In three visual-world eye-tracking studies, we investigated the individual and combinatory use of lexico-semantic information of the verb and gender marking of the determiner to predict the referent of an upcoming object among L1 and L2 German speakers. Importantly, in our design random, the linguistic input randomly varied in terms of the type of information (lexico-semantic, morphosyntactic, or a combination thereof) that could be used as predictive cue, most closely reflecting real-world language input compared to most previous studies. Cluster-based permutation analyses revealed prediction effects based on lexico-semantic cues in both L1 and L2, although for placement verbs, effects were only found in L1. Gender cues and the combination of lexico-semantic and gender cues also led to prediction effects solely in L1. The results not only confirm more limited predictive processing in L2 compared to L1 reported in previous studies, but also suggest a moderating role of type of cue in L1 predictive processing. These findings are discussed in the light of the good-enough approach (Ferreira et al., 2002; Ferreira & Patson, 2007) and the shallow structure hypothesis (Clahsen & Felser, 2006, 2018).
Building similarity graph...
Analyzing shared references across papers
Loading...
Pernelle Lorette
University of Mannheim
Building similarity graph...
Analyzing shared references across papers
Loading...
Pernelle Lorette (Wed,) studied this question.
www.synapsesocial.com/papers/69f6e5ac8071d4f1bdfc643c — DOI: https://doi.org/10.1016/j.actpsy.2026.106948