When People are Floods: Analyzing Dehumanizing Metaphors in Immigration Discourse with Large Language Models

2025-02-18

AI Summary

This study investigates the distribution of dehumanizing metaphors (e.g., “flood,” “pest”) in social media discourse on immigration and their associations with political ideology and user engagement. We propose the first multi-granular metaphor quantification framework integrating word-level semantic similarity, document-level topic modeling, and large language model–based metaphor detection to systematically measure the strength of seven conceptual metaphor categories. Empirical analysis of 400,000 U.S. immigration-related tweets reveals that conservative users deploy certain dehumanizing metaphors more frequently; however, biologically grounded metaphors (e.g., “virus,” “pest”) significantly increase retweet rates—and this effect is stronger among liberal authors—demonstrating a non-monotonic relationship between metaphor use and ideological orientation. Our work contributes a reproducible methodological paradigm and empirical evidence for the political semantics of metaphor in digital public discourse.

Technology Category

• Natural Language Processing: Discourse, Pragmatics & Argument Mining
• Application Domains: Humanities & Computational Social Science
• Cognitive Modeling & Cognitive Systems: Analogy

Application Category

• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Social Networks and Social Media: Computational social science
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Metaphor, discussing one concept in terms of another, is abundant in politics and can shape how people understand important issues. We develop a computational approach to measure metaphorical language, focusing on immigration discourse on social media. Grounded in qualitative social science research, we identify seven concepts evoked in immigration discourse (e.g."water"or"vermin"). We propose and evaluate a novel technique that leverages both word-level and document-level signals to measure metaphor with respect to these concepts. We then study the relationship between metaphor, political ideology, and user engagement in 400K US tweets about immigration. While conservatives tend to use dehumanizing metaphors more than liberals, this effect varies widely across concepts. Moreover, creature-related metaphor is associated with more retweets, especially for liberal authors. Our work highlights the potential for computational methods to complement qualitative approaches in understanding subtle and implicit language in political discourse.

Problem

• Analyzing dehumanizing metaphors in immigration discourse
• Measuring metaphorical language in social media
• Linking metaphor use to political ideology and engagement

Innovation

• Large Language Models analysis
• Word and document-level signals
• Metaphor measurement technique