2.16.1 Tensor Infinite Basis Context

Tensor Infinite Basis Context is the study of what a basis of a vector space means, how it is guaranteed to exist, and how it behaves once that vector space is infinite-dimensional, together with the consequences this has for building bases of tensor spaces constructed from it. Where a finite-dimensional space admits a basis that can be written down explicitly as a finite list of vectors, an infinite-dimensional space requires a more careful notion of basis, called a Hamel basis, whose existence is guaranteed only abstractly and whose properties differ sharply from the finite case in ways that ripple through every tensor built on top of it.

The Hamel Basis

Definition

A Hamel basis of a vector space V is a subset B of V that is linearly independent and spans V in the strict algebraic sense: every vector in V is expressible as a finite linear combination of elements of B, with all but finitely many coefficients equal to zero. This "finite linear combination" requirement is retained unchanged from the finite-dimensional definition of a basis; what changes is only that B itself is now an infinite set.

Existence and Non-Constructiveness

Every vector space, of any dimension, possesses a Hamel basis, but this fact is a consequence of the axiom of choice (equivalently, Zorn's lemma applied to the poset of linearly independent subsets ordered by inclusion). For most infinite-dimensional spaces encountered in analysis, such as the space of continuous functions on an interval, no explicit description of a Hamel basis is known or constructible; its existence is purely an existence theorem.

Cardinality Considerations

Hamel Bases Are Typically Uncountable

For familiar infinite-dimensional spaces such as function spaces, the Hamel basis has the cardinality of the continuum or larger, strictly greater than the countably infinite cardinality one might naively expect from an "infinite sequence" of basis vectors. This is a crucial distinction from the more familiar countable Schauder-type bases used in analysis, which rely on convergent infinite series rather than finite linear combinations and are not Hamel bases in the algebraic sense.

with strict inequality to a larger cardinal in essentially every infinite-dimensional space that arises in practice, where B denotes the Hamel basis.

Dual Space Cardinality

Because the algebraic dual space V* consists of all functions from the Hamel basis B to the field F, not just those with finite support, its cardinality is at least |F| raised to the power |B|, which strictly exceeds |B| itself whenever B is infinite. This cardinality gap is the underlying reason V* cannot be identified with V once V is infinite-dimensional.

Building a Tensor Basis From a Hamel Basis

Basis of the Tensor Product

Given a Hamel basis B = {e_α} of V indexed by an arbitrary index set A (possibly uncountable), a basis for the algebraic tensor product V ⊗ V is given by the collection of simple tensors e_α ⊗ e_β for all pairs (α, β) drawn from A × A. Every element of V ⊗ V remains a finite linear combination of these simple tensors, since the algebraic tensor product construction only ever forms finite sums, regardless of how large the index set A is.

where N is finite, even though the index set A from which the α_k and β_k are drawn may be infinite or uncountable.

No Escape From Finiteness at the Algebraic Level

A common point of confusion is expecting the tensor product basis to somehow "sum over the whole basis at once." The algebraic construction never permits this: any single tensor, no matter how the underlying space is structured, is built from only finitely many basis pairs. Infinite combinations require moving outside the purely algebraic framework into a topological one, where convergence can be defined.

Diagram Comparing Finite and Hamel Bases

Consequences for Tensor Theory

Basis-Dependent Statements Require Care

Any statement about tensors that is proved "by choosing a basis" in the finite-dimensional theory must be re-examined in the infinite-dimensional context, since the basis being chosen is a Hamel basis whose existence is nonconstructive. Results that depend on explicit enumeration, ordering, or finiteness of the basis, such as inductive arguments over basis elements, typically fail to transfer directly and must be replaced by arguments that work uniformly over an arbitrary index set.