The Horizonal Structure of Visual Experience

ABSTRACT

How is it that we can visually experience complete three-dimensional objects despite being limited, in any given perceptual moment, to perceiving the sides facing us from a specific spatial perspective? To make sense of this, such visual experiences must refer to occluded or presently unseen back-sides which are not sense-perceptually given, and which cannot be sense-perceptually given while the subject is occupying the spatial perspective on the object that they currently are—I call this the horizonality of visual experience. Existing accounts of these horizonal references are unsatisfactory. In providing a satisfactory account, this paper argues that the content and structure of the visual experience of complete three-dimensional objects is as follows: the object is presented as being perceptible from yet-to-be-determined alternative points of view. As part of the content of visual experience, this motivates non-propositional attitudes of anticipation. Explicating this proposal is the central positive aim of this paper.

KEYWORDS:

• Visual Experience
• Horizon
• Perspective
• Action

Acknowledgements

I would like to thank Peter Poellner for reading over previous drafts of this material. I would also like to acknowledge the constructive comments and feedback of two anonymous referees at the Australasian Journal of Philosophy.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1 For similar statements of the problem see Kelly 2004: 98; Noë 2004: 60; Schellenberg 2007: 604, 613; Nanay 2009: 307–9; Church 2011: 36.

2 For discussion of this idea in classical phenomenology see Merleau-Ponty 1945: 6; Gurwitsch 1957: Part 4; Husserl 1973: §8, 1997: §16, §19 and §24, 1982: §41 and §44. For a discussion of Merleau-Ponty’s view see Kelly 2004: 74—110. For an overview of Husserl’s view see Drummond 1990: ch. 8.

3 Moore (1918: 220–241) expresses this view. Kelly (2004: 79–80) attributes it to Husserl, although cf. Poellner (2007: 440, fn. 48).

4 See Kelly 2004; Briscoe 2011; with a more empirical focus see Nanay 2022.

5 See Evans 1982: 123–4, on the belief-independence of perception.

6 For some empirical discussion which suggests that our perceptions of three-dimensional shape are insulated from beliefs see Keane et al., 2012. Also, Ekroll et al., (2018) showed that perceptual illusions involving (‘erroneous’) amodal completion persist even when subjects know the true shape of the object.

7 There is a different view that the relevant ‘hypotheses’ are states which figure at the subpersonal processing level of the visual system, and do so early on rather than in central cognition—this kind of view would likely be able to accommodate such cases of recalcitrance, although states which figure early on in visual processing are not usually thought to be belief-like (such an idea is found in predictive processing accounts of perception; see Hohwy 2007; cf. Marvan and Havlik 2021 for criticism of the idea that such accounts can non problematically be taken as theories of perceptual consciousness). However, as noted in the introduction, my interest is in perceptual experiences and their conscious components, as concerned with the way perceptual experiences might include horizonal references to non-facing sides.

8 See Church 2011: ch.2 for this active imagining view, aspects of which are also in Kant 1781, A120, fn. a, Strawson 1974, and Dummett 1993: 112. Contrastingly, see Brewer 1998: 23–4, who appeals only to a subject’s capacity to imagine such alternative perspectives. Bence Nanay (2009: 239–54), defends a different version, appealing to (non-phenomenal) representation of the relevant mental imagery. See Gregory 2017: section 5, for further critical discussion.

9 In separate work I consider a fuller range of objections.

10 See Schellenberg 2007. The idea that there is a constitutive connection between perception and action finds expression in Husserl and Merleau-Ponty (see fn.1), but also in Gibson’s ecological theory of perception (1979) and more recently in Evans 1982, Hurley 1998, and Noë 2004.

11 For discussion of empirical research which suggests that three-dimensional shape perception is closely tied to movement and the possibility of viewing an object from multiple points of view see Todd 2004.

12 Arguably subject-movement is not necessary to acquire the relevant sensorimotor profile; it would suffice for the subject to have remained at a fixed spatial location but the relevant object to have moved in such a way as to reveal the previously concealed sides, say by being rotated in physical space (see Noë 2004: 117, 119). I discuss this disjunctive conception of the acquisition of sensorimotor knowledge, and the issues it creates for the view, in separate work.

13 See Stanley and Williamson 2001 for criticism of the distinction between knowledge-that and knowledge-how. Whether this specific distinction holds, there is undoubtedly a significant difference between a subject possessing an ability to Φ (e.g., the competency to ride a bike) and a subject possessing and being able to express propositional knowledge concerning Φ-ing (e.g., bike-riding theory).

14 See Bennett 2012; 2016 for a similar argument against SMK views based on empirical research in vision science.

15 Noë’s (2004: 63) distinction between conceptual skills and more basic ‘perceptual presence’ might be a slightly different way of formulating these ideas and could be deployed in a similar way to deal with the objection. However, I take the presentation in the text to be a clearer formulation of this strategy.

16 One interesting upshot of reflection on highly determinate complex spatial-types like scutoids is that it gives the lie to any suggestion that when it comes to working out what is involved in the formation of some supposed maximally general sensorimotor profile for three-dimensional objects per se that sensorimotor expectations of spatial symmetry of non-facing sides with facing-sides we can see is essential.

17 Schellenberg (2007) distinguishes her view from the one which Kelly (2004) develops (and attributes to Merleau-Ponty) according to which in three-dimensional visual perception objects are perceived from points of view other than one’s own, but as different actual (rather than possible) viewpoints on the object. This proposal has the problem of explaining how these actual viewpoints, perceived in visual experience, can be unified into an overall perception of a particular three-dimensional object. There is also the following issue. According to Kelly (2004: 91) it is one’s perception of background objects, and their actual ‘points of view’ on the object (from their spatial location), that accounts for the horizonality of three-dimensional object perception. Yet this implausibly hamstrings three-dimensional object perception to there being other objects in one’s visual field. As will be clear in section 5, appeal to the spatial background is central to a plausible account of the phenomenon at issue, but this notion is developed in a different direction.

18 For discussion of transparency see Tye 2002 and Martin 2002.

19 So developed the proposal bears a similarity to ideas developed by Dominic Gregory concerning perceptual expectations in visual experience. Gregory rejects the idea that we should think of the experienced ‘externality’ of an object as turning on a reference to the way that things would look in the course of later visual sensations and rather appeals to the idea that there is a reference within the content of ordinary visual sensations to ‘ways that things actually then look from various perspectives’ (2015: 5), writing latter that ‘the presence of apparently external items within vision corresponds to our possession of expectations concerning the way that things look from other viewpoints’ (Gregory 2015: 17). Gregory (2017) goes onto develop similar ideas with reference to the connection between perceptual expectations and surprise, although relies less on an appeal to modal content.

20 Consider somewhat analogously, the role that understanding plays concerning the meaning of a declarative sentence about some empirical states of affairs, such as ⟨it is raining outside⟩. In such cases I plausibly have a consciousness of understanding, which takes the form of a kind of background ability consciousness. I am aware that I could verify what is said or could at least envisage what it would take to verify it, but I don’t actually have to do or envisage anything in order to understand the sentence.

21 Note: is it of course plausible that in most cases such non-propositional practical attitudes become sedimented in visual experience given past instances in which a subject in fact moved to occupy what then became a determinate alternative point of view. Past familiarity of moving through space to occupy alter-ego points of view undoubtedly accounts (at least partly) for having acquired the spatial know-how which constitutes a practical grasp of space.

22 We take it that its visual potential will not be exhausted by any particular perspectival appearance of it from a determinate location.