The link between language and cognition is evident in the first months of life. As early as three months of age, a parent labeling an object that the infant is looking at influences the infants’ thoughts about those objects (Ferry et al., 2010). Prior research investigating this link between language and cognition has found successful categorization with 6- and 12-month-olds using face-to-face methodologies via preferential-looking paradigms (Fulkerson & Waxman, 2007). Due to COVID, we sought to validate an online counterpart of the categorization task. In Experiment 1, we found that language labels facilitated object categorization for 10- to 12-month-old infants. In contrast, a control condition that presented the same labeling phrases played in reverse did not facilitate categorization. In Experiment 2, we found that labels did not facilitate object categorization among 6-month-old infants. In sum, the advantage of online infant testing may be restricted to older ages.

There are some puzzling findings concerning the early development of relational cognition. There is evidence that infants can carry out analogical generalization and form abstract relations such as same (X,X) (Anderson et al., 2018; Ferry et al., 2015). Yet, 3.5-year-olds have difficulty with these relations (Christie & Gentner, 2014; Hochmann et al., 2017; Walker & Gopnik, 2016). Further, Walker et al. (2016) find a decline from 1.5 years to 3.5 years on another same-different task (see also Carstensen et al., 2019) while prior studies have found gains in relational reasoning during this period (e.g., Gentner & Rattermann, 1991). To reconcile these findings, we are re-examining this developmental trajectory with a new task. Children must point one direction for same pairs and another for different pairs. The results revealed that 4-year-olds learned and generalized same-different relations after four exemplars and are discussed in the context of a competence-performance distinction.

The ability to generalize previous knowledge to new contexts is a key aspect of human cognition and relational learning. A well-known learning maxim is that breadth of training predicts breadth of transfer. When examples vary in their surface features, this provides evidence that only the common relational structure is relevant. However, there is some evidence suggesting that the above maxim may not apply well in early relational learning. Here, we present a further test whether the maxim holds for young infants. We find that 3-month-old infants perform better with a narrow, perceptually similar training set than with a broad, perceptually variable set. We argue that lower-level perceptual similarities can prompt comparison processes that facilitate relational abstraction. These findings cohere with research arguing relational learning depends on relational alignment.

Relational ability—the ability to compare situations or ideas and discover common relations – is a key process in higher-order cognition that underlies transfer in learning and creative problem solving. For this reason, it has generated intense interest both among developmentalist and in cross-species comparative studies. The gold standard for evaluating relational ability is the Relational-Match-to-Sample (RMTS) task (Premack, 1983). Current work in cognitive development has produced inconsistent results as to when children are able to pass the RMTS, with Christie and Gentner (2014) finding earlier success than Hochmann et al. (2017) and Kroupin and Carey (2022). In this research, we attempt to resolve this issue. We first describe two studies that bear out and extend Christie and Gentner’s (2014) findings. We then discuss factors that might explain the discrepancy between the findings.

The link between language and thought lies at the core of what it means to be human. This link is evident in the first months of life. As early as three months of age, a parent labeling an object that the infant is looking at influences how that infant thinks about those objects (Ferry et al., 2010). Prior research used face-to-face methodologies. Considering the constraints posed by the COVID-19 pandemic, we sought to validate an online categorization task. The results show that language labels facilitated category formation for 9- to 12-month-old infants. In contrast, a control condition that presented the same labeling phrases in reverse did not facilitate categorization. These findings suggest that infants can perform categorization using an online protocol such as Zoom and this paradigm provides a promising new avenue for studying the intersection between language and cognition.

Violating infants expectations about solid objects (e.g., a ball passing through a wall) leads to increased exploration andlearning about the objects properties (Stahl & Feigenson, 2015). How limited is this type of learning? Infants can anticipatehow non-solid substances behave and interact (Hespos et al., 2009; 2016), but the non-cohesive nature of substances meansthat they have less predictable shapes and boundaries. Across four trials, we presented 12- to 14-month-olds with itemsthat looked solid or liquid. For half the trials, the items behavior was consistent with its appearance, so, for example, itlooked solid and remained cohesive. For the other half, the behavior was inconsistent. Infants spent significantly moretime exploring the inconsistent items, whether solid or non-solid, F(1, 57) = 24.00, p = .001, p =.29. These results suggestthat infants preference for learning from violations might be a general mechanism responsible for new knowledge.

What makes humans smart? We think it has to do with our analogical ability—our ability to perceive common relational patterns between different objects, events, or ideas. It is a cornerstone of our higher-order reasoning ability and may have its origins in a relational processing mechanism that allows us to abstract relations using comparison. In this symposium, we present research on the nature of our relational learning ability. We show that findings from infants, preschoolers, elementary school students, and adults reveal common signatures across a wide variety of contexts. These studies provide evidence about how to improve learning of abstract relational concepts and provide insight about how to build better AI models.

Researchers have recently found that 3-month-old infants are capable of using analogical abstraction to learn the same or different relation, given the right conditions (Anderson et al. 2018). Surprisingly, seeing fewer distinct examples led to more successful learning than seeing more distinct examples. This runs contrary to the prediction of standard learning theories, which hold that a wider range of examples leads to better generalization and transfer, but is compatible with other findings in infant research (Casasola 2005; Maguire et al. 2008). Anderson et al. (2018) propose that this is due to interactions between encoding and analogical learning. This paper explores that proposal through the lens of cognitive simulation, using automatically encoded visual stimuli and a cognitive model of analogical learning. The simulation results are compatible with the original findings, thereby providing evidence for this explanation. The assumptions underlying the simulation are delineated and some alternatives are discussed.

This research tests whether analogical processing ability is present in 3-month-old infants. Infants are habituated to a series of analogous pairs, instantiating either same (e.g., AA, BB, etc.) or different (e.g., AB, CD, etc.), and then tested with further exemplars of the relations. If they can distinguish the familiar relation from the novel relation, even with new objects, this is evidence that for analogical abstraction across the study pairs. In Experiment 1, we did not find evidence of analogical abstraction when 3-month-olds were habituated to six pairs instantiating the relation. However, in Experiment 2, infants showed evidence of analogical abstraction after habituation to two alternating pairs (e.g., AA, BB, AA, BB...). Further, as with older groups, rendering individual objects salient disrupted relational learning. These results demonstrate that 3-month-old infants are capable of analogical comparison and abstraction. Our findings also place limits on the conditions under which these processes are likely to occur. We discuss implications for theories of relational learning.