Abstract
Bare plural generics ('generics' for short) attribute a feature to members of a category without specifying how many actually possess the feature (e.g., 'Belgians love fries'). Generics are often used to perpetuate stereotypes and misinformation, yet researchers from many different fields disagree about how people decide what justifies a generic. This has led to a wide variety of theories on how people reason with generics. Pragmatic theories state that people find generics justifiable if they express knowledge that is useful (e.g., for survival or efficient transmission of knowledge). Statistical theories state that people find generics justifiable if the distribution of the features in the involved categories satisfies certain criteria. We compared the predictions of several influential theories in a registered study where participants, in each trial, saw the distribution of a feature in two fictitious groups. Participants then judged the justifiability of a generic that attributes the feature to one of the two groups. We independently manipulated the dangerousness of the features (a factor of relevance in pragmatic theories), absolute prevalence in the target group, and relative prevalence in the target group (factors of relevance in statistical theories). All experimental manipulations were within-participants. The order of conditions and the combinations of stimulus materials were fully randomized. Because our design allowed to predict specific, divergent patterns of main and interaction effects from each theory, this registered study allowed us to examine their relative merit and to further elucidate the cognitive mechanisms underlying generic justifiability.