In Dreamless Sleep, the Embodied Mind, and Consciousness, Evan Thompson retraces the steps of the classical Indian debate between the Advaitins and the Nyaiyāyikas on the occurrence of conscious experience during dreamless sleep (see also Thompson 2014, chap. 8). The classical Indian debate is important, according to Thompson, because if the Advaita Vedānta and Yoga claims about the persistence of consciousness during dreamless sleep are correct, the default view of consciousness as that which disappears during dreamless sleep is false and requires revision. In this section, I briefly reconstruct Thompson’s main arguments and sharpen the precise points of agreement and disagreement in the classical Indian debate, as well as their overlap with questions raised in cognitive science and contemporary philosophy of mind. I also introduce three challenges to Thompson’s view.
Thompson’s reconstruction of the classical Indian debate starts out from a deceptively simple question: How, after awakening from sleep, do we know that we have slept peacefully? The Yoga and Advaita Vedānta schools argue that retrospective reports of having slept peacefully are memory reports: we directly and non-inferentially remember (and hence are able to report) a state in which we were phenomenally conscious, but did not experience any particular thoughts or images. Dreamless sleep experience is, in this view, devoid of intentional content; it is a state of knowing nothing and at least in principle, it can be remembered and accurately reported upon awakening. The Nyāyas disagree, arguing that reports of having slept peacefully are inferential. Their point is that if dreamless sleep involves a particular form of ignorance, or of not-knowing, this not-knowing cannot itself be known, either during sleep or retrospectively. Because the means for knowledge are lacking during dreamless sleep, we can at best infer, when we wake up feeling refreshed and remember nothing, that we must have slept peacefully.
As Thompson (sec. 3) points out, the classical debate about conscious experience during dreamless sleep has to be seen in the larger context of how these schools construe the relationship between consciousness and the self. For the Nyāyas, consciousness is an adventitious property of the self, meaning that the self can persist throughout sleep even when consciousness ceases. They also claim that cognition always involves taking something as its object, where this object is necessarily distinct from the cognitive state itself. This view is compatible with the occurrence of object-directed thought and dream-related imagery during sleep, but prohibits the occurrence of objectless cognitive states. For the Advaitins, the situation is different. Because for them, the self is pure, reflexive (or self-luminous) consciousness, they cannot allow that consciousness can disappear altogether even during sleep, because this would entail a disappearance of the self. Unlike the Nyāyas, the Advaitins do not, however, take consciousness to be necessarily object-directed. Instead, they regard the essentially reflexive and self-luminous character of consciousness and the self as separate from and indeed as the very condition of object-directed thought. A prediction would be that “pure” cases of reflexive, self-luminous consciousness should occur even in the absence of object-directedness, for instance during sleep.
Despite these differences, the debate on dreamless sleep experience unfolds before a background of mutual agreement. Both schools agree, for instance, that object-directed consciousness can (and does, for instance in the form of dreams) occur during sleep, but also that it does not persist throughout sleep. Both also agree that dreamless sleep is a state in which object-directedness is lost. And finally, both agree that the self persists throughout dreamless sleep, even in the absence of object-directedness. Their disagreement thus hinges, first, on what exactly it means to say that the self persists during dreamless sleep, understood in the sense of a state in which object-directed thought is lost, and second, on how to construe the relationship between consciousness, the self, and memory reports. Both points are relevant, as we will see, for assessing the relationship between the Indian debate and contemporary research as well.
How, then, to adjudicate between the two sides in the debate? Thompson (p. 6) reconstructs the Nyāya claim that our knowledge of dreamless sleep is inferential as involving a five-step syllogism. His discussion of the Nyāya syllogism is already so clear that nothing would be gained from rehearsing it once more here. Instead, I want only to recall to readers’ attention that Thompson’s reconstruction of the Adavaitin response shows the Nyāya syllogism to be inherently fallacious: it is either circular or results in an infinite regress. In order to infer from the fact that I was in a special state that I knew nothing in this state, I must first have a reason for saying that I was indeed in a special state; and if this reason is that I knew nothing in this state, I am presuming what is supposed to be shown and the argument is circular. Alternatively, if I say that the means for knowledge were lacking in this special state, for instance because the mental faculties and the senses were inactive, then this further claim has to be backed up by independent evidence. Saying that I felt refreshed upon awakening will not do—for in order to know that feeling refreshed after awakening is correlated with the inactivity of the mental faculties and the senses during sleep, I would either once more have to appeal to memory (which, on pains of circularity, I cannot do), or I would be headed for an infinite regress. Thompson sums up his critique of the Nyāya syllogism by formulating a general principle:
More generally, the only way I can know that the means for knowledge were absent in deep sleep is by knowing that there was no knowledge present in this state. Only by knowing the effect—my not knowing anything—can I infer the cause—the absence of the means for knowledge. So unless I already know what the inference is trying to establish—that I knew nothing—I cannot establish the reason on which the inference relies. (p. 7)
The Advaitin view offers an easy way out. As Thompson points out, it can be reconstructed as involving the phenomenological claim
that when I wake up from a dreamless sleep, it seems that I can sometimes knowingly say I have just emerged from a dreamless sleep, and this saying seems to be a reporting of my awareness, not the product of having to reason things out. (p. 8)
At least in principle, the subjective impression of having awakened from dreamless sleep can be reflected in veridical reports of awareness during dreamless sleep.
It is important to see that Thompson’s assessment of the Indian debate does not lead to a whole-hearted endorsement of the Advaitin view; the view he promotes is in fact much more subtle, and also more humble. Thompson’s main goal is to establish the logical possibility of dreamless sleep experience. For this, it is sufficient that veridical memories of having slept dreamlessly are possible in principle (p. 5, p. 9). He also explicitly allows that there could be cases in which one’s memory of having slept peacefully and dreamlessly is mistaken. Thompson’s view is also weaker than the Advaitin position in that it is not committed to the persistence of conscious experience throughout sleep, but leaves room for periods of unconsciousness during sleep. According to Thompson, the mere possibility of dreamless sleep experience challenges the default view and highlights the need for a refined taxonomy of sleep states, because such a refined taxonomy is the condition for investigating dreamless sleep experience experimentally (p. 3).
To be sure, Thompson also offers some factual evidence for thinking that dreamless sleep experience actually exists: experienced meditators report witnessing or becoming lucid during dreamless sleep, and they show a changed pattern of EEG activity during slow wave sleep. Meditative training may, as Thompson suggests, facilitate cognitive access to the state of dreamless sleep (p. 11) and with it, more accurate reports. But his main point is that conceptual and empirical questions about dreamless sleep experience are well worth asking and that in order to do so, prominent theories of sleep, but also of consciousness (such as Tononi’s Integrated Information Theory; see Tononi 2008) should at least make room for the possibility of its occurrence and require revision.
While I find Thompson´s case for the logical possibility and conceptual coherence of dreamless sleep experience compelling, I worry that its humility makes it vulnerable to three related objections. A proponent of the default view could acknowledge that veridical reports of dreamless sleep experience are logically possible but could insist that unless such veridical reports are identifiable and can be distinguished from nonveridical ones, such reports cannot be used for the experimental investigation of dreamless sleep experience, or only in a very small and admittedly special group of highly trained subjects. Thompson’s own suggestions for the future investigation of dreamless sleep experience assume that this basic problem has been solved. For instance, he proposes that because dreamless sleep experience is supposed to be devoid of intentional objects, asking participants to report anything that was going through their minds before awakening, which is a question about the objects of awareness or the contents of consciousness, might be poorly suited to the target phenomenon. A good alternative, he suggests, would be to direct participants’ attention to the phenomenal character of sleep itself, for instance by asking them to report any feelings or any qualitative states experienced before awakening (p. 12). Here, the proponent of the default view might object that this strategy falls short of a methodology for investigating dreamless sleep experience: In order to use reports of dreamless sleep experience as evidence, some rationale for distinguishing veridical reports from nonveridical ones is needed. Without this, the large-scale revision of standard sleep-state taxonomy demanded by Thompson may seem premature; Thompson’s case for the mere possibility of dreamless sleep experience lacks the empirical grounding and research methodology to justify such a move.
A related problem is that in order to empirically investigate the occurrence of dreamless sleep experience, it is not enough to identify veridical reports of such experiences and distinguish them from nonveridical ones. Instead, in order to determine the frequency of dreamless sleep experience, one has to determine whether subjects can reliably report not just the presence of dreamless sleep experience, but also its absence. This problem is especially pronounced because Thompson’s claim is not that experience persists throughout sleep. As we saw earlier, his view departs from the Advaitin claim in that he thinks that dreamless sleep experience occurs only occasionally and contrasts with periods of genuine unconsciousness during sleep. A report-based methodology for investigating dreamless sleep experience will consequently have to assume not only that reports of dreamless sleep experience reliably indicate the presence of such experience during the preceding sleep period, but also that the absence of such experiences can be reliably reported, or at least that it can be inferred from the absence of reports of dreamless sleep experience. Unless this second condition is fulfilled, reports of dreamless sleep experience could be highly reliable in that they occur only when dreamless sleep experience was in fact present during the preceding sleep period, but could nonetheless fail to be sensitive to its actual frequency, for instance by only following a small proportion of such sleep experiences (for a discussion of the reliability and sensitivity of first-person reports, see Fink unpublished manuscript).
Thompson himself shies away from both commitments. In fact, he casts doubt on the assumption, common in cognitive neuroscience, “that a content of consciousness is a reportable content, and that reportable contents are ones that can be attentionally selected, held in working memory, and used to guide thought and action” (p. 12). Relatedly,
the general point that retrospective oblivion does not prove a prior lack of consciousness must be kept in mind whenever we are tempted to infer that consciousness is absent in deep sleep because people report not being able to remember anything when they are woken up. (p. 11)
Here, he might be read as effectively denying the possibility of using retrospective reports as a source of evidence for the scientific investigation of dreamless sleep experience. Moreover, given these doubts about the reliability and sensitivity of retrospective reports, Thompson’s (p. 17) proposal that meditation makes positive occurrences of dreamless sleep experience accessible to verbal report is not enough; a proponent of the default view could object that expertise of the relevant type is acquired only if meditation enables periods of unconscious sleep to be retrospectively reported as well (or at least to be measured indirectly through the inability to report conscious experiences from the preceding sleep period).
Finally, a proponent of the default view might grant that reports of expert meditators are more trustworthy than those of laypeople in both respects: meditators can report both when dreamless sleep experience was present and when it was absent.[4] Yet, it could still be objected that the example of expert meditators is simply too remote to justify the large-scale revision of sleep-state taxonomy that Thompson has in mind. For all practical purposes, or so the objection might go, the default view of consciousness and dreamless sleep as diametrically opposing and mutually exclusive states stands.
To be clear, I do not think these objections are particularly worrisome; but I do think they help set the agenda for how best to develop Thompson’s view, defend it against skeptical objections, and place it on broader empirical grounding. The first step, taken in the next section, is to introduce a stronger defense of the trustworthiness of reports of dreamless sleep experience, as well of reports of its absence. If successful, this provides a sound methodological basis for the experimental investigation of dreamless sleep experience. The second step is to provide a broader theoretical and empirical basis by proposing a conceptual framework of dreamless sleep experience as well as additional candidates for its future investigation.