If what I have said so far is on the right track, then the question of whether dreamless sleep, at least on occasion, involves phenomenal experience is open to empirical investigation, and progress towards answering it can be made by applying the methods already used in scientific dream research, for instance by combining timed awakenings in the sleep laboratory with questionnaires that are carefully calibrated to direct participants’ attention towards the relevant features of such experiences and facilitate their reportability. Even occasional reports of dreamless sleep experience will support the claim that dreamless sleep experience exists. The next step towards turning the question of dreamless sleep experience into a scientifically tractable research project is to draw a more precise conceptual map of the territory. Sketching at least the rough outlines of such a conceptual map is my aim in this section.
Thompson’s reconstruction of the classical Indian debate as well as his own positive proposals for how to study dreamless sleep experience provide a helpful point of departure. To begin with, as Thompson points out, the concept of dreamless sleep itself requires phenomenological refinement (p. 13). If dreamless sleep experience exists, then it is not enough to characterize dreamless sleep by the absence of dreaming or its electrophysiological correlates. Rather, dreamless sleep can now be seen to be a blanket term covering different types of conscious and nonconscious mental activity. Some forms of conscious mental activity that are commonly contrasted with dreaming (and in this simple sense can be said to occur in dreamless sleep), such as hypnagogic imagery during sleep onset or repetitive and non-progressive types of sleep thinking, are not candidates for the kind of dreamless sleep experience described in the Indian debate. Dreamless sleep experience in this narrow sense, if it exists, is a form of phenomenal experience characterized by nonintentional awareness (Thompson 2015, p. 2): “When we’re deeply asleep […] we don’t cognize anything—there’s no object being cognized and no awareness of the ‘I’ as knower. Nevertheless, [...] we feel this absence while we sleep and remember it upon awakening” (Thompson 2015, p. 238). Dreamless sleep experience is not just characterized by the absence of certain object-directed forms of conscious experience, but by the fact that this is an experienced absence. Moreover, it is not just the objects of experience that are absent, but also the subject of experience, or the “I”. A very basic experiential feature, namely that of being an epistemic agent or a potential possessor of knowledge, has been lost (cf. Metzinger 2013 for a fuller discussion of the term of an “epistemic agent model”). Dreamless sleep experience is characterized by a dissolution of subject-object duality, or, to put a more contemporary gloss on this, by a breakdown of even the most basic form of the self-other distinction (Windt et al. 2014).
This last point is important because it suggests a way of differentiating between dreaming and dreamless sleep experience. Many different definitions of dreaming exist—indeed, the lack of a uniform definition is an important desideratum for theoretical and experimental work on dreaming—but work on dreaming in philosophy of mind often focusses on a structural feature of dream experience. The assumption that dreaming involves the experience of a self in a world marks a point of convergence for philosophers of different stripes, ranging from contemporary philosophers of mind working towards an empirically informed theory of dreaming (Metzinger 2003; Revonsuo 2006) to authors working in the tradition of classical phenomenology (Husserl 2006; Conrad 1968).[13] Studies have shown that an overwhelming majority of dream reports describe the presence of a dream self (Strauch & Meier 1996) though the precise way in which the dream self is represented is variable (Occhionero et al. 2005; McNamara et al. 2007). The description of dreams as involving not just a self in a world, but an intersubjective world has even informed theories on the functions of dreaming (see for instance Revonsuo et al.’s 2015 theory of dreaming as a simulation of social reality). Importantly, the description of dreaming as the experience of a self in a world also informs Thompson’s own work on dreaming. In Waking, Dreaming, Being, he tells us that “the core feature of full-blown dreaming is the experience of immersion in the dream world” (Thompson 2014, p. 127), and also that this immersive quality is exactly what distinguishes hypnagogic imagery during sleep onset from dreaming (pp. 135ff.). The hypnagogic state is a state of absorption, in which attention is fully captured by a series of dynamically changing and often short-lived images; “the hypnagogic state blurs the boundaries between inside and outside, self and world” (p. 124).
This description coincides nicely with my own theoretical work on dreaming. Elsewhere, I have argued that the analysis of self-experience is the key towards understanding not just different types of dreaming (Windt 2010, 2015, chap.s 11 and 12), but also the relationship between dreaming and waking experience. In this view, the common denominator underlying different types of dreams, such as lucid and nonlucid dreams, but also nightmares and false awakenings is their immersive quality. Even in simple forms of dreaming, there is still a sense of presence, a phenomenal here, or the sense of being located at a specific point in space, as well as a sense of duration centered on a phenomenal now. This basic structure is preserved even when the features that characterize a majority of dreams, such as interaction with non-self dream characters, objects, emotions, or even visual imagery are lost. In such minimal dreams, phenomenal selfhood takes the form of pure spatiotemporal self-location, arising independently of more complex forms of phenomenal selfhood that involve the experience of being a thinking self and embodied agent. There may even be the experience of phenomenal disembodiment, or of lacking a body, and the dream self may be experienced (and later described) as an abstract, undefined volume of indeterminate extension or even as an unextended point in space. Even though this sense of identification with a phenomenal here and now involves a drastically reduced form of phenomenal selfhood, it is still sufficient to ground retrospective claims of having had a self in dream reports. The basic structural feature of a self that is experienced as distinct from and located at a precise point within the world is preserved. To be sure, the locus of self-location and self-identification is more fluid in dreams than in wakefulness—the phenomenal here is subject to sudden shifts, and sometimes, we identify with a dream character or even a series of dream characters that are quite distinct from our waking self (Rosen & Sutton 2013). Yet, as long as there still is a world experienced as distinct from the self, at least a basic form of the self-other distinction continues to exist.
Within this framework, immersive spatiotemporal hallucination, or self-location with respect to a largely nonveridical, spatiotemporal reference frame, marks the cutoff line between dreaming and nondreaming. It also helps isolate and empirically ground minimal phenomenal selfhood (Blanke & Metzinger 2009), or the simplest conditions under which the experience of being or having a self arises. Here, I would like to suggest that this framework can be extended to dreamless sleep experience as well. A very basic point is that we can now sharpen the claim that dreamless sleep experience is a selfless state. Within the present framework, in order for dreamless sleep experience to count as selfless, even the basic form of self-other distinction that underlies spatiotemporal self-location must be lost. The next step is to consider the spatial and the temporal characteristics of self-location independently of each other and ask whether either of them, considered on their own, would be sufficient to give rise to phenomenal selfhood. An affirmative answer would mean that we had not yet identified the phenomenal signature of dreamless sleep experience; an even more simplified account would be needed.
Considering the spatial and temporal aspects of self-location separately, there seems to be a strong conceptual link between the phenomenal here and the sense of being located in and relative to a larger spatial expanse. A spatial reference frame, according to the present theory, turns into an experienced world when it is centered on a phenomenal here, which in turn is identified as the self. The spatial variant of presence thus seems to have the self, or some rudimentary form of self-other distinction, written into it. Moreover, the attempt to conceive of an experience characterized by a phenomenal here but lacking any temporal characteristics whatsoever strains the limits of conceivability. Speaking of an experience that is both instantaneous, lacking any temporal extension, and fails to have temporal location seems to be a contradiction. It is not clear how this could count as an experience at all, and even less how it could count as a reportable one.
By contrast, the phenomenal now does not appear to carry the same conceptual commitments. At least intuitively, the notion of a form of temporal experience that is independent of and perhaps more basic than the experience of being or having a self seems more acceptable than that of an immersive but nonetheless selfless form of spatial experience. Moreover, we can at least conceive, it would seem, of a phenomenal now that fails to be differentiated from or clearly located relative to a larger temporal reference frame.[14] And we can also, it would seem, conceive of an experience characterized only by temporal but not by spatial characteristics. Thinking, for instance, is not always experienced as having spatial location (as in thoughts occurring in one’s head), but it certainly has temporal dynamics.[15] Spatiality does not seem to be essential to phenomenality in quite the same way as temporality.
Note that I do not intend these conceptual considerations to carry too much weight. In the framework I am working towards, conceptual distinctions are informed by differences in the structure of phenomenal experience and such differences should at least in principle be memorable and describable, for instance in dream reports or reports of dreamless sleep experience.[16] I also think that the most empirically plausible view will allow for gradual transitions between states involving a phenomenal self and those retrospectively described as selfless; and the same may also be true for the emergence of the simplest forms of phenomenal experience. If this is correct, then we should expect there to be a certain amount of uncertainty when dealing with borderline cases. Where exactly to draw the cutoff line for minimal phenomenal selfhood in a given case may well be hard (if not impossible) to determine; but even so, it might still be useful to introduce a conventional cutoff line (for instance by saying that minimal phenomenal selfhood involves both the spatial and the temporal aspects of self-location) if this helps pick out a theoretically meaningful transition in the structural features of experience and guides future research in a constructive manner. We will also expect such a theoretical conception to be reasonably well aligned with the way such experiences are described in retrospective reports.[17] I think that both types of considerations support the claim that spatiotemporal self-location can be meaningfully described as a minimal form of phenomenal selfhood, or at least as a theoretically salient point of transition on the trajectory from states described as selfless to states involving self-experience in a fuller sense. By contrast, the phenomenal now, when it arises independently of spatial self-location, is a candidate for a structural feature of phenomenal experience that provides the conditions of possibility for self-experience but that when occurring on its own is still prior to it. I would like to suggest, then, that pure subjective temporality is a candidate for minimal phenomenal experience; it is a condition for but still more basic than minimal phenomenal selfhood. It can be described as subjective only because it involves phenomenal experience; yet, it does not involve the additional experience of being a self, or a separate entity having the experience.[18]
There is, of course, a rich philosophical debate on the nature of time experience, as well as a large empirical discussion (for an introduction, see Dainton 2010; Arstila 2014; LePoidevin 2015). I cannot begin to do justice to this literature here, but want only to focus on one specific aspect. This is the idea, which we find in William James as well as in Husserlian phenomenology, but also in the neuroscience of time consciousness (see for instance Pöppel 2003), that even the smallest unit of temporal experience, the temporal now, is extended rather than instantaneous.[19] Following this conception, a rudimentary form of duration would be intrinsic to the phenomenal now; and neuroscientific work seems to suggest that this temporal now is itself variable (Wykowska & Arstila 2014). The window of simultaneity, or the maximum time-frame within which two different events are experienced as occurring now, is modality-specific. The cutoff line for two stimuli being experienced as simultaneous is, for instance, larger for visual stimuli than for auditory ones. As Wykowska & Arstila (2014, p. 443) note,
it might be that a relatively broad window of simultaneity is actually beneficial. The human brain needs to exhibit some degree of tolerance to asynchronous stimuli in order to be able to bind different sensory inputs into one event. The window of simultaneity can be seen as an integration window for stimuli and, as such, is a necessary mechanism for binding signals from different pathways into one single object or event.
Human temporal resolution is flexible, it is easily affected by attentional processes as well as by training and expertise (Wykowska & Arstila 2014). Duration perception might be state-dependent as well, showing characteristic changes in altered states of consciousness and psychiatric disorders (Noreika et al. 2014); and perhaps the same is true for the degree to which the experienced now itself is stretched in time. There also seems to be a close relationship between changes in time perception and alterations in self-experience. When the self becomes the focus of attention, when we attend to our current mental or emotional state, or to bodily sensations (such as hunger or pain), time seems to slow down; by contrast, when we are thoroughly absorbed in an activity, time contracts and seems to move faster (Wittmann in press). When self-experience is lost, as in selfless states, the loss of a reference point may be associated with feelings of timelessness (Wittmann in press); the phenomenal now is stretched indefinitely. There is a sense of duration, but the sense of succession, of there being a chain of present moments, has been lost.
Importantly, this way of thinking about subjective temporality and the experienced now is one which Thompson (2015) endorses. He explicitly appeals to the Husserlian conception of time experience in his defense of retrospective reports of dreamless sleep experience. Here, he suggests that memories of dreamless sleep experience may be grounded by retentional awareness, “the holding onto the just-past as an intentional content belonging to our consciousness of the passage of time, including our mental lives as flowing in time” (p. 9). Because temporal experience has the retention of the immediate past and protention, or the anticipation of the next moment, written into it, the moment after awakening still carries with it the traces of dreamless sleep experience: “Immediately, the ego sense appropriates the lingering impression or retention of not-knowing and associates this retention with itself, thereby generating the retrospective thought, ‘I did not know anything’” (p. 10).
In Mind in Life, Thompson (2010) endorses a version of Husserl’s conception of time-consciousness according to which the streaming, flowing character of subjective experience is both the “condition of possibility for every other kind of consciousness, but is not itself made possible by some other, still deeper level of consciousness” (p. 324). This absolute flow of consciousness is self-constituting (p. 324); it is also prior to and essential for phenomenal selfhood. As Thompson (2010) puts it,
to be aware of phenomena across time, consciousness must be retentionally and protentionally aware of itself across time. Therefore, time-consciousness entails prereflective self-awareness. In other words, our being conscious of external temporal phenomena entails that our temporally enduring experiences of those phenomena are self—aware. Inner time-consciousness is thus nothing other than prereflective self-awareness. (p. 328)
This prereflective awareness that consciousness has of itself (its self-luminousness, reflexivity, or self-acquaintance[20]) is not yet the same has being or having a phenomenal self in the sense used here. Rather, this minimal form of phenomenal experience is the condition for the emergence of minimal phenomenal selfhood.
My suggestion, then, is that we can enrich our theoretical conception of dreamless sleep experience by applying Thompson’s account of how we remember dreamless sleep experience (namely with the help of retentional awareness) to the description of dreamless sleep experience itself. Dreamless sleep experience involves pure subjective temporality that is not yet structured around intentional objects, including a phenomenal self. As Thompson (2015) puts it, “although deep sleep creates a gap or a rupture in our consciousness, we often feel the gap immediately upon awakening. […] We are aware of the gap from within our consciousness” (p. 4). Just as upon awakening, I am directly aware that it was I who was asleep and unknowing, I am typically aware that a certain (though perhaps indefinite) amount of time has passed. Following Proust’s more poetic formulation in the passage quoted by Thompson,
a sleeping man holds in a circle around him the sequence of the hours, the order of the years and world. He consults them instinctively as he wakes and reads in them in a second the point on the earth he occupies, the time that has elapsed up to his waking; but their ranks can be mixed up, broken. (p. 3)[21]
We might even say that metaphorically speaking, subjective temporality provides a reference frame that is still empty, but poised to integrate and lend temporal structure to intentional contents such as thoughts, objects and events, but also the self, as they arise—for instance by imposing sequential order on them and representing some of them as simultaneous, and others as successive. Yet, this form of temporality is more basic than the events it later integrates; it predates them and provides a space in which they can appear.
Incidentally, this idea fits in nicely with the Vedantan view that, “deep sleep is a kind of ‘ground state’ of consciousness, a lowest-energy state from which the ‘excited states’ of dreaming and waking arise” (Thompson 2014, pp. 260-261). Again, deep sleep is the baseline, the causal source from which other conscious states arise; it is also called “seed sleep”, because it is thought to contain the seeds of both dreaming and waking consciousness. Perhaps we can begin to make sense of this idea by saying that dreamless sleep experience, understood as pure subjective temporality, is a candidate for minimal phenomenal experience.[22]
How can we make progress on identifying real-world cases of dreamless sleep experience? Importantly, if the account of dreamless sleep experience defended here is even remotely correct, we should not expect dreamless sleep experience to be restricted to experienced meditators. Instead, dreamless sleep experience might be fairly prevalent even in people without any formal training in contemplative traditions. This approach requires disambiguating between at least two variants of the target phenomenon. Note that within the Indian conception of dreamless sleep, we can distinguish between an insight component and a more basic experiential component. The insight component refers to the ability to become aware, during sleep, of the nature of this state. This is not necessarily a conceptually mediated form of knowing that you are currently sleeping dreamlessly, but rather consists “in being able to witness the state of dreamless sleep and recall its phenomenal clarity upon awakening” (Thompson 2015, p. 15). Still, even this nonconceptual form of witnessing is not epistemically neutral, but can lead (or fail to lead) to veridical retrospective reports. To be sure, this form of insight or awareness itself can have a particular phenomenal feel—it bears the phenomenal signature of knowing (Metzinger & Windt 2014, 2015), the feeling of just having become aware of the nature of one’s ongoing state—but importantly, this type of phenomenal experience carries with it epistemic commitments. My feeling of knowing can be true or false. It also seems plausible, as suggested by Thompson, that meditation facilitates this type of lucid dreamless state, or perhaps could even be a way of inducing it systematically.
But the model of dreamless sleep experience as pure subjective temporality also points to a more basic experiential component that as such bears no obvious connection to an epistemic state of knowing or of being aware of the nature of the state one is currently experiencing. Dreamless sleep experience in this primary phenomenological reading refers to a kind of experience during sleep; but this does not require the ability to conceptualize this as a form of sleep experience. In principle, you can have dreamless sleep experience without realizing that you are asleep: dreamless sleep experience is a form of experience occurring in sleep, but it is not necessarily an experience of sleep as a state of sleep. It might enable us to estimate how long we have slept, but it can also be misleading, maybe even leading us to misjudge whether we have slept at all. This is particularly obvious if dreamless sleep experience is construed as an answer to the question of how we know, upon awakening, that we slept peacefully (Thompson 2015, p. 4). Thompson’s reconstruction of the Indian debate, taken together with my analysis, suggests that because this state is characterized only by its temporal character, we have the sense of there being a gap between two periods of wakefulness; and because this gap is devoid of intentional objects, we describe it as peaceful. Yet, this does not seem to require that we were aware of (or took ourselves to be aware of) the nature of this state while it was occurring, namely during sleep. If any sophisticated epistemological reading of insight were indeed crucial to dreamless sleep experience, the experience of having slept peacefully would have to be reserved for special subject groups, such as experienced meditators—and it would be quite mysterious why clearly, it is not.
Perhaps we can model the relationship between the epistemic and the phenomenological components of dreamless sleep experience on the relationship between lucid and nonlucid dreaming. Thompson (2015, p. 15) himself explicitly contrasts lucid dreaming, or knowing that you are dreaming while you are dreaming, with lucid dreamless sleep. Given this suggestion, a good place to begin the project of broadening the investigation of dreamless sleep experience beyond expert meditators is to consider reports from experienced lucid dreamers.