0
$\begingroup$

First Isomorphism Theorem

If $\varphi : R \rightarrow S$ is a ring homomorphism with $ker\varphi=I$, then there is an isomorphism $R/I \rightarrow im\varphi$ given by $r+I\rightarrow\varphi(r)$.

I have trouble to understand the second part of the theorem

then there is an isomorphism $R/I \rightarrow im\varphi$ given by $r+I\rightarrow\varphi(r)$.

I can't image how it looks like. Can anyone explain it more easily? Thank you.

$\endgroup$
  • $\begingroup$ I am writing an answer. I request you to wait. $\endgroup$ – астон вілла олоф мэллбэрг Jun 15 '18 at 8:04
  • $\begingroup$ Thank you! I'm moving on to the next Chapter. $\endgroup$ – user8314628 Jun 15 '18 at 8:05
  • $\begingroup$ Just as a comment, homomorphism is an epimorphism onto its image(trivially). İt is a monomorphism if kernel is 0, and R/I identifies kernel as equivalent to 0. Hence on R/I, it is an isomorphism. $\endgroup$ – Atbey Jun 15 '18 at 8:21
  • $\begingroup$ The terms make me more confused... $\endgroup$ – user8314628 Jun 15 '18 at 8:23
1
$\begingroup$

See, $R/I$ is a ring, whose elements are equivalence classes under the relation $x \sim y$ if $x - y \in I$. That is, if I ask you "name an element of $R / I$", then you would tell me the name of some element in $R$, say $r \in R$, and I will take the equivalence class of $r$, which is denoted $r+I$, as the element of $\frac RI$ which you are referring to.

Let's take an example. Take the ring $\mathbb Z $ under usual addition and multiplication, along with the ideal $3\mathbb Z = \{3n : n \in \mathbb Z\}$, the multiples of $3$.

Then, what are the elements of $\frac{\mathbb Z}{3\mathbb Z}$? You name any element of $\mathbb Z$, and I will interpret the equivalence class it belongs to, as the element you are referring to. For example, if you say $73$, then $73 + 3\mathbb Z$ is an element of $\frac{\mathbb Z}{3 \mathbb Z}$. Similarly, so is $65 + {3 \mathbb Z}$.

Now, every element of $\frac{\mathbb Z}{3 \mathbb Z}$ is of the form $r + {3 \mathbb Z}$ where $r \in \mathbb Z$.


So what is the isomorphism given to you actually doing? Given an element $E$ of $\frac RI$, it is an equivalence class, therefore there is some $r$ such that $E = r + I$. Now, find where that $r$ goes under $\phi$, and send $E$ to $\phi(r)$.


For example, if $\phi : Z \to S$ (where $S$ is some ring) is a map whose kernel is exactly $3 \mathbb Z$, then the isomorphism between $\frac{\mathbb Z}{\mathbb 3Z}$ and $S$ is like this :

The isomorphism would take $73 + {3\mathbb Z}$ to $\phi(73)$. It would take $65 +{3 \mathbb Z}$ to $\phi(65)$. It would take $1 + {3\mathbb Z}$ to $\phi(1)$. And so on.


Now, I hope at least you know what the isomorphism is in the general case. But there is a deeper issue.

If I give you an equivalence class, $E$, in $\frac RI$, I told you there is an element such that $E = r+I$. But $r$ need not be unique! For example, $1 +3 \mathbb Z$ is equal to $73 + \mathbb 3Z$, because $73 - 1 = 72 = 3 \times 24 \in \mathbb 3Z$, so $1 \sim 73$, and related elements belong to the same equivalence class, so $1 + 3 \mathbb Z = 73 + 3\mathbb Z$.

But then, what will $\phi$ do when confronted with this? If I give it $1 + 3\mathbb Z$, then it will return $\phi(1)$, but if I give it $73 + \mathbb Z$, then it will return $\phi(73)$.

How can $\phi $ return two elements, $\phi(1)$ and $\phi(73)$, while taking the same equivalence class $1 + 3\mathbb Z = 73 + \mathbb Z$ as input? Every function has a single input and returns a single output, but it seems there could be two outputs here, $\phi(1)$ and $\phi(73)$!


This is not the case : actually, $\phi(1) = \phi(73)$, by the fact that $\phi(72) = 0$(what is the kernel of $\phi$?) and $73 = 72 +1$ so $\phi(73) = \phi(72 + 1) = \phi(72) + \phi(1) = 0 + \phi(1) = \phi(1)$. So, $\phi$ is not doing anything wrong. More formally, we say it is well-defined in this situation.


Putting the deeper issue to rest, I hope you now see :

  • what the isomorphism from $\mathbb R/I$ to the image of $\phi$ is doing.

  • why it is well-defined (I gave an example, but you can generalize and work it out).


As to why it is an isomorphism, well, that I shall explain only if you require it to be done.

$\endgroup$
0
$\begingroup$

Define $ \psi: R/I \to im \varphi$ by

$$ \psi(r+I):=\varphi(r).$$

Then show that $ \psi$ is well defined (this means that $r+I=s+I $ implies $\varphi(r)= \varphi(s)$) and show that $\psi$ is a ring isomorphism.

$\endgroup$

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.