14.21.5 Tensor Map Product Structural Role

The Tensor Map Product Structural Role defines how tensor maps interact, shaping algebraic structures through composition and transformation in multilinear algebra.

Tensor Map Product Structural Role is the function the tensor product of maps performs in making the tensor product of vector spaces into a coherent monoidal structure, by ensuring that the structural isomorphisms of that monoidal structure, namely associativity and the unit isomorphisms, transform correctly and consistently whenever the underlying vector spaces are replaced by other vector spaces via linear maps.

Extending Structure Maps to Act on Morphisms

The Associator as a Structural Isomorphism

For any three vector spaces $V, W, U$ , the associator is the canonical isomorphism

α_{V, W, U} : (V \otimes W) \otimes U \to V \otimes (W \otimes U)

sending $(v \otimes w) \otimes u$ to $v \otimes (w \otimes u)$ . The structural role of the tensor product of maps is what allows this single family of isomorphisms, defined once for all triples of spaces, to interact correctly with linear maps between different triples of spaces.

Naturality of the Associator

For linear maps $f : V \to V^{'}$ , $g : W \to W^{'}$ , and $h : U \to U^{'}$ , the structural role of the tensor product of maps is the statement that

α_{V^{'}, W^{'}, U^{'}} \circ ((f \otimes g) \otimes h) = (f \otimes (g \otimes h)) \circ α_{V, W, U}

so that regrouping first and then mapping gives the same result as mapping first and then regrouping; this square commutes precisely because the tensor product of maps is built directly from its action on simple tensors, matching how the associator itself acts on simple tensors.

Extending the Unit Isomorphisms

Left and Right Unitors

The base field $K$ acts as a unit for the tensor product, with canonical isomorphisms $λ_{V} : K \otimes V \to V$ and $ρ_{V} : V \otimes K \to V$ , sending $α \otimes v$ to $α v$ and $v \otimes α$ to $α v$ respectively.

Naturality of the Unitors

The structural role again ensures compatibility with linear maps, in the form

λ_{V^{'}} \circ ({id}_{K} \otimes f) = f \circ λ_{V}

for every linear map $f : V \to V^{'}$ , and the analogous identity for $ρ$ , confirming that the unit isomorphisms are unaffected in a structurally consistent way by which linear map is applied to the non-unit factor.

Coherence

The Pentagon and Triangle Conditions

Beyond individual naturality squares, a monoidal structure requires coherence conditions, the pentagon identity relating different ways of reassociating a fourfold tensor product and the triangle identity relating the associator to the unit isomorphisms, to hold as equalities of composite isomorphisms built from associators, unitors, and identity maps. The structural role of the tensor product of maps ensures these coherence conditions are compatible with morphisms as well: applying a triple or quadruple tensor product of maps before or after the coherence isomorphisms produces the same result, since both sides reduce, upon expansion, to the same explicit formula on simple tensors.

Why Coherence Matters for the Tensor Product of Maps Specifically

Without this compatibility, different, equally valid ways of parenthesizing a repeated tensor product of maps, such as $(f \otimes g) \otimes h$ versus $f \otimes (g \otimes h)$ , could in principle disagree once transported across the associator; coherence guarantees they agree, which is what justifies dropping parentheses altogether and writing $f \otimes g \otimes h$ without ambiguity.

Summary of the Structural Contribution

Underpinning Every Other Algebraic Role

The functional, form, operator, and representation construction roles all implicitly rely on being able to manipulate tensor products of maps freely, associating and reassociating factors, identifying tensoring with the base field as trivial, and composing repeated tensor products without tracking parenthesization; the structural role is precisely what licenses this freedom, by guaranteeing that all such manipulations are compatible with the tensor product of maps at every step.