Proceedings of the Annual Meeting of the Cognitive Science Society

Children learn language from the speech in their home envi-ronment. Recent work shows that more infant-directed speech(IDS) leads to stronger lexical development. But what makesIDS a particularly useful learning signal? Here, we expandon an attention-based account first proposed by R ̈as ̈anen etal. (2018): that prosodic modifications make IDS less pre-dictable, and thus more interesting. First, we reproduce thecritical finding from R ̈as ̈anen et al.: that lab-recorded IDS pitchis less predictable compared to adult-directed speech (ADS).Next, we show that this result generalizes to the home lan-guage environment, finding that IDS in daylong recordings isalso less predictable than ADS but that this pattern is muchless robust than for IDS recorded in the lab. These results linkexperimental work on attention and prosodic modifications ofIDS to real-world language-learning environments, highlight-ing some challenges of scaling up analyses of IDS to largerdatasets that better capture children’s actual input.