Jam vs. Jelly: How Children Reason Analogically with Semantically Similar Words

Children’s ability to name an object by multiple semantically similar labels (i.e. synonyms) has been documented by the age of three; however, the ability to reason using semantically similar labels remains sparsely investigated. The present study expands our understanding of children’s ability to utilize semantically similar labels in two relational reasoning tasks – a semantic completion task (Experiment 1) and a semantic substitution task (Experiment 2). In Experiment 1, participants were presented a base pair of related words (e.g., castle:rock), then given a partially completed target word-pair (castle: ?) to complete with a label that made the target word-pair relationally identical to the base word-pair (e.g., stone). Additional response options included a label that was thematically related (king) and an unrelated label (milk). In Experiment 2, the same semantic relationships were explored by presenting the labels in stories with prompt questions. Results indicated that four-year-olds exhibited difficulty with semantically similar labels in the semantic completion but not in the semantic substitution task. Theoretical implications of these results are discussed.