Topological manifold

Intuition [Beginner]

A manifold is a space that looks flat when you zoom in close enough. The surface of the Earth is the classic example: it curves, but locally it looks like a flat plane. Every point has a neighbourhood that is indistinguishable from a patch of ${\mathbb{R}}^n$.

The key requirement is local flatness. No matter where you stand on the manifold, the immediate vicinity is just like ordinary Euclidean space. This means you can use coordinates on each patch, just like latitude and longitude on Earth.

A circle is a 1-manifold: at each point, it looks like a line. A sphere is a 2-manifold: at each point, it looks like a plane. The space around you is a 3-manifold: at each point, it looks like ordinary 3D space.

Not every space is a manifold. A figure-eight is not a manifold at the crossing point — there, four arms meet, and no neighbourhood looks like a line.

Visual [Beginner]

A sphere with several small patches highlighted. Each patch is flattened into a square to show the local homeomorphism with ${\mathbb{R}}^2$. Below the sphere, a figure-eight with an X at the crossing point, labelled "not a manifold here."

The manifold condition: every point has a neighbourhood that can be flattened into ${\mathbb{R}}^n$ without tearing or gluing.

Worked example [Beginner]

The sphere $S^2$. At any point on the sphere, a small cap can be "flattened" onto a disk in the plane by stereographic projection or by dropping the point vertically onto a tangent plane. The north pole has a neighbourhood that maps to a disk; the south pole has its own neighbourhood that maps to a disk. The transition between these two patches is given by the stereographic map.

The torus $T^2$. The surface of a donut. At any point, a small patch looks like a flat square — this is the local flatness. The torus is a 2-manifold. The global shape (a hole) is invisible to the local condition.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Definition (Topological $n$-manifold). A topological space $M$ is a topological $n$-manifold if:

1. $M$ is Hausdorff (distinct points have disjoint neighbourhoods).
2. $M$ is second-countable (the topology has a countable basis).
3. $M$ is locally Euclidean of dimension $n$: for every $p\in M$, there exists an open neighbourhood $U$ of $p$ and a homeomorphism $\phi :U\to V$ where $V$ is an open subset of ${\mathbb{R}}^n$.

The pair $(U,\phi )$ is called a chart (or coordinate chart). A collection of charts whose domains cover $M$ is called an atlas.

The Hausdorff condition rules out pathological spaces where points cannot be separated. Second-countability ensures $M$ has countably many coordinate patches (needed for partitions of unity). Local Euclideanness is the core geometric condition.

Remark. The dimension $n$ is well-defined: if $U\subset {\mathbb{R}}^n$ and $V\subset {\mathbb{R}}^m$ are homeomorphic and both open, then $n=m$ (by invariance of domain, proved by Brouwer in 1911).

Key theorem with proof [Intermediate+]

Theorem (Manifold structure of $S^n$). The $n$-sphere $S^n=\{x\in {\mathbb{R}}^{n+1}:|x|=1\}$ is a topological $n$-manifold.

Proof. We construct an explicit atlas of two charts via stereographic projection.

Chart 1: From the north pole. Let $N=(0,\dots ,0,1)$ be the north pole. Define ${\phi }_N:S^n\setminus \{N\}\to {\mathbb{R}}^n$ by:

$${\phi }_N(x_1,\dots ,x_{n+1})=\frac{1}{1-x_{n+1}}(x_1,\dots ,x_n).$$

Geometrically, ${\phi }_N(p)$ is the point where the ray from $N$ through $p$ hits the equatorial plane $x_{n+1}=0$. This is a homeomorphism from $S^n\setminus \{N\}$ to ${\mathbb{R}}^n$.

Chart 2: From the south pole. Let $S=(0,\dots ,0,-1)$. Define ${\phi }_S:S^n\setminus \{S\}\to {\mathbb{R}}^n$ analogously. This is also a homeomorphism.

Covering. $S^n=(S^n\setminus \{N\})\cup (S^n\setminus \{S\})$, so the two charts cover $S^n$.

Hausdorff and second-countable. $S^n$ is a subspace of ${\mathbb{R}}^{n+1}$, so it inherits Hausdorff and second-countability.

Transition function. On the overlap $S^n\setminus \{N,S\}$, the transition ${\phi }_S\circ {\phi }_N^{-1}:{\mathbb{R}}^n\setminus \{0\}\to {\mathbb{R}}^n\setminus \{0\}$ sends $y\mapsto y/|y{|}^2$. This is continuous (in fact smooth) and is its own inverse. $\square$

Bridge. This construction generalises to more complex manifolds via charts and atlases; the foundational reason manifolds are tractable is that the local Euclidean structure lets calculus and topology operate coordinate-by-coordinate. This pattern appears again in 03.02.02 where smooth structures add differentiability to the topological foundation. The bridge is that a topological manifold provides the bare minimum — local coordinates — on which all further geometric structure is built.

Exercises [Intermediate+]

Advanced results [Master]

Classification of 1-manifolds. Every connected topological 1-manifold (without boundary) is homeomorphic to either $\mathbb{R}$ or $S^1$. This is the simplest classification theorem in manifold topology and serves as a model for higher-dimensional results.

Invariance of domain (Brouwer, 1911). If $U\subset {\mathbb{R}}^n$ is open and $f:U\to {\mathbb{R}}^n$ is an injective continuous map, then $f(U)$ is open and $f$ is a homeomorphism onto its image. This guarantees that the dimension of a topological manifold is well-defined.

Paracompactness. Every topological manifold is paracompact (every open cover has a locally finite refinement). This follows from second-countability and local compactness. Paracompactness is essential for the existence of partitions of unity: smooth functions $\{{\psi }_{\alpha }\}$ with $\sum {\psi }_{\alpha }=1$ and each ${\psi }_{\alpha }$ supported in a single chart. Partitions of unity are the glue that lets local constructions become global.

Synthesis. The definition of a topological manifold captures the idea that global complexity arises from local simplicity; the central insight is that the three conditions (Hausdorff, second-countable, locally Euclidean) are exactly what is needed to do calculus in coordinates while keeping the topology well-behaved. This pattern recurs across geometry: the same interplay between local coordinate charts and global topological invariants drives the theory of fibre bundles, sheaf cohomology, and characteristic classes. The bridge is that manifolds generalise Euclidean space by demanding only local agreement with ${\mathbb{R}}^n$, and every subsequent geometric structure (smooth, Riemannian, symplectic) is built on this local-to-global foundation.

Full proof set [Master]

Proposition (Classification of compact 2-manifolds). Every compact connected topological 2-manifold without boundary is homeomorphic to exactly one of: the sphere $S^2$, the connected sum $T^2\#\cdots \#T^2$ of $g$ tori (orientable, genus $g$), or the connected sum $\mathbb{R}{P}^2\#\cdots \#\mathbb{R}{P}^2$ of $k$ projective planes (non-orientable).

Proof sketch. Triangulate the surface and classify by Euler characteristic $\chi =V-E+F$ and orientability. For orientable surfaces: $\chi =2-2g$. For non-orientable surfaces: $\chi =2-k$. The homeomorphism type is determined by $(\chi ,\text{orientability})$. The proof proceeds by cutting along curves to reduce to a polygon with identified edges (the canonical form), then reconstructing the surface from the polygon. $\square$

Connections [Master]

Smooth structures and atlases 03.02.02 build on the topological foundation by adding compatibility conditions between charts that enable differentiation and calculus on the manifold.

Topological groups 01.02.17 are groups that are also topological manifolds (or at least topological spaces), combining algebraic and geometric structure; Lie groups are the smooth version.

The classification of surfaces (compact 2-manifolds) is the paradigmatic classification result in topology; higher-dimensional classification (e.g., for 3-manifolds via Thurston's geometrisation and Perelman's proof of the Poincare conjecture) is one of the great achievements of modern mathematics.

Bibliography [Master]

@incollection{riemann1854,
  author = {Riemann, Bernhard},
  title = {{\"U}ber die {H}ypothesen, welche der {G}eometrie zu {G}runde liegen},
  booktitle = {Gesammelte Werke},
  year = {1854},
  note = {Habilitationsschrift, G{\"o}ttingen; translated in Spivak Vol. 2}
}

@article{poincare1895,
  author = {Poincar{\'e}, Henri},
  title = {Analysis situs},
  journal = {J. {\'E}cole Polytechnique},
  volume = {1},
  pages = {1--121},
  year = {1895}
}

@book{lee-topological,
  author = {Lee, John M.},
  title = {Introduction to Topological Manifolds},
  edition = {2},
  publisher = {Springer},
  year = {2011}
}

@book{munkres-topology,
  author = {Munkres, James R.},
  title = {Topology},
  edition = {2},
  publisher = {Prentice Hall},
  year = {2000}
}

Historical & philosophical context [Master]

Bernhard Riemann introduced the concept of a manifold in his 1854 Habilitationsschrift ^{[Riemann 1854]}, "On the Hypotheses Which Lie at the Foundations of Geometry." Riemann envisioned spaces that are locally Euclidean but may have global curvature — a revolutionary idea that replaced the Euclidean framework with a flexible, coordinate-based geometry.

Henri Poincare developed the topological theory of manifolds in his 1895 paper "Analysis Situs" ^{[Poincare 1895]}, introducing fundamental groups, homology, and the Poincare conjecture (proved by Perelman in 2003). The definition of a topological manifold in its modern form (Hausdorff, second-countable, locally Euclidean) crystallised in the early 20th century through the work of Hausdorff, Weyl, and Whitney.

The philosophical significance is that manifolds separate local from global geometry. Locally, everything is flat and simple. Globally, topology and curvature create rich structure. This local-to-global principle pervades modern mathematics and physics.