Planning Beyond Text: Graph-based Reasoning for Complex Narrative Generation
Hanwen Gu, Chao Guo, Junle Wang, Wenda Xie, Yisheng Lv · 2026
Read Planning Beyond Text: Graph-based Reasoning for Complex Narrative Generation by Hanwen Gu, Chao Guo, Junle Wang, Wenda Xie, Yisheng Lv (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-21253.txt.
Introduction
Hanwen Gu and four colleagues plan stories on graphs instead of raw text, aiming to fix coherence breaks and flat characters in fluent but wandering tales. Their PLOTTER framework runs an evaluate-plan-revise cycle over event and character graphs, fixing topological flaws under strict logical constraints before full text is made. Optimizing causality and the story skeleton first pays off across diverse cases, where PLOTTER clearly beats rival baselines. The work argues that structural planning is key to long-context reasoning in complex story generation.
Catalog record
| CREATOR | Hanwen Gu, Chao Guo, Junle Wang, Wenda Xie, Yisheng Lv |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84039 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.21253
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Planning Beyond Text: Graph-based Reasoning for Complex Narrative Generation Hanwen Gu, Chao Guo, Junle Wang, Wenda Xie, Yisheng Lv Abstract While LLMs demonstrate remarkable fluency in narrative generation, existing methods struggle to maintain global narrative coherence, contextual logical consistency, and smooth character development, often producing monotonous scripts with structural fractures. To this end, we introduce PLOTTER, a framework that performs narrative planning on structural graph representations instead of the direct sequential text representations used in existing work. Specifically, PLOTTER executes the Evaluate-Plan-Revise cycle on the event graph and character graph. By diagnosing and repairing issues of the graph topology under rigorous logical constraints, the model optimizes the causality and narrative skeleton before complete context generation. Experiments demonstrate that PLOTTER significantly outperforms representative baselines across diverse narrative scenarios. These findings verify that planning narratives on structural graph representations—rather than directly on text—is crucial to enhance the long context reasoning of LLMs in complex narrative generation. 1 1 footnotetext: These authors contributed equally. 2 2 footnotetext: Corresponding authors. 1 Introduction Although LLMs demonstrate remarkable proficiency in storytelling, long-form complex narrative generation remains challenging. It requires not only linguistic fluency but also the construction of a logically grounded narrative structure ( 16 ; 3 ; 8 ) . To realize coordinated planning for long-form narratives, recent work has introduced iterative hierarchical planning ( 33 ; 32 ; 29 ) or role-playing strategies ( 3 ; 10 ) via LLM agents. However, these methods still struggle to maintain global narrative coherence, contextual logical consistency, and smooth character development, often producing monotonous scripts with structural fractures such as insufficient character motivation and a lack of necessary twists ( 14 ; 20 ; 26 ) . We argue the primary reason is that narrative planning directly on text representations is inefficient. Without explicit modeling of plot dependencies, such systems cannot effectively reason about the underlying cause-and-effect web or the evolving relationships among characters and events, ultimately limiting their ability to produce rigorous narrative structures ( 23 ; 34 ) . To address this issue, we introduce PLOTTER , which stands for PL anning Bey O nd T ext: Graph-based Reasoning for Complex Narra T ive Gen ER ation. Unlike prior methods that plan on text directly, we plan narratives on graph structures, transforming script generation from a sequence planning problem into a dynamic graph generation and refinement problem. By applying the Evaluate-Plan-Revise cycle onto the narrative topology, PLOTTER enables the model to diagnose and repair structural inconsistencies at the causal level before textual realization. We represent narrative dynamics through two interacting structures: an Event Graph , capturing the causal structure and skeleton of the plot, and a Character Graph , modeling inter-character relationships. This design draws inspiration from classic narratological theories regarding action logic ( 2 ) and character networks ( 17 ) . Consequently, the challenge of script writing is decomposed into atomic graph-editing operations—adding, deleting, or re-linking nodes and edges—to resolve causal gaps and character inconsistencies. To enforce structural integrity, we deploy a graph-grounded refinement cycle with a multi-agent critique module to audit symbolic structures rather than raw text. These agents identify weaknesses—such as disconnected causal paths—and formulate revision strategies. A Constrained Graph Editor then executes atomic operations to repair the narratives. We evaluate our framework across diverse narrative scenarios using multiple mainstream LLM backbones. Results show that our graph-based narrative reasoning approach significantly outperforms existing methods. Our main contributions are as follows: • We propose PLOTTER , the first framework to perform narrative planning on structural graph representations rather than the direct sequential text, enhancing long context causal reasoning in complex narrative generation. • We construct an LLM agent system for the iterative structural refinement of event and character graphs. By coordinating specialized agents to execute the Evaluate-Plan-Revise cycle on the graph topology, we achieve precise structural diagnosis and repairs that are inaccessible to text-based narrative planners. • Experiments show that PLOTTER significantly outperforms strong existing methods, verifying that elevating narrative planning from direct text representations to explicit graph representations is crucial for enhancing long-range reasoning in complex narratives. 2 Related Work The landscape of automated narrative generation has evolved from sequence-to-sequence generation to LLM-based planning. We categorize existing literature into two streams: outline-based and role-play based narrative planning. 2.1 Outline Based Narrative Planning To improve the coherence in long narrative contexts, most approaches adopt a “Plan-and-Write” paradigm and decompose the generation task into hierarchical stages guided by the planned outline. Frameworks like Re3 ( 33 ) utilize recursive prompting for context maintenance, while Detailed Outline Control (DOC) ( 32 ) and DOME ( 25 ) impose strict constraints to align generation with high-level outlines. Recent advancements such as CONCOCT ( 27 ) further refine this by dynamically evaluating outline pacing. However, because these methods operate sequentially, logical inconsistencies in early steps often propagate downstream, causing cascading errors ( 33 ) . Furthermore, treating the outline as a rigid constraint restricts the flexibility required for complex revisions while avoiding extensive rewriting of preceding contexts ( 32 ) . 2.2 Role-Play Based Narrative Planning Recent research has shifted focus from static planning to dynamic multi-agent simulation, where frameworks like HoLLMwood ( 3 ) , Agents’ Room ( 11 ) , IBSEN ( 10 ) , and StoryWriter ( 28 ) assign specialized personas (e.g., Director, Actor) to distinct LLM instances. While these role-playing paradigms excel at stylistic diversity and dialogue richness, their coordination relies predominantly on unstructured natural language. Research indicates that such purely textual critique loops are susceptible to ambiguity and semantic drift over long contexts ( 1 ) . Consequently, without a shared symbolic state to ground collaboration, instructions from high-level planners are often misinterpreted by downstream agents, leading to hallucinations that contradict established narrative facts. 2.3 Graph-Based LLM Reasoning Recent research integrates graph structures to enhance multi-hop reasoning for LLMs. For instance, Reasoning with Graphs (RwG) ( 9 ) structures implicit context into explicit graphs, Think-on-Graph (ToG) ( 22 ) utilizes knowledge graphs for deep inference, and TG-LLM ( 30 ) employs temporal graphs to bolster chronological understanding. In the specific domain of LLM story generation, recent work, such as R 2 ( 12 ) , extracts graphs from a complete source text to build static external memories for generation reference. These works utilize graph structures as fixed contextual references for generation, while PLOTTER performs dynamic narrative planning via graph-structured representations of events and characters. Unlike prior work that performs story planning directly on text representations, we use LLMs to plan narrative structure on graph-based representations. Specifically, we ensure logical and relational consistency through iterative generation and editing of event and character graphs via atomic edit operations, thereby improving plot coherence, engagement, and character development. Figure 1: Overview of the PLOTTER framework. (1) Graph-Based Script Planning initializes the narrative backbone comprising Event ( G_{e} ) and Character ( G_{c} ) graphs. (2) Iterative Graph Refinement employs a Multi-Agent Critic ( \mathcal{C} ) to diagnose structural issues, which are resolved by a Constrained Graph Editor ( \mathcal{R} ) to produce an optimized graph ( G^{*} ). (3) Graph-Grounded Script Synthesis serializes the graph via Event Serialization ( \mathcal{H} ) and Persona Expansion ( \mathcal{E} ), conditioning a State-Aware Generator to synthesize the