8
$\begingroup$

Let $X \subset \mathbb{C}^n$ be an irreducible affine algebraic curve with coordinate ring $$\mathbb{C}[X] = \mathbb{C}[z_1, \ldots, z_n] / (f_1, \ldots, f_m ) $$ with each $f_i \in \mathbb{Z}[z_1, \ldots, z_n]$. Define $R[X] = R[z_1, \ldots, z_n] / (f_1, \ldots, f_m )$ for any subring $R \subseteq \mathbb{C}$ containing $\mathbb{Z}$.

Now, if $f\in \mathbb{C}[X]$ can be represented by a polynomial in with integer coefficients, then $\mathbb{C}[X]$ (resp. $R[X]$) can be regarded as a module over $\mathbb{C}[f]$ (resp. $R[f]$), where $\mathbb{C}[f]$ is the $\mathbb{C}$-algebra generated by $f$ in $\mathbb{C}[X]$. Note that $\mathbb{C}[f]$ is the image of the induced homomorphims $f^\ast \colon \mathbb{C}[\mathbb{A}^1] \to \mathbb{C}[X]$.

Suppose $\mathbb{C}[X]$ and $\mathbb{Z}[X]$ are finite over $\mathbb{C}[f]$ and $\mathbb{Z}[f]$. For a ring $R$ as above, let $\mu_R(f)$ denote the minimal number of generators for $R[X]$ as a module over $R[f]$. Then $$\mu_\mathbb{Z}(f) \geq \mu_\mathbb{Q} ( f) = \mathrm{rank}_{\mathbb{Q}[f]} \mathbb{Q}[X]$$ where the equality follows from the fact that $\mathbb{Q}[X]$ is a finitely generated torsion free module over a PID. The inequality can be strict, however.

What is the appropriate geometric interpretation of the case when $\mu_\mathbb{Z} ( f) > \mu_\mathbb{Q} (f)$? What does this say about $f$ when thinking of it as a regular function on $X$? Does this inequality tell us anything geometric about $X$?

From a few Overflow posts (like this and this), Corollary A3.3 in Eisenbud's Commutative Algebra, and glancing through Hartshorne, it appears that this has to do with algebraic vector bundles on $\mathbb{A}^1$ and/or coherent sheaves, but I can't figure out how the regular function $f$ fits into the picture or what intersecting down to $\mathbb{Z}$ does specifically.

Edit: I believe the module structure of $\mathbb{C}[X]$ over $\mathbb{C}[f]$ corresponds to the push-forward of the trivial line bundle over $X$. This gives a vector bundle of rank $\deg f$ over $\mathbb{A}^1$. Is this correct?

Note: this is my first Overflow question and I previously asked this question on StackExchange where it was upvoted a few times, but had no responses. If it is an inappropriate question for this site, please let me know and I will delete it.

$\endgroup$
2
  • $\begingroup$ For rings, the term finitely generated [as an algebra] usually refers to ring extensions that can be written as $R[x_1,\ldots,x_n]/I$. However, when you write that $\mathbb C[X]$ is finitely generated over $\mathbb C[f]$, it seems that you mean that $\mathbb C[X]$ is finite over $\mathbb C[f]$ (i.e. finitely generated as a module). $\endgroup$ Apr 13, 2016 at 6:14
  • $\begingroup$ Yes, that's exactly what I mean. $\endgroup$ Apr 13, 2016 at 13:55

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.