Abstract
Our primary research goal is to determine whether alphanumeric characters and higher-level linguistic stimuli can be processed in parallel, without capacity limitation. There is an abundance of evidence for simultaneous processing of multiple alphanumeric characters (Erkison & Spencer, 1969; Shiffrin & Gardner, 1972; Pashler & Badgio, 1987). Whether the identification of higher-level linguistic stimuli benefits from covert shifts of attention is more controversial. In the current set of experiments we replicate Pashler and Badgio's (1987) highest-digit identification task and extend it to semantically-equated number words and American Sign Language number signs. This allows us to hold constant the underlying semantic content of targets and distractors (numerosity), the relationship between target and distractors, and the target selection criterion (highest number) while varying only the external surface structure (digit vs. word vs. sign) of the stimulus. Moreover, it allows us to assess digit, word, and sign processing in an understudied population of users with expertise in distributing visual attention over a variety of high-level linguistic stimuli. In hearing non-signers (n = 4) we find no processing advantage for successive vs. simultaneous identification of digits, but significant improvements in accuracy for successively-presented words. In Deaf signers (n = 5) we find identical patterns to the hearing non-signers, and additionally find that sign identification parallels that of word identification. These data are inconsistent with unlimited-capacity parallel processing models, and speak to the importance of the surface form of objects during multiple-object identification tasks. These results represent critical first steps in our goals to explore attentional capacity limitations at play during the perception of a wider range of real-world images and objects.