Input optimization: Verb-argument constructions in English textbooks in Türkiye

Abstract

Based on usage-based approaches to second language acquisition, studies point at a statistically significant correlation between type-token frequency, skewed distribution of items, and faster learning. Madlener (2016) shows a positive correlation between a Zipfian distribution of items in a German construction and faster, more accurate learning of the construction. While there are many studies that analyze input in English language teaching materials, no study has scrutinized selected constructions from an input optimization perspective, following Madlener (2016). Thus, using TAASSC (Kyle, 2016), the present paper analyzes four constructions, V in N, V about N, V for N, V with N, and the ditransitive constructions in the high school English textbooks in Türkiye. The results indicate that the input available for these constructions are not viable for generalizations to occur, leaving learners with unproductive one-time instantiations of the constructions, and low token frequency of these constructions also suggest that little to no entrenchment might take place. As such, the study proposes adopting a more corpus-based approach to English teaching materials.