Let $R$ and $R'$ be rings and let $\phi: R\mapsto R'$ be a ring homomorphism and $N$ an ideal of $R$.

Show that $\phi[N]$ is an ideal of $\phi[R]$, and give an example to show that $\phi[N]$ need not be an ideal of $R'$.

Let $N'$ be an ideal of either $\phi[R]$ or R', and show that $\phi^{-1}[N']$ is an ideal of R.

I just started learning ideals so I am having a lot of trouble with this. I know that the kernel of $\phi$ is an ideal, but I don't know how I can use this.

  • $\begingroup$ I guess you mean $\phi(N)$ need not to be ideal of $R'$. $\endgroup$ – mesel Mar 5 '14 at 16:49
  • $\begingroup$ that's correct, sorry. $\endgroup$ – terrible at math Mar 5 '14 at 16:54
  • $\begingroup$ $\phi$ is from $R$ to $R'$ , how can you talk about $\phi(R^{'})$ $\endgroup$ – mesel Mar 5 '14 at 16:57
  • $\begingroup$ sorry I made lots of mistakes typing this out, edited. $\endgroup$ – terrible at math Mar 5 '14 at 17:00

In order to show that $\phi(N)$ is an ideal we must show that it is an additive group and closed under multiplication in $\phi(R)$

Showing $\phi(N)$ is an additive group


So as $0\in N$ and $\phi$ is a ring homomorphism we have that $\phi(0)=0\in \phi(N)$


Now take $x,y\in \phi(N)$ then by definition of $\phi(N)$ there must exists $a,b\in N$ such that $\phi(a)=x$ and $\phi(b)=y$. Now as $N$ is an ideal we have that $a+b\in N$ and so $\phi(a+b)\in \phi(N)$, then as $\phi$ is a ring homomorphism we have that $\phi(a+b)=\phi(a)+\phi(b)=x+y\in \phi(N)$.


If we have $x\in \phi(N)$ then by definition there must be an $a\in N$ such that $\phi(a)=x$ now as $N$ is an ideal we have that $a^{-1}\in N$ and so $\phi(a^{-1})\in \phi(N)$. Now as $\phi$ is a ring homomorphism we have that $\phi(a^{-1})=\phi(a)^{-1}=x^{-1}\in \phi(N)$

So we have that $\phi(N)$ is an additive subgroup of $\phi(R)$

Now to show that it is an ideal we have to show that if we have $r\in \phi(R)$ then $r\phi(N)\subset \phi(N)$ and $\phi(N)r\subset \phi(N)$

Showing $\phi(N)$ is closed under multiplication of $\phi(R)$

Now if $r\in \phi(R)$ then there must be an $r'\in R$ such that $\phi(r')=r$

Then as $r'N\subset N$ we have that $\phi(r'N)=\phi(r')\phi(N)\subset \phi(R)$ (same argument for $r$ on the right)

Example of image of an ideal is not an ideal

We can define the inclusion map $f:\mathbb{Z}\rightarrow \mathbb{Q}$ and then take an ideal in $\mathbb{Z}$ say $3\mathbb{Z}$ then $f(3\mathbb{Z})$ is not an ideal in $\mathbb{Q}$.

Suppose that it was, then noting that $1\not\in f(3\mathbb{Z})$. But by the definition of ideal we must have, $\frac{1}{3}\in\mathbb{Q}$, $\frac{1}{3}\times 3\in f(3\mathbb{Z})$ which gives a contradiction.

Note: as $\mathbb{Q}$ is a field then it has only the two trivial ideals so it follows directly from this.

Another example is given here: The image of an ideal under a homomorphism may not be an ideal

  • $\begingroup$ just reading the first couple lines -- is it clear that $\phi(R)$ is a ring? I mean, to show we have an ideal of a ring we need to first establish that we have a ring at all. It's probably true that the image of any ring homomorphism is always a ring, but I just want to be sure. Never hurts in math. $\endgroup$ – terrible at math Mar 5 '14 at 17:07
  • $\begingroup$ @terribleatmath as $R\lhd R$ if we show that $\phi(N)$ is an ideal then this is the same as showing $\phi(R)$ is a ring, pretty much. I mean it is basically what I have done above- you should be able to adapt it pretty easily $\endgroup$ – hmmmm Mar 5 '14 at 17:09
  • $\begingroup$ It's easy to check directly: $\phi(a)\phi(b)=\phi(ab),\phi(a)-\phi(b)=\phi(a-b)$. $\endgroup$ – Kevin Carlson Mar 5 '14 at 17:11
  • $\begingroup$ @hmmmm: if you find an example such that $\phi(N)$ is not ideal of $R^{'}$ then your answer is complete. :) $\endgroup$ – mesel Mar 5 '14 at 17:11
  • 1
    $\begingroup$ @terribleatmath it is just running essentially the same argument over and over that things in the image of $\phi$ must have a pre-image in $R$. For these sorts of questions when you say "I don't think I know anything about the inverse image"- you do. You need a bit more confidence to stick with it for longer. You know its the inverse image of a ring homomorphism, so the only thing that you can use is that information- you then just need to play around with it for a while and you will get there :) $\endgroup$ – hmmmm Mar 5 '14 at 23:58

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.