Overriding a moral default for honesty or dishonesty

There is a long-standing paradox concerning the cognitive nature of honesty: Is it a matter of “will” or “grace” (1)? The will hypothesis assumes that honesty requires cognitive control to suppress temptation to cheat, while dishonest behavior to serve self-interest is people’s automatic response. In contrast, the grace hypothesis assumes that honesty flows automatically without active resistance to temptation, while dishonest behavior is realized by cognitive control to override honest impulses. The previous findings related to this debate are mixed: Some studies have empirically supported the will hypothesis (2, 3), but others have empirically supported the grace hypothesis (4, 5). In an ambitious study in PNAS, Speer et al. (6) provide reconciliation between these two competing hypotheses, indicating that the prefrontal network could orchestrate both the honesty of individuals who are generally dishonest and the dishonesty of those who are generally honest through cognitive control, which depends on the individual’s moral default. Speer et al. suggest a reconciliation between the evidence supporting the will and grace hypotheses by focusing on individual differences in honesty, using the newly developed experimental paradigm. In the study of self-serving dishonesty, researchers often engage participants in tasks that enable them to be honest or to cheat to obtain monetary rewards (7). For example, in tasks that reward participants according to their self-reported accuracy of a private prediction of random coin flip, some of the participants report higher-than-chance accuracy, suggesting that they cheat to increase monetary gain. Although this kind of task design is useful to determine whether a participant shows improbably high levels of self-reported accuracy, from which researchers can infer their dishonest behavior, it does not allow the identification of individual lies: Whereas some of the cheating trials involve actual lying, others can be won honestly. This situation is problematic, especially for … [↵][1]1Email: abe.nobuhito.7s{at}kyoto-u.ac.jp. [1]: #xref-corresp-1-1