Testing predictions of competing models of semantics in bilingualism

• Applicant: Professorin Dr. Katharina von Kriegstein, since 1/2022
• Subject Area: Human Cognitive and Systems Neuroscience
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 468554903

With the ever increasing globalization of countries and cultures, the learning of foreign languages (L2) has become necessary for the movement of goods, people, and information. A central aspect of L2 learning is acquiring the meanings, or semantics, of novel words. My previous work has demonstrated that brain structures housing L2 semantic representations show experience-dependent neuroplasticity following L2 training. Other work suggests that L2 representations may be less rich or detailed in terms of their semantic features than native language (L1) representations, potentially due to reduced L2 experience relative to L1. My proposal draws these two lines of work together by asking (i) how semantics are similarly and/or differently represented in L2 in comparison to L1, and (ii) what sort of L2 training, if any, could lead L2 semantic representations to become more similar to L1 representations.Answering the question of how semantics are represented in L2 versus L1 is important because the language we use may impact the decisions we make. Behavioral studies show that individuals tend to make more rational choices in L2 relative to L1, rate L2 as less emotional than L1, and produce fewer errors when performing cognitive tasks in L2 relative to L1. Thus, differences in semantic processing in L1 and L2 may not only have an impact on overt language comprehension but may also have more implicit consequences for everyday behavior. This question also has implications for clinical interventions, such as training on L2 as a means of influencing L1 aphasia, and pedagogy, such as using different teaching strategies to instruct students in their L1 versus an L2.My hypotheses and experimental approach are driven by three competing models of bilingual semantic representations: shared, hybrid, and distinct models. Shared models view semantic representations as shared across L1 and L2, hybrid models view semantic representations for L1 and L2 as mostly distinct but with some overlapping features, and distinct models view semantic representations for L1 and L2 as entirely separated. I plan to use semantics for action words (e.g., combing), which are associated with specific cortical, autonomic, and muscle responses, as a test bed for dissociating model predictions. My research questions necessitate a multi-method approach involving fMRI, electromyography, heart rate recording, EEG, and highly-controlled L2 training protocols. This innovative approach overcomes several previous limitations including the inability to dissociate semantic- versus word-form-mediated brain responses to L2 words and difficulty disentangling multiple parameters related to L2 learning experience (e.g., proficiency and age of acquisition).

• DFG Programme: Research Grants
• Ehemaliger Antragsteller: Brian Mathias, Ph.D., until 1/2022