The Student Room Group

Prove cos2(x)+sin2(x) = 1

Original post by ben-smith
Someone needs to catalogue the awesomeness in this thread, perhaps in the style of the STEP solution threads (eg).


You mean something like this?
EDIT: Work is really boring today.

Dadeyemi
Rotation Matrices

generalebriety
Unit Circle

Simba
f'(x) = 0

SsEe
Power Series

nuodai
Power Series / Complex Numbers
Euler's Formula
*cough*
*cough* (again)

L'art pour l'art
Addition Formula (Cosine)

Jtking3000
Maclaurin series

joe17292
Simple Harmonic Motion

ziedj
So concise I missed it!

IrrationalNumber
Integration

Zuzuzu
Not sure what to call this...
Addition Formula (Sine)
arccos / arcsin
Osborne's Rule

Farhan.Hanif93
Vectors - Dot/Cross Product

Perpetuality,
Feat. Klein Group: Finite Simple Group (Of Order Two)

Mr M
Trig Ratios, but no Pythag!
Integral cosh(t)

Reminisce
Euler's Formula

DFranklin
...Mathematical Stuff

ukdragon37
Integration

TwilightKnight
Hyperbolic Trig

EDIT2: I want to include tony's posts here, since they make this thread so much more hilarious...
EDIT3: Oooooooh lunchtime.
(edited 12 years ago)

Scroll to see replies

Reply 1
Well this is just for fun,
I proved it using some matrices stuff but i wanna know if any1 else has an interesting proof.

Prove cos2(x) + sin2(x) = 1

see how many different proofs we can get.

my proof is added as an attachment.
Circle: x^2 + y^2 = 1. Some point P has coordinates (x, y) where y/x = tan theta, and so x = cos theta, y = sin theta, and use Pythagoras. Standard proof.
Reply 3
Let f(x) = cos²x + sin²x.

f'(x) = -2sinxcosx + 2sinxcosx = 0.

However, f(0) = 1.

Therefore, f(x) = 1.

QED :smile: .
Reply 4
Here's a proof from the power series definition. Certainly not the neatest way but it assumes comparatively little real analysis.

Firstly we need to know how to multiply series:

Assuming absolute convergence (which the series for sin and cos satisfy!):

(n=0an)(n=0bn)=n=0cn\displaystyle \left( \sum_{n=0}^{\infty} a_n \right) \left( \sum_{n=0}^{\infty} b_n \right) = \sum_{n=0}^{\infty} c_n

where cn=k=0nakbnk\displaystyle c_n = \sum_{k=0}^{n} a_kb_{n-k}

Next we need two sums which crop up in the multiplication of the sin and cos series:

(1+x)2n=k=02n(2nk)xk\displaystyle (1+x)^{2n} = \sum_{k=0}^{2n} {2n \choose k} x^k

(1x)2n=k=02n(2nk)(1)kxk\displaystyle (1-x)^{2n} = \sum_{k=0}^{2n} {2n \choose k} (-1)^k x^k

Adding these we get the cancellation when k is odd and doubling up when k is even. Putting x=1 (and being careful when n=0) we have:

k=0n(2n2k)={1      n=022n1 n>0\displaystyle \sum_{k=0}^{n} {2n \choose 2k} = \left\{ {1\ \ \ \ \ \ n=0 \atop 2^{2n-1}\ n>0} \right.

Doing the same with 2n+2 replacing 2n and subtracting we get the even terms cancelling and the odd terms doubling up and so:

k=0n(2n+22k+1)=22n+1\displaystyle \sum_{k=0}^{n} {2n+2 \choose 2k+1} = 2^{2n+1}

(disclaimer: No guarantee those are the easiest proofs. Just the first ones I thought of!!)

cos(x)=n=0(1)nx2n(2n)!\displaystyle \cos(x) = \sum_{n=0}^{\infty} (-1)^n \frac{x^{2n}}{(2n)!}

cos2(x)=n=0cn\displaystyle \cos^2 (x) = \sum_{n=0}^{\infty} c_n

where cn=k=0n(1)kx2k(2k)!(1)nkx2n2k(2n2k)!\displaystyle c_n = \sum_{k=0}^{n} (-1)^k \frac{x^{2k}}{(2k)!} (-1)^{n-k} \frac{x^{2n-2k}}{(2n-2k)!}

=(1)nx2n(2n)!k=0n(2n2k)={1(1)nx2n(2n)!22n1\displaystyle = \frac{(-1)^n x^{2n}}{(2n)!} \sum_{k=0}^{n} {2n \choose 2k} = \left\{ {1 \atop \frac{(-1)^n x^{2n}}{(2n)!} 2^{2n-1}} \right.

And similarly with sin:

sin2(x)=n=0cn\displaystyle \sin^2(x) = \sum_{n=0}^{\infty} c_n

where cn=k=0n(1)kx2k+1(2k+1)!(1)nkx2n2k+1(2n2k+1)!\displaystyle c_n = \sum_{k=0}^{n} (-1)^k \frac{x^{2k+1}}{(2k+1)!} (-1)^{n-k} \frac{x^{2n-2k+1}}{(2n-2k+1)!}

which comes to:

(1)nx2n+2(2n+2)!22n+1\displaystyle \frac{(-1)^n x^{2n+2}}{(2n+2)!} 2^{2n+1}

(using the second binomial sum above)

Finally, cos2(x)+sin2(x)=1+n=1(1)nx2n(2n)!22n1+n=0(1)nx2n+2(2n+2)!22n+1\displaystyle \cos^2(x) + \sin^2(x) = 1 + \sum_{n=1}^{\infty} \frac{(-1)^n x^{2n}}{(2n)!} 2^{2n-1} + \sum_{n=0}^{\infty} \frac{(-1)^n x^{2n+2}}{(2n+2)!} 2^{2n+1}

And if we re-index the second sum we get:

n=0(1)nx2n+2(2n+2)!22n+1=n=1(1)n1x2n(2n)!22n1\displaystyle \sum_{n=0}^{\infty} \frac{(-1)^n x^{2n+2}}{(2n+2)!} 2^{2n+1} = \sum_{n=1}^{\infty} \frac{(-1)^{n-1} x^{2n}}{(2n)!} 2^{2n-1}

And the sums cancel giving the result.

(I'll be impressed if anyone bothers reading through this properly!!)
That was an excellent reply !!!

Thanks. I was looking for that demonstration in order to use it with characteristic function (f(t)=E(itx)) and show that |f(t)|<1.
Reply 6
For what it's worth, here's another (which also comes from the power series definition, just with loads of the working cut out!).

sinx=eixeix2i\sin x = \dfrac{e^{ix}-e^{-ix}}{2i} and cosx=eix+eix2\cos x = \dfrac{e^{ix}+e^{-ix}}{2}, so

sin2x=e2ix2+e2ix4, cos2x=e2ix+2+e2ix4\sin^2 x = -\dfrac{e^{2ix}-2+e^{-2ix}}{4},\ \cos^2 x = \dfrac{e^{2ix}+2+e^{-2ix}}{4}

Adding them together gives sin2x+cos2x=2+24=1\sin^2 x + \cos^2 x = \dfrac{2+2}{4}=1
wow..........old thread.
Reply 8
Original post by boromir9111
wow..........old thread.


I should have noticed that when I saw that generalebriety and Simba had replied :rolleyes:

@diegofrantic: why not create a new thread for your new question rather than digging up a thread from 2007?
1=cos(0)=cos(xx)=cosxcosx+sinxsinx=cos2x+sin2x.\begin{aligned}1 & = \cos(0) \\& = \cos(x-x) \\& = \cos{x}\cos{x}+\sin{x}\sin{x} \\& = \cos^2{x}+\sin^2{x}.\end{aligned}
Reply 10
Woah. I must have been bored on that day... :tongue:
Original post by L'art pour l'art
1=cos(0)=cos(xx)=cosxcosx+sinxsinx=cos2x+sin2x.\begin{aligned}1 & = \cos(0) \\& = \cos(x-x) \\& = \cos{x}\cos{x}+\sin{x}\sin{x} \\& = \cos^2{x}+\sin^2{x}.\end{aligned}


Very elegant.
Reply 12
Here's another.

cos2x+sin2x=(cosx+isinx)(cosxisinx)=eixeix=e0=1\cos^2 x + \sin^2 x = (\cos x + i\sin x)(\cos x - i\sin x) = e^{ix}e^{-ix} = e^0 = 1

Maybe diegofrantic digging this thread up was worthwhile :p:
Original post by nuodai
Here's another.

cos2x+sin2x=(cosx+isinx)(cosxisinx)=eixeix=e0=1\cos^2 x + \sin^2 x = (\cos x + i\sin x)(\cos x - i\sin x) = e^{ix}e^{-ix} = e^0 = 1

Maybe diegofrantic digging this thread up was worthwhile :p:


bee-you-tea-full :yy:
Original post by SsEe
Here's a proof from the power series definition. Certainly not the neatest way but it assumes comparatively little real analysis.

Firstly we need to know how to multiply series:

Assuming absolute convergence (which the series for sin and cos satisfy!):

(n=0an)(n=0bn)=n=0cn\displaystyle \left( \sum_{n=0}^{\infty} a_n \right) \left( \sum_{n=0}^{\infty} b_n \right) = \sum_{n=0}^{\infty} c_n

where cn=k=0nakbnk\displaystyle c_n = \sum_{k=0}^{n} a_kb_{n-k}

Next we need two sums which crop up in the multiplication of the sin and cos series:

(1+x)2n=k=02n(2nk)xk\displaystyle (1+x)^{2n} = \sum_{k=0}^{2n} {2n \choose k} x^k

(1x)2n=k=02n(2nk)(1)kxk\displaystyle (1-x)^{2n} = \sum_{k=0}^{2n} {2n \choose k} (-1)^k x^k

Adding these we get the cancellation when k is odd and doubling up when k is even. Putting x=1 (and being careful when n=0) we have:

k=0n(2n2k)={1      n=022n1 n>0\displaystyle \sum_{k=0}^{n} {2n \choose 2k} = \left\{ {1\ \ \ \ \ \ n=0 \atop 2^{2n-1}\ n>0} \right.

Doing the same with 2n+2 replacing 2n and subtracting we get the even terms cancelling and the odd terms doubling up and so:

k=0n(2n+22k+1)=22n+1\displaystyle \sum_{k=0}^{n} {2n+2 \choose 2k+1} = 2^{2n+1}

(disclaimer: No guarantee those are the easiest proofs. Just the first ones I thought of!!)

cos(x)=n=0(1)nx2n(2n)!\displaystyle \cos(x) = \sum_{n=0}^{\infty} (-1)^n \frac{x^{2n}}{(2n)!}

cos2(x)=n=0cn\displaystyle \cos^2 (x) = \sum_{n=0}^{\infty} c_n

where cn=k=0n(1)kx2k(2k)!(1)nkx2n2k(2n2k)!\displaystyle c_n = \sum_{k=0}^{n} (-1)^k \frac{x^{2k}}{(2k)!} (-1)^{n-k} \frac{x^{2n-2k}}{(2n-2k)!}

=(1)nx2n(2n)!k=0n(2n2k)={1(1)nx2n(2n)!22n1\displaystyle = \frac{(-1)^n x^{2n}}{(2n)!} \sum_{k=0}^{n} {2n \choose 2k} = \left\{ {1 \atop \frac{(-1)^n x^{2n}}{(2n)!} 2^{2n-1}} \right.

And similarly with sin:

sin2(x)=n=0cn\displaystyle \sin^2(x) = \sum_{n=0}^{\infty} c_n

where cn=k=0n(1)kx2k+1(2k+1)!(1)nkx2n2k+1(2n2k+1)!\displaystyle c_n = \sum_{k=0}^{n} (-1)^k \frac{x^{2k+1}}{(2k+1)!} (-1)^{n-k} \frac{x^{2n-2k+1}}{(2n-2k+1)!}

which comes to:

(1)nx2n+2(2n+2)!22n+1\displaystyle \frac{(-1)^n x^{2n+2}}{(2n+2)!} 2^{2n+1}

(using the second binomial sum above)

Finally, cos2(x)+sin2(x)=1+n=1(1)nx2n(2n)!22n1+n=0(1)nx2n+2(2n+2)!22n+1\displaystyle \cos^2(x) + \sin^2(x) = 1 + \sum_{n=1}^{\infty} \frac{(-1)^n x^{2n}}{(2n)!} 2^{2n-1} + \sum_{n=0}^{\infty} \frac{(-1)^n x^{2n+2}}{(2n+2)!} 2^{2n+1}

And if we re-index the second sum we get:

n=0(1)nx2n+2(2n+2)!22n+1=n=1(1)n1x2n(2n)!22n1\displaystyle \sum_{n=0}^{\infty} \frac{(-1)^n x^{2n+2}}{(2n+2)!} 2^{2n+1} = \sum_{n=1}^{\infty} \frac{(-1)^{n-1} x^{2n}}{(2n)!} 2^{2n-1}

And the sums cancel giving the result.

(I'll be impressed if anyone bothers reading through this properly!!)


I would be really surprised as well. What is the point in going in for really complicate proofs when really simple ones exist. I would go for a simple right-angled triangle labelled O, A and H. Then apply pythagoras theorem, followed by dividing each term by H^2.
Looks like this thread should be titled "Look how many ways I can prove sin^2(x) + cos^2(x) = 1".
Original post by tonyhawken
Looks like this thread should be titled "Look how many ways I can prove sin^2(x) + cos^2(x) = 1".


A very similar thing was said by the OP himself.

SsEe's post was excellent. why u hatin?
Reply 17
Original post by tonyhawken
I would be really surprised as well. What is the point in going in for really complicate proofs when really simple ones exist. I would go for a simple right-angled triangle labelled O, A and H. Then apply pythagoras theorem, followed by dividing each term by H^2.


It depends on how you define sin and cos. In a real analysis course, it'll be defined by its power series. But it's not obvious that the power series definition and the angles in a triangle definition are the same. I went from the power series. There are many hidden assumptions in the shorter proofs. Some of them rely on things you don't see until you've done about 4 months at university. My proof only needed about 4 weeks!

Besides. The thread did ask for as many different proofs as possible.
Someone should actually make a "How many ways could we prove..." thread and have a well known identity or theorem for a few weeks before changing it :smile:
Original post by ben-smith


SsEe's post was excellent. why u hatin?


:ditto:

Quick Reply

Latest