4

Suppose $\,ab \equiv 0 \pmod n\,$ and $\,0<a,b < n \ $ (so $n\mid ab,\ n\nmid a,b)$

Does it follow that either $a | n$ or $b | n$?

For this question is this a suitable answer: No it doesn't. $3$ and $4$ are both less than $6$. $$3 \times 4 \equiv 0 \pmod {6}$$

but $4 \not \mid 6.\,$ However, can such $\,a,b\,$ always be found?

Bill Dubuque
  • 272,048
Megan
  • 65

3 Answers3

4

No, your answer is incorrect.

Either/Or means it's sufficient that one of them abides the condition.

And in the specific example at hand, one of them indeed does, as $3|6$.

A proper counterexample would be $8\times9\equiv0\pmod{12}$.


$$[a=8,b=9,n=12]\implies[ab\equiv0\pmod{n}]\wedge[a<n]\wedge[b<n]\wedge[a\not|n]\wedge[b\not|n]$$

barak manos
  • 43,109
1

$$10 \dot \ 12 \equiv 0 \mod 15$$

But neither $10|15 $, nor $12 |15$.

Edit:

If you consider $n$ prime, then $\gcd(n , a) =1$ for $a = 1, \ldots, n-1$. Let any $\overline{a} \in \mathbb Z_n - \{\overline{0}\}$ then there exist integers $b$ and $t$ such that

$$ab + nt =1 \Rightarrow \overline{ab + nt} = \overline{1} \Rightarrow \overline{a}\ \ \overline{b} + \underbrace{\overline{n}}_{=\ 0}\ \ \overline{t} = \overline{1} \Rightarrow \overline{a}\overline{b} = \overline{1}$$

then $\overline{a}$ is invertible, it follows that $\mathbb Z_n$ is a field (every element is invertible), therefore a integral domain.

Aaron Maroja
  • 17,571
1

Zero divisors in $\Bbb Z_n$ needn't be factors of $\,n,\,$ e.g. $\,6\mid 4\cdot 9\,$ but $\,4,9\nmid 6.\,$ Yes, we can force it true by choosing smaller factors of each factor, as follows. $ $ Suppose that $\!\bmod n\!:\ ab\equiv 0\,$ for $\,a,b\not\equiv 0.\ $ Let $\,k =\,$ additive order of $\,b.\,$ Then $\,a\cdot b\equiv 0\equiv n\cdot b\,\Rightarrow\, k\mid a,n\,\Rightarrow\,k\mid (a,n),\,$ so $\,(a,n)b\equiv 0.\,$ Similarly we deduce $\,(a,n)(b,n) \equiv 0,\,$ where, indeed, $\,(a,n)\mid n,\ (b,n)\mid n;\,$ e.g. applying this to the above we get $\, 6\mid 4\cdot 9\,\Rightarrow\, 6\mid(4,6)(9,6) = 2\cdot 3,\,$ where $\,2,3\mid 6$.

Thus if $\,n\,$ fails Euclid's Lemma (Prime Divisor Property) $\,n\mid ab\,\Rightarrow\,n\mid a\,$ or $\,n\mid b,\,$ the above yields a constructive proof that $\,n\,$ is composite, by constructing a proper factor $\,(a,n)\,$ of $\,n.\,$ Indeed, $\,(a,n)\ne n\,$ else $\,n\mid a,\,$ and $\,(a,n)\ne \color{$c00}1\,$ else $\,n\mid (a,n)(b,n)=(b,n)\,\Rightarrow\,n\mid b.\,$ For example, given any polynomial $\,f(x)\in\Bbb Z_n[x]\,$ with more roots than its degree, we can quickly compute a nontrivial factor of $\,n\,$ via a $\rm\,gcd.\,$ Said failure of Eclid's Lemma is sometimes described with the language: $\,ab\,$ is a witness that $\,n\,$ is composite, i.e. $\,n\mid ab,\ n\nmid a,b$.

Said algebraically: not prime $\,\Rightarrow\,$ reducible or, contrapositively, irreducible $\,\Rightarrow\,$ prime. Since the converse always holds, we conclude: $\,\Bbb Z/n\,$ is a domain $\iff n\,$ is prime $\iff n\,$ is irreducible.

This idea is sometimes used in integer factorization algorithms to deduce a factor of $\,n\,$ from a "nontrivial" factor of a multiple of $\,n.$

The above constructive proof also generalizes to certain ideals, namely

Theorem $ $ If ideal $I\ne 1$ satisfies: ideal $\,J \supset I \Rightarrow J\,|\,I\,$ then $I$ not prime $\Rightarrow I\,$ reducible (properly).

Proof $\ $ $I$ not prime $\Rightarrow$ exists $a,b \not\in I$ and $ab \in I.$ $\ A := (I,a)\supset I \Rightarrow A|I,$ say $I = AB;$ wlog we may assume $b \in B$ since $A(B,b) = AB\,$ via $Ab = (I,a)b \subset I = AB.$ The factors $A,B$ are proper: $A = (I,a),\, a \not\in I;\,\ B \supset (I,b),\, b \not\in I.\quad$ QED

Bill Dubuque
  • 272,048
  • But I thought that Barak's example was when it "doesn't follow" – Megan Feb 01 '15 at 16:11
  • @Megan The point is that every counterexample can be transformed into an example as described above. – Bill Dubuque Feb 01 '15 at 16:12
  • So what you are basically saying is that every example that does not follow can be forced into saying that it does follow? – Megan Feb 01 '15 at 16:13
  • Can you also explain each step that you took. I am having trouble seeing how this is true for my problem. – Megan Feb 01 '15 at 16:18
  • @Megan Look at the example. Although the property does not hold for $,12\mid 8\cdot 9,$ we can mechanically transform it to $,12\mid 4\cdot 3,$ where it does hold, which is often a useful way of finding factors of $,n.,$ – Bill Dubuque Feb 01 '15 at 16:22
  • But doesnt 12 divide 72? – Megan Feb 01 '15 at 16:28
  • Wait nevermind I get what you are saying. You made it to 4 and 3 so now we can have 12 divides 4 and 12 divides 3. Right? – Megan Feb 01 '15 at 16:29
  • @Megan Yes, that's the point, we can always transform it that way by taking gcds as shown. This idea often proves useful, esp. when viewed in the language of "orders" of elements as in the proof. The structure at the heart of the matter will become clearer if you study (cyclic) groups (but you needn't know that to understand the above). – Bill Dubuque Feb 01 '15 at 16:31
  • @Megan I added some remarks about how it relates to the property of being an (integral) domain. – Bill Dubuque Feb 01 '15 at 17:16