A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.

Parallel Transport

Research Depth 86 in the knowledge graph ☐ I know this ☆ Set as goal
32topics build on this
482prerequisites beneath it
See this on the map →
Connections and Covariant DerivativeVector Fields→→Curvature TensorGeodesics
parallel-transport holonomy covariant-derivative curvature

Core Idea

Parallel transport moves a tangent vector along a curve while keeping it "as constant as possible" according to a connection — formally, the covariant derivative along the curve vanishes. On a flat manifold, parallel transport is path-independent: the transported vector depends only on the endpoints. On a curved manifold, parallel transport around a closed loop generally rotates the vector, and the amount of rotation is directly related to the curvature enclosed by the loop. This path-dependence of parallel transport is the geometric essence of curvature.

Explainer

In Euclidean space, moving a vector from one point to another "without changing it" is obvious — you translate it, keeping its components constant. On a curved manifold, there is no such natural translation. Different paths from p to q might "hold the vector constant" in different ways, depending on the curvature encountered. Parallel transport resolves this ambiguity given a connection: it provides a specific rule for moving vectors along a specific curve.

Given a connection ∇ and a curve γ : [a,b] → M, a vector field V along γ is parallel if ∇_{γ'(t)} V(t) = 0 for all t. In coordinates, this is the system of ODEs dVᵏ/dt + Γᵏᵢⱼ (dγⁱ/dt) Vʲ = 0. The theory of linear ODEs guarantees a unique solution for any initial condition V(a) = v₀ ∈ T_{γ(a)}M. The map P_{a→b} : T_{γ(a)}M → T_{γ(b)}M sending v₀ to V(b) is the parallel transport map along γ. It is a linear isomorphism between tangent spaces. When the connection is the Levi-Civita connection, parallel transport is an isometry — it preserves lengths and angles.

The holonomy of a connection is what happens when you parallel transport around a closed loop. Starting with a vector v at a point p, transporting it around a loop γ back to p, you generally get a different vector P_γ(v) ≠ v. The set of all such rotations forms the holonomy group Hol_p(∇) ⊂ GL(TpM). On a Riemannian manifold, holonomy lies in O(n) (the orthogonal group) because parallel transport preserves the inner product. The holonomy group encodes global geometric information: for instance, Berger's classification of holonomy groups constrains the possible geometries of irreducible Riemannian manifolds.

The connection between parallel transport and curvature is the deepest insight. For an infinitesimal parallelogram spanned by vectors X and Y at a point p, transporting a vector v around the parallelogram produces a rotation approximately equal to R(X,Y)v, where R is the curvature tensor. Curvature measures the infinitesimal holonomy — the rotation per unit area of parallel transport around small loops. Zero curvature means parallel transport is locally path-independent: the vector you get at the end depends only on the endpoints, not on the path. Nonzero curvature means the manifold is genuinely curved, and the geometry felt by a vector depends on the path it takes.

Practice Questions 4 questions

Prerequisite Chain

Understanding Zero → The Number Zero → Counting to Five → Counting to 10 → Counting to 20 → Counting a Set of Objects Up to 20 → Cardinality: The Last Number Counted → Matching Numerals to Quantities → Subitizing Small Quantities → Addition Within 10 → Number Bonds to 10 → Addition Within 20 → Doubles and Near Doubles → Doubles Facts Within 10 → Near Doubles Facts Within 20 → Mental Math Strategies for Addition → Mental Math: Adding and Subtracting Tens → Addition Within 100 → Repeated Addition as Multiplication → Multiplication as Equal Groups → Multiplication: Arrays → Basic Multiplication Facts (0s, 1s, 2s, 5s, 10s) → Multiplication Facts Within 100 → Division as Equal Sharing → Division as Grouping (Measurement Division) → Division: Grouping (Repeated Subtraction) Model → Division: Fair Sharing Model → Division as Equal Sharing → Division as Grouping → Basic Division Facts → Division Facts Within 100 → Multiplication and Division Fact Families → Relationship Between Multiplication and Division → Division Facts as Inverse of Multiplication → Remainders and Quotients in Division → Division Word Problems → Multi-Step Word Problems → Solving Multi-Step Word Problems → Multiplication Word Problems → Division Word Problems → Introduction to Long Division → Factors and Multiples → Prime and Composite Numbers → Equivalent Fractions → Relating Fractions and Decimals → Decimal Place Value → Integers and the Number Line → Comparing and Ordering Integers → Absolute Value → Adding Integers → Subtracting Integers → Multiplying Integers → Introduction to Exponents → Order of Operations → Integer Order of Operations → Variable Expressions → The Distributive Property → Variables and Expressions Review → Introduction to Polynomials → Adding and Subtracting Polynomials → Multiplying Polynomials → Factorial → Permutations → Combinations → Counting Principles: Addition and Multiplication Rules → Introduction to Graph Theory → Propositional Logic Foundations → Logical Equivalences → De Morgan's Laws → Negation of Quantified Statements → Proof by Contradiction → Topological Spaces: Definition and Examples → Open Sets in Topological Spaces → Neighborhoods and Open Sets → Open Sets in Topological Spaces → Basis for a Topology → Neighborhoods and Local Properties → Limit Points and Convergence → Separation Axioms: T₀, T₁, and T₂ (Hausdorff) → Hausdorff Spaces → Introduction to Topological Manifolds → Smooth Manifolds → Tangent Vectors and Tangent Spaces → Vector Fields → Lie Brackets → Connections and Covariant Derivative → Parallel Transport

Longest path: 87 steps · 482 total prerequisite topics

Prerequisites (2)

Leads To (2)