Is this alternative proof of Theorem 3.7 ("Baby" Rudin, Ch. 3) correct and, if so, well written?

249 Views Asked by At

Rudin, in his Principles of Mathematical Analysis, proves the following theorem:

The subsequential limits of a sequence $\{p_n\}$ in a metric space $X$ form a closed subset of $X$.

I've tried to come up with a different proof, but I have a few doubts about its correctness:

Let $E$ be the set of the subsequential limits of $\{p_n\}$ and $q$ a limit point of $E$. We have to prove that $q\in E$.

Since $q$ is a limit point of $E$, there exists a sequence $\{a_n\}$ in $E$ that converges to $q$ (this is a theorem that Rudin proves earlier in the book). Let $V_n$ be the open ball of center $a_n$ and radius $d(q,a_n)$.

Since there is a subsequence of $\{p_n\}$ which converges to $a_n$, $V_n \cap p(\mathbb{N})$, where $p(\mathbb{N})$ is the range of the sequence $\{p_n\}$, is nonempty for every $n\in \mathbb{N}$; that is, $V_n$ always contains at least one term of the sequence.

Therefore, by the axiom of countable choiche, there exists a sequence $\{q_n\}$ such that $q_n \in V_n\cap p(\mathbb{N})$ for every $n$. This happens to be a subsequence of $\{p_n\}$.

Now it is left to prove that $q_n$ converges to $q$: let $\epsilon >0$ be given; then there is $N$ such that $d(q,a_n)<\epsilon$ if $n>N$. We also find that $d(a_n,q_n)<d(q,a_n)$; thus, by the triangle inequality, $d(q,q_n)\leq d(a_n,q_n)+d(q,a_n)<2\epsilon$. Hence the thesis.

I realize that it's definitely more complicated that the one offered by the author, but is it sound nonetheless? Is it written in a fashion that would be accepted in an exam?

1

There are 1 best solutions below

0
On BEST ANSWER

Here is how to prove it without the axiom of choice:...[0]:.. Let $q$ be a limit point of $E$...[1]:Observe that for any member of $E$ there is a strictly increasing $f : N \to N$ such that $(p_{f(n)})_{n \in N}$ converges to it...[2]:.. For all $r>0$, the set $S(r)= \{ n \in N : p_n \in B(q,r) \}$ is infinite because...[2a]:.. B(q,r) contains some $q^* \in E$ and,by [1], $q^*$is the limit of a sequence $(p_{f(n) : n \in N})$ for some strictly increasing $f:N \to N$, and ...[2b]:.. There exists, by the Triangle Inequality, some $s>0$ for which $B(q^*,s) \subset B(q,r)$, so... [2c}:..The infinite set $\{ n \in N : p_{f(n)} \in C \}$ is a subset of $S(r)$....[3]:..Now define $g:N \to N $ recursively by letting $g(0)= \min \{ k: p_k \in B(q,1) \}$ and $g(n+1)= \min \{ k>g(n) : p_k \in B(q,2^{-n-1} \}$. Then $g$ exists by [2] and is strictly increasing and $(p_{g(n)})_{n \in N}$ converges to $q$.