Even though it is likely that no ‘original’ function of dreaming can be acknowledged, but rather a multiplicity of functions depending on specific research questions and segmentations of the dream space, one aspect of dreaming might distinguish TST and SST from other functions of sleep and dreaming, including other simulation functions: obliviousness of the avatar about being in a dream. Impaired insight into the own state of mind is a hallmark of normal dreaming, (Dresler et al. in press). The well-known exception of this symptom of most dreams is the case of lucid dreaming (Dresler et al. 2015), which in turn can be used to test whether state obliviousness is indeed a characterizing feature of TST and SST when compared with other dream functions.
There is no obvious reason why obliviousness about the dream state would be necessary for the memory function of sleep and dreaming. For procedural memory consolidation, lucid dreaming has even been suggested as a state that allows for a hyper-realistic mental training of recently learned motor skills (Erlacher & Chapin 2010). Several studies support this idea: lucidly dreamed training of coin tossing (Erlacher & Schredl 2010) or a finger tapping task (Stumbrys et al. in press) has been demonstrated to be effective, and a considerable number of professional athletes use lucid dreams to practice sports skills, with most of them having the impression that their performance is thereby improved (Erlacher et al. 2011). For the creativity and insight function of sleep and dreaming, obliviousness regarding the current state of mind is no prerequisite, and lucid dreaming has explicitly been suggested and shown to be used as a tool to increase creative processes (Stumbrys & Daniels 2010; Schädlich & Erlacher 2012; Stumbrys et al. 2014). As with non-lucid dreaming, lucid dreaming is associated with defocussed attention and flat association hierarchies—lucid dreams have even been reported to include even more uncommon and bizarre elements than non-lucid dreams (McCarley & Hoffman 1981). At the same time, regained reflective capabilities enable the creative dreamer to evaluate new associations and ideas, a step in the phase model of creativity that for non-lucid dreams is reserved for subsequent wakefulness. This mechanism is illustrated by two interesting case studies: Barrett (2001) describes the case of a painter who in his lucid dreams visited galleries, and then searching for interesting motifs to be painted soon after awakening from the lucid dream. A comparable strategy was used by one of our own study participants (Dresler et al. 2011, 2012), a music composer: when he aimed to compose a new piece of music, he turned on a radio in his lucid dreams and changed radio stations until he heard a composition that he considered interesting. He then woke himself up and wrote the new composition down. In line with these data, questionnaire studies reported that frequent lucid dreamers might be more creative than less-frequent lucid dreamers (Blagrove & Hartnell 2000).
For the emotion regulation function of sleep and dreaming the situation is less clear, however here there is also some evidence indicating that obliviousness is not generally necessary: for the case of positive affects, subjects often report that lucid dreams are associated with particularly positive emotions. And for negative affects, the successful use of lucid dreaming as a therapeutic tool in affective disorders indicates that dream lucidity does not interfere with the emotion regulation function of dreaming (Holzinger 2014).
In contrast, for those cases where a general emotion regulation function of dreaming overlaps with the TST, the necessity of staying ignorant about the true state of consciousness becomes obvious: to successfully serve as an authentic simulation of a threat, the dreamer has to take the threat as real and thus be oblivious towards his true state of mind. The cognitive insight that everything encountered consists only of hallucinated dream imagery and thus cannot harm the dreamer in reality immediately takes the sting out of the threatening experience. This mechanism has been successfully utilized for recurrent nightmares, where lucid dreaming has been demonstrated to be of therapeutic value (Spoormaker et al. 2003, 2006; Dresler et al. 2015; Rak et al. in press). Thus, for the threat simulation function of dreaming, obliviousness regarding the current state of mind is essential.
For SST, several lines of evidence indicate that obliviousness regarding the current state of mind is a prerequisite for social simulation to be effective. During normal dreams, non-self dream characters are attributed with feelings and thoughts just like in waking life (Kahn & Hobson 2005). Being oblivious about the true nature of these dream characters might ensure that non-perfect social simulations are also taken as autonomous agents instead of mere puppets controlled by the dreamer: dream characters are often implausible compared to their real-life waking counterparts (Kahn & Hobson 2003) , however, are nevertheless recognized and identified without major puzzlement (Kahn et al. 2000, 2002b). During a lucid dream, implausible dream characters might be treated less seriously by the dreamer, rendering the social simulation much less effective. This is illustrated by a recent study demonstrating that being tickled by an intentionally-controlled non-self dream character during a lucid dream was comparably ineffective as self-tickling during wakefulness, whereas being unexpectedly tickled by another dream character felt more ticklish (Windt et al. 2014). Non-self dream characters lead to different predictions depending on their perceived autonomy, and their respective simulation thus serves different functions. Lucid dreaming frequency correlates with the amount of control over the dream (Wolpin et al. 1992; Stumbrys et al. 2014), implying that frequent lucid dreamers would conceive dream characters as less autonomous than less frequent lucid dreamers. Thus, although non-self dream characters appear to have quasi-independent mental lives during lucid dreams (Tholey 1989), convincing training of social skills would require the dreamer to be oblivious to the fact that dream characters are not real, but hallucinated.
In summary, in contrast to other functions of sleep and dreaming, TST and SST essentially depend on state obliviousness of the dreamer. State obliviousness in dreaming might therefore be seen as a prime example of an epistemic constraint of phenomenal experience that leads to new and beneficial functional properties (Metzinger 2003). While both TST and SST (and other functions of sleep) might be applicable to humans and other social animals alike, state obliviousness might be a function that specifically developed in humans: it is unlikely that animals without sophistic language skills possess the ability to reflect on their current state of mind and compare it to alternative mind-states. In turn, such animals do not need a differential mechanism switching state reflectiveness on and off depending on the current vigilance state. Of note, neural correlates of state reflectiveness, i.e. lucid dreaming, strikingly mirror brain differences seen in humans vs. non-human primates (Dresler et al. 2013).