Describing your group's good behavior in broad character terms but their group's identical behavior in narrow, situational terms.
Imagine your friend shares their toy — you say 'she's so kind!' But if a kid from the other class shares their toy, you say 'he just handed it over.' When your friend does something bad, you say 'she bumped into him,' but when the other kid does something bad, you say 'he's mean.' You use big, personality words for the good stuff your group does and the bad stuff the other group does, but tiny, specific words for everything else.
Linguistic Intergroup Bias operates through systematic differences in the level of linguistic abstraction people use when describing behaviors of in-group versus out-group members. When an in-group member does something positive or an out-group member does something negative, people encode the behavior using abstract trait terms (e.g., 'generous,' 'aggressive') that imply a stable disposition. Conversely, when an in-group member does something negative or an out-group member does something positive, people default to concrete, descriptive action verbs (e.g., 'handed money to the man,' 'opened the door') that frame the behavior as a one-time, situational event. This asymmetry makes positive in-group attributes and negative out-group attributes seem enduring and characteristic, while inconvenient counter-examples are linguistically quarantined as exceptions. The effect operates largely outside of conscious awareness, making it a powerful vehicle for the transmission and perpetuation of stereotypes across conversations and generations.
The same glitch looks different depending on the terrain. Finance, medicine, a relationship, a team — same mechanism, different costume.
Analysts and financial media tend to describe successful investments by in-group firms using abstract, dispositional language ('innovative company,' 'visionary leadership') while describing identical successes by competitor firms with concrete action descriptions, subtly reinforcing brand loyalty and industry tribalism.
Healthcare providers may unconsciously use abstract language when describing non-compliant behavior by patients from demographic out-groups ('she is irresponsible about her health') while using concrete language for similar behavior by in-group patients ('he missed his Tuesday appointment'), reinforcing stereotypes about patient populations.
Anne Maass, Daniela Salvi, Luciano Arcuri, and Gün R. Semin, 1989, published in the Journal of Personality and Social Psychology. Built upon Semin and Fiedler's Linguistic Category Model (1988).
Maintaining a coherent, favorable representation of one's own group would have enhanced coalition stability, trust, and cooperative coordination in ancestral environments. Encoding in-group virtues as stable traits and out-group threats as enduring dispositions helped early humans quickly identify reliable allies and persistent dangers, supporting survival in intergroup competition.
Language models trained on large text corpora absorb and reproduce the linguistic intergroup biases present in training data, systematically using more abstract and dispositional language when describing certain demographic groups' negative behaviors and more concrete language for their positive behaviors. This can amplify stereotypes in AI-generated text, summaries, and sentiment analysis outputs.
This page taught you the name. The deck turns the name into reflex. 1,100+ swipeable scenarios, 1,100+ defenses, 650+ detection prompts — spaced-repetition Swipe Deck, unlimited Spot-the-Bias Quiz, Defense Playbook, Pre-Flight, My Blindspots, Cheat Sheets, Field Guide e-book. $39.53$59.