Predicting the Birth of a Spoken Word

Roy, B. C., Frank, M. C., DeCamp, P., Miller, M., & Roy, D. (2015).Predicting the birth of a spoken word. Proceedings of the NationalAcademy of Sciences, 112(41), 12663-12668.

Early on, young children are surrounded by words, but they do not yet know what these words mean, or how to use the words themselves. Despite not yet knowing the meanings of words, the experiences and interactions that surround the usage of these words may drive the emergence of word knowledge. Laboratory studies, however, are not ideal for examining real-world experiences. In this novel study, the researchers examined the combination of factors that were most predictive of word knowledge in a naturalistic environment.

To collect these data, the researchers conducted a large-scale, longitudinal observation of a single, typically-developing child. They collected audio and video recording from all rooms of the child’s house from birth to age 3 years, adding up to 200,000 hours of data. They analyzed the space, time, and language context for each word spoken to the child. For instance, the word “breakfast” was most often spoken in the kitchen, between 8:00am to 10:00am, and was often spoken with other words such as “chew”, “yum”, and the child’s name. The main outcome measure was the age at which individual words were spoken by the child. They found that words produced in distinct spatial, temporal, and linguistic contexts were produced earlier by the child, suggesting they were easier to learn. These three factors were also strongly correlated with one another, and were stronger predictors of the age at which the child first produced a word than how frequently the word was said to the child.

Taken together, this rich observational study supports the notion that words produced in more distinctive contexts are learned earlier. It is important to note that because these data came from one child, the generalizability of these findings needs to be established. Nevertheless, this study demonstrates how multimodal data collection is valuable in understanding language acquisition. The causal structure of language acquisition is a complex puzzle, but these data provide some evidence for revealing part of this process.

Blogger: Nicolette Noonan is a PhD student in Psychology, supervised by Drs. Lisa Archibald and Marc Joanisse.