Why side-effect outcomes do not affect intuitions about intentional actions: Properly shifting the focus from intentional outcomes back to intentional actions

Over the last decade, many articles have suggested that the "badness" of side-effect outcomes influences perceivers' intuitions about intentionality, contradicting the traditional notion that mental state inferences lead to moral judgments rather than the reverse. Challenging this assertion, we argue that typically, consideration of intentionality involves thinking about "intentional actions" (things people do) rather than unintended outcomes. Across several studies, we offer an explanatory framework describing why side-effect asymmetries emerge. We first establish that people differentiate actions, outcomes, goals, and side effects, associating intentions with goals but intentionality with actions in furtherance of goals, and that each of these components is readily identified in side-effect scenarios. We then demonstrate that when relationships among actions, goals, and side effects are available for consideration in response options, side-effect effects disappear. We additionally show that, because actions are not explicitly referenced, people reinterpret questions about the intentionality of side effects-particularly for harmful outcomes-as asking about intentional actions that caused side effects, creating a mismatch between participants' pragmatic and researchers' literal interpretations. Finally, we demonstrate how harmful side effects shift perceivers' attention toward considering agents' knowledge/awareness, whereas beneficial side effects focus attention on intentions/motives, which serves a useful social purpose. We discuss how perceptions of intentionality are not influenced by side-effect valence, although, because of structural differences in how people view harm versus benefit, outcomes influence which mental states perceivers consider important when answering questions that are typically asked in side-effects research. Beyond intentionality, we consider how these findings may shed light on trait attribution processes, more generally.