The American philosopher Eric Switzgebel has made a name for himself in recent years by challenging a lot of what people think they know, and, in particular, by challenging what people think they alone know—things about their own conscious mental experience. For example, he has challenged the notion that mental images are picture-like, which a lot of people believe. People confidently state that they “had an image in their heads” of this or that and discuss the contents of it (oh, I was looking at an image in my head of Grandma’s house as it appeared back when I was a kid in ’85), but when you press them on this, their image report turns out to be extremely unreliable in a way that pictures–real pictures–aren’t. If I ask someone to visualize a candle, he or she will report having done this. He or she might say something like, “I had a picture of it in my head,” but when you press the person on this: what was the candle sitting on? What was the color of the candle? How far down had it melted? Was any of the candle still liquid or molten at its base? What was in its immediate background? Was the flame flickering? etc., people don’t know the details that they would know if they had seen an actual picture. Switzgebel tells people to “picture a house,” and they report having done this, but then they can’t tell whether the house has a chimney, which they could easily do if they were consulting something actually “picture-like.”
In a similar vein, the psychologist Elizabeth Loftus famously showed that when people remember events, they think they are simply reporting a recording in the brain of what happened when, in fact, their brains actively construct memories from a combination of the few things from that past time that actually got recorded into long-term memory and a lot of general knowledge about the world, including things that they were later falsely told about this past. Memory is not playback, though it seems like that to the person having the memory. It is reconstruction, and extremely unreliable reconstruction at that.
I think that it might be the case that something similar occurs with dreaming. This is a fairly radical proposal, and I haven’t encountered it before in work by others. We tend to think that the events in dreams play out over time, typically in real time, that they have duration. Event A occurs, then Event B, then Event C, and so on. But what if, instead, the same story construction mechanism that occurs in memory occurs when we call upon a memory of a dream? Suppose that an executive conscious portion of the brain
takes random stuff that is currently being processed by the brain and is therefore currently available;
automatically, swiftly, constructs a narrative, with duration and a protagonist and other narrative elements, from that material; and
presents the result to consciousness?
And, furthermore, suppose that throughout REM sleep, this conscious executive part of the brain is checking in, in this way, from time to time, finding slightly changed inputs each time and requesting correspondingly changed narratives that are automatically, swiftly supplied? One might call this the “Multiple Drafts Theory of Dreaming.” The idea bears similarities to Daniel Dennett’s Multiple Drafts Theory of Consciousness presented in his book Consciousness Explained.
If what I am suggesting is true, we don’t actually experience events with duration during our sleep. Instead, we construct, from the random stuff our brain is processing, events with the qualities that events have, including duration and other narrative elements like conflicts and antagonists, and then those events are presented, as recent memory, to conscious memory, and we think that we experienced them in the dream because we are remembering them, just as we think, all the time, that we experienced events in our own pasts that our brains have partially or wholly simply constructed as part of the active making, each time we remember, of the memory.
And if all this is true, what is happening in the morning when we “remember a dream,” is that the conscious part of the brain is calling for a memory of the dream, which in turn tells the unconscious part, “Construct a narrative memory of a dream from the most recent brain state and present it to me as an actual memory” (not as a hypothetical or construction). This shouldn’t be surprising, for this is just what happens in memory of life events. But it is a bit disconcerting because it challenges what we think we know about ourselves and our experiences.
Ofc, this ability to construct narratives, automatically, swiftly, has a readily available evolutionary explanation in that it would enable prediction of the behavior of others–the essential survival skill of playing out into the future of a creature’s theory of other minds. It pays to have an idea what that hippo one has suddenly come upon might do.
NB: One support of the theory is the iffiness of duration in dream reports–in dreams, there are often leaps in time and compressions or extensions of it; so, it’s unreliable and more like duration in stories than in active, waking life.