Interval¶
An order-convex subset containing every ambient element between any two of its members.
Core Idea¶
An interval is a subset of a linearly ordered set with no missing point between two of its members. Formally, if \(a,b\) lie in \(I\) and an ambient element \(x\) satisfies \(a\leq x\leq b\), then \(x\) also lies in \(I\). This order-convexity rule is the defining structure. The familiar real sets \([a,b]\), \((a,b)\), rays, singletons and the whole real line are its customary forms; the empty set is also included under a common convention.[1][2]
The same rule defines an integer interval such as \(\{2,3,4,5\}\), because every integer between its members is present. It does not say that this discrete set is connected in the real topology. Conversely, for a nonempty subset of the real line in its usual topology, being an interval is equivalent to being connected. This equivalence explains the intermediate-value role of real intervals, but it depends on properties of \(\mathbb R\), not on the general order-convexity definition alone.[1]
Structural Signature¶
Sig role-phrases:
- Ambient linear order — Establishes what “between” means. The interval test depends on the ordered universe; it cannot be applied to an untyped collection of points.[2]
- Chosen subset — Specifies which members are proposed to form the interval. The entire ordered set can be an interval, but so can a proper subset.[1]
- Between-members closure — Every ambient element between two chosen members must also be chosen. This is the constitutive no-gap test.[1][2]
- Boundary convention — In \(\mathbb R\), open/closed/half-open brackets state whether finite limiting points are included. A ray is unbounded in one direction; the infinity symbol denotes that direction, not an included real endpoint.[1]
- Possible topological consequence — Nonempty real intervals are exactly connected real subsets. Other ordered sets may lack that real-line equivalence.[1]
Condensed: ambient order + chosen subset + no missing between-point → interval; endpoint notation and topology depend on the setting.
What It Is Not¶
- Not merely a pair of written endpoints. Open intervals omit their boundary points, rays are unbounded, and intervals in general linear orders need not have two endpoints that are elements of the order.[1][2]
- Not every connected set in every ordered space. The cited equivalence is specifically for nonempty subsets of \(\mathbb R\) with its usual topology.[1]
- Not Interval Arithmetic. That live entry performs enclosure computations using intervals; the mathematical interval is the order-convex set it operates on.
- Not an Interval Order. That is a partial-order representation involving intervals, a different object from one interval itself.
- Not generic geometric convexity. The intrinsic condition is betweenness in a linear order, though real one-dimensional intervals also satisfy vector-space convexity.
Scope of Application¶
Intervals organize real analysis because a continuous real function sends a nonempty real interval to another real interval: its image is connected, and nonempty connected real subsets are intervals. This is the structural core behind intermediate-value reasoning. The fact that nonempty real intervals are connected does not make every interval in \(\mathbb Z\) connected under the subspace topology it inherits from \(\mathbb R\). One must keep the ambient order and topology explicit.[1]
In discrete orders, integer ranges still satisfy the no-gap criterion relative to the integers. In general linear orders, the criterion remains meaningful even where a subset's boundary is an unattained cut rather than an element that can be written in a bracket. Intersections of intervals remain intervals because a point between two members of the intersection lies in both originals; arbitrary unions can leave a gap.[2]
Clarity¶
For the real line, \([0,1)\) contains \(0\) but not \(1\), and it contains every real strictly between them. The bracket notation records inclusion of finite bounds, not the reason the set is an interval. The reason is the universal between-members rule. Likewise, \(\{2,4\}\) is not an interval in \(\mathbb Z\) because \(3\) lies between its members but is missing, whereas \(\{2,3,4\}\) is an interval in that order.[1][2]
The empty set and singleton cases make the rule vacuously true: there are not two distinct members between which a gap could occur. Whether to use the word “interval” for empty sets varies by convention, so a theorem about nonempty intervals should not silently import the empty case. The source topology notes explicitly classify the empty set as an interval under their definition and state the connectedness equivalence with a nonempty qualifier.[1]
Manages Complexity¶
One condition classifies an entire family of real sets: bounded intervals of four bracket types, unbounded rays, singletons, the empty set by convention and \(\mathbb R\) itself. The rule makes proofs economical. Instead of checking each bracket form separately, prove that an operation preserves between-members closure, then conclude it preserves intervals.[1][2]
This compression should not erase assumptions. A theorem about connected subsets of \(\mathbb R\) cannot be applied directly to \(\mathbb Z\) just because both have intervals. A guaranteed numerical enclosure in Interval Arithmetic additionally requires correct operations and rounding rules; the mere fact that inputs are intervals does not supply those guarantees.
Abstract Reasoning¶
To decide if \(I\) is an interval, specify the ambient linear order. Take any two included elements \(a\leq b\), then test every ambient \(x\) with \(a\leq x\leq b\). One counterexample shows a gap. For \(\{2,4\}\subseteq\mathbb Z\), \(x=3\) refutes interval status. For \(\{2,3,4\}\), no intermediate integer is omitted.[2]
To reason about a real continuous function, first use connectedness of the nonempty input interval and continuity of the map; then use the real-line equivalence to conclude the image is an interval. This is a theorem with hypotheses, not a direct consequence of order-convexity for arbitrary functions or orders.[1]
Knowledge Transfer¶
The no-gap criterion transfers from real to integer intervals and to other linear orders. Endpoint symbols and connectedness results do not transfer without additional structure. This is a useful example of a formal abstraction that preserves one invariant—order-convex membership—while changing its analytic consequences with the surrounding order and topology.[1][2]
The name remains domain-specific because it refers to a particular mathematical subset type. Its broader skeleton, betweenness closure, is recognizable elsewhere but does not make every “range” in ordinary speech a mathematical interval.
Examples¶
Closed real interval \([0,1]\)¶
The ambient order is the usual order on \(\mathbb R\). Every real number between two members of \([0,1]\) also lies in the set; both finite endpoints are included. It is connected in the usual topology.[1]
Mapped back: order = real order; subset = \([0,1]\); closure = all intervening real points included; boundary = both endpoints included.
Consecutive integers \(\{2,3,4,5\}\)¶
The ambient order is the usual order on \(\mathbb Z\). There is no omitted integer between two members, so the set is an interval in that order. In the topology inherited from the real line, it is not connected; this tests the limit of the real-line theorem.[2]
Mapped back: order = integer order; subset = four consecutive integers; closure = every ambient between-point included; topological consequence = not the real connectedness theorem.
Near miss: \(\{2,4\}\subseteq\mathbb Z\)¶
Both 2 and 4 are selected, but 3 lies between them and is missing. Thus the set fails the intrinsic interval test even though it can be written as two apparent bounds.[2]
Structural Tensions¶
The formal interval identity has no intrinsic two-sided design tradeoff. The ambient order is a hypothesis and bracket notation is a representation, not a pair of opposed benefits and costs. Two boundary cautions matter: order-convexity transfers to integer or other linear orders, while nonempty interval connectedness is a real-topological result; and bracket notation can conceal singleton, empty, ray or unattained-cut cases. Diagnostic: specify the ambient order and any topology, then test whether every point between two included points is included, regardless of the displayed brackets.[1][2]
Structural–Framed Character¶
Interval is a strongly structural mathematical object with explicit ambient assumptions. The no-gap rule is formal and does not depend on a human judgment about whether points look close. The framing choices are the ambient linear order, the convention on the empty set, and any topology placed on that order. Its vocabulary comes from mathematics; “open” and “closed” describe boundary membership or topology, not social evaluation. Human practice chooses notation and a domain of discourse, but cannot override a missing between-point once those are fixed. Imported into a discrete order, the structural identity remains while real connectedness does not. Its character is an exact order-theoretic subset type whose consequences depend on the ambient structure.
Structural Core vs. Domain Accent¶
The skeleton is closure under betweenness: membership of two bounds forces membership of everything between them. The domain accent is mathematical ordering, endpoint classification and, on \(\mathbb R\), topology and intermediate-value reasoning. The live Order prerequisite makes betweenness definable but does not itself assert closure; related Convexity is not the recorded parent. Whether betweenness closure has a distinct cross-domain identity is a future-prime question. A generic “continuous span” metaphor is too loose, and Interval Arithmetic is a separate computation built on the set type. The named interval remains domain-specific because its exact no-gap test requires a specified mathematical order.
Instantiates / Related Primes¶
This entry presupposes Order.
Composition prerequisite: Order (presupposes, strict). An interval's betweenness closure cannot be stated without an ambient order, while an order is not itself an interval. The differently scoped Η set identity is not a genus merely because intervals can be sets. Convexity is a related skeleton, not a parent; arbitrary order intervals need not be topologically connected or convex in real vector space. Interval Arithmetic is a distinct operation.
Relationships to Other Abstractions¶
Current abstraction Interval Domain-specific
Parents (1) — more general patterns this builds on
-
Interval presupposes Order Prime
Order intervals presuppose an order relation.Betweenness closure is defined by comparisons in an ordered carrier; many orders are not intervals and the interval is not the order itself.
Hierarchy paths (3) — routes to 3 parentless roots
- Interval → Order → Comparison → Self Checking
- Interval → Order → Relation
- Interval → Order → Set and Membership
Neighborhood in Abstraction Space¶
Interval 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 — Formal Models & Logical Foundations (33 abstractions)
Nearest neighbors
- Filtration (algebra) — 0.82
- Ordered geometry — 0.82
- Ring Ideal — 0.82
- Linear order — 0.81
- Subspace Topology — 0.81
Computed from structural-signature embeddings · 2026-10-08
Not to Be Confused With¶
Interval Order names a relation among objects represented by intervals; Interval Arithmetic names computations with interval enclosures; an interval is the underlying no-gap subset. An unbounded real interval does not contain the number \(\infty\) as a real endpoint, and a finite integer interval is not thereby a connected real set.[1][2]
References¶
[1] Leinster, University of Edinburgh Topology notes, §C2, Definition C2.1, Lemma C2.2 and Proposition C2.6. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q
[2] University of Toronto order/topology notes, exercise 10.2, p. 26. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m