Abstract:
Adult humans readily learn to respond to relations, but it is normally assumed that their ability to verbalize relations plays a critical role. To study relational learning in absence of verbalization, we developed a new technique using a multiple-object tracking task. In this task, participants are told to
track four out of eight objects cued at the beginning of the trial. At the end of the trial, a single object is cued, and participants respond whether they tracked it (yes/no task). The display contained two
strips of different width but participants were not informed about their presence.
The participants were randomly assigned to Informative and Random conditions. In Informative condition, the location of object cued at the end of the trial predicted the correct response. If the answer was "yes", then the cued object was located next to the narrower strip; otherwise, it was located next
to the wider strip (or vice versa). In Random condition, the cued object was located next to either strip, so that its location was not predictive of the correct answer. Postexperimental questionnaire showed that participants in Informed condition were not aware of predictive role of object location;
nonetheless, they were more accurate than participants in random condition, providing evidence of implicit relational learning in this new experimental paradigm. Our results suggest that ability to verbalize relations may not be essential for demonstrating relational learning.
