Tracking Transitional Probabilities in Infancy : Word Segmentation from Dual Language Input
MetadataShow full item record
During language acquisition, infants use cognitive tools to dissect language: a heightened sensitivity to phonetic and prosodic cues and statistical learning. Research has shown that statistical learning paired with the aforementioned acoustic cues aids in successful word segmentation. Infants reared in a bilingual environment exhibit enhanced statistical learning abilities to efficiently employ and track sequential probabilities at the syllabic unit in speech. The current study examined the influence of transitional probabilities on word segmentation from dual language input. Monolingual infants were placed in a bilingual linguistic environment to investigate their response in a dual language word segmentation task. Participants were 60 fullterm 12-month-old monolingual infants reared in an English-speaking environment (in which participants were exposed to at least 80% English). Participants were exposed to an artificial speech stream comprised of English- and Finnish-accented languages prior to the testing trials, during which infants responded to words and part-words of the two languages. The length of time participants stared at the screen while hearing the word or part-word indicated proper (short time period) or improper (long time period) word segmentation. The current study is ongoing due to a dearth of participants. Preliminary results suggest that infants successfully tracked transitional probabilities from the dual language input for English word segmentation and potentially for the Finnish task. The findings revealed a recency effect, in which the infants achieved better word segmentation in the language that they heard last in the speech stream.