A traditional view of understanding other’s mental states is that early in ontogeny infants start from a me-first position, and through themselves learn to understand others. Here we propose an opposite developmental trajectory, where infants might start out highly altercentric (Southgate, 2020) and through development increasingly rely on their own point of view. In a pre-registered cross-sectional study we present 1-6-year-old children with a task where altercentric modulation has been found with 14-month-old infants (Kampis & Kovács, 2020). There, infants tended to search longer in a box when another person believed an object to be present, than when she believed it was empty. We predicted this tendency to decrease with age. Preliminary results with n=191 children included based on their search in baseline trials, show a decrease in altercentric modulation with age (r=-.174, p= .016), which will be discussed together with the relationship between altercentrism and the development of self-concept.

We constantly interpret and predict others’ decisions and behaviors when interacting with the world. Do our predictions of others’ decisions differ depending on whose behavior we are trying to predict? In four experiments, we probed how adults and children (N = 144, 6- to 8-years old) predict different agents’ intuitive numerical decisions. We varied the phylogenetic and ontogenetic history of the agents: an adult, a child, an infant, a chimpanzee, and an ant. We found that both adult and child participants predicted the adult and child agents to reliably choose the larger amount. Both adult and child participants predicted the infant and ant to choose at random. However, while adult participants attribute numerical decisions similar to adults and children to chimpanzees, children seem to think chimpanzees behave similarly to infants and ants. Overall, these results suggest that adults and children share similar intuitions about others’ numerical decisions based on agent identity.

Adults (Keysar et al, 2000) and children (Epley et al, 2004) sometimes commit egocentric errors when interpreting otherscommunication, if the self-perspective differs from the speakers perspective. Training imitation inhibition reduces egocen-tric error in adults (Santiesteban et al., 2011), presumably because it makes salient the distinction between self and other.As managing the self-other perspective difference may undergo developmental changes during preschool years (South-gate, in press), we tested whether a social imitation inhibition training may reduce egocentric mistakes in 3-6-year-oldchildren. Results with n=47 (of n=50 preregistered) children show that the imitation inhibition group selected the object towhich the speaker referred more often than children in a control condition (F(1,35)=5.346, p=.026). However, there wasan interaction with age (F(2,35)=3.805, p=.032): only 4-year-olds, but neither 3- nor 6-year-olds, were more accurate inthe inhibition group. Childrens reaction times and hesitation will be analyzed on the final sample.

Most infant studies on location memory involve an agent hiding or retrieving the object. Recent work indicates that, foryoung infants, the presence of other agents enhances encoding of the targets of their actions and perceptions, and in apilot study we did not find evidence for location memory with a paradigm where we removed agency cues. Here, wesystematically compared whether 7.5-months-old infants remembered the location of an object better when it was placedthere by an agent compared to a highly similar but non-social setting where a conveyor belt transports the object. Locationmemory was tested through infants’ looking times in response to outcomes showing unexpected vs. expected absences ofthe object. Contrary to our preliminary results, at n=58/64 of this preregistered study we see no main effects of conditionand outcome, as well as no interaction between them (all 95% credible intervals contain 0).

The N400 is an ERP component sensitive to semantic violations. Here, we probed if infants’ detection of a mismatch is subject to an altercentric bias: a bias by another person’s perspective. We hid the object in a box at the time of labelling, so its representation had to be maintained. Exp.1 established that 14-month-olds detect when an occluded object is labelled incorrectly (t(27)=2.12, p=0.043). Exp.2 included a perspective mismatch, and labelling was always incongruent for the infant, but in 50% congruent for the other. We found a reduced N400 for the congruent-for-other trials (t(33)=-2.19, p=.036), indicating an altercentric influence on infants’ semantic mismatch detection. Exp.3 probes whether infants detect any incongruence from their perspective when the labelling is always consistent for the other. Preliminary analyses (n=22 out of 34 pre-registered) show no effect of congruency from infants’ perspective, suggesting that congruency for the other may override infants’ own mismatch detection.

We test for infants’ proposed altercentric learning bias (Southgate, 2020). By tracking others’ attention, infants build their own models of the environment based on the well developed models of the adults. If true, an event will be better remembered if witnessed together with someone else than if attended alone. Eight-month-old-infants (6 conditions, n = 32/group) saw an object being hidden first in one, then in another location. At the end of each trial, one of the locations was revealed, always empty. Participants correctly remembered the location and looked longer if it was empty (conditions 1-2). When an agent attends the first hiding but not the second, infants misremember the object in the first location (condition 3), providing initial evidence for the altercentric bias. Contrary to our predictions, when the agent only attends to the last hiding location, or both, infants have no expectation of the object’s whereabouts (conditions 4-6).

As adults, we do not only choose what we prefer, but we also tend to adapt our preferences post-hoc to fit our previous choices, even when these were blind. This is thought to result from cognitive dissonance as an effort to reconcile our choices and preferences. It has been argued to rely on the self-concept, but has also been found in preschoolers and monkeys. In a preregistered study, we therefore investigated when blind choice-induced preferences emerge and whether they are related to self-concept development in the second year of life. Results in N=200 children aged 16-36 months provide strong evidence that blind-choice induced preferences develop between 1 and 2 years (BF10=12.5). Two-year-olds avoided an object that they had previously discarded in a blind choice, whereas 1-year-olds did not show any such preferences yet. Further, we found substantial evidence against a relation with measures of self-concept development (BF10=0.2).

Humans are ultra-social: they spontaneously incorporate others’ mental states into their action-planning (Kaminski et al.,2008), and altercentric: their behavior is influenced by others’ perspectives, even perspectives irrelevant to their instrumental goal (Kampis & Southgate, 2020). Recent evidence suggests that similarly to human infants, non-human great apes anticipate others’ actions based on their beliefs (Krupenye et al.,2016; Kano et al, 2019); raising the critical question whether altercentrism is uniquely human. In two experiments, we tested chimpanzees, bonobos, and orangutans in a manual search paradigm adapted from Mendes et al. (2008). These experiments replicated findings demonstrating apes’ first-person object-tracking abilities. Experiment 1 found no evidence for altercentrism: apes’ search behavior was not spontaneously modulated by another agent’s beliefs (unlike 14-month-old human infants; Kampis & Kovács, 2020). Experiment 2 found tentative evidence that apes inferred a person’s actions based on her beliefs, and adapted their own actions based on this information.