local parameter on an irreducible affine algebraic curve

606 Views Asked by At

On page 14 of Shafarevich's Basic Algebraic Geometry 1, it is stated that for an irreducible affine algebraic curve $X: f(x,y) = 0$, and a nonsingular point $P \in X$, there is a regular function $t$ (called the local parameter) that vanishes at $P$, and such that for every rational function $u$ that is not identically $0$ on $X$, $u = t^k v$ for some $k$ and some regular function $v$ where $v(P) \neq 0$.

Now, it is proved in the case where $P = (0,0)$ and where $\dfrac{\partial f}{\partial y} \neq 0$, and in this case we get $t = x$.

He calls $k$ the multiplicity of the zero of $P$ on $u$.

My questions are:

(1) How would our function $t$ change if we didn't have $P = (0,0)$ or $\dfrac{\partial f}{\partial y} \neq 0$?

(2) What is the significance of the number $k$? In particular, is it related at all to the multiplicity of a point on $X$ ($m_P(X)$ as in Fulton's Algebraic Curves)?

(3) In the absence of any commutative algebra machinery (e.g. local rings), how can I make sense of any of this?!

For the answers, it would be nice if the use of things like local rings, etc..., could be kept to a minimum. I am more or less familiar with just pages 1-14 of Shafarevich.

EDIT: I realize this is asking for a lot, so any partial answers will get votes, and in time, I'll select the best one.

1

There are 1 best solutions below

2
On BEST ANSWER

Some quick answers:

  1. We can always perform an affine-linear change of coordinates to make the assumptions $P=(0,0)$ and $\partial f / \partial y \neq 0$ true in the new coordinates. (Note that $P$ is a nonsingular point of the curve, at least one of $\partial f / \partial x$ and $\partial f / \partial y$ must be nonzero there.) If you want to translate back to the original coordinates, the function $t$ will then have the form $t=ax+by+c$ for some constants $a,b,c$.

  2. No, the multiplicity of a rational function is not related to the multiplicity of a point. (Indeed, we are talking about a nonsingular point $P$ on $X$, so $P$ always has multiplicity one.) As Shafarevich's terminology indicates, you are supposed to think of this as the order of vanishing of the function $u$ at the point $p$. It might help to recall the case of functions on $\mathbf C$: a meromorphic function $f$ can be expanded in a Laurent series around 0 of the form $$\sum_{k=N}^\infty a_k z^k $$ with $a_N \neq 0$; the standard definition is that $f$ vanishes to order $N$ at $0$. (Note that $N$ might be negative!) To see the similarity with your definition, rewrite the sum as $$z^N \left(\sum_{k=0}^\infty a_{k+N} z^K \right)$$ and note that the term in parentheses in nonzero at $0$.

  3. I hope that my answer to 2. helps with this.