2
$\begingroup$

Show that there is no right triangle whose legs are rational numbers and whose hypotenuse is $\sqrt{2022}$. My tries:

  1. I used Pythagoras' Theorem to get: $$\sqrt{2022}^2=a^2+b^2 \implies a^2+b^2 = 2022$$ where $a$ and $b$ are the legs of the triangle. I don't know what to do next: Is there another formula I could use? I know that $a+b>\sqrt{2022}$ but I don't think this is going to help us much.

hope one of you can help me! thank you!

$\endgroup$
8
  • $\begingroup$ If you were to rewrite $a$ and $b$ with common denominator $k$, you then need to show that $2022\,k^2$ is not the sum of two squares of integers for any positive integer $k$. You might end up using $2022=2 \times 3\times 337$ $\endgroup$
    – Henry
    Commented Jan 3, 2023 at 10:10
  • $\begingroup$ i dont really understand how you rewrite them with k $\endgroup$
    – user1104319
    Commented Jan 3, 2023 at 10:22
  • $\begingroup$ Note that $a$ and $b$ are fractions, and then multiply out the denominators of $2022 = a^2 + b^2$. We have a lot more tools for dealing with equations of integers than we do for dealing with equations of rational numbers, so getting rid of all fractions is a quite natural step. $\endgroup$
    – Arthur
    Commented Jan 3, 2023 at 10:24
  • $\begingroup$ We know $a$ and $b$ are rational. So, we can represent them as $a = \frac xk$ and $b = \frac yk$ for some natural numbers $x$,$y$ and $k$. Then the equation becomes:$2022 = \frac{x^2}{k^2}+\frac{y^2}{k^2} \implies 2022k^2 = x^2+y^2$. $\endgroup$
    – D S
    Commented Jan 3, 2023 at 10:35
  • $\begingroup$ Pythagoras, not "pitagora". $\endgroup$ Commented Jan 3, 2023 at 10:49

4 Answers 4

4
$\begingroup$

Quickly ruling out $\ a\ $ being even, we try $\ a\ $ and $\ b\ $ both being odd:

$$ (2k_1+1)^2 + (2k_2+1)^2 = 2022\quad k_1,k_2\in\mathbb{Z}$$

$$ \implies 4({k_1}^2 + {k_2}^2 + k_1 + k_2) + 2 = 2022 $$

$$ \implies {k_1}^2 + k_1 + {k_2}^2 + k_2 = 505, $$

which is impossible, since $\ {k_i}^2 + k_i\ $ is even for $\ i=1,2.$

$\endgroup$
5
  • $\begingroup$ can you plz look at the comments above and do a solution as it says......o rewrite a and b with common denominator k.... i tried doing it that way......anyways, tysm $\endgroup$
    – user1104319
    Commented Jan 3, 2023 at 11:49
  • $\begingroup$ the question asks for rational and not necessarily integral $a$ and $b$. $\endgroup$
    – D S
    Commented Jan 3, 2023 at 12:38
  • $\begingroup$ Ah - I didn't see the question says rational. Apparently, a lot of other people missed this also! $\endgroup$ Commented Jan 3, 2023 at 13:02
  • $\begingroup$ @AdamRubinson please see my answer $\endgroup$
    – D S
    Commented Jan 3, 2023 at 13:07
  • $\begingroup$ and also @BobDobbs answer $\endgroup$
    – D S
    Commented Jan 3, 2023 at 13:08
2
$\begingroup$

Let $a=\frac{x}{k}$, $b=\frac{y}{k}$, where $x,y,k$ are positive integers and $\gcd(x,y,k)=1$. Then, the equation becomes $x^2+y^2=2022k^2$. From this equation, it is clear that $x$ and $y$ must have the same parity. If they are both even, then $k$ is odd and the equation has no solution since $4\not|2022$. Therefore, $x$ and $y$ must be odd. If $x,y$ are odd then since square of an odd number modulo $8$ is $1$, we get $k^2\equiv -1\pmod 8$ which is not possible.

$\endgroup$
4
  • $\begingroup$ I don't understand how $x,y$ must be odd... $\endgroup$
    – D S
    Commented Jan 3, 2023 at 13:28
  • $\begingroup$ $k$ can be even. $\endgroup$
    – D S
    Commented Jan 3, 2023 at 13:28
  • $\begingroup$ @DS We can assume that $\gcd(x,y,k)=1$. $\endgroup$
    – Bob Dobbs
    Commented Jan 3, 2023 at 13:31
  • $\begingroup$ see encyclopediaofmath.org/wiki/Legendre_theorem I like the treatment in Cassels, Rational Quadratic Forms, pages 80-82, especially the note middle page 82 $\endgroup$
    – Will Jagy
    Commented Jan 3, 2023 at 17:34
1
$\begingroup$

An alternative answer: From the comments, I have shown that $$x^2+y^2 = 2022k^2$$ So, it remains to show that $2022k^2$ can not be represented as the sum of squares of 2 integers. For contradiction, assume that $2022k^2$ can be represented as the sum of $2$ squares. From the Sum of two squares theorem, if $2022k^2$ can be represented as the sum of squares of $2$ integers, it will have prime factors of the form $4n+3$ raised to an even power. $3$ is of the form $4n+3$. Let $$k = 3^pq$$ where $p \ge 0$ and $q$ are integers such that $3$ does not divide $q$. Then, $k^2 = 3^{2p}q^2$. Thus, $$2022k^2 = 2\times 3 \times 337 \times 3^{2p}q^2 = 3^{2p+1} \times 2 \times 337 \times q^2$$We know that $q^2$ does not contain $3$ in its prime factorisation, since $q$ does not contain $3$ in its prime factorisation. So, $2022k^2$ has $3^{2p+1}$ in its prime factorisation and no other $3$. But, $2p+1$ is an odd number. Thus, we get a contradiction, $2022k^2$ cannot be represented as the sum of squares of $2$ integers, meaning $2022$ cannot be represented as the sum of squares of $2$ rational numbers.

$\endgroup$
8
  • $\begingroup$ Sum of two squares theorem is interesting. $\endgroup$
    – Bob Dobbs
    Commented Jan 3, 2023 at 13:39
  • 1
    $\begingroup$ I came across it (Theorem 3.13) here: math.uchicago.edu/~may/VIGRE/VIGRE2008/REUPapers/Bhaskar.pdf $\endgroup$
    – D S
    Commented Jan 3, 2023 at 14:06
  • $\begingroup$ @BobDobbs The more interesting result in my opinion (although irrelevant in this situation) is that the converse holds for primes: If the prime factorization of an integer has an even number of each $4n+3$ prime factor and this integer also is prime (which is to say, if this integer is a prime of the form $4n+1$), then it can be written as the sum of two squares. $\endgroup$
    – Arthur
    Commented Jan 3, 2023 at 14:49
  • $\begingroup$ I think there is a formula for $r_2(n)$. $\endgroup$
    – Bob Dobbs
    Commented Jan 3, 2023 at 15:01
  • 1
    $\begingroup$ @Arthur I'm not disagreeing. Just thought I should add a link to the Wikipedia page for curious readers. $\endgroup$
    – D S
    Commented Jan 4, 2023 at 9:57
1
$\begingroup$

I think this answer is better than my previous one.


I was doing an unrelated question when I thought of this question and a solution struck me too (don't ask me how). So here it goes:
Write the equation as (using different variables this time): $$\left(\frac pr\right)^2 +\left(\frac qs\right)^2 = 2022$$where these are the fractions in their lowest terms. We get:
$$\color{green}{(ps)^2+(qr)^2 = 2022(rs)^2}$$Both sides are integers, one is divisible by an even power of $3$, the other by an odd power. Hence, there is no solution. (Because in LHS, both terms can only be $0,1 \pmod 3$. Let $ps = 3^mk$ and $qr = 3^nl$ such that $k,l \neq 0 \pmod 3$. Then LHS is divisible by $3^{2m}$ or $3^{2n}$, whichever is lesser. So, LHS is divisible by an even power of $3$. On the other hand, RHS is divisible by an odd power of $3$ since $2022 = 3 \cdot 2 \cdot 337$. )

$\endgroup$

You must log in to answer this question.