3.11.2 Tensor Dual Basis Cardinality

Tensor Dual Basis Cardinality is the precise count, or in the infinite-dimensional case the precise cardinal number, of elements in a basis of the dual space V*, contrasted against the cardinality of a basis of V itself. In finite dimensions this cardinality is simply the shared integer n, but the question becomes considerably more delicate once V is infinite-dimensional, where the cardinality of a basis of V* is governed by a specific theorem relating it to the cardinality of the field and the cardinality of a basis of V.

Finite Cardinality

Counting the Dual Basis Directly

When V has a finite basis e_1, ..., e_n, the associated dual basis e^1, ..., e^n, defined by the Kronecker delta relation, has exactly n elements, one for each basis vector of V. This count is immediate from the explicit construction: the dual basis is indexed by precisely the same index set used for the original basis, giving a bijection between the two index sets and hence equal finite cardinalities.

No Ambiguity in the Finite Case

Because a finite-dimensional vector space has a well-defined dimension, an ordinary non-negative integer, and every basis of a finite-dimensional space has exactly that many elements, the cardinality of the dual basis in the finite case is unambiguous and requires no further qualification beyond stating the shared value n.

Infinite Cardinality and the Erdős–Kaplansky Theorem

The General Result

For an infinite-dimensional vector space V with a basis of infinite cardinality κ, over a field F of cardinality λ, the algebraic dual space V* has dimension, and hence any basis of V* has cardinality, equal to

This result, known as the Erdős–Kaplansky theorem, shows that the dual basis cardinality in the infinite-dimensional case is governed by cardinal exponentiation rather than by a simple matching count.

Why the Cardinality Strictly Increases

By Cantor's theorem on cardinal exponentiation, λ^κ is always strictly greater than κ whenever κ is infinite and λ is at least 2, which holds for any field. Consequently, the cardinality of a basis of V* is always strictly larger than the cardinality of a basis of V whenever V is infinite-dimensional, a sharp qualitative difference from the finite case.

Illustrative Case: A Countably Infinite Basis

Setting Up the Example

Suppose V has a countably infinite basis, so κ = ℵ_0, and F is the real numbers, so λ equals the cardinality of the continuum, 2^{ℵ_0}. The Erdős–Kaplansky theorem gives the dimension of V* as

using the cardinal arithmetic identity ℵ_0 × ℵ_0 = ℵ_0, so the dual basis in this case has cardinality equal to the continuum itself, strictly greater than the countably infinite cardinality of the original basis of V.

Interpreting the Result

This means an infinite-dimensional space with a countable basis has a dual space requiring uncountably many basis elements, an entirely different regime from the finite-dimensional matching cardinality, and one that has no simple analogue when reasoning intuitively from finite-dimensional experience.

Consequences for the Construction of a Dual Basis

The Naive Dual Construction Falls Short

Attempting to mimic the finite-dimensional dual basis construction directly, defining e^i via e^i(e_j) = δ^i_j for each basis vector e_i of an infinite basis, produces only a linearly independent set inside V*, not a spanning set, since these functionals can only ever act nontrivially on finitely many basis vectors at once, while V* in general contains functionals with unrestricted, non-finitely-supported behavior across the whole basis.

A Genuine Basis Requires the Axiom of Choice

Establishing that V* has a basis at all in the infinite-dimensional case, of whatever cardinality the Erdős–Kaplansky theorem specifies, relies on the fact that every vector space has a basis, a consequence of the axiom of choice; no explicit, constructive listing of such a basis is generally possible.

Practical Relevance

Why Tensor Algebra Usually Restricts to Finite Dimensions

The dramatic cardinality mismatch in the infinite-dimensional case is one of the principal reasons that introductory and applied treatments of tensor algebra restrict attention to finite-dimensional vector spaces, where the dual basis cardinality matches the original basis cardinality exactly and all of the standard tensor space constructions remain finite-dimensional and well-behaved.

Alternative Duals in Infinite Dimensions

In infinite-dimensional settings such as function spaces, the issue is typically sidestepped by working instead with a topological dual, consisting only of continuous linear functionals, which is often isomorphic to a more manageable space and avoids the enormous cardinality growth associated with the full algebraic dual.

Diagrammatic Summary

The diagram summarizes the sharp contrast between matched cardinality in finite dimensions and the strictly larger cardinality of the dual basis governed by the Erdős–Kaplansky theorem in infinite dimensions.