The Student Room Group

Cube root of -8

Most people would agree (so I think) that when considering the reals:
(8)13=2(-8)^{\frac{1}{3}}=-2

So:
(8)26=2(-8)^{\frac{2}{6}}=-2
(8)26=(82)16(-8)^{\frac{2}{6}}=(-8^{2})^{\frac{1}{6}}
(82)16=6416(-8^{2})^{\frac{1}{6}}=64^\frac{1}{6}

But 6416=264^\frac{1}{6}=2

Therefore:
(8)13=2(-8)^{\frac{1}{3}}=2

Discuss
Reply 1
When you take the root apqa^{\frac{p}{q}} you require pq\dfrac{p}{q} to be in its lowest terms (specifically, you require the gcd of p and q to be 1).
nuodai
When you take the root apqa^{\frac{p}{q}} you require pq\dfrac{p}{q} to be in its lowest terms (specifically, you require the gcd of p and q to be 1).

Heh, saw you reply and knew you'd own my post...is that by definition? Never knew that...
Reply 3
nuodai
When you take the root apqa^{\frac{p}{q}} you require pq\dfrac{p}{q} to be in its lowest terms (specifically, you require the gcd of p and q to be 1).


Duh!

PS: Wow, you're really good at maths...
Reply 4
milliondollarcorpse
Heh, saw you reply and knew you'd own my post...is that by definition? Never knew that...

It's sort of by definition, but not really. I mean, clearly we must have (8)2/6=(8)1/3(-8)^{2/6} = (-8)^{1/3} otherwise it wouldn't make sense. However, the rule ap/q=apq=(aq)pa^{p/q} = \sqrt[q]{a^p} = (\sqrt[q]a)^p only necessarily holds when pq\frac{p}{q} is in its lowest terms. It follows from a similar reasoning that gives the result that if you say that x2=1x=1x^2=1 \Rightarrow x=1 then you lose a root of the equation (namely x=1x=-1).

Now... who negged me?
nuodai
It's sort of by definition, but not really. I mean, clearly we must have (8)2/6=(8)1/3(-8)^{2/6} = (-8)^{1/3} otherwise it wouldn't make sense. However, the rule ap/q=apq=(aq)pa^{p/q} = \sqrt[q]{a^p} = (\sqrt[q]a)^p only necessarily holds when pq\frac{p}{q} is in its lowest terms. It follows from a similar reasoning that gives the result that if you say that x2=1x=1x^2=1 \Rightarrow x=1 then you lose a root of the equation (namely x=1x=-1).

Ah I see. I knew TSR would solve this pretty quickly...stomped the maths teachers I asked in my school :p:
nuodai
It's sort of by definition, but not really. I mean, clearly we must have (8)2/6=(8)1/3(-8)^{2/6} = (-8)^{1/3} otherwise it wouldn't make sense. However, the rule ap/q=apq=(aq)pa^{p/q} = \sqrt[q]{a^p} = (\sqrt[q]a)^p only necessarily holds when pq\frac{p}{q} is in its lowest terms. It follows from a similar reasoning that gives the result that if you say that x2=1x=1x^2=1 \Rightarrow x=1 then you lose a root of the equation (namely x=1x=-1).

Now... who negged me?

You're a sub aren't you? Surely you can see who negged you anyway?
Reply 7
Farhan.Hanif93
You're a sub aren't you? Surely you can see who negged you anyway?

I've not been a sub for aaaaages!

Anyway OP, here's a more thorough explanation.

Suppose that we have x=ap/qx=a^{p/q}, then xq=apx^q=a^p, so far so good. Suppose that qq is even, so that xq=xqx^q = |x|^q, so that xq=ap|x|^q=a^p. We can't then deduce that x=ap/q|x|=a^{p/q}, since that would imply that x=xx=|x| (which isn't true when x<0x<0).

So, although we can say that x=(8)1/3x3=8x=83x=(-8)^{1/3} \Rightarrow x^3=-8 \Rightarrow x = \sqrt[3]{-8}, we can't say x=(8)2/6x6=(8)2=64x=646x=(-8)^{2/6} \Rightarrow x^6 = (-8)^2=64 \Rightarrow x = \sqrt[6]{64} and then take the positive root.
nuodai
I've not been a sub for aaaaages!

Oh. Your avatar is still large though so that probably threw me. :s-smilie:
I think I'll stop derailing now. :ninja:
nuodai
I've not been a sub for aaaaages!

Anyway OP, here's a more thorough explanation.

Suppose that we have x=ap/qx=a^{p/q}, then xq=apx^q=a^p, so far so good. Suppose that qq is even, so that xq=xqx^q = |x|^q, so that xq=ap|x|^q=a^p. We can't then deduce that x=ap/q|x|=a^{p/q}, since that would imply that x=xx=|x| (which isn't true when x<0x<0).

So, although we can say that x=(8)1/3x3=8x=83x=(-8)^{1/3} \Rightarrow x^3=-8 \Rightarrow x = \sqrt[3]{-8}, we can't say x=(8)2/6x6=(8)2=64x=646x=(-8)^{2/6} \Rightarrow x^6 = (-8)^2=64 \Rightarrow x = \sqrt[6]{64} and then take the positive root.

Very good explanation thanks, I'll pos rep you to balance out the neg one someone gave :p:
Reply 10
nuodai
When you take the root apqa^{\frac{p}{q}} you require pq\dfrac{p}{q} to be in its lowest terms (specifically, you require the gcd of p and q to be 1).


The only good resolution to this paradox is to impose arbitrary limitations. Essentially, it's only safe to assume (ap)q=apq(a^p)^q = a^{pq} if p and q are both integers or a is positive. (Consider 1\sqrt{-1}. Clearly, 11=1\sqrt{-1} \sqrt{-1} = -1. But 1111=1\sqrt{-1} \sqrt{-1} \ne \sqrt{-1 \cdot -1} = 1.) The bottom line is that for the general complex number, non-integer powers are not well-defined and depend on an arbitrary choice of a branch of what really ought to be considered a multi-valued function.

Latest