Alphabeta Math
Pipeline-generated
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

✓ 3 results · all verified · 2 also independently AI-judged
Every result on this page is machine-checked by a proof checker and read in full and owner-audited; the judge is an additional, independent cross-model AI review of the proofs. The 1 not AI-judged were verified by owner audit (typically over a confirmed judge false positive), not failures.

Pontryagin Duality for Locally Compact Abelian Groups — Examples

1 · Prerequisites

2 · Summary

These examples and counterexample exercise the duality interface of the companion page on concrete groups, and they are where the convention that the dual carries the compact-open topology can be seen to do real work.

The first example computes annihilators inside Euclidean space, where the dual is again Euclidean space with the pairing x↦e2πi ξ⋅x. It compares the annihilator of a linear subspace, which is the coarse orthogonal complement, with the annihilator of a lattice: for the degenerate inclusion Zk×{0}n−k⊆Rn the annihilator is Zk×Rn−k, so it is a lattice only in the full-rank case, and the quotient-dual identification is read off on standard coordinates as the discrete Fourier pairing. The matrix case H=AZn gives H⊥=A−TZn. The second example computes the bidual maps on the circle and the integers under the published dual identifications: evaluation reproduces exactly the integer n or the point z it started from, so the abstract biduality identification is the natural one on these two groups. The counterexample then shows that the algebraic character group of Z equipped with the discrete topology fails biduality: a Q-linear map of the line obtained from a Hamel basis induces a discontinuous algebraic character of the circle, so the group of continuous characters is strictly larger than Z. With the compact-open topology the same group is compact and biduality holds.

3 · Logical flowchart

4 · Definitions, theorems and proofs

None yet.

5 · Examples, counterexamples and false statements

ExampleConstruction: Literature-sourcedVerification: AI-adaptedprecheck passOpen item page →

Annihilators of closed subgroups of Euclidean space

Example

Let n≥0 and 0≤k≤n be integers. Coordinates are indexed by 0≤i<n (The standard list e:n→Fn with ei(i)=1F and ei(j)=0F for j≠i is an ordered basis of Fn; hence dim⁡FFn=n, and F0 is the zero space with basis ∅ and dimension 0). Work in G=Rn with the dual identified with Rn by ξ↦(x↦e2πi ξ⋅x), computed below from the classification of characters of the line and the product-dual identification (Continuous characters of the real line are exponentials, Duals of finite products and of discrete direct sums). No choice principle is used by the computations below; the general quotient-dual and closed-subgroup identifications cited in the dependency list give the context of clause (2), whose coordinate content is computed directly here.

(1) If V≤Rn is a linear subspace, then V⊥={ξ:ξ⋅v∈Z for all v∈V}; when V=Rk×{0}n−k this is {0}k×Rn−k, the coarse orthogonal complement, whereas for the lattice H=Zk×{0}n−k it is Zk×Rn−k. For k=n the annihilator of the full-rank lattice Zn is again the lattice Zn, while for k<n the annihilator contains the line R ek and is not discrete; so an annihilator is not in general an orthogonal complement.

(2) For H=Zk×{0}n−k the quotient Rn/H is identified with (R/Z)k×Rn−k, and its characters are computed on the standard coordinates by the discrete Fourier pairing: its characters are exactly the maps x+H↦e2πi ξ⋅x with ξ∈Zk×Rn−k, whose pullbacks are precisely the characters of Rn trivial on H.

(3) If A is an invertible n×n real matrix and H=AZn, then H⊥=A−TZn with A−T:=(A−1)T=(AT)−1.

Facts & Assumptions

Given: The group Rn with its standard topology and the dual identification constructed from the character classification of the line and the product-dual identification (Continuous characters of the real line are exponentials, Duals of finite products and of discrete direct sums).

[F1]

Every continuous homomorphism φ:Rn→T is φ(x)=e2πi ξ⋅x for a unique ξ∈Rn, and ξ↦φξ is an isomorphism of topological groups Rn→Rn^: the classification of characters of the line supplies the algebraic bijection, and finite-product duality reduces the topology check to the line. On the line, compact K is bounded, say ∣t∣≤M; continuity of s↦e2πis at 0 makes e2πiξt uniformly close to e2πiξ0t on K when ∣ξ−ξ0∣ is small. Conversely, if ∣ξ−ξ0∣≥ϵ, then t=1/(2∣ξ−ξ0∣)∈[−1/ϵ,1/ϵ] gives ∣e2πi(ξ−ξ0)t−1∣=2, so the uniform ball of radius 1 on this compact interval forces ∣ξ−ξ0∣<ϵ. Uniform balls are compact-open neighbourhoods by The compact-open character group is a Hausdorff topological abelian group, and compactness and boundedness of these line sets follow from Heine-Borel in Rn: with the Euclidean metric a subset of Rn is compact if and only if it is closed and bounded, and the proof by bisection uses no choice principle; the same holds on the real line. The value at ±πi is −1 by exp⁡(x+iy)=ex(cos⁡y+isin⁡y), ∣exp⁡(x+iy)∣=ex, and eiπ+1=0. At n=0 both groups are trivial. (Continuous characters of the real line are exponentials, Duals of finite products and of discrete direct sums, The Pontryagin dual with the compact-open topology)

[F2]

H⊥={γ:γ(h)=1 for every h∈H}, and e2πit=1 exactly when t∈Z, while e2πi(t+k)=e2πit for k∈Z. (The annihilator of a subgroup, ker⁡(exp⁡)=2πiZ, and exp⁡z=exp⁡w exactly when z−w∈2πiZ, The integers as equivalence classes of pairs of naturals)

[F3]

The quotient Rn/H for H=Zk×{0}n−k is (R/Z)k×Rn−k with the product topology: each projection R→R/Z is continuous and open, since the saturation of an open set is the union of its integer translates. The finite product of these maps and identity maps is therefore continuous, open and surjective, with kernel H, and the induced quotient bijection is continuous and open. A continuous character on Rn constant on the cosets of H factors through the quotient map, uniquely and continuously; conversely characters of the quotient pull back to characters of Rn that are trivial on H. (The quotient group G/N and coset product (gN)(hN)=ghN, The discrete, indiscrete, cofinite, cocountable, particular-point and Sierpinski topologies, For a quotient map q:X→Y, a map out of Y is continuous iff its composite with q is; a continuous map on X constant on the fibres of q factors uniquely through q; and a composite of quotient maps is a quotient map, Homeomorphism, open map, closed map, embedding, and what it means for a property to be topological)

[F4]

For a real matrix A, a vector ξ and z∈Zn one has (ATξ)⋅z=ξ⋅(Az) and A−T=(AT)−1; the standard basis vectors e0,…,en−1 lie in Zn, and y∈Rn satisfies y⋅z∈Z for all z∈Zn exactly when y∈Zn. (Transpose is linear and involutive, and (AB)T=BTAT, The transpose AT of a matrix, Invertible matrices and the general linear group GL⁡n(F), The standard list e:n→Fn with ei(i)=1F and ei(j)=0F for j≠i is an ordered basis of Fn; hence dim⁡FFn=n, and F0 is the zero space with basis ∅ and dimension 0, The integers as equivalence classes of pairs of naturals)

Verification

1.1F1F2F4

For a linear subspace V≤Rn, a character γξ lies in V⊥ exactly when ξ⋅v∈Z for every v∈V, by [F1] and [F2]. When V=Rk×{0}n−k and v=t ei for 0≤i<k and t∈R, the condition t ξi∈Z for all real t forces ξi=0 (otherwise take t=1/(2ξi)); the remaining coordinates are unconstrained, so V⊥={0}k×Rn−k.

1.2F1F2F4

For H=Zk×{0}n−k, the condition ξ⋅h∈Z for all h∈H reads ∑0≤i<kξihi∈Z for all integers h0,…,hk−1. Taking h=ei gives ξi∈Z for 0≤i<k, and the remaining coordinates are unconstrained; hence H⊥=Zk×Rn−k. In particular H⊥ is not discrete when k<n, since the nonzero vectors m−1ek lie in it and converge to 0 as positive integers m→∞, and the full-rank case k=n gives (Zn)⊥=Zn.

1.3F1F2F4

For H=AZn with A invertible, γξ∈H⊥ if and only if ξ⋅(Az)∈Z for all z∈Zn, that is (ATξ)⋅z∈Z for all z∈Zn, which by [F4] holds exactly when ATξ∈Zn, that is ξ∈(AT)−1Zn=A−TZn.

2.1F2F3step 1.2

For H=Zk×{0}n−k, the quotient identified in [F3] has the standard coordinates (x0,…,xk−1) mod Z and x′∈Rn−k. A character γξ with ξ∈Zk×Rn−k satisfies γξ(h)=e2πi∑0≤i<kξihi=1 for all h∈H, so it is constant on cosets and factors through the quotient, where it takes the value e2πi(∑0≤i<kξixi+ξ′⋅x′) on the class of x; these are exactly the characters of the quotient, which is the stated discrete Fourier pairing on the torus factor.

3.1step 1.1step 1.2step 1.3step 2.1∎

Clauses (1), (2) and (3) are proved in steps 1.1 and 1.2, step 2.1 and step 1.3; the example also records that "the annihilator of a lattice is a lattice" holds only in the full-rank case of clause (3), not for the degenerate subgroups Zk×{0}n−k with k<n.

ExampleConstruction: Literature-sourcedVerification: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)Open item page →

The bidual map on the circle and the integers

Example

Assume the Axiom of Choice (The Axiom of Choice) and Dependent Choice (The axiom of dependent choice: a relation in which every element is related to something admits an N-indexed chain); the two identifications of values below are direct computations from the classifications of the characters of Z and T given in the Facts, and the continuity statement is exactly Pontryagin biduality: the evaluation map is a topological isomorphism. Under the identifications Z^≅T, z↦(n↦zn), and T^≅Z, k↦(z↦zk), the evaluation map ΦZ:Z→Z^^ is the identity of Z, and ΦT:T→T^^ is the identity of T. In particular the abstract biduality identification is the natural one on these two groups, not merely an abstract isomorphism.

Facts & Assumptions

[F1]

A homomorphism γ:Z→T is determined by γ(1), and every z∈T occurs; since Z carries the discrete topology every such homomorphism is continuous. Hence z↦(n↦zn) is an algebraic isomorphism T→Z^. Compact subsets of the discrete space Z are finite (their cover by singletons has a finite subcover), so the preimage of each compact-open subbasic set is a finite intersection of open conditions on powers of z; the inverse is evaluation at 1, whose preimages of open sets are subbasic open sets. Thus both directions are continuous. (The integers as equivalence classes of pairs of naturals, The discrete, indiscrete, cofinite, cocountable, particular-point and Sierpinski topologies, The Pontryagin dual with the compact-open topology, Topological group: multiplication and inversion are continuous)

[F2]

Every continuous homomorphism χ:T→T is χ(z)=zk for a unique k∈Z. Indeed, writing T≅R/Z through the unit-circle isomorphism, χ corresponds to a continuous homomorphism ψ:R→T with ψ(1)=1; by the classification of characters of the line there is a unique ξ∈R with ψ(t)=e2πiξt, and ψ(1)=1 forces e2πiξ=1, that is ξ∈Z by the kernel of the complex exponential. The dual of the compact circle is discrete by Compact groups have discrete duals and discrete groups have compact duals, so this bijection from the discrete group Z is a topological isomorphism. (The multiplicative unit circle is a compact metrizable topological abelian group, Continuous characters of the real line are exponentials, ker⁡(exp⁡)=2πiZ, and exp⁡z=exp⁡w exactly when z−w∈2πiZ, The Pontryagin dual with the compact-open topology)

[F3]

For every locally compact Hausdorff abelian group G the evaluation map ΦG(x)(γ):=γ(x) is an isomorphism of topological groups G→G^^. (Pontryagin biduality: the evaluation map is a topological isomorphism)

Verification

1.1F1F2

Under the identification [F1], the element z∈T corresponds to the character n↦zn of Z, so for n∈Z one has ΦZ(n)(z)=zn; reading this character of T through the identification of [F2], which assigns to k∈Z the character z↦zk, gives the integer n. Hence ΦZ is the identity of Z.

1.2F1F2

Under the identification of [F2], the integer k corresponds to the character z↦zk of T, so for z∈T one has ΦT(z)(k)=zk; reading this character of Z through the identification [F1] gives back z. Hence ΦT is the identity of T.

2.1F3step 1.1step 1.2∎

By [F3] the maps ΦZ and ΦT are topological isomorphisms, and steps 1.1 and 1.2 identify them with the identity maps of Z and T; so the biduality identification is the natural one on these groups, which is the example.

CounterexampleConstruction: AI-generatedVerification: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)Open item page →

Forgetting the compact-open topology destroys Pontryagin duality

Statement refuted

The biduality conclusion of Pontryagin duality holds for every algebraic character group when the group is equipped with the discrete topology: if X is the algebraic character group Hom⁡(Z,T) of the discrete group Z with the discrete topology, then the evaluation map Z→Hom⁡cts(X,T) is an isomorphism.

Facts & Assumptions

[F1]

The map z↦(n↦zn) is an isomorphism of topological groups T→Z^, because a homomorphism out of Z is determined by its value at 1 and Z is discrete. Compact subsets of the discrete space Z are finite (their cover by singletons has a finite subcover), so the preimage of each compact-open subbasic set is a finite intersection of open conditions on powers of z; the inverse is evaluation at 1, whose preimages of open sets are subbasic open sets. Thus both directions are continuous. Every continuous homomorphism χ:T→T is χ(z)=zk for a unique k∈Z: writing T≅R/Z, the character corresponds to a continuous homomorphism ψ:R→T with ψ(1)=1, the classification of characters of the line gives ψ(t)=e2πiξt for a unique real ξ, and ψ(1)=1 forces ξ∈Z by the kernel of the complex exponential. (The integers as equivalence classes of pairs of naturals, The discrete, indiscrete, cofinite, cocountable, particular-point and Sierpinski topologies, The multiplicative unit circle is a compact metrizable topological abelian group, Continuous characters of the real line are exponentials, ker⁡(exp⁡)=2πiZ, and exp⁡z=exp⁡w exactly when z−w∈2πiZ, The Pontryagin dual with the compact-open topology)

[F2]

Assuming the Axiom of Choice, R has a Hamel basis B over Q: every real is a finite Q-linear combination of basis elements in exactly one way, a Q-linear map R→R is determined by its values on B, and the coefficient of any basis element is a well-defined Q-linear map. (Assuming the Axiom of Choice, R has a Hamel basis over Q: there is B⊆R such that every real is a finite Q-linear combination of elements of B in exactly one way, and each basis vector carries a well-defined Q-linear coefficient map)

[F3]

A group homomorphism on a discrete group is continuous, and every function on a discrete space is continuous; composition of continuous homomorphisms is a continuous homomorphism, and Z⊆R consists of the integer multiples of 1. (The discrete, indiscrete, cofinite, cocountable, particular-point and Sierpinski topologies, Monoid homomorphism and group homomorphism, Topological group: multiplication and inversion are continuous, The integers as equivalence classes of pairs of naturals)

[F4]

The assumed AC implies DC (AC implies DC implies countable choice), so biduality applies with its full choice hypotheses. With the compact-open topology the group X=Z^ is compact, because the dual of a discrete abelian group is compact under the Axiom of Choice, and Pontryagin biduality identifies its dual with Z through the evaluation isomorphism. (Compact groups have discrete duals and discrete groups have compact duals, Pontryagin biduality: the evaluation map is a topological isomorphism, Open cover, subcover, and compact topological space; a compact subset is a subspace that is compact in its own right, The Pontryagin dual with the compact-open topology)

Counterexample

1.1F1F3

Identify X=Hom⁡(Z,T) with T by [F1] and equip it with the discrete topology. Since a homomorphism on the discrete group X is automatically continuous by [F3], the group of continuous characters Hom⁡cts(X,T) is exactly the group Hom⁡alg(T,T) of all algebraic homomorphisms. The evaluation map Z→Hom⁡cts(X,T) is n↦(z↦zn); with the compact-open topology on T, [F1] says these power maps are exactly its continuous characters.

1.2F2

By [F2] choose a Hamel basis B and expand 1 in it. Choose a basis element b0 whose rational coefficient r=Λb0(1) is nonzero, and put T=Λb0/r:R→R. Then T is Q-linear with T(1)=1, hence T(n)=n for every integer n. The complement clause of the Hamel-basis supplier gives another basis vector b1≠b0; then T(b1)=0 while b1≠0. Consequently T cannot be multiplication by a real scalar: its value at 1 would force that scalar to be 1, contradicting its value at b1. Since T(Z)⊆Z, it induces the well-defined homomorphism fT(x+Z)=T(x)+Z of R/Z. This uses the supplied coordinate map and does not assume the arbitrary Hamel basis contains 1.

2.1F1F2F3step 1.2

Let fT be the homomorphism induced by T of step 1.2. As a homomorphism from the discrete group X it belongs to Hom⁡cts(X,T) by [F3]. If it were continuous for the compact-open topology on T, then by [F1] there would be k∈Z with fT(z)=zk for all z, that is T(x)−kx∈Z for all x∈R. The map G(x):=T(x)−kx is then Q-linear with G(R)⊆Z; for every x and every n≥1 one has G(x)=n G(x/n)∈nZ, so G(x)∈⋂n≥1nZ={0} and T=k id, contradicting the choice of T. Hence fT is an algebraic homomorphism continuous on the discrete X but not continuous on the compact-open circle.

3.1F1F3step 2.1

The homomorphism fT lies in Hom⁡cts(X,T)=Hom⁡alg(T,T) but is not one of the maps z↦zk, so it is not in the image of the evaluation map Z→Hom⁡cts(X,T), whose image is exactly {z↦zn:n∈Z} by [F1] and is a copy of Z. The evaluation map is injective, since z↦zn determines n, but it is not surjective: the discrete algebraic character group X fails the biduality conclusion, and no topological constraint on X beyond discreteness is available to repair it.

4.1F4step 3.1∎

By contrast, with the compact-open topology the same group X=Z^≅T is compact and [F4] gives Hom⁡cts(X,T)≅Z through the evaluation map. Thus it is the discarded compact-open topology, not the algebraic character group, that carries Pontryagin biduality; the statement refuted above is false.

Sources