Preschoolers’ inferences about pointers and labelers: The modality matters

• We investigated the magnitude of children’s inferences about pointers and speakers.
• Children see an accurate and inaccurate actor who either speak or point.
• Children inferred that accurate pointers would point, not speak, accurately later.
• Children inferred that accurate speakers would also point accurately later.
• Children make more limited inferences about those who point than those who speak.

Preschoolers seek out and endorse new labels from informants who have previously provided accurate labels over those who have previously provided inaccurate labels (e.g., Koenig et al., 2004). In Study 1, we show that 4-year-olds also prefer a previously accurate pointer over a previously inaccurate one as a source of information about newly hidden objects. In Study 2, we show that they do not expect that a previously accurate pointer will necessarily be a good labeler, though they do expect a previously accurate labeler will be a good pointer. This asymmetry suggests that the scope of inferences children draw about the knowledge of informants can be influenced by the modality in which they communicate. By sharing semantic information (e.g., object names), labelers demonstrate generalizable knowledge; by sharing episodic information (e.g., an object’s location), pointers demonstrate more limited knowledge.