269 | Sahar Heydari Fard on Complexity, Justice, and Social Dynamics

0:00:01.3 Sean Carroll: Hello, everyone, and welcome to the Mindscape podcast. I'm your host, Sean Carroll. One of the great things about having a podcast is that I am not a university, in particular, I do not have departments, I do not have to worry about hiring people, I can just talk to people about ideas, whatever the ideas might be, as long as I think that they're interesting. So as a result of that, I don't need to worry about which department somebody fits in. You know that I'm back now as a real professor at Johns Hopkins. When I was at Caltech, I was a research professor, which was great for doing research, but you're less integrated into the wider university life than I am now. So not only am I a professor now, but I am both a professor in philosophy and physics, and I have attachments to other parts of the university, etcetera, so I'm thinking a lot about these issues of how people fit in. And it's so frustrating, so annoying, that some people might be brilliant scholars, but don't fit easily into any one department. Here at Mindscape, we don't need to worry about that.

0:01:04.6 SC: And today's guest is a wonderful example. Sahar Heydari Fard is a philosopher, she is in the Philosophy department at Ohio State University. No doubt about that. Her work touches, as you'll hear in the episode, on ideas put forward by people like Cailin O'Connor and Elizabeth Anderson, who are also both philosophers who are previous guests, but it also touches on ideas that are similar to what Herb Gintis talked about. And Herb was also difficult to fit into a category, but if anything, he'd be called an economist. And a lot of what Sahar thinks about is political science, justice, political kinds of philosophy. You'll hear names like Hobbes and Hume come up in our conversation. People like John Rawls and others have thought a lot about what is a just society? The specific angle that Sahar brings to this is, guess what? Complexity theory. Even more so than people like Cailin O'Connor, Elizabeth Anderson, she's not only thinking about society and its dynamics, but she's explicitly doing so from the lens of complex systems research. Clearly, if anything is going to be complex in the world, all of society is going to be pretty complex, right?

0:02:20.4 SC: And there's a question you're perfectly willing to ask, willing to wonder about, if you're a complexity theory skeptic, which is, do these techniques that go under the rubric of complex systems research have any specific applicability to the individual fields? If society is complex, if the economy is complex, if the internet is complex, but also if an organism is complex, if an individual cell is complex, are there really features that are common to these that are worth thinking about versus just thinking about each individual thing in its own right? And I think that the answer that comes out of the conversation you're about to listen to is that there is something gained by thinking about society as a complex system that we can learn about from thinking about complex systems in general. We'll think about phase transitions, we'll think about game theory, we'll think about networks, and that not only helps you analyze the structure of society, which maybe is a perfectly obvious thing, if you think this thing might work.

0:03:29.0 SC: The interesting thing is, it helps you philosophize about what a good society would be like. Thinking about society as a complex system, as Sahar will try to make the case for it, and I think that she's convinced me anyway. Makes you think about what society should shoot for in a different way. It's maybe in retrospect not surprising, the better you understand something, the more likely you are to have good ideas about how to optimize that thing. So whatever your personal values are, thinking about society accurately might suggest ways to achieve those personal values in the organization of society, and the specific fact that society is a complex system will... As you will see, I don't wanna give it away, but as you will see, we'll make certain suggestions about that. So glad that I'm not a university and I don't need to worry about fitting people into departments, I can just talk to people because they're interesting. With that, let's go.

[music]

0:04:46.7 SC: Sahar Heydari Fard, welcome to the Mindscape podcast.

0:04:49.2 Sahar Heydari Fard: Thank you for having me, Sean.

0:04:50.8 SC: I have to start with a conversation that we had just a little while ago, we were at a workshop together, and I asked you if as a philosopher who is thinking about society and political questions, do you ever collaborate with political scientists? And roughly speaking, you said yes, and in those collaborations, your job is to run the model, run the computer model of what they're thinking about.

0:05:14.4 SF: That's true. [chuckle]

0:05:16.2 SC: And it made me laugh because as a physicist who talks to philosophers, I often let people know, talking to philosophers is super useful, but you don't go to them to calculate a Feynman diagram. So, just give us a teaser or an overview, how do you end up being the philosopher that people go to to run their computer simulations?

0:05:38.3 SF: Well, that's a lovely question, thank you for starting with it. The short answer is that perhaps because of the things that you have mentioned in your conversation with Cailin O'Connor in another episode of this podcast...

0:05:55.4 SC: Yeah.

0:05:58.2 SF: About just-so stories. The idea is that we have some assumptions about how society works. Based on those assumptions, we come up with some system for how things can be prevented from going the wrong, or can be maintaining a stable functioning level. And then based on those assumptions, we also make a lot of prescriptions for what shouldn't be done, or how can we make things more optimal or more effective? And then follows some normative claims about, what is just, what is moral, what is fair, blah, blah. But those assumptions sometimes are faulty assumptions, or assumptions that might come from our intuitions about our interactions at the local level in a very small scale, but it might not hold when you are relying on those assumptions at a population level. Or when the population has five times the size that philosophers who've been theorizing this [chuckle] could possibly imagine. And the modeling part comes in to test out the ideas that if we start with those assumptions of how these interactions work and then expand the size, would those intuitions still be a reliable guide for us to know what we should do or how we should make things the way that we wanted them, or who should be the person who we prioritize their wantings or their needs, in what way?

0:07:37.4 SC: That's actually a great answer because it's very...

0:07:40.0 SF: I'm full of great answers.

[chuckle]

0:07:41.0 SC: It's a great answer, because it parallels very nicely what happens between philosophy and physics, because physicists, as I very often say, just love to get the right answer, and they're very happy to get the right answer for the wrong reasons, if it gives them the right answer. But then the problem with that is that when you extrapolate beyond the regime in which you know you get the right answer, your faulty reasoning comes to bite you a little bit. And based... In some sense, to over-simplify, you're saying same thing is true in political theory?

0:08:13.2 SF: No, absolutely. But also when you started before elaborate, but the connection you're referring to between philosophy and physics, I was thinking about other kind of connections that have put physics in a place that does the modeling for philosophy. Not thinking about the Aristotelian way of motion works, or the way that we moved from analytic descriptions of the world to more statistics-friendly [chuckle] because we figured out like, well, there are details that you can just dismiss and it's just not helpful, and you can talk about averages and they're not only better, but also in some sense, give you a better understanding of this ontological structure of what you're talking about.

0:09:07.5 SC: Yeah.

0:09:09.5 SF: And then, I think complexity is another big step that you're making that, "Oh, okay." So, it is true that the analytics solutions might not give us everything. It is also true that sometimes statistics or those averaging models are the best way to cope up things. But sometimes you have this kind of organized complexity in the middle that, you cannot just all the way jump in and say, "Oh, I talk about the average, and then I've gathered everything that I wanted." And I think that's exactly what I do. Philosophers or our social scientists like political scientists, to distinguish these and distinguish the tools that they're using to talk about each of them.

0:09:52.6 SC: Well, that's also a great answer, I gotta say, because physicists are a little bit nervous about complexity for exactly the reasons that you mentioned, they like things where you could average things out. The classic example being getting fluid mechanics from a bunch of atoms, and then you just get still a very simple explanation at the end of the day, fluid mechanics. But complex systems are different. We've talked about this a lot in the podcast, but let's pretend we haven't. What is your personal way of thinking about what complexity is and why it's interesting?

0:10:25.5 SF: Why do you have to start with the hardest questions? [laughter]

0:10:28.1 SC: They're all gonna be hard, they're never gonna get easy as we go.

[laughter]

0:10:31.4 SF: Well, the way that I'm thinking about complexity is a kind of phenomenon that has some sort of stability at the aggregate level, but that stability is not dependent on the homogenous components who have the same kind of direction, incentive property, however you're thinking about it. In the social world is, If you're thinking individuals as rational actors with diverse incentives, if you don't wanna think of them as rational actors, just think of them as actors who are...

[laughter]

0:11:01.6 SF: Acting for diverse reasons and diverse needs and diverse situations. But at the same time, you see that some kind of stability is emerging from this kind of many actor condition, that some of them are more influential, some of them are less, but it requires coordination among so many different level at so... Different individuals at so many different levels to get that kind of stability. And without that you cannot have any explanation whatsoever, you cannot just assume that like, "This is just a unit," like the public opinion or political opinion, it just occurs and happens to oscillate. I don't know why.

[laughter]

0:11:49.0 SF: And you see that it oscillates over time...

0:11:52.4 SC: Sure.

0:11:53.1 SF: Sometimes in response to external stimuli, but sometimes endogenously, sometimes because people are re-orienting their political views or connections or who they're talking to, who they're trusting, and you see that kind of oscillation over time. And if you want to explain that kind of thing, it seems like, again, going back to individuals and their incentives is too much information, people do things for all sorts of reasons, you cannot take a survey of everyone why they're doing what they're doing. But thinking of them as their average of the public opinion, is too little information get lost a lot. The way that I'm thinking about complexity is, first, a phenomena in which you have many actors, that coordination among them is necessary to generate some level of stability, but you don't have any central organizer that is doing the work for you. We have something that people call the emergence of the stability that also constrains the behavior.

0:12:54.6 SC: Well, I'm gonna become a broken record, but that was a great answer because...

[laughter]

0:12:57.8 SC: It's very different, it's, you're highlighting a different aspect of complexity that we talked about before, this stability over time. It reminds me of Erwin Schrödinger's little book. You know Schrödinger, the physicist, wrote this little book called "What Is Life?" and he has a line in there that answers the question where he says life is something that keeps on moving long after it should have stopped. And no one pays attention to that line in the book where he actually answers the question in the title, but it is fascinating to me, and as a physicist, I'm trying to figure out where these come from. You have the luxury of saying, "Okay, I'm gonna look at society." They're there. Do you wanna use the tools of complexity, I guess, to think about how that stability happens, and presumably how it might change? They're not perfectly stable, like our current democracy, for example.

0:13:52.1 SF: That's true. That's true. And that's a very, very important point to think about that, one way of thinking about stability is that average kind of mentality, that it's just, it's stable. I don't know, just let me explain what's going on. And the kind of phenomena that I'm seeking to explain in the world is the kind of phenomena that is stable. So I start from that assumption of a stability, and then I can bring in some functional definitions, or some other assumptions that make the individuals perhaps a function of that stability at the social level, or their property is a function of that, and then explain a lot of things. But then you want to incorporate change. And what happens is that, the only way to conceptualize that is that we should burn everything to the ground...

[chuckle]

0:14:40.3 SF: And start from scratch to be able to have change. The avenues for gradual change or even rapid change that is not revolutionary in a sense that it doesn't make us completely unstable, but help us to go from one equilibrium or relatively more stable state to the other, another, or just have a stable situation that drifts over time or obsolete, or have some other kind of dynamic behavior. But it's not like setting metaphysical stone, that this is it and I prioritize this metaphysically and explain everything else based on the stability. And you see that in social theory and political science, this a lot happen... Happens a lot, because when you recognize that there is some higher level of stability and you don't need to figure out how individuals interacted to generate this, then when you are trying to use that to theorize how change happens, you end up saying things that people don't want to do, like live through a revolution or make it possible, or afterwards in a way that I did.

0:15:51.5 SC: Very, very... Okay, good. That brings in another aspect of your work, which is that you do want to make the world a better place, or at least theorize...

0:16:00.5 SF: That's true.

0:16:00.6 SC: About how to make the world a better place, right. I just hope the world becomes a better place, but my work is not gonna help it along, so that is an important distinction. But it's nice because what you just said is, the theory of change aspect of it is important, it reminds us of punctuated equilibrium in biology where population is doing pretty well, but then mutation comes along that makes them do even better, and suddenly there's a pretty rapid change. That's the kind of thing I'm gonna suggest, and you'll correct me if I'm wrong, that complexity theory is geared up to talk about in maybe ways that other theories are not.

0:16:36.9 SF: Absolutely. But I think on top of that, the complexity theory can help us think about how punctuated equilibrium happens, sometimes, given that there are so many other inter-related things that would keep that equilibrium very, very stable. Sometimes, this comes up in discussions about economics, people talk about norm change in the same sense of punctuated equilibrium. You have the practice of foot binding, it started because a king in, I don't know, somewhere, very much liked one of his dancers biting her foot in a certain way, and then that became a way for other women to make themselves desirable for either the king or the people who are close to the thinking. And then it became like a form of status to have a wife or a partner or whatever, who are concubine, who has that property, and then incentivize women to do that more and more often, to be able to marry up with a lot of income inequality or wealth gap or whatever feudal systems [chuckle] deal with. It became like some method for people to marry their kid up. You have a stable situation that a parent is stuck, because if they do this while they're hurting their child, the child cannot walk anymore, it's very painful and it's very...

0:18:12.7 SC: Yeah.

0:18:13.8 SF: I don't know whether you've seen pictures of it, it's horrendous. But if they don't do that to their child, their child doesn't have the possibil... Their daughter doesn't have the possibility to marry, and in an environment in which there's no other way for these women to support themselves other than marrying. Well, you have harmed that child. So everyone has incentive to keep things in place. But centuries passes and it's now a practice that everyone does, so you don't even do this to marry up or down, [chuckle] but it's just like a stable feature of the culture.

0:18:48.4 SC: Locked in. Yeah.

0:18:49.8 SF: Yeah. And then people are realizing that this is a problematic thing, like, "We shouldn't do this," but no one has the incentive to deviate from it. But what happened is that, a group of people who are wealthier get together and signed a pledge that they're not gonna let their child marry to a person who has done this to their foot, or would not do that to their own children. And then it had a trickle effect, overnight, or if not overnight... A very short period of time, that practice stopped. It only stopped, but it was a negative view of those who do that. But this is a situation in which the equilibrium is kind of separate or decomposable from many other things...

0:19:40.0 SF: Exist kind of in isolation, even though it feeds into the marriage market and income and equality and so on and so... But there are so many other kind of equilibrium states that their stability is not just because people repeat them or they don't have incentive to change, but because they are so interrelated to other social norms and social practices or laws or whatever, that even if you change them overnight, other social practices reproduce or recreate this kind of phenomena all over again. I'm coding this from Elizabeth Anderson who has this example of school segregation, that in some measures, at least in some parts of the country, school segregation is worse than what it was in 1960s, not because the law is generating it, but because the law or the practice of doing that was interrelated to so many other things like taxing for houses, how funding for schools come about, where people live, how... Many, many other things that... And then when you take the law out of the equation or make it even illegal to do that, impose a cost on people who want to segregate schools, other factors recreate and reintroduce the very same phenomena all over again. So that kind of punctuated equilibrium is not gonna help us completely understand how you can get out of the situation. At the end, it is a punctuated equilibrium.

0:21:15.3 SC: Yeah.

0:21:16.4 SF: But when you take account of those interaction, it seems like you are... If you think of a landscape of the choices people make, it's like creating a path, in that... A new path in that landscape that allow people to come and use that path to get to another point.

0:21:35.5 SC: Elizabeth Anderson, of course another former Mindscape guest. I do appreciate you name-checking all the former guests. Now, that's very interesting because, just to back up, you used the word "landscape," that's a word that is used sometimes by evolutionary biologists, although I know that others worry about it, that they always worry, people worry. And the whole issue there is that, for the biological case, for natural selection, it's not teleological, it's just the mutations are random. And if you are separated by a barrier from, or a valley, I suppose, the biologists would say, from an even better equilibrium, it might be very difficult to get there. I think what you're pointing at is that we humans are supposed to be better than that. [chuckle] We can see that there's a better state there, but they're still a collective action problem. Maybe it would be better if we all went there. But if one of us goes there, it's still bad. What can we do about that?

0:22:35.5 SF: Lovely question. I think the collective action problem is serious. I agree that it relies at least in its traditional form, it relies on very restrictive assumptions in terms of self-interest, rationality, perfecting from and so forth. But even when you drop those things, you see that in biological systems, we don't have any of that, you see similar kind of phenomena. And at the same time, you see that groups do have problems to act collectively. If you just think that, "Oh, we are not all rational, or we have other altruistic motives," still there should be some mechanism to explain why we fail to do things, that even collectively we realize that they're beneficial. So, you know that this alternative is good for us, but we have trouble getting there. And sometimes that comes with lack of assurance. I'm not sure that if I go, everyone else will follow and this is costly for me. For instance, if I don't do... I don't bind my child's foot and everyone else keeps doing what they're doing, and the age passes, well, my child has to pay the price for this.

0:23:56.7 SF: If I want to be self-interested, I'm better off doing something that I actually don't wanna do, namely, binding my foot. But the way that we do it in ways that doesn't require... Like you mentioned, that we are better than biological organisms, I think it's partly true. But at the same time, the complexity of the problem that we are trying to solve is very high, so it has many variables and these variables are highly interdependent. Regardless of how smart we are, it is possible that we can get it wrong, and it's possible that we try to do something and not everyone in the society will follow, generates counter-movements or backlash or so many other layers of complication that can mess things up. But when you're looking at the history of movements or changes that happen rapidly, not overnight necessarily, but reasonably fast, in a decade or so, you see that public opinion, people's practices, or many things relevant to that change, is that you generate counterpublics.

0:25:06.5 SC: Okay.

0:25:08.8 SF: You generate an environment in which deviating from what's the standard or the equilibrium or the norm is less costly for individuals.

0:25:17.9 SC: Ah.

0:25:20.0 SF: And those counterpublics, you can experiment and see whether this alternate way of living has any plausibility and does it work or not? And then if it works, well, people around you might be motivated to follow and copy. And if you have some other means to destabilize the equilibrium in the rest of the society, then if this is working well, it will bring a lot of people in. I'm saying a lot of abstract things, I'm happy to break it down, but I just wanted to give a general picture of what I'm thinking.

0:26:00.8 SC: No, that's great, because, are we thinking here of the existence of sub-cultures within a diverse society or giving us ways to experiment? I know that in the federal system for a government that's supposed to be what states are supposed to do, individual states experiment, they're the laboratories of democracy. It sometimes works, it sometimes doesn't. But we're all philosophers here, we're thinking about what is conceivable in principle. Is that basically what you have in mind? You have examples?

0:26:30.8 SF: Kind of, but in a more informal way.

0:26:37.3 SC: Okay. Yeah.

0:26:38.4 SF: And a more fluid way, because when you create states, you're also creating very outcome-oriented ways or, you're changing people's incentive when they are participating, that kind of collective thinking. But sometimes these things, this kind of change when it comes at the cultural level or, we want to change practices that are ingrained in everyday people's lives, well, you might need their input. And... It's true.

[laughter]

0:27:17.3 SC: Fair enough.

0:27:18.3 SF: So you cannot just ask them to vote and tell you whether it works or not.

0:27:22.8 SC: So you've gone this far, we've gone this far without actually saying the words "game theory." But it sounds like they're behind some things you were saying.

0:27:32.1 SF: Yes, that's true.

0:27:32.3 SC: We mentioned Cailin O'Connor and she definitely uses game theory. Herb Gintis was also in the podcast, and he does it. When we're talking about these equilibria, these interactions, is it useful too, and do we use game theory to model them?

0:27:50.1 SF: Absolutely.

0:27:51.1 SC: Okay, good.

0:27:51.9 SF: Because the thing here is that we are talking about a lot of interactions, and the way that people need to cooperate with each other in order to achieve certain goals or have some outcome. And game theory is a very helpful way to formalize and simplify those interactions in a way that we can see how these dynamics can work, by gamifying them in a sense. You are thinking that like, well, when we're talking about, let's say, discrimination or group-based disadvantages, what we're talking about is that we are observing that in a consistent way, people are getting the short end of the stick in their social interactions, or there's some accumulation of this now clustering of disadvantage or clustering of some social problems that otherwise these things don't have similar-ish mechanisms to be clustered around the same group of people. And when you want to think about, how is it even possible to be ending up in this kind of situation? Game theory is a very useful way of thinking about the basics. Do you want me to say more about what kind of game-based is this?

0:29:23.5 SC: I do. I do. But I think that everyone has heard the phrase "game theory." Some people know what the prisoner's dilemma is, but for some people might even know what Nash equilibria are and things like that. But yeah, to make a little bit more specific example would be great.

0:29:37.7 SF: So, a game like a prisoner's dilemma is composed of some main components. One is the players who are interacting and for better or worse, we often talk about two players. I think it's a problem.

[laughter]

0:29:58.3 SC: It's a limitation. Yeah.

0:30:00.2 SF: Yeah. But it does the job done for at least a big class of problems that it's helpful to maintain it. We have two players coming to a table and they have strategies that they can choose or they can... They have choices, let's say, for how they want to interact with the other person or the other player. And there are some outcomes that come out of those decisions. But each of them might not know what is the decision that the other person is going to make.

0:30:37.1 SC: Sure.

0:30:37.3 SF: So, the classic example is two prisoners who are in jail, but the police doesn't know whether they actually have committed something or not. So they're playing a trick on them, that like, "If you come... " To each of them, "If you come and tell me what you guys have done, I'll let you go, but the other person will stay here and pay the price by going to jail for a longer time." They're giving both of these actors an incentive to be mean or tell on the other person to gain some personal benefit. Each of them have the choice to defect or to tell the other person and gain the advantage of being free and not worrying about the crime they've committed, or to both cooperate. If they both stay quiet, the police doesn't have actually any evidence. But maybe because they just want to keep them in jail because they don't have any suspect, they'll get a very, very short time, and they can be released.

0:31:49.3 SF: So, what this game is trying to show is that it is possible for us to stay in or be in a situation that the structure of the world, or the world has structured our choices, such that there is a tension between what is good for us, if we could cooperate with each other. Comparing to what is good for us, if we knew what is the decision that the other person is going to make, and what's the effect of this not knowing what the other person's gonna... So, each of them have the incentive to defect, because even though they won't benefit from being completely clean and the police not knowing what they've done, but they're preventing themselves from paying the cost for the other person, so going to jail for 10 years instead of going to jail for six months. So, both of them gonna defect, even though they both had the option to stay in jail only for a month.

0:32:55.9 SC: So they're collectively best off if they both cooperate, but individually, they're both better off if they defect, so sadly they're gonna end up defecting and the world is a terrible, terrible place.

0:33:04.6 SF: Yes. Yes. And when you... Going back to the conversation about the collective action problem, a collective action problem is a similar dynamic but with N people. All of them have this worry that they might do something that commits them to the strategy or this choice, but if other people don't follow suit, they're gonna be the people who are gonna pay the price for it, and the price is gonna be high. And this price can be high enough that will incentivize almost everyone or enough people to not follow suit. Those who are worried about collective action problem say it's practically impossible for this group of people to achieve any collective good. Yeah.

0:33:52.4 SF: The physicist in me wants to think of this like statistical mechanics, and imagine a huge number of people constantly bumping into each other with their strategies at some temperature, which means that their strategies can change over time and things like that, and maybe there are different equilibria depending on the strength of the coupling and the temperature and stuff like that. There must be people doing exactly that. No?

0:34:16.9 SC: Yes, yes. I don't consider myself a physicist, but the undergrad physicist Sahar...

[laughter]

0:34:25.9 SF: No, undergrad level physicist Sahar, is thinking about this dynamic as a very similar picture that you describe, but this caveat that you're not talking about an ideal gas, so individuals are not doing what they think they should do independent of their social connections or interaction with the other. Instead of an ideal gas, you have a bunch of molecules who are kind of tied to one another, but not perfectly, everyone has some significant interaction with the one on the right or the left. When you add this kind of interdependence, the collective action problem changes its structure because then there are ways for people to solve their collective action problem because they have some repeated interaction, because they have some more information about each other, or because through these interactions, they can come up with some strategy that they think of as the norm. It's the salient option now. Going back to O'Connor's episode, she talks about dividing a pie or you start with the pint of ice cream, and the mom who has decided, giving the children this rule that, "If you both claim the same amount or claim less than what is available, you can get what you claim equal or less." But if you claim more than what is available, if it was my mom, I'll throw away the ice cream...

[laughter]

0:36:15.2 SF: So, you both know that I shouldn't do that. When you are giving that kind of option, people have the choice of both claiming half of the ice cream pint or claiming less than what's available to be cautious, or if they have some information about what the other side claim, claim more, but not high enough that there's no way for the other person to get anything. And then they talk about some equilibrium solutions, that either dividing in half or dividing, one dividing, taking much less, one taking much more, will become the evolutionary stable outcome of social interactions that are repeated over time. Some of them might be with the same people, sometimes you don't have to even make that assumption, but just keep that interaction going. So it's not a one-shot prisoner's dilemma, it's a...

0:37:12.4 SF: Yeah.

0:37:13.0 SF: Prisoner's dilemma that happens over time. But the game that I described is not actually a prisoner's dilemma, it's a coordination game, but it's similar enough in the idea. But, all I wanted to say is that the difference between what prediction, the prisoner... No, collective action problem is giving us, which is very sad that, like, "Oh, you guys cannot do anything together done," and the conclusion that, "Well, look at the world, we can do things, so you're missing," is similar to the comparison you made, or some of the comparison I made based on your example of thinking of an ideal gas in which individuals are acting in isolation, comparing to an environment in which interaction is dominant, or the interaction is the one that is doing things.

0:38:03.0 SC: Good. And you can see where philosophical implications come in here a little bit. But again, as usual, correct me if I'm wrong, but it sounds like we can find ourselves in a situation where bad things happen, even if nobody is to blame. [chuckle]

0:38:18.1 SF: That's so true. Yeah.

0:38:20.9 SC: And maybe that changes our philosophical view on how to either think about or even deal with bad situations.

0:38:26.8 SF: That's absolutely true. For instance, O'Connor and Bruner, and a lot of other people use this kind of allusion and game theoretic mechanism to explain how in the same way that we expect fair divisions to become dominant strategy of a population that's trying to divide their resources and go about their days. You can see that unfair divisions become the dominant strategy. Why? Because you can have some structural components, as like Axelrod too, right? The... Robert Axelrod...

0:39:01.6 SC: Oh yeah.

0:39:01.7 SF: The famous book on 'Evolution of Cooperation'. The idea is that, if you have some added layer of information about what your opponent is gonna do, while you can base your strategy on that added information, and that can be sometimes group-based membership, that like, "I treat everyone in my group fairly, but everyone outside of my group unfairly, or I wanna be greedy to them." And then you see that other kind of symmetry breakage can make us end up with an unstable, or no, a stable outcome that is not actually fair or not equal at this in terms of the outcome it generates.

0:39:52.0 SC: But we do have, unlike the prisoner's dilemma, where a crucial feature is the two prisoners can't talk to each other. Right? And...

0:39:57.6 SF: That's true.

0:39:58.8 SC: And so, one of the things that's interesting about the differences between the physical system or the evolutionary genetics and humans, is that we can anticipate and imagine other possible futures, but another one is we can talk to each other. How much is that part of the models when you're making these models that people can try to persuade each other to act in a certain way that helps the collective good?

0:40:28.8 SF: That's a very good point. I think whenever you... Let's assume that in all these conditions people actually can talk to each other, but they don't have any way to hold each other accountable...

0:40:39.8 SC: Right.

0:40:40.9 SF: If the other person lied. Even if you can talk to each other, talking is not gonna do anything because I can come and tell you that, "Give me your watch or give me something relatively expensive, and I promise that I'll bring it back to you tomorrow."

[chuckle]

0:41:00.3 SF: But I can just lie. And in the absence of any other mechanism that help you think that whether or not I am gonna actually stick to my promise, it's the same as no communication.

0:41:12.7 SC: Yeah. Okay.

0:41:13.5 SF: It might give you some information about the other person. But if that information is not giving you a way to hold that person accountable, but that doesn't mean anything. So, yes, you can add as much communication as you want, but it's still, at its face, the same kind of dynamic.

0:41:31.7 SC: But we do interact more than once, so doesn't that add some complication there?

0:41:35.5 SF: Sure. But I think what does a lot of work is what network theories actually help us to see, that sometimes if I'm a complete rando and I'm asking for your watch, the dynamic is very different comparing to a situation that you and I both know Cailin O'Connor, and we both respect her, and I'm worried that if I don't bring back your watch, well, that gonna affect my relationship With Cailin too.

0:42:06.7 SC: Right.

0:42:07.6 SF: And you know that I care about this relationship, and that is enough for you to have some re-enforcement mechanism that will hold me accountable. And that repeated interaction also is a similar dynamic that if I don't bring your watch back to you tomorrow, then next time I'm asking you for something, well, you're just not gonna give it to me. And if that's the case, this worry about the future is a way that you can use to hold me accountable and make me do what I promise. Communication becomes relevant again. But when you are adding by saying that people can communicate is not anything other than just repeated interaction, or being embedded in a network is doing that work. That's why people don't talk about communication as much, and try to offload the...

[overlapping conversation]

0:43:05.9 SC: I guess so. I don't know. I worry, I wonder whether we're not being fair to people or maybe we're being too fair to them, thinking of them as too rationally, self-interested, whereas people might take pride in being honest or being trustworthy or something like that. Right?

0:43:24.2 SF: But you should know that.

0:43:26.2 SC: Yeah. Okay.

0:43:26.4 SF: If I don't, I wouldn't take pride.

[laughter]

0:43:29.0 SF: And that won't come without some other layers of information or some other layers of getting this information that this person not gonna lie regardless of...

0:43:43.6 SC: Fair enough. Okay.

0:43:44.6 SF: What is the situation.

0:43:45.7 SC: Let's be super philosophical here. There's different ways that philosophers have thought about justice, so we've done a wonderful job laying the groundwork of some complexity theory, game theory, network theory, even though we didn't quite use that, but that was involved, and you're saying that we're not an ideal gas. Right?

0:44:01.5 SF: Yeah, yeah, yeah.

0:44:03.8 SC: There are certain connections that exist and certain ones that don't. How is that different than, I don't know, Hobbes and Rousseau, and Locke, or even John Rawls or whatever, the traditional theorists of political justice?

0:44:17.2 SF: Well, I think it's similar in many regards, but different in important ways too. Similar in a sense that, for instance, Hobbes and Rousseau and others want to think about what keeps us a coherent functioning society. And even though, maybe a lot of us are altruistic people, but we have enough people who are not altruistic and might be self-interested that can ruin it for everyone else. And they are thinking about, what are the conditions that we can put in place that we don't have to depend on everyone being altruistic for the society to work? And someone like Hobbes want to say that we have to have a leviathan. And what is that leviathan doing is, it's making sure that whoever wants to deviate has reason to be scared of doing it. That, if I say, I promise to you that I'll bring back your watch and I don't, there's a leviathan, an all-powerful thing...

0:45:27.9 SC: Distinct.

0:45:28.0 SF: That come... Yeah, and hold me accountable, put me to jail and make me pay the fine or whatever. And if that's the case, and you know that the leviathan exist and I know that the leviathan exist, then I don't have an incentive to not bring back your watch. Right?

0:45:44.6 SC: Yep.

0:45:45.3 SF: So, doesn't matter whether I'm a honest person or not, I'm an altruistic person or not, even if I am a very self-interested, selfish self-centered person, still a society would work if you have that kind of powerful entity that people can depend on, to make sure that deviating from what they agreed on has a cost. But at the same time, what someone like Hobbes is missing is that, well, that's not all we need, first. Second, that is too high of a price to pay to make sure that someone brings back my watch, to give all the power to one entity who walks around and make sure that everyone sticks to their promise. And people like Hume or others talk about culture, or other social practices that help us hold each other accountable or cooperate or do things that we need to do together to benefit from the fruits of that cooperation. And allusion and game theory is a way to talk about that kind of cultural practices.

0:47:03.5 SC: And it's interesting to me because it goes back to the persistence and stability question that you first brought up when thinking about complex systems, and so I am asking myself questions about how non-equilibrium steady state systems require low entropy energy, they require free energy, and then they dissipate and make the world a more random place. And in some sense, there's a parallel here, it sounds vaguely like it. I don't know. Is there some equivalent of free energy that we need here? Is there some resource that these stable structures require?

0:47:39.6 SF: Yeah. Think about a society in which everyone actually follows the rules and they don't even need some higher level things like... Or a leviathan or whatever. It's a very stable society, everyone follows the norm. But pandemic happens, I don't know, something happens that disturbs the structure or dynamics that they were relying on very religiously to keep things together, and they don't have any alternative already existing there. Well, the society not gonna last. You need that kind of diversity and variation and this kind of energy that moves around...

0:48:27.4 SC: Yeah. Okay.

0:48:29.8 SF: That keeps things stable and on their foot. And if something happens, like those countercultures we talked about, sometimes they generate change, but a lot of times they help us remain stable. Some norms or practices are irrational to have in many situations. I had this example that I haven't ever thought about this, so it might not work, so bear with me. That, I had a...

0:48:56.2 SC: We'll bear with you.

0:48:57.4 SF: I had a partner who was a hoarder, many years ago, and it was very annoying to me...

[laughter]

0:49:04.5 SF: How much of a hoarder this person is. And for many reasons, he would never be able to find something he needs because he has so much stuff. And a lot... It was expensive to move them around, it was expensive to keep on to them, there was not much utility. But then pandemic happened, you cannot go out and buy whatever you need on demand.

0:49:27.2 SC: Aha.

0:49:29.3 SF: And the hoarding quality...

0:49:31.3 SC: The revenge of the hoarder. Yeah.

0:49:33.2 SF: Was so good. We had balls of rubber bands that we used to make masks.

[laughter]

0:49:40.8 SF: I never thought having that many rubber bands would be even useful. And we could just even give it to others who didn't have rubber bands...

0:49:47.7 SC: Sure.

0:49:49.1 SF: To make...

0:49:50.0 SC: Basically, you're making an argument for diversity in some sense, it's not just moral good to be diverse in the cause of being fair to people. But you're saying if you wanna be the best society and find the right solutions, sometimes you might do something that might seem to you to be sub-optimal, you need to have a mixed strategy in game theory terms.

0:50:18.0 SF: Mm-hmm. But at the same time, if you have just a closed society, that closed society is often have a lot of conformity pressure. A closed society over time makes everyone look the same, because of the social pressures of doing so. But an open society that has an influx of people who are coming from different parts of the world, different cultures or whatever, is a way to keep this kind of stability. That, sometimes that is just the influx of information or different ways of living, some other times it is just having actual people who are coming to generate that kind of diversity. You don't get that kind of diversity over a long time, at least in a way that, game theoretically we can think about societies, and maintain your diversity, unless you are very intentional about keeping things separate. Yeah.

0:51:21.7 SC: It actually reminds me of the foot binding example, in the sense that you can have a strategy or a norm or customer, whatever, that was optimal under some conditions. The conditions go away, but you're left with the strategy, and if you don't have this influx of other people, maybe you forget how to justify, you forget the reasons why you were doing it that way.

0:51:45.8 SF: No, fair, fair, fair. Yeah, yeah, yeah. No, no, I really like the analogy. Yeah.

0:51:49.4 SC: And so it leads to... Are there... Let me ask it this way. I can see this in the terms of the game theory network, very complexity picture, this interest in diversity and constantly probing our weaknesses and etcetera. Is that new and different in this discussions of society and how to set them up? Is that an insight that we got from thinking in these ways?

0:52:15.9 SF: Well, I think... By these ways you mean like focusing on complexity?

0:52:20.0 SC: Yeah.

0:52:21.5 SF: Okay. The way that I think complexity distinguishes or thinking from the standpoint of complexity distinguishes itself from other viewpoints is that, if we, again, go back, you can tell how much of a fan of Cailin O'Connor I am. I... [laughter]

0:52:38.6 SC: That's fine.

0:52:38.7 SF: Constantly go back to her. But the evolutionary ways of talking about social norms is at the end of the day, trying to find some stable outcome. Right?

0:52:51.4 SC: Okay.

0:52:53.3 SF: So it's a stability-based kind of explanation, that gives you some insight of how that can be possible or how it... The complexity approach is helping us to not take that stability to be our end goal, and think about how change can happen. Not only can happen, but it happens all the time. And our ability to explain things is not dependent on finding some fixed features of the world. I don't have to offload the robustness of the causal connections I find in the world onto unchanging things. And if I can do that, I can show that how I can, perhaps, guide the changes that are already happening [chuckle] to a direction that are in a better... Help us to get to a better outcome, without assuming that I know what is this better outcome. Right?

0:54:01.8 SC: Yeah. And it reminds me once again of Elizabeth Anderson, who was the one who...

0:54:07.0 SF: That's true.

0:54:08.1 SC: Convinced me for the first time in my life that ideal theory was not the best way to think about...

0:54:12.6 SF: Absolutely.

0:54:13.3 SC: Political justice, rather than...

0:54:14.1 SF: That's true.

0:54:15.0 SC: To me, as a physicist maybe, it always made sense, just think about the perfect society and work to try to get in that direction. And she points out, number one, you don't know what the perfect society is, and number two, conditions are changing. But it's much easier just to try to improve a little bit. So asking about the local derivative is a more sensible thing than asking about the final goal.

0:54:35.6 SF: Absolutely. And if you want to improve things, and you started with this assumption that the problem that we're solving is too complex for any of us, or for most of us, or even for, I don't know, an institution that it's responsible for gathering information, to have a bird eye view of this landscape to tell us what is the better outcome or what is the better equilibrium. We want to know how possibly we can make improvements. And the way to think of that, about that is to think about, first, how knowledge is produced, how these practices come to existence, how are they generating outcomes that are sub-optimal in ways that the evolutionary models help us to understand, but also how we can minimize the error in finding the alternatives. Let's go back to the landscape kind of metaphor that, you see that in scientific communities too.

0:55:38.4 SC: Sure.

0:55:39.8 SF: There are different ways of exploring this unknown in the rugged epistemic landscape, and it has a lot of local optima that makes you think that, "Oh. We found the best possible solution, that," I don't know, we have flow just in or whatever...

[laughter]

0:55:58.3 SF: [0:55:58.3] ____ That's very good. But you want people to explore alternatives, and find whether there is a better way to go about solving this question or not. And people model this as like a, let's say a group of scientists who have different kind of strategies, some of them want to find what is... Or follow what is the consensus, and to stick with it. Some of them want to find what's the consensus and do the opposite, because they felt like it. And some of them don't care about anyone else and just focus on their own work, and they want to be a good scientist that proves everything from the basic axioms to the conclusion that they wanna drive. And they show that when you have this kind of dynamic and let people follow or learn from each other, exploring this epistemic landscape is not something that you can plan and tell people that like, "Look, this strategy is bad, this strategy is good," even though intuitively, doing something that the scientific community says, "Well, we have consensus over this" and do the opposite is not a promising way. But they show that having people who do that help us to maximize our ability to reach out and find what is the better equilibriums or what are the better answers or better outcomes? And the structure of who is talking to who is another important way to think about it. So those models...

0:57:41.0 SF: Don't think about that. I have some other work. I'm trying to make sense of, when you compare chapter-based movements, for instance, to hierarchical movement, chapter-based movements try to figure out how they solve a problem locally...

0:57:56.4 SC: Okay.

0:57:57.0 SF: In their own, I don't know, neighborhood or city or state or whatever. But they have ways to communicate with other chapters, and ask them, "How well you're doing, what did you do that worked? How can I not waste important resources to reinvent the wheels?" And if they try something, it doesn't work, they can tell others, but at the same time, they take into account that something that works for me in Columbus, Ohio, might not work in Baltimore, and might not work in Arizona or whatever. And then, they are changing the structure of their communication or their collective action, as a way to minimize their chance of making mistakes and maximize their chance of finding solutions that work for them. Not assuming that the solution even works for everyone, or there is a way for everyone to do. So, it's a way of incorporating that situated knowledge, knowing that we don't know what is the alternative necessarily that we can get to, we know that along the way, we might figure out that we were wrong all along. Some people make this argument in terms of the feminist resistance to some conceptions of a family, a nuclear family, and then this change in people's attitude when same-sex marriage became legalized, and the tension between these two ways of thinking about what is the right way of solving the problems that people were facing. And realizing that, well, we might have been wrong.

[laughter]

0:59:47.2 SF: In fact, it's too costly to ask people to not live in nuclear families, given everything else. Maybe in an alternative world, it would have worked, but there is a better outcome that we can seek, and it can improve the lives of many who we care about. And otherwise, they would be in trouble or not benefit from those outcomes.

1:00:09.3 SC: But this example of science or of academia more broadly, reminds us that diversity can be hard to achieve. If you have a certain academic subfield that has different approaches to the problem, and people think, well, a certain approach is 80% likely to be right. But you're only hiring people in that subfield once every 10 years. You're gonna hire one of those consensus people 100% of the time, and you're not gonna give 20%, you don't have room.

1:00:37.8 SF: That's true.

1:00:39.5 SC: So, you need to make some extra special effort to nurture the diversity in that sense.

1:00:45.1 SF: That is absolutely true. And when you, again, look at scientific community, you see that when there is a critical mass who is kind of negating the dominant or the mainstream way of thinking about things, they often start generating their own journals, they're creating departments that is more likely to hire. So, hire people who have this kind of tendency because that's the only way that you can protect these more minority groups. And I'm not saying minority to say being minority here is always right...

1:01:19.8 SC: Sure.

1:01:20.7 SF: Or being minority here is... I'm just saying, it's good to preserve the diversity here, and the pressure for conformity is real, and you want ways to allow this minority to persist, just in case the majority was wrong. So you have some people who have explored the alternative and you can borrow from them. But they create their own enclaves that helps them to survive or even be successful in their own smaller community. But it also protects them from the pressure that they are feeling from the mainstream. And in social movements or social change literature, you see a lot of that kind of dynamic that people generate groups to protect them from the cost of deviation, but at the same time amplifies their voice, so they have a chance to be heard by the rest.

1:02:19.7 SF: I love this example of Rosa Parks, that if you're thinking that she is just a tired person who didn't want to change her seat because she is sick of the racist rules and practices in her town. Well, you're missing... That's partly true, but you're missing a lot of important information about what actually helps to make an action like hers to be a milestone or a significant driver of something like a civil rights movement. That she is a local organizer of NAACP, her husband is a youth organizer for the same organization. She has been thinking about doing this for a long time. They have tried other people who wanted to not change their seat and see what happens. And she had an army of other organizers and activists who, when that happened, reassured her that the cost is not something that she's the only one who bears it. It's not like she will lose her job and not gonna have any support whatsoever, and she's alone in bearing the consequences, but also they amplified her voice.

1:03:46.7 SC: Right. They were primed to help out to support... Yeah.

1:03:49.7 SF: Yeah. Yeah, yeah. And she's not the first person who has ever decided to change, not change her seat or not follow the racist laws or practices in her hometown. The difference is that, the connections that she has, and the counter-community that protects her from just doing something that makes her go to jail and lose her job, as it would be the case for everyone else. And the same, I think, goes for the scientific...

1:04:21.2 SC: Yeah, but the Rosa Parks example is a good one because this was the last thing I wanted to talk about, we've segued perfectly to the end of podcast here...

1:04:28.2 SF: Okay. Good.

1:04:29.4 SC: Because you write a lot about social movements and their importance, and here is a perfect example of one. And, someone who hadn't thought about this very carefully might just say, "Why do we need a social movement? We have a democracy, everyone's gonna vote, they're gonna get what they want." But these collective behavior kind of issues really do matter. Even if we live in a perfect democracy, which we don't, but that's a whole another podcast.

1:04:53.7 SF: Yeah, yeah, yeah.

1:04:55.5 SC: But even if we did this kind of coordination, game theory, network theory kind of thing, becomes really important to making the society you live in closer to its own ideals.

1:05:06.6 SF: That is true, because when you're thinking about a democratic society, one of the distinctive features is that people have roughly equal say in how things go, so it's a self-governing society. But it might be the case in an ideal gas situation...

[laughter]

1:05:26.9 SC: Yeah.

1:05:28.1 SF: That each molecule has the same effect on the temperature, roughly speaking. But we don't live in an ideal gas situation, and we have so many other mechanisms to first make collective action possible, but also generate power, generate influence, generate mechanisms that can advantage some people and disadvantage others. But the very same kind of mechanisms can be used to amplify change, it's like you are increasing the number of connections between people. So, something that would happen to someone and no one else would hear about because that person is very disconnected or have very poor resources or whatever. When you're talking about social movements, you are in a, this, heightened level that everyone is more or less prime to think about this problem first, but second, they might have extra connections that make them hear about this kind of news. When a shooting happens or a killing happens to a member of a population or by a member of a group, everyone else gonna know about it in a way that it wouldn't happen before. So the networks can help us explain how change is way more likely and stability is... Or the basin of attraction is shrunk, because people are more connected, because people are more prime even to think about the lack of connections that they had to a problem, or the lack of connections that would prevent them from hearing about some phenomena.

1:07:18.4 SC: So if I wanted to start my own social movement, would insights from complexity theory and evolutionary game theory help me build a more effective social movement, or is it more describing things after the fact?

1:07:31.5 SF: Lovely question. I think my conclusion or what I've learned from social or complexity theory is not a recipe for starting a social movement, it's a recipe for knowing that when change is going to happen, I have an influence way beyond what I thought I have, because I'm not just myself, I'm also an extended network of people who know me or they might take what I say, with more confidence or accept it with more confidence than they would at someone else. And if I am at the state that like a movement, like a wave is approaching me, it's not like I can just disconnect if I am feeding into it, I can amplify it by a lot. Or I can be, I wait for it to dampen it down. My conclusion from complexity theory is that, first, Sahar, start rethinking...

[chuckle]

1:08:42.1 SF: What you think you as an individual can do, and stop thinking of yourself as a molecule and an ideal gas.

[laughter]

1:08:53.5 SF: And that the addition of your action with a bunch of others will generate change, because I can have cascading effects. Anyone can have cascading effect with the right kind of... But also, if you want to start change, mind your connections. Go find others who are like-minded, support them, be supported by them. And know that without that, not much gonna come out of...

1:09:20.7 SC: Well, you just gave me a little epiphany here, so I very much appreciate this because, of course, physicists love to talk about phase transitions, which is highly analogous, if not exactly the same as social transitions of various sorts.

1:09:35.1 SF: Exactly. Yeah.

1:09:39.3 SC: But to stand up for my physicist friends, we don't only think about ideal gases, sometimes we think about a lattice or a solid, which are very primitive network kind of models. And something you notice when you have a lattice that it has a physical system on it that's going to do a phase transition, is that far away from the phase transition, correlations between what's going on at nearby lattice points might be either only infinite range or only short range, it's exactly at the critical point, at the phase transition that you have both, short-range correlation and long-range correlation. And you just explained why that's important to a social movement, so I think that's great.

1:10:19.2 SF: You're welcome.

[laughter]

1:10:23.5 SC: Whether you planned it or not, you have very much helped me launch my social movement, when I decided to do.

1:10:27.4 SF: Good. Good.

[laughter]

1:10:30.8 SC: Sahar Heydari Fard, thanks so much for being on the Mindscape podcast.

1:10:32.7 SF: Thank you so much for having me. It was a delight.

[music]