How Low Can the Influence of Meaning Go? Does Activation in the Semantic System Influence Object Detection?
AuthorSkocypec, Rachel M.
AdvisorPeterson, Mary A.
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractInteracting with the objects we perceive in our environment is an essential aspect of daily life. Therefore, it is important to understand the factors that influence object detection. This dissertation, comprising three chapters, investigated whether object meaning (i.e., semantics) influences object detection. Figure assignment served as the index of object detection in all experiments in all chapters. Masked displays in which a portion of a familiar object was sketched on one side of the central border – the critical or “familiar configuration” side – were shown for either 90 ms or 100 ms; the longer exposure duration allowed more time for the familiar object in the displays to activate semantics. Familiar objects were depicted in both upright and inverted orientations; orientation-dependent detection is expected if configured representations of objects are involved. Chapter 1 investigated whether semantic expectations initiated by words denoting objects influence object detection. An unmasked word was presented before each test display to pre-activate the semantic network; the words denoted either the basic-level (BL) category of the objected sketched on the critical side of the display or an unrelated object from a different superordinate (natural vs. artificial) category (UNRdsc). Detection accuracy and detection RTs from experiments with words were compared to control experiments in which words did not precede test displays. For both 90-ms and 100-ms displays, object detection accuracy was higher than control following BL words but was unaffected by UNRdsc words. In the 100-ms condition, conflict emerged between the semantics activated by the UNR dsc word and the object in the display. Chapter 1 demonstrated that object detection is not only influenced by semantic activation, but it entails semantic activation in that object detection does not occur until conflict within the semantic system is resolved. Chapter 2 investigated whether semantic conflict in our paradigm emerged earlier in time when words in the UNR condition denoted objects in the same superordinate category (UNRssc) as the object sketched in the display. Semantic networks for objects in different superordinate categories have relatively few overlapping properties and are more distant from one another in cognitive and neural space than networks for objects in the same superordinate category. Conflict may emerge earlier in time for displays preceded by UNRssc words because the semantic networks have greater overlap and greater connectivity. Consistent with Chapter 1, object detection accuracy was higher than control for both 90-ms and 100-ms displays following BL words but was unaffected by UNRssc words. As predicted, conflict emerged earlier in time, with 90-ms displays, following UNRssc words. Together, the findings from Chapters 1 and 2 suggest that the amount of conflict in the semantic system at a given point in time varies as a joint function of the amount of semantic activation initiated by the display and the distance in cognitive and neural space between that display-generated semantic activation and the semantic activation initiated by the UNR word. Chapter 3 investigated whether unconsciously presented words affect object detection. BL and UNRdsc words were presented below conscious awareness threshold using sandwich masking in which the briefly presented word was immediately preceded and followed by a string of random letters. Results suggest that semantic activation from unconsciously presented words do not influence object detection. Together, the three chapters broaden our understanding of how, when, and where in the visual hierarchy semantic expectations influence object detection.
Degree ProgramGraduate College