Skip to content

Mathematics · Ch 8 — Differentiation

Derivatives of Standard Inverse Trigonometric Functions

8.2.4

Derivatives of Standard Inverse Trigonometric Functions

This section derives the standard derivative of each inverse trigonometric function, by writing y=(inverse trig of x)y=(\text{inverse trig of }x) as x=(trig of y)x=(\text{trig of }y), differentiating implicitly with respect to yy, and converting the result back into xx using a Pythagorean identity — with the sign fixed by which quadrant the principal-branch angle yy can lie in.

1. y=sin⁡−1xy=\sin^{-1}x, −1≤x≤1-1\le x\le1, −π2≤y≤π2-\frac{\pi}{2}\le y\le\frac{\pi}{2}. Then x=sin⁡yx=\sin y. Differentiating w.r.t. yy: dxdy=cos⁡y=±1−sin⁡2y=±1−x2\dfrac{dx}{dy}=\cos y=\pm\sqrt{1-\sin^2y}=\pm\sqrt{1-x^2}. Since −π2≤y≤π2-\frac\pi2\le y\le\frac\pi2, yy lies in the 1st or 4th quadrant, where cos⁡y≥0\cos y\ge0, so the positive sign is taken: dxdy=1−x2\dfrac{dx}{dy}=\sqrt{1-x^2}. By the inverse-function theorem, dydx=11−x2\dfrac{dy}{dx}=\dfrac{1}{\sqrt{1-x^2}}, for ∣x∣<1|x|<1.

2. y=cos⁡−1xy=\cos^{-1}x, 0≤y≤π0\le y\le\pi: left as an exercise for the student to prove by the same method (writing x=cos⁡yx=\cos y); the standard result is dydx=−11−x2\dfrac{dy}{dx}=-\dfrac1{\sqrt{1-x^2}}.

3. y=cot⁡−1xy=\cot^{-1}x, x∈Rx\in\mathbb R, 0<y<π0<y<\pi. Then x=cot⁡yx=\cot y. Differentiating w.r.t. yy: dxdy=−cosec2y=−(1+cot⁡2y)=−(1+x2)\dfrac{dx}{dy}=-\text{cosec}^2y=-(1+\cot^2y)=-(1+x^2). So dydx=1−(1+x2)=−11+x2\dfrac{dy}{dx}=\dfrac{1}{-(1+x^2)}=-\dfrac1{1+x^2}.

4. y=tan⁡−1xy=\tan^{-1}x, −π2<y<π2-\frac\pi2<y<\frac\pi2: left as homework by the same method (writing x=tan⁡yx=\tan y, dxdy=sec⁡2y=1+tan⁡2y=1+x2\dfrac{dx}{dy}=\sec^2y=1+\tan^2y=1+x^2); the standard result is dydx=11+x2\dfrac{dy}{dx}=\dfrac{1}{1+x^2}.

5. y=sec⁡−1xy=\sec^{-1}x, ∣x∣≥1|x|\ge1, 0≤y≤π0\le y\le\pi, y≠π2y\ne\frac\pi2. Then x=sec⁡yx=\sec y. Differentiating: dxdy=sec⁡ytan⁡y=±sec⁡ysec⁡2y−1=±xx2−1\dfrac{dx}{dy}=\sec y\tan y=\pm\sec y\sqrt{\sec^2y-1}=\pm x\sqrt{x^2-1}. The sign is fixed by noting sec⁡y\sec y and tan⁡y\tan y are both positive in Quadrant I and both negative in Quadrant II, so their product sec⁡ytan⁡y\sec y\tan y is positive throughout this branch; matching that against x>0x>0 (Quadrant I, giving xx2−1>0x\sqrt{x^2-1}>0) and x<0x<0 (Quadrant II, giving −xx2−1>0-x\sqrt{x^2-1}>0) shows dxdy=xx2−1\dfrac{dx}{dy}=x\sqrt{x^2-1} when x>1x>1 and dxdy=−xx2−1\dfrac{dx}{dy}=-x\sqrt{x^2-1} when x<−1x<-1. Reciprocating: dydx=1xx2−1\dfrac{dy}{dx}=\dfrac1{x\sqrt{x^2-1}} for x>1x>1, and dydx=−1xx2−1\dfrac{dy}{dx}=-\dfrac1{x\sqrt{x^2-1}} for x<−1x<-1.

Note 1. A function is increasing where its derivative is positive and decreasing where its derivative is negative. Note 2. The derivative of sec⁡−1x\sec^{-1}x is always positive, because the graph of sec⁡−1x\sec^{-1}x is always increasing (on each piece of its domain).

Figure 1Fig. 1.2.2 — the graph of y = sec⁻¹ x: the branch for x ≥ 1 rises from (1, 0) toward the asymptote y = π/2, and the branch for x ≤ −1 descends from π toward y = π/2 (principal values in [0, π], y ≠ π/2).
Fig. 1 — Fig. 1.2.2 — the graph of y = sec⁻¹ x: the branch for x ≥ 1 rises from (1, 0) toward the asymptote y = π/2, and the branch for x ≤ −1 descends from π toward y = π/2 (principal values in [0, π], y ≠ π/2).

Drawn by us to help you understand the concept clearly, and verified to make sure it's accurate. For exams, practice from your textbook's own diagram.

What this figure shows. A diagram of the coordinate plane's four quadrants used to justify the sign taken for sec y . tan y while deriving the derivative of sec^{-1} x: it marks that y (the angle whose secant is x) lies only in the first or second quadrant under the principal branch 0 <= y <= pi, y not equal to pi/2, and shows that sec y and tan y are both positive together in the first quadrant and both negative together in the second quadrant, so their product sec y tan y is always positive there — which is what fixes the sign of dx/dy …

…