Differential geometry is the mathematics of smooth curved spaces. It combines calculus, linear algebra and topology to study manifolds, tangent spaces, curvature, geodesics and geometric structures that vary from point to point.
Euclidean geometry works in flat space. Differential geometry asks how lengths, angles, directions and shortest paths behave when the underlying space is curved. This makes it central to general relativity, mechanics, robotics, computer graphics, optimisation and modern geometry.
Series route: Mathematics Learning Hub → How Mathematics Works → Differential Geometry.
1. What differential geometry is
Differential geometry studies smooth geometric spaces using derivatives and local linear approximations. The main objects are manifolds: spaces that may curve globally but locally resemble ordinary Euclidean space.
2. Manifolds are locally simple and globally rich
The surface of a sphere is curved globally, yet a small neighbourhood around any point looks approximately like a flat plane.
This local-flatness principle lets calculus operate on curved spaces.
3. Coordinate charts provide local descriptions
A coordinate chart maps a region of a manifold to ordinary Euclidean coordinates. Different charts may cover different regions.
Where charts overlap, transition maps tell us how one coordinate system changes into another.
4. Geometry should not depend on coordinates
Coordinates are descriptions, not the geometric object itself. Differential geometry distinguishes intrinsic quantities that remain meaningful across coordinate changes from artefacts of one coordinate choice.
5. Tangent spaces linearise curved geometry
At each point on a smooth manifold, the tangent space contains the possible instantaneous directions through that point.
Although the manifold may be curved, each tangent space is a vector space. This lets linear algebra act locally on nonlinear geometry.
6. Tangent vectors represent local motion
A tangent vector can be defined through derivatives of curves passing through a point or as a derivation acting on smooth functions.
Either way, it represents first-order directional behaviour.
7. Cotangent spaces encode linear measurements
The cotangent space is the dual vector space of linear functionals on tangent vectors.
Differentials of functions live naturally here, linking calculus with geometric structure.
8. A metric measures lengths and angles
A Riemannian metric assigns an inner product to each tangent space smoothly from point to point.
Once a metric is given, tangent-vector lengths, angles, curve lengths and local geometric quantities become measurable.
9. Curve length accumulates local speed
The length of a smooth curve is obtained by integrating the norm of its velocity vector along the curve.
This is a direct synthesis of calculus and metric geometry.
10. Geodesics generalise straight lines
Geodesics are curves that locally follow the straightest possible path according to the connection or metric structure.
On a sphere, great circles are geodesics. A long-distance flight can therefore appear curved on a flat map while following a geodesic on Earth.
11. Shortest and straightest are related but distinct
Geodesics are locally extremal paths, but a geodesic need not remain globally shortest over arbitrary distances.
This distinction becomes important near cut loci and on spaces with nontrivial global topology.
12. Connections compare directions at different points
Tangent vectors at different points live in different tangent spaces, so subtracting them directly is not naturally defined.
A connection provides a rule for differentiating vector fields and comparing directions across nearby points.
13. Parallel transport moves directions along curves
Parallel transport carries a tangent vector along a curve while keeping it parallel according to the chosen connection.
On curved spaces, transporting a vector around a closed loop can return it rotated. That failure to return unchanged reveals curvature.
14. Curvature measures deviation from flatness
Curvature captures how a space differs locally from Euclidean geometry. Different curvature tensors or scalars encode different aspects of that deviation.
Positive, zero and negative curvature produce different geodesic and triangle behaviour.
15. Gaussian curvature is intrinsic
For surfaces, Gaussian curvature can be computed entirely from the metric, without referring to how the surface sits in a higher-dimensional ambient space.
This is one of geometry’s profound discoveries: bending without stretching can preserve intrinsic curvature information.
16. Curvature changes triangle geometry
In flat Euclidean geometry, triangle angles sum to 180°. On positively curved surfaces the sum can exceed 180°, while on negatively curved spaces it can be less.
Familiar geometric rules are therefore local consequences of flatness, not universal truths about every space.
17. Vector fields describe distributed direction
A vector field assigns a tangent vector to each point. Velocity fields, force fields and flows are examples.
Differential geometry provides the coordinate-independent language needed to study them on curved manifolds.
18. Differential forms integrate over curves and surfaces
Differential forms are antisymmetric multilinear objects designed for integration on manifolds.
They provide a unified language for line integrals, surface integrals, flux and orientation.
19. Exterior derivatives generalise grad, curl and divergence
The exterior derivative maps differential forms to higher-degree forms and satisfies d²=0.
Many vector-calculus identities become one structural statement in the language of forms.
20. Stokes’ theorem unifies integration theorems
The generalised Stokes theorem relates the integral of a differential form over the boundary of a manifold to the integral of its exterior derivative over the manifold.
The fundamental theorem of calculus, Green’s theorem and classical Stokes’ theorem are all descendants of this architecture.
21. Differential geometry and topology interact
Topology determines global structure while differential geometry adds smoothness, metric and curvature.
The Gauss–Bonnet theorem famously connects curvature, a local geometric quantity, to the Euler characteristic, a global topological invariant.
22. Differential geometry and relativity
General relativity models spacetime as a curved manifold. Matter and energy influence curvature, while curvature influences the motion of matter and light.
The physics is specific, but the mathematical language is differential geometry.
23. Differential geometry and optimisation
Some optimisation problems have variables constrained to manifolds such as spheres or matrix groups.
Riemannian optimisation follows gradients and geodesic-like updates while respecting the curved constraint space.
24. Differential geometry and robotics
Robot orientations and rigid-body poses live naturally on manifolds and Lie groups rather than ordinary flat coordinate spaces.
Geometric methods help plan motion and combine transformations without losing structural constraints.
25. A worked mechanism: shortest paths on a sphere
Suppose two points lie on Earth’s surface.
- Model Earth approximately as a sphere.
- Restrict motion to the spherical surface.
- Geodesics on the sphere are great-circle arcs.
- The shorter great-circle arc gives the locally and globally shortest route unless antipodal ambiguities arise.
The “straight line” is curved in three-dimensional space but straight relative to the surface’s intrinsic geometry.
26. Common failure modes
- Coordinate fixation: treating coordinates as the geometry itself.
- Extrinsic/intrinsic confusion: mixing properties of an embedding with properties of the manifold.
- Flat-space intuition: applying Euclidean angle or distance rules on curved spaces.
- Tangent-space confusion: subtracting vectors based at different points without a connection.
- Geodesic/shortest-path overclaim: assuming every geodesic is globally shortest.
27. Differential geometry and linear algebra
Every tangent space is a vector space. Metrics are bilinear forms. Curvature and connections are represented through linear maps and tensors locally.
28. Differential geometry and analysis
Geometric PDEs, variational problems and smooth maps require analytical tools. Geometry determines the space; analysis determines how functions and fields behave on it.
29. Differential geometry as a machine
Manifold → Local Coordinates → Tangent/Cotangent Spaces → Metric/Connection → Curvature/Geodesics → Global Geometric or Physical Interpretation.
30. What mastery looks like
- distinguish manifolds from coordinate descriptions;
- work with tangent spaces and vector fields;
- understand metrics as local measurement structures;
- interpret geodesics and parallel transport;
- read curvature intrinsically;
- use differential forms and Stokes-type theorems;
- connect local geometry to global topology;
- apply geometric language to physical and computational systems.
31. Conclusion
Differential geometry works by applying calculus and linear algebra locally on spaces that may be curved globally. Tangent spaces provide local linear models. Metrics define measurement. Connections compare directions. Curvature measures deviation from flatness. Geodesics generalise straight lines. Differential forms unify integration over curved spaces.
Topology tells us what a space is globally. Differential geometry tells us how that space bends, measures and moves locally.
How Mathematics Works | Batch 05
- How Mathematics Works | Combinatorics
- How Mathematics Works | Graph Theory
- How Mathematics Works | Measure Theory
- Differential Geometry — this article
Return to the Mathematics Learning Hub or the How Mathematics Works root.