The following is a rough transcript which has not been revised by The Jim Rutt Show or David Krakauer. Please check with us before using any quotations from this transcript. Thank you.
Jim: Today’s guest is David Krakauer, a longtime good friend and president of the Santa Fe Institute, and he’s agreed to be our prototype of the WorldView podcast. Brave man. Right? Welcome, David.
David: Start at the first rung of the ladder. That’s how I think about it. But, anyway, it’s wonderful to be with you.
Jim: So, David, you woke up this morning. A few seconds later, you were more or less conscious. Who was this guy that woke up this morning?
David: Someone with a viral infection given to me by my brother over the Christmas holidays. So there’s that. Well, I woke up thinking about this conversation, not knowing what we were going to talk about. I think I’m a pattern seeker. That’s kind of boundaryless for me. I tend to procrastinate in the mornings. That’s what I do in the mornings. And I procrastinate by pulling random books off shelves, typically, not always. I mean, you know, and reading something that I think by some weird synchronicity is going to prove to be useful in my thinking. And I genuinely believe that, by the way.
I have a commitment to chance, so I think maybe that’s part of who I am. I think I try not to exclude things based on strong opinions of what will be valuable or not. I have no confidence in that.
Jim: So you’re casting a broad fishing rod out into the waters and, you know, hope that the yield is worth the time.
David: I’m kind of an amoeba. Here’s a good one. I wake up as a gas. I go through the day and become a liquid, and then I go to sleep as solids. Maybe that’s the right way to put it.
Jim: And what do you think about when the morning, are you the same person that you were before you went to bed? We don’t know for sure. Right? It’s always an interesting question.
David: I tell you what’s more important than sleep for me, and that’s, you know, we walk the dog in the morning, and I can say this almost unequivocally, unconditionally. Nearly all of my best ideas come to me when I’m moving. And I don’t have to be self-propelled. Right? I could be in a train or whatever. Typically, I am self-propelled. And it’s that gas point. Right? It’s not particularly focused. I’m not being restrained or constrained by someone else’s limited ideas.
And I wake up in expansive mode. I think that’s what I am. And in that sense, right, actually, to your question, nearly identity-less, right, less identity than later in the day, where things tend to congeal through a system of interactions.
Jim: And so the felt sense of David Krakauer is weaker in the morning.
David: Yeah. Weaker. More interesting.
Jim: That’s very interesting, actually. Now the universe you woke up in, what is this universe? You think it’s real? You think it’s only in your head? Is it a simulation or something entirely else?
David: No. I think it’s real. I’m a realist. Yeah. I don’t have the imagination to simulate everybody else and the universe. Right? So I choose to believe that my limited capacity implies the existence of reality. And, also, it would be a perverse mind, right, that only disclosed to me truths slowly, partially, imperfectly, or not at all. Right? Why not just tell me? I find it so implausible.
So, yes, I believe there is a real world out there. I think it’s a real universe that we barely understand. I mean, that’s something that I probably share with you, but not with a lot of physicists, which is that I don’t think we’ve scratched the surface. I believe them.
Jim: Yep. I’m with you there that it may not even all be knowable. Right? As we know, there are limits to knowledge, both practical and fundamental, but I’m with you. I’m a realist. Actually, my old age, I’ve become a little bit philosophical. I’ve concluded that the stance of realism is necessary even though we can’t prove it.
David: Yeah. I like that. It’s a good operational principle to get work done.
Jim: Exactly. I wrote a nice essay on that called, minimum viable metaphysics, and that was the first point. We can’t prove it, because I like to tell people. I cannot prove the universe didn’t flick into existence five seconds ago and will flick out of existence five seconds from now with all of our memories in place for that 10 seconds, fossils in the ground, airplanes flying. Can’t disprove that. Yeah. But if you’re going to have any traction in the world, at least it sure seems that taking a realist stance is the stance that gives you the most traction.
David: I agree. I mean, realism is like parsimony. Solipsism is just too much work. Right? It’s like, not only would I have to work out the world for myself, but I have to translate it into some weirdo version of the world that you have. Right? Despite the fact that you are my invention. I mean, the whole thing is so laborious, not worth pursuing.
Jim: And then also you have to explain, how is it that you and I, even though of your creation, can more or less have an open-ended conversation about things we don’t necessarily know about. Right? You know, it just becomes insane. So let’s move on with the journey of David and who he is and what he thinks. You’ve said a couple times at least that one of the strongest things you remember about your intellectual formation was when you were 12 or 13, you had an epiphany about the idea of the idea. Talk about that a little bit.
David: Yeah. You know, it’s funny. Let me put a little preamble to that, and that is that I have or I still have, but I had very strongly when I was young, a perverse attraction to the arcane, to the esoteric. And I don’t mean the occult. Right? I mean the difficult. And when I was growing up, there were still bookshops. Right? And in London, particularly at that time, Foyles, you know, as you know, up in Charing Cross Road there and Tottenham Court Road. And I loved books whose covers suggested something outrageously difficult or exclusive intellectually.
And so I would buy books, Jim, that I had no chance of reading. I mean, no way. I mean, you know, something on the psychoanalytic theories of France between 1971 and 1972 or something. And I would keep these books by my side almost as talismans. And what’s going on, right, is a kind of almost idolatrous relationship, not with the book itself, right, but with the ideas that it contains. And I found that extremely so that is kind of pornography, where it wasn’t, you know, nude bodies. It was difficult concepts. And I still, to this day, have a kind of weird physiological relationship to ideas. Right?
Jim: And, of course, ideas vary in quality. Your French psychoanalytical theories from 1971 to ’72, probably mostly utter garbage.
David: Probably. But I would never have known that. That’s the point. Right? And so that’s the beginning of it. And then, of course, very early, actually, my teens, this kind of coalesces into a set of core interests that have never left me, right, which are, on the one hand, the evolution of order and, to include the evolution of ideas, through the instrument of the mind and brain. And I don’t view those as separable.
Most people, by the way, nowadays will get into all this rubbish on LLMs and all this. Actually, the strong functionalists, the immaterialists, who are basically modern dualists, don’t take matter seriously enough as far as I’m concerned. So I consider evolution integral to any conception of intelligence and consequently to the world of ideas. So my career decisions were very directly connected to this early obsession, I guess.
Jim: Interesting. And so the idea of the idea led you to think about what does it mean for there to be an idea? What is it honed in? What are the attributes of something that’s differentiated from noise, for instance. Right?
David: Exactly. And where does it come from, and is it everywhere? Is it in things, nonliving things? And, of course, once you go down that path, which seems a bit silly, you realize there are entire edifices built on this conception. Right? Platonism and on it goes. And then you make contact with a world of other people who have had the same pornographic fixations as you. Right? That is with the idea. Then you develop a kind of virtual community.
So that’s a very important point. Who am I? I am a person whose community is mainly not alive. I live in the graveyard of ideas or of books, much more than of community, despite the fact that I’m in community. But the history of ideas is as alive.
Jim: And how would you say you have held true to that spark since your 13-year-old self or 12-year-old self? And maybe to what degree have you drifted a little bit from that initial idea of the idea?
David: Well, as you know, because you’ve been such an important adviser to me in terms of my professional deformation into a director of an institute, there’s a little bit of a digression, diversion. Although you could argue through some subterfuge sophistry that it’s achieving the objective indirectly. I think, intellectually, I’m extremely true to a vision of what I only now understand to be the complex world. And I didn’t have a name for it then.
But if you think about it—I mean, we have these shared interests. What is this interest in combinatorial games, evolution, intelligence? What is that? And we now slowly it dawned on us that, oh, wait a minute. It’s a real phenomenology, and it gets called complexity. And somehow, universities that was a digression for me. Right? In other words, institutional education, not that it was bad or anything, but it was definitely stepping off the tracks and heading in the wrong direction. And I had to get back on track, actually, much later.
Jim: I look back at my own life, and, you know, I ended up jumping out of science and jumping into just being a mischief maker in the world. And I now realize if complexity science had existed in 1975, I would have been all in. Right? I still recall taking a small course seminar for Paul Samuelson himself on macroeconomics, 15 students. And I came away from that saying, this is the biggest load of horseshit I’ve ever seen in my life. Right?
And if they had tried to come in with a complexity perspective and link micro, which there’s at least a little sense to it, and macro and explained that, okay, we can’t necessarily tie all the pieces together because of deterministic chaos and lots of other things. But that would have been interesting. But instead of just like making up these high level formulas that I go, what the fuck? This ain’t science. This is like reading chicken entrails or something. Right?
David: Well, it’s interesting. Right? Because this gets to something that I think I know we both worry about, which is the risk averse character of most scholars. Right? I mean, scientists are supposed to be deeply reverent. Right? I mean, that’s the game. It’s a revolutionary game. But, unfortunately, once it was institutionalized, probably sometime in the eighteenth century, god knows, it attracted people wanting careers. And there’s nothing wrong with career. I don’t mean to, you know, we all need them.
But I think that there’s something, in fact, zero sum about the pursuit of the career and the pursuit of revolutionary new ideas. I think that they’re actually probably at some deep level incompatible. Right? Because you’re going to end up alienating everyone, of course, by virtue of the novelty of your concept. And I think that’s the very difficult balance. Maybe that’s something I also think about a lot, actually. Not when I wake up, but when I try to go to sleep. Oh, shit.
Jim: Of course, you know, that’s the whole point of Santa Fe Institute is to somehow subvert that dynamic. And you know what? You’re absolutely right. I’ve been on various governing boards of various scientific institutions, and I will say that in most cases, it is certainly the careerist dynamics which drive them. But you maybe talk a little bit about how you and the Santa Fe Institute tradition have attempted to fight back against that.
David: It’s really hard because it’s not a naturally stable equilibrium. It needs to be driven by energy. We’re a big bull on a very, very steep precipice being held up by a few 100 Sisyphuses. You know? That’s how I view it. And it could fall at any moment, and I think, anyway, that makes it exciting, but it’s also challenging.
I think one is knowing what we’re discussing, being a little bit honest about that more than most people are prepared to be. Because there’s nothing wrong with this fact. Right? I mean, everyone knows this. I mean, you know that from your career. You started off hot blooded, iconoclast, and slowly, you were assimilated in the Borg collective. I mean, it just happened. It’s all of us.
Jim: Never.
David: No. Okay. Well, I’m trying, but you know what I mean. I think for most of us, it’s very difficult. Yep. And I think you have to know that. And the worst thing that happens, right, is people who dissimulate and say, no. No. I am the CEO of this paper production company, and I remain as radical as I ever was, especially when I’m having to report out to the shareholders. So that’s number one.
And two, and then put in place, to be honest, habits, principles, practices that are aligned with this ambition to be more revolutionary somehow. And we do that all the time by not, you know, part of it is like a negative principle. Like, we will not have the, we will not do this. We will not do this. We will not do that. So SFI is blissfully free of the denitis that universities have, all that intermediate bloat that is a, I mean, might serve an important allocated function. I am not convinced, but, okay, we don’t have any of that. Right? So it’s basically researchers and then a little pimple on top like me and a few others who are essentially ignored. So it all works.
Jim: Yeah. SFI has been quite remarkable how long they’ve been able to maintain that metastable balance. Right? Been some challenges as we both know. But absolutely. So far so good. Let’s move on to something else, which I did not know you were involved in, so I did a little research this morning, and that is this Ouroboros idea of yours. The idea that, you know, I always knew you were an anti foundationalist as am I, but didn’t quite know how radically anti foundationalist you’ve become in your old age. Right? Did you develop this with Wolpert, I think I read?
David: Yeah.
Jim: Yeah. Why don’t you lay out this idea? This is really, it might be a complete crock, but it’s very interesting.
David: Okay. Let me explain. So here’s the standard model that everyone gets brainwashed into believing. And it’s the standard model, actually. Right? It says that the most fundamental of all the disciplines is physics because it deals with the most fundamental ontological components, elementary particles. And by aggregating them and taking various limits in various regimes, you slowly explain collective forms of matter. Right? I mean, ultimately culminating in the living state. And all of these approaches to understanding these more aggregative collectives are called the special sciences. It’s sort of like, you know, special ed. Physics gets to be foundational.
So this has led to two kinds of silliness. Right? One kind of silliness is greedy reductionist, what our colleague Dan Dennett called greedy reductionism, which is, ultimately any truly legitimate explanation has to be expressed in terms of the most fundamental constituents. So eventually, the poetry department will be expressed as the library of Alexandria of, you know, quantum gravity or whatever your preferred foundation is, which is, you know, patently ludicrous. Okay. The for me, at least, silliness that it’s produced is the structure of the academy in a way, which is the proliferation of departments and specialties that might not have an ontological basis that are historically contingent. They’re accidents.
Jim: Right? Accidents of power. Right? Music theory departments. Right?
David: So the point is that there is this linear conception of reality that’s a kind of a scale theory starting with very little things that we call fundamental, increasingly aggregated things that we call effective, and that’s the story. And what the Ouroboros says is is a bit more radical. It says, you know what? Here’s a funny fact. You start with language and mathematics. You conceptualize the universe in terms of mathematical principles, fields, forces, particles. You develop physics. You use your physical theory to develop chemistry, which explains some elements of biology and so on. So you started with the thing that should be the most derived state, mind, and you derive the most fundamental state, physics.
So that you could start with biology. You could say, well, look. Everything starts with evolution after all. We start as prokaryotic things, single celled organ. They become multicellular. They develop tissue specialization, nervous systems, eventually material culture and language, eventually theories. Out pops physics, you know, and the circle continues. So you can kind of what David and I was saying is that epistemologically, and this gets an important point, there is no such thing as fundamental. You can establish your foundations anywhere you like, and then you complete the circle. Right? And you come back and eat your own tail at some point. Right?
You could start with language department. Say, after all, you have to start with language before you can have mathematics, and then you get mathematics, you get physics, and chemistry, and biology, and psychology, on it goes. So the Ouroboros is, I think, just a true statement about the nature of the way epistemology works.
Now there’s a deeper question, and this is where you think it might be probably shit, which is, is it ontologically possible? Now according to idealists, it’s obviously true because everything starts in the mind of god. And so it starts with an idea. According to Platonists, including our friend Roger Penrose and others, it might not start with god, but it starts with some peculiar domain of the absolute from which everything else is an approximation. So there are many metaphysical concepts out there which aren’t realist materialist. So that’s one thing to point out.
And secondly, there are contemporary ways of reasoning, like simulation theory that you kind of indirectly alluded to in your opening questions, which would make programming, right, in some sense fundamental and everything else derivative. So I don’t think it’s entirely out of left field, this worldview, but I think our primary argument is that epistemologically speaking, you don’t have a right to claim that physics is more fundamental than mathematics or than language. Ontologically, there’s something to be argued about.
Jim: Now here was my reaction when I read this for the first time this morning, which is maybe there’s a measurement that could be made to establish a direction. And that is the degree of differences of the substrate in the emergent stack, right? In an emergent stack near particle physics, the particles are identical to the degree we can tell. Electrons, electrons, electron, photons, photons, photons, they may be in different states, but their rest attributes are the same. So there’s zero approximately differentiation amongst the components, which lead to chemistry.
Chemistry, as we know, especially as the atoms get bigger, are into all kinds of strange complex motions that we can no longer calculate, maybe in principle, but not in practice. And so as soon as you get to any kind of four or five atom chains of chemistry, you’re in a range where all the components, even if they’re the same chemical molecule, species of molecule are different because they’re flipping around and doing all kinds of things. And so, and you go up and up and up, you get the cells. Every cell is different. All your red blood cells, even though they’re the same class, very different at the microdynamics levels. Yes, they are. They have the same attractors. They have the same general architecture. They got mitochondria, they got ribosomes, they got the nucleus and all. So they have mechanisms that they share, but the actual state of them are radically different. And it goes up and up and up and up. And so I was thinking that that might actually be a principled way to anchor the traditional stories, how we say.
David: Well, I don’t disagree with that. I mean, what you just articulated is an emergent story based on broken symmetries. And I think ontologically, I think it’s quite compelling. I don’t disagree with that. But things get a bit messy. Let me give an example of the, you know, the slowly dying pseudoscience of string theory. So here is an area of inquiry which is massively overfitted to reality. Right? I mean, there are so many vacuum states it’s not even worth thinking about.
Jim: 10 to the five hundredth. Right?
David: Right. So our universe falls out of one of them. And, you know, this is the ultimate post hoc mathematization of physical reality. It’s not clear to me, Jim, what it is. Right? In other words, it’s clearly a branch of mathematics which is ingenious. There’s no denying the capabilities of its practitioners. To what reality does it refer? I mean, because of this degeneracy problem, it’s not going to be predictive in the way that we’ve become accustomed to from late, say, Newtonian mechanics.
So there are fields of inquiry which, to my mind at least, blur the boundary, right, between epistemology and ontology. I can’t quite wrap my head around it. Now it turns out, actually, that is a property of complex systems, right, because they’re so observer dependent in interesting ways, like natural selection or whatever. So even if I don’t have a strong commitment to the ouroboros as an ontological object, I still think it’s a bit of a mess. It’s at least throwing a noose around the phenomenon, right, so that we can wrestle with it a bit.
Jim: Yep. That makes sense. Have to think that one through. It was a completely new idea to me. Interesting, to say the least. I’m going to jump ahead to a question I have later with this. You mentioned string theory. Right? The great black hole that sucked in 60% of the physics PhDs in the early double odds. And as far as we can tell, to produce nothing of physical value. But to your point, produced a whole bunch of mathematics of which many of the practitioners claim is beautiful mathematics. To what degree do you think beauty is at all a useful north star in thinking through, particularly the interface between reality and mathematics?
David: Yeah. I think that’s a pure obfuscation. It’s a self-aggrandizing obfuscation. There are good reasons why minimality in representation has worked in the physical world, and you actually just articulated them. Because down there, everything’s the same. Right? And so the model fit can be quite simple. And since most of the interactions are quite simple, they’re captured quite effortlessly in different functional forms. That’s what mathematics does. Right? And so parsimony is not an aesthetic decision. Parsimony is the correct representation of simple interactions. There’s no mystery here.
And what happened, presumably out of some need to communicate the quasi-divine status of one’s own discipline, you start using this language of aesthetics, which I consider very important, by the way, in different ways. Hopefully, we’ll get there. But that that should be the defense of your physical theory seems to me disingenuous, especially given that it hasn’t done any good work in the physical world, which is what it’s funded for. And so and also, I would say, and I’m with Roger Penrose on this, there’s nothing beautiful about a model that has that many dimensions. It’s a sort of that’s a bit of a sleight of hand, isn’t it? And so you go in to buy, you know, a Ferrari, and you’re given, like, a Rube Goldberg horse and cart. And they say, oh, it’s more beautiful. You think, well, could have fooled me.
Jim: Interesting. That makes good sense to me. Unfortunately, I never had a chance to dig into string theory. Too many other things in the world. Would have liked to have spent three months of, just learning the math a little bit to see if there really is something strikingly beautiful about it, but I have my doubts. But Jim, one thing I do want to say is,
David: You know, because I’ve had many string theory friends, especially when I was at the Institute for Advanced Study, and what they will say because they’re much too clever, right, to be so easily dismissed. What they will say is that post facto, it’s the only framework that we have had that provides some resolution to integrating the continuous with the discrete formalisms. And it doesn’t do testable prediction ex ante, but there really isn’t anything else that stitches together the known quite as well. Right? The physically known. And there’s probably something to that. But to your point, this should be, given its performance fringe, not mainstream. And I think what happened, by the way, and this is very important, and it’s something that we don’t talk about enough in the history of ideas, is the importance of charisma, personality, and the cult of genius. Because you get figures like Ed Witten who are kind of weightlifters, and there are other physicists who are impressed by the number of mathematical blocks they can lift at once. And that’s not the same as performing well at solving a problem. And I think that thing actually became a misdirection for the entire field, that the cult of mathematical capability.
Jim: As we all know, science is not just science, but it’s also a sociology.
David: Exactly.
Jim: And why don’t you talk a little bit about the sociology of science?
David: I mean, we’re human beings. Right?
Jim: Some of us.
David: Yes. Yeah. And so it’s kind of inseparable. I think I have a kind of a scaling theory with a sweet spot on the sociology of science because it’s nice to have community. Right? There is such thing as collective intelligence. But at a certain scale, the constraints overwhelm the advantages of the division of labor or however you want to articulate the benefit. And I think much of the sociology of science that, say, Thomas Kuhn talked about in the framing of paradigms, were resistant to change, was about having excessive scale that made it almost impossible for individuals or small groups to make a difference. So I’m a kind of a small science guy because I think that’s the sociology that works.
Jim: What do you think is a good maximum size for a lab?
David: I mean, it sort depends on the problem. Right?
Jim: Yeah. CERN, you’re going to need a shitload of people.
David: Right? You just need it. It’s just no way around it. I think the problem defines it. And but in terms of human interaction, technology, equipment, instrument, measurement aside, it’s probably less than 10.
Jim: Yeah. I often find that five to six is an optimal crew size. Right? That, you know, the information flow could be dense, and yet there’s workarounds when there’s communications problems, and I’ve always found that to be an interesting size.
David: Yeah. Yeah. I tend to agree with that.
Jim: All right. Let’s go back to something you mentioned in passing, which as you know is a strong interest of mine, and that is your view of emergence as being closely associated or perhaps fundamentally associated with broken symmetries. If you could tell us what that means and give us an example or two.
David: So, I mean, this really is the central argument of Phil Anderson’s 1972 “More is Different” paper, right, which is that the symmetries of the laws don’t necessarily correspond to symmetries of configurations. And you gave the example, actually, of molecules that flip between states. And what Phil said, you know, is as those molecules got bigger and bigger, the rate of flipping got slower, exponentially slower and slower. And so to know what the state was, it was more informative to know the initial condition than the law.
I call this the Wigner reversal. Eugene Wigner made the point in a very beautiful paper on laws of nature and invariances. Physics essentially proceeds by if you think about two draws, one is called the law draw, and one is called the initial condition or boundary condition draw. Said what physics does, it takes everything it can out of the parameters that you can’t derive, the initial conditions, boundary conditions, and you put them in the law draw. And your ideal would be there’s nothing left in the free parameter draw. Everything is foundationally derived.
And but what Phil said is, unfortunately, as systems scale, because symmetries are broken, that is the state of the system depends on its initial condition, you start having to move more and more out of the law drawer and more and more into the boundary condition drawer until eventually you end up with natural history, right, where all you have is description. And what Phil said, which was really profound, is under that condition, you have no option but to move up a level because you try and stay at the law level, the foundational law level, as they would describe it. You’re going to be in this weird paradoxical situation where your field, which is so proud of its parsimony and elegance, actually becomes bird spotting. Right?
And so what he suggested was you have to now aggregate the phenomena in such a principled way that they can be law like, but with new laws. Right? The laws of markets, the laws of genetics, the laws of cognitive science, the law and so on, biophysical laws. And that’s a very profound move, and it’s another reason, by the way, why we believe in the Ouroboros, because if in the end you’re trying to make the natural world intelligible by presenting it in sufficiently low number of dimensions, which is the, after all, the ultimate aim of the physical law, string theory aside, then you might be better off doing psychological theory. It might look more like physics.
I mean, that’s the kind of the surprising twist, and that’s the surprising twist of complexity science. Complexity science says, we also, like physics, like these intelligible laws, but we realize that the only way to do it would be to coarse grain in principled ways and recover the law like status of emergent phenomena. And we can talk about emergent. But and I think that’s that’s the essential move.
Jim: Okay. Let’s do that. Let’s take the move to emergence, which is coming back as a topic of study after sort of being ignored for quite a while, and one that I’m particularly interested in, done a little work in, and I’ve been reading many of the papers. I did dig into a little bit of your stuff, and I think discovered were probably different perspectives, but in a similar range, which is reject both weak and strong emergence.
David: Yeah.
Jim: Why don’t you start there and then give your view on what is emergence and how does emergence or how is emergence implicated in these higher level laws?
David: Yeah. So there’s a perfectly reasonable, modest statement of emergence, which is a feature of all theories, not just physics or just complex systems or the humanities or what have you. In other words, it simply says that there are concepts that do useful work, and these concepts describe phenomena at different levels of aggregation. And we call those concepts typically in the domain of natural science effective theories. Right? So for example, you could say the law of supply and demand. It’s kind of right. Right? I mean, you know, I mean, it gets violated. But if you offer me a computer for a dollar, I’m probably going to buy it. If it’s a good one, whereas you offer for a million, I’m not. And so that’s an effective law, and there’s an underlying effective theory that we know, that relates to that.
So emergence tries to understand when concepts are causally justified. That’s how I would put it. Such that an economic theory would be every bit as compelling as a chemical theory. Now the problem is, right, is that these concepts or effective theories or effective variables or effective degrees of freedom, which would be a slightly more technical way to say it, are approximate. Right? And very difficult to find.
So you think about here’s a good example from our colleague’s work, Jeffrey West’s work on scaling. So they have these nice results, right, in biology that suggest that metabolic rate scales as body mass to the three quarters power. They’re not photons and electrons and quarks. Right? This is lots of tissues and metabolic rate. These are very, very aggregate phenomena, and yet they’re very stable. And so it turns out that they have real causal efficacy. And if I know your mass or I know your metabolic rate, there’s all sorts of other things I can predict about you, and Jeffrey would say your longevity, for example, on average. So these are instrumentally useful, effective degrees of freedom, and emergence is about them. It’s about finding them. It’s about explaining why they work. And, also, by the way, it’s about rooting out the fake versions that don’t work.
Jim: Give me an example of a fake version. That sounds interesting.
David: Well, you know, phlogiston or, you know—
Jim: Oh, yeah. Yeah.
David: Yeah. You know, utility, you know, in economics. There are lots of concepts which are entirely questionable that I don’t think are truly emergent causal factors. And I think the failure, by the way, of a lot of fields is the failure of reckoning with spurious emergent phenomena. You could argue it’s very interesting if you look at psychoanalytic theory. So I’m a big fan of Freud and Jung as literary theorists, right, as authors and writers. But I’m not a fan of them as theorists of mind and psychology. And I think that applied to the brain, these things actually have the character of being spurious, effective theories. They’re not dealing with truly emergent phenomena. They’re invented. Doesn’t mean they’re not very literature, which is interesting in itself.
Jim: I always talked about that about Freud. I said, not a scientist. Great novelist. Right?
David: Yeah. Well and there’s nothing wrong with that. It’s just that there was a kind of a miscategorization. Should be on a different bookshelf.
Jim: Yeah. Exactly. Now perhaps where thinking about emergence in this way starts to pay off, is where we start to be able to measure or at least get indicators of stronger versus weaker, downward causality. Yeah. Maybe you could explain for the audience what downward causality is and how that fits in with your thinking.
David: Yeah. So here’s the simple version. So the standard causality fits that linear story of causality that we described earlier in relation to the ouroboros, that you have particles. They get aggregated into molecules, molecules into tissues, and so on. And the idea, right, is that what is fundamentally causal is that which is fundamental, and everything else is an approximate expression of collective modes of behavior. Alright. Downward causality takes it the other way. It says, mind states, for example, expressed in language, can’t be causal of brain states because that’s going the wrong way. Because, surely, the physical interaction, the true kind of Newtonian causality, has to live at the level of the brain. The mind is just this efficient theoretical encoding of brain. And so it would be weird to talk about causality going the other way.
I think it’s a big mistake. And where this comes from, by the way, is this notion of coarse graining. So you start with all the lots of particles. You average and average and average, and you get these other states. But I have this conception of what I call micrograining, and I’ll explain how it works. When Jim, when you program your computer, you’re articulating a concept in a high level language or an assembly or whatever you like to use. Assembler. And that translates through a system of compilations and microcode into states of transistors. So we have built engineered devices that can take these high level, very low dimensional, in some sense, concepts, objects, and do information expansion to the extent of setting the states of transistors.
I think that is what complex systems do all the time because complex systems have evolved to do that well. That, for me, is the legitimate version of downward causality. There’s nothing mysterious about it. I don’t think, by the way, it exists outside of complex systems. I do not think it’s a property of the physical universe, the abiotic universe. It’s a property of agents, and that’s actually the only thing that makes life possible. Right? It’s what’s making this communication that we’re having now over Zoom possible because I’m setting brain states in you as you are in me. And that that’s micrograining. And because the study of emergence grew out of really rigorously the connection between statistical mechanics and thermodynamics, which is all about coarse graining, in the physical world, this other version, which is very natural to the evolved world, has been somewhat neglected. So I sometimes call that the theory of compilation of emergence because we use them all the time.
Jim: I’m going to push back a little bit on the abiotic versus biotic. Just hit me.
David: Okay.
Jim: In my current paper that I’m working on on emergence, I use as a intuition pump a traffic jam on a superhighway that goes up a hill, some trucks slow down and propagates, etcetera. Now I write the thing as if it’s humans driving the cars and the trucks, but I just realized they could be Waymo’s. Right? And the emergence of the traffic jam that comes into being starts to constrain the behavior downward to the individual elements and then gradually dissipates is, you know, a small form of emergence and, doesn’t seem to require, biotics at all. It’s but it does require agency to your point.
David: Right. But no just right. That’s interesting because I think, you know, you could argue that some of the phenomena you’re describing are properties of spin glasses, right, or magnets, where you have the particular state of a spin at a particular lattice point being a function of the average field. But actually, that average field is actually an epistemological construct because it really is the interaction among many, many particles. So I would suggest that in the example you gave, if you stripped out the agentic part, you could express what you’re you’re thinking of the downward constraint as simply a pattern of global interactions that you could describe microscopically.
But I think once they’re functional, once you have a kind of teleology with then I think they’re engineered or evolved, and my argument kicks in because you’ve programmed the reaction of the individual component to the collective. That’s the point because you don’t want to have an accident or a pile up.
Jim: And you probably didn’t write a specific routine for that. Well, I know you didn’t write a specific routine for that traffic jam. You have some general parameters that operate together and come up with good decisions, basically, and should probably even do a better job than humans, if not today than in a few years. Anyway, want to think about that one a little bit more.
David: Maybe we should talk a little bit about our mutual interest in puzzles and games. I since that’s what I’ve been working on for the last two years, in a which is relates to everything we’ve talked about because it’s the world of rule system, which we’ve talked about for 20 years, generative rule systems, and progress in that area, and where that comes from, where where at least my interest, foot is worth, comes from.
Jim: That sounds good. As you know, I’ve been a gamer since I was 10, back when they were little cardboard pieces on Avalon Hill games. And one of my final projects was writing a pretty successful mobile game called Network Wars. People go spend your 99¢ on the App Store. It’s still quite popular, surprisingly. So what is it about games? You know, this is the thing about games that humans create is the level of combinatorics that can emerge. Right? And some of them, checkers lower, chess higher, go higher. But some of these multi-turn games where you have many different things you can move simultaneously, I’ve calculated they’re vastly bigger than Go. So, anyway, with that as an opener, what do you think about games, and why are they an insightful probe into this reality we’ve thrown ourselves into?
David: So I think two ways in. One is because we’re very interested in ideas like local rule, global pattern. We’re interested in simple iterative procedures like Turing machines, natural selection, reinforcement learning, et cetera. We’re really interested in configuration spaces, right, and how you search them. So all of that, which is so central to all of our interests, whether it’s in economics, evolution, technology, physics, engineered physics, present in these combinatorial games, which are like model systems for analyzing them. So that’s my interest. They’re like the hydrogen atom or the drosophilids of rule systems with very high configuration spaces that should be NP hard to search, and yet we use heuristics to find solutions in finite time. Right? So they have all that stuff we care about.
And here’s another side, and this is more personal. So when I started playing Go badly, you learn these concepts. Right? And part of the fun of learning Go is like pretending that you’re speaking Japanese. Right? So there are these concepts like sente or goate or seki, right, or atari or thickness. Okay. And it slowly dawned on me, Jim, that these are just synonyms for concepts that I was using in my science, that AG is just cryptic variation. Goate is just neutral variation. Right? Anyway, on it goes.
And then you take the next move, and then you realize, oh, maybe the true grand unified theory, which—and I think our project is much more grand unified theory than physics, by the way, because we’re not just doing physics. We’re doing all these fields—is realizing that all these kinds of combinatorial solution spaces that have to be searched to discover functions share common properties. And my interest in those games has been as a window into the grand unified theory of complexity, quite frankly. So that’s where it comes from.
Jim: Interesting. And when I started getting involved a little bit in the AGI community in 2014 and combined that with my interest in games, I had a bit of an epiphany, which I’ve done no more than give to people, which is the word you used, heuristics. You know, the world is way too high dimensional to actually calculate as we know, and LLMs are one way to be a dimensional reducer by brute force. But heuristics are an even more powerful one. And if somebody could do the work on heuristic induction from reality, that might be the golden road to AGI.
David: Yeah. No. I think that’s really important. In our work recently, we’ve been working mainly on physical artifacts on Rubik’s cube, pocket cube, Soma cube, and all these kinds of things, the abacus, and introducing artifact-specific properties that we call the principles of materiality, which reduce search spaces in principled ways. And so the heuristic in that case, by the way, is induced by the physical artifact.
Jim: Yeah. So the physical artifact is the heuristic. Right?
David: It is the heuristic. Exactly. So we can express precisely what is meant by the heuristic in that particular case. So that’s one side to this. The other side that we haven’t dealt with, and I do want to mention briefly, when I was a grad student at Oxford, the two people whose offices I would visit and whose talks I would go to that were outside of my own domain of interest, outside of evolution and all that stuff—well, on the one hand, PMS Hacker, the Wittgenstein scholar, who was a kind of adversary in his style of thinking. And the other one was Isaiah Berlin. And my entire worldview in terms of political economy and the history of ideas comes from Berlin.
Jim: I was going to say he seems to be the wisest political thinker of them all.
David: I think so. Despite being this quite heterodox, nondogmatic thinker on intellectual history. But in particular, I want to mention two ideas which connect to combinatorics. One is the idea of which we know from his book on two concepts on liberty, positive and negative freedoms. So I am a very, very—who is David Krakauer?—a strong believer in negative liberties. I should be free to do what I like without you getting in my way. And as opposed to positive liberties, things I impose on you. Positive liberty was, I choose to impose my religious orthodoxy on the world. Negative liberty is a set of freedom of speech, freedom of thought, freedom of love, freedom of whatever you want. So I’m very interested in negative liberties.
Notice negative liberties are systems of constraints, and constraints are generative. That’s what they do, like in poetry or in mathematics or in computer code. So the second connection, which I’ve been very interested in, is Isaiah Berlin’s interest in value pluralism. You know, he believed very strongly that like paradigms, cultural values were nontranslatable. They were irreducible. And there isn’t one super enlightened political system that should rule the world, which is something I also agree with. And value pluralism is another way of talking about historicity. It’s another way of talking about contingency and branching paths, high dimensional configuration spaces where the state you end up in depends on your history in a really important way.
So when we talk about grand unified theory, I actually think that ideas that I’ve always shared personally from Isaiah Berlin in terms of my disposition towards the world, in terms of political economy, are actually related to the game of go and game of chess. And that there are underlying principles that could span the two.
Jim: Very cool. One of the things about these worldview podcasts, I’ve decided to make them exactly sixty minutes or very close rather than my open ended ninety to two hundred minutes, which my other ones would run. So I got a couple of quick hit questions for you. You know my obsession with the Fermi paradox. The current crop of UAPs, any of them actual aliens?
David: No.
Jim: Okay. Quite great. Are there actually technologically capable aliens in our galaxy?
David: No. You’ve changed since last time I talked to you about that. I believe that there is life, non-terrestrial life in the universe. I think your statement—maybe I should be agnostic, but I think the Fermi paradox would suggest to me some evidence. So I’m going to punt on it maybe.
Jim: Okay. Yeah. It’s funny. I started out as typical 14-year-old nerds saying, of course, there have to be tens of thousands of them. But I’m now purely agnostic. Like, how the hell did we go for protests to eukaryotes? Is that a once in the history of the universe weird ass thing? We just don’t know. But we will know in a few thousand years as we start looking at exoplanet atmospheres and looking at Enceladus and, you know, who knows what all.
Let’s see. Just so you can piss some people off. God’s spirits and all that, any of it real?
David: No.
Jim: Yet people seem to love this stuff. Twenty-three hundred years ago, let’s call it, the Enlightenment gave us permission to start to reject unjustified authority. And yet three hundred years later, we still have a world full of people who seem to love authority.
David: Well, let me tell you one thing on Isaiah Berlin, which is a third dimension, which I think bears on this, which I’m interested in, which is this. If you remember, towards the end of his life, he was writing all these essays on the thinkers of the counter-Enlightenment, Vico, Haman, Herder, Demaestra, and so forth. And here’s something I kind of learned, which I believe, which is that the rational is the appropriate approach to acquiring knowledge. The irrational is the right way to live your life. And there are domains where excessive rationality is deleterious. I love movies. I do not expect them to make sense. You know? And, now that’s not the same thing as positing the existence of a god. But I would like to defend the importance of the irrational.
Jim: And also, the one we talked about before is the artistic. You know, if religion were thought of as literature, I actually like the Bible, the Old Testament, at least.
David: Absolutely.
Jim: As lit, it’s great literature, but, you know, just you don’t have to swallow the metaphysics.
David: No. And that’s exactly right. So and that doesn’t—since great literature means it’s an important source of meaning and understanding. Absolutely. It’s just you don’t need exactly. You don’t need the metaphysics. Exactly.
Jim: I want to thank David Krakauer for what I am going to declare to be a successful prototype of the WorldView podcast.
David: Well, it was wonderful to see you. This is great fun.
Jim: Actually. It really was.