# Would like references to a counterexample in nilpotent Lie Algebras (Wikipedia error?)

Discussion in 'Math Research' started by Jeremy Henty, Jun 18, 2008.

1. ### Jeremy HentyGuest

I hope this isn't too trivial for sci.math.research !

claim that a Lie Algebra is nilpotent if and only if its Killing form
vanishes. This appears to be false: it is easy to exhibit a
non-nilpotent Lie Algebra with a zero Killing form. (Example at the
end of this post.)

I have been searching for references to show the Wikipedia editors but
found nothing. Presentations of Cartan's criterion usually refer to
solvable and semisimple cases only. Some mention in passing that a
nilpotent Lie Algebra has zero Killing form, but I've seen no
discussion of the converse. So, does anyone have any references? Is
this one of those obvious things that everyone knows and noone ever
writes down?

[1] http://en.wikipedia.org/wiki/Lie_algebra
[2] http://en.wikipedia.org/wiki/Cartan's_criterion

Regards,

Jeremy Henty

Counterexample.

Let L be a 3-dimensional complex vector space with basis a, b, c .
There's a unique alternating bilinear form [,] on L such that [ab] = b
, [ac] = i.c , [bc] = 0 . The only non-trivial case of the Jacobi
identity is [a[bc]] + [b[ac]] + [c[ab]] = 0 , which follows since each
of the three terms vanishes individually. Hence L is a Lie algebra.

Let * be the Killing form. Now, [L,L] = <b,c> and <b,c> is abelian,
x) = 0 , so b*x = c*x = 0 . Also, a*a = tr((ad a)^2) =
tr(diag(0,1,i)^2) = tr(diag(0,1,-1)) = 0 , so * vanishes.

Finally, L is not nilpotent. Eg. it has trival centre (since ker(ad
a) = <a> , ker(ad b) = ker(ad c) = <b,c> , so the kernel of the
adjoint representation is trivial). Hence the lower central series is
identically zero. Alternatively, observe that the [L,L] = <b,c> and
[L,<b,c>] = <b,c> , so the upper central series stabilises at <b,c> .

Jeremy Henty, Jun 18, 2008

2. ### victor_meldrew_666Guest

You are (or at least can be) a Wikipedia editor. As you are
right you can correct the articles yourself. I would suggest
putting the counterexample (as you've written it here) in the
Lie algebras talk page.

Victor Meldrew
"I don't believe it!"

victor_meldrew_666, Jun 19, 2008