5.13.2 Tensor Product Universal Factorization

Tensor Product Universal Factorization is the property stating that every bilinear map defined on the Cartesian product of two vector spaces (or modules) V × W factors uniquely through the tensor product V ⊗ W via a linear map. This factorization is the precise formal expression of the universal property that characterizes the tensor product up to canonical isomorphism, and it is the mechanism by which bilinear problems are systematically converted into linear ones.

Statement of the Factorization

The Canonical Bilinear Map

Let V and W be vector spaces over a field F, and let ⊗: V × W → V ⊗ W denote the canonical bilinear map sending (v, w) to the simple tensor v ⊗ w. This map is itself bilinear, meaning it is linear in each argument separately.

The Factorization Property

Given any bilinear map β: V × W → Z into an arbitrary vector space Z, there exists a unique linear map f: V ⊗ W → Z such that:

Equivalently, for every pair (v, w) ∈ V × W:

Why the Factorization Exists

Compatibility with the Defining Relations

The factorization is possible precisely because β satisfies the same bilinear relations that were used to construct V ⊗ W as a quotient of a free module. Since β treats sums and scalar multiples in each argument exactly as the relations dictate, assigning f(v ⊗ w) = β(v, w) on simple tensors respects the equivalence classes of the quotient and extends consistently to all of V ⊗ W.

Well-Definedness on Sums of Simple Tensors

Because a general element of V ⊗ W is a finite sum of simple tensors, f is defined by linear extension:

The bilinear relations guarantee this value does not depend on which particular representation of the element as a sum of simple tensors is chosen.

Uniqueness of the Factoring Map

Determined on Simple Tensors

Since the simple tensors v ⊗ w span V ⊗ W, any linear map agreeing with β on all simple tensors is completely determined. There can be no second linear map g ≠ f satisfying g ∘ ⊗ = β, because linear maps are fully specified by their action on a spanning set.

Canonical Isomorphism Consequence

The uniqueness of the factorization is what makes the tensor product unique up to canonical isomorphism: if two objects T₁ and T₂ both satisfy the universal factorization property with respect to bilinear maps out of V × W, then the identity bilinear maps of each factor through the other, and the resulting pair of linear maps are mutually inverse isomorphisms.

Categorical Interpretation

Representing the Bifunctor of Bilinear Maps

The universal factorization expresses V ⊗ W as the object representing the functor that sends a vector space Z to the set of bilinear maps Bil(V × W, Z). Concretely:

Initial Object Among Bilinear Factorizations

Equivalently, the pair (V ⊗ W, ⊗) is an initial object in the category whose objects are pairs (Z, β) consisting of a vector space Z together with a bilinear map β: V × W → Z, and whose morphisms are linear maps compatible with the bilinear structure. Initiality in this category is exactly the categorical restatement of the universal factorization property.

Practical Consequences

Converting Bilinear Problems into Linear Ones

The factorization property is the reason tensor products are useful in practice: any construction or proof involving a bilinear map can be replaced by a construction or proof involving a single linear map on V ⊗ W, which is often easier to analyze using standard linear algebra techniques such as matrix representations, rank, and kernel computations.

Extension to Multilinear Maps

The same factorization pattern extends to multilinear maps of several variables, where a multilinear map V₁ × V₂ × ... × Vₙ → Z factors uniquely through a linear map on the iterated tensor product V₁ ⊗ V₂ ⊗ ... ⊗ Vₙ, generalizing the two-factor factorization described above.