An empirical study on word order in predicates: on syntax, processing and information in native and learner English

Abstract

This study focuses on the linearisation of verbal complements (or arguments) and adjuncts (or modifiers) in learner advanced English. The analysis is based on minimal pairs of sequences of complements and adjuncts in predicates, as in He will investigate [the construction] [in a somewhat strange way which will lead to odd results] versus He will investigate [in a somewhat strange way which will lead to odd results] [the construction]. Constituent linearisation has been claimed to be potentially subject to lexical, syntactic, processing and informative determinants. This paper analyses the influence of the verbal heads and the principles ‘complements-first’, ‘end-weight’ and ‘given-new’ on the production of predicates containing two dependents. The study has two objectives: first, to determine in which areas the ordering of such constituents in English by non-native speakers is particularly influenced by their first language (Spanish); second, to assess the plausibility of the Interface Hypothesis in a pattern which affects the so-called internal and external interfaces. Results show that internal constraints such as the length of the dependents (end-weight) and the learner’s source language (i.e. compliance with complements-first) are strong predictors of ordering choices.