1
$\begingroup$

This is actually from an Analysis text but i feel its a set theory question.

Proposition for ever rational number $\epsilon > 0$ there exists a non-negative number x s.t $x^2 < 2 < (x+ \epsilon )^2 $

It provides a proof that im having trouble understanding.

Proof: let $ \epsilon >0$ be rational. Suppose for contradiction sake that there is no on-negative rational number x that $x^2 < 2 < (x+ \epsilon )^2 $ holds.

ie when every $ x^2 < 2$ the statement $(x+ \epsilon )^2 <2 $

It states by a previous proposition that $(x+ \epsilon )^2 $ cannot equal 2.

Then it states "Since $0^2 < 2$ we thus have $ \epsilon ^2 < 2$ which then implies that $ (2\epsilon )^2 < 2$ and indeed a simple induction shows that $ (n\epsilon )^2 < 2$ for every natural number n." Which is what i cant understand.

The rest of the proof is strange as well im fine with the statement $ \epsilon ^2 < 2$ as it clearly follows that $ \epsilon ^2 < (x+ \epsilon )^2 $ as x is positive and $ \epsilon ^2$ is on both sides of the expression.

If i was proving it then i would rewrite

$ \epsilon ^2 = n $ $ \epsilon' $ s.t $n \in \mathbb {N} $ and $ \epsilon' \in \mathbb {Q} $

i would then use the Archimedean property to prove this is a contradiction.

If anyone can follow/explain what the bold text means i would greatly appreciate it.

$\endgroup$
  • 4
    $\begingroup$ "i feel its a set theory question." It isn't. $\endgroup$ – Andrés E. Caicedo Aug 25 '16 at 22:20
  • $\begingroup$ if you like, i proved it for you in a different way so that you can see a more transparent reasoning behind that statement. $\endgroup$ – Francesco Alem. Aug 25 '16 at 23:08
4
$\begingroup$

Before the bolded passage, you've concluded that if the statement you're trying to prove fails, then it must be the case that $x^2<2$ implies $(x+\epsilon)^2<2$.

Now, just take $x=0$. $0^2=0<2$, so we must have $(0+\epsilon)^2=\epsilon^2<2$.

So $\epsilon^2<2$. Now take $x=\epsilon$. Since $\epsilon^2<2$, we have $(\epsilon+\epsilon)^2=(2\epsilon)^2<2$.

Now take $x=2\epsilon$ . . .

More generally, having shown that $(n\epsilon)^2<2$, we may now take $x=n\epsilon$ and conclude that $(x+\epsilon)^2=((n+1)\epsilon)^2<2$. So, by induction, $(n\epsilon)^2<2$ for every natural number $n$. Since $\epsilon>0$, this contradicts the Archimedean property.

From your question, I think you're a little confused about what exactly is being proved by induction here. The goal is to prove $$(*)\quad \mbox{For each natural number $n$, $(n\epsilon)^2<2$.}$$ Note that this is not an induction on $\epsilon$ - it's an induction on the coefficient of $\epsilon$. The base case is $n=1$ (that is, showing $\epsilon^2<2$), and the inductive step is as described above.

$\endgroup$
  • $\begingroup$ Your a genius thanks. $\endgroup$ – Faust Aug 25 '16 at 22:16
2
$\begingroup$

The statement is true. here's a constructive proof for ya.

let $\epsilon\in \mathbf{Q}\,|\, \epsilon >0$

choose $k \in \mathbf{N}$ such that $k>2\epsilon$ so that $\frac{\epsilon}{k}<\frac{1}{2}$

choose $x \in [\sqrt 2 -\frac{\epsilon}{k};\,\,\sqrt2 - \frac{\epsilon}{2k}] \cap\mathbf{Q}$ arbitrarily (such an $x$ exists since rationals are dense in the reals, a consequence of the archimedean property).

$x$ is chosen in such a way that certainly makes it positive and lower than $\sqrt 2$ since $x<\sqrt 2 -\frac{\epsilon}{2k}<\sqrt 2$, hence it must hold $x^2<2$

also $x+\epsilon$ is certainly above $\sqrt 2$ since $x+\epsilon>\sqrt 2 -\frac{\epsilon}{k}+\epsilon>\sqrt 2$ hence it must hold $(x+\epsilon)^2>2$

it follows that $x^2<2<(x+\epsilon)^2$

$\endgroup$
1
$\begingroup$

We assumed that if $x^2<2$ then $(x+ϵ)^2<2$ (*).

Lets start with $x=0$: $x^2 = 0^2 = 0 < 2$, so we conclude that $(x+ϵ)^2 = ϵ^2<2$.

Now we have learned that $ϵ^2<2$, so we can use (*) again, with $x=ϵ$. We conclude that $(x+ϵ)^2=(ϵ+ϵ)^2=(2ϵ)^2<2$

Now we have learned that $(2ϵ)^2<2$, so we can use (*) again ... and again .. and again.

$\endgroup$

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.