1

You can't use geometric sums, minimal polynomials, pentagon, and exact values with radicals.

All the five, fifth-roots of unity are :$1,\left(\cos \left(\frac{2 \pi}{5}\right)+i \sin \left( \frac{2\pi}{5}\right)\right),\left(\cos \left(\frac{4 \pi}{5}\right)+i \sin \left(\frac{4 \pi}{5}\right)\right),\left(\cos \left(\frac{6 \pi}{5}\right)+i \sin \left(\frac{6 \pi}{5}\right)\right), \left(\cos \left(\frac{8 \pi}{3}\right)+i \sin \left(\frac{8 \pi}{5}\right)\right)$

UmbQbify
  • 889
  • 3
    Did you try using trigonometric identities? For instance, $\cos(4\pi/5) = \cos(2\pi/5 + 2\pi/5) = ...$ – Stockfish Jul 21 '20 at 13:44
  • 2
    There ought to be an algebraic solution which mirrors why the roots cancel geometrically; in other words, the complex solutions have imaginary parts which cancel in two pairs leaving a real total of $1$ then the remaining root of $-1$ sums to zero. Don't fight the geometry, let it guide the algebra. – James S. Cook Jul 21 '20 at 14:03

5 Answers5

3

If you want a simple trigonometric proof, recall $$\sin a-\sin b=2\sin\frac{a-b}2\cos\frac{a+b}2.$$ Therefore \begin{align} 0&=\sin(a+2\pi)-\sin a\\ &=(\sin(a+2\pi)-\sin(a+8/\pi/5))+(\sin(a+8/\pi/5)-\sin(a+6/\pi/5))\\ &+(\sin(a+6/\pi/5)-\sin(a+4/\pi/5))+(\sin(a+4/\pi/5)-\sin(a+2/\pi/5))\\ &+(\sin(a+2/\pi/5)-\sin a)\\ &=2(\sin\pi/5)\left[\cos(a+9\pi/5)+\cos(a+7\pi/5)+\cos(a+\pi)+\cos(a+3\pi/5)+\cos(a+\pi/5)\right] \end{align} and $$\cos(a+9\pi/5)+\cos(a+7\pi/5)+\cos(a+\pi)+\cos(a+3\pi/5)+\cos(a+\pi/5)=0.$$ Taking $a=-\pi/5$ gives $$\cos(8\pi/5)+\cos(6\pi/5)+\cos(4\pi/5)+\cos(2\pi/5)+1=0.$$ and taking $a=-7\pi/10$ gives $$\sin(8\pi/5)+\sin(6\pi/5)+\sin(4\pi/5)+\sin(2\pi/5)+0=0.$$ These are the real and imaginary parts of the sum of the fifth roots of zero.

Of course, this method works for other values of $5$.

Angina Seng
  • 158,341
3

Let $S$ be the set of roots of $z^5-1$, then $$e^{\frac{2\pi i}{5}}S=\{e^{\frac{2\pi i}{5}}, e^{\frac{4\pi i}{5}},e^{\frac{6\pi i}{5}},e^{\frac{8\pi i}{5}},e^{\frac{10\pi i}{5}}\}=\{e^{\frac{2\pi i}{5}},e^{\frac{4\pi i}{5}},e^{\frac{6\pi i}{5}},e^{\frac{8\pi i}{5}},1\}=S$$

The sets are equal, so $$e^{\frac{2\pi i}{5}}\sum_{z\in S}z=\sum_{z\in S}z\\\Rightarrow \sum_{z\in S}z=0$$


More generally

Let $G$ be a multiplicative group of order $n$. Then $$\sum_{z\in G}z=0$$

Proof: Let $g\in G$ such that $g\neq e$. Let $\phi_g(x)=gx$, then $\phi_g\in\text{Aut}(G)$ since:

  • injective: If $\phi_g(x)=\phi_g(y)$ then $gx=gy\Rightarrow x=y$.
  • surjectivity: If $a\in G$, then $\phi_g(g^{-1}a)=a$.

This means that the sets $gG,G$ are equal. And so,

$$\sum_{z\in G}gz=\sum_{z\in G}g\iff (g-e)\sum_{z\in G}z=0\\\Rightarrow \sum_{z\in G}z=0$$

cansomeonehelpmeout
  • 12,782
  • 3
  • 22
  • 49
2

You don't need "geometric sum" to prove $$ \sum_{j=0}^{n-1} \cos(\theta+j\phi)=\frac{\sum_{j=0}^{n-1}\cos(\theta+j\phi)\sin(\phi/2)}{\sin(\phi/2)}=\frac{\sum_{j=0}^{n-1}\left[\sin(\theta+(j+\tfrac12)\phi)-\sin(\theta+(j-\tfrac12)\phi)\right]}{\sin(\phi/2)}=\dots $$ and similarly $\sum_{j=0}^{n-1}\sin(\theta+j\phi)$.

user10354138
  • 33,239
1

Good question. Shall we try?

So we have the five roots: $$1,\left(\cos \left(\frac{2 \pi}{5}\right)+i \sin \left( \frac{2\pi}{5}\right)\right),\left(\cos \left(\frac{4 \pi}{5}\right)+i \sin \left(\frac{4 \pi}{5}\right)\right),\left(\cos \left(\frac{6 \pi}{5}\right)+i \sin \left(\frac{6 \pi}{5}\right)\right), \left(\cos \left(\frac{8 \pi}{3}\right)+i \sin \left(\frac{8 \pi}{5}\right)\right)$$

I am just going to write $\frac{2\pi}{5}$ as $\alpha$. It's going to save time and make this easier to read. In this notation, the roots are : $1, \ \cos (\alpha) + i\cdot\sin(\alpha), \ \cos (2\alpha) + i\cdot\sin(2\alpha), \ \cos (3\alpha) + i\cdot\sin(3\alpha)$ and $\cos (4\alpha) + i\cdot\sin(4\alpha)$

OK let's add them all up, we'll start with the imaginary part because that is going to be easy. We have $$i \times \left[\sin \alpha+\sin 2\alpha+\sin 3\alpha+\sin 4\alpha\right]$$ Well we know that $5\alpha = 2\pi$ and so $\sin \alpha = \sin(2\pi-\alpha) = \sin(-4 \alpha) =-\sin (4\alpha$). Also $\sin(2 \alpha)=-\sin(3 \alpha)$ for similar reasons and so all of these terms cancel to zero.

In the real part, we have $$1+\cos(\alpha)+\cos(2\alpha)+\cos(3\alpha)+\cos(4\alpha)$$ $\cos(\alpha)=\cos(2\pi-\alpha)=\cos(4\alpha)$ Also $\cos(2 \alpha)=\cos(3 \alpha)$ for similar reasons. So we just need to show that $$1+2\cos(\alpha)+2\cos(2\alpha)=0$$

How hard can this be? We know, from the trig identities that $\cos(2\alpha)=2 \cos^{2} (\alpha)-1$. So now we need $1+2 \cos(\alpha)+4 \cos^{2}(\alpha)-2=0$ or $$4 \cos^{2}(\alpha)+2 \cos(\alpha)-1=0$$

Trigonometry tells us that $\cos \alpha = \frac{\sqrt 5-1}{4}$. If we pop this in then we get.

$$\frac{6-2 \sqrt 5}{4}+\frac{\sqrt 5-1}{2}-1=0$$

Hooray and Phew.

UmbQbify
  • 889
  • But that's my issue. You had to substitute an exact value which I don't think is common knowledge especially to high school students. Or is it supposed to be? – Allan Ray Jul 21 '20 at 15:01
  • OK that's a good point. It's interesting isn't it? because we only needed to use an exact value for the real part. The imaginary part just cancelled to zero because there was so much symmetry operating there. If we had taken sixth roots (or any even number root) then there would be symmetry in the real part as well. In fact we don't really need an exact value. The proof that $\cos(\alpha)=(\sqrt 5 - 1)/4$ (or even ($4 \cos^{2} \alpha + 2 \cos \alpha -1 =0)$ can be crunched out using trig formulas and polynomial factorisation. – Simon Terrington Jul 21 '20 at 15:36
0

$$z^5=1$$ $$z^5-1=0$$ $$(z-1)(z^4+z^3+z^2+z+1)=0$$ Note that the parenthesis with 5 terms are the roots but as we know the complex is a filed no zero divisors thus the sum of the roots is zero! Plug in $e^{\frac{2\pi i }{5}}=z$ $$(e^{\frac{2\pi i }{5}}-1)(e^{\frac{8\pi i }{5}}+ e^{\frac{6\pi i }{5}}+ e^{\frac{4\pi i }{5}}+ e^{\frac{2\pi i }{5}}+1)=0$$

IrbidMath
  • 3,187
  • 12
  • 27