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 am trying to understand the following proof:

Given $f$ is continuous, prove that for every convergent sequence $(x_n) \to a$ that $\lim_{k\to \infty}f(x_k) = f(a)$

So the prove goes like this (by contradiction and contrapositive?). I found this proof in Spivak's which the whole content will be a quote

If $\lim_{k\to \infty}f(x_k) = f(c)$ were not true, there would be some $\epsilon > 0$ such that for every $\delta$ there is an $x$ with

$0 < |x - a| < \delta$, but $|f(x) - f(a)| > \epsilon$

In particular, for each $n$ there would be a number $x_n$ such that

$0 < |x_n - a| < \delta$, but $|f(x_n) - f(a)| > \epsilon$

Now the sequence $(x_n)$ clearly converges to $c$, but since $|f(x_n) - f(a)| > \epsilon$ for all $n$, the sequence $(f(x_n))$ does not converge to $f(a)$. Which is a contradiction.

When they were doing the contrapositive. They say $\forall \delta >0$, s.t. $\exists \epsilon >0$ etc etc...

Does the epsilon here now depend on $\delta$? If so doesn't that mean it wouldn't make much sense to pick a changing $\delta$? If $\epsilon$ here doesn't change, was the goal just to find some $\delta$ (a decreasing one) as to contradict the non-changing $\epsilon$ ball around $f(x_n)$?

share|cite|improve this question
It is a proof by contradiction. If $\lim_{k \rightarrow \infty} f(x_k) \neq f(c)$, then it means that there exists an $\epsilon$ (which we now fix) such that we can always find a sequence $x_{N_k}$, such that $f(x_{N_k})$ is $\epsilon$ far away from $f(c)$. This in turn shows that $f$ is not continuous. – Calvin Lin Dec 30 '12 at 19:48
So my original doubt was correct right? – Hawk Dec 30 '12 at 19:58
Yes. his proof seems written in a strange way, with the "for every "\ldelta" there is an x with 0 < |x-a| < \delta$, since we do not exactly have an $x$, but only the $x_k$. The idea of building a sequence is correct though. – Calvin Lin Dec 30 '12 at 20:08
up vote 1 down vote accepted

Suppose we have two logical statements $A$ and $B$.

Proof by contradiction: $$\neg A \implies \bot$$

Proof by contrapositive: $$ \neg A \implies \neg B$$

The proof you gave was done by contradiction.

Original statement $A$:

for every $\epsilon > 0$, there exists a $\delta(\epsilon) > 0$...

Negation of $A$:

there exists $\epsilon >0$, such that for every $\delta(\epsilon)>0$...

In other words, if $A$ was true, then for every $\epsilon$, it was possible to find $\delta(\epsilon)$ such that the subsequent conditions held. The negation of $A$ claims that there is at least one case where the subsequent implications do not hold, meaning we can find a black sheep $\epsilon$ such that every $\delta(\epsilon)$ will not lead us to the desired implication.

share|cite|improve this answer
Very helpful! thanks – Hawk Dec 30 '12 at 20:16

Definition of continuity:

If $f:X\subseteq \mathbb{R}\to \mathbb{R}$ and $a\in X$ then $f$ is continuous at $a$ if $$\forall \epsilon>0\ \exists \delta>0\ :\forall x\in X\ \left|x-a\right|<\delta\implies \left|f(x)-f(a)\right|<\epsilon$$

The negation: If $f:X\subseteq \mathbb{R}\to \mathbb{R}$ and $a\in X$ then $f$ is discontinuous at $a$ if $$\exists\epsilon>0:\ \forall \delta>0\ \exists x\in X:\ \left|x-a\right|<\delta\text{ and } \left|f(x)-f(a)\right|\ge\epsilon$$

So $\epsilon$ can be thought as constant (independent of $\delta$), $\delta$ is arbitrary (changing) and $x=x_{\epsilon,\delta}$

To your last question, yes. If you pick $\delta=1,\frac12,...,\frac1n,...$ you have a decreasing $\delta$ that gives $x_n\to a$ and serves as your contradiction.

Notes: I don't know why mr. Spivak writes $0<\left|x-a\right|$. Most definitions of continuity don't contain this (not that it changes too much).

Second, the argument: "for each $n$ there would be a number $x_n$ such that..." is a bit dark. To make it rigorous consider $\delta=1$, $\delta=\frac12$ etc. and construct such a sequence $x_n$

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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