One consequence of not distinguishing between perspectival phenomena is that the notion of a first-person perspective becomes ambiguous. One can clearly see this ambiguity in descriptions of the role of first-person perspective in the multisensory stimulation protocols developed in recent work on the neuroscience of bodily self-consciousness. These protocols all involve participants being touched on their torso whilst visually observing a body-shape (either the body of another person, a mannequin, or a virtual body) being touched on its torso. The protocols differ along two dimensions: the side of the torso stimulated and the location of the origin of the participants’ line of sight with respect to the body being observed. In one protocol, the body-swap illusion, participants are stroked on their chest whilst they look at a body being stroked on its chest from a position located where its head would be (cf. Ehrsson 2007; see Petkova et al. 2011b; Petkova & Ehrsson 2008; Petkova et al. 2011a). In another protocol, the full-body illusion, participants are stroked on their back, whilst they observe a body from behind being stroked on its back from a position entirely removed from its location (Ionta et al. 2011; Lenggenhager et al. 2007; Pfeiffer et al. 2013). The body-swap illusion protocol is often distinguished from the full-body illusion protocol as involving first-person perspective as an independent variable (Petkova et al. 2011a). However, recent work on the full-body illusion has demonstrated effects that the authors describe as changes in first-person perspective (Pfeiffer et al. 2014): Participants lain prone whilst feeling and observing strokes on the back report experiences of either looking up or down at the body they observe (Ionta et al. 2011). These variations in report seem to depend upon the individual’s relative weighting of vestibular and visual gravitational cues (Pfeiffer et al. 2013).
Admitting the differentiation of perspectival phenomena allows us to make sense of the differences in use of the term first-person perspective. In the terms introduced in the previous section, the first-person perspective in the body-swap illusion is an origin perspective. It presents the typical view of one’s own body with a line of sight originating in the head. The first-person perspective in the full-body illusion is an egocentric perspective. It forms the centre of an egocentric frame of reference, according to which the observed body occupies a location in a particular egocentric direction (up or down). Distinguishing these forms of first-person perspectival experience reveals that each of these protocols facilitates manipulation of a distinct form of perspectival experience. It also sheds light on the fact that the differences in vestibular and somatosensory processing between these forms of perspectival experience have yet to be compared.
One way of conducting such a comparison would be to use virtual reality display techniques to present an individual with two avatars in series, whilst measuring time-locked vestibular evoked potentials via scalp EEG.
Experiment 1: Participants are stroked on both their chest and their back whilst supine, whilst wearing a head-mounted display. In the meantime, participants observe either the chest of Avatar 1 being stroked on its chest, presented from a position corresponding to the avatar’s head, as in the body-swap illusion, or they observe Avatar 2 being stroked on its back, as in the full-body illusion. Ideally, the two avatars are presented in the same viewing, such that the participant views one avatar and then in a continuous movement shifts their gaze to view the other.[6]
I have claimed that each of the two protocols conjoined in this proposed experiment facilitates manipulation of different forms of perspectival experience. If this is correct, then finding significant differences in vestibularly-evoked potentials between observation of Avatar 1 and Avatar 2 would be a first step in determining differences in vestibular processing between these forms of perspectival experience.
As noted earlier, there do seem to be individual differences in the contents of egocentric perspectival experience in the full-body illusion. This would suggest that some individuals, those who are more heavily dependent upon vestibular gravitational cues to determine orientation, would experience themselves as looking upwards at Avatar 2. Whereas if the right visual gravitational cues were provided, some individuals may experience themselves as looking downwards at Avatar 2 (Ionta et al. 2011; Pfeiffer et al. 2013). This might allow the investigation of the relationship between egocentric perspectives and egomotion perspectives, by incorporating a second phase into a new experiment:
Experiment 2: Phase 1: experiment 1, described above. Phase 2: Participants continue to be stroked on their back and chest. Participants fixate upon Avatar 2 and observe it rotating about a horizontal axis, whilst being visibly stroked on its back and chest. Both reports of experienced orientation (upward vs. downward) and reports of experienced egomotion are gathered.
Participants may experiences themselves as rotating around a horizontal axis in just the way they observe Avatar 2 rotating. Alternatively, they might experience themselves as revolving around Avatar 2. In particular, what would be of interest would be the way in which any resultant illusory experiences of egomotion might correlate with experienced egocentric orientation (upward vs. downward). Moreover, individual differences in experienced egocentric orientation might even predict the contents of experienced egomotion. This would be a major step in determining both the relative influence of vestibular processing on these forms of perspectival experience and the relationship between these forms of perspectival experience.