3

I have proved that $\mathbb{R}[x]/(x^2-1) \simeq \mathbb{R} \oplus \mathbb{R}$. This fact is a corollary of the generalized C.R.T.

I have proved also that $\mathbb{R}[x]/(x^2+1) \simeq \mathbb{C}$. The isomorphism is given by a map $ax+b \mapsto b + ia$.

I can see why $\mathbb{R}[x]/(x^2+1) \not\simeq \mathbb{R}[x]/(x^2-1)$. This is because a homomorphism always maps $0$ to $0$, and $x^2-1=(x+1)(x-1)$ but $x^2+1$ is irreducible over $\mathbb{R}$.

For the same reason $\mathbb{R}[x]/(x^2) \not\simeq \mathbb{R}[x]/(x^2+1)$. But I still need to prove that $\mathbb{R}[x]/(x^2) \not\simeq \mathbb{R}[x]/(x^2-1)$. I would appreciate any help.

Sergey Filkin
  • 1,340
  • 9
    There is a nilpotent element in $\mathbb R[x]/(x^2)$, not in $\mathbb R[x]/(x^2-1)$. Do you see it ? – Lierre Feb 07 '12 at 12:19
  • @Lierre that is $x$. I think I got it. $\mathbb{R}/(x^2)$ has nilpotent, but $\mathbb{R}/(x^2-1)$ has no, thus there is no map with zero kernel between them. Am I right? – Sergey Filkin Feb 07 '12 at 12:25
  • Yes you are ! More generally, isomorphic rings share all property expressible with the ring operations. And the property $$\exists x : x^2 = 0$$ is such a property. And by the way, you should right $\mathbb R[x]/(x^2)$ instead of $\mathbb R/(x^2)$, which, as is, is meaning less. – Lierre Feb 07 '12 at 12:33
  • Also: the "maps $0$ to $0$" argument seems like it could use some filling out. Another way to talk about the difference between these two rings is to note that $\mathbb R[x]/(x^2 - 1)$ has non-zero zerodivisors, which is obvious from your use of the Chinese remainder theorem. – Dylan Moreland Feb 07 '12 at 12:34
  • @DylanMoreland thank you for your comments and for your edits to my post. I had $\mathbb{R}[x]$ in mind of course. – Sergey Filkin Feb 07 '12 at 12:41
  • @DylanMoreland now I see that the map $ax+b \mapsto -a+ib$ is not suitable. The right map is $ax+b \mapsto b+ia$. Thank you for pointing that. – Sergey Filkin Feb 07 '12 at 12:50

1 Answers1

3

Hint: $\ $ if $\,{\rm char}(F) \ne 2\ $ then by CRT $\,F[x]/(x^2-1)\, \cong\, F[x]/(x-1) \oplus F[x]/(x+1)\, \cong\, F^2\:$ has nontrivial $(\neq 0,1)$ idempotents, e.g. $\,(0,1),\,$ but $F[x]/(x^2)\:$ does not (as one easily verifies).

user26857
  • 52,094
Math Gems
  • 19,574