Joint optimization of informational and reputational goals in moral narratives

For any given situation (e.g., after committing a wrongdoing), different people are likely to provide different accounts about what happened. Why do people decide what to say, and why do they tell different narratives about the same situation? We asked whether individual narrators’ moral narratives can be explained by how they balance conflicting informational and reputational goals. Across 15 experiments and 2 sets of paradigms (economic game setting vs. hypothetical vignette), participants provided naturalistic narratives about their own moral actions (morally questionable vs. morally positive) to an audience. We find that given a conflict between being informative and preserving their reputation, the variability in narratives, as well as variability of self-reported strategies, increase. Notably, these narrators find diverse ways to jointly optimize both informational and reputational goals, becoming under-informative to simultaneously not lie and not reveal the (undesirable) truth. More generally, we demonstrate how even complex, open-ended narratives can be understood within the popular rational speech act (RSA) framework.