Effects of types of input frequency distribution on second language construction learning

Abstract

English is taught and learned as a foreign language in Korea, which results in an inadequate amount of natural input, which, in turn, makes it imperative that language input be manipulated in ways that would lead to better learning outcomes. In this quasi-experimental study, 91 Korean elementary school students were randomly assigned into one of the five input frequency distribution conditions (one balanced distribution condition, three skewed-first distribution conditions with three different levels of skewedness, and one control condition) and exposed to a novel construction in English. The results showed overall that frequency did work and that the more skewed the distribution was, the more learning outcomes were observed (in terms of the extent of construction learning as well as the generalizability of the learned knowledge). The findings are discussed in light of the input-driven usage-based and frequency-based approaches to language learning.