Random Noise Is Our Most Valuable Resource

Yesterday I suggested that it will be surprisingly easy to build “creative” computers, because what we think of as “creativity” is just a temporary suppression of our mind’s inbuilt tendency to avoid unusual thoughts that violate its accustomed pattern. A computer could just not have that tendency – or, if that tendency is useful for some reason, it can make it an adjustable parameter that it can relax for “brainstorming” and then turn back up for rigorous thought.

Since we can’t do that, we search for something outside ourselves to break us out of the accustomed patterns. I mentioned how I get a surprising amount of inspiration simply from misunderstanding what other people are saying, and commenters agreed:

In partner dancing I sort of stumble onto inventing a lot of new moves by fucking up a standard move. But then it’s kinda hard to remember the new move since it was based on a mistake. The only times I remember a new move is if I make the same mistake at the same point in the step constantly.

Neil Gaiman noted, in his commencement address that’s all over the Internet, that he once misspelled “Caroline” as “Coraline” and he went “that’s interesting,” and saved it for later.

I remember reading that Neuromancer’s great opening line “The sky above the port was the color of television, tuned to a dead channel” was intended just to describe a blue sky, because Gibson’s television showed a blue screen on dead channels, but that most people read it and picture black-and-white static, which makes the line a lot cooler and more memorable

The part where you said you mishear other people, misinterpret it and then have a new idea – shares similarity with Harold Bloom’s ideas about Shakespeare’s soliloquies, wherein the character speaks aloud, mishears himself, misinterprets what he misheard and then allows that new interpretation to change him away from his original stance, allowing growth.

Something else commenters brought up was the idea of deliberately seeking random noise by, for example, rearranging sentences in Markov chains, picking random (tarot) cards, or using “brainstorming strategies” that randomly suggest directions to take something in.

In practice these have never really worked well for me. I think it has to be a very special kind of noise. In fact, as a commenter yesterday pointed out, “noise” may be the wrong word. Maybe “disruption”? Or even “difference”. Something that takes my thoughts processes in a direction different than the ones they would usually go.

This isn’t just about coming up with great works of art or new scientific ideas, it’s about the grunt work of evaluating theories. If I believe A, it probably means I’m in a rut where all the arguments in favor of A are available to me, but all the arguments against A are in some unreachable part of thoughtspace. Just having someone tell them to me might not work – the rut might be so deep that I round them to the nearest cliche or fail to appreciate them correctly. So the goal of looking for “the right kind of noise” is not only to become more creative but to become more correct.

I like this idea of “looking for” the noise. It stands in contradiction to the common conception of creativity as something that happens when you sit down and think really hard. This might work a little, but in general I think of creativity as a resource that needs to be mined somewhere. I don’t expect it to show up by sitting and thinking any more than oil will show up if you sit and think about it. If you’ve done your homework, you can refine your thoughts anywhere, the same way you can refine oil anywhere. But you’ve got to find it first.

Just as geologists know where to look for oil, so there should be some heuristics about where to look for original thoughts that will expand your ideaspace. The main rule seems to be: anywhere with people whose thoughts have diverged from your own a lot.

Cross cultural studies seem like the most fertile source. Eastern philosophy, while not as different from western philosophy as some people like to believe, is still deeply surprising to people who have only learned the Western tradition. My favorite example of someone using foreign cultures to overcome limitations in creativity is Harry Turtledove’s Striking The Balance series. The alien race therein is one of the most convincing and genuinely alien species in all of science fiction. It’s also transparently based on the Chinese Empire. Sitting in an armchair and trying to think of the most shockingly different extraterrestrials you can imagine still gets you something less alien (to American eyes) than simply adopting China for the purpose.

A second example of this: the languages made by beginning conlangers (people who invent constructed languages) are often really laughably bad. I remember one person who just went through an English dictionary, came up with a different word for every word (the = bla, cat = mred, are = nam, on = zig, mat = phlurd) and then combined them with English grammar and syntax (the cats are on the mats = bla mreds nam zig bla phlruds) and thought it was a language. You don’t get any idea how much languages can vary until you study a couple of others. Even then, if you’ve only studied Indo-European languages, the ones you invent – even the ones which in your constructed universe are spoken by creatures of pure energy and meant to be maximally bizarre – still have a clear Indo-European feel to them, because it’s difficult to imagine the dimensions upon which languages can vary until you’ve seen them. As far as I know, no conlanger has ever invented the dual case before they learned it was a real thing. This is another example where someone copying Chinese would get a whole lot weirder than someone trying to sound maximally alien.

But in some cases that can be too foreign, so foreign I can’t understand and integrate it at all. I have always had more luck with the past. I think this is the broader point beyond what I was talking about in Read History Of Philosophy Backwards. The goal of reading old philosophers is to expand your concept-space, realize that ideas you thought were almost tautologically correct actually have strong alternatives it is almost impossible for you to think of on your own. I now realize that part of my failure to understand MacIntyre was because my brain was totally incapable of understanding a certain concept of “community” that was the default in the ancient world. I thought I obviously understood what “community” meant, that there was no other possible meaning, that a lot of people who used it in weird ways were just confused. I was wrong, I never would have figured it out on my own, and eventually being bombarded with past resources helped me figure out things that seemed perfectly obvious to people back then.

Another potent source of intellectual disruption is talking to smart people you disagree with. If they’re smart enough that you know they’re not just making a stupid error, you can be pretty sure their cognitive ruts are different (maybe opposite) to yours. That gives you an exciting opportuntity to explore them and add new areas to your mental terrain. I find that taking the smartest people I can find, believing the most seemingly ridiculous thing, and latching onto them and not shutting up until what they’re saying makes sense to me is just about the quickest way to add new explored areas to my mental map possible. And since they don’t have my ruts, once I have theirs I can be one of the first people to possess two different mental tools at the same time, and then have the ability to synthesize them into exciting new ideas.

The last and most important source of disruption is, of course, reality, which is almost never what you think it is and usually pretty weird.

48 thoughts on “Random Noise Is Our Most Valuable Resource

  1. Ken Arromdee

    The blue sky reference is backwards. He intended it to refer to something resembling gray snow. Advances in technology then changed how televisions behave, so now a nonexistent channel is clear blue and the meaning of the line changes in a way that was not intended when it was written.

    This example is useless in the way you used it.

  2. Ghatanathoah

    This reminds me of an Orson Scott Card essay I read where he advised new writers that, when they make continuity mistakes, sometimes instead of going back and editing them out they should keep them in and develop an elaborate explanation for them. He says this is because since mistakes are not planned they often add creativity to the story.

    The personal example Card gives is when he made a mistake drawing a map of a fantasy city and drew a gate where there was no room for one in the wall. This caused him to come up with a bizarre and elaborate subplot about magical gates.

  3. Mike Johnson

    I was thinking of your idea about random noise in the shower, one of my most creative times of the day.

    … then it hit me: maybe random auditory noise (‘white noise’) provides or somehow boosts some of the creative randomness you’re talking about? I.e., maybe one reason showers are a hotbed of creativity is the noise, and silent showers wouldn’t be nearly as creativity-inducing.

    A test of this theory: brainstorming in a completely quiet room, vs brainstorming when a fan is on.

    1. Anonymous

      I think it has more to do with the fact that bathrooms tend to be free of distractions, and a shower is involved enough that you can’t just zone out and fall asleep and will not quickly become bored (that’s the most important part), but not so involved as to require any sort of actual cognitive effort.

      1. OldCrow

        I’m tempted to agree with Anonymous. I experience the same creativity boost while smoking a cigarette that I get in the shower, and Anon’s description hits on the biggest similarities between the two activities (plus the obvious chemical effect). But since I smoke outside, there’s usually enough background noise that it isn’t a good control. I guess I need to find a quiet, low-stimulation room where nobody will be bothered by the smell.

        What about knitting? Does anyone who knits find they’re more creative while doing so?

  4. Armstrong For President 2020

    This ground has probably already been tread in the last post, but there seems to be an equivocation here between creativity-as-novelty and creativity-as-insight.

    Novelty is easy to generate, relatively speaking; you’re just taking bits and rearranging them any way that seems to fit so any new piece will make exponentially more novel combinations. While most won’t stand, enough inevitably will hold together that in the end you have a new structure. But these cyclopean concepts are typically ugly and full of holes for the same reason they’re easy to make; there is no central element, just putting the bits wherever they happen to fit and not topple over immediately.

    Insight is more like a mosaic; you need to start with a central inspiration and build it up slowly out of the pieces while keeping the full picture in mind at all times. It’s much more time consuming and precise work, but in the end the product is a creation worthy of the time spent. This kind of creativity seems to be the least ammenable to simply adding new peices or jumbling things up to see what shakes together.

    I’d argue that, unless you’re deluberately trying for the bizzare and unsettling like in postmodern art or philosophy, it is always a better use of your time to look for unifying principles than novel curiousities.

    (Also, pardon my spelling as my phone despises wordpress and makes editing exceptionally difficult)

  5. gwern

    Reminds me of how Gwern noticed that limited exposure to foreigners has been a useful inspiration for thinkers historically. I can’t think of the title of the essay or paragraph but its likely still up on his site.

    That would be http://www.gwern.net/Notes#cultural-growth-through-diversity

    I still think there’s something to this idea of cultural walled-gardens building up ideas which can be harvested by limited mingling with other walled-gardens, but I have no idea how to rigorously pursue such an idea. It’s easy to point to instances like cursory accounts of China helping to inspire Malthus and the Enlightenment, but harder to look at such a claim quantitatively or investigate globalization consequences.

    1. Oligopsony

      Michael Vassar has a similar model of “scholarship” (as distinct from “science”) where circles of scholars work on problems in parallel but are at least somewhat epistemically segregated from another, and other scholars find convergent conclusions among the segregated groups and employ it as evidence.

      1. Multiheaded

        This sounds really neat. (Perhaps △we△ should do something like this when/if the △secret club treehouse△ is ready to grow/split!)

  6. a person

    Also here’s another example for the “mishearing” thing:

    “Cobain came up with the song’s title when his friend Kathleen Hanna, at the time the lead singer of the riot grrrl band Bikini Kill, spray painted ‘Kurt Smells Like Teen Spirit’ on his wall. Since they had been discussing anarchism, punk rock, and similar topics, Cobain interpreted the slogan as having a revolutionary meaning. What Hanna actually meant, however, was that Cobain smelled like the deodorant Teen Spirit, which his then-girlfriend Tobi Vail wore. Cobain later claimed he was unaware that it was a brand of deodorant until months after the single was released”

  7. a person

    In your previous post you mentioned dreams and drugs. I wonder if these two are especially good because they have the ability to bring subconscious processes to light. For art this makes sense because art is about exploring your feelings (huge oversimplification but yeah), and for more practical things this can make sense because sometimes I’ll come up with the solution to a problem and think “the answer was in the back of my mind the whole time”. Also drugs and dreams tend to carry with them an illusory sense of profound importance, which might help somehow. I like your theory overall but it might be a bit of a reduction to write off the creative aspects of drugs and dreams as just random noise.

  8. Pingback: Intellectual Disruption | CURATIO Magazine

  9. Mai La Dreapta

    As far as I know, no conlanger has ever invented the dual case before they learned it was a real thing.

    This might be true, but conlangers (experienced conlangers, not the sorts who make transparent relexes with English syntax) are constantly coming up with new and bizarre ideas and then trying to see if they already exist in the natural world. In fact, we have a name for this phenomenon: ANADEW (A Natlang Already Does Except Weirder). A very common occurrence on the CONLANG list is for someone to introduce their freaky bizarre and incomprehensible creation and then ask “ANADEW?” And somebody immediately responds with an example from Quechua or Warlpiri or East Sudanese Arabic. (There is a LOT of stored linguistic knowledge on the CONLANG list.)

    Conclusion: it is possible to invent very weird things without having heard of them before, but natural languages usually get there first.

    The broader point stands, though, in that people are only able to brainstorm these new oddities after they’ve been exposed to a very wide variety of natural oddities and have an appreciation of how vast the scope of natural language variation is.

    1. nydwracu

      Right, and then you have the trigger alignment, where no one’s sure if it really appears in natural languages or not but the general consensus seems to be that it doesn’t, and which was first developed because people kept misunderstanding Austronesian alignment.

      (also lol @ the conlang/neoreaction overlap)

  10. somnicule

    Brain firmware upgrades, like the Sequences, a lot of mathematics and statistics, programming skills, and other things which can be drastically new ways of thinking, always seem to boost my own creativity, but it tends to be temporary. Whether that’s because I exhaust the nearby obvious ideas, or I don’t bother looking for new ones after a bit, or whether new ideas based on those upgrades don’t feel creative any more I don’t know, but it seems potentially relevant.

    Maybe if one made a deck of things like “prediction markets” and “dimension reduction” and “Bayes theorem” and “recursive algorithms” and so on, and chose selections of two or three of those cards at a time, and for each one spent one minute

    Actually, that reminds me of Feynman’s principle of keeping a dozen important problems in mind whenever encountering a new idea.

    1. Lesser Bull

      I’ve had the same experience. It’s a cliche for lawyers that you learn all these new kinds of concepts and ways of thinking in law school and you start applying them to other areas of life in a way that is creative and exciting, but after a while you get used to the ideas and they stay inside their silo.

  11. Deiseach

    Talking about jolting oneself out of the ruts, and using “noise”, makes me wonder if this is why traditionally there has been perceived to be a connection between mental instability and creativity:

    “Great wits are sure to madness near allied,
    And thin partitions do their bounds divide”,
    (John Dryden, Absalom and Achitophel)

    Perhaps our creative computers will have a case of machine insanity? I don’t know; I think one could do a lot of harm to oneself by seeking a shortcut to creativity by putting one’s brain under that kind of strain.

    Then again, I’m about as creative as a brick (years back, I did an introductory art course and proved that yep, no imaginative/artistic ability whatsoever) so I may perhaps be a little too concretely stuck in the rut to appreciate the benefits along with the risk.

    1. Andy

      I’ve always had a problem with the “insanity/genius” conflation, because most people with mental illness spend a lot of time suffering and relatively little time creating and a lot of time just trying to survive their demons. Lord knows I’ve had enough problems in that area myself, and I’m no Van Gogh, either in terms of creativity or illness.

      1. Deiseach

        That’s been my problem with this romantic notion of creativity; the Tortured Artist who must flee his bourgeois responsibilities and head off to Tahiti because Art Is All.

        Or the notion of Chatterton, the unrecognised young genius, committing suicide. Or the glamour about rock stars overdosing.

        The idea that creativity both excuses and demands shitty behaviour, which can all be blamed on ones’ “inner demons”.

        In that case, I tend to go more with what Chesteron said in an essay on Whistler:

        The artistic temperament is a disease that afflicts amateurs. It is a disease which arises from men not having sufficient power of expression to utter and get rid of the element of art in their being. It is healthful to every sane man to utter the art within him; it is essential to every sane man to get rid of the art within him at all costs. Artists of a large and wholesome vitality get rid of their art easily, as they breathe easily, or perspire easily. But in artists of less force, the thing becomes a pressure, and produces a definite pain, which is called the artistic temperament. Thus, very great artists are able to be ordinary men – men like Shakespeare or Browning. There are many real tragedies of the artistic temperament, tragedies of vanity or violence or fear. But the great tragedy of the artistic temperament is that it cannot produce art.

  12. Kaj Sotala

    I think that all the examples that have been presented are illustrations of the fact that creativity is about taking existing patterns and the recombining them in an interesting way, and the more patterns you have access to, the easier it is to get novel combinations out of them. “Noise” is sometimes useful because introducing noise produces minor variations of existing patterns, and thus expands our store of patterns – the “mistakes in dancing” example is a particularly good indication of this.

    And because we judge new inventions as creative if they’re interesting and surprising, going outside your ordinary cultural sphere helps you get access to patterns that haven’t yet been combined with the patterns you and your audience are familiar with. If you’re only combining the patterns you’re familiar with, you’re less likely to get anything surprising.

    That said, having a high brainstorming toggle on the AI doesn’t necessarily help as much as you think, because randomly combining patterns is the easy part. The hard part is coming up with combinations in a way that actually forms interesting and coherent patterns. Notch up the randomness and you just get Time Cube, which is moderately entertaining but probably not considered very creative by most people. To get what we’d consider creativity, the system needs to have an elaborate set of additional mechanisms which somehow evaluate the potential usefulness of different proto-ideas and pick the most promising ones to be developed further. In the dance pattern example, I guess that someone consistently making the same mistake at a certain part of the dance routine is an indication that there’s a sensible alternative dance pattern that feels natural enough for them that they’re consistently drawn towards it: whereas just trying to randomly make mistakes in the dance would just produce a bad dance.

    This is probably why it’s so useful to go to different cultures: because you’re tapping into a rich source of patterns that have already been vetted for basic sanity. They’re not just random noise, they form an internally consistent whole that has its own basic logic that can then be extracted and transported into a different context.

    (Compare Eliezer at How to Seem (and Be) Deep: Transhumanism also has cached thoughts about death. Death: complete the pattern: “Death is a pointless tragedy which people rationalize.” This was a nonstandard cache, one with which my listeners were unfamiliar. I had several opportunities to use nonstandard cache, and because they were all part of the developed philosophy of transhumanism, they all visibly belonged to the same theme. This made me seem coherent, as well as original.

    I suspect this is one reason Eastern philosophy seems deep to Westerners—it has nonstandard but coherent cache for Deep Wisdom. Symmetrically, in works of Japanese fiction, one sometimes finds Christians depicted as repositories of deep wisdom and/or mystical secrets.)

    1. Hainish

      “Death is a pointless tragedy which people rationalize.” This was a nonstandard cache, one with which my listeners were unfamiliar.

      Wait, this is non-standard and unfamiliar? (Where have I been living??)

      1. Rob

        “Violent or ‘untimely’ death of non-evil people is a pointless tragedy that people rationalise” is pretty standard. What’s non-standard is “Any death at any time of any person is a pointless tragedy that people rationalise. This includes 110-year-olds who lived full lives dying painlessly surrounded by their loved ones”.

  13. Jaskologist

    Ever read any of C.J. Cherryh’s work? She is especially good at making aliens alien. The main point of the Foreigner series is to present Atevi psychology; the plots are really just to stitch that together. Forty Thousand in Gehenna was also pretty interesting from that standpoint.

  14. Izaak Weiss

    Your mention of conlangers and dual case reminded me of a personal anecdote that sort of contradicts what you had said; I reinvented clusivity for one of my conlangs before I knew it was a real thing, having only studied English, Spanish, and Hebrew to any real extent.

    1. Nick

      I’ve reinvented weird grammar while conlanging too, but I have to admit Scott picked a good example, because I don’t think I ever would have pursued a dual number.

    2. Luke Somers

      I made one with both clusivity and dual number (specified in consecutive paragraphs no less), and only found out that these were features of real languages later on.

      There are some features of that language that I still haven’t seen in any real language

      – a preposition system incorporating orientation and/or proximity (both are optional, and if neither is specified it is a very generic connection such as a logical connection)

      – a grammatical form for controlling a characteristic, like how ‘redden’ would be for simply making something more red.

    3. Lesser Bull

      Clusivity is one of those things where you run across a gap in the language, though. Everyone has had to clarify who they mean by ‘we’. But the dual case is much less obvious. It’s cool because it suggests a genuinely different way of thinking about things.

  15. Aaron T

    This reminds me a lot of one of Venkatesh Rao’s posts here: http://www.ribbonfarm.com/2014/06/11/the-deliberate-practice-of-disruption/

    He winds up talking a lot about the difference between practicing to achieve a greater sense of mastery within a particular domain, and practicing to have experiences in which you can notice things not quite fitting, and poke on them to see just how long they can hold together moving outside of what you’re supposed to do there.

    Particularly related-seeming:

    If the reward for effective sustaining metacognition is a sense of your own inner sacredness, experienced as flow, the reward for effective disruptive metacognition is a sense of snowballing absurdity and paradox that miraculously does not unravel. Effective awkwardness that inspires irreverent laughter rather than reverent awe. Instead of approval from honored figures, you get the slightly vicious pleasures of desecration.

  16. Amanda L.

    I now realize that part of my failure to understand MacIntyre was because my brain was totally incapable of understanding a certain concept of “community” that was the default in the ancient world… I was wrong, I never would have figured it out on my own, and eventually being bombarded with past resources helped me figure out things that seemed perfectly obvious to people back then.

    Ooh. Could you possibly sidestep the process for us and give a brief summary of what community meant back then?

    Alternatively, what are the best books to read to get this understanding?

      1. Emile

        Thirded, I was going to ask for that too (it’s not that clear whether Scott’s link is supposed to point to “that time where I misunderstood community” or “that time where I explained how I had misunderstood community” – the linked article hardly mentions the word).

        1. Andy

          Neither do the other two articles Scott wrote on After Virtue. And if “community” means something other than you thought it meant, then that definition might be a good thing to have.

    1. Scott Alexander Post author

      Hard to explain. I got some of it from The Righteous Mind, but that was just something that allowed a few things to click into place for me. The reactionaries helped as well, as did (for some reason) reading about monarchomachs and trying to figure out why this was a controversial theory.

  17. Benquo

    The last and most important source of disruption is, of course, reality, which is almost never what you think it is and usually pretty weird.

    No, YOU’RE weird!

    (I feel like I’m allowed to do this because I’m a part of reality.)

  18. Hank

    In the context of tarot, the “disruption” you’re looking for is forcing the most salient aspects (eg, the symbology, archetypical aspects, etc) of a random spread of cards into a coherent narrative, that fits with the constraints of whatever you’re trying to read. This narrative-forcing generates certain side effects (bits of the generated narrative connoted by secondary aspects of the symbology, which are not totally supported, but not totally disallowed by first-order evidence), which provides insights. It also forces you to constrain the external facts into the first-order narrative provided by the cards, which can provide some insights into alternate ways of thinking about whatever you’re trying to read.

    1. Trent Fowler

      Well said. I had my first tarot experience a few years back, and was immediately aware that my previous skepticism of it was over-reaching; it clearly has the ability to generate insights. I’d call it a source of semi-noise, not exactly random, but random enough that it’ll probably take you down some interesting paths.

  19. Darcey Riley

    I tend to find that thought-ruts correspond to activity-ruts. If I do the same thing every day, then I will think the same thoughts while doing that thing. If I make a point of doing new things or going new places, I find that I’m more likely to have new thoughts. (Here I’m talking less about creativity, and more about overcoming those neurotic, stressful, repetitive thoughts that continually circle through my head. But maybe it works for creativity too? I dunno.)

    Anyway, I think it’s underappreciated how much of an affect our environment has on our thoughts. Habits (including thought-habits) are contextual: they’re prompted by some signal in the environment (e.g. sitting at one’s desk, the time being roughly 6PM, etc.). I feel like a really good way to break out of old habits/thoughts is just to avoid familiar contexts for a while.

    1. Sam Rosen

      I completely agree. “It is easier to behave your way into a new way of thinking than it is to think your way into a new way of behaving.”

  20. Andy

    So are you going to return to MacIntyre and revise/ rewrite your blog posts? Because I thought those were some really good writing.

    1. Scott Alexander Post author

      No. So far I only have the vaguest ghost of an idea of what appreciating MacIntyre would look like. If I ever get more, I might write about it. But right now I can see that there’s a there there, but not really make out what it would look like. Still useful for some other things, though.

Comments are closed.