[Paper Review] Complicating the Social Networks for Better Storytelling: An Empirical Study of Chinese Historical Text and Novel

Chenhan Zhang|arXiv (Cornell University)|Aug 25, 2020

Topic Modeling37 references8 citations

TL;DR

This paper uses NLP and social network analysis to compare the social networks and sentiment portrayals of characters in the historical text 'Records of the Three Kingdoms' and the novel 'Romance of the Three Kingdoms.' It finds that the novel constructs a significantly more complex and dynamic social network than the historical text, with heightened emotional sentiment and narrative richness, demonstrating how literary genre enhances story complexity through network elaboration.

ABSTRACT

Digital humanities is an important subject because it enables developments in history, literature, and films. In this paper, we perform an empirical study of a Chinese historical text, Records of the Three Kingdoms ( extit{Records}), and a historical novel of the same story, Romance of the Three Kingdoms ( extit{Romance}). We employ natural language processing techniques to extract characters and their relationships. Then, we characterize the social networks and sentiments of the main characters in the historical text and the historical novel. We find that the social network in extit{Romance} is more complex and dynamic than that of extit{Records}, and the influence of the main characters differs. These findings shed light on the different styles of storytelling in the two literary genres and how the historical novel complicates the social networks of characters to enrich the literariness of the story.

Motivation & Objective

• To investigate how narrative complexity differs between a historical text and a historical novel of the same story using digital humanities methods.
• To extract and compare the social networks of main characters in 'Records of the Three Kingdoms' and 'Romance of the Three Kingdoms' via NLP techniques.
• To analyze sentiment differences in character portrayals, particularly for Cao Cao and Liu Bei, across the two texts.
• To quantify topological features of the social networks, such as centrality, density, and small-world properties, to assess structural differences.
• To demonstrate how literary genre influences narrative grandeur through enhanced social network complexity and emotional sentiment.

Proposed method

• Employed BERT-based NLP models for named entity recognition to identify characters and their relationships in both texts.
• Constructed social networks from extracted character interactions, treating each interaction as a directed edge between named entities.
• Applied network science metrics including small-worldness, scale-free properties, centrality, and clustering coefficient to characterize network topology.
• Used SentiWordNet to quantitatively score sentiment polarity of evaluative words associated with key characters like Cao Cao and Liu Bei.
• Generated word clouds and sentiment score comparisons to visualize and analyze subjective portrayals in the two texts.
• Built a five-slice dynamic network model to represent narrative progression, with potential for finer-grained temporal subdivision.

Experimental results

Research questions

• RQ1 How do the topological structures of social networks differ between the historical text 'Records of the Three Kingdoms' and the novel 'Romance of the Three Kingdoms'?
• RQ2 To what extent does the historical novel construct a more complex and dynamic social network than the historical record?
• RQ3 How do sentiment portrayals of major characters like Cao Cao and Liu Bei differ between the two texts?
• RQ4 What is the quantitative relationship between narrative complexity and character influence in the two works?
• RQ5 How do sentiment scores from SentiWordNet reflect the authors' subjective evaluations of key characters in the two genres?

Key findings

• The social network in 'Romance of the Three Kingdoms' is significantly more complex and dynamic than in 'Records of the Three Kingdoms,' with higher centrality and clustering.
• The network in the novel exhibits small-world and scale-free properties, indicating a tightly connected, hierarchical structure conducive to narrative engagement.
• Cao Cao receives significantly more negative sentiment words in 'Romance' than in 'Records,' while Liu Bei is portrayed more positively in the novel.
• SentiWordNet scores confirm that Cao Cao's sentiment score is lower in 'Romance' than in 'Records,' while Liu Bei's score is higher in the novel, reflecting a shift in narrative tone.
• Both characters receive higher overall sentiment scores in 'Romance' than in 'Records,' indicating a more emotionally charged narrative in the novel.
• The study demonstrates that literary elaboration in historical novels enhances story vividness by deliberately complicating social networks and emotional portrayals.