In their first years of life infants face a large variety of new situations, and in every new situation they are challenged to grasp the structure of never experienced scenes in order to learn and be able to form future expectations. The ease and speed with which infants figure out a new problem are indicative of a general-purpose cognitive machinery that guarantees a large flexibility unexplained by domain-specific adaptations. We have proposed that an inferential system operating according to the rules of deductive logic might be a good candidate for such a machinery.
Which kind of logical concepts are available to infants, which kind of logical inferences can they perform? In order to answer and systematically study these questions, Prelog was structured around two main axes. One of the axes focused on the i.) identification of logical primitives available early in development (e.g. disjunctive relations, negation etc.), ii.) modal concepts (hypothesizing that the representation of multiple, mutually exclusive possibilities is central to learning and problem solving already in the first years of life), iii.) the origins of compositional thought in ontogeny. Our findings documented that infants do engage in logical reasoning possibly relying on the powerful inferential device characteristic to the human mind. Our studies revealed the early presence of a reasoning system with the potential to generate conclusions that can be integrated with other processes. For instance, preverbal disjunctive inferences generate conclusions that are sufficiently solid to contribute to a high inferential productivity by interfacing with other computational systems dedicated to distinct cognitive domains (e.g. the physical domain and the social domain). For instance, in Cesana-Arlotti, et al, 2020 14-month-old infants successfully used the outcome of the disjunctive inference for identifying the preference of an agent.
In a different series of studies, along the second main axis of Prelog, we explored the representational preconditions of quantification in infants and the quantificational performance of preschool children. In the latter studies we have targeted preschoolers’ interpretations of generic statements, and these differ from quantifiers.
Methodologically, these studies involved rather diverse paradigms: some of them required the development of specific testing paradigms (e.g. gaze contingent interactivity controlled by an eye tracker) some others used classical methods (e.g. looking-time based analysis, or behavioral responses). The findings were presented to the scientific community at various conferences, some were already published (e.g. Cesana-Arlotti et al, 2018; Science; Cesana-Arlotti, Kovács & Téglás, 2020, Nature Communications) and we have engaged in communication to the general public as well. Contrary to the consolidated position according to which logical reasoning follows a prolonged maturation, our findings suggest that, already early in life, preverbal forms of logical reasoning function as a reliable source of evidence that can support learning by offering a logical route for knowledge acquisition.
Overall, our results together document the availability of a small, but efficient set of logical representations in infancy, before language is acquired. Research targeting questions along these lines helps us better understand the nature of the cognitive mechanisms that are at the origin of human rationality supporting infants and children in coping with problems never experienced before and have a crucial role in the acquisition of new information.
