- seCtIon FIVe: BoRDeRLAnDs
monitoring of phenomenal experience, which together plan and co-create the
dream: ‘As it develops, the unconscious planning simultaneously becomes the con-
scious syntactic organization of the dream’ (Cicogna and Bosinelli, 2001, p. 34).
The idea is that ‘the iterative feedback mechanism constructs successive drafts
of the dream [. . .] at an unconscious level with only the end product, the dream
itself, being accessible to awareness’ (p. 34). The top-down processes make the
first draft, the memory elements activate or inhibit other elements, and so dif-
ferent versions are created. Although the authors agree with Dennett in reject-
ing the idea of a central controller, they still describe the operations involved in
dream generation as ‘unconscious’, and so end up having to ask what the function
is of the conscious processes in the dream. And because only the final product is
‘conscious’, we must also still ask what it is that makes the difference.
Returning to the question of timing, there is still an interesting conflict in the find-
ings described above. On the one hand, we know that dreams occur in real time;
on the other, we know that people often wake from dreams in which the event
that woke them fits the end of a long dream story. How can this be?One way of explaining this, very much in the spirit of multiple drafts theory, is the
retro-selection theory of dreams (Blackmore, 2004). During REM sleep numerous
brain processes are going on at once, none of which is either in or out of conscious-
ness. On waking up, a story is concocted by selecting one out of a vast number of
possible threads running through the multiple and confusing scraps of memory
that remain. The chosen story is woven backwards to fit the timing, but is only one
of many such stories that might have been selected had a different event woken
the dreamer up. The important point is that there is no version of the story that
counts as the actual dream, consciously experienced at the time. This theory can
resolve the peculiar conflict described above, but it means accepting that there is no
right answer to the question ‘what did I really dream about?’ The theory is testable.
For example, since lots of brain events are going on during sleep, these should be
observable using the methods developed by Tomoyasu Horikawa and others (2013).
And it ought to be possible to wake dreamers using different types of stimulus and‘it is not like anything
to dream, although it is
like something to have
dreamed’
(Dennett, 1976, p. 138)
FIGURE 15.4 • According to the retro-selection theory (Blackmore, 2004), dreams are not conscious experiences. They are
concocted retrospectively on waking by selecting from the myriad trains of thoughts and images that were going
on in parallel in the dreaming brain. So on waking, this dreamer might recall that he had ripped some flowers
from their pot, rushed off on skis to escape retribution, arrived in a forest and had a picnic and a bottle of wine
under a pine tree. With many more parallel processes going on than are shown here, a very large number of
potential dreams are possible, and alarm clocks ringing or other sounds on waking might easily influence which
was selected.