Geometry of Abstract Structure

Often described as “the mathematics of mathematics”, category theory is simply the study of connections, structures, and processes. Traditional set theory looks inside an object to see what elements it contains. Category theory stands back and looks at the outside—how objects relate to each other through arrows.

Formally, a category $\mathcal{C}$ is defined as a tuple $(\text{Ob}(\mathcal{C}), \text{Hom}(\mathcal{C}), \circ, \text{id})$ consisting of:

• Objects: A collection of objects $\text{Ob}(\mathcal{C})$.
• Morphisms: For each pair of objects $A, B \in \text{Ob}(\mathcal{C})$, a set of morphisms $\text{Hom}_\mathcal{C}(A, B)$ (or arrows $f : A \to B$).
• Composition: A binary operation $\circ : \text{Hom}(B, C) \times \text{Hom}(A, B) \to \text{Hom}(A, C)$ mapping $f : A \to B$ and $g : B \to C$ to $g \circ f : A \to C$.

subject to three foundational axioms:

1. Closure under Composition: $$\forall f \in \text{Hom}(A, B), g \in \text{Hom}(B, C) \implies (g \circ f) \in \text{Hom}(A, C)$$

2. Identity Laws: For every object $A \in \text{Ob}(\mathcal{C})$, there exists a unique identity morphism $\text{id}_A \in \text{Hom}(A, A)$ such that for any $f : A \to B$: $$f \circ \text{id}_A = f = \text{id}_B \circ f$$

3. Associativity of Composition: For all $f : A \to B$, $g : B \to C$, and $h : C \to D$: $$(h \circ g) \circ f = h \circ (g \circ f)$$

We gain an intuitive grasp of these abstract rules by replacing mathematical symbols with geometric shapes ($\bigcirc, \square, \triangle$) as our objects and shape transformations as our arrows.

In this shape universe, we ignore the internal pixels or vertices of an object and focus entirely on directed, structure-preserving transformations: a circle transforming into a square ($f : \bigcirc \to \square$), or a square transforming into a triangle ($g : \square \to \triangle$).

For these shapes and arrows to form a valid category, the categorical axioms must hold strictly. When we have a transformation $f : \bigcirc \to \square$ and another transformation $g : \square \to \triangle$, there must exist a composite arrow $g \circ f : \bigcirc \to \triangle$ representing the direct route from circle to triangle. Furthermore, every shape $A$ possesses an identity morphism $\text{id}_A$—a “do nothing” transformation—leaving the shape invariant. Whether we group successive transformations as $(h \circ g) \circ f$ or $h \circ (g \circ f)$, associativity guarantees that the final resulting shape is identical.

By constraining or expanding the structure of our shape universe, specific mathematical concepts emerge naturally. Consider a category containing only one object—a single, universal shape $S$—accompanied by many arrows. This single-object category is a Monoid. The arrows represent self-transformations ($S \to S$) such as rotating ninety degrees, scaling by two, or inverting colors, while composition chains these actions and the identity arrow acts as do_nothing. In software engineering, familiar abstractions like string concatenation or integer addition with zero mirror this exact single-object categorical structure.

Conversely, if we consider a category where between any two shapes $A$ and $B$ there is at most one arrow, we obtain a Preorder or partial order (such as $\le$). An arrow exists from shape $A$ to shape $B$ if and only if $\text{Size}(A) \le \text{Size}(B)$. Composition becomes transitivity: if $\text{Small} \le \text{Medium}$ and $\text{Medium} \le \text{Large}$, then an arrow guarantees $\text{Small} \le \text{Large}$. Identity satisfies reflexivity, as every shape is at least as large as itself.

Moving beyond single categories, a Functor $F : \mathcal{C} \to \mathcal{D}$ acts as a structure-preserving mapping from one category of shapes to another. It translates every object $A$ in $\mathcal{C}$ into an object $F(A)$ in $\mathcal{D}$, and maps every arrow $f : A \to B$ in $\mathcal{C}$ into an arrow $F(f) : F(A) \to F(B)$ in $\mathcal{D}$, preserving identity ($F(\text{id}_A) = \text{id}_{F(A)}$) and composition ($F(g \circ f) = F(g) \circ F(f)$). Imagine Category $\mathcal{C}$ containing simple plain shapes and Category $\mathcal{D}$ containing framed container shapes; a functor $F$ maps plain shapes to container shapes without tearing down the underlying relationships between them. In programming, container types like List<T>, Maybe<T>, or Result<T> are functors whose map function lifts shape transformations $f : A \to B$ into $F(f) : F(A) \to F(B)$.

When we need to map between two different functors operating over the same categories, we arrive at a Natural Transformation $\alpha : F \Rightarrow G$. A natural transformation transforms shape containers without disturbing the underlying shape transformations. For every shape $A$, $\alpha_A$ supplies an arrow $F(A) \to G(A)$ satisfying the commutative diagram property: $G(f) \circ \alpha_A = \alpha_B \circ F(f)$. If $F$ produces filled shapes ($\text{FilledShape}$) and $G$ produces outlined shapes ($\text{OutlinedShape}$), $\alpha$ converts filled shapes into outlined shapes. Transforming a filled circle into a filled square and then outlining it yields the exact same outlined square as outlining the filled circle first and then transforming it to an outlined square. Polymorphic functions like headOption : List<A> -> Option<A> or flatten : List<List<A>> -> List<A> are real-world natural transformations.

Category theory also reconstructs familiar algebraic data types purely through universal shape interactions. A Product ($A \times B$) pairs two objects into a single compound shape $(\text{Shape}_A, \text{Shape}_B)$ with projection arrows (fst, snd) to extract each component, corresponding to records or structs.

A Coproduct ($A + B$) defines a choice between shapes (Either ShapeA ShapeB) with injection arrows ($\text{inl}, \text{inr}$), mirroring sum types, enums, or tagged unions.

Viewed through geometric shapes and transformations, category theory demonstrates that software architecture is not an arbitrary collection of patterns—it is geometry in abstract space.