1 Definition and basic construction
A functor category is a category built from two fixed categories \(C\) and \(D\). Its objects are functors from \(C\) to \(D\), and its morphisms are natural transformations between such functors. This construction packages an entire family of maps into a single categorical setting, making it possible to study diagrams, parameterized structures, and families of objects in a uniform way.
Functor categories are often denoted \([C, D]\) or \(D^C\). The notation reflects the idea that the category behaves like an “internal hom” when the ambient category supports such a structure. In practice, functor categories are used whenever one wants to treat all \(C\)-shaped diagrams in \(D\) as objects of a category themselves.
1.1 Functors as objects
An object of a functor category is a functor \(F: C \to D\). Such a functor assigns to each object of \(C\) an object of \(D\), and to each morphism in \(C\) a morphism in \(D\), in a way that preserves identities and composition.
From the standpoint of the functor category, each functor is regarded as a single object, even though it encodes a whole diagram in \(D\). This perspective is useful because it allows constructions on diagrams to be discussed using categorical language rather than elementwise arguments.
1.2 Natural transformations as morphisms
A morphism between two objects \(F, G: C \to D\) is a natural transformation \(\eta: F \Rightarrow G\). For every object \(c\) in \(C\), the transformation provides a morphism \(\eta_c: F(c) \to G(c)\) in \(D\), and these component maps must satisfy the usual naturality condition.
The naturality requirement ensures compatibility with the morphisms of the source category. In effect, a morphism in a functor category is not merely a collection of arrows in \(D\), but a coherent family that respects the structure of the diagram.
1.3 Notation and conventions
Different authors use different notations for functor categories, though the underlying construction is the same. The choice of notation often depends on whether one emphasizes exponentiation-like behavior or the categorical perspective of diagrams and natural transformations.
1.3.1 Exponential notation
The notation \(D^C\) is used when the functor category is viewed as an exponential object. This convention is especially common in contexts where cartesian closed structure is important, since it suggests that functors from \(C\) to \(D\) play a role analogous to functions from one set to another.
1.3.2 Bracket notation
The notation \([C, D]\) is standard in many category theory texts. It highlights the fact that the category consists of morphisms from \(C\) to \(D\) viewed as objects, together with natural transformations between them. This notation is often preferred when discussing functor categories in a general categorical setting.
1.4 Identity morphisms and composition
The identity morphism on an object \(F: C \to D\) is the identity natural transformation, whose components are the identity morphisms \(\mathrm{id}_{F(c)}\) for each object \(c\) in \(C\). Composition of morphisms is given by vertical composition of natural transformations, defined componentwise.
These operations satisfy the axioms of a category because the identities and composition in \(D\) do so pointwise. As a result, the collection of functors \(C \to D\) and natural transformations between them forms a genuine category.
2 Examples
Functor categories appear naturally in many familiar settings, often without being named explicitly. They provide a systematic way to organize families of objects or maps indexed by a category.
2.1 Functor categories from small categories
When \(C\) is a small category and \(D\) is any category, the functor category \([C, D]\) consists of all \(C\)-shaped diagrams in \(D\). For example, if \(C\) is a finite indexing category, objects of \([C, D]\) are finite diagrams and morphisms are maps of diagrams.
This case is especially important because many categorical constructions are defined by specifying how they act on arbitrary diagrams. Functor categories provide the natural home for those diagrams.
2.2 Presheaf categories
A presheaf category is a special kind of functor category in which the source is replaced by its opposite category. Such categories are central in modern category theory and form one of the most important classes of functor categories.
2.2.1 Covariant and contravariant presheaves
A contravariant presheaf on \(C\) is a functor \(C^{\mathrm{op}} \to \mathbf{Set}\). A covariant presheaf is a functor \(C \to \mathbf{Set}\). The contravariant version is the one most commonly used in geometry, topology, and topos theory.
Both kinds of presheaf organize data assigned to objects of \(C\), together with restriction or transport maps induced by morphisms. The contravariant form is particularly well suited to local-to-global constructions.
2.2.2 Sets-valued presheaves
When the target category is \(\mathbf{Set}\), presheaves become especially concrete. They may be thought of as collections of sets varying over the objects of a category, with structure maps determined by morphisms in the source.
Sets-valued presheaves are fundamental because many other presheaf-like structures can be expressed by enriching or modifying this basic case. They also serve as the prototype for sheaves and related objects.
2.3 Diagram categories
Functor categories are often called diagram categories when their objects are interpreted as diagrams of a fixed shape. If \(C\) is an indexing category, then an object of \([C, D]\) is a diagram in \(D\) indexed by \(C\).
This viewpoint is useful for describing commutative squares, chains, cones, and other structured collections of objects in a category. Diagram categories make it possible to treat families of related objects as single entities.
2.4 Discrete source categories
If \(C\) is a discrete category, then a functor \(C \to D\) is simply a choice of object of \(D\) for each object of \(C\). Natural transformations between such functors are just families of morphisms in \(D\), one for each object of \(C\).
In this case, the functor category reduces to a product of copies of \(D\), one for each object of the discrete category. This gives a simple model for understanding how functor categories generalize products.
2.5 Group actions as functors
A group can be viewed as a category with one object whose morphisms are the group elements. A functor from such a category to another category \(D\) encodes an action of the group on an object of \(D\), together with coherent structure maps.
When \(D\) is \(\mathbf{Set}\), \(\mathbf{Vect}\), or another familiar category, these functors describe ordinary group actions, representations, or similar structured actions. Natural transformations then correspond to equivariant maps.
3 Categorical properties
Functor categories inherit many structural features from the target category, often in a pointwise manner. This makes them highly tractable and explains why they are so widely used in advanced category theory.
3.1 Existence of limits and colimits
If the target category \(D\) has certain limits or colimits, then the functor category \([C, D]\) typically has them as well, computed object by object. This pointwise behavior is one of the most important features of functor categories.
3.1.1 Pointwise computation
A limit or colimit in a functor category is obtained by applying the corresponding limit or colimit in \(D\) at each object of \(C\). In other words, one forms the limit or colimit of the values of the relevant functors at each index separately.
This method works because naturality ensures compatibility among the component constructions. It turns many abstract categorical problems into families of more familiar ones in \(D\).
3.1.2 Preservation of structure
Structural properties of \(D\), such as the existence of products, equalizers, or coproducts, often transfer to \([C, D]\). The preservation is not automatic in every setting, but for many standard constructions the functor category reflects the same organizing principles as the target.
This feature is especially useful in algebra and topology, where complex objects can be handled one component at a time.
3.2 Completeness and cocompleteness
If \(D\) is complete, then \([C, D]\) is complete; if \(D\) is cocomplete, then \([C, D]\) is cocomplete, under standard hypotheses. Thus functor categories often inherit a large supply of categorical limits and colimits from their target.
This makes them suitable environments for solving universal mapping problems. It also supports many constructions in homological and homotopical contexts.
3.3 Functor categories as categories of diagrams
Functor categories can be seen as categories whose objects are structured families indexed by \(C\). Morphisms are then compatible transformations between those families, rather than arbitrary pointwise maps.
This interpretation clarifies why functor categories are so useful in studying commutative diagrams, diagram-chasing arguments, and systematic parameter dependence. Many constructions on diagrams become ordinary categorical constructions in the corresponding functor category.
3.4 Terminal and initial objects
If \(D\) has a terminal object, then the constant functor at that object is terminal in \([C, D]\). Similarly, if \(D\) has an initial object, the corresponding constant functor is initial.
These objects are determined pointwise. Their existence illustrates a general theme: many categorical properties of the target category extend to the functor category by applying them uniformly across all objects of the source.
3.5 Isomorphisms and equivalences
Isomorphisms in a functor category are natural isomorphisms. That is, a natural transformation is an isomorphism precisely when each component morphism is an isomorphism in \(D\).
Equivalences between functor categories can arise from equivalences between the source or target categories, though the precise relationship depends on how the functors interact with the indexing structure. This makes functor categories sensitive to both the shape of diagrams and the ambient category in which they live.
4 Pointwise behavior
Many constructions in functor categories are performed pointwise, meaning they are defined separately at each object of the source category and then assembled into a functor or natural transformation. This is one of the most powerful simplifying principles in the subject.
4.1 Pointwise limits
A limit of a family of functors \(F_i: C \to D\) is often obtained by taking the limit of the objects \(F_i(c)\) in \(D\) for each \(c\) in \(C\). The resulting assignment \(c \mapsto \lim_i F_i(c)\) can usually be made into a functor.
This pointwise process is compatible with the action on morphisms because the limit cones are natural in the indexing object. It provides a direct way to construct limits of diagrams of functors.
4.2 Pointwise colimits
Colimits behave similarly: one computes the colimit of the values \(F_i(c)\) in \(D\) for each object \(c\) in \(C\). The resulting objectwise colimit often defines the colimit functor in \([C, D]\).
This method is especially effective in categories such as \(\mathbf{Set}\), \(\mathbf{Ab}\), and \(\mathbf{Vect}\), where many familiar colimits exist and can be described concretely.
4.3 Evaluation functors
For each object \(c\) of \(C\), there is an evaluation functor \(\mathrm{ev}_c: [C, D] \to D\) sending a functor \(F\) to its value \(F(c)\). On morphisms, it sends a natural transformation to its component at \(c\).
Evaluation functors are central tools because they connect the global structure of a functor category with the local data at individual objects of the source category. Many arguments in functor categories reduce to checking statements after evaluation.
4.4 Pointwise monomorphisms and epimorphisms
A natural transformation is often a monomorphism or epimorphism in the functor category precisely when each of its components has the corresponding property in \(D\), under suitable assumptions. This pointwise criterion is one of the advantages of working in functor categories.
The result reflects the fact that categorical notions of injectivity and surjectivity-like behavior can be verified locally on each object of the source. It is especially useful in categories of diagrams and presheaves.
4.5 Pointwise adjunctions
Adjunctions between functor categories are frequently induced pointwise from adjunctions between target categories. If a pair of functors between categories has an adjoint relationship, that relationship may extend to diagram categories by applying the adjunction at each object of the source.
This produces a flexible framework for transporting universal properties through functor categories. It also underlies many constructions in algebraic and homotopical settings.
5 Relations to other constructions
Functor categories are closely related to several major categorical constructions. Their connections to presheaves, exponentials, enriched categories, opposite categories, and products reveal their role as a unifying concept.
5.1 Presheaf categories and sheaves
Presheaf categories are among the most important examples of functor categories. Sheaves arise as subcategories of presheaf categories defined by additional gluing conditions, so the functor category provides the ambient structure in which sheaf theory is developed.
This relationship is foundational in modern geometry and topology. The presheaf category supplies the raw data, while the sheaf condition selects the objects that behave well under local-to-global reconstruction.
5.2 Exponential objects in cartesian closed categories
In a cartesian closed category, functor categories help explain the meaning of exponential objects. The notation \(D^C\) suggests that the functor category behaves like a categorical function space.
This analogy is not merely notational. Under suitable conditions, functor categories embody the universal property expected of exponentials, making them central to the study of cartesian closed structure.
5.3 Enriched functor categories
When categories are enriched over another monoidal category, one can define enriched functor categories whose hom-objects carry extra structure. These categories generalize ordinary functor categories by replacing sets of natural transformations with structured hom-objects.
Enriched functor categories are important in contexts where morphisms carry more information than plain sets can encode. Examples include categories enriched over abelian groups, chain complexes, or topological spaces.
5.4 Opposite categories
If \(C\) is replaced by its opposite category \(C^{\mathrm{op}}\), the corresponding functor category changes from covariant to contravariant diagrams. This is the basis of presheaf theory and many duality arguments in category theory.
Using opposite categories often reveals a dual version of a construction with the arrows reversed. Functor categories provide a flexible way to express both directions in a unified framework.
5.5 Product categories
When the source category is discrete, the resulting functor category is essentially a product category. More generally, functor categories can be viewed as organized collections of objects from \(D\) indexed by the objects and arrows of \(C\).
This relationship shows how products arise as a special case of diagram categories. It also illustrates how the source category controls the shape of the resulting categorical structure.
6 Advanced topics
Functor categories support a rich collection of deeper results and constructions. Many of these depend on universal properties and interact strongly with standard tools such as the Yoneda lemma and Kan extensions.
6.1 Yoneda lemma in functor categories
The Yoneda lemma can be formulated naturally in the language of functor categories. It identifies morphisms out of representable functors with elements or natural transformations satisfying a precise universal property.
This perspective is essential in understanding why presheaves are so powerful. Representable functors sit inside the presheaf category as canonical objects that encode the behavior of the source category itself.
6.2 Natural transformations and transformations of diagrams
Natural transformations can be interpreted as morphisms between diagrams of the same shape. This makes them a categorical analogue of coherent pointwise comparison.
In practice, such transformations are used to compare constructions that vary over an indexing category. Their compatibility conditions ensure that the comparison respects all structure maps in the diagram.
6.3 Kan extensions
Kan extensions are among the most important constructions associated with functor categories. They generalize the process of extending a functor along another functor while preserving universal properties.
Left and right Kan extensions can often be described as colimit-like or limit-like constructions indexed by comma categories. They play a central role in categorical algebra, sheaf theory, and homotopical methods.
6.4 Limits in functor categories
The existence and computation of limits in functor categories are frequently used to prove more general theorems. Since limits are often taken pointwise, one can reduce many proofs about diagram categories to corresponding proofs in the target category.
This strategy is particularly valuable when dealing with infinite diagrams or iterated constructions. The functor category provides a controlled setting for organizing such arguments.
6.5 Adjunctions induced by functor categories
Adjunctions can often be lifted to functor categories by applying them objectwise. This yields a powerful mechanism for transferring universal properties and constructing new adjoint pairs.
Such induced adjunctions appear throughout category theory and its applications. They are especially useful when working with limits, colimits, and extension problems in categories of diagrams.
7 Applications
Functor categories have broad applications across mathematics. They serve as a common language for studying families of algebraic, geometric, and logical structures.
7.1 Homological algebra
In homological algebra, functor categories organize chain complexes, derived functors, and diagrams of modules or abelian groups. Many constructions are naturally functorial, so the functor category framework clarifies how they behave under maps between indexing categories.
This setting also supports diagram chasing and the systematic treatment of exactness properties. Functor categories provide a convenient environment for comparing homological constructions across different objects.
7.2 Representation theory
Representation theory often interprets groups, algebras, or other algebraic structures as functors into categories such as \(\mathbf{Vect}\). In this form, representations become objects in a suitable functor category.
Natural transformations then correspond to intertwining maps, making functor categories a natural language for studying morphisms between representations. This categorical viewpoint helps unify many classical examples.
7.3 Topos theory
Topos theory makes extensive use of presheaf categories and sheaves, both of which arise from functor categories. These categories provide models of generalized spaces and support a rich internal logic.
The ability to treat varying data as objects of a category is essential in this area. Functor categories supply the ambient framework in which local information can be assembled into global geometric or logical structures.
7.4 Algebraic topology
In algebraic topology, functor categories appear in the study of systems of invariants such as homology and cohomology. Diagrams of spaces, groups, or chain complexes are often best handled categorically.
They also arise in the organization of simplicial objects and other structured diagrams. This makes functor categories useful in both classical and modern homotopical methods.
7.5 Logic and semantics
In logic, functor categories support categorical semantics for variable-dependent or context-dependent constructions. Presheaf categories, in particular, are used to model changing information and interpret logical systems in a categorical setting.
The functorial viewpoint helps formalize how meanings vary with context. It also plays a role in the semantics of type theories and related foundational frameworks.