Trialgebras and Leibniz 3-Algebras: Lie Lie Ann Lie
Trialgebras and Leibniz 3-Algebras: Lie Lie Ann Lie
Trialgebras and Leibniz 3-Algebras: Lie Lie Ann Lie
12, 2006
J. M. CASAS
1. Introduction
It is well-known that there exists a functor [−, −] from Ass (the category
of associative algebras) to Lie (the category of Lie algebras) which endows
an associative K-algebra A with a structure of Lie algebra by means of the
bracket [x, y] = x · y − y · x and that this functor is right adjoint to the universal
enveloping algebra functor U(−) : Lie → Ass. In order to fix the notation, K
denotes a fixed ground field throughout the paper.
A non skew-symmetric version of Lie algebras, called Leibniz algebras, was
introduced by Loday [8, 9] as K-vector spaces equipped with a bilinear bracket
which satisfies the Leibniz identity
[x, [y, z]] = [[x, y], z] − [[x, z], y]
If we factorize a Leibniz algebra g by the two-sided ideal spanned by the
elements [x, x], x ∈ g, then we obtain the Lie algebra denoted by gLie . The
kernel of the canonical map g → gLie is denoted by gann . Thus we have defined
a functor (−)Lie : Leib → Lie which is left adjoint to the inclusion functor from
Lie to Leib which considers a Lie algebra as a Leibniz algebra.
Moreover Loday [10] introduced a type of algebras, called dialgebras, which
are K-vector spaces endowed with two associative operations a and ` (left and
right) satisfying three axioms (see below). We denote by Dias the category
corresponding to these objects. This kind of algebras is closely related to
binary trees [10] and they play a similar role with respect to Leibniz algebras
as associative algebras with respect to Lie algebras, that is, a dialgebra D
can be functorially endowed with a structure of Leibniz algebra by means of
the bracket [x, y] = x a y − y ` x and this functor is right adjoint to the
universal enveloping dialgebra functor Ud(−) : Leib → Dias (see Proposition
1.9 in [10]).
165
166 J. M. CASAS
When Loday and Ronco were studying ternary planar trees [13], they found
a type of algebras, called trialgebras, which are K-vector spaces equipped
with three binary associative operations a, ⊥, and ` (left, middle, and right)
satisfying eight axioms (see below).
The goal of this paper (see section 2) is to endow trialgebras functorially with
a structure of Leibniz 3-algebra (Leibniz n-algebras [2] are K-vector spaces
endowed with an n-ary bracket satisfying the fundamental identity (2.4) below)
by means of the bracket
[x, y, z] = x a (y ⊥ z − z ⊥ y) − (y ⊥ z − z ⊥ y) ` x.
Moreover we construct the universal enveloping trialgebra functor UT (−) from
3 Leib (the category of Leibniz 3-algebras) to Trias (the category of trialgebras)
which is left adjoint to the functor described previously. These adjoint functors
are related with the adjoint functors Ud(−) a [−, −] : Leib → Dias by means
of the commutative diagram (2.12) below.
Section 3 is devoted to introducing the category 3 QLie of QuasiLie 3-
algebras as Leibniz 3-algebras for which the following identity holds:
[x, y, y] = 0
for all x, y. This kind of algebras plays a similar role with respect to Leibniz
3-algebras as Lie algebras with respect to Leibniz algebras. Concretely, we
functorially endow an associative algebra with a structure of QuasiLie 3-
algebras by mean of the bracket
[x, y, z] = x · y · z − x · z · y − y · z · x + z · y · x = x · (y · z − z · y) − (y · z − z · y) · x
We construct the universal Quasi-Lie 3-algebra functor U3 (−) as the left adjoint
to the functor [−, −, −] : Ass → 3 QLie and, finally, we relate this pair of adjoint
functors with the adjoint pair UT (−) a [−, −, −] : 3 Leib → Trias by means of
the commutative diagram (3.4) below.
Finally, in section 4, we construct the universal enveloping algebra of a
Leibniz 3-algebra and we prove that the category of representations of a Leibniz
3-algebra L is equivalent to the category of right-modules on the universal
enveloping algebra U3 L(L). Also we prove the typical properties of a universal
enveloping algebra for U3 L(L) (see Th. 1, p. 152 in [5]).
2. Trialgebras
Our goal in this section is to construct a pair of adjoint functors between
the categories of Leibniz 3-algebras and trialgebras similar to the adjoint pair
between the categories of Leibniz algebras and dialgebras. We start recalling
a kind of algebras with three associative operations, called trialgebras, which
were introduced by Loday and Ronco [13] when they studied ternary planar
trees.
Definition (2.1). An associative trialgebra is a K-vector space A equipped
with 3 binary associative operations: a, ⊥, `: A ⊗ A → A (called left, middle,
and right, respectively), satisfying the following relations:
1. (x a y) a z = x a (y ` z) = x a (y ⊥ z)
2. (x ` y) a z = x ` (y a z)
TRIALGEBRAS AND LEIBNIZ 3-ALGEBRAS 167
3. (x a y) ` z = x ` (y ` z) = (x ⊥ y) ` z
4. (x ⊥ y) a z = x ⊥ (y a z)
5. (x a y) ⊥ z = x ⊥ (y ` z)
6. (x ` y) ⊥ z = x ` (y ⊥ z)
A morphism between two associative trialgebras is a linear map which is
compatible with the three operations. We denote by Trias the category of
associative trialgebras.
Examples (2.2).
i) An associative K-algebra A endowed with the binary operations x a y =
x ⊥ y = x ` y = x.y, for all x, y ∈ A. This operation defines a functor from
Ass to Trias which has as left adjoint the functor (−)Ass : Trias → Ass which
maps a trialgebra T to the associative algebra TAss , which is the factorization
of T by the ideal (in sense of trialgebras) spanned by the elements of the form
x ` y − x ⊥ y; x a y − x ⊥ y; x, y ∈ T .
ii) Let A be an associative K-algebra. Take T = A ⊗ A ⊗ A and define the
following operations:
a ⊗ b ⊗ c a a0 ⊗ b0 ⊗ c0 := a ⊗ b ⊗ ca0 b0 c0
a ⊗ b ⊗ c ` a0 ⊗ b0 ⊗ c0 := abca0 ⊗ b0 ⊗ c0
a ⊗ b ⊗ c ⊥ a0 ⊗ b0 ⊗ c0 := a ⊗ bca0 b0 ⊗ c0
Extending these formulas by linearity on A ⊗ A ⊗ A one obtains product
applications a, `, ⊥ which satisfy the trialgebra axioms.
iii) Opposite Trialgebra: The opposite trialgebra of a trialgebra (T, a, `, ⊥)
is the trialgebra T op with the same underlying vector space and product given
by
x a0 y := y ` x; x `0 y := y a x; x ⊥0 y := y ⊥ x
iv) For other examples we refer to [13].
In order to show the role which trialgebras play with respect to Leibniz
3-algebras we start recalling few well-known material.
Leibniz algebras [8, 9] are a non-skew symmetric version to Lie algebras,
that is, they are K-vector spaces g equipped with a bilinear bracket [−, −] : g ⊗
g → g satisfying the Leibniz relation [x, [y, z]] = [[x, y], z] − [[x, z], y]. We
denote by Lie and Leib the categories of Lie algebras and Leibniz algebras,
respectively.
A dialgebra [10] is a K-vector space equipped with two associative opera-
tions: a, ` (called left and right, respectively), satisfying the following rela-
tions:
1. (x a y) a z = x a (y ` z)
2. (x ` y) a z = x ` (y a z)
3. (x a y) ` z = x ` (y ` z)
A morphism between two associative dialgebras is a linear map which pre-
serves the two operations. We will denote by Dias the category of associative
dialgebras.
168 J. M. CASAS
(2.3) 6U(-)
[-,-] Ud(-) 6 [-,-]
? (−)Ass ?
Ass ,→ Dias
From the diagonal composition of adjoint functors in diagram (2.3) one derives
the isomorphism Ud(g)Ass ∼ = U(gLie ) (see Lemma 4.8 in [11]).
On the other hand, a dialgebra D can be functorially endowed with a struc-
ture of trialgebra by means of one of the inclusion functors:
1. Taking the dialgebra operations a, ` and defining x ⊥ y := x a y.
2. Taking the dialgebra operations a, ` and defining x ⊥ y := x ` y.
Conversely, for any trialgebra T , let T1Dias be the factorization of T by the
three-sided ideal I1 spanned by the elements of the form x ⊥ y − x a y, and
let T2Dias be the factorization of T by the three-sided ideal I2 spanned by the
elements of the form x ⊥ y − x ` y. It is clear that x ⊥ y = x a y in the first
case and x ⊥ y = x ` y in the second one. Therefore, TiDias , (i = 1, 2), is an
associative dialgebra. Moreover, the factorization application T → TiDias , (i =
1, 2) is universal for applications from T to any associative dialgebra, that
is, the dialgebrization functor (−)iDias : Trias → Dias is left adjoint to the
inclusion functor inci : Dias → Trias, (i = 1, 2).
Proof.
{{x, y, z},a, b} − {{x, a, b}, y, z} − {x, {y, a, b}, z} − {x, y, {z, a, b}}
= [[x, y · z], a · b] − [[x, a · b], y · z] − [x, [y, a · b] · z] − [x, y · [z, a · b]]
= [[x, y · z], a · b] − [[x, a · b], y · z] − [x, [y, a · b] · z + y · [z, a · b]]
= [[x, y · z], a · b] − [[x, a · b], y · z] − [x, [y · z, a · b]]
The last term vanishes thanks to the Leibniz identity.
Lemma (2.9). If P is a non-commutative Leibniz-Poisson algebra , then P op
is also a non-commutative Leibniz-Poisson algebra. Here P op has the same
Leibniz algebra structure as P , but the associative algebra structure in P op is
x∗y =y·x
Lemma (2.10). If A is an associative trialgebra, then (A, ⊥, [−, −]) is a non-
commutative Leibniz-Poisson algebra, where
[a, b] = a a b − b ` a .
Corollary (2.11). Let A be a trialgebra, then A is a Leibniz 3-algebra in
two ways:
1. with respect to the bracket: [x, y, z]1 = x a (y ⊥ z) − (y ⊥ z) ` x;
2. with respect to the bracket: [x, y, z]2 = (z ⊥ y) ` x − x a (z ⊥ y) .
Proof. Apply Lemmas (2.8), (2.9) and (2.10).
Let us observe that [x, y, z] in Proposition (2.5) is equal to [x, y, z]1 +[x, y, z]2 .
We will use the structure in Proposition (2.5) since it is the structure which
guarantees the commutativity of square (2.12) below.
Proposition (2.5) gives us a functor [−, −, −] : Trias → 3 Leib which has
as left adjoint the universal enveloping trialgebra functor UT(−) : 3 Leib →
Trias. This functor assigns to a Leibniz 3-algebra L the trialgebra UT (L)
defined by
⊕n≥1 K[Pn ] ⊗ L⊗n
UT (L) =
I
where ⊕n≥1 K[Pn ] ⊗ L⊗n is the free associative trialgebra over the underlying
vector space L [13] and I = h{x a (y ⊥ z) − (y ⊥ z) ` x − x a (z ⊥ y) + (z ⊥ y) `
x − [x, y, z] | x, y, z ∈ L}i.
3. QuasiLie 3-algebras
The goal of this section is to construct a diagram similar to diagram (2.3) in
the category 3 Leib. In order to achieve our goal we need to introduce a new
kind of ternary algebras.
Definition (3.1). A QuasiLie algebra of order 3 or QuasiLie 3-algebra is a
Leibniz 3-algebra L for which the following identity holds:
[x, y, y] = 0
for all x, y ∈ L.
Obviously, a homomorphism of QuasiLie 3-algebras is a linear map such
that preserves the bracket. We denote by 3 QLie the category of QuasiLie
3-algebras.
Examples (3.2). i) Lie 3-algebras [4, 14].
ii) Lie triple systems [7] are K-vector spaces equipped with a trilinear
bracket which satisfies the identity (2.4) and, instead of skew-symmetry, sat-
isfies the conditions
[x, y, z] + [z, x, y] + [y, z, x] = 0
and
[x, y, y] = 0
This is an example of QuasiLie 3-algebras which are not Lie 3-algebras.
iii) Let A be a K-associative algebra equipped with a K-linear map D : A →
A satisfying
D(a · Db) = Da · Db = D(Da · b)
for all a, b ∈ A. If we define the bracket
[a, b, c] = a · b · Dc − a · Dc · b − b · Dc · a + Dc · b · a
then we have a Leibniz 3-algebra, when D is an endomorphism of algebras
such that D2 = D or D is a derivation such that D2 = 0, which is a QuasiLie
3-algebra, for instance, in particular case of D = Id.
iv) The particular case D = Id in example iii) shows a K-associative algebra
A endowed with a structure of Quasi-Lie 3-algebra by means of the bracket
[x, y, z] = x · y · z − x · z · y − y · z · x + z · y · x = x · (y · z − z · y) − (y · z − z · y) · x
x, y, z ∈A. This QuasiLie 3-algebra is not a Lie 3-algebra.
If we factorize a Leibniz 3-algebra L by the three-sided ideal spanned by all
brackets of the form [x, y, y], x, y ∈ L, then we obtain the QuasiLie 3-algebra
denoted by LQ Lie . The canonical morphism L → LQ Lie is universal for every
morphism from L to a QuasiLie 3-algebra, that is, the functor (−)Q Lie : 3 Leib →
3 QLie is left adjoint to the inclusion functor. We denote the kernel of the
canonical morphism L → LQ Lie by LQ ann .
172 J. M. CASAS
On the other hand, the functor [−, −, −] : Ass → 3 QLie described in Ex-
ample (3.2) iv) has as left adjoint the universal enveloping QuasiLie 3-algebra
functor. The universal enveloping QuasiLie 3-algebra functor assigns to a
QuasiLie 3-algebra L the factorization of the free QuasiLie 3-algebra QL(L)
by the ideal spanned by the elements of the form [x, y, z] − x · y · z + x · z · y +
y · z · x − z · y · x, for all x, y, z ∈ L.
Proposition (3.3). There exists the free QuasiLie 3-algebra over a set X.
2. For 1 ≤ k ≤ n
[ρ1 (ln , . . . , l2n−2 ), ρk (l1 , . . . , ln−1 )]
n−1
X
= ρk (l1 , . . . , li−1 , [li , ln , . . . , l2n−2 ], li+1 , . . . , ln−1 )
i=1
Thanks to relation iii) we have that the subalgebra spanned by the elements
lx⊗y , x ⊗ y ∈ L ⊗ L, is isomorphic to U((L ⊗ L)Lie ).
Let d : (L ⊗ L)l ⊕ (L ⊗ L)m ⊕ (L ⊗ L)r → U((L ⊗ L)Lie ) be the K-linear
application defined by d(lx⊗y ) = −x ⊗ y = −x ⊗ y + J , where J = h{[x ⊗ y, x ⊗
y] | x ⊗ y ∈ L ⊗ L}i, d(mx⊗y ) = 0, d(rx⊗y ) = 0. One extends d to an algebra
homomorphism from T ((L ⊗ L)l ⊕ (L ⊗ L)m ⊕ (L ⊗ L)r ) to U((L ⊗ L)Lie ) which
TRIALGEBRAS AND LEIBNIZ 3-ALGEBRAS 175
Acknowledgements.
The author was supported by MCYT, Grant BFM2003-04686-C02-02 (Eu-
ropean FEDER support included) and Xunta de Galicia, Grant
PGIDIT04PXIC37101PN.
I would like to express my gratitude to Teimuraz Pirashvili who suggested
to me the definition and the application in this context of Noncommutative
Leibniz-Poisson algebras.
References
[1] J. M. Casas, Homology with trivial coefficients of Leibniz n-algebras, Comm. in Algebra 31
(3) (2003), 1377–1386.
[2] J. M. Casas, J.-L. Loday and T. Pirashvili, Leibniz n-algebras, Forum Math 14 (2) (2002),
189–207.
[3] J. M. Casas and T. Datuashvili, Noncommutative Leibniz-Poisson algebras, Comm. in
Algebra 34 (7) (2006) 2507–2530.
[4] V. T. Filippov, n-Lie algebras, Sib. Mat. Zh. 26 (6), (1985), 126–140.
[5] N. Jacobson, Lie algebras, Dover Publications, Inc., New York, 1979.
[6] I. V. Kanatchikov, On field theoretic generalizations of a Poisson algebra, Rep. Math. Phys.
40 (1997), 225–234.
[7] W. Lister, A structure theory of Lie triple systems, Trans. Amer. Math. Soc. 72 (1952), 217–
242
[8] J.-L. Loday, Cyclic homology, Grundl. Math. Wiss. Bd. 301, Springer (1992).
[9] J.-L. Loday, Une version non commutative des algèbres de Lie: les algèbres de Leibniz,
L’Enseignement Mathématique, 39 (1993), 269–292.
[10] J.-L. Loday, Algèbres ayant deux opérations associatives (digèbres), C. R. Acad. Sci. Paris, t.
321, Série I, P. (1995), 141–146.
178 J. M. CASAS
[11] J.- L. Loday, Dialgebras, In “Dialgebras and related operads”, 7-66, Lecture Notes in Math.,
1763, Springer-Verlag (2001).
[12] J.-L. Loday and T. Pirashvili, Universal enveloping algebras of Leibniz algebras and
(co)homology, Math. Ann. 296 (1993), 139–158.
[13] J.-L. Loday and M. O. Ronco, Trialgebras and families of polytopes, “Homotopy Theory: Re-
lations with Algebraic Geometry, Group Cohomology, and Algebraic K-theory”, Contemporary
Math. 346 (2004), 369–398.
[14] A. P. Pojidaev, Solvability of finite-dimensional n-ary commutative Leibniz algebras of
characterictic 0, Comm. in Algebra 31 (1) (2003), 197–215.
[15] J.-P. Serre, Lie algebras and Lie groups, Lect. Notes in Math. 1500, Springer (1992).