Poster #62 - Infants’ Moral Trait Inferences in Protective Third-party Intervention Paradigms

• In Event: Poster Session 18

Abstract

By representing other agents’ moral traits, humans can predict the future moral behaviour of those agents. There is some evidence that these representations develop as early as infancy. When infants and young toddlers are presented with harm causing or hindering agents, they suspend expectations that these agents will follow fairness norms (Ting & Baillargeon, 2021; Surian et al., 2018). These findings, however, are slightly limited as infants only suspended expectations for fairness but did not expect immoral agents to behave unfairly. To better understand the development of moral trait inferences in infancy, we tested infants with a novel paradigm - the protective third-party intervention (PTPI) paradigm. In PTPI, an aggressor repeatedly hits a victim before a third-party intervenes to protect the victim. Kanakogi et al. (2013, 2017) find that infants (a) prefer victims over aggressors and (b) prefer interveners over non-interveners. We examined whether infants would form predictions regarding how agents in a PTPI paradigm would behave in a separate resource distribution situation.

Using a violation-of-expectation paradigm, 92 infants aged 12 to 24 months (Mage = 18.1 months, 49 female) were tested online over zoom. Infants were randomly assigned to three conditions, aggressor, intervener, or baseline. Infants in the aggressor and intervener conditions were familiarized to four repetitions of a PTPI paradigm (Figure 1A); infants in the baseline condition were familiarized using a random movement control. In the aggressor condition, infants saw two test trials, one in which the aggressor distributed resources equally, and another in which the aggressor distributed resources unequally (Figure 1B). In the test trials for the intervener condition, the intervener made the two resource distributions (Figure 1C), and in the baseline condition, one of the three agents made resource distributions. A mixed-factor ANOVA revealed a significant interaction between condition and test trial type (F(2, 89) = 6.78, p = .001; figure 2). Post-hoc t-tests demonstrated a significant difference in looking time towards fair and unfair test trials in the aggressor (t(31) = 2.45, p = .02) and intervener (t(27) = 2.34, p = .018) conditions, but not in the baseline condition (t(31) = 1.84, p = .076). Infants looked longer when the aggressor distributed resources fairly, but when the intervener was making the resource distributions, infants looked longer at the unfair resource distribution. Although it was only marginally significant, infants in the baseline condition looked longer at the unfair test trial, consistent with research on infants’ expectations for fairness norms (Ziv & Sommerville, 2017). These results demonstrate that infants formed expectations about how agents in a PTPI situation would behave in a separate moral domain, being surprised when aggressors behaved fairly, and when interveners behaved unfairly.

The results of this research provide evidence that infants use information that they have gained about an agent's moral leaning in one context to inform their expectations about how that agent will behave in future moral contexts. This contributes to a growing body of literature suggesting that moral trait inferences are an important component of infants’ socio-moral development.

Authors

• Norman J Zeng, University of Toronto
  Presenting Author

• Inderpreet Kaur Gill, University of Toronto

• Jessica A Sommerville, University of Toronto