1

How can we prove using basic notions of algebra, that there are no quotients of $ \mathbb{Z}[x] $ which are fields with 72 elements, without using for example the fact that a finite field has a power of a prime number as cardinality?

Thank you!

Rick88
  • 415
  • 5
    The fact that a finite field has a power of a prime as its cardinality is proven using only basic notions of algebra... – TheSilverDoe May 20 '21 at 16:31
  • Yep, this is quick: https://math.stackexchange.com/questions/72856/order-of-finite-fields-is-pn – morrowmh May 20 '21 at 16:37
  • Yes, but I would like a demonstration that did not make use of this notion. For example, I think it is sufficient to prove that the only maximal ideals are of type $(p,q(x))$, where $p$ is a prime integer and $q(x)$ is irreducible in $\mathbb{Z}_p(x)$, is this true? – Rick88 May 20 '21 at 17:05
  • What counts as basic notions of algebra? – lhf May 21 '21 at 16:37

1 Answers1

2

You can avoid linear algebra, though you'll be dancing carefully around it the whole time.

Suppose $\mathbb{Z}[x]/I$ is a field with 72 elements. Consider $J = \mathbb{Z} \cap I$. This is an ideal in $\mathbb{Z}$, so $J = n\mathbb{Z}$ for some $n$. The surjective map $\mathbb{Z}[x] \to \mathbb{Z}[x]/I$ descends to a surjection $\mathbb{Z}[x]/\langle n\rangle \to \mathbb{Z}/I$. Since $\mathbb{Z}[x]/\langle n\rangle \cong \mathbb{Z}/n\mathbb{Z}[x]$, we have a surjection $\mathbb{Z}/n\mathbb{Z}[x] \to \mathbb{Z}[x]/I$. If $n$ were composite, running the two factors through this map would force one of them to be zero, so $n=p$ is prime and $\mathbb{Z}/n\mathbb{Z} = \mathbb{F}_p$ is a field of prime order. Now $\mathbb{F}_p[x]$ is a PID, so the kernel of $\mathbb{F}_p[x] \to \mathbb{Z}[x]/I$ is some principal ideal $\langle f\rangle$. Hence we have that $\mathbb{F}_p[x]/\langle f\rangle$ is a field with 72 elements. We may suppose $f$ is monic of degree $d$.

What is the order of $\mathbb{F}_p[x]/\langle f\rangle$? Consider the polynomials of the form $\{c_0 + c_1 x + \cdots + c_{d-1} x^{d-1} : c_i \in \mathbb{F}_p\}$. There are $|\mathbb{F}_p|^d = p^d$ of these. None of these polynomials are equal in the quotient since their difference has too small of a degree to be divisible by $f$. Moreover, every polynomial of degree $d$ or higher can be reduced to one of lower degree since $f$ is monic of degree $d$. So these polynomials are representatives of the quotient, and $p^d = 72$. This is a contradiction since $72$ is not a power of a prime.

  • Perfect! Thank you! – Rick88 May 20 '21 at 17:34
  • If $n=ab$ were composite, then the product of the image of $a$ and $b$ in the field would be $0$, so at least one of the images would be 0, but then $a$ or $b$ would be in $J$, making $J$ larger than it actually is. – Joshua P. Swanson May 20 '21 at 17:54
  • @Rick88 Your question claims you seek a proof w/o use of knowledge of cardinality of finite fields, but that is exactly what is used (in fact re-proved) above. You should remove that criterion from your question if you really don't mean to require it – Bill Dubuque May 20 '21 at 20:53
  • @BillDubuque My answer does not use the fact that finite fields have prime-power order or re-prove it. It shows that any quotient of $\mathbb{Z}[x]$ which is a field has prime-power cardinality, which is a weaker statement. It does effectively re-prove the classification of maximal ideals in $\mathbb{Z}[x]$. – Joshua P. Swanson May 20 '21 at 21:46
  • @Joshua Your argument essentially repeats (most of) one of the common proofs about such cardinaslity. – Bill Dubuque May 20 '21 at 22:15
  • @BillDubuque That does not seem to be true. For instance, none of the answers to the linked question use quotients in polynomial rings, let alone the popular answers. They all use vector spaces or group theory. If you show that finite fields are characterized as splitting fields, then I suppose you can use the end of my argument to conclude the general case, but the whole point seemed to be to avoid field theory. If you disagree, I think you'll have to be more specific and provide some reference. – Joshua P. Swanson May 20 '21 at 23:12