The Role of Individual Differences in L1 and L2 Processing of Bridging and Predictive Inferences

Abstract

Second language acquisition (SLA) and psycholinguistic researchers are interested in how inferences are processed in first language (L1) and second language (L2) reading to better understand how learners negotiate for meaning within a text. Preliminary evidence from brain-based studies suggests that bridging inferences (i.e., connecting pronouns and referents) may rely more heavily on short-term memory (STM) (Burkhardt, 2006) while predictive inferences (i.e., forecasting upcoming events from textual information) may rely more heavily on cognitive control (CC) (Jin et al., 2009). However, despite these advances, it is still not clear whether or not L2 readers have access to the necessary linguistic structures (Horiba, 1996) or, instead, if proficiency level is a confounding variable (Cohen, 1998; Guerrero, 2005; Leontiev, 1981).

To address these gaps, an L1 English reading study and an L2 English reading study were carried out to explore whether or not inferences are processed differently than non-inferences, whether or not bridging and predictive inferences are processed differently from one another, and which cognitive resources are recruited during the processing of bridging and predictive inferences. The L1 and L2 data were compared to see if inference processing differs based on language background. L1 English speakers (n = 50) and L2 English-L1 Spanish speakers (n = 23) completed a cumulative self-paced reading task in a distraction paradigm; participants read scenarios that required them to make a bridging inference, a predictive inference or no inference with interference (while completing a secondary task that taxed STM or CC) or without interference. Independent measures of STM and CC were also collected to see whether these scores were associated with reading performance.

The results indicated that L1 and proficient L2 readers process inferences similarly. Participants relied on CC for bridging and predictive inferences, supporting a multi-component framework of attentional control capacity (Barnes & Jones, 2000). These results are consistent with research on individual difference in SLA (e.g., Dornyei & Skehan, 2003), in that they suggest that CC may influence L2 classroom learning and account for the variation in L2 reading acquisition. Additionally, the findings have implications for L2 testing, as they pose a threat to the internal validity of standardized reading tests and ultimately the grade-level placement of L2 readers.