I am unable to understand why the derivative of : $$y = \sin^{-1}(2x\sqrt {1-x^2}), {-1\over\sqrt{2}}<x<{1\over\sqrt{2}} $$

is: $$ 2\over\sqrt {1-x^2} $$

and why it cannot be / is not : $$ -2\over\sqrt {1-x^2} $$

well, if we take $x$ as $\cos\theta$ and proceed...

$$ y = \sin^{-1}(2 \cos\theta\sqrt {1-\cos^2\theta}) $$

$$ \Rightarrow y =\sin^{-1}(2\sin\theta \cos\theta) $$

$$ \Rightarrow y = \sin^{-1}(\sin2\theta)$$

$$ \Rightarrow y = 2\theta $$

$$ \Rightarrow y = 2\cos^{-1}x $$

the derivative comes out to be: $$ -2\over\sqrt {1-x^2} $$

Now, at first, I thought that it's due to the domain inequality ${-1\over\sqrt{2}}<x<{1\over\sqrt{2}}$ (if we take the $\cos^{-1}$ of this inequality, we get an absurd solution. $({3\pi \over 4} < \cos^{-1}x < {\pi \over 4})$.

Certainly, ${3\pi \over 4} < {\pi \over 4}$ is False. So I thought this was the reason why we cant take $x$ as $\cos\theta$. ( we take $x$ as $\sin\theta$ and get the right answer $2\over\sqrt {1-x^2}$

Then, the next question was to find the derivative of:

$$ y = \sec^{-1}{1 \over (2x^2 - 1)} , 0<x<{1\over\sqrt{2}} $$

Here taking $x$ as $\cos\theta$ gives the absurdity again, BUT gives the answer too... So, after all, what is the correct way to solve these problems?

  • $\begingroup$ math.stackexchange.com/questions/672575/… $\endgroup$ – lab bhattacharjee Jun 4 at 4:07
  • 2
    $\begingroup$ When $x=\cos\theta$, the continuous branch of $y=\sin^{-1}\sin(2\theta)$ is not $2\theta$, but $\pi-2\theta$. You can see this by trying $x=0$, so $\theta=\pi/2$ and $y=\sin^{-1}(\sin\pi)=0$. $\endgroup$ – user10354138 Jun 4 at 4:09
  • 1
    $\begingroup$ $\sqrt {1-\cos^2 \theta} = \sqrt {\sin^2 \theta} = |\sin \theta|.$ In this case $-\frac {\pi} {4} < \theta < \frac {\pi} {4}.$ In this range of $\theta,$ $\sin \theta \not\geq 0.$ Therefore $|\sin \theta| \neq \sin \theta.$ $\endgroup$ – Dbchatto67 Jun 4 at 4:10
  • $\begingroup$ Please share the details of the last absurdity $\endgroup$ – lab bhattacharjee Jun 4 at 4:20
  • $\begingroup$ @labbhattacharjee to solve the second problem, when we put x as $cos\theta$ and then take the cos inverse, we get pi < pi/4 .... $\endgroup$ – J Shelly Jun 4 at 4:36

Always keep in mind the principal values of


$-\dfrac1{\sqrt2} <x<\dfrac1{\sqrt2}$

actually implies $\dfrac{3\pi}4>\arccos x=y>\dfrac\pi4$ as $\arccos(x)$ is decreasing in $[-\dfrac\pi2,\dfrac\pi2]$


But $-\dfrac\pi2\le u=\arcsin(\sin2y)\le\dfrac\pi2$

So,$u=\pi-2y$ for the above range of $y$

  • $\begingroup$ I did not understand this... $\endgroup$ – J Shelly Jun 4 at 4:34
  • $\begingroup$ @JShelly, please pinpoint the confusion $\endgroup$ – lab bhattacharjee Jun 4 at 4:46
  • $\begingroup$ the 2nd line "actually implies...." $\endgroup$ – J Shelly Jun 4 at 5:04
  • $\begingroup$ @JShelly, If $x<\dfrac1{\sqrt2},\cos^{-1}x=\arccos(x)>\dfrac\pi4$ right? $\endgroup$ – lab bhattacharjee Jun 4 at 5:14
  • $\begingroup$ yes, you're right $\endgroup$ – J Shelly Jun 4 at 6:51

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.