Five-Month-Old Infants Attribute Inferences Based on General Knowledge to Agents

Abstract

To make sense of others’ actions, we generally consider what information is available to them. This information may come from different sources, including perception and inference. Like adults, young infants track what information agents can obtain through perception: If an agent directly observes an event, for example, they expect her to have information about it. However, no investigation has yet examined whether young infants also track what information agents can obtain through inference, by bringing to bear relevant general knowledge. Building on the finding that by 4 months of age, most infants have acquired the physical rule that wide objects can fit into wide but not narrow containers, we asked whether 5-month-olds would expect an agent who was searching for a wide toy hidden in her absence to reach for a wide as opposed to a narrow box. Infants looked significantly longer when the agent selected the narrow box, suggesting that they expected her (a) to share the physical knowledge that wide objects can fit only into wide containers and (b) to infer that the wide toy must be hidden in the wide box. Three additional conditions supported this interpretation. Together, these results cast doubt on two-system accounts of early psychological reasoning, which claim that infants’ early-developing system is too inflexible and encapsulated to integrate inputs from other cognitive processes, such as physical reasoning. Instead, the results support one-system accounts and provide new evidence that young infants’ burgeoning psychological-reasoning system is qualitatively similar to that of older children and adults.

Can 5‐month‐old infants consider the perspective of a novel eyeless agent? New evidence for early mentalistic reasoning

“…Our research demonstrates that infants in the first half‐year of life engage in perspective taking even with a novel eyeless agent: When the agent's representation of a scene is less complete than their own, they use the agent's incomplete representation, non‐egocentrically, to interpret its actions. These results extend those described in the Introduction involving agents with eyes (Choi & Luo, 2015; Hamlin et al, 2013; Kim & Song, 2015; Liszkowski et al, 2007, 2008; Luo & Baillargeon, 2007; Luo & Johnson, 2009; Meristo & Surian, 2013; Ting et al, 2019, 2021; Tomasello & Haberl, 2003; Vouloumanos et al, 2014), and they provide converging evidence that early epistemic reasoning is both mentalistic and non‐egocentric. Together, these results make clear that mentalistic accounts provide a more accurate characterization of early epistemic reasoning than do teleological accounts.…”

“…In our task, 5‐month‐old infants blocked out an object that an eyeless agent could not detect. In other tasks, 5‐ to 12‐month‐olds blocked out objects that an agent with eyes could not see (Kim & Song, 2015; Luo & Baillargeon, 2007; Luo & Johnson, 2009; Ting et al, 2021). In yet other tasks, infants behaved as though they recognized that objects or agents might be hidden from them .…”

“…To assess infants’ expectations, we built on two prior findings from preference tasks involving agents with eyes. The first finding is that infants aged 5 months and older attribute a preference for object‐A to an agent who consistently approaches object‐A as opposed to object‐B, and they expect the agent to continue acting on this preference when the objects’ positions are switched (Ting et al, 2021; Woodward, 1998). The second finding is that this effect is eliminated if the agent cannot see object‐B: If object‐B is hidden from the agent by a barrier, infants refrain from attributing a preference for object‐A to the agent (Kim & Song, 2015; Luo & Johnson, 2009).…”