Nightmares

The definition treats nightmares as perception rather than production, which is the whole of its departure from ordinary accounts. Nothing is being invented. What arrives is a glimpse of earthbound reality that the endofilter would normally exclude, and the filter relaxes during sleep. Marsh needs an endofilter precisely so that a person can conduct their lives without being distracted by the ghostly endoselves without, so the filter is doing necessary work and its relaxation is not a malfunction but a lapse in a service. The word earthbound is doing specific work too, since it points at the ring of decoupled but undispersed animus rather than at anything general, which is the same territory as ghosts, spirits attached to places and to the living, and the emotional wills that occupy people. Her broader position is that hallucinations and dreams are very likely representations of reality that the brain is not effectively filtering rather than fabrications, and imagination is glimpses of conscious outside personal perspective, often appearing in dreams or a trance state. So nightmares, dreams and imagination form one family in her account, distinguished by what is being glimpsed and how unpleasant it is. Whether the reader accepts the ontology or not, the structural claim is separable and clear. What appears during sleep is filtered differently rather than generated differently.

Dream research offers several accounts, with the activation-synthesis model treating dreams as the cortex making sense of random brainstem activity, threat simulation theory proposing that nightmares rehearse danger and the continuity hypothesis holding that dream content reflects waking concerns. Clinical work treats nightmare disorder and its association with trauma, with imagery rehearsal therapy as a standard intervention. All of these treat dream content as generated by the sleeping brain. Marsh's account reverses the direction and treats it as received, which is a metaphysical claim rather than a psychological one and would not be accepted within the field, though her structural point that the difference lies in filtering rather than in production has a counterpart in predictive processing accounts of perception.