Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

For the projective space, we have the very well known Euler sequence $$0 \to \Omega_{P^n_k} \to \mathcal{O}_{P^n_k}(-1)^{n+1} \to \mathcal{O}_{P^n_k} \to 0.$$ Are there any generalizations to other rational homogeneous spaces?

share|improve this question

1 Answer 1

up vote 7 down vote accepted

Here is how it works for the (complex) Grassmannian. I will leave you the pleasure to extend this point of view to others homogeneous spaces (for instance complete and incomplete flag manifolds).

Firs of all, let me give a slightly different point of view for the Euler exact sequence on the projective space. Let $V$ be a complex vector space of dimension $n+1$ and $P(V)$ its corresponding projective space of lines. Call $\underline V=P(V)\times V$ the trivial rank $n+1$ vector bundle with fiber $V$ on $P(V)$. Then you have the short exact sequence \begin{equation} 0\to\mathcal O(-1)\to\underline V\to H\to 0, \end{equation} where $\mathcal O(-1)$ is the tautological line bundle on $P(V)$ and $H$ is the quotient rank $n$ vector bundle $\underline V/\mathcal O(-1)$.

Now, looking at the differential $d\pi_x\colon V\to T_{P(V),[x]}$ of the projection $\pi\colon V\setminus\{0\}\to P(V)$, it is straightforward to see that it defines a canonical isomorphism $$ T_{P(V)}\simeq H\otimes\mathcal O(1). $$ Tensoring the short exact sequence above by $\mathcal O(1)$ and using this canonical isomorphism you get the Euler exact sequence $$ 0\to \mathcal O\to\underline V\otimes\mathcal O(1)\to T_{P(V)}\to 0 $$ (to get your version just notice that $\underline V\otimes\mathcal O(1)\simeq\mathcal O(1)^{\oplus(n+1)}$, then take the dual exact sequence).

Now, let $V$ be of dimension $d$ and $G_r(V)$ be the Grassmannian of $r$-codimensional vector subspaces of $V$. Consider the tautological subbundle $S\subset \underline V=G_r(V)\times V$ and the associated short exact sequence $$ 0\to S\to\underline V\to Q\to 0, $$ where $Q=\underline V/S$. Similarly as before, it is straightforwardly seen that there is a canonical isomorphism $$ T_{G_r(V)}\simeq S^*\otimes Q\simeq\operatorname{Hom}(S,Q). $$ Tensoring by $S^*$, you get your Euler sequence for the Grassmannian, that is $$ 0\to S^*\otimes S\to\underline V\otimes S^*\to T_{G_r(V)}\to 0. $$ Of course, it reduces to the standard one whenever $r=d-1$, and playing with the short exact sequence you can give it your favorite shape (taking duals, tensoring with other vector bundles...).

Observe finally that $$ \det S^*\simeq\det Q $$ is the (very) ample line bundle on $G_r(V)$ which gives the Plücker embedding into projective space and whose first Chern class generates $H^2(G_r(V),\mathbb Z)$.

share|improve this answer
    
Can you elaborate about how we derive $T_{P(V)}\simeq H\otimes \mathcal{O}(1)$ from the differential of the projection map? Are we to understand $d\pi_x$ as the linear map on a fiber of a bundle map $H\otimes \mathcal{O}(1)\to T_{P(V)}$? –  Joe Hannon Sep 30 '13 at 3:37
    
The fiber of $H\otimes \mathcal{O}(1)$ is $V/[x]\otimes [x]^*.$ So let $\sigma\in [x]^*$ and $v+[x]\in V/[x].$ Define the map to be $\sigma\otimes(v+[x])\mapsto \sigma(x)\cdot d\pi_x(v).$ Then show that this is well-defined as a function on $P(V)$, i.e. $\sigma(x)\cdot d\pi_x(v) = \sigma(\lambda x)\cdot d\pi_{\lambda x}(v).$ But is there a more elegant way to say it, i.e. not in terms of an arbitrarily chosen point $x$ in $V$? –  Joe Hannon Sep 30 '13 at 4:11
    
Your proof is more or less the proof I knew... –  diverietti Sep 30 '13 at 9:19

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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