All the proofs I've seen are geometrical, assuming that $A+B$ is less than $90$ degrees. How can you prove this identity for $A+B$ greater than $90$ degrees, or more generally, any arbitrary value?
-
I'm sure you can just modify the usual arguments, even if the sum is greater than 90 - you'll just have to be careful with the signed lengths. – Akiva Weinberger Aug 27 '14 at 07:05
-
1Easiest way, though, would probably be to mess with coordinates and the unit circle. – Akiva Weinberger Aug 27 '14 at 13:26
-
See this recent answer. – JimmyK4542 Aug 27 '14 at 20:28
2 Answers
If you don't want a geometrical proof, then you need to indicate how you are defining $\cos$ and $\sin$. One way to define them is that $\cos(x)$ and $\sin(x)$ are the real and imaginary parts of $\exp(ix)$, and use the property that $$\exp(i(a+b)) = \exp(ia)\exp(ib)$$ Then $$\begin{align} \exp(i(a+b)) &= \exp(ia)\exp(ib)\\ &= (\cos(a) + i\sin(a))(\cos(b) + i\sin(b))\\ &= \cos(a)\cos(b) - \sin(a)\sin(b) + i(\sin(a)\cos(b) + \cos(a)\sin(b)) \\ \end{align}$$ And from this we obtain two trig identities at once: $$\cos(a+b) = \text{Re}(\exp(i(a+b))) = \cos(a)\cos(b) - \sin(a)\sin(b)$$ $$\sin(a+b) = \text{Im}(\exp(i(a+b))) = \sin(a)\cos(b) + \cos(a)\sin(b)$$
-
But that requires knowing the Taylor series of sine, which requires knowing the derivative of sine... which involves knowing the summation formula. Besides, I think that this is much more heavy machinery than needed. I'm guessing that the OP wants a proof that is geometrical, but isn't restricted to $0\le\theta<90^\circ$. – Akiva Weinberger Aug 27 '14 at 06:59
-
May be you could remove the reference to the power series that, in any manner, you don't use in your good answer. – Claude Leibovici Aug 27 '14 at 07:42
-
@ClaudeLeibovici But how else do you prove that sine and cosine are the real and imaginary parts of $\exp(ix)$? I know know way that doesn't involve differentiating the trig functions. – Akiva Weinberger Aug 27 '14 at 13:27
-
@ClaudeLeibovici I just realized - you don't need to prove that $\exp(ix)$ is anything. You only need to prove De Moivre - that $\arg(wz)=\arg(w)+\arg(z)$. This can be done more easily, by noting that rotation is linear. – Akiva Weinberger Aug 27 '14 at 14:18
-
-
Hint:
Suppose that the two unit circles in the diagram are centered at the origin, with the second one rotated clockwise by $\measuredangle A$. Note that the two red line have equal length. Find the coordinates of $P,Q,R,\text{ and } S$ (but especially $Q$). Use this information to derive $\sin(A+B)$ and $\cos(A+B)$.

- 7,790
-
-
-
point $P$ has coordinates $(\cos A, \sin A)$. What are $Q,R,$ and $S$'s coordinates? – John Joy Aug 29 '14 at 12:33
-
1