A Scalable Tool for Measuring Manner and Result Verbs in Developmental Language Research

2026-06-25 · Source: Paper Index on ACL Anthology · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Advanced, quick

Summary

A new computational tool addresses the challenge of measuring manner and result verbs at scale for developmental language research. Researchers developed an approach that leverages linguistically informed prompts to generate sentence-level annotations using large language models (LLMs) over data from MASC and InterCorp. This process significantly extends coverage from previously annotated portions of VerbNet to 436 classes. A RoBERTa-based classifier was then trained on these LLM-generated annotations and evaluated on three held-out gold-standard datasets, including a new expert-annotated set. The model demonstrated promising performance, achieving an average accuracy of up to 89.6%, positioning it as a scalable measurement tool for future verb semantics research.

Key takeaway

For developmental language researchers or NLP engineers studying verb semantics, this tool offers a scalable method to classify manner and result verbs. You can leverage its 89.6% accuracy to analyze large datasets, extending beyond traditional VerbNet coverage to 436 classes. Consider integrating this RoBERTa-based classifier into your workflow to accelerate research on early verb learning and other linguistic phenomena, while validating borderline cases.

Key insights

A scalable computational tool uses LLMs and RoBERTa to classify manner and result verbs for linguistic research.

Principles

• Linguistically informed prompts enhance LLM annotation quality.
• Computational methods can scale verb classification for research.

Method

Generate sentence-level annotations using LLMs with linguistically informed prompts on MASC/InterCorp data. Train a RoBERTa-based classifier on these annotations, then evaluate on gold-standard datasets.

In practice

• Apply LLM-generated annotations to expand linguistic datasets.
• Utilize RoBERTa for fine-grained verb classification tasks.

Topics

• Manner and Result Verbs
• Developmental Language Research
• Large Language Models
• RoBERTa
• Verb Semantics
• Computational Linguistics

Best for: AI Scientist, NLP Engineer, Research Scientist

Related on AIssential

• A Scalable Tool for Measuring Manner and Result Verbs in Developmental Language Research — A RoBERTa classifier, trained on LLM-generated annotations, accurately distinguishes manner and result verbs at scale.
• Can Machines Tell Whether Gamers Love or Hate a Game? — Fine-tuning Transformer models like RoBERTa significantly improves sentiment analysis accuracy on imbalanced, complex human language datasets.
• Datasets for Verb Alternations across Languages: BLM Templates and Data Augmentation Strategies — This work addresses the underexplored ability of large language models (LLMs) to capture cross-sentence paradigmatic patterns, specifically verb alternations.
• GENIE: A Fine-Grained Measure for Novelty — GENIE offers a fine-grained metric to measure task-specific novelty in LLM outputs, surpassing holistic evaluations.
• A Lexicon-Grammar of Brazilian Portuguese Predicative Adjectives — A new machine-readable lexicon for Brazilian Portuguese predicative adjectives was created with 36 syntactic properties.
• MORPHOGEN: A Multilingual Benchmark for Evaluating Gender-Aware Morphological Generation — MORPHOGEN evaluates LLM gender-aware morphological generation across French, Arabic, and Hindi, revealing performance gaps and biases.

Open in AIssential →

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