Abstract
The object-file framework suggested by Kahneman, Treisman, and Gibbs (1992) explains how object continuity is maintained as the objects move, and how successive percepts are construed as belonging to the same object. A core tenet of this framework is that when we focus our attention on an object, we automatically retrieve this object’s recent history. The main supporting evidence comes from the object-specific preview benefit (OSPB), which refers to the finding that participants are faster to name the letter appearing in the second of two successive displays when a preview of the letter has appeared in the same vs. different object in the first display. Although this framework has been very influential, replicating the OSPB has proved difficult, a difficulty that was attributed to the fact that observers could ignore the initial preview-display letters. To address this problem, a modification to the task was suggested, in which participants have to match the target letter to the ones that appeared at the preview display. This paradigm has become the standard paradigm used to characterize object-file representations. However, by making retrieval of the objects’ history task-relevant, it does not capture the automaticity of this process, which is the heart of the object-file account. In the current research, we suggest an alternative go/ no-go object-reviewing task that resolves this problem: it requires participants to attend to the preview display yet does not require retrieval of the object history. Using our new paradigm, we reliably replicate the OSPB. Additionally, as a proof of concept, we show that the finding that the OSPB is long-lasting reported using the modified paradigm does not hold when using our novel paradigm.