SPONTANEOUS ATTENTION TO PRIMED AND NONPRIMED INPUTS

The literature is ambiguous with respect to whether attention is drawn spontaneously to expected or unexpected items in mixed arrays. Severa l studies from our own laboratory indicate that even though expected w ords are more localizable than unexpected words in unmixed four-word a rrays, showing a baseline advantage for expected words, unexpected wor ds are sometimes more localizable than their expected companions in mi xed arrays, suggesting that unexpected words attract attention (see, e .g., Johnston & Schwarting, 1996). By contrast, Dark, Vochatzer, and V anVoorhis (1996) observed that expected words were more reportable tha n their unexpected companions in mixed, two-word arrays. However, beca use the Dark et al. research did not include arrays in which both word s were expected, it is not clear whether their findings reflect an att entional effect over and above a baseline advantage of expected words. The present study added some additional controls in order to assess t his possibility. The superior reportability of expected words was even greater in mixed arrays than in unmixed arrays, suggesting that expec ted words in mixed arrays attract attention. Following Johnston and Ha wley (1994), the conflicting effects of expectancy on spontaneous atte ntion are taken as further evidence that the mind/brain system is bias ed simultaneously toward both what it most expects and what it least e xpects to perceive.