I was wondering if there is any work on the number of distinct finite and infinite fields (upto field isomorphism) for a particular characteristic $p$.

In particular, questions of the form: If $W_p := \{F \;|\; F \text{ is a field and } \text{char}(F) = p\}$, and for any $F \neq F'$, it holds that $F \not\cong F'$. Then what is $|W_p|$?

For example, for characteristic $p = 0$, there are no finite fields. There are no fields with characteristic 1. (I'm defining fields to have $0 \neq 1$). For characteristic $p > 1$, there is only 1 finite field (upto isomorphism) for that characteristic.

So my question is two-fold:

  1. What do we know about infinite fields of characteristic $0$? Are there infinitely many of them? How about uncountably infinite? i.e. What is the cardinality of $W_0$?
  2. What do we know about fields of characteristic $p > 1$. There is only $1$ finite field but how many distinct infinite fields of char $p$ exist?

It seems to me that there should be at least $|\mathbb{R}|$ fields of char $0$ just from field extensions of $\mathbb{Q}$. But is the cardinality of $W_0$ larger than that of $\mathbb{R}$? If they are equal then can we construct a bijection?

For infinite fields with char $p > 1$, I've seen several explicit constructions (In particular, one example is the field of rational functions with coefficients in $\mathbb{F}_p$). I've only seen a handful examples though so I'm not sure there are an infinite number of infinite fields with char $p > 1$.

(Edit) I forgot about fields of order $p^k$ so clearly there are an infinite number of finite fields with char $p$.

  • 2
    $\begingroup$ If you don't limit the cardinality, there are class-many fields of characteristic $p$ - $W_p$ is not a set (neither is $W_0$). There are certainly infinitely many - an easy infinite family of countable ones is $\mathbb F_p$ adjoined $n$ indeterminates for each natural number $n$. $\endgroup$ – user208649 Aug 13 '20 at 18:50
  • 1
    $\begingroup$ A finite field of characteristic $p$ has cardinality $p^n$ for some positive integer $n$. Furthermore, for every prime $p$ and positive integer $n$, there is a unique (up to isomorphism) field of cardinality $p^n$. $\endgroup$ – Andreas Blass Aug 13 '20 at 18:54
  • $\begingroup$ @AndreasBlass I'm not taking of the cardinality of a field but rather the cardinality of the set of fields of char p $\endgroup$ – John White Aug 13 '20 at 18:55
  • 2
    $\begingroup$ Nevertheless, my comment answers your question for the case of finite fields, for it immediately gives that there are exactly $\aleph_0$ nonisomorphic finite fields of characteristic $p$. $\endgroup$ – Andreas Blass Aug 13 '20 at 18:56
  • 2
    $\begingroup$ As indicated in the answer by diracdeltafunk, there's a proper class of nonisomorphic infinite fields of any given characteristic. Instead of the Löwenheim-Skolem argument in that answer, you can start with the $p$-element field (or $\mathbb Q$) and adjoin any cardinal number of independent transcendental elements. The resulting fields are all nonisomorphic, and there are as many of them as there are cardinals, i.e., a proper class. $\endgroup$ – Andreas Blass Aug 13 '20 at 19:10

In fact, there are more isomorphism classes of fields than can fit in any set! Thus, $W_p$ has no well-defined cardinality.

By the Löwenheim–Skolem Theorem, there exists at least one field of each possible characteristic and each infinite cardinality. If you wish to avoid this kind of abstract model-theoretic argument, it's not too hard to actually construct fields of arbitrary characteristic and arbitrary infinite cardinalities (although not super obvious – let me know if you'd like me to expand on this).

For finite fields, your understanding has a gap: there are infinitely many finite fields of each prime characteristic $p$! The cardinality of the set of isomorphism classes of finite fields of characteristic $p$ is $\aleph_0$.

  • $\begingroup$ Woops. I forgot about fields with order $p^k$. Is there an explicit bijection that shows this? $\endgroup$ – John White Aug 13 '20 at 19:00
  • $\begingroup$ Given $k>0$ and prime $p$, all fields of order $p^k$ are isomorphic. $\endgroup$ – Somos Aug 13 '20 at 21:25
  • $\begingroup$ Yes, and to show that two finite fields of the same order are isomorphic, we just note that any finite field of order $p^k$ must be the splitting field of $x^{p^k} - x$ over $\mathbb{F}_p$. $\endgroup$ – diracdeltafunk Aug 13 '20 at 23:48

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.