Given the Iwahori-Hecke algebra $H_{n}$ (over some ring commutative ring $R$ with identity) with generators $\{T_{1},\ldots T_{n-1}\}$, we know it has basis $\{T_{w}\}_{w\in S_{n}}$, where $S_{n}$ is the permutation group and for some reduced expression $w=s_{i_{1}}\cdots s_{i_{r}}$, we have $T_{w}=T_{i_{1}}\cdots T_{i_{r}}$. This basis can be pictured as the set of all positive braids corresponding to each $w\in S_{n}$. E.g. for $w=s_{i_{1}}\cdots s_{i_{r}}$, we write $\sigma_{w}=\sigma_{i_{1}}\cdots \sigma_{i_{r}}$.

That is, if we consider the isomorphism $\varphi:R[B_{n}]\backslash{I}\to H_{n}$ (where $I$ is the ideal generated by the Hecke relation $(T_{i}-q_{1})(T_{i}-q_{2}) \ \forall i$ with units $q_{1}, q_{2}\in R$), have $\varphi(\sigma_{w})=T_{w}$ whence we may identify the two.

We obtain the annular Skein algebra $V$ by taking the braid closure of all elements in $H_{n}$ (so $V$ is the vector space consisting of $R$-linear combinations of closed $n$-braids modulo Reidemeister-II/III moves and the Skein relation (i.e. Hecke relation)).


What is a basis for $V$? Consider the linear map $\pi:H_{n}\to V$ (which we can think of as braid closure). $V$ should have the same basis as $H_{n}$ "modulo closure" i.e. the problem reduces to when does $\pi(T_{w})=\pi(T_{w'})$? By Markov's theorem, this should be iff $\sigma_{w}$ is conjugate to $\sigma_{w'}$ (since only Markov-I moves are relevant here). So this appears to be equivalent to finding the conjugacy classes amongst the positive braids $\{\sigma_{w}\}_{w\in S_{n}}$. Note that $V\cong H_{n}\backslash{[.,.]}$ (since $\pi(T_{w_{1}}T_{w_{2}})=\pi(T_{w_{2}}T_{w_{1}})$).

Pursuing an 'Algebraic' Solution:

In [Big05], section 5 is devoted to determining such a basis. This is shown to be the image of braids $b_{\lambda}$ under $\pi\circ\varphi$ where $b_{\lambda}=b_{(\lambda_{1})}\sqcup\cdots\sqcup b_{(\lambda_{k})}$ with $b_{(m)}=\sigma_{m-1}\cdots\sigma_{1}$ given a partition $\lambda=(\lambda_{1},\ldots,\lambda_{k})$ of $n$. This is done through geometrical/topological arguments. Can we do this algebraically instead?


From Bigelow's result, we can conclude that dim $V=p(n)$ (the $n^{th}$ partition no.) and $\pi$ is a linear projector. This tells us that the basis of $V$ is $\{T_{\lambda}\}_{\lambda\in\mathcal{P}}\subseteq \{T_{w}\}_{w\in S_{n}}$, where $\mathcal{P}$ is the set of conjugacy classes in $S_{n}$ (which we know to be the classes of different $\lambda$-cycles), and the chosen representative $\lambda\in S_{n}$ is the 'canonical' permutation corresponding to the $\lambda=(\lambda_{1},\ldots,\lambda_{k})$-cycle: that is, in cycle-notation, we fill out $k$ brackets from left-to-right with the numbers $1$ to $n$, where the $i^{th}$ bracket has length $\lambda_{i}$.

Essentially, his result tells us that $\pi(T_{w})=\pi(T_{w'})$ iff $w$ is conjugate to $w'$ in $S_{n}$. By our above development of the problem, does this mean that the algebraic route to Bigelow's result is to show that:

$\sigma_{w}$ is conjugate to $\sigma_{w'}$ in $B_{n}$ iff $w$ is conjugate to $w'$ in $S_{n}$ ?

I don't know if this final statement about conjugacy of positive braids is true, but if it is and the above reasoning is fine, then surely this is a valid route. If this isn't true, what's the best algebraic route to the solution?

Your Answer


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.