5
$\begingroup$

In a previous, and quite popular, question it was discussed about whether or not $\pi$ contains all finite number combinations.

Let us assume for a moment that $\pi$ does in fact contain all finite combinations of numbers. What prevents $\pi$ from also containing all infinite sets?

It seems that at some point one would also see the first couple digits of, for example, $e$ (2.71828). But why does it need to stop there, couldn't it contain a bunch of digits of $e$? Perhaps even an infinite number of digits of $e$?

My understanding is that $e$ could also be replaced by $\sqrt2$ or any other irrational number, so long as that irrational number contained all finite sets of number combinations. Which might imply that somewhere along the way, $e$ contains a number of digits of $\pi$. Implying this ridiculous situation where within $\pi$ we see $e$, and then within $e$ we again begin to see $\pi$ again. Then all the universe collapses into a singularity. Or maybe someone can just explain why one infinite sequence can't contain another infinite sequence, and perhaps why we have not defined some type of super-infinity that can.

To reiterate the primary question: What prevents $\pi$, or other infinite irrational number that contains all finite sets of numbers, from also containing all infinite sets?

$\endgroup$
10
$\begingroup$

If the decimal expansion of $\pi$ contained the string of digits '$000\ldots$' then it only has a finite number of non-zero digits (some subset of the digits before the string of $0$s starts) and so $\pi$ is a rational number. Contradiction.

$\endgroup$
  • $\begingroup$ That makes a lot of sense. So clear, I almost feel silly for not having realized it. $\endgroup$ – mwjohnson Sep 11 '13 at 1:35
  • 3
    $\begingroup$ The point is that infinite strings don't have an end (on the right as we write them) so there are very few infinite strings actually contained in $\pi$ - only those of the form $[10^n\cdot\pi]$ for $n\geq -1$ where $[\bullet]$ is the 'fractional part' function. $\endgroup$ – Dan Rust Sep 11 '13 at 1:38
  • 3
    $\begingroup$ Rationality and irrationality aside, if it contains 000000... it certainly cannot also contain 111111.... $\endgroup$ – Nate Eldredge Sep 11 '13 at 2:10
  • 1
    $\begingroup$ @NateEldredge Yeah there's a bunch of natural contradictions you can reach. $\endgroup$ – Dan Rust Sep 11 '13 at 2:15
  • $\begingroup$ Well, you should carefully distinguish between "set" and "sequence". It does contain many different sequences: for instance, 1415926..., 415926..., 15926..., etc, where each one is some "tail" of all the ones before it. But it cannot contain both of two sequences where neither is a tail of the other. $\endgroup$ – Nate Eldredge Aug 6 '18 at 16:26
7
$\begingroup$

There are an uncountable number of infinite digit strings, and $\pi$ contains only countably many infinite digit strings-namely all the ones that are the digits after some point in the expansion. It does contain infinite strings, just not most of them.

$\endgroup$
  • 1
    $\begingroup$ This is a stronger answer because it also disposes of a lot of variations such as "what if we only want each infinite sequence of digits to appear at some set of positions that form an arithmetic progression?" $\endgroup$ – Henning Makholm Sep 11 '13 at 13:05
  • $\begingroup$ @user500668: that depends what you mean by includes. You could start after the fifth decimal place-would you say it includes that sequence? You could take every other term. You could do a lot of things. $\endgroup$ – Ross Millikan Aug 6 '18 at 16:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.