3
$\begingroup$

Prove that $\sqrt{17}$ is irrational by using the Well-ordering property of the natural numbers.

I've been trying to figure out how to go about doing this but I haven't been able to.

$\endgroup$
4
  • 1
    $\begingroup$ Try using contradiction, with $p/q$ in lowest terms. $\endgroup$
    – Clayton
    Commented Aug 20, 2013 at 1:01
  • 1
    $\begingroup$ May I please ask how this is a Calculus problem? $\endgroup$ Commented Aug 20, 2013 at 1:12
  • 1
    $\begingroup$ @AhaanRungta This is the sort of thing one may encounter in the opening sections of Spivak's Calculus or equivalent texts. (I agree that formally it is not calculus, and other tags would be more appropriate.) $\endgroup$ Commented Aug 20, 2013 at 1:13
  • $\begingroup$ Ah, yes, I see what you mean. Calculus books always open up with non-calculus. $\endgroup$ Commented Aug 20, 2013 at 1:15

4 Answers 4

10
$\begingroup$

In general, suppose $n$ is an integer that is not a perfect square. I will show that $\sqrt{n}$ is irrational. Note: This proof is not original, although I did come up with it independently.

Let $m = \lfloor \sqrt{n} \rfloor$. Since $n$ is not a perfect square (this is where we use that assumption), $m < \sqrt{n}$.

Suppose $\sqrt{n}$ is rational, so $\sqrt{n} = \frac{a}{b}$ for some positive integers $a$ and $b$.

At this point, we can go in two different directions. We can either assume that $b$ is the smallest positive integer such that $b\sqrt{n}$ is an integer, and deduce a contradiction by finding a smaller positive integer $d$ such that $d\sqrt{n}$ is an integer, or we can show that there is a smaller positive integer $d$ such that $d\sqrt{n}$ is an integer and thus create a contradiction using infinite descent.

I will use the first method.

Suppose $b$ is the smallest positive integer such that $b\sqrt{n}$ is an integer. Let $a = b\sqrt{n}$ As above, let $m = \lfloor \sqrt{n} \rfloor$, so that $m < \sqrt{n} < m+1$.

Then (watch closely - the fingers never leave the hands)

$\begin{align} \frac{a}{b} &=\sqrt{n}\\ &=\sqrt{n}\frac{\sqrt{n}-m}{\sqrt{n}-m}\\ &=\frac{n-m\sqrt{n}}{\sqrt{n}-m}\\ &=\frac{n-m(a/b)}{a/b-m}\\ &=\frac{nb-ma}{a-mb}\\ \end{align} $

so that $(a-mb)\sqrt{n} = nb-ma$ is also an integer.

Since $m < \sqrt{n} < m+1$, $m < a/b < m+1$ or $mb < a < mb+b$, or $0 < a-mb < b$.

Therefore $a-mb$ is a positive integer smaller than $b$ and $(a-mb)\sqrt{n}$ is an integer. This contradicts the definition of $b$ as the smallest such integer.

You can put $n = 17$ or $n=2$ (which is where I first saw this proof and generalized it to what you see here), or any non-perfect square and show that $\sqrt{n}$ is irrational this way.

$\endgroup$
8
$\begingroup$

Let $b$ be the smallest positive integer whose product with $\sqrt{17}$ is an integer (if $\sqrt{17}$ is rational then, by well-ordering, such a $b$ exists). Then $c=(\sqrt{17}-4)b$ is a smaller positive integer whose product with $\sqrt{17}$ is an integer, contradiction, hence $\sqrt{17}$ is irrational.

$\endgroup$
3
  • $\begingroup$ Short, simple, and strong! $\endgroup$
    – awllower
    Commented Sep 14, 2014 at 4:58
  • $\begingroup$ Where did the -4 come from? $\endgroup$
    – user157000
    Commented Oct 1, 2014 at 23:34
  • 1
    $\begingroup$ @user, I want $c$ to be smaller than $b$, so I need to multiply by something less than 1; I want $c$ to be positive, so I need to multiply by something positive; and I want $c$ to be an integer, so I need to multiply by $\sqrt{17}+r$ for some integer $r$. The only integer $r$ that makes $0<\sqrt{17}+r<1$ is $r=-4$, since $4<\sqrt{17}<5$. $\endgroup$ Commented Oct 1, 2014 at 23:56
5
$\begingroup$

HINT:

Suppose by contradiction the root of $17$ was rational, and without loss of generality positive. Take the least $p$ such that for some $q$ the rational $\frac pq$ is that root. Now derive a contradiction by showing that $p$ isn't minimal with respect to that property.

$\endgroup$
1
  • 4
    $\begingroup$ Further hint: $\sqrt{17}=17/\sqrt{17}$ so $p/q=(17q-4p)/(p-4q)$. $\endgroup$ Commented Aug 20, 2013 at 1:13
3
$\begingroup$

Suppose there exist integers $a_0,b_0$, with $a_0,b_0\gt 0$, such that $\left(\frac{a_0}{b_0}\right)^2=17$.

Then $a_0^2=17b_0^2$. Since the prine $17$ divides $17b_0^2$, it divides $a_0^2$, and therefore $17$ divides $a_0$. Let $a_0=17b_1$. Then $17^2b_1^2=17b_0^2$, and therefore $b_0^2=17b_1^2$. Let $a_1=b_0$.

We conclude that $a_1^2=17b_1^2$. It is easy to see that $b_1\lt b_0$.

Similarly, we can produce positive integers $a_2,b_2$ such that $b_2\lt b_1$ and $a_2^2=17b_2^2$, with $b_2\lt b_1$.

Continuing in this way, we can produce two infinite sequences $a_n,b_n$ such that $a_n^2=17b_n^2$ and $b_0\gt b_1\gt b_2\gt \cdots$.

In particular, we can produce an infinte descending sequence $b_0,b_1,b_2,\dots$ of positive integers.

This contradicts the Well-Ordering Principle.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .