Skip to content

Inner measure

A set function giving lower size estimates through measurable subsets and satisfying inner-measure axioms.

Version
v1 · 2026-09-28 · History
Domain-specific #
7674
Domain group
Formal Sciences
Origin domain
Mathematics
Subdomain
Measure Theory → Mathematics

Core Idea

An inner measure on a set \(X\) is a function \(\varphi:2^X\to[0,\infty]\) that assigns every subset a lower size subject to inner-measure axioms.[1] It gives the empty set value zero, is superadditive on disjoint sets, is continuous along decreasing sequences when the first set has finite value, and requires an infinite value to be approached by subsets of arbitrarily large finite value.[2] These conditions make its estimates coherent under inclusion and limiting operations without requiring every subset to be measurable.

The standard construction begins with a measure \(\mu\) on a sigma-algebra \(\Sigma\).[3] For any \(T\subseteq X\), its induced inner measure is

\[ \mu_*(T)=\sup\{\mu(S):S\in\Sigma\text{ and }S\subseteq T\}. \]

Thus the value of \(T\) is approached from within by measurable subsets.[4] The supremum operation is constitutive: an arbitrary lower estimate, or a size inferred from measurable supersets, is not this induced inner measure.[5]

Paired with the corresponding outer measure, inner measure also supplies a measurability boundary. For a finite measure, subsets on which inner and outer values agree form the completed sigma-algebra, and their common value extends the original measure.[6] The abstraction is therefore a controlled lower approximation to set size, not simply another name for a measure defined on all subsets.

Structural Signature

Sig role-phrases:

  • the ambient set — a set X supplies the universe whose entire power set is assigned extended nonnegative values.
  • the set functional — φ : 2^X → [0, ∞] assigns a lower-size value even when its argument is not measurable.
  • the null guarantee — the empty set receives value zero.
  • the disjoint-union guarantee — disjoint sets satisfy superadditivity rather than the countable additivity required of a measure.
  • the decreasing-tower guarantee — a decreasing sequence beginning at finite inner value has intersection value equal to the limit of its values.
  • the approachable-infinity condition — an infinite value must contain subsets whose finite inner values exceed every prescribed positive bound.
  • the measurable inner family — for an induced inner measure, measurable subsets S ⊆ T supply admissible approximations from within an arbitrary target set T.
  • the supremal construction — μ_*(T) retains the supremum of μ(S) over those measurable inner approximants.
  • the monotonic lower bound — every admissible S ⊆ T guarantees μ_*(T) ≥ μ(S), and enlarging T cannot decrease its inner value.
  • the completion extension — for finite μ, equality μ_*(T) = μ^*(T) admits T to the completed sigma-algebra with their common value.
  • the characteristic limitation — the scalar inner value need not be attained, recover approximating-set geometry, make every subset measurable, or define a measure on all of 2^X.

What It Is Not

  • Not an arbitrary lower bound on set size. An inner measure is a set functional on all subsets that obeys its null, superadditive, continuity, and approachable-infinity conditions; an isolated underestimate has none of that structure.
  • Not an outer measure. The induced inner construction takes a supremum over measurable subsets contained in the target, whereas an outer measure approaches the target through measurable supersets.[7]
  • Not ordinarily a measure on the full power set. Its disjoint-union guarantee is superadditivity, not the countable additivity that would make the functional a measure on every subset.[8]
  • Not the completion of a measure. The inner functional helps identify the completed sigma-algebra when it agrees with the corresponding outer measure; it is an input to that extension test, not the completed measure itself.
  • Not a guarantee that the supremum is attained. The value records the best lower size approached by admissible measurable subsets, even when no single subset realizes it.
  • Not a reconstruction of a set's internal geometry. Equal inner values can arise from different families of approximants, because the scalar output retains their extremal measure rather than their shapes or arrangement.
  • Not permission to assign infinity without approximation. If an inner value is infinite, subsets of arbitrarily large finite inner value must occur; an unapproachable infinity violates the defining regime.

Scope of Application

Inner measure applies throughout measure theory wherever a set functional on all subsets obeys the inner-measure axioms, or where a measure on a sigma-algebra induces lower size by the supremum of measurable subsets contained in a target. Its literal reach requires that typed measurable family and approximation-from-within rule; an arbitrary lower estimate or approximation by supersets is outside the instrument.

  • Axiomatic set-function analysis — functions on 2^X are tested for the null, disjoint-superadditive, decreasing-tower, and approachable-infinity conditions that define an inner measure.[9]
  • Induced lower-size estimation — an arbitrary subset receives the supremum of the original measure over its measurable inner approximants, whether or not a maximizing subset exists.
  • Measure completion — for a finite measure, equality of induced inner and outer values identifies sets admitted to the completed sigma-algebra and supplies their extended measure.
  • Measurability-boundary diagnosis — a gap between inner and outer values records that available measurable approximations do not determine a common size for the target set.
  • Measure-theoretic probability — probability spaces use the same sigma-algebra, containment, and supremal construction to assign lower probability to events not yet in the completed measurable family.
  • Nested-set limit arguments — decreasing sequences beginning at finite inner value are analyzed through continuity to the intersection rather than by an unconstrained limiting estimate.
  • Infinite-value regimes — sets assigned infinite inner value are checked for contained subsets of arbitrarily large finite inner value, preserving the required approachability condition.

Clarity

Naming an inner measure makes the direction of approximation explicit. It estimates an arbitrary set from measurable subsets contained inside it, whereas the corresponding outer measure estimates from measurable supersets. The inner value is therefore not just any numerical lower bound, and an induced inner measure need not itself be a measure on the entire power set. Superadditivity, rather than the countable additivity expected of a measure, is the relevant structural signal.

The concept licenses a precise question for a subset \(T\): How large can a measurable subset of \(T\) be under the original measure? It also makes the extension test legible: equality of the induced inner and outer values identifies the sets admitted to the completed sigma-algebra. Keeping approximation direction and domain of additivity visible prevents “inner measure,” “outer measure,” and “completed measure” from being treated as interchangeable names.

Manages Complexity

An arbitrary subset of X can have combinatorial detail far beyond the sigma-algebra on which a measure is defined. The induced inner measure compresses that inaccessible structure to one extremal quantity: among measurable subsets S contained in T, retain only the supremum of μ(S). Instead of cataloguing every internal approximation, the analyst tracks the admissible family, containment direction, and best attainable lower value. Monotonicity then orders nested sets, superadditivity combines disjoint pieces, and continuity along decreasing finite-valued towers controls a principal limiting branch.

Paired with outer measure, two scalar bounds organize the extension problem. If the inner and outer values agree, the set enters the completed measurable branch and receives their common value; if they differ, the gap records exactly what the available measurable approximations do not settle. Infinite values require a separate branch in which arbitrarily large finite inner values must be attainable. This compression does not reconstruct the geometry of the approximating subsets, identify which subset realizes a supremum, or make every set measurable. It isolates the size information that survives approximation from within while leaving the unresolved set structure visible as a boundary rather than hiding it.

Abstract Reasoning

For an inner measure induced by μ, the primary approximation move runs from measures of eligible subsets S ⊆ T to the sharp lower size μ_(T) by taking their supremum. Any exhibited measurable S proves μ_(T) ≥ μ(S); a sequence of internal approximants with values tending upward can therefore establish the inner value even when no single subset attains the supremum.

An interventionist set move runs from changing containment or disjoint composition to a constrained change in inner measure. Enlarging T cannot lower its inner value because every old admissible subset remains available, while adjoining a disjoint set invokes superadditivity. For a decreasing tower beginning at finite inner value, the limiting intersection is predicted to inherit the limit of the tower's values rather than an unrelated lower estimate.

A boundary move runs from the pair (μ_(T), μ^(T)) to the completion decision. Equality places T in the completed sigma-algebra and licenses their common value as its measure; a gap leaves measurability unresolved by this extension. Infinite inner value uses its own regime—arbitrarily large finite values must be approached—and neither that condition nor a scalar inner value reconstructs the internal approximating sets or turns the inner measure into a countably additive measure on all of 2^X.

Knowledge Transfer

Within measure theory, inner measure transfers literally across choices of underlying set, sigma-algebra, and measure. The same construction replaces the detailed family of measurable subsets of T by sup{μ(S) : S ∈ Σ, S ⊆ T}; the direction “from within,” the inner/outer comparison, and the decreasing-tower and infinite-value regimes remain available when the ambient measure changes. These ingredients carry into measure-theoretic probability as well, because a probability measure supplies the same typed objects and operations rather than merely a visual resemblance.

Its principal beyond-context reach is (C) an instrument or measure: μ_* is a numerical lower-size functional that can be carried into any setting that genuinely supplies a measure on a sigma-algebra and arbitrary subsets to be assessed. A limited (B) shared mechanism also appears in lower-envelope or approximation-from-below reasoning, but that broader mechanism is not by itself an inner measure. What transfers is the extremal construction, its axioms, and the diagnostic comparison μ_*(T) = μ^*(T); what remains home-bound is the measure-theoretic typing of measurable subsets, sigma-algebras, and completion. Informal claims that something is estimated “from the inside” are only (A) analogy. The transfer stops when admissible inner approximants or the required inner-measure axioms are absent, and the scalar value never transfers the geometry of the approximating sets.

Examples

Canonical

Let μ be a measure on a sigma-algebra Σ over X, and let T be an arbitrary subset of X. The induced inner measure is obtained by considering every measurable S ∈ Σ contained in T and retaining sup{μ(S) : S ⊆ T}. If one such subset has measure 3, then it certifies μ_*(T) ≥ 3; a nested succession whose measures approach 5 certifies an inner value of at least 5, even if no single subset attains that value.[10] The construction records only the sharp lower size supplied by admissible inner approximants, not their shapes.[11]

Mapped back: Here the ambient set is X, the set functional is μ_*, the measurable inner family is the collection of Σ-measurable subsets of T, and the supremal construction retains their largest approached measure. Each exhibited approximant supplies the monotonic lower bound, while possible nonattainment exemplifies the characteristic limitation.

Applied / In Practice

Suppose μ is finite and a subset T is not in the original sigma-algebra but has equal induced inner and outer values. The equality μ_*(T) = μ^*(T) admits T to the completed sigma-algebra and assigns it their common value.[12] In the especially clear null-set case, any subset of an original measurable null set has inner and outer value zero and is thereby included by completion, even if it was absent from the starting measurable family.[13]

Mapped back: The original measured space supplies the ambient set and the measurable inner family; the equality of the two extremal values performs the completion extension. The zero-valued subset respects the null guarantee, while the fact that admission follows from agreement rather than from inner measure alone preserves the characteristic limitation.

Structural Tensions

T1: Sharp lower value versus unattained approximant (an extremum without a witness). Taking the supremum over measurable subsets gives the strongest lower estimate supported from within, yet no single admissible subset need realize that value. Requiring attainment would make the construction depend on compactness or closure properties it does not assume; ignoring nonattainment can turn a limiting argument into a nonexistent maximizing set. The inner measure therefore certifies how closely measurable subsets can approach the target's size while deliberately withholding a canonical internal representative. Diagnostic: Has the argument established only a supremal value through an approximating family, or does it separately prove that some measurable subset attains that value?

T2: Superadditive lower size versus additive measure (measure-like behavior short of measurability). Inner measure rewards disjoint combination through superadditivity, making contained pieces jointly certify a lower bound. But this is not the countable additivity that would make the functional a measure on every subset. Treating it as additive overstates what arbitrary nonmeasurable sets permit; treating it as an unconstrained underestimate loses the coherence supplied by its axioms. The object occupies a productive middle position: it quantifies all subsets while reserving full measure behavior for a smaller family. Diagnostic: Does the intended step use only the inner-measure guarantees, or has ordinary measure additivity been imported without first establishing measurability?

T3: Inner–outer agreement versus unresolved measurability (extension and boundary in the same pair). Pairing inner and outer measures can admit a set to the completed sigma-algebra when their values agree. The same apparatus exposes a gap when approximation from within and without does not settle one size. Treating equality as automatic makes every subset measurable; treating disagreement as failure of the tools misses its value as a precise boundary signal. Completion expands the measurable family, but it does so only where the two extremal constructions converge. Diagnostic: Do the inner and outer values coincide under the stated finite-measure setting, and if not, which claimed conclusion about the target set must remain unavailable?

T4: Finite continuity versus approachable infinity (one functional, two limiting regimes). Continuity along a decreasing tower is guaranteed when the first inner value is finite, whereas an infinite value is controlled by the requirement that arbitrarily large finite values occur inside it. Applying the finite rule without its hypothesis can produce invalid limit conclusions; accepting infinity without approachability allows a disconnected label with no finite witnesses. Maintaining two regimes complicates the definition but prevents either finiteness or infinity from becoming formally empty. Diagnostic: Is the argument operating in the finite decreasing-tower regime, or has it supplied the arbitrarily large finite subsets required to justify an infinite inner value?

T5: Inner-measure autonomy versus reduction to Function (Mapping). Every inner measure is a strict specialization of the exact parent Prime Function (Mapping) (Function (Mapping)): it assigns exactly one value in [0, ∞] to every member of the domain 2^X under a fixed rule. Reduction preserves that total single-valued mapping, but loses empty-set normalization, monotonicity, superadditivity, the inner-approximation construction where applicable, and the inner–outer measurability boundary. Treating inner measure as wholly autonomous would hide its complete function structure, while calling it ordinary Measure would impose unsupported countable additivity.
Diagnostic: Is there merely a total single-valued assignment, or does it also satisfy the specific inner-measure axioms and approximation direction?

Structural–Framed Character

Inner measure is mixed-structural on the structural–framed spectrum: it is a rigorously formal set functional built on a portable mapping, while its defining axioms and approximation direction remain specifically measure-theoretic.

On evaluative_weight, the label is neutral: it states how a lower size is assigned rather than approving the set, functional, or result. On human_practice_bound, mathematicians choose and formalize the axioms, but an inner measure's identity is fixed by its domain, codomain, and satisfaction of those axioms rather than by an ongoing social practice. On institutional_origin, no institution constitutes an instance, although mathematical convention governs the name and admissible definition. On vocab_travels, function, domain, codomain, supremum, containment, and monotonicity retain broad formal meanings, whereas measurable subset, inner approximation, superadditivity, decreasing-tower continuity, and inner–outer agreement are measure-theoretically typed. On import_vs_recognize, any set functional satisfying the stipulated axioms can be recognized directly as an inner measure; a generic lower estimate or approximation from within becomes one only by importing those additional commitments.

The smallest reviewed portable skeleton is Function (Mapping): every subset in 2^X is assigned exactly one value in [0, ∞] under a specified rule. That cross-domain reach belongs to the Function (Mapping) Prime. Inner measure remains in situ because the null, disjoint-superadditive, decreasing-tower, approachable-infinity, and, for induced instances, measurable-inner-supremum conditions supply its distinct recognition and collapse tests.

Its character: a mixed-structural formal instrument whose total mapping skeleton is portable but whose identity is fixed by measure-theoretic axioms and approximation from within.

Structural Core vs. Domain Accent

This decomposition shows why Inner Measure is a domain-specific abstraction rather than a Prime.

What is skeletal (could lift toward a cross-domain prime). The carrier is a specified domain of inputs and codomain of outputs, the operation assigns exactly one output to each input, and identity is preserved by the resulting total single-valued rule. This is the complete skeleton inherited by strict subsumption from Function (Mapping): for inner measure the inputs are all subsets of an ambient set and the outputs are extended nonnegative values. Remove the measure-theoretic restrictions and that mapping remains; remove the assignment or permit one subset to receive incompatible values and neither the parent skeleton nor an inner measure remains.

What is domain-bound. Inner Measure requires the power set 2^X, the codomain [0, ∞], empty-set normalization, disjoint superadditivity, decreasing-tower continuity in the finite regime, and the approachable-infinity condition. An induced inner measure adds the measurable-inner-subset family and the supremum-from-within construction; paired inner–outer agreement can then mark completion. Those constraints supply the recognition and failure tests: a general function, arbitrary lower estimate, outer approximation, or countably additive measure on a sigma-algebra does not become an inner measure merely by returning a nonnegative number.

Why this does not clear the prime bar. The complete power-set, extended-nonnegative, inner-axiom, measurable-subset, and approximation-from-within signature does not recur literally in three unrelated domains such as software engineering, constitutional law, and thermodynamics. Those fields instantiate Function (Mapping) when they have typed single-valued assignments, but they do not thereby instantiate Inner Measure; informal talk of an estimate from inside is analogical. Portable reach therefore belongs to Function (Mapping), while the specialized axioms and measurable-set semantics remain measure-theoretic. Removing that domain accent leaves a mapping, not Inner Measure. Conversely, retaining words such as inner, subset, or measure while removing the operative total assignment and its inner-measure guarantees destroys the candidate-level abstraction.

This entry is a kind of Function (Mapping).

Instantiates — Function (Mapping) (Function (Mapping)). An inner measure has the full mapping carrier: its domain is the power set 2^X, its codomain is [0, ∞], and its axioms specify one determinate value for every subset. Removing the measure-theoretic accent leaves a total single-valued function; removing that assignment destroys the inner measure. Strict subsumption therefore holds even though the specialized rule is superadditive rather than countably additive.

Related to — Measure (Measure). An induced inner measure takes a supremum of values of measurable subsets under an ordinary measure, and inner–outer equality can help extend that measure. But an inner measure on all subsets is not generally countably additive on a sigma-algebra and need not arise from a particular prior measure, so Measure is a construction source and comparison standard rather than a strict superclass.

Relationships to Other Abstractions

Local relationship map for Inner measureParents 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.Inner measureDOMAINPrime abstraction: Function (Mapping) — is a kind ofFunction(Mapping)PRIME

Current abstraction Inner measure Domain-specific

Parents (1) — more general patterns this builds on

  • Inner measure is a kind of Function (Mapping) Prime

    An inner measure has the full mapping carrier: its domain is the power set 2^X, its codomain is [0, ∞], and its axioms specify one determinate value for every subset.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Inner measure sits in a sparse region of the domain-specific corpus (63rd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Point-Set Topology & Measure Structures (14 abstractions)

Nearest neighbors

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

Not to Be Confused With

  • Outer measure. An outer measure is a set functional that approaches a target from measurable supersets and supplies an upper size; it is the directional counterpart, not the same construction. Tell: inspect whether the defining extremum is an infimum over supersets or a supremum over measurable subsets contained in the target.
  • Measure. A measure is countably additive on a sigma-algebra, whereas an inner measure is defined on the full power set and has the stated superadditive and continuity guarantees. Tell: check whether disjoint sets must satisfy countable additivity on a declared measurable family or the inner-measure axioms on every subset.
  • Measure completion. Completion enlarges a sigma-algebra and extends a measure to newly admitted sets; inner measure is one of the comparison functionals used to decide that admission. Tell: equality of inner and outer values produces the completed-set decision, while the inner functional alone remains only the lower approximation.

References

[1] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[2] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[3] Encyclopedia of Mathematics, Measure, EMS Press (accessed 2026-09-13). registry ↩

[4] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[5] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[6] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[7] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[8] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[9] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[10] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[11] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[12] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[13] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩