PERCEPTUAL ABSTRACTION AND INTERACTIVITY IN REPEATED ORAL READING - WHERE DO THINGS STAND

Whittlesea (1990) reports context specificity in oral rereading. Repetition benefit is larger when a random scrambling of a text is read twice than when two different scramblings are read, indicating an "episodic" or "text-level" contribution to repetition benefit. We examine Whittlesea's data by focusing on variability across stimulus materials in the magnitude of context specificity, which ranges from -0.1% to 694.7% of the repetition benefit attributable solely to lexical overlap. We describe similar variability in other recent research, including new work of our own. The central question in the episodic-abstractionist debate is no longer whether context specificity can or cannot be found in rereading but what factors govern the relative contributions of text-level and lexical-level repetition. We offer a "level of focal attention" hypothesis to supplant the "task change" hypothesis of Carr, Brown, and Charalambous (1989). The new hypothesis is a strategy-oriented version of transfer-appropriate processing.