13.9.3 Tensor Inner Product Component Summation

Tensor Inner Product Component Summation involves summing products of tensor components across specified indices to compute a scalar result.

Tensor Inner Product Component Summation is the explicit componentwise computation that carries out an inner product contraction once specific numerical values have been assigned to the components of the tensors involved, expanding the abstract summation convention into a concrete sum of individual products taken over every value of the shared indices. It describes the mechanical arithmetic underlying the inner product, expressed in terms of actual numbers rather than symbolic index notation.

Conceptual Basis

From Symbolic Notation to Explicit Terms

The inner product of two tensors is written compactly using the Einstein summation convention, with repeated indices implicitly summed. Component summation is the process of expanding this compact notation into the full explicit sum of products, one term for each combination of index values permitted by the summation.

Necessity of a Fixed Basis

Component summation can only be carried out once a specific basis has been chosen for the vector spaces involved, since it requires concrete numerical values for the tensor components and, where applicable, for the metric tensor used to pair indices of matching variance.

Relationship to the Underlying Contraction

Component summation is not a separate mathematical operation from contraction itself, but rather its computational unfolding: the abstract statement that an index is summed becomes, in component summation, an actual list of additions performed over that index's full range.

Formal Description

Vector Inner Product Expansion

For two vectors with contravariant components $u^{i}$ and $v^{j}$ in an $n$ -dimensional space, and a metric $g_{i j}$ , component summation expands the inner product as:

⟨ u, v ⟩ = \sum_{i = 1}^{n} \sum_{j = 1}^{n} g_{i j} u^{i} v^{j}

with a separate term contributed for every pair of index values $(i, j)$ .

Simplification for a Diagonal Metric

If the metric is diagonal, meaning $g_{i j} = 0$ whenever $i \neq j$ , the double sum collapses to a single sum:

⟨ u, v ⟩ = \sum_{i = 1}^{n} g_{i i} u^{i} v^{i}

reflecting the familiar componentwise product-and-sum structure of the standard Euclidean dot product when the metric is the identity.

Numerical Example

For a three-dimensional Euclidean space with the identity metric and vectors $u = (2, -1, 3)$ and $v = (4, 0, 1)$ , component summation gives:

⟨ u, v ⟩ = (2 \times 4) + (-1 \times 0) + (3 \times 1) = 11

Properties

Term Count Growth With Rank and Dimension

The number of individual product terms contributing to the component summation grows with the number of summed index pairs and with the dimension of the underlying space, since each additional summed index multiplies the number of terms by the dimension.

Order Independence of Addition

Because ordinary addition is commutative and associative, the terms contributing to a component summation may be added in any order without altering the final result, allowing the sum to be grouped or reorganized for computational convenience.

Agreement With the Abstract Contraction

Regardless of the specific basis used to perform the component summation, the resulting scalar agrees with the value obtained in any other basis, consistent with the basis independence of the underlying inner product contraction.

Practical Role

Numerical Implementation

In computational settings, component summation corresponds directly to the loop structure used to evaluate an inner product numerically, iterating over each combination of index values and accumulating the corresponding products.

Verification of Symbolic Results

Carrying out component summation explicitly on a small example provides a direct means of verifying that a symbolic tensor expression involving an inner product has been set up correctly, since any indexing error typically produces a numerical result inconsistent with the intended computation.

Bridge to Matrix and Vector Arithmetic

For low-rank tensors, component summation reduces the inner product contraction to familiar operations from matrix and vector arithmetic, situating tensor inner products as a direct generalization of dot products and matrix inner products already used in linear algebra.