Overview

  • Does every finite group occur over the rational numbers?
  • How does a regular cover produce number fields?
  • What is the difference between a realization and a generic polynomial?
  • Which groups are known to occur?

Four realization questions

Fix a field KK and a finite group GG.

  • Inverse Galois problem. Does some finite Galois extension L/KL/K have group GG?
  • Regular inverse Galois problem. Does some finite Galois extension E/K(T)E/K(T) have group GG and satisfy E∩K‾=KE\cap\overline K=K?
  • Parametric polynomial. Is there one polynomial family whose specializations give every GG-extension of KK as a splitting field?
  • Generic polynomial. Is there such a family that works over every extension field K′/KK'/K?

These are different requirements. The word “generic” in generic Galois group only refers to a group over a function field; it does not assert that the polynomial parametrizes every extension.

References 2

Regular extensions and specialization

Let KK be a number field and let E/K(T)E/K(T) be a regular Galois extension with nontrivial finite group GG. Hilbert irreducibility produces specializations with group GG. More precisely, for every finite extension L/KL/K, one can choose a specialization field linearly disjoint from LL. Repeating this construction gives infinitely many linearly disjoint GG-extensions.

Regularity is essential for this conclusion. If E=L(T)E=L(T) is a constant extension, every unramified specialization gives the same field LL.

One approach is therefore to construct a regular cover first and then specialize. Critical points, invariant theory, branch cycles, rigidity, and embedding problems provide different constructions.

References 2

Examples

In characteristic different from 22, every quadratic extension of a field KK is K(a)K(\sqrt a) for some nonsquare a∈Ka\in K. Thus x2−Tx^2-T is a generic polynomial for C2C_2.

For a cyclic cubic over Q\mathbb Q, take

x3+x2−2x−1.x^3+x^2-2x-1.

Neither 11 nor −1-1 is a root, so the polynomial is irreducible. Its discriminant is 4949, hence its group is A3=C3A_3=C_3. A root is ζ7+ζ7−1\zeta_7+\zeta_7^{-1}, so its splitting field is the real cubic subfield of Q(ζ7)\mathbb Q(\zeta_7).

References 2

Abelian groups

There is a direct cyclotomic construction for finite abelian groups. Write G=Cm1×⋯×CmrG=C_{m_1}\times\cdots\times C_{m_r}. Choose distinct primes pi≡1(modmi)p_i\equiv1\pmod{m_i}. Then

Gal⁡(Q(ζp1⋯pr)/Q)≅∏i(Z/piZ)×\operatorname{Gal}(\mathbb Q(\zeta_{p_1\cdots p_r})/\mathbb Q)\cong\prod_i(\mathbb Z/p_i\mathbb Z)^\times

has GG as a quotient, because each factor is cyclic and mi∣pi−1m_i\mid p_i-1. The fixed field of the kernel realizes GG. This argument uses Dirichlet's theorem to choose the primes.

References 1

Known realizations

Hilbert proved that every symmetric group SnS_n and alternating group AnA_n occurs over Q\mathbb Q. Regular realizations are also known for these groups. Shafarevich's theorem realizes every finite solvable group over Q\mathbb Q. Many finite simple groups and their extensions have also been realized, but no theorem realizes every finite group over Q\mathbb Q.

The methods differ. Solvable realizations use embedding problems and arithmetic of number fields. Many nonsolvable realizations come from regular covers and rigidity. A list of computed polynomials supplies examples for particular groups; it does not settle the unrestricted problem.

References 3

The degree of a realizing polynomial

A Galois extension L/KL/K with group GG can be described by many polynomials. If H≤GH\le G and α\alpha generates LH/KL^H/K, then the minimal polynomial of α\alpha has degree [G:H][G:H]. Its normal closure is LL precisely when

⋂g∈GgHg−1=1.\bigcap_{g\in G}gHg^{-1}=1.

This intersection is the core of HH. Equivalently, the action of GG on the cosets of HH must be faithful. Thus a degree-nn realization asks for a faithful transitive action of degree nn, in addition to the abstract group.

References 2

Parametrization

Realizing a group once need not give a generic polynomial. For example, C8C_8 occurs over Q\mathbb Q, but has no generic polynomial over Q\mathbb Q, even with several parameters. For one parameter, the nontrivial groups admitting a generic polynomial over Q\mathbb Q are exactly C2,C3,S3C_2,C_3,S_3. The article Generic polynomials and invariant fields gives the definitions, examples, and precise classification.

References 1

Questions

OpenSpecialization and inverse Galois

Does every finite group occur as a Galois group over Q?

Scope. Finite Galois extensions of Q; no fixed degree bound

This is the classical inverse Galois problem. The cited monographs describe known realizations. For a specified group, one can also ask for a realization in a prescribed transitive action.

References 1
Discuss this question →
OpenSpecialization and inverse Galois

Does every finite group admit a regular realization over Q(T)?

Scope. Regular finite Galois extensions of Q(T)

This is Serre’s Conjecture 4.1.1. A positive answer would imply the ordinary inverse Galois problem by specialization. The conjecture concerns finite groups, not arbitrary profinite groups.

References 1
Discuss this question →

Related topics

References

  1. Generic Polynomials: Constructive Aspects of the Inverse Galois Problem

    Jensen, Christian U. and Ledet, Arne and Yui, Noriko. Mathematical Sciences Research Institute Publications 45, Cambridge University Press, 2002.

    Read the source
  2. On parametric and generic polynomials with one parameter

    Dèbes, Pierre and König, Joachim and Legrand, François and Neftin, Danny. Journal of Pure and Applied Algebra, 225(10), Article 106717, 2021.

    Read the source
  3. Topics in Galois Theory

    Serre, Jean-Pierre. Edition 2. A K Peters/CRC Press, Research Notes in Mathematics 1, 2007.

    Read the source
  4. Inverse Galois Theory

    Malle, Gunter and Matzat, B. Heinrich. Edition 2. Springer Monographs in Mathematics, 2018.

    Read the source
  5. Galois groups of cubics and quartics (not in characteristic 2)

    Conrad, Keith. Author-hosted expository note.

    Read the source
  6. Explicit Galois realization of transitive groups of degree up to 15

    Klüners, Jürgen and Malle, Gunter. Journal of Symbolic Computation, 30(6), 675–716, 2000.

    Read the source