Given how much you have learned of the techniques of rationality, of Bayesian updates and standard of evidence, of curiosity being the first virtue and being willing to update your beliefs... have any of your dreams been affected by them?


The reason I ask; I'm reading the entirely of the Sequences, and am about an eighth of the way through. And I've just woken from a dream whose plot was somewhat unusual. I had noticed some mildly strange animals and/or people, and upon trying to find out what was going on, discovered a small riverside camp of people who fell well outside what I understood to be the realm of human variation. The person I had started investigating with then claimed to be a god, or if I preferred, a vastly powerful and intelligent alien entity, and offered to do something to prove it to me. I remembered that I had once established for myself a standard of evidence for exactly this sort of question - the growth of a new, perfectly functional limb, in a way outside of present medical understanding... and in a few moments, my dream-self was the possesser of a nice, long tail. I had not been expecting that to happen, and noticed I was extremely confused, and deliberately raised my estimate of the probability that I really was talking to a god-like figure by some number of decibans. At the end of the dream, said deity-figure said that he would offer to split us off from his 'main project', on a few conditions - one of which was 'no more clues', since he had given us 'more than enough to figure out what's going on'... ... whereupon I questioned a few things, and immediately woke up.

I don't recall having a dream of anything like that sort before - and I dream in understandable narrative plots so often that I sometimes dream sequels. So I'm curious; is this a normal sort of thing that happens to LessWrongians?

New to LessWrong?

New Comment
42 comments, sorted by Click to highlight new comments since: Today at 2:57 AM

I frequently have dreams about whatever I'm currently obsessing over, be it video games, romantic interests, or whatever.

This is particularly fun when you're spending a month designing a lovecraftian Solstice ceremony.

I am actually feeling bizarrely awful about this (I'm in the habit of exaggerating whenever I'm telling a story that shouldn't have major consequences, but it feels wrong on a forum about truthseeking and I feel guilty for getting karma for it).

My statement about lovecraftian dreams is technically true but not nearly as interesting as I made it sound. I didn't actually have dreams about the world ending in tentacles or anything. I did wake up in the middle of the night and feel compelled to write a song entitled "Mindspace is deep and wide" that briefly features alien gods, but there was not actually a cool dream preceding it, and while I think I had vaguely interesting dreams relating to the solstice that month, I can't recall them now.

I HAVE had extremely powerful dreams that were clearly inspired by my narrative influences at the time, so the important point still stands.

Would you feel better if we took some of that karma away?

This. You dream about your experiences, not your beliefs. I had a dream about Mafia (the game) last night, because I found a place to play it online and played it for several hours.

Actually, it's only very rarely that I can relate my dreams to any of my experiences, recent or otherwise.

I find it rather amazing how many people's dreams, when they describe them, sound basically mundane, like things that could more or less happen in real life. For me, dreams have always been, not just too weird to possibly mistake for reality, but too weird to make sense of after the fact.

I have one dream in particular which I remember because it was probably the happiest dream I've ever had, and left me in a great mood when I woke up, but if I tried to describe it I wouldn't even be able to explain why it was happy.

The emotions invoked within a dream are a further fact about the dream, beyond the narrative and the sensory experience. Once a person has finished telling you all the mundane, nonsensical details of their dream, they have failed to impart the powerful emotions, the experiential gestalt, the sense of awesome import that moved them to recount the dream in the first place.

I tell them, "There was a toy bunny on a white field ... it was horrifying." The listener is unimpressed by this, because seeing a toy bunny is not a reason to feel horror. In this way, stories about dreams are not like stories about real life. I could say, "There was a toy bunny on a white field, and this was the most terrifying thing in the universe." But that is confusing the order of things: I saw a toy bunny, I experienced a profound horror, and I projected that horror onto the bunny.

The dreams that I tell people are the ones with the coolest-sounding stories, not the dreams that mean the most to me.

Hm. I can almost always trace my dreams to my experiences. That's interesting.

I find it rather amazing how many people's dreams, when they describe them, sound basically mundane, like things that could more or less happen in real life.

At least twice as a child, I actually completed homework assignments in dreams. Or at least I thought I'd completed them upon waking.

I wonder if those experiences have something to do with why I react to non-realism with such hostility.

It's interesting that you say that. I've always treated non-realism with sympathy, and it took a while for the concept of "beliefs that correspond to reality" to anchor itself. It is possible that this correlates with my interesting and not-like-real-life dreams...

Not sure which way, if at all, the causation acts though.

I'm like this too, except that I can make sense of the weird dreams: just not with normal reasoning. I wake up from dreams with weird and complicated plotlines with the absolute conviction that the plotline was /consistent in dream-logic/. I usually make the effort to remember an interesting dream when I wake up though, so I have a good database of information to work from.

Dream logic involves things like: if you willed something to happen hard enough, it happens (so you can jump from buildings if you just willed a soft landing into happening), every building has secret passageways (and in a pinch you could always dive into a cupboard first and will the secret passageways into existence after), if in a situation which resembles a video game puzzle, it's always possible to visualise a level mini-map, never run down stairs because they frequently terminate into an abyss without warning (and then you're forced to wake up)... Oddly, I rarely realise I'm dreaming (if I do I normally force myself to wake up - into another dream). I simply act in my dream as if these things are simply "true".

I don't have the unusual emotional response though, my normal emotional responses just get multiplied by a variable factor. (I don't feel happy about anything I wouldn't feel happy, to some degree, about when awake.)

But then again, I've always had a fairly weak grasp on the real world. I have a particularly poor memory for details, and as a result it's hard to trust my own sensory inputs. It's quite interesting when I'm talking to my OH, because while I find it odd how he finds it so hard to remember his dreams, he finds it odd how I find it so hard to remember what I said and did this time last year. It feels like these two things are related; I'm not sure how though.

I could never understand people who had boring dreams. If I could pick between living in the dream world (my dream world is usually internally consistent between dreams) and the real world, I'd pick the dream world. If I wake up in the morning and don't have anything pressing to do, I close my eyes and go back to my dream (since about half an hour of real time dreaming corresponds to about 2-3 hours worth (or "about one story's worth" of dream-time activities). As a result, I feel like people who have mundane dreams are really missing out...

I'll think that my dreams make sense while I'm having them, sometimes I've even written down plot elements from my dreams when I woke up in the conviction that they were brilliant and I would have to reuse them. Unfortunately, things that seem brilliant when I'm still half asleep tend to look anywhere from stupid to lunatic once I'm properly awake.

My dreams do have recurring consistent elements, but I've actually lost the ability to make things happen in my dreams by willing them hard. For a while when I was a kid I could, but after some point I found that if I tried it, I would simply wake up instead, and this is still the case.

It often seems as if my dreams are being created by someone who's in an adversarial relationship with me, and if I learn anything about using dream logic to my advantage, they will eventually catch up and change it on me.

To clarify: my dreams certainly don't make sense in regular logic. However, I've found they usually make sense if I replace certain axioms and resulting theorems about reality with alternative ones which apply to dreams.

Causation dissolves somewhat in dreams: it's not that "things happen /because/ I will them to", because "because" doesn't make sense in dream logic. It wouldn't work if I believed, in the dream world, that wishing things to happen causes them to happen (I have tried this). It's more "when I want things to happen, much more often than not they do, though the precise mechanic by which this happens is indescribable using the rules of the real world". There's a very specific degree of "willing, but not insisting" which allows the events to happen almost always.

It seems to me that the mechanism is perfectly well describable: in my dreams, my desires and my perceived environment are both almost entirely influenced by a single common cause with a short causal path, and therefore tend to correlate. Outside of my dreams, my perceived environment is influenced by many other things which take a much longer causal path to affect my desires, and therefore tend to correlate less well.

Well, this is certainly how it works when viewed from the real world. A bit difficult to work out inside my dreams though, when causality is no longer a thing for that precise reason. Compounded with the fact that I still feel like I have free will in dreams, this makes the dream-environment (where I'm not aware that I'm dreaming, but have internalised the dream axioms) quite different from the real world.

Ah! I see. Yes, agreed, the causal mechanisms I believe exist while inside my dream are often hard to describe. (Indeed, often there aren't any such believed-in mechanisms to describe.)

Whenever I'm doing something visually stimulating for a while before bed, it ends up going in my head while I'm in the process of falling asleep. Tetris, MarioKart, any puzzle game involving moving pieces. Sometimes if I look at a lot of .gifs (really short animations) I get semi-dreams in the form of a lot of little clips. This is all especially weird because most of my waking thoughts are verbal, but my dreams are nearly all visual.

It's happened at least two or three times that I can recall that I've come home late after a night out, gone to bed, and had dreams where I was still out. Also, the first time I lucid dreamed I was woken up by an argument on the street outside my room - I fell back asleep and had a lucid dream, and didn't realize until someone told me the next day that the fight had actually been real and it had merged seamlessly into my dream.

I think these are exceptional though - otherwise I don't think I tend to have dreams based on my experiences.

I very rarely recall any dreams, but I do remember one time, during a summer I spent playing a lot of MUD (Internet text-based game, primitive ancestor to World of Warcraft), that I had a dream in text.

I've dreamed in scrolling text before due to extensive MUD playing. I even wrote an essay about it, though it's a bit embarrassing to read now.

I've dreamed in scrolling text before due to extensive MUD playing.

So have I. Which MUD, out of interest?

I spent spent literally years playing WOTmud, playing the Gaidin Wedrifid, from whom I took this name. (By 'years' I refer to actual logged in hours that can be measured in years, not years passing during which I played MUD.) I ask because 'Rain' was also a prominent character on that MUD - a wolfbrother or something impressive like that - I think he stabbed folks a lot.

I played WOT only a little; I even had a wolfbrother; but I was not Rain.

I primarily played on Sojourn / TorilMUD, where Brad McQuaid also played before creating EverQuest. Others I've played: Realms of Despair, various God Wars, Vampire Wars, Dark and Shattered Lands, The Last Sunrise, CoreMUD, RetroMUD, Shadowrun: Denver, Shadowrun: Detroit, and Metro: Toronto by Night.

May I ask for more details? For example, do you recall whether you were observing text on a screen, or if you were looking at a visual field with just text, or some combination of imagining typing while visualizing what was being described, or the like?

It has been more than a decade since then. All I have left are the less-reliable memories-of-memories of the dream. Having said that, I recall the dream being of text coloured like the MUD I was playing, but I am pretty sure that there was only the text. I don't even recall anything that happened in the dream or if I previously did and have forgotten.

Isn't it normally hard to read in dreams? How did that work out?


I'm able both to read and to type in "day residue" dreams: those that are primarily based on prior events while awake. I've often had dreams which involve me reading and posting to internet message boards.

For dreams occurring on a deeper, more symbolic level, I haven't had the same experience, though I couldn't say whether that means I can't read in those dreams, or only that I haven't remembered doing so.

Perhaps individual circumstances matter. I was hyperlexic in childhood and to this day my preferred learning style consists of reading and/or rote copying, which I've learned is unusual.

I am not sure about reading, but writing is impossible for me, in dreams. When I write something, and I read it again, I see that I actually wrote something else. So I try to correct it, and again it is wrong; and that is very frustrating. Doing calculations, even as simple as counting my fingers, is also impossible.

From what I read online on topic of lucid dreaming, different people have different "abilities" and "disabilities" in dreams, but some of them are rather common (for example the inability to count one's fingers). If you make a habit of testing these skills a few times a day, it may help you notice when you are dreaming.

I can't help you there. Not enough detail has survived the years.

That's a very common claim; but it runs counter to my own experience in dreaming.

I had a dream where I was flying by incrementing my own x and y coordiates. Somewhat related to simulated worlds, but also to straight programming.

I had not been expecting that to happen, and noticed I was extremely confused, and deliberately raised my estimate of the probability that I really was talking to a god-like figure by some number of decibans.

As this example shows, it should chiefly raise your credence for being in a dream and e.g. having the ability to levitate regardless of what your 'companion' does. Though if you happen to remember that next time, do recall the possibility of an expected value calculation as well.

While this predates my reading Less Wrong, I've certainly had more lucid dreams, and somewhat fewer where I took my sensations as a refutation of the dream hypothesis, since I noticed that this fails to hold.

Three downvotes already? Interesting.

I wonder if it would help if I mentioned that I'm completely aware that while I phrased this post in the form of a question, a significant part of my motivation for posting was to try to tell an interesting story about myself which I hoped would mildly improve local readers' impression of me... Nah, probably not. :)

On a related note, I'm still trying to figure out how my dream-self was able to feel the sensation when somebody grabbed hold of my/his new tail, a body-part which my brain's sensory homunculus wouldn't seem to have a space for... Any thoughts?

Two possibilities come to mind.

First, that the limbs of your "dream self" aren't mapped to your brain's sensory homonculus in the same way that your limbs are at all. (I find that likely.)

Second, that the limbs of your "dream self" are mapped to your brain's sensory homonculus in the same way that your limbs are, but that your brain constructed the narrative of the dream in such a way that the signals coming from the portion of your sensory homonculus ordinarily mapped to, say, your arm are interpreted as coming from a tail instead.

What interests you about three downvotes?

Given what I felt (or 'felt', as the case may be), your first suggestion seems more likely than your second; though given what little I do know of the brain, the second seems more plausible than the first. I don't have enough data to think one is significantly more likely than the other, and I'm not sure where I might find or create such data.

As for the downvotes, I'd made a private bet to myself before posting, that I was sufficiently oblivious to and unaware of local social norms that I'd make /some/ social error causing my post to be downvoted to oblivion, and that I wouldn't know why unless someone explicitly told me, which I expected to be unlikely. I'd thought I'd recalled LessWrong's "oblivion" was five downvotes rather than three, but was wrong about that.

As for the downvotes, I'd made a private bet to myself before posting, that I was sufficiently oblivious to and unaware of local social norms that I'd make /some/ social error causing my post to be downvoted to oblivion

Here's one.

Re: sensory narratives... what knowledge of the brain makes the latter seem more plausible than the former?

Re: downvotes... congratulations on winning your bet!

Re sensory: From what I've read, the dreaming brain takes advantage of many existing brain-structures - eg, when you see things, parts of the visual cortex light up. So it seems plausible that that would also work for physical sensations.

I suppose one possible resolution is that my particular sensory homunculus is already primed for a tail somehow, though, again, I'm not sure how to find evidence to help support or reject that theory.

My experience is that when I'm working on training my capacity for sustained attention (which I do for a few months every few years), I frequently experience dreams in which I either start being critical of the dream-logic (thereby typically ending the dream), or have two simultaneous independent tracks of narrative, one of which follows dream logic and one of which doesn't. (For example, I will frequently have dreams in this state where I'm in a social setting and interacting with other people, where my interactions depend on not knowing certain things which I somehow know that I don't know. I've often tried to reproduce this dual-narrative state in my waking life, with no success.)

The relationship between this and LW-rationality is not clear to me. If I had to guess, I'd say they're related only by being in the same cluster of interests.

I've had a lot more dreams where I noticed that I was confused about what was happening, leading me to conclude that I'm hallucinating (of which dreaming is a subset). When I use that rationality technique in real life, am I also accidentally training a lucid-dream technique? Time will tell, since I've only had limited success in staying aware long enough to manipulate the dream environment.

I've had some effects, in a roundabout way. Joining LessWrong got me to start introspecting more, and (I think as a result) I'm much more aware of sensations in my dreams. I've always dreamed in color and sound, but in the past year or so I've also been experiencing temperature, wetness, taste, the sensation of "that's gross", and limited metacognition in dreams.