Does Knowledge of Prior Moral Behaviour Influence Infants’ Expectations of Future Moral Behaviour?

• In Event: Inferences drawn from moral actions in infants and children

Abstract

Adults recognize that behavior in one moral subdomain can have implications in another: for example, they expect someone who helped someone cross the street to also return extra money they received by mistake to a cashier (Harris et al., 2016). Here, we investigate infants’ ability to generalize inferences across the most central subdomains of morality: help/harm and fairness (i.e., Lucca et al., 2021; Woo et al., 2022).
A variety of evidence suggests that infants prefer helpers over hinderers (Hamlin et al., 2007, 2013), expect fair resource distributions (Schmidt & Sommerville, 2011; Geraci & Surian, 2011), and use this information to guide their social evaluations (Geraci & Surian, 2011; Van de Vondervoort & Hamlin, 2016). Limited evidence suggests infants may also recognize the consequences of behavior in one sub-domain for another: 14-month-old infants suspend expectations for fair resource distributions after learning that the distributing agent had previously hindered (Surian et al., 2018). We investigated whether and when infants expect agents to engage in subsequent moral transgressions after committing a moral transgression in another domain (i.e., act unfairly after hindering and vice versa).
We tested 14- to 24-month-old infants (Experiment 1, Mage = 17.98, n = 48; Experiment 2, Mage = 17.96. n = 51). In Experiment 1, we aimed to replicate whether infants link help/harm with fairness (cf. Surian et al., 2018): infants were familiarized to a helper or hinderer; on test, infants saw the same agent perform fair and unfair distributions. Replicating prior work (i.e., Surian et al., 2018), following familiarization to helping behaviour, infants look longer to unfair behaviour, M = 23.02, SE = 2.09 than fair behaviour, M = 8.90, SE = 2.18; t(18) = -4.52, p < .001. Yet, following familiarization to hindering behaviour, infants looked equally to both test events (MFair = 17.37, SE = 2.47 and MUnfair = 16.90, SE = 2.3; t(26) = .139, p = .891). In Experiment 2, following familiarization to fair or unfair distributions, infants watched the same agent help or hinder. Infants familiarized to unfair behavior looked longer to helping, M = 22.12, SE = 1.80, than hindering, M = 15.69, SE = 2.60, t(23) = 2.06, p = .051, yet infants familiarized to fair behavior looked equally to both test events (MHelp = 15.67, SE = 2.25 and MHinder = 14.89, SE = 1.92; t(24) = .258, p = .799).
Together, these results suggest that infants generalize bidirectionally. Given prior work suggesting that infants possess a baseline expectation for fairness (see Sommerville & Enright, 2018; Sommerville & Ziv, 2018), but not helpfulness (Hamlin et al., 2007; Hamlin & Wynn, 2011), our findings illustrate that this generalization manifests in different ways across each moral subdomain – leading infants to eliminate expectations for fair behavior following hindering, and generate expectations for hindering following unfair behavior. Moreover, our results suggest that moral transgressions provide a more powerful means for generalization than adherence to moral norms. In ongoing work, we are investigating the basis for this asymmetry in generalizations from moral transgressions versus adherence.

Authors

• Inderpreet Kaur Gill, University of Toronto
  Presenting Author

• Jessica A Sommerville, University of Toronto