The Catch-22 of Forgetfulness: Responsibility for Mental Mistakes

ABSTRACT

Attribution theorists assume that character information informs judgments of blame. But there is disagreement over why. One camp holds that character information is a fundamental determinant of blame. Another camp holds that character information merely provides evidence about the mental states and processes that determine responsibility. We argue for a two-channel view, where character simultaneously has fundamental and evidential effects on blame. In two large factorial studies (n = 495), participants rate whether someone is blameworthy when he makes a mistake (burns a cake or misses a bus stop). Although mental state inferences predict blame judgments, character information does not. Using mediation analyses, we find that character information influences responsibility via two channels (Studies 3–4; n = 447), which are sensitive to different kinds of information (Study 5; n = 149). On the one hand, forgetfulness increases judgments of responsibility, because mental lapses manifest an objectionable character flaw. On the other hand, forgetfulness decreases judgments of state control, which in turn decreases responsibility judgments. These two channels cancel out, which is why we find no aggregate effect of forgetfulness on responsibility. Our results challenge several fundamental assumptions about the role of character information in moral judgment, including that good character typically mitigates blame.

KEYWORDS:

• responsibility
• character
• mental action
• forgetfulness
• blame

Acknowledgements

We presented an earlier version of this paper at the Australasian Experimental Philosophy Group and the University of Michigan. We are grateful to everyone who participated in these discussions, especially Justin Sytsma, Jonathan Weinberg, Chandra Sripada, Peter Railton, and Laura Soter. For comments on our draft, we thank Jordan Bridges, Elise Dykhuis, Santiago Amaya, Rebecca Stangl, Nate Adams, and anonymous reviewers at this journal. Most importantly, we thank Sheisha Kulkarni, without whom this paper would have failed at multiple stages. She gave us key empirical advice (visualize your descriptive statistics!!!), philosophical feedback, and inspired the conscientious character in our Catch-22 of Forgetfulness.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1 We focus on responsibility as accountability, as do our experiments. Some individual is responsible in the accountability sense when their conduct makes them an appropriate target of negatively-valanced reactive attitudes, such as resentment [Shoemaker 2015]. Thus, accountability involves some culpability-imputing judgement that is non-trivially connected to moral emotions. Moreover, while some forms of responsibility mark out deserved credit or praise for right action, we limit our discussion to blame and the reactive attitudes associated with blameworthiness.

2 Previous research in the moral psychology of character has largely ignored trait control. This is part of a general trend. Past research has neglected character traits that concern one’s capacity for self-governance (e.g., forgetfulness), and instead focused on other-directed traits (e.g., fairness and kindness). We argue that both kinds of character traits affect responsibility, but in different ways (see discussion).

3 Participants had to infer trait information based on short text descriptions. Previous research has shown that people reliably infer dispositions from small amounts of information [Willis and Todorov 2006], including short narrative descriptions or single terms [Peabody and Goldberg 1989]. Moreover, these inferences are relatively stable within social groups [Stolier et al. 2020].

4 To carry out our mediation analysis, we used a bootstrapping procedure with 5000 samples to compute bias-corrected confidence intervals. Unsurprisingly, the overall effect of state control on blame was significant (p < .001, b = 1.46, 95% CI [1.02, 1.90]). Importantly, the indirect effect, mediated by causal judgements, was not significant (p = .065) whereas the direct effect was significant (p < .001, b = 1.41, 95% CI [0.97, 1.85]). This is the opposite of what one would predict if causal judgements drove the effect of state control on responsibility.

5 The Masters label is a performance-based distinction given to Mechanical Turk workers who demonstrate exemplary performance. Masters workers must maintain a high level of performance to retain the label.

6 We used a bootstrapping procedure with 5000 samples to compute bias-corrected 95% confidence intervals. We again found significant overall (p < .001, b = .92, 95% CI [.50, 1.29]) and direct effects (p < .001, b = .96, 95% CI [.56, 1.35]) of state control on blame. The indirect effect of state control on blame, mediated by causal judgements, was not significant ($p=.267$).

7 We modelled state control as a mediator, rather than a covariate, because we hypothesized that state control judgments should depend on our character manipulation.

8 Using a multiple mediation analysis, we confirmed that neither the direct nor indirect effects of character on blame were fully mediated by judgments of causal responsibility. We present these additional analyses in the OSF page for the project: https://osf.io/eqb2f/.

Additional information

Funding

This project was supported by University of Virginia’s 3 Cavaliers program, grant #164 (‘Harnessing the Wandering Mind’), the John Templeton Foundation, grant #60845 (‘Getting better at simple things’), and Duke University’s Summer Seminars in Neuroscience and Philosophy.