0

Prove if $n\in N$ is composite then there exists an equivalence class $a\bmod n$ that does not have a multiplicative inverse

I know that the if $n=j*k$ with $j,k<n$ that $j\bmod n$ or $k\bmod n$ probably don't have inverses, but I don't remember the theorem for it.

Arturo Magidin
  • 398,050
user8714896
  • 535
  • 3
  • 16
  • 1
    If $j^{-1}j\equiv1$ and $jk\equiv0$ then $k=1k\equiv j^{-1}j*k\equiv0$ – J. W. Tanner Nov 07 '19 at 20:10
  • $0$ is always an equivalence class without an inverse $\pmod N$ even if $N$ is prime. If $N$ is composite then you can find a non-zero equivalence class without an inverse. – lulu Nov 07 '19 at 20:25
  • You don't need "to remember the theorem for it" when you start with an example of a composite number $n$, say $n=4$. – Dietrich Burde Nov 07 '19 at 20:31
  • @J.W.Tanner I sort of see where you're going with your solution, but how does that prove that k has no inverse mod n? – user8714896 Nov 07 '19 at 20:48
  • 1
    If $k\equiv0$ then $k$ times anything is $0$, so nothing multiplies by $k$ to get $1$ ((unless $0=1$)) – J. W. Tanner Nov 07 '19 at 21:03
  • @J.W.Tanner. Brilliant! Thanks! – user8714896 Nov 07 '19 at 21:04
  • @J.W.Tanner I think there also needs to be something written for when both j and k don't have inverses like n = 4 and j,k=2 then $2*1(mod)4 \equiv 2$, which doesn't have an inverse but mod4 doesn't isn't equivalent to 0 – user8714896 Nov 07 '19 at 21:41
  • 2
    We’re asked to prove there’s an equivalence class without an inverse. If both $j$ and $k$ don’t have inverses, then we’re done – J. W. Tanner Nov 07 '19 at 21:49

1 Answers1

0

Per lulu's comment, $0$ is an equivalence class without a multiplicative inverse $\pmod n$.

Here is a demonstration that there is a non-zero equivalence class

without a multiplicative inverse $\pmod n$ when $n$ is composite.

Say $n=jk$ with $j$ and $k$ strictly between $0$ and $n$.

Then $jk\equiv0\pmod n$ but $j,k\not\equiv0\pmod n$.

Assume, aiming for contradiction, there is $k^{-1}$ such that $kk^{-1}\equiv1\pmod n$.

Then $0\equiv0\times k^{-1}\equiv (jk)k^{-1}\equiv j(kk^{-1})\equiv j\times1 \equiv j.$

This contradiction shows that $k$ is a non-zero equivalence class

without a multiplicative inverse $\pmod n$.

Swapping the role of $j$ and $k$ would show that $j$ is also not invertible $\pmod n$.

J. W. Tanner
  • 60,406