I did this calculation manually the first time through—some years ago. It goes a lot easier with a computer algebra system. Today, I'm going to use Maxima to solve the problem.
An identity which comes up in the development of the complex numbers, gives a convenient way of expressing sinnθ in terms of combinations of sinθ and cosθ.
De Moivre's formula:
rn(cosθ+isinθ)n=rn(cosnθ+isinnθ).
So, if we take r=1 and n=5, we can do a binomial expansion and easily isolate the sinθ and cosθ. In a Maxima cell, I type
eq: (cos(%theta) + %i*sin(%theta))^5 = cos(5*%theta) + %i*sin(5*%theta);
realpart(eq);
imagpart(eq);
After executing, I obtain the fifth angle identities (if you want to call them that).
cos(5θ)=5cos(θ)sin(θ)4−10cos(θ)3sin(θ)2+cos(θ)5
sin(5θ)=sin(θ)5−10cos(θ)2sin(θ)3+5cos(θ)4sin(θ)
To eliminate cos2θ from the sine identity we use
subst([cos(%theta)=sqrt(1-sin(%theta)^2)],%);
expand(%);
and obtain
sin(5θ)=16sin(θ)5−20sin(θ)3+5sin(θ).
Taking θ=18°, using x to stand in for sin18°, and rearranging, we have
0=16x5−20x3+5x−1.
Factoring this one manually was made easier by recognizing x=1 as a root. After that I tried forming a square free polynomial and found that there was in fact a squared polynomial involved. In general, this helps in a couple ways (potentially):
- Reduces the degree of the polynomial to be factored.
- Sometimes produces fully reduced factors as a by-product of forming the square free polynomial.
- This happens when each repeated factor is of different degree; that is, occurs a different number of times.
(Square free means no factors left that are raised to some power—each factor occurs exactly once.) Doing this manually involves taking the greatest common divisor between a polynomial and it's derivative. But today, let's let Maxima do it.
factor(0=16*x^5-20*x^3+5*x - 1);
0=(x−1)(4x2+2x−1)2.
At this point, you really do have to be exceedingly lazy to not continue with the rest of the work manually. It's just a quadratic formula to be applied now (we discard x=1 as an artifact). To use a computer algebra system for the rest is...well...okay, it is Saturday [at original publication], so...
solve(0=4*x^2+2*x-1,[x]);
[x=−√5+14,x=√5−14]
We reject the negative solution (which is sin−54° and, interestingly, is the negative of half the Fibonacci number) on the grounds that the range of the sine function on our interval is positive. Hence,
sin18°=√5−14.