Let $\omega$ be primitive $n$-th root of unity. Can we determine all tuples of integers $(c_1, c_2,\ldots,c_n) $ such that $$c_1+c_2 \omega + c_3 \omega^2+\cdots+ c_n \omega^{n-1}=0 \,?$$

It is clear to me that if $ n$ is prime, then this means $ \omega$ is a root of polynomial $$c_1 + c_2x + c_3x^2+\cdots+c_nx^{n-1} =0 \,,$$ which implies $c_1 = c_2 = c_3\cdots = c_n $ as minimal polynomial of $\omega$ in this case is $${1+x+x^2+\dotsb+x^{n-1}}\,.$$ But if $ n$ is not prime and $\phi(n) $ divides $(n-1) $ then other solutions are also possible. Does this become highly dependent on $ n$? Or can we still say something for general $ n$?

  • 1
    $\begingroup$ Obviously, we're going to answer it. But what are your thoughts regarding the question. $\endgroup$
    – SarGe
    Jul 22, 2020 at 6:19
  • 3
    $\begingroup$ Hint: This isn't as much about roots of unity as it is about vectors in the plane. $\endgroup$
    – Arthur
    Jul 22, 2020 at 6:20
  • 1
    $\begingroup$ @Arthur is correct. Notice that $n^{th}$ roots of unity lie on vertices of $n$-sided regular polygon with one side along real axis. $\endgroup$
    – SarGe
    Jul 22, 2020 at 6:26
  • $\begingroup$ Answer depends on $n$. Easy case is when $n$ is prime. Can you do that case, Aditya? By the way, perhaps you want to specify that $\omega$ is a primitive $n$th root of unity? $\endgroup$ Jul 22, 2020 at 6:28
  • 1
    $\begingroup$ @Arthur Is it really? It’s basically asking for any linear relations within the $n$-th roots of unity and linear algebra alone doesn’t seem to get you far without knowing a thing or two about the roots first, does it? $\endgroup$
    – k.stm
    Jul 22, 2020 at 6:31

1 Answer 1


Many papers have been written on this question. I'd suggest having a look at

  1. Conway & Jones, Trigonometric diophantine equations, Acta. Arith. 30 (1976) 229-240,

  2. U Zannier, Vanishing sums of roots of unity, Rend. Sem. Mat. Univ. Pol. Torino 53 (1995) No. 4, 487-495,

  3. Lam & Leung, On vanishing sums of roots of unity, Journal of Algebra 224 No. 1 (2000) 91-109,

  4. Gary Sivek, On vanishing sums of distinct roots of unity, Integers 10 (2010) 365-368 #A31.

  • $\begingroup$ This should be written in comments. $\endgroup$
    – SarGe
    Jul 24, 2020 at 14:54
  • $\begingroup$ If you have a better answer, @SarGe I invite you to post it. $\endgroup$ Jul 25, 2020 at 3:16
  • $\begingroup$ Henry B Mann , ON LINEAR RELATIONS BETWEEN ROOTS OF UNITY ,1965 This paper seems also related to the question. $\endgroup$
    – Chao H
    Jul 26, 2020 at 17:07
  • $\begingroup$ @Alex, yes, it's probably in the references of the other papers so I figured anyone who followed the links I gave would come across the Mann paper. Do you have a link to it online? $\endgroup$ Jul 26, 2020 at 23:10
  • $\begingroup$ [link](dx.doi.org/10.1112/S0025579300005210 ) $\endgroup$
    – Chao H
    Jul 26, 2020 at 23:48

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .