Recall, the Segre embedding is a map $\sigma: \mathbb{P}^m \times \mathbb{P}^n \to \mathbb{P}^{(m+1)(n+1)-1}$ given by \begin{equation} ([x_0:\cdots:x_m],[y_0:\cdots:y_n]) \mapsto [x_0y_0:x_0y_1: \cdots: x_iy_j:\cdots: x_my_n] \end{equation}

Call the image of $\sigma$ the Segre variety $\Sigma_{m,n}$. I vaguely remember reading somewhere that no Segre variety is contained in a hyperplane in $\mathbb{P}^{(m+1)(n+1)-1}$. This certainly feels true enough (since the defining equations for $\Sigma_{m,n}$ are quadratic, not linear), but writing down a proof in the general case seems difficult.

In fact, we may consider the following generalization. Let $X \subseteq \mathbb{P}^m$ and $Y \subseteq \mathbb{P}^n$ be varieties. The Segre product of $X$ and $Y$ is the image $\sigma(X \times Y)$ in $\mathbb{P}^{(m+1)(n+1)-1}$. Can someone supply a proof or reference that if $X$ and $Y$ are nonempty, then $\sigma(X \times Y)$ is never contained in a hyperplane? If this turns out to be false, what conditions on $X$ and $Y$ are needed to guarantee this is true?

Edit: It seems that this is false if either $X$ or $Y$ consists of points only. So perhaps one needs to require that the varieties be at least one-dimensional.


a) It is indeed true that $\Sigma_{m,n}$ is not contained in a hyperplane and, contrary to your fears, writing down a proof is not difficult. Actually it is child's play:

A hyperplane in $\mathbb{P}^{(m+1)(n+1)-1}$ has equation $\sum a_{ij}z_{ij}=0$ and to say that it contains $\Sigma_{m,n}$ means that $\sum a_{ij}x_iy_j=0 \:$ for all $(x,y)\in \mathbb{P}^m \times \mathbb{P}^n$.
But this forces every $a_{ij}$ to be zero [do you see which pair $ (x_i,y_j) \in \mathbb{P}^m \times \mathbb{P}^n$ shows that?], which is impossible for the equation of a hyperplane.

b) However it is quite possible that $\sigma(X \times Y)$ is contained in a hyperplane for subvarieties $X \subseteq \mathbb{P}^m$ and $Y \subseteq \mathbb{P}^n$:

Just take for $X$ the hyperplane $x_0=0$ of $\mathbb{P}^m$and for $Y$ the whole of $\mathbb{P}^n$.
The image $\sigma(X \times Y)$ is then contained in each of the hyperplanes $z_{00}=0,z_{01}=0,\cdots,z_{0n}=0 $ .

  • $\begingroup$ Thank you for your help. (a) was indeed much simpler than I thought and your example for (b) shows that it seems nothing much can be said for Segre products that do not lie in some hyperplane. $\endgroup$ – JHF May 10 '13 at 23:18

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.