0

Is it in general true that for $f(x)$ continuous also $\frac{f(x)}{f'(x)}$ is continuous?

If not, are there certain circumstances under which it is?

mucciolo
  • 2,950
Vazrael
  • 2,281
  • You need to say $f$ is differentiable to start with, then the continuity of $f$ is automatic. – zhw. Nov 09 '17 at 21:45

3 Answers3

3

You can have $f(x)$ continuous and $f'(x)$ not. A simple example is $f(x)=|x|$ where $f'(x)$ is discontinuous and undefined at $x=0$. Any curve with "corners" will suffer this fate. Any time $f'(x)$ is continuous and nonzero, $\frac {f(x)}{f'(x)}$ will be continuous. The fact that $f'(x)$ exists is sufficient to guarantee that $f(x)$ is continuous.

Ross Millikan
  • 374,822
2

Just because $f$ is continuous, it doesn't need to be differentiable, i.e. $f'$ might not be defined. In that case $\frac{f(x)}{f'(x)}$ is not defined either, much less continuous.

$f$ can also be differentiable with a non-continuous derivative.

Other questions on this site contain plenty of example of functions that are weird in any way imaginable, just search for it.

  • If a derivative (of a real function) exists on some open interval, doesn't it have to be continuous on that interval? – CiaPan Nov 09 '17 at 21:48
  • @CiaPan If "it" refers to the function, then yes. If "it" refers to the derivative of the aformentioned function, then no. –  Nov 09 '17 at 22:25
  • @G.Sassatelli I'm afraid both my imagination and my English are not good enough for that. Luckily I managed to find an example: $f(x)=x^2\cdot\sin(1/x); f(0)=0$ is differentiable with $f^\prime(x)$ oscillating between $-1$ and $+1$ in each $(-\epsilon,+\epsilon)$ interval whilst $f^\prime(0)=0$, hence $f^\prime$ exists in $\mathbb R$ but is discontinuous at $x=0$. Same for $g(x)=x^2\cdot\sin(1/x^2); g(0)=0$ and even more: $g^\prime(0)=0$ but $g^\prime(x)$ oscillates between $-\infty$ and $+\infty$ in each neighborhood of $x=0$... – CiaPan Nov 10 '17 at 08:13
1

The answer is NO.

As others already said, $f$ being continuous does not imply $f$ is differentiable, hence $\frac {f(x)}{f'(x)}$ needs not be even defined.

For example the Weierstrass function is continuous, but not differentiable. In this case $f/f'$ is not even defined, let alone continuous...!.

Additionally, even if $f'(x)$ exists it may be zero, in which case $\frac {f(x)}{f'(x)}$ is undefined, too.

Let $f(x)=1$. It is continuous on $\mathbb R$. Then $f^\prime(x)=0$ is continuous, too.
However, $$\frac{f(x)}{f^\prime(x)}=\frac 10$$ does not exist, either

Finally, as Henrik points out in the answer, even if $f'(x)$ exists and is non-zero, it may be discontinuous, hence $\frac {f(x)}{f'(x)}$ needs not be continuous.

CiaPan
  • 13,049