LANGUAGE ACQUISITION: RESEARCH METHODS
Much language acquisition research is longitudinal. Three general approaches can be found:
Theory-driven. The researcher adopts a theoretical framework; then seeks support for it in their data. This approach is particularly favoured by those who subscribe to Chomskyan theory, and who seek (for example) to trace evidence of infants setting parameters in the direction of the target language. The children might be asked to make grammaticality judgements, indicating whether they regard a sentence as grammatically acceptable or not.
Observational, analysing the data without prior assumptions. Diary studies have proved informative– though their disadvantage is that they do not preserve a record of the actual speech event. Video and cassette recordings have been widely used. They are obtained during regular meetings between researcher and infant; or by use of a timer.
Experimental. It is obviously difficult to engage very young subjects in experimental tasks. Two types of method have proved useful:
the high amplitude sucking procedure. An infant’s sucking on a teat settles down to a regular rhythm if there is little in the environment to distract the child– but speeds up markedly if something novel engages its attention. This phenomenon can be used to establish the extent to which infants discriminate between similar linguistic features, a change of sucking rate showing that they have identified a sound or rhythm that differs from an earlier one.
the operant headturn procedure, in which an infant is trained to turn its head when it encounters a novel stimulus. An independent observer notes when the infant moves its head through at least 308, and this is taken to show that the infant has noticed a change in the signal. A variant is the Headturn Preference Procedure, where different stimuli are played from different sides (e.g. an utterance in the ambient language on the left followed by an utterance in another language on the right). The infant’s headturns are monitored as an indication of which stimuli it finds more interesting.
For older infants, ‘listen and repeat’ tasks have been used to investigate areas such as phonological memory and lexical segmentation. In addition, many experimenters (especially those researching morphol ogy acquisition) have used simple picture-based tests. The researcher might, for example, show the infant a picture of an object and name it a wug; then show the infant a picture of two and ask for a response.
Language acquisition researchers have the benefit of a rich archive in the Child Language Data Exchange System (CHILDES) database collected by Brian MacWhinney and Catherine Snow, which brings together data from nearly a hundred research projects in a variety of languages.
Further reading: Foster (1990: Chap. 6); Jusczyk (1997: Appendix)
