Search results

Query: faculty: "FNWI" and publication year: "2011"

AuthorJosefine Karlsson
TitleSelf-monitoring in early language development
SupervisorsC.C. Levelt, P. Boersma
Year2011
FacultyFaculty of Science
ProgrammeFNWI MSc Brain and Cognitive Sciences
Keywords-
AbstractIn early language acquisition, both the child‟s phonological perception and production are at first prone to errors. The lexical representation, underlying both perception and production, is thought to become more detailed over time. However, the lexical representation is not necessarily the underlying cause of speech errors in children; these errors could also occur later in de
production process. One purpose of the present study was to investigate whether the often claimed discrepancy between the detailed storage of a word and its reduced production can be demonstrated in a production experiment.

Research has shown that children as young as two years old can be encouraged to repeat a production. In the second production, they can change and improve information or grammatical structure compared to the initial production. In addition, spontaneous self-repairs in children reveal that they can make use of self-monitoring on a phonological level. We wanted to investigate whether or not toddlers can be induced to make self-repairs, leading to changes or improvements on a phonological level. We targeted the children‟s phonological knowledge and self-monitoring by prompting self-repairs in words with onset consonant clusters, which are typically difficult at this age.

In our first experiment we recorded word pairs from Dutch 24-month olds and 30-month olds. Thereof, we conducted an error analysis of the onset clusters by comparing the distribution of correct and
incorrect segments between the first (P1) and the second production (P2). We found a significant difference between P1 and P2 in target clusters starting with an /s/ in children aged 24 months.
These results provide evidence that the initial erroneous production did not depend on the underlying representation entirely, but that the error sometimes lies in the production process.
Additionally, the results show that children can be induced to use self- monitoring on a phonological level.

In our second experiment we presented the recorded child speech to Dutch adults. Their task was to decide which production in each word pair was more adult-like. The results showed that there was no
overall preference for either P1 or P2. We assumed that the adults used multiple acoustic cues in their assessment. Indeed, the duration of the entire word and of the stressed vowel both influenced the adults‟ decision and it was not possible to tell these factors apart. Additionally, the adults noticed a difference between P1 and P2 in target clusters starting with an /s/. There was a significant positive correlation between the adults preference for P2 and
target clusters starting with an /s/. Hence, the results from the first and the second experiments corroborate.
LanguageEngels
Document type scriptie master
Download paper