Locally Discrete Collection¶

A family of subsets of a topological space for which every point has an open neighborhood meeting at most one family member.

Version: v1 · 2026-10-07 · History
Domain-specific #: 13933
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomains: General Topology, Metrization Theory → Mathematics
Aliases: Discrete Collection of Sets

Core Idea¶

A locally discrete collection is a family of subsets of a topological space in which every point has an open neighborhood meeting at most one member of the family. The condition concerns a chosen family, not whether all points of the space are isolated. A point outside every member must still pass the neighborhood test, so a sequence of disjoint sets accumulating at an omitted point can make the family fail.^[1]

R. H. Bing's 1951 paper calls this a discrete collection and states an equivalent closure condition. It is stronger than pairwise disjointness and stronger than local finiteness. One such family may be useful in an open-cover refinement or as one layer of a base, but it need not cover the space or form a base by itself.^[1]

Structural Signature¶

Ambient topological space: a set X with specified open sets τ; changing τ can change which neighborhoods witness the condition.^[1]
Chosen family of subsets: its members are the sets whose local encounters are counted. They need not be singletons or themselves open, though Bing's two application families are open.^[1]
Pointwise witness: for each point x of X, some open neighborhood U of x meets at most one distinct member of the family. The witness can vary with x.^[1]
At-most-one bound: meeting two members in every neighborhood of even one point is a decisive failure, including when that point belongs to neither member.^[1]
Family-level consequence: the condition implies pairwise disjoint members and local finiteness, but those weaker properties alone do not imply the condition.^[1]

The empty family passes vacuously; useful cover and base applications impose further coverage conditions. Index labels do not alter the topological test, which counts distinct family members.^[1]

What It Is Not¶

A discrete space has every singleton open. A locally discrete collection may live in a non-discrete, even nonmetrizable, space. Pairwise disjoint sets may crowd toward a limit point, leaving every neighborhood there meeting many members. A locally finite family only requires finitely many encounters near each point, possibly two or more. A countable union of locally discrete layers is a different construction: its combined family need not remain locally discrete.^[1]

Nor is the mere existence of one locally discrete family a metrization theorem. Bing's criterion for a regular space uses a countable sequence of discrete open families with a local-basis property. In a discrete space a single family of open singletons can be a base; the countable-layer design is needed in general, not because one layer is logically forbidden to suffice.^[1]

Scope of Application¶

The notion applies to families of subsets of any specified topological space. Bing constructs discrete open families at fixed stages in a metric-space argument and also gives a nonmetrizable lower-limit-line example where an open cover has a discrete open refinement. Both are positive family cases, though only the metric construction is part of the countable-basis necessity argument for metrizability.^[1]

The relation is relative to the topology and to the selected family. It does not assert that every open cover in every space has such a refinement, or that a family in one topology has the same status when the open sets are changed.^[1]

Clarity¶

Bing's printed definition says a discrete collection has pairwise disjoint closures and that the union of the closures of every subcollection is closed. The point-neighborhood formulation used here is a mathematical equivalence, not his verbatim sentence. In one direction, a neighborhood meeting at most one member prevents a point from lying in two member closures; for a point outside the closure union of a subcollection, its local witness can be reduced to avoid the possible sole member, proving that union closed.^[1]

Conversely, Bing's closure properties let a point avoid every member except possibly one: if it lies in one closure, avoid the closed union of all the other closures; if it lies in none, avoid the closed union of all closures. The resulting open neighborhood meets at most one member. This argument shows why the condition must be checked at points outside the union too.^[1]

Manages Complexity¶

A base or an open-cover refinement can contain many sets. Testing all pairwise intersections misses accumulation at external points. The local condition compresses that global-looking behavior into one witness at each point and allows a large construction to be organized into separately discrete layers. The shorthand must retain the layer boundary: Bing's countable sequence of families has a stronger basis role than any single arbitrary family.^[1]

His nonmetrizable Example D is a useful warning against overreading the shorthand. Even a space where every open cover has a discrete open refinement need not satisfy the countable local-basis condition of his metrization theorem.^[1]

Abstract Reasoning¶

To test a family, fix an arbitrary point, including one outside all members. Find an open neighborhood of that point and count how many distinct members it meets. A witness with zero or one passes at that point; if every neighborhood meets two, the whole family fails. Pairwise disjointness is only a preliminary check, not a substitute.^[1]

When a theorem mentions a sequence of discrete families, test each indexed layer separately and then verify its extra coverage or basis hypothesis. Do not silently replace those statements with one discrete family or infer metrizability from one successful local test.^[1]

Knowledge Transfer¶

The family test transfers from a metric space to a nonmetrizable topological space because it asks only for open neighborhoods, not distances. Bing's metric construction and lower-limit-line example exhibit the same local condition under unlike ambient topologies. What does not transfer automatically is the larger theorem's countable-basis property.^[1]

The live Topological Space entry is a strict ambient prerequisite: without an open-set structure there is no neighborhood condition. This is a composition/presupposes edge, not a claim that a family is itself a kind of space. The live Discreteness Prime names a broader separated-state idea, but its isolated-point reading does not establish a strict all-instance parent when a chosen family is discrete in a non-discrete space.^[1]

Examples¶

Canonical: one metric-space layer in Bing's construction¶

In the necessity proof of Bing's metrization theorem, at each selected scale and index he constructs a discrete collection of small closed sets and corresponding discrete open neighborhoods. Choose one of those open families, rather than the union across all stages. Its members are subsets of the metric space; Bing's discreteness claim gives each point a neighborhood meeting at most one of them.^[1]

Mapped back: the metric space supplies the ambient topology; one fixed open family supplies the members; its discrete property supplies pointwise witnesses and the at-most-one bound. The countable sequence of such layers and its basis property are separate requirements of the theorem, not part of the identity of one locally discrete family.^[1]

Bing's Example D uses a regular separable lower-limit topology on the line that he shows is not metrizable. For an open cover, he describes an open refinement that he identifies as discrete. Select that refinement as the family: its ambient space is nonmetrizable, but the same local at-most-one test holds.^[1]

Mapped back: the lower-limit-line topology supplies different open neighborhoods; the chosen refinement supplies the member subsets; Bing's discrete-refinement statement supplies the pointwise witnesses. The example demonstrates that a discrete family does not imply metrizability or a countable discrete base. It does not imply that every disjoint open family there is discrete.^[1]

Structural Tensions¶

The source establishes no intrinsic pair of competing costs for one locally discrete collection. Strong local separation is the defining condition, while coverage or being a base is an additional property a construction may seek. In some spaces a single family can be both locally discrete and a base; in Bing's general metric construction countably many separately discrete layers supply the desired basis. Treating their distinction as a universal tradeoff would be false.^[1]

The useful diagnostic is whether a claim concerns one family, a cover refinement, or a sequence of families whose union is a base. Those are different mathematical objects and hypotheses, not opposing poles that every family must optimize.^[1]

Structural–Framed Character¶

Evaluative weight: discrete is a formal condition here, not praise for a sparse or efficient design. Human-practice dependence: the mathematical definition does not depend on a particular institution, although mathematicians choose families for a proof. Institutional origin: Bing uses the term in general topology and metrization theory; neither a journal nor a construction procedure is constitutive. Vocabulary travel: everyday “discrete” and the live Discreteness Prime are broader; the local family condition does not follow from the word alone. Import versus recognition: verify neighborhoods in the actual topology rather than import a picture of separated intervals.^[1]

The portable skeleton may be a local at-most-one encounter relation among chosen members, but whether it extends literally beyond topological neighborhoods is a future-Prime question. Its character: a formal, topology-relative family property that remains meaningful in metric and nonmetric spaces but presupposes the ambient open-set structure.^[1]

Structural Core vs. Domain Accent¶

The core is each point has an open neighborhood meeting at most one member of a chosen family. The domain-bound differentia is the topological meaning of point and open neighborhood. Remove the topology and the test is undefined; replace “at most one” with “finitely many” and one has local finiteness instead. Metric versus lower-limit-line topology, open refinements versus base layers, and chosen scales are case accents.^[1]

The named entry does not clear the Prime bar by calling all forms of separation “discrete.” Its precise local relation presupposes Topological Space, and a general at-most-one pattern across domains would need separate literal cases and a different candidate identity. The strict edge to the ambient space captures an actual prerequisite now; the broader cross-domain relation remains a future-Prime question rather than a claimed parent.^[1]

This entry presupposes Topological Space.

A locally discrete collection, in every case, depends on a Topological Space. Every instance needs a pair (X, τ) so that its neighborhoods and closures exist, and a topological space can exist without any designated family. The collection is not a kind of space: a family is not a space. Nor is the space a part of it: the space is the required ambient parameter, not a member inside the collection.^[1]

Discreteness is thematically related, but it is not established as broader than this entry. Indexed Family supplies optional addressed organization, while Bing's family condition concerns sets regardless of how they are labeled. Collectionwise Normal Space is a property of spaces that uses discrete families as inputs, not a genus of the family. These comparisons leave Topological Space as the one established relation.^[1]

Relationships to Other Abstractions¶

Current abstraction Locally Discrete Collection Domain-specific

Parents (1) — more general patterns this builds on

Locally Discrete Collection presupposes Topological Space Domain-specific

A locally discrete collection requires an ambient topology to define each point's open neighborhood and test encounters with family members.

Hierarchy paths (5) — routes to 3 parentless roots

Locally Discrete Collection → Topological Space → Closure

Show alternative paths (4)

Neighborhood in Abstraction Space¶

Locally Discrete Collection sits in a sparse region of the domain-specific corpus (76^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Point-Set Topology Foundations (17 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With¶

Discrete topology: every point isolated, not a condition on one selected family. Pairwise disjoint family: members may still accumulate at a point. Locally finite family: a neighborhood may meet more than one member. Sigma-discrete collection or base: a countable union of discrete families, not necessarily one discrete family. Metrizability: requires Bing's full regularity and countable local-basis hypotheses. A disjoint refinement: not automatically discrete without the local test.^[1]

References¶

[1] R. H. Bing, “Metrization of Topological Spaces,” Canadian Journal of Mathematics 3 (1951), pp. 175–186, especially p. 176 (discrete collection definition), p. 179 (Theorem 3 necessity construction), and p. 181 (Example D); original journal PDF. https://www.cambridge.org/core/services/aop-cambridge-core/content/view/48C1A50A9E249D05BD7054529F93BAA1/S0008414X00030923a.pdf/metrization-of-topological-spaces.pdf registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s ↩t ↩u ↩v ↩w ↩x ↩y ↩z ↩27 ↩28 ↩29 ↩30 ↩31 ↩32 ↩33