5.19.1 Tensor Product Vector Space Pair

Tensor Product Vector Space Pair is the ordered pair (V, W) of vector spaces treated as the fundamental input data of the tensor product construction, considered as a unit whose joint properties, matching or differing dimensions, matching or differing base fields, and any additional compatible structure such as inner products, determine the properties of the resulting space V ⊗ W. Framing V and W together as a pair, rather than examining each space in isolation, emphasizes that the tensor product is fundamentally a binary operation on vector spaces, and the output depends jointly on both inputs in ways that cannot be deduced from either space alone.

The Pair as an Ordered Input

Order Matters for the Construction, Not for the Result up to Isomorphism

As established in the treatment of factor order, the pair (V, W) and the pair (W, V) produce tensor products V ⊗ W and W ⊗ V that are canonically isomorphic but not identical as constructed objects, meaning the input pair should be understood as ordered even though the isomorphism class of the output does not depend on this order.

The Pair Must Share a Common Field

As discussed under scalar field compatibility, forming V ⊗ W requires both members of the pair to be vector spaces over the same field F; the pair (V, W) is only a valid input to the tensor product construction once this shared-field condition is confirmed.

Joint Properties of the Pair Determining the Tensor Product

Dimension of the Pair

If V and W are finite-dimensional with dim(V) = m and dim(W) = n, the pair jointly determines the dimension of the tensor product via the product formula dim(V ⊗ W) = mn, a quantity that cannot be recovered from either m or n alone without knowing both.

Special Pairs: Equal Factors

When the pair consists of the same space twice, (V, V), the tensor product V ⊗ V acquires extra structure absent for a generic pair, namely the swap automorphism described in the factor order discussion, which splits V ⊗ V into symmetric and antisymmetric parts, a decomposition with no analogue when V ≠ W.

Special Pairs: One Factor Is the Base Field

When one member of the pair is the field F itself, viewed as a one-dimensional vector space over itself, the tensor product simplifies dramatically:

illustrating that the pair (F, V) behaves as an identity-like input for the tensor product operation.

The Pair Under Linear Maps

Pairs of Morphisms Acting on Pairs of Spaces

Given a pair of linear maps φ: V → V' and ψ: W → W', matching the structure of the vector space pair (V, W) mapping to (V', W'), there is an induced map φ ⊗ ψ: V ⊗ W → V' ⊗ W', showing that the tensor product operation extends from acting on pairs of spaces to acting on pairs of morphisms between spaces, consistent with its role as a bifunctor.

Functoriality in Both Components of the Pair

This extension is functorial in both components of the pair simultaneously: composing pairs of maps and then tensoring produces the same result as tensoring first and then composing, a compatibility condition central to treating the tensor product as an operation on the category of vector-space pairs rather than merely a set-level construction.

Diagram of a Vector Space Pair and Its Tensor Product

Comparing Pair-Level Reasoning to Individual-Space Reasoning

Why the Pair Cannot Be Decomposed for Analysis

Because the tensor product's dimension, and much of its structure, depends multiplicatively and jointly on both V and W, statements about V ⊗ W generally cannot be deduced by separately analyzing V and W in isolation and then combining the conclusions in an obvious way; the pairing itself introduces genuinely new structure, such as rank and entanglement-like phenomena, that exists only at the level of the combined space.

Analogy to Ordered Pairs in Other Bilinear Contexts

The treatment of (V, W) as a joint unit of input parallels how an ordered pair of numbers is the natural input to multiplication, or how a pair of matrices is the natural input to matrix multiplication, reinforcing that the tensor product should be conceptually understood as a genuine binary operation on the category of vector spaces, not as a repeated unary construction applied to each space separately.

Broader Significance

Extending the Pair Concept to Tuples

The same pair-based reasoning extends naturally to tuples of more than two vector spaces, (V₁, ..., Vₙ), whose joint tensor product V₁ ⊗ ... ⊗ Vₙ depends multiplicatively on all n dimensions simultaneously, generalizing the binary pair analysis presented here to the fully multilinear setting.

Foundation for Categorical Product Structures

Viewing vector space pairs as the objects on which the tensor product bifunctor acts situates this construction within the broader categorical framework of monoidal categories, where the tensor product provides exactly the kind of binary operation on objects, together with associated coherence isomorphisms like the swap map, that defines a monoidal structure on the category of vector spaces.