Abstract
Concepts of positive and negative valence are metaphorically structured in space (e.g., happy is up, sad is down). In fact, coupling a conceptual task (e.g., evaluating words as positive or negative) with a visuospatial task (e.g., identifying stimuli above or below fixation) often gives rise to metaphorical congruency effects. For instance, after reading a positive concept, visual target processing is facilitated above fixation. However, it is possible that tasks requiring upwards and downwards attentional orienting artificially strengthen the link between vertical space and semantic valence. For this reason, in the present study the vertical axis was uncoupled from the response axis. Participants made eye movements along the horizontal axis after reading positive or negative affect words, while their saccade movement trajectories were recorded. Based on previous research on saccade trajectory deviation, we predicted that fast saccade trajectories curve towards the salient segment of space, whereas slow saccade trajectories would curve away from the salient segment. Examining saccadic trajectories revealed a pattern of deviations along the vertical axis consistent with the metaphorical congruency account, although this pattern was mainly driven by positive concepts. These results suggest that semantic processing of valence can automatically recruit spatial features along the vertical axis.
Acknowledgements
This study was supported by Discovery grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) to JP and ALC, a graduate NSERC scholarship awarded to DGG, and an NSERC undergraduate research scholarship awarded to AC. We would like to thank Naseem Al-Aidroos for very helpful suggestions.
Notes
1 With black distractors, we found that fast incorrect saccades tended to be more upwards than slow incorrect saccades, direction-specific proportion of errors for fast and slow saccades: 3% vs. –1%, t(11) = 2.17, SE=0.02, p=.052. By contrast, with white distractors, slow incorrect saccades tended to be more upwards than fast incorrect saccades, although this pattern was not significant, direction-specific errors: 0% vs. 2%, t(11) = 1.12, SE=0.024, p=.28