Statistical word learning: Examining changes in learning behaviour due to differences in language environment across time

• Applicant: Tanja Charlotte Römbke, Ph.D.
• Subject Area: General, Cognitive and Mathematical Psychology
• Term: since 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 505754094

Statistical word learning—the ability to acquire word meanings based on their co-occurrence with referents across ambiguous situations—is a key mechanism in language acquisition. It is commonly studied using the cross-situational word learning paradigm, in which learners must infer word-meaning-mappings across multiple ambiguous trials. While most research has focused on simple one-to-one mappings, natural language often involves more complex relationships, such as synonyms or, for bilinguals, translations. How such complexity affects learning remains poorly understood. In the first funding phase, we investigated, inspired by bilingualism, in several experiments how differences in the language environment can impact statistical word learning. More specifically, we studied how mapping complexity (e.g., multiple words sharing a referent), participants’ language backgrounds (monolingual vs. bilingual), and language context (e.g., language switching) shape statistical word learning. We found that both the structure of word-meaning-mappings and significantly influence learning outcomes. Notably, we could observe that while words with multiple meanings are harder to learn, they can also be more easily remapped to a new meaning. In addition, we observed learning differences between monolingual and bilingual participants. The proposed second funding phase builds directly on these findings of the first funding phase. We will pursue three key objectives: First, we will investigate how different levels of referential ambiguity affect the learning and retention of simple and complex word mappings. Second, we will examine the role of attentional mechanisms during acquisition and retention using both behavioural and eye-tracking measures. Third, we will explore how long-term bilingual experience shapes statistical word learning through a systematic review and meta-analysis. This continuation expands the original project in critical ways. First, we will systematically vary referential ambiguity to understand its impact on learning trajectories—an aspect previously held constant. Second, we will assess retention across time, moving beyond single-session designs to examine the durability and integration of learned words. Third, by including unambiguous conditions, we will link our work to broader language learning research not limited to cross-situational word learning paradigms.

• DFG Programme: Research Grants
• Co-Investigator: Professor Dr. Iring Koch