Much is at stake in debates about rationalization. To rationalize is to take one's non-rational action and “[concoct] the beliefs or desires that would have made it rational” (target article, sect. 1.1, para. 1). If rationalization were widespread, we would have to abandon a premise of much discourse: that our beliefs and actions are generally based on reasons. Imagine a person arguing for climate policies by stating: “I believe humans have caused global warming, but my belief is not based on reasons.” The absurdity of the statement exemplifies how human discourse usually assumes that rationalization is a curious aberration, not our default mode of operation. In support of this assumption, we propose that rationalization is rare and reasoning is pervasive.

Cushman's account of rationalization begins with the claim that people “rationalize all the time” (sect. 1.1, para. 2) and “are never fully certain of what to believe and what to value” (sect. 1, para. 3). We term this the ubiquity hypothesis. Cushman supports the ubiquity hypothesis with two kinds of evidence: negative evidence that many behaviors cannot be based on reasoning and positive evidence that many behaviors are caused by something other than reasoning. (We assume that Cushman views his own arguments as reasons that support his theory, not as rationalizations.)

We contend that the reviewed research offers little evidence for rationalization or its ubiquity. As negative evidence of rationalization, Cushman discusses a study of vacation preferences (Sharot et al. Reference Sharot, Velasquez and Dolan2010). Each participant rated how they would feel, from unhappy (=1) to extremely happy (=6), about various vacation destinations. Next, researchers picked pairs of destinations that the participant had rated identically and told the participant, falsely, that they had chosen one destination over another, for instance, Thailand over Greece. On average, participants who were told that they had chosen Thailand increased their post-choice rating of Thailand by 0.08 points (Sharot et al. Reference Sharot, Velasquez and Dolan2010, Fig. 1). Cushman describes the participants’ responses as “gross errors” and “rationalization” (sect. 1.1, para. 4). However, ratings of, say, 5.00 and 5.08 on a 6-point scale represent near-identical attitudes (Krosnick Reference Krosnick1999): Either rating means that participants were very happy to travel. Insofar as the ratings are practically indistinguishable, choosing one over the other would not be an error, let alone a gross error. Moreover, the study did not document that participants concocted beliefs or desires to explain their shifts in ratings, as would have been required by the definition of rationalization.

Indeed, non-rational processes have mostly been found to influence choices among near-equivalent options (Krueger & Funder Reference Krueger and Funder2004; Turiel Reference Turiel2010). For instance, incidental disgust can make ratings of others’ actions slightly more negative, say from 3 to 14 on a 100-point scale (Wheatley & Haidt Reference Wheatley and Haidt2005). In contrast, incidental disgust rarely, if ever, causes categorical shifts in judgments (e.g., from “okay” to “wrong,” Landy & Goodwin Reference Landy and Goodwin2015; Pizarro et al. Reference Pizarro, Inbar and Helion2011). For more consequential choices, for instance, about saving instead of taking a life or vacationing in York instead of New York, evidence for non-rational influences is scant (Dahl et al. Reference Dahl, Gingo, Uttich and Turiel2018; Royzman & Hagan Reference Royzman, Hagan, Bonefon and Trémolière2017).

We find the positive evidence for rationalization similarly unconvincing. Cushman considers three “non-rational” influences on behavior: norms, innate instincts, and learned habits (for critiques of innate vs. learned behaviors, see Dahl Reference Dahl2019; Spencer et al. Reference Spencer, Blumberg, McMurray, Robinson, Samuelson and Tomblin2009). On Cushman's account, people often follow norms because “it just feels like the right thing to do” (sect. 3.4, para. 2). Discussing a study on food avoidance among pregnant and nursing women in Fiji (Henrich & Henrich Reference Henrich and Henrich2010), Cushman writes that the women avoid certain foods “simply because of the norm” and are unaware that the foods “pose special risks to fetuses and infants” (sect. 3.4, para. 4). However, interviews with the mothers contradict this interpretation: Most believed that the foods did pose risks to their fetuses and infants, and many had additional reasons for avoiding the foods (e.g., social disapproval; Henrich & Henrich Reference Henrich and Henrich2010, supplementary materials). The study did not show that people followed norms merely because it felt right – rather, participants expressed reasons for their actions.

Psychological research shows that reasoning plays a pervasive role in human decisions (Dahl et al. Reference Dahl, Gingo, Uttich and Turiel2018). By reasoning, we mean the formation of beliefs or decisions in accordance with considerations that people articulate and endorse (Dahl & Killen Reference Dahl, Killen, Wixted and Ghetti2018). Reasoning is evident in many areas of psychology, operating quickly or slowly and often in concert with emotions (Adler & Rips Reference Adler and Rips2008; Ajzen & Fishbein Reference Ajzen and Fishbein2005; Harman Reference Harman1986). Our work has shown that children and adults provide reasons that largely explain their social and moral judgments (see Dahl & Killen Reference Dahl, Killen, Wixted and Ghetti2018; Turiel & Dahl Reference Turiel, Dahl, Bayertz and Boughley2019). In recent research on moral judgments, we interviewed adolescents and adults about pushing one person to his death in order to save five others (Dahl et al. Reference Dahl, Gingo, Uttich and Turiel2018). In Study 1, 81% said it would be wrong to push the one person; nearly of half of them reasoned that the victim had no involvement in the situation. Study 2 manipulated victim involvement by stipulating that the person to be pushed had tried to kill the five persons; now, only 27% thought pushing the person was wrong. As in numerous studies, participants provided reasons for their judgments that, in experimental manipulations, proved to guide their judgments.

The abundant evidence for reasoning, and the limited evidence for rationalization, lead to our conclusion that reasoning is pervasive and rationalization is rare. This conclusion is both reassuring and forward-looking: reassuring because it supports the assumption that our beliefs and actions are largely guided by reasons, and forward-looking because it points to new areas of research on rationalization. We expect that rationalization is particularly common in ambiguous or challenging situations in which individuals struggle to decide among comparable options (Dahl & Waltzer Reference Dahl and Waltzer2018; Turiel & Dahl Reference Turiel, Dahl, Bayertz and Boughley2019). Studying rationalization in ambiguous situations is a valuable topic of inquiry, as long as researchers recognize how many of our everyday situations are far less ambiguous (Dahl Reference Dahl2017). More research is needed to identify the contexts in which rationalization occurs, and to separate reasoning and rationalization through careful definitions and operationalizations.