Ann Arbor Times

A recent study conducted by researchers at the University of Michigan and the University of Chicago explores the significant role metaphors in tweets play in shaping public perception of immigration. The study analyzed over 400,000 immigration-related tweets from 2018 to 2019 by individuals aligned with two political parties in the United States.

The study conducted a linguistic analysis focusing on seven metaphorical concepts: animal, vermin, parasite, physical pressure, water, commodity, and war. According to the researchers, these metaphors can either be explicit, such as referring to immigrants as “animals,” or more implicit, like describing them using terms like “swarms” or “flocking.”

Ceren Budak, associate professor at the U-M School of Information, noted, “When dehumanizing metaphors are used, it can increase discrimination and harsh policies.” The findings reveal that those with strong conservative views tend to employ more metaphors compared to moderates. Conversely, liberals, especially those on the far left, use fewer metaphors about water or commerce but employ more creature-related metaphors.

Lead author Julia Mendelsohn, a postdoctoral scholar at the University of Chicago, explained that metaphor usage varies between the political parties. Mendelsohn observed that tweets containing creature metaphors are more likely to be retweeted, particularly among liberals. “So, the difference between liberals’ tweets with and without creature metaphors is greater than the difference between conservatives’ tweets with and without creature metaphors,” she said.

The research utilized large language models, notably ChatGPT, to identify metaphors, highlighting the potential of these models for detailed political discourse analysis. However, Mendelsohn emphasized the need for diligent study design and human oversight in the evaluation of these language models.