3

This is rather an intuitive question, in the sense that indeed a real number raised to the power of an irrational doesn't make sense but I wanted to know what does it mean intuitively. If $2^7=2*2*2*2*2*2*2, 2^{\frac{1}{2}}=\frac{1}{\sqrt{2}}$ and $a^b=a*a*a*a*a*...*a\, \, b$ many times, I'm curious as to what does $2^π$ mean. Obviously it's equal to $8.8249778271...$ but what's the meaning in the aforementioned scenario?

Ski Mask
  • 1,988

7 Answers7

10

There is no meaning, if you use that definition of exponentiation (not that they teach you that in high shcool when they introduce irrational exponents; they usually just power through and expect you to accept it, because there is so much curriculum to get through and so little time. Kudos to you for being sceptical and actually thinking about it).

But that's an important part of what mathematics is about: Take something that seems like it ought to have a meaning, but doesn't quite have one, and see if you can't force it to work anyways (it's called "generalization"). There are three common ways to mend your issue:

  1. Instead define $a^b$ as $$\exp(b\ln a):= 1 + (b\ln a) + \frac12(b+\ln a)^2 + \frac16(b+\ln a)^3+\cdots$$ where the denominators of the fractions are the factorial numbers. The logarithm of $a$, $\ln a$ is the one number such that $\exp(\ln a) = a$, and $\pi\cdot \ln 2$ makes sense, since (I hope) you know how to make sense of multiplication of two real numbers. To prove: $\ln 2$ and the $\exp$ function actually exist and are well-defined.

  2. Define the value of irrational exponents as the limit of a sequence using rational exponents. In other words, take the sequence $$ 2^3, 2^{3.1}, 2^{3.14},2^{3.141},\ldots $$ and define $2^\pi$ to be the limit of this sequence. To prove: Different sequences of rational exponents that all converge to $\pi$ give the same result when you apply the above.

  3. Take a function $f: \Bbb R$ to $\Bbb R$ such that

    • $f$ is continuous
    • $f(a + b) = f(a)\cdot f(b)$
    • $f(1) = 2$

    then define $2^x := f(x)$, and evaluate $f(\pi)$. To prove: $f$ exists and is unique.

So, all of these approaches have things that need to be proven before they can be used effectively, and those things are by no means easy (and proving that they give the same result is not easy either). But as I said above, that's what mathematics is about.

Arthur
  • 204,511
  • Very nice answer! I think approach $2$ is best for intuition. – jonsno Jan 07 '18 at 13:53
  • @samjoe I agree, as long as evaluation is all you care about. But I think approach 3 stays truest to exponentials as you learn them in school (you can see the $a^{b+c} = a^b\cdot a^c$ rule standing there clear as day, for instance). Alas, I fear that approach 1 is the most common, and it's only really the best one if you want to differentiate things (or expand into bases with complex numbers; for real base with complex exponent approach 3 still works). – Arthur Jan 07 '18 at 13:58
  • At the beginning of some courses in real analysis, you will actually engage in answering questions like these. It is not trivial that elementary arithmetic extends to the reals (and that limits of certain infinite sequences exist). “Baby Rudin” actually offers a proof that n-th roots of real numbers exist in his construction of the reals (which is similar to your question). – BRSTCohomology Jan 07 '18 at 14:16
  • @MarcusAurelius That would be university / college courses. This problem will appear for people in high school. That's what I was referring to there. I'll make it clearer. – Arthur Jan 07 '18 at 14:18
  • Yeah, I was mostly highlighting the importance of such questions – BRSTCohomology Jan 07 '18 at 14:28
  • @samjoe: approach 2 is intuitive but hard to justify with rigor. Approach 1 is easiest to deal with precisely because of existence of very well behaved functions like $\exp, \log$ and is the favorite among many analysis texts. – Paramanand Singh Jan 08 '18 at 18:48
  • @ParamandSingh If I were to develop this with rigor, i would start with approach 1, and once that is done prove that 2 and 3 work and are equivalent. However, if rigor is not needed (like in, say, high school, where in my opinion things ought to be justified and explained for the sake of motivation and intuition, but the formal rigor isn't that important, or just when appealing to intuition in general) then I still prefer 3. – Arthur Jan 08 '18 at 19:19
6

Start with $2^3=8$. Then think about what $2^{3.1}$ is:

$$2^{3.1} = 2^{31/10} = \sqrt[10]{2^{31}} \approx 8.574187700.$$

Then consider

$$2^{3.14} = 2^{314/100} = \sqrt[100]{2^{314}} \approx 8.815240927.$$

Since the sequence $3, 3.1, 3.14, 3.141, \ldots$ approaches $\pi,$, the sequence

$$2^3, 2^{3.1}, 2^{3.14}, 2^{3.141}, \ldots$$

should approach $2^\pi.$ So it's probably best to think of $2^\pi$ as a limit.

B. Goddard
  • 33,728
2

$a^{k}$ is $a*a*...*a$, k times.
$a^{1/n}$ is the n-th root of $a$.
If $r$ is irrational, then $a^r=\lim\limits_{x \to r}a^x=\lim\limits_{k/n \to r} a^{k/n}$.

Botond
  • 12,134
2

We can make sense of numbers on the form $2^x$ when $x$ is rational using the way you describe: if $x = \frac{m}{n}$ with $m,n$ integers then $2^x = (2^{m})^{1/n}$ so it's the $n$'th root of $2\cdot 2\cdots 2\cdot 2$ ($m$ factors).

This way of looking at it does not apply to irrational $x$. But any irrational $x$ can be approximated arbitrarily well by rational numbers, i.e. we can find integer sequences $m_k$ and $n_k$ such that $x = \lim_{k\to \infty} \frac{m_k}{n_k}$ (for example for $\pi$ we can for example take $m_k=\{3, 31, 314, 3141,\cdots\}$ and $n_k = \{1,10,100,\cdots\}$). Now the usual way of defining $2^x$ is simply as the limit $2^x \equiv \lim_{k\to\infty}2^{\frac{m_k}{n_k}}$.

What this procedure is doing is to extend the power-function $2^x$ from the rationals (where we know how to compute it) to all the reals by demanding it to be a continuous function.

Winther
  • 25,313
1

$2^\pi = e^{\pi \ln(2)} = \sum_{n=0}^{\infty}\frac{(\pi\ln(2))^n}{n!}$...

Oria Gruber
  • 13,035
1

You know how to calculate powers when exponents are rational. Namely $$2^{\frac{m}{n}} = \sqrt[n]{2^m}$$

$2^\pi$ can then be defined as:

$$2^\pi = \sup\{2^q : q \in \mathbb{Q}, 0 < q < \pi\}$$


An equivalent definition would be $$2^\pi = \exp(\pi \ln2) = \lim_{n\to\infty} \left(1 + \frac{\pi \ln 2}{n}\right)^n$$

where $\exp$ is defined to be $$\exp x = \lim_{n\to\infty}\left(1 + \frac{x}{n}\right)^n$$ for $x \in \mathbb{R}$.

mechanodroid
  • 47,570
1

Consider the sequence $2^3, 2^{3.1}, 2^{3.14}, \dots$. We define $2^\pi$ to be the limit of this sequence.

We can compute, for example, $2^{3.14} = 2^{314/100}=\sqrt[100]{2^{314}}$. Practically, this is rediculous since it means we have to compute the $100^{\text{th}}$ root of

$$33374797436264220037422214158899251790667258161822699530422525122222183215322508594108782608384$$

But, theoretically, it makes perfect sense and there are tricks that might make such a calculation more reasonable to compute.