Ep143 "How do things last?" Part 1: neurons to civilizations

in ancient history and how did it completely disappear? Today we're going to talk about the concept of persistence, from sharks to Roman concrete to DNA. So get ready for a great brain stretch. Welcome to Inner Cosmos with me David Eagelman. I'm a neuroscientist and author at Stanford and in these episodes we dive deeply into our three pound universe to understand some of the most surprising aspects of

our world. Today, we're talking about how things persist. I want to start today's episode by talking about Greek fire, which I'm obsessed with. This story begins with a refugee named Kalanikos who arrived in Constantinople in the late six hundreds and he handed the emperors a weapon that he'd invented, and this completely changed warfare in the Mediterranean. His weapon was known as Greek fire. Think of it like a flamethrower, a thousand years before the invention of the modern flamethrower.

Greek fire was a thick liquid. It was petroleum based, and you'd hurl this in pots or you'd blast it from siphons and it would burn on the water. And the key is you could incinerate the enemy's ships this way, and no one had ever seen anything like this, So the Byzantine army immediately leveraged this as a defense that broke the Arab sieges that were attacking Constantinople, and Greek fire quickly became the most important weapon of the age.

The Byzantines were able to turn naval battles completely on their head and control the waters. Now here's the thing. The recipe for how to make Greek fire was ferociously secretive. The recipe was compartmentalized, and it was known only to the emperor's circle and Klinikos's genetic line. And this worked out well. No one else knew how to make it, and over centuries the Byzantine chronicles credit Greek fire for saving the capitol, and by extension, the entire empire. But

the price of extreme secrecy is fragility. They had a small bus factor. Now I'm not sure how Cammon this term is. We use this in Silicon Valley a lot. The idea of a bus factor is how many people in your company would have to get hit and killed by a bus before no one knows how to do something anymore. You want to make sure that you always have a bus factor bigger than one like only Bob knows the password to our servers, because if Bob gets hit by a bus, then your company grinds to a halt.

The smaller the bus factor, the more likely the company is not going to make it. So you always need to make sure you have redundancy. So in the case of Greek Fire, you've got a very small bus factor for secrecy reasons, and all you need are some political convulsions or the failure of a single lineage, and suddenly no one knows how to make Greek fire anymore. And that's what happened after the Fourth Crusades sack of Constantinople.

After that, the Byzantines all remembered the weapon, but nobody remembered how to make it anymore. So the lesson of Greek fire is that critical knowledge can easily evaporate. And really, when we look across history, we find that's a common story. Things or organizations get started or invented, and eventually they

die out. So this kind of story arc how we once knew something and then totally lost it has always made me wonder how anything survives for a long time, how anything persists because the universe is slipping away moment by moment. But everywhere we look we do find systems that come into existence that we fuse to let things go. Brains hold on to a continuous sense of self, cultures hold on to very old traditions, religion's last hundreds of generations.

The body plans of sharks has continued essentially unchanged for hundreds of millions of years. So today's episode is about persistence, about things lasting through time. This is something I've always been fascinated with. So today we're going to weave together several paths to get to the bottom of a question. What makes things last from neurons to civilizations, and what do very different lasting things have in common? And this

is part one. Next week, in part two, we're going to talk about thinking about our current moment on a ten thousand year timescale, we'll talk with Alexander Rose, the former executive director of the Long Now Foundation, where we'll hear about what organizations last through generations and why. For today, we're going to start in the brain with very brief windows of persistence, and then we'll move out to see some of these same principles across decades and centuries and millennia.

So let's dive into the inner cosmos. Every sensory system faces a fundamental problem, which is that the world arrives in small brief pieces. The brain solves this by holding on, by making the signals persist longer than they would otherwise. For example, when you're hearing my voice, every sound I make is only in the outside world between us for some tens of milliseconds. But your brain holds onto those sounds, it retains them internally, so they're around long enough to

be combined with whatever follows. So, in other words, all that comes through your speakers at any moment is a single phoneme, a little unit of speech. So you hear one phoneme and then another, and then another and then another. And the key is that it wouldn't mean anything to you, except they are able to hold them in your brain and replay them so that each little moment is heard with the others kept in mind, so you hear speech.

So with these little sounds of speech, these phonemes, the brain preserves each of them long enough to be able to understand sounds in context. If the sound appeared and disappeared as quickly as it does from the world, nothing would have any meaning. And I'll give you another example which will make this even clearer. Music Music also depends on persistence in the brain. You can understand a melody only because earlier notes remain present in your brain when

the later notes arrive. Without that, music would collapse into isolated blips with no rhythm and meaning. Why Because a single pitch is just an air compression wave. It arrives, it vibrates the ear drum, and then it physically disappears on its own. It doesn't carry much meaning. But when notes arrive in sequence, a melody appears. But the melody only exists because of the brain's ability to keep information alive. As each note reaches your ears, the previous notes are

still present in neural activity. They remain in there shooting around in the neurons, and they do this long enough to overlap with what comes next. This allows your brain to create relationships across time. What you hear at any moment is shaped by what you heard in the moments before. So, just like speech, music depends on this internal persistence. A melody requires comparison, just like a word or sentence does. When we talk about resolving tension in music, this is

only possible because the tension has been sustained. Even understanding a simple beat depends on your brain's capacity to predict the next thump by preserving the timing of at least the last two thumps. The stuff of the outside world has to be held onto and maintained. So let me drive this point home with an example that I love. A tune that you recognize has nothing to do with the individual notes. Instead, it's all about the relationship between

the ne notes. And this is why you can transpose a tune to any key, so that all the notes are different, but their relationship is the same. So here's happy Birthday and the key of C. But if I transpose this to another key, you'll see that the individual notes are different, but you have no trouble recognizing the melody. Why can you still recognize the tune? Because it's not about the individual notes. A piece of music is about the relationship between the notes, and that relationship depends on

holding in mind what just came before. So here's my assertion for today. I suggest that music would not be understood by any the alien species that doesn't keep signals alive internally in their nervous systems for a little while. If some space alien didn't have a way to make signals persist, but instead they were just reactive to the moment, then they would have an immediate sensory experience for each note, but they wouldn't be able to understand the notes as

having a relationship. Music only makes sense when you think about it as the tension between what is being held and what arrives next. Same with language, of course, if the alien heard each of our phonemes and then forgot it before the next phoneme arrived, language couldn't mean anything.

Everything would just be disconnected. Sounds. Meaning depends on holding earlier signals in mind while later ones unfold, so the brain is constantly retaining signals and giving them a law longer life than they actually have in the outside world, and we see this across all the senses. Take vision. Imagine you're sitting in a dark room and I flash an LED light on and off. It's so brief, let's say twenty five milliseconds, that it's already gone before you

can register it. But here's the wacky part that my lab has done a lot of experiments on. The flash seems to you to last a little while, like one hundred milliseconds. It's not existing in the world anymore, but it exists for even longer directly in your conscious experience. The neurons that represented in awareness continue firing. That brief flash is stretched in time. This phenomenon is called visual persistence, and it's a version of the same idea, but this

time directly in conscious awareness. That perception makes things last longer than they exist in the world. And visual persistence is the only reason that movies work, because we're able to perceive continuity and motion from looking at still frames flashed one after the other. In other words, because of persistence, successive images overlap in the mind. The brain integrates pictures over time, blending what just happened with what is happening now,

and the result is motion from flashed still frames. You can see this visual persistence with a simple demonstration. If you flash a light on and off, and on and off and on and off really quickly, it will fuse into what looks like a steady light. The speed at which this happens is called the flicker fusion threshold, and it's normally about fifty five hurts, meaning you do this on off, on off fifty five times every second. So if a light flickers fast enough, the flicker goes away

and it seems like a solid light now. One of the things my lab has studied over the years is the way we can leverage that as a diagnostic tool, because it turns out that not all brains make things persist for the same amount of time, and specifically, people who are suffering with schizophrenia have a longer window of persistence. A brief flash lingers in their brain for a longer time.

What this means is that the visual world has more overlap, So, for example, a person with schizophrenia has a lower flicker fusion threshold. What that means is that I can flash a light on and off more slowly and the person will see that as a steady light. In other words, persistence lasts longer. Now I'm telling you that little detail about schizophrenia because this allowed me to build a purely visual test for diagnosing schizophrenia, and this is something I'm

currently running tests on at Stanford. If you're interested, I'm linking my papers on this to the show notes. So let's zoom out to what we've seen so far about how the brain makes very short signals persist. The brain constructs what the psychologist William James a century ago called this specious present, by which he meant that the present moment is not actually an instant, but it's a short time window. This little sliding window of now causes the

present moment to have a certain thickness. Everything sticks around for a bit in the brain so that events in time can get compared. In other words, your neural circuits refuse to let the world disappear too quickly. And we see this kind of temporal integration in every sensory system, which allows the immediate past to remain active in the present, whether the brain is stitching notes into music or still frames into motion. It relies on short lived memory to

transform fragments into meaning. So persistence is something we talk about in neuroscience, but today I want to explore the notion of persistence from animals to nations to civilizations. And in thinking about this, I now think there are five main ways that things persist. So let's dive in. The First reason something persists is that it's optimized on some

fitness landscape, meaning that it can't really get better. So generally, when we think about evolution, we picture a constant transformation of species, with new forms replacing old ones, like the cartoon of Homo sapiens evolving from a common ancestor with great apes until he's walking upright. But what's interesting is that some animals haven't changed in hundreds of millions of years, like sharks. Little details have shifted, but the overall body

plan has persisted. Why Well, because they work, they're streamlined, they have very efficient propulsion, and their sensory systems are perfectly tunfoer hunting, so the design continues to succeed. So this first reason is obvious. Some things persist because they are at or near a local optimum, and if they were to change, that would probably introduce risk, as in, you're more likely to muck up the system rather than to improve it. So stability becomes the winning strategy, and

persistence here is just a signal of success. And of course we see this with everything. Technologies persist when they meet needs efficiently. We can see these kind of technological sharks everywhere. The bicycle rolling down the street still has the triangular frame that was designed over a century ago. It's a design that balances strength and weight and efficiency so well that most attempts to reinvent the bicycle these remain curiosities. Or look at steel shipping containers. These stack

and ports around the world. They have standardized dimensions and that locks ships and cranes and trucks and warehouses into a single, coordinated system that would only get messed up if you tried to change it. Or look at this screw with its spiral thread which translates rotation into linear force. This solves its little mechanical problems so cleanly that its basic form has persisted for centuries. Or take the four

tined fork. This is centuries. There are variations, with the basic form of a fork stays the same because it accomplishes spearing and lifting efficiently across lots of food types. So these technologies persist because they reached a place of alignment with performance and costs and usability. Just like this shark, they solve their ecological problems efficiently, and once a form settles into a valley where it works well, staying there

often proves a better decision than constant reinvention. So, like I said, I think that's the obvious reason why some things persist. But I was thinking about this, and I think there are four more interesting angles that we could take on this about other reasons why things last. For the first one, let's turn to something I think about often. Why you can recognize a friend that you haven't seen

in like ten years. This happened to me at my high school reunion, where I saw people whose hair had changed entirely, maybe they didn't have hair anymore, Their face had changed shape a bit, their body had changed shape, and yet I was able to mostly recognize them, so something about them remained unmistakable. How do our brains allow someone's identity to persist for so long? And, by the way, I've wondered the same thing when I look at Google

or Apple photo albums. In these photo albums, they recognize you or your kids through decades of time, even though you can see that everyone's faces have changed a lot, everyone's gotten older. So what's going on with brains and algorithms? Okay, for both of these, it's simply that they're figuring out

which aspects of a face are worth carrying forward. They're sensitive to deep structure rather than the surface detail, so things like the spacing of features, how how far apart your eyes are, and how that relates to your nose, and the details of your mouth and your smile and your filterum, that's the vertical groove that connects to the bottom of the nose to your upper lip. All that geometry remains stable enough to anchor down your identity even

while everything else changes. This is why airport security cameras can recognize the face even if you dye your hair or wear colored contacts or whatever. Recognition works because certain information persists, and the algorithms can ignore certain kinds of change, like shifts in the lighting or changes in the hairstyle. They filter all that stuff away. And for your brain, in particular, it's training on faces from infancy onward. It sees them again and again, and it learns which features

persist across situations. Over time, it builds an internal model that figures out what matters enough to preserve. This question of what information to keep is asked by memory. More generally, most of your world passes through you and you don't remember it. Your memory only retains some experiences and decides the rest just doesn't matter. And in fact, culture works this way too, It repeats practices that continue to matter while letting most others drift into obscurity. What these all

have in common is that persistence depends on abstraction. What survives is something a little removed from the details. So the brain's ability to recognize a familiar face after decades is one example of a universal strategy, which is preserving identity by learning what to ignore. In other words, persistence is about knowing what to hold on to. And nowhere does this apply more than to the notion of your own self. You consider you you, even though you have

changed enormously over the years. You're not the same person that you were five years ago, and certainly all your cells have changed. I refer you to Inner Cosmos episode eighty two, about the continuity of the Self and the

ship of theseus for more on this front. But The point I want to emphasize now is simply that you change through time, but you have certain things in common, like your name and the stories you like to tell, and where did you go to college and so on, and you end up with the illusion of a stable self that persists through time. Now here's the thing that might come as a surprise. The persistence of your self is something that you have to learn. This is what

we call a cognitive development. So look at very young children. A young child can recognize her reflection in a mirror pretty early in life. She notices that when she smiles, the mirror image smiles. She notices that when she touches her face, so does the mirror image. So that's good. But this kind of self recognition doesn't yet work across time. If you show that same child a video of herself

from a year earlier, something surprising happens. She might point to the screen and say that's a girl, or she calls the child by another name. The idea that this earlier version of her belongs to herself hasn't yet taken hold. So selfhood across time is something we have to figure out. What happens over the first few years of life is that children gradually assemble a sense of persistence. They come to understand that the person who fell down yesterday is

the same one standing here right now. This depends on memory and languag whig and narrative all developing together, so the self becomes a story that can be updated while retaining a central thread. Now, one consequence of this is we all have what's called childhood amnesia, meaning we don't remember anything from the first few years of our lives. Now, it's not that you didn't have memory during those first few years, because you were remembering a ton of stuff

and learning from it. But what was missing was a durable narrative to scaffold everything. You had experiences, but there was no persistent identity that you could stitch them into. As you grew, you began to tag experiences as belonging to you. You learned which details madded enough to carry forward and which didn't really matter. So your identity emerged as a stable pattern extracted from constant chain, just like

with face recognition algorithms. And now on to the next way that some things persist, and that has to do with self repair. So, for example, if you walk along an ancient Roman Harbor, you can still find concrete structures that have endured wave after wave for two thousand years. These blocks were sitting in seawater, exposed to salt and

stress and erosion, but they only got better with time. Now, how in the world does Roman concrete shrug off centuries of battering while our modern concrete would spawl and corrode under these circumstances, Roman concrete persists because it can self repair. This is probably accidental, but it's baked into the chemistry. In twenty twenty three, a team at MIT performed all kinds of imaging and found that Roman builder hot mixed their concrete with quicklime, which left tiny lime clasts scattered

through the mortar. And these act like self healing capsules. So when micro cracks form and water seeps in, calcium dissolves and reprecipitates this calcium carbonate, which seals the crack. So, whether purposeful or not, the Romans achieved persistence of their structures.

So even though the concrete lasts, how did this knowledge fade? Well, it's partly because history re optimizes, so various other recipes for cement took over, most notably Portland cement by the eighteen hundreds, mostly because Portland cement cures quickly and it pairs well with steel. In other words, we switched goals to speed and early strength and skyscrapers and so on,

and so the recipe changed. But just go stand in an aqueduct and run your fingers over the mortar and you'll see a material that knows how to persist across centuries by self repair. And now I want to move on to a fourth way that things persist. And by the way, none of these methods are mutually exclusive. To understand this, one walk into an old cathedral and look closely at the windows. The glass often appears thicker at

the bottom than at the top. Why well, what was explained to me when I was touring a cathedral in Europe a while ago, was that the glass behaves like a very slow liquid. You give it enough centuries and it gradually flows downward under its own weight. What's cool about this idea is that it invites you to imagine time as a force so powerful that even solid matter eventually yields. And by the way, I've heard this explanation of cathedral glass a dozen times since then, But that

explanation is totally wrong. Glass at room temperature behaves as a solid, and its viscosity is so high that any flow would require time scales far longer than the age of the universe. As it turns out, the uneven thickness of old windows comes from manufacturing methods. Medieval glassmakers shaped pains unevenly, and installers generally placed the thicker edge at the bottom for stability. It has nothing to do with

gravity pulling on it, so the glass hasn't moved. But the myth goes on, and the reason the myth persists is more interesting than the myth itself. The reason the myth persists is that ideas follow rules of survival that differ from the rules governing matter. An idea can last because it feels explanatory, or it resonates in some way. Ideas persist not necessarily because they're true, but because they

have some other quality that makes them sticky. One might hypothesize that the glass myth persists because vast spans of time are awe inspiring, and or because the story translates the abstract concept of time in something visible, and or because it makes the person who says it sounds smart. Whatever the reasons that it became sticky. Once an idea is embedded in memory and culture, everyone starts to repeat

it and so it gets carried forward. So what this tells us is that with human culture, persistence doesn't require accuracy, just good transmission from brain to brain. And we see this pattern everywhere, like with urban legends. With an urban legend, a story begins somewhere, maybe with a little grain of plausibility, and then it travels. There are urban legends about contaminated Halloween candy, or vanishing hitchhikers, or a friend of a

friend who narrowly escaped disaster. And by the way, when you travel around the world, you see that urban legends shift their details according to local fears and cultural contexts, but the core structure stays the same. Why do these sorts of stories persist. In part, it's because urban legends attach themselves to emotion, usually some combination of surprise and danger, and this emotional component buffs up the recall and people love retelling it, and the stories are shaped and refined

for memorability the way that good jokes are. So urban legends unmask the same principle of persistence that the myth about medieval Glass does, which is that ideas can survive just by embedding in shared memory. Once a community starts to repeat a story, that repetition becomes a stabilizing force, and that lets it endure across time. Now beyond explanatory stories about class and urban legends, this is true of cultural beliefs of all sorts. Takes something like religion. Now.

I happen to be of the opinion that religions can be very useful for people and communities. But I also don't think, with two thousand religions on the globe, that one of them is true and all the others are false. But for today's purpose, what I think is fascinating is the persistence of religions, the way their stories continue to

be told and retold across tens or hundreds of generations. Now, like the glass or the urban legend, some of the particular ideas and religions survive because they're well told stories, and mostly because they strive to explain some meaningful mystery in our lives. And I think there's a social flywheel effect such that if your parents told you a story, it feels very lovely to pass that on to your

own children. On top of that, you have ritual repetition, you have texts, you have music, you have great architecture. And religious systems also endure because they distribute information across a huge swath of people and places, and this redundancy creates resilience. We'll come back to that in a moment. So put all this together, and religions are persistent engines. They solve the problem of how to keep an idea

alive across centuries of change. In this way, cultural ideas like religions can often look like a cultural habit, in the sense that once they're established, they keep on trucking along automatically even when their origins fade. Sticky religions persist because they prove effective at capturing attention and moving from one mind to another. By the way, the reason I say sticky religions is because there are tens of thousands

of other religions that we don't even see anymore. They lasted for a long time, and people gave up their life for those deities, but the religions weren't sufficiently well constructed to persist to the present. So we see that religions and many type of cultural stories last because they were optimized for transmit and that lets us revisit Greek fire, which I mentioned at the beginning. Greek Fire illustrates the converse. I told you that Greek fire was a weapon so

powerful that as shaped to the fates of empires. But the formula vanished precisely because there wasn't enough redundancy for carrying it forward. The bus factor was too small. Greek fire was super powerful, but power didn't translate into endurance. So persistence isn't just about impact, but it depends on transmission. So back to the big picture. Some things persist because

they're optimal, or at least locally optimal, like sharks. Other things endure because they're strong and self healing, like Roman concrete, and we just saw that. Other things persist because they're structured just right to be memorable. They're sticky, and once an idea digs in, it can outlast the originators in any evidence that got it started in the first place. And the idea's mechanism of transmission is moving from one

brain to another. And now I want to turn to the fifth reason why some things last, And for that, think about when you were a kid and you played rockstis or paper, and one of the rules was that paper beats rock. I was thinking about this recently in terms of the way that information gets written down. Until recently, almost two thousand years ago, people used to carve proclamations into stone, and stone has the advantage of feeling really permanent.

Then paper got introduced almost too millennia ago in China and in Europe about a thousand years ago. The problem with paper is that it's very fragile by comparison. It tears, and it burns and it disintegrates. But what's happened cross history is that paper has outlasted stone. Why it's because monuments erode. Wherever you find any ancient inscription, you'll see that time has softened it and it's nearing disappearance. You see that ancient statues lose their faces and their name plaques.

So as material, stone persists, but the information carved into it gradually fades away. Now, the surprising contrast is that ideas written on paper have much longer lives. Why it's because they're copied and recopied and translated and rewritten. Here's the key. The individual fragile pages vanished, but the pattern survives. It seems so crazy that paper could beat rock, but it does because it can be easily replaced. So what does this mean? It means that your ability turns out

to matter less than replicability. A medium that can be reproduced easily travels farther through time than one that demands physical permanence. What survives is what can be regenerated when it breaks. And once you start looking for this principle, you'll see it everywhere in nature. Take genes. The wild part about the DNA code is how long it's survived

and how massively successful it is. This simple little codebook has survived across billions of years, and the key is that, just like with paper, its longevity doesn't come from physical toughness. Individual DNA molecules are unbelievably fragile. They break easily when you subject them to heat or other kinds of radiation or chemicals they don't like. But the genetic code keeps

on trucking. Why DNA persists because it is copied. Every time a cell divides, its genetic information is rewritten onto new base pairs. And whenever you do that, you check for errors and try to correct mistakes, and you scan for damage and fix it. But the key is that the original molecule doesn't need to survive as long as the pattern survives. This makes DNA less like stone and more like paper. The medium is expendable and the information

is what matters. Even though the molecule is fragile, the information continues. And this is the same principle with memories persisting in your brain. You can still remember, for example, your third grade teacher's name. Now, why is that weird? It's because every cell in your body is getting all its pieces and parts replaced every moment of your life, such that you're an entirely new person. Some years later, none of the the molecules that you had in your

body in the third grade are there anymore. It's all new stuff. But you still have your memories and your knowledge and your perception because those are stored in the constantly rebuilding structure that persists. Your whole brain is made of very fragile stuff, just little cells, but it keeps the structure even as the details get replaced again and again. So back to religions and cultural stories. Even more generally, I argued in a minute ago that cultural stories last

because they're useful, they're memorable. They propose answers to questions that people have, and they give rituals and social coherence. But the interesting part here is that you can think of something like the Catholic Church living on top of generations of humans that are born. They absorb the religious dogma from the elders, they pass it to their children, and they pass away. The Catholic Church is something like

information living on top of constantly recycling humans. Culture persists because ideas are reenacted and recopied across generation, just like the way your memory of your third grade teacher lives on top of constantly changing neurons. So across the spectrum

from brains to civilizations, persistence sometimes depends on renewal. Unlike something that lasts because it's optimized or strong, Persistence very often emerges from something fragile, like paper or DNA or memories or humans, and you keep passing the message along. In other words, what you want is not necessarily the thing that doesn't get damaged. You want fragile elements organized into replicating patterns. So let's wrap up. Today is a meditation on how things last. And what we saw is

that across all scales from neurons to cultures. Persistence follows just a few patterns. Systems that rely on a single instance those vanish easily, but systems that distribute themselves endure, and systems that can be repaired or self repair outlast. Systems that demand perfection. Now, coming back to the beginning, what I find interesting is that we're always told to live in the present, But the fact is that the world you inhabit is not just the knife edge of

the present. It's a dense accumulation of held moments. Every habit you have, every story you tell, every building that you live in, the culture that you are embedded in. These are all solutions to the same problem, how to keep things alive for longer. Even podcasts like this one participate in that process. Sounds held in your brain for seconds become ideas held for years. A story that you hear becomes part of your internal landscape, maybe for the

rest of your life. The cosmos outside continues to change every second in its vast motion, but the job of your inner cosmos is to stitch moments together to make some things in the world persist. And now we're well positioned for next week's episode, where I'm going to talk with Alexander Rose, the former director of the Long Now Foundation, and we're going to dive into his research on organizations that outlast their founders might say over a thousand years.

What makes them last? Join me next week to find out. Go to eagleman dot com slash pod cast for more information and to find further reading. Join the weekly discussions on my substack and check out and subscribe to Inner Cosmos on YouTube for videos of each episode and to leave comments. Until next time. I'm David Eagleman, and this is Inner Cosmos