We propose that machine-learned computational models (MLCMs), in which the model parameters and perhaps even structure are learned from data, can complement extant approaches to the study of text and discourse. Such models are particularly useful when theoretical understanding is insufficient, when the data are rife with nonlinearities and…

This experiment investigated interactions between lower- and higher-level processing when reading in a second language (L2). We conducted an eye-tracking experiment with the within-subject manipulation inconsistency (to tap higher-level coherence-building) crossed with a within-subject manipulation of word-processing difficulty (to alter the ease…

This study aims to establish whether connectives can create referential expectations in discourse, and, if so, what these expectations are based on: connective semantics or frequency distributions in language use. This was tested by comparing the processing of the connectives "and" and "but" in Dutch and Russian by means of an…