4

I was attempting to evaluate the following infinite fraction:

$$\frac{-1}{\frac{-1}{\frac{-1}{\dots}}}$$

So I let $x=\frac{-1}{\frac{-1}{\frac{-1}{\dots}}}$, thus $x=\frac{-1}{x}$ and we arrive at $x^2=-1$, so $x=\pm i$.

Is this correct?

user7530
  • 49,280
frog1944
  • 2,357
  • 3
    You need to define precisely the meaning of the dots. It is usually defined as the limit of the sequence $x_{n+1}=\frac{-1}{x_n}$, provided this limit exists. What you proved is: IF the limit exists, then it is either $i$ or $-1$. – Taladris Oct 24 '17 at 07:42
  • So your infinite fraction is an interesting way of writing down the limit of the sequence $-1,1,-1,1,-1,\dots$ which doesn't exist – user340297 Oct 24 '17 at 07:48
  • 2
    No, because the limit of a real sequence, if existent, is unique and cannot be an element of $\mathbb{C}\setminus\mathbb{R}$. The map $x\to-\frac{1}{x}$ has no real fixed points. – Jack D'Aurizio Oct 24 '17 at 10:54
  • @Taladris You mean $-i$. – Simply Beautiful Art Oct 24 '17 at 21:29
  • To add onto @JackD'Aurizio even with an initial complex seed, it may not converge. – Simply Beautiful Art Oct 24 '17 at 21:30

1 Answers1

11

All of these and similar "tricks" for evaluating infinite expressions are really arguments of the form,

"If the expression converges to some number $x$, then the following must be true about $x$..."

So you have indeed proven that if your expression converges, it does so to either $i$ or $-i$, and not to some other value. But you haven't proven that the expression converges in the first place (and it is easy to see that it does not).

user7530
  • 49,280