As part of the course's assignments, we received a task to prove the following sentence using only Bézout identity:

Every common divisor of $a, b$ divides the gcd $(a, b)$.

I tried the following proof: By Bézout's identity, we know that $$\text{gcd}(a,b) = ax + by\tag{*}$$ for some integers $x,y$.

Let $c$ be the common divisor of $a$ and $b$. By definition, since $c$ divides $a$, we know that there exists $k_1$ so that $ck_1 = a$. The same is true for $b$, so $k_2$ exists so that $ck_2 = b$. Replace $a$ and $b$ in an equation $(*)$ and received:

(Edited) $$\text{gcd}(a,b) = cxk_1+cyk_2 = c(xk_1+yk_2)$$ Therefore we obtained that $c$, which is any common divisor of $a$ and $b$ divides gcd $(a, b)$.

Is proof enough? To my mind, it seems too simple. I'm pretty new in the elementary number theory world, I'd be happy to have another opinion. Thanks

  • 2
    $\begingroup$ You haven't used $x$ and $y$ at all, so it cannot be literally true, but it has the right ideas in it. $\endgroup$ – darij grinberg Apr 22 at 20:17
  • $\begingroup$ Oops, you're right. I meant to write like this: + gcd (a, b) = ax + by = cxk1 + cyk2 = c (xk1 + yk2) And the rest of the proof remains true $\endgroup$ – StevenU Apr 22 at 20:31
  • $\begingroup$ Yep, this way it's true. $\endgroup$ – darij grinberg Apr 22 at 20:32
  • 1
    $\begingroup$ This yields the fundamental GCD Universal Property $$ d\mid a.b,c\iff d\mid \gcd(a,b,c)$$ The dual for LCM is $$ a.b.c\mid m\iff {\rm lcm}(a,b,c)\mid m$$ These are in fact the definitions of GCD & LCM in more general rings. $\endgroup$ – Bill Dubuque Apr 22 at 20:58

I would define instead

$$k_1:= \frac{a\color{brown}x}{c}\implies a\color{brown}x=k_1c\qquad\qquad k_2:=\frac{b\color{brown}y}{c}\implies b\color{brown}y=k_2c$$

Therefore $$\text{gcd}(a,b)=a\color{brown}x+b\color{brown}y=k_1c+k_2c=c(k_1+k_2)\implies \color{red}{c\mid \text{gcd}(a,b)}$$

The proof doesn't make sense with your definitions of $\rm k_1$ and $\rm k_2$.


Apparently, your definition of greatest common divisor is “the largest common divisor”.

Assuming Bézout's identity $\gcd(a,b)=ax+by$, your proof is correct. If $c$ is a common divisor of $a$ and $b$, then by definition $a=ck_1$ and $b=ck_2$, for some integers $k_1$ and $k_2$, so you have $$ \gcd(a,b)=ax+by=ck_1x+ck_2y=c(k_1x+k_2y) $$ and this implies $c\mid\gcd(a,b)$.

Good job!

Linguistic remark: one should say “Let $c$ be a common divisor”, not “Let $c$ be the common divisor”.


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.