Databases
By Stephen Shimanek . Last update : 10 July 2007. trans. Stephen & Caroline Shimanek
Case studies have revealed a general and yet serious issue: researchers each have their own system for encoding the information they collect. Many have an idiosyncratic system for representing different aspects of the information they gather and, as a result, their research material becomes difficult or even impossible for other researchers to use. As such, the need for a standardized method was quickly felt. In the 1980s, two well-known language acquisition specialists, Brian MacWhinney and Catherine Snow (1985) suggested putting together a computerized database on child language production that would be available to the entire university community. They created CHILDES, the Child Language Data Exchange System, to which it is possible to add data in a standardized format. There is data available from nearly thirty languages today, to which one can gain access with a simple click of the mouse. Hundreds of scholars have used the database, often citing CHILDES as their primary source of data in their published articles. It is important not to overestimate a database like CHILDES. Its ease of use makes it tempting to rely on this resource and thereby dispense with constituting one’s own data collection, but such a practice would have scholars working only on written transcription without ever having heard the speech of living subjects. In our view, though, being directly in contact with child language through observation or experimentation is absolutely fundamental for all those who wish to study language acquisition.
Choice of data collection methodology
This article was written from work done in collaboration with Mireille Brigaudiot and presented at the colloquium of the Societas Linguisticae Europaea, ENS-LSH - Lyon September 2004
