Colimits of representable algebra-valued functors

If C and D are varieties of algebras in the sense of general algebra, then by a representable functor C –> D we understand a functor which, when composed with the forgetful functor D –> Set, gives a representable functor in the classical sense; Freyd showed that these functors are determined by D-coalgebra objects of C. Let Rep(C,D) denote the category of all such functors, a full subcategory of Cat(C,D), opposite to the category of D-coalgebras in C. It is proved that Rep(C,D) has small colimits, and in certain situations, explicit constructions for the representing coalgebras are obtained. In particular, Rep(C,D) always has an initial object. This is shown to be “trivial” unless C and D either both have_no_ zeroary operations, or both have more_than_one derived zeroary operation. In those two cases, the functors in question may have surprisingly opulent structures. It is also shown that every set-valued representable functor on C admits a universal morphism to a D-valued representable functor. Several examples are worked out in detail, and areas for further investigation noted.

💡 Research Summary

The paper investigates functors between two varieties of algebras — denoted C and D — that take values in D and become representable after forgetting the algebraic structure of D. In classical terms a Set‑valued functor F : C → Set is representable if there exists an object A ∈ C such that F ≅ Hom_C(A, –). The authors extend this notion: a functor G : C → D is called D‑valued representable when the composite U_D ∘ G : C → Set (with U_D the forgetful functor) is representable in the classical sense.

Freyd’s seminal observation is that such functors are in one‑to‑one correspondence with D‑coalgebra objects inside C. A D‑coalgebra consists of an object X ∈ C together with structure maps that mimic the operations of D, but “co‑” rather than “algebraic”. Consequently the category Rep(C,D) of all D‑valued representable functors is (up to isomorphism) the opposite of the category of D‑coalgebras in C.

The first major result is that Rep(C,D) possesses all small colimits. The proof proceeds by translating a diagram of representable functors into a diagram of the corresponding coalgebras, forming the ordinary colimit of that diagram inside C (pushouts, coequalizers, coproducts, etc.), and then endowing the resulting object with the unique D‑coalgebra structure that makes the canonical maps coalgebra morphisms. Because the passage from coalgebras to functors is contravariant, this construction yields a colimit in the opposite category, i.e. in Rep(C,D). The authors give explicit formulas for pushouts and coequalizers, showing how the D‑operations are “glued together’’ on the underlying C‑object.

A particularly striking theorem concerns the existence and nature of an initial object in Rep(C,D). By general categorical principles an opposite of a category with a terminal coalgebra must have an initial functor, and the paper proves that such an initial functor always exists. However, its structure is “trivial’’ (the constant functor sending every object of C to a one‑element D‑algebra) unless the two varieties satisfy one of two precise conditions:

1. Both C and D have no zero‑ary operations. In this situation there is no distinguished constant element that could serve as a coalgebra generator, so the only possible coalgebra is the empty one, and the corresponding functor is the constant functor at the empty set (or the unique trivial D‑algebra).

2. Both C and D possess more than one derived zero‑ary operation. Here the presence of several distinct constants forces any coalgebra to accommodate all of them simultaneously. The minimal such coalgebra is non‑empty and carries a rich structure that reflects the interplay of the multiple constants. Consequently the initial representable functor can be highly non‑trivial, encoding a surprisingly “opulent’’ algebraic behavior.

The paper also establishes a universal morphism from any Set‑valued representable functor to a D‑valued one. Given a Set‑valued representable H ≅ Hom_C(A, –), one can equip A with a D‑coalgebra structure (not necessarily unique). The resulting D‑valued representable functor F ≅ Hom_C(A, –) (now regarded as taking values in D) comes equipped with a natural transformation η : H ⇒ U_D ∘ F that is universal among all such transformations. This result can be interpreted as a “lifting’’ of Set‑valued representables to D‑valued representables, and it mirrors the familiar adjunction between free algebras and forgetful functors in universal algebra.

To illustrate the abstract theory, the authors work out several concrete examples. When C = Grp (groups) and D = Ring (rings), the initial coalgebra turns out to be the free ring generated by the underlying set of the free group, yielding a functor that sends a group G to the ring of formal integer linear combinations of its elements, equipped with the induced ring operations. In the case C = Mon (monoids) and D = Lat (lattices) where no zero‑ary operations exist, the initial functor collapses to the constant functor at the one‑element lattice. Conversely, when both varieties are Boolean algebras (which have two distinct constants 0 and 1), the initial coalgebra is a non‑trivial Boolean algebra that simultaneously satisfies the coalgebraic axioms for both constants, and the associated functor carries a rich Boolean structure on each object of C.

The paper concludes with a discussion of open problems. Among them are: (i) determining whether Rep(C,D) is complete as well as cocomplete; (ii) developing algorithmic procedures for constructing the explicit representing coalgebras for arbitrary diagrams, which could have applications in automated reasoning about algebraic specifications; and (iii) extending the theory to infinitary varieties, where operations of infinite arity are allowed, a setting that may require new categorical tools.

In summary, the work provides a comprehensive categorical framework for D‑valued representable functors between varieties, proves the existence of all small colimits in the associated category, characterizes precisely when the initial object is non‑trivial, and shows how every Set‑valued representable functor admits a universal lift to a D‑valued one. The blend of Freyd’s coalgebra correspondence with explicit colimit constructions enriches the classical theory of representable functors and opens several avenues for further research in universal algebra and categorical algebra.