MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am interested in finding formulas for commutators of symmetrized monomials in a universal enveloping algebra. Let $C(x_1,\ldots, x_n)= (1/n!)\sum x_{\sigma(1)}\cdots x_{\sigma(n)}$ where the sum runs over all permutations, and $x_i \in L$ for some LIe algebra $L$. This is an element of $UL$.

Now, reasonable combinatorics shows that, in $UL$, we have,

$ [C(x_1,\ldots, x_n), l] = \sum_{i=1}^n C(\ldots,[x_i,l],\ldots) $

for $l\in L$.

I am looking for formulas for $[C(x_1,\ldots, x_n), C(y_1,\ldots, y_m)]$ in terms of symmetrized monomials and brackets. Even for $n=m=2$ the number of terms gets fairly large. If anyone knows where I can find such things I would be very grateful.

share|cite|improve this question
What is exact formulas for the case $n=m=2$? – Melania Feb 8 '11 at 6:54
Presumably you are working over an arbitrary field of characteristic 0 here. As Melania comments, one should try to answer the question first for products of length 2. Anyway I can't visualize an illuminating general answer, since the symmetrization process already depends so heavily on the specific nature of the bracket in the Lie algebra. Maybe in very special cases a good formula could be written down? – Jim Humphreys Feb 8 '11 at 14:09

This is probably not yet a final answer but may shine some additional light on the problem: For simplicity, I assume that $L$ is finite-dimensional and defined over the reals (for some other field of char $0$, the following should still work).

The symmetrization map can be viewed as a linear map \begin{equation} \sigma\colon \mathrm{S}(L) \longrightarrow \mathrm{U}(L) \end{equation} from the symmetric algebra over $L$ into the universal envelopping algebra. It is now possible (essentially via PBW) to show that this is a fitration compatible linear bijection. Thus it allows to pull-back the product of $\mathrm{U}(L)$ to $\mathrm{S}(L)$. The result is the star product of Gutt / Drinfel'd (both in 1983, I guess). A further canonical isomorphism yields that the symmetric algebra is nothing else than the poylnomials on the dual $L^*$ (suppose $L$ is finite-dimensional for convenience) Thus your question is equivalent to the following task:

What is the Gutt star product commutator of two (homogeneous) polynomials on $L^*$?

Gutt has computed many properties of this star product and fan almost explicit formula. However, it essentially involved the full BCH series of $L$, so my guess is that a complete answer might be as complicated as computing BCH.

The Gutt star product can be characterized nicely as follows: take $x, y \in L$ and view them as linear polynomials on $L^*$ as usual. Then form the formal exponential functions $e_{\hbar x} (\alpha) = \exp(\hbar \alpha(x))$ and similarly for $e_{\hbar y}$ where $\hbar$ is your formal parameter. Then $\star_{\mathrm{Gutt}}$ is uniquely determined by \begin{equation} e_{\hbar x} \star_{\mathrm{Gutt}} e_{\hbar y} = e_{\mathrm{BHC}(\hbar x, \hbar y)} \end{equation} I hope I got the signs right :) You can find this formulas also in Section 8 of q-alg/9707030 (published in Commun. Math. Phys.). There are many more papers on the Gutt star product, so a little MRsearch will probably give some addition info.

The solution of your problem is now obtained by differentiating the above equation with respect to $\hbar$ sufficiently often and use polarization afterwarts. But as I sad, you need to know BCH quite well to efficiently do that. In the end you take commutators\ldots

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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