I'll use "affinization" to describe the natural map of schemes $X \rightarrow \text{Spec}(\Gamma(X, \mathcal{O}_X))$. For quasi-affine varieties $X$ this is an open embedding.

Let $G$ be a reductive linear group and $\omega_1, \ldots, \omega_n$ a choice of fundamental weights. We can get a "Plucker embedding" $G/B \hookrightarrow \Pi_{i=1}^n \mathbb{P}(V_{\omega_i})$ which sends a Borel to its eigenspace in each representation. We can also take an embedding: $$G/N \hookrightarrow \Pi_{i=1}^n (V_{\omega_i} - 0) \hookrightarrow \Pi_{i=1}^n V_{\omega_i}$$ This is an open embedding, so $G/N$ is quasiaffine.

The question is: what is the affinization of $G/N$? I have a guess, which is that it is the closure of the image under the embedding. Is this true? If it's true, why (some general fact about the affinization of quasiaffine varieties, or a more specific argument)?

Also, there should be a description of this set that goes something like: collections of vectors $(v_1, \ldots, v_n) \in V_{\omega_1} \times \cdots \times V_{\omega_n}$ such that the $v_i$ are eigenvectors of the same Borel, or zero. I don't really have an argument for this either, though.

  • $\begingroup$ Presumably, you mean the closure of the image of the product map, not the product of the closures. $\endgroup$
    – Ben Webster
    Apr 28 '14 at 8:41

The functions on $G/N$ are a ring of the form $\oplus_{\lambda}V_{\lambda}^\star$ where $\lambda$ ranges over all dominant integral weights. For me, $V_\lambda$ has a fixed highest weight vector $v_\lambda$, and the map is given by $v\mapsto \langle v,gv_\lambda\rangle$ as a function of $g$. The multiplication is given by Cartan product (the unique projection $V_{\lambda_1}^\star\otimes V_{\lambda_2}^\star\to V_{\lambda_1+\lambda_2}^\star$ which sends $v_{\lambda_1+\lambda_2}\to v_{\lambda_1}\otimes v_{\lambda_2}$).

In particular, $\oplus V_{\omega_i}^\star$ are generators for this ring, showing the affinization embeds in the corresponding affine space $\oplus V_{\omega_i}$ as the closure of the image of $G/N$. Furthermore, the form of the map shows that the coset $g$ is sent to $(gv_{\omega_1},\dots, gv_{\omega_n})$, so indeed the affinization is the closure of this set.


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.