Terminology for a certain generalization of central extensions of Lie algebras

16 Views Asked by At

I’m currently reviewing some Lie algebra theory, in particular extension of Lie algebras. I’m searching for the name of a certain class of extensions, which generalize central extensions.

Let $\mathfrak{h}$ and $I$ be two Lie algebras. An extension of $\mathfrak{h}$ by $I$ is a short exact sequence $$ 0 \to I \xrightarrow{\;\iota\;} \mathfrak{g} \xrightarrow{\;\pi\;} \mathfrak{h} \to 0 $$ of Lie algebras, i.e. a short exact sequence of vector spaces in which each term is a Lie algebra, and were each map is a homomorphism of Lie algebras. Up to equivalence of extensions we may assume that $$ \mathfrak{g} =  \mathfrak{h} \oplus I $$ on the level of vector spaces, that $\iota$ is the canonical inclusion and that $\pi$ is the canonical projection. The Lie bracket of $\mathfrak{g}$ is then of the form $$ [(x,c), (y,d)] = ( [x,y], \kappa(x,y) + \theta(x)(d) - \theta(y)(c) + [c,d] ) $$ for a unique bilinear map $\kappa \colon \mathfrak{h} \times \mathfrak{h} \to I$ and a unique linear map $\theta \colon \mathfrak{h} \to \operatorname{Der}(I)$.

How are those extensions called which arise – up to equivalence – in the case of $\theta = 0$?

If we would also require the Lie algebra $I$ to be abelian then these would be precisely the central extensions of $\mathfrak{h}$ by $I$. Here I’m wondering about the case in which $I$ isn’t required to be abelian.

Moreover, how can these extensions be described without referring to $\kappa$ and $\theta$?

What I have in mind is again the special case of central extension. These can on the one hand be described in terms of $\kappa$ and $\theta$: the Lie algebra $I$ must be abelian, and $\theta = 0$. But they can also be described more intrinsically: the image of $I$ in $\mathfrak{g}$ must be contained in the center of $\mathfrak{g}$.