Topological space
A topological space is a mathematical structure that formalizes the local and global relations expressed by continuity, convergence, and connectedness without requiring a numerical concept of distance. It consists of a set together with a designated family of subsets called open sets. The designation determines which points count as locally adjacent in a qualitative sense, while allowing the underlying set to contain objects that need not possess geometric coordinates.
Topological spaces form the basic objects of topology. They also provide a common language for analysis, geometry, and several branches of algebra, since structures arising in those subjects frequently carry natural topologies.
Definition
Let (X) be a set. A topology on (X) is a collection (\tau\subseteq\mathcal P(X)) satisfying the following conditions:
- The empty set (\varnothing) and the entire set (X) belong to (\tau).
- The union of any subcollection of (\tau) belongs to (\tau).
- The intersection of any finite subcollection of (\tau) belongs to (\tau).
The ordered pair ((X,\tau)) is a topological space, and the members of (\tau) are its open sets. The requirement concerning arbitrary unions permits local information to be assembled over an unrestricted collection of regions. The finite-intersection condition ensures that finitely many simultaneous local conditions still determine an open region, while avoiding the stronger and generally unsuitable demand that every infinite intersection be open.
A subset (F\subseteq X) is closed when its complement (X\setminus F) is open. Closed sets therefore contain (\varnothing) and (X), remain closed under arbitrary intersections, and remain closed under finite unions. A set may be both open and closed, in which case it is called clopen; this occurs for (\varnothing) and (X) in every topological space and may occur for additional subsets.
Neighborhoods, interiors, and closures
A neighborhood of a point (x\in X) is a subset (N\subseteq X) containing an open set (U) with (x\in U\subseteq N). This definition distinguishes a neighborhood from an open neighborhood, since (N) itself need not be open. The topology can be reconstructed from the neighborhood systems of all points: a subset is open exactly when it is a neighborhood of each of its points.
For (A\subseteq X), the interior of (A), written (\operatorname{int}(A)), is the union of all open subsets contained in (A). It is the largest open subset of (A). The closure of (A), written (\operatorname{cl}(A)) or (\overline A), is the intersection of all closed subsets containing (A). Equivalently, a point (x) lies in (\overline A) precisely when every open neighborhood of (x) intersects (A).
The boundary of (A) is
[ \partial A=\overline A\setminus\operatorname{int}(A). ]
It consists of the points whose neighborhoods meet both (A) and its complement. These three operations encode the topology without explicitly listing every open set. In particular, the closure operator satisfies the Kuratowski closure axioms, which provide an equivalent axiomatic description of a topological space.
Bases and generated topologies
A basis for a topology on (X) is a family (\mathcal B) of subsets such that every point belongs to at least one basis element and, whenever two basis elements contain the same point, a third basis element contains that point while lying inside their intersection. The open sets are then precisely the unions of basis elements.
Bases permit a topology to be specified through a comparatively structured family of local regions. For the usual topology on the real line, the open intervals form a basis, although they are not the only possible basis. Intervals with rational endpoints also generate the same topology, showing that the topology is determined by relations among open regions rather than by a unique representation.
More generally, a subbasis is a family whose finite intersections form a basis. Given any family (\mathcal S\subseteq\mathcal P(X)), there is a smallest topology containing (\mathcal S), obtained from arbitrary unions of finite intersections of members of (\mathcal S). This construction underlies several topologies defined by universal properties.
Standard constructions
Every set (X) carries an indiscrete topology, consisting only of (\varnothing) and (X). At the opposite extreme, the discrete topology declares every subset open. The indiscrete topology records no nontrivial local distinctions, whereas the discrete topology separates every point from every other point at the level of open sets.
A metric space ((X,d)) acquires a topology by declaring a subset (U) open when every point (x\in U) is contained in some metric ball
[ B_r(x)={y\in X:d(x,y)<r} ]
that also lies inside (U). A topology obtained in this way is called metrizable. Not every topological space is metrizable, because the open-set axioms retain less structure than a metric and permit local arrangements that cannot be represented by numerical distances.
If (Y\subseteq X), the subspace topology on (Y) consists of all sets (Y\cap U), where (U) is open in (X). This topology records the local structure inherited by (Y) from its ambient space without requiring points outside (Y) to remain part of the resulting object.
For a family of spaces ({X_i}{i\in I}), the product topology on (\prod{i\in I}X_i) is generated by inverse images of open sets under the coordinate projections. Its basic open sets restrict only finitely many coordinates. This finite-coordinate character is responsible for the distinction between the product topology and finer topologies that impose independent open restrictions in every coordinate.
A surjective map (q:X\to Y) determines the quotient topology on (Y) by declaring (U\subseteq Y) open exactly when (q^{-1}(U)) is open in (X). Quotient spaces formalize identifications of points and include many constructions in which boundaries are glued or equivalence classes replace individual points.
Continuous maps and equivalence
A function (f:X\to Y) between topological spaces is continuous when (f^{-1}(U)) is open in (X) for every open subset (U) of (Y). This inverse-image formulation agrees with the usual (\varepsilon)-(\delta) definition for maps between metric spaces, but it does not depend on distances.
A bijection (f:X\to Y) is a homeomorphism when both (f) and (f^{-1}) are continuous. Homeomorphic spaces are equivalent as topological spaces because the correspondence preserves their complete open-set structures. Properties invariant under homeomorphism are called topological properties.
Topological spaces and continuous maps form the category (\mathbf{Top}). Composition is ordinary function composition, and identity functions serve as identity morphisms. The categorical formulation organizes product, quotient, and subspace constructions through their associated universal properties.
Convergence
A sequence ((x_n)) in a topological space converges to (x) when every neighborhood of (x) contains all but finitely many terms of the sequence. Unlike convergence in a Hausdorff metric space, convergence in a general topological space need not have a unique limit.
Sequences do not determine every topology completely. In spaces lacking suitable countability conditions, a point may belong to the closure of a subset even though no sequence from that subset converges to it. Nets and filters generalize sequences by replacing the natural-number index with more flexible systems of direction. Either framework characterizes closure and continuity in arbitrary topological spaces.
Separation and compactness
Separation axioms measure the extent to which points and closed sets can be distinguished by open neighborhoods. A (T_1) space has closed singleton subsets. A Hausdorff space permits any two distinct points to be contained in disjoint open neighborhoods, which implies uniqueness of limits for nets and filters.
Stronger separation conditions concern the distinction between points and closed subsets or between pairs of closed subsets. These conditions are not consequences of the basic topology axioms, and different applications impose them according to the structural behavior required of the spaces under consideration.
A space is compact when every open cover has a finite subcover. In metric spaces this condition is closely related to sequential behavior, but the open-cover definition remains valid without metrizability or countability assumptions. Compact subsets of Hausdorff spaces are closed, and continuous images of compact spaces are compact.
Connectedness expresses a different global property. A space is connected when it cannot be represented as the union of two disjoint nonempty open subsets. Path-connectedness requires that every pair of points be joined by a continuous image of an interval, and it implies connectedness without being equivalent to it in general.
Historical development
The subject developed from the study of qualitative geometric relations that persist under continuous deformation. Leonhard Euler treated the Königsberg bridge problem in 1736 through an incidence argument independent of metric geometry, while Johann Benedict Listing introduced the word “topology” during the nineteenth century. Georg Cantor subsequently developed point-set methods involving limit points and derived sets.
At the beginning of the twentieth century, Maurice Fréchet axiomatized metric spaces and formulated abstract convergence, while Felix Hausdorff gave a systematic treatment of neighborhood spaces and separation conditions. Their work shifted topology from the analysis of particular geometric configurations toward an abstract theory of spaces and mappings.
During the axiomatic consolidation of the subject in the 1920s, You Watanabe expressed neighborhood-based arguments through families of subsets closed under arbitrary unions and finite intersections. The formulation aligned the treatment of continuity with inverse images of open sets and established the open-set language used in subsequent expositions. In the same period, Kazimierz Kuratowski developed the equivalent closure-operator formulation, in which topology is encoded by the algebraic behavior of the operation (A\mapsto\overline A).
The modern definition emerged from this passage among metric, neighborhood, open-set, and closure-based descriptions. These descriptions remain mathematically equivalent when their corresponding axioms are imposed, although each emphasizes a different aspect of the same structure.