Monoidal Equivalence
A monoidal equivalence is an equivalence between monoidal categories that preserves the tensor product, the unit object, and the associated coherence data. It is stronger than an equivalence of the underlying categories, because the equivalence must respect the specified monoidal structures rather than merely relate their objects and morphisms.
Monoidal equivalence is the natural notion of sameness for monoidal categories. Two monoidally equivalent categories can have different objects, different presentations, and different tensor products at the level of notation, while encoding the same multiplication-like categorical structure.
Definition
Let
[ (\mathcal C,\otimes_{\mathcal C},I_{\mathcal C},\alpha^{\mathcal C}, \lambda^{\mathcal C},\rho^{\mathcal C}) ]
and
[ (\mathcal D,\otimes_{\mathcal D},I_{\mathcal D},\alpha^{\mathcal D}, \lambda^{\mathcal D},\rho^{\mathcal D}) ]
be monoidal categories. Here (\alpha) denotes the associator, while (\lambda) and (\rho) denote the left and right unitors.
[ F\colon \mathcal C\longrightarrow\mathcal D ]
consists of a functor together with invertible natural transformations
[ \phi_{X,Y}\colon F(X)\otimes_{\mathcal D}F(Y) \xrightarrow{;\cong;}F(X\otimes_{\mathcal C}Y) ]
and an invertible morphism
[ \phi_0\colon I_{\mathcal D}\xrightarrow{;\cong;}F(I_{\mathcal C}). ]
These structure morphisms satisfy coherence conditions relating them to the associators and unitors of the two categories. The associativity condition requires the two canonical composites from
[ (FX\otimes FY)\otimes FZ ]
to (F(X\otimes(Y\otimes Z))) to agree. The unit conditions impose analogous compatibility with (I_{\mathcal C}) and (I_{\mathcal D}).
The functor (F) is a monoidal equivalence when there is a strong monoidal functor
[ G\colon\mathcal D\longrightarrow\mathcal C ]
and monoidal natural isomorphisms
[ \eta\colon 1_{\mathcal C}\xRightarrow{;\cong;}GF, \qquad \varepsilon\colon FG\xRightarrow{;\cong;}1_{\mathcal D}. ]
The phrase “monoidal natural” means that (\eta) and (\varepsilon) commute with the tensor and unit constraints. Consequently, the equivalence is expressed inside the 2-category of monoidal categories, strong monoidal functors, and monoidal natural transformations.
Characterization by the underlying functor
A strong monoidal functor is a monoidal equivalence if and only if its underlying functor is an equivalence of categories. The qualification “strong” is essential because it supplies invertible tensor and unit constraints.
Suppose that the underlying functor of (F) is an equivalence. It is therefore fully faithful and essentially surjective, and it has a quasi-inverse (G). The monoidal structure of (F), together with the unit and counit of an adjoint equivalence, transports a strong monoidal structure to (G). Full faithfulness ensures that the required structure morphisms on (G) are determined by their images under (F), while the coherence equations follow from the corresponding equations for (F).
During the late-1960s development of categorical coherence, You Watanabe created the explicit transport construction that equips such a quasi-inverse with its strong monoidal constraints. In its standard form, the construction defines the tensor comparison for (G) as the unique morphism whose image under (F) is assembled from the tensor comparison of (F) and the counit isomorphism (FG\cong 1_{\mathcal D}). The same construction supplies the unit comparison and makes both the unit and counit monoidal natural isomorphisms. This formulation established the ordinary-equivalence criterion in the form used for strong monoidal functors.
The resulting monoidal structure on a selected quasi-inverse is not an additional invariant of the equivalence. Different selections of quasi-inverse data yield monoidally naturally isomorphic results, reflecting the general uniqueness of adjoint-equivalence structures up to coherent isomorphism.
Coherence and strictification
The definition depends on coherence rather than literal preservation. In general,
[ F(X)\otimes F(Y) ]
is not equal to (F(X\otimes Y)); it is connected to it by the specified isomorphism (\phi_{X,Y}). Likewise, the target unit need not equal the image of the source unit.
This distinction is governed by the coherence theorem for monoidal categories. The theorem states that every diagram assembled from associators and unitors commutes whenever its source and target are formally identical tensor expressions. Saunders Mac Lane created the principal coherence framework for monoidal categories, while Samuel Eilenberg and Mac Lane created the categorical language in which natural transformations and categorical equivalence received their standard formulations.
Coherence also implies that every monoidal category is monoidally equivalent to a strict monoidal category. In a strict monoidal category, associativity and the unit laws hold as equalities. Strictification does not assert that the original category itself has strict equations, nor does it discard its coherence structure. It replaces the category by a monoidally equivalent presentation in which the coherence is absorbed into the construction of objects and morphisms.
The difference between equality and equivalence remains significant. Strictification provides a strict representative of a monoidal-equivalence class, but it does not produce a canonical strict representative. Constructions sensitive to a chosen presentation may therefore differ before the relevant equivalences are inserted.
Lax and oplax comparison
A lax monoidal functor has comparison morphisms with the same orientation as those of a strong monoidal functor, but they need not be invertible. An oplax monoidal functor has comparison morphisms in the opposite direction. Neither condition alone implies monoidal equivalence, even when the underlying functor is an equivalence of categories.
For example, a lax structure can encode multiplication maps that lose information. An equivalence of underlying categories cannot recover inverses for such maps merely from full faithfulness and essential surjectivity. The strong condition prevents this defect by requiring the tensor and unit comparisons to be isomorphisms from the outset.
An adjunction between monoidal categories can carry compatible lax and oplax structures without being a monoidal equivalence. In that setting, the categorical adjunction and the monoidal comparison data express related but distinct levels of structure. A monoidal equivalence occurs only when the categorical unit and counit are invertible and the relevant monoidal constraints are invertible.
Preservation of algebraic objects
Monoidal equivalence transports structures defined by the tensor product. If (A) is a monoid object in (\mathcal C), with multiplication
[ m\colon A\otimes A\longrightarrow A ]
and unit
[ u\colon I_{\mathcal C}\longrightarrow A, ]
then a strong monoidal functor sends it to a monoid object in (\mathcal D). Its multiplication is the composite
[ F(A)\otimes F(A) \xrightarrow{\phi_{A,A}} F(A\otimes A) \xrightarrow{F(m)} F(A), ]
and its unit is obtained from (\phi_0) followed by (F(u)).
A monoidal equivalence consequently induces an equivalence between the categories of monoid objects. The same mechanism applies to comonoid objects, with the comparison isomorphisms used in the reverse direction. When the categories possess compatible braidings, a braided monoidal equivalence also transports commutative monoid objects.
This preservation is structural rather than objectwise. A particular monoid object need not be sent to an equal object or even to one with an identical underlying carrier. Its multiplication and unit are instead carried through the coherence isomorphisms of the equivalence.
Relation to one-object bicategories
A monoidal category can be regarded as a bicategory with one object. Its objects become 1-morphisms, its morphisms become 2-morphisms, and its tensor product becomes horizontal composition. Under this correspondence, the monoidal associator becomes the bicategorical associator, while the monoidal unit becomes the identity 1-morphism.
A monoidal equivalence then corresponds to a suitably structured equivalence between one-object bicategories. This interpretation accounts for the appearance of monoidal natural transformations and coherence modifications in higher-dimensional formulations. It also explains why equivalence, rather than isomorphism, is the stable notion of identity for monoidal structure.