Captions and learnability factors in learning grammar from audio-visual input

Abstract

This study explores the effects of extensive audio-visual input with three captioning modes - unenhanced captions, textually enhanced captions, and no captions - on learning a variety of L2 grammatical constructions and examines the effects of three learnability factors: construction type, frequency, and recency. A total of 112 participants watched ten full-length TV series episodes over a period of five weeks. The study targeted 27 frequently occurring grammatical constructions categorized as fully-schematic, partially-filled, and fully-filled. The design included a pretest, an immediate posttest to measure the effects of recency, and a delayed posttest. The results indicated mixed effects of captioning: textually enhanced captions - a more salient condition - led to immediate learning outcomes while unenhanced captions resulted in higher long-term effects. A limit to the amount of different textually enhanced constructions presented in the input for effective learning is suggested. In general, unenhanced captions appear sufficient for successful grammar construction learning.