Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a problem with this exercise:

How many sequences of rational numbers converging to 1 are there?

I know that the number of all sequences of rational numbers is $\mathfrak{c}$. But here we count sequences converging to 1 only, so the total number is going to be less. But is it going to be $\mathfrak{c}$ still or maybe $\aleph _0$?

share|cite|improve this question
If $x_n\rightarrow0$, then $1\pm x_n\rightarrow 1$ for any choice of signs. – David Mitra Jan 9 at 14:02
So the number of the sequences converging to 1 is equal to number of the sequences converging to 0. But what now? – Bartłomiej Sługocki Jan 9 at 14:07
encode in $x_n$ a sequence of bits, so that the answer is $\ge$ than $2^{\aleph_0} = \aleph_1$. then considering a general sequence of rationnal numbers (not constrainted to converge to $1$) show that the answer is $\le$ than $2^{\aleph_0} = \aleph_1$ – user1952009 Jan 9 at 14:07
@user1952009 $2^{\aleph_0}$ isn't necessarily $\aleph_1$ - that equality is the continuum hypothesis, which is independent of ZFC. ($2^{\aleph_0}$ does however equal $|\mathbb R|$) – Milo Brandt Jan 9 at 17:20
How many infinite subsets of $\{1+\frac1n:n\in\Bbb N_+\}$ are there? – Akiva Weinberger Jan 10 at 22:05
up vote 4 down vote accepted

We have that non-repeating (injective) sequences of elements in $\{\,1+1/n:n\in\mathbb{N}\,\}$ form a continuum, and all of them have limit $1$, so our set is at least a continuum. Since also all rational sequences form a continuum, our set is also at most a continuum.

share|cite|improve this answer

The number of sequence of rational numbers converging to $1$ is not countable. Suppose you get all squences by $(a_n^{(1)})_{n\in \mathbb{N}}, (a_n^{(2)})_{n\in \mathbb{N}}, (a_n^{(3)})_{n\in \mathbb{N}}, \dotsc$ Define a sequence $(b_n)_{n\in \mathbb{N}}$ by $$b_{k}:=\begin{cases}1 & a_k^{(k)}\neq 1\\ 1+\frac{1}{n}& a_k^{(k)}= 1 \end{cases}$$ Then $\lim_{n\to\infty}b_n=1$ but $(b_n)_{n\in \mathbb{N}}\neq (a_n^{(i)})_{n\in \mathbb{N}} $ for all $i\in\mathbb{N}$.

share|cite|improve this answer
Is it possible to prove it's $2^{\aleph_0}$ without CH? – Kevin Jan 10 at 6:34
@Kevin: Yes. At position $n$ in the sequence, you can have $1+\frac 1n$ or $1$. That is in bijection with the reals in $(0,1)$ expressed in binary. – Ross Millikan Jan 10 at 16:51

For each real number $x$ define $$a_n(x):= 1+ \frac{ \lfloor xn \rfloor}{n^2}$$

Show that this is a one-to-one function from $\mathbb R$ to the set of sequences converging to $1$.

share|cite|improve this answer
@DavidRicherby That means that the function is not ONTO... I never said anything about that ;) – N. S. Jan 10 at 19:04
Oh. Duh. I hate the term "one-to-one" because, to me, it sounds much more like it means "bijective" than "injective" so I always make that mistake. – David Richerby Jan 10 at 21:14

I believe there is a continuum number of such sequences. For ease of writing, let's try to equivalently count the number of sequences converging to $0$. First note that there are countably many rational numbers in the interval $[-a,a]$. Now set for every sequence we will consider $x_i \in [\frac{-1}{i},\frac{1}{i}]$, so that each sequence of $x_i$'s will converge to $0$. The number of such sequences is equivalent to the number of functions from $\mathbb{N}$ to itself, which is provably uncountable.

share|cite|improve this answer
the question was indeed how to prove that the number of such sequences is the number of functions from $\mathbb{N}$ to itself ? – user1952009 Jan 9 at 17:43
You have demonstrated there is at least a continuum of such sequences. Can you provide an upper bound? – Jan Dvorak Jan 10 at 14:02
@JanDvorak: OP correctly said the number of all sequences of rationals is $\mathfrak c$. The ones converging to $0$ are a subset of these, so $\mathfrak c$ is an upper bound. – Ross Millikan Jan 10 at 16:47

$q \pm \frac{1}n$ is a sequence converging to $q$ for any sequence of $\pm$ signs.

That is a lower bound matching the stated upper bound. This solves the problem in the sense that the Schroeder Bernstein principle applies. Finding a specific 1-1 correspondence between rational convergent sequences and 0-1 sequences is more complicated.

share|cite|improve this answer
Isn't simply $\mathbb{N}^\mathbb{N}$ a continuum, whence also $\mathbb{Q}^\mathbb{N}$ is? – yo' Jan 10 at 21:53
The convergent sequences are a superset of a copy of 2^N from the +- signs, and a subset of a Q^N. But this is a Schroeder-Bernstein argument, not a direct bijection to 2^N. – zyx Jan 15 at 0:33

Suppose you have a sequence of rationals $(a_i)$ that converges strictly monotonically to $1$ from above, that is, for all $i$, $a_i > a_{i+1}$. Consider an arbitrary sequence of non-negative integers $(c_k)$. Construct a subsequence of $(a_i)$ by skipping the first $c_0$ elements of $(a_i)$ and set $b_0$ to the next element of $(a_i)$; then skip another $c_1$ elements of $(a_i)$ and set $b_1$ to the next element of $(a_i)$; then skip another $c_2$ elements of $(a_i)$, and so forth. This produces a sequence $(b_k)$ which is a subsequence of $(a_i)$.

Any two distinct sequences of integers $(c_k)$, $(d_k)$ will produce two distinct subsequences of $(a_i)$ via this procedure. For if $m$ is the least integer such that $c_m \neq d_m$, the first $m$ elements of the subsequences produced for $(c_k)$ and $(d_k)$ will be equal, but the next elements will be different, since we skip different numbers of elements to find the next element of each subsequence, and no two elements of $(a_i)$ are equal.

That is, following this procedure, for every sequence of non-negative integers there is a unique sequence of rationals that converges to $1$. Let $A$ be the set of all sequences of non-negative integers and $S$ be the set of all sequences of rationals converging to $1$; we have shown that $\left\vert{A}\right\vert \leq \left\vert{S}\right\vert$.

But of course every sequence of rationals converging to $1$ is a sequence of rationals, so if $B$ is the set of all sequences of rationals, $\left\vert{S}\right\vert \leq \left\vert{B}\right\vert$.

But we also know that $\left\vert{A}\right\vert = \left\vert{B}\right\vert = \mathfrak{c}$. That is, $\mathfrak{c} \leq \left\vert{S}\right\vert$ and $\left\vert{S}\right\vert \leq \mathfrak{c}$. It follows that $\left\vert{S}\right\vert = \mathfrak{c}$.

share|cite|improve this answer

Well, by Cantor's diagonal it can't be $\aleph_0$. (If it were, list all the sequences; change the i-th term of the i-th sequence; the resulting sequence converges to 1 but wasn't on the list.)

So the question is, is it going to be any cardinality between $\aleph_0$ and $\mathfrak{c}$.

hmm, don't think that's possible. But I'm weak on my theory. I'd think if $\{a_n\} \rightarrow r \in \mathbb R$ then $a_n/r \le b_n \le a_n/r + 1/n; b_n \in \mathbb Q$, then $\{b_n\} \rightarrow 1$. So I think it must be $\mathfrak{c}$.

share|cite|improve this answer

protected by Community Jan 11 at 5:49

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.