Effects of robotic agents' appearances on users' interpretations of the agents' attitudes: towards an expansion of "uncanny valley" assumption.