Abstract
Do semantic expectations activated by words aid object detection? In two experiments, participants viewed black and white (B/W) test displays divided into two equal-area regions by a central border. One side of the border depicted a familiar object (upright or inverted); the complementary side depicted a novel object. Objects are typically more likely to be detected on the familiar side. Before each test display, a word appeared denoting either the basic level (BL) of the familiar object or an unrelated different-category (natural/artificial) object. Participants reported where they perceived an object relative to the border. Semantic expectation effects would be demonstrated if objects are detected on the familiar side of the border more often following BL than unrelated words. In Experiment 1, subjects categorized unmasked 500-ms words as natural or artificial; 200 ms afterwards, a 90-ms B/W display appeared and was masked. Object detection on the familiar side of the border was reduced for both orientations following unrelated words, but not enhanced following BL words, p < .03. This pattern indicates that semantic expectations affect object detection by interfering when features of the expected object mismatch those of the presented object. In Experiment 2, 50-ms words were rendered unconscious by pre- and post-masking; 120 ms later, a test display appeared. Semantic effects from masked words are notoriously unreliable, and are tuned up/down by task set. To enhance semantic processing of the masked words we randomly intermixed test words requiring semantic categorization with the B/W displays. Word categorization responses were accurate (97%), although response times were ~300 ms longer than in single-task experiments. However, under these task uncertainty conditions the exposure duration of the B/W test display was too short to observe familiarity effects, p = .90, so word type effects could not be assessed. Longer display exposure durations are currently being tested.
Meeting abstract presented at VSS 2018