For a function $f(x)$ the definition of its derivative is $$f'(x) = \lim \limits_{\Delta x \to 0} \frac{f(x+\Delta x)-f(x)}{\Delta x}.$$

The derivative $f'(x)$ is supposed to be the same for $\Delta x>0$ and $\Delta x<0$. Which means that if we now set $\Delta x>0$, $$\lim \limits_{\Delta x \to 0} \frac{f(x+\Delta x)-f(x)}{\Delta x}=\lim \limits_{\Delta x \to 0} \frac{f(x-\Delta x)-f(x)}{-\Delta x}$$ needs to be true.

But how do you show that the above expression is true?

I thought of the example where $f(x)=e^x$, and let $\Delta x=3$. At the point $x=0$, $$\frac{f(x+\Delta x)-f(x)}{\Delta x}=\frac{e^3-1}{3}$$and$$\frac{f(x-\Delta x)-f(x)}{-\Delta x}=\frac{e^{-3}-1}{-3}.$$ They are not equal and intuitively I thought that as $\Delta x \to 0$, they will approximately equal to each other but never exactly equal, since the graph of $f(x)=e^x$ is not symmetrical about $x=0$.

Is the statement that the derivative is independent of the sign of $\Delta x$, ie the statement $$\lim \limits_{\Delta x \to 0} \frac{f(x+\Delta x)-f(x)}{\Delta x}=\lim \limits_{\Delta x \to 0} \frac{f(x-\Delta x)-f(x)}{-\Delta x}$$ just an approximation then?

  • $\begingroup$ Hint: for the "negative" direction, instead of $$ \frac{f(x-\Delta x) - f(x)}{-\Delta x} $$ I would write $$ \frac{f(x) - f(x-\Delta x)}{-\Delta x} $$ which better reflects the direction of the slope. $\endgroup$ – Matti P. Jul 4 '18 at 6:05
  • $\begingroup$ @MattiP. the $\Delta x$ below in your second expression should not have a negative sign then, if the $\Delta x $ itself is positive. $\endgroup$ – Taenyfan Jul 4 '18 at 6:07
  • $\begingroup$ The main problem here is that $\Delta x=3$ is too large. The growth rate of this function changes too quickly. Try $\Delta x= 0.1$ to see what you wanted to see. Another problem is that you seem to think $f(0)=0$. But in fact $f(0)=e^0=1$. $\endgroup$ – Jyrki Lahtonen Jul 4 '18 at 6:13
  • $\begingroup$ If the derivative of f as you approach x from from the right weren't equal to the derivative as you approach x from the left, the function would have a sharp kink in it (it wouldn't be differentiable at that point) and $f'(x)$ would be discontinous at x. For your example of $f(x) = e^x$, $f'(x) = e^x$ is definitely continuous. $\endgroup$ – David Diaz Jul 4 '18 at 6:15
  • $\begingroup$ @JyrkiLahtonen thanks, I edited my question. $\endgroup$ – Taenyfan Jul 4 '18 at 6:16

What is equal is the limit indeed

$$\lim \limits_{\Delta x \to 0} \frac{f(x+\Delta x)-f(x)}{\Delta x}=\lim \limits_{\Delta x \to 0} \frac{e^{\Delta x}-1}{\Delta x}=1$$


$$\lim \limits_{\Delta x \to 0} \frac{f(x-\Delta x)-f(x)}{-\Delta x}= \lim \limits_{\Delta x \to 0} \frac{e^{-\Delta x}-1}{-\Delta x}=\lim \limits_{\Delta x \to 0} \frac1{e^{\Delta x}}\frac{e^{\Delta x}-1}{\Delta x}=1$$

  • $\begingroup$ thank you, I understand that the limits equal. Do you know how to show that the two limits will equal for any continuous function $f(x)$? $\endgroup$ – Taenyfan Jul 4 '18 at 6:26
  • $\begingroup$ Recall that continuity is not sufficient (e.f. |x| at x=0). For f differentialble it can be shown by a change of variable. As an alternative recall the geometrical meaning of the derivative as the slope of the tangent at a point, and if f is differentiable the slope of the secant line approaches the same value from both sides. $\endgroup$ – gimusi Jul 4 '18 at 6:41

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.