3.11.5 Tensor Dual Dimension Tensor Type Relation

Exploring how dual dimensions relate tensor types in algebraic structures through dual space mappings and dimensionality constraints.

Tensor Dual Dimension Tensor Type Relation is the precise dependence of a general tensor space's dimension on the combined rank p + q of its type, rather than on how that rank is split between contravariant and covariant factors, a fact that follows directly from the dual dimension equality dim(V*) = dim(V) = n. Because V and V* contribute identically to the dimension count, tensors of type (2, 0), (1, 1), and (0, 2) all inhabit spaces of exactly the same dimension n^2, even though they represent structurally different kinds of multilinear objects.

The Dimension Formula for General Tensor Types

Building the Tensor Space from Basis Elements

The space T^p_q(V) of type (p, q) tensors is spanned by all tensor products of the form

e_{i_{1}} \otimes \dots \otimes e_{i_{p}} \otimes e^{j_{1}} \otimes \dots \otimes e^{j_{q}}

where each of the p vector indices and each of the q covector indices ranges independently over 1 to n. Since there are n choices for each of the p + q factors, and these choices are made independently, the total number of basis tensor products, and hence the dimension of T^p_q(V), is

dim (T_{q}^{p}, (V)) = n^{p + q}

Why Only the Total Rank Matters

The Role of Dual Dimension Equality

This formula depends on V contributing a factor of n for each contravariant slot and V* also contributing a factor of n for each covariant slot, and this second contribution is only equal to n precisely because dim(V*) = dim(V) = n. If V* had a different dimension m ≠ n, as can happen in infinite dimensions, the dimension formula would instead read n^p m^q, and the total dimension would depend on the specific split between p and q, not merely their sum.

Type (2, 0), (1, 1), and (0, 2) Compared

For a fixed total rank of 2, the three possible finite-dimensional tensor types all have the identical dimension n^2:

dim (T_{0}^{2}) = dim (T_{1}^{1}) = dim (T_{2}^{0}) = n^{2}

Yet these three types represent conceptually distinct objects: a (2, 0) tensor is a bilinear map on V* x V*, a (1, 1) tensor is equivalent to a linear endomorphism of V, and a (0, 2) tensor is a bilinear form on V. Dimension alone does not distinguish these roles; only the placement of upper and lower indices does.

Isomorphisms Implied by the Shared Dimension

Vector Space Isomorphism, Not Structural Equivalence

Because T^p_q(V) and T^{p'}_{q'}(V) share the same dimension whenever p + q = p' + q', they are isomorphic as abstract vector spaces; some linear bijection between them always exists. However, this isomorphism is generally not natural or canonical, and it does not respect the distinct roles the two tensor types play, such as how they act on vectors and covectors or how they transform under contraction with other tensors.

A Concrete Example of a Non-Canonical Isomorphism

A (1, 1) tensor can be converted into a (0, 2) tensor, and vice versa, only by choosing an identification between V and V*, such as a metric; without this additional structure, there is no natural way to reinterpret a linear endomorphism as a bilinear form, even though the two spaces have matching dimension.

Consequences for Counting and Classifying Tensors

Counting Independent Components

The formula n^{p+q} gives the number of independent numerical components needed to fully specify a general tensor of type (p, q) relative to a basis, which is a direct and practically useful consequence of the underlying dimension relation, used whenever storage requirements or degrees of freedom for a tensor quantity must be estimated.

Symmetric and Antisymmetric Subspaces

When a tensor space is further restricted to its symmetric or antisymmetric subspaces, such as the space of alternating (0, q) tensors, the dimension of these subspaces is computed using binomial coefficients built from n, but the starting point for any such refinement remains the full dimension n^{p+q} established by the basic dual dimension relation.

Diagrammatic Summary

The diagram shows three distinct tensor types of total rank 2 sharing the identical dimension n^2, a direct consequence of the dual dimension equality between V and V*.