4
$\begingroup$

The Grothendieck ring of the abelian category $Rep(\mathfrak{sl}_2)$ of finite-dimensional representations of $\mathfrak{sl}_2$ is, according to Bakalov-Kirillov's Lecture notes on tensor categories and modular functors (Ch.2, page 32), isomorphic to the ring of palindromic Laurent polynomials, via the map $$K(Rep(\mathfrak{sl}_2)) \to \mathbb{Z}[x,x^{-1}]^{S_2} \qquad , \qquad V \mapsto \mathrm{tr}_V \ x^h$$ were $h=\mathrm{diag}(1,-1) \in \mathfrak{sl}_2$.

What do the authors mean by $\mathrm{tr}_V \ x^h$? I'd like to see for myself that this is a ring isomorphism, but I do not really see how this map is defined.

$\endgroup$

1 Answer 1

3
$\begingroup$

This is a version of the formal character.

It will be easier to first think about a less formal version of the character. Let's work over $\mathbb{C}$. If $V$ is a finite-dimensional representation of $\mathfrak{sl}_2(\mathbb{C})$ then it integrates to a finite-dimensional representation of $SL_2(\mathbb{C})$, and for any such representation we can compute its character in the group-theoretic sense. If the representation is given by $\rho : SL_2(\mathbb{C}) \to GL(V)$ then its character is

$$\chi_V(g) = \text{tr}_V(\rho(g)).$$

This is a continuous class function on $SL_2(\mathbb{C})$. Since diagonalizable matrices are dense it is determined by its behavior on diagonalizable matrices and hence on diagonal matrices (the maximal torus). Every such matrix has the form $\begin{bmatrix} x & 0 \\ 0 & x^{-1} \end{bmatrix}$ for some $x \in \mathbb{C}^{\times}$, so in fact the character is determined by its restriction to matrices of this form, where it becomes some function of $x$, which by an abuse of notation I will denote $\chi_V(x)$. For example, the character of the defining representation is

$$\chi_{\mathbb{C}^2}(x) = \text{tr} \begin{bmatrix} x & 0 \\ 0 & x^{-1} \end{bmatrix} = x + x^{-1}.$$

More generally you can compute that the character of $S^n(\mathbb{C}^2)$ in this sense is $x^n + x^{n-2} + \dots + x^{-n+2} + x^{-n}$.

This is pretty much the formal character already, except that $x$ is taken to be a formal variable rather than an arbitrary element of $\mathbb{C}^{\times}$. The reason for the funny notation $x^h$ is to suggest the intuition that we are invoking a formal version of the exponential map: namely, we could also have written an arbitrary diagonal matrix in $SL_2(\mathbb{C})$ as an exponential

$$\begin{bmatrix} e^z & 0 \\ 0 & e^{-z} \end{bmatrix} = \exp \left( z \begin{bmatrix} 1 & 0 \\ 0 & -1 \end{bmatrix} \right) = e^{zh}, z \in \mathbb{C}$$

of an element $zh$ of the Cartan subalgebra of $\mathfrak{sl}_2(\mathbb{C})$ given by diagonal matrices. In relation to our previous parameterization we have $x = e^z$, so you can think of "$x^h$" as shorthand for the exponential $e^{zh}$. A more abstract interpretation of this notation where we really try to make $x$ a formal variable is also possible, I think. Basically $x^h$ should be the linear map with the property that if $v$ is an eigenvector of $h$ with eigenvalue $\lambda$ then $x^h v = x^{\lambda} v$.

$\endgroup$
3
  • $\begingroup$ Thank you for your elaborated answer. Could you please clarify what you mean by $S^n (\mathbb{C}^2)$? $\endgroup$
    – Minkowski
    Commented Jul 10 at 22:18
  • 1
    $\begingroup$ @Minkowski: are you familiar with the classification of irreducible representations of $\mathfrak{sl}_2$? This is the $n^{th}$ symmetric power of $\mathbb{C}^2$. Explicitly it can be described as homogeneous polynomials of degree $n$ in two variables. This is the unique irreducible representation of dimension $n+1$. $\endgroup$ Commented Jul 10 at 22:25
  • $\begingroup$ As a quantum topologist my mind can only recognise for that at first sight the notation $V_n$... - thank you for the clarification! $\endgroup$
    – Minkowski
    Commented Jul 10 at 22:31

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .