Effects of perceived age on prosodic judgments of child and adult speech.

Abstract

Speakers’ ability to effectively use prosody (e.g., pitch and timing) to convey information improves over developmental time. This research investigated effects of perceived talker age on subjective prosody judgments, which are often central to developmental prosody research. Speech was subjected to modification in which formants and fundamental frequency were lowered, for 5‐year‐old children, or raised, for adult caregivers. The manipulations were expected to affect talkers’ perceived ages, while leaving critical acoustic‐prosodic correlates intact (e.g., fundamental frequency contour, speech rate). In experiment 1, eight analysts used a prosodic labeling system to indicate phrasal boundaries and disfluencies in unaltered and spectrally modified speech. In experiment 2, listeners rated unaltered and spectrally modified speech regarding rate, fluency, and estimated age. Preliminary data revealed that the manipulation caused children and adults to be judged older or younger, respectively, as expected. Importantly, the manipulation affected ratings of rate and fluency (experiment 2), but not rates with which analysts identified phrasal boundaries or disfluencies (experiment 1). Results suggest that prosodic labeling systems may provide more reliable means of studying prosody than ratings, due to reduced effects of confounding speaker characteristics (e.g., perceived age) on judgments, a finding with implications for studying speech development. [Work funded by NIH‐NICHD grant 1R01HD061458‐01.]