Pseudometric space¶

Equip a set with a symmetric, nonnegative, triangle-inequality distance that may assign zero separation to distinct points, with metric quotient obtained by identifying zero-distance classes.

Version: v1 · 2026-08-30 · History
Domain-specific #: 2567
Origin domain: mathematics
Subdomain: generalized metric spaces
Aliases: Semimetric space (zero-separation convention), Pseudo-metric space

Core Idea¶

A pseudometric space is a pair \((X,d)\) in which \(d:X\times X\to[0,\infty)\) satisfies \(d(x,x)=0\), symmetry \(d(x,y)=d(y,x)\), and the triangle inequality \(d(x,z)\le d(x,y)+d(y,z)\), but may have \(d(x,y)=0\) for distinct \(x\) and \(y\). The missing identity-of-indiscernibles axiom is the whole distinction from a metric space. Open balls still define a topology, although points at zero distance have exactly the same neighborhoods and cannot be topologically separated.^[1]

Define \(x\sim y\) when \(d(x,y)=0\). Symmetry and the triangle inequality make this an equivalence relation. The function descends to a genuine metric \(\bar d([x],[y])=d(x,y)\) on the quotient \(X/{\sim}\), because the triangle inequality ensures the value is independent of representatives and zero distance now means the classes coincide. Many pseudometrics arise from a seminorm \(p\) by \(d(x,y)=p(x-y)\), from ignoring selected coordinates, or from comparing observable features rather than underlying objects.^[2]

Terminology is not uniform: some authors use semimetric for a distance lacking the triangle inequality, while others use it for what is here called a pseudometric. A pseudometric is not merely an approximate metric and its zero-distance pairs are not measurement mistakes. It does not induce a Hausdorff topology unless zero-distance points are already identical. A family of pseudometrics may nevertheless separate points collectively and generate a Hausdorff uniformity or topology. Quotienting changes the carrier and must be stated rather than silently replacing original points by classes.^[3]

Structural Signature¶

Carrier set. Arbitrary points supply objects to be compared.
Distance-like map. A nonnegative real value is assigned to each ordered pair.
Reflexive zero. Every point has zero self-distance.
Symmetry. Reversing the pair leaves the value unchanged.
Triangle inequality. Indirect comparison cannot undercut the declared bound on direct separation.
Zero-distance kernel. Distinct points may lie in the same indistinguishability class.
Ball topology. Open balls generate neighborhoods that cannot distinguish kernel-equivalent points.
Metric quotient. Collapsing zero-distance classes restores identity of indiscernibles.

What It Is Not¶

Not a metric space. A metric additionally forces zero distance to imply point equality.
Not an approximate metric. All stated axioms are exact; only separation is relaxed.
Not a dissimilarity score. Arbitrary scores need not be symmetric or satisfy a triangle inequality.
Not a seminorm. A seminorm is a function on vectors; it can induce a pseudometric by difference.
Not a quotient metric. The quotient is a derived space whose points are equivalence classes.
Not one universal semimetric convention. Authors use semimetric for several inequivalent axiom packages.

Scope of Application¶

The abstraction is literal wherever practitioners can identify the same constitutive roles, apply the same boundary tests, and obtain the same kind of output. The following habitats are uses of Pseudometric space itself, not metaphors based only on resemblance.

Functional analysis. Turning seminorms and locally convex families into distance-like comparisons.
Quotient construction. Identifying objects observationally indistinguishable under the chosen comparison.
Probability. Comparing random variables modulo almost-sure equality or distributions modulo selected tests.
Computer science. Comparing states or behaviors while abstracting from unobserved components.
Topology. Generating neighborhoods before asking whether separation axioms hold.
Data analysis. Encoding feature-level proximity that deliberately ignores nuisance variation.

Clarity¶

A clear account of Pseudometric space must preserve the recognition invariant stated in the Core Idea rather than rely on the title alone. List every distance axiom used and state explicitly that distinct zero-distance points are allowed. Declare the source's semimetric convention rather than treating the word as a fixed synonym. Compute the zero-distance equivalence relation before making separation claims. When quotienting, prove representative independence and distinguish original points from equivalence classes. These declarations are not editorial extras: each changes what observations count, which transformations are licensed, and what conclusion can be drawn. A reader should be able to reconstruct the input, the operative rule, the output, and at least one defeater from the account without consulting an implementation or guessing an unstated convention.

Manages Complexity¶

Pseudometric space manages complexity by replacing a diffuse field of observations or possible operations with a bounded role structure: carrier set supplies arbitrary points supply objects to be compared.; distance-like map supplies a nonnegative real value is assigned to each ordered pair.; reflexive zero supplies every point has zero self-distance.; symmetry supplies reversing the pair leaves the value unchanged.; triangle inequality supplies indirect comparison cannot undercut the declared bound on direct separation.. The compression is useful because it localizes disagreement. One can ask whether the input was properly formed, whether a constitutive relation held, whether an alternative explanation defeats the inference, or whether the output was overinterpreted. The same compression can mislead when its discarded detail is exactly what the decision requires. A reference-grade use therefore reports both the invariant retained and the information intentionally lost.

Abstract Reasoning¶

Specify the carrier and formula for the distance-like comparison.
Verify nonnegativity, self-zero, symmetry, and the triangle inequality.
Search for distinct pairs at zero distance to determine whether the space is nonmetric.
Form the zero-distance equivalence relation and prove transitivity using the triangle inequality.
Analyze open balls and identify which points share all neighborhoods.
Descend the formula to equivalence classes and verify a genuine metric.
State whether the original or quotient space is used in later arguments.
Test the candidate interpretation against the nearest named confusable rather than accepting a shared surface feature.
State the conclusion at the same scope as the source conditions, and retain uncertainty or nonuniqueness where the construct does not remove it.

Knowledge Transfer¶

The strict upward abstraction is Comparison. Pseudometric Space instantiates Comparison because its function places every pair of points in a shared nonnegative scale and reads off their distance relation, specialized by metric-style axioms without point separation. Within generalized metric spaces, the full mechanism transfers literally when the same roles and boundary tests recur. Beyond that domain, only the parent-level skeleton should travel. Reusing the label Pseudometric space after removing its constitutive vocabulary would hide a change of mechanism behind an analogy. The honest transfer rule is therefore two-stage: recognize the domain-specific pattern first, then lift only the parent relation that remains invariant under a substrate change.

Examples¶

Canonical¶

On \(\mathbb R^2\), define \(d((x_1,x_2),(y_1,y_2))=|x_1-y_1|\). The function is nonnegative, symmetric, and obeys the triangle inequality. Points with the same first coordinate have zero distance even when their second coordinates differ, so it is not a metric. The quotient by zero distance can be identified with \(\mathbb R\), where the descended metric is ordinary absolute difference.

Mapped back: input and conventions → constitutive role test → bounded output → explicit interpretation and defeater check.

Applied / In Practice¶

Two programs are compared only by the sequence of publicly observable outputs they produce on a declared test set. Implementations with different internal states can receive pseudodistance zero when all tested outputs agree. The metric quotient treats those implementations as one observational class. That is useful for the stated abstraction, but it cannot support claims about internal equivalence or behavior on untested inputs.

Mapped back: field observation or problem → candidate recognition → confusable and limit checks → appropriately scoped conclusion.

Structural Tensions¶

T1: Zero distance versus identity. Ordinary geometric intuition treats zero as equality. Diagnostic: Exhibit the kernel classes before calling points distinguishable.
T2: Topology versus Hausdorff separation. Open balls exist even when distinct points share all neighborhoods. Diagnostic: Check the T0 or Hausdorff axiom explicitly rather than inferring it from balls.
T3: Original carrier versus quotient. Collapsing kernel classes changes what counts as a point. Diagnostic: Track brackets around equivalence classes throughout the argument.
T4: One pseudometric versus a separating family. Individual comparisons can collapse points while a family separates them. Diagnostic: Test whether for every distinct pair some family member has positive distance.
T5: Terminology versus axioms. Semimetric has incompatible meanings across sources. Diagnostic: Quote or restate the complete axiom package.
T6: Autonomy versus Comparison. Comparison supplies a shared-frame relation; pseudometric adds numeric axioms and a quotient mechanism. Diagnostic: Remove symmetry, triangle inequality, and zero-kernel quotient and see whether the object remains.

Structural–Framed Character¶

Pseudometric Space is strongly structural: exact axioms determine membership, topology, kernel, and quotient, while the choice of observed features frames which distinctions are intentionally discarded. The five framing criteria point in a consistent direction. Evaluative weight is limited to whether the defining conditions are met, not whether the outcome is desirable. Human practice matters to the extent that experts choose conventions, instruments, or reporting thresholds, but those choices do not make every verdict arbitrary. Institutional history explains the name and standard use; it does not replace the recognition rule. The operative vocabulary travels within the home field and closely adjacent subfields, while transfer farther away requires translation to the parent prime. Thus recognition remains disciplined even where interpretation is defeasible.

Structural Core vs. Domain Accent¶

What is skeletal. Pseudometric Space instantiates Comparison because its function places every pair of points in a shared nonnegative scale and reads off their distance relation, specialized by metric-style axioms without point separation. This is the part that can be expressed without the candidate's specialist nouns.

What is domain-bound. The irreducible accent is a nonnegative pair function, symmetry, triangle inequality, failure of identity of indiscernibles, zero-distance equivalence classes, ball topology, and metric quotient. Remove those elements and the result is no longer Pseudometric space; it is only the parent relation or a loose analogy.

Why this does not clear the prime bar. The name does not recur with unchanged diagnostics across three independent domains. What transfers is already represented by prime:comparison. The candidate remains autonomous because its in-domain recognition rule, failure modes, and consequences are stable, but its vocabulary and interventions do not float free of the home substrate.

Pseudometric Space instantiates Comparison because its function places every pair of points in a shared nonnegative scale and reads off their distance relation, specialized by metric-style axioms without point separation.

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

Relationships to Other Abstractions¶

Current abstraction Pseudometric space Domain-specific

Parents (1) — more general patterns this builds on

Pseudometric space is a kind of Comparison Prime

Pseudometric Space instantiates Comparison because its function places every pair of points in a shared nonnegative scale and reads off their distance relation, specialized by metric-style axioms without point separation.

Hierarchy path (1) — routes to 1 parentless root

Pseudometric space → Comparison → Self Checking

Neighborhood in Abstraction Space¶

Pseudometric space sits in a sparse region of the domain-specific corpus (79^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Metric Geometry & Approximation (13 abstractions)

Nearest neighbors

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

Not to Be Confused With¶

Metric space. Adds identity of indiscernibles and therefore a Hausdorff ball topology.
Semimetric space. May mean a triangle-inequality-free distance in other conventions.
Quasimetric space. Usually relaxes symmetry and can have directional distance.
Premetric. Often denotes a weaker distance-like function with fewer axioms.
Seminormed space. A vector space with a homogeneous subadditive function inducing a translation-invariant pseudometric.
Uniform space. A general uniform-neighborhood structure that may be generated by a family of pseudometrics.

References¶

[1] Kelley, J. L. (1955). General Topology. Van Nostrand. Reprint: Springer, 1975. ISBN 978-0-387-90125-1. registry ↩

[2] Willard, S. (1970). General Topology. Addison-Wesley. Reprint: Dover, 2004. ISBN 978-0-486-43479-7. registry ↩

[3] Howes, N. R. (1995). Modern Analysis and Topology. Springer. https://doi.org/10.1007/978-1-4612-0831-0 registry ↩