5.13.3 Tensor Product Unique Linear Map

Tensor Product Unique Linear Map is the specific linear transformation guaranteed by the universal property of the tensor product, defined on V ⊗ W and induced by a given bilinear map on V × W. It is called "unique" because the universal property does not merely assert the existence of some linear map compatible with a bilinear map, it asserts that exactly one such linear map exists, and this uniqueness is what allows the tensor product to serve as a faithful linear replacement for bilinear data.

Formal Definition

Setting Up the Correspondence

Let V and W be vector spaces over a field F, and let β: V × W → Z be a bilinear map into a vector space Z. The Tensor Product Unique Linear Map is the linear map:

satisfying the defining condition on simple tensors:

for every v ∈ V and w ∈ W.

Existence Guaranteed by Construction

The existence of f follows directly from how V ⊗ W is built as a quotient of the free module on V × W by the submodule of bilinear relations. Since β already respects those relations, it descends to a well-defined map on the quotient, giving f.

Why the Map Is Unique

Simple Tensors Span the Whole Space

The uniqueness of f rests on the fact that the simple tensors v ⊗ w span all of V ⊗ W, even though not every element of V ⊗ W is itself a simple tensor. Every element can be written as a finite sum:

No Freedom Remains Once Values on Simple Tensors Are Fixed

Because a linear map is completely determined by its values on a spanning set, fixing f(v ⊗ w) = β(v, w) for every simple tensor leaves no additional freedom: the value of f on any sum of simple tensors is forced to equal the corresponding sum of β-values. If a second linear map g agreed with β on all simple tensors, then f and g would agree on a spanning set and therefore be identical everywhere.

Relationship to the Universal Property

The Map as the Output of Universality

The Tensor Product Unique Linear Map is precisely the object whose existence and uniqueness together constitute the universal property of the tensor product. Stating that V ⊗ W "has the universal property" is equivalent to stating that for every bilinear map β, this unique linear map f exists.

Naturality in the Choice of Target Space

The assignment sending a bilinear map β: V × W → Z to its associated unique linear map f: V ⊗ W → Z is itself natural in Z, meaning it is compatible with composing β by further linear maps out of Z. This naturality is the basis for the isomorphism between the space of bilinear maps and the space of linear maps out of the tensor product.

Constructing the Map in Practice

Defining on a Basis

When V and W are finite-dimensional with bases {eᵢ} and {fⱼ} respectively, the unique linear map f is fully determined by its values on the basis tensors eᵢ ⊗ fⱼ, since these form a basis of V ⊗ W. In practice, this reduces the construction of f to specifying a matrix or array of values β(eᵢ, fⱼ).

Well-Definedness Independent of Representation

Even though an element of V ⊗ W can often be written as a sum of simple tensors in more than one way, the value assigned by f is guaranteed to be independent of the chosen representation, precisely because f was constructed to respect the bilinear relations that identify these different representations as equal in V ⊗ W.

Consequences of Uniqueness

Tensor Product as a Faithful Encoding of Bilinear Data

Since every bilinear map corresponds to exactly one linear map on V ⊗ W, no information about the original bilinear map is lost or duplicated in passing to the tensor product. This makes the tensor product a faithful and lossless linear encoding of bilinear structure.

Basis for Functorial Behavior

The uniqueness of the induced linear map is also what allows the tensor product operation to be extended functorially to linear maps between vector spaces, since any pair of linear maps V → V' and W → W' induces, via composition with the canonical bilinear map, a unique linear map V ⊗ W → V' ⊗ W'.