Serena Lecce

Affective word meanings are often treated as stable lexical properties, although their organization may change across development. Using representational content analysis, we examined how semantic and form-based lexical information predicts valence and arousal judgments from childhood to adulthood. Across two experiments, Spanish affective ratings from children aged 7, 9, 11, and 13 years were modeled within and across age groups and compared with adult judgments. Semantic representations were the strongest predictors of both dimensions, but valence and arousal showed distinct developmental patterns. Valence remained comparatively stable across childhood and showed strong correspondence with adult judgments already at age 7. Arousal, despite being systematically predictable within each child group, showed weaker cross-age generalization, with semantic transfer declining as developmental distance increased. Moreover, semantic information predicted word-specific age differences more accurately for arousal than for valence. Child-adult correspondence in arousal shifted from negative at age 7 to positive at ages 11 and 13, suggesting increasing similarity in lexical-affective mappings across development. These findings indicate that developmental variation in arousal does not reflect an absence of systematic organization in younger children but rather age-related differences in how lexical meaning relates to activation. Overall, affective lexical development combines a relatively stable evaluative structure with a more flexible organization of arousal.

Metaphor is a powerful tool of verbal reasoning that helps individuals understand complex phenomena and plays a central role in climate change communication. However, children’s still-developing metaphor comprehension abilities may limit the effectiveness of figurative environmental messages. To address this issue, we extended the MetaCom training by developing MetaCom-Green, a novel intervention combining metaphor training with climate change education. Each session revolved around a metaphor describing an environmental phenomenon (e.g., The ocean is a soup or Animals are eating bites of fire). A group of fourth graders were assigned either to MetaCom-Green or to a control condition receiving an equivalent, metaphor-free environmental program. Pre-training assessments indicated that the two groups were comparable on linguistic, socio-cognitive, and most environmental measures. Post-training, children in the MetaCom-Green group showed greater improvements not only in metaphor comprehension, but also in climate change knowledge. Moreover, they developed richer conceptualizations of climate change phenomena and showed changes in environmental attitudes, with greater hope, lower despair about the future, and more pro-environmental behavior. These findings suggest that metaphor constitutes a powerful “learning-by-thinking” tool, fostering reasoning, conceptual reframing, and possibly embodied experiences that may enhance children’s engagement with environmental challenges.

Psychological measures are commonly reduced to aggregate scores, such as questionnaire totals, accuracy scores, or human-annotated measures. Although useful, these scores may hide information about the configuration of participants’ responses. The present study took mentalization (i.e., the ability to infer others’ mental states) as a working example to investigate how this domain can be characterized through vector-based response representations in addition to classical scoring. One hundred and sixty-four university participants completed the Interpersonal Reactivity Index (IRI), the Reading the Mind in the Eyes Test (RMET), and the Triangles task. For each task, we constructed participant-by-participant distance matrices based on classical scores and on vector representations: IRI item-response profiles, RMET response-choice patterns, and sentence embeddings of open-ended Triangles ToM descriptions. Classical score correlations showed limited convergence across tasks, with only RMET accuracy and Triangles score showing a small positive association. Representational similarity analyses showed that vector-based representations reliably converged with classical score distances within each task, although the associations were modest. Cross-task analyses further indicated that score-based and vector-based representations produced partly different prediction patterns. Overall, these findings suggest that classical mentalization measures organize individual differences differently depending on whether responses are represented as aggregate scores or full response configurations, with implications for how psychologists conceptualize and validate social-cognitive assessment.Keywords: mentalization; representational similarity analysis; individual differences.

Background: RIDInet-Reading Trainer 2 (RT-2) is a web-platform for the remote treatment of developmental dyslexia (DD) which has been shown to improve reading performance. However, no previous studies have investigated stability in reading improvement after RT-2 training and the influence of a previous diagnosis of developmental language disorder (DLD) and of participants’ age on stability. Objectives: In a sample of 52 Italian-speaking children with DD who participated in a 3-month home-based treatment with RT-2, we aimed (1) to assess the stability in reading improvement after RT-2 training at a 3-month follow-up and the potential moderating role of DLD and age; and (2) to evaluate the impact of RT-2 training in reading comprehension. Results: By implementing linear mixed model analysis, our findings confirmed reading improvement after RT-2 training in word and text reading in DD. Moreover, we observed an overall stability in single-word and text reading speed performances after three months, regardless of the diagnosis of DLD and the age of the participants. Conversely, accuracy showed an overall stability for single-word reading, while it was significantly stable only in the younger participants in text reading. The improvement was educationally relevant as it impacted reading comprehension. Conclusions: The current study supports the use of remotely delivered DD interventions among school-aged children.

Children’s theory of mind (ToM) is critical for academic success, yet its associations with reading comprehension and mathematics remain unclear. This pre-registered meta-analysis synthesizes findings from 53 studies, including 12,347 participants aged 2.5 – 17 years. Results revealed a significant overall association between ToM and academic achievement ( r = .32), which: (i) did not differ across domains (i.e., reading, mathematics, general achievement); (ii) was independent of verbal ability, non-verbal ability, executive function, and socioeconomic status; and (iii) appeared bidirectional. These longitudinal findings should be interpreted with caution, as they are largely based on zero-order associations without baseline controls. Sample (i.e., age, gender and country) and task characteristics (i.e., level of mentalizing, ToM stimulus type, mental state category, response type, and sensitivity to individual differences) did not moderate the relation between ToM and academic achievement, although weaker effects emerged in European samples. Our results demonstrate a consistent association between ToM and academic achievement underscoring the potential contribution of social–cognitive understanding to children’s academic development. Future research should clarify the mechanisms underlying this association, its domain specificity, and the direction of causal influence.

This longitudinal study investigates theory of mind (ToM) development in middle childhood and its demographic and linguistic predictors. In this study, 194 Italian primary school children (Mage = 9.4 years; 47% girls; 28% with one or both Italian parents) were assessed four times over 2 years using a multi-modal ToM battery. Results confirmed longitudinal measurement invariance of the ToM battery. Latent growth modeling showed ToM growth from Time 1 to Time 3, followed by a plateau. Analyses revealed significant inter-individual variability at Time 1 but no variance in the rate of development. Higher verbal ability and having siblings predicted higher initial ToM, while non-native speakers began lower but showed greater gains. Findings extend existing accounts on ToM development.

Attempts to describe the early acquisition of metaphor skills led to inconsistent results. We address this issue by examining developmental trajectories across metaphor types and along the literal–figurative and physical–mental dimensions. To this end, we implemented the Physical and Mental Metaphors task in a multiple-choice format (with literal, unrelated, figurative physical, and figurative mental interpretation options) and administered it to 176 typically developing children aged 4-6, alongside linguistic, cognitive, and Theory of Mind tests. We observed a curvilinear trend, with age 6 marking a significant improvement in physical, but not mental, metaphors, driven by a drop in literal interpretations. For mental metaphors, children at this age still prefer physical interpretations. Vocabulary supported physical metaphors, while Theory of Mind aided mental metaphor comprehension. Results highlight that overcoming literal and physical preferences are two important stages in metaphor development, interacting with verbal and socio-cognitive factors.

Understanding how Theory of Mind (ToM), the ability to attribute mental states to others, develops and differs across individuals requires robust and easily applicable measures. Tasks involving open ended verbal responses such as the Triangles Task offer valid indicators of ToM, but their manual scoring is time-consuming and demands substantial training. In this study, we test the validity and reliability of TrAIngles, an automatic scoring software for the Triangles Task developed through fine-tuning of large language models (LLMs). TrAIngles was trained on a dataset of N ~ 11,000 responses collected from 581 Italian-speaking participants aged between 8 and 81 years. LLM-based ToM scores showed high accuracy and reliability metrics (Cohen’s kappa, ICC, Krippendorff’s alpha, Spearman’s rho) comparable to human ratings and consistently exceeding commonly accepted thresholds for inter-rater agreement (i.e., > .80). Performance remained robust across age groups (children, adolescents, young adults, adults, and older adults). Our findings demonstrate that reliable and valid ToM indexes can be obtained from the Triangles Task using this new automated scoring system based on LLMs. TrAIngles thus provides an innovative tool for lifespan ToM assessments relying on open-ended responses, highlighting the potential of training automated scoring systems on archival, human-rated datasets. TrAIngles is openly available for use both as code and via a user-friendly graphical interface.