“…While some studies have shown intermediate and lower intermediate learners rely heavily on bottom-up processing, breaking down the sound stream into meaningful units, for interpreting the incoming speech stream (Hansen & Jensen, 1994), many more have shown top-down processing, drawing on the context and background knowledge, is widely used (Mueller, 1980;Voss, 1984;Wolff, 1987;Koster, 1987;Mack, 1988;Long, 1989) and that it could be misleading if learners failed to correct their initial predictions based on their background knowledge in a timely fashion (Long, 1990;Lund, 1991;Hansen & Jensen, 1994;Tsui & Fullilove, 1998). However, more recent studies have underlined that, depending on the circumstance, one or the other strategy might be brought to the front and this may well be different from learner to learner (Field, 2004).…”