Cubical Set¶
Represent a combinatorial or homotopical object as a contravariant set-valued functor on a declared cube category, with functorial face, degeneracy, and optional richer cube operations.
Core Idea¶
Fix a cube category Box whose objects are formal n-dimensional cubes and whose morphisms describe the allowed ways of inserting endpoints, deleting or identifying dimensions, permuting coordinates, or adding connection operations. A cubical set is a presheaf
The set X_n=X(Box^n) is the collection of abstract n-cubes of X. A cube map alpha:Boxm→Boxn induces a structure map X(alpha):X_n→X_m. The reversal of arrows is load-bearing: probing an n-cube along a lower-dimensional cube map produces its lower-dimensional face or specialization.
Scope of Application¶
Algebraic homotopy theory. Cube categories that are test categories let their presheaves present classical homotopy types after an appropriate localization. Cubical boundaries, horns or boxes, fillers, and geometric realization supply combinatorial analogues of paths and homotopies.
Homotopy type theory. Cartesian and De Morgan cubical sets provide interval objects and path semantics. Awodey develops a cartesian cubical model of homotopy type theory, while Cohen, Coquand, Huber, and Mörtberg use a richer cubical model to give constructive computational meaning to univalence and to handle function extensionality and selected higher inductive types.
Clarity¶
The shortest recognition test asks three questions: What is Box? What is X(Box^n) in each dimension? What maps does X assign to each morphism of Box? If the answer gives only cubes but no contravariant structure maps, it is not yet a cubical set.
Manages Complexity¶
Higher-dimensional attachment data can require many separate compatibility equations. The functor packages them into one law: identity maps act as identities and composite cube maps act by composite restrictions in reverse order. The cubical identities are no longer ad hoc bookkeeping; they are instances of functoriality.
Abstract Reasoning¶
Use this protocol:
- Declare Box and list its generating morphisms and equations.
- Specify X_n for each needed dimension.
- Give the induced face, degeneracy, and optional richer operations.
- Verify their cubical identities, or derive them from functoriality.
- Define morphisms as natural transformations commuting with every operation.
- Add fillers, labels, basepoints, realizations, or model structures only as separately declared layers.
Knowledge Transfer¶
Literal transfer occurs when topology, concurrency, type theory, or higher category theory retains the same presheaf grammar. The meanings of a 2-cell may differ, but Box, contravariance, dimension fibers, and functorial boundary maps remain recognizable.
Transfer from cubical to simplicial methods needs an explicit adjunction, triangulation, nerve, or comparison functor. Calling squares triangles does not preserve their face combinatorics. Transfer from a cubical set to a topological space likewise passes through geometric realization and may be considered only up to the declared equivalence.
Relationships to Other Abstractions¶
Current abstraction Cubical Set Domain-specific
Parents (1) — more general patterns this builds on
-
Cubical Set is a kind of Functor Domain-specific
Functor. A cubical set is literally a contravariant functor from a cube category to Set.
Hierarchy paths (4) — routes to 4 parentless roots
- Cubical Set → Functor → Category → Associativity → Invariance
- Cubical Set → Functor → Function (Mapping)
- Cubical Set → Functor → Category → Closure
- Cubical Set → Functor → Category → Associativity → Symmetry
Neighborhood in Abstraction Space¶
Cubical Set sits in a sparse region of the domain-specific corpus (69th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Unclustered & Miscellaneous (1565 abstractions)
Nearest neighbors
- Simplicial Presheaf — 0.85
- Aztec Diamond — 0.85
- McKay Graph — 0.84
- Synthetic differential geometry — 0.84
- Algebraic stack — 0.84
Computed from structural-signature embeddings · 2026-09-08