Skip to content

Emotional language in online discourse

Emotional language—text expressing or evoking emotional states such as anger, sadness, joy, disgust, or fear—is a key predictor of information virality and social influence. Emotionally charged content spreads faster and farther than neutral content, a phenomenon called "emotional contagion."

Key dimensions:

Emotional arousal — High-arousal emotions (anger, excitement, surprise) increase information diffusion more than low-arousal emotions (sadness, contentment). Negative emotional arousal (anger, anxiety) tends to increase sharing more than positive arousal in many contexts.

Valence effects — The positivity or negativity of emotional content interacts with topic and audience. For divisive political issues, negative moral-emotional language (expressing outrage) spreads within in-groups. For consensual issues (e.g., LGBTQ+ inclusion), positive emotional language spreads more.

Moral-emotional language — When emotional language is paired with moral judgment ("This is immoral" expressed with anger or disgust), the effect on diffusion is stronger than either emotion or morality alone. Moral-emotional words increase message reach by ~20% per word.

Discrete emotions — Different emotions have different effects. Anger and disgust (high-arousal moral emotions) amplify moral messaging. Sadness (low-arousal) suppresses sharing. Context determines which emotions drive diffusion (anger drives climate change sharing; positive sentiment drives LGBTQ+ content sharing).

Key papers

  • Brady et al. (2017) — Emotion shapes the diffusion of moralized content: analyzes 563,312 tweets on gun control, same-sex marriage, and climate change; shows moral-emotional words increase retweet counts by ~20% per word; demonstrates valence effects are topic-specific (negative emotions drive climate change diffusion; positive emotions drive same-sex marriage diffusion); shows sadness consistently suppresses diffusion while anger is context-dependent.

Connections