2
$\begingroup$

Hi I am having certain doubts about the sum of uncountable numbers

In my class on functional analysis we proved that if $\sum_{\alpha \in A} X_{\alpha}$ converges then $X_{\alpha}$ is not equal to zero at most a countably many times. I know that there is a similar question on the Math Stack exchange website but I am having serious troubles understanding the logic behind the proof in the class

The proof given by @Benji makes complete sense.

The sum of an uncountable number of positive numbers

We had a very similar proof in the class which was partly discussed in the comment section .

We used the definition that A sequence $\sum_{\alpha \in A} X_{\alpha}$ is convergent iff $\sup_{F \subset A} \sum_{\alpha \in F}X_{\alpha} < \infty$ where $F$ refers to a finite subset of $A$

The proof is as follows:

Proof: By Contradiction. We assume that $\sum_{\alpha \in A} X_{\alpha}$ converges and $\{ \alpha |X_{\alpha}>0 \}$ is uncountable

we defined sets $S_n=\{\alpha |X_\alpha >\frac{1}{n}\}$ but the we have $\bigcup_{n \in \mathbb{N}} S_n=\{\alpha |X_\alpha >0\}$ But by assumption the RHS is uncountable and since LHS is countable union of sets , then it implies that $\exists N \in \mathbb{N}$ such that $S_N$ is uncountable

(Please note in the following whenever I use $F$ , it is a finite subset of A)

Now the prof concludes that $\sup_{F \subset A} \sum_{\alpha \in F}X_{\alpha} = \infty$ And from the definition in bond it follows that $\sum _{\alpha \in A}X_{\alpha}$ does not converge and we get a contradiction

Now if I consider the set $S_N=\{\alpha_1,\alpha_2,\dots \dots\}$ defined as before i.e $X_{\alpha_j}>\frac{1}{N}$ $\forall \alpha_j \in S_N$ and consider a generic $F=\{\alpha_1,\alpha_2,\dots ,\alpha_k \}$ Then

$\sum_{\alpha \in F \subset S_N} X_{\alpha}>\frac{k}{N}$ for any k as cardinality of F needs to be finite

So therefore I have that $\sup_{F \in \mathcal{F}} \sum_{\alpha \in F} X_{\alpha} <\sup_{k \in \mathbb{N}} \frac{k}{N}=\infty$

where $\mathcal{F}$ is the set of all finite subsets of $A$

**I don't understand why does $S_N$ need to be uncountable for this thing thing to work. I mean wouldn't it work as long as it is infinite(countably)? ** Please let me know if this question is not clear.

$\endgroup$
3
$\begingroup$

The argument would work if $S_N$ was countably infinite, but the point is that this is the only way to get that $S_N$ is infinite. If we have assumed $\{\alpha : X^{\alpha} >0\}$ countable then each $S_n$ could easily be finite.

But your issue is correct: $S_N$ needs only be infinite, nor necessarily uncountable.

$\endgroup$

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.