Toddlers’ looking behaviours during referent selection and relationships with immediate and delayed retention

• Does referent selection (RS) involve a process-of-elimination or a novelty bias? • Converging gaze and point data indicated that toddlers use a process-of-elimination. • Assessed the role of target focus during RS and ostensive naming (ON) on retention. • With ostensive naming, greater target focus predicted better immediate retention. • Without ostensive naming, greater target focus predicted better delayed retention. The current study investigates whether children’s attempts to solve referential ambiguity is best explained as a process-of-elimination or a novelty bias. We measured 2.5-year-old children’s pointing and eye movements during referent selection trials and assessed whether this changes across repeated exposures. We also tested children’s retention of novel words and how much focusing on novel targets during referent selection supports immediate and delayed retention as well as the effect of hearing the words ostensively named after referent selection. Time course analyses of children’s looking during referent selection indicated that soon after noun onsets, in familiar target trials there was a greater focus on targets relative to chance, but in novel target trials, children focussed on targets less than chance, suggesting an initial focus on competitors. Children also took longer to focus on and point to novel compared to familiar targets. Thus, this converging evidence suggests referent selection is best described as a process-of-elimination. Ostensive naming also led to faster pointing at novel targets in subsequent trials and better delayed retention than the non-ostensive condition. In addition, a greater focus on novel targets during referent selection was associated with better immediate retention for the ostensive naming condition, but better delayed retention for the non-ostensive condition. Therefore, a focus on novelty may supplement weaker encoding, facilitating later retention.