Voice onset time in consonant cluster production by children and adults

Voice onset time (VOT) characteristics of stops in initial clusters in American English words produced by children and adults were studied. Words were spoken in isolation and in sentence context by eleven three-and four-year-old children, and by a male and female adult. Spectrograms were made and VOT duration measurements taken. Three experienced listeners transcribed the isolated words and sentences. Analyses showed that overall timing characteristics were similar for children and adults. Speakers differed in their [±voice] boundary and there was no absolute time distinction between [±voice] stops; [+voice] stops showed less variability than [−voice]. VOT generally increased from labial to dental to velar clusters, and was shorter in sentence context and longer in clusters than in singletons. Children's VOT averages were generally, but not significantly, longer than adults' in all contexts, and coarticulation constraints affected the accuracy with which children produced the stop and liquid portion of a particular cluster.