Individual differences in the earliest stages of second language morphosyntactic learning

Although late second language (L2) learning is often assumed to draw on explicit learning mechanisms, there is evidence that adults can acquire the grammar of novel languages under incidental learning conditions (Rebuschat et al., 2021) and that the acquisition of this knowledge can be implicit in nature (Rebuschat & Williams, 2012). Nevertheless, participants tend to vary markedly in their learning outcomes, with part of this individual variation stemming from a range of learner-internal and learner-external variables and their interactions. This thesis presents three artificial language learning studies revisiting the question of whether adults can learn grammar in incidental learning conditions, while also exploring the modulating effects of learners’ first language (L1) background and metalinguistic awareness (study 1), individual differences in cognitive abilities (study 2) and redundant linguistic cues (study 3) at the initial stages of L2 acquisition. Learners in studies 1 and 2 were exposed to Kepidalo, an artificial language that had case marking and flexible word order, with a canonical Subject-Object-Verb (SOV) and a non-canonical Object-Subject-Verb (OSV) order. Study 1 comprised two experiments. Experiment 1 consisted of five online sessions, during which native speakers of English received vocabulary training while being incidentally exposed to grammar, and were, subsequently, asked to complete a series of picture-selection and grammaticality judgment tasks (GJT) assessing their grammatical knowledge. In the second experiment, the amount of incidental exposure was increased by two additional sessions whilst, this time, participants were native speakers of German. Moreover, testing was followed by a metalinguistic awareness questionnaire. Built on the first experiment of Study 1, Study 2 examined how individual differences in cognitive abilities, specifically explicit learning (EL), implicit statistical learning (ISL), and sustained attention, contribute to early L2 learning outcomes. To examine the relative effects of these abilities, participants were presented with a different cognitive ability task at the end of each of the first 3 sessions. Finally, in study 3, three groups of participants were presented with different variants of Kesadalo, a simplified artificial language that had a fixed Verb-Object-Subject word order. Importantly, for two out of the three groups, thematic role assignment was additionally determined by a low-salient or a high-salient redundant case marker, allowing to test the role of redundancy during initial novel morphosyntactic learning and the moderating effect of perceptual salience. The results of these studies underscore adult L2 learners' difficulty with acquiring case marking as opposed to aspects of word order under incidental exposure conditions. In addition, the findings showcase the facilitative roles for metalinguistic awareness and EL, as well as, to a lesser extent, ISL in early incidental L2 acquisition, and highlight the time-varying nature of their contributions, with the latter being a determinant of learner’s success in the very early sessions of exposure and the former two contributing to learning at a later time point. Furthermore, results from study 3 indicate that redundancy has a rather hindering effect on early adult L2 acquisition, which can, however, vary depending on the salience of the redundant cues. Taken together, the results reported in this thesis speak to the debate about the limitations of adult L2 learning pointing towards the presence of a threshold, which appears to be tightly linked to learners’ biases from their prior L1 experience (NC. Ellis, 2006a; MacWhinney, 2005) and the emergence of explicit knowledge, thereby providing new insights on the study of incidental L2 learning.