Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $x,y \in \mathbb{R}$ such that the set $\{\cos{(n\pi x)} + \cos{(n\pi y)} | n \in \mathbb{N} \}$ is finite. Show that $x$ and $y$ are rational.

I have been trying to consider a graph of this set and maybe an argument by contradiction....

share|cite|improve this question

Lemma. If $\alpha\in\mathbb R$, $m\in\mathbb N$, there exist $a,b\in\mathbb Z$ with $1\le a\le m$ and $|a\alpha-b|<\frac 1m$.

Proof: Partition the interval $[0,1)$ into $m$ subintervals $\left[\frac km,\frac{k+1}m\right)$, $0\le k<m$. By the pigeon-hole principle, there exist $0\le i<j\le m$ such that $i\alpha-\lfloor i\alpha\rfloor$ and $j\alpha-\lfloor j\alpha\rfloor$ are in the same subinterval, hence with $a=j-i$ and $b=\lfloor j\alpha\rfloor-\lfloor i\alpha\rfloor$, the result follows.$_\square$

Corollary. If $\alpha,\beta\in\mathbb R$, $\epsilon>0$, there exist $r,s,n \in \mathbb Z$ such that $n\ge 1$, $|n\alpha-r|<\epsilon$ and $|n\beta-s|<\epsilon$.

Proof: Select $N>\frac1\epsilon$. By the lemma, there exist integers $a,b$ such that $1\le a\le N^2$ and $|a\cdot \alpha-b|<\frac 1{N^2}$. Again by the lemma, there exist integers $c,d$ such $1\le d\le N$ and $$|d\cdot a\beta-c|<\frac1N<\epsilon.$$ But we also have $$|da\alpha-db|<\frac d{N^2}\le \frac1N<\epsilon.$$ Thus the result follows with $n=da$, $r=db$, $s=c$. $_\square$

Now for the main event:

By continuity of $\cos$, for any $\epsilon>0$ there exists $\delta>0$ such that $\cos t>1-\frac\epsilon2$ for all $t$ with $|t|<\delta$. We find $r,s,n$ as in the corollary such that $$\left\vert n\cdot\frac x{2}-r\right\vert< \frac{\delta}{2\pi}\qquad\text{and}\qquad\left\vert n\cdot\frac y{2}-s\right\vert<\frac{\delta}{2\pi}.$$ It follows that $n\pi x$ and $n\pi y$ differ from $2\pi r$ and $2\pi s$ by less than $\delta$, hence $$\cos (n\pi x)+ \cos(n\pi y)>2-\epsilon.$$

Now assume that $S:=\{\cos(n\pi x)+\cos(n\pi y)\mid n\in\mathbb N\}$ is finite. Then $\max S<2$ would lead to a contradiction if we let $\epsilon = 2-\max S$. Therefore $\max S=2$, i.e. there exists $n\in\mathbb N$ with $$\cos (n\pi x)+ \cos(n\pi y)=2,$$ hence $\cos(n\pi x)=\cos(n\pi y)=1$ and $nx,ny$ are even integers, hence $x,y\in\mathbb Q$. $_\square$

share|cite|improve this answer

If $x$ is irrational, the values $e^{in\pi x}$ are dense on the unit circle. Similarly for $y$.

share|cite|improve this answer
I am sorry, but I don't see how this answers the question. When considering this, we have to consider sine as well, yes? – Math2012pc Nov 26 '12 at 0:58
If those complex numbers are dense on the unit circle, then their real parts, which are the cosines, are dense in the interval $[-1,1]$. No need to consider the sines. – Gerry Myerson Nov 26 '12 at 4:35
Ah, I see. But in order to show that the complex numbers are indeed dense on $S^1$, do we need to use a topological argument or can we argue with number theory? – Math2012pc Nov 26 '12 at 5:08
The sum of two dense sequences need not be dense! – Hagen von Eitzen Nov 26 '12 at 7:25

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.