---
title: 'Yamaguti Algebras: Theory & Applications'
url: https://www.emergentmind.com/topics/yamaguti-algebras
type: topic
---

# Yamaguti Algebras: Theory & Applications

Yamaguti algebras, in the recent sense developed under the name **associative-Yamaguti algebras**, are vector spaces equipped with one binary operation and two ternary operations satisfying a system of eleven “same-order” identities that act as an associative-side analogue of Lie-Yamaguti theory. They were introduced as enveloping objects for Lie-Yamaguti algebras, because a suitable skew-symmetrization of their operations yields a Lie-Yamaguti algebra. At the same time, a substantial part of the older literature uses “Yamaguti” primarily through **Lie-Yamaguti algebras**, the binary–ternary structures that generalize Lie algebras and Lie triple systems and arise from differential-geometric constructions involving torsion and curvature [2509.03648][2510.03148].

## 1. Terminology, origin, and geometric motivation

The historical core of the subject lies in **Lie-Yamaguti algebras**. Recent work recalls that Lie-Yamaguti algebras arose from Nomizu’s work on affine connections with parallel torsion and curvature, were axiomatized by Yamaguti, and were later known as Lie triple algebras before the now standard designation Lie-Yamaguti algebras became established [2510.03148]. In this older and still standard usage, the “Yamaguti” component refers to a nonassociative structure with a binary and a ternary operation whose compatibility encodes generalized Lie-theoretic and geometric data.

Their geometric meaning is explicit. For a manifold with affine connection, one has
\[
[x,y]=-T(x,y), \qquad [x,y,z]=-R(x,y)z,
\]
where \(T\) is torsion and \(R\) is curvature; this gives a Lie-Yamaguti algebra on vector fields [2310.05360]. In the reductive homogeneous-space setting, Lie-Yamaguti algebras arise naturally on the smooth sections of the tangent bundle of a reductive homogeneous space, and the conditions
\[
\nabla R = 0,\qquad \nabla T = 0
\]
are equivalent to the manifold being locally a reductive homogeneous space with its canonical connection [2408.10815]. In that geometric dictionary, the binary operation comes from torsion and the ternary operation comes from curvature.

The more recent notion of **Yamaguti algebra** narrows the terminology in a different direction. In the operadic and associative literature, “Yamaguti algebra” denotes the associative-side envelope of a Lie-Yamaguti algebra, with operations whose arguments always remain in the same order. This newer meaning does not replace the older one; rather, it sits above Lie-Yamaguti theory as a non-skew precursor whose symmetrization recovers the Lie-Yamaguti layer [2510.03148].

## 2. Lie-Yamaguti algebras as the foundational layer

A Lie-Yamaguti algebra is a vector space \(g\) with a skew-symmetric bilinear bracket \([\, ,\,]\) and a trilinear product \(\{\, ,\, ,\,\}\), skew-symmetric in the first two variables, satisfying
\[
[x,y] = -[y,x],\qquad \{x,y,z\}=-\{y,x,z\},
\]
together with the compatibility identities
\[
\sum_{x,y,z}\big([[x,y],z]+\{x,y,z\}\big)=0,
\]
\[
\sum_{x,y,z}\{[x,y],z,a\}=0,
\]
\[
\{a,b,[x,y]\}=[\{a,b,x\},y]+[x,\{a,b,y\}],
\]
\[
\{a,b,\{x,y,z\}\}
=
\{\{a,b,x\},y,z\}
+\{x,\{a,b,y\},z\}
+\{x,y,\{a,b,z\}\}.
\]
These identities place Lie-Yamaguti algebras strictly between Lie algebras and Lie triple systems: if the ternary operation vanishes, one recovers a Lie algebra; if the binary operation vanishes, one recovers a Lie triple system [2403.17015].

This layer already supports a substantial structural theory. The free Lie-Yamaguti algebra on a set \(\mathcal C\) is described as
\[
\LYA(\mathcal{C}) \cong \Alg(\mathcal{C}; *, [\![, , ]\!]) / I_{\LYA},
\]
and admits an explicit basis built from generators, ordered binary products, and Lie-triple-Hall-type ternary products [2408.10815]. On the deformation side, Lie-Yamaguti algebra structures can be encoded as Maurer-Cartan elements in a graded Lie algebra, and the resulting differential graded Lie algebra realizes Yamaguti cohomology as the deformation differential [2310.05360]. These developments explain why the associative-side notion of Yamaguti algebra is treated as an envelope rather than as an unrelated construction.

## 3. Associative-Yamaguti, or Yamaguti, algebras

An **associative-Yamaguti algebra**—called simply a **Yamaguti algebra** in the 2025 associative treatment—is a quadruple
\[
(A,\cdot,\{\, ,\, ,\,\},\{\!\{\, ,\, ,\,\}\!\})
\]
consisting of a vector space \(A\), a binary operation \(\cdot\), and two ternary operations \(\{\, ,\, ,\,\}\) and \(\{\!\{\, ,\, ,\,\}\!\}\), subject to the identities
\[
(a\cdot b)\cdot c-a\cdot(b\cdot c)+\{a,b,c\}-\{\!\{a,b,c\}\!\}=0 \tag{AY1}
\]
\[
\{a\cdot b,c,d\}=\{a,b\cdot c,d\} \tag{AY2}
\]
\[
\{a,b,c\cdot d\}=\{a,b,c\}\cdot d \tag{AY3}
\]
\[
\{\!\{a\cdot b,c,d\}\!\}=a\cdot\{\!\{b,c,d\}\!\} \tag{AY4}
\]
\[
\{\!\{a,b\cdot c,d\}\!\}=\{\!\{a,b,c\cdot d\}\!\} \tag{AY5}
\]
\[
a\cdot\{b,c,d\}=\{\!\{a,b,c\}\!\}\cdot d, \tag{AY6}
\]
together with five higher ternary identities \((AY7)\)–\((AY11)\) governing iterated applications of the two ternary operations [2509.03648].

Two features are structurally decisive. First, these are **same-order** analogues of Lie-Yamaguti identities: the variables do not undergo the skew permutations characteristic of Lie-type structures. Second, the identities are **nonsymmetric operadic identities**: the authors of the operadic treatment emphasize that the arguments always appear in the same order, so the governing operad is nonsymmetric [2510.03148].

The passage to Lie-Yamaguti algebras is by skew-symmetrization. From a Yamaguti algebra one defines
\[
[a,b]=a\cdot b-b\cdot a
\]
and
\[
\llbracket a,b,c\rrbracket
=
\{a,b,c\}-\{b,a,c\}-\{\!\{c,a,b\}\!\}+\{\!\{c,b,a\}\!\}.
\]
The resulting structure
\[
(A,[\, ,\,],\llbracket\, ,\, ,\,\rrbracket)
\]
is a Lie-Yamaguti algebra [2509.03648]. This is the precise sense in which Yamaguti algebras serve as envelopes of Lie-Yamaguti algebras, a point emphasized again in the operadic paper [2510.03148].

## 4. Constructions and enveloping associative algebras

The associative-Yamaguti framework contains several natural subclasses and sources of examples [2509.03648].

| Source | Induced operations | Outcome |
|---|---|---|
| Associative algebra \((A,\cdot)\) | \(\{a,b,c\}=\{\!\{a,b,c\}\!\}=(a\cdot b)\cdot c=a\cdot(b\cdot c)\) | associative-Yamaguti algebra |
| Reductive associative algebra \(A=A_0\oplus A_1\) | operations on \(A_1\) defined via \(\operatorname{pr}_{A_0}\) and \(\operatorname{pr}_{A_1}\) | associative-Yamaguti structure on \(A_1\) |
| Associative triple system of the first kind | binary product trivial, \(\{\!\{a,b,c\}\!\}=\{a,b,c\}\) | associative-Yamaguti algebra |
| Diassociative algebra \((D,\dashv,\vdash)\) | \(a\cdot b=a\dashv b+a\vdash b\), with ternary operations defined from \(\dashv,\vdash\) | canonical associative-Yamaguti algebra |

For reductive associative algebras \(A=A_0\oplus A_1\) with
\[
A_0A_0\subset A_0,\qquad A_0A_1\subset A_1,\qquad A_1A_0\subset A_1,
\]
the induced operations on \(A_1\) are
\[
a\bullet b=\operatorname{pr}_{A_1}(a\cdot b),\quad
\{a,b,c\}=(\operatorname{pr}_{A_0}(a\cdot b))\cdot c,\quad
\{\!\{a,b,c\}\!\}=a\cdot(\operatorname{pr}_{A_0}(b\cdot c)).
\]
For a diassociative algebra, the canonical formulas are
\[
a\cdot b=a\dashv b+a\vdash b,
\]
\[
\{a,b,c\}=-(a\dashv b)\vdash c=-(a\vdash b)\vdash c=-a\vdash(b\vdash c),
\]
\[
\{\!\{a,b,c\}\!\}=-a\dashv(b\vdash c)=-a\dashv(b\dashv c)=-(a\dashv b)\dashv c.
\]

A major structural theorem states that **every associative-Yamaguti algebra admits an enveloping associative algebra**. The construction uses the associative algebra
\[
\mathcal M(A)\subseteq \operatorname{End}(A)\oplus \operatorname{End}(A),
\]
spanned by
\[
\Delta(a,b)=(\sigma_{a,b},\tau_{a,b}),
\qquad
\sigma_{a,b}(c)=\{a,b,c\},\qquad
\tau_{a,b}(c)=\{\!\{c,a,b\}\!\},
\]
and forms the reductive associative algebra \(\mathcal M(A)\oplus A\) with product
\[
(\xi,a)\circledast (\eta,b)
=
(\xi*\eta+\Delta(a,b),\ \xi\triangleright b+a\triangleleft \eta+a\cdot b).
\]
This product is associative and reductive, so every associative-Yamaguti algebra is induced from a reductive associative algebra [2509.03648].

## 5. Operads and the noncrossing-partition model

The nonsymmetric operad of Yamaguti algebras, denoted \(\Yam\), admits a concrete combinatorial realization in terms of noncrossing partitions without singleton blocks [2510.03148]. For each \(n\), let \(B(n)\) be the set of noncrossing partitions of
\[
\{0,1,2,\dots,n\}
\]
such that every block has cardinality at least \(2\). The vector space \(\mathscr B(n)\) with basis \(B(n)\) forms a nonsymmetric operad \(\mathscr B\), and in fact a nonsymmetric cyclic operad.

The partial composition
\[
\circ_i : \mathscr{B}(m)\otimes \mathscr{B}(n)\to \mathscr{B}(m+n-1)
\]
is defined by a two-term rule. One glues the disk for \(\nu\in B(n)\) into the \(i\)-th segment of \(\pi\in B(m)\). The first term is the resulting glued partition. The second term is obtained by cutting the new block along the gluing line, splitting it into two blocks; if this creates a singleton block, the second term is omitted. The unit is the unique basis element in arity \(1\), written \({1}\).

A key proposition states that \(\mathscr B\) is generated by the unique binary partition
\[
{2}\in \mathscr B(2),
\]
and the two ternary partitions
\[
{3_c},\qquad {3_b}\in \mathscr B(3).
\]
The operad morphism
\[
\psi:\Yam\longrightarrow\mathscr{B}
\]
sends the binary generator to \({2}\), the first ternary generator to \(-\,{3_c}\), and the second ternary generator to \(-\,{3_b}\). The main theorem is that
\[
\psi:\Yam \xrightarrow{\sim} \mathscr{B}
\]
is an isomorphism of nonsymmetric operads [2510.03148].

This identification has several explicit consequences. The basis of \(\Yam(n)\) is indexed by noncrossing partitions of \(\{0,\dots,n\}\) without singleton blocks; \(\dim \Yam(n)\) is therefore the corresponding Riordan number; the operad is cyclic; and the defining relations exhibit Gröbner-basis behavior. A plausible implication is that the nonsymmetric nature of the Yamaguti identities is not merely formal: it is rigid enough to admit a complete combinatorial normal form.

## 6. Cohomology, deformations, and splitting structures

The associative-Yamaguti theory includes a dedicated \((2,3)\)-cohomology used for formal deformations and abelian extensions [2509.03648]. A representation of an associative-Yamaguti algebra \(A\) on a vector space \(M\) consists of structure maps with one \(M\)-input satisfying the AY identities with one variable in \(M\), and \(A\oplus M\) is an associative-Yamaguti algebra if and only if \(M\) is such a representation.

A \((2,3)\)-cocycle is a triple
\[
(\mu,F,G),\qquad \mu:A^{\otimes 2}\to M,\quad F,G:A^{\otimes 3}\to M,
\]
satisfying the AY-type cocycle identities. The quotient
\[
\mathcal H^{(2,3)}(A,M)=\mathcal Z^{(2,3)}(A,M)/\mathcal B^{(2,3)}(A,M)
\]
is the \((2,3)\)-th cohomology group. For a formal deformation
\[
\mu_t=\cdot+t\mu_1+t^2\mu_2+\cdots,\quad
F_t=\{\, ,\, ,\,\}+tF_1+\cdots,\quad
G_t=\{\!\{\, ,\, ,\,\}\!\}+tG_1+\cdots,
\]
the infinitesimal
\[
(\mu_1,F_1,G_1)
\]
is a \((2,3)\)-cocycle in the adjoint representation, and equivalent deformations have infinitesimals differing by a coboundary. Abelian extensions
\[
0\to M\to E\to A\to 0
\]
are classified by the same cohomology group:
\[
\mathrm{Ext}(A,M)\cong \mathcal H^{(2,3)}(A,M).
\]

The same paper also recasts the theory operadically through **Yamaguti multiplications** on a nonsymmetric operad and introduces **dendriform-Yamaguti algebras** as splitting objects for associative-Yamaguti algebras. A dendriform-Yamaguti algebra carries two binary operations \(\prec,\succ\), three ternary operations \(\{\, ,\, ,\,\}_{[i]}\), and three corresponding double-brace ternary operations \(\{\!\{\, ,\, ,\,\}\!\}_{[i]}\), subject to 58 identities. Summing the split operations produces an associative-Yamaguti algebra. Relative Rota-Baxter operators provide the bridge: if \(R:M\to A\) is a relative Rota-Baxter operator, then \(M\) inherits a dendriform-Yamaguti structure, and conversely every dendriform-Yamaguti algebra arises from such an operator, namely the identity map on its total associative-Yamaguti algebra [2509.03648].

## 7. Generalizations and current structural directions

The Lie-Yamaguti side of the subject continues to expand in several directions. **Hom-Lie-Yamaguti algebras** are twisted versions
\[
(L,*,\{\,,\, ,\,\},\alpha)
\]
of Lie-Yamaguti algebras; they are multiplicative Hom-algebras, reduce to ordinary Lie-Yamaguti algebras when \(\alpha=\mathrm{Id}\), are closed under twisting by self-morphisms, and connect to Hom-Malcev algebras when the ternary operation is expressed through the binary one by a Yamaguti-type formula [1012.0445]. Their representation and cohomology theory supports infinitesimal deformations and classifies abelian extensions by the \((2,3)\)-cohomology group [1503.06392].

Bialgebraic and symmetry-theoretic developments are equally active. The **classical Lie-Yamaguti Yang-Baxter equation** is defined by
\[
[r,r]=0,\qquad [r,r,r]=0,
\]
and a skew nondegenerate solution corresponds to a relative Rota-Baxter operator with respect to the coadjoint representation; matched pairs, Manin triples, and double construction Lie-Yamaguti bialgebras are equivalent in this setting [2210.17066]. For finite-dimensional Lie-Yamaguti algebras, a universal algebra \(\mathcal A(\mathfrak L)\) and its Hopf envelope \(\mathcal H(\mathfrak L)\) provide a universal coacting bialgebra and a universal coacting Hopf algebra, from which one recovers the automorphism group and classifies abelian group gradings [2506.01328].

Recent work also treats the factorization and complement problem. If a Lie-Yamaguti algebra \(E\) strongly factorizes through \(\mathfrak g\) and \(\mathfrak h\), then
\[
E\cong \mathfrak g\Join \mathfrak h
\]
for a matched pair, and every other \(\mathfrak g\)-complement is obtained from \(\mathfrak h\) by a deformation map \(r:\mathfrak h\to\mathfrak g\). These deformation maps generalize homomorphisms, derivations, crossed homomorphisms, and relative Rota-Baxter operators, and they admit a Maurer-Cartan characterization through a governing \(L_\infty\)-algebra [2605.25576]. In the Hom setting, isoclinism has also been developed for regular Hom-Lie Yamaguti algebras: every isoclinism family contains a stem algebra, and for finite-dimensional regular Hom-Lie Yamaguti algebras of the same dimension, isoclinism is equivalent to isomorphism [2508.01631].

Taken together, these results place Yamaguti algebras in a two-level landscape. On one level, Lie-Yamaguti algebras remain the geometric and Lie-theoretic nucleus of the subject. On the other, associative-Yamaguti algebras supply an envelope with explicit operadic, combinatorial, cohomological, and splitting structures. The present literature treats these levels not as competing definitions, but as adjacent realizations of the same binary–ternary paradigm.

Source: https://www.emergentmind.com/topics/yamaguti-algebras