Polyphasic Sleep, Lucid Dreaming, Critical Thought, Thought Loops
Last month, I was thinking a lot about polyphasic sleep schedules. I didn't manage to discipline myself in such a schedule yet, and my feeble attempts only combined with procrastination into reducing my sleep time, to my ultimate psychological downfall. A more reasonable target will be to shoot for simple biphasic sleep, with a regular nap in the afternoon after a regular lunch: when at ENS, I met someone who did quite well this way, which is also the way it was done for all children in my kindergarten. When I manage that, it will be time to add more naps and do less sleep.
However, short sleep schedules have the power to bring more lucid dreaming. And with my reading about lucid dreaming, and my recurring disappointment since I was a kid at not being able to take notes in my dreams that I could leave to my waking self, I had this most interesting meta lucid dream.
Lucid dreams, at least to me, happen when something goes wrong; Since I was twelve or so, I've unlearnt to have real nightmares, and when something goes wrong, I instead enter a more lucid dream -- the wronger it gets, the more lucid the dreaming. Well, in my dreams, I have for a long time been frustrated by not being able to manipulate written documents as in real life: you know, as a thing that you can read multiple times and remains the same between readings, so you can read in multiple passes to get the idea then the details, ponder things slowly and thoroughly, and transmit messages to your future self. Well, during a dream about otherwise boring (to you) worries of mine, I wrote something on a piece of paper. But as usual, the piece of paper would not read twice the same. And the frustration triggered lucid dreaming.
I became aware I was in a dream, and reflected about the inconsistency of imagined written documents. And then I realized quite clearly that dreams activate concepts at various abstraction level, independently from the underlying phenomena at lower abstraction levels that usually activate the concept. Details at these lower levels are made up on the fly out of limited interaction with critical modules of the brain. Thus, I may imagine that I write something; and because this something involves a sheet of paper with letters, I may make up the paper: that's easy, just the few concepts for a wavily folded soft-touched white rectangle with black signs on it. The signs are characters; since I'm not interested in the details of how they look like, I don't invent a handwriting; rather, when my critical modules insist on having a clear image of the sheet, I make up a sequence of monospace computer characters. And the characters don't have to match the text being written, because the critical modules don't do all the consistency check in a lot of detail across multiple levels of indirection; and anyway, it can't spellcheck words as fast as I make up words, with only the available peripheral attention; only the most prominent letters may be right, and the approximate length of words, the rest is random rubbish characters.
As I thought of it, I could vividly visualize the sheet, with at the same time a simple meaningful text being written, and the semi-random characters that made up the words for the text. I could see clearly the inconsistency without disbelieving either the text being written or the details of characters that didn't match. But even the text was not something completely consistent; it was more of the idea that there is a text that has this kind of content, with globally inconsistent details filled up as critical modules locally required them. Therefore, as I wanted to focus on a text simple enough that I could fully understand it while visualizing the inconsistency, I imagined my name on the sheet, and the random characters also; except that somehow at another level, I also saw that it was my name due to its being spelled correctly when the critical module's cursory glance went over that. (Speaking of visualizing, VISUALIZE WHIRLED PEAS said a bumper sticker on a car parked on Mass Av next to MIT.) That was quite an interesting dream; after that startling revelation, I ended up assuming that I had woken up from dreaming, whereas I was still dreaming and had only escaped the first dream into a meta-dream.
I'm sure that such dreams can tell us something about the structure of our minds, in terms such as in Minsky's The Society of Mind, or as what has been called homuncular functionalism. Our rational abilities are a kind of interactive critical cursor that locally checks active concepts for consistency with respect to some admitted rules; it rewards the consistent ones and punishes the inconsistent ones; the concepts in turn try to justify themselves when facing criticism by making up justifications -- which is easy when the concept was activated by a pattern of lower-level sensory concepts -- but less so when doing more abstract thinking. In this evolutionary struggle for life, only survive the more stable concepts, with respect to internal criticism.
Of course, depending on the person, on the circumstances, etc., the consistency requirement attached to each concepts may vary, be looser or tighter, or wholly different. The cursory strategy by which the critical module explores, examines and checks requirements, rewards or punishes concepts, may also vary. Also, concepts the consistency test of which always succeed or always fail lack any relevant discriminating power, at least in the considered context; they are redundant or useless, and should be avoided or short-circuited, least the critical mind should enter an endless loop of self-rewarding self-congratulating evidence. Thus, the mind must have some kind of loop detection routine, that is able to detect that some rule keeps succeeding or keeps failing, and trigger a feeling of boredom or otherwise inadequacy, lllso as to put that rule in the background.
Now, the loop detection routines can not possibly be perfect either. Thus, while they can easily detect some simple loops, and isolate obviously useless concepts, there is a vast domain of more elaborate loops that they cannot detect, including loops that involve repeated detection of simple loops. Obsessive and compulsive behaviour might be such mishandled loops; the mystical feeling of a god as a perfect concept that succeeds all consistency tests is another such mishandled loop.
There is a lot we can learn from how the brain works by carefully observing ourselves.