Zorn's Lemma Implies Zermelo's Well-Ordering Theorem/Mistake
Jump to navigation
Jump to search
Source Work
1984: Gerald B. Folland: Real Analysis: Modern Techniques and their Applications:
- Prologue
- $2$. Orderings: $(\text P.3)$
Mistake
- Every nonempty set $X$ can be well-ordered.
- Consider the collection $\WW$ of well orderings of subsets of $X$.
- Such well orderings may be regarded as subsets of $X \times X$, so $\WW$ is partially ordered by inclusion.
- It is easy to verify that the hypotheses of Zorn's lemma are satisfied, so $\WW$ has a maximal element.
Correction
It is insufficient for the well-orderings to be ordered by inclusion $\subseteq$.
We additionally need some kind of stability of the initial segment through a chain, for example as established in the main proof.
Proceeding with just $\subseteq$ as partial ordering of $\WW$, consider the sequence of well-ordered sets:
- $W_n = \set {-i: 0 < i \le n} \subseteq \Z$
endowed with the standard ordering.
Then $\sequence {W_n}_{n \mathop \in \N}$ is a chain in $\WW$, but there is no upper bound, for:
- $\ds \bigcup_{n \mathop \in \N} W_n = \set {-i: i \in \N_{>0} } = \Z_{<0}$
has no smallest element with respect to $<$.
Hence the hypotheses of Zorn's Lemma are not satisfied.
Sources
- 1984: Gerald B. Folland: Real Analysis: Modern Techniques and their Applications: $(\text P.3)$