Local ring of a projective curve at a point and order of a rational function at that point

866 Views Asked by At

This is from Silverman's Arithmetic of Elliptic Curves. I'm confused by some of the definitions. Let me first lay down the relevant definitions.

Let $V$ be an affine variety. The ideal of $V$ is $I(V) = \{f\in \bar K[X] \mid f(P)=0\;\forall P\in V \}$. If $I(V)$ is generated by elements in $K[X]$, the $V$ is said to be defined over $K$ and define $I(V/K) = I(V)\cap K[X]$. Define the coordinate rings $$ \bar K[V] = \bar K[X] / I(V) \quad \text{ and } \quad K[V] = K[X]/I(V/K). $$

The function field $K(V)$ is the field of fractions of $K[V]$, and the function field $\bar K(V)$ is the field of fractions of $\bar K[V]$.

Let $P$ be a point on the affine variety $V$. The ideal $M_P$ in the coordinate ring $\bar K[V]$ of $V$ is defined by $M_P = \{ f\in \bar K[V]\mid f(P) = 0\}$. So $M_P$ is a maximal ideal in $\bar K[V]$ since $\bar K[V]/M_P$ is isomorphic to the field $\bar K$ via $f \mapsto f(P)$.

The local ring of $V$ at $P$, denoted by $\bar K[V]_P$, is

$\bar K[V]_P = \{ F \in \bar K[V] \mid F =f/g$ for some $f,g \in \bar K[V]$ with $g(P) \neq 0\}$.

Change of notation. Let $V$ be a projective variety and $P$ be a point on $V$. The function field $K(V)$ is defined to be $K(V\cap \mathbb{A}^n)$. Similarly, $\bar K(V) = \bar K(V\cap \mathbb{A}^n)$. Here the author notes that it doesn't matter which $\mathbb{A}^n$ we choose: for different choices of $\mathbb{A}^n$, the different $K(V)$ are canonically isomorphic. I'm guessing this canonical isomorphism is defined by homogenizing an element of one $K(V)$ and then dehomogenizing it to another $K(V)$???

Choose an affine space $\mathbb{A}^n$ such that $P \in \mathbb{A}^n$. The local ring of $V$ at $P$ is the local ring of $V\cap \mathbb{A}^n$ at $P$.

Let $C$ be a curve, i.e. projective variety of dimension 1.

It's noted above that $M_P = \{ f\in \bar K[C]\mid f(P) = 0\}$ is a maximal ideal in $\bar K[V]$. But the same $M_P$ is also the unique maximal ideal in $\bar K[C]_P$. (Am I right here?)

Now the confusing part.

Let $P$ be a smooth point on a curve $C$. The (normalized) valuation on $\bar K[C]_P$ is

$$ \mathrm{ord}_P: \bar K[C]_P \to \{0, 1, 2,\dots \} \cup \{\infty\},\\ \mathrm{ord}_P(f) = \mathrm{sup}\{d\in\mathbb{Z}\mid f \in M_P^d \}. $$

Using $\mathrm{ord}_P(f/g) = \mathrm{ord}_P(f) -\mathrm{ord}_P(g)$, we extend $\mathrm{ord}_P$ to $\bar K(C)$, $\mathrm{ord}_P: \bar K(C) \to \mathbb{Z}\cup\{\infty\}$.

My questions: $M_P^d$ is a subset of $\bar K[C]$, but $\mathrm{ord}_P$ (before extending) is defined on $\bar K[C]_P$, which is a subset of $\bar K(C)$. So how does the initial defn of $\mathrm{ord}_P$ work?

What does it mean for $\mathrm{ord}_P(f) = 0$ or $\mathrm{ord}_P(f) = \infty$? I don't know what $M_P^0$ or $M_P^\infty$ even means.

The next defn is more confusing.

Let $f \in \bar K(C)$. If $\mathrm{ord}_P(f) > 0$, $f$ has a zero at $P$. If $\mathrm{ord}_P(f) < 0$, $f$ has a pole at $P$. If $\mathrm{ord}_P(f) \ge 0$, $f$ is regular (defined) at $P$ and we can evaluate $f(P)$. Otherwise, $f$ has a pole at $P$ and we write $f(P) = \infty$.

The part about $f$ having a zero or is regular at $P$ seems to me like a theorem/proposition rather than a defn. In fact, in proposition 2.1, the author seems to use the fact that if $\mathrm{ord}_P(f) = 0$, where $f\in \bar K(C)$, then $f(P)$ is defined and $f(P) \neq 0$.

Here a rational function $f \in \bar K(C)$ is defined/regular at $P$, by defn, if $f \in \bar K[V]_P$, i.e. $f$ has a representative whose denominator is non-zero at $P$.

1

There are 1 best solutions below

0
On BEST ANSWER

Firstly, on the canonical isomorphism for different choices of $\mathbb{A}^n$: observe that for two different choices of $\mathbb{A}^n$ you find a point $P$ in the intersection of those two choices which is contained in an affine variety openly contained in the intersection of those two choices. Thus, for each choice the local rings at $P$ are canonically isomorphic. Then convince yourself that $K(V)$ is just the fraction field of $K[V]_P$ for any point $P \in V$.

Second: the ring $K[V]_P$ is the localization of $K[V]$ at the maximal ideal $M_P \subseteq K[V]$. Thus, the maximal ideal in $K[V]_P$ (which we also denote by $M_P$) is given as $\{ \frac{f}{g} \mid f \in M_P \subseteq K[V], g \in K[V] \backslash M_P \} = \{\frac{f}{g} \mid f(P)=0, g(P) \neq 0\} $. Think of it as the functions on $V$ that are defined at $P$ and vanish at $P$.

Thirdly: in the definition of the order function, by $M_P$ we mean the unique maximal ideal in $K[C]_P$ (which is just the localization of $M_P \subseteq K[C]$ in $K[C]_P$).

Fourthly: we have for a ring $R$ and $I \subseteq R$ an ideal, one sets $I^0 = R$. Thus, $M_P^0 = k[C]_P$. That means, $f \in K[C]_P$ is of order $\mathrm{ord}_P(f)=0$ if and only if $f \in K[C]_P$ is a unit (i.e. not contained in the maximal ideal of the local ring), or, in other words, if $f$ does not vanish at $P$ (n.b., by definition then $\frac{1}{f}$ is also an element of $K[C]_P$).

Fiftly, $\mathrm{ord}(f)= \infty$ just means $f \in M_P^k$ for all $k \geq 0$. But in a discrete valuation ring we have $\bigcap_{k \in \mathbb{N}} M^k=0$. So $\mathrm{ord}(f) = \infty$ is true if and only if $f=0$.

Sixthly, the definition on poles and zeroes actually only is a definition. A very geometrically intuitive one though (and that is why it might seem to you that there is some statement hidden in there).

Seventhly, let us check that $\mathrm{ord}_P(f) = 0$ if and only if $f(P)$ is defined and $f(P) \neq 0$ for $f \in K(C)$. Pick a uniformizer $t$ at $P$ (i.e. an element $t$ generating the maximal ideal $M_P$). Write $f= g/h$ with $g,h \in K[C] \subseteq K[C]_P$. Then $\mathrm{ord}_P(g) = \mathrm{ord}_P(h)$, so we have $g= g't^m$, $h=h't^m$ for some $m=\mathrm{ord}_P(g)$ and $g', h'$ of $\mathrm{ord}_P(g')=\mathrm{ord}_P(h')=0$. But then, $f= g'/h'$, and by (Fourthly) we see that $1/h' \in K[C]_P$, so we get that $f \in K[C]_P$. On the other hand, $g' \notin M_P$ (because $\mathrm{ord}_P(g')=0$), so $f(P) \neq 0$.