9
$\begingroup$

In a Lie algebra $\mathfrak{g}$ the Jacobi identity $\newcommand{\bracket}[2]{\left[#1\,#2\right]} \bracket{x}{\bracket{y}{z}} + \bracket{z}{\bracket{x}{y}} + \bracket{y}{\bracket{z}{x}} = 0$ holds. In the quantized enveloping algebra $\mathrm{U}_q(\mathfrak{g})$ where we define $\bracket{x}{y}_q := xy-qyx$ is there a "nice" most general $q$-analogue to the Jacobi? I've found that $$ \bracket{x}{\bracket{y}{z}_q} \;\;=\;\; \bracket{\bracket{x}{y}}{z} _q + \bracket{y}{\bracket{x}{z}}_q \quad\text{ and }\quad \bracket{x}{\bracket{y}{z}_q} + \bracket{z}{\bracket{x}{y}_q} + \bracket{y}{\bracket{z}{x}_q} =0 $$ (which work of any power of $q$ really) but in general, allowing for twistings by different powers of $q$, the cleanest equation I've been able to write is that for any $m,n,k \in \mathbf{Z}$, $$ \bracket{x}{\bracket{y}{z}_{q^n}}_{q^m} + q^{m+k}\bracket{z}{\bracket{x}{y}_{q^{n-k}}}_{q^{-m-k}} + q^{m}\bracket{y}{\bracket{z}{x}_{q^{n-m-k}}}_{q^k} =0\,. $$ But that's not very pretty or symmetric! Is there a better way to write this thing? Or a better way to think about what this $q$-analogue should be? (I'll keep playing with this off-and-on, and edit if I figure anything out)

$\endgroup$

2 Answers 2

16
$\begingroup$

There are various deformations of the Jacobi identity that can be found scattered in the literature. As far as i know, using the definition: $[A,B]_q=AB-qBA$, one of the most general ones (though i do not now if this is "symmetric" enough for your purposes) is the following one: $$ \big[A,[B,C]_{q_1}\big]_{q_2}+q_2\big[B,[C,A]_{q_1}\big]_{q_2^{-1}}+\big[C,[A,B]_{q_1q_2}\big]=0 $$ which is valid for arbitrary values of the parameters $q_1$, $q_2$. See [1].

Notice that both identities stated by the OP are compatible with it:

  • The first one can be recovered for: $q_1=q$, $q_2=1$,
  • while a particular case (for $k=-m$) of the second one for: $q_1=q^n$, $q_2=q^m$.

Edit: I was thinking that the second identity of the OP might be combined with the 2-parameter deformation given here to construct a further 3-parameter deformation of the Jacobi identity. But i am not sure on its exact form, i 'll try to think a little on it and come back if something interesting comes up.

Edit 2: I was thinking about the OP's last paragraph (asking for something more pretty or symmetric). Well, i did some further search and it turns out that we can get some symmetric generalization if we consider a two parameter deformation of the commutator. Using $[A,B]_{(p,q)}=pAB-qBA$ we have the following 3-parameter deformation of the Jacobi identity: $$ {\small \big[A,[B,C]_{(q_1,q_1^{-1})}\big]_{(q_3/q_2,q_2/q_3)}+\big[B,[C,A]_{(q_2,q_2^{-1})}\big]_{(q_1/q_3,q_3/q_1)}+\big[C,[A,B]_{(q_3,q_3^{-1})}\big]_{(q_2/q_1,q_1/q_2)}=0 } $$ See: [2].

References:
[1]. $q$-deformed Jacobi identity, $q$-oscillators and $q$-deformed infinite-dimensional algebras, Chaichian, Kulish, Lukierski, Phys. Let. B, v.237, 3-4, p. 401-406, 1990
[2]. On $q$-deformed infinite-dimensional $n$-algebra Lu Ding, Xiao-Yu Jia, Ke Wu, Zhao-Wen Yan, Wei-Zhong Zhao, arXiv:1404.0464v3 [hep-th]

$\endgroup$
4
$\begingroup$

This question might be better understood in the context of the well-known problem of whether there exists a ``quantum Lie algebra'' inside $U_q(\frak{g})$. See this old M.O. question for a discussion.

While there is no consensus about what a ``quantum Lie algebra'' should be, one interesting proposed solution is Majid's notion of a braided Lie algebra, as discussed for example in his Quantum Groups Primer book. Braided Lie algebras are also discussed in the old answer M.O. linked above.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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