The research investigates whether speakers use less redundant language with more proficient interlocutors with a focus on infant-directed speech. Both the communicative efficiency framework and language development literature have suggested that speech directed toward younger infants should be more redundant compared to speech directed toward older infants. To test this, the researchers employed an innovative method by quantifying redundancy in infant-directed speech using entropy rate—an information-theoretic measure that reflects the average degree of repetitiveness in speech.
“Infant-directed speech is often described as repetitive, but existing research typically looked at different language “ingredients” separately. For example, we know that parents repeat words more often when talking to younger infants and vary their lexicon more as children grow older. In our study, we provide a novel and holistic measure of redundancy in speech directed to children to explore how redundant it is overall,” explained Dr. Shira Tal. “The idea was to see whether we use more redundant speech to younger infants, who are still in the early stages of language learning, and allow ourselves to be less redundant to the older and, hence, more proficient the children get.” The study used speech recordings of children at different ages to compare the entropy rates of speech directed to children across development. The results showed that parents indeed use less redundant speech when talking to older children, demonstrating that perceived interlocutor proficiency significantly affects redundancy.
Prof. Eitan Grossman highlighted the significance of the findings: “The developmental decrease in redundancy not only reflects a reduction in lexical repetition. We found that it is also impacted by a decrease in repetitions of multi-word sequences. This underscores the importance of larger sequences in early language learning.”
The study’s findings are particularly relevant for understanding how natural language input adjusts in response to a child’s growing linguistic abilities. By using the entropy rate as a measure, the researchers were able to capture the subtleties of how speech patterns evolve with a child’s development, offering valuable insights for both theoretical and practical applications in language acquisition research.
Prof. Inbal Arnon emphasized the theoretical implications: “These findings demonstrate that babies are exposed to language that is adapted to the way we perceive them, much like the way we adapt our language to different people we talk to. It also adds a piece to the puzzle of unraveling the special characteristics of the input from which children learn their first language.”
This study represents a significant advancement in the understanding of language acquisition and the adaptive nature of human communication.
The research paper titled “Infant-directed speech becomes less redundant as infants grow: Implications for language learning” is now available in Cognition and can be accessed here.
Researchers:
Shira Tal1,2, Eitan Grossman3, Inbal Arnon4
Institutions:
- Department of Linguistics and English Language, University of Edinburgh
- Department of Cognitive Sciences, The Hebrew University of Jerusalem
- Department of Linguistics, The Hebrew University of Jerusalem
- Department of Psychology, The Hebrew University of Jerusalem