🔢 mathematics

Computable Structuralism: A Categorical Rewrite Calculus of Mythic Variants

This paper proposes "Computable Structuralism," a formal categorical framework that models mythic and narrative variants as typed rewrite programs on a dual-register state, thereby translating Lévi-Straussian structural analysis into a computable system capable of testing coherence, diagnosing structural failures, and quantitatively bridging anthropology with cultural analytics across diverse storytelling traditions.

Original authors: Juan J. Segura

Published 2026-01-22

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Juan J. Segura

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a giant library of stories: fairy tales, religious myths, superhero movies, and long TV franchises. Usually, when scholars compare these stories, they look for similarities: "Oh, both Cinderella and Spider-Man have a sad childhood!" or "Both have a magical helper!"

But this paper argues that looking for surface similarities is like judging two cars only by their paint color. It misses how the engines actually work. The authors, led by Juan J. Segura, propose a new way to compare stories that treats them like mathematical puzzles or computer code, but one that still keeps the human meaning of the story intact.

Here is the breakdown of their idea using simple analogies:

1. The Two-Register Machine (The "Social" vs. The "Rules")

The authors suggest that every story runs on a machine with two distinct "registers" (like two tracks on a cassette tape or two lanes on a highway):

Register X (The Everyday Lane): This is the messy, physical world. It has the characters, the fights, the tools, the money, and the physical actions. Example: A wolf eating a grandmother, or a superhero punching a villain.
Register Y (The Rulebook Lane): This is the invisible world of laws, taboos, prophecies, and contracts. It's the "why" that makes the action matter. Example: A mother's warning not to talk to strangers, a law requiring superhero registration, or a divine curse.

The Insight: In most stories, the physical action (X) only becomes a "story" when it bumps up against the rules (Y). If a wolf eats a grandma, that's just a tragedy. But if he does it because he broke a taboo, or if the grandma survives because of a prophecy, the story gains its structure. The authors claim you can't understand the story unless you track both lanes simultaneously.

2. The "Transformation" Test (The Magic Trick)

Structuralists (like the famous Claude Lévi-Strauss) believe that stories are related not because they look alike, but because one can be transformed into another through specific rules (swapping roles, inverting values).

The authors turn this into a math test. They imagine two ways to update a story:

Update A (The Direct Way): You change the story, but you keep the roles the same. (The hero is still the hero; the law is still the law).
Update B (The Canonical Way): You perform a "magic trick" where you swap things around. Maybe the thing that was a "rule" becomes the "villain," or the "hero" becomes the "victim."

The "Coherence Check":
The paper asks: If I swap the roles (Update B), does the story still make sense mathematically?
They use a concept called a "Natural Transformation" (a fancy math term) which acts like a quality control stamp.

Pass: If the story still holds together after the swap, the transformation is "legal." The two stories are structurally related.
Fail: If the story falls apart (e.g., the "law" suddenly acts like a "tool" and breaks the logic), the transformation is "illegal." This failure isn't a mistake; it's a diagnosis. It tells the researcher, "You tried to compare these two things, but they don't actually fit together under these rules."

3. The "Key" (The Story's Fingerprint)

Stories happen in a specific order. If you swap the order of events (e.g., the hero gets the sword after the battle instead of before), the story changes completely.

To capture this, the authors assign every story a "Key" (labeled A through E). Think of this like a fingerprint or a barcode for the story's sequence.

It doesn't tell you what happens, but how the events are arranged.
This allows them to quickly compare 80 different stories to see if they share the same "structural rhythm," even if one is about a frog prince and the other is about a space wizard.

4. The Experiment: 80 Stories

The team tested this system on 80 stories:

20 Folktales (e.g., Cinderella)
20 Religious Myths (e.g., Adam and Eve)
20 Superhero stories (e.g., Batman)
20 Franchises (e.g., Star Wars)

What they found:

The "Rulebook" is real: In 74% of the stories, the "Rulebook Lane" (Y) was explicitly named. Whether it was a "taboo," a "law," a "prophecy," or a "contract," the stories relied on these invisible rules to drive the plot.
The System Works: They could successfully map these different genres onto the same two-track system. A "fairy godmother" in a fairy tale and a "registration law" in a superhero movie both function as the "mediator" or "constraint" in the same mathematical slot.
It Catches Errors: When they tried to force a story into the wrong slot (like calling a "law" a "tool"), the math broke. This proved the system is strict enough to catch bad comparisons.

The Bottom Line

This paper doesn't try to replace human storytelling or tell us which stories are "better." Instead, it builds a translation tool.

It turns the vague idea of "structural similarity" into a testable, reproducible math problem. It allows researchers to say: "These two stories are related not just because they look alike, but because they follow the same underlying logic of how physical actions interact with social rules."

It's like taking a messy pile of different colored LEGO bricks and realizing they all snap together using the exact same hidden mechanism. The authors have just written down the manual for that mechanism.

Technical Summary: Computable Structuralism

Problem Statement
Comparative analysis of myths and narratives faces a methodological tension between two impulses: the insistence that meaning is local and context-dependent (social, ritual, historical conditions) and the structuralist view that narratives are comparable via systematic transformations (substitutions, inversions, role exchanges). While structural anthropology (e.g., Lévi-Strauss) established that myths function as systems of transformation, contemporary cultural analytics requires methods that are explicit enough to be reproduced, contested, and scaled. Current approaches often rely on resemblance-based metrics (e.g., motif overlap, embedding similarity) which quantify proximity but fail to express "lawful transport" under explicit operator constraints. The paper identifies a gap in operationalizing Lévi-Straussian transformation: there is no formal framework that renders structural transformation as mathematics while remaining readable as narrative analysis and capable of falsifiable constraints.

Methodology
The authors propose a formal framework modeling narratives as typed rewrite programs acting on a coupled two-register state $(X, Y)$ .

Two-Register Configuration ( $Z = (X, Y)$ ):
- Register X (Everyday/Social): Tracks agents, institutions, survival constraints, political conflict, and material conditions.
- Register Y (Symbolic/Legitimation): Tracks law, taboo, prophecy, covenant, ritual, moral constraints, and cosmological order.
- The model posits that transformations in $X$ are consequential only when recognized/authorized in $Y$ , and disruptions in $Y$ are real only when they reorganize $X$ .
Encoding Schema:
Each narrative is encoded into a four-slot functional schema $(a, b, x, y)$ :
- $a$ : Focal agent (carrier of transformation).
- $b$ : Opposing/perturbing force (antagonist, chaos, institution).
- $x$ : Mediator enabling transformation (tool, pact, helper).
- $y$ : Constraining/legitimating order (law, taboo, rule).
Update Policies and Natural Transformation:
The framework defines two systematic update policies (endofunctors) on configurations:
- Direct Policy ( $U$ ): Updates preserve register roles (mediator remains mediator; constraint remains constraint).
- Canonical Policy ( $V$ ): Implements a structured swap + inversion move (e.g., a constraint becomes an object of conflict; a conflict is reframed as a constraint).
- Coherence Condition: A canonical transformation is defined as the existence of a natural transformation $\eta: U \Rightarrow V$ . For a proposed transport $h$ between variants, the diagram must commute: $V(h) \circ \eta_Z = \eta_{Z'} \circ U(h)$ .
- Falsifiability: If the naturality equation fails, it diagnoses a specific error: a mis-specified opposition, an illegal operator transport, or an incorrect register assignment. This turns structural analysis into a "corpus-facing type check."
Context Sensitivity:
Context is internalized via structured operator choice. Admissible rewrite operators and transport rules are specific to narrative families (e.g., folktales vs. franchises). This prevents the "anything maps to anything" collapse by making specific operator transports illegal if they violate the declared context.
Order-Signature Invariant (Key):
To account for sequential order effects (e.g., trickster crossings, time loops) where "swap then invert" $\neq$ "invert then swap," each narrative is assigned a Key (archetypes A–E). These keys summarize order effects using braid archetypes and computable invariants, serving as a compact structural fingerprint.

Corpus and Experimental Design
The study utilizes a stratified corpus of 80 narratives divided equally into four media families:

20 Folktales (e.g., Cinderella, Little Red Riding Hood)
20 Religious Myths (e.g., Adam and Eve, Noah)
20 Superhero narratives (e.g., Batman, Spider-Man)
20 Franchises (e.g., Star Wars, Harry Potter)

The design is balanced to prevent genre dominance and ensure the method travels across media. A two-pass coding audit on a subset ( $n=20$ ) demonstrated high inter-rater reliability ( $\kappa = 0.95$ for $x$ and $y$ ), with disagreements concentrated at the $x/y$ boundary (mediator vs. constraint), which the model explicitly targets.

Key Results

Operational Coherence: The framework successfully models narratives as typed rewrites. The naturality condition serves as a diagnostic tool; in a worked example (Little Red Riding Hood), a plausible mis-typing (swapping the hunter's role from mediator $x$ to constraint $y$ ) caused a naturality failure, localizing the structural incoherence.
Constraint Prevalence: In 59/80 (74%) of the narratives, the $y$ register explicitly names a normative constraint (law, taboo, contract, prophecy). This supports the validity of the two-register abstraction.
Cross-Media Stability: The functional roles $(a, b, x, y)$ remain intelligible across media. While realizations differ (e.g., "Royal Court" in folktales vs. "Registration Law" in superheroes), the structural function of $y$ as a constraint system persists.
Key Sensitivity: A micro-test ( $n=10$ ) showed that Keys are stable under conservative normalization (8/10 cases) but change under "worst-case" stress tests (swapping $x$ and $y$ ), confirming that the invariant is sensitive to the specific structural logic of the narrative.
Constraint Typology: Analysis of the $y$ register reveals distinct "constraint-type signatures" across corpora (e.g., Superheroes/Franchises lean toward legal/institutional constraints; Religious Myths toward divine/cosmic constraints).

Significance and Claims
The paper claims to provide a testable bridge between structural anthropology and cultural analytics. Its primary contributions are:

Formalization: It renders Lévi-Straussian canonical transformation into a testable coherence condition (natural transformation) rather than a loose template.
Operationalization: It shifts comparison from "resemblance" to "lawful transformation," making structural claims explicit, transportable, and refutable.
Diagnostic Power: Failures in the model are not defects but diagnoses that locate exactly where an analysis has collapsed (e.g., confusing a mediator with a constraint).
Interpretability: The framework maintains interpretability for close reading (via the $a,b,x,y$ schema and Keys) while enabling corpus-level computation.

The authors explicitly state that this is a demonstration set for operationalization, not a claim about global narrative prevalence. The framework is designed to make transformation claims explicit and contestable, separating what is encoded, what is inferred, and what is tested. Future work is identified as extending the encoding from single-shot summaries to episode-level graphs and publishing a replication package.