Loading field map
Differential geometry began with a tension: the familiar calculus of flat space could describe curves and surfaces embedded in three-dimensional space, but the shapes themselves—twisting, bending, and curving—resisted a purely flat description. How could one measure curvature from within a surface, without appealing to the surrounding space? That question drove a succession of frameworks, each rethinking what geometry is, what tools it uses, and what problems it can solve.
The Classical Differential Geometry of Curves and Surfaces (1686–Present) grew out of the calculus. Its practitioners—from Leibniz and the Bernoullis through Euler, Monge, and Gauss—studied curves and surfaces as objects embedded in Euclidean space. They developed formulas for arc length, curvature, and torsion, and they classified surfaces by their Gaussian curvature. The decisive turning point came in 1827, when Gauss proved his Theorema Egregium: the Gaussian curvature of a surface depends only on measurements made within the surface itself, not on how it sits in space. That result revealed that curvature is an intrinsic property, not merely a feature of embedding.
Riemannian Geometry (1854–Present) absorbed this intrinsic insight and transformed it into a general theory. In his 1854 habilitation lecture, Bernhard Riemann abandoned the assumption that space must be Euclidean. He defined an n-dimensional manifold—a space that locally resembles Euclidean space—and equipped it with a smoothly varying inner product on tangent vectors, now called a Riemannian metric. This metric allowed one to measure lengths, angles, and, crucially, the curvature of the manifold from within. Riemannian geometry replaced the extrinsic, surface-by-surface approach of classical differential geometry with a unified intrinsic framework for spaces of any dimension. The classical framework did not disappear; it remained active for concrete calculations on surfaces, but Riemannian geometry became the conceptual foundation for everything that followed.
Tensor Calculus (1901–Present) gave Riemannian geometry a computational language. Developed by Ricci and Levi-Civita, tensor calculus provided a systematic notation for expressing geometric quantities—metrics, curvatures, covariant derivatives—in a coordinate-independent way. It was not a rival to Riemannian geometry but an infrastructure that made Riemannian calculations tractable. Einstein used tensor calculus to formulate general relativity, and the framework has persisted ever since as the standard computational tool for Riemannian and pseudo-Riemannian geometry.
While Riemann was rethinking curvature, another line of inquiry focused on symmetry. Lie Group and Homogeneous Space Geometry (1872–Present), rooted in Sophus Lie's theory of continuous transformation groups and Felix Klein's Erlangen Program, treated geometry as the study of properties invariant under a group of transformations. A homogeneous space is a manifold on which a Lie group acts transitively, so that every point looks like every other. This symmetry-first view contrasted with Riemann's metric-first approach: Riemannian geometry began with a metric and derived symmetry as a special case, while the Lie-group approach began with symmetry and derived geometry from it. The two frameworks coexisted and later merged in the theory of connections.
Contact Geometry (1870–Present) emerged from the study of geometric structures defined by a maximally non-integrable hyperplane field. Contact geometry is often described as the odd-dimensional sibling of symplectic geometry: where symplectic geometry studies closed nondegenerate 2-forms on even-dimensional manifolds, contact geometry studies 1-forms whose exterior derivative is nondegenerate on the hyperplane field. The two frameworks share a structural duality—contact manifolds are the boundaries of symplectic fillings—but contact geometry developed independently, driven by problems in optics, mechanics, and differential equations.
Exterior Differential Systems (1899–Present) provided a unified method for studying differential equations on manifolds. Developed largely by Élie Cartan, this framework uses differential forms and the exterior derivative to encode systems of PDEs in a coordinate-free way. It absorbed and generalized earlier work on Pfaffian systems and gave geometers a powerful tool for analyzing integrability conditions. Exterior differential systems became the language in which Cartan formulated his theory of connections.
Method of Moving Frames (1900–Present), also due to Cartan, replaced the static coordinate systems of classical differential geometry with a flexible, frame-by-frame approach. Instead of fixing a global coordinate chart, one attaches an orthonormal frame to each point of a manifold and studies how the frame changes as one moves along a curve. The method of moving frames made it possible to compute curvature and torsion efficiently and, crucially, to treat connections as infinitesimal transformations of frames. This framework directly influenced Cartan's later theory of connections, and it remains a standard technique in Riemannian geometry, Lie group theory, and geometric mechanics.
Riemannian geometry itself soon proved too restrictive for some applications. Pseudo-Riemannian Geometry (1908–Present) relaxed the requirement that the metric be positive-definite, allowing indefinite signatures. This generalization was driven by general relativity: spacetime is modeled as a four-dimensional pseudo-Riemannian manifold with signature (3,1). Pseudo-Riemannian geometry preserved the core apparatus of Riemannian geometry—connections, curvature, geodesics—but adapted it to a setting where the metric distinguishes timelike, spacelike, and null directions. It coexists with Riemannian geometry as a parallel framework for Lorentzian and other indefinite signatures.
Finsler Geometry (1918–Present) generalized Riemannian geometry in a different direction. Instead of defining length via an inner product on tangent vectors, Finsler geometry defines length via a norm that varies smoothly from point to point but need not come from an inner product. This makes Finsler geometry more flexible—it can model anisotropic media and certain physical theories—but also more computationally demanding. The framework narrowed in practice: most geometers continue to work in the Riemannian or pseudo-Riemannian setting, reserving Finsler methods for problems where asymmetry or anisotropy is essential.
Cartan Connections (1920–Present) unified and generalized the connection concepts of Riemannian geometry, Lie group geometry, and projective and conformal geometry. Cartan realized that a connection could be understood as an infinitesimal version of Klein's homogeneous spaces: a Cartan connection on a manifold M is modeled on a homogeneous space G/H, and it encodes how M deviates from that homogeneous model. This framework absorbed the method of moving frames and extended it to a wide class of geometric structures. Cartan connections remain active today in conformal geometry, CR geometry, and parabolic geometry.
Global Differential Geometry (1920–Present) shifted attention from local curvature to the relationship between local geometry and global topology. Early Riemannian geometry had focused on local invariants—curvature tensors, geodesics, and their behavior in small neighborhoods. Global differential geometry asked: what can the curvature of a complete manifold tell us about its topological shape? The Bonnet–Myers theorem, the Hopf–Rinow theorem, and later the sphere theorem and the Cheeger–Gromoll splitting theorem exemplified this approach. Global differential geometry did not replace local Riemannian geometry; it absorbed it as a foundation and added topological methods, including Morse theory and comparison geometry.
Complex Differential Geometry (1933–Present) adapted Riemannian and Hermitian geometry to complex manifolds. A complex manifold is a manifold that locally looks like complex Euclidean space, and complex differential geometry studies metrics, connections, and curvature that are compatible with the complex structure. Kähler geometry—the intersection of Riemannian, symplectic, and complex geometry—became a central subfield, with applications from algebraic geometry to string theory. Complex differential geometry preserved the Riemannian framework but added new invariants, such as the Ricci form and the first Chern class, that link curvature to complex topology.
Fiber Bundle and Connection Theory (1935–Present) reorganized differential geometry around the concept of a fiber bundle: a space that locally looks like a product of a base manifold and a fiber, but may be globally twisted. Connections on fiber bundles generalize the Levi-Civita connection of Riemannian geometry and the Cartan connection of homogeneous geometry. This framework provided a unified language for describing gauge fields, characteristic classes, and holonomy. It did not replace earlier connection theories; it absorbed them as special cases and supplied a common infrastructure.
Symplectic Geometry (1939–Present) studies even-dimensional manifolds equipped with a closed nondegenerate 2-form. Symplectic geometry grew out of Hamiltonian mechanics, where the phase space of a mechanical system carries a natural symplectic structure. For decades it was a niche framework within differential geometry, but it exploded after the 1980s with the development of Gromov's pseudoholomorphic curves and Floer homology. Symplectic geometry now stands alongside Riemannian geometry as a major active framework, with its own methods, invariants, and applications. It shares a deep duality with contact geometry: every contact manifold is the boundary of a symplectic filling, and many symplectic results have contact analogues.
Chern-Weil Theory (1944–Present) connected curvature to topology in a new way. Chern and Weil showed that certain polynomials in the curvature of a connection on a vector bundle yield closed differential forms whose cohomology classes are independent of the connection. These characteristic classes—Chern classes, Pontryagin classes, the Euler class—became powerful topological invariants. Chern-Weil theory did not replace fiber bundle theory; it used bundles and connections as its raw material and extracted global information from local curvature data. It remains a cornerstone of modern differential geometry and topology.
Gauge Theory in Differential Geometry (1954–Present) turned fiber bundles and connections into a field-theoretic framework. Inspired by Yang–Mills theory in physics, gauge theory studies connections on principal bundles and their curvature, interpreted as fields. The Yang–Mills equations, which generalize Maxwell's equations, are PDEs for connections that minimize a certain energy functional. Gauge theory absorbed the fiber bundle and connection framework and added variational methods, analysis, and a deep interplay with topology. It led to spectacular results in four-manifold topology (Donaldson's theorems, Seiberg–Witten theory) and remains a vibrant area of research.
Geometric Analysis (1970–Present) uses tools from partial differential equations to solve geometric problems. Where earlier frameworks studied curvature as a given datum, geometric analysis treats curvature as a quantity that can be deformed, evolved, or optimized. The Ricci flow, introduced by Hamilton and famously used by Perelman to resolve the Poincaré conjecture, is a paradigmatic example: it evolves a Riemannian metric by its Ricci curvature, smoothing out irregularities and revealing topological structure. Geometric analysis also includes minimal surface theory, harmonic maps, and the study of geometric PDEs like the Yamabe equation and the Einstein equations. It does not replace Riemannian or global differential geometry; it adds a powerful set of analytic methods that have transformed both fields.
Today, no single framework dominates differential geometry. Riemannian geometry remains the default language for metric geometry and general relativity. Global differential geometry and geometric analysis together drive much of the research in curvature and topology. Symplectic and contact geometry form a thriving ecosystem with their own journals, conferences, and methods. Complex differential geometry is essential to algebraic geometry and string theory. Gauge theory continues to produce deep results in low-dimensional topology. These frameworks agree on the basic apparatus of manifolds, tangent spaces, and differential forms, and they share a common reliance on connections and curvature. They disagree on which structures are primary: Riemannian geometry privileges the metric, symplectic geometry privileges the 2-form, complex geometry privileges the complex structure, and gauge theory privileges the connection. This pluralism is a sign of health—each framework illuminates a different aspect of curved space, and the most interesting problems often lie at their intersections.