Picture Detection in RSVP: Features or Identity?

A pictured object can be readily detected in an RSVP sequence when the target is specified by a superordinate category name such as animal or vehicle. Are category features the initial basis for detection, with identification of the specific object occurring in a second stage (Evans & Treisman, 2005), or is identification of the object the basis for detection? When two targets in the same superordinate category are presented successively (lag 1), only the identification-first hypothesis predicts lag 1 sparing of the second target. The results of two experiments with novel pictures and a wide range of categories supported the identification-first hypothesis and a transient-attention model of lag 1 sparing and the attentional blink (Wyble, Bowman, & Potter, 2009).