A topological space is a set equipped with a collection of "open sets" satisfying three simple rules, giving a rigorous notion of nearness and continuity without requiring any actual measurement of distance. This abstraction is the foundation of topology — the branch of mathematics often described as "rubber sheet geometry," studying properties that survive stretching, bending, and twisting, but not tearing or gluing.
Imagine a doughnut made of very stretchy rubber. You can squash it, stretch it, and bend it into a coffee cup shape (with the doughnut's hole becoming the cup's handle) — and topologists say these two shapes are "the same," because one can be continuously deformed into the other without cutting or gluing. What can't you turn a doughnut into? A solid ball — because you'd have to somehow get rid of the hole, and that requires cutting.
Topology studies which properties of a shape survive continuous deformation (stretching, bending, twisting) but not properties that depend on exact measurements (like length, angle, or straightness). The number of holes in a shape, whether a shape is connected in one piece, and whether a loop can be shrunk to a point — these are topological properties. Length, area, and being perfectly circular are not.
A topological space formalises "nearness" using collections of "open sets" — regions without a hard boundary edge (think of an open disk that doesn't include its circular boundary line). This turns out to be exactly the right abstract structure needed to define continuity, convergence, and connectedness without ever needing to measure an actual distance between points.
A topological space is a pair (X, τ), where X is a set and τ (the topology) is a collection of subsets of X called "open sets," satisfying:
A set is closed if its complement is open. Notably, sets can be both open and closed, or neither — the everyday intuitions about "open" and "closed" from basic geometry don't always transfer directly.
Every metric space (a set with a distance function) generates a natural topology: define open sets as unions of open balls {y : d(x,y) < r}. This means every notion of continuity and convergence you learned in calculus, based on distances, is actually a special case of the more general topological framework — but topology also covers spaces where no natural distance function exists at all.
In calculus, continuity uses epsilon-delta (distance-based) definitions. In topology, a function f : X → Y is continuous if the preimage of every open set in Y is open in X — a definition using no distances whatsoever, yet exactly matching the calculus definition when X and Y are ordinary metric spaces like ℝⁿ.
Two topological spaces are homeomorphic ("topologically the same") if there is a continuous bijection between them with a continuous inverse. Homeomorphic spaces share all topological properties — this is the precise mathematical sense in which a doughnut and a coffee cup are "the same shape."
A space is connected if it cannot be split into two disjoint non-empty open sets — informally, it's "in one piece." A space is compact if every open cover has a finite subcover — a technical condition generalising the intuitive idea of being "closed and bounded" (as made precise for ℝⁿ by the Heine-Borel theorem).
Topology's roots trace to Euler's 1736 solution of the Königsberg Bridge Problem and his 1750s polyhedron formula (V − E + F = 2), both early examples of results depending only on connectivity, not precise measurement. The field of "analysis situs" (the position of things, an early name for topology) was developed by Poincaré in a series of papers starting in 1895, which first systematically introduced concepts like the fundamental group. The modern abstract axiomatic definition of a topological space, in terms of open sets, was given by Felix Hausdorff (1914) in Grundzüge der Mengenlehre, and refined by Kazimierz Kuratowski (1922) using the closure operator as an alternative equivalent starting point.
Not all topological spaces behave "nicely" — some are too abstract to match everyday intuition. Mathematicians classify spaces using separation axioms (T0 through T4 and beyond), which impose increasingly strong conditions on how well distinct points or sets can be separated by open sets. A Hausdorff space (T2) requires that any two distinct points have disjoint open neighbourhoods — a condition satisfied by all familiar spaces (like ℝⁿ) but not by all topological spaces in general, and often assumed as a "reasonableness" condition in advanced topology.
Algebraic topology assigns algebraic structures (groups, rings) to topological spaces in a way that respects homeomorphism — meaning homeomorphic spaces get isomorphic algebraic invariants. This allows genuinely hard geometric questions ("are these two spaces really different shapes?") to be resolved via easier algebraic questions ("are these two groups isomorphic?"). Key invariants include the fundamental group (encoding loop structure) and various homology and cohomology groups (encoding higher-dimensional hole structure).
Topology has grown into several major sub-branches: point-set topology (the foundational open-set framework discussed here), algebraic topology (using algebraic invariants), and differential topology (studying smooth manifolds using calculus-based tools). Each brings a different toolkit to bear on questions of shape and continuity, and major results — such as the Poincaré Conjecture — often require combining insights across multiple sub-branches.