2

Definition: We define $\displaystyle \sin x = \sum_{n=0}^{\infty}\frac{(-1)^n x^{2n+1}}{\left ( 2n+1 \right )!}, \; x \in \mathbb{R} $ and $ \displaystyle \cos x = \sum_{n=0}^{\infty}\frac{(-1)^n x^{2n}}{\left ( 2n \right )!}, \; x \in \mathbb{R}$.

Well I want to prove using these definitions that:

  • $\displaystyle \sin^2 x+\cos^2 x =1 $
  • $\displaystyle \left ( \sin x \right )'=\cos x, \; (\cos x)' =-\sin x$

I was able to prove most of the properties that these functions have using these definitions but not those two.

For the first one I started as follows: $$\begin{aligned} \sin^2 x +\cos^2 x &=\left ( \sum_{n=0}^{\infty}\frac{(-1)^n x^{2n+1}}{\left ( 2n+1 \right )!} \right )^2+\left ( \sum_{n=0}^{\infty}\frac{(-1)^n x^{2n}}{\left ( 2n \right )!} \right )^2 \\ &= \sum_{m=0}^{\infty}\sum_{n=0}^{\infty}\frac{\left ( -1 \right )^{m}x^{2m+1}}{\left ( 2m+1 \right )!}\frac{\left ( -1 \right )^{n}x^{2n+1}}{\left ( 2n+1 \right )!}+ \sum_{m=0}^{\infty}\sum_{n=0}^{\infty}\frac{\left ( -1 \right )^m x^{2m}}{\left ( 2m \right )!}\frac{(-1)^n x^{2n}}{\left ( 2n \right )!} \end{aligned}$$

then I don't know how to proceed .

As for the second one despite I differentiated the series , I cannot get the series of cosine and vice versa the minus series of sine.


I also have another question regarding the definitions.

The classic definition of the trigonometric functions involves the unit circle. Suppose we define $\sin x, \; \cos x$ as above. Are these two definitions equivelant? Similary if we define $\displaystyle \sin x = \frac{e^{ix}-e^{-ix}}{2i}, \; \cos x =\frac{e^{ix}+e^{-ix}}{2}$ are these three definitions equivelant?

Tolaso
  • 6,886
  • I don't know how to prove that... !! – Tolaso Jan 26 '15 at 21:26
  • To use series to prove that the derivative of the sine is the cosine is kind of odd. After all, to establish the terms of the series of for example the sine, you are already taking its derivatives in order to establish each next coefficient...And the complex sine and cosine are not definitions. They have been derived. – imranfat Jan 26 '15 at 21:26
  • @imranfat I guess if you start with the given definition, we are following another equivalent rule for establishing the next coefficient which does not include the derivatives of sine and cosine. – Alice Ryhl Jan 26 '15 at 21:28
  • I see what you mean, point well taken. I just wanted to point out that one should not end up in a "cyclic" form of reasoning. That is, to show that the derivative of a sine would be a cosine using some tools in which it is already understood that the derivative of a sine is a cosine... – imranfat Jan 26 '15 at 21:32
  • Differentiating the series is completely straightforward, so it's hard to guess where you went wrong. The first term of the series for $\sin x $ is $x$; the derivative of this term is $1$, which is the first term of the series for $\cos x $, The second term of the series for $\sin x $ is $-\frac {x^3}6$; the derivative of this term is $-\frac {x^2}2$, which is the second term of the series for $\cos x $. What went wrong when you tried this? – MJD Jan 26 '15 at 21:32
  • Actually what went wrong was the differentiation... ! I messed up a little bit by hurry.. That part is done.. how about the other one? I don't want to mix differential equations or something like that.. but if it is unavoidable , then that is fine – Tolaso Jan 26 '15 at 21:34
  • 1
    @KristofferRyhl I disagree with your first comment. The geometric definitions are troublesome and require a lot of faith. – Git Gud Jan 26 '15 at 21:35
  • @GitGud yeah, I've thought a bit more about this now, and I agree with that. – Alice Ryhl Jan 26 '15 at 21:38
  • @imranfat you can get the power series from a definition of $\sin$ and $\cos $ as solutions of a certain differential equation without knowing the derivatives ahead of time, so it's not completely crazy to ask for a proof that the power series have the usual properties you expect $\sin $ and $\cos $ to have. – MJD Jan 26 '15 at 21:39
  • I'm almost covered...but one question remains... ARE THE DEFINITIONS EQUIVELANT?

    If so, how can we prove that?

    – Tolaso Jan 26 '15 at 21:42
  • This answer of mine remarks on how the power series representations coincide with the geometric ones. The connection is made with a little combinatorics and the limit $(\sin x)/x \to 1$ as $x\to 0$ (which is effectively a given, since you start with the power series). – Blue Jan 26 '15 at 21:51
  • 1
    @Blue That was a great answer. Nicely said. – Tolaso Jan 26 '15 at 23:03
  • Related: https://math.stackexchange.com/questions/1048/different-definitions-of-trigonometric-functions – Hans Lundmark May 26 '21 at 20:13

1 Answers1

2

Justify that you can differentiate the series term by term and find, for all $x\in \mathbb R$, $$\sin'(x)=\sum \limits_{n=0}^\infty\left(\dfrac{(-1)^nx^{2n}}{(2n)!}\right)=\cos(x).$$

The first equality is straightforward, I can't imagine what you're missing. Let me know if you need it and I'll try to clarify.

Similarly $\cos'=-\sin$.

To prove that $\forall x\left((\sin(x))^2+(\cos(x))^2=1\right)$, consider the function $x\mapsto (\sin(x))^2+(\cos(x))^2$, differentiate, etc, etc.

"The classic definition of the trigonometric functions involves the unit circle. Suppose we define $\sin x, \; \cos x$ as above. Are these two definitions equivalent?"

Yes, if one accepts the geometric definitions as something meaningful, then they are equivalent.

"Similary if we define $\displaystyle \sin x = \frac{e^{ix}-e^{-ix}}{2i}, \; \cos x =\frac{e^{ix}+e^{-ix}}{2}$ are these three definitions equivalent?"

Note that $$e^{ix}+e^{-ix}=\sum \limits_{n=0}^\infty\left(\dfrac{i^nx^n}{n!}\right)+\sum \limits_{n=0}^\infty\left(\dfrac{(-1)^ni^nx^n}{n!}\right).$$

Now consider the cases $n\equiv k\pmod 4$ with $k\in \{0,1,2,3\}$.

It's not hard to see that some stuff will cancel (the imaginary parts, that is, when $n\equiv \pm 1\pmod 4 $) and others will add up (double up actually) and the definition of $\cos$ given at the top of the question appears.

Hence they are all equivalent.

Git Gud
  • 31,706
  • Oh.. I like your approach for the first question. You mean differentiate term by term the series , prove that the derivative is constant and then plug in $x=0$, right?

    As for the differentiation , yes.. i messed up a little bit.. by hurry.

    – Tolaso Jan 26 '15 at 21:37
  • @Tolaso if you prove that $\sin'x=\cos x$ and $\cos'x=-\sin x$ first, it's even easier to do the first. – Alice Ryhl Jan 26 '15 at 21:39
  • Good! I'll give it a go! – Tolaso Jan 26 '15 at 21:41
  • @Tolaso Yes, that's what I mean. – Git Gud Jan 26 '15 at 21:42
  • @Tolaso I answered the rest of your questions. They had escaped me at first. Edit: By the way, it's 'equivalent' not 'equivelant'. – Git Gud Jan 26 '15 at 21:53
  • Thank you very much... I had totally forgotten how "equivalent" was written.. – Tolaso Jan 26 '15 at 22:01
  • @Tolaso I'm afraid I was too fast in replying to your first comment. That wasn't what I meant, that's why I started by the second question. If you start with the second question, you don't need the series anymore, you just use the properties $\sin '=\cos$ and $\cos '=-\sin$. – Git Gud Jan 26 '15 at 22:22
  • You lost me.. ! So, basically I prove the derivative and then apply the chain rule? Right? – Tolaso Jan 26 '15 at 22:57
  • Yes, that's all. Only when plugging $x=0$ do you return to the series definition. – Git Gud Jan 26 '15 at 22:58