Abstract
We describe transposed Poisson structures on generalized Witt algebras \(W(A,V,\langle \cdot ,\cdot \rangle )\) and Block Lie algebras L(A, g, f) over a field F of characteristic zero, where \(\langle \cdot ,\cdot \rangle \) and f are non-degenerate. More specifically, if \(\dim (V)>1\), then all the transposed Poisson algebra structures on \(W(A,V,\langle \cdot ,\cdot \rangle )\) are trivial; and if \(\dim (V)=1\), then such structures are, up to isomorphism, mutations of the group algebra structure on FA. The transposed Poisson algebra structures on L(A, g, f) are in a one-to-one correspondence with commutative and associative multiplications defined on a complement of the square of L(A, g, f) with values in the center of L(A, g, f). In particular, all of them are usual Poisson structures on L(A, g, f). This generalizes earlier results about transposed Poisson structures on Block Lie algebras \(\mathcal {B}(q)\).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Poisson algebras originated from the Poisson geometry in the 1970s and have shown their importance in several areas of mathematics and physics, such as Poisson manifolds, algebraic geometry, operads, quantization theory, quantum groups, and classical and quantum mechanics. One of the popular topics in the theory of Poisson algebras is the study of all possible Poisson algebra structures with fixed Lie or associative part [1, 9, 10, 17]. Recently, Bai, Bai, Guo, and Wu [2] have introduced a dual notion of the Poisson algebra, called transposed Poisson algebra, by exchanging the roles of the two binary operations in the Leibniz rule defining the Poisson algebra. They have shown that a transposed Poisson algebra defined this way not only shares common properties of a Poisson algebra, including the closedness under tensor products and the Koszul self-duality as an operad, but also admits a rich class of identities. More significantly, a transposed Poisson algebra naturally arises from a Novikov-Poisson algebra by taking the commutator Lie algebra of the Novikov algebra. Thanks to [3], any unital transposed Poisson algebra is a particular case of a “contact bracket” algebra and a quasi-Poisson algebra. Later, in a recent paper by Ferreira, Kaygorodov, and Lopatkin a relation between 1/2-derivations of Lie algebras and transposed Poisson algebras has been established [7]. These ideas were used to describe all transposed Poisson structures on Witt and Virasoro algebras in [7]; on twisted Heisenberg-Virasoro, Schrödinger-Virasoro and extended Schrödinger-Virasoro algebras in [19]; on oscillator algebras in [3]; Witt type Lie algebras in [12]. It was proved that each complex finite-dimensional solvable Lie algebra has a non-trivial transposed Poisson structure [13]. The \(\textrm{Hom}\)- and \(\textrm{BiHom}\)-versions of transposed Poisson algebras and transposed Poisson bialgebras have been considered in [14, 15]. The algebraic and geometric classification of 3-dimensional transposed Poisson algebras is given in [4]. For the list of actual open questions on transposed Poisson algebras see [3].
The first non-trivial example of a transposed Poisson algebra was constructed on the Witt algebra with the multiplication law \([e_i,e_j]=(i-j)e_{i+j}\) for \(i,j \in \mathbb Z\) (see, [7]). This attracted certain interest to the description of transposed Poisson structures on Lie algebras related to the Witt algebra. Thus, all transposed Poisson structures on the Virasoro algebra [7], Block type Lie algebras and Block type Lie superalgebras [11], Witt type Lie algebras [12] have been described. In the last years, the concept of Witt type and Block type Lie algebra has been enlarged and generalized by various authors, such as Kawamoto, Osborn, Đoković, Zhao, Xu, Passman, Jordan, etc. (see, for example, [11, 12] and references therein). In the present paper, we study transposed Poisson structures on the class of generalized Witt algebras defined by Đoković and Zhao in [6] and general Block algebras defined by Block in [5]. We use the standard method of characterization of transposed Poisson algebra structures on a fixed Lie algebra \(\mathfrak {L}\) based on the description of the space of \(\frac{1}{2}\)-derivations of \(\mathfrak {L}\).
Our work consists of two main parts. Section 2 is devoted to a description of \(\frac{1}{2}\)-derivations and transposed Poisson structures on generalized Witt algebras \(W(A,V,\langle \cdot ,\cdot \rangle )\), which result in the following theorem.
Theorem A
(Theorem 13). Let \(W(A,V,\langle \cdot ,\cdot \rangle )\) be a generalized Witt algebra with non-degenerate \(\langle \cdot ,\cdot \rangle \) and \(\textrm{char}(F)=0.\)
-
(i)
If \(\dim (V)>1\), then all the transposed Poisson algebra structures on \(W(A,V,\langle \cdot ,\cdot \rangle )\) are trivial.
-
(ii)
If \(\dim (V)=1\), say, \(V={{\,\textrm{span}\,}}_F\{v\}\), then the transposed Poisson algebra structures on \(W(A,V,\langle \cdot ,\cdot \rangle )\) are exactly mutations of the product \((a\otimes v)\cdot (b\otimes v)=(a+b)\otimes v\).
In Sect. 3 we study the same questions on general Block Lie algebras L(A, g, f) and obtain the following result.
Theorem B
(Theorems 22 and 29). Let L(A, g, f) be a general Block Lie algebra with non-degenerate f and \(\textrm{char}(F)=0.\)
-
(i)
If \(g=0,\) then there is only one non-trivial transposed Poisson algebra structure \(\cdot \) on L(A, 0, f). It is given by \(u_0 \cdot u_0=u_0.\)
-
(ii)
If \(g\ne 0\) and \((g(a),h(a))\ne (0,-1)\) for all \(a\in A\), then all the transposed Poisson algebra structures on L(A, g, f) are trivial.
-
(iii)
If \(g\ne 0\) and there is \(a\in A\), such that \((g(a),h(a))\ne (0,-1)\), then the transposed Poisson algebra structures on L(A, g, f) are usual Poisson algebra structures that are extensions by zero of commutative associative products \(*\) on the complement \(V={{\,\textrm{span}\,}}_F\{u_a\mid g(a)=h(a)+2=0\}\) of [L, L] with values in \(Z(L)={{\,\textrm{span}\,}}_F\{u_a\mid g(a)=h(a)+1=0\}\).
2 Definitions and Preliminaries
All the algebras below will be over a field F of characteristic zero and all the linear maps will be F-linear, unless otherwise stated.
Definition 1
Let \({\mathfrak {L}}\) be a vector space equipped with two nonzero bilinear operations \(\cdot \) and \([\cdot ,\cdot ].\) The triple \(({\mathfrak {L}},\cdot ,[\cdot ,\cdot ])\) is called a transposed Poisson algebra if \(({\mathfrak {L}},\cdot )\) is a commutative associative algebra and \(({\mathfrak L},[\cdot ,\cdot ])\) is a Lie algebra that satisfies the following compatibility condition
Transposed Poisson algebras were first introduced in a paper by Bai, Bai, Guo, and Wu [2].
Definition 2
Let \(({\mathfrak {L}},[\cdot ,\cdot ])\) be a Lie algebra. A transposed Poisson algebra structure on \(({\mathfrak {L}},[\cdot ,\cdot ])\) is a commutative associative multiplication \(\cdot \) on \(\mathfrak L\) which makes \(({\mathfrak {L}},\cdot ,[\cdot ,\cdot ])\) a transposed Poisson algebra.
Definition 3
Let \(({\mathfrak L}, [\cdot ,\cdot ])\) be an algebra and \(\varphi :{\mathfrak {L}}\rightarrow {\mathfrak {L}}\) a linear map. Then \(\varphi \) is a \(\frac{1}{2}\)-derivation if it satisfies
Observe that \(\frac{1}{2}\)-derivations are a particular case of \(\delta \)-derivations introduced by Filippov in [8] (see also [20] and references therein). The space of all \(\frac{1}{2}\)-derivations of an algebra \(\mathfrak L\) will be denoted by \(\Delta (\mathfrak L).\)
Definitions 1 and 3 immediately imply the following key Lemma.
Lemma 4
Let \(({\mathfrak {L}},[\cdot ,\cdot ])\) be a Lie algebra and \(\cdot \) a new binary (bilinear) operation on \({\mathfrak {L}}\). Then \(({\mathfrak {L}},\cdot ,[\cdot ,\cdot ])\) is a transposed Poisson algebra if and only if \(\cdot \) is commutative and associative and for every \(z\in {\mathfrak {L}}\) the multiplication by z in \(({\mathfrak {L}},\cdot )\) is a \(\frac{1}{2}\)-derivation of \(({\mathfrak {L}}, [\cdot ,\cdot ]).\)
The basic example of a \(\frac{1}{2}\)-derivation is the multiplication by a field element. Such \(\frac{1}{2}\)-derivations will be called trivial.
Theorem 5
Let \({\mathfrak {L}}\) be a Lie algebra without non-trivial \(\frac{1}{2}\)-derivations. Then all transposed Poisson algebra structures on \({\mathfrak {L}}\) are trivial.
Given a Lie algebra \((\mathfrak {L},[\cdot ,\cdot ])\) denote by \(Z(\mathfrak {L})\) its center, i.e. \(Z(\mathfrak {L})=\{a\in \mathfrak {L}\mid [a,b]=0,\ \forall b\in \mathfrak {L}\}\), and by \([\mathfrak {L},\mathfrak {L}]\) its square, i.e. \([\mathfrak {L},\mathfrak {L}]={{\,\textrm{span}\,}}_F\{[a,b]\mid a,b\in \mathfrak {L}\}\). Fix a complement V of \([\mathfrak {L},\mathfrak {L}]\) in \(\mathfrak {L}\). Then any commutative associative product \(*:V\times V\rightarrow Z(\mathfrak {L})\) defines a transposed Poisson algebra structure \(\cdot \) on \(\mathfrak {L}\) by means of
where \(a_1,b_1\in V\) and \(a_2,b_2\in [\mathfrak {L},\mathfrak {L}]\). Indeed, the right-hand side of (1) is zero, because \(z\cdot x,z\cdot y\in Z(\mathfrak {L})\), and the left-hand side of (1) is zero by (3), because \([x,y]\in [\mathfrak {L},\mathfrak {L}]\). We say that \(\cdot \) is the extension by zero of \(*\). Observe that \(\cdot \) is at the same time a usual Poisson structure on \((\mathfrak {L},[\cdot ,\cdot ])\).
3 Transposed Poisson Structures on Generalized Witt Algebras
3.1 Generalized Witt Algebras
Đoković and Zhao [6] introduced the following generalization of the classical Witt algebra.
Definition 6
Let F be a field, \((A,+)\) a non-trivial abelian group, \(V\ne \{0\}\) a vector space and \(\langle \cdot ,\cdot \rangle :V\times A\rightarrow F\) a map linear in the first variable and additive in the second one. Denote \(W:=FA\otimes _F V\) and define the product \([\cdot ,\cdot ]\) on W by means of
Then \((W,[\cdot ,\cdot ])\) is a Lie algebra called a generalized Witt algebra.
When it is necessary to specify A, V and \(\langle \cdot ,\cdot \rangle \), one writes \(W=W(A,V,\langle \cdot ,\cdot \rangle )\). We assume that \(\langle \cdot ,\cdot \rangle \) is non-degenerate, i.e.
We also assume that \(\textrm{char}(F)=0\). Then it follows from 5 that A is torsion-free.
The algebra \(W(A,V,\langle \cdot ,\cdot \rangle )\) is a generalization of the so-called Witt type Lie algebra V(f) (corresponding to an additive map f) introducedFootnote 1 by Yu in [18]. We recall its definition using the notation from the present paper. Given an abelian group A, a field F and a function \(f:A\rightarrow F\), define V(f) to be a vector space with basis \(\{e_a\}_{a\in A}\) and multiplication
Without loss of generality, one assumes that \(f(0)=0\). Then V(f) is a Lie algebra if and only if
for all \(a,b\in A\). Observe that in general one does not require that f be additive. However, it turns out to be so if \(|f(A)|\ge 4\) by [18, Lemma 4.6].
Lemma 7
Let \(\dim (V)=1\) and \(\langle \cdot ,\cdot \rangle \) be non-degenerate. Then \(W(A,V,\langle \cdot ,\cdot \rangle )\) is isomorphic to the Witt type Lie algebra V(f) for some additive injective \(f:A\rightarrow F\) with \(|f(A)|=\infty \).
Proof
Choose \(v\in V\setminus \{0\}\). Then \(\{a\otimes v\mid a\in A\}\) is a basis of W. Define an additive map \(f:A\rightarrow F\) by \(f(a)=\langle v,a \rangle \) and a bijective linear map \(\varphi :W\rightarrow V(f)\) by \(\varphi (a\otimes v)=e_a\). Then by (6) and (4)
Observe that f is injective by (5), because \(\langle v,a \rangle =0\Leftrightarrow \langle V,a \rangle =\{0\}\). Since A is torsion-free, then \(|A|=\infty \), whence \(|f(A)|=\infty \) as well. \(\square \)
Since transposed Poisson structures on V(f) were described in [12], we only need to deal with the case \(\dim (V)>1\).
Lemma 8
Let \(\dim (V)>1\). If \(a\ne 0\), then there exist two linearly independent \(v',v''\in V\) such that \(\langle v',a \rangle \ne 0\ne \langle v'',a \rangle \).
Proof
Denote \(V_0=\{v\in V\mid \langle v,a \rangle =0\}\). If \(V_0=\{0\}\), there is nothing to prove. Otherwise, choose \(0\ne v_0\in V_0\). Since \(a\ne 0\), by (5) there is \(v'\in V\) such that \(\langle v',a \rangle \ne 0\). Observe that \(v'\) and \(v_0\) are linearly independent, since otherwise \(v'=kv_0\) and \(\langle v',a \rangle =k\langle v_0,a \rangle =0\). Then \(v'\) and \(v'+v_0\) are also linearly independent and \(\langle v'+v_0,a \rangle =\langle v',a \rangle \ne 0\). So, we may choose \(v''=v'+v_0\).
3.2 \(\frac{1}{2}\)-Derivations of Generalized Witt Algebras
Observe that W is an A-graded algebra, namely
For all \(a\in A\) and \(v\in V\) denote, for simplicity,
Any linear map \(\varphi :W\rightarrow W\) decomposes as
where \(\varphi _a:W\rightarrow W\) is a linear map such that \(\varphi _a(W_b)\subseteq W_{a+b}\) for all \(b\in A\). In particular, \(\varphi \in \Delta (W)\) if and only if \(\varphi _a\in \Delta (W)\) for all \(a\in A\). We write
where \(d_a:W\rightarrow V\).
Lemma 9
Let \(\varphi _a:W\rightarrow W\) be a linear map satisfying 8. Then \(\varphi _a\in \Delta (W)\) if and only if for all \(x,y\in A\) and \(v,w\in V\)
Proof
and
\(\square \)
Lemma 10
Let \(\dim (V)>1\), \(a\ne 0\) and \(\varphi _a\in \Delta (W)\) satisfying (8). Then \(\varphi _a=0\).
Proof
Substitute \(y=0\) into (9):
that is
Then setting \(x=a\) in (10) we obtain
By 8 there are two linearly independent \(v',v''\in V\) such that \(\langle v',a \rangle ,\langle v'',a \rangle \ne 0\). Choosing consecutively \(v=v'\) and \(v=v''\) in (11) we have
By the linear independence of \(v'\) and \(v''\),
for all \(w\in V\). It follows from (10) that
If \(x\ne a\), then \(\langle w,a-x \rangle \ne 0\) for some \(w\in V\) by (5). Thus,
Now substitute \(x=a\) and \(y=-a\) into (9) and use (12):
Since A is torsion-free, then \(a\ne -a\), so \(d_a(w_{-a})=0\) by (13). Taking consecutively \(w=v'\) and \(w=v''\) we have
whence
by the linear independence of \(v'\) and \(v''\). Combining (13) and (14), we conclude that \(\varphi _a=0\). \(\square \)
Lemma 11
Let \(\varphi _0\in \Delta (W)\) satisfying (8) with \(a=0\). Then \(\varphi _0\in {{\,\textrm{span}\,}}_F\{\textrm{id}\}\).
Proof
For \(a=0\), equality (9) takes the form
Then \(y=0\) in (15) gives
If \(x\ne 0\), then choosing \(w\in V\) with \(\langle w,x \rangle \ne 0\), we obtain
In particular, \(d_0(v_x)=d_0(v_{-x})\) for all \(x\ne 0\). On the other hand, taking \(y=-x\ne 0\) in (15), we have
which for \(w=v\) gives
Choosing \(v\in V\) with \(\langle v,x \rangle \ne 0\) and applying (16), we conclude that
If \(x\ne 0\) and \(\langle v,x \rangle =0\), then, thanks to (16), equality (17) becomes
Choosing an arbitrary \(w\in V\) with \(\langle w,x \rangle \ne 0\), we arrive at
Combining (18), (19) and (8), we finally prove the desired fact. \(\square \)
Proposition 12
If \(\dim (V)>1\), then \(\Delta (W)={{\,\textrm{span}\,}}_F\{\textrm{id}\}\).
Proof The inclusion \(\Delta (W)\subseteq {{\,\textrm{span}\,}}_F\{\textrm{id}\}\) is Lemmas 10 and 11. The converse inclusion is trivial.
Theorem 13
Let \(\textrm{char}(F)=0\) and \(\langle \cdot ,\cdot \rangle \) be non-degenerate.
-
(i)
If \(\dim (V)>1\), then all the transposed Poisson algebra structures on \(W(A,V,\langle \cdot ,\cdot \rangle )\) are trivial.
-
(ii)
If \(\dim (V)=1\), say, \(V={{\,\textrm{span}\,}}_F\{v\}\), then the transposed Poisson algebra structures on \(W(A,V,\langle \cdot ,\cdot \rangle )\) are exactly mutations of the product \((a\otimes v)\cdot (b\otimes v)=(a+b)\otimes v\).
Proof
(i) is an immediate consequence of Proposition 12 and [7, Theorem 8], while (ii) follows from Lemma 7 and [12, Proposition 26].
\(\square \)
4 Transposed Poisson Structures on General Block Lie Algebras
4.1 General Block Lie Algebras
Another generalization of the Witt algebra is the class of Lie algebras studied by Block in [5].
Definition 14
Let F be a field, \((A,+)\) a non-trivial abelian group, \(g:A\rightarrow F\) an additive map and \(f:A\times A\rightarrow F\) an anti-symmetric biadditive map. The Block algebra L(A, g, f) is the F-vector space with basis \(\{u_a\}_{a\in A}\) and product
It is known [5] (and, in fact, easy to see) that L(A, g, f) is a Lie algebra if and only if either \(g=0\) or there exists an additive map \(h:A\rightarrow F\) such that for all \(a,b\in A\):
We will write \(L=L(A,g,f)\) for the simplicity of notation. We will also assume that \(\textrm{char}(F)=0\) and f is non-degenerate in the sense that
Then, as in Sect. 2.1, this implies that A is torsion-free.
Observe that L(A, g, f) is a generalization of the Block Lie algebra \(\mathcal {B}(q)\) studied in [11] (it had been introduced in [16] under slightly different assumptions on q and on the basis). Recall that \(\mathcal {B}(q)\), where \(q\in \mathbb {C}\), is the complex Lie algebra with a basis \(\{L_{m,i} \mid m, i \in \mathbb {Z}\}\), where
for all \(i, j, m, n \in \mathbb {Z}\). It is immediately seen that \(\mathcal {B}(q)=L(A,g,f)\), where \(F=\mathbb {C}\), \(A=\mathbb {Z}\times \mathbb {Z}\),
for \((m,i),(n,j)\in \mathbb {Z}\times \mathbb {Z}\). If \(q\ne 0\), then the corresponding map h from (21) can be chosen to be
Observe that f is non-degenerate, because f((m, i), \( (0,-1))=m\) and \(f((m,i),(1,0))=i\).
We will need descriptions of Z(L) and [L, L] in the general case.
Lemma 15
Let f be non-degenerate.
-
(i)
If \(g=0\), then \(Z(L)={{\,\textrm{span}\,}}_F\{u_0\}\). Otherwise, \(Z(L)={{\,\textrm{span}\,}}_F\{u_a\mid g(a)=h(a)+1=0\}\).
-
(ii)
If \(g=0\), then \([L,L]={{\,\textrm{span}\,}}_F\{u_a\mid a\ne 0\}\). Otherwise, \([L,L]={{\,\textrm{span}\,}}_F\{u_a\mid g(a)\ne 0\text { or }h(a)+2\ne 0\}\).
Proof
(i). Let \(g=0\). The inclusion \({{\,\textrm{span}\,}}_F\{u_0\}\subseteq Z(L)\) is trivial. Conversely, if \(x=\sum x_au_a\in Z(L)\) and \(x_a\ne 0\) for some \(a\ne 0\), then choose \(b\in A\) such that \(f(a,b)\ne 0\) (it exists due to non-degeneracy of f) and calculate
Let \(g\ne 0\). If \(g(a)=h(a)+1=0\), then for all \(b\in A\)
so \([u_a,u_b]=0\). This proves the inclusion \({{\,\textrm{span}\,}}_F\{u_a\mid g(a)=h(a)+1=0\}\subseteq Z(L)\). Conversely, assume that \(x=\sum x_au_a\in Z(L)\). Then \([x,u_0]=0\) implies \(g(a)=0\) for all a with \(x_a\ne 0\). Consequently, \([x,u_b]=-g(b)\sum x_a(h(a)+1)u_{a+b}\). Choosing \(b\in A\) with \(g(b)\ne 0\), we conclude that \(h(a)+1=0\) whenever \(x_a\ne 0\). Thus, \(Z(L)\subseteq {{\,\textrm{span}\,}}_F\{u_a\mid g(a)=h(a)+1=0\}\).
(ii). Let \(g=0\). If \([u_a,u_b]\ne 0\), then \(b\ne -a\), since otherwise \(f(a,b)=f(a,-a)=0\). Hence, \([u_a,u_b]=f(a,b)u_{a+b}\in {{\,\textrm{span}\,}}_F\{u_a\mid a\ne 0\}\). This proves \([L,L]\subseteq {{\,\textrm{span}\,}}_F\{u_a\mid a\ne 0\}\). Conversely, for any \(a\ne 0\) there exists \(b\in A\) such that \(f(a,b)\ne 0\). Then \([u_{a-b},u_b]=f(a-b,b)u_a=f(a,b)u_a\ne 0\), whence \(u_a\in [L,L]\).
Let \(g\ne 0\). If \([u_a,u_b]\ne 0\), then either \(g(a+b)\ne 0\) or \(h(a+b)+2\ne 0\), since otherwise
Hence, \([L,L]\subseteq {{\,\textrm{span}\,}}_F\{u_a\mid g(a)\ne 0\text { or }h(a)+2\ne 0\}\). Conversely, take \(a\in A\) with \(g(a)\ne 0\) or \(h(a)+2\ne 0\). If \(g(a)\ne 0\), then \([u_a,u_0]=g(a)u_a\ne 0\), so \(u_a\in [L,L]\). Otherwise, \(h(a)+2\ne 0\) and
so choosing \(b\in A\) with \(g(b)\ne 0\) we have \([u_{a-b},u_b]=-g(b)(h(a)+2)u_a\ne 0\), whence \(u_a\in [L,L]\). \(\square \)
We will also need the following technical lemma.
Lemma 16
Let \(\alpha ,\beta :A\rightarrow F\) two non-zero additive functions. Then there exists \(a\in A\) such that \(\alpha (a)\ne 0\ne \beta (a)\).
Proof
Assume that for any \(a\in A\) either \(\alpha (a)=0\) or \(\beta (a)=0\). Then \(A=\ker \alpha \cup \ker \beta \). Since \(\ker \alpha \) and \(\ker \beta \) are subgroups of A, then either \(\ker \alpha \subseteq \ker \beta \), in which case \(A=\ker \beta \), or \(\ker \beta \subseteq \ker \alpha \), in which case \(A=\ker \alpha \). Hence, either \(\alpha =0\) or \(\beta =0\), a contradiction. \(\square \)
4.2 \(\frac{1}{2}\)-Derivations of General Block Lie Algebras
It follows from (20) that \(L=\bigoplus _{a\in A} Fu_a\) is an A-grading, so any linear map \(\varphi :L\rightarrow L\) decomposes into the direct sum of linear maps
where \(\varphi _a(u_b)\in Fu_{a+b}\) for all \(b\in A\). Moreover, \(\varphi \in \Delta (L)\) if and only if \(\varphi _a\in \Delta (L)\) for all \(a\in A\). As usual, we write
where \(d_a:L\rightarrow F\).
Lemma 17
Let \(\varphi _a:L\rightarrow L\) be a linear map satisfying (25). Then \(\varphi _a\in \Delta (L)\) if and only if for all \(x,y\in A\)
Proof
and
\(\square \)
4.2.1 The case \(g=0\)
Assume first that \(g=0\).
Lemma 18
Let \(a\ne 0\) and \(\varphi _a\in \Delta (L)\) satisfying (25). Then \(\varphi _a=0\).
Proof
Taking \(y=-x\) in (26) and using anti-symmetry of f, we have
Hence,
Now, substitute \(y=-a\) into (26):
whence
Since \(f(a,a+x)=f(a,x)\), the latter is equivalent to
On the other hand, \(y=a\) in (26) gives
If \(f(a,x)\ne 0\), then using (28), we come to \(4d_a(x)=d_a(x)+2d_a(a)\). Consequently,
However, \(f(a,x)\ne 0\Leftrightarrow f(a,-x)\ne 0\), so replacing x by \(-x\) in (29) and taking into account \(\textrm{char}(F)=0\), we have
Combining (27) and (30), we conclude that
Now assume that \(f(a,x)=0\). Since \(a\ne 0\), by (22) there exists \(y\in A\) such that \(f(a,y)\ne 0\). Observe that \(f(a,x+y)=f(a,y)\ne 0\). Then \(d_a(y)=d_a(x+y)=0\) thanks to (31), so (26) takes the form
By Lemma 16 applied to \(f(a+x,-)\) and \(f(a,-)\), whenever \(x\ne -a\), the element y can be chosen in a way that \(f(a+x,y)\ne 0\ne f(a,y)\). Thus, we have proved
Finally, taking \(y=-a-x\) in (26) we see that the right-hand side is zero, while the left-hand side equals \(2f(a,x)d_a(-a)\). Choosing \(x\in A\) such that \(f(a,x)\ne 0\), we show that \(d_a(-a)=0\). Combining this with (31) and (32), we get the desired fact. \(\square \)
Lemma 19
Let \(\varphi _0\in \Delta (L)\) satisfying (25) with \(a=0\). Then \(\varphi _0(x)=\varphi _0(y)\) for all \(x,y\ne 0\).
Proof
Write (26) with \(a=0\):
Consequently,
Observe that \(f(x,y)=f(x+y,-y)\), so applying (33) with (x, y) replaced by \((x+y,-y)\), we have
Combining (33) and (34), we come to
However, \(f(x,-y)=-f(x,y)\), so replacing y by \(-y\) in (35), we obtain
It follows from (35) and (36) that \(d_0(y)=d_0(-y)\), so
because \(\textrm{char}(F)=0\).
Now let \(x,y\ne 0\). By Lemma 16 applied to \(f(x,-)\) and \(f(y,-)\) there exists \(z\in A\) such that \(f(x,z)\ne 0\ne f(y,z)\). Then (37) gives
as needed. \(\square \)
Lemma 20
The linear map \(\alpha :L\rightarrow L\) given by
is a \(\frac{1}{2}\)-derivation of L.
Proof
Observe by Lemma 15(i) that \(\alpha (L)\subseteq Z(L)\), so the right-hand side of (2) is always zero for \(\varphi =\alpha \). Now, \(\alpha ([L,L])=\{0\}\) by Lemma 15(ii). Thus, the left-hand side of (2) is always zero as well. \(\square \)
Proposition 21
We have \(\Delta (L)={{\,\textrm{span}\,}}_F\{\textrm{id},\alpha \}\).
Proof
The inclusion \(\Delta (L)\subseteq {{\,\textrm{span}\,}}_F\{\textrm{id},\alpha \}\) is Lemmas 18 and 19. The converse inclusion is Lemma 20.
Theorem 22
Let \(\textrm{char}(F)=0\) and f be non-degenerate. Then there is only one non-trivial transposed Poisson algebra structure \(\cdot \) on L(A, 0, f). It is given by
Proof
Let \(\cdot \) be a non-trivial transposed Poisson algebra structure on L(A, 0, f). By Proposition 21 and Lemma 4 for any \(a\in A\) there are \(k_a,l_a\in F\) such that
Since \(|A|>2\) (A is torsion-free), for any \(a\ne 0\) there exists \(b\not \in \{0,a\}\). Then by (39) and commutativity of \(\cdot \) we have \(k_a u_b=u_a\cdot u_b=u_b\cdot u_a=k_b u_a\). Consequently, \(k_a=0\) for \(a\ne 0\). Similarly, \((k_a+l_a)u_0=u_a\cdot u_0=u_0\cdot u_a=k_0 u_a\) gives \(k_0=l_a=0\) for \(a\ne 0\). Thus, the only non-zero product \(u_a\cdot u_b\) is \(u_0\cdot u_0=l_0u_0\). So, up to isomorphism, \(\cdot \) is of the form (38).
Conversely, in view of Lemma 15 the product (38) is of the form (3), so \((L,\cdot ,[\cdot ,\cdot ])\) is a transposed Poisson (and usual Poisson) algebra. \(\square \)
Remark 23
Consider \(\mathcal {B}(0)\) as the complex Block algebra \(L(\mathbb {Z}\times \mathbb {Z},0,f)\), where f is given by (23). Then we obtain the description of transposed Poisson algebra structures on \(\mathcal {B}(0)\) given in [11, Theorem 2.14] as a particular case of Lemma 22.
4.2.2 The case \(g\ne 0\)
In this case, as it was commented above, there exists an additive map \(h:A\rightarrow F\) such that (21) holds.
Lemma 24
Let \(a\ne 0\) and \(\varphi _a\in \Delta (L)\) satisfying (25). If \(g(x)f(a,x)\ne 0\), then \(\varphi _a(x)=0\).
Proof
Consider first \(y=0\) in 26
Then
Replacing x by \(-x\), we obtain
On the other hand, \(y=-x\) in (26) gives
Multiplying both sides of (42) by \(g(a-x)g(a+x)=g(a)^2-g(x)^2\) and using (41) and (40), we get
We have
Since
the coefficient of \(d_a(0)\) on the right-hand side of (43) equals
Subtracting the coefficient of \(d_a(0)\) on the left-hand side of (43), we obtain
Thus, under the assumption \(g(x)f(a,x)\ne 0\), (43) is equivalent to
Case 1. \(f(a,x)\ne 3g(x)\). Then (44) gives
Case 1.1. \(g(a)\ne g(x)\). It follows from (45) and (40) that \(d_a(x)=0\).
Case 1.2. \(g(a)=g(x)\). Then \(g(a+x)=2\,g(x)\ne 0\), so \(d_g(-x)=0\) by (41). Moreover, \(-f(a,x)+g(a+2x)=-f(a,x)+3g(x)\ne 0\), so (45) and (42) yield \(d_a(x)=0\).
Case 2. \(f(a,x)=3g(x)\ne 0\). Then (40) becomes
Since \(f(a,x)=3g(x)\) is invariant under the replacement of x by kx, then (46) implies
On the other hand, \(y=2x\) in (26) gives
Multiplying both sides of this equality by \(g(a-x)g(a-2x)g(a-3x)\) and using (47) we get
Comparing the coefficients of \(g(a)^ig(x)^jd_a(0)\), \(0\le i+j\le 4\), in (49), we see that (49) is equivalent to \(36g(x)^4d_a(0)=0\). Hence, we again have (45).
Case 2.1. \(g(a)\ne g(x)\). Then (45) and (40) yield \(d_a(x)=0\).
Case 2.2. \(g(a)=g(x)\). Then \(g(a-kx)=(1-k)g(x)\ne 0\) for \(k\ne 1\), so \(d_a(2x)=d_a(3x)=0\) by (47). Moreover, \(5g(x)+g(a)=6g(x)\ne 0\), so (45) and (48) imply \(d_a(x)=0\). \(\square \)
Lemma 25
Let \(a\ne 0\) and \(\varphi _a\in \Delta (L)\) satisfying 25. If \(g(x)f(a,x)=0\), then \(d_a(x)=0\), unless \(g(a)=g(x)=0\), \(h(a)=1\) and \(h(x)=-2\).
Proof
Case 1. \(g(a+x)\ne 0\). Since \(a\ne 0\), by Lemma 16 applied to \(f(a,-)\) and g there exists \(y\in A\) such that \(f(a,y)\ne 0\ne g(y)\). Observe that for any \(k\ne 0\) we have
because \(\textrm{char}(F)=0\). We affirm that k can be chosen in a way that
Indeed, \(f(a,y)\ne 0\), so there exists at most one k such that \(f(a,x)+kf(a,y)=0\). If \(f(a+x,y)-g(y)=0\), then \(f(a+x,ky)+g(a+x-ky)=g(a+x)\ne 0\) for all k. Otherwise, there exists at most one k such that \(k(f(a+x,y)-g(y))+g(a+x)=0\). Thus, (51) and (50) hold for infinitely many integer \(k\ne 0\) (recall that \(\textrm{char}(F)=0\)). Then \(d_a(ky)=d_a(x+ky)=0\) for any such k by Lemma 24. Consequently, applying (26) with y replaced by ky and using (51) we prove \(d_a(x)=0\).
Case 2. \(g(a+x)=0\). Then
Case 2.1. \(h(a+x)+1\ne 0\). Then the same argument as in Case 1 gives \(d_a(x)=0\).
Case 2.2. \(h(a+x)+1=0\). Then \(0=g(x)f(a,x)=g(a)^2=g(x)^2\), so (26) simplifies to
Observe that \(h(a)\ne 0\), since otherwise \(f(a,y)=0\) for all \(y\in A\) contradicting (22). Hence,
On the other hand, since \(g(a)=0\), then substituting \(x=0\) in (26) we get
Observe that \(g(x+y)=g(y)\) whenever \(g(x)=0\), so (53) and (52) yield
It follows that \(d_a(0)=0\), unless \(h(a)+1=0\). In any case, (53) implies
Write (26) replacing x by \(x-y\), where \(g(y)\ne 0\). Since \(d_a(y)=d_a(x-y)=0\) by (54), we come to
Thus, \(d_a(x)=0\), unless \(h(x)=-2\) (in which case \(h(a)=1\)). \(\square \)
Lemma 26
Let \(\varphi _0\in \Delta (L)\) satisfying (25) with \(a=0\). Then \(\varphi _0\in {{\,\textrm{span}\,}}_F\{\textrm{id}\}\).
Proof
Writing (26) with \(a=0\), we get
Case 1. \(g(x)\ne 0\). Then put \(y=0\) in (55) to get:
Case 2. \(g(x)=0\) and \(h(x)\ne -1\). Then \(f(x,y)=-g(y)h(x)\), and (55) is equivalent to
Choose \(y\in A\) such that \(g(y)\ne 0\). Then \(d_0(y)=d_0(x+y)=d_0(0)\) by (56). Hence, (57) yields \(d_0(x)=d_0(0)\).
Case 3. \(g(x)=0\) and \(h(x)=-1\). Let us use (55) with (x, y) replaced by \((x+y,-y)\). Observe that \(f(x+y,-y)=-f(x,y)=-g(y)\), so we get
Choosing \(y\in A\) with \(g(y)\ne 0\) we have \(d_0(x+y)=d_0(-y)=d_0(0)\) by (56). Thus, \(d_0(x)=d_0(0)\) due to (58). \(\square \)
Given \(\lambda ,\mu \in F\), we introduce the following notation:
Lemma 27
Let \(a\in A_{(0,-2)}\) and \(b\in A_{(0,-1)}\). Then the linear map \(\alpha _{(a,b)}:L\rightarrow L\) given by
is a \(\frac{1}{2}\)-derivation of L.
Proof
Observe by Lemma 15(i) that \(\alpha _{(a,b)}(L)\subseteq Z(L)\), showing that the right-hand side of (2) is always zero for \(\varphi =\alpha _{(a,b)}\). Furthermore, \(\alpha _{(a,b)}([L,L])=\{0\}\) by Lemma 15(ii), and the left-hand side of (2) is always zero as well. \(\square \)
Proposition 28
We have \(\Delta (L)={{\,\textrm{span}\,}}_F(\{\textrm{id}\}\cup \{\alpha _{(a,b)}\mid a\in A_{(0,-2)}\text { and }b\in A_{(0,-1)}\})\).
Proof
The fact that any \(\varphi \in \Delta (L)\) is a linear combination of \(\textrm{id}\) and \(\alpha _{(a,b)}\) follows from Lemma 24–26. Conversely, the inclusion \(\textrm{id}\in \Delta (L)\) is trivial, and \(\alpha _{(a,b)}\in \Delta (L)\) is Lemma27. \(\square \)
Theorem 29
Let \(\textrm{char}(F)=0\), \(g\ne 0\) and f be non-degenerate. If \((g(a),h(a))\ne (0,-1)\) for all \(a\in A\), then all the transposed Poisson algebra structures on L(A, g, f) are trivial. Otherwise, the transposed Poisson algebra structures on L(A, g, f) are exactly extensions by zero of commutative associative products \(*\) on the complement \(V={{\,\textrm{span}\,}}_F\{u_a\mid g(a)=h(a)+2=0\}\) of [L, L] with values in \(Z(L)={{\,\textrm{span}\,}}_F\{u_a\mid g(a)=h(a)+1=0\}\).
Proof
Let \(\cdot \) be a transposed Poisson algebra structure on L(A, g, f). If (g(a), \( h(a))\ne (0,-1)\) for all \(a\in A\), then \(A_{(0,-1)}=\emptyset \), so by Proposition 28 we have \(\Delta (L)=\{\textrm{id}\}\). It follows from [7, Theorem 8] that \(\cdot \) is trivial.
Assume that \((g(a),h(a))=(0,-1)\) for some \(a\in A\). Then \((g(2a),h(2a))=(0,-2)\), so both \(A_{(0,-1)}\) and \(A_{(0,-2)}\) are non-empty. By Proposition 28 and Lemma 4 for any \(a\in A\) there are \(k_a\in F\) and \(\{l_a^{(x,y)}\}_{x\in A_{(0,-2)},y\in A_{(0,-1)}}\subseteq F\), such that
Since \(\textrm{char}(F)=0\) and A is torsion-free, then \(A{\setminus } A_{(0,-1)}\) and \(A{\setminus } A_{(0,-2)}\) are infinite (indeed, if \(a\in A_{(\lambda ,\mu )}\) with \((\lambda ,\mu )\ne (0,0)\), then \(ka\not \in A_{(\lambda ,\mu )}\) for all \(k\ne 1\)). So, for any \(a\not \in A_{(0,-2)}\) there exists \(b\not \in A_{(0,-2)}\), \(b\ne a\). Then by (59) and commutativity of \(\cdot \) we have \(k_a u_b=u_a\cdot u_b=u_b\cdot u_a=k_b u_a\). Consequently,
Now let \(a\in A_{(0,-2)}\) and \(b\not \in A_{(0,-1)}\cup A_{(0,-2)}\). Then \(k_bu_a+\sum _{y\in A_{(0,-1)}}l_b^{(a,y)}u_y=u_b\cdot u_a=u_a\cdot u_b=k_a u_b\) gives
It follows from (60) and (61) and the commutativity of \(\cdot \) that
Thus, \(\cdot \) is of the form (3) for the commutative associative product \(u_a*u_b=\sum _{y\in A_{(0,-1)}}l_a^{(b,y)}u_y\in Z(L)\) on \(V={{\,\textrm{span}\,}}_F\{u_a\mid a\in A_{(0,-2)}\}\).
Conversely, in view of Lemma 15 the product (38) is of the form (3), so \((L,\cdot ,[\cdot ,\cdot ])\) is a transposed Poisson (and usual Poisson) algebra.
Remark 30
Consider \(\mathcal {B}(q)\) with \(q\ne 0\) as the complex Block algebra \(L(\mathbb {Z}\times \mathbb {Z},g,f)\), where g, f and h are given by (23) and (24). Then \((g(m,i),h(m,i))=(0,-2)\Leftrightarrow (m,i)=(0,-2q)\) and \((g(m,i),h(m,i))=(0,-1)\Leftrightarrow (m,i)=(0,-q)\), so we again obtain the description of transposed Poisson algebra structures on \(\mathcal {B}(q)\) given in [11, Theorem 2.14] as a particular case of Theorem 29.
Data Availability
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.
Notes
Notice that ‘V’ in V(f) is not the same space V from Definition 6.
References
Albuquerque, H., Barreiro, E., Benayadi, S., Boucetta, M., Sánchez, J.M.: Poisson algebras and symmetric Leibniz bialgebra structures on oscillator Lie algebras. Journal of Geometry and Physics 160, 103939 (2021)
Bai, C., Bai, R., Guo, L., Wu, Y.: Transposed Poisson algebras, Novikov-Poisson algebras, and 3-Lie algebras. Journal of Algebra 632, 535–566 (2023)
Beites P., Ferreira B. L. M., Kaygorodov I.: Transposed Poisson structures, arXiv:2207.00281
Beites P., Fernández Ouaridi A., Kaygorodov I.: The algebraic and geometric classification of transposed Poisson algebras, Revista de la Real Academia de Ciencias Exactas, Físicas y Naturales. Serie A. Matemáticas, 117 (2023), 2, 55
Block, R.: On torsion-free abelian groups and Lie algebras. Proceedings of the American Mathematical Society 9, 613–620 (1958)
Đoković, D., Zhao, K.: Derivations, isomorphisms, and second cohomology of generalized Witt algebras. Transactions of the American Mathematical Society 350(2), 643–664 (1998)
Ferreira, B.L.M., Kaygorodov, I., Lopatkin, V.: \(\frac{1}{2}\)-derivations of Lie algebras and transposed Poisson algebras, Revista de la Real Academia de Ciencias Exactas. Físicas y Naturales. Serie A. Matemáticas 115, 142 (2021)
Filippov, V.: \(\delta \)-Derivations of Lie algebras. Siberian Mathematical Journal 39(6), 1218–1230 (1998)
Jaworska-Pastuszak, A., Pogorzały, Z.: Poisson structures for canonical algebras. Journal of Geometry and Physics 148, 103564 (2020)
Kaygorodov, I., Khrypchenko, M.: Poisson structures on finitary incidence algebras. Journal of Algebra 578, 402–420 (2021)
Kaygorodov, I., Khrypchenko, M.: Transposed Poisson structures on Block Lie algebras and superalgebras. Linear Algebra and Its Applications 656, 167–197 (2023)
Kaygorodov, I., Khrypchenko, M.: Transposed Poisson structures on Witt type algebras. Linear Algebra and its Applications 665, 196–210 (2023)
Kaygorodov, I., Lopatkin, V., Zhang, Z.: Transposed Poisson structures on Galilean and solvable Lie algebras. Journal of Geometry and Physics 187, 104781 (2023)
Laraiedh I., Silvestrov S.: Transposed \({\rm Hom}\)-Poisson and \({\rm Hom}\)-pre-Lie Poisson algebras and bialgebras, arXiv:2106.03277
Ma, T., Li, B.: Transposed \({\rm BiHom}\)-Poisson algebras. Communications in Algebra 51(2), 528–551 (2023)
Xia, C., You, T., Zhou, L.: Structure of a class of Lie algebras of Block type. Communications in Algebra 40(8), 3113–3126 (2012)
Yao, Y., Ye, Y., Zhang, P.: Quiver Poisson algebras. Journal of Algebra 312(2), 570–589 (2007)
Yu, R.W.T.: Algèbre de Lie de type Witt. Communications in Algebra 25(5), 1471–1484 (1997)
Yuan, L., Hua, Q.: \(\frac{1}{2}\)-(bi)derivations and transposed Poisson algebra structures on Lie algebras. Linear and Multilinear Algebra 70(22), 7672–7701 (2022)
Zusmanovich, P.: On \(\delta \)-derivations of Lie algebras and superalgebras. Journal of Algebra 324(12), 3470–3486 (2010)
Funding
Open access funding provided by FCT|FCCN (b-on).
Author information
Authors and Affiliations
Contributions
All authors contributed to the study, conception and design. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Conflict of interest
There is no potential conflict of ethical approval, conflict of interest, and ethical standards.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The work was supported by FCT UIDB/MAT/00212/2020, UIDP/MAT/00212/2020, 2022.02474.PTDC and by CMUP, member of LASI, which is financed by national funds through FCT — Fundação para a Ciência e a Tecnologia, I.P., under the project with reference UIDB/00144/2020.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Kaygorodov, I., Khrypchenko, M. Transposed Poisson Structures on Generalized Witt Algebras and Block Lie Algebras. Results Math 78, 186 (2023). https://doi.org/10.1007/s00025-023-01962-y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00025-023-01962-y