Comparing several examples of a target category leads to better generalization than single presentations. It has been argued that the more children know about the world the more likely they will discover new conceptually relevant dimensions (Gentner & Hoyos, 2017). From another point of view, the comparison process is underpinned by control processes such as executive functions (Augier & Thibaut, 2013). This study contributes to this debate with a correlational approach in which we assessed, word generalization in a comparison design, inhibition, cognitive flexibility, working memory, and vocabulary. 120 five-years-old children performed the generalization task with unfamiliar objects and 120 other five-years-old children performed the generalization task with familiar objects. Whatever the familiarity of the objects in the generalization task, we found positive and significant correlations between categorization and flexibility but not with vocabulary. It suggests that executive functions are more at play in the comparison process than word knowledge.

In this experiment, we investigated the role of dimensional distinctiveness on the generalization of novel names for unfamiliar objects. In a comparison design, we manipulated the sequence of trials difficulty, starting either with more difficult trials or with easier trials. To achieve this, we manipulated the dimensional distinctiveness of the first comparison trials and of the, later, transfer trials. Results showed that high-distinctiveness (easy) stimuli increased children’s later performance in the low-distinctiveness (difficult) condition whereas low-distinctiveness early training led to no later improvement in easier trials. Last, a correct answer for the first trial in the first learning part predicted the level of performance in the second learning part. We interpret these findings in terms of differential costs of comparison for varying levels of distinctiveness and level of abstraction from one condition to another.

In this study, 4-year-old children were tested in an object name generalization task with a stimulus comparison design. Performance in the generalization task was correlated with performance in a vocabulary test and three executive function tasks assessing inhibition, flexibility, and working memory. Correlational analyses revealed a significant association with flexibility but not with inhibition, working memory or vocabulary test. We interpret the results in terms of a capacity to flexibly generate novel dimensions rather than inhibiting irrelevant dimensions. Individual differences in working memory and inhibition did not significantly influence performance in the word extension task. Moreover, the absence of correlation with the vocabulary performance supports the idea that children did not rely on existing knowledge to find out the relevant dimension.