José Alemán Bañón. Foto: Pia Nordin
José Alemán Bañón. Foto: Pia Nordin

Webbimarium i Zoom

Om Zoom.


Research on language comprehension suggests that native speakers of a language use a variety of linguistic and nonlinguistic cues to anticipate likely continuations at all levels of linguistic representation (see Kuperberg & Jaeger, 2016). In contrast, the role of anticipatory mechanisms in second language “L2” processing is less clear cut. Some proposals assume that L2 learners have a reduced ability to generate predictions across all domains of grammar, even at high levels of proficiency (e.g., Grüter et al., 2012, 2017; Martin et al., 2013). Others claim that prediction is qualitatively similar in the L1 and the L2, although factors such as low proficiency or cross linguistic differences can obscure L2 learners’ predictive abilities (see Kaan, 2014).

In this talk, I will present preliminary results from two ERP (event related potentials) studies investigating predictive processing in English at the level of the discourse, a domain of grammar that remains largely unexplored with respect to predictive processing (Rohde & Horton, 2014). In Study 1, we examined the extent to which native speakers of English and advanced L1 Spanish L2 English learners used information structure constraints (i.e., knowledge that an it cleft construction assigns Focus) to anticipate certain properties of an upcoming Focus phrase. Our results suggest that, although both groups generated predictions, they did so differently, possibly due to cross linguistic differences. In Study 2, we followed up on this possibility by examining a case in which the activated representation (i.e., what is predicted) encoded feature information that was either similar or different in the learners’ L1 (Spanish or Swedish). Our results revealed a native like ERP pattern for cases where the activated representation encoded similar features in the L1 and the L2.

Overall, these preliminary results suggest that, although L2 learners can use information structure and contextual constraints to generate predictions at the level of the discourse, these are impacted by differences in how the L1 and the L2 realize the activated representation.

Referenser: se pdf