The uncanny valley phenomenon can be explained by categorization failure rather than categorization difficulty

Abstract

We perceive non-human objects that resemble humans to be eerie (uncanny valley). Prior research has explained this phenomenon based on animacy perception or cognitive processes not related to animacy (categorization-based hypothesis). This study removed biological properties from the stimuli to assess more fully the categorization-based hypothesis. Supporting this hypothesis, we found an uncanny valley-like effect with geometrically morphed figures; specifically, figures that took longer to categorize were rated as less likeable compared to quickly categorized figures (Experiment 1). The lower likeability of the morphed figures did not improve with perceptual priming (Experiment 2) but improved somewhat with “categorization label cues” (Experiment 3). Interestingly, categorization label cues (Experiment 3) and cognitive load (Experiment 4) did not increase/decrease the likeability of the least likeable morphed figures. These results suggest that categorization failure, due to stimulus characteristics rather than difficulty associated with processing fluency, is a potential factor underlying the uncanny valley phenomenon.