Narrative, Translation & Transmission

Translation and adaptation do not simply preserve or distort a source. They reveal what persists across transformations and what changes under the pressure of language, culture, medium, or method.

With Jon Chun, Elkins has developed computational methods that make stories comparable as trajectories while returning every pattern to the passages, languages, and interpretive traditions that give it meaning.

The research problem

Narrative Form Across Versions

What survives translation?

What changes when a story crosses cultures?

Can narrative form be measured without reducing meaning?

Answering these questions requires more than a single scale of reading. Emotional trajectories locate patterns across complete works; close analysis of the original-language passages determines what those patterns mean. Model selection, smoothing, scale, and segmentation remain interpretive choices rather than neutral preprocessing.

Language and literary form

Translation & Literary Style

In Search of a Translator: Using AI to Evaluate What’s Lost in Translation

Frontiers in Computer Science 6, 2024 · DOI 10.3389/fcomp.2024.1444021

The Proust project compares a literary original with three English translations across whole-narrative time. It asks which dimensions of sentence structure, style, readability, syntax, lexical variation, sentiment, and emotional experience persist, flatten, or change.

Translation becomes a model of the larger representational question: what elements of literary form survive movement between systems, and which transformations alter interpretation?

Publication page →Proust and philosophy of mind →

Adaptation across cultures

Cultural Transmission & Variation

The Shapes of Cinderella: Emotional Architecture and the Language of Moral Difference

Humanities 14(10), 198, 2025 · DOI 10.3390/h14100198

The Cinderella research extends the persistence-and-change question from translation into cultural transmission. Original-language, clause-level trajectories compare Chinese, French, and German variants, asking how recognizable narrative structures survive while moral, emotional, and cultural features change.

The use of clauses matters: languages divide the same story into sentences differently, so sentence-level analysis can manufacture cross-language artifacts.

Publication page →Scholarly reception →

Comparison without reduction

Computational Philology

Computational philology combines computational comparison with close attention to textual variation, allowing researchers to trace how language, structure, and meaning change across versions. Technical choices are treated as interpretive decisions.

Middle reading

Computational trajectories locate structure; close reading of the passage decides what it means.

SentimentArcs

An ensemble pipeline for diachronic sentiment analysis, using dynamic time warping to compare narrative arcs of unequal length.

MultiSentimentArcs

An extension across media that compares coherence between a film’s visual and dialogue trajectories.

Segmentation

The choice of textual unit; clause-level comparison avoids artifacts introduced by different sentence boundaries across languages.

2019
Can Sentiment Analysis Reveal Structure in a “Plotless” Novel?
Introduced middle reading and joined quantitative trajectories to close reading.
2021–2023
SentimentArcs and explainable story analysis
Established ensemble sentiment trajectories, dynamic time warping, shared coordinates, and sentence-level explainability.
2022
The Shapes of Stories
Developed the methodology for narrative sentiment analysis and middle reading.
2024–2025
Translation, cross-modal comparison, and cultural transmission
Extended comparison to Proust translations, film, plotless narrative, and original-language Cinderella variants.

The Shapes of Stories →Beyond Plot →Technical research context →

A current extension

From Narrative Transmission to Cultural Preservation

Translation moves meaning between languages. Computational representation moves meaning between human-readable and machine-readable forms. In both cases, the central questions remain: what survives representation, what becomes newly visible, and what is flattened or lost?

This persistence-and-change framework now informs Archival Intelligence, where AI-assisted preservation raises new problems of context, local knowledge, and cultural transmission.

Research →About →