Definition (As written in Michael Spivak's Calculus)

The function $f$ approaches a limit $l$ near $a$ means: for every $\epsilon >0$ there is some $\delta > 0$ such that, for all $x$, if $0<|x-a|<\delta$, then $|f(x)-l|<\epsilon$.

my question is: why can't it be: $$0<|x-a|\leq \delta,|f(x)-l|\leq \epsilon$$ After looking at limits of functions for a long time just to grasp it's meaning and using the definition quite a lot solving homework I realized I keep writing the same inequality without really understanding why.

The only explanation given in Spivak's book for this part of the definition goes over it without explaining the inequality. I tried looking for an explanation myself but wasn't really able to find anything wrong with it. Is it also possible to write the definition like that or is there a problem with that?

(first non-homework related question :p)

  • 2
    $\begingroup$ Excellent first non-homework related question! $\endgroup$
    – Nameless
    Dec 23 '12 at 15:11
  • 1
    $\begingroup$ I just want to stress that it is important that $|x-a|$ is strictly greater than 0. $\endgroup$
    – Yury
    Dec 23 '12 at 17:06
  • $\begingroup$ Yeah about that part I know, every book and class did even explain it (rather then just giving it as is) $\endgroup$
    – Nescio
    Dec 23 '12 at 17:50

Excellent question. In fact, it turns out that the definition with $\leq$ is equivalent.

Suppose that $f$ approaches the limit $l$ near $a$ using the standard definition (the one with $<$), so that for every $\epsilon >0$ there is some $\delta > 0$ such that, for all $x$, if $0<|x-a|<\delta$, then $|f(x)-l|<\epsilon$.

Then it is also the case that for any $\epsilon>0$, there is some $\gamma>0$ such that, for all $x$, if $0<|x-a|\leq\gamma$, then $|f(x)-l|\leq\epsilon$. What is the $\gamma$ we can produce? Well, taking the $\delta$ we know exists for this $\epsilon$ from the previous paragraph, we can choose $\gamma=\frac{\delta}{2}$, because $$0<|x-a|\leq\frac{\delta}{2}\implies 0<|x-a|<\delta,$$ and we know that any $x$ with $0<|x-a|<\delta$ satisfies $|f(x)-l|<\epsilon$, so certainly any such $x$ will also satisfy $|f(x)-l|\leq\epsilon$.

(In case it's not clear, the fact that we've used the letter $\gamma$ here is of no consequence; I only did so to help keep the two situations mentally separate.)

Thus, if $f$ approaches the limit $l$ near $a$ using the $<$ definition, it will also approach the limit $l$ near $a$ using the $\leq$ definition.

Showing the other implication is similar.

  • $\begingroup$ I actually just commented asking how exactly would you prove it, since I got into a bit of complications of getting rid of the cases of $|x-a|\leq \delta$ and $|f(x)-a|\leq \epsilon$ in the each direction respectively. This clarifies everything nicely. $\endgroup$
    – Nescio
    Dec 23 '12 at 15:20
  • $\begingroup$ Glad to help!$\,$ $\endgroup$ Dec 23 '12 at 15:23
  • $\begingroup$ @ZevChonoles, how does $0\lt|x-a|\le\epsilon$ imply $|f(x)-l|\le\epsilon$? I must be missing some basic inequality rule. $\endgroup$ Jun 15 '15 at 7:10

Let's look at the definitions more carefully. Suppose $f:\mathbb{R}\to \mathbb{R}$ and at point $a$ the "limit" is $L$.

  1. $$(\forall \epsilon>0)(\exists \delta>0)(\forall x\in \mathbb{R})\ (0<\left|x-a\right|<\delta\implies \left|f(x)-L\right|<\epsilon)$$
  2. $$(\forall \epsilon>0)(\exists \delta>0)(\forall x\in \mathbb{R})\ (0<\left|x-a\right|\le \delta\implies \left|f(x)-L\right|\le\epsilon)$$

The two are equivalent. Indeed I will prove $1\implies 2$ and $2\implies 1$

Suppose 1. Let $\epsilon>0$. Then $\exists \delta>0$ so that $\forall x\in \mathbb{R}$, $$\left|x-a\right|<\delta\implies \left|f(x)-L\right|<\epsilon$$ Choose $0<\delta^{\prime}<\delta$. Then, $$0<\left|x-a\right|\le \delta^{\prime}\implies 0<\left|x-a\right|<\delta\implies \left|f(x)-L\right|<\epsilon\implies \left|f(x)-L\right|\le\epsilon$$

Suppose 2. Let $\epsilon>0$ and choose $0<\epsilon^{\prime}<\epsilon$. Then $\exists \delta>0$ so that $\forall x\in \mathbb{R}$, $$\left|x-a\right|\le \delta\implies \left|f(x)-L\right|\le \epsilon^{\prime}$$ Then, $$0<\left|x-a\right|<\delta\implies \left|x-a\right|\le\delta\implies \left|f(x)-L\right|\le \epsilon^{\prime}\implies \left|f(x)-L\right|<\epsilon$$

Note: As Zev point above we can always choose $\delta^{\prime}=\frac{\delta}2$ and $\epsilon^{\prime}=\frac{\epsilon}2$. You may also want to ask why there are only implications and not equivalencies in the 2 definitions


The two ways of writing the definition you have written above are equivalent.

Suppose for every $\epsilon>0$, there is a $\delta>0$ such that $0<|x-a|<\delta$ implies $|f(x)-l|<\epsilon$. Then $0<|x-a|\leq\frac{\delta}{2}$ implies $|f(x)-l|\leq\epsilon$, with $\frac{\delta}{2}>0$.

Conversely, suppose for every $\epsilon>0$, there is a $\delta>0$ such that $0<|x-a|\leq\delta$ implies $|f(x)-l|\leq\epsilon$. Then for every $\epsilon>0$, there is a $\delta>0$ such that $0<|x-a|\leq\delta$ implies $|f(x)-l|\leq\frac{\epsilon}{2}$, since $\frac{\epsilon}{2}>0$; hence $0<|x-a|<\delta$ implies $|f(x)-l|<\epsilon$.


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.