Voicing of obstruent clusters in Dutch past tense acquisition

Language acquisition theory predicts young children to show a preference for unmarked structures. Voiced obstruents as well as obstruent clusters are marked contexts. What follows is that children should initially prefer voiceless obstruent clusters over voiced obstruent clusters. To provide empirical evidence for this claim, an experiment was designed that presented Dutch participants with past tenses, because Dutch verbs with voiceless or voiced obstruent-final verb stems yield past tenses with voiceless or voiced obstruent clusters, respectively. The present experiment was run with six-year-old and nine-year-old children as well as adults. The results show a clear preference for the voiced past tense suffix [-dǝ] by the nine-year-old group against a preference for the voiceless past tense suffix [-tǝ] by the six-year-old group and the adults. Other important findings are that the orthographical representation results in less incorrect inflections and that phonotactically illegitimate pseudo-verbs trigger more incorrect inflections