i read in mark wildon book , an introduction to lie algebras, in page 22 say that : Suppose that dim $L$ = 3 and dim $L'$ = 2. We shall see that, over $\mathbb{C}$ at least, there are infinitely many non-isomorphic such Lie Algebras. Take a basis of $L'$, say $\{y,z\}$ and extend it to a basis of $L$, say by $x$. To understand the Lie algebra $L$, we need to understand the structure of $L'$ as a Lie algebra in its own right and how the linear map $ad x : L \rightarrow L$ acts on $L'$.

I dont understand this part "To understand the Lie algebra $L$, we need to understand the structure of $L'$ as a Lie algebra in its own right and how the linear map $ad x : L \rightarrow L$ acts on $L'$."

Can someone explain more about this part, how can we understand Lie algebra $L$ with the linear map $ad x : L \rightarrow L$ acts on $L'$? and for the case take in book, only case $x \notin L'$ , why we didnt consider $x \in L'$?

Thanks for your kindness

i have problem now to prove part b of lemma 3. $ad x : L' \rightarrow L'$. i stuck to show $ad x : L' \rightarrow L$ is homomorphism. How to show $ad x([y,z]) = [ad x(y), adx(z)]$ where $y,z \in L'$.

  • 1
    $\begingroup$ I assume that $L' = [L,L]$. To understand the Lie bracket, we need to know what $[x,y]$, $[x,z]$, and $[y,z]$ are. The first two are encoded in the adjoint map, and the last one comes from the structure of $L'$. $\endgroup$ – user27126 Nov 8 '12 at 19:07
  • $\begingroup$ @sanchez thanks for your explanation sir,i understand your explanation, what the use we see linear map $adx : L \rightarrow L$ acts on $L′$.? $\endgroup$ – Ongok Ongoksepen Nov 9 '12 at 5:19
  • $\begingroup$ @OngokOngoksepen If you have a new question, please post as such and don't edit old questions. $\endgroup$ – Julian Kuelshammer Dec 12 '12 at 12:04
  • $\begingroup$ @JulianKuelshammer ok sir, thanks for your suggestion. here the new link math.stackexchange.com/questions/257147/… $\endgroup$ – Ongok Ongoksepen Dec 12 '12 at 15:55

We have that $L=L'\oplus \langle x\rangle_k$ as vectorspaces. We want to understand $L$ as a Lie algebra, i.e. we want to know $[\ell,m]$ (for all $\ell, m\in L$). Now write $\ell,m\in L$ as $\ell=\ell'+\lambda x$ with $\ell'\in L'$ and $\lambda\in k$ and similarly $m=m'+\mu x$. Then we have $\begin{align} [\ell,x]&=[\ell'+\lambda x,m'+\mu x]\\ &=[\ell',m']+\lambda[x,m']+\mu[\ell',x]+\lambda\mu[x,x] \end{align}$ Now we understand $[\ell',m']$ if we understand $L'$ as a Lie algebra. We have $[x,m']=\operatorname{ad} x(m')$ and $[\ell',x]=-[x,\ell']=-\operatorname{ad}x(\ell')$ and $[x,x]=0$. The last two equalities follow from anticommutativity. Thus to understand $L$ as a Lie algebra, we have to know how $L'$ is as a Lie algebra and how $\operatorname{ad}x$ acts on $L'$.

  • $\begingroup$ thanks for your answer sir, but what is $<x>_k$? pardon me, if this is low quality question. The case in the book only two , $x \notin L$ and ad $x$ is diagonalizable , case two $x \notin L$ and ad $x$ is not diagonalizable. Why we didnt consider $x \in L$. So the conclusion, if we want to classify 3 dimensional algebra with dim $L'$ = 2 with isomorphism. we just see matrix representasion of $ad x$. if their matrix representation isomorphic , then the lie algebra is isomorphic? $\endgroup$ – Ongok Ongoksepen Nov 9 '12 at 15:32
  • $\begingroup$ $\langle x\rangle_k$ is the $1$-dimensional vector space spanned by $x$. $x$ is chosen, so that $x\notin L$. That way we have $L=L'\oplus \langle x\rangle_k$ as vector spaces. They determine every Lie algebra $L'$ with $\dim L'=2$ before they determine the $3$-dimensonal ones. $\endgroup$ – Julian Kuelshammer Nov 9 '12 at 15:36
  • $\begingroup$ thanks for your kindness sir. So, what it the conclusion that can we get from lie algebra 3 dimensional with 2 dimensional derived lie algebra?, i still didnt get the conclusion after read that classify the complex Lie algebras in the book sir. $\endgroup$ – Ongok Ongoksepen Nov 9 '12 at 15:52
  • $\begingroup$ The conclusion is if you fix the derived Lie algebra and the action of $\operatorname{ad}x$ up to conjugacy (choosing a different representative of the conjugacy class just results in a different choice of basis for $L'$) you can classify all $3$-dimensional Lie algebras up to isomorphism. $\endgroup$ – Julian Kuelshammer Nov 9 '12 at 16:16
  • 1
    $\begingroup$ Suppose you have two Lie algebras $L$ and $\mathfrak{g}$ with $\dim L=\dim \mathfrak{g}=3$ and $\dim L'=\dim\mathfrak{g}'=2$. Then they are isomorphic iff $L'\cong \mathfrak{g}'$ and $\operatorname{ad} x$ is congruent to $\operatorname{ad} g$, where $x\in L$, but $x\notin L'$ and $g\in \mathfrak{g}$, but $g\notin\mathfrak{g'}$. $\endgroup$ – Julian Kuelshammer Nov 10 '12 at 17:12

As a first step, show that $L^{\prime}$ is abelian. Then you have $[y,z]=0$ and $[y, x]=a y+bz$, $[z,x]=cy+dz$. So, you have one such Lie algeba for each conjugacy class in $GL(2, \mathbb{C})$.

  • $\begingroup$ what the use we see map $ad x : L \rightarrow L$ acts on $L'$ for understand lie algebra L? and in that book , second step is prove $ad x: L' \rightarrow L'$ is an isomorpishm $\endgroup$ – Ongok Ongoksepen Nov 9 '12 at 5:22

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.