Skip to content

3-Category

A higher category with objects and composable morphisms in dimensions one through three, under specified coherence laws.

Version
v1 · 2026-09-28 · History
Domain-specific #
7811
Domain group
Formal Sciences
Origin domain
Mathematics
Subdomain
Higher Category Theory → Mathematics

Core Idea

A 3-category extends categorical organization through a third level of morphisms. Objects are related by 1-morphisms, parallel 1-morphisms by 2-morphisms, and parallel 2-morphisms by 3-morphisms; composition at these levels obeys compatibility and coherence laws. That final layer—not the numeral in the name—is the feature a merely two-dimensional category lacks.

Strict, Gray/semistrict, and weak 3-categories impose different equations or coherent replacements. A strict 3-category can be described through enrichment over 2Cat, while the Gray tensor provides a semistrict account. These variants should not be collapsed, even though coherence results relate appropriate weak forms to Gray presentations.

How would you explain it like I'm…

Paths Between Paths Between Paths

Picture towns as dots, and roads going from town to town. Now imagine ways to slide one road into another road between the same two towns, and then ways to change one of those slides into another slide. A 3-category keeps track of all these layers at once, with careful rules for joining things together at every layer.

Three Layers of Arrows

Mathematicians sometimes draw things as dots with arrows between them; with rules for joining arrows end to end, that's called a category. A 2-category adds arrows between arrows, and a 3-category adds one more level: arrows between those arrows. So there are dots, arrows, arrows-between-arrows, and arrows-between-those. The important part is that extra third level of arrows, not the number 3 itself. At every level, things can be joined together, and the joinings have to agree with each other.

Category With 3-Morphisms

A category has objects and morphisms (arrows) that compose. A 3-category builds this up two more times: 1-morphisms go between objects, 2-morphisms go between parallel 1-morphisms (ones with the same start and end), and 3-morphisms go between parallel 2-morphisms. Composition works at every level and must obey compatibility and coherence laws so that different ways of combining things agree. The third layer of morphisms is exactly what a 2-category lacks. There are different versions, strict, semistrict (Gray) and weak, depending on whether certain laws hold exactly or only up to a specified higher morphism, and they should not be treated as the same thing.

 

A 3-category extends category theory by a third level of morphisms: objects, 1-morphisms between objects, 2-morphisms between parallel 1-morphisms, and 3-morphisms between parallel 2-morphisms, with composition at each level subject to compatibility and coherence laws. The distinguishing feature relative to a 2-category is this top layer, not the numeral in the name. Variants impose different amounts of strictness. A strict 3-category can be defined as a category enriched over 2Cat (the category of strict 2-categories), so all laws hold as equations. A Gray (semistrict) 3-category is built using the Gray tensor product instead, and weak 3-categories replace many equations with coherent higher morphisms. Coherence results relate appropriate weak forms to Gray presentations, but these variants should not be treated as interchangeable.

Structural Signature

Sig role-phrases:

  • Objects — Provide the 0-dimensional entities on which higher morphisms are arranged. It is constitutive. Counterfactual: Without objects there is no carrier for the 1-morphisms and their higher cells.
  • 1-morphisms — Relate objects and serve as boundaries for 2-morphisms. It is constitutive. Counterfactual: Remove them and the structure loses its category-like first layer.
  • 2-morphisms — Relate parallel 1-morphisms and provide boundaries for 3-morphisms. It is constitutive. Counterfactual: Without them a 3-morphism has no properly typed 2-cell endpoints.
  • 3-morphisms — Relate parallel 2-morphisms and distinguish the third categorical dimension from a 2-category. It is constitutive. Counterfactual: Omitting this layer collapses the described structure to at most a 2-category.
  • Composition and coherence — Specify how cells compose within and across dimensions, strictly or up to coherent comparison cells. It is constitutive. Counterfactual: A mere list of three levels with no compositional laws is not a 3-category.

What It Is Not

  • Not a category with exactly three objects. The number refers to morphism dimension, not object count.
  • Not merely a 2-category renamed. A genuine third level of cells and its composition must be specified.
  • Not one fixed strictness convention. Strict, semistrict, and weak versions differ in how composition laws hold.
  • Not any three-layer hierarchy. Typing, composition, and coherence distinguish the categorical object from an ordinary hierarchy.
  • Closest near-miss. A 2-category has the lower levels and their composition, but lacks the constitutive third morphism level.

Scope of Application

  • Higher category theory. Distinguish strict, Gray, and weak constructions by their coherence requirements.
  • Categories of 2-categories. 2Cat exhibits how transformations and modifications add a third morphism level.
  • Algebraic topology. Fundamental 3-groupoids organize points, paths, and higher homotopies.
  • Coherence comparison. Use equivalence results to relate weak and semistrict presentations without identifying them as identical objects.

Clarity

The name is made precise by enumerating the typed levels and the compositions they support. A 3-morphism goes between 2-morphisms; it is not simply a third arrow between objects. Likewise, saying that weak and Gray forms are related by coherence does not make their definitions word-for-word interchangeable. The strictness convention should accompany any purported example.

Manages Complexity

Higher-categorical constructions contain many interacting compositions and identity laws. The 3-category framework groups these into levels and coherence constraints, allowing a reader to ask which cells compose and how competing compositions compare. Strict presentations simplify equations; Gray or weak presentations retain nontrivial coherence. The framework organizes complexity, but it does not make the coherence theorems automatic or remove size and typing issues.

Abstract Reasoning

  1. Identify the objects and the first-, second-, and third-level morphisms.
  2. Check that each higher morphism has correctly typed parallel lower-dimensional boundaries.
  3. Specify composition and identities at every relevant level.
  4. Declare whether the structure is strict, Gray/semistrict, or weak.
  5. Test the relevant interchange and coherence requirements rather than assuming all composites are equal.
  6. When comparing presentations, state whether the result is literal identity or an appropriate equivalence.

Knowledge Transfer

The term transfers literally among mathematical settings with objects, three typed morphism levels, and valid coherence—such as categories of 2-categories and fundamental 3-groupoids. A project team with three management levels or a database with three relation types is only an analogy unless categorical composition and higher cells are actually defined. The broader portable idea is layered relations, not this exact mathematical structure.

Examples

Canonical

Baez identifies 2Cat as the 3-category formed from 2-categories. Its objects are 2-categories; 2-functors serve as 1-morphisms; transformations between 2-functors provide the next layer; modifications between those transformations supply a third. The example shows why a category of lower-dimensional categories naturally adds one more dimension. The exact strictness convention must be specified when choosing the kind of transformation.

Mapped back: Objects → 2-categories; 1-morphisms → 2-functors; 2-morphisms → transformations between 2-functors; 3-morphisms → modifications; Composition and coherence → composition at each typed level under the chosen convention.

Applied / In Practice

Baez's dimensional-ladder outline gives the fundamental 3-groupoid of a topological space, specifically Π₃(S²), as a semistrict 3-category example. Points, paths, homotopies of paths, and higher homotopies supply successive levels, with invertibility appropriate to a groupoid. This is a topological instance of higher-categorical organization, not simply a synonym for 'sphere' or a generic 2-category.

Mapped back: Objects → points of S²; 1-morphisms → paths; 2-morphisms → homotopies of paths; 3-morphisms → higher homotopies; Composition and coherence → semistrict composition appropriate to the fundamental 3-groupoid.

Structural Tensions

T1 — Strict Equations versus Faithful Higher Structure. Strict equality makes composition easier to state and calculate, but can suppress higher coherence information. Weak or semistrict formulations retain more structural variation while requiring explicit coherence data.

Diagnostic: Does the application need strict identities, or only coherent equivalence between composites?

T2 — General Definition versus Tractable Presentation. A weak tricategory allows the most general third-level composition, while Gray categories provide a semistrict presentation related by coherence. Treating a convenient presentation as literal equality can lose the equivalence-versus-identity distinction.

Diagnostic: Is a comparison an equality of presented objects or an equivalence under a coherence theorem?

T3 — Named Third Level versus Higher-Category Umbrella. A 3-category has a distinctive third morphism level and coherence problem, while the broader idea of layered morphisms extends to n-categories. Calling every layered hierarchy a 3-category would drop the exact typing and composition requirements.

Diagnostic: Are there genuine 3-morphisms with categorical composition, or only three informal layers?

Structural–Framed Character

A 3-category is structural-leaning within formal mathematics, not a purely substrate-free hierarchy. Evaluative weight: satisfying the typed composition and coherence laws is a mathematical membership test, not praise for a structure with three levels. Human-practice-bound: once the axioms are fixed, their consequences do not depend on a user's preference, but choosing strict, Gray, or weak conventions is part of mathematical practice. Institutional origin: no agency makes the cells compose; higher-category theory supplies the definitions and proof standards by which a proposed example is recognized. Vocabulary travels: objects, higher morphisms, and composition transfer literally to 2Cat or a fundamental 3-groupoid when the cells and laws are specified, but not to an ordinary management hierarchy. Import versus recognize: an informal three-tier organization imports categorical language by analogy; a new mathematical model with genuine 3-morphisms is another recognition of the same formal kind.

The portable skeleton is typed higher-order composition: relations between relations can themselves be related under compatibility laws. No current strict parent prime is asserted for that full skeleton; it remains a future-prime candidate rather than an excuse to attach this node to a merely related idea such as general Composition. The objects, three morphism dimensions, and selected coherence regime remain the domain-bound identity. Its character: a precise higher-categorical structure whose formal transfer requires the whole typed apparatus.

Structural Core vs. Domain Accent

Skeletal core. Relationships between relationships can themselves be related under compatible composition. Domain-bound accent. A 3-category specifically requires objects, 1-, 2-, and 3-morphisms, typed boundaries, and strict or coherent composition laws. Remove these and the residual is only a layered-relations analogy. Why not a prime. The exact formalism and its coherence tests live in category theory; a generic layered-pattern intuition travels more broadly but is not the named 3-category.

  • Current DAG placement. The 3-morphism layer and its coherence conditions distinguish this entry from a 2-category. The frozen placement leaves it unparented pending a reviewed higher-category genus that covers those conditions.

  • Related, not asserted parent. A 2-category supplies the lower layers, but the third morphism level and its coherence are the differentia of this entry.

Neighborhood in Abstraction Space

3-Category sits in a crowded region of the domain-specific corpus (39th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Category Theory & Higher Structures (18 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With

  • 2-category. Tell: Has objects, 1-morphisms, and 2-morphisms; ask whether a third typed morphism level is present.
  • Gray category. Tell: Is a semistrict kind or presentation of 3-category; ask whether the Gray tensor and its weaker interchange are intended.
  • Tricategory. Tell: Names a weak 3-category formulation; ask what equalities are replaced by coherent data.
  • Three-level hierarchy. Tell: May have nested classes but no higher morphism composition; ask for typed cells and coherence laws.

References

  • John Baez, This Week's Finds in Mathematical Physics, Week 49, description of 2Cat as a 3-category with transformations and modifications.
  • John Baez, The Dimensional Ladder, part 3, strict, semistrict, and weak 3-categories, including the fundamental 3-groupoid Π₃(S²).
  • Frozen Wikipedia discovery revision: https://en.wikipedia.org/wiki/3-category (revision 1371090047).
  • Preserved source candidate: http://www.ams.org/memo/0558
  • Preserved source candidate: http://www.tac.mta.ca/tac/volumes/10/1/10-01abs.html
  • Preserved source candidate: https://math.ucr.edu/home/baez/trimble/tetracategories.html
  • Preserved source candidate: https://ncatlab.org/nlab/show/Gray-category
  • Preserved source candidate: https://ncatlab.org/nlab/show/strict+3-category
  • Preserved source candidate: https://ediss.sub.uni-hamburg.de/handle/ediss/6382
  • Preserved source candidate: http://pantodon.jp/index.rb?body=Gray-tensor_product

The frozen Wikipedia revision is discovery provenance. The retained source set was reviewed for identity, formal or operational relation, and scope. The encyclopedia's structural synthesis is bounded to those claims; a thin authority surface is recorded as a nonblocking source-strengthening repair rather than concealed.