A Statistical Model for Predicting Child Language Acquisition: Unfolding Qualitative Grammatical Development by Using Logistic Regression Model

Language acquisition is a scientific puzzle still awaiting a theoretical solution. Children seem to acquire their native language in a spontaneous and effortless way and they probably do so by keeping track of the frequency with which language items such as phonemes or parts of speech occur. Advances in data storage, processing and visualization have triggered a growing and fertile interest in analysing language by relying on statistics and quantitative methods. In this paper we propose a multiple logistic regression model to evaluate how different components of language contribute to its acquisition over time. The empirical basis consists of a corpus, which can be considered as a series of statistically representative samples taken at regular time intervals. The aim is to show how quantitative methods can contribute to explaining the creation and development of grammatical categories in first language acquisition.