A Galois group is a group of symmetries of a field extension: its elements are automorphisms of the larger field that leave every element of the smaller field fixed. For a polynomial, the term usually means the Galois group of its splitting field over its coefficient field. Such automorphisms permute the polynomial’s roots while preserving all algebraic relations among them. The group therefore records not merely which roots exist, but how those roots are related over the chosen base field.
Named after Évariste Galois, Galois groups provide the central objects of Galois theory. Their classical importance lies in translating questions about equations into questions about groups: a polynomial is solvable by radicals precisely when its Galois group satisfies an appropriate solvability condition. Their modern significance extends far beyond formulas for roots. Finite and infinite Galois groups organize intermediate fields, encode the behavior of prime numbers in extensions, and act on geometric objects. They consequently connect group theory, number theory, and algebraic geometry.
Definition and the role of the base field
Let be an extension of fields. A -automorphism of is a bijection preserving addition and multiplication and satisfying for every . The set of these maps forms a group under composition:
For a Galois extension, this group is denoted . Some authors use the same notation for an arbitrary extension, although the fundamental correspondence between subgroups and intermediate fields requires additional hypotheses. The identity automorphism is the group identity, and the inverse of a field automorphism is again a field automorphism.
An algebraic extension is Galois when it is both normal and separable. Normality means that every irreducible polynomial over having one root in splits completely in . Separability means that the minimal polynomial of every element of has distinct roots in an algebraic closure. Every algebraic extension in characteristic zero is separable; inseparability becomes an issue in characteristic . A finite extension is Galois exactly when it is the splitting field of a separable polynomial over its base field.
For any finite extension,
where the degree is the dimension of as a vector space over . Equality holds exactly when the extension is Galois. This criterion separates the degree of an extension from the number of its symmetries. For example, has degree three but only the identity automorphism. Its other two conjugate roots are nonreal and do not belong to that field. Likewise, the purely inseparable extension has degree and trivial automorphism group.
The base field is indispensable to the definition. Complex conjugation is a nontrivial automorphism of , but it does not fix all of . More generally, enlarging the base field inside a fixed Galois extension replaces the original group by a subgroup. Thus a polynomial does not have a Galois group independently of its coefficient field: the same polynomial can exhibit different symmetry groups over , , a finite field, or an intermediate algebraic field.
Permutations of roots and elementary examples
Let be separable, and let be its splitting field, the smallest extension containing all its roots. If is a root and , then
Every automorphism therefore permutes the roots. Since those roots generate , their images determine the automorphism uniquely. For a degree- separable polynomial, this gives an injective homomorphism
where is the symmetric group. Relabeling the roots conjugates the embedded subgroup, so the particular subgroup presentation depends on labeling, whereas its conjugacy class does not.
Not every permutation of the roots necessarily arises. A valid permutation must preserve every polynomial relation over , including relations involving several roots simultaneously. This distinction explains why the Galois group is often much smaller than . The resulting group action is transitive precisely when is irreducible. More generally, its orbits correspond to the irreducible factors of a separable polynomial over the base field.
For over , the splitting field is . There are two automorphisms: the identity and the map sending to . Its group is the cyclic group . For the polynomial , the splitting field is , and the signs of the two square roots may be changed independently. The resulting group is , the Klein four-group. Its action on four roots is not transitive because the polynomial has two irreducible quadratic factors.
A more informative example is . Writing and , its splitting field is
The field has degree six over . An automorphism sends to and fixes ; complex conjugation fixes and sends to . They satisfy
Thus . The same polynomial over has Galois group , while over its splitting field is and its group is . These examples distinguish the roots, their ambient field, and the base field fixing the permitted symmetries.
The correspondence between subgroups and intermediate fields
The fundamental theorem of Galois theory states that, for a finite Galois extension with group , intermediate fields correspond bijectively to subgroups of . To an intermediate field it assigns
the automorphisms fixing . To a subgroup it assigns the fixed field
These constructions are inverse to one another. In particular, , while the fixed field of the identity subgroup is .
The correspondence reverses inclusion. More fixed elements mean fewer allowable automorphisms:
It also converts field degrees into group orders and indices:
Here the index counts cosets. Consequently, the lattice of subgroups supplies a complete description of all intermediate fields, including their possible degrees and containment relations. Field intersection corresponds to the subgroup generated by the associated subgroups, while the compositum of two intermediate fields corresponds to the intersection of their associated subgroups.
Normality has an especially important translation. The intermediate extension is Galois exactly when is a normal subgroup of . In that case restriction of automorphisms gives a surjective homomorphism
with kernel , and hence
Without normality, the full quotient group is not defined as a group, and not every automorphism of preserves . More precisely, the automorphism group of is isomorphic to , where is the subgroup normalizing .
For the splitting field of , the unique subgroup of order three is normal in ; its fixed field is , a quadratic Galois extension. The three order-two subgroups fix three degree-three fields generated respectively by , , and . None is normal over , reflecting the nonnormality of the corresponding subgroups. Thus the failure of a cubic field to contain all conjugates of one of its generators appears directly in the subgroup structure of .
Solvability by radicals and geometric construction
A polynomial is solvable by radicals when its roots can be obtained from the base field by repeatedly adjoining elements whose positive integer powers lie in the preceding field. Over a characteristic-zero field, a polynomial is solvable by radicals if and only if its Galois group is a solvable group. This is a structural criterion: it concerns the existence of radical expressions, not whether those expressions are short, convenient, or numerically stable.
A group is solvable if its derived series eventually reaches the identity. Beginning with , define
where the bracket denotes the subgroup generated by commutators. Equivalently, a solvable group admits a finite series with abelian successive quotients. The connection with radicals becomes visible after adjoining suitable roots of unity: extensions obtained by taking individual roots then have cyclic symmetry groups, and a tower of such steps produces a group assembled from abelian quotients. Conversely, the subgroup structure of a finite solvable Galois group can be used, with appropriate roots of unity adjoined, to construct a radical extension containing the required roots.
The groups , , and are solvable, accounting for the existence of general radical formulas through degree four. The group is not solvable because its derived subgroup is the nonabelian simple group . This explains the obstruction behind the Abel–Ruffini theorem: there is no general radical formula for equations of degree five. The obstruction persists for general equations of higher degree. It does not mean that every quintic is unsolvable; for example, is explicitly solvable by radicals and has a solvable Galois group.
A concrete nonsolvable quintic is . Modulo it is irreducible, so it is irreducible over . Modulo it factors as
The corresponding Frobenius cycle types yield a 5-cycle and an element of cycle type ; cubing the latter gives a transposition. A transitive permutation group of prime degree containing a transposition is the full symmetric group. Its Galois group is therefore , ruling out expressions for its roots by radicals.
Straightedge-and-compass construction imposes a stronger restriction. A real algebraic number is constructible precisely when it belongs to a tower of quadratic extensions of . Equivalently, its normal closure has Galois group of order a power of two. Merely having degree a power of two is necessary but not sufficient. These distinctions explain why radical solvability, geometric constructibility, and the algebraic degree of a number are related but different properties.
Finite fields, roots of unity, and abelian extensions
Galois groups over finite fields have an especially uniform description. For a prime power , every finite extension of is isomorphic to for some positive integer , and
A generator is the Frobenius automorphism . Its th power is the identity on , and no smaller positive power fixes every element. Finite fields are perfect, so their finite extensions are separable, and every finite extension is normal.
For every divisor , there is exactly one intermediate field of size . It is the fixed field of the subgroup generated by . Thus the subgroup lattice of a cyclic group becomes the divisibility lattice of finite subfields. If an irreducible polynomial over has degree and root , its roots are exactly
Frobenius therefore describes both the automorphism group and the conjugacy of polynomial roots. This explicit structure is central to finite-field algorithms and to arithmetic performed in extension fields.
Over , roots of unity produce another fundamental family. Let be a primitive th root of unity. The cyclotomic field is the splitting field of , and its automorphisms are exactly
Consequently,
and the extension degree is Euler’s totient . The multiplication of residue classes corresponds to composition of automorphisms, making modular arithmetic an explicit model of field symmetry.
Cyclotomic groups are abelian but not always cyclic. For instance, the field has group , whereas has group . Complex conjugation corresponds to the residue class ; its fixed field is the maximal real subfield when . The Kronecker–Weber theorem states that every finite abelian extension of is contained in a cyclotomic field. Generalizing the arithmetic description of abelian Galois groups over other number fields is a principal achievement of class field theory.
Determining groups through arithmetic and computation
Computing a Galois group usually means determining its permutation action on roots, rather than writing every field automorphism explicitly. The first steps are to factor the polynomial over the base field and test separability. Irreducibility establishes transitivity. The polynomial discriminant then supplies an important parity test. For a separable polynomial over a field of characteristic different from two,
Indeed, the product of pairwise root differences changes sign under odd permutations, while its square lies in the base field. A nonsquare discriminant therefore guarantees the presence of odd permutations.
Reduction modulo primes gives further constraints. Let be monic and separable, and let not divide its discriminant. If its reduction modulo factors into distinct irreducible factors of degrees , then its Galois group over contains an element with disjoint cycle lengths . An irreducible reduction produces an -cycle; a factorization into one linear and one irreducible quadratic factor produces a transposition in degree three. Several primes can quickly eliminate most candidate transitive subgroups.
This rule is an instance of the arithmetic of Frobenius elements. For a finite Galois extension of number fields and an unramified prime, choosing a prime above it determines a Frobenius element in a decomposition group. Changing that choice conjugates the element, so the extension naturally associates a conjugacy class to the original prime. Its action on residue fields is a power map. In a polynomial realization, its orbits on the roots explain the observed factor degrees.
The Chebotarev density theorem strengthens this connection: if is a conjugacy class in , the unramified primes with Frobenius class have density
It follows, for example, that primes splitting completely in the extension have density . Thus prime factorization patterns are not isolated coincidences; their frequencies reflect the distribution of conjugacy classes in the group.
Nevertheless, collecting factorization patterns alone need not distinguish every candidate group. Exact algorithms also use resolvent polynomials, invariant expressions in roots designed to detect membership in particular subgroups. A resolvent with a root in the base field can force the Galois group into a subgroup or one of its conjugates. Modern computer algebra systems combine modular arithmetic, subgroup databases, numerical root approximations, and exact certification. Numerical approximations assist with identifying expressions, but the final group determination depends on exact algebraic tests.
Infinite Galois groups and local arithmetic
For an infinite Galois extension , the automorphism group carries essential topological structure. The Krull topology has as a neighborhood basis of the identity the subgroups , where ranges over finite Galois subextensions. With this topology,
It is therefore a profinite group: an inverse limit of finite discrete groups, and in particular compact, Hausdorff, and totally disconnected. The topology records which automorphisms become indistinguishable when restricted to any fixed finite subextension.
The infinite fundamental theorem pairs intermediate fields with closed subgroups, not with all abstract subgroups. A nonclosed subgroup has the same fixed field as its closure, so omitting the topology would destroy injectivity of the correspondence. Closed normal subgroups correspond to intermediate extensions that are Galois over the base, and their quotient groups carry the corresponding quotient topology. Open subgroups correspond precisely to finite-degree intermediate fields.
Fixing a separable closure gives the absolute Galois group
It organizes all finite separable extensions of simultaneously. For a finite field, , the profinite completion of the integers, topologically generated by Frobenius. Over it is , and over an algebraically closed field it is trivial. The group , by contrast, has an exceptionally rich structure whose finite quotients include the Galois groups of all finite Galois extensions of .
The p-adic numbers provide local versions of this structure. For a finite Galois extension of local fields , reduction induces a surjection onto the Galois group of the residue-field extension:
The kernel is the inertia group. It measures ramification, the part of the extension not visible in the residue fields. Its wild inertia subgroup is a -group when the residue characteristic is ; the quotient by wild inertia describes tame ramification. For the absolute group, the quotient by inertia is .
Local groups also occur inside global ones. Choosing a prime of an algebraic closure of a number field determines a decomposition subgroup of its absolute Galois group, isomorphic to the absolute Galois group of the corresponding completion. Different choices give conjugate subgroups. This places local arithmetic—residue fields, ramification, and completions—within the symmetry group of a global field.
Geometric monodromy and Galois representations
Galois groups have a geometric interpretation when polynomial coefficients depend on parameters. Consider a polynomial over . Away from values where roots collide or the degree drops, its roots define a finite covering of the parameter space. Analytically continuing the roots around a closed loop permutes them. These permutations form a monodromy group, and for the associated connected algebraic covering this permutation group agrees with the Galois group of the splitting field over .
For the equation , a loop once around sends a chosen root to its product with . The resulting cyclic monodromy reflects the cyclic Galois group of . More complicated branched coverings produce nonabelian groups. Branch points describe where the covering ceases to be locally unramified, and the local permutations around them generate the global monodromy, subject to relations imposed by the topology of the base.
The étale fundamental group extends this correspondence to algebraic geometry over arbitrary fields. It classifies finite étale coverings in a way analogous to the classification of topological coverings by the ordinary fundamental group. For the spectrum of a field , its étale fundamental group is . For a geometrically connected variety , there is an exact sequence
This separates geometric covering symmetries from symmetries arising from the arithmetic of the coefficient field.
A second major connection is through Galois representations, continuous homomorphisms from Galois groups to groups of invertible matrices. For example, if is an elliptic curve and is prime to the characteristic, its -torsion points over a separable closure form a group isomorphic to . The Galois action gives
Passing compatibly through powers of a prime yields an action on the Tate module, a free rank-two module over .
Such representations make an enormous profinite group accessible through linear algebra. At a prime of good reduction with residue field of size , distinct from , an arithmetic Frobenius element acting on the elliptic curve’s Tate module has characteristic polynomial
Thus a matrix invariant of a Galois action is determined by counting points on a reduced curve. Relations between these representations and modular forms are central to modern arithmetic, including the proof of Fermat’s Last Theorem.
The inverse Galois problem and realization of groups
Ordinary Galois theory starts with a field extension and determines its group. The inverse Galois problem reverses the question: given a finite group and a field , does a finite Galois extension exist with ? The answer depends strongly on the base field. Over a finite field, only cyclic groups occur. Over , no nontrivial finite extensions occur. Over , it remains unknown whether every finite group occurs.
Every finite group can be realized if the base field is allowed to vary. One construction begins with a field and algebraically independent variables indexed by the elements of . The group acts faithfully on the rational function field by permuting variables through its regular action. If is the fixed field, then the rational function field is a finite Galois extension of with group . This demonstrates that the main difficulty over is not an abstract restriction on possible field symmetries, but the requirement of a specific arithmetic base.
A major result over , due to Igor Shafarevich, establishes that every finite solvable group occurs as a Galois group. Symmetric groups and alternating groups also have realizations, as do many families of finite simple groups. These successes do not establish the general conjecture: there is no known construction covering all finite groups over .
One important strategy first constructs an extension of the rational function field . If the algebraic closure of inside that extension is exactly , the extension is called regular. Hilbert’s irreducibility theorem then provides suitable rational specializations whose specialized extensions retain the intended Galois group. This turns a geometric construction involving a parameter into an arithmetic extension of .
Over , every finite group occurs, through the theory of branched coverings of the projective line and the Riemann existence theorem. Descending such a cover to , however, introduces additional arithmetic conditions. Methods involving prescribed conjugacy classes of branch permutations, especially rigidity methods, have produced many realizations. Related problems ask whether a given Galois extension can be enlarged so that its group fits a prescribed group extension. These embedding problems examine not just which finite groups occur, but how finite quotients of an absolute Galois group fit together.
Historical development of the concept
The origins of Galois groups lie in attempts to solve polynomial equations by explicit formulas. Sixteenth-century formulas for cubic and quartic equations showed that surprisingly elaborate radical expressions were possible. In 1770–1771, Joseph-Louis Lagrange studied how permutations of roots affect auxiliary expressions, now called resolvents. He identified the importance of permutation patterns in explaining the existing formulas, although he did not obtain a general criterion for radical solvability.
Paolo Ruffini published an attempted proof of the impossibility of a general radical formula for the quintic in 1799. Niels Henrik Abel established the impossibility result in 1824 and developed a fuller treatment afterward. This settled a central existence question, but did not yet provide the general group-theoretic classification of individual equations. Some quintics are solvable by radicals and others are not, so a theory was needed that could distinguish equations by their internal algebraic structure.
Galois’s work of 1830–1832 supplied that distinction. He associated an equation with a group of permutations preserving its algebraic relations and related radical solvability to a decomposition of that group. His memoir on solvability by radicals was submitted to the Paris Academy in 1831 but was not accepted in its submitted form. Joseph Liouville recognized the importance of the work and published Galois’s mathematical writings in 1846. The central insight subsequently became a foundation of abstract algebra: the apparent complexity of root formulas is governed by the structure of a symmetry group.
The modern presentation in terms of fields and automorphisms emerged gradually during the nineteenth century. Richard Dedekind and Heinrich Weber helped develop field-theoretic language, while subsequent treatments separated the theory from dependence on explicit formulas and numerical roots. Emil Artin’s twentieth-century formulation emphasized fixed fields and automorphism groups. In this framework, a finite group of automorphisms of a field determines a Galois extension over its fixed field, with extension degree equal to the group’s order.
During the twentieth century, Emmy Noether and others integrated Galois theory into a broader structural approach to algebra. Wolfgang Krull developed the topological treatment of infinite Galois groups, while class field theory organized abelian extensions through arithmetic data. Later developments in étale coverings and Galois representations expanded the subject from equations to geometric and cohomological structures. In present usage, the same term encompasses a concrete finite permutation group attached to a polynomial, a profinite absolute group attached to a field, and groups acting continuously on arithmetic invariants.