3

Every ideal in $K[x]$ is of the form $(p(x))$ for some $p(x) \in K[x]$. But the result is not valid for the ring $K[x_1,...x_n]$.

Comments:

I was able to solve the first case using the division algorithm. But I have difficulty justifying the second part, I know that the ideal $ (x_1, x_2) $ is an ideal that is not generated by $p(x)$.

Bill Dubuque
  • 272,048
Croos
  • 1,809

3 Answers3

5

Hint $ $ If it is a PID then $(x_1,\cdots,x_n) = (a)$ so $\,a\,$ is a common divisor of nonassociate primes $x_i\,$ so $a$ is a unit, so evaluating at all $\,x_i = 0\,$ implies $\,0\,$ is a unit in the field $K,$ contradiction.

Remark $ $ More conceptually, it is easy to show that PIDs have dimension $\le 1,\,$ i.e. nonzero prime ideals are maximal, but here we have $\,(0)\subsetneq (x_1) \subsetneq (x_1, x_2)\subsetneq (1)$ hence the nonzero prime $(x_1)$ is not maximal, so the ring is not a PID. In fact PIDs are precisely the UFDs of dimension $\le1.$

Bill Dubuque
  • 272,048
1

More generally:

If $R[X]$ is a PID, then $R$ is a field.

Indeed:

Take $a\in R$ with $a\ne0$. Consider the ideal $I=(a,X)$. Then $I=(f)$ for some $f$. Now, $a=fg$ implies that $f$ is a constant. Finally, $X=fh$ implies that $f$ is a unit. Thus, $I=R[X]$. Write $1=ap+Xq$ and evaluate at $X=0$. You get $1=ap(0)$, which implies that $a$ is a unit.

Apply this result to $R=K[x_2,\dots, x_n]$, which is not a field, to conclude that $K[x_1,x_2,\dots, x_n]=K[x_2,\dots, x_n][x_1]$ is not a PID.

lhf
  • 216,483
0

Let $x,y \in K[x,y]$ such that $gdc(x,y)=1$, i.e. there are $p(x,y),q(x,y) \in K[x,y]$ with $1=xp(x,y)+yq(x,y)$, left of this relation there is a polynomial with term known $1$, and right there is a polynomial with term known $0$, but

$$(x,y)= \left \{ xp(x,y)+yq(x,y) : p,q \in K[x,y] \right \}=\left \{ f(x,y) : f(0,0)=0 \right \}$$

and $(x,y)$ can not be principal.

user288972
  • 2,360