Abstract
Mojica & Peterson (VSS 2013) tested whether semantic priming affected figure assignment. Subjects viewed bipartite displays suggesting a familiar object on one side of a central border and reported where they saw the figure. Masked non-words (control primes) or words naming a different object than the object suggested in the display (experimental primes) appeared before displays. Words named an object in the Same Category (natural or artificial) as the object in the display (SC) for half the subjects and an object in a Different Category (DC) for the other half. SC participants saw the familiar object as figure more often (74%) than DC participants (67%), p=0.05. However, for both groups experimental and control means were equal. We investigated whether these equal means masked differential effects of experimental and control primes at different points during the experiment. In SC, familiar figure reports on experimental trials were originally low due to the basic-level mismatch between the prime and the object in the display yet increased to above control trials in the second half (p=.015; regression line slope = .49, p=.005). Over time, subjects likely learned to attend to superordinate category (natural/artificial) information in the prime and to devalue basic category information, ultimately resulting in semantic priming. In DC, familiar figure reports were initially low on both experimental and control trials perhaps because the mismatch between prime and display on experimental trials led subjects to devalue semantics regardless of source (prime or display). Familiar figure reports increased on both experimental and control trials (slope = .623 and .514, ps<.004) because DC subjects likely learned to revalue semantic information from the displays. Rapid learning of the relationship between primes and displays over a 64-trial experiment influenced the weighting of semantic information for figure assignment, and was required for semantic priming to emerge in SC.
Meeting abstract presented at VSS 2014