ribbonfarm

This Captain’s Log blogchain has unintentionally turned into an experiment in memory and identity. The initial idea of doing a blogchain without meaningful headlines or fixed themes — partly inspired by twitter and messenger/Slack/Discord modes of writing — was partly laziness. I was tired of thinking up sticky and evocative headlines, plus I was getting wary of, and burned-out by, the unconsciously clickbaity nature of headlined longform.

I couldn’t remember anything of what I’ve written here, so I just went back and read the whole series, all 20 parts, and it’s already slipped away from my mind again. Names are extraordinarily strong memory anchors and without them we barely have textual memories at all. I can recall the gist of many posts written over a decade ago given just the name or a core meme, but for this blogchain, even having re-read it five minutes ago I couldn’t tell you what it was about. The flip side is, it wasn’t actively painful to reread the way a lot of my old stuff is (which is why I rarely re-read). In some ways it was kinda surprising and interesting to review. The lack of names means a lack of fixed mental models of what posts were about. It’s weird to be able to “cold read” my own posts. It’s like simulated Alzheimer’s or something, and it’s almost scary. It would be terrible to go through life with this level of non-recall.

The amnesiac effect of lack of names is reinforced by the lack of narrative, which is a function of lack of theme (or more concretely, lack of memetic cores). Over the 20 parts so far, I’ve wandered all over the place, with no centripetal force driving towards coherence. The parts were also far enough apart, there was no inertia from being in the same headspace between parts. It’s been a random walk of my mind.

This feels weird. It’s easy to remember at least a few highlights of themed blogchains, even if they lack a proper narrative throughline. I have a (very) vague sense of the ideas I’ve covered in the Mediocratopia or Elderblog Sutra blogchains for instance. Even if there isn’t a necessary order and sequence to the writing, a themed series grows via a web of association. So if you recall one thing, you remember some other things.

But order matters too. We remember things more easily when there is a natural and necessary order to them. This was reinforced for me in this blogchain in dealing with a bug The series plugin I use screwed up and indexed several of the posts out of order, which I took 5 minutes to fix. But reading the posts out of order made zero difference. Since they are not related, either by causation or thematic association, order is neither necessary, nor useful. It’s like how chess players have uncanny recall of meaningful board positions that can actually occur in a game, but not of boards with randomly placed pieces. It’s more than a mnemonic effect though. There is intrinsically higher randomness to a record of unnamed thoughts. The only order here is that induced by me and the world getting older.

This all seems like downsides. Recall is far worse, coherence is far worse. For the reader, the readability is far worse. Is there any upside to writing in this way? I’m not sure. It does seem to tap into a sort of atemporal textual subconscious. It also makes for a very passive mode of writing. A name is a boundary that asserts a certain level of active selfhood. A theme is a sort of grain to the interior contents. A narrative is a sequence to the contents. Each of the three elements acts as a filter to what part of the outside world makes it into the writing. When you take down all three, the writing occupies something like an open space where ideas and thoughts can criss-cross willy-nilly. It is homeless writing, with all the attendant unraveling and disintegration of the bodily envelope (I wrote about this in a paywalled post on the Ribbonfarm Studio newsletter).

A named idea space is a space with a wall. A named and themed idea space is a striated space with a wall (in Deleuze and Guattari sense). A named, themed, and narrativized space is a journey through an arborescence. A nameless, themeless, storyless space develops in a rhizomatic way, reflecting the knots and crooks of the environment. It is not just homeless writing, it is writing where there’s nobody home. It’s the textual equivalent of the “nobody home” affect of far-gone mentally unravelled homeless people.

Another data point for this effect. I just finished a paper noteboook I started just before the pandemic. So it’s taken me about 2 years to fill up. Back in grad school, 20 years ago, I used to be very diligent with paper notes. There was a metacognitive process to it. I’d summarize every session’s notes, and keep a running table of contents. I’d progressively summarize every dozen or so sessions. My notes were were easy and useful to review. Now I’m lazy, I don’t do anything of that sort. It’s just an uncurated stream of consciousness. With just a few pages left in the notebook, I tried to go back and reconstruct a table of contents (thankfully I was at least dating the start pages of each session) but it was too messy, hard, and useless, so I gave up. Progressive summarization ToC-ing is only useful and possible when you do it nearly real time. Naming and headlining work only when you name and headline as you work. So what I have with this latest filled notebook is just one big undifferentiated idea soup that’s nearly impossible to review. It’s worse than Dumbledore’s pensieve. It’s something of a memory blackhole. It is recorded but not in a usefully reviewable way. But arguably, not doing the disciplined thing led to different notes being laid down. I thought and externalized thoughts I would otherwise not have thought at all. I can’t prove this, but it feels true. And while it’s harder to review, perhaps the process of writing made it more transformative?

About the only thing I’ve been able to do with both this blogchain and the paper notebook, in terms of review, is go back (with a red pen or the editor) and underline key terms/phrases, and maybe tabulate them elsewhere into an index. I can trace the evolution of my thought through the index phrases. These nameless memories are indexable, but not amenable to structuring beyond that. It’s the part of your mind that you can Google but not map (this is the real “googling yourself”). These are demon notebooks. It’s dull to review now, but in a few years perhaps, it will be interesting to review as a record of what I was thinking through the pandemic. Maybe latent themes will pop.

Twitter of course is the emperor of such demon notebooks, though shared with others. I’ve taken to calling the nameless structures that emerge in my tweeting threadthulhus. These blog and paper demon notebooks though, are not threadthulhus. They are more compact and socially isolated. They are lumps of textual dark matter. They are pre-social, more primitive. They lack the identity imposed by mutualism.

With both this blogchain and my unreviewable demon paper notebook, I think I’ve kinda explored what names/headlines, target themes, and narratives do in writing: they alienate you from your own mind by allowing you to create a legible map of your thoughts as you think. Anything you structure with a name/theme/narrative (the alienation triad) is a thing aside from yourself that you can sort of distance from yourself and point to as an object, and let go off, and even meaningfully sell or give away to others. Alienation is packaging for separation. Anything that you don’t do those things to remains a part of you. This is not a bad thing. Not everything you can think is ready to be weaned from your mind. Even if you’re willing to share it with the world, it does not mean you are able to separate it from yourself. Just because you make second brains doesn’t mean first brains disappear. Exploring them is a distinct activity.

This sort of writing is arguably indexical writing. Writing as self-authorship. What doesn’t have its own name, theme, and narrative is part of you. In fact the only thing holding it all together is the fact that you’re writing it. This is a self-reinforcing effect. The act of writing in that mode sort of encourages those least detachable thoughts in your head from emerging and making themselves available to hold and be.

There is a paradox here. The most indexical writing is also the most open-to-the-world writing since it lacks filters. So it is both a self-authoring process and a self-dissolution process. What comes out is both most truly you, and not you at all. Self-authorship and self-dissolution are two sides of the same coin. Being is unbecoming. To be homeless is for there to be nobody home.

You could argue that it is the process of giving names, boundaries, and thematic and narrative structure to thoughts to externalize them that is a highly unnatural and strange process. Like mutilating your brain by carving out chunks of it to push out. I am not sentimental enough about the writing process to actually feel that way, but I kinda get now what angsty poets must feel.

I think this is the key difference between diary-writing or journaling and “writing.” The lack of traumatic separation and self-alienating packaging.

This experiment hasn’t yet run its course, and I might keep it going indefinitely, but I think I finally understand the point of it, and why I unconsciously wanted to do it and why I feel it helps the other writing.

Where do you go from this kind of writing? Well, if you continue down this course — and I already see this happening a bit — you head towards increasingly commodity language. You seek to avoid evocative turns of phrase, stylistic flourishes, and individual signature elements — anything that asserts identity. You seek to make the writing unindexable, not just unmappable. You seek to go beyond individual self-authorship and channel a larger vibe or mood. Or maybe you try to fragment your own mind into a bunch of authorly tulpas. Or maybe you mind-meld with GPT-3 and write in some sort of transhuman words-from-nowhere mode. Ultimately you get to various sorts of automatic writing. I don’t necessarily want to go there, but it’s interesting to see that that’s where this path leads. This is the death of the author as an authorial stance as opposed to a critical readerly stance. It’s a direction that naturally ends in a sort of textual suicide. At the level I’m playing it, it’s merely a sort of extreme sport. Textual base-jumping perhaps. But this direction has strong tailwinds to it. Increasingly large amounts of public text in the world form this sort of featureless mass that’s grist for machine-learning mills, and increasingly, no identity of its own.

You might say the natural end point of this kind of writing is when it becomes indistinguishable from its GPT-3 extrapolations and interpolations.

Or going the other way, there are potential experiments in radical namefulness. Everything is uniquely identifiable, memorable, evocative, and nameable, and has a true name. Narrative coherence is as strong as possible. Thematic structure and causal flow is as tight as possible. Un-machine-learnable texts. I’m not sure that kind of text is even possible.

People who have a literal-minded interest in matters that extend beyond their own lives, and perhaps those of a couple of generations of ancestors and descendants, are an odd breed. For the bulk of humanity, these are zones for mythological imaginings rather than speculative-empirical investigation, if they are of interest at all. For the typical human, what happened in 1500 AD, or what might happen in 2500 AD, are questions to be answered in ways that best sustain the most psyche-bolstering beliefs for today. And you can’t really accuse them of indulging in self-serving cognitions when the others who might be served instead are either dead or unborn.

As a result, mythology is popular (by which I mean any heavily mythologized history, such as that of the founding of the United States, not just literal stories of Bronze Age gods and demons), but history is widely viewed as boring. Science fiction is popular, but futurism (of the wonky statistical trends and painstakingly reality-anchored scenario planning variety) is widely viewed as boring.

But if you think about it, it is history and futurism that are the highly romantic fields. Mythology and science fiction are pragmatic, instrumental fields that should be classified alongside therapy and mental healthcare, since they serve practical meaning-making purposes in the here and now, in a way that is arguably as broadly useful as antibiotics.

History proper is rarely useful. The only reason to study it is the romantic notion that understanding the past as it actually unfolded, even if only 10 people in your narrow subfield pay attention and there are no material consequences in the present, is an elevating endeavor.

Similarly, long-range futurism proper (past around 30 years say) is rarely useful. Most political and economic decisions are fully determined or even overdetermined by much shorter range incentives and considerations. There is also usually crippling amounts of uncertainty limiting the utility of what your investigations reveal. And humans are really bad at acting with foresight that extends past about a year anyway, even in the very rare cases where we do get reliable glimpses of the future. So the main reason to study the future is the romantic notion that it is an elevating endeavor.

Who exactly is it that believes these endeavors are elevating, and why should their conceits be respected, let alone potentially supported with public money?

Well, people like you and me for one, who read and write and argue about these things, and at least occasionally try to rise above mythologizing and science-fictional instincts to get a glimpse of the past and future as they really were or will be, with high plausibility. And I can’t say I have good arguments for why our conceits should be respected or supported. Fortunately, they are also very cheap conceits as conceits go. All we need is time and an internet connection to indulge them, and a small cadre of researchers in libraries and archives generating fodder.

How do we even know when we’ve succeeded? Well of course, sometimes history at least is dispositive, and we find fragments that are no longer subject to serious revisionism. And sometimes the future is too — we can predict astronomical events with extraordinary certainty, for instance.

But that’s just the cerebral level of success. At a visceral, emotional level, when either history or futurism “work” in the romantic sense that interests me, the result is a slight shrinkage in anthropocentric conceptions of the world.

Every bit of history or futurism that actually works is like a micro-Copernican revolution.

When they work, history and futurism remind us that humans are no more at the “center” of time, in the sense of being at the focal point of unfolding events, than we are at the “center” of the universe. The big difference between space and time in this regard is that decentering our egos in historical time is a matter of patient, ongoing grinder-work, rather than one of a few radical leaps in physics.

In 2018, historian Michael McCormick nominated 536 AD as the worst year to be alive. There was a bit of fun coverage of the claim. This sort of thing is what I think of as “little history.” It’s the opposite of Big History.

Big History is stuff like Yuval Noah Harari’s Sapiens, Jared Diamond’s Guns, Germs, and Steel, or David Graeber’s Debt. I was a fan of that kind of sweeping, interpretive history when I was younger, but frankly, I really dislike it now. Historiographic constructs get really dubious and sketchy when stretched out across the scaffolding of millennia of raw events. Questions that are “too big to succeed” in a sense, like “why did Europe pull ahead of China,” tend to produce ideologies rather than histories. It’s good for spitballing and first-pass sense-making (and I’ve done my share of it on this blog) but not good as a foundation for any deeper thinking. Even if you want to stay close to the raw events, you get sucked into is-ought historicist conceits, and dangerously reified notions like “progress.” Yes, I’m also a big skeptic of notions like Progress Studies. To the extent that the arc of the moral universe has any coherent shape at all, to proceed on the assumption that it has an intelligible shape (whether desired or undesired), is to systematically blind yourself to the strong possibility that the geometry of history is, if not random, at least fundamentally amoral.

About the only Big History notion I have any affection for is Francis Fukuyama’s notion that in an important sense, Big History has ended (and I’m apparently the last holdout, given that Fukuyama himself has sort of walked back the idea).

Little histories though, are another matter. Truly little histories are little in both space and time, and much of history is intrinsically little in that sense, and fundamentally limited in the amount of inductive or abductive generalization it supports once you let go of overly ambitious historicist conceits and too-big-to-succeed notions like “Progress.” But some little histories are, to borrow a phrase from Laura Spinney’s book about the Spanish Flu, Pale Rider, “broad in space, shallow in time.” They allow you to enjoy most of the pleasures afforded by Big Histories, without the pitfalls.

Whether or not the specific year 536 was in fact the worst year, and whether or not the question of a “worst year” is well-posed, the year was definitely “broad in space, shallow in time” due to the eruption of an Icelandic volcano that created extreme weather world-wide. The list of phenomena capable of creating that kind of globalized entanglement of local histories is extremely short: pandemics, correlated extreme weather, and the creation or destruction of important technological couplings.

The subset of little histories that are “broad in space, shallow in time” — call them BISSIT years (or weeks, or days) — serve as synchronization points for the collective human experience. Most historical eras feature both “good” and “bad” from a million perspectives. Sampling perspectives from around the world at a random time, and adjusting for things like class and wealth, you would probably get a mixed bag of gloomy and cheery perspectives that are not all gloomy or cheery in the same way. But it is reasonable to suppose that at these synchronization points, systematic deviations from the pattern would emerge. Notions of good and bad align. Many of us would probably agree that 2020 sucked more than most years, and would even agree on the cause (the pandemic), and key features of the suckage (limitations on travel and socialization). Even if there were positive aspects to it, and much needed systemic changes ensue in future years, for those of us alive today, who are living through this little history, the actual experience of it kinda sucks.

The general question of whether the human condition is progressing or declining to me is both ill-posed and uninteresting. You get into tedious and sophomoric debates about material prosperity versus perceived relative deprivation. You have people aiming gotchas at each other (“aha, the Great Depression was actually more materially prosperous than optimistic Gilded Age 1890s!” or “there was a lot of progress during the Dark Ages!”).

The specific question of whether a single BISSIT year should be tagged with positive or negative valence though, is much more interesting, since the normal variety of “good” and “bad” perspectives temporarily narrows sharply. Certainly, BISSIT years have a sharply defined character given by their systematic deviations, and sharp boundaries in time. They are “things” in the sense of being ontologically well-posed historiographic primitives that are larger than raw events, but aren’t reified and suspect constructs like “Progress” or “Enlightenment.” There is a certain humility to asking whether these specific temporal things are good or bad, as opposed to the entire arc of history. Two such things need not be good or bad in the same way. History viewed as a string of such BISSIT beads need not have a single character. Perhaps there are red, green, and blue beads punctuating substrings of grey beads, and in your view, red and green beads are good, while blue beads are bad, and so on. We don’t have to have a futile debate about Big History that’s really about ideological tastes. But we can argue about whether specific beads are red or blue. You can ask about the colors of historical things. We can take note of the relative frequencies of colored versus grey beads. And if you’re inclined to think about Big History at all, you can approach it as the narrative of a string of beads that need not have an overall morally significant character. Perhaps, like space, time can be locally “flat” (good or bad as experienced by the people within BISSIT epochs) but have no moral valence globally. Perhaps we live in curved historical event-time, with no consistent “up” everywhere.

A week ago, for the first time in decades, I spent several days in a row doing many hours of hands-on engineering work in a CAD tool (for my rover project), and noticed that I had particularly vivid dreams on those nights. My sleep data from my Whoop strap confirmed that I’d had higher REM sleep on those nights.

These were dreams that lacked narrative and emotional texture, but involved a lot of logistics. For example, one dream I took note of involved coordinating a trip with multiple people, with flights randomly canceled. When I shared this on Twitter, several people replied to say that they too had similar dreams after days of intense but relatively routine information work. A couple of people mentioned dreaming of Tetris after playing a lot, something I too have experienced. High-REM dreaming sleep seems to be involved in integrating cognitive skill memories. This paper by Erik Hoel, The Overfitted Brain: Dreams evolved to assist generalization, pointed out by @nosilver, argues that dreaming is about mitigating overfitting of learning experiences, a problem also encountered in deep learning. This tracks for me. It sounds reasonable that my “logistics” dreams were attempts to generalize some CAD skills to other problems with similar reasoning needs. REM sleep is like the model training phase of deep learning.

This got me interested in the idea of tapping into the unconscious towards pragmatic ends. For example, using the type of dreams you have to do feedback regulation of the work you do during the day. I made up a half-assed hypothesis: the type of dream you have at night depends on 2 variables relating to the corresponding daytime activity — the level of conflict in the activity (low/high) and whether or not the learning loop length is longer or shorter than 24 hours. If it is shorter, you can complete multiple loops in a day and the night-time dream will need to generalize from multiple examples. If it’s less than a complete loop, there is no immediate resolution, so instead you get more dreams that integrate experiences in an open-loop way, using narrative-based explanations. If there’s no conflict and no closed loops, you get low dreaming. There’s nothing to integrate, and you didn’t do much thinking that day (which for me tends to translate to poor sleep). I made up the 2×2 above for this idea.

I have no idea whether this particular 2×2 makes any sense, but it is interesting that such phenomenology lends itself to apprehension in such frameworks at all. I rarely remember dreams, but I think even I could maintain a dream journal based on this scheme, and try to modulate my days based on my nights.

This also helps explain why people in similar situations might have similar dreams (such as all the “corona dreams” that many of us were having during early lockdown months). It also lends substance to the narrative conceits of stories like H. P. Lovecraft’s Call of Cthulhu, which begins with people having widespread similar dreams (where it turns into science fiction is in the specificity of the shared motifs and symbols that appear).

You don’t need to buy into dubious Jungian or Freudian theories of individual and collective dreaming to think about this stuff in robust ways. The development of deep learning, in particular, offers us a much more robust handle on this phenomenology. Dreams are perhaps our journeys into our private latent spaces, undertaken for entirely practical purposes like cleaning up our messy daytime learning (there’s other theories of dreaming too of course, like David Eagleman’s theory that we dream to prevent the visual centers from getting colonized by other parts of the brain at night, but we’re only hypothesizing contributing causes, not determinative theories).