Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

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

Suppose that $f$ is a real function and that $f'(a)$ exists:

$$ f'(a) = \lim_{h \to 0} \frac{f(a+h) - f(a)}{h}. $$

By replacing all instances of $h$ with $-h$, I can get an equivalent definition

$$ f'(a) = \lim_{-h \to 0} \frac{f(a-h) - f(a)}{-h}. $$


Is it right to say that I can replace $\lim_{-h \to 0}$ with $\lim_{h \to 0}$ in the second definition because $-h$ tending to $0$ is the same as saying $h$ tends to $0$ from the left/bottom?

Since $f'(a)$ exists, the left limit must be equal to the right limit and hence the switch is valid?

share|cite|improve this question
Well remember that for $\lim\limits_{h\to 0}$ to exist, both $\lim\limits_{h\to 0^+}$ and $\lim\limits_{h\to 0^-}$ should both exist and be equal. With that, both your forward difference and backward differences will exist if the corresponding directional derivatives exist... – J. M. Aug 14 '12 at 11:24
well, $\lim_{h \rightarrow 0}$ normally means you allow $h$ to approach 0 from either side (i.e. both tiny positive numbers and tiny negative numbers), so yeah, same thing. – uncookedfalcon Aug 14 '12 at 11:24
Thank you so much for the help! – Legendre Aug 14 '12 at 11:26
I've up-voted Brian M. Scott's answer, but I think mine also covers it and is simpler than his. See below$\ldots\,{}$. – Michael Hardy Aug 14 '12 at 17:00
Hmmm...sorry if this wasn't clear in my question: I was trying to replace only the (-h -> 0) in the limit notation, instead of replacing both (-h -> 0) in the limit notation and (-h) in the fraction. – Legendre Aug 15 '12 at 16:18
up vote 4 down vote accepted

Neither statement involves a one-sided limit. In both statements $h$, whether as $h$ or as $-h$, can be either positive or negative.

Both statements are shorthand. The first says that for each $\epsilon>0$ there is a $\delta>0$ such that $$\left|\frac{f(a+h)-f(a)}h-f'(a)\right|<\epsilon\tag{1}$$ whenever $0<|h|<\delta$. The second says that for each $\epsilon>0$ there is a $\delta>0$ such that $$\left|\frac{f(a-h)-f(a)}{-h}-f'(a)\right|<\epsilon\tag{2}$$ whenever $0<|-h|<\delta$. Now $|-h|=|h|$, so in both cases you're simply saying that a certain inequality involving $f'(a)$ is true whenever $-\delta<h<0$ or $0<h<\delta$. There is no implication that $h$ is positive or that $-h$ is negative. The question is whether saying that $(1)$ is true whenever $-\delta<h<0$ or $0<h<\delta$ is the same thing as saying that $(2)$ is true whenever $-\delta<h<0$ or $0<h<\delta$.

The answer is yes. Let $$q_1(h)=\frac{f(a+h)-f(a)}h$$ and $$q_2(h)=\frac{f(a-h)-f(a)}{-h}\;.$$

A little algebra shows that $q_1(-h)=q_2(h)$ for all non-zero $h$. As $h$ takes on all non-zero values between $-\delta$ and $\delta$, so does $-h$; it just does so in reverse order, so to speak. Thus, saying (as in $(1)$) that $|q_1(h)-f'(a)|<\epsilon$ whenever $-\delta<h<0$ or $0<h<\delta$ is exactly the same as saying that $|q_1(-h)-f'(a)|<\epsilon$ whenever $-\delta<h<0$ or $0<h<\delta$, which is the same as saying (as in $(2)$) that $|q_2(h)-f'(a)|<\epsilon$ whenever $-\delta<h<0$ or $0<h<\delta$.

share|cite|improve this answer

You have $\lim\limits_{h\to0}\cdots\cdots$.

Let $j=-h$. Notice that as $h\to0$, you also get $j\to0$. Since $h=-j$, you can put $-j$ wherever you had $h$, and $j\to0$ where you had $h\to0$.

Then, since it doesn't matter by which name you call a variable, you can call it $h$ instead of $j$. And there you have it.

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.