I am trying to prove that for type $A$ , rational smoothness of Schubert varieties implies smoothness.

So suppose we are in Type $A_{n-1}$, so let $G=Sl(n,\mathbb C)$, $B=$ the group of upper triangular matrices in $G$ , $T=$ group of diagonal matrices in $G$. $W=N_G(T)/T \cong \mathfrak S_n$. Let $R=\{e_i -e_j : 1\le i, j\le n, i\ne j\}$ be our root system and $R^+=\{e_i-e_j : 1\le i<j\le n\}$ be positive root system.

Let $\le$ be the Bruhat order on $W$.

For $w\in W$ , let $\Gamma(w)$ be the Bruhat graph of $w$, whose vertex set is $\{u \in W : u\le w\}$ and between any two vertices $x$ and $y$, there is a edge if and only if $x=ys$ for some reflection $s$ (not necessarily simple).

Now suppose it is given that every vertex of the Bruhat graph $\Gamma (w)$ has degree equal to $l(w)$ (i.e. equivalently the Schubert variety $X(w)$ is rationally smooth ). Then how to show that

$l(w)=$ # $ \{\alpha \in R^+| s_{\alpha} \le w\}$ ?

[Note that if I can show my claim then I am done because $l(w)=\dim X(w)$ and dimension of the tangent space of $X(w)$ at $e_{id}$ is $\dim T(w,e_{id})=$# $ \{\alpha \in R^+| s_{\alpha} \le w\}$, so the claim would imply $X(w)$ is smooth at $e_{id}$ , hence smooth.]


1 Answer 1


Unless I'm missing something obvious:

Consider the degree of the identity element $e$ in $\Gamma(w)$. On the one hand, by the assumption you've made it is $\ell(w)$. On the other hand, it is clearly equal to the number of reflections $s_{\alpha}$ less than $w$, because the only things $e$ will be connected to are of the form $e\cdot s_{\alpha}=s_{\alpha}$.

EDIT: By the way, the fact that rational smoothness implies smoothness for Schubert varieties in the simply laced types (and I think in exactly the simply laced types) is a result of Dale Peterson. See e.g. these notes: http://www.math.ubc.ca/~carrell/SPTVGB.pdf

  • $\begingroup$ Thanks a lot ! And yes I know that result ... I am only concerned with the classical types $A,B,C,D$ in which case, rational-smooth ness implies smoothness only for Type A and D... the Type A case is easier which is what I was trying to prove by hand first $\endgroup$
    – user102248
    May 7, 2019 at 22:05

Your Answer

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

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