of the Kashiwara–Vergne equations in an explicit way, and the resulting solutions are the symmetric Kashiwara–Vergne solutions arising from associators (Kuno, 3 Apr 2025). The construction occupies the interface between associator theory, the Grothendieck–Teichmüller group, tangential automorphisms of completed free Lie algebras, and deformation quantization; it also underlies the injective map from $\mathsf{GRT}_1$ to $\mathsf{KRV}$, later operadic and topological reformulations, and the role of the Alekseev–Torossian associator in formality theory (Dancso et al., 2022).

1. Algebraic setting

The Kashiwara–Vergne side of the construction is formulated in the degree-completed free Lie algebra $Lie(x,y)$, its completed universal enveloping algebra $A$, and the cyclic quotient

$cyc \coloneqq A/[A,A].$

A tangential derivation is a derivation $u\in tder$ such that

$u(x)=[x,u_1],\qquad u(y)=[y,u_2]$

for some $u_1,u_2\in Lie(x,y)$. Tangential derivations exponentiate to tangential automorphisms $TAut$. The noncommutative divergence is

$j:tder\to cyc,$

and its integrated version is the Jacobian

$\mathsf{KRV}$0

with cocycle property

$\mathsf{KRV}$1

A Kashiwara–Vergne solution is then a pair $\mathsf{KRV}$2 satisfying

$\mathsf{KRV}$3

and

$\mathsf{KRV}$4

In this form, the Kashiwara–Vergne problem is recast as a problem about tangential automorphisms and a Jacobian cocycle rather than in its original formulation (Dancso et al., 2023).

On the associator side, a Drinfeld associator is a solution of the pentagon and hexagon equations in a completed algebraic setting. The Grothendieck–Teichmüller group $\mathsf{KRV}$5 acts freely and transitively on the set of Drinfeld associators, so any two associators differ by a unique element of $\mathsf{KRV}$6. Its Lie algebra is $\mathsf{KRV}$7, and $\mathsf{KRV}$8 because the group is pro-unipotent (Felder, 2016). This torsor structure is the symmetry-theoretic background for the Alekseev–Enriquez–Torossian passage from associators to Kashiwara–Vergne data.

2. Explicit construction from associators to Kashiwara–Vergne solutions

A standard later formulation states that the combination of results of Alekseev–Torossian and Alekseev–Enriquez–Torossian demonstrates that solutions to the Kashiwara–Vergne equations can be explicitly constructed from Drinfeld associators. In this terminology, the Kashiwara–Vergne solutions produced by the construction are the symmetric Kashiwara–Vergne solutions, namely those arising from associators via the Alekseev–Enriquez–Torossian map (Dancso et al., 2023).

At the level of symmetry groups, the construction yields an injective group homomorphism

$\mathsf{KRV}$9

This is the precise formula quoted in later topological treatments of the subject (Dancso et al., 2022). The formula expresses the passage from a Grothendieck–Teichmüller element to tangential-automorphism data normalized by the constraint $Lie(x,y)$0. A common structural reading is that associators form a bi-torsor under $Lie(x,y)$1, Kashiwara–Vergne solutions form a bi-torsor under $Lie(x,y)$2, and the Alekseev–Enriquez–Torossian construction links the two torsor pictures by an explicit map of symmetry groups (Dancso et al., 2022).

The computational significance of the construction is repeatedly emphasized in later work. The Kashiwara–Vergne equations live in smaller algebraic spaces than associator equations, so degree-by-degree work on the Kashiwara–Vergne side can be used to simplify the computation of associators through the Alekseev–Enriquez–Torossian bridge. This suggests a practical strategy: construct or extend symmetric Kashiwara–Vergne solutions and then use the correspondence with associators (Dancso et al., 2023).

3. Operadic and topological reformulations

A later moperadic reformulation treats the construction as an isomorphism problem between the moperad of parenthesized braids with a frozen strand and the moperad of chord diagrams. In that language, any isomorphism between $Lie(x,y)$3 and $Lie(x,y)$4 gives rise to a family of genus zero Kashiwara–Vergne solutions operadically generated by a single classical Kashiwara–Vergne solution. Conversely, any symmetric Kashiwara–Vergne solution gives rise to a morphism from the moperad of parenthesized braids with a frozen strand to the moperad of tangential automorphisms of free Lie algebras, and this morphism factors through the moperad of chord diagrams if and only if the associated Kashiwara–Vergne associator is a Drinfeld associator (Dancso et al., 22 Jul 2025).

In this reformulation, the frozen strand supplies the extra cyclotomic-type generator needed to extract Kashiwara–Vergne data from an associator-valued operadic isomorphism. The binary solution $Lie(x,y)$5 obtained from the frozen-strand generator is shown to satisfy the Kashiwara–Vergne equations, and the higher genus-zero solutions are produced by operadic composition. A plausible implication is that the original construction is not merely a formula on generators of $Lie(x,y)$6, but an operadic extension principle in which one binary solution generates a full genus-zero family (Dancso et al., 22 Jul 2025).

A complementary topological and diagrammatic reinterpretation uses homomorphic expansions. Homomorphic expansions of parenthesized braids are in one-to-one correspondence with Drinfeld associators, while homomorphic expansions of $Lie(x,y)$7-foams are in one-to-one correspondence with Kashiwara–Vergne solutions. The associated automorphism groups satisfy

$Lie(x,y)$8

and the induced map

$Lie(x,y)$9

recovers the Alekseev–Enriquez–Torossian map $A$0 (Dancso et al., 2022). In this interpretation, the construction is encoded by a functorial passage from braids to $A$1-foams and then by comparison of the induced actions on associated graded objects.

4. The Alekseev–Torossian associator as midpoint of an interpolating family

The Alekseev–Torossian associator is the most visible explicit associator emerging from the broader Alekseev–Enriquez–Torossian / Alekseev–Torossian framework. A one-parameter family of flat connections

$A$2

on the trivial principal $A$3-bundle over $A$4 is constructed from configuration-space integrals and depends polynomially on $A$5. It interpolates between the Knizhnik–Zamolodchikov connection at $A$6, the Alekseev–Torossian connection at $A$7, and the anti-KZ connection at $A$8. The corresponding regularized associators are obtained by parallel transport for $A$9 along the standard path between the boundary configurations

$cyc \coloneqq A/[A,A].$0

and satisfy

$cyc \coloneqq A/[A,A].$1

Thus $cyc \coloneqq A/[A,A].$2 is the midpoint of a continuous bridge connecting $cyc \coloneqq A/[A,A].$3, $cyc \coloneqq A/[A,A].$4, and $cyc \coloneqq A/[A,A].$5 (Rossi et al., 2014).

This interpolating family settles a conjecture of Pavel Etingof in a split form. If $cyc \coloneqq A/[A,A].$6 is the unique element sending $cyc \coloneqq A/[A,A].$7 to $cyc \coloneqq A/[A,A].$8 and $cyc \coloneqq A/[A,A].$9 is the unique element sending $u\in tder$0 to $u\in tder$1, then the weak conjecture

$u\in tder$2

is true, while the strong conjecture

$u\in tder$3

is false. More precisely, there is a family $u\in tder$4 and odd-degree elements $u\in tder$5 such that

$u\in tder$6

and

$u\in tder$7

Hence $u\in tder$8 lies in the Lie subalgebra generated by the odd-degree pieces, but the path-ordered exponentials are not symmetric enough to force $u\in tder$9 (Rossi et al., 2014). This rules out the interpretation of the Alekseev–Torossian element as a square root of the KZ-to-anti-KZ element.

5. Formality morphisms, graph cocycles, and deformation quantization

The same interpolating parameter $u(x)=[x,u_1],\qquad u(y)=[y,u_2]$0 governs stable formality morphisms. For a Kontsevich graph $u(x)=[x,u_1],\qquad u(y)=[y,u_2]$1, the weight form is

$u(x)=[x,u_1],\qquad u(y)=[y,u_2]$2

and the stable formality morphism is

$u(x)=[x,u_1],\qquad u(y)=[y,u_2]$3

These morphisms satisfy

$u(x)=[x,u_1],\qquad u(y)=[y,u_2]$4

where $u(x)=[x,u_1],\qquad u(y)=[y,u_2]$5 is a family of graph cocycles. The homotopy class of $u(x)=[x,u_1],\qquad u(y)=[y,u_2]$6 corresponds to the associator $u(x)=[x,u_1],\qquad u(y)=[y,u_2]$7; in particular, $u(x)=[x,u_1],\qquad u(y)=[y,u_2]$8 corresponds to $u(x)=[x,u_1],\qquad u(y)=[y,u_2]$9, $u_1,u_2\in Lie(x,y)$0 corresponds to $u_1,u_2\in Lie(x,y)$1, and the paper explicitly states that the “logarithmic” Kontsevich formality morphism corresponds to the Knizhnik–Zamolodchikov associator (Rossi et al., 2014).

The graph-complex side is equally explicit. One has

$u_1,u_2\in Lie(x,y)$2

and the cocycles $u_1,u_2\in Lie(x,y)$3 represent the odd generators of $u_1,u_2\in Lie(x,y)$4 under the isomorphism

$u_1,u_2\in Lie(x,y)$5

The family $u_1,u_2\in Lie(x,y)$6 is a graph cocycle for all $u_1,u_2\in Lie(x,y)$7, has only odd-degree pieces, and maps to the tangent element $u_1,u_2\in Lie(x,y)$8 governing the associator flow (Rossi et al., 2014). This places the Alekseev–Torossian midpoint simultaneously inside associator theory, graph cohomology, and deformation-quantization formality.

Related work shows that the Alekseev–Torossian associator is also the choice that identifies Kontsevich’s and Tamarkin’s formality morphisms up to homotopy. Kontsevich’s $u_1,u_2\in Lie(x,y)$9-morphism extends to a $TAut$0- and then $TAut$1-morphism, and it is homotopic to Tamarkin’s formality morphism provided one uses the Alekseev–Torossian associator in the little disks operad formality (Willwacher, 2011). In another direction, the compatibility between cup products and Shoikhet’s main result yields a rewriting of the Kontsevich product on $TAut$2 by means of the Alekseev–Torossian flat connection, and a similar formula follows directly from the Kashiwara–Vergne conjecture (Rossi, 2012). These results show that the construction is not confined to associator classification: it enters directly into explicit formality and star-product formulas.

6. Structural consequences, computations, and later extensions

A recurring misconception is that it has been proved that all Kashiwara–Vergne solutions arise from associators. The later degree-by-degree study states this only as the Alekseev–Torossian conjecture and says that it has been verified at least up to degree $TAut$3 (Dancso et al., 2023). What is proved there is the surjectivity of the truncation maps

$TAut$4

so any Kashiwara–Vergne solution modulo degree $TAut$5 can be extended to degree $TAut$6. The same work proves that the associated graded Lie algebra of $TAut$7 is canonically isomorphic to $TAut$8,

$TAut$9

thereby identifying the graded symmetry algebra controlling the extension tower (Dancso et al., 2023).

The Alekseev–Torossian associator is sufficiently explicit for coefficient-level study. Its coefficients can be calculated as linear combinations of iterated integrals of Kontsevich weight forms of Lie graphs, by rewriting its parallel transport as a path-ordered exponential of a differential operator built from those weight forms (Furusho, 2017). A particularly sharp arithmetic result concerns the coefficient of $j:tder\to cyc,$0: $j:tder\to cyc,$1 Under the standard conjecture that the weight-$j:tder\to cyc,$2 multiple zeta value space has $j:tder\to cyc,$3-basis

$j:tder\to cyc,$4

this implies that the Alekseev–Torossian associator is not rational (Felder, 2016).

The Kashiwara–Vergne side of the construction also interfaces with other Lie algebras related to multiple zeta values. The double shuffle Lie algebra $j:tder\to cyc,$5 injects into the Kashiwara–Vergne Lie algebra $j:tder\to cyc,$6 (Schneps, 2012). Depth-graded and elliptic variants further extend the picture: a linearized algebra $j:tder\to cyc,$7 and an elliptic algebra $j:tder\to cyc,$8 are defined, with injective Lie morphisms

$j:tder\to cyc,$9

(Raphael et al., 2017). A weakened linearized variant goes further: the emergent Drinfeld equations define a space $\mathsf{KRV}$00 canonically isomorphic to the symmetric part of the KV Lie algebra $\mathsf{KRV}$01 via

$\mathsf{KRV}$02

and the usual map $\mathsf{KRV}$03 factors through $\mathsf{KRV}$04 (Kuno, 3 Apr 2025). A functorial reformulation reproduces the Alekseev–Torossian injection $\mathsf{KRV}$05 by constructing Lie bialgebra brackets and cobrackets from operad modules and extends the mechanism to higher-genus groups (Navarro-Betancourt, 14 Nov 2025).

Taken together, these developments fix the mathematical profile of the Alekseev–Enriquez–Torossian construction. It is an explicit associator-to-Kashiwara–Vergne passage, a symmetry-level injection $\mathsf{KRV}$06, an operadic and topological functor from braid-type objects to tangential automorphisms, a midpoint mechanism centered on the Alekseev–Torossian associator, and a source of graph-complex, formality, arithmetic, and deformation-quantization structures.