Algebra in the complex domain

using x = π in Euler’s formula:

e[sup]xi[/sup] = cos(x) + isin(x)

e[sup]πi[/sup] = -1

square both sides:

e[sup]2πi[/sup] = 1

putting both sides to the power of ln(5)/2πi
1[sup]ln(5)/2πi[/sup] = (e[sup]2πi[/sup])[sup]ln(5)/2πi[/sup] = e[sup]ln(5)[/sup] = 5

Now this looks incorrect as I have a power of 1 that is not equal to 1 (though it is a complex power).

Now I’ve been told that (a[sup]b[/sup])[sup]c[/sup] = a[sup]bc[/sup] doesn’t hold in the complex domain , but the other well-known identity:

i[sup]i[/sup] = e[sup]-π/2[/sup]

can seemingly be derived from the x = π case of Euler’s formula using this rule.

Of course using simlair reasoning you get this silly result:

e[sup]2πi[/sup] = 1 =>

2πi = ln(1) = 0 =>

-4π[sup]2[/sup] = 0

but all this suggests to me is that ln(x) is multi-valued.

Anyway basically I want to know whether:

1[sup]ln(5)/2πi[/sup] = 5

is a correct or incorrect equality.

“Now this looks incorrect as I have a power of 1 that is not equal to 1 (though it is a complex power).”

Why should that be surprising, (for instance) 1 has three cube roots, only one of them (1) is wholly real.

(PS: I didn’t check your calculations, just wanted to observe that I saw no surprises).

Yes, I suppose, but it’s a real answer, but then again 1 has two square-roots one of which isn’t one and is real.

I suppose I should of rephrased that into it looks incoorect as it has an answer that isn’t 1, -1, i or -i.

If x is complex, as in this case, then yes, ln(x) is a multi-valued function.

If one has e[sup]lnz[/sup]=z, then because of the periodicity of the function e[sup]lnz + 2n(pi)i[/sup], the logarithm function in the complex plane is multi-valued, which explains your strange results - ln is not uniquely defined.

1[sup]1/3[/sup] has three cubed root, namely,
1
-cos60 + isin60
and
-cos60 - isin60

Only one of which (1) is in the set {1, -1, i, -i}, I’m missing your point, obviously.

It is, believe it or not, a correct equality.

The reason is that exponentiation in the complex plane is defined as follows: a[sup]b[/sup] = exp(b ln(a)). But as you’ve noted, ln() is a multi-valued function when applied to complex numbers. Therefore, exponentiation is also multi-valued in the complex domain.

Now consider 1[sup]ln(5)/2πi[/sup]. By definition, this must equal exp( (ln(5)/2πi) ln(1)). If we let ln(1)=0, then we get exp(0) which equals 1, which we expect. But since we’re considering complex numbers, ln(1) also equals 0+2πi, in which case 1[sup]ln(5)/2πi[/sup] = exp(ln(5)) = 5. In the same fashion, we can show that 1[sup]ln(5)/2πi[/sup] = exp(2 ln(5))=25, and so on.

If we restrict ourselves to the “primary branch” of the logarithm function…that is, if we arbitrarily require that the imaginary part of ln(z) always lies between -πi and πi…then some of these problems go away. But that in turn causes other problems, like breaking the rule that ln(ab) = ln(a)+ln(b). In short, complex numbers are weird.

Thanks Orbifold while your here can you solve another seemingly paradoxiacal result in a simlair vein:

e[sup]-π/2[/sup] = i[sup]i[/sup] =>

e[sup]-πi/2[/sup] = i[sup]-1[/sup] =>

e[sup]πi/4[/sup] = i[sup]1/2[/sup] =

(e[sup]πi[/sup])[sup]1/4[/sup] = i[sup]1/2[/sup] =

1[sup]1/4[/sup] = i[sup]1/2[/sup]

= 1, -1, i, -i

You have a small error in your second-to-last step. exp(pi i) = -1, not 1.

Oops it should of been:

e[sup]-π/2[/sup] = i[sup]i[/sup] =>

e[sup]-πi/2[/sup] = i[sup]-1[/sup] =>

e[sup]πi/4[/sup] = i[sup]1/2[/sup] =

(e[sup]2πi[/sup])[sup]1/8[/sup] = i[sup]1/2[/sup] =

1[sup]1/8[/sup] = i[sup]1/2[/sup]

= 1, -1, i, -i

but I see what’s wrong with that anyway now, and all I’ve really proved is that the square root of i is equal to the sqaure root of i.

So th question that I really wanted to ask is: what is the square root of i?

By DeMoivre’s Theorem they are
cos45 + isin45
and
-cos45 - isin45

i has two square roots, (1+i)/sqrt(2) and (-1-i)/sqrt(2). Those are also the two possible numbers you get when evaluating i[sup]1/2[/sup] via the formula exp(ln(i)/2).

If the square root of 1 = 1 and -1,
maybe the square root of i = i and -i?

WAIT…

i^2=-1
-i^2= (-1)(i)(-1)(i)=(-1)^2(i)^2=1*-1=-1

HUMPH!

Generally if you express your complex number in polar terms (so r<t is complex number of modulus r and angle t)

then the nth roots of r<t are

(r[sup]1/n[/sup]) <(t/n + 360kn) where k = 1, 2,… …n-1

Which is pretty damned gorgeous (and one of the few things I can still remember about complex numbers).

Psst, The Great Unwashed. That’s supposed to be 360k/n, right?

Might be. It could be a new theory of mine (all mine), and it works perfectly well if you don’t care for all the roots. So nyah!

Actually if one of the mods could correct that it’d be appreciated – okay it’ll make MikeS look pretty silly, but he’s been here two years and only posted forty-six times (pulls out calculator), that’s less than one post per fortnight! Why do have to pick on mine?

Snaffen-raffen MikeS.

You also need k to go from 0 to n-1, or from 1 to n. If you only go from 1 to n-1, you miss one of the roots.

snaffen-raffen Achernar

Really, how could I be so right, and yet so wrong?

Right, except, that I multiplied where I should have divided, and that the range of values for k was “out by one”, how else did the Romans mess me up?

snaffen-raffen Achernar

Really, how could I be so right, and yet so wrong?

Right, except, that I multiplied where I should have divided, and that the range of values for k was “out by one”, how else did the Romans mess me up?

Oh yes, they made me double post, but besides that…

Sorry to resurect this thread, but something is bothering me about this, if

1[sup]ln(5)/2π[/sup] = 5

then

5[sup]2π/ln(5)[/sup] = 1

but surely if

x[sup]y[/sup] = 1

then

|x| = 1

I must of gone wrong somewhere, but where?