Level-two visual perspective taking as spontaneous perceptual simulation

Visual perspective taking lies at the heart of social interactions, allowing people to readily compute what another person can see, or how they see it (Flavell et al., 1981). This ability not only links to the unique human propensity to infer others’ beliefs, desires, and goals (Schurz et al., 2015), but also to more general spatial abilities such as navigation skill (Kozhevnikov et al., 2006) and action planning (Creem-Regehr et al., 2013). What is unclear is how others’ perspectives are represented. While several studies have in the past implied that others’ perspectives are represented in a similar way as our own viewpoint (e.g. Freundlieb et al., 2016; Sampson et al., 2010; Surtees et al., 2016) none have demonstrated the form that this overlapping perspective takes, and how these altercentric representations interfere with our own. The experiments in this thesis provide the first known evidence that when taking another’s perspective, representations of their visual input take a (quasi)-perceptual form, and can drive perceptual decision making in the same way as own input (e.g. Roelfsema & de Lange, 2016). Using a modified mental rotation paradigm (Shepard & Metzler, 1971), we simply asked participants to state whether alphanumeric characters appearing at varying orientations away from upright were canonical or mirror inverted (e.g. “R” vs. “Я”). In some scenes another person was present, seen sitting at 90 degrees either to the left or the right of the character. We first confirmed the classic mental rotation effect, where response times increase gradually as a function of the item’s increasing angular disparity away from upright. Crucially, we also found that the angular disparity relative to the inserted persons also produced this response pattern, and found that participants responded surprisingly quickly to items that would normally be difficult when they were facing the other person. These effects occurred spontaneously, and specifically in the presence of a human (but not a lamp), even when the other person was task irrelevant. Subsequent studies showed these effects were not sensitive to the looking direction of the other person, but instead reflected their position in space more generally. Finally, we found that a person’s perceived ability to move (and therefore motorically emulate the position of the other person) did not influence these effects. These findings show that (1) others’ perspectives can indeed ‘stand in’ for our own input, taking the form of a perceptual simulation, which paints the (imagined) perspective of another person onto our own perception, (2) these simulations reflect what someone could – in principle – see from their position in space, and speculate that (3) VPT emerged primarily to support navigation of, and interaction with, the environment and that more sophisticated Theory of Mind may then capitalise on the epistemic insight derived through more primitive forms of VPT.