Galois Extensions
Given a polynomial equation, when can we understand all of its symmetries by counting a single number? This is the concrete question that Galois extensions answer. A field extension carries a group of automorphisms fixing pointwise. In general this group can be far too small to see anything interesting. It may even be trivial. The extension is called Galois precisely when this group is as large as possible. This happens when . This equality holds exactly when is both normal and separable. This equality is not a curiosity. It is the hinge on which the entire structure of Galois theory turns. Once we know , we can ask a sharper question. Which subfields of correspond to which subgroups of ? The fixed-field characterization of Galois extensions states this precisely. is Galois if and only if is exactly the fixed field of , no larger. This is the precise statement that makes this correspondence a genuine bijection, rather than a rough analogy. The big picture is a translation device. Field-theoretic questions become group-theoretic questions once we know we have a Galois extension. These questions include: does this polynomial factor, is this extension solvable by radicals, what are the intermediate fields. They translate into: subgroup structure, normal subgroups, solvability of a finite group. Every other theorem in the subject presupposes this equality and this fixed-field statement. This includes the Fundamental Theorem, the classification of cyclotomic and Kummer extensions, and the unsolvability of the quintic. This article proves, completely and from first principles, the equivalence of five conditions that define a Galois extension. The two headline results are proved in full: the counting equality and the fixed-field characterization. It assumes familiarity with Normal Extensions, Separable Extensions, and The Galois Group of an Extension. It also sets up, without yet proving, the correspondence developed in The Fundamental Theorem of Galois Theory.
- A finite extension is Galois iff it is normal and separable.
- Equivalently: — the automorphism group is exactly as large as degree allows.
- Equivalently: the fixed field of is precisely , with no room to spare.
- Equivalently: is a splitting field over of some separable polynomial.
- When Galois, we write , the Galois group of the extension.
- Key idea: normality plus separability is exactly the condition that makes automorphism-counting saturate the degree, turning symmetry into a faithful mirror of the field extension.
Where this lives today
The equivalence proved here is the mechanism behind the Fundamental Theorem of Galois Theory — see The Fundamental Theorem of Galois Theory — which upgrades the counting equality into an order-reversing bijection between subgroups of and intermediate fields. Galois extensions of generated by roots of unity, the cyclotomic extensions, are a central family whose Galois groups are computed directly (see Cyclotomic Extensions) using exactly the separable-splitting-field criterion proved above. In algebraic number theory, Galois extensions of number fields organize ramification and decomposition of primes, with the Galois group replaced by decomposition and inertia subgroups. In cryptography, the Galois theory of finite fields — as in Example 2 — underlies the arithmetic of used in elliptic-curve and pairing-based cryptosystems, where Frobenius automorphisms provide fast field arithmetic. In coding theory, Galois groups of finite field extensions control the structure of BCH and Reed–Solomon codes. In computer algebra, algorithms for polynomial factorization and for computing Galois groups of specific integer polynomials rely on precisely detecting normality and separability as in this article. In differential algebra, an analogous "differential Galois theory" replaces finite automorphism groups by algebraic groups classifying symmetries of solutions to linear differential equations.
:::quiz[If is finite and , which further conclusion is justified?]
- need not be separable
- is automatically both normal and separable ✓
- is normal but its separability cannot be determined
- must be algebraically closed
:::quiz[Why does fail to be Galois over ?]
- Because is inseparable over
- Because is not finite
- Because the minimal polynomial does not split inside ✓
- Because is not a perfect field
:::faq[Why does require BOTH normality and separability, rather than just one?] Separability alone guarantees that the number of embeddings over equals (this is the content of the Lemma). But an embedding need not be an automorphism of — its image might land partly outside , inside the ambient algebraic closure. Normality is exactly the condition that forces every such embedding's image to be contained in (hence equal to, by dimension count) itself, converting embeddings into automorphisms. Drop normality and you may have many embeddings but few automorphisms (Example 1: embeddings of , only an automorphism). Drop separability and even the embedding count falls short of , so no repair is possible (the counterexample). Both hypotheses attack different halves of the same inequality chain .
Ask the tutor: Can you walk through why the embedding-count lemma needs the extension-of-isomorphisms theorem, and what would go wrong if weren't algebraically closed?
Let be a field extension, and assume it is algebraic. Let
For a subgroup (or any subset closed under the group operations is not needed — any subgroup) , define the fixed field
The extension is called a Galois extension if it is finite and
When is Galois we write in place of , and call it the Galois group of the extension. (This is the fixed-field definition; the theorem below shows it is equivalent to " finite, normal, and separable," and to ".")
Intuition
Think of as measuring how much "room to move" the elements of have while the ground field stays perfectly still. If is not normal, some roots of minimal polynomials get left behind in an extension field, so automorphisms of can't reach them — the group is too small. If is not separable, a minimal polynomial has a repeated root, and any automorphism is forced to send that root to itself (there's only "one root's worth" of image to hit even though the polynomial has higher degree) — again the group is too small. Galois-ness says: neither obstruction occurs, so every root of every minimal polynomial is both present in and distinct, giving the automorphism group the maximum conceivable size, . The fixed-field statement is the same idea viewed from the other side. If were a proper subfield of , there would be "extra" elements of that every automorphism is powerless to move — symmetries blind to structure that isn't really there in . Galois-ness rules this out: the group sees everything that distinguishes from , and fixes nothing more than itself.
Let be a finite field extension. The following are equivalent: 1. is normal and separable. 2. . 3. is the splitting field over of some separable polynomial . 4. (i.e. is Galois, per the definition above). Consequently, "Galois," "normal and separable," and "" are interchangeable descriptions of a finite extension.
Let be fields with finite, and let be an algebraic closure of containing . Let be the inclusion. The number of field embeddings extending , call it , satisfies: (a) If for a single algebraic element with minimal polynomial , then equals the number of distinct roots of in . (b) In general, , with equality if and only if has no repeated roots for every — equivalently, iff is separable. (c) (Multiplicativity) For a tower , , where each embedding of is first extended to and then extended further to in ways (this count is independent of which extension of is chosen, by an isomorphism-extension argument).
(a) An embedding extending is determined by , since every element of is a polynomial in with coefficients in . Since fixes and is a ring homomorphism, applying to gives (coefficients of lie in , hence are fixed). So must be a root of in . Conversely, for any root of , the map , , is a well-defined field homomorphism (since , this factors through the quotient by the ideal ), and composing with the isomorphism gives an embedding sending . This is a bijection between embeddings and roots, proving (a).
(b) Induct on . If , , , trivial. Otherwise pick , so . By (a), (number of distinct roots of ) , with equality iff is separable (has no repeated root in ; note irreducible over , and separability of doesn't depend on which algebraically closed field we count roots in, since a repeated root of is detected by over itself). By induction on the tower , . Each of the embeddings of extends to some embedding of (since is algebraically closed and is algebraic, an isomorphism-extension theorem guarantees this), and for each such extension the number of further extensions to is by definition, independent of the chosen extension because any two embeddings of into over differ by an automorphism of over (both being generated by roots of the same ), and post-composing with such an automorphism permutes the extensions to bijectively. Hence
by the tower law. Equality holds iff both factors are equalities: separable for every choice of generator of a simple subextension in the chain, and by induction this happens for the whole tower iff every element of has a separable minimal polynomial over the relevant intermediate field. Chasing through the induction (or invoking that separability of means every has separable minimal polynomial over , and using that minimal polynomials over intermediate fields divide those over , hence share the separability property, since -separability of forces its irreducible factor over any extension field to remain separable) shows equality in the whole chain holds if and only if is separable. This proves (b).
(c) This is exactly the multiplicative bookkeeping described in the equality case of the argument in (b): every embedding of over extends to , and the extensions of that embedding to number regardless of which extension to was used, by the same automorphism-of- argument. Summing over the embeddings of gives .
∎Proof of the Theorem (equivalence of (1)–(4)). We prove , and separately , closing the cycle. Throughout, fix an algebraic closure of containing (possible since is finite, hence algebraic).
. Assume is normal and separable. By the Lemma (b), since is separable, , where counts embeddings over . We claim every such embedding is in fact an automorphism of , i.e. its image is itself; this converts the embedding count into an automorphism count. Let be an -embedding. For any with minimal polynomial , the argument in Lemma (a) shows is a root of in . Since is normal, splits completely in (this is the defining property of a normal extension applied to : the minimal polynomial of every element of splits in ). Hence every root of in already lies in , so . As was arbitrary, . Since is an injective -linear map (it is a field homomorphism, automatically injective) on the finite-dimensional -vector space , and , dimension-counting forces : an injective linear self-map of a finite-dimensional vector space is surjective. Thus restricts to an automorphism of fixing , i.e. . Conversely every element of is in particular an -embedding (compose with inclusion ), and distinct automorphisms give distinct embeddings. So the embeddings over are in bijection with , giving
. Assume . First we show is separable. By Lemma (b), always, with equality iff separable. But always (every automorphism of over , followed by inclusion into , is one of the embeddings, and distinct automorphisms give distinct embeddings — this uses no normality, just injectivity of the map "automorphism embedding"). Hence
forcing , so by Lemma (b), is separable. Next we show is normal, and identify a defining separable polynomial. Since is finite, write (any finite extension is finitely generated by algebraic elements — e.g. an -basis of ). Let be the minimal polynomial of , and set after removing duplicate irreducible factors so that has no repeated irreducible factors; note each is separable (as a factor of the polynomial defining a separable extension — indeed and separable means every element, in particular , has separable minimal polynomial) and distinct irreducible polynomials share no roots, so is a separable polynomial (no repeated roots in ). We claim is the splitting field of over . Since each , and we must show all roots of each lie in (splitting), consider any automorphism ; it permutes the roots of that lie in among themselves (as in the proof of (a) (b), is again a root of , and automatically since maps ). We must further show all roots of (not just those already known in ) lie in . Consider the orbit ; this is a finite subset of roots of lying in . We show is all the roots of in , by an embedding-counting argument: extend and use that . Every root of in gives an -embedding , (Lemma (a)); since has embeddings into total, and by the multiplicativity of Lemma (c) these decompose as extensions of the -many embeddings of (one for each root of , since separable means all roots are distinct and give distinct embeddings) — each extending in exactly ways. In particular for every root of , there is at least one embedding over with . But we showed above (in 's argument, which only used normality — so we cannot yet invoke it circularly here) — instead, argue directly: was already established in this direction (both equal ), meaning every -embedding has image itself (else the images would be embeddings not coming from automorphisms, but we've shown the count of automorphisms already saturates the total embedding count , so no embedding can have image other than , since automorphisms are automatically embeddings and there is no room for extra ones). Hence , so . Thus every root of lies in : splits in . As this holds for each , splits in , and is generated over by the 's, which are among the roots of ; hence is generated over by roots of and splits in , so is precisely the splitting field of over . This proves (3), with separable as shown.
. Assume is the splitting field over of a separable polynomial . That is normal is a standard fact about splitting fields (a splitting field of any polynomial is a normal extension of the base field — every splitting field is normal because any -embedding of into must send roots of to roots of , hence stays within the subfield generated by those roots, which is itself, forcing every embedding to be an automorphism; equivalently, minimal polynomials of elements of , being factors of powers of up to the structure of , split in ). We take this standard characterization of splitting fields as previously established under Normal Extensions and do not reprove it here. It remains to prove is separable, i.e. every has separable minimal polynomial over . Since is generated over by the roots of (with , all distinct since is separable), it suffices to show that any subfield generated over by roots of a separable polynomial is separable over ; this again is a standard closure property of separable extensions (separability is preserved under composita and subextensions generated by separable elements), which we invoke from Separable Extensions. Hence is separable, completing . This closes the cycle .
. Assume ; by the cycle just proved this is equivalent to (1), so is normal and separable. Let , the fixed field; clearly since every fixes by definition of . We must show , i.e. . Regard as an extension of the intermediate field ; every automorphism in fixes pointwise by definition of fixed field, so . Also every automorphism fixing certainly fixes the smaller field , so . Hence , of size . Now is normal (since is normal: every has minimal polynomial over splitting in ; its minimal polynomial over the larger field divides this splitting polynomial's factor, hence also splits in — a standard fact that normality persists over intermediate fields) and separable (a subextension of a separable extension is separable, since minimal polynomials over divide those over , and divisors of separable polynomials are separable). So by already proved, . Combining with gives
Since and (tower law) with finite and nonzero, we get , i.e. . This proves (4).
. Assume . We show is separable, then normal. Separability: Let be arbitrary, and let be its finite orbit under (finite because is a finite group — it injects into the set of embeddings , which is finite of size by Lemma (b), as is finite). List the distinct elements of as , and form
Every permutes (since maps the orbit of to the orbit of , which is the same orbit, as is a group acting on itself's orbits — more precisely, composing with any with gives , still in the orbit of ). Hence permutes the roots , so it permutes the linear factors of , and therefore fixes each coefficient of (symmetric functions of the are permutation-invariant). This holds for every , so every coefficient of lies in by hypothesis (4). Thus , , and has distinct roots by construction (they were listed as distinct). Since the minimal polynomial of over divides (as vanishes at ), and has no repeated roots, has no repeated roots either. Hence is separable over . As was arbitrary, is separable. Normality: With the same notation, we show , which shows splits in (its roots are the ), giving normality since was arbitrary. We know . Conversely, the identity map shows is a root of ; and for each , since for some , applying to (with fixed coefficient-wise by ) gives . So every is a root of , meaning divides (as has all distinct roots , each simple since... to be careful: requires the to be roots of with at least their multiplicity in , which is each; since all are distinct roots of , divides ). Combined with and both monic, . Since splits completely in by construction (its roots are ), splits in . As was arbitrary, every element of has minimal polynomial splitting in : is normal. This proves , closing all implications and establishing the full equivalence of (1), (2), (3), (4).
∎Worked examples
Example 1: A genuinely non-Galois extension, diagnosed and repaired. Let and let , and set . The minimal polynomial of over is (irreducible by Eisenstein at ), so . Is Galois? Check normality: the other roots of are and where is a primitive cube root of unity. These are complex (non-real) numbers, while . So does not split in : is not normal. Directly counting automorphisms confirms this. Any must send a root of lying in (by Lemma (a)'s argument, since is a root of , and trivially as is an automorphism of ). The only real root of is itself (the other two are genuinely complex, since and ). Since , the only possible image is , forcing (as is determined by its value on the generator ). So , giving . Consistently, the fixed field of is all of , not : . All four conditions of the theorem fail simultaneously, exactly as the equivalence predicts. To repair this, enlarge to its normal closure , the splitting field of over . Now (since , and adjoining , a root of the irreducible-over- quadratic — irreducible because contains no non-real root — multiplies the degree by ). Since has characteristic coefficients, it is automatically separable (its formal derivative shares no root with , as is not a root of ). By condition (3) of the theorem, is Galois, so , and one checks directly this group is isomorphic to (permuting the three roots ), matching exactly.
Example 2: A Galois extension over a field of positive characteristic — finite fields. Let and consider and (see Finite Fields). We have , since is the (unique, up to isomorphism) field of order and is a -dimensional -vector space. Every finite field is a splitting field: is precisely the splitting field over of , since the roots of in an algebraic closure of are exactly the elements of (the multiplicative group has order , and every nonzero element satisfies , hence ; together with , all elements of satisfy , and since there can be no other roots). Moreover is separable: its formal derivative is in characteristic (since ), which is a nonzero constant, so and has no repeated roots. By condition (3), is Galois, so . We can identify explicitly: the Frobenius map , , is a field automorphism (it is a ring homomorphism since in characteristic — the cross terms in the binomial expansion carry factors of that vanish — and it is injective, hence bijective on the finite set ), and it fixes pointwise (Fermat's little theorem: for all ). So . Its powers are distinct: if for , every element of would satisfy , giving at most roots for the polynomial , contradicting that all elements of would need to be roots. Since exactly, we conclude
generated by the Frobenius automorphism — a fact that generalizes to for every prime and every .
The theorem requires finite with the further hypothesis, in condition (1), of both normal and separable. Drop separability and keep normality: let be a prime, let be an indeterminate, and consider
Here is a root of . This polynomial is irreducible over (an application of Eisenstein-type criteria in the ring , viewing as a "prime" of , localized appropriately — or directly: any factorization of over would force a root in by a valuation argument on the -adic valuation, and has no -th root in by degree considerations), so . Over , since we are in characteristic , the Frobenius identity gives
(using in characteristic , since all the intermediate binomial coefficients for are divisible by and vanish). So splits completely in — with a single root of multiplicity . Since is generated by this one root, is the splitting field of over , hence is normal. But has repeated roots (all of its roots coincide at ), so is not separable, and in fact 's minimal polynomial is inseparable: is not separable. Now check the automorphism count. Any sends to a root of in ; but has only the single root (with multiplicity , but as a set of roots it is ). So forcibly, giving . Hence
Normality alone was not enough: without separability, the automorphism count collapses to regardless of the degree , and the fixed field as well. This shows separability cannot be dropped from condition (1), and correspondingly the equality in condition (2) and the fixed-field statement in (4) both fail the instant it is removed.
Scope, edge cases & common traps
| Situation | What holds | What can fail |
|---|---|---|
| finite, normal, separable | All four conditions equivalent; well-defined, | — |
| normal but not separable (char ) | Normality alone | ; fixed field strictly bigger than |
| separable but not normal | Separability alone; (embedding count) | ; not every embedding is an automorphism |
| infinite algebraic (e.g. ) | Analogous Galois theory exists (Krull topology) | The finite-group counting argument here does not directly apply; requires profinite methods, outside this scope |
| perfect (e.g. , finite fields, algebraically closed fields) | Every algebraic extension is automatically separable | Normality can still fail (Example 1) — perfectness only removes the separability obstruction |
The mathematician behind it
Origins · 1 minÉvariste Galois and how this was discovered
Galois developed the group-theoretic viewpoint on polynomial solvability as a teenager, submitting memoirs to the French Academy of Sciences that were repeatedly lost or rejected. He introduced the idea of associating a group of permutations of the roots of a polynomial to the polynomial itself, and showed that solvability of the equation by radicals corresponds to a solvability property of this group — the origin of the modern term "solvable group." He died at twenty from wounds sustained in a duel, having written down much of his mathematics the night before in a letter to a friend.
Is the fixed field of any finite group of automorphisms automatically giving a Galois extension?
Yes, in the following precise sense (a fact closely tied to, but distinct from, the theorem proved here): if is any finite group of automorphisms of a field , and is its fixed field, then is automatically Galois with . This is sometimes called Artin's theorem on fixed fields, and it is the reverse construction to the one in this article — instead of starting with an extension and asking whether it is Galois, one starts with a group acting on and manufactures a Galois extension by definition. It is a companion result to, but not a restatement of, the equivalence proved above, and is used heavily in the proof of The Fundamental Theorem of Galois Theory.
Does every field extension have a "Galois closure"?
Yes, for any finite separable extension . One takes the splitting field over of the product of minimal polynomials of a generating set of ; since is separable, this product can be chosen separable (after removing repeated irreducible factors), and is then a splitting field of a separable polynomial, hence Galois over by condition (3) of the theorem. This is called the Galois closure (or normal closure) of ; it is the smallest Galois extension of containing . Example 1 illustrates this: is not Galois over , but its Galois closure is , of degree .
Why is separability automatic in characteristic and for finite fields, but not in general?
A field is called perfect if every algebraic extension of is separable; this holds automatically when (the derivative test for separability never produces the degenerate case, since can only signal repeated roots via characteristic- phenomena in the derivative) and when is finite (every element of a finite field satisfies , and the Frobenius map is an automorphism, which forces every irreducible polynomial to be separable). Function fields like are not perfect: the element has no -th root in , and adjoining one, as in the counterexample, produces a genuinely inseparable extension. This is why the counterexample above needed a transcendental base field rather than or a finite field.
Can an infinite extension be "Galois" in this sense?
The definition given here — via — extends verbatim to infinite algebraic extensions, and conditions (1), (3), (4) (normal+separable, separable splitting field of a possibly infinite set of polynomials, fixed field equality) remain equivalent. However condition (2), the numeric equality , needs reinterpretation since both sides can be infinite cardinals and a naive cardinality match is too weak to carry the same information; the correct infinite-dimensional replacement uses the Krull topology on and profinite group theory. This article's counting proof is specific to the finite case and does not directly generalize; the infinite theory is a separate, more advanced topic.
How does this relate to separable but non-algebraic (e.g. purely transcendental) extensions?
It doesn't directly apply — the entire framework assumes is algebraic (indeed finite), since minimal polynomials and splitting fields are meaningless for transcendental elements. A purely transcendental extension like has an enormous automorphism group (all Möbius-type substitutions with , for instance) but is not algebraic at all, so the notion of "Galois" as defined here simply does not classify it; a different framework (transcendence degree, differential Galois theory for function fields) is needed for such extensions.