Let $\mathfrak{g}$ be a complex simple Lie algebra with bracket $[x,y]$. For which $z\in \mathfrak{g}$ does the formula $$ \mu(x,y)=ad (z)([x,y])=[z,[x,y]] $$ define another Lie bracket on the same vector space ? For $\mathfrak{g}=\mathfrak{sl}(2,\mathbb{C})$ this holds for every $z$. Explicit computation seems to show that for $\mathfrak{g}=\mathfrak{sl}(n,\mathbb{C})$ with $n\ge 3$ only $z=0$ is possible. Does this hold for all simple Lie algebras of rank at least $2$, and is there a proof somewhere ?
In the literature, often the Jacobi identity for $\mu(x,y)=[z,[x,y]]$ is studied for all $z$. Then it is known that only $\mathfrak{sl}(2,\mathbb{C})$ is possible. Another reference are simple Hom-Lie algebras, but this is different. For nonassociative algebras a new multiplication $x\circ_{z} y$ on the same vector space induced by an element $z$ is called a homotope.

What is known for simple Lie algebras ?

EdiT: I saw that $[x,y]_R:=[R(x),y]+[x,R(y)]$ is a Lie bracket iff $R$ is a classical $R$-matrix. For $R=ad(z)$ we have $[x,y]_R=[z,[x,y]]$. So the question is, for which $z$ the map $R=ad(z)$ is a classical $R$-matrix of a simple Lie algebra.


In the paper Derivation Double Lie Algebras the following result is proved:

Theorem $3.2$: Let $\mathfrak{g}$ be a simple Lie algebra of rank $r\ge 2$ over an algebraically closed field $K$ of characteristic zero, and $z\in \mathfrak{g}$. Suppose that $R={\rm ad}(z)$ is a classical $R$-matrix. Then $z=0$ and $R=0$.

The proof uses the Jordan-Chevalley decomposition for $z=n+s$, the Jacobson-Morozov theorem, arguments concerning roots and algebraic group actions etc. This gives $s=0$ and ${\rm ad}(z)^3=0$, and finally $z=0$. For the case of rank $1$ we have the following result:

Proposition $3.1$: Let $\mathfrak{g}=\mathfrak{sl}(2,\mathbb{C})$. Then $R={\rm ad}(z)$ is a classical $R$-matrix for all $z\in \mathfrak{g}$. The Lie algebra $\mathfrak{g}_R$ is isomorphic to $\mathfrak{r}_{3,1}(\mathbb{C})$ for all $z\neq 0$.

Here $\mathfrak{r}_{3,1}(\mathbb{C})$ is the $3$-dimensional solvable Lie algebra given by $[e_1,e_2]=e_2$ and $[e_1,e_3]=e_3$, and $\mathfrak{g}_R$ has bracket $[x,y]_R=[z,[x,y]]$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.