I've been given an interesting proof question and have attempted it myself. Although I roughly understand how to go about the proof, I have one critical doubt, which I will expound upon later.

The following is what needs to be proven:


$\lim_{x\to a^-} f(x)= \infty $

$\lim_{x\to a^+} f(x)= -\infty $

use only the precise definitions of the limit and infinite limit to prove that

$\lim_{x\to a} {1\over f(x)}= 0 $

My attempt

Suppose that:

1) For all $M>0$, there exists $\delta_1 >0$ such that:

$0<a-x<\delta_1\implies f(x)>M$

2)For all $N<0$, there exists $\delta_2 >0$ such that:

$0<x-a<\delta_2\implies f(x)<N$

What we need to show is:

For all $\epsilon>0$, there exists a $\delta>0$ so that

$0<|x-a|<\delta\implies |{1\over f(x)}|<\epsilon$

Consider RHS:

$|{1\over f(x)}|<\epsilon$

$ \implies |f(x)|>{1\over \epsilon}$

Choose $\delta=min(\delta_1, \delta_2)$


$0<|x-a|<\delta\implies f(x)>M \land f(x)<N$

i.e. $0<|x-a|<\delta\implies M<f(x)<N$

Choose $M={1\over \epsilon}$

Therefore, from our assumptions, we have shown there exists a $\delta$ such that

$0<|x-a|<\delta\implies |{1\over f(x)}|<\epsilon$

I would like your help in assessing whether this proof is valid. I spot one issue in it myself: how can the following be?


We know $M>0$ and $N<0$. I wonder what logical missteps I took to reach this conclusion.

Thank you.


You have got the right idea and there is slight problem at the end. You start with the following statements which are correct:

  1. For every $M > 0$ there is a $\delta_{1} > 0$ such that $f(x) > M$ whenever $0 < a - x < \delta_{1}$.
  2. For every $N < 0$ there is a $\delta_{2} > 0$ such that $f(x) < N$ whenever $0 < x - a < \delta_{2}$.

Now we have to use these two statements above to prove the truth of the following statement:

For any $\epsilon > 0$ there is a $\delta > 0$ such that $|1/f(x)| < \epsilon$ whenever $0 < |x - a| < \delta$.

We observe that $|1/f(x)| < \epsilon$ means that $|f(x)| > 1/\epsilon$ and this is equivalent to "$f(x) < -1/\epsilon$ or $f(x) > 1/\epsilon$". Choose $M = 1/\epsilon, N = -1/\epsilon$ and get $\delta_{1}, \delta_{2}$ based on $M, N$ (this is possible because of the two statements mentioned in the beginning. Let $\delta = \min(\delta_{1}, \delta_{2})$ and this will ensure that if $0 < |x - a| < \delta$ then either $f(x) < -1/\epsilon$ or $f(x) > 1/\epsilon$ and therefore $|1/f(x)| < \epsilon$.

| cite | improve this answer | |
  • $\begingroup$ Thank you. The handling of $1\over f(x)$ at the end was what I was most confused about! The 'or'-ing of the statement naturally comes from the modulus - this I did not realize. $\endgroup$ – Sar T1 Sep 25 '16 at 17:08
  • $\begingroup$ @SarT1: you are right. I am glad that my answer helped you to identify your mistake. $\endgroup$ – Paramanand Singh Sep 25 '16 at 19:26

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.