Dependency Distance Effects on Eye-Tracking Measures in Brazilian Portuguese

2026-04-12 · Source: Paper Index on ACL Anthology · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Computational Linguistics · Depth: Expert, short

Summary

Research presented at PROPOR 2026 investigates how dependency distance and its directionality influence eye-tracking measures in Brazilian Portuguese. Utilizing the RastrOS corpus, which is augmented with surprisal and syntactic annotations, the study found that the absolute dependency distance significantly enhances the prediction of first fixation durations. This finding supports memory-based theories of sentence processing. Conversely, the direction of the dependency, whether the dependent precedes or follows its head, exhibited less consistent and weaker effects. The results suggest that the magnitude of syntactic distance is crucial for early lexical retrieval, while later reading processes related to integration are less impacted, underscoring the combined importance of syntactic distance and surprisal in modeling reading behavior.

Key takeaway

For NLP engineers developing language models for Portuguese, your focus should include absolute dependency distance as a critical feature. This research indicates that incorporating this metric can improve the accuracy of models predicting early lexical retrieval and overall reading behavior, potentially leading to more human-like processing simulations. Consider refining your model's feature set to explicitly account for the magnitude of syntactic dependencies.

Key insights

Absolute dependency distance significantly predicts early lexical retrieval in Brazilian Portuguese eye-tracking.

Principles

• Memory-based accounts explain sentence processing.
• Syntactic distance complements surprisal in reading models.

Method

The study used the RastrOS corpus, enriched with surprisal and syntactic annotations, to analyze the effect of dependency distance and its directionality on eye-tracking measures in Brazilian Portuguese.

In practice

• Integrate syntactic distance into NLP models.
• Consider dependency magnitude for language processing.

Topics

• Dependency Distance
• Eye-Tracking Measures
• Brazilian Portuguese
• RastrOS Corpus
• Sentence Processing

Best for: AI Scientist, NLP Engineer, Research Scientist

Related on AIssential

• Global vs. Local Sentence Embeddings for Brazilian Portuguese: Revisiting Monolingual Models in the Age of Foundation Models — Optimal model choice for PT-BR sentence embeddings balances vocabulary coverage and adaptation flexibility.
• Probing for Reading Times — Early language model layers predict initial human reading effort better than scalar surprisal, which excels at later stages.
• Enhanced Universal Dependencies in the Wild: Evaluating Portuguese EUD Parsing in Realistic Scenarios — Portuguese-specific EUD rules outperform universal rules, but automatic basic syntax severely impacts performance.
• What They Saw, Not Just Where They Looked: Semantic Scanpath Similarity via VLMs and NLP metric — VLMs and NLP metrics can quantify semantic similarity in eye-tracking scanpaths, complementing traditional spatial analysis.
• AfriSUD: A Dependency Treebank Collection for Evaluating Models on African Languages — AfriSUD provides the first large-scale, native-speaker verified dependency treebanks for nine diverse African languages, revealing a significant syntax gap in current NLP models.
• Disparities In Negation Understanding Across Languages In Vision-Language Models — VLMs exhibit affirmation bias, with negation understanding varying significantly across typologically diverse languages.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.