Centre for Research in Language & Literacy (Vocabulary Studies) Seminar: From input to output: Insights of vocabulary learning from developmental corpora of reading and writing

From input to output: Insights of vocabulary learning from developmental corpora of reading and writing

External speaker: Yaling Hsiao (University of Oxford)

Book language differs from everyday speech in the richness and diversity of words present. This unique source may contribute to children’s subsequent development in literacy, including writing. Through corpus analysis of several developmental corpora of child-directed speech (CHILDES-UK) and child-directed text (picture book corpus and a reading book corpus), we measured lexical richness and structural properties. We found that children’s book language is lexically denser, more lexically diverse, and comprises a larger proportion of rarer word types compared to child-directed speech. They are also later acquired, more abstract, and more emotionally arousing than the words more common in child-directed speech. Similar analyses were conducted for a large corpus of children’s writing (BBC 500 Words Competition, 2019). There was a clear increase in lexical richness with age. Older children also produced words that were increasingly similar to words representative of book language and dissimilar to those representative of speech. These words were also later acquired and more abstract.  The findings suggest that written language provides unique linguistic input, which children learn to produce themselves as they grow.