1.3 Tensor Structure Foundations

Tensor Structure Foundations is the body of concepts that describe how tensors are organized as algebraic objects: the vector spaces and dual spaces from which they are built, the operations that combine them, the algebraic structure imposed on the collection of all tensors over a given space, and the identities and symmetries that constrain how their components may be arranged. Where the foundations of tensor definition establish what a single tensor is, its type, its components, and their transformation behavior, tensor structure foundations establish how tensors relate to one another as a coherent algebraic system.

The Underlying Algebraic Data

Vector Space and Dual Space

Every tensor is built from a single finite-dimensional vector space V over a field F, together with its dual space V*, the space of linear functionals on V. All of the structural apparatus of tensor algebra is generated from these two pieces of data and the operations available on them.

The Tensor Product Construction

The fundamental operation binding tensor structure together is the tensor product, ⊗, which combines two vector spaces into a new space whose elements represent all possible bilinear combinations of elements from the original spaces. Repeated application of the tensor product to copies of V and V* generates the spaces T^p_q(V) of type (p, q) tensors for every choice of p and q.

The Total Tensor Algebra

Direct Sum of All Types

Collecting the spaces T^p_q(V) for every nonnegative integer p and q and forming their direct sum produces the full tensor algebra over V, a single algebraic structure that contains scalars, vectors, covectors, and every higher-rank tensor as distinguished substructures.

Grading by Rank

The tensor algebra is naturally graded by rank, p + q, meaning it decomposes into a sequence of subspaces indexed by rank, and the tensor product operation respects this grading by mapping a tensor of rank r1 and a tensor of rank r2 to a tensor of rank r1 + r2, giving the tensor algebra the structure of a graded algebra.

Structural Operations on Tensors

Tensor Product of Tensors

Given a tensor of type (p1, q1) and a tensor of type (p2, q2), their tensor product is a tensor of type (p1 + p2, q1 + q2), formed by juxtaposing all of the upper indices and all of the lower indices of both factors while multiplying their components.

Contraction

Contraction is the structural operation that pairs one upper index with one lower index of a tensor, sums over that shared index, and produces a new tensor of rank two less than the original, reducing type (p, q) to type (p - 1, q - 1).

Raising and Lowering Indices

When a vector space is equipped with a nondegenerate bilinear form, such as a metric tensor, its components and their inverse can be used to convert an upper index into a lower index or vice versa, an operation called raising or lowering an index, which relates the type (p, q) and type (p - 1, q + 1) or (p + 1, q - 1) descriptions of what is regarded as fundamentally the same tensor.

Symmetry Structures Within the Algebra

Symmetric and Exterior Subalgebras

Within the tensor algebra, the subsets of totally symmetric tensors and totally antisymmetric tensors form their own closed structures under appropriate products: the symmetric algebra, generated by the symmetric product, and the exterior algebra, generated by the wedge product, each inheriting a grading by rank from the ambient tensor algebra.

Decomposition of General Tensors

Every rank-two tensor decomposes uniquely into a symmetric part and an antisymmetric part, and this pattern generalizes, with higher-rank tensors decomposable into pieces transforming under different symmetry types corresponding to the irreducible representations of the permutation group acting on the indices.

Structural Role of the Component Transformation Law

The Defining Consistency Condition

The transformation law relating a tensor's components in one basis to its components in another is not an incidental fact but the structural condition that makes the entire edifice of tensor algebra basis-independent: every operation defined on tensors, tensor product, contraction, symmetrization, must be verified to respect this law so that the result of the operation is itself a well-defined tensor rather than a basis-dependent artifact.

Diagrammatic Summary

The diagram shows the tensor algebra T(V) arising as the graded structure generated from repeated tensor products of the vector space V and its dual V*, with contraction, symmetrization, and index raising and lowering acting as the principal operations connecting the different graded pieces.