Marlene M. Meyer, Marina Proft and Hannes Rakoczy
Affiliation: University of Göttingen
Category: Psychology
Keywords: metacognition, uncertainty, modal cognition, possibility, social context, curiosity
Date: Tuesday 2nd of September
Time: 15:00
Location: Room 232 (232)
View the full session: Metacognition
Metacognition and modal cognition are fundamental for higher cognition, learning and critical thinking. Prior developmental research has yielded converging evidence that children master explicit forms of both cognitive capacities surprisingly late, not before they reach age 4 to 6. In the case of metacognition, this protracted development reveals itself in younger children’s failure to acknowledge their own uncertainty in light of several possibilities. For instance, when asked in so-called partial ignorance tasks whether they know which one of several previously shown items is hidden inside a container, children younger than 6 years of age incorrectly claim to know (e.g., Rohwer et al., 2012; Kloo et al., 2017). Similarly, a growing body of literature on the development of modal thought reveals that young children fail to consider alternative possibilities. For instance, when asked to make a rational choice between one certain and two uncertain options, only children older than 4 years of age demonstrate competent modal reasoning (e.g., in the so-called 3-cups-task; Mody & Carey, 2016; Leahy & Carey, 2020; Leahy et al., 2022). In both cases, earlier and implicit forms of the respective cognitive capacity can be acknowledged (e.g., in dual-system accounts). But in light of the robust empirical data on children’s performance in direct tasks, the standard in the field has it that explicit metacognition and modal thought emerge rather late, well into children’s school age years (e.g., Perner, 2012; Proust, 2019; Goupil & Proust, 2023; Leahy & Carey, 2020).
However, the underlying assumption that metacognitive and modal reasoning abilities are primarily useful for individual reflection and decision-making has been challenged: Recent theoretical accounts emphasize the social origins and functions of higher human cognition, in particular metacognition (Shea et al., 2014; Frith, 2012; Nagel, 2015). This view yields interesting predictions for the study of cognitive development. Prior research has mainly focused on young children’s competencies in individual settings. Both the partial ignorance task and 3-cups-task have studied children’s metacognitive judgments and modal reasoning in individual, standardized experimenter-child environments, confronting the participants with academic test questions. However, the social framework of metacognition would predict that explicit metacognitive abilities should be most relevant and thus most easily (and earlier) detectable in socio-communicative settings.
Here, we build on recent empirical approaches that have tested these predictions and revealed a striking pattern: On the one hand, testing children in a socio-communicative context enhanced their ability to provide accurate explicit metacognitive judgments. Already 3-year-old children acknowledged their uncertainty in a social partial ignorance task, i.e., they correctly stated that they did not know which of several items was hidden inside a container when a cooperative partner relied on their advice (Meyer et al., submitted). On the other hand, no such social boost was found for children’s reasoning about uncertainty in the 3-cups-task: Comparing children’s performance between a standard individual and a social task version revealed no significant differences, as 3-year-old children failed to make rational choices despite the socio-communicative test context (Meyer et al., in prep).
These findings thus present a mixed picture: Strengthening the social nature and context of a task does reveal earlier competence in explicit metacognition; but no analogous effect was found for modal cognition. Why is this so? Theoretically, one potential explanation is that this reflects deep differences in the cognitive competencies: Metacognition but not modal cognition might become primarily functional in socio-communicative contexts; thus, a social approach would yield positive results only in the respective domain. Alternatively, the diverging results may merely reflect superficial methodological differences: First, the content of the uncertainty judgments in question differs, as the test questions concern “what is X” (partial ignorance task) vs. “where is X” (3-cups-task). Second, the complexity of measures and thus structural demands for competently solving the task differ. In the partial ignorance task, children are confronted with a local uncertainty judgment that requires them to represent uncertainty about the content of a single unknown box. However, succeeding in the 3-cups-task requires making a contrastive choice: Children have to represent the contrast in (un-)certainty between a known box and two unknown boxes, and reach an integrative decision of choosing the certain option.
The present project aims to disentangle these alternative explanations. To this end, we systematically tested these structural differences (content of uncertainty judgment and complexity of measure) with 3-year-old children (N = 136) in a mixed 2x2 design: Children were randomly assigned to either a social partial ignorance (“what” uncertainty) or a social 3-cups-task (“where” uncertainty). Both tasks included explicit metacognitive test questions (potentially less complex local uncertainty judgment measure) and required children to decide between certain and uncertain options (potentially more complex contrastive choice measure). The overall joint goal of children and their cooperative partner was to collect farm animals (see Figure 1 for details on the experimental design and procedure). Children participated in two test trials per measure and two control trials (for the local uncertainty judgment only), in counterbalanced order (total of six trials).
Preliminary analyses (on data of n = 122 children) suggest that either factor alone cannot fully explain children’s diverging performances across previous metacognitive and modal reasoning tasks. First, children provided more accurate local uncertainty judgments in the test conditions of the partial ignorance task (“what” uncertainty) as compared to the 3-cups-task (“where” uncertainty), but chose the certain option significantly above chance level only in the latter. However, children’s overall performance in their local uncertainty judgments was comparably lower than expected from previous studies adopting a socio-communicative context (Meyer et al., submitted; Meyer et al., in prep). This pattern warrants further investigation of a potential interaction between uncertainty judgment and measure.
Altogether, the current project raises important further questions that will be discussed in the talk: What other factors may play into children’s decision-making progress in uncertain situations? One promising candidate is curiosity: Understood as a metacognitive feeling (Goupil & Proust, 2023), children’s inherent drive to explore and seek information may interfere with their “rational” choices in paradigms like the 3-cups-task.