Webbinarium i Zoom


Om Zoom.


Two key issues in first (L1) and second (L2) language processing research have been the extent to which L2 processing can become nativelike (Clahsen & Felser, 2006, 2018; Cunnings, 2017; Hopp, 2018; McDonald, 2006), and whether purported L1-L2 differences can be explained by individual differences (Hopp, 2014, 2015). In this talk, I will examine these two issues from the perspective of garden-path sentences like (1), where the temporarily ambiguous phrase 'the baby' may initially be interpreted as the direct object of the verb 'dressed' during incremental processing, when it is in fact the subject of 'played'. Although it has long been shown that garden-paths cause processing difficulty during L1 and L2 reading (Frenck-Mestre & Pynte, 1997; Juffs & Harrington, 1996; Roberts & Felser, 2011), L1/L2 differences have been observed such that L2 learners have increased difficulty in recovering from garden-paths (Jacob & Felser, 2016; Pozzan & Trueswell, 2016). Individual differences may also influence L2 syntactic ambiguity resolution (e.g. Hopp, 2014, 2015).

(1) After Anna dressed the baby in the cot played happily.
I will present results from a series of studies examining these issues. I will argue that L1/L2 differences in garden-path recovery primarily relate to L2ers' difficulty in erasing initial misinterpretations from memory, rather than an inability to construct the globally correct syntactic structure of a garden-path sentence. In terms of individual differences, I will highlight methodological concerns regarding whether psycholinguistic tasks systematically measure individual differences, as a case study of wider debate in the use of cognitive tasks from experimental psychology as individual differences measures (Hedge et al., 2018; James et al., 2018; Parsons et al., 2019). I will argue that before examining individual differences in L2 (and L1) processing, we must first consider whether commonly used psycholinguistic tasks consistently measure systematic individual differences to begin with.