Cartesian monoidal category

In mathematics, specifically in the field known as category theory, a monoidal category where the monoidal ("tensor") product is the categorical product is called a cartesian monoidal category. Any category with finite products (a "finite product category") can be thought of as a cartesian monoidal category. In any cartesian monoidal category, the terminal object is the tensor unit. Dually, a monoidal finite coproduct category with the monoidal structure given by the coproduct and unit the initial object is called a cocartesian monoidal category, and any finite coproduct category can be thought of as a cocartesian monoidal category.

Cartesian categories with a Hom functor that is an adjoint functor to the product are called Cartesian closed categories.^[1]

Properties

Cartesian monoidal categories have a number of special and important properties, such as the existence of diagonal maps Δ_x : x → x ⊗ x and augmentations e_x : x → I for any object x. In applications to computer science we can think of Δ as ‘duplicating data’ and e as ‘deleting’ data. These maps make any object into a comonoid. In fact, any object in a cartesian monoidal category becomes a comonoid in a unique way.

Examples

Cartesian monoidal categories:

• Set, the category of sets with the singleton set serving as the unit.
• Cat, the bicategory of small categories with the product category, where the category with one object and only its identity map is the unit.

Cocartesian monoidal categories:

• Vect, the category of vector spaces over a given field, can be made cocartesian monoidal with the "tensor product" given by the direct sum of vector spaces and the trivial vector space as unit.
• Ab, the category of abelian groups, with the direct sum of abelian groups as monoidal product and the trivial group as unit.
• More generally, the category R-mod of (left) modules on a ring (commutative or not) becomes a cocartesian monoidal category with the direct sum of modules as tensor product and the trivial module as unit.

In each of these categories of modules equipped with a cocartesian monoidal structure, finite products and coproducts coincide (in the sense that the product and coproduct of finitely many objects are isomorphic), or more formally that if f : X₁ ∐ ... ∐ X_n → X₁ × ... × X_n is the "canonical" map from the n-ary coproduct of objects X_j to their product, for a natural number n, in the event that the map f is an isomorphism, we say that a biproduct for the objects X_j is an object $X = \bigoplus_{j \in {1,\ldots,n}} X_j.$ isomorphic to $X = \coprod_{j\in {1,\ldots,n}}X_j$ and $X = \prod_{j\in {1,\ldots,n}}X_j$ together with maps i_j : X_j → X and p_j : X →  X_j such that the pair (X, {i_j}) is a coproduct diagram for the objects X_j and the pair (X, {p_j}) is a product diagram for the objects X_j , and where p_j ∘ i_j = id_Xj. If in addition the category in question has a zero object, so that for any objects A and B there is a unique map 0_A,B : A → 0 → B, it often follows that p_k ∘ i_j = : δ_ij, the Kronecker delta, where we interpret 0 and 1 as the 0 maps and identity maps of the objects X_j and X_k, respectively. See pre-additive category for more.

See also

• Cartesian closed category

References

1. Cartesian monoidal category in nLab