1 Definition and Universal Property
A colimit is a way to assemble the data of a diagram into one object that best captures all of its compatible identifications and combinations. The definition is categorical and depends on a chosen indexing category, but the guiding idea is familiar: one takes many pieces and merges them according to specified arrows. Colimits are determined by a universal property, so when they exist they are unique up to unique isomorphism.
1.1 Diagrams and Cocones
A colimit begins with a diagram, meaning a functor from an indexing category into the category under study. The diagram selects a family of objects and morphisms that are to be combined. To compare candidate combinations, one uses cocones, which are maps from every object in the diagram into a single target object, with compatibility along the diagram’s arrows.
1.1.1 Functors as Indexing Schemes
The indexing category encodes the shape of the construction. A functor from that category to a category \(\mathcal{C}\) assigns an object of \(\mathcal{C}\) to each index and a morphism to each arrow in the shape. In this sense, the functor acts as a blueprint for how the pieces are related.
1.1.2 Cocones and Their Morphisms
A cocone on a diagram consists of a target object together with compatible morphisms from each diagram object into that target. A morphism between cocones is a map between their apex objects that commutes with all the structure maps. This makes cocones into a category, within which the colimit is an initial object.
1.2 Colimit as a Universal Object
The colimit is the cocone that factors every other cocone in a unique way. This universal characterization is what gives the concept its force across many categories, since it does not depend on element-based reasoning. Once a colimit exists, it is canonical up to unique isomorphism.
1.2.1 Uniqueness Up to Unique Isomorphism
Any two colimits of the same diagram are uniquely isomorphic. Moreover, there is only one isomorphism compatible with the cocone structure. This makes the resulting object essentially independent of arbitrary choices.
1.2.2 Universal Mapping Property (UMP)
The universal mapping property says that for every cocone from the diagram to another object, there is a unique mediating morphism from the colimit to that object. This mediating morphism makes all triangles commute. The UMP is the defining feature of colimits and is often the most useful way to verify them.
2 Basic Examples of Colimits
Many common constructions in mathematics are special cases of colimits. These examples illustrate how the abstract definition captures familiar operations such as combining objects, imposing identifications, and gluing spaces together.
2.1 Coproducts
Coproducts are the categorical analogue of a disjoint combination of objects. They are the simplest nontrivial example of a colimit and arise from a diagram with no nonidentity arrows.
2.1.1 Disjoint Union Viewpoint
In categories like sets, a coproduct is formed by taking disjoint unions. Each summand remains distinguishable inside the total object. This reflects the fact that coproducts combine objects without identifying their elements unless the category’s structure forces that behavior.
2.1.2 Universal Property Formulation
A coproduct comes with injections from each object into the combined object. Given maps from each summand into any target, there is a unique map from the coproduct extending them. This is exactly the colimit property for a discrete diagram.
2.2 Coequalizers
Coequalizers identify elements or substructures that two parallel arrows send to the same place. They are a standard method for imposing relations and for forming quotient objects in a categorical way.
2.2.1 Quotient-by-Relation Intuition
A coequalizer can be viewed as the result of forcing two morphisms to agree. In concrete categories, this often resembles taking a quotient by the smallest relation compatible with the given maps. The construction preserves precisely the information that remains after the imposed identification.
2.2.2 Constructing Coequalizers via Universal Arrows
The coequalizer is the universal arrow out of the codomain of the parallel pair that equalizes them. Any other arrow that also equalizes the pair factors uniquely through it. This universal description is what makes coequalizers central in algebra and logic.
2.3 Pushouts
Pushouts combine two objects that share a common part. They provide a formal version of gluing along a shared subobject and are among the most useful colimits in geometry and algebra.
2.3.1 Gluing Along a Shared Subobject
A pushout takes two morphisms with a common domain and identifies their images in a minimal universal way. In spaces, this often corresponds to attaching one space to another along a chosen map. The result records both pieces while respecting the specified overlap.
2.3.2 Universal Property and Commuting Squares
The pushout is characterized by a commuting square that is universal among all such squares with the same left side. Any compatible pair of maps out of the two objects factors uniquely through the pushout. This property makes pushouts the categorical form of amalgamation.
2.4 Sequential and Directed Colimits
Directed colimits combine a family of objects arranged by a directed system. They are especially important in algebra and topology, where one builds large objects from increasing stages.
2.4.1 Direct Limits (Inductive Constructions)
Direct limits are colimits over directed indexing categories. They arise when objects are assembled along a compatible family of maps that becomes progressively larger or more refined. Such limits are often used to define infinite objects by stages.
2.4.2 Colimits of Chains
A chain is a sequence of objects and maps arranged linearly. Its colimit captures the eventual union of the chain in categories where such unions make sense. Chains provide one of the most accessible forms of directed colimits.
3 Colimits in Specific Categories
The abstract notion of colimit takes different concrete forms depending on the category. In each setting, the universal property remains the same, but the underlying objects and maps change the interpretation.
3.1 Set
In the category of sets, colimits are built from standard set-theoretic constructions. The category is especially helpful because it provides a concrete model for the general theory.
3.1.1 Coproducts as Disjoint Unions
Coproducts in Set are disjoint unions. Each set embeds into the total set as a separate component. This realizes the idea of combining sets without accidental overlap.
3.1.2 Coequalizers as Quotients
Coequalizers in Set are quotient sets by the smallest equivalence relation that identifies elements mapped together by the parallel arrows. The quotient map is universal among maps that respect that identification. This is the set-theoretic prototype for many other quotient constructions.
3.2 Groups and Monoids
In algebraic categories, colimits often encode generators, relations, and amalgamated constructions. They are useful because they package algebraic presentations in a compact universal form.
3.2.1 Free Product and Coproducts
In groups, the coproduct is the free product. It combines groups while preserving their internal structures and imposing no extra relations beyond those already present. For monoids, a similar universal combination exists, though its behavior is distinct from the group case.
3.2.2 Presentations via Coequalizers
Group and monoid presentations can be described using coequalizers. One starts with freely generated objects and then identifies relations by quotienting through a universal construction. This viewpoint clarifies how generators and relations fit into categorical language.
3.3 Topological Spaces
In topology, colimits often appear as quotient or identification spaces. They formalize the process of attaching spaces along maps and are widely used in geometric constructions.
3.3.1 Colimits and Quotient/Identification Spaces
Many topological colimits are formed by taking a disjoint union and then imposing identifications dictated by the diagram. The resulting space carries the final topology with respect to the canonical maps. This makes colimits a natural language for gluing.
3.3.2 Behavior Under Different Category Structures
The categorical colimit in Top depends on the topology placed on the underlying set after gluing. The resulting construction may differ from the set-theoretic colimit because continuity conditions must be preserved. Thus the topology plays an essential role in the final universal object.
3.4 Functor Categories
Colimits in categories of functors are often computed objectwise. This makes them especially manageable and shows how universal constructions can be transferred across many contexts.
3.4.1 Computing Colimits Pointwise
If a functor category has the relevant colimits, they are typically formed at each object of the source category separately. One computes the colimit in the target category for each component. This pointwise behavior is a major practical advantage.
3.4.2 Natural Transformations and Componentwise Constructions
Natural transformations ensure that the pointwise colimits assemble coherently into a new functor. Compatibility across components is what makes the construction functorial. The resulting colimit reflects the structure of the entire diagram of functors, not just isolated pieces.
4 Duality with Limits
Colimits are the dual notion to limits. Many statements about one have a mirror image for the other, with arrows reversed and universal properties exchanged.
4.1 Limits as Dual Constructions
Limits gather compatible maps into a single object that maps into each member of a diagram. Colimits reverse the direction of these comparison maps. The duality is not merely formal; it organizes much of category theory into paired concepts.
4.1.1 Reversing Arrows and Index Categories
To obtain a limit statement from a colimit statement, one reverses the arrows in the indexing diagram and in the target maps. This arrow reversal turns cocones into cones. The same pattern appears throughout categorical duality.
4.1.2 Consistency of Universal Properties
Universal properties remain stable under dualization, but their direction changes. Initial objects become terminal objects, and colimit constructions correspond to limit constructions in the opposite category. This consistency is one reason category theory can be efficiently organized by dual pairs.
4.2 Common Consequences of Duality
Duality allows many theorems to be stated twice, once for limits and once for colimits. It also reveals structural parallels between constructions that may look unrelated at first glance.
4.2.1 Relation to Ends and Coends
Ends and coends are higher-level categorical constructions related respectively to limits and colimits. A coend may be regarded as a generalized colimit with additional identifications built into its variance. This places it naturally on the colimit side of the duality picture.
4.2.2 Understanding Colimits via Limit Theorems
Many facts about colimits can be inferred by applying known limit results in the opposite category. This method is common in proofs and helps reduce duplication. Duality thus serves as a conceptual and technical tool.
5 Existence and Preservation of Colimits
Not every category has every colimit. Existence depends on structural properties of the category, while preservation concerns how colimits behave under functors.
5.1 Conditions for Existence
To say that a category admits certain colimits means that those universal constructions exist whenever the corresponding diagrams are given. Some categories are especially rich in colimits, while others support only limited forms.
5.1.1 Cocompleteness of Categories
A category is cocomplete if it has all small colimits. This is a strong structural property and is often assumed in applications. Many familiar categories of algebraic and geometric objects are cocomplete.
5.1.2 Smallness Conditions on Indexing
The size of the indexing category matters. A “small” diagram is one indexed by a small category, and many existence theorems are stated only for such diagrams. Size restrictions help avoid set-theoretic difficulties.
5.2 Preservation and Reflection
Functors can interact with colimits in several ways. Some preserve them exactly, while others reflect their presence or structure.
5.2.1 Colimit-Preserving Functors
A functor preserves colimits if it carries the colimit of a diagram to the colimit of the image diagram. Such functors are often called left exact only in the broader sense of categorical behavior, though more precisely they are cocontinuous when preserving relevant colimits. Preservation is important because it lets constructions survive passage between categories.
5.2.2 Right/Left Adjoint Theorems
Left adjoint functors typically preserve colimits. This is one of the central links between universal constructions and adjunctions. It explains why free constructions, which are often left adjoints, naturally commute with colimit formation.
5.2.3 Exactness-Style Behavior
In many algebraic contexts, colimits interact predictably with other operations such as quotients, sums, and tensor-like constructions. This behavior is often summarized by exactness properties. While not every exactness notion is identical across categories, colimit compatibility is a recurring theme.
6 Operations and Calculations with Colimits
Colimits can often be combined, rearranged, or computed through related constructions. These operations make them practical as well as conceptual tools.
6.1 Composition of Colimits
Colimits may be formed in stages. Under suitable hypotheses, iterated constructions can be simplified or reorganized.
6.1.1 Iterated Colimits
An iterated colimit is a colimit taken after another colimit has already been formed. In many settings, such constructions can be exchanged or reduced to a single colimit over a combined indexing category. This flexibility is useful in complex calculations.
6.1.2 Changing Index Categories
Sometimes a diagram can be reindexed without altering its colimit. If the new indexing category is suitably related to the old one, the resulting universal object remains equivalent. This allows one to choose the most convenient shape for a computation.
6.2 Colimits of Functorial Constructions
When diagrams themselves vary in more than one variable, colimits can be taken in a structured way. Functoriality ensures that these constructions remain coherent.
6.2.1 Colimits in Multiple Variables
A multivariable diagram may admit colimits taken in one variable at a time or jointly. Under suitable conditions, these approaches agree. Such interchange principles are important in advanced categorical arguments.
6.2.2 Natural Transformations Between Diagrams
A natural transformation between diagrams induces a comparison between their colimits when the relevant universal objects exist. The transformation may produce a map from one colimit to another, reflecting the way the diagrams are related. This provides a categorical method for comparing constructions.
6.3 Coends and Other Related Colimit Constructions
Some constructions generalize colimits by incorporating variance and parameterization. Coends are among the most prominent examples.
6.3.1 Coend as a Canonical Colimit
A coend can be expressed as a quotient of a coproduct indexed by objects of a category, subject to relations coming from morphisms. This makes it a canonical colimit-like construction. It often arises when combining a functor of two variables in a balanced way.
6.3.2 Relationship to Enriched Contexts
In enriched category theory, colimits and coends are formulated relative to a monoidal base. The same universal ideas persist, but hom-objects replace hom-sets. This extension broadens the scope of colimit methods while preserving their conceptual core.
7 Colimits in Theory of Kan Extensions
Kan extensions provide a systematic framework for extending functors along other functors. Left Kan extensions are especially closely tied to colimits.
7.1 Left Kan Extensions as Colimit Expressions
A left Kan extension can often be computed by taking a colimit over a suitable comma category. This expresses an extension problem as an aggregation problem. The resulting construction is universal among all extensions with the same source data.
7.1.1 Computing Extensions via Colimits
In concrete terms, the value of a left Kan extension at an object is frequently built from a colimit of values over all ways of mapping into that object. This local-to-global recipe is a powerful computational tool. It reduces an extension problem to a family of colimit calculations.
7.1.2 Universal Properties in Extension Problems
The left Kan extension is defined by a universal property analogous to that of a colimit. It is the best approximation to extending a functor along another functor while preserving the relevant structure. This link explains why colimits are central in the theory.
8 Applications
Colimits are widely used to construct algebraic objects, describe glued spaces, and simplify diagrammatic arguments. Their utility comes from the fact that they package complex identifications into a single universal result.
8.1 Building Algebraic Structures
Many algebraic objects are defined by generators and relations, which align naturally with colimit constructions. Colimits thus provide an efficient formal language for building and presenting structures.
8.1.1 Presentations and Quotients via Coequalizers
Presentations can often be encoded by coequalizers of free objects. Generators arise from free constructions, while relations are imposed by the identifying arrows. The resulting quotient captures exactly the intended algebraic structure.
8.1.2 Gluing Constructions via Pushouts
Pushouts are used to combine algebraic or geometric pieces along shared substructures. They are a natural choice when an object is formed by attaching one component to another. This makes them especially valuable in recursive constructions.
8.2 Diagrammatic Reasoning
Colimits help replace elaborate networks of morphisms with a single universal object. This can make proofs shorter and conceptual.
8.2.1 Replacing Complex Diagrams by Universal Objects
Instead of manipulating many compatible arrows separately, one can often pass to the corresponding colimit. The universal object summarizes the entire diagram in a single target. This reduction simplifies both statements and proofs.
8.2.2 Compositionality in Categorical Proofs
Because colimits are defined by universal properties, they compose well with other categorical constructions. Proofs can often be organized by showing that one universal property implies another. This compositional style is one of the main reasons category theory is effective as a unifying language.