Category Theory

Primitives: Object and Morphism

Category theory reduces all of mathematics to two irreducible primitives:

• Object — a node, a “thing” with no internal structure visible to the theory
• Morphism (arrow) — a directed relationship between objects: $f: A \to B$

The fundamental insight is that structure lives entirely in the arrows. What an object is does not matter; what matters is how morphisms compose through it.

The Hierarchy of Forms

Category theory is itself organized in a three-level hierarchy of form:

This is “the equation of morphism” — the form of a relationship between forms.

Notable categorical forms:

• Morphism $f: A \to B$ — the irreducible directed relationship
• Functor — a structure-preserving map between categories
• Adjunction — the canonical form of a “best approximation” between two worlds

Axioms and Universal Constructions

A category $\mathcal{C}$ is defined by exactly two laws:

1. Identity — for every object $A$, a morphism $\mathrm{id}_A: A \to A$ such that $f \circ \mathrm{id}_A = f$ and $\mathrm{id}_A \circ g = g$.
2. Associativity — $(h \circ g) \circ f = h \circ (g \circ f)$ whenever compositions are defined.

Functors

A functor $F: \mathcal{C} \to \mathcal{D}$ must satisfy:

$F(\mathrm{id}_A) = \mathrm{id}_{F(A)}, \qquad F(g \circ f) = F(g) \circ F(f)$

The Yoneda Lemma

The deepest theorem derivable from the axioms alone:

$\mathrm{Nat}(\mathrm{Hom}(A, -),\, F) \cong F(A)$

An object is completely determined by the morphisms into and out of it — identity is relational, not intrinsic.

Universal Properties

All standard constructions — products, coproducts, limits, colimits — are defined purely by their morphism-relationships, not by what they “contain.” The adjunction $F \dashv G$ captures the universal notion of optimal translation between two categories.

Procedures

Category theory’s algorithmic content is largely in verification and construction — checking that maps respect structure, and building universal objects.

Verifying a Functor

Given a proposed map $F: \mathcal{C} \to \mathcal{D}$:

1. Check each object $A \in \mathcal{C}$ maps to an object $F(A) \in \mathcal{D}$.
2. Check each morphism $f: A \to B$ maps to $F(f): F(A) \to F(B)$.
3. Verify $F(\mathrm{id}_A) = \mathrm{id}_{F(A)}$ — identity preserved.
4. Verify $F(g \circ f) = F(g) \circ F(f)$ — composition preserved.

Finding Limits and Colimits

1. Fix a diagram — a functor from a small index category.
2. Define cones (commuting triangles converging to a tip) over the diagram.
3. The limit is the terminal cone — the most economical object from which the whole diagram factors.

Adjoint Functors

The adjunction $F \dashv G$ is established by finding a natural bijection:

$\mathrm{Hom}_{\mathcal{D}}(F(A), B) \cong \mathrm{Hom}_{\mathcal{C}}(A, G(B))$

Adjunctions are ubiquitous: free/forgetful, product/diagonal, suspension/loop space.

Category as a System

A category is a dynamical system of structure: a collection of objects (the stock) connected by morphisms (the flows), governed by two laws that enforce a stable equilibrium.

• Objects (stocks) — stable nodes; their “identity” is entirely their position in the web of arrows.
• Morphisms (flows) — directed transformations between objects; the system’s transactions.
• Composition (the process) — the mechanism by which flows chain: $g \circ f$ routes through an intermediate object.
• Identity morphisms (the balancing loop) — every object has a self-loop that leaves any incoming or outgoing flow unchanged.

Functors are system-level transformations: maps that transport an entire category to another while preserving the connectivity structure. A functor that forgets structure (the forgetful functor) and one that builds it freely (the free functor) form an adjoint pair — a feedback relationship between two systems.

Natural transformations are the equilibria between functors: a systematic family of morphisms that interpolate between two ways of transporting one category into another, without disturbing the relational fabric.