4
$\begingroup$

It is possible to see euclidean rings with lattices. For instance

enter image description here

from : https://en.wikipedia.org/wiki/Gaussian_integer

or

enter image description here

from : https://en.wikipedia.org/wiki/Eisenstein_integer

my question is the following :

We love euclidean rings. They have lattices regular enough to create an euclidean function. When we lower hypothesis, we can easily fall and lose the property to be factorial. For instance, $ \mathbb Z [i \sqrt 5 ] $ isn't factorial. Would it be possible to plot the set of those integers and "see" why specific integers are irreducible but not prime ? Which is a necessary condition to be factorial.

I don't know how to plot graphs like the one I put above, and I'm even worse to interpret them. That's why I asking you if you could give a hand.

thanks a lot!

$\endgroup$
1
$\begingroup$

Let $\alpha\in\Bbb{C}$ be some quadratic algebraic integer and consider the ring $\Bbb{Z}[\alpha]$, and let $N$ denote the norm function on $\Bbb{Z}[\alpha]$. Then $\Bbb{Z}[\alpha]$ is Euclidean w.r.t. $N$ if and only if for all $x,y\in\Bbb{Z}[\alpha]$ with $y\neq0$ there exist $q,r\in\Bbb{Z}[\alpha]$ with $N(r)<N(y)$ such that $x=qy+r$. In $\Bbb{Q}(\alpha)$ we can express this as $$\frac xy-q=\frac ry,$$ where $N(\tfrac ry)<1$. This shows that $\Bbb{Z}[\alpha]$ is Euclidean if and only if for all $z\in\Bbb{Q}(\alpha)$ there exists $q\in\Bbb{Z}[\alpha]$ such that $N(z-q)<1$. Geometrically speaking, if and only if the unit discs centered at the lattice points of $\Bbb{Z}[\alpha]$ cover the plane.

And indeed, drawing unit discs around the lattice points in your pictures confirms that $\Bbb{Z}[i]$ and $\Bbb{Z}[\omega]$ are Euclidean domains. And for $\Bbb{Z}[\sqrt{-5}]$ you will see that the lattice points are too far apart in the vertical direction; drawing unit discs around the lattice points leaves narrow horizontal bands that are not covered. This means that the norm map is not a Euclidean function on $\Bbb{Z}[\sqrt{-5}]$.

Note that this does not imply that $\Bbb{Z}[\sqrt{-5}]$ is not Euclidean; there may still be some other Euclidean function on $\Bbb{Z}[\sqrt{-5}]$. It certainly does not imply that $\Bbb{Z}[\sqrt{-5}]$ is not factorial. It turns out that this isn't a factorial ring, but I doubt there is any geometrical way to see this.

The reason for my doubt is that for $\Bbb{Z}[\tfrac12(1+\sqrt{-19})]$ the same argument as for $\Bbb{Z}[\sqrt{-5}]$ shows that the norm is not a Euclidean function on this ring; note that $\tfrac12\sqrt{19}$ and $\sqrt{5}$ are very close to eachother. But $\Bbb{Z}[\tfrac12(1+\sqrt{-19})]$ is factorial!

$\endgroup$
  • $\begingroup$ Your answer is really instructive but you re not dealing with the core of my question, irreducibility and primality. Am I missing something? $\endgroup$ – Marine Galantin Apr 6 at 16:23
  • 1
    $\begingroup$ My answer is not conclusive. If I understand correctly, you ask for a geometrical way to see whether a ring $\Bbb{Z}[\sqrt{-d}]$ is factorial or not from such a picture, and perhaps even to visually identify elements that are irreducible but not prime. My last two paragraphs suggest that this is not possible; the pictures for $\Bbb{Z}[\sqrt{-19}]$ and $\Bbb{Z}[\sqrt{-5}]$ are very similar, but one ring is factorial and the other is not. $\endgroup$ – Inactive - Objecting Extremism Apr 6 at 17:16
  • $\begingroup$ In fact, there are only nine integers $d>0$ for which $\Bbb{Z}[\sqrt{-d}]$ is factorial; five of them are Euclidean, four of them are not. I doubt any picture is helpful. $\endgroup$ – Inactive - Objecting Extremism Apr 6 at 17:18

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.