0

$\DeclareMathOperator{Spec}{Spec}\DeclareMathOperator{support}{support}$ Let $R$ be a commutative ring with unity. Let $a, b \in R$. If $(a) = (b)$, then we wish to show that there is some unit $u \in R^\times$ such that $a = ub$.


Let us use the geometric viewpoint for intuition. Let $R$ be the ring of smooth functions on a compact smooth manifold $M$. Thus, $a, b \in R$ are smooth functions on $M$.

$(a) = (b)$ implies the support of $a$ equals the support of $b$.

For a proof, since $a \in (a) = (b)$, we have $a \in (b)$, ie, there is a function $f \in R$ such that $a = fb$. Thus, $a(p) \neq 0 \implies f(p)b(p) \neq 0 \implies b(p) \neq 0$. Thus, $\support(a) \subseteq \support(b)$. By symmetry, we get $\support(b) \subseteq \support(a)$. In total, we have $\support(a) = \support(b)$. This allows us to create a function $u$:

$$ u(p) \equiv \begin{cases} b(p)/a(p) & a(p) \neq 0 \\ 0 & a(p) = 0 \end{cases} $$

We see that $b = ua$. If $u$ is supported everywhere on $M$ (ie, $support(u) = M$), then $u$ is a unit in $R$ such that $(a) = (b)$. $\blacksquare$


I wish to translate the above geometric intuition into algebraic geometry. I believe I should be able to treat $a$ and $b$ as functions on $\Spec(R)$, and repeat the argument as above. Here's my attempt:

Since $(a) = (b)$, this implies that $a = fb$ for some element $f \in R$. Consider a point (prime ideal) $\mathfrak p \subseteq R$. Since $a \equiv_{\mathfrak p} fb$, if $a \not \equiv_{\mathfrak p} 0$, this means $fb \not \equiv_{\mathfrak p} 0$, or that $f \not \equiv _{\mathfrak p} 0$ and $b \not \equiv_{\mathfrak p} 0$. But how do we proceed further?

  1. To create $u_{\mathfrak p} \equiv_{\mathfrak p} b/a$, we need some way to invert the $a \in R/\mathfrak p$. How do we do this?
  2. Even if we succeed in creating all the $u(\mathfrak p)$, how do we glue the solutions together to get an element $u \in R$?

I feel knowing how to translate the geometric argument above into algebraic geometry will really help me see the power of (affine) schemees.

  • 6
    But the claim that $(a)=(b)$ implies $a=ub$ for a unit $u$ is not true in general, for a commutative ring - see here. – Dietrich Burde Jan 16 '22 at 19:28
  • 1
    Your calculation and method will work if you take $R$ to be an integral domain. – Jake Levinson Jan 18 '22 at 00:19
  • @JakeLevinson Could you elaborate in an answer how this can be made to work on an integral domain? I'd really like to see the details to understand concretely how to translate the geometric proof to AG! – Siddharth Bhat Jan 18 '22 at 04:14

0 Answers0