An abelian group is a group in which the order of combining two elements does not affect the result: for all elements and . Its operation also satisfies the group axioms of associativity, an identity element, and inverses. Addition of integers is a basic example. Abelian groups form a fundamental class of objects in abstract algebra. (jmilne.org)
The theory connects group operations with arithmetic and linear methods. Every abelian group is naturally a module over the integers, and every finitely generated abelian group can be decomposed into cyclic groups. Beyond finite generation, torsion, divisibility, and freeness distinguish substantially different kinds of infinite groups. These structures also supply the groups used in homology, linking algebraic classification to the study of topological spaces. (arxiv.org)
Definition, notation, and commutativity
An abelian group, or commutative group, is a nonempty set equipped with a binary operation satisfying the group axioms and an additional commutative law. The operation is closed: belongs to whenever . It is associative, so . There is an identity element satisfying , and every has an inverse satisfying . The additional requirement is
Associativity concerns parentheses; commutativity concerns order. Neither replaces the other in the definition. (ocw.mit.edu)
Abelian groups are commonly written in additive notation: the operation is , the identity is , and the inverse of is . Thus and ; subtraction means . These symbols describe the abstract operation, not necessarily ordinary numerical addition. For a positive integer , the expression denotes the sum of copies of ; one defines and . In multiplicative notation, the corresponding expressions are , , , and , with and . Multiplicative notation does not imply that a group is noncommutative. (math.berkeley.edu)
The adjective “abelian” commemorates the Norwegian mathematician Niels Henrik Abel (1802–1829). It specifies commutativity, not the stronger requirement of being a cyclic group, meaning a group generated by a single element. Every cyclic group is abelian because , but the converse fails. For example, , with componentwise addition, is abelian but not cyclic: the multiples of one pair cannot include both and . (www2.math.uconn.edu)
Commutativity can also be expressed through the center of a group , defined in multiplicative notation by
The center is a subgroup consisting of the elements that commute with every element of . Consequently, is abelian exactly when . Merely having a nonidentity element in the center is a weaker condition. (jmilne.org)
A second characterization uses commutators. With the convention , the equality holds exactly when . Thus is abelian if and only if every commutator is the identity, equivalently if its commutator subgroup—the subgroup generated by all commutators—is . Some authors instead use ; the criterion that all commutators are trivial is unchanged. (jmilne.org)
Elementary examples and nonexamples
The integers , rational numbers , real numbers , and complex numbers are abelian groups under ordinary addition. In each case, sums remain in the specified set, the identity is , and the inverse of is ; associativity and commutativity are inherited from numerical addition. The operation is essential to the specification: under multiplication is not a group, since has no integer multiplicative inverse. Likewise, none of these four complete number systems forms a group under multiplication, because has no multiplicative inverse. (math.mit.edu)
For every positive integer , the residues form an abelian group under addition in modular arithmetic. Writing for the residue class of , the operation is , with identity and inverse . This is a cyclic group: repeatedly adding produces all elements. For example, in , . The case is the one-element group. More generally, every cyclic group is abelian, because any two powers of a generator satisfy . (jmilne.org)
The Klein four-group supplies an abelian example that is not cyclic. It can be represented as
with componentwise addition modulo . Thus , in either order. Every nonidentity element added to itself gives , so each generates only a two-element group; no element generates all four elements. This distinguishes from the cyclic group , despite their equal sizes. A multiplicative realization consists of the four diagonal real matrices , where each is or . (ocw.mit.edu)
For any field , the nonzero elements form an abelian group under multiplication. Its identity is , and the inverse of is ; excluding zero ensures that every element is invertible. Examples include , , and . In the finite field , the multiplicative group has elements : the successive powers of modulo are , so it is cyclic of order . These multiplicative groups must not be confused with the additive groups of the same fields. (math.mit.edu)
By contrast, the symmetric group , consisting of the six permutations of three objects under composition, is not abelian. With the rightmost permutation applied first,
The two products differ, although composition remains associative and every permutation has an inverse. Thus associativity does not imply commutativity. (jmilne.org)
Invertible matrices under multiplication provide another nonexample. The general linear group contains
both with determinant , but direct multiplication gives
Matrix-based groups are not necessarily nonabelian—the diagonal realization of above is abelian—but invertibility alone does not make multiplication commutative. (jmilne.org)
Subgroups, quotients, and homomorphisms
A subgroup of an abelian group is itself abelian, since its operation is inherited from . It is also a normal subgroup: in additive notation, conjugation has the form , so every element of leaves invariant under conjugation. Consequently, every subgroup defines a quotient group . Its elements are cosets , with addition
This operation is independent of the representatives: replacing and by and changes their sum by . The identity is , the inverse of is , and the quotient remains abelian. For example, identifies integers whose difference is divisible by , giving five cosets. (math.mit.edu)
A homomorphism between additively written abelian groups preserves the operation:
It follows that , , and for every integer . Its kernel is , and its image is ; these are subgroups of and , respectively. The kernel records precisely which differences are lost: if and only if . Thus is injective exactly when , and surjective exactly when . An isomorphism is a bijective homomorphism; its inverse also preserves addition. (jmilne.org)
The first isomorphism theorem expresses the relationship between these constructions:
The displayed map is well-defined because elements of the same coset have the same image, injective because equal images imply equal cosets, and surjective onto by definition. For a concrete example, let send to . Its kernel is , while its image consists of the three residue classes . Hence this image is isomorphic to , although the stated codomain has twelve elements. (math.columbia.edu)
For an arbitrary group , now written multiplicatively, the commutator subgroup is generated by all elements . It is normal, and the quotient
is called the abelianization of . In this quotient every commutator becomes the identity, forcing all cosets to commute. More precisely, is the smallest normal subgroup for which is abelian. If is already abelian, then , so its abelianization is naturally isomorphic to . (jmilne.org)
Abelianization has a universal property: every homomorphism into an abelian group factors uniquely through the quotient map . Indeed, , so , and the required map is
Its uniqueness follows from the surjectivity of . Thus “largest abelian quotient” describes a factorization property, not a comparison of cardinalities: every homomorphism from to an abelian group is completely determined by a homomorphism from . (gt.postech.ac.kr)
Direct sums, products, and limiting constructions
For a family of abelian groups , the direct product consists of all tuples , with . Its addition is componentwise: ; the zero tuple and componentwise negatives supply the identity and inverses. The direct sum is the subgroup
The finite set of nonzero coordinates is the tuple’s support. Adding two finite-support tuples again gives finite support, since the support of their sum is contained in the union of their supports. Neither construction requires a topology or a notion of convergence. (lib.ysu.am)
When is finite, every tuple has finite support, so the direct sum and product coincide canonically. Their distinction for infinite families is also expressed by their universal properties: homomorphisms correspond to families of homomorphisms , whereas homomorphisms correspond to families . In the latter case, the induced map sends to , a well-defined finite sum. Thus the direct sum is the coproduct in the category of abelian groups; no general infinite summation operation is needed. (lib.ysu.am)
For example, contains precisely the integer sequences that are eventually zero, while contains every integer sequence, including . These groups are not merely different subgroups of the same ambient product: they cannot be isomorphic. Indeed, the sum is countable, being the union of the countable sets of sequences supported in . The product is uncountable because it contains all binary sequences ; its cardinality is . This cardinality argument directly establishes the failure of the finite sum–product identification in this example. (lib.ysu.am)
A direct limit assembles groups connected by forward homomorphisms. For a sequence , it can be constructed as
where denotes the inclusion into the direct sum. These relations identify each element with its image at the next stage. Every limit element has a representative at one stage, and two representatives agree exactly when their images agree at some common later stage. For inclusions , the direct limit is the increasing union . In particular, adjoining a zero coordinate gives . (arxiv.org)
An inverse limit instead uses backward maps and retains compatible sequences:
Addition remains componentwise. For a prime , the reduction maps yield the additive group of the -adic integers,
An element therefore records a residue at every precision, each reducing to the preceding residue. For , the compatible sequence represents . Ordinary integers embed by taking their residues at every level; injectivity follows because an integer divisible by every must be zero. (math.mit.edu)
Integer modules and free abelian groups
For an abelian group , integer scalar multiplication is determined by its addition: for , define as the sum of copies of , set , and put . Commutativity ensures that
These are precisely the axioms of a module over the ring . Conversely, the additive group of any -module is abelian, and its scalar multiplication must be this repeated-addition operation. Every group homomorphism satisfies , so it is automatically -linear. Thus abelian groups and -modules describe the same objects and maps, not merely analogous structures; their subgroups are exactly their submodules. (faculty.niu.edu)
A free abelian group is an abelian group with a subset , called an integer basis, such that every element has a unique expression
with only finitely many nonzero coefficients. Equivalently, generates and is linearly independent over : any finite integer relation among distinct basis elements has all coefficients zero. Given an arbitrary set , the free abelian group on can be constructed as the group of finitely supported functions , with pointwise addition. Its basis elements are the functions taking value at one specified point and elsewhere, and it is denoted . The finite-support condition remains essential when is infinite. (faculty.niu.edu)
This construction has a universal mapping property: every function , where is any abelian group, extends uniquely to a homomorphism . Explicitly,
Finiteness makes the sum meaningful, and uniqueness of coordinates makes it well-defined. No independence condition is imposed on the chosen images . For example, specifying the images of the two standard basis elements of determines a homomorphism to any abelian group. More generally, taking to be the underlying set of gives a surjective homomorphism from a free abelian group onto . (faculty.niu.edu)
An integer basis differs from a vector-space basis because integers do not form a field: division by a nonzero scalar is not generally available. For example, and are integer-linearly independent in , but they do not generate it, since their first coordinates are always even. The same pair is a basis of the vector space , where multiplication by is permitted. Thus an independent set need not extend to an integer basis. Nor does every abelian group possess one: in , for , every element is annihilated by , whereas a nonzero element of a free abelian group cannot satisfy such a relation, by coordinate uniqueness. (faculty.niu.edu)
Every subgroup of a free abelian group is itself free abelian, including groups with infinite bases under the usual axiom of choice. In finite rank, an induction illustrates the mechanism. For , project onto the last coordinate. Its image is either zero or for some . The kernel is a subgroup of , hence free by induction. If the image is nonzero, choose with last coordinate ; subtracting suitable multiples of gives . The subgroup’s basis need not extend to one of the ambient group: has basis , but neither nor generates . (arxiv.org)
Classification of finitely generated abelian groups
An abelian group is finitely generated if some finite list expresses every element as an integer combination . The fundamental classification theorem states that every such group has an isomorphism
Here , and denotes the direct sum. The integer , called the free rank, and the ordered sequence , called the invariant factors, are uniquely determined by . Consequently, two finitely generated abelian groups are isomorphic exactly when these data agree. Empty lists are allowed: gives , while gives the trivial group. Uniqueness concerns the isomorphism types of the factors, not a preferred choice of generators or internal summands. (jmilne.org)
The finite cyclic summands together constitute the subgroup of all finite-order elements, whereas the quotient by that subgroup is isomorphic to . Thus is finite precisely when ; in that case its order is , with the empty product equal to . A finitely generated group with no nonzero finite-order elements is therefore free. For example, has free rank and six finite-order elements, but is itself infinite. The finite-generation hypothesis is essential to these conclusions. (crypto.stanford.edu)
An equivalent elementary-divisor decomposition replaces the finite cyclic summands by cyclic groups of prime-power order:
Only finitely many summands occur, and the multiset of prime powers is unique. Conversion from invariant factors uses the Chinese remainder theorem: if , reduction modulo the prime-power factors gives
To reverse the conversion, arrange each prime’s exponents in nondecreasing order, pad shorter lists on the left with zeros, and multiply the prime powers in each column. The resulting integers form a divisibility chain. For instance, elementary divisors give invariant factors , so . (jmilne.org)
For order , the theorem yields exactly three abelian groups up to isomorphism:
These correspond to the exponent lists , , and . Their largest element orders are respectively , , and , which also directly distinguishes their isomorphism types. This list classifies only the abelian groups of that order, not all groups of order . (crypto.stanford.edu)
More generally, for a fixed prime , each integer partition , with , determines
Its order is ; existence and uniqueness of elementary divisors show that every abelian group of this order occurs exactly once. Hence the number of isomorphism classes is the partition number , independent of . For an arbitrary positive order , the choices for distinct primes are independent, giving classes. For example, gives abelian groups up to isomorphism. (crypto.stanford.edu)
Presentations and Smith normal form
An abelian presentation specifies generators and integer relations , for . Commutativity is understood; these are relations in a free abelian group, not an unrestricted group presentation. Using the relations as columns produces an integer matrix representing a homomorphism . The presented group is its cokernel:
Thus two integer combinations of generators represent the same element exactly when their coefficient vectors differ by an integer combination of the relation columns. Every finitely generated abelian group admits such a finite presentation. (ocw.mit.edu)
The permitted changes of presentation are integral changes of basis. A square integer matrix is unimodular if its determinant is or , equivalently if its inverse also has integer entries. Replacing by , with unimodular matrices and , preserves its cokernel up to isomorphism: changes coordinates in the generator lattice, while changes the basis used to express the relations. Elementary operations are swapping rows or columns, multiplying a row or column by , and adding an integer multiple of one row or column to another. The Euclidean algorithm supplies the repeated division steps underlying this reduction. (math.mit.edu)
The resulting Smith normal form is a rectangular diagonal matrix
where is the rank of over . The positive entries are uniquely determined. Consequently,
An entry contributes a trivial group; an entry greater than contributes a finite cyclic factor; each remaining coordinate in the target contributes a free generator. This makes the classification theorem computational rather than merely existential. Henry John Stephen Smith’s treatment of integer systems appeared in his 1861 paper “On systems of linear indeterminate equations and congruences.” (people.reed.edu)
For example, presents generators with and . Although already diagonal, it is not in Smith normal form because . Applying the allowed operations gives the following explicit calculation:
The steps are , , swapping columns, , and finally followed by a row-sign change. Hence the cokernel is , consistent with the Chinese remainder theorem. (math.mit.edu)
Determinantal data provide an independent check. If is the positive greatest common divisor of all minors and , then for . In the example, and , giving . For a nonsingular square matrix, the cokernel is finite of order . Unlike Gaussian elimination over a field, integral reduction cannot divide a row by an arbitrary nonzero pivot: dividing by , for example, is not an automorphism of the integer lattice. Over , the example reduces to the identity and has zero cokernel, losing precisely the finite-group information that Smith reduction retains. (math.mit.edu)
Torsion groups and primary decomposition
An element of an abelian group is a torsion element if for some positive integer ; its order is the least such . The torsion subgroup is
Commutativity ensures closure: if and , then . The quotient group is torsion-free: if , then is torsion, so for some , forcing . A group is a torsion group when ; this does not imply finiteness. For example, is infinite, but every coset is annihilated by . (arxiv.org)
For each prime , define the -primary component
Primary decomposition gives the canonical internal direct sum
Thus a torsion group equals the sum of its primary components, and every element has a unique decomposition with only finitely many nonzero components. The mechanism is the Chinese remainder theorem: factoring an annihilator into coprime prime powers supplies integer multiples of the element that isolate each primary part. For example, in , the -primary component is generated by , the -primary component by , and . The components are intrinsic subgroups, preserved by every homomorphism. (arxiv.org)
A group has bounded exponent if a single positive integer annihilates all its elements; the least such is its exponent. This uniform bound is stronger than torsion, where the annihilator may depend on the element. Direct calculations illustrate the distinction: is infinite with exponent , whereas is torsion with unbounded element orders. In the latter group, every element has finite support and hence finite order, although the orders of the summand generators grow without bound. These examples follow directly from the finite-support definition of a direct sum and the torsion criterion. (arxiv.org)
Within a -primary group , element height measures divisibility inside , rather than element order. Define a decreasing filtration, indexed by ordinals, by
for a limit ordinal . A nonzero element has height when ; an element surviving every stage is assigned height . For finite , membership in means that is solvable. In particular, can contain nonzero elements even in a reduced group, so finite heights alone need not capture its structure. (math.berkeley.edu)
Writing , the Ulm invariant at stage is
The quotient is a vector space over the field . These dimensions generalize cyclic-summand multiplicities: for , one has . Helmut Ulm’s 1933 classification theorem states that two countable reduced abelian -groups are isomorphic if and only if all their Ulm invariants agree. Here “reduced” means having no nonzero divisible subgroup, where divisibility requires solvability of for every and every . Countability is essential to this statement: it does not assert classification of arbitrary uncountable -groups by these invariants. Without reducedness, the divisible part must also be specified, since its Ulm invariants all vanish. (arxiv.org)
Divisible groups and reduced complements
An abelian group is divisible if, for every and every positive integer , the equation has a solution in . Equivalently, every multiplication map is surjective; solutions need not be unique. The additive group of rational numbers is divisible because . Its quotient group is also divisible: maps to under multiplication by . Unlike , this quotient consists entirely of finite-order elements, showing that divisibility does not imply torsion-freeness. More generally, quotients of divisible groups are divisible. (math.ucla.edu)
For each prime , the Prüfer -group, denoted or , has the concrete realization
The displayed cyclic subgroups have orders and are nested. Multiplication by is surjective because maps to . Multiplication by an integer coprime to is an automorphism on every finite cyclic stage; together these facts establish divisibility for every positive integer. Thus is an infinite divisible torsion group, despite having only finite cyclic groups at its individual stages. (math.ucla.edu)
The classification theorem states that every divisible abelian group has an isomorphism
where the multiplicities and are uniquely determined cardinals, possibly zero or infinite. Parenthesized exponents mean direct sums, not unrestricted products. If denotes the torsion subgroup, then
Here is naturally a rational vector space, while each Prüfer summand contributes one dimension to . In particular, , with exactly one summand for each prime. (math.ucla.edu)
Every abelian group has a unique largest divisible subgroup : it is the sum of all divisible subgroups of . This sum is divisible because each element is a finite sum, whose terms can be divided separately. The splitting mechanism is the equivalence between divisibility and injectivity for abelian groups: every homomorphism from a subgroup into a divisible group extends to . Applying this property to the identity map of gives a retraction , with , and hence
Thus the divisible subgroup is always a direct summand, not merely a subgroup or quotient. (math.ucla.edu)
A group is reduced if its only divisible subgroup is zero. The complement above is reduced, since any divisible subgroup of would also lie in . Although is canonical, the embedded complement need not be: only its isomorphism type, , is determined. For an explicit illustration, in , the largest divisible subgroup is , but every subgroup
is a reduced complement. Indeed, uniquely, and distinct values of give distinct embedded copies of . This directly exhibits the nonuniqueness permitted by the splitting theorem. (math.colorado.edu)
Infinite torsion-free groups and classification limits
For a torsion-free abelian group , the equation , with a nonzero integer, implies . Its torsion-free rank is
where the tensor product is a vector space over the rational numbers. Equivalently, rank is the cardinality of a maximal subset that is linearly independent over : every finite integer relation among its elements has all coefficients zero. Rank measures independence, not the minimum number of generators. A maximal independent subset need not generate ; it becomes a vector-space basis only after rationalization. For torsion-free , the natural map is injective, so a group of finite rank can be represented as a subgroup of . (math.stonybrook.edu)
Every finitely generated torsion-free abelian group is free, hence isomorphic to , but finite rank alone does not imply finite generation. The additive groups and
for a prime , both have rank one and are not free. Indeed, any finite collection of rational numbers has a common denominator , so its integer span lies in . This cannot exhaust ; for , finitely many generators have bounded powers of in their denominators and cannot generate the next power. Both groups require countably infinitely many generators. Since a free group of rank one is infinite cyclic, neither is free. Nevertheless, every finitely generated subgroup of either is cyclic. Thus even being an increasing union of finite-rank free groups does not guarantee freeness. (arxiv.org)
Nonzero torsion-free groups of rank one are precisely the groups isomorphic to subgroups of the additive group of rational numbers. Their classification records divisibility prime by prime. For , define its -height by
allowing the value . Two height sequences have the same type when they have identical infinite coordinates and differ at only finitely many finite coordinates. All nonzero elements of a rank-one group have the same type, and two such groups are isomorphic exactly when their types agree. For the element , all heights in are zero; in , the -height is infinite and all other prime heights are zero; in , every prime height is infinite. These invariants distinguish groups whose rationalizations are all one-dimensional. (karlin.mff.cuni.cz)
Whitehead’s problem, attributed to J. H. C. Whitehead, asks whether an abelian group must be free if every short exact sequence
splits. Here splitting means that a homomorphism satisfies , yielding compatibly with the sequence. The condition is conventionally written ; groups satisfying it are called Whitehead groups. Free groups satisfy it because lifts of basis elements define a section. Conversely, every countable Whitehead group is free, and every Whitehead group is torsion-free. The difficulty therefore lies in uncountable groups. (shelah.logic.at)
In his 1974 paper Infinite abelian groups, Whitehead problem and some constructions, Saharon Shelah proved independence from ZFC already for groups of cardinality , the least uncountable cardinal. Under the constructibility axiom , every Whitehead group of that cardinality is free; under Martin’s axiom together with the negation of the continuum hypothesis, a nonfree Whitehead group of cardinality exists. Assuming ZFC is consistent, both settings are consistent with it. Consequently, ZFC cannot prove either the unrestricted assertion that every Whitehead group is free or its negation. This is a proved limitation of the specified axioms, not an ordinary unresolved conjecture awaiting a proof within them. (shelah.logic.at)
Exact sequences, extensions, and categorical structure
A sequence of homomorphisms is exact at a group when the image of the incoming map equals the kernel of the outgoing map. Thus a short exact sequence
means that is injective, is surjective, and . It identifies with a subgroup of and with the quotient ; the group , together with these maps, is an extension of by . The sequence splits if there is a homomorphism satisfying . In that case gives an isomorphism compatible with the sequence. (stacks.math.columbia.edu)
Exactness alone does not imply splitting. For every integer , reduction modulo gives
This sequence is exact because the kernel of reduction is . If a section existed, the image of the residue class would satisfy both and , which is impossible. By contrast, the inclusion and projection in give a split extension with the same end groups. (math.uchicago.edu)
In categorical terms, abelian groups and their homomorphisms form the category . Each set is itself an abelian group under pointwise addition, , and composition distributes over this addition. The trivial group is a zero object, and finite direct sums serve simultaneously as products and coproducts. A map has kernel and cokernel , while its coimage is naturally isomorphic to its image . These properties make an abelian category. That term describes a category with additive and exactness structure, not a single group with a commutative operation; categories of modules over rings are further examples. (math.uchicago.edu)
The tensor product is generated by symbols , subject to additivity in each variable. Its universal property says that every map additive separately in both arguments factors uniquely through a homomorphism . For example, . Tensoring is right exact: it preserves the final part of an exact sequence, but need not preserve the initial injection. Tensoring the multiplication-by- sequence above with turns its injective map into the zero map between two nonzero cyclic groups. (stacks.math.columbia.edu)
The functor measures this failure of tensoring to preserve injections. It is computed by tensoring a free resolution of with and taking first homology. The resolution yields
Thus it detects elements annihilated by , rather than the quotient produced by tensoring itself. Over , this Tor group vanishes whenever either argument is torsion-free. (math.uchicago.edu)
The Ext functor has a complementary extension-theoretic interpretation: classifies short exact sequences , up to isomorphisms of middle groups inducing the identity on both end groups. Its addition, the Baer sum, combines two extensions using the diagonal map on and the addition map on ; its zero class consists precisely of split extensions. Applying to the same cyclic resolution gives
In particular, , whereas for every free abelian group : every extension with free quotient splits. (stacks.math.columbia.edu)
Roles in topology, number theory, and harmonic analysis
In algebraic topology, integral homology is constructed from free abelian groups generated by continuous maps of -simplices into a space . Boundary homomorphisms satisfy , allowing the quotient
Its elements are cycles considered equivalent when their difference is a boundary. Cohomology instead uses cochains , with an abelian coefficient group , and takes cocycles modulo coboundaries. Continuous maps induce homomorphisms in the same direction on homology and the reverse direction on cohomology; both constructions are invariant under homotopy. (math.mit.edu)
For the torus ,
The two generators represent loops around its two circle factors; a class records their integer multiplicities. When integral homology is finitely generated, its free rank counts independent infinite-order classes—the corresponding Betti number. A torsion class of order is represented by a cycle that is not a boundary although its -fold multiple is. For example, : twice a generating cycle bounds. Torsion therefore records information absent from rational homology. (math.mit.edu)
In algebraic number theory, let be a number field and its ring of integers. Nonzero fractional ideals form an abelian group under multiplication, and principal fractional ideals form a subgroup . The ideal class group
is finite; its order is the class number. Nonzero integral ideals factor uniquely into prime ideals even when elements do not factor uniquely into irreducibles. The class group measures the obstruction to ideals being principal: it is trivial precisely when is a principal ideal domain, equivalently a unique factorization domain. (jmilne.org)
The unit group , by contrast, consists of invertible elements and describes multiplicative arithmetic within the ring. Dirichlet’s unit theorem gives
where is the finite cyclic group of roots of unity in , counts real embeddings, and counts conjugate pairs of complex embeddings. Thus real quadratic fields have unit rank , whereas imaginary quadratic fields have rank . Units also describe the ambiguity in generators: for nonzero elements exactly when for a unit . (jmilne.org)
In harmonic analysis, a locally compact abelian group carries a Hausdorff topology making addition and inversion continuous, with a compact neighborhood at every point. This is additional structure, not part of the abstract group axioms. A continuous character is a homomorphism , where . Characters form the dual group under pointwise multiplication, equipped with the compact-open topology. Pontryagin duality states that
is an isomorphism of topological groups. In particular, it recovers both the group operation and the topology. (math.uh.edu)
The dual of discrete is the compact circle , since a character is determined by ; conversely, the circle’s characters give its discrete dual . The additive real line is self-dual, with characters . More generally, duality exchanges compact and discrete abelian groups. For an integrable function and a chosen Haar measure, the Fourier transform takes the form
Accordingly, Fourier series on the circle have integer-indexed frequencies, while the Fourier transform on the real line has real-indexed frequencies. (people.math.harvard.edu)
References
- 703 Modern Algebra, Groupsocw.mit.edu
- Additive notation for commutative semigroups and abelian groupsmath.berkeley.edu
- Group Theoryjmilne.org
- Mathematics 315: People documentswww2.math.uconn.edu
- Niels Abel (1802–1829) – Biographymathshistory.st-andrews.ac.uk
- Commutator subgroupen.wikipedia.org
- Algebra fact sheetmath.mit.edu
- RES.18-011 (Fall 2021) Full Lecture Notes: Algebra I Student Notesocw.mit.edu
- Section 19.10: Grothendieck's AB conditions—The Stacks projectstacks.math.columbia.edu
- Lemma 13.33.6—The Stacks projectstacks.math.columbia.edu
- Abelian Groupsarxiv.org
- 782 Introduction to Arithmetic Geometry, Lecture #4math.mit.edu
- Modules | Department of Mathematical Sciences | Northern Illinois Universityfaculty.niu.edu
- Group Theory — Finitely Generated Abelian Groupscrypto.stanford.edu
- Finitely Generated Abelian Groupssites.millersville.edu
- Smith Normal Form and Combinatoricsmath.mit.edu
- RES.18-012 (Spring 2022) Full Lecture Notes: Algebra II Student Notesocw.mit.edu
- Instructions for presentationsmath.ucsd.edu
- Math 361 lecture for Friday, Week 8people.reed.edu
- Paper, 'On systems of linear indeterminate equations and congruences' by Henry J [John] Stephen Smithcatalogues.royalsociety.org
- The Smith normal formbruno.stonek.com
- Translating the class of abelian p-groups into an elementary first-order theorymath.berkeley.edu
- The Isomorphism Problem for Computable Abelian p-Groups of Bounded Lengtharxiv.org
- Classification theorems for p-groups and modules over a discrete valuation ringscispace.com
- Abstract Algebra — Chapter 12 Homological algebra · Romyar Sharifimath.ucla.edu
- Your title heremath.ucla.edu
- 32 also define D_max to be the problemweb.math.wisc.edu
- 12 MARCOS MAZARI-ARMIDAmath.colorado.edu
- Lecture Notes on Morse and Conley Theorymath.stonybrook.edu
- Abelian groups with a minimal generating setkarlin.mff.cuni.cz
- Infinite abelian groups, Whitehead problem and some constructionsshelah.logic.at
- Infinite abelian groups, whitehead problem and some constructions — The Hebrew University of Jerusalemcris.huji.ac.il
- Section 12.5: Abelian categories—The Stacks projectstacks.math.columbia.edu
- Section 12.6: Extensions—The Stacks projectstacks.math.columbia.edu
- Section 13.27: Ext groups—The Stacks projectstacks.math.columbia.edu
- Section 10.12: Tensor products—The Stacks projectstacks.math.columbia.edu
- WOMP 2006: Homological Algebra, Part 2math.uchicago.edu
- Lectures on Algebraic Topologymath.mit.edu
- Algebraic Number Theoryjmilne.org
- Pontryagin Dualitymath.uh.edu
- Outline of Pontrjagin dualitypeople.math.harvard.edu