I don't understand the $3$rd line in the following proof, why is there an ideal $N$ s.t. $JL=sN$. I thought ''to contain means divides'' is valid only in some special rings, for example Dedekind rings (whereas the converse is true for any commutative unital rings)
Why is there such an ideal?
118 Views Asked by Bumbble Comm https://math.techqa.club/user/bumbble-comm/detail AtThere are 3 best solutions below
On
We have $(s) \supseteq IL$ for principal ideals, so $(s)$ divides $IL$. Hence there exists an ideal $M$ with $JL=(s)M=sM$. Now $JL$ is a subset of $sM$ and an ideal, hence of the form $sN$ (see quasicoherent's answer).
Reference for "to contain is to divide" with principal ideals: P. Clark, Proposition $2$, page $2$.
On
I think we can show this directly; let the set $N \subset R$ be defined as
$N= \{ n \in R \mid sn \in JL \}; \tag 0$
then it is evident that
$sN \subset JL; \tag 1$
since
$JL \subset (s), \tag 2$
every $t \in JL$ is of the form $t = sa$, $a \in R$; since such $a \in N$, this shows that
$JL \subset sN, \tag 3$
so by (1) and (3) we know that
$JL = sN. \tag 4$
We show $N \subset R$ is an ideal: suppose
$n_1, n_2 \in N; \tag 4$
then
$sn_1, sn_2 \in JL, \tag 5$
and since $JL$ is an ideal of $R$,
$s(n_1 - n_2) = sn_1 - sn_2 \in JL, \tag 6$
and so
$n_1 - n_2 \in N, \; \forall n_1, n_2 \in N; \tag 7$
likewise for
$sn = t \in JL, \; n \in N \tag 8$
we have, again since $JL$ is an ideal,
$s(nr) = (sn)r = tr \in JL, \; r \in R, \tag 9$
which shows that
$nr \in N, \; \forall n \in N, \; r \in R; \tag{10}$
combining (7) and (10) shows that $N \subset R$ is an ideal in $R$.

Let me extract the fact you'd like to prove.
Lemma: Let $R$ be a domain, $(a)$ be a principal ideal and $I$ be an ideal with $I \subseteq (a)$. Then there exists an ideal $J$ such that $I = aJ$.
Proof: Take $$ J = (I: (a)) = \{r \in R \mid r (a) \subseteq I\} = \{r \in R \mid ra \in I\} $$ to be the colon ideal. We claim that $I = aJ$.
$\supseteq$: Given $aj \in aJ$, then by definition of $(I: (a))$ we have $aj \in I$.
$\subseteq$: Given $b \in I \subseteq (a)$, then $b = ar$ for some $r \in R$. Since $ar = b \in I$ then $r \in (I : (a)) = J$. Thus $b = ar \in aJ$.
Returning to your question, the above lemma shows that we can take $N = (JL : (s))$.