Let $k$ be a field. Let $F$ be a covariant functor on the category of $k$-algebras to the category of sets. Assume that the opposite functor $F^{op}$ on the category of affine $k$-schemes is a sheaf. (This last assumption might not be necessary.)

Now, suppose that $F(R)$ is a finitely generated $R$-module for all $k$-algebras $R$ and that

$F(R) = F(k)\otimes_k R$.

It seems that in this case $F$ is "representable" by the vector space $$F(k).$$

I don't know what this means and I'm probably missing some very standard construction here. It should mean that $F^{op}$ is representable by a finite type $k$-scheme.

I know one can associate to a finite-dimensional vector space $E$ over $k$ the affine variety $\mathrm{Spec} \ \mathrm{Sym}(E).$ Is it clear that in this case the functor $F$ is representable by the affine variety $\mathrm{Spec} \ \mathrm{Sym}(E)$?


Since $F(k) \cong k^n$ for some $n$, we have $F(R) \cong R^n$, naturally in $R$. But $R^n \cong (\mathbb{A}^n)(R)$, so that $F$ is just the scheme $\mathbb{A}^n$. This is also isomorphic to $\mathrm{Spec}(\mathrm{Sym}(F(k))$.

EDIT: More generally and coordinate free: When $S$ is some base scheme and $\mathcal{E}$ is some quasi-coherent module over $S$, then $\mathbb{V}(\mathcal{E}^\*) := \mathrm{Spec}(\mathrm{Sym}(\mathcal{E}))$ represents the contravariant functor which sends an $S$-scheme $p : T \to S$ the set of $\mathcal{O}_S$-homomorphisms $\mathcal{E} \to p_* \mathcal{O}_T$. When $S=\mathrm{Spec}(k)$ is affine, this is the set of $k$-module homomorphisms $\Gamma(S,\mathcal{E}) \to \Gamma(T,\mathcal{O}_T)$. For example, $\mathbb{V}(\mathcal{O}_S^n) = \mathbb{A}^n_S$.

Now if $F$ is as in your question, we see that $F$ represents the same functor as $\mathbb{V}(F(k))$.

| cite | improve this answer | |
  • 3
    $\begingroup$ I think normally we take Spec(Sym(E*)) (E* means E dual). That convention will make morphisms from Speck to this scheme the same as elements of E, if E is a vector space. $\endgroup$ – 36min May 12 '12 at 16:02
  • $\begingroup$ Right, I've fixed it. $\endgroup$ – Martin Brandenburg May 12 '12 at 17:45

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.