295 | Solo: Emergence and Layers of Reality

0:00:00.2 Sean Carroll: Hello, everyone. Welcome to the Mindscape Podcast. I'm your host, Sean Carroll. Longtime Mindscape listeners know that the idea of "emergence" emerges over and over again. And in various contexts, for example, you might hear someone claim that consciousness is not a new fundamental category but rather an emergent category from underlying physics. This is a useful idea of emergence in the sense that we're talking about it here. It's talked about a lot, both in science context and in philosophy contexts, but it's also a little frustrating. It's frustratingly vague about what exactly is meant. Part of this is because different people use it to mean different things. Part of it is because different people want it to mean different things. And part of it is because people are not clear about what it is that they actually are meaning. In particular, the thing that gets me is I just really would like clarity when some people talk about higher level emergent things about whether or not they are imagining that the existence of these higher level emergent things violates or demands a modification of the lower level laws of physics.

0:01:16.0 SC: Maybe it does, maybe it doesn't. But I do think that as a respectable thinker about these things, we should be clear about what exactly is going on. And so, I'm not interested so much in policing the definition. I'm not here to say what the right definition of emergence is, but I do think that it's important to clarify all the possible ways in which we might profitably talk about something being emergent without first judging what exactly is going on. And indeed, this is a longstanding ambition of mine. I finally got a chance to write a paper about it with Achyuth Parola, who's a student here at Johns Hopkins. We wrote a paper called "What Emergence Can Possibly Mean." This is actually the second title of the paper. The original title when we were working on it was called "Emergence Without Judgments." That was supposed to [chuckle] represent our frustration at people who tried to define emergence but used words that were themselves ill-defined, like "novelty" or "surprise" or things like that.

0:02:20.8 SC: So, one way or the other, what we're trying to do is just to bring super-duper clarity to this. And today's podcast episode is about exactly talking about this, going over what we did in the paper and more generally commenting on how emergence goes. So, you may or may not agree on some of my judgment calls along the way, but hopefully the classification system for different kinds of emergence will be useful to anybody no matter what kind of emergence you actually think exists. Now, I should of course mention that for contemporary listeners, those that are listening to the podcast soon after it was released, we are in the aftermath less than the week away of a presidential election in the United States, which was held last week, where Donald Trump beat Kamala Harris to become the president-elect. This is a big deal, and people have asked and I did contemplate, whether or not it would be appropriate to devote a solo episode to that.

0:03:17.1 SC: The emergence episode was in my brain for a while, so I was planning on that, and I knew that Trump absolutely had a chance to win. That's what it means to be a coin flip in an election. But I decided not to; I decided not to devote this episode to talking about that. I mean, I do think it's a huge deal. I think that Trump winning is a blow to anyone who is against racism and corruption, and in favor of democracy and the international order and economic stability and things like that. Just as one little tiny thing among many, we now have an incoming vice president who has proclaimed in a speech that universities are the enemy. So, you might imagine that I am not in favor of the incoming administration. But I decided not to devote this podcast to it for a couple reasons.

0:04:07.9 SC: One is that it is still fairly recent, and I have talked about related things, and I'm not sure I have anything really new and interesting to say right now. I know that there's always a rush to judgment to say, "Well, this is the reason why that happened. If only they had listened to me," [laughter] or to be sort of doomerist about it and saying, "Oh no, let's just tear our hair out about how bad it's going to be," or to be prescriptivist like, "Here's what we have to do right now." I'm very sympathetic to all of these impulses. I have them myself, but I don't need to foist them on you. Okay? I do think it's okay to think about these things, to really decide what it means, what we're supposed to do going forward. I don't think that the second Trump administration is going to be like the first one. I think it'll be worse in any number of ways. But we'll have to actually see and respond accordingly.

0:05:04.8 SC: The other reason is because I absolutely believe that we have to fight against encroachments on rights and democracy and so forth. We can't just resign ourselves to the worst things that can possibly happen. But part of that fight is keeping going, preserving those aspects of our lives that make life worth living. As always on these podcasts, I will essentially mention the importance of considering our lives as more than merely survival. And of course, depending on one situation in the world, that can be easier or harder to do, right? Easier said than done for some people. But I do think that we have to remember that there's more to life than politics. I think politics is important. I think that people who think that politics is distracting or annoying are part of the problem, but it's not everything. And that politics is existing in some part in order to carve out space for the other things that are actually super-duper important. And for me, one thing that is bringing meaningfulness and purpose to life is the ongoing quest to better understand our universe.

0:06:16.0 SC: So, I think we need a balance of political struggle and discussion, and work and concentration, and for that matter play and enjoyment concerning the rest of existence. So, there's a role for an entirely abstract, not very directly useful kind of discussion like this one, because we can do both. We are not tied to only fighting the political struggle or being ivory tower academics. We are grownups who can do all of those things at once. And this is part of that. This is part of the reminder, the urge to keep our lives moving forward, even as we're temporarily depressed about the political state of the world. And who knows? Maybe it'll help us sleep a little bit better, give us some ideas. I'd love to hear feedback as always on the solo podcasts. It's me talking out loud in some sense, in a way that, about something I think I know something about. But there's certainly more questions to be answered. So, let's go.

[music]

0:07:36.0 SC: These days, I am busily working on volume three of The Biggest Ideas in the Universe series, and its subtitle is going to be called Complexity and Emergence. So, as you might imagine, I've been thinking about these things. Complexity and Emergence, to some extent is a grab bag title. I'm gonna be doing lots of things like thermodynamics and cosmology, which are related to those subjects, but then also digging into what it means to say something is emergent, something is complex. The theme of those books is always to be uncontroversial, non-speculative, to actually do the things that people will agree on and continue to agree on for hundreds of years from now. So, I won't be going into the specific kind of categorization I talk about today. But I did open the book; I do open the book, will open the book, with a thought experiment that is worth keeping in mind.

0:08:25.0 SC: You can do it as a real experiment if you want or as just a thought experiment. Very simple. Take a piece of paper, crumple it up into a ball, a piece of paper you don't really care about what's written on it. Hold it in your hand and then toss it into the air a short distance so that you can catch it. Okay? Imagine having done that or actually doing it. It doesn't really matter. Hopefully you know what it would be like to do that. I want to claim that you have just demonstrated one of the most remarkable and important features of the natural world as we experience it. In fact, you've demonstrated something that relies on several very important features of the natural world, just so we keep them in mind. One of them is gravity is a weak force, right? The entire gravitational field of all of the mass of the Earth is pulling down on that little ball, and your arm is able to toss it in the air even though you are much tinier than the Earth.

0:09:23.2 SC: That's because your arm's forces come from electrochemical reactions, which are enormously stronger than gravity on a particle-by-particle basis. That might very well be a non-accidental important feature of the world. There's something called the weak gravity conjecture, that suggests that gravity has to be the weakest force, and it has to do with entropy and quantum gravity. And that's very interesting. We're not gonna talk about it today. Another interesting feature that you have implicitly demonstrated is that the world is predictable. That there are laws of physics, right? When you throw that ball into the air, you are able to use your brain to figure out, to predict where it's gonna come down. It does not just randomly go up into the air, right? Does not randomly go right or left. There is a pattern; there's a predictability to how the world works. Obviously a crucially important feature of reality.

0:10:19.0 SC: And for that matter, the world is intelligible, right? It's not just that there is a pattern to how the ball goes up and down, but that you can figure out what that pattern is and use it to make the catch. Good for you. All of these are worth book-length treatments and have received them all by themselves. But what we're gonna talk about is something that is so intrinsic in how we think about the world that we take it for granted. Namely, these days, when you think about what that piece of paper is, it's made of molecules which are made of atoms, which are made of elementary particles, which are described by the rules of quantum field theory.

0:10:57.2 SC: Many times in the podcast we've mentioned the idea of Laplace's demon. Now, you can be specific about whether or not you're in the classical approximation, or you really wanna do the whole quantum field theory if you want to. But one way or the other, the idea of Laplace's demon is that if you knew exactly the state of that piece of paper plus the environment that it was bumping into, then the laws of physics would tell you, would allow you to predict how it would behave, okay? You would have the position and momentum of all the atoms that were in the piece of paper, and you could tell how it would go up and would go down. You're not Laplace's demon. You're never going to be Laplace's demon. We all know that. You don't know all those positions and velocities. You have vastly incomplete information about the specific microscopic state of the atoms in that piece of paper. And nevertheless, you are perfectly capable of saying what it's gonna do when you throw it up into the air. That's kind of amazing, okay?

0:11:58.7 SC: It's not just that. Well, it is, but it's not just that you can throw away some of the information contained in the microscopic description of the system, but that we know exactly what information we can throw away. All you basically need to know is something about the center of mass of that little ball of paper that you have crumpled up, and maybe something about the environment that it's in, if it's windy or something like that. But relatively few pieces of information give you a very good handle on what's going to happen in the macroscopic world. That is emergence at work. It is one example of emergence. Some people would not even define that as emergence. Like I said, it's a contentious definition that we can battle over, but to me it's exactly what I'll be talking about here, or one of the examples I'll be talking about.

0:12:49.3 SC: The idea is that there are multiple levels of description of the world. This is a theme in my earlier book, The Big Picture, where I talked about poetic naturalism. There's only one world, but there are many ways to talk about it. So, in this case, there is a micro level or a lower level as we usually talk about it. That's the level where we can describe the piece of paper as a collection of atoms or elementary particles or whatever you want to do. And then there is a higher level, a macro level, where you have pieces of paper and you have people and they have hands, and they can throw the pieces of paper up in the air and catch them, okay? And the crucially, amazingly wonderful non-trivial fact about the world is that you don't need to know about the lower level to navigate the higher level.

0:13:38.2 SC: You don't need to know anything about the atoms of which the paper was made. In fact, people could do this exercise of taking a piece of paper, crumpling it up, throwing it up in the air and catching it, long before they knew about atoms and molecules, etcetera. This higher or emergent level where you just had the macroscopic things like the paper and you and so forth, is really descriptive all by itself. It captures something real, something that Daniel Dennett, former Mindscape guest called "real patterns" in the underlying dynamics of the system. To me, that's the basis of emergence, the idea that you have something that has many, many things going on, but you don't need to keep track of all the things going on in order to make useful predictions. There are certain kinds of predictions you can make based only on macroscopic data, coarse-grained data, limited data, whatever you want to call it, compressed data about the system that you're thinking about.

0:14:41.0 SC: And we'll get into exactly what that means. Let me first though, point out, there's a question here. Is the existence of such higher level patterns robust? Is it a generic feature? Is it true that for almost anything the underlying laws of physics could have been, that you would get some emergent higher level descriptions that it would be possible to make predictions with vastly incomplete data? Or is that very, very special? Is there some feature or set of features about the specific rules that we have in nature around us, that helps us by allowing us to discuss it at this higher emergent level? I do not know the answer to that one. It would be really nice to have some theorems or some demonstrations about when emergence happens and when it doesn't; not just, if this happens, then emergence is possible. But in the space of all possible underlying theories, how many of them, what proportion, what fraction, how likely is it that you will have any kind of higher level emergent theory?

0:15:45.1 SC: I truly don't know the answer to that one at all. I can speculate, but that's up to future researchers to figure that out. Okay. So, that's the general idea. The general idea of Emergence is you have a lower level, you have a higher level that related in some way. And part of the relationship is that the higher level uses less data. It does not require that you know everything that there is to know about the lower level, and you can talk about each level for its own sake in its own terms, using its own vocabulary. So, there's a couple of problems associated with this very natural setup. One is that when people actually do this, they are very vague. People are vague. It's very frustrating to me. As I said in the intro, I don't mind disagreeing about the substance of how you think the world works. That's fine.

0:16:37.8 SC: But this is a strange part of philosophy and science where people are almost intentionally vague about what exactly it is they mean, or at least maybe they're not intentionally vague, but they're satisfied with using terminology that is just not that clear what is meant when they say it. And I don't see any reason for that. I would like to be clear in what I'm trying to say, whether or not people agree with it. So, very often in descriptions of emergence, one of the aspects that you'll find is that people say that the higher level properties or higher level behaviors or dynamics are novel or new or surprising or cannot be derived from the microscopic level. What does any of that mean? Or at least if it does mean something, if we think it means something, what precisely does it mean? Is it subjective when you say something is surprising or novel about a higher level theory that is not there, not obvious from the point of view of the lower level theory?

0:17:39.0 SC: Well, what if it's not surprising to me? [laughter] Could it be surprising to one person and not surprising to somebody else? Likewise, even if you say the behavior cannot be derived even though it is implicitly there from the underlying theory, what do you mean it can't be derived? Like, I haven't derived it yet? There are some things which I haven't derived yet, but maybe in the future I will derive. Does emergence sort of time-dependent? Once you are able to derive it, it's no longer emergent anymore? So, this was a big motivating factor in writing the paper that a Achyuth Parola and I wrote, the idea of removing all subjective judgment-like words. I do think that these ideas are related to real objective features. It's just a matter of spelling out explicitly what those real objective features are. One paper that is a classic in the discussion of emergence at least among physicists is Phil Anderson, the condensed matter physicist who passed away a few years ago, wrote a paper in the '70s called "More is Different."

0:18:43.1 SC: And his point in the More is Different paper is, as I said before, if he's a condensed matter physicist and he's trying to study superconductivity or something like that, as he often did, he doesn't need to know all the microscopic specific details about things. He doesn't need to know about the Higgs boson or the top quark to study superconductivity. You don't even necessarily need to know about... We certainly don't need to know about quarks and gluons. Maybe you don't even know about atoms at a certain level of abstraction, okay? But in that paper... So, people take that paper, the title of which is "More is Different." And they, I think... [chuckle] I don't like to be unfair. But I think they read the title, and I don't think they read the paper. Because from the title, you get the impression, well, once you have many things and they're coming together and they're interacting in a certain way, different things happen, right? Things that you would not maybe have predicted from the underlying stuff.

0:19:43.0 SC: But Anderson is very clear. He says multiple times, and I'm sort of sympathetic to his way of writing it because I've done this before myself where I know I'm saying something that is gonna be misunderstood so I say it as clear as possible multiple times. Anderson says very clearly, reductionism is true. That is to say, when he's studying his superconductors or whatever, he has no doubt that there is a way of describing what's going on that is at a lower level, that is in terms of atoms and forces and quantum field theory and whatever. And what happens at the higher level is 100% compatible with that and indeed in principle 100% predicted by it.

0:20:27.3 SC: He's not denying that in any possible way. What he's denying is not reductionism but what he calls "constructivism." He's denying that the right way to study the behaviors of superconductors or higher level things is to start thinking at the lower level. And that's I think pretty undeniable. Certainly if you go much further than superconductors up to biology or economics, there's zero insight gained by studying the lower level description in terms of particle physics, right? We have the core theory. We have a very successful micro-understanding of the world in the everyday life regime, the standard model of particle physics plus gravity, okay?

0:21:11.1 SC: That's not very helpful when you wanna study biology. But that's zero reason to think, following Anderson's perspective, that biology is somehow not in principle entailed by the standard model of particle physics plus general relativity. Anyway, all of this is to say that people need to be clear about what their words mean. When I had my debate with Philip Goff over panpsychism, etcetera, a few... Oh, maybe I guess it's over a year ago now. I said that the one thing... I gave advice to people. I said if you're talking to a panpsychist, the one question you have to ask them is, okay, you're imagining that consciousness is everywhere or everything is consciousness. You're promoting some view of consciousness that it is something more than just a convenient way of talking about the collective behavior of underlying particles and fields. Fine. Does your understanding of panpsychist consciousness imply that we need to modify the underlying laws of physics as we know them now, or is it completely compatible with those?

0:22:22.5 SC: Either one is worth discussing. I'm happy to discuss the possibility we don't need to do it. I'm happy to discuss the possibility we do need to do it. But tell me whether we need to do it or not. And Philip's answer was, "No, I don't have to tell you that. I'm not really there. Maybe it is. Maybe it isn't." And in some sense, that's fine. It's fine to not have your theory developed well enough that you can answer every possible question. But then don't claim that your theory is very good. My point is that this is a very, very fundamental basic feature of your theory of consciousness. Is it compatible with the known microscopic laws of physics or not? That's super-duper important. That's not like, "Oh, we'll figure that out someday. But meanwhile, we have a very successful theory." If you haven't figured that out, you do not have a successful theory. And this is part of the quasi-intentional vagueness that rubs me the wrong way.

0:23:15.2 SC: The other thing that gets in the way of talking about emergence is that even when they're not being vague, people do use the word differently depending on where they come from, what they were exposed to in the early point of their education and so forth. And this is perfectly fair because the meaning of the word has grown and changed over time. And that's just a reason to try to be explicit about what you mean, not just use the word "emergence" and then assume everyone is going along with you, but to just say what kind of emergence you're talking about. If you look up or you talk to philosophers or you look up in philosophy references, the word "emergence" came into popularity from a group of people who are known as the British emergentists. I believe that John Stuart Mill was actually in some sense the founder of this perspective.

0:24:03.7 SC: And it's a way of thinking about consciousness, especially. It could be spread more widely to other kinds of ideas. But they cared about consciousness. People like C. D. Broad and Samuel Alexander and so forth talked about the emergence of consciousness, and they were emphasizing in their view that there really was something truly new at the emergent level. So, if I understand correctly, the perspective of the British emergentists, they were on the side of things that said the kind of emergence they care about is something where you could not predict the higher level behavior just on the basis of the underlying theory. Of course, they had no idea that we would have such an effective underlying theory like we have now. But still, the perspective is perfectly okay.

0:24:52.9 SC: And these days, we use emergence sometimes to mean that and sometimes to mean something much more down to earth, just like what Phil Anderson meant, okay? And both ways are fine but they are very, very different especially in their philosophical implications. So, once again, I would advocate being clear about this. And finally, I guess I should say just to be super clear because not everyone is a physicist or a philosopher. The word "emergence" absolutely has a connotation of something happening or playing out over time, right? The emergence of a chick from its egg or the emergence of worms from the ground after it rains or something like that. Zero about what I'm talking about in this solo podcast is about that kind of emergence. Our kind of emergence is a relationship between two theories, a micro theory and a macro theory.

0:25:44.7 SC: And that relationship is supposed to be true at all times, okay? It's not when we say "consciousness emerges," that that particular phrase does not refer to the fact that things come to be conscious over time. That might also be true. But that's not what we mean by "consciousness emerging" in this discussion that we're having right now. Okay. Given that there are all these different definitions, different ways of talking, and I'm trying to get people to be clear, let me be clear about the fact that there is a very famous and somewhat useful distinction between strong emergence and weak emergence. You've heard of this distinction probably if you've heard of emergence discussions at all. It's the first thing people talk about. And again, I think the British emergentists came first, and they would be in the realm of what we call strong emergence and only later did the idea of weak emergence become so obviously useful that they started calling it emergence also.

0:26:43.0 SC: There's a famous paper by philosopher Mark Bedau that is simply entitled Weak Emergence. I don't know if that's the first time the idea of weak emergence was explicitly spelled out in the philosophy community, but it's certainly been a very influential paper since then. And Bedau's point is that you might have a situation where there's a microscopic theory and a macroscopic theory. And it might be true that the higher level properties, the way that we talk about the macroscopic theory, is not obvious from just thinking about the microscopic theory. But in principle, in his version of weak emergence, the idea is that it does arise from the lower level dynamics.

0:27:24.0 SC: And the nice thing about Bedau's paper is that he operationalizes this idea. So, when you say, okay, the higher level dynamics arises from the lower level dynamics. What exactly do you mean by that? That's the kind of tedious thing we're gonna be talking about during this whole podcast. What exactly do you mean by that? And he answered it. He said, "What I mean by 'arises from' is that in principle, I could take the description of the system I'm talking about, I could cast it in the language of the lower level of the microscopic theory, and then I could put it on a computer." That's the step that he did that was really very useful. He said, rather than using philosophy words like "governing" or "supervening" or "determining" or whatever, which are fine but then you have to negotiate what you mean by them, he was very down to earth. He said, "Look, if I really believe I have a theory of the atoms in, let's say, a crumpled up piece of paper I'm throwing up in the air, I could put the state of the crumpled up piece of paper and the dynamical laws of the lower level theory on a computer, run the simulation, and the simulation would make a prediction, and it would be correct." That's what he means by weak emergence.

0:28:37.8 SC: So, we may or may not be able to derive the higher level description from the lower level ones. But in principle, it's there, as is made clear by the idea of the possibility of putting on a computer and asking. As opposed to this, we have strong emergence. Strong emergence is where the higher level properties are not only not obvious, but they are fundamentally new. They are even in principle not predictable from the lower level. There's another famous paper by David Chalmers, former Mindscape guest, where he talks about both strong and weak emergence. And he's very clear in the sense that he says that the higher level properties in strong emergence are not deducible even in principle from the lower level properties.

0:29:26.2 SC: And this "in principle" thing is very important because a lot of people one of the ways, one of the favorite ways to be vague about emergence is to blur the distinction between what is in principle impossible [chuckle] and what is merely kind of difficult. Okay? So, he's saying that in strong emergence, even in principle, even by putting on a computer, you could not figure out what was happening in the macro theory. So, that's fine. That's a relatively clear conception. But there's an obvious problem with it. Not a problem with it, but a puzzle that arises when you take that conception seriously. And this puzzle is sometimes but not very often confronted, which is the following:

0:30:08.0 SC: If you say I have a lower level theory, I have a theory of atoms and how they bump into each other and how they combine chemically and whatever, and I also have a higher level theory of like a conscious person with a brain and the brain is made of atoms, okay? Then in principle, I think I could imagine taking the state of the brain, the person or whatever it is, and putting it on a computer and making a prediction based on the lower level theory. So, are you saying that the lower level theory is wrong? I think that people don't want to say that. It's very, very rare that strong emergentists will actually say, "Yes, I'm saying the lower level theory is just wrong in the domain where you have a person with a brain." I think what they would like to say is that the lower level theory is just incomplete, right? That somehow there's wiggle room in the lower level theory.

0:31:04.1 SC: So, when you apply it to a person who has a brain, there were gaps. There were lacunae in the lower level description that are filled in by some higher level strongly emergent properties. I think that's just wrong. I think that in the case of a good lower level theory, if the higher level theory is supposed to literally describe a system that is made of lower level things, the lower level theory says something, and you're either gonna agree with that or disagree with it. And therefore, if you disagree with it, if you think that there is strong emergence going on, then the lower level theory was never the right lower level theory in the first place. So, the problem in this sense with strong emergence is just that it raises the possibility that it's not a deep feature of the world. It's just that you made a mistake when you invented that lower level theory, right? Maybe, maybe not. That's what we're gonna talk about.

0:32:00.1 SC: So, our goals are first to distinguish between these various varieties of weak emergence. So, even aside from the "weak versus strong" dichotomy, there's a lot going on just in the world of weak emergence. And I think that one of the stumbling blocks to clear communication here is that people haven't really been clear about the different subdivisions of weak emergence. And I wanna do that in a way that doesn't use these sort of subjective judgmental words like "novelty" or "derivability" or whatever. And then the other goal is to explain how strong emergence could be a real thing without just reducing it in some simple sense to the microscopic theory or your idea of what the microscopic theory is was just wrong, right? Is there any way you could have some version of strong emergence and still think that in its own domain, the microscopic theory is perfectly right? And I wanna offer that, yes, there's a potential answer yes to that question.

0:33:00.7 SC: Now because I've gotten feedback, polite feedback from some people who've looked at the paper, I wanna be very clear that that's all we're trying to do. We're trying to figure out the different varieties of emergence, okay? That's why the title of the paper is What Emergence Can Possibly Mean. There is another maybe arguably more important thing to do which is to ask, when does emergence happen and how can you discover it? That is to say, given some microscopic theory, can you figure out whether there are higher level emergent ways of talking about it. And if so, can you literally construct what those higher level ways are? People do work on this. Former Mindscape guests like Anil Seth or David Krakauer have written interesting papers about exactly this problem. Super-duper important, super-duper interesting, not what we're talking about. So, you're welcome to be interested in that problem, but it's just not the fish that we're trying to fry here today. Okay.

0:34:02.7 SC: So, how are we gonna do this? How are we gonna figure out the different ways that there could be some kind of emergence and that we can define what we mean very clearly without using fuzzy subjective language? The short answer is we're gonna think like physicists. That is to say, we're going to imagine that there are systems that we're interested in, and those systems have two elements. They have states that you can be in. In classical mechanics, you have constituents with positions and momenta. In quantum mechanics, you have wave functions. In biology, maybe you have cells or organisms or whatever. There's some space of states.

0:34:38.2 SC: And then you have some dynamical rules for how those states evolve with time. And the idea is that if you give me the state of the system, maybe the microstate, maybe some macrostate, whatever, you can use the rules to make predictions about what will happen next. And the rules need not be either super precise or even deterministic. Maybe the rules are there's a volcano and it will erupt a certain fraction of the time or something like that, right? There's a probability per unit year that it will erupt. That's fine. That's a theory. We're not demanding that your theories be perfect. We're just saying that they exist. We're making one assumption that is non-trivial, which is that the theories are what is called Markovian. Well, this is sort of a slight abuse of terminology because Markovian is usually invoked as a concept when you have non-deterministic theories. We're gonna include deterministic theories as well.

0:35:33.9 SC: And the idea of Markovian is just that the theory makes a prediction for what will happen next based on the state at a current time. It doesn't need to know the past history of the system and what it was doing to make a prediction about what comes next. If you have a baseball flying through the air, you can predict where it will go from knowing how it left the bat, its position and its velocity. But if you look halfway through, and you tell me its position and velocity halfway through its trajectory, you can predict its future without knowing its past. Okay? That's because the dynamics are Markovian. I don't think this is a real restriction because I think that there's sort of a matter of convenience here. I could always define a space of states to include enough information about the past right now to make that Markovian version of the theory perfectly adequate. Okay? So, I don't think this is a restriction on what we're looking at, but I just wanted to flag that as something that we do actually assume. It makes the later discussion much, much simpler.

0:36:36.0 SC: And then with this setup, we have spaces of states and we have an evolution law over time. So, we're not worrying about complicated questions in quantum gravity, about the emergence of time, or anything like that. 'Cause time is fundamental for us, okay? Then the basic idea of emergence is pretty straightforward. First, there is a map, the emergence map from the micro theory to the macro theory, which will be throwing away information. The map will be coarse graining. It will be many to one. There are many different microstates that map onto the same macrostate, okay?

0:37:12.9 SC: That's not the only thing you can have if you have two different theories. There are certainly examples in physics where you have two different theories that are secretly equivalent to each other, right? So, one state in one version of the theory maps onto one state in the other one. That's fine. You can have that. I'm just saying that that's not under our definition of emergence. You throw away information in the emergence map. So, you map from a collection of many, many atoms in a piece of paper to just the crumpled-up ball of paper. You throw away an enormous amount of information when you make that map. There are many different individual microscopic states that could amount to that kind of ball of wadded up paper, okay?

0:37:56.4 SC: And then secondly, you retain the ability to predict even though you've thrown away all that information. That's the other aspect of emergence, that the macroscopic theory by itself has sensible dynamics. I don't need to know about the atoms to throw the ball in the air and to catch it, okay? So, those are the elements you have. You have a microscopic theory with its space of states and its evolution law. You have a macroscopic theory with its space of states and its evolution law. And technically, what we're saying is that there are two maps. One is the map from microstates to macrostates. The other is the map of the state at one time to the state at a future time. And the requirement that all of this setup describe emergence be that the diagram commutes.

0:38:48.9 SC: That is math speak for saying that I can start in the micro theory at some time, I can evolve it forward in time using the laws of physics, and then I can do the emergence map to a macrostate. Or, I can start with the microscopic state, immediately map it to a macroscopic state, and then evolve it forward in time with its laws of physics, and I will get to the same place. I will get to the same future state of the macroscopic system. That's what it means for the diagram to commute. We can walk through the diagram. The diagram has four nodes in it. Microscopic state at time 1, microscopic state at time 1, microscopic state at time 2, macroscopic state at time 2. And it has arrows, upward arrows for the time evolution, arrows going left to right for the coarse-graining emergence map, and the diagram commutes. You can chase through those arrows either way.

0:39:42.0 SC: So, that's what it means for us roughly to get emergence one way or another. Now, something we're not going to pursue but is very interesting is, and I'm not even sure this is a thing. I'm not sure this is important or it's just obvious. But in many examples of emergence, not only can you throw away an enormous amount of information in this coarse-graining emergence map, but the info you do throw away tends to be... Sorry. The info that you keep, I should say, the opposite of the info you throw away, the information that is still there in the macroscopic description tends to be observable from the macroscopic point of view, right? So, if I have the little ball of wadded up piece of paper, not only is what I need to know about it, roughly its position and roughly its macroscopic momentum. But I can see those things by looking at them, right? That's conceptually not exactly the same thing. It is as a practical matter, super important to the fact that higher level emergent theories are useful. But I just don't really know if it's sort of obviously always going to be the case or if we just get lucky.

0:40:52.3 SC: It's certainly the case that there are things that are not observable that might be useful to me in predicting the future, right? If I have a theory of volcanoes erupting, it might be useful to me to be able to observe microscopic features of what's going on within the volcano that give me much more precise ability to predict rather than just a certain rate of eruption. But I can't observe those. That's not part of my macroscopic theory. If I could read someone's mind, if I could look into the state of their neurons and know what they were gonna do next, that'd be super useful. But the laws of physics don't let me do that. Okay? So, there's some feature of, there's some relationship between what is observable at the macro level and what you need to build a decent emergent theory that is fascinating to me, but I don't have anything to say about, I'm just letting you know that it's out there. Okay.

0:41:45.0 SC: With this setup, micro theory, macro theory, space of states, emergence map, evolution laws, we can start classifying all the different kinds of emergence. And we are being so general that some of what we call emergence won't be called emergence by anyone else. So, we had in our paper something called "type-zero emergence" or "featureless emergence." And that's where you have literally nothing but what I just gave you. So, there's no extra structure. So, later on for other kinds of emergence, we're gonna be talking about locations of things in space, right? Things are gonna have structure in space and they're gonna interact locally. That's gonna be very important.

0:42:22.9 SC: But you notice I didn't say anything about that? I didn't say that there were wholes made of parts, wholes, W-H-O-L-E-S, right? Very often in discussions of emergence, you will instantly leap to a discussion of little things coming together to constitute big things. But I didn't use any of those words when I said micro theory, macro theory. So, it's about the theories and their spaces of states. And at the level of type-zero emergence, there's no talk of parts, wholes, spatial structure, or anything like that. Roughly speaking, I have one example that is very important, very close to my heart of type-zero emergence, which is the classical limit of quantum mechanics. Arguably, there are other limits. There's other examples of, general relativity mapping onto Newtonian gravity or something like that.

0:43:12.2 SC: But basically what you have is some very rich description at the fundamental level, like quantum mechanics says you have these wave functions for many, many particles. But let's just look at one particle, okay? Let's not worry about entanglement or anything like that. When we learn about quantum mechanics in class, we are told how to take the classical limit of a big heavy quantum object. If a quantum mechanical object is very massive, then you can make its uncertainty in both position and velocity, relatively small, right? So, that's where you get a classical limit and you literally take the average value within the wave function of the position and of the velocity, and you show using Ehrenfest's theorem that those average values, those expectation values of position and velocity will satisfy the classical equations of motion. So, in this case, you have a very, very explicit emergence map from a wave function to classical point-and-phase base, position and velocity.

0:44:18.0 SC: And it is clearly a many-to-one map because all the information you're keeping is the location of the center, the expectation value of the position and of the velocity or the momentum, okay? You're losing, you're throwing away a lot of specific details about the shape of the wave function. As long as it's relatively localized, there's more than one way to be relatively localized. There's many, many ways to be relatively localized, but you don't care. All you care is where the centroid of that wave function is. So, to us, that is absolutely emergence. The classical limit is an emergent description from quantum mechanics. And notice that it is not valid in all regimes, right?

0:45:01.7 SC: Whenever you have a situation where quantum mechanics is necessary, like if you talk about the double slit experiment, right? Where you send a wave function through two slits and they interfere with each other, that is outside the classical limit. So, that's fine. All of these emergence maps are generally going to be well-defined in some domain of applicability that's perfectly legit. So, no one else calls the classical limit of quantum mechanics emergence, but we specifically and intentionally do so because we want to start with things that are easy to understand, right? Start with things that are unambiguous, that are perfectly clear what is going on while still building up to what ordinary people would use or have in mind when they invoke the idea of emergence.

0:45:48.6 SC: So, we didn't, in type-zero emergence, what we called featureless emergence, make use of the whole being made outta parts. A quantum particle, an electron that classically is described by a position and a velocity, and quantum mechanically is described by a wave function, doesn't have the property that somehow that position and velocity are made up of little bits of wave function, right? That's just not how it works. But still, there's a many-to-one map, which is a more general idea. But okay, admittedly, more often we're talking about situations where the big macroscopic things are made up of many different little microscopic things. When we take the piece of paper and we crumple it up, the idea of emergence there is that the piece of paper is made of atoms, okay? And the emergence map takes many, many different configurations of the atoms and combines them into a single configuration of the piece of paper. So, if we move on to type-1 emergence, which we call local emergence, that's exactly what we have. Local emergence, you'll not be surprised to hear, relies on a notion of locality. That is to say there is space, three dimensional space typically for you and me, but physicists will let their imaginations wander a little bit. But, okay, three dimensional space. And there's a notion of locality.

0:47:10.1 SC: Things have locations in three dimensional space. And locality refers both to what you are and how you interact, or at least I should say it can refer to both of those things. So, the word "locality" is sort of doing double duty here. And I think it's important to be careful. It means both that objects have extent and position in space in a way, which by the way, they kind of don't in quantum mechanics, right? The way function is defined all over space or in field theory for that matter. But once you get to classical particle dynamics, etcetera, things have locations in space and they also interact locally in space. Or they can interact locally in space, which is a way of saying that billiard balls bump into each other. They bump into each other, not when they have the same velocities or opposite velocities, but when they are near or at the same position in space, okay? That's what it means for interactions to be local.

0:48:10.8 SC: In quantum field theory, interactions are local. The field at any one point is affected by values of the field only right there at that point, or at least really, really nearby. Specifically, derivatives of the values of the field in space can affect what the field does, but there's no direct influences from far away. You can have something happen far away in field theory, but either classical field theory or quantum field theory. You poke a field. Let's say your field is a very down-to-earth thing like the surface of water on a pond, okay? You throw a stone into the pond and little waves ripple out in all directions. The waves can travel quite away, but what's happening is that the wave at one point in space is affecting the point right next to it. And those effects just accumulate over space and time to give you the impression of waves rippling out in all directions. That is local interactions.

0:49:11.9 SC: Even though it's not exactly what we're talking about here, let me be super careful 'cause I know there's a lot of quantum mechanics fans out there in the Mindscape audience. In quantum mechanics as we know it, in something like the core theory, the dynamics are local at the level of the Schrodinger equation. So, what we call the unitary dynamics of quantum field theory are perfectly 100% local. Said in yet other ways, when we are not measuring the theory, when it's just doing its thing without its wave function collapsing, that's all 100% local. Non-localities come in when you measure the system, and that's of course Bell's theorem and entanglement and EPR and all that fun stuff. All that's great, none of it we're gonna be talking about in this discussion today.

0:50:01.2 SC: So, in local emergence, we have a situation where macroscopic objects are made up of collections of microscopic objects, okay? So, locality is doing a lot of work. Locality is respected in both the microscopic theory and the macroscopic theory. Pieces of paper are made up of atoms, organisms are made of cells, whatever you want to say. All of these are different versions of big things being made up of little things, and local emergence, type-1 emergence respects that. If you wanna be a little bit more specific, famous examples are let's say the center of mass motion of planets in the solar system, right? And I bring this one up specifically because it goes back to Isaac Newton. When Isaac Newton first wrote down his law of gravity, and he's thinking about in the Principia Mathematica how to derive the motions of the planets around the sun, Isaac Newton was pretty smart. He didn't just say, "Let's idealize a planet as a point." He knew that planets had size.

0:51:09.0 SC: And so he put in the work to show that as long as the planet is spherically symmetric, its behavior in a gravitational field reduces to the behavior of a single particle with the appropriate mass and location of the center of mass and momentum of the center of mass. So, when we talk about the motion of the planets in the sky or in the solar system, we don't need to track all of the planets, all of the particles of which all these planets are made of. The Earth has something like 10 to the power of 50 atoms in it. Newton didn't know that. We still don't care about it for purposes of flying a rocket to the moon or anything like that. We can reduce this description of many, many, many, many particles to a relatively small number of variables. The position and velocity of the center of mass of the various planets.

0:52:02.8 SC: And again, some people would say that doesn't even count as emergence because it's too easy to understand; [laughter] it's too simple. I think maybe this is my physics background talking, that having some examples where you do understand things perfectly is actually good, not bad. A more standard example is of course, the emergence of fluids from what we call kinetic theory or atomic theory, right? The air in the room around you is made of atoms and molecules. You don't need to know that the air is made of atoms and molecules in order to predict what's gonna happen to the air. You would do a very, very good job if you knew the density, the pressure, the velocity, and the temperature of the air at each point in space. You might very quickly worry, there's a reasonable worry but not a difficult worry, that even though there's a lot of atoms around you, right? A lot of molecules in the air, the fluid description of the air has an infinite number of variables because there's an infinite number of points of space, and there's a value for these collective variables, the density and pressure and so forth at every point in space. They're smooth, continuous fields.

0:53:15.0 SC: Doesn't that look like not a coarse-graining map? Doesn't that look like we've increased the amount of information? But that's a little bit fake, because if you wanted to do the Mark Bedau thing and put the thing on a computer, you would actually discretized space. You would put the space around you and think of it as a lattice. So, in other words, rather than specifying pressure and density, etcetera, at literally every point in space, you would divide up space into little tiny boxes, maybe a millimeter across or a nanometer across or whatever. There's still going to be a large number of atoms in those little boxes. And in those little boxes, you go from all the atoms to just a small number of variables: Pressure, density, velocity, temperature, okay?

0:54:00.0 SC: So, for all practical purposes, you are actually greatly decreasing the number of variables. And there you get a description, which is a useful example to contrast with the center of mass motion example. And this is why the fluids coming from particles or atoms is a more common example of emergence, because the higher level description, the emergent description, talks a different language than the lower level description does. When you have particles that make up the Earth and you just reduce them to their center of mass motions, at both the microscopic level and the macroscopic level, you have particles obeying Newton's laws, okay? In the atoms going to fluids or to gases, example, at the lower level you have particles, and at the higher level you have fluids. That's a different kind of thing, okay? So, it's a fun example of emergence.

0:54:58.0 SC: But it's still, and you know, nothing in the definition that we gave said that the kind of theory you have at the macroscopic level has to be the kind of theory you had at the microscopic level. They can be completely different-looking theories. Some people would still say that the fluids coming from atoms is too cheap of an example. It's too easy. And the reason why is because we know exactly what the map is. We have a formula. We have equations for saying, given some collection of atoms doing certain things, here's what the pressure is, here's what the emergent temperature and density and velocity are. And the existence of such a simple relation, such a simple explicitly write-down-able emergence map, rubs some people the wrong way. Like I just said, I actually think it's a feature, not a bug.

0:55:48.8 SC: It is all still, by the way, remarkable that it exists. That you can throw away all of that information, throw away what all the atoms in the air around you are doing, and only keep track of some macroscopic variables. So, why is it like that? [laughter] As I already said, I don't know why it's like that. I'm not gonna tell you why it's like that. It was not part of our paper. But I still do wanna pause to kind of take a breath and reflect on the fact that the lower level laws that we have in the world allow for the existence of these higher level patterns, these emergent laws of physics. I don't know exactly why it's like that. I do think it has something to do with the locality of the underlying laws. And in some very real sense, quantum mechanics is less local than classical mechanics, but still, there are things called positions in space, and there are things, there's a feature in quantum mechanics or classical mechanics, there's a sense in which things interact when they're at or near the same position in the space.

0:56:58.0 SC: And those features of physics, as we know it, are so built in to how we think about the world, that we sometimes forget it didn't have to be that way. If I imagine that I have a certain set of what physicists called "degrees of freedom," whether they be atoms or values of a field at different points in space, the degrees of freedom are the things you have to specify to tell me what point you're at in the state space of the theory, okay? So, if we imagine we have some very large set of degrees of freedom, and we imagine relatively generic laws of physics in the space of all possible laws of physics that we could invent, very, very few of those laws of physics have any notion of locality built into them at all.

0:57:46.0 SC: Locality means that if I have an atom here, it's gonna interact with its nearest neighbors but not with all of the gajillion atoms somewhere else. Okay? That's a huge restriction on what can happen in the world. And one of the things that I'm interested in, in wearing my fundamental physicist hat, is why are the laws of nature like that? Did that notion of locality arise dynamically somehow, or was it just built in? This rubs up against questions of what are the laws of physics? Are you humane or anti-humane, etcetera? In a very real sense, honestly, I almost hate to bring this up, but it's a kind of fine-tuning of the laws of physics, the existence of locality itself, in the sense that generic laws of physics wouldn't be like that. And, look, maybe the answer is just, the world is like that; the world has space, space exists; there's never any option for it to be anything other than that. But I would like to think a little bit more deeply about that.

0:58:45.8 SC: Okay, that was all a little bit of an aside. All I'm trying to do is give you some examples about the most straightforward kind of emergence. And this is definitely weak emergence. What we're calling type-1 emergence is indisputably weak by the conventional characterization. But notice that we never mentioned derivability or anything like that. We never said that the higher level laws could or could not be derived. We just said that they existed and that's all that you need to have this kind of type-1 local emergence. Having said that, we should admit that some examples of type-1 local emergence are more straightforward than others. So, in the paper, we suggested a sub-categorization into what we called "direct emergence versus incompressible emergence." Type 1a and type 1b, if you wanna call them that, okay? So, direct emergence is basically the case that we've already talked about, either center of mass motion or atoms to fluids, where you have a very explicit formula that gives you the emergence map.

0:59:55.1 SC: And again, we didn't wanna say that having a formula was the important thing, 'cause maybe we don't have a formula now but we invent it, okay? But what we can say relatively objectively is whether or not the map from microstates to macrostates is algorithmically simple or algorithmically complex in the sense of Kolmogorov complexity. You've probably heard of this idea of Kolmogorov complexity, that you have all the possible computer programs that would output a certain string, and the Kolmogorov complexity of the string is the length of the shortest computer program that does this. A couple times recently we've mentioned Charlie Bennett's version of logical depth, which is not the length of the program but the time it takes to run the program to predict that. And you could use either one for this particular purpose.

1:00:43.8 SC: The idea is do you have to work hard to specify the map from the microstates to the macrostates, or can you state it in a very short compact expression? So, you get direct emergence like fluids from atoms when the map is algorithmically simple. Incompressible emergence is when you have to work hard to specify the map, the emergence map from microstates to macrostates. So, arguably, and I don't have a knockdown version of an incompressible emergence but I do wanna be open to the possibility. The example that Mark Bedau uses in his paper is the Game of Life, Conway's Game of Life. John Conway, the mathematician, famously had this cellular automaton, a two-dimensional cellular automaton with on and off sites in a square lattice and rules for the sites propagating over time. And what you find in the Game of Life is that there are certain very specific configurations of sites being on and off that persist over time. You can get a glider that has a certain shape, and it moves up and to the left or up and to the right, depending on how you've pointed it. You can get a glider gun, a configuration that sort of produces gliders and spits them out and as well as many other more complicated things.

1:02:03.2 SC: And roughly speaking, the only way to tell me whether you have a glider or a glider gun or whatever is to give me the explicit expression for what sites are turned on and what sites are turned off. It's not like in a little integral formula that you can just write down and plug it in. So, I think arguably that's a case of what we're calling incompressible emergence, where there is a higher level way of talking. It's not super good because I don't know how comprehensive the higher level language of gliders and glider guns would be in talking about the dynamics, but it's there. And the only way to specify it that I know is to just give the explicit formula in an incompressible map from microstates to macrostates.

1:02:47.9 SC: So, to me, that idea of whether or not the map from the micro theory to the macro theory is algorithmically simple or complex is a better thing to keep in mind than whether or not the macroscopic behavior is surprising or novel or unexpected or whatever. Those seem to me to be in the eye of the beholder, whereas the Kolmogorov complexity of the map is something that in principle you could figure out what it actually is. Which reminds me that I should also draw a distinction for those real emergence fans out there. There is a distinction in emergence talk which is not quite the "strong versus weak" emergent distinction, but it is the "epistemological versus ontological" distinction. Okay?

1:03:34.0 SC: So, epistemological emergence would be, well, you have a higher level theory which is just a way of talking about, a way of knowing things about and describing lower level things, versus ontological emergences, "Oh, no, there really are new things at the higher level," right? Ontology being the discussion of what really exists. To me, I don't quite vibe with this distinction because I'm on the train of people like Dan Dennett who talked about real patterns. Also James Ladyman, another former Mindscape guest. He and Don Ross have done a lot of exploration of an idea; it's not original to them, but they've been championing it for a while, called "structural realism." Where the idea is that what really exists out there in the world, we don't know, right? We don't know the fundamental theory of everything yet, and therefore you might worry it's impossible to say that anything is real. We used to think that atoms are real, but now maybe they're just manifestations of a wave function. Does that mean that atoms don't really exist? Ladyman and Ross in structural realism more generally say, no, there are patterns in the behavior that are preserved as you go from a higher level theory to a lower level theory, and those patterns are real and they count as really existing. So, ontological structural realism, I should label this as.

1:05:02.5 SC: So, what this does is, it licenses you to do exactly what you want to do and should do, which is to say tables and chairs are real. To me, if your definition of reality is not broad enough to encompass tables and chairs, then I think you're probably barking up the wrong tree or at least you haven't chosen the most useful definition of reality. And structural realism and the real patterns point of view from Dennett both point you in a direction of saying, "There's absolutely something real about these higher level emergent things." They might be epistemologically useful, but they're also ontologically worth taking seriously. So, I'm not dwelling on that distinction between epistemological and ontological, but I'm just letting you know that to me, this is ontological emergence that we're talking about.

1:05:49.8 SC: Okay, so that's the easy part. [laughter] Those are the easy parts of emergence. And we only have two more categories left, and they're action packed, but I do think that it's very helpful to think in the terms that we introduce in this paper. So, the next category, given that type-1 emergence was local emergence, you're not gonna be surprised that type-2 emergence is non-local emergence. And we had to think a lot. It took us a bit of discussing and writing things down and noodling about what exactly it means to say that there is non-locality in the emergence idea. Is it that the objects are not local? Or is it that the rules are not local? Or what is going on? This is something that... And so the short answer is both. [laughter] You can have objects at your higher level of emergence that are not locally defined. That is to say they are not simply made up of little locally collected pieces of your microscopic theory, and/or you could have influences between objects in your higher level that sort of extend over space in ways that ordinary physical things like billiard balls bumping into each other do not.

1:07:13.7 SC: And this is real. This is something that absolutely is plausible. Certainly in biology or in the social sciences where very often both your microscopic theory... Microscopic theory doesn't mean to be fundamental physics. You can absolutely have a case where your microscopic theory is, for example, human beings, and your macroscopic theory is political structures or the economy or something like that, right? So, we want our conception of emergence to be rich enough to include all of those different possibilities. We're not physics chauvinists here. And so, it might be very often true in biology or the social sciences that we have very important emergent entities that are not localized in space or time. Maybe the US Constitution. Now, the US Constitution is written on a piece of paper, but the Constitution itself doesn't have a location in space. A contract or an obligation more generally is not localized in space, but it might play a super important explanatory role at some higher emergent level. Right? He didn't do this because there's a contract that says that he can't do it. That's the idea of the higher level emergent theory, is supposed to be able to tell you what happens in some reliable way. And therefore the pieces of it are going to have some causal power. And if they do, in other words, the existence of something or non-existence of it affects what happens. Okay?

1:08:45.1 SC: And if they do, they're part of that higher level theory, and very often they're just not local. If you think that consciousness is in some sense emergent, there's not one neuron in your brain where the consciousness is located. It's not even exactly right to say it's localized in your brain, although we could have a discussion about that. Simon DeDeo, who's another former Mindscape guest, read our paper and gave us a wonderful example of this non-local emergence that is as close to fundamental physics as I think we can get, which is the jerk. This is the jerk in the mathematical sense. Sometimes you'll be told that the first derivative of position with respect to time is the velocity. The second derivative is the acceleration. The third derivative is the jerk. And then there are higher level derivatives with "snap," "crackle," "pop." I swear to God, these are the words that are attached to the higher derivatives of position with respect to time. But if you know anything about Newton's laws of classical mechanics, F equals MA is the second law, force is mass times acceleration.

1:09:54.1 SC: So, it's a nice feature of classical mechanics that the state space requires you to give me positions and velocities. It does not require you to give me accelerations because accelerations are given to you by a formula, F equals MA. And once you know the position and velocity and you figured out the acceleration from F equals MA, that is enough to determine the entire future evolution of the system. That's what Laplace's Demon does. So, what you notice in that discussion is the jerk, the third derivative of position with respect to time, nowhere appears. It is not part of your fundamental description. You could figure it out, you could calculate it. Once someone gave you the positions and velocities, and you used F equals MA to predict the future evolution, you could calculate what the jerk was. But it didn't play any causal role. The causal role was already filled up by the other things, the position and the velocity and the acceleration. But what Simon points out in a nice little paper that he wrote, is that you can feel the jerk as a person in a car or an elevator. The jerk is the rate of change of acceleration. So, you not only feel acceleration, but you would swear to God you could feel that acceleration was changing. Right?

1:11:12.9 SC: What is going on there? Why are you able to have jerk as a part of your useful description of the higher level world? It could even play some causal role. And the answer is, roughly speaking, because it's non-local emergence. Because you at the higher level are kind of averaging over things that happen at slightly different points of time. Right? The reason why you think that you know what the jerk is, is because your conscious experience of the world is not really instantaneous. It has some length of time over which you're sort of integrating what you're feeling and then you can tell people, you can report back on what you have experienced. And from our point of view, you could if you wanted to trade in that finite discrete interval of time for a finite discrete region of space. Because at the fundamental level, if you're trying to map this higher level description onto what's going on fundamentally, everything that happens to you in the elevator or in the car or whatever, is determined by what happens in your past light cone. So, whatever you count as a moment before, there is a region of space that is not that big but is pretty big, in which if you knew everything that happened, it would completely determine what you call the jerk going forward.

1:12:37.9 SC: But it does it in a non-local way. You need to know not just what happens at what point in space, but what is happening over a certain region of space. So, when you go from the microscopic theory of just Newton's laws or whatever, to the macroscopic theory of your human experience, there is a bit of non-locality that has seeped in. And I think that's a wonderful example of exactly this kind of type-2 non-local emergence. It doesn't happen that much. It doesn't happen in any real noticeable way if both your microscopic theory and your macroscopic theory, I should say, if both your lower level theory and your higher level theory are still pretty low level. [laughter] So, the jerk example is because your higher level theory was human beings, which are very high level. If your lower level theory is the standard model of particle physics and your higher level theory is some theory of atoms and electrons in a superconductor, that's not very far removed, right? Your so-called higher level theory is still pretty low level, and in that case, the speed of light really is a fundamental limitation, and so you're not going to get emergent non-locality in that case.

1:13:49.1 SC: So, emergent non-locality basically has the possibility of being relevant when you're coarse graining so much that the speed of light is just not a limitation. As far as you and I are concerned, when I see you in a room a few meters away, I'm basically seeing you now, right? Technically, I'm seeing you in the past 'cause it takes time for light to get to me, but on the time scales that I personally move around and react to things and think about things, I'm seeing you instantaneously. So, because we all move so much more slowly than the speed of light, in a fairly reasonably sized region of the universe, the speed of light is not there and we can think non-locally that non-local effects can be very relevant to how we talk about the world. So, it's an interesting sort of changeover from emergence at the super micro to the merely slightly micro level, versus emergence at the micro level to the truly macro level.

1:14:49.0 SC: Now, why are we bothering about this? Why do we really distinguish between type-2 emergence and type-1 emergence? Because the possibility of type-2 emergence, non-local emergence, opens up an interesting possibility that wasn't there in local emergence. And we didn't give an especially fascinating name or label to this possibility, but we called them "filter functions." We should have worked harder on giving the name 'cause maybe it would be an idea that would catch on better if we had come up with a snappier label for it. The idea of the filter function is this. Imagine you have some micro dynamics that is, appears to you to be local and perfectly well understood. The standard model particle physics, the core theory for example. But let's imagine that you are convinced that when the global configuration, that is to say the state of the universe or the state of some non-trivially sized part of the universe is a certain way, then the dynamics that you thought were good enough at the microscopic level fail.

1:15:58.7 SC: That is to say new features kick in at the macro level but only for certain configurations that are global that you need to give me information spread out over space to specify. So, when we do particle physics for example, when we do experiments to learn about the standard model particle physics, what kind of experiments are we doing? We're typically at a particle collider, we're at the Large Hadron Collider, we're smashing particles together. It all happens in a very tiny region of space with relatively few particles colliding to each other. Okay? So, we've developed the ideas of the core theory in a fairly simple set of circumstances, small regions of space-time, small number of particles. The possibility exists that how electrons behave is different if that electron is in the human brain versus when it is in a particle accelerator, when it is in the detector, the ATLAS detector at the Large Hadron Collider or something else. This is not what you expect from the core theory. This is changing the core theory.

1:17:12.6 SC: We're being very, very explicit about it. But we're saying that you could do all of the experiments you want in this sort of localized small number of particles regime and never notice the new dynamics. So, the idea of the filter function is, there is an equation telling you how the particles or how the constituents of your micro theory behave, and that there's a set of terms in that equation that refer to what happens locally. In other words, there's a set of terms that say, "In some region of space, I care about the particles bumping into me or whatever. I don't care about what happens far away." And then the filter function says plus [laughter], there are additional terms that do care about the global configuration. There are things that affect the dynamics of electrons. And again, I'm saying this hypothetically, I don't think this is true, I'm just allowing for the possibility. But there could be things that affect the dynamics of electrons when the broader context the electron is in has the form of a human brain, that are not noticeable if that broader context is not there.

1:18:23.0 SC: So, in type-2 non-local emergence, you have the possibility of new dynamical considerations that you would not have noticed by simply speaking the language and doing the experiments that you thought were appropriate for your micro theory. So, to be clear, you're not changing the micro theory. What you're doing is you're explaining why the micro theory was incomplete. These new dynamics governed by the filter function were always there and they could in principle be captured in terms of the micro theory. Okay? So, you're not adding new ontology, you're not adding new consciousness particles or whatever. Let's say the micro theory is the core theory. It's the same quantum fields that you had all along. They just have certain dynamical properties that were not evident from the sort of microscopic perspective. So, it's natural to miss them. We tend to only probe a certain domain and then extrapolate.

1:19:22.7 SC: And in quantum field theory, we have good reason to do that. We can give you an argument why based on things like locality, that is more than good enough. But in the space of all possible theories, maybe quantum field theory is not right. Okay? Maybe there are true differences. So, if that's true, if that possibility is realized in nature, this fits into our type-2 dynamics, type-2 emergence. We're not introducing new stuff, we're introducing new dynamics for the old stuff. But the new dynamics depends on global non-local considerations. So, you might have missed it. So, we're trying to explain how you could simultaneously think that you believe everything that particle physicists are telling you about the behavior of quantum fields, and yet we need a theory that changes those dynamics in certain situations like human brains. Okay?

1:20:20.0 SC: So, again, I think that this is possible, but I don't think it's actually true in the case where your lower level is literally fundamental physics and the core theory, because I think that locality, the constraints of locality there are very strong. I think that non-local emergence is much more plausible when your micro theory is already a little bit macro. But we're open to the possibility. So, if someone who believes in strong emergence wants to take up that challenge, we have an equation. [laughter] We have an equation that tells you how in principle you could modify the core theory to allow for this kind of dynamics, and we encourage you to speak the language of that equation. Okay?

1:21:03.9 SC: Now, in the paper, Achyuth and I mention that this possibility could open up the idea of counterfeit downward causation. Downward causation is the idea that there are things that happen in a higher level theory that play a causal role in the lower level theory, okay? That you need to know about something that sounds purely emergent in order to completely explain something that is going down at the microscopic level. And in what we call type-2 non-local emergence, you can think that's what's happening because you have an electron that is in a human brain, and maybe you are smart enough to do an experiment that actually showed that this electron behaves differently in a human brain than in a rock or in a particle accelerator. Okay? So, that would look like downward causation. That would look like the existence of a brain has been affecting the motion of the electron. But in our definition of type-2 emergence, and by the way, there will be a type-3 where this is not true, but in what we're calling type-2 emergence, this kind of downward causation is merely counterfeit.

1:22:14.0 SC: And the reason why is because the micro level theory and indeed, good theories more generally are causally closed. They are sufficient to describe what happens perfectly well in their own domain of applicability. There is a rule that I talk about in the big picture, where I say you shouldn't bounce willy-nilly between levels of description. The levels are independent, okay? If you have a good macro description of some emergent phenomena, the whole point is it shouldn't rely on what's going on at the microscopic level and vice versa, what's going on at the microscopic level shouldn't rely on the macro theory. So, that doesn't mean you don't have both descriptions, but you should only talk one language at a time. So, for example, I hate to bring up this example, but one might ask, why are the Philadelphia 76ers off to a truly awful start [chuckle] in this basketball season 2024/'25, which they are, sadly. This is dwelling on my mind and bringing me down.

1:23:21.2 SC: What kind of explanation could you offer for a bad start? Well, maybe it's because there's too many new faces, right? They let a lot of the players go from the previous year and they have a lot of new people, they're still working hard to incorporate them into the offense and the defense, the schemes, okay? Or you could say maybe it's 'cause of injuries. Joel Embiid has been injured, Tyrese Maxey just got injured. Paul George, all their best players have struggled with injuries in this short beginning into the season. That's another possible explanation. And you can very legitimately debate which of these factors is more important, or maybe they both matter, maybe that there's a third thing we haven't thought of that. That's a perfectly sensible debate to have.

1:24:02.9 SC: There's another possible answer you could give. Why are the Sixers off to a truly awful start? Well, it's because of the initial conditions of the universe and the laws of physics, right? That's an answer you could give. That's an answer you could give to any question of the form. Why is this particular thing the way it is right now? Because of the initial conditions of the universe and the laws of physics. That's not a wrong answer. The initial conditions of the universe plus the laws of physics really do in some tangible sense explain why the current situation is what it is. All I'm trying to say is you shouldn't mix them together. Once you've chosen to speak the language of the microscopic theory and you've said, "Why are the Sixers not doing well? Well, because of the initial conditions of the universe plus the laws of physics, plus this coarse-graining map that fits the record of the basketball team into that language." You don't also say because of injuries or because there's too many new faces or something like that. That is over-attributing what is going on. Because in principle, from the initial conditions of the universe and the laws of physics, you can derive the fact that they're injured or they have too many new faces or whatever. You only get to give the explanation once.

1:25:18.0 SC: And so, that's the reason why it is counterfeit downward causation. Because in this type-2 picture, there are higher level non-local influences, but ultimately you can give a perfectly good lower level explanation. You can explicitly include those influences in the microscopic theory. You don't need to invoke the higher level. You don't need to talk about influences at the emergent level exerting causal influence on the micro level. So, examples of downward causation, if you look them up, you might find something like, why is a certain hydrocarbon molecule in a certain physical location of the world? And people will say, well, if you realize that what you're actually talking about is a little hydrocarbon molecule in a gas tank of a car, then you can give an explanation for why that molecule is there based on ideas of the internal combustion engine and the modern economy and the need to get from place to place in a suburban environment or whatever.

1:26:20.1 SC: And they would go so far as to claim that unless you give that explanation, you have not answered the question; you've not actually accounted for why that molecule is where it is without using these higher level emergent ideas. To me, I think that's just a mistake. I think that's just wrong. I think that you can, in principle, not in practice obviously, but in principle you can perfectly account for the location of that molecule purely at the micro level, right? There's no one saying in the context of type-2 emergence that you can't account for the dynamics that led that molecule to be there, purely in terms of the local dynamics of the underlying core theory, right?

1:27:06.0 SC: You have a much more efficient, informative explanation at the micro level. That's a very common thing. That's the wonderful thing about emergence, is that you can have these explanations based on far fewer pieces of information than you need at the micro level, but that's not the same as saying you needed it. It might be more convenient. You can talk either language you want, one language might be better for the purposes that you have in front of you. Using one description or another might be useful relative to certain purposes, but it isn't demanded. So, I think that at that level, if we're talking a type-2 language where the microscopic states, the microscopic entities are in principle all you need, but they might have some global influences on their dynamics that you missed by doing microscopic experiments, still you don't need to actually talk the higher level emergent language in order to account for what happens at the microscopic level.

1:28:06.3 SC: So, within type-2 or non-local emergence, you can have apparent downward causation 'cause global properties matter, but it's not honest. You could have given a complete account purely within the causally closed micro theory. And having said all that, of course, you'll be completely unsurprised to hear there's one final type of emergence that we talk about, which is labeled type-3 or augmented emergence. I think this is what people have in mind when they talk about strong emergence. I think sometimes when they talk about strong emergence, they're really just talking about what we call type-2. Sometimes they're talking about what we're gonna call type-3. That's why we have these extra specifications so people can be clear when they're talking to each other. The idea of augmented type-3 emergence is to admit that the micro theory really is simply incomplete. In other words, that there are regimes of applicability for the micro theory and a regime of applicability for the macro theory, and one is not a subset of the other.

1:29:09.8 SC: There are places where the macro theory works, where the micro theory is just wrong in its own terms. But it's not just that you made a mistake, we're using the same kind of filter function language we introduced in type-2 to help provide an understanding of why you might have believed the micro theory in the first place. In the domain where you test and get empirical information about the micro theory, the micro rules work, and these new macro interaction or new macro influences don't play a role. But in actual type-3 emergence, we allow for the existence of truly new ontological entities that are strictly global. Maybe you think that way about consciousness; maybe consciousness is just not ever reducible even in principle to the to-ings and fro-ings of microscopic particles. Maybe there's something extra about it: Consciousness juice or something, the spirit, the soul, the geist.

1:30:12.4 SC: Or maybe in a slightly more subtle way, you think that there's a teleological aspect to the laws of nature, right? Maybe you think that, yeah, when you look at particles bumping into each other, it all looks random. But in fact, to explain the origin and evolution of life, you need to include an extra effect having to do with certain things are more likely than others. If they ultimately lead to some biological happenings down the road. That would be a truly new thing, a truly non-local aspect of reality. So, in type-3, these new things, these augmentations of the micro theory are real. They do affect the micro-dynamics, but they only do so when the micro configurations are under certain global conditions. So, again, when an electron is part of a brain, is it affected by these new ontological features? That's what the filter function is supposed to tell you.

1:31:12.8 SC: The filter function is there both in type-2 and type-3. The difference is that in type-2, the filter function says when you're in a brain, you turn on new interactions between the existing microscopic lower level features. In type-3, when the filter function turns on, you allow for new influences that are simply not describable in terms of the microscopic features: Consciousness, the spirit or whatever, or the, teleologically, the future goal of the universe. So, that's supposed to help explain, it's supposed to help reconcile the way in which the microscopic theory is not complete. It is not correct in some sense, but you might have been tricked into thinking that it was correct because that's where you probed it, right?

1:32:05.7 SC: I personally do not think that anything like type-3 emergence happens when the lower level theory is particle physics, is the core theory, okay? When we are talking about lower levels that are really fundamental physics as they're currently understood, we have very, very good reason to think that the dynamics are truly local. But the thing that we're trying to do in this paper is be as explicit and clear as possible as to how that expectation could go wrong. So, again, when people say consciousness is strongly emergent, etcetera, that's not my way of thinking, but it's a free country and I could be wrong. I want people to pursue other possibilities. What I object to is just needless vagueness. I think that people can be explicit about the ways in which these things come to pass. And so, I think that again, we have an equation in our paper, which gives you a template for showing how truly new ontological features of the world could become relevant but only globally.

1:33:14.0 SC: That would be a case where the core theory of particle physics did not completely and correctly account for how human beings behave. That is obviously a logical possibility. But if you wanna explore it, you gotta do more than wave your hands. You gotta be explicit about how the equations change. We have given you a template for changing those equations. So, I think that's a good place to end up because I've been very, very explicit about what I think is the way the world actually works. But I don't know everything. [laughter] I could certainly be wrong about some things. I think that in these very difficult questions, it's perfectly good to be open-minded. Hopefully we will have given people a little bit of a roadmap, a little bit of an example of knowing what to look for, right? Knowing what it would mean to have these things be relevant to our best understanding of consciousness or life or anything else, and maybe a bit of vocabulary to distinguish between other options that are a little bit less traumatic than that. I know that thinking this through has helped me anyway. Whether or not our jargon catches on or not, I can be a lot more clear when I'm talking to other people about what I mean by different kinds of emergence. So, what more can we ask from that? With that, thanks for listening. Thanks for supporting Mindscape. Talk to you next time.

[music]