8.2 Tensor Indexed Notation Areas

Tensor Indexed Notation Areas is the set of distinct mathematical and applied fields in which tensor index notation is used as the standard working language, each area drawing on the same core notational rules — superscript and subscript indices, the summation convention, contraction, and raising and lowering — while adapting them to its own characteristic objects and conventions. It surveys where the notation is put to use rather than the rules of the notation itself.

Differential Geometry

Coordinate-Based Tensor Calculus

Differential geometry uses tensor index notation to express the components of tensor fields defined on a manifold relative to a coordinate chart, including the metric tensor g_{ij}, the Christoffel symbols Γ^{i}_{jk} that describe how a coordinate basis changes from point to point, and curvature tensors built from derivatives of the Christoffel symbols. Index notation here is essential because the objects of study are fields whose components vary continuously over the manifold and whose derivatives must themselves be expressed in index form.

Covariant Differentiation

The covariant derivative, written ∇ᵢ or with a semicolon as T^{j}_{;i}, is expressed entirely in index notation as an ordinary partial derivative corrected by Christoffel-symbol terms, allowing differentiation of tensor fields to be carried out consistently across a curved coordinate system.

General Relativity and Physics

The Metric and Stress-Energy Tensor

General relativity expresses its central equation, the Einstein field equation, entirely in tensor index notation, relating the curvature of spacetime, encoded in the Ricci tensor R_{μν} and Ricci scalar R, to the distribution of matter and energy, encoded in the stress-energy tensor T_{μν}, with Greek indices conventionally reserved for the four spacetime coordinates.

Electromagnetism and Field Theory

The electromagnetic field is packaged into the field-strength tensor F^{μν}, and Maxwell's equations, ordinarily written as several separate vector equations, reduce in tensor index notation to a compact pair of tensor equations, illustrating how the notation unifies formulas that appear disparate in non-tensorial form.

Continuum Mechanics and Engineering

Stress and Strain Tensors

Continuum mechanics describes the internal forces and deformations of a material using the stress tensor σᵢⱼ and strain tensor εᵢⱼ, both symmetric rank-two tensors whose index notation directly encodes how force or deformation in one coordinate direction relates to another, and whose contraction with a surface normal vector nⱼ gives the traction vector σᵢⱼ nⱼ acting on that surface.

Elasticity Relations

Constitutive relations linking stress and strain, such as generalized Hooke's law σᵢⱼ = Cᵢⱼₖₗ εₖₗ, are written using a rank-four elasticity tensor Cᵢⱼₖₗ contracted against the strain tensor, a formulation that would be unwieldy without the compactness afforded by index notation and the summation convention.

Diagram of Notation Areas

Machine Learning and Numerical Computing

Multi-Dimensional Array Operations

Modern machine learning frameworks organize data and model parameters as multi-dimensional arrays commonly called tensors, and operations on these arrays — batched matrix multiplication, attention mechanisms, convolutions — are frequently specified using explicit index notation, sometimes formalized as Einstein-summation-style notation directly in software libraries, allowing a single compact expression to describe a complex multi-axis contraction.

Difference in Rigor from the Geometric Setting

While the same summation convention and index bookkeeping is used, tensors in this computational area typically lack the transformation law tied to a change of coordinate basis that defines a tensor in the geometric and physical sense; the notation is borrowed for its combinatorial convenience in specifying array contractions, which is a narrower use than the full geometric notion of a tensor.

Common Thread Across These Areas

Shared Grammar, Different Vocabulary

Every one of these areas relies on the same underlying grammar: indices as superscripts or subscripts, the summation convention for repeated indices, and contraction as the basic operation for combining tensors, while each area supplies its own vocabulary of named tensors — metric, stress, field-strength, elasticity — suited to its subject matter.

Portability of Skill Across Areas

Because the notational rules are identical across these areas, facility with tensor index notation gained in one area, such as differential geometry, transfers directly to reading and manipulating expressions in another, such as continuum mechanics or general relativity, since the underlying index bookkeeping does not change even though the physical or geometric meaning of the specific tensors involved does.