Prove arccos(-x)+arccos(x) = pi

GreyNoise
Gold Member
Messages
32
Reaction score
6

Homework Statement


From an old edition of the Anton calculus text, I am asked to prove
cos-1(-x) + cos-1(x) = π
or equivalently
cos-1(-x) = π - cos-1(x)

Homework Equations

The Attempt at a Solution


Earlier I proved that sin-1(x) was odd by noting sin-1(sin(x)) = -sin-1(sin(-x)), so I have tried to start the problem from the same approach, but none of my attempts have been insightful.
I started with cos-1(cos(x)) = x and cos-1(cos(-x)) = -x but I am not going anywhere with it. I'm stuck. It makes geometric sense that the sum cos-1(-x) + cos-1(x) is in fact π when viewing a graph of cos-1(x). But how do I prove this one? It appears that I am not supposed to use derivatives of the inverses to solve this (they are covered after this section in the text). It is from Anton's Brief Edition 1981, p 462. Any hints or the like are appreciated.
 
Physics news on Phys.org
:blushing:
 
Last edited:
GreyNoise said:

Homework Statement


From an old edition of the Anton calculus text, I am asked to prove
cos-1(-x) + cos-1(x) = π
or equivalently
cos-1(-x) = π - cos-1(x)

Homework Equations

The Attempt at a Solution


Earlier I proved that sin-1(x) was odd by noting sin-1(sin(x)) = -sin-1(sin(-x)), so I have tried to start the problem from the same approach, but none of my attempts have been insightful.
I started with cos-1(cos(x)) = x and cos-1(cos(-x)) = -x but I am not going anywhere with it. I'm stuck. It makes geometric sense that the sum cos-1(-x) + cos-1(x) is in fact π when viewing a graph of cos-1(x). But how do I prove this one? It appears that I am not supposed to use derivatives of the inverses to solve this (they are covered after this section in the text). It is from Anton's Brief Edition 1981, p 462. Any hints or the like are appreciated.

Try calling ##\alpha = \cos^{-1}(-x)## and ##\beta = \cos^{-1}(x)##. Convince yourself that the angles on both sides of your equation are in ##[0,\pi]## and then check that they have equal cosines.
 
Thank you epenguin and LCKurtz. I thought over your hints, but I was unable to wrap my head around them; I tried the following.

(1) \alpha and \beta are angles of a right triangle : given

(2) \alpha + \beta + \frac{\pi}{2} = \pi : sum of angles is \pi for a triangle

(3) \cos^{-1}(\cos(\alpha)) + \cos^{-1}(\cos(\beta)) + \frac{\pi}{2} = \pi : substitute definition for inverse

(4) \cos^{-1}(\cos(\alpha)) + \cos^{-1}(\cos(\pi/2-\alpha)) + \frac{\pi}{2} = \pi : because \beta = \frac{\pi}{2} - \alpha

(5) \cos^{-1}(\cos(\alpha)) + \cos^{-1}(\sin(\alpha)) + \frac{\pi}{2} = \pi : because \cos(\pi/2-\alpha) = \sin(\alpha)

From step (3), I also tried
(4) \cos^{-1}(\cos(\alpha)) + \cos^{-1}(\cos(-\beta)) + \frac{\pi}{2} = \pi : because \cos(x) is even

But setting

-\cos(\alpha) = \cos(-\beta))

seems to beg the question and I don't know how get through that. Can you tell me what I am missing here?
 
GreyNoise said:

Homework Statement


From an old edition of the Anton calculus text, I am asked to prove
cos-1(-x) + cos-1(x) = π
or equivalently
cos-1(-x) = π - cos-1(x)

Homework Equations

The Attempt at a Solution


Earlier I proved that sin-1(x) was odd by noting sin-1(sin(x)) = -sin-1(sin(-x)), so I have tried to start the problem from the same approach, but none of my attempts have been insightful.
I started with cos-1(cos(x)) = x and cos-1(cos(-x)) = -x but I am not going anywhere with it. I'm stuck. It makes geometric sense that the sum cos-1(-x) + cos-1(x) is in fact π when viewing a graph of cos-1(x). But how do I prove this one? It appears that I am not supposed to use derivatives of the inverses to solve this (they are covered after this section in the text). It is from Anton's Brief Edition 1981, p 462. Any hints or the like are appreciated.

For notational convenience below, say you want to prove ##\arccos(a) + \arccos(-a) = \pi##. Draw a picture: locate the points ##P_1, P_2## on the unit circle whose x-coordinates are ##x= a## and ##x = -a##. Now stare at the picture.
 
@GreyNoise: You are trying to show ##\alpha = \pi - \beta##. Take the cosine of both sides, written just like that and simplify it.
 
Ray Vickson said:
For notational convenience below, say you want to prove ##\arccos(a) + \arccos(-a) = \pi##. Draw a picture: locate the points ##P_1, P_2## on the unit circle whose x-coordinates are ##x= a## and ##x = -a##. Now stare at the picture.

I think I got it! The unit circle will carry x=a to x=-a if we sweep \pi rads (thnx Ray V.), so the angles \cos^{-1}(x) and \cos^{-1}(-x) are supplementary (thnx LCKurtz). So I proceeded as follows:

(1) \alpha + \beta = \pi
(2) \cos^{-1}(\cos\alpha) + \cos^{-1}(\cos\beta) = \pi : definition of inverse
(3) \beta = \pi - \alpha : subtract \alpha from (1)
(4) \cos(\pi - \alpha) = \cos\pi\cos\alpha+\sin\pi\sin\alpha : trig identity
(5) \cos(\pi - \alpha) = -\cos\alpha : simplify (4)
(6) \cos\beta = -\cos\alpha : sub (3) into (5)
(7) \cos^{-1}(\cos\alpha) + \cos^{-1}(-\cos\alpha) = \pi : sub (6) into (2)
(8) x = \cos\alpha so
(9) \cos^{-1}(x) + \cos^{-1}(-x) = \pi

Wella
 
GreyNoise said:
I think I got it! The unit circle will carry x=a to x=-a if we sweep \pi rads (thnx Ray V.), so the angles \cos^{-1}(x) and \cos^{-1}(-x) are supplementary (thnx LCKurtz). So I proceeded as follows:

(1) \alpha + \beta = \pi

That's what you are trying to prove. You don't usually start with that, unless you are claiming it is true from the geometry. But in that case you don't need the stuff below.

(2) \cos^{-1}(\cos\alpha) + \cos^{-1}(\cos\beta) = \pi : definition of inverse
(3) \beta = \pi - \alpha : subtract \alpha from (1)
(4) \cos(\pi - \alpha) = \cos\pi\cos\alpha+\sin\pi\sin\alpha : trig identity
(5) \cos(\pi - \alpha) = -\cos\alpha : simplify (4)
(6) \cos\beta = -\cos\alpha : sub (3) into (5)
(7) \cos^{-1}(\cos\alpha) + \cos^{-1}(-\cos\alpha) = \pi : sub (6) into (2)
(8) x = \cos\alpha so
(9) \cos^{-1}(x) + \cos^{-1}(-x) = \pi
So you ended up with what you started with.

So, one more time, try taking the cosine both sides of ##\alpha =\pi - \beta##. If the angles have equal cosines in that range then the angles are equal. It is two simple steps and you don't have to use any geometry.
 
LCKurtz said:
So, one more time, try taking the cosine both sides of α=πβ

Is this what you have in mind for me LCK?

\alpha = \pi - \beta
\cos\alpha = cos(\pi - \beta)
\cos\alpha = -cos\beta\hspace{10mm}:by difference trig identity so
\cos^{-1}(\cos\alpha) = \cos^{-1}(-cos\beta)

Now sub back for \alpha and \beta at the top I get

\cos^{-1}(-cos\beta) = \pi - \cos^{-1}(cos\beta)

which has the form

\cos^{-1}(-x) = \pi - \cos^{-1}(x)
 
  • #10
What about differentiating arcos(-x) and arcos(x) separately and noting that the derivatives are the equal?
 
  • #11
GreyNoise said:
Is this what you have in mind for me LCK?

\alpha = \pi - \beta
\cos\alpha = cos(\pi - \beta)
\cos\alpha = -cos\beta\hspace{10mm}:by difference trig identity so
\cos^{-1}(\cos\alpha) = \cos^{-1}(-cos\beta)

Now sub back for \alpha and \beta at the top I get

\cos^{-1}(-cos\beta) = \pi - \cos^{-1}(cos\beta)

which has the form

\cos^{-1}(-x) = \pi - \cos^{-1}(x)
No. You are still overcomplicating it and starting with what you are trying to prove. We are starting by calling$$
\cos^{-1}(-x) = \alpha,~\cos^{-1}(x) = \beta$$
You are trying to show ##\alpha = \pi - \beta## by showing they have equal cosines. So you start by taking the cosines of both sides. The left side gives you ##\cos\alpha## and the right side gives you ##\cos(\pi - \beta) = -\cos(\beta)##, as you have shown. So far so good. So now if ##\cos\alpha = -\cos(\beta)## you are done. From your given, what is ##\cos\alpha## and what is ##\cos\beta##? Give the answers in terms of ##x##. Does ##\cos\alpha=-\cos\beta##?
 
  • #12
AfterSunShine said:
What about differentiating arcos(-x) and arcos(x) separately and noting that the derivatives are the equal?

This section calls for to not use differentiation; I've done it that way before (long time ago), but this is a new restriction, no derivatives.
 
  • #13
LCKurtz said:
No. You are still overcomplicating it and starting with what you are trying to prove. We are starting by calling$$
\cos^{-1}(-x) = \alpha,~\cos^{-1}(x) = \beta$$
You are trying to show ##\alpha = \pi - \beta## by showing they have equal cosines. So you start by taking the cosines of both sides. The left side gives you ##\cos\alpha## and the right side gives you ##\cos(\pi - \beta) = -\cos(\beta)##, as you have shown. So far so good. So now if ##\cos\alpha = -\cos(\beta)## you are done. From your given, what is ##\cos\alpha## and what is ##\cos\beta##? Give the answers in terms of ##x##. Does ##\cos\alpha=-\cos\beta##?
LCKurtz said:
So, one more time, try taking the cosine both sides of α=π−β
Yes, I did get the sinking feeling I was begging the question after I posted
So how about this?

given:\hspace{10mm}\cos^{-1}(-x) = \alpha and \cos^{-1}(x) = \beta
show:\hspace{10mm}\alpha = \pi - \beta

from given
\cos\alpha = -x
\cos\beta = x\hspace{5mm} so
\cos\alpha = -cos\beta

\cos\alpha = \cos(\pi - \beta) = -cos\beta\hspace{10mm}:trig diff id
\cos\alpha = -cos\beta\hspace{10mm}:same as given

so
\alpha = \pi - \beta
 
  • #14
GreyNoise said:
Yes, I did get the sinking feeling I was begging the question after I posted
So how about this?

given:\hspace{10mm}\cos^{-1}(-x) = \alpha and \cos^{-1}(x) = \beta
show:\hspace{10mm}\alpha = \pi - \beta

from given
\cos\alpha = -x
\cos\beta = x\hspace{5mm} so
\cos\alpha = -cos\beta

\cos\alpha = \cos(\pi - \beta) = -cos\beta\hspace{10mm}:trig diff id
\cos\alpha = -cos\beta\hspace{10mm}:same as given
so
\alpha = \pi - \beta
Yes, you have now got all the steps, but that is a very awkward arrangement of steps for an argument. Since you seem to now have it figured out, I will suggest a more reasonable way to write it up:

Let ##\alpha=\cos^{-1}(-x)## and ##\beta = \cos^{-1}x##. To show ##\alpha = \pi -\beta##.

##\alpha## and ##\pi - \beta## are both in ##[0,\pi]##, so if their cosines are equal, the angles are equal.

##\cos\alpha = -x##
##cos(\pi - \beta) = \cos\pi\cos\beta+\sin\pi\sin\beta = -\cos\beta = -x##
Therefore ##\cos\alpha = \cos(\pi -\beta)## so ##\alpha = \pi - \beta##, and the proof is finished
 
  • #15
LCKurtz said:
Yes, you have now got all the steps, but that is a very awkward arrangement of steps for an argument. Since you seem to now have it figured out, I will suggest a more reasonable way to write it up:

Let ##\alpha=\cos^{-1}(-x)## and ##\beta = \cos^{-1}x##. To show ##\alpha = \pi -\beta##.

##\alpha## and ##\pi - \beta## are both in ##[0,\pi]##, so if their cosines are equal, the angles are equal.

##\cos\alpha = -x##
##cos(\pi - \beta) = \cos\pi\cos\beta+\sin\pi\sin\beta = -\cos\beta = -x##
Therefore ##\cos\alpha = \cos(\pi -\beta)## so ##\alpha = \pi - \beta##, and the proof is finished

Got it LCK, thnx!
 
Back
Top