Young children’s disambiguation across the senses

• Avoidance of label overlap across the senses depended on label retrieval and discovery-based interference.
• Four-year-olds avoided label overlap more than 3-year-olds.
• Avoidance of label overlap was related to judgments of label knowledge.

When asked to find the referent of a novel label, young children usually select a novel object rather than an object that has a known label. However, children did not show this so-called disambiguation effect in a situation that required cross-modal generalization of the known label (Scofield et al., 2009). In three experiments, children learned a label for an object that they could see, but not touch, then examined two objects that they could touch, but not see. One of these tactile objects was novel, whereas the other was an exemplar of the just-trained label. On the critical trials, children were asked to decide which object was the referent of a novel label. Neither 3- nor 4-year-olds favored the novel object unless they were first asked to choose the one that was the referent of the just-trained label (both age groups) or choose the one that was the same as the visual training object (4-year-olds only). Children’s tendency to disambiguate across the senses was associated with how accurately they judged their own knowledge of object labels. These findings are consistent with the claim that the cross-modal disambiguation effect can be undermined by children’s reactions to discovering the cross-modal match and by their failure to retrieve the known label for this matching object when considering whether a novel label applies to it.