A number of recent studies suggested that visuo-spatial perspective taking (VSPT) occurs spontaneously when viewing either a human body or an action by an agent. However, it remains unclear whether VSPT is caused by the observation of an (potential) action or occurs because the observer infers from certain cues that another mind is present (whether actions/action cues are displayed or not). To examine whether action cues are necessary for VSPT, we presented to participants human faces without a torso, that is: stimuli that indicated the presence of another mind, without providing action (effector) cues. Furthermore, we examined whether "relevance" of understanding the other mind would influence VSPT, where relevance was manipulated by comparing the effects of observing a fearful versus a neutral facial expression on VSPT. Results showed that spontaneous VSPT occurs when a face with a fearful, but not with a neutral, expression is perceived. This indicates that spontaneous VSPT occurs (at least more robustly) under circumstances, where VSPT is of "relevance" for understanding the situation. Furthermore, directly observing actions, or action cues, does not appear to be a prerequisite for VSPT. (Contains 2 tables and 1 figure.)