0

Using Zorn's lemma is the approach I've seen. Would like feedback for a different approach I tried.

Similar/related question: Existence of minimal prime ideal contained in given prime ideal and containing a given subset

Proof?: Let $I$ be an ideal and $P$ be a prime ideal of the ring $R$ such that $I \subset P$. Let $$\mathscr{F}:=\{P'\subseteq R \vert (\text{P' is a prime ideal}) \land (I \subsetneq P' \subseteq P))\}$$ . Since $I \subset P \subseteq P$, then $P \in \mathscr{F}$. So $\mathscr{F}\neq \emptyset$. We assert that $\bigcap \mathscr{F} \neq \emptyset$ and that it's a prime ideal. Since $\forall P' \in \mathscr{F}(I \subsetneq P')$, then $I \subsetneq \bigcap\mathscr{F}$. Note $I \subsetneq \bigcap\mathscr{F}$ implies $\bigcap\mathscr{F}\neq \emptyset$. If $ab \in \bigcap\mathscr{F}$ and $a \notin \bigcap\mathscr{F}$, then $\forall P' \in \bigcap\mathscr{F}(ab \in P')$ and $a \notin P'$. Since $\forall P'\in\mathscr{F}$, $P'$ is a prime ideal, then $ab \in P'$ and $a \notin P'$ imply $b \in P'$, for all $P' \in \mathscr{F}$. Thus, $b \in \bigcap\mathscr{F}$, as desired. Lastly, we show $\bigcap\mathscr{F}$ is the minimal prime ideal of $I$. Suppose $\overline{P} \in \mathscr{F}$ such that $\overline{P} \subseteq \bigcap\mathscr{F}$. Since $\overline{P} \in \mathscr{F}$ imply $\bigcap\mathscr{F} \subseteq \overline{P}$, then $\bigcap\mathscr{F} = \overline{P}$. This equivalently, means there exist no prime ideal $\overline{P}$ where $I \subsetneq \overline{P} \subseteq \bigcap\mathscr{F}$, unless $\overline{P} = \bigcap\mathscr{F}$. Therefore, $\bigcap\mathscr{F}$ is the minimal prime ideal of $I$ contained in $P$.

  • 1
    The intersection of proper supersets need not be a proper superset. – Hagen von Eitzen May 26 '19 at 08:18
  • I have follow up question. The corollary of this theorem states: "Every proper ideal of the ring $R$ contains at least one minimal prime ideal." The proof proceeds with saying "Since every proper ideal is contained in a maximal ideal and every maximal ideal is prime, we can apply Theorem 5.15(the proposition of the post above.)." My question is, why was the necessity of an identity element in the ring $R$ left out? I thought this is a strict necessity for the ring, to make all maximal ideals also prime. – TheLast Cipher May 26 '19 at 09:21
  • Your proof is not correct : you go from $a\notin \bigcap\mathscr{F}$ to $\forall P'\in\mathscr{F}, a\notin P'$, which isn't correct. That's why we use Zorn's lemma : it becomes almost true if $\mathscr{F}$ is linearly ordered, and it's "true enough" to ensure the presence of $b$ in the intersection – Maxime Ramzi May 26 '19 at 10:16
  • @TheLastCipher that corollary is false though. In $F[x]/(x^3)$, the ideal generated by $(x^2)$ is a proper ideal containing no prime ideal. (There is exactly one prime ideal, the ideal generated by $x$.) – rschwieb May 26 '19 at 11:19
  • Thank you guys! :) – TheLast Cipher May 26 '19 at 13:35

0 Answers0