NARRATIVE OPERATORS
Abstract
A mathematical abstraction earns confidence not from a single application but from surviving translation into a second, unrelated one. This paper provides that second application for Edwardian Algebra. We interpret the semantic domain S as the space of narrative-canon facts (equivalence classes of statements true in a fictional world), and reasoning operators R as authorial writing moves — transformations of canon by a single scene, revision, or edit. We verify the concretization criteria from PRR-2026-002 §11 (reproducibility, composability, verifiability) hold for this domain independently of whether the author is a person or an AI system, then map each structure identified in PRR-2026-006 onto an attested narrative-craft phenomenon: the trivial subgroup onto inert scene-setting prose; finitary operators onto bounded local revisions, with the alternating/full-symmetric split corresponding to the parity of a plot twist's constituent fact-reversals; unrestricted G onto fully nonlinear restructuring edits; the trajectory trichotomy onto converging revision, revision-thrash (the workshop phenomenon of oscillating between two unsatisfying versions), and genuinely never-ending serialized or branching narrative; and closure operators onto continuity-bible canon-locking. We close by discussing what it means that the same theorems hold, unmodified, in a domain with no reference to computation at all.
Keywords. narrative theory; concretization; applied algebra; canon continuity; revision dynamics; plot-twist parity; Edwardian algebra
1. Introduction
PRR-2026-002 argued that AI reasoning systems instantiate R because they satisfy three criteria: reproducibility (the same input yields consistent output), composability (outputs chain and pool), and verifiability (outputs can be checked against S). Nothing in that argument is specific to computation. A human author revising a scene, or an editorial process moving a manuscript from draft to final, satisfies the same three criteria: a given revision move, applied to a given canon state, produces a consistent result; revisions chain (apply one pass, then another) and pool (two editors' independent notes are reconciled); and the result can be checked against established canon. This paper takes that observation seriously and builds out the second concretization in full, rather than leaving it as a remark.
2. The Narrative Concretization
Let SN be the Lindenbaum–Tarski algebra (PRR-2026-002 Def. 2.3) of a countable language of narrative statements — sentences describing facts about a fictional world — quotiented by logical equivalence. Interpretively: an element [φ] \in SN is an equivalence class of canon claims, e.g. "the King is dead," identified with every logically equivalent phrasing of the same fact.
A narrative operator is any function \rho : S_N \to S_N — a single authorial move (a scene, a revision pass, an editorial cut) mapping the current canon state to a new one. The set of all narrative operators is R_N = S_N^{S_N}, identical in structure to R.
Reproducibility. A given authorial move applied to a given canon state produces a well-defined resulting state — the scene, once written and accepted into canon, says what it says; re-reading it does not change what happened. Composability. Scenes sequence (\circ): chapter two's canon state is chapter one's state passed through chapter two's operator. Parallel threads pool (\oplus): two POV chapters advancing different subplots are read together and their combined canon is the disjunctive pooling of either advancing. Verifiability. A scene's consistency with established canon is exactly what a continuity check verifies — the same operational content as PRR-2026-002's verifiability criterion, applied to manuscripts instead of model outputs.
All three criteria are satisfied without reference to who or what performs the operator — human author, editor, or AI story-generation system (the last of which is this project's own operating domain, and was already covered by the first concretization). The claim of this paper is narrower and more useful than "AI-written fiction is an instance of R" — it is that authorship itself, independent of whether the author is a person or a system, already lives in R.
3. Subgroup Structure Applied: What Kind of Edit Is It?
PRR-2026-006 §2 established that the invertible operators G have exactly four normal subgroups: \{\rho_0\}, the finitary alternating group, the finitary symmetric group, and G itself. Each corresponds to a distinct, recognizable class of editorial move.
| Algebraic structure | Narrative-craft phenomenon |
|---|---|
| \{\rho_0\} — trivial subgroup | Inert transitional/atmospheric prose. Changes zero canon facts; present in every narrative, structurally negligible, and correctly so — this is what description and scene-transition prose should be, algebraically speaking. |
| Finitary operators (Def. 2.1, PRR-2026-006) | A bounded local revision: a single character's arc turn, a subplot resolution — an edit that alters finitely many canon facts and leaves the rest of the story-world exactly as it was. |
| Finitary alternating subgroup specifically | A plot twist realized as an even composition of elementary fact-reversals — the double-reversal mystery, where a second twist restores an apparent status quo while adding a layer the reader did not have (e.g. "the traitor was actually the hero all along, but that reveal was itself staged"). |
| Finitary symmetric group, outside the alternating subgroup | A single clean reveal — an odd elementary reversal with no compensating second flip (the classic single-culprit whodunit resolution). |
| G itself (unrestricted) | A fully nonlinear restructuring edit: reordering the entire told sequence of events (a scrambled chronology, a Rashomon-style multi-vantage retelling) while remaining reversible in principle — the original telling is recoverable from the rearranged one. |
Reading. This is not an analogy chosen to fit; parity of permutations is a real, well-defined property once "the letters being permuted" are identified with canon facts, and the even/odd split of Theorem 2.2 (PRR-2026-006) genuinely bisects bounded plot revisions into two structurally distinct craft categories that working editors already informally distinguish (a "twist" versus a "double-twist") without a name for the distinction.
4. Trajectory Trichotomy Applied: What Kind of Revision Loop Is It?
PRR-2026-006 §3 proved every reasoning trajectory is exactly one of: stabilizing, eventually periodic without stabilizing, or aperiodic. Applied to iterative revision of a single passage:
| Trajectory type | Revision-practice phenomenon |
|---|---|
| Stabilizing | Ordinary healthy revision: each pass changes less than the last, and eventually a pass changes nothing. The scene is done. |
| Eventually periodic (period \ge 2) | Revision-thrash: alternating between two (or more) unsatisfying versions — "darker," "too dark, make it funnier," "too flip, make it darker again." Not slow convergence; a genuine cycle. PRR-2026-006's reading applies verbatim: the fix is detecting the cycle (comparing against more than the immediately prior draft), not revising longer. |
| Aperiodic | A narrative that keeps generating genuinely new canon forever without repeating or settling — the defining structural property of an open-ended serialized story or a perpetually branching interactive narrative, where each new reader action is, by construction, a fresh state never previously visited. |
The aperiodic case is not a metaphor for this project's own operating domain — a collaboratively-branching interactive fiction tree, where every submitted action produces a new page distinct from every prior one, is a direct, literal instance of Proposition 3.2(iii) (PRR-2026-006): an infinite family of pairwise-distinct atoms, each reachable from the last by a single operator application, none ever repeating.
5. Closure Operators Applied: The Continuity Bible
PRR-2026-006 §4 established a bijective correspondence between closure operators with finite image and finite meet-closed families of S — sets of facts closed under conjunction and containing \top.
A continuity bible — the industry-standard practice in serialized and collaborative fiction of maintaining a fixed reference document of settled canon facts that all future material must remain consistent with — is exactly such a family. The closure operator \rho_F of PRR-2026-006 Theorem 4.2, which maps any proposed canon state to its consistent completion under the bible F, is exactly the operational content of a continuity check: given a proposed new scene, find the nearest canon-consistent state that does not contradict anything already settled. The order-reversing correspondence (Proposition 4.3, PRR-2026-006) — larger bibles give smaller, more restrictive closure operators — matches the familiar experience that a more detailed continuity bible narrows, rather than widens, what a future writer is free to do.
6. Discussion and Conclusion
No new abstract theorem appears in this paper. Its content is the concretization itself, and what follows from the fact that it works. An abstraction that only ever fits the domain it was invented for is not yet distinguishable from a description of that domain. Edwardian Algebra was built from the study of AI reasoning systems; every structure identified since — the trivial center, the normal subgroup classification, the trajectory trichotomy, closure operators — was proved with no reference to narrative, authorship, or craft. That the same structures land on recognizable, independently-attested phenomena in a domain with no computation in it at all is evidence, though not proof, that the algebra captured something about reasoning-as-such rather than something about AI systems specifically.
A natural further direction, not pursued here: legal reasoning and case-based argument admit a similar reading (precedent as canon, an opinion as an operator, overturning as an irreversible non-inverse), as does game design (a rule-change as an operator on the state of play). Neither is developed in this paper; both are left as candidates for a future applied companion, in the same spirit this paper was written.
References
Edwards, R. (2026). Edwardian Algebra. Perspectivity Research Reports, Vol. 1, No. 2, PRR-2026-002.
Edwards, R. (2026). Reasoning Dynamics Algebra. Perspectivity Research Reports, Vol. 1, No. 6, PRR-2026-006.
Forster, E. M. (1927). Aspects of the Novel. Edward Arnold & Co. (General reference for narrative-structural vocabulary; no specific claim of this paper is attributed to it.)
Lindenbaum, A., & Tarski, A. (1936). Über die Beschränktheit der Ausdrucksmittel deduktiver Theorien. Ergebnisse eines mathematischen Kolloquiums, 7, 15–22.
Peer Review
This is a genuinely disciplined applied paper — it resists the temptation to smuggle in new abstract content under the cover of an application, and says so explicitly in the founding note. The plot-twist parity reading in §3 is the strongest result: it is not merely evocative, it is a precise consequence of Theorem 2.2 (PRR-2026-006) once the permuted set is identified correctly, and the paper is careful to say so rather than leave it as loose analogy. Recommend acceptance.
Section 4's identification of the aperiodic trajectory case with this project's own branching-narrative structure is honestly framed as a literal instance rather than a metaphor, which is the correct and more defensible claim. I would flag the "further direction" paragraph in §6 (legal reasoning, game design) as appropriately left undeveloped — gesturing at unproven extensions without claiming them is the right amount of restraint for a paper that has otherwise been careful throughout. Recommend acceptance.