Help with basic transfinite induction proof

  • #1
psie
261
32
TL;DR Summary
I'm working on a basic transfinite induction proof, which I'd need some help with. It concerns the construction of the Borel ##\sigma##-algebra.
Some definitions:

Let ##\mathcal{F}_0## be the set of open subsets of ##\mathbb R##.

Successor For any ordinal ##\alpha##, let ##\mathcal{F}_{\alpha+1}## be the set of countable unions of sets ##E_n\in\mathcal{F}_\alpha## and their complements ##E_m: E_m^c\in \mathcal{F}_\alpha##.

Limit For any limit ordinal ##\lambda##, put ##\mathcal{F}_{\alpha}=\cup_{\alpha<\lambda} \mathcal{F}_\alpha##.

Together, these form a nested sequence of sets for all ordinals, that is, ##\alpha<\beta\implies \mathcal{F}_\alpha\subset\mathcal{F}_\beta##. The Borel sigma algebra is ##\mathcal{F}_{\omega_1}##, where ##\omega_1## is the first uncountable ordinal.

The following statement has been left as an exercise in transfinite induction in a handout.

##\mathcal{F}_{\omega_1}\subset\mathcal{G}## for any ##\sigma##-algebra ##\mathcal{G}## containing ##\mathcal{F}_0##, i.e. ##\mathcal{F}_{\omega_1}## is the minimal ##\sigma##-algebra containing ##\mathcal{F}_0##.

I'm looking at Wikipedia and am trying to follow their outline:

1. Show it for the base case, i.e. that ##\mathcal{F}_{0}\subset\mathcal{G}##.

This is, however, trivial, since we said that ##\mathcal{G}## should contain ##\mathcal{F}_{0}##.

2. Show it for successor ordinals, that is, if ##\mathcal{F}_\alpha\subset\mathcal{G}##, then ##\mathcal{F}_{\alpha+1}\subset\mathcal{G}##.

This follows from the fact that ##\mathcal{G}## is a ##\sigma##-algebra and thus it is closed by countable unions and complements, and since ##\mathcal{F}_\alpha\subset\mathcal{G}##, this gives ##\mathcal{F}_{\alpha+1}\subset\mathcal{G}##. However, looking at Wikipedia, I'm unsure if this is all I have to do in the successor step. On Wikipedia, they claim that, if necessary, show also that if ##\mathcal{F}_\alpha\subset\mathcal{G}##, then ##\mathcal{F}_\beta\subset\mathcal{G}## for all ##\beta<\alpha##. I'm not sure if have to or how to show this.

3. Show it for limit ordinals, that is, if ##\mathcal{F}_\alpha\subset\mathcal{G}## for all ##\alpha<\lambda## where ##\lambda## is a limit ordinal, then ##\mathcal{F}_\lambda\subset\mathcal{G}##.

I'm not sure how to do this.

Appreciate any help.
 
Last edited:
Physics news on Phys.org
  • #2
psie said:
2. ... looking at Wikipedia, I'm unsure if this is all I have to do in the successor step. On Wikipedia, they claim that, if necessary, show also that if ##\mathcal{F}_\alpha\subset\mathcal{G}##, then ##\mathcal{F}_\beta\subset\mathcal{G}## for all ##\beta<\alpha##. I'm not sure if have to or how to show this.
By that "if necessary" they mean that, if you can't prove
$$\mathcal{F}_{\alpha+1}\subset\mathcal{G}$$
by assuming
$$\mathcal{F}_{\alpha}\subset\mathcal{G}$$
then try proving it by assuming the stronger precedent:
$$\forall \beta \leq \alpha:\ \mathcal{F}_{\beta}\subset\mathcal{G}$$

Since you were able to do the first, you don't need to do the second. Using a weaker precedent (first case) means you have a stronger result.
 
  • Like
Likes psie and FactChecker
  • #3
Thanks. I guess 3. is not so hard after all.

If ##\mathcal{F}_\alpha\subset\mathcal{G}## for all ##\alpha<\lambda##, then ##\mathcal{F}_\lambda=\cup_{\alpha<\lambda}\mathcal{F}_\alpha\subset\mathcal{G}##. The conclusion follows from the following implications: $$E\in \cup_{\alpha<\lambda}\mathcal{F}_\alpha\implies E\in\mathcal{F}_\alpha \text{ for some } \alpha<\lambda \implies E\in\mathcal{F}_\alpha\subset\mathcal{G} \text{ by assumption.}$$ Also the nested property ##\alpha<\beta\implies \mathcal{F}_\alpha\subset\mathcal{F}_\beta## for any ordinals ##\alpha,\beta## is not hard to show. Either ##\beta## is a successor ordinal or a limit ordinal. If it is a successor ordinal, note that ##E\in\mathcal{F}_\alpha\implies E\in\mathcal{F}_{\alpha+1}## for any ##\alpha## since every ##\mathcal{F}_\alpha## contains ##\emptyset## and ##\mathbb R## (and thus by complements and countable unions, so will ##\mathcal{F}_{\alpha+1}##). If ##\beta## is a limit ordinal, then ##E\in\mathcal{F}_\alpha## implies it is in the union of all ##\mathcal{F}_\alpha## for ##\alpha<\beta##, so ##E\in \cup_{\alpha<\beta}\mathcal{F}_\alpha=\mathcal{F}_\beta##.
 
Last edited:

Similar threads

Back
Top