Morality Between the Lines: Detecting Moral Sentiment In Text

Expressions of moral sentiment play a fundamental role in political framing, social solidarity, and basic human motivation. Moral rhetoric helps us communicate the reasoning behind our choices, how we feel we should govern, and the communities to which we belong. In this paper, we use shortpost social media to compare the accuracy of text analysis methods for detecting moral rhetoric and longer form political speeches to explore detecting shifts in that rhetoric over time. Building on previous work using word count methods and the Moral Foundations Dictionary [Graham et al., 2009], we make use of pre-trained distributed representations for words to extend this dictionary. We show that combining the MFD with distributed representations allows us to capture a cleaner signal when detecting moral rhetoric, particularly with shortform text. We further demonstrate how the addition of distributed representations can simplify dictionary creation. Finally, we demonstrate how capturing moral rhetoric in text over time opens up new avenues for research such as assessing when and how arguments become moralized and how moral rhetoric impacts subsequent behavior.