Narrative network¶
A network representation of a narrative in which selected characters, events, places, concepts, or states become nodes and textually evidenced interactions, sequence, co-occurrence, causation, or discourse relations become edges under an explicit extraction and perspective model.
Core Idea¶
A narrative network is a graph representation of a bounded narrative in which selected characters, events, places, concepts, or states are nodes and textually evidenced interactions, sequence, co-occurrence, causation, transitions, or discourse relations are edges under a declared perspective and extraction model. Construction is interpretation. Construction is interpretation.
Scope of Application¶
Narrative networks are used in digital humanities, narratology, literature, film and television studies, oral history, historical chronicles, journalism, game narratives, folklore, and computational social science. Use it with corpus/version/medium and narrator, node ontology and alias/coreference, edge semantics/direction/weight/sign/layers, narrative versus discourse order and time slicing, manual/NLP extraction and validation, annotation agreement and uncertainty, construction and metrics/null models, sensitivity to thresholds/granularity, visualization, quoted evidence and return to close reading. Distinguish the graph from the represented world, cast lists, plot summaries, general knowledge graphs, and claims that centrality equals narrative importance.
- Character relations. Maps interaction/co-presence.
- Event chains. Represents sequence and causality.
- Perspective. Compares narrators or adaptations.
- Corpus analysis. Finds recurring structures.
- Visualization. Makes patterns inspectable.
Clarity¶
Report corpus/version/medium and scope, narrator/perspective, node types and identity/coreference, edge definition/window/direction/weight/sign/layers, narrative versus discourse order, temporal slicing, manual/NLP extraction and code/model/version, annotation agreement/precision/recall, uncertain or absent relations, graph construction, metrics/null models, sensitivity to thresholds/granularity, visualization choices, passages supporting findings, and limits on real-world/causal inference. The closest near miss sets the boundary: A character network is nearest but narrower; narrative networks can instead center events, causal/temporal relations, places, or mixed entities.
Manages Complexity¶
The graph compresses long sequential narratives into relational structure, enabling pattern detection while discarding language, focalization, chronology, ambiguity, and meaning unless deliberately encoded. The central relational overview–textual nuance tradeoff is this: Graphs reveal large patterns while stripping voice and ambiguity. A second reproducible coding–interpretive plurality tension matters because Explicit rules aid replication while legitimate readings differ.
Abstract Reasoning¶
Use three linked moves: define narrative question, corpus, version, and perspective; design node and edge ontology tied to evidence; extract and validate relations with temporal/uncertainty handling. As a collapse test, the representation fails when extraction choices are hidden, edges lack evidence, temporal order is erased against the question, or graph metrics are interpreted without returning to the narrative. A fourth check is to analyze structure against null and alternative encodings. A final check is to return every interpretation to the text and representational losses.
Knowledge Transfer¶
Network techniques transfer across novels, films, and oral accounts only after remapping evidence units, identity resolution, relation semantics, sequence, and medium-specific perspective. No canonical parent prime is currently asserted; broader structural comparisons remain related-prime analogies until separately adjudicated in the DAG. Formal representation, but exact parentage needs live-signature review. Evidence domain and meaning frame.
Relationships to Other Abstractions¶
Current abstraction Narrative network Domain-specific
Parents (1) — more general patterns this builds on
-
Narrative network is a kind of, conditional Formal Model Domain-specific
Supported when the network is an explicit formal representation of narrative entities and relations.
Condition / exception Supported when the network is an explicit formal representation of narrative entities and relations.
Hierarchy path (1) — routes to 1 parentless root
- Narrative network → Formal Model → Representation → Abstraction
Neighborhood in Abstraction Space¶
Narrative network sits in a crowded region of the domain-specific corpus (38th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.
Family — Narrative Structure & Storytelling Devices (24 abstractions)
Nearest neighbors
- Narrativity — 0.89
- Dialogical analysis — 0.89
- Episodic-like memory — 0.88
- Writing system — 0.87
- Outside Context Problem — 0.87
Computed from structural-signature embeddings · 2026-10-08