This paper examines figurative expressions, such as metaphors, and their association with emotional language in political rhetoric. We compare figurative and literal language employed by US governors by applying a deep learning-based model to their State of the State speeches from 1995 to 2022. Predictions are clustered into nine socioeconomic topics relevant to policy. We then document their use as part of strategic communication. Our results show that partisanship predicts the types of expressions used. Democratic governors are more likely to use highly figurative expressions when discussing environmental issues compared to their Republican counterparts. In contrast, Republicans use them relatively more when talking about moral values such as freedom and liberty. In addition, we show that figurative language on these topics is associated with more emotional language, with significant heterogeneity among them. These results suggest that figurative language may be used as emotional appeals to discuss key political issues.
Figurative and literal language in US governor speeches. // CCF Transactions on High Performance Computing. – 2026. – DOI: 10.1186/s41937-026-00155-y. – URL: https://link.springer.com/article/10.1186/s41937-026-00155-y
(2026). Figurative and literal language in US governor speeches. CCF Transactions on High Performance Computing. https://doi.org/10.1186/s41937-026-00155-y
“Figurative and literal language in US governor speeches.” CCF Transactions on High Performance Computing, 2026. doi:10.1186/s41937-026-00155-y.
@article{article34152,
title = {Figurative and literal language in US governor speeches},
journal = {CCF Transactions on High Performance Computing},
year = {2026},
doi = {10.1186/s41937-026-00155-y},
url = {https://link.springer.com/article/10.1186/s41937-026-00155-y}
}
This paper describes the microdata underlying the Swiss Consumer Price Index (CPI) and provides a fully reproducible pipeline for their use in empirical research. It documents the features most relevant for research—product hierarchy,...
We use administrative data on income and wealth of taxpayers in the canton of Bern to impute consumption based on the budget constraint. We validate the imputed consumption and use the panel data on consumption, income and wealth to...
This paper presents a weekly GDP indicator for Switzerland, which addresses the limitations of existing economic activity indicators by using alternative high-frequency data created in response to the COVID-19 pandemic. The indicator is...
Using data from the only large-scale randomized controlled trial promoting prolonged exclusive breastfeeding, we study how the intervention affected child health and why. The intervention increased weight-for-age in infancy, with...
The proposed revision of the Swiss Banking Act introduces a public liquidity backstop (PLB) for distressed systemically important banks (SIBs), in part to facilitate resolution. We examine the impact of the PLB on fiscal balances,...