Verified by 
Most dreams seem unconnected to one another, but a few seem to carry on themes from one dream to another. Does this continuity of content come from the dreams themselves, or is continuity of content due to a continuity of experiences in our waking life?
There can be little doubt that most continuity of content is related to a continuity of our waking experiences—but I want to entertain the possibility that some continuity of content is due to the dreams themselves. For that to be possible, it seems that we would need to entertain the possibility that dreams either draw on a specialized memory system only available to the dreaming brain, or that dreams can access selected memory systems that are typically unavailable to the waking brain.
If there is a specialized memory system available only to the dreaming brain then it would follow that dreams very likely have specialized processing outputs as well, and if that were the case, it would be easier to suppose that dreams have a special function, possibly related to that specialized memory system that enables the specialized cognitive output we experience in dreams. The argument would be that dreams produce a special product for the mind/brain that the mind/brain cannot attain from any other cognitive system. That specialized product may not be necessary for our physical survival but it may be necessary for the optimal functioning of our mind/brain.
If dreams have sole access to a dream-specialized memory system, then it should not be surprising that dreams can cumulate content over time that is not related merely to daytime waking consciousness. Dream memories would build up over time—just like memories in systems that are accessed via waking consciousness. The models we have cumulating content in memory systems available to waking consciousness are systems like autobiographical memory, semantic memory, and episodic memory. (Procedural memories also cumulate over time but these are less accessible to awareness.) While the semantic memory system appears to be organized by meaning-relatedness criteria among “ideas," autobiographical memory appears to be organized according to the salience or significance of specific personal historical episodes. Memories within the dream memory system, if such a system exists at all, would likely be organized along similar lines.
Is there any evidence for dream-specific memory systems? When we collect longitudinal series of dreams from a single individual, there is simply no doubt that themes, ideas, episodes, characters, images, and emotions recur consistently—and frequently—across these dream series. This is the case for dream series considered across years and dream series considered across a single night of sleep. The crucial question is whether any of this recurring content is due to dream-specific content/occurrences versus daytime-waking content/occurrences. That is, can we ascribe any content from Dream 2 to content occurring in Dream 1 and not to content occurring in waking life between Dream 1 and Dream 2?
In my opinion, we can.
For example, many people have dreams in which they perform (in the dream) some function that they are unfit for, or unable to do, in waking life—and they get progressively better at performing that function over the dream series. This progressive improvement in performance at something the dreamer is unable to do in waking life cannot always be reduced to mere wish fulfillment. For example, I had a series of dreams over several years in which I got progressively better at searching through a landscape involving both mountains and urban settlements while being chased by vaguely threatening characters who never quite caught up to me. The landscapes I passed through in these dreams were always the same, with the exception of new areas opening up in each dream. I did not want to return to those landscapes, as I felt threatened and scared. I do not now and did not then believe that the dream series was merely an elaborate metaphor for something I was going through in my life—I am still going through experiences that should give rise to such metaphor, but no such metaphors are appearing in my current dreams. Instead, it seems more plausible and true to me and my own experience of this dream series that the dreams, so to speak, were interested in exploring that landscape and “I” had to oblige whether I wanted to or not!
Whatever the case may be, the science of dreams needs to explore this question of whether dreams draw upon a specialized dream-specific memory system when constructing scenarios.
