Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Probing the neural processing of orthographic gender cues on German nouns: An ERP study on L1 sentence reading
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
This poster is part of the Sandbox Series.
Rasmus Sinn1, Ana Rita Sá-Leite1; 1Johann Wolfgang Goethe University Frankfurt
Grammatical gender in German is a challenging aspect of linguistic theory and experimental research. Most German nouns are assigned non-semantic gender, a process largely characterized as arbitrary. On the other hand, scholars have identified a range of form-based regularities, demonstrating correlations between sublexical characteristics of nouns and their gender value (e.g., the word ending -ung and feminine gender). It has since been a subject of debate whether and to what extent these regularities affect the access and retrieval of grammatical gender. Pioneering work in other languages with similar situations have led to the proposal of a dual-route model of gender processing. This model assumes (1) a lexical route, via which gender is simply accessed as an abstract representation attached to whole-word entries from the mental lexicon, and (2) a sublexical route, where form-based regularities on nouns are used to retrieve gender. The sublexical route is thought to exploit transparent noun endings, i.e., endings that are highly predictive of a particular gender value. Previous research in Spanish and Italian provided support for this idea, showing that brain responses differed for transparent nouns ending in -a (feminine) or -o (masculine) compared to opaque or irregular nouns. The generalization to less gender transparent languages, however, has not yet been addressed with brain data. Therefore, we aim to replicate the existing event-related potential (ERP) findings in German. Native speakers are presented with sentences word-by-word in a rapid serial visual presentation (RSVP) and instructed to decide on the sentences’ grammatical status. Sentences can either include a gender agreement error on article–noun pairs or not. In addition to this violation paradigm, we manipulate the level of gender transparency of the critical nouns by using three conditions: nouns can either (a) display a transparent pseudosuffix cue, -e for feminine or -er for masculine, (b) a transparent suffix cue, such as -ung for feminine, or (c) display no gender cue at all, thus being opaque. Across these three transparency conditions, nouns are controlled on various dimensions like syllabic structure, frequency, concreteness, and neighborhood measures. Participants’ continuous electroencephalogram (EEG) is recorded and ERPs are time-locked to the onset of the critical nouns. Based on prior research, we preregistered to analyze four consecutive time windows from 100 ms to 900 ms post-stimulus onset with linear mixed-effects models (LMMs). First, we expect a biphasic response to gender agreement errors consisting of a left anterior negativity (LAN) followed by a P600. Second, we expect gender transparency to independently influence the ERP amplitudes in the following way: we predict the most negative amplitudes for transparent pseudosuffixes, followed by transparent suffixes, and lastly opaque endings. Such a staircase pattern would indicate that neural processing is sensitive to gender cues on nouns during lexical access and support the idea that this sensitivity is a function of the cues’ strength. This study provides a crucial test for the dual-route model of gender processing, with the intricacies of the German gender system allowing for a fine-grained insight into the neural activity during lexical access.
Topic Areas: Reading, Morphology