0

Let's start with the couple of definitions:

Let's $R$ be an integral domain:

(1): Let $a,b \in R$. We say that a divides b or $a \mid b$, if there exists $c \in R$, such that: $b = ac$.

(2): An element $p \in R$ is called a prime element, if the following implication holds:

$p \mid ab \implies p \mid a$ or $p \mid b$.

Let $R$ be an unique factorization domain:

(3): Polynomial $f \in R[x]$ is called primitive if all coefficients of f are coprime.

(4): Eisenstein's criterion: Let's $R$ be an integral domain and $f = \sum_{i=0}^{n} a_ix^i$ be a primitive polynomial from $R[x]$. If there exists prime element $p \in R$, such that $p \mid a_0$, $p \mid a_1$, $\ldots$, $p \mid a_{n-1}$ and $p^2 \nmid a_0$, then polynomial f is irreducible in $R[x]$.

We know that all polynomials in the form: $(x^n - 2)$, $n \in \mathbb{N}$ are irreducible in $\mathbb{Q}[x]$.

$\mathbb{Q}$ is unique factorization domain/integral domain and therefore we can apply Eisenstein's criterion.

Let's choose for example: $(x^2 - 2)$. I've seen examples of choosing prime element $p = 2$, saying that $p \mid 2$ but $p^2 \mid 2$, therefore $(x^2 - 2)$ is irreducible in $\mathbb{Q}[x]$ via Eisenstein's criterion.

However, we can write $p^2 = 4 = 2.2$, $2 \in \mathbb{Q}$, therefore $p^2 \mid 2$, so $p = 2$ cannot be used to check irreducibility in this case.

I would like to understand where I am making a mistake.

Thanks.

meerkat
  • 371
  • 2
    Use Eisenstein's criterion over $\mathbb Z$. Then use Gauss Lemma, which states that an irreducible polynomial over $\mathbb Z[x]$ is also irreducible over $\mathbb Q[x]$. – Deif Mar 10 '24 at 12:29
  • @Deif thanks! You are right, but I am missing one thing. How do I check that x^2 - 2 is primitive also in rational numbers? (might be trivial as leading coefficient is 1, just want to be precise). I guess I cannot use Euclidean algorithm when working with rational coefficients. – meerkat Mar 10 '24 at 13:49
  • Oh I guess the answer is this: "In a field every nonzero element is unit: this means that every nonzero polynomial is primitive." https://math.stackexchange.com/questions/2306176/how-to-determine-whether-a-polynomial-is-primitive-in-mathbbqx so should be ok. – meerkat Mar 10 '24 at 13:52
  • 1
    The argument Deif sketches doesn't require you to check that anything is primitive over the rationals. The steps are "it's irreducible over $\Bbb Z$ by Eisenstein (requiring primitivity over $\Bbb Z$), hence it's irreducible over $\Bbb Q$ by Gauss". However you are right that nonzero polynomials are always trivially primitive over a field. – Izaak van Dongen Mar 10 '24 at 13:56
  • @IzaakvanDongen yes thanks. You're right. From wiki: "A non-constant polynomial in Z[X] is irreducible in Z[X] if and only if it is both irreducible in Q[X] and primitive in Z[X]". So primitivity is only required in field Z[X]. – meerkat Mar 10 '24 at 14:10
  • 1
    We apply Eisenstein with $,R = \Bbb Z,,$ not $,R = \Bbb Q,$ (there are no primes $,p,$ in a field). $\ \ $ – Bill Dubuque Mar 10 '24 at 18:46

0 Answers0