What should the sensorimotor enactivist say about dreams?

ABSTRACT

Dreams provide a compelling problem for sensorimotor enactivists like Alva Noë: they seem to replicate our perceptual experiences without sensorimotor interaction with distal sensory stimuli. Noë has responded by saying that dreams actually fail to replicate perceptual experiences in virtue of their lack of detail and stability. Noë's opponents have replied by pointing out that some dreams are richly detailed and stable, and that instability and a lack of detail in dreams can anyway be explained in terms of the underlying neural activity. In this paper, I develop how the sensorimotor enactivist should respond: dreams fail to replicate perceptual experiences because they are exhausted by what shows up at a given moment in phenomenal consciousness, while perceptual experiences go beyond this to include everything accessible via sensorimotor exploration. This difference permeates all levels of experience, so that dreams can't even replicate perceptual experiences of simple shapes and colors. Further, unlike detail and stability, there are not obvious neural explanations of this phenomenal difference.

KEYWORDS:

• Dream phenomenology
• sensorimotor enactivism
• presence
• neural internalism

Acknowledgments

This paper owes much to an anonymous referee (or referees) at this journal, who gave multiple detailed sets of helpful comments, the first of which lead to substantial revisions. These comments helped me both clarify my ideas and better situate them with respect to Noë's work. Earlier drafts of this paper were read by Mark Fortney and Mohan Matthen, who both gave helpful feedback. An earlier draft of this paper was also presented at Dreams, Hallucinations and Imagination, a workshop by the Centre for the Study of Perceptual Experience at the University of Glasgow (February 23–24, 2019). Thanks to the participants and organizers of this workshop, especially Melanie Rosen, for their helpful feedback.

Disclosure statement

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

Notes

1 By ‘sensorimotor interaction’ (or sometimes: ‘sensorimotor exploration’) I mean physical actions directed at a distal sensory stimulus mediated by a subject's implicit grasp of how movement affects their phenomenal consciousness. Shifting my focus from my pencil's tip to its eraser in order to bring the eraser into better view is a sensorimotor interaction, and so is turning the pencil to see its backside, running my thumb over its barrel to feel its surface, and moving my head to see how the 2D projection of the pencil in my field of view changes in order to ascertain the pencil's 3D shape.

2 Some relationalists (e.g. Campbell 2002; Fish 2009) claim that perceived objects are constituent parts of the phenomenal character of the experiences in which they are perceived, implying that swapping one object for a qualitatively identical twin will necessarily change phenomenal character (see Mehta 2014; French and Gomes 2019). While Noë thinks that perceived objects are constituent parts of the perceptual experiences in which they are perceived (see Noë 2007, 465), he doesn't generally talk as if these constituent objects affect phenomenal character. He says (Noë 2007, 464) “What makes it the case that our experience has the [phenomenal] character it does are the regularities governing our manner of interaction with … objects.” These regularities concern an object's properties, not its particularity (i.e. its identity).

3 Despite this talk of constitutive dependence, Noë (2016) says that perceptual experiences are not identical to, and don't reduce to, skilled sensorimotor interactions. So presumably he takes them to emerge out of, or supervene on, these interactions. It's not clear how to square all of Noë's claims.

4 There are representationalist views which similarly make phenomenal character dependent on what's outside the head (e.g. Dretske 2003), but don't face any challenge from dreams because the extraneural substrate posited on these views (a history of causal interaction with sensory stimuli) is still realized while dreaming (Block 2005, 264).

5 Perhaps it's then a challenge for the sensorimotor enactivist to explain how we could have perceptual experiences from within dreams, but this challenge would be distinct from the one we're addressing.

6 Noë is not the first to make this claim. Austin (1962, 42) famously wrote that “we all know that dreams are throughout unlike waking experiences”.

7 Noë also suggests (Noë 2007, 472) that perhaps dreams can replicate perceptual experiences, but only because of previous sensorimotor interaction with the environment while awake. The problem, as I explain below, is that Noë's account entails that phenomenal character depends on sensorimotor interaction with the things currently being experienced. Past interactions don't help.

8 Following Windt (2013), I assume that dream reports are reliable.

9 On reflection, the dream was probably more incomplete than this report suggests. I recall that the trash I picked up included styrofoam bowls which were missing their right halves, as if they had been cleanly cut. While I was aware of what was being said, I didn't actually auditorily experience most of it; instead, I simply somehow just knew what was being said. Experience of my own body was lacking as well. As I walked I didn't feel my body or any sense of exertion. At no point did I visually see (say) my arms and hands, or my nose, either.

10 Noë says (Noë 2007, 471) that “… dream experiences, whatever their nature, are not of the same basic kind as perceptual experiences. Phenomenologically, perceiving is for us an encounter with situations and things; it is not, for us, an encounter with mental images or some other kind of interior data of sense.” My proposal is a way to understand these brief remarks. The completeness of perceptual experiences and the incompleteness of dreams explains why the one phenomenally presents itself as an encounter with the world and the other phenomenally presents itself as an encounter with mental images.

11 This idea is from G.E. Moore's famous off-hand remarks on transparency (Moore 1903) and is adopted by both some neural representationalists (e.g. Dretske 2003, 67) and relationalists (e.g. Fish 2009, 10). The idea is that, when we try to describe what it's like to experience a stimulus, all we're able to do is describe the stimulus itself. Noë's explanation of phenomenal character in terms of objective stimuli features suggests that he endorses this idea (Noë 2004, 82–86, 123).

12 Braun et al. (1998, 91), recording during REM sleep, found depressed activity in the primary visual cortex, while Horikawa and Kamitani (2017b, 3), recording during sleep-onset (NREM sleep), were able to decode dreams from this area. Presumably the area is more active at sleep-onset.

Additional information

Funding

This work was completed during at postdoctoral fellowship at the University of Toronto, Mississauga supported by Canada Research Chairs.

Notes on contributors

Michael Barkasi

Michael Barkasi is a philosophy instructor at York University. His research focuses on how perception connects us to the external world and the relationship between experience and the brain.