Differential Geometry

Differential geometry is the study of smooth shapes using calculus.

15 frameworks, from 1750 to today · Geometry · Mathematics

still a working view ended · each leads to its entry below

Overview

Differential geometry is the study of smooth shapes using calculus1. Its objects are curves (paths traced by a moving point), surfaces (two-dimensional skins like a sphere or a saddle), and their higher-dimensional generalizations, manifolds—spaces that look flat in small enough patches but may curve globally, the way the Earth looks flat from a field but is round from orbit123. What is at stake is how to measure and describe curvature, length, angle, and shape in a way that does not depend on how you draw a coordinate grid on the object. That question turned out to be one of the deepest in mathematics, and it became the language in which physicists describe gravity and the geometry of the universe.

Curves and Surfaces: The Classical Era

Differential geometry began in the seventeenth and eighteenth centuries as the calculus of curves. Descartes and Fermat had already connected curves to equations; Newton and Leibniz supplied the derivatives needed to talk about tangents and rates of change1. The decisive step was to attach a moving frame to a curve—a tangent direction, a normal perpendicular to it, and a measure of how sharply the curve bends, the curvature4. By the early eighteenth century the curvature of a plane curve was understood as the reciprocal of the radius of the circle that best fits it at a point.

The study of surfaces came next. Leonhard Euler in the 1760s investigated the curvature of surfaces and the special curves on them, and Gaspard Monge around 1800 developed the geometry of surfaces as a practical and theoretical subject, including the idea of developable surfaces—surfaces that can be flattened onto a plane without stretching, like a rolled sheet of paper. Monge's work, taught at the École Polytechnique, made surface theory a standard part of the mathematical curriculum.

The great synthesis came from Carl Friedrich Gauss1. In his 1827 Disquisitiones generales circa superficies curvas, Gauss introduced a systematic way to describe a surface through two parameters and defined what is now called Gaussian curvature: at each point, the product of the curvatures of the surface in two perpendicular directions. His Theorema Egregium ("remarkable theorem") showed that this curvature can be computed purely from measurements of lengths and angles made on the surface itself, without reference to how the surface sits in three-dimensional space12. A sheet of paper can be bent into a cylinder but not into a sphere without tearing, because bending preserves the intrinsic curvature (zero for both the flat sheet and the cylinder) while a sphere has positive curvature. This distinction between intrinsic geometry (what an inhabitant of the surface could measure) and extrinsic geometry (how the surface is embedded in space) is the founding insight of the whole subject15.

Riemann's Revolution: Spaces of Any Dimension

Gauss's student Bernhard Riemann pushed the idea much further1. In his 1854 habilitation lecture, published in 1868 as "On the Hypotheses Which Lie at the Foundations of Geometry," Riemann proposed that geometry need not be confined to flat space or to surfaces in three dimensions12. One could define a space of any number of dimensions in which the distance between nearby points is given by a formula—a metric—that may vary from place to place1. Such a space is now called a Riemannian manifold12. Riemann's framework absorbed Euclidean and non-Euclidean geometry (the hyperbolic and spherical geometries developed earlier in the century by Lobachevsky, Bolyai, and others) as special cases, and it made curvature a local, computable quantity in any dimension.

Riemann's lecture was dense and largely programmatic. The machinery to make it work came from Elwin Bruno Christoffel and, above all, from Gregorio Ricci-Curbastro and his student Tullio Levi-Civita1. In 1900 they published a systematic account of what became known as the absolute differential calculus or tensor calculus: a bookkeeping system for quantities that transform in prescribed ways under changes of coordinates16. The central object is the connection (later called the Levi-Civita connection), which tells you how to compare vectors at nearby points—how to differentiate a vector field along a curve27. The Riemann curvature tensor, built from the connection, measures the failure of parallel transport around a small loop to return a vector to its starting direction. This tensor calculus was the technical heart of intrinsic Riemannian geometry: everything could be expressed in coordinates, yet the results were independent of the coordinates chosen1.

Einstein, Relativity, and Pseudo-Riemannian Geometry

The tensor calculus might have remained a specialist's tool had not Albert Einstein needed exactly it1. In developing general relativity (1915), Einstein required a geometry in which the metric is not positive-definite: the "distance" between nearby events can be positive, negative, or zero, reflecting the causal structure of spacetime, where light travels along null directions. This is pseudo-Riemannian geometry, and the four-dimensional case with one time and three space directions is Lorentzian geometry. Einstein's field equations relate the curvature of spacetime to the distribution of matter and energy. The collaboration with Marcel Grossmann, who pointed Einstein toward Ricci and Levi-Civita's work, made differential geometry a branch of physics as well as mathematics. From then on, pseudo-Riemannian geometry and general relativity developed in tandem, and the subject gained a permanent constituency of physicists.

Global Geometry and the Rise of Topology

Through the nineteenth and early twentieth centuries, differential geometry was largely local: it studied what happens in a small patch around a point. The global questions—what shapes can carry which kinds of curvature, how local curvature constrains the overall topology of a space—became central in the twentieth century. The Gauss–Bonnet theorem, in its global form, is the model: for a closed surface, the total of the Gaussian curvature over the whole surface equals a constant times a topological invariant, the Euler characteristic. A sphere must have total curvature 8π no matter how it is dented; a torus must have total curvature zero. Local geometry is constrained by global shape.

Heinz Hopf and others in the 1920s and 1930s studied global questions about surfaces and higher-dimensional manifolds. The introduction of fiber bundles—spaces that look locally like a product of a base space and a fiber, but may be twisted globally (the Möbius band is the classic example)—gave a language for global phenomena8. Hassler Whitney's work on sphere bundles and characteristic classes in the late 1930s, and the development of characteristic classes by Whitney, Eduard Stiefel, and others, tied the twisting of bundles to topological invariants. Shiing-Shen Chern and André Weil in the 1940s showed how to compute these classes from curvature using differential forms—Chern–Weil theory—a bridge between the differential and the topological that remains fundamental9.

Comparison Geometry and Curvature–Topology Theorems

A major strand of twentieth-century global geometry asks: if a manifold's curvature is bounded above or below, what can you say about its shape and topology? The Bonnet–Myers theorem (S. B. Myers, 1941, building on Ossian Bonnet's earlier work) says that a complete Riemannian manifold whose curvature is everywhere sufficiently positive must be compact and have finite diameter—it cannot stretch out indefinitely. The Cartan–Hadamard theorem says that a complete simply connected manifold of nonpositive curvature is topologically a Euclidean space. These are comparison theorems: they compare a curved manifold to a model space (sphere, Euclidean space, hyperbolic space) of constant curvature and draw conclusions from the comparison.

The most celebrated curvature–topology result is the sphere theorem: a simply connected manifold whose curvature is pinched sufficiently close to a positive constant must be homeomorphic (and, in stronger versions, diffeomorphic) to a sphere. Rauch, Berger, Klingenberg, and others developed the comparison techniques—Rauch comparison, Toponogov's theorem—that made such results possible. The field of comparison geometry became a major industry, with the work of Jeff Cheeger and Detlef Gromoll in the 1970s on manifolds of nonnegative curvature, and the soul theorem describing their structure.

Geometric Analysis: PDE Meets Geometry

From the mid-twentieth century, differential geometry increasingly merged with the theory of partial differential equations. Geometric analysis treats geometric problems as PDE problems: finding a metric with prescribed curvature, for instance, becomes solving a nonlinear elliptic equation110. The Yamabe problem—can every metric on a closed manifold be conformally deformed to one of constant scalar curvature?—was posed in 1960 and resolved in stages by Trudinger, Aubin, and Richard Schoen (1984), drawing on hard analysis. The Calabi conjecture, that a compact Kähler manifold with vanishing first Chern class admits a Ricci-flat metric, was proved by Shing-Tung Yau in 1977; the resulting Calabi–Yau manifolds became central to string theory. Geometric analysis also produced the Atiyah–Singer index theorem (1963), which computes the dimension of solution spaces of elliptic differential operators in terms of topological data—a vast generalization of the Gauss–Bonnet theorem11.

Curvature Flows

A distinct and spectacular development was the study of curvature flows: evolving a shape by a rule that moves each point according to its curvature1213. The Ricci flow, introduced by Richard Hamilton in 1982, deforms a metric in the direction of its Ricci curvature, smoothing it out much as heat diffuses. Hamilton used it to prove major results about manifolds of positive curvature. The flow can develop singularities—regions where curvature blows up—and understanding these singularities was the obstacle13. Grigori Perelman in 2002–2003 introduced surgery techniques to continue the flow past singularities and used Ricci flow to prove the Poincaré conjecture (that every simply connected closed three-manifold is a sphere) and the more general geometrization conjecture of William Thurston. The mean curvature flow, which evolves a surface by moving it in the direction of its mean curvature, is a related and actively studied flow, with connections to minimal surfaces and image processing.

Symplectic Geometry and Mathematical Physics

Symplectic geometry studies spaces equipped with a closed, nondegenerate 2-form—a structure that assigns to each pair of directions a number measuring the "area" they span11415. It arose from the Hamiltonian formulation of classical mechanics, where the state space of a mechanical system carries a natural symplectic structure1415. In the twentieth century it became an independent field, with deep results on the rigidity of symplectic embeddings and the existence of periodic orbits (the Arnold conjectures, proved in various forms by Conley–Zehnder and Floer). Gauge theory—the study of connections on fiber bundles with symmetry groups—became the language of the Standard Model of particle physics, and mathematicians such as Simon Donaldson and Michael Atiyah used gauge-theoretic ideas (instantons, the Yang–Mills equations) to prove striking results about the topology of four-dimensional manifolds in the 1980s.

Finsler Geometry and Exterior Differential Systems

Two further strands deserve mention. Finsler geometry, named for Paul Finsler (1918), generalizes Riemannian geometry by allowing the length of a vector to depend on direction as well as position—the metric is a function on the tangent bundle, not just a quadratic form. It arises naturally in problems of optimal control and in models of anisotropic media. Exterior differential systems, developed by Élie Cartan in the early twentieth century, is a method for studying systems of equations expressed in differential forms; Cartan's method of moving frames and his theory of prolongation remain powerful tools for local equivalence problems and for the geometry of submanifolds.

Discrete Differential Geometry

In recent decades, discrete differential geometry has emerged as a field in its own right: the study of discrete analogues of smooth geometric notions—discrete curves, polyhedral surfaces, discrete curvature, discrete Laplacians—with the aim of preserving geometric structure under discretization. It is motivated both by computer graphics and by pure questions about which smooth theorems have exact discrete counterparts. The discrete Gauss–Bonnet theorem, for instance, holds exactly for polyhedra: the sum of angle defects at vertices equals 2π times the Euler characteristic.

The Present

Differential geometry today is a broad, interconnected field. Intrinsic Riemannian geometry and comparison geometry remain active, with major open questions about the structure of manifolds under curvature bounds. Geometric analysis and curvature flows are central, with Ricci flow and mean curvature flow still under intense study and with applications ranging from topology to general relativity. Symplectic geometry has grown into a large field with its own conjectures and techniques. Gauge theory and mathematical physics continue to feed both directions, with string theory, mirror symmetry, and quantum field theory posing geometric questions that drive new mathematics1. Discrete differential geometry connects the classical subject to computation16. What unites these strands is the original insight: that curvature is an intrinsic, measurable property of space, and that understanding it reveals the deep connection between local geometry and global shape1.

References

  1. Differential geometry (introduction, After calculus (1600–1800), Intrinsic geometry and non-Euclidean geometry (1800–1900), Intrinsic versus extrinsic, Geometric analysis, Symplectic geometry, Modern differential geometry (1900–2000)), Wikipedia
  2. Riemannian manifold (introduction, History), Wikipedia
  3. Differentiable manifold, Wikipedia
  4. Differentiable curve, Wikipedia
  5. Differential geometry of surfaces, Wikipedia
  6. Ricci calculus, Wikipedia
  7. Levi-Civita connection, Wikipedia
  8. Fiber bundle, Wikipedia
  9. Chern–Weil homomorphism, Wikipedia
  10. Geometric analysis, Wikipedia
  11. Atiyah–Singer index theorem, Wikipedia
  12. Gauss curvature flow, Wikipedia
  13. Mean curvature flow, Wikipedia
  14. Symplectic geometry, Wikipedia
  15. Symplectic manifold, Wikipedia
  16. Discrete differential geometry, Wikipedia

The frameworks of Differential Geometry

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1750-1900
  2. 1848-Present

    Curvature–Topology Theorems

    How does curvature constrain the overall shape of a space?

    Fly to its tower in Nousfarer

  3. 1868-Present

    Riemannian geometry

    How can distance and curvature be defined in any dimension?

    Fly to its tower in Nousfarer

  4. 1900-Present

    Exterior differential systems and moving frames

    How can differential equations be solved geometrically using moving frames?

    Fly to its tower in Nousfarer

  5. 1900-1960

    Ricci calculus

    How can geometric quantities be computed without depending on coordinates?

    Fly to its tower in Nousfarer

  6. 1915-Present

    Pseudo-Riemannian manifold

    What geometry describes spacetime with time and space directions?

    Fly to its tower in Nousfarer

  7. 1918-Present

    Finsler manifold

    What if length depends on direction as well as position?

    Fly to its tower in Nousfarer

  8. 1935-Present

    Global differential geometry and fiber bundles

    How do local patches fit together into a global whole?

    Fly to its tower in Nousfarer

  9. 1941-Present

    Comparison geometry

    What does a curvature bound tell us about a space's shape?

    Fly to its tower in Nousfarer

  10. 1944-Present

    Chern-Weil Theory

    How can curvature compute topological invariants?

    Fly to its tower in Nousfarer

  11. 1960-Present

    Geometric analysis

    How can partial differential equations solve geometric problems?

    About Geometric analysis
    Geometric analysis is the discipline where methods from partial differential equations (PDEs), especially nonlinear elliptic and parabolic equations, are used to answer deep questions in differential geometry and topology. It emerged in the mid-twentieth century as geometers realized that many geometric objects—minimal surfaces, harmonic maps, Einstein metrics, Ricci flows—are naturally described as solutions to PDEs. Distinct from Riemannian geometry's focus on curvature from a metric, and from gauge theory's emphasis on connections, geometric analysis treats differential equations themselves as the primary tool for constructing and deforming geometric structures. Its core curriculum revolves around a set of canonical problems, each linking a PDE to a geometric or topological property.
    Core Curriculum
    The first-course spine of geometric analysis typically covers these topics in order:
    1. Minimal surfaces – surfaces of zero mean curvature, modeled by the Euler–Lagrange equation for area, arising from Plateau's problem.
    2. Harmonic maps – maps between Riemannian manifolds that minimize energy, generalizing geodesics; governed by a semilinear elliptic system.
    3. Yamabe problem – find a conformal metric of constant scalar curvature; leads to a critical exponent nonlinear PDE.
    4. Ricci flow – a parabolic evolution equation for Riemannian metrics, used by Hamilton to deform metrics toward canonical forms.
    5. Mean curvature flow – evolution of hypersurfaces by their mean curvature, a parabolic PDE with applications to topology and singularity analysis.
    6. Yang–Mills connections – analogues of harmonic maps for connections on principal bundles, governed by a nonlinear second-order PDE (the Yang–Mills equation).
    7. Kähler–Einstein metrics – Einstein metrics compatible with a complex structure; existence reduces to solving a complex Monge–Ampère equation.
    8. Positive mass theorem – uses minimal surface techniques and the inverse mean curvature flow to prove non-negativity of ADM mass in general relativity.
    9. Gauge theory – study of moduli spaces of connections, e.g., instantons and anti-self-dual connections, via PDE methods (Donaldson theory).
    10. Pseudoholomorphic curves – maps from Riemann surfaces to symplectic manifolds satisfying the Cauchy–Riemann equation, fundamental to Gromov–Witten theory.
    11. Geometric measure theory – tools for studying surfaces with singularities, such as rectifiable currents and varifolds, used in existence and regularity theory.
    12. Spectrum of the Laplacian – analysis of eigenvalues of the Laplace–Beltrami operator and their relation to geometry (e.g., Cheeger's inequality, Lichnerowicz theorem).
    Intellectual History
    The roots of geometric analysis lie in the classical calculus of variations. Tibor Radó and Jesse Douglas independently solved Plateau's problem for minimal surfaces in the 1930s, proving existence of area-minimizing disks. John Nash's 1956 embedding theorem used a nonlinear PDE to isometrically embed Riemannian manifolds into Euclidean space, showcasing the power of analytic methods. In the 1940s–50s, Louis Nirenberg's work on the Minkowski problem and the Weyl problem solved geometric questions by treating them as existence problems for PDEs. Around the same time, Aleksandr Aleksandrov and Aleksei Pogorelov developed powerful convex geometry and Monge–Ampère equation techniques.
    The modern era of geometric analysis began in earnest in the late 1970s and 1980s. Karen Uhlenbeck's foundational work on Yang–Mills connections and harmonic maps, Clifford Taubes's applications of gauge theory to topology, Shing-Tung Yau's proof of the Calabi conjecture (existence of Kähler–Einstein metrics), and Richard Schoen's work on the Yamabe problem and positive mass theorem established a thriving program. Richard Hamilton introduced the Ricci flow in 1982 and initiated a program to deform metrics toward constant curvature, culminating in Grigori Perelman's 2002–2003 proof of the Poincaré conjecture via Ricci flow.
    Geometric analysis coexists with its peer frameworks—Riemannian geometry, gauge theory, symplectic geometry—often importing their objects (metrics, connections, symplectic forms) and applying PDE techniques to construct or deform them. It has absorbed many results from global analysis and the calculus of variations, and it remains a central engine of progress in differential geometry.
    The 12 concepts its tower teaches
    • Gauge theory
    • Geometric measure theory
    • Harmonic maps
    • Kähler–Einstein metrics
    • Mean curvature flow
    • Minimal surfaces
    • Positive mass theorem
    • Pseudoholomorphic curves
    • Ricci flow
    • Spectrum of the Laplacian
    • Yamabe problem
    • Yang–Mills connections

    Fly to its tower in Nousfarer

  12. 1970-Present

    Symplectic geometry

    What is the geometry of phase space and Hamiltonian motion?

    Fly to its tower in Nousfarer

  13. 1975-Present

    Gauge theory

    What geometry underlies the fundamental forces of nature?

    Fly to its tower in Nousfarer

  14. 1982-Present

    Curvature flows

    How do shapes evolve smoothly under their own curvature?

    Fly to its tower in Nousfarer

  15. 1990-Present

    Discrete differential geometry

    How can smooth geometry be captured on discrete meshes?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Classical differential geometry of cu…Curvature–Topology TheoremsRiemannian geometryExterior differential systems and mov…Ricci calculusPseudo-Riemannian manifoldFinsler manifoldGlobal differential geometry and fibe…Comparison geometryChern-Weil TheoryGeometric analysisSymplectic geometryGauge theoryCurvature flowsDiscrete differential geometry

Beyond the field

Where Differential Geometry's ideas came from in other fields, and what they led to there: the strongest 12 of 102 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Differential Geometry is a planet in Nousfarer

Its city has a tower for each of these 15 frameworks. Land, walk in, and climb one concept at a time.

Land on Differential Geometry

Related fields

Fields its ideas are linked to

More in Geometry

See where Differential Geometry sits on the Map of Human Knowledge