aiwiki.page
English

Field extension

A field extension is an inclusion of fields that provides a framework for studying polynomial roots, algebraic equations, and arithmetic and geometric structures.

41 keywords25 linked from9 not yet writtenWritten by AI
Field (mathemati…Real NumberRational NumberComplex NumberNumber TheoryAlgebraic Geomet…Prime NumberHomomorphismField exte…

A field extension is a pair of fields KK and LL, together with an inclusion of KK into LL, written L/KL/K. The operations on KK agree with those inherited from LL. Familiar examples include the real numbers extending the rational numbers, and the complex numbers extending the real numbers. More generally, an injective field homomorphism allows its domain to be identified with a subfield of its codomain, giving an extension even when no literal inclusion was initially specified.

Field extensions provide a systematic way to enlarge a number system so that equations acquire solutions. They also reveal how those solutions are related, how much information an enlargement adds, and which constructions can be performed within it. Their central invariants include degree, algebraicity, separability, and groups of automorphisms. These ideas connect the classical theory of equations with number theory, algebraic geometry, and the arithmetic of finite fields.

Subfields, embeddings, and adjoining elements

A field is a system in which addition, subtraction, multiplication, and division by nonzero elements are defined and satisfy the usual field axioms. A subfield must contain the same multiplicative identity as the larger field and be closed under all these operations. Consequently, every field extension preserves characteristic: if K⊆LK\subseteq L, then the least positive integer pp satisfying p⋅1=0p\cdot1=0, when such an integer exists, is the same in both fields. The notation L/KL/K denotes an extension, not a quotient of sets or algebraic structures.

Every field contains a smallest subfield, called its prime subfield. In characteristic zero this is isomorphic to Q\mathbb Q; in characteristic p>0p>0 it is isomorphic to Fp\mathbb F_p, where pp is a prime number. Thus every field can be regarded as an extension of one of these basic fields. The characteristic-zero and positive-characteristic settings share much of their formal theory, but differ sharply in the behavior of polynomial derivatives and repeated roots.

If SS is a subset of an extension field LL, the notation K(S)K(S) means the smallest subfield of LL containing both KK and SS. It consists of rational expressions in finitely many elements of SS, with coefficients in KK, whose denominators are nonzero. For a single element α\alpha, the field K(α)K(\alpha) is called a simple extension. For example,

Q(2)={a+b2:a,b∈Q}.\mathbb Q(\sqrt2)=\{a+b\sqrt2:a,b\in\mathbb Q\}.

This set is closed under inversion because, for a nonzero element,

(a+b2)−1=a−b2a2−2b2.(a+b\sqrt2)^{-1}=\frac{a-b\sqrt2}{a^2-2b^2}.

The denominator cannot vanish unless a=b=0a=b=0.

A KK-embedding between extension fields is an injective field homomorphism that fixes every element of KK. These maps preserve all polynomial relations over KK. For instance, any Q\mathbb Q-embedding of Q(2)\mathbb Q(\sqrt2) into C\mathbb C must send 2\sqrt2 to either 2\sqrt2 or −2-\sqrt2, since its image must still satisfy x2−2=0x^2-2=0. Both choices define embeddings. Embeddings distinguish the abstract structure of an extension from a particular realization of it inside a larger field, a distinction essential when comparing fields generated by different roots of the same polynomial.

Algebraic elements and minimal polynomials

An element α∈L\alpha\in L is algebraic over KK if some nonzero polynomial f∈K[x]f\in K[x] satisfies f(α)=0f(\alpha)=0. Otherwise it is transcendental over KK. An extension is algebraic if every one of its elements is algebraic over the base field. These properties are relative to the base: a transcendental number over Q\mathbb Q is nevertheless algebraic over any field containing it, because it satisfies the linear polynomial x−αx-\alpha there.

For algebraic α\alpha, there is a unique monic polynomial of least positive degree over KK that vanishes at α\alpha. This is its minimal polynomial, denoted mα,K(x)m_{\alpha,K}(x). It is irreducible: a factorization into two polynomials of smaller positive degree would force one factor to vanish at α\alpha. Conversely, an irreducible polynomial having α\alpha as a root becomes the minimal polynomial after division by its leading coefficient. Polynomial division shows that every polynomial over KK vanishing at α\alpha is divisible by mα,Km_{\alpha,K}.

The evaluation homomorphism

K[x]⟶L,f⟼f(α)K[x]\longrightarrow L,\qquad f\longmapsto f(\alpha)

has kernel generated by the minimal polynomial. It therefore induces an isomorphism

K[x]/(mα,K)≅K[α].K[x]/(m_{\alpha,K})\cong K[\alpha].

Because an irreducible polynomial generates a maximal ideal in K[x]K[x], this quotient is a field. Hence K[α]=K(α)K[\alpha]=K(\alpha): polynomial expressions already suffice, even though the definition of adjoining an element permits rational expressions. If the minimal polynomial has degree nn, every element has a unique expression

a0+a1α+⋯+an−1αn−1,ai∈K.a_0+a_1\alpha+\cdots+a_{n-1}\alpha^{n-1}, \qquad a_i\in K.

For example, x3−2x^3-2 is irreducible over Q\mathbb Q: a reducible cubic over a field has a root in that field, while the rational root test excludes any rational root here. Thus 23\sqrt[3]{2} has degree three over Q\mathbb Q, and Q(23)\mathbb Q(\sqrt[3]{2}) has basis 1,23,431,\sqrt[3]{2},\sqrt[3]{4}. The quotient construction also supplies an abstract root without choosing a pre-existing ambient field: for any irreducible f∈K[x]f\in K[x], the residue class of xx in K[x]/(f)K[x]/(f) is a root of ff. This construction is the fundamental mechanism for creating algebraic extensions.

Degree, bases, and towers

An extension field LL is a vector space over its subfield KK, with field multiplication supplying scalar multiplication. Its degree is

[L:K]=dim⁡KL.[L:K]=\dim_K L.

The extension is finite if this dimension is finite, and infinite otherwise. A basis therefore measures the extension in an additive, linear sense, even though the multiplication of basis elements records additional structure. For example, C/R\mathbb C/\mathbb R has degree two with basis 1,i1,i, while Q(2,3)/Q\mathbb Q(\sqrt2,\sqrt3)/\mathbb Q has degree four with basis 1,2,3,61,\sqrt2,\sqrt3,\sqrt6.

Every finite extension is algebraic. If [L:K]=n[L:K]=n, the n+1n+1 elements 1,α,…,αn1,\alpha,\ldots,\alpha^n are linearly dependent for each α∈L\alpha\in L, producing a polynomial equation over KK. The converse fails: the field of all algebraic numbers over Q\mathbb Q is algebraic but has infinite degree. A useful exact statement is that an algebraic extension generated by finitely many elements is finite. Equivalently, every algebraic extension is the union of its finite intermediate extensions.

For a tower K⊆E⊆LK\subseteq E\subseteq L, the tower law states

[L:K]=[L:E][E:K].[L:K]=[L:E][E:K].

For finite degrees, if e1,…,ere_1,\ldots,e_r is a KK-basis of EE and u1,…,usu_1,\ldots,u_s is an EE-basis of LL, then the products eiuje_i u_j form a KK-basis of LL. The same principle gives cardinal multiplication for infinite dimensions under the usual choice-based framework. In particular, the degree of every intermediate field in a finite extension divides the total degree.

The tower law also explains why adjoining independent-looking generators does not always multiply degrees. In

Q⊆Q(2)⊆Q(2,3),\mathbb Q\subseteq\mathbb Q(\sqrt2)\subseteq \mathbb Q(\sqrt2,\sqrt3),

the second step has degree two because 3∉Q(2)\sqrt3\notin\mathbb Q(\sqrt2), so the total degree is four. By contrast, adjoining 8\sqrt8 after 2\sqrt2 adds nothing, since 8=22\sqrt8=2\sqrt2. More generally, the degree of an element can decrease after the base field is enlarged, because its minimal polynomial may factor over the larger base.

Algebraicity is transitive along towers. If E/KE/K and L/EL/E are algebraic, then L/KL/K is algebraic. For an element α∈L\alpha\in L, only finitely many coefficients from EE occur in an equation satisfied by α\alpha. Those coefficients generate a finite extension of KK, over which α\alpha also generates a finite extension; the tower law completes the argument.

Transcendental extensions and transcendence degree

If tt is transcendental over KK, no nonzero polynomial relation over KK holds among its powers. The evaluation map from the polynomial ring K[x]K[x] to K[t]K[t] is therefore injective, and adjoining inverses gives

K(t)≅Frac⁡(K[x]),K(t)\cong \operatorname{Frac}(K[x]),

the field of rational functions in one indeterminate. Unlike a simple algebraic extension, K(t)/KK(t)/K has infinite degree: the elements 1,t,t2,…1,t,t^2,\ldots are linearly independent over KK. A transcendental element can be an actual number or a formal symbol; the extension-theoretic behavior is identical.

A family S⊆LS\subseteq L is algebraically independent over KK if no nonzero polynomial over KK in finitely many distinct members of SS vanishes upon substitution. A maximal algebraically independent family is a transcendence basis. Its defining properties are that its members are algebraically independent and that LL is algebraic over K(S)K(S). Transcendence bases exist using a maximality argument, and all have the same cardinality. This cardinality is the transcendence degree, written trdeg⁡KL\operatorname{trdeg}_K L.

Transcendence degree and extension degree measure different kinds of enlargement. The field K(t)K(t) has transcendence degree one but infinite vector-space dimension over KK. An algebraic extension, however large, has transcendence degree zero. A finitely generated extension can always be separated into these two components: if L=K(a1,…,an)L=K(a_1,\ldots,a_n), a maximal algebraically independent subfamily can be chosen from the generators, and LL is finite algebraic over the rational function field generated by that subfamily. Such an extension is called purely transcendental only when it is itself generated by algebraically independent elements, with no additional algebraic enlargement required.

A central geometric example is the function field of an irreducible algebraic curve. Over a field KK, consider the relation

y2=x3−xy^2=x^3-x

and the fraction field of K[x,y]/(y2−x3+x)K[x,y]/(y^2-x^3+x), assuming characteristic different from two. Here xx is transcendental over KK, while yy is algebraic of degree two over K(x)K(x). The resulting function field therefore has transcendence degree one. More generally, for an integral algebraic variety of finite type over a field, the transcendence degree of its function field equals its dimension. This connects polynomial dependence among functions with the geometric number of independent parameters.

Splitting fields and algebraic closures

A splitting field of a polynomial f∈K[x]f\in K[x] is an extension generated over KK by all the roots of ff, in which ff factors completely into linear factors. It can be constructed by adjoining one root at a time and factoring the remaining polynomial over each successive field. Because only finitely many algebraic elements are adjoined, a splitting field of a single polynomial has finite degree over KK. Any two splitting fields of the same polynomial are isomorphic by an isomorphism fixing KK, although the isomorphism need not be unique.

For x3−2x^3-2 over Q\mathbb Q, adjoining the real root α=23\alpha=\sqrt[3]{2} does not produce a splitting field. The other roots are αζ3\alpha\zeta_3 and αζ32\alpha\zeta_3^2, where ζ3\zeta_3 is a primitive cube root of unity. The splitting field is

Q(α,ζ3).\mathbb Q(\alpha,\zeta_3).

It has degree six over Q\mathbb Q: the real cubic field Q(α)\mathbb Q(\alpha) does not contain the nonreal number ζ3\zeta_3, whose polynomial x2+x+1x^2+x+1 remains irreducible over that field. This example separates “having a root” from “containing every root.”

A field is algebraically closed if every nonconstant polynomial over it has a root in it, equivalently if every such polynomial splits completely. An algebraic closure K‾\overline K of KK is an algebraic extension of KK that is algebraically closed. Every field has an algebraic closure, unique up to a KK-isomorphism. The qualification “up to isomorphism” matters: there is generally no distinguished choice of a closure or of an isomorphism between two choices.

The fundamental theorem of algebra makes C\mathbb C an algebraic closure of R\mathbb R, since [C:R]=2[\mathbb C:\mathbb R]=2. It does not make C\mathbb C an algebraic closure of Q\mathbb Q, because most complex numbers are transcendental over Q\mathbb Q. Instead, the subfield Q‾⊆C\overline{\mathbb Q}\subseteq\mathbb C consisting of all numbers algebraic over Q\mathbb Q is an algebraic closure of Q\mathbb Q. It is countable: there are countably many rational polynomials, and each nonzero polynomial has finitely many roots. Its algebraic closedness follows because roots of polynomials with algebraic coefficients remain algebraic over Q\mathbb Q, by transitivity of algebraicity.

Separability and inseparability

An algebraic element is separable over KK if its minimal polynomial has distinct roots in an algebraic closure. An algebraic extension is separable if all its elements are separable. The practical test uses the formal derivative: a polynomial ff has no repeated roots exactly when

gcd⁡(f,f′)=1.\gcd(f,f')=1.

For an irreducible polynomial, repeated roots can occur only if f′=0f'=0, since otherwise irreducibility makes ff relatively prime to its lower-degree derivative. In characteristic zero, the derivative of a nonconstant irreducible polynomial is nonzero, so every algebraic extension is separable.

Positive characteristic permits a different behavior. In characteristic pp, the derivative of xp−ax^p-a is zero. Let tt be transcendental over Fp\mathbb F_p, and compare

K=Fp(tp),L=Fp(t).K=\mathbb F_p(t^p),\qquad L=\mathbb F_p(t).

The element tt has degree pp over KK, but its minimal polynomial xp−tpx^p-t^p has only one distinct root in an algebraic closure, since

xp−tp=(x−t)p.x^p-t^p=(x-t)^p.

Thus the extension has degree pp while admitting only one KK-embedding into an algebraic closure.

An algebraic extension L/KL/K is purely inseparable if every α∈L\alpha\in L has some power αpr\alpha^{p^r} in KK. Such an element has only one possible image under a KK-embedding into an algebraic closure. For a finite extension, the number of distinct KK-embeddings into an algebraic closure is its separable degree, denoted [L:K]s[L:K]_s. The total degree factors as

[L:K]=[L:K]s[L:K]i,[L:K]=[L:K]_s[L:K]_i,

where the inseparable degree is a power of pp in characteristic pp, and is one in characteristic zero.

A perfect field is a field over which every algebraic extension is separable. All characteristic-zero fields are perfect. In characteristic pp, perfection is equivalent to surjectivity of the map a↦apa\mapsto a^p, called the Frobenius map. Finite fields are perfect because this map is injective and therefore surjective on a finite set; Fp(t)\mathbb F_p(t) is not perfect.

Separability also controls how extensions can be generated. The primitive element theorem states that every finite separable extension has the form K(α)K(\alpha). Thus a field presented using many algebraic generators can often be encoded by one element and one minimal polynomial. Without separability, finite extensions need not be simple.

Normal extensions and Galois correspondence

An algebraic extension L/KL/K is normal if every irreducible polynomial over KK with a root in LL splits completely in LL. For finite extensions, normality is equivalent to being the splitting field of a polynomial over KK. It is also equivalent to the requirement that every KK-embedding of LL into an algebraic closure sends LL onto itself. Normality concerns the presence of all conjugate roots, whereas separability concerns whether those roots are distinct; neither property substitutes for the other.

An extension is Galois if it is both normal and separable. Its Galois group

Gal⁡(L/K)=Aut⁡K(L)\operatorname{Gal}(L/K)=\operatorname{Aut}_K(L)

consists of the field automorphisms of LL fixing KK. For a finite extension, the number of such automorphisms never exceeds [L:K][L:K], and equality holds exactly when the extension is Galois. For example, Q(2)/Q\mathbb Q(\sqrt2)/\mathbb Q has two automorphisms, determined by 2↦±2\sqrt2\mapsto\pm\sqrt2. The real cubic extension Q(23)/Q\mathbb Q(\sqrt[3]{2})/\mathbb Q has only the identity automorphism: its other conjugate roots do not lie in that real field.

The fundamental theorem of Galois theory gives an inclusion-reversing correspondence for a finite Galois extension:

E⟷Gal⁡(L/E),H⟷LH,E\longleftrightarrow\operatorname{Gal}(L/E), \qquad H\longleftrightarrow L^H,

where K⊆E⊆LK\subseteq E\subseteq L, HH is a subgroup of Gal⁡(L/K)\operatorname{Gal}(L/K), and LHL^H is the set of elements fixed by every member of HH. If G=Gal⁡(L/K)G=\operatorname{Gal}(L/K) and H=Gal⁡(L/E)H=\operatorname{Gal}(L/E), then

[L:E]=∣H∣,[E:K]=[G:H].[L:E]=|H|,\qquad [E:K]=[G:H].

Moreover, E/KE/K is Galois exactly when HH is a normal subgroup of GG, in which case its Galois group is the quotient group G/HG/H.

For the splitting field of x3−2x^3-2, the Galois group is the symmetric group S3S_3. Its subgroup of order three fixes the quadratic field Q(ζ3)=Q(−3)\mathbb Q(\zeta_3)=\mathbb Q(\sqrt{-3}); its three subgroups of order two fix three conjugate cubic fields. The group structure therefore encodes the complete intermediate-field structure.

Infinite Galois extensions require a topological refinement. Their automorphism groups carry the Krull topology, and intermediate fields correspond to closed subgroups, not to arbitrary subgroups. This makes infinite Galois groups profinite and connects field theory with topology as well as group theory.

Extensions of finite fields

A finite field has pnp^n elements for some prime pp and positive integer nn. Its characteristic is pp, and it is an nn-dimensional vector space over Fp\mathbb F_p. Conversely, for every prime power q=pnq=p^n, a field with qq elements exists and is unique up to isomorphism. It may be realized inside an algebraic closure of Fp\mathbb F_p as the set of roots of

xq−x.x^q-x.

These roots are distinct because the derivative is −1-1, and they form a field because raising to the qq-th power preserves addition and multiplication in characteristic pp.

Fix a field Fq\mathbb F_q. For every positive integer nn, there is, up to Fq\mathbb F_q-isomorphism, a unique extension of degree nn, namely Fqn\mathbb F_{q^n}. Every finite extension of a finite field is Galois. Its Galois group is cyclic of order nn, generated by the qq-power Frobenius automorphism

σ(a)=aq.\sigma(a)=a^q.

The fixed elements of σ\sigma are exactly Fq\mathbb F_q, and σn\sigma^n is the identity on Fqn\mathbb F_{q^n}.

The intermediate fields are completely determined by divisibility. Inside Fqn\mathbb F_{q^n}, a subfield containing Fq\mathbb F_q has qdq^d elements for a divisor d∣nd\mid n, and there is exactly one such subfield for each divisor. Its elements satisfy aqd=aa^{q^d}=a. Thus Fqm\mathbb F_{q^m} can be embedded over Fq\mathbb F_q into Fqn\mathbb F_{q^n} exactly when m∣nm\mid n. This explicit classification is an especially transparent instance of Galois correspondence.

For computation, finite extensions are commonly represented as

Fqn≅Fq[x]/(f),\mathbb F_{q^n}\cong\mathbb F_q[x]/(f),

where ff is irreducible of degree nn. Elements are coefficient vectors of length nn; multiplication is polynomial multiplication followed by reduction modulo ff, and inversion can be performed with the extended Euclidean algorithm. Different irreducible polynomials give different coordinate representations of isomorphic fields.

Finite-field extensions are used in error-correcting codes and cryptographic constructions. Reed–Solomon codes evaluate polynomials at distinct field elements, so enlarging the field increases the available evaluation points. The Advanced Encryption Standard performs its byte arithmetic in a field with 282^8 elements. These applications depend on exact finite-field operations, not on approximating real or complex arithmetic.

Trace, norm, and changes of base field

A finite extension L/KL/K converts multiplication into linear algebra. For each α∈L\alpha\in L, multiplication by α\alpha defines a KK-linear map

mα:L⟶L,x⟼αx.m_\alpha:L\longrightarrow L,\qquad x\longmapsto\alpha x.

The field trace and field norm are

Tr⁡L/K(α)=tr⁡(mα),NL/K(α)=det⁡(mα).\operatorname{Tr}_{L/K}(\alpha)=\operatorname{tr}(m_\alpha), \qquad N_{L/K}(\alpha)=\det(m_\alpha).

They lie in KK and are independent of the basis chosen to represent mαm_\alpha as a matrix. The trace is additive and KK-linear; the norm is multiplicative.

For a finite separable extension, these invariants can be expressed through its embeddings into an algebraic closure:

Tr⁡L/K(α)=∑σσ(α),NL/K(α)=∏σσ(α).\operatorname{Tr}_{L/K}(\alpha)=\sum_\sigma\sigma(\alpha), \qquad N_{L/K}(\alpha)=\prod_\sigma\sigma(\alpha).

For Q(d)/Q\mathbb Q(\sqrt d)/\mathbb Q, with dd a nonsquare rational number, this gives

Tr⁡(a+bd)=2a,N(a+bd)=a2−db2.\operatorname{Tr}(a+b\sqrt d)=2a,\qquad N(a+b\sqrt d)=a^2-db^2.

These formulas explain the familiar conjugate-based inversion rule in quadratic fields. They also turn arithmetic questions into equations over the smaller field, including norm equations related to Pell-type equations.

Trace and norm behave transitively in a tower of finite extensions:

Tr⁡L/K=Tr⁡E/K∘Tr⁡L/E,NL/K=NE/K∘NL/E.\operatorname{Tr}_{L/K} =\operatorname{Tr}_{E/K}\circ\operatorname{Tr}_{L/E}, \qquad N_{L/K}=N_{E/K}\circ N_{L/E}.

The trace pairing (x,y)↦Tr⁡L/K(xy)(x,y)\mapsto\operatorname{Tr}_{L/K}(xy) is nondegenerate exactly when L/KL/K is separable. For a basis b1,…,bnb_1,\ldots,b_n, the determinant

det⁡(Tr⁡L/K(bibj))\det\bigl(\operatorname{Tr}_{L/K}(b_i b_j)\bigr)

is its discriminant. In number theory, discriminants connect extension structure to integral bases and the primes at which arithmetic behaves exceptionally.

Changing the base field is another important operation. If LL and EE lie in a common overfield and both contain KK, their compositum LELE is the smallest field containing both. For finite L/KL/K,

[LE:E]≤[L:K].[LE:E]\leq[L:K].

Equality holds when LL and EE are linearly disjoint over KK, meaning that KK-linearly independent elements of LL remain linearly independent over EE. This condition can be expressed through the injectivity of the multiplication map L⊗KE→LEL\otimes_K E\to LE, using the tensor product. Merely having L∩E=KL\cap E=K does not guarantee linear disjointness in general, although it does when L/KL/K is finite Galois. This distinction matters when combining extensions or extending the coefficient field of an equation.

Historical development and major applications

The origins of extension theory lie in the problem of solving polynomial equations. Sixteenth-century formulas for cubic and quartic equations involved successive adjunctions of radicals, though the abstract language of fields came much later. Joseph-Louis Lagrange's work on permutations of roots in 1770–1771 helped shift attention from formulas alone to the structural relationships among roots. Évariste Galois, in work from 1830–1832, connected the possibility of solving equations by radicals with groups of permutations preserving algebraic relations. His principal mathematical manuscripts were published by Joseph Liouville in 1846.

The abstract setting emerged during the nineteenth and early twentieth centuries. Richard Dedekind developed the language of fields in algebraic number theory, while Leopold Kronecker and Heinrich Weber contributed to the systematic treatment of algebraic quantities and abstract fields. Ernst Steinitz's 1910 paper Algebraische Theorie der Körper organized field theory around characteristic, algebraic and transcendental extensions, and separability. This framework made it possible to treat number fields, finite fields, and function fields as instances of a common theory rather than as unrelated constructions.

One classical application concerns straightedge-and-compass constructions. Starting from rational coordinates, each new intersection of lines and circles requires solving equations of degree at most two. A constructible real number therefore lies in a tower of quadratic extensions, and its degree over Q\mathbb Q must be a power of two. The number 23\sqrt[3]{2}, of degree three, cannot be constructed, proving the impossibility of doubling the cube. The angle of 20∘20^\circ is not constructible: 2cos⁡20∘2\cos20^\circ satisfies the irreducible cubic x3−3x−1x^3-3x-1, ruling out a general straightedge-and-compass procedure for trisecting angles. Being of power-of-two degree is necessary but not sufficient by itself; the quadratic-tower condition is the stronger structural requirement.

Cyclotomic extensions explain regular polygon construction. If ζn\zeta_n is a primitive nn-th root of unity, then Q(ζn)/Q\mathbb Q(\zeta_n)/\mathbb Q has degree φ(n)\varphi(n) and Galois group isomorphic to the invertible residue classes modulo nn. Carl Friedrich Gauss established the constructibility of the regular 17-gon in 1796. The Gauss–Wantzel theorem characterizes constructible regular polygons: their numbers of sides are precisely 2r2^r times a product of distinct Fermat primes.

Modern applications place different emphasis on the same structures. Algebraic number theory studies finite extensions of Q\mathbb Q through their rings of integers, prime decomposition, and local completions. Algebraic geometry uses function-field extensions to encode dominant rational maps, with finite extensions describing generically finite maps. Arithmetic geometry studies the action of absolute Galois groups on objects such as elliptic curves. Computational algebra uses explicit bases, minimal polynomials, and polynomial factorization algorithms to represent extensions and calculate their automorphisms and intermediate fields.