## Beware of square roots and exponentials!

Posted 19th November 2010 by Holger

In the last post I asked you the question, what is wrong with the following proof.

As some of you noticed, the answer lies in the fact that the identity

is only true for positive real numbers x and y. In general one has to be careful with the identity

(*x**y*)^{a} = *x*^{a}*y*^{a}

for non-integer a and non-positive x and y.

I posted this question because, although it ought to be, it is not always taught to students when complex numbers are introduced. Once students are taught complex numbers they feel that, now that they are free to take the square root of any real number, they don’t have to pay attention anymore to what they are doing. They apply the rules blindly, resting in false comfort that everything is now fine.

Another pitfall with complex numbers, that even educated people are not always aware of, is the fact that

if y and z are complex numbers, **even for positive, real x**. If it were we, could prove the following (here I assume that you are familiar with the complex exponential function).

Let’s start with the following identity

This makes

(*e*^{1 + 2iπn})^{2iπn} = (*e*)^{2iπn} = 1

But, if the above identity were true, the last line would be equal to

Which, in general, is a positive real number not equal to 1.

That 1 = -1 proof is one of my favorites.

I also like

0 = (1 + -1) + (1 + -1) + (1 + -1) + …

0 = 1 + (-1 + 1) + (-1 + 1) + (-1 + 1) + …

0 = 1 + 0

0 = 1

The problem of course being associativity doesn’t hold in the infinite case.

Good one… I didn’t realize that about the complex numbers… I will keep that in mind.