Cognitive markers of the uncanny valley effect

Abstract

In his essay “The Uncanny Valley", Masahiro Mori notes the following: People will experience negative feelings towards a humanoid object if some of its features reveal its mechanical nature. When the nature of the object and its features are consistent, there is no such abrupt negative response. This transition to negative affect has been called the "uncanny valley" effect. One of the most popular explanations for this effect is the evolutionary hypothesis. Its main concept is the «pathogen avoidance system», an evolutionary mechanism responsible for the induction of aversion in response to the sight of a sick member of the species. Like the authors, we hypothesise that the same system is involved in the formation of the uncanny valley effect. In our study, subjects were asked to perform a task in which they had to match images of faces in three categories (healthy, infected, "uncanny") with one of two categories (alive/not alive). The result was that healthy faces were categorised with less accuracy and speed than uncanny faces. Our results provide partial support for the evolutionary hypothesis. In a subsequent experiment, we decided to clarify the influence of the atypicality of the faces presented on our earlier results by comparing the recognition success of three groups of faces: unusual/uncanny/typical. It was shown that unusual faces were recognised significantly more accurately than uncanny or typical faces.