Skip to content

Dual System

Pair two vector spaces by a nondegenerate bilinear form so each separates points of the other and determines weak, polar, and compatible locally convex structures.

Version
v3 · 2026-09-06 · History
Domain-specific #
1723
Origin domain
mathematics
Subdomain
functional analysis
Aliases
Dual pair, Paired vector spaces, Separating duality

Core Idea

A dual system or dual pair is a triple \((X,Y,\langle\cdot,\cdot\rangle)\) consisting of vector spaces over the same scalar field and a bilinear pairing that separates points on both sides: every nonzero \(x\in X\) is detected by some \(y\in Y\), and every nonzero \(y\in Y\) is detected by some \(x\in X\). Thus (Y) embeds into the algebraic dual of (X), and (X) embeds into the algebraic dual of (Y), without requiring either to be the other's entire algebraic dual.[1]

The pairing is the governing relation. It generates annihilators, polars, weak topologies, and a family of locally convex topologies having exactly the selected continuous dual.

Structural Signature

  • Two vector spaces (X) and (Y) over a common field.
  • A scalar-valued bilinear pairing.
  • Point separation of (X) by (Y).
  • Point separation of (Y) by (X).
  • Canonical injections into opposite algebraic duals.
  • Annihilator and polar operations reversing inclusion.
  • Weak topologies σ((X,Y)) and σ((Y,X)).
  • Compatible locally convex topologies with prescribed continuous duals.
  • A strongest compatible Mackey topology under the standard theory.
  • Explicit real, complex-bilinear, or sesquilinear convention.

What It Is Not

It is not any pair of objects called dual, nor categorical duality. It is not necessarily an inner-product space: symmetry, positivity, norm, completeness, and an identification (X=Y) are absent. It is not necessarily the canonical pair of a topological vector space with its full continuous dual, and nondegeneracy does not make the pairing surjective onto either full algebraic dual.

Scope of Application

Dual systems organize locally convex spaces, weak and weak-star convergence, convex polarity, distribution-test-function pairings, and optimization duality. Trèves uses paired spaces to control weak topologies and compatible dual structures in distribution theory.[2] Narici and Beckenstein develop polar topologies and the Mackey–Arens theorem from the same pairing data.[3]

Clarity

State the scalar field, whether the pairing is bilinear or sesquilinear, which argument is linear, and whether both separation conditions hold. Distinguish the algebraic dual, continuous dual under a stated topology, and the designated partner (Y). Name the topology when using boundedness, continuity, closure, or compactness.

Manages Complexity

The abstraction lets one choose exactly the observables needed to distinguish states. Rather than commit immediately to a norm or maximal dual, it starts with evaluation data and derives the weakest topology making those observations continuous. Stronger compatible topologies can then be compared without changing the continuous dual.

Abstract Reasoning

  1. Select spaces (X) and (Y) and define their pairing.
  2. Verify that each side separates the other.
  3. Embed each space in the opposite algebraic dual.
  4. Form annihilators or absolute polars of relevant subsets.
  5. Generate the weak topology from point evaluations by the partner space.
  6. Determine equicontinuous, bounded, and weakly compact sets.
  7. Compare compatible locally convex topologies between the weak and Mackey bounds.
  8. Use bipolar and separation results only under their stated convexity and topology conditions.

Bourbaki's treatment shows how polarity and duality supply the organizing language for topological vector spaces.[4]

Knowledge Transfer

The portable pattern is choose a family of reciprocal probes that separates points, then let those probes determine admissible topology and closure. It transfers to state-observable pairings. The proposed immediate parent is Duality.

Examples

A locally convex space (E) and its continuous dual (E'), with evaluation ⟨(x,f)⟩=(f(x)), form a dual system when continuous linear functionals separate points. A Banach space (E) paired with (E') yields weak topology σ((E,E')); the reversed pair yields the weak-star topology on (E').

The finitely supported sequences can be paired with all scalar sequences by the finite sum ∑(x_ny_n). The partner spaces need not carry the same topology or be norm-duals of one another.

Structural Tensions

  • Chosen partner versus full algebraic dual.
  • Weak observability versus stronger compatible topology.
  • Algebraic separation versus topological completeness.
  • Symmetric notation versus asymmetric topological roles.
  • Bilinear convention versus complex sesquilinear convention.

Structural–Framed Character

Reciprocal point separation and induced topology are structural. Vector spaces, linear functionals, polars, and locally convex compatibility are constitutive. The identity is domain-specific.

Structural Core vs. Domain Accent

The structural core is two carriers -> reciprocal probes -> separation -> induced weak structure. The domain accent is bilinear functional-analytic duality.

Duality is the proposed immediate parent. Relation, Observation, Separation, and Topology are related primes. Dot Product and Graph Duality are narrower or differently framed domain-specific neighbors.

The prospective queue contains one strict edge to prime:duality. No live DAG mutation is authorized.

Relationships to Other Abstractions

Local relationship map for Dual SystemParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Dual SystemDOMAINPrime abstraction: Duality — is a kind ofDualityPRIME

Current abstraction Dual System Domain-specific

Parents (1) — more general patterns this builds on

  • Dual System is a kind of Duality Prime

    Duality is the proposed immediate parent.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Dual System sits in a sparse region of the domain-specific corpus (84th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

  • Algebraic dual space.
  • Continuous dual without a stated topology.
  • Inner product or dot product.
  • Dual object in category theory.
  • Pontryagin duality.
  • A primal–dual optimization pair.

References

[1] Helmut H. Schaefer and Manfred P. Wolff, Topological Vector Spaces, 2nd ed. (Springer, 1999), pp. 122–128, doi:10.1007/978-1-4612-1468-7. registry ↩

[2] François Trèves, Topological Vector Spaces, Distributions and Kernels (Dover, 2006), pp. 368–377. registry ↩

[3] Lawrence Narici and Edward Beckenstein, Topological Vector Spaces, 2nd ed. (CRC Press, 2011), pp. 225–273, doi:10.1201/9781584888674. registry ↩

[4] Nicolas Bourbaki, Topological Vector Spaces: Chapters 1–5 (Springer, 1987), doi:10.1007/978-3-642-61715-7. registry ↩