13

Possible Duplicate:
Why are the solutions of polynomial equations so unconstrained over the quaternions?

Coudl someone explain me the following: Why should $x^2+1=0$ have uncountable infinite many solutions $x\in\mathbb H$?

In my opinion it has only 4 solutions, namely $i^2=j^2=k^2=ijk=-1$ ?

Community
  • 1
Voyage
  • 919
  • 1
    If $a^2+b^2+c^2=1$ then $ai+bj+ck$ is solution of this equation. – Hanul Jeon Jan 28 '13 at 14:18
  • 8
    Math is not really about opinions. If you think there are only four solutions, try and prove that every solution is one of those four. If you get stuck it might be a good start on a counterexample, or a proof that there are uncountably many solutions after all. If you succeed, then it's not an opinion anymore (beware of false proofs, though). – Asaf Karagila Jan 28 '13 at 14:19

2 Answers2

13

Let $a, b, c$ be real numbers satisfying $a^2 + b^2 + c^2 = 1$ and let $x = ai + bj + ck$. Then $$ x^2 = (ai + bj + ck)(ai + bj + ck) = -(a^2 + b^2 + c^2) = -1. $$

Tunococ
  • 10,303
8

Let $0 \leq p \leq 1$, and consider numbers of the form $$x = \sqrt p i + \sqrt{1-p} j$$ Taking the square, we have $$x^2 = (\sqrt p i + \sqrt{1-p} j)^2 = pi^2 + (1-p)j^2 + \sqrt{p(1-p)}(ij + ji)$$ The third term is zero, since $ij = -ji$. The first two terms sum to -1, so any number of the given form is a solution to $x^2 + 1 = 0$. Since there are uncountably many $p \in [0,1]$, there are uncountably many solutions.

Jonathan Christensen
  • 3,870
  • 15
  • 21
  • Thanks, this seems clear to me. Was it intuition considering numbers of the form $x = \sqrt p i + \sqrt{1-p} j$$ – Voyage Jan 28 '13 at 14:18
  • @Voyage I first tried the form $pi + (1-p)j$, and when I saw where that went wrong it was clear that taking square roots would make it work. – Jonathan Christensen Jan 28 '13 at 14:19