8

The title is self-explanatory. I know it's irreducible but is it a prime? How to prove these primality and/or irreducibility of $1+\sqrt{5}$.

Can you just briefly state how a prime is defined under $\mathbb{Z}[{\sqrt{5}}]$? I know that it will only be divisible by its associates and the unity. But please tell about the norm conditions and other properties or the complete rigorous definition of primes.

Mathejunior
  • 3,354

5 Answers5

6

Definitions are indeed very important. Two answerers on this one question have been led astray thinking you're asking about $\mathbb{Z}[\sqrt{-5}]$, which is a big but understandable mistake, one that many askers on this site have made.

There are at least two different things you might mean by $\mathbb{Z}[\sqrt{5}]$. One is the ring of algebraic integers of the form $a + b \sqrt{5}$, with $a, b \in \mathbb{Z}$; this includes numbers like $1 + \sqrt{5}$ and $-3 + 2 \sqrt{5}$. Or you could mean the ring of all algebraic integers in $\mathbb{Q}(\sqrt{5})$, which could potentially include numbers like $$\frac{3}{4} + \frac{7 \sqrt{5}}{3},$$ but I'm getting ahead of myself.

If you mean the former, the answer is clear: $1 + \sqrt{5}$ is not prime. Since $(1 + \sqrt{5})(1 - \sqrt{5}) = -4$ but $1 + \sqrt{5}$ is not a divisor of $-2$ nor $2$, it follows that $1 + \sqrt{5}$ is irreducible but not prime (I'm taking your word for it that the number is irreducible).

If you mean the latter, then you have overlooked a couple of things, because $1 + \sqrt{5}$ is actually composite. Verify that $$\frac{1 + \sqrt{5}}{2}$$ is an algebraic integer with minimal polynomial $x^2 - x - 1$ (ever hear of a thing called the golden ratio?) Then $$2 \left(\frac{1 + \sqrt{5}}{2}\right) = 1 + \sqrt{5}.$$

Mr. Brooks
  • 1,156
5

In a ring $R$, we say a non-unit, nonzero $p \in R$ is prime if for any $a, b \in R \setminus \{ 0 \}$ such that $p | ab$, we have either $p | a$ or $p | b$. Mathmo123 showed $1 + \sqrt{5}$ wasn't prime.

AJY
  • 9,086
  • You need $p$ to be non-zero and not a unit. Do we need $R$ to be a domain? – Mathmo123 Nov 17 '17 at 13:53
  • @Mathmo123 I don't think we need $p$ a non-unit – AJY Nov 17 '17 at 13:55
  • So $1\in \mathbb Z$ is a prime for you? – Mathmo123 Nov 17 '17 at 13:55
  • Sure, why not. $R / (1)$ is an integral domain. – AJY Nov 17 '17 at 13:57
  • I guess this comes down to conventions. We don't want $1$ to be a prime, and we don't want $(1)$ to be a prime ideal for all manner of reasons. As such, we often define an integral domain to be a non-zero ring (see Wikipedia for instance). – Mathmo123 Nov 17 '17 at 14:00
  • Okay. Editing accordingly. – AJY Nov 17 '17 at 14:00
  • @Mathmo123 though I don't see why, given the OP's question was in part what primality means in an arbitrary ring, you couldn't have put it in your own answer. – AJY Nov 17 '17 at 14:02
  • I added it in the comments. My reasoning was that despite being equivalent to the usual definition, this really is the "wrong definition" in that it completely misses the point about why primes are important. That, and because thinking about divisibility in an arbitrary ring makes my head hurt. I'd rather think about ideals, and the OP should learn to do that too if they want to learn about number theory. – Mathmo123 Nov 17 '17 at 14:06
  • @Mathmo123 to second your comment, outside of the ufd context, the only "unique prime decomposition" I can think of in an arbitrary commutative ring is primary decomposition, which deals exclusively with ideals. – Andres Mejia Nov 17 '17 at 22:45
  • Sorry to pile on the criticism, but could the non-primality of $1 + \sqrt{-5}$ change in the future? I would have written "isn't" rather than "wasn't." Still, upvote from me. – Robert Soupe Nov 18 '17 at 19:16
  • @RobertSoupe Nope. The US Congress decreed $1 + \sqrt{5}$ to be prime between 123's answer and mine. I don't make the rules. – AJY Nov 18 '17 at 19:20
4

No, it's not prime. Compare $6$. Is $6$ prime in $\textbf Z$? If it was, we'd see that whenever $6 \mid ab$, either $6 \mid a$ or $6 \mid b$. Yet $6 \mid 3 \times 4$ but $6 \nmid 3$ nor $4$.

Likewise, in $\textbf Z[\sqrt{-5}$, $(1 + \sqrt{-5}) \mid 2 \times 3$ but...

David R.
  • 1,316
3

One would usually define a non-zero element $a$ of a ring $R$ to be prime if the ideal it generates is a prime ideal.

Under this definition, $1+\sqrt 5$ is not prime, since $4 =2\times2= (1+\sqrt5)(\sqrt5-1) \in \langle1+\sqrt 5\rangle$, but $2\notin\langle1+\sqrt 5\rangle$.

You should note that $\mathbb Z[\sqrt 5]$ is not the ring of integers of $\mathbb Q(\sqrt 5)$. Rather $\mathbb Z[\frac{1+\sqrt 5}2]$ is, and $1+\sqrt 5$ is prime in this ring.

Mathmo123
  • 23,718
  • I think you're mistaken. It's $4=(1+\sqrt{5})(1-\sqrt{5})$ – Mathejunior Nov 17 '17 at 13:35
  • @Mathbg thanks! I've corrected the mistake, but the point still stands – Mathmo123 Nov 17 '17 at 13:36
  • Unfortunately, I don't have any idea about ideals. I think I got the idea from one of the above comments. If possible, please suggest me where to learn about integer rings. I only have a vague idea from some 2-3 page handwritten note I got from a friend. – Mathejunior Nov 17 '17 at 13:44
  • 1
    @Mathbg Before learning about integer rings, you should first learn about rings! I don't have a good source off the top of my head, but look for some lecture notes from an undergraduate course. In particular, you will need to know about rings, ideals, PIDs, UFDs, etc. – Mathmo123 Nov 17 '17 at 13:47
  • @Mathbg You could equivalently define $a\in R$ as being prime if $a\ne 0$, $a$ is not a unit and whenever $b,c\in R$ and $a \mid bc$, then $a\mid b$ or $a \mid c$. Here divisibility means divisibility in $R$. But this doesn't really get to the heart of the issue. – Mathmo123 Nov 17 '17 at 13:48
  • According to that, okay, it's not a prime. Is it reducible? I think it is, since it's not a prime. – Mathejunior Nov 17 '17 at 13:56
  • This is where the ring theory comes in. It is irreducible but it is not prime. A ring in which every irreducible is prime is called a UFD (unique factorisation domain). In $\mathbb Z$ these concepts coincide, and the usual definition of a prime that we learn at school is actually the definition of being irreducible. – Mathmo123 Nov 17 '17 at 13:57
  • 2
    I don't know, learning about rings before learning about integer rings could be confusing rather than enlightening. Kind of like how ideals are made out to be much more difficult than they actually are by talking about ideals in general before talking about principal ideals. – Robert Soupe Nov 18 '17 at 19:12
  • @RobertSoupe If you want to just view integer rings as nice examples of rings, then fine, although to define them, you need to know what "integral closure" means. However, if you want to start learning the number theory behind these rings (in particular, Dedekind domains, unique factorisation of ideals, class groups etc), then ring theory is very much a prerequisite. – Mathmo123 Nov 19 '17 at 14:04
  • 1
    @Mathmo123 Math dilettantes like myself are different from professional mathematicians like you. Something might be a prerequisite to you but to me is just some boring, seemingly unmotivated thing that I can hold off on learning about. Such gaps in my knowledge have no consequence on my practical math needs (percentages and such). – Robert Soupe Nov 19 '17 at 18:53
2

We need to agree on definitions before giving a meaningful answer. If you asked Ethan Bolker fifty years ago, he'd say yes, $1 + \sqrt{-5}$ is indeed prime in $\mathbb Z[\sqrt{-5}]$. It's there in black and white on page 102 of his classic Elementary Number Theory: An Algebraic Approach:

note that $1 + \sqrt{-5}$ is prime in $\textbf A(-5)$ because its norm is 6, which is prime in $\textbf B(-5)$

Dover reprinted the book in 2007, and Bolker wrote a new preface and corrected some errors, but also suggested he would have written the book very differently today.

Fifty years ago, he was using "prime" to mean what today most of us call "irreducible." That is, if $p$, which is not a unit, is divisible only by units and associates, it is irreducible. For example, 7 is irreducible in $\mathbb Z$ since it's only divisible by $-7, -1, 1, 7$.

In infinite domains, we can narrow down where to look by looking only at numbers that are in some sense "smaller" than $p$. In imaginary rings like $\mathbb Z[\sqrt{-5}]$, we can take "smaller" to mean closer to 0, or having a smaller norm.

To call it prime, we now state the additional requirement that whenever $p \mid ab$, $p$ must also divide one of $a$ or $b$, maybe both. In domains like $\mathbb Z$ or $\mathbb Z[\sqrt{-2}]$, the additional requirement is irrelevant: all irreducibles are prime.

But in domains like $\mathbb Z[\sqrt{-5}]$, it makes for a very important distinction. To go back to our example of 7, we see that it is irreducible but not prime in $\mathbb Z[\sqrt{-5}]$, since $7 \mid (3 - \sqrt{-5})(3 + \sqrt{-5})$ but $7 \nmid (3 - \sqrt{-5})$ and $7 \nmid (3 + \sqrt{-5})$ either.

Likewise, we see that $1 + \sqrt{-5}$ is irreducible, since it's divisible only by $-1 - \sqrt{-5}, -1, 1, 1 + \sqrt{-5}$. But it's not prime, because $(1 + \sqrt{-5}) \mid 2 \times 3$ but $(1 + \sqrt{-5}) \nmid 2$, $(1 + \sqrt{-5}) \nmid 3$. And 2 and 3 are also irreducible but not prime in this domain.

And then the norm is a stronger indication of primality. Given a number $z \in \mathbb Z[\sqrt{-5}]$ with nonzero imaginary part, if $N(z)$ is prime in $\mathbb Z$, then $z$ is prime in $\mathbb Z[\sqrt{-5}]$. Quick exercise: verify $3 + 2 \sqrt{-5}$ is irreducible and prime.

As for a purely real positive odd number $p$ that is prime in $\mathbb Z$ and irreducible in $\mathbb Z[\sqrt{-5}]$, this is how you can tell tell if it's also prime: try to solve $x^2 \equiv p - 5 \pmod p$. If that has no solutions, then $p$ is indeed prime.

For example, with 7 we see that $3^2 \equiv 2 \pmod 7$ and $4^2 \equiv 2 \pmod 7$, so 7 is not prime, readily leading us to find that $N(3 + \sqrt{-5}) = 14$ and $N(4 + \sqrt{-5}) = 21$ (there are more, of course, but these are enough to make the point that 7 is irreducible but not prime). On the other hand, 11 is irreducible and prime since $x^2 \equiv 6 \pmod{11}$ has no solutions.


These remarks can be generalized to other imaginary quadratic integer rings. For instance, you want to see if $x^2 \equiv p - 10 \pmod p$ has solutions to determine primality in $\mathbb Z[\sqrt{-10}]$.

One more thing: it might be worthwhile to say that 5 is neither irreducible nor prime in $\mathbb Z[\sqrt{-5}]$, since $(-1)(\sqrt{-5})^2 = 5$.

Robert Soupe
  • 14,999
  • 2
    I'm surprised and pleased to see the reference to my book. Thanks. You;ve made the distinction between prime and irreducible clear - noting that I didn't. One small point about this particular answer: the OP asked about $\mathbb{Z}[5]$, not $\mathbb{Z}[-5]$. – Ethan Bolker Nov 18 '17 at 19:15
  • 1
    You're right, that's actually a big point. So now I wonder if he was thinking about $\mathbb Z[\phi]$ ($\phi$ being the golden ratio). – Robert Soupe Nov 18 '17 at 19:28
  • An understandable but huge mistake. – Mr. Brooks Nov 18 '17 at 22:17