AUTHOR=Maoz Uri , Sita Kellienne R. , van Boxtel Jeroen J. A. , Mudrik Liad TITLE=Does It Matter Whether You or Your Brain Did It? An Empirical Investigation of the Influence of the Double Subject Fallacy on Moral Responsibility Judgments JOURNAL=Frontiers in Psychology VOLUME=10 YEAR=2019 URL=https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2019.00950 DOI=10.3389/fpsyg.2019.00950 ISSN=1664-1078 ABSTRACT=

Despite progress in cognitive neuroscience, we are still far from understanding the relations between the brain and the conscious self. We previously suggested that some neuroscientific texts that attempt to clarify these relations may in fact make them more difficult to understand. Such texts—ranging from popular science to high-impact scientific publications—position the brain and the conscious self as two independent, interacting subjects, capable of possessing opposite psychological states. We termed such writing ‘Double Subject Fallacy’ (DSF). We further suggested that such DSF language, besides being conceptually confusing and reflecting dualistic intuitions, might affect people’s conceptions of moral responsibility, lessening the perception of guilt over actions. Here, we empirically investigated this proposition with a series of three experiments (pilot and two preregistered replications). Subjects were presented with moral scenarios where the defendant was either (1) clearly guilty, (2) ambiguous, or (3) clearly innocent while the accompanying neuroscientific evidence about the defendant was presented using DSF or non-DSF language. Subjects were instructed to rate the defendant’s guilt in all experiments. Subjects rated the defendant in the clearly guilty scenario as guiltier than in the two other scenarios and the defendant in the ambiguously described scenario as guiltier than in the innocent scenario, as expected. In Experiment 1 (N = 609), an effect was further found for DSF language in the expected direction: subjects rated the defendant less guilty when the neuroscientific evidence was described using DSF language, across all levels of culpability. However, this effect did not replicate in Experiment 2 (N = 1794), which focused on different moral scenario, nor in Experiment 3 (N = 1810), which was an exact replication of Experiment 1. Bayesian analyses yielded strong evidence against the existence of an effect of DSF language on the perception of guilt. Our results thus challenge the claim that DSF language affects subjects’ moral judgments. They further demonstrate the importance of good scientific practice, including preregistration and—most critically—replication, to avoid reaching erroneous conclusions based on false-positive results.