Let $k$ be a field and $ X = \mathbb{P}^n_k $. I have read that every vector field on $ X $, i.e. every global section of $ \mathcal{T}_X = \mathbf{Hom}_{\mathcal{O}_X}(\Omega^1_{X/k},\mathcal{O}_X) $ is induced by a $ k $-derivation $ E : k(x_0,\dots, x_n) \to k(x_0,\dots, x_n) $ with $ E(x_i) $ being homogeneous linear for all $i$.

Indeed given such a derivation on $ D_+( x_j ) \subset \mathbb{P}^n_k$ we can define a derivation on $$ k[x_0/x_i,\dots,x_n/x_i] \to k[x_0/x_i,\dots,x_n/x_i] \\ \frac{x_i}{x_j} \to \frac{x_j E(x_i) - x_i E(x_j)}{x_j^2} $$ and these glue together. However I don't know exactly how to argue that all vector fields are induced this way.

Given a vector field $V$ and $ f/g \in k(x_0,\dots,x_n)^0 $ ($f,g$ are homogeneous of same degree) let $ A = k[x_0,\dots,x_n,1/g]^0 $. Then $D_+(g) = \text{Spec}(A)$. We have $$ V(D_+(g)) \in \text{Der}_k(A,A) $$

We can see this for all $g$. Hence $V$ is induced by a $k$-derivation of $k(x_0,\dots,x_n)^0$. But why is this one induced by a derivation $ E $ of $ k(x_0,\dots,x_n) $ and why is $E(x_i) $ linear homogeneous?


From the Euler Sequence $$0 \rightarrow \mathcal{O}_{\mathbb{P}^n}\rightarrow \mathcal{O}_{\mathbb{P}^n}^{\oplus n+1}(1) \rightarrow \mathcal{T}\mathbb{P}^n\rightarrow 0$$ and ${\rm H}^1(\mathbb{P}^n,\mathcal{O}_{\mathbb{P}^n})=\{0\}$ it follows that every global section of $\mathcal{T}\mathbb{P}^n$ comes from a $(n+1)$-uple of linear homogeneous polynomials $(A_0, \dots, A_n)$ modulo the radial $(x_0, \dots, x_n)$ i.e. $$(B_0, \dots, B_n) = (A_0 + cx_0 , \dots, A_n + cx_n) $$ define the same section for every $c\in k$. Therefore, you can define your derivation setting $E(x_i) = A_i$ and vice versa.


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.