Language and Task Effects on the Inferencing Processes of Skilled Readers in L1 and L2 for Two Academic Tasks

Although there is a substantial body of inferencing research, most of our knowledge comes from narrative texts, but text type affects inferencing processes (Horiba, 2013; Lorch, 2015; Narvaez et al., 1999; van den Broek et al., 2001), as does reading purpose (Carver, 1997; Kintsch, 1998; Lorch, 2015; Perfetti & Fukaya, 2015). Further, differences and similarities of L1 and L2 readers are even less clear (Horiba, 2000, 2013; Nassaji, 2007). Inference models based on narrative texts do not necessarily apply to expository texts; elaboration of current models is required for expository reading research (Lorch, 2015). This presentation details an extended model of narrative-based inferences (see Cook & O’Brien, 2014; Graesser et al., 1994; Trabasso et al., 1989; van den Broek et al., 2001; Zwaan & Singer, 2003) accounting for expository differences such as textual complexity, reading purpose, and decreased background knowledge (Lorch, 2015). Framed with this expanded model, the study investigated the inferencing processes of skilled, tertiary-level L1 (German) and L2 (English) readers for language and task effects on inference types, frequency and distribution. Participants read two expository texts (L1 & L2) in two academic task conditions: a summary task and a position-paper task. Tasks were chosen for theoretical and authenticity reasons; reading to summarise should foster different inferencing processes than a position-paper (Kintsch, 1998) and both are common tertiary-level tasks. Think-aloud protocol and stimulated recall data were recorded, transcribed, coded, and compared quantitatively and qualitatively across tasks and languages. The results showed task and language effects on inferencing processes. When reading for a summary, inferencing processes differed across languages but not for the position-paper task. Inferencing processes differed significantly across tasks in L2, but not in L1. The results suggest that skilled readers strategically inference based on task, but the transfer from L1 to L2 is not complete.