Morphosyntax in predictive L2 sentence processing: A Visual World eye-tracking study of German verb morphology

Koch, E. M. (Speaker), Godfroid, A. (Contributor), Housen, A. (Contributor), Bulté, B. (Contributor)

Activity: Talk or presentationTalk or presentation at a conference


In sentence comprehension, L1 speakers constantly integrate lexical, morphosyntactic, pragmatic and contextual information to predict upcoming input (Huettig, Rommers, & Meyer, 2011), which facilitates efficient communication (Kutas, DeLong, & Smith, 2011). Adult L2 learners, however, have been shown to have difficulty exploiting grammatical cues to generate such predictions (Kaan, 2014). Using visual-world eye-tracking, Hopp (2013, 2015) found that L2 learners of German did not use gender marking on articles and adjectives for predictive processing, possibly because they were relying more on lexical-semantic information. Our study further investigates the issue of L2 morphosyntactic prediction by testing whether adult learners of German can use the grammatical information encoded by German verb inflection to anticipate upcoming input during real-time listening. Specifically, we investigate the ability to exploit syntactic number information, as contained in the German present tense verb form, to predict whether the upcoming referent will be singular or plural. We examine both productive (regular verb conjugation, providing morphosyntactic information through affixation) and unproductive German verb morphology (strong verb conjugation, using stem-vowel alternations in addition to affixation). We compare the data of 20 advanced learners of German to those of 20 L1 speakers in a visual-world eye-tracking experiment. The participants are exposed to two pictures, varying in the number of referents depicted (singular vs. plural), combined with oral sentences in which the suffix or stem-vowel of the verb represents the first cue for referent number. Successful exploitation of the morphosyntactic number cue is measured as anticipatory eye-movements towards the correct picture. In addition, we assess the participants’ explicit knowledge of German verb inflection by means of written and oral production tasks, requiring the participants to conjugate the verbs and to state how confident they are about the correctness of their productions. The findings may shed light on the extent to which L2 learners are able to exploit (un)productive morphological features during real-time auditory sentence comprehension, and they may help to identify processing differences between L2 and L1 speakers. Comparing the eye-tracking results to the participants’ explicit knowledge may enable us to distinguish between the presence of knowledge and its actual usage in comprehension.
Period21 Sep 2019
Event titleSecond Language Research Forum
Event typeConference
Conference number2019
LocationEast Lansing, United States, Michigan