7

Let us assume $a$ is an integer which is rational which implies $a=p/q$ (where $p$ and $q$ are integers and $q$ not equal to $0$). If $p$ and $q$ are not coprime, let us simplify the fraction so this it is (I don't know how to talk like mathematicians).

Which implies, $$a=b/c$$ (where $b$ and $c$ are coprime integers). Squaring on both sides, \begin{align} a^2&=b^2/c^2\\ a^2c^2&=b^2 \end{align} So $a^2$ is a factor of $b^2$, and also of $b$, due to the uniqueness of the fundamental theorem of arithmetic. So, \begin{align} b &=a^{2}d \tag{where $d$ is an integer}\\ b^2 &= a^{4}d^{2} \end{align} But $b^2=a^2c^2$ So, \begin{align} a^2c^2 &= a^4d^2\\ c^2 &= a^2d^2 \end{align}

So, $a^2$ is a factor of $c^2$ and $c$ due to the fundamental theorem of arithmetic. So $b$ and $c$ have $a^2$ as a common factor. But this contradicts the fact that $b$ and $c$ are coprime. This is because we have taken $a$ as a rational integer, so $a$ cannot be a rational integer.

What's wrong here (genuinely asking)?

  • 12
    "a^2 is a factor b^2 and b" You are right $a^2$ divides $b^2$, but why would it divide $b$? – Wojowu Jul 07 '19 at 16:06
  • 7
    First error I spotted is that $a^2$ need not be a factor of $b$ just because it is a factor of $b^2$. Indeed, if $a=2$ and $b=2$, then $a^2=4$ is a factor of $b^2=4$ but $a^2=4$ is not a factor of $b=2$. – Dave Jul 07 '19 at 16:06
  • 3
    Also, it is clear that integers are rational because given any integer $a$ we can write $a=\frac{a}{1}$. – Dave Jul 07 '19 at 16:08
  • You should probably check out https://math.meta.stackexchange.com/questions/5020/mathjax-basic-tutorial-and-quick-reference to help you format your future posts. For example, it's hard to read "This is because we have taken a as a rational integer so a cannot be a rational integer" as opposed to "This is because we have taken $a$ as a rational integer so $a$ cannot be a rational integer". – Hendrix Jul 07 '19 at 16:20
  • In the time of typing this, could you not have gone through an example like $a=2 = 2/1$ and noticed that $2^2$ being a factor of $2^2$ does not imply that $2^2$ is a factor of $2$? Genuinely asking. – Torsten Schoeneberg Jul 07 '19 at 16:39
  • "Also, it is clear that integers are rational because ...." That does not answer the question "what is wrong with this proof"? It's obvious that integers are rational. But the OP has a proof that they are not. So the proof must be wrong. So the OP is guenuinely asking where the error is. Pointing out that integers are obviously rational does nothing to answer that. – fleablood Jul 07 '19 at 16:49
  • 4
    If b and c are coprime, and a=b/c, then what does c have to be? – pokep Jul 07 '19 at 17:33
  • 6
    @Toolazytothinkofaname It is rather simple to spot the error. Substitute the variables by actual integers and find where the proof is wrong. – miracle173 Jul 08 '19 at 14:18

3 Answers3

44

The problem in the proof is that $a^2|b^2\nRightarrow a^2|b$. For instance, take $a=2$ and $b=6$. Clearly, $4|36$ but $4\nmid 6$.

Anand
  • 3,588
7

I think you are confusing that if $p$ is prime and $p$ divides $b^k$ then $p|b$. That is true if $p$ is prime.

Actually it's also true for a composite $a|b^k$ then $a|b$ if $a$ has no square factors. But if $a$ as any prime factors to a power greater than $1$ it need not be true.

And in fact its obviously not true as $a^2$ divides $a^2$ but $a^2$ doesn't divide $a$ (unless $a = 1$).

Read on....

It most certainly is not true if $a|b^k$ that $a|b$ It means that the prime factors of $a$ are prime factors of $b$. And it means that the powers of those prime factors of $a$ are at most equal to $k$ times the powers of the same prime factors of $b$ but because $k$ is larger than .....

Oh let me put it this way.

Suppose $a = \prod p_i^{m_i}$ be the prime factorization of $a$. Suppose $a|b^k$. Then that means that $p_i$ are prime factors of $b$ and that $b = d\prod p_i^{j_i}$. And it means that $b^k = d^k \prod p_i^{k*j_i}$.

And as $a|b^k$ that means each $m_i \le k*j_i$. But that does not mean $m_i \le j_i$ which would mean $a|b$.

You statement $a|b^k$ means $a|b$ if $a$ has square free and all the prime factor powers were $1$ but not other wise.

Simple example if $a = 12 = 2^2*3$ and $b= 90 = 2*3^2*5$. Now $a|b^2 = 8100 = 2^2*3^4*5^2$.

This means the prime factors of $a$ ($2,3$) are also prime factors of $b$. And it means that the powers of the prime factors of $a$ ($2\mapsto 2; 3\mapsto 1$) are less or equal to $2$ times the powers in $b$ ($2\mapsto 1$ and $2 \le 2*1$ and $3\mapsto 2$ and $1 \le 2*2$) but it doesnt mean the are less than or equal to the powers of $b$. (In $a; 2\mapsto 2$ but in $b; 2\mapsto 1$ and $2 \not \le 1$).

So $12 \not \mid 90$.

It's certainly can't be the case that $a|b \implies a^2| b^2 \implies a^2|b$! That would mean every time you have $a|b$ you can just keep squaring and reducing to get $a^{m}|b$ for any power of $m$.

That would mean if $3|6$ then $3^2|6$ and $3^4|6$ and $3^{2048}|6$ and so on.

Or in this case as $a = b$ (and $c=1$.... because $a$ is an integer) you would have $a|a$ so $a^2|a$? And $a^4|a$. That's .... simply not true.

fleablood
  • 124,253
  • No, $,p\mid b^k,\Rightarrow, p\mid b,$ is true $\iff p,$ is squarefree. Follow the link for a handful of characterizations of squarefree integers. – Bill Dubuque Jul 07 '19 at 17:44
  • Why do you say "no"? That is exactly what I said. – fleablood Jul 07 '19 at 17:52
  • Because the first paragraph was incorrect. Now it is correct after your edit. – Bill Dubuque Jul 07 '19 at 17:54
  • 1
    Okay. I used "only" colloquially. My bad. I'm pretty sure the OP was confusing the FTA with Euclid's lemma. So I said that only works if $p$ is prime. Colloquially that doesn't mean $p$ being prime is required and it is false otherwise. It means you can only cite that lemma if $p$ is prime. There are many other ways $a|b^k$ and $a|b$ can both be true but citing Euclid's lemma is usually reserved for $a$ prime. But in math I shouldn't have used the loaded word "only". – fleablood Jul 07 '19 at 18:56
2

Basic facts missing $ac=b$ is a lot easier to use. $a^2$ does not need to divide $b$. A fraction sharing no common factor other than 1, between the number on top ( numerator), and the number on the bottom ( denominator), is said to be in lowest terms .

Anyways starting from $a={b\over c}$ we get $ac=b$ showing c divides b, sharing no factor other than 1, and therefore, $c=1$, implying $a=b$ so $a={a\over 1}$ it Also can be used to show :$a={-a\over -1}$