Let $\mathfrak{g}$ be a finite dimensional Lie algebra over $\mathbb{C}$ and let $\operatorname{Rad}(\mathfrak{g})$ denote the radical of $\mathfrak{g}$; the unique maximal solvable ideal of $\mathfrak{g}$.

If $\mathfrak{h}$ is an ideal of $\mathfrak{g}$, I want to know about the relation between $\operatorname{Rad}(\mathfrak{g/h})$ and $\operatorname{Rad}(\mathfrak{g})/\operatorname{Rad}(\mathfrak{h})$; what are the necessary and sufficient conditions for $\operatorname{Rad}(\mathfrak{g/h})$ to be isomorphic to $\operatorname{Rad}(\mathfrak{g})/\operatorname{Rad}(\mathfrak{h})$? Is $\operatorname{Rad}(\mathfrak{g/h})$ always isomorphic to $\operatorname{Rad}(\mathfrak{g})/\operatorname{Rad}(\mathfrak{h})$?

  • $\begingroup$ How do you know that ${\rm Rad } (\mathfrak{h})$ is an ideal in ${\rm Rad } (\mathfrak{g})$? $\endgroup$ – Dietrich Burde Mar 21 '17 at 11:46
  • $\begingroup$ In characteristic zero the radical is a characteristic ideal, so that will be the case Dietrich. $\endgroup$ – David Towers Mar 21 '17 at 12:04
  • $\begingroup$ Ah, very good. This should be added in the text. $\endgroup$ – Dietrich Burde Mar 21 '17 at 12:08

To simplify the notation write $R$ for ${\rm Rad}$ and let the Lie algebra be $L$ and the ideal $I$. Put $X/I=R(L/I)$. Then $R(L)\cap I=R(I)$, so we have $(R(L)+I)/I\cong R(L)/R(I)$. Clearly $R(L)+I\subseteq X$, so, if $L=R(L)\dot{+} S$ is the Levi decomposition of $L$, $X=R(L)+ X\cap S$. But $X\cap S$ is a semisimple ideal of $S$, and so $X\cap S= (X\cap S)^{(n)} \subseteq I$ for some $n \in \mathbb{N}$. Hence $X \subseteq R(L)+I$ and $R(L)/R(I)\cong (R(L)+I)/I= X/I=R(L/I)$.

| cite | improve this answer | |

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.