Contact:
Jonathan Potts
412-268-6094

For immediate release:
October 20, 2005

World's Largest Bilingual Child Language Database Added To Carnegie Mellon's Groundbreaking CHILDES System

Psychology Professor Brian MacWhinney created CHILDES in 1984, and it is now the most sophisticated of 13 databases known as TalkBank.

PITTSBURGH—Carnegie Mellon University's innovative Child Language Data Exchange System (CHILDES) now includes the Hong Kong Bilingual Child Language Corpus, which is the world's largest video-linked database of children learning two first languages.

CHILDES is an online, searchable collection of interactions that demonstrate how children learn language in a variety of settings. Carnegie Mellon Psychology Professor Brian MacWhinney created CHILDES in 1984, and it is now the most sophisticated of 13 databases known as TalkBank. Carnegie Mellon and the University of Pennsylvania launched TalkBank in 1999 with a grant from the National Science Foundation. TalkBank includes video and audio recordings, all searchable on the Web, from a variety of disciplines, including law and medicine. Audio recordings of oral arguments made before the U.S. Supreme Court are included on TalkBank.

The Hong Kong database—created by Professor Virginia Yip of The Chinese University of Hong Kong and her husband, Dr. Stephen Matthews from The University of Hong Kong,—features 170 hours of audio and video files of four families raising their children bilingually in Cantonese and English. The project, which includes transcripts and searchable video and audio segments, took 10 years to compile.

"Researchers around the world are now relying on software developed at Carnegie Mellon to study the acquisition of child language. In earlier decades, this work relied on tape-recordings and notebooks," MacWhinney said. "However, today these projects are using video recording of children while interacting with their caretakers in the home as a way of understanding the real process of language learning in context."

The Hong Kong database has already been the data source for several undergraduate and graduate dissertations in Hong Kong, and it is the basis of a book written by Yip and Matthews to be published by Cambridge University Press. The database focuses on children who are bilingual in English and Cantonese and who learned to speak two languages through the one-parent, one-language approach. Using that method, one parent speaks to the child in one language, and the other parent speaks to the child in another.

"One of our main objectives was to produce a multimedia state-of-the-art database," Matthews said. "Researchers in many fields can use it for their own special interests."

CHILDES and TalkBank can be accessed at http://childes.psy.cmu.edu/.