In a larger proof of some theorem in algebraic geometry, an argument is used which I don't really understand. Giving the full proof is clearly over the top, so I tried to minimize the information needed (although I'm not sure if more details are needed). I think it boils down to the following:

Let $f,g \in K[x_1^{\pm1},\dots,x_n^{\pm1}]$ be Laurent polynomials in multiple variables, where $K$ is a field and $f$ is irreducible. Suppose that $\langle f,g \rangle \cap K[x_1^{\pm1},\dots,x_{n-1}^{\pm1}] = \{0\}$. Then, since $f$ is irreducible, $g$ is a multiple of $f$.

Here, "$\langle f,g \rangle$" denotes the ideal generated by $f$ and $g$. I don't clearly see the argument behind it. Can somebody tell me why this needs to hold?

Thanks in advance!

EDIT: Another piece of information that might be helpful is that $f$ has the following special form: When regarding $f$ as polynomial in $x_n$ with coefficients in $K[x_1^{\pm1},\dots,x_{n-1}^{\pm1}]$, the coefficients are all monomials of the form $\displaystyle cx_1^{u_1} \cdots x_{n-1}^{u_{n-1}}$ for $c \in K$ and $u_1, \dots, u_{n-1} \in \mathbb Z$.


After multiplying by an appropriate power of $x_{n}$, we may assume that $f,g \in K[x_{1}^{\pm 1} , \dotsc , x_{n-1}^{\pm 1} , x_{n}]$ and that $f$ has nonzero constant term when viewed as a polynomial in $x_{n}$ with coefficients in $K[x_{1}^{\pm 1} , \dotsc , x_{n-1}^{\pm 1}]$ (this implies $f$ is irreducible in $K[x_{1}^{\pm 1} , \dotsc , x_{n-1}^{\pm 1} , x_{n}]$). Now we can apply the following lemma to the case $A = K[x_{1}^{\pm 1} , \dotsc , x_{n-1}^{\pm 1}]$:

Let $A$ be a UFD. Let $f,g \in A[x]$ be polynomials such that $f$ is irreducible in $A[x]$ and $(f,g) \cap A = (0)$. Then $f$ divides $g$ in $A[x]$.

Proof: We may assume that $g \not\in A$. Suppose $f$ does not divide $g$ in $A[x]$. Set $F := \mathrm{Frac}(A)$. Then $f$ is irreducible in $F[x]$ as well by Gauss' lemma. Moreover $f$ does not divide $g$ in $F[x]$ (since $A[x]$ is a UFD and irreducibles in $A$ remain irreducible in $A[x]$). Since $F[x]$ is a PID, there exist $a,b \in F[x]$ such that $af+bg = 1$. Clearing the denominators of $a,b$ gives a nonzero element of $(f,g) \cap A$, contradiction.


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.