TL;DR
This study investigates how children's intonation patterns develop to reflect narrative structure, finding that prosodic cues to discourse organization become more adult-like between ages 5 and 7. The research highlights that while intonation complexity increases, the steepness of pitch contours within utterances remains surprisingly consistent.
What Problem Does This Paper Address?
Adults use intonation to signal the hierarchical structure of their narratives, marking how ideas are organized and connected. Children, however, tend to produce narratives with a flatter structure, both linguistically and prosodically. While we know children's sentence complexity and discourse marking develop significantly, their prosodic development in signaling narrative structure is less understood. This paper aims to fill that gap by longitudinally tracking how children's intonation patterns evolve over time to reflect discourse organization, specifically looking for parallels with the development of linguistic complexity.
Methods
The researchers collected spontaneous narratives from 60 children over three years, starting when they were between 5 and 7 years old. They used the "Frog Story" picture books to elicit these narratives, recording them with a wireless microphone. The audio was then processed to extract individual utterances, and fundamental frequency (F0) was tracked at 10-millisecond intervals.
Key prosodic measures were derived:
- Slope Steepness and Linearity: The steepness and linearity of the F0 contour within each utterance were measured. These are known correlates of syntactic and semantic complexity in adult speech.
- Slope Direction: Pitch contours were categorized as falling or rising.
- Inter-utterance Continuity: The degree of continuity between the F0 contours of adjacent utterances was assessed, looking for "pitch resets" that signal discourse boundaries.
Statistical analyses, including linear mixed-effects models and logistic regression, were used to examine the effects of age and study year on these prosodic measures, controlling for individual speaker variability.
Datasets
The study utilized a longitudinal corpus of spontaneous narratives from 60 typically developing children. These narratives were collected annually over a three-year period. The children ranged in age from 5;2 to 7;10 at the start of the study. No new benchmark datasets were introduced.
Key Insights and Ideas
- Prosody Catches Up to Language: Children's intonation patterns gradually become more adult-like in their ability to signal narrative structure, mirroring the development seen in their linguistic complexity.
- The Critical Window: Ages 5-7: The most significant developmental changes in prosodic features related to discourse organization occur between the ages of 5 and 7. After age 7, these patterns appear to stabilize.
- Intonation's "Flatness" Persists: While other prosodic features evolve, the steepness of F0 contours within individual utterances surprisingly did not change systematically with age or development. This suggests that this specific measure might be more tied to utterance length or speech preparation than to discourse structure itself.
- Continuity vs. Discontinuity: Younger children (around age 5) produced fewer continuous intonation contours between utterances compared to older children. This finding aligns with previous research on discourse continuation but is somewhat counterintuitive given the "event-chaining" narrative style often observed in young children.
Key Results
- Slope Linearity: Linearity of F0 contours decreased significantly from Year 1 to Year 3 for 5-year-olds, suggesting increased pitch variation within utterances as they developed. This trend slowed down after age 7.
- Slope Direction: The proportion of falling pitch contours increased with age, with a noticeable shift between ages 5 and 9. This aligns with adult narrative patterns.
- Inter-utterance Continuity: A significant effect of study year was found, indicating that children produced more continuous contours between utterances over time. However, age itself did not have a significant effect, likely due to the plateauing of this measure after age 7.
- Slope Steepness: This measure showed significant effects of age and study year, but the patterns were not systematic across groups, leading the authors to question its direct link to discourse structure in this context.
Limitations and Future Work
The authors acknowledge that the systematic change in slope steepness was not observed, and they speculate that this measure might be more influenced by utterance length or the difference between spontaneous and prepared speech. They suggest that further investigation is needed to understand why this specific feature behaved differently. The study also notes that developmental changes in discourse prosody appear to plateau around age 7, suggesting future research could focus on the period beyond this age to understand continued refinement or maintenance of these skills.
Summary updated