# Algebraically why is that $\cos(0) =1$? [closed]

I have been thinking about this for a day now but it seems every way is predicated on using either $\cos(0)=1$ and $\sin(0)=0$ as a giving to justify the argument. Is it possible to show this to be true without prior knowledge of sine and cosine at zero?

For example, if we use the identity $\cos^2(0) + \sin^2(0) = 1$, we have to assume knowledge of either $\cos(0)$ or $\sin(0)$, maybe I missed something. Then I thought about using the Taylor series at $x=0$ but not assuming it to be giving. By doing so, I end up with \begin{align} \cos(x) &= \cos(0) - x\sin(0) - \frac{x^2}{2!}\cos(0) + \frac{x^3}{3!}\sin(0) + \cdots\\ &= \cos(0)\sum_{n=0}^{\infty}(-1)^n\frac{x^{2n}}{(2n)!} - \sin(0)\sum_{n=0}^{\infty}(-1)^n\frac{x^{2n+1}}{(2n+1)!}\tag{1}\\ \cos(0) &= \cos(0)\cdot 1 - \sin(0)\cdot 0\\ \cos(0) &= \cos(0) \end{align} Therefore, I haven't reached $\cos(0) = 1$ unless I use it as a giving. Then I thought about the exponential $e^{ix} = \cos(x) + i\sin(x)$. Now this would work $$e^{i\cdot 0} = \cos(0) + i\sin(0) = 1 + 0\cdot i$$ so the imaginary part has to be zero, but if we derive the power series and not take Euler's formula for granted, we get \begin{align} e^{ix} &= \cos(0)\biggl[\sum_{n=0}^{\infty}(-1)^n\frac{x^{2n}}{(2n)!}+i\sum_{n=0}^{\infty}(-1)^n\frac{x^{2n+1}}{(2n+1)!}\biggr]\\ &+ \sin(0)\biggl[\sum_{n=0}^{\infty}(-1)^n\frac{x^{2n}}{(2n)!}-\sum_{n=0}^{\infty}(-1)^n\frac{x^{2n+1}}{(2n+1)!}\biggr]\\ e^0 &= \cos(0)[1+0\cdot i]+\sin(0)[1-0]\\ 1 &= \cos(0) + \sin(0) \end{align} Now, I need to use $\sin(0)=0$ as a giving but the point is to not take these identities for granted.

For why I accept the derivative of $\cos$ to be $\sin$, the power series would be $$\sin(x) = \sin(0)\sum_{n=0}^{\infty}(-1)^n\frac{x^{2n}}{(2n)!}+\cos(0)\sum_{n=0}^{\infty}(-1)^n\frac{x^{2n+1}}{(2n+1)!}\tag{2}$$ Then derivative of $(2)$ is $(1)$ after a change of variables.

## closed as unclear what you're asking by anomaly, Hans Lundmark, RE60K, Jonas Meyer, MicahMar 6 '15 at 21:08

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

• It's a definition. You can define cosine as the solution to $y''=-y$ satisfying $y(0)=1$. – symplectomorphic Mar 6 '15 at 18:03
• How do you defined "algebraically" the cosine and the sine ? – user171326 Mar 6 '15 at 18:04
• I think it would help if you told us what axioms you would feel comfortable using. You have to start with a definition of cosine somehow, otherwise you can't prove anything about it because it doesn't mean anything. – zeldredge Mar 6 '15 at 18:05
• @dustin Then whats your definition of cosine? – Marm Mar 6 '15 at 18:06
• That Taylor series is not the definition. It's just the expanion of an arbitary (presumably analytic) function $f$ at $0$, with the coefficient $f^{(n)}(0)$ replaced by $\pm \cos(0)$ and $\pm \sin(0)$. In other words, you're taking $\cos x$ to be the function with $y'(x) = -\sin x$ and $y''(x) = -\cos x$...except that that definition is obviously circular. If you want to take $\cos x$ to be the function with $y''(x) = -y(x)$, then it's still not going to be uniquely determined unless you set boundary conditions, such as specifying $y(0) = \cos 0$ and $y'(0) = \sin 0$. – anomaly Mar 6 '15 at 18:21

If you define $\cos$ by $\cos x = \sum_{n=0}^{\infty}(-1)^n\frac{x^{2n}}{(2n)!}$, then $\cos 0 = 0^0 = 1$ (all other terms are $0$). If you define $0^0$ differently, you are not actally talking about the $\cos$ function. If you define $\cos x = \Re(e^{ix})$, then you get the same.

If you want to define $\cos$ via the differential equation:

$$\cos''(x) = -\cos(x)$$ then you need an initial value. Again, if you don't take $\cos(0)=1$, then you are not talking about the $\cos$ function. The choice is $\cos(0)=1$ is canonical, the only other canonical choice is mapping $0$ to $0$, but this is what $\sin$ already does.

Also, if $e^{ix} = \cos(x) + i\sin(x)$, then $1 + i0 = 1 = e^0 = e^{i0} = \cos(0) + i\sin(0)$, so by taking the real value: $\cos(0) = 1$.

• Start with a function called $\cos(x)$. We don't know its definition so we say hey let's determine the taylor series. We get $f(0) = \cos(0)$, $f'(0) = -\sin(0)$ and so on. The power series then has cosine and sine involved. Unless we know something about the values at zero, the Taylor series doesn't lose the values. – dustin Mar 6 '15 at 18:13
• @dustin How? In what way do you get the taylor series? Derivatives? Then where did the definition of the derivative of cosine come from? proofwiki.org/wiki/Derivative_of_Sine_Function Each proof somehow relies on the value of at least one of cos(0) or sin(0), knowledge of the other's derivative, or from a richer definition of cosine than you are using. – JMoravitz Mar 6 '15 at 18:17
• @JMoravitz see edit. No knowledge is assumed to determine they are derivatives. – dustin Mar 6 '15 at 18:23
• I also stated not to take Euler's formula for granted. – dustin Mar 6 '15 at 18:29
• @dustin You are mistaken. If you don't "know" $\cos(x)$, then how are you going to reason about it, derive it's taylor series for instance? Let me define the term "bleeeerg". No, I won't tell you what "bleeeerg" means, but as you can "obviously" see, "bleeeerg" is green, wears pants and likes eating candy. Doesn't make sense? Of course not! Now replace "bleeeerg" by $\cos(x)$ and "is green, wears pants and likes eating candy" by "has taylor expansion ..." and we are back to our initial situation. – Stefan Perko Mar 6 '15 at 20:16

For me, cosine was defined analytically using the infinite sum $$\cos(x):=\sum_{n=0}^\infty \frac{(-1)^nx^{2n}}{(2n)!}=1-\frac{x^2}{2!}+\frac{x^4}{4!}-\ldots$$

from which you can see that $\cos(0)=1$

Which definition do you want to use?

• Derive that definition without assuming $\cos(0)=1$ and you have a bunch of $\cos(0)$ hanging around. – dustin Mar 6 '15 at 18:08
• It's a definition, and therefore cannot be derived. We can define $\cos$ in this way, in which case it is apparent that $\cos(0) =1$, or we can define $\cos$ in terms of triangles, in which case it is also apparent that $\cos(0) = 1$. – gj255 Mar 6 '15 at 18:08
• What do you mean 'derive that definition'? This is my definition. – Tyr Curtis Mar 6 '15 at 18:08
• This agrees with the analysis definition of the cosine used here: proofwiki.org/wiki/Definition:Cosine – JMoravitz Mar 6 '15 at 18:19

you know that $\lim_{x\rightarrow \infty}\frac{sinx}{x}=1$ geometrically. from here you know that $\lim_{x\rightarrow \infty}\sin x=0$. if you assume that the sine function is continuous, then you have that $sin(0)=0$, and from here it follows from $\sin^2(0)+\cos^2(0)=1$ that $cos(0)=\pm 1$. if you also assume that the cosine function is continuous, then $cos(0)$ couldn't be negative, because $\cos x\geq 0$ for $0\leq x\leq \frac{\pi}{2}$, which is defined geometrically.

so if $\cos(0)$ can't be negative, we've established that $\cos(0)=1$

• cos0=0????????? – RE60K Mar 6 '15 at 19:39
• here, happy? ;) – tzoorp Mar 6 '15 at 20:15
• You have write @name otherwise adg won't know you said anything. – dustin Mar 6 '15 at 21:14

If we define cos and sin on the unit circle as x and y coordinates on the periphery.. then at length 0 along the circle ("angle" 0) we will be at coordinate (1,0) so cos will be 1 and sin will be 0.

• I want to show this algebraically not relying on geometry. – dustin Mar 6 '15 at 18:02
• Ok so you are fine with trigonometrical one i.e. sum of squares is 1. Then the taylor expansion of sin(0) will be 0 so cos(0) must be either -1 or 1 according to that sum of squares = 1. – mathreadler Mar 6 '15 at 18:05
• this is in no way wrong! – RE60K Mar 7 '15 at 6:46

Anything you do would be useless, since all this is defined once and everything is based upon it. Even Euler's Formula and Taylor series use the definition that $\cos0=1,\sin0=0$. The following could be taken as a starting point while neglecting any geometrical consequence: $$y''=-y$$ And the solution will be of the form of: $$a\sin x+b\cos x$$ (where a,b are variable constants)

Suppose someone unknown about trigonometric functions wanted to define them, he would take two starting functions: $$\sin_2x=c\sin x+d\cos x\\\cos_2x=e\sin x+f\cos x$$ (where c,d,e and f are constants)

And then to him the solution of above equation would be: $$a_2\sin_2x+b_2\cos_2x$$

Now if did took $(c,d,e,f)\equiv(1,2,3,4)$, then $\sin_20$ would be $2$ and $\cos_20=4$.

Now the special solution we use is $(1,0,0,1)$ , because it follows a nice property: $$\sin^2x+\cos^2x=1$$ but his does: $$(\cos_2x-2\sin_2x)^2+\left(\frac{3\sin_2x-\cos_2x}2\right)^2=1$$ There are also other unmentioned advanted, one involving range too, since $|\sin x|,|\cos x|<1$ but $|\sin_2x|<\sqrt{5},|\cos_2x|<5$. There is surely an advantage with that solution if considering without any geometrical basis. When geometry comes into play, there's no choice left...

• Prior to having the knowledge we have today, mathematicians had to find a way to prove it to be true without saying I know this to be true. How was it done in this case, because unless I know something useful about $\cos(0)$ or $\sin(0)$, making the conclusion $\cos(0)=1$ doesn't seem to come so naturally. Once $\cos(0)=1$ is known the other will follow. – dustin Mar 6 '15 at 18:52
• @dustin that follows from geometrical ease. Also you could have $\cos_2x=\sin x$ an $\sin_2x=\cos x$then $\cos_2(0)=0$ and all formulas are interchanged, it's just an easy definition which when defined like this become the functions related to triangles. – RE60K Mar 6 '15 at 19:05
• I found a solution from ancient Greece. – dustin Mar 6 '15 at 19:06