1
$\begingroup$

I've got a monster of a proof by contradiction going where to finish I need to show that there is no $\alpha\in \mathbb{Q}(x)$ (the field of fractions of $\mathbb{Q}[x]$) such that $\alpha^2=x$. Obviously this is equivalent to showing that $\sqrt{x}\not\in\mathbb{Q}(x)$, which seems trivial enough, but I'm stuck.

Here's what I've done so far:

If $\sqrt{x}=\frac{p(x)}{q(x)}$ where $q(x)=a_0+...+a_nx^n$, then $p(x)=a_0x^{1/2}+...+a_nx^{n/2}$. It seems like this is a contradiction, but we are of course not prevented from choosing $a_0,...,a_n$ to be nonzero only for even terms greater than two, so it isn't.

Is there a quick algebraic proof of this fact?

$\endgroup$
5
  • 5
    $\begingroup$ Hint: try to adapt the proof that $\sqrt 2\not \in \mathbb Q$. Start with $xq^2(x)=p^2(x)$. $\endgroup$
    – lulu
    Apr 5, 2021 at 15:31
  • 1
    $\begingroup$ rational function where the polynomials are real? $\endgroup$
    – Asinomás
    Apr 5, 2021 at 15:31
  • 8
    $\begingroup$ More simply, one could just note that, if you had suitable $p,q$, then $\sqrt 2 =\frac {p(2)}{q(2)}\in \mathbb Q$. $\endgroup$
    – lulu
    Apr 5, 2021 at 15:33
  • $\begingroup$ You can prove by using the algebraic properties that a rational function whose square is $x$ must satisfy, or by looking at the growth rates of rational functions. $\endgroup$
    – Asinomás
    Apr 5, 2021 at 15:38
  • $\begingroup$ Assuming it is, you can also get a contradiction about the degrees of the polynomials. $\endgroup$
    – leo
    Apr 5, 2021 at 15:45

2 Answers 2

5
$\begingroup$

If $p(x)$ and $q(x)$ are polynomials, let $m=\deg p(x)$ and let $n=\deg q(x)$. Then$$\lim_{x\to\infty}\frac{\sqrt x}{p(x)/q(x)}=\begin{cases}\infty&\text{ if }m-n\leqslant0\\0&\text{ if }m-n>0.\end{cases}$$Therefore, $\sqrt x$ is not a rational function; if it was equal to $\frac{p(x)}{q(x)}$, that limit would be $1$.

$\endgroup$
2
  • $\begingroup$ Can you answer my little question? Is my claim correct? If $\text{deg}~f(x)>\text{deg}~g(x)$, then $\text{deg}~f'(x)>\text{deg}~g'(x)$. Thank you! $\endgroup$ Apr 5, 2021 at 17:41
  • 1
    $\begingroup$ @lonestudent Yes, for non-constant polynomials. $\endgroup$ Apr 5, 2021 at 17:42
4
$\begingroup$

$$ \text{------------------------------------------------------------------------------------------------------} $$ Suppose there exists $\alpha\in\mathbb{Q}(x)$ such that $|\alpha(x)|=\sqrt{x}$ for all $x\in\mathbb{Q}$ with $x\ge 0$.

Then in particular, we have $|\alpha(2)|=\sqrt{2}$, contradiction, since $\alpha(2)$ can't be irrational. $$ \text{------------------------------------------------------------------------------------------------------} $$ Here's another way . . .

Suppose there exists $\alpha\in\mathbb{Q}(x)$ such that $\alpha^2=x$.

Then in particular, we have $\bigl(\alpha(-1)\bigr)^2=-1$, contradiction, since $\bigl(\alpha(-1)\bigr)^2$ can't be negative. $$ \text{------------------------------------------------------------------------------------------------------} $$ Here's still another way, which has the advantage of working for $K(x)$, where $K$ is any field . . .

Let $K$ be a field, and suppose there exists $\alpha\in K(x)$ such that $\alpha^2=x$.

Then writing $\alpha={\large{\frac{p}{q}}}$ for some nonzero $p,q\in K[x]$, we get $$ p^2=xq^2 $$ contradiction, since $\deg(p^2)$ is even, whereas $\deg(xq^2)$ is odd. $$ \text{------------------------------------------------------------------------------------------------------} $$

$\endgroup$

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.