This is taken from Timmermann's Invitation to Quantum Groups and Duality.

Hi folks I am struggling a little with a small calculation in the above text.

I will just get right into it.

Lemma 3.2.5 Let $(A,\Delta)$ be a Hopf $*$-algebra with a normalised integral $h$, let $\chi_V$ and $\chi_W$ be corepresentations on finite-dimensional vector spaces $V$ and $W$, respectively, and let $Q\in \hom(V,W)$. Denote by $X$ and $Y$ the corepresentation operators corresponding to $\chi_V$ and $\chi_W$, respectively, and define $R,T\in\hom(V,W)$ by $$R:=(I\otimes h)(Y^{-1}(Q\otimes 1)X)\,\,\,\textit{ and }\,\,\, T:=(I\otimes h)(Y(Q\otimes 1)X^{-1}).$$ Then $R,T\in\hom(\chi_V,\chi_W)$. If $Q\in\hom(\chi_V,\chi_W)$, then $R=T=Q$.

Proof Since $Y$ and $X$ are corepresentations, the composition

$$\begin{align} Y^{-1}(R\otimes 1)X&=Y^{-1}(I\otimes h)(Y^{-1}(Q\otimes 1)X)X \\&=(I\otimes h\otimes I)\left(Y_{[13]}^{-1}Y_{[12]}^{-1}(Q\otimes 1\otimes 1)X_{[12]}X_{[13]}\right) \end{align}$$

Question: Where does the second equality come from?

The Rest of the Proof

This second expression can be rewritten in the form $$(I\otimes h\otimes I)((I\otimes \Delta)\left(Y^{-1}\right)(Q\otimes 1\otimes 1)(I\otimes\Delta)(X)).$$

By right-invariance of $h$, this composition is equal to

$$(I\otimes h)\left(Y^{-1}(Q\otimes 1)X\right)\otimes 1=R\otimes 1.$$

Thus $(R\otimes 1)X=Y(R\otimes 1)$. A similar calculation shows that $(T\otimes 1)X=Y(T\otimes 1)$ (using left-invariance of $h$). A previous result shows that $R$ and $T$ are then elements of $\hom(\chi_V,\chi_W)$.

If $Q\in\hom(\chi_V,\chi_W)$, then $Y^{-1}(Q\otimes 1)X=Q\otimes 1=Y(Q\otimes 1)X^{-1}$ by that previous result and so $R=Q=T$.

Further Details

Normalised Integral: A linear functional $h:A\rightarrow \mathbb{C}$ is a normalised integral if it is both left-and right-invariant:

$$(I\otimes h)\Delta(a)=h(a)1_{\mathcal{M}(A)}\,\,\,\text{ and }\,\,\,(h\otimes I)\Delta(a)=h(a)1_{\mathcal{M}(A)},$$

(where $\mathcal{M}(A)$ is the multiplier algebra of the quantum group $A$), as well as normalised


Corepresentation Operator: Where $\chi:V\rightarrow V\otimes A$ is a corepresentation of $A$ on a vector space $V$, the corepresentation operator associated with $\chi$ is an $X\in\hom(V)\otimes A$ such that $$X=\sum_{i,j}e_{ij}\otimes a_{ij},$$

where the $a_{ij}$ are elements of $A$ given by

$$\chi(e_j)=\sum_i e_i\otimes a_{ij},$$

so that $\chi(v)=X(v\otimes 1_A)$.

The inverse of $X$ is given by $X^{-1}=(I\otimes S)(X)$ where $S$ is the antipode.

Leg Notation: The leg notation is given by $X_{[12]}=X_0\otimes 1_A$ and $X_{[13]}=(I\otimes\tau)(X\otimes 1_A)$ where $\tau:A\otimes B\rightarrow B\otimes A$ is the flip map and we can show that $$(I\otimes\Delta )X=X_{[12]}X_{[13]}.$$

Intertwiners: An intertwiner is a map $\varphi:V\rightarrow W$ such that $$\chi_W\varphi=(\varphi\otimes 1_A)\chi_V.$$ The space of intertwiners from a corepresentation $\chi_V:V\rightarrow V\otimes A$ to a corepresentation $\chi_W:W\rightarrow W\otimes A$ is denoted by $\hom(\chi_V,\chi_W)$.

My Thoughts

...are few and far between. If anything I think it should be $$Y^{-1}(R\otimes 1)X=Y^{-1}((I\otimes h)(Y^{-1}(Q\otimes 1)X)\otimes 1)X$$

I had previously tried to unravel this lemma but Timmermann called $R$ by $S$ which I confused with the antipode eventually to experience "death by a million linear maps".

Further information available of course on request.


The formula should be

$$\begin{align} Y^{-1}(R\otimes 1)X&=Y^{-1}(I\otimes h)(Y^{-1}(Q\otimes 1)X)X \\&=(I\otimes h\otimes I)\left(Y_{[13]}^{-1}Y_{[12]}^{-1}(Q\otimes 1\otimes 1)X_{[12]}X_{[13]}\right) \end{align}$$

i.e. the $Q$ in the first term should be an $R$.

The first equality is just the substitution of the definition of $R$ into the expression $Y^{-1}(R\otimes 1)X$. In the second equality the Haar state $h$ gets pulled to the outside of the expression and the resulting triple tensor product expression is re-written in leg notation (since the $Y^{-1}$ and the $X$ on the outside of the previous expression now act on the first and third factor in the tensor product).

I hope this helps.

  • $\begingroup$ I fixed the typo and thank you for your help. $\endgroup$ – JP McCarthy Nov 6 '13 at 12:20

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.