Gains and Losses: Is Figure Ground Perception Influenced by Motivation or Learned Value?

摘要

Can learned value influence figure-ground perception? Participants learned to associate probabilistic gains and losses with novel shapes. Shapes were presented in fixed pairs: two win pairs, two loss pairs, and two control pairs. Within win and loss pairs, one shape had a p(80%) value (win: +10?￠; loss: -10?￠) the other had a p(20%) value. The complement of the p(80%)-shape in one win(loss) pair was the p(20%)-shape in the other win(loss) pair. Subjects learned to maximize their earnings by choosing the p(80%)-win shape and the p(20%)-loss shape. Next, participants reported on perceived figure-ground organization of bipartite displays in which an 80% win(loss) shape competed for figural status with a 20% shape in the same win(loss) condition. If learned value influences figure assignment, the relatively high-value shape should be seen as the figure (80% win shape, 20% loss shape). Instead, participants were more likely to perceive the low-value shape as figure, evident across eight repetitions of figure-ground displays in the win condition (57%, p .04), but only four repetitions in the loss condition (58%, p .02). Thus, the expected effect of value was not observed: When a low-value shape competed for figural status with a high-value shape, the low-value shape won. This is the shape participants were motivated to avoid [p(-10?￠)=80% in loss pairs and p(+10?￠)=20% in win pairs]. The high-value shape that participants were motivated to choose during learning (because it maximized earnings) lost the competition for figure assignment. Our results are the first to show that motivation, particularly motivation to avoid, influences figure assignment. Why does motivation to avoid rather than motivation to choose influence figure assignment? The decision-making literature shows that losses loom greater than gains; our results are consistent with that human tendency provided that we consider a low p(gain) a loss relative to a high p(gain).