ATTENTION TO FORM AND MEANING: LEARNING WITHOUT AWARENESS? AN INTERPRETABLE AND UNINTERPRETABLE APPROACH

Abstract

ATTENTION TO FORM AND MEANING: LEARNING WITHOUT AWARENESS? AN INTERPRETABLE AND UNINTERPRETABLE APPROACH

While attentional models of SLA postulate attention as crucial for learning to take place (e.g., Robinson, 1995b; Schmidt, 1990, 2001; Tomlin & Villa, 1994), the role that awareness or lack thereof plays remains debatable. Indeed, the studies that have empirically addressed the construct of unawareness reveal conflicting results (e.g., Williams, 2005 vs. Hama & Leow, 2010). Interestingly, within the recent minimalist syntactic approach to L2 learning, the notion of detectability in meaningful contrasts (Lardiere, 2009) is being postulated to play a role. In this approach, language variability responds to the specification of two types of features: interpretable and uninterpretable, and there is a debate as to whether both features are available for processing. (cf. Tsimpli & Dimitrakopoulou, 2007; Lardiere, 2009). To provide a bridge between these two areas, this dissertation sought to address (1) whether there was a distinction between the performances of 62 aware and unaware participants (advanced students of Spanish) on implicit form-meaning mappings during an incidental reading task, (2) whether type of feature [+/- animate], [+/- count], [+/- feminine] influenced the performances, and (3) whether participants' assignment of gender was based on [+/- feminine] noun gender or if type of noun ending ruled gender assignment. Participants completed immediate and one-week delayed posttests, including recognition, controlled written production, and comprehension tasks. The data were analyzed both quantitatively and qualitatively.

The results revealed that awareness seems to be beneficial for further processing and learning of the target items; however, level of awareness appears to be a factor that needs to be considered, especially at the delayed posttest stage. A disassociation between unawareness and further processing or learning was found, more in line with Schmidt's noticing (1990, 2001). The qualitative analysis of the think-aloud protocols revealed that participants adopted two different strategies. While both data-driven and conceptually-driven processes (Leow, 1998; Robinson, 1995b; Williams, 1999) were found, it was revealed that another variable, namely, type of feature, seemed to have contributed to the usage of one or the other. Finally, awareness or lack thereof did not appear to play a role in participants' reliance on ending cues for gender assignment.