Human relationships are often described in the language of “chemistry” — does that make the beliefs and attitudes of individuals a kind of “physics”? It is, at least, a fascinating avenue of inquiry. In particular, the field of statistical mechanics offers potent tools for understanding how exactly people form their views and change their minds. From this perspective, everyone is a dynamic network of opinions and values, in a tense and ever-changing balance both with others and ourselves. The “chemistry” of social life, then, arises from multilevel interactions in our noisy minds and how they influence each other.
Welcome to Complexity, the official podcast of the Santa Fe Institute. I’m your host, Michael Garfield, and every other week we’ll bring you with us for far-ranging conversations with our worldwide network of rigorous researchers developing new frameworks to explain the deepest mysteries of the universe.
In this conversation, we speak with SFI Postdoc Jonas Dalege about how his research uses physics models to understand the emergence of higher-level behaviors from lower-level behaviors, both within and between people. We discuss the role of entropy in the formation of individual beliefs; statistical approaches to the study of ambivalence and cognitive dissonance; the wisdom (and challenge) of tolerating ambiguity; and the social consequences when we try to minimize internal conflict…
If you value our research and communication efforts, please rate and review us at Apple Podcasts, and/or consider making a donation at santafe.edu/podcastgive. You can find numerous other ways to engage with us at santafe.edu/engage. Thank you for listening!
Join our Facebook discussion group to meet like minds and talk about each episode.
Podcast theme music by Mitch Mignano.
Follow us on social media:
Twitter • YouTube • Facebook • Instagram • LinkedIn
Key Links:
Jonas’s Website | Google Scholar Page
Related Papers, Talks, and Complexity Podcast Episodes:
[Video] Explosive Proofs of Mathematical Truths by Simon DeDeo
Coarse-graining as a downward causation mechanism by Jessica Flack
9 - Mirta Galesic on Social Learning & Decision-making
33 - The Future of the Human Climate Niche with Tim Kohler & Marten Scheffer
42 - Carl Bergstrom & Jevin West on Calling Bullshit: The Art of Skepticism in a Data-Driven World
55 - James Evans on Social Computing and Diversity by Design
This transcript was generated by machine at podscribe.ai with help from editor Aaron Leventman. If you would like to volunteer to help us edit transcripts, please email michaelgarfield[at]santafe[dot]edu. Thanks and enjoy!
Jonas Dalege (0s): People want to reduce inconsistency and dissonance in their own belief system, but they also want to reduce dissonance at the social level. And of course, people can differ than in how important they judge these different levels. So for some people that might be more important. Okay. That's their own belief system is very consistent. Other people will be a bit more inclined to agree with the social network and so on. And what I think is also very interesting about this is that there's kind of indications that what happens in a single humans mind is actually quite similar to what happens in a group of humans.
So for example, if a group has kind of an inclined leaning toward a given issue, let's say it's about abortion. And most of the individuals in the group are pro choice. And then they discuss about this. And then after discussing this, people will actually move to being even more pro-choice at the same time, when you ask people to think about the attitudes, the can kind of the same happens. So if I ask someone who is, let's say moderately, pro-choice I think a bit about your attitude. Then after a few minutes, this person will actually move a bit more towards the extreme
Michael Garfield (1m 25s): Human relationships are often described in the language of chemistry. Does that make the beliefs and attitudes of individuals, a kind of physics? It is at least a fascinating Avenue of inquiry. In particular, the field of statistical mechanics offers potent tools for understanding how exactly people form their views and change their minds. From this perspective, everyone is a dynamic network of opinions and values in a tense and ever-changing balance both with others and ourselves.
The chemistry of social life then arises from multi-level interactions in our noisy minds and how they influence each other. Welcome to complexity, the official podcast of the Santa Fe Institute. I'm your host, Michael Garfield, and every other week, we'll bring you with us for far ranging conversations with our worldwide network of rigorous researchers, developing new frameworks, to explain the deepest mysteries of the universe. In this conversation, we speak with SFI post-doc Yonas.
about how his research uses physics models to understand the emergence of higher level behaviors from lower-level behaviors, both within and between people. We discuss the role of entropy in the formation of individual beliefs, statistical approaches to the study of ambivalence and cognitive dissonance, the wisdom and challenge of tolerating ambiguity and the social consequences when we try to minimize internal conflict. If you value our research and communication efforts, please rate and review us at Apple Podcasts or make a donation at santafe.edu/podcastgive.
Give you can find numerous other ways to engage with us at Santafe.edu/engage. Thank you for listening.
Jonas the Lega. It's a pleasure to have you on complexity podcast.
Thank you. It's a pleasure to be here. So I like to start these conversations by humanizing the scientist a little bit here, talking about how you got into your work and what animated or inspired the kinds of questions that you're interested in asking in your research. So give us a little bit of a backstory on your entry into science.
Jonas Dalege (3m 46s): Yeah, sure. I kind of was always very interested in people's beliefs. I know sometimes people are quite a mystery to me and some of the beliefs people have, and it might also sound a bit like a cliche, but also being a German kind of inspired me our history and how people can have these extremely horrible beliefs and then act on them. So that is one of the great mysteries that are in our world. I don't come from a academic background and not from an academic family. So I didn't really know what science was like, but at some point I kind of decide, okay, I'm really interested in this.
I should just study psychology. And that then was what I did. And then I got deeper into these questions. How do people's beliefs form, how also form extreme beliefs. And then during my PhD, I also discovered complexity science aspects to the study of beliefs that we then let me to the research that I'm currently doing, that I also did during my PhD, which is based mostly on the idea that we can understand how beliefs form and change through studying them as networks.
Michael Garfield (4m 47s): I feel like there were a lot of core ideas that we need to define in order to really get into the meat of your work. And I'm one of those would just be the way that you understand attitudes in the first place and beliefs and how you're defining those things formally in a way that's somewhat different from the sort of common use of those terms, not entirely, but you're making some nuanced distinctions here that I think are key to understanding how you actually go about creating models for this stuff.
Jonas Dalege (5m 21s): The definition of attitudes, the general definition that I use in my, in my work that is mostly comes from very classic ideas in mostly social psychology and attitudes, and generally just liking or disliking of an editor object, which can basically be anything. It can be a person, an idea, a product and so on. This is more like this global evaluation, but you also have these more nuanced elements of attitudes, which generally come in the form of beliefs. So for example, if you think of a person that you really like, you might think about the person, this person is a very, very intelligent person, fun to talk to and honest and so on.
But of course there's also more feelings involved in his attitude. So do you really like this person? You have sympathy to what this person, and then there's also of course, behaviors involved. So for example, you enjoy spending time with this person. And this is one of the classic ideas in social psychology that extrudes consists of these three different components of attitudes. And then what my research most focused on is on the interactions between these different elements. Because now it's not the thing that you, you basically just have all these different beliefs, feelings, and behaviors in isolation, but they all influence each other, having one positive belief about the person we're making more likelier to other positive beliefs.
Michael Garfield (6m 36s): We're definitely going to get to that, the network nature of this stuff. But before we do, I think one of the key points to stress in this discussion is the fact that you're bringing in terms and models from statistical mechanics and using those to analyze this. So I think it's worth assuming zero knowledge on the part of our listeners, which is probably deeply unfair, but nonetheless building our ladder at the ground. Let's talk about what it is that you mean when you say attitude and entropy and the different kinds of entropy that figure into this in terms of the way that you just described it about these micro beliefs, leading to a person's attitude about something. That's really key.
Jonas Dalege (7m 19s): Exactly. So as you said, the important point here is that there's basically two levels. And I mean, of course it's a, it's a bit of a simplification because also beliefs can also be a bit more nuanced, more global, but for the second simplicity, we treated now mostly as having two levels. So having these more fine rent beliefs, and then this general attitude, and I think the framework or statistical mechanics then lends very well to this because this is really what I think is statistical. My case is an essence about, about connecting these lower level properties to higher level emerging properties.
And then of course also entropy comes into play. And that is often also one critique I get on my work because people think, okay, entropy that comes from, from physics, from Nesser science, what does it have to do with psychology? But the interesting thing about entropy is that it's a statistical concept. So in this sense, it might be, there's a place also very well to attitudes because it's just much easier actually to have a very inconsistent and unclear attitude. Because I mean, when you think again about the person and there's many different ways, you can be ambivalent about the person, but just basically one way to feel very positive about this person so that you have all these positive beliefs and positive feelings.
And so on. This then directly relates to Boltzmann entropy, which is basically that how many of these micro properties can lead to this global emerging pattern.
Michael Garfield (8m 38s): A couple of examples that you give in this paper, this is the one that you coauthored with Borsune Von Hariveld and VanderMoss in psychological inquiry, the attitudinal entropy framework, as a general theory of individual attitudes. You give a couple of examples. One I like is the slot machine. There's a lot of ways to lose. There's only a few ways you can get three or four or five in a row. And then another one is a little bit more on the psychological end is about snakes, which I think given the amount of work on the formation of political opinions that you and colleagues have done at SFI might be a little on the nose, but there are many, many different reasons to like, or not like someone.
And yet you end up voting for a given candidate along with a lot of people with whom you disagree on most points, but this is all kind of assuming that all of these different beliefs are uncorrelated. And as you already said, it's not that case. You're not equally likely to have two different beliefs about a person. This is where the network model comes in and the way that all of these different beliefs influence one another enters the picture. So can you talk a little bit about that and how you're applying network models to these sort of ecologies of belief.
Jonas Dalege (9m 55s): It's really based on this basic idea that you don't form these beliefs in isolation. When you, for example, you observe a person to act in an honest manner. It's not just the case that you just kind of know this and gonna save it up in your mind and say, "Oh yeah, this person's honest." And of course it primes you to also think about this person other ways. So for example, you will assume, I suppose, this person is probably also caring. It's a nice person. And so on. You might also infer that this person is intelligent and then there's implies that all these different beliefs influences others, but also beliefs each other in different amounts.
For example, judging someone who's honest will have a high impact on judging. This person is caring, but a lesser impact on judging this person that's intelligent. So I think this is really why I think that applying network analysis to attitudes makes a lot of sense because these beliefs form a network and the connections between beliefs are also not trivial. So I think we can actually learn quite a bit from studying attitudes as networks and can, for example, see which belief is most central to a person's belief system, which can also give to answer information, okay, how might these beliefs be checked
Michael Garfield (10m 58s): In this paper that you and your coauthors use, The Icing Model, which I think most fans of complex systems research know from inquiry into the alignment of magnetic particles, but here you're doing something a little different with it. And I'd like to link that to a statement that we kind of glossed on the way here, which is what it is that a person's brain is doing with inconsistent beliefs and how the macro state of a person's attitude is more or less unstable due to inconsistencies in the micro States of their beliefs.
There's a trend or a bias in us towards seeking consistency in the same way that, you know, you put two magnetic North ends together and you know, the magnets flip, they reorient according to the influence, that's a real, simple, naive explanation of this. And I'd love to hear you go into a little more depth about the way that the icing model is being used in this and the way that you might consider a person's cognition as seeking the lower entropy macro state, if you will.
Jonas Dalege (12m 9s): Yep. Yep. And upon point to add to this is that it would also depend on the circumstances how much you are motivated to seek this consistency. So for example, if the attitudes or the belief system is very important to you, for example, you have to make a decision based on this. You have to decide between the houses you want to buy. Then you are very motivated to arrive at a decision and arriving at this decision will be much easier if you have this somewhat consistent attitude. And on the other hand, if you don't really care about a given topic, maybe you are, for example, if you are not interested in politics at all, then you will probably be fine with having inconsistent attitudes towards given petitions.
We use this icing model to autumn model, these differences in the icing model in general, and Cisco physics, you have this idea of temperature, which actually will influence how strongest drive towards consistency is. So basically if you really care about something, this cannot be modeled using the Icing model with having a low temperature attitude, which will then go to a more extreme attitude. But if you don't really care, you come all this go for high temperature attitudes, which then is a bit more all over the place and will also likely not have that much impact, for example, on your behavior. Also relating to this there's pretty cool research actually by Dan Simon of the university of Southern California, where he gives people a few facts about the legal case. And then the first thing that people don't know that these effects are related and they just read them. And then at some point they learn, okay, this is about the legal case. And you have to decide whether the person is guilty or not. And then all these different facts of, for example, that this person was seen at the scene of the crime and that maybe it was a robbery or something, and that this person has financial problems.
These different statements then become correlated in people's minds if they belong to the same case, which really underlines the point that you want to come to a decision, all the different aspects of the system become dependent on each other.
Michael Garfield (14m 6s): This is a, maybe a bit of a Baroque constellation here. I'm going to try to put together, but it's listening to you speak about all this stuff, just to take a dip out for a moment of this paper and propose some links to other work. I'm reminded of Emory university human development, professor James Fowler, who wrote a book in 1981 called Stages of Faith, where he was looking at child and adult psychological development, and how as a person learns to take the structures of their mind as objects as they grow older and wiser, he basically looked at this massive survey of interviews and survey information, and he basically described wisdom as the tolerance for ambiguity.
And if you think about development in terms of moving the subject of one's identity into the position of an object, you learn to take your beliefs as objects and manipulate them through metacognition. These become less imminent and less important to you. You gain distance from them. And so you're allowed to remain in a position of unknowing about them. And so to link that back into SFI research, I'm reminded of the conversation I had with Tim Koehler and Martin Scheffer about climate change as a collective action problem and how this is something that is so abstract to many, many people has been abstracted, even in cases where the extreme weather events are urgent, palpable and immediate.
It's really unclear how we would all align in a position that results in concerted action towards one strategy or one outcome over another. It's interesting because what determines the stakes and what determines therefore, the urgency that a person feels to settle into one state or another in this model seems sometimes to have to do with wisdom and maturity and sometimes to do with ignorance and or misaligned incentives.
So I don't know what your thoughts are on all of that, but it just seemed worth mentioning that these are two cases where it seems like your model is very easily applied, but they kind of point in different directions.
Jonas Dalege (16m 32s): This quote, what was it? "A wisdom is the tolerance of ambiguity." I think that's very good quote, and I think that's also a problem. I think there's this general tendency. I think it's, it's very human thing to do. It might almost be automatic to reduce ambiguity and arrive at a more simple representation than what the word actually is. It's also that we have a lot of pressure, I think, in, in our society that you should have very clear opinions. It's often kind of consider that a bad sign. If you say, "Oh, I don't know."
I mean, for many, many topics, your answer actually should be. "Yeah, I don't really know." I mean, that's also what drives science acknowledging that for the most part, we just don't know. And I think people in general don't really like acknowledging this, but yeah, I think it often really helps then also going through issues like climate change it's portrayed and of course rightfully it as a very important issue, but this will also then motivate people to reduce the ambiguity in their belief system even more. And this will also then induce resistance actually so that some people will just reduce a bigger team by thinking, okay, climate change is just not real.
And so I think it can often cause problems if we make a issue too important, which of course, it's a bit of a conundrum there because I mean, it's, it is very important, but there's a fine line that you have to walk there.
Michael Garfield (17m 48s): Certainly as with so many in the complex systems world, it seems like we're really seeking is a kind of Aristotelian golden mean between two virtues that are horribly destructive if taken to the extreme. And in that sense, another piece I'd love to link to your work is the piece that as a researchers, Vicky Yang, Tamara vendor dose, and Henrik Olson just published in PLS one today, falling through the cracks.
Jonas Dalege: I didn't know that.
Michael Garfield: Yeah. Modeling the formation of social category boundaries. So like what you and your coauthors are saying in this piece is that it's psychologically uncomfortable to remain in this sort of gaseous state about your convictions and that this is not just true at the individual level, but this is true when it comes to the formation of identity. And therefore it's a dynamic that makes for these overly simple voting blocks and political categories and categories of identity where moderates and independence and non-binary gendered individuals. And so on just sort of disappear. I'm curious, based on this model of yours, what hope you think we have for creating a space where more nuance and more categories is not such a gruesome thing to entertain, is not such a difficult thing to actually hold. This is where we can get into some of the other implications and predictions of your model about what happens when we try to manipulate the weights and a network model of attitudes.
Jonas Dalege (19m 33s): First of all, I think it's important to actually acknowledge that, that it's quite a natural drive for humans to make complex issues more simple, but also by acknowledging this, I think this also then leads us to realize that we shouldn't make this even more pronounced, right? So for example, having a two party system where people actually even have to register with a party, you will just fuel this natural tendency. You already have to reduce your inconsistency to be able to vote. No political party or no political candidate will exactly be what you want.
So there must also be already some reduction of this inconsistency there, but then having people be reminded all the time. Okay. Yeah. You're either a Democrat or Republican really must make this much more pronounced, right? So that's basically the whole time you will be even more, less tolerant of this complexity. So I think that would be important to acknowledge this and then derive strategies, how we can actually deal with this. And I think it's it's of course not as easy to just move from a two party system to a multi party system, but trying to reduce this identification would I think help already, at least a bit,
Michael Garfield (20m 40s): One of the insights that spills out of this model has to do with why it is and you of hinted at this just a moment ago, why it is that it is so incredibly difficult to persuade people with evidence and why attempts to persuade people so often backfire. And so, you know, I'd love to ground this in something that hopefully helps people navigate their personal relationships, their organizational bureaucracy is the work that they do as lobbyists and activists. Why is this happening, that our persuasion so often backfires? And then what can we do about that?
Jonas Dalege (21m 26s): Pretty direct implication of our model is that what basically happens, if you have a discussion with a person you don't agree with, there's probably two things going on. Let's assume this person will listen to you. So it's not you, you don't hate each other, but you might be colleagues or friends that you just have different opinions talking this. This will have basically two consequences. So first of all, you will communicate some information you have to this other person. And this other person will also, at least to some extent will receive this information. But what will also happen is that because of this conversation, the other person will of course direct attention to the issue.
And this in our model directly leads you to your beliefs, becoming more correlated and going to more extremes. So what then can happen is even if this person hears the information from your side, it might actually be that increase in attention and importance will actually move this person further away from you because this person's existing beliefs become more correlated and therefore this person becomes more extreme in the other direction. And so this already is kind of the default thing, what will happen. But again, they're evaluating this, that this will happen already implies. You don't want to accelerate this.
And I think for example, if you would tell a person that basically, well, you're wrong, this will basically just lead this person to become even more entrenched in his or her former beliefs. And so basically having an open mind and signaling, yeah, I have an open mind also for what would you think would probably reduce this tendency at least somewhat, and then maybe you can actually really have a good discussion and actually exchange information instead of just entrenching your beliefs even more,
Michael Garfield (23m 2s): Again, to link this to some other SFI research I'm reminded of Simon Dayo, his presentation on explosive proofs of mathematical truths, rather than looking at a mathematical proof as a linear, if then series of arguments that if you map it as a network, that you can actually kick a lot of the legs out from under that table without the table falling over. One of the predictions that you just spoke to in this model is that if you can reduce the dependence parameter of a network, that's actually what allows for a less stable, a less resistant attitude.
And so, you know, I'm thinking about this in terms of, you know, a tried and true approach to formal debate, which is again, not to attack the resolution itself, not to attack the macro state and not to attack the components of it that are the most stable, but to look for little ways to erode and undercut this piece by piece and thereby differentiating for your debate opponent in the way that, you know, Simon talks about this, that these are not all sort of logically consequential from one another.
And you make this distinction in, in, in this paper also that there's a difference between a discreet shift in a person's attitude and a continuous shift. You don't necessarily have to change a person's mind entirely. You can just move them a little closer to the center of the argument. Does this analogy hold?
Jonas Dalege (24m 40s): Yeah, I think so. But of course also it relates again, to what extent this attitude is then very important to this person because having these discrete shifts happening that is something that's much more likely to happen if it's very important and this more continuous shift, it's more likely if the person is not too involved in this issue. So basically having debates strategy where you're at at the same time, don't make this argument too involved, but also getting your argument across will probably be the most effective. But of course, it's also, it's also very difficult to actually do that,
Michael Garfield (25m 11s): Even though I feel like you've touched on this already. I'd like to specifically highlight ambivalence and cognitive dissonance. And in particular you differentiate between attitudinal ambivalence and felt ambivalence. And also, I, I know, so, so many people in the modern world that are suffering from really profound cognitive dissonance at this time. So yeah, just, I think, you know, just again, grounding your model and its predictions and implications in something that helps people get a better sense for why it is that this is such a, an epidemic experience in this time.
I'd love to hear your thoughts on that. Is it simply the fact that we are exposed to so many different perspectives in a global social networks or what is going on and what, if any bomb can we apply to these wounds that isn't just about entrenching ourselves in some sort of retrogressive future shock kind of move into these counterfactual positions, just to feel a little bit more united within ourselves?
Jonas Dalege (26m 22s): I, I would think that it might be actually two reasons. Of course we have much more information about all different things than we used to have, but at the same time, especially in the political arena, it's also, there's this much, much more pressure actually to be involved in this with the last president of the U S constantly being on Twitter. I mean, it's people getting reminded all the time. Okay. There's on the one side, people who really, really liked this president, but they also remind of this constantly. And then there were other people who really disliked this president and they were also constantly reminded of this.
And so at the same time, you have this extreme flow of information. And on the other hand, you also kind of all the time pressured to make up your mind. And I think this is kind of a toxic combination because then you would just kind of have to really simplify all the information you get. And instead of just accepting, okay, there's just a lot of different things going on. And for some things I just don't know. Actually, after the election of 2016, Stephen called Bayer, actually put this very well, how we should deal with it. And he said something like informed, yes, but not in it all the time.
And I think this is something that we should actually approach these issues because in the last years, I think everyone was basically in it all the time and it's just too much. And I think this would actually reduce the likelihood that people will also accept with different our opinions. And that is fine. But I also want to stress that I think polarization, for example, is not always a bad thing. So for example, the turnout in the last election was the highest turnout in, I think, 60 years or so. And I think their polarization actually had a very positive impact, but then of course we don't want people hating other people that disagree with them.Disagreeing is great and fine, but hate of course is not the way to go.
Michael Garfield (28m 3s): You've just reminded me of a couple different things. One of which is a paper I love bringing up whenever people are talking about world peace, which is a paper David Krakauer and Jessica Flack coauthored with Eleanor Brush in science advances, conflicts of interest, improve collective computation of adaptive social structures. And so, yeah, this is one of those sort of awkward realities, if you will, that comes out of complex systems research, which is, if everyone's looking in the same direction, then that's actually an extremely fragile, brittle attention ecology.
You're going to get hit in the back of the head with the Dodge ball, all of you. Whereas finding ways to accommodate a certain amount of polarization, actually results in a smarter society over the long-term. And I, you know, I was just talking about this on the show with James Evans a couple episodes ago, where he was talking about the overwhelming evidence, that if you look at the, you know, the science of science, and again, this is just maybe an advertisement for the kind of non disciplinarity going on at SFI, but the incentives of science, they tend to lead to these silos, which are extremely focused and therefore missing important things.
And that massive scientific advancements it's like discreet shifts, scientific revolutions tend to come from naive outsiders that are willing to call to attention, something on the periphery of a given field. We spoke about this on the show with Carl Bergstrom and Jevon West about the importance of calling bullshit, the importance of finding ways to accommodate in our societies and in our organizations, the child that can call out the emperor for wearing no clothing.
There's not really a question attached to that, but it just seems to me, it just seems like there is this sense in which as awkward as it is, it is good for us to accept some friction for all of these reasons, which include reasons like leaving a little slack in workday, evidence that taking some time off from a particular question, leads to a creative solution.
Jonas Dalege (30m 17s): Great idea. Also, when we apply this model to creativity, the more you think about the issue, the more you will follow a given path. In some cases, this will be fine if you're on the right path, but sometimes you just have to take a step back and then just taking a walk outside might actually lead into having a fresh mind and then taking this problem from a different side and also on a more societal or a more on a group level. I mean, what you just said. I mean, that's also a bit about Koons writing about revolutions in science. If science discipline and science is on the right path, then this given paradigm works quite well.
Right? So then you can make a lot of discourage that, but at some point, of course, you hit an end and that you need this outside influence and really something or someone to actually shake this up a bit so that you, that also the discipline as a, as a whole can actually tackle this other problems from a different angle. I think this is definitely one of the things that makes science so successful, actually that it's not only people who are very at the core of a given field. Not only they can influence the speed, but also someone coming from a different field or a young person's just coming in with new ideas. The test is given sense of anarchy insights.I think it's really makes it much stronger than if we would just have this more hierarchical structure.
Michael Garfield (31m 26s): To link that into some of the insights from evolutionary biology that we discussed with David Krakauer in episode 29 on mass extinctions. There's a part of the transmission series when we were talking about system level shocks and the strategies, you know, higher, low mutation rates viruses are basically practicing what David called a high beta investment strategy because they're changing the context all the time. They're constantly being sneezed and do a different organism.
They have a much more mutable identity. It calls to mind this question about the proliferation of cognitive dissonance in the 21st century. Would you agree that the rate of change and the sort of flexibility of a person's attitude, do you see it sort of obeying the same dynamics that, you know, a larger society plays by these rules of like the genetics of a large continental population versus a small Island population, then that it's easier to change your mind in a smaller community because you're trying to peg your beliefs and behaviors to a smaller number of people.
Whereas on the mainland, you can get stuck in a sort of suboptimal position because you're trying to satisfy too many different constraints at the same time.
Jonas Dalege (32m 50s): That's also a nice extension actually, of what we mostly talk about until now basically of how your own beliefs are related to each other. But of course your own beliefs, don't only depend on your other beliefs you have, but also on the beliefs of the people around you. It's an interesting idea too, if you have this larger society, I mean, my guess would actually be that in the last society, the beliefs, you have follow much more constraints than if you live in a very small community, but I mean, it can also be, of course they have this very small community where everyone is very highly dependent on everyone else.
That of course could also lead you to being better constraint in how you can change your beliefs. But in general, I would think of course, the, the, the people around you put a high constraints on what you actually can believe and also what is actually okay for you to believe or what you think is okay for you to believe. Of course, there's a lot of disagreements in, in our society, but of course these follow a specific game rules. There's of course a lot of things we basically don't even consider. For example, even for something like democracy to work, it has to be that basically our beliefs all have to be constrained to a given level.
Right? And I think that is something that actually society kind of accomplices that people are kind of, okay, you can, you can argue within this, given a set of beliefs, but not outside of that. Yeah.
Michael Garfield (34m 5s): Out of intellectual integrity. I think it's important for us to mention that we're standing directly in the middle of a pre-print that you wrote with Mirta Galesic and Henrik and Tamara and Dan Stein, integrating social and cognitive aspects of belief dynamics towards a unifying framework, as someone who loves the, the powers of 10 multi-scale zoom that allows us to consider things like civilization as a superorganism and so on. I think it'd just be worth acknowledging this particular piece, just to give you the opportunity to go into a little bit more detail about that piece that we haven't already addressed.
Jonas Dalege (34m 45s): Let's get also what this piece is about is it's also, it's very late to what my main project here at SFI is actually, and that is combining these individual or, or more personal beliefs that you have, and the dynamics of this, which my work has until now mostly focused on with more social belief dynamics, basically with the idea, okay, if there's people at the same time, want to reduce inconsistency and dissonance in their own belief system, but they also want to reduce dissonance at the social level. And of course, people can differ then in how important they judge these different levels.
So for some, some people that might be more important, okay, that's their own beliefs dismissed, very consistent. Other people will be a bit more inclined to agree with the social network and so on. And what I think is also very interesting about this is that quite a lot of indications that actually what happens in a single humans mind is actually quite similar to happens in a group of humans. So for example, there's a lot of classic studies on group polarization. So basically if a group has kind of an inclined leaning toward a given issue, let's say it's about abortion.
And most of the individuals in the group are pro-choice and then they discuss about this. And then after discussing this, people will actually move to being even more pro choice. And at the same time, there's also studies on single individuals. When you ask people to think about the attitudes, the can kind of the same happens. So if I ask someone who is, let's say moderately pro-choice and I asked this person, they have to think a bit about your attitude. Then after a few minutes, this person will actually move a bit more towards the extreme. For me, this is really, really very interesting. It's kind of similar dynamics going on at these different levels.
That is always, I think fascinating also lends us then to integrating these different dynamics into the same model. The paper you just mentioned is actually one of our first steps in doing this, and we're actually working much more on this.
Michael Garfield (36m 34s): Minds, me of Jessica Flax paper on core screening as downward causation and how, you know, if you think about what a society is doing by aggregating and evaluating input from all of these local measurements and, you know, local attitudes about reality at the level of a social creature. That's intimately related to what each of us are doing in terms of modeling everyone else's opinions. And so you end up with the social contract or, you know, to use a sort of less endorsed term, the Holy ghost of a society is something that emerges at the intersections of people. And again, to link that to when we had Mirta on the show in episode nine, and she was talking about how you see this in voting, you were just talking about how people care about what their friends and family think, and that can outweigh your own personal convictions about who you should be voting for, but something that I, you know, I wonder again, in terms of how modern society differs from the kind of societies sub Dunbar number of societies that we evolved in, you know, smaller tribes, clans villages, et cetera.
I think you kind of spoke to this when we were talking about Island biogeography, that affluence, that arises in subpopulations as a consequence of economies of scale, there's something about the way that privilege creates the opportunity for us to care less about each other's opinions and therefore erodes the social contract. And I think us back to the point about what it takes for people to regard a particular subject as urgent and imminent, whereas you talk to people that are like, "Oh, I'm not worried about the collapse of society because I've got enough money that I'll just be able to wall myself off from that."
You can see the shifting in the weights between what I believe in what I believe about what other people believe and how much importance I place on other people's beliefs. And so there's this weird what to me was a counterintuitive consequences of all of that, which is that the more affluent we become as a society, generally, the less we actually function as a society. Does this accord with everything that you've come to?
Jonas Dalege (38m 52s): I definitely agree that there's definitely too dangerous there. I mean, meta actually mentioned this in her podcast. It's much easier to find people that agree with you that used to be right. I mean, if you live in a small village with just a given number of people, of course you have, your options are limited. And now basically you can just go online and find people who agree with you. It's also this reports on people in this QANON movement or whatever you want to call it. That really, they cut all the links to their family members, to their old friends, and just talk to two other QANON members that's that's of course, a point where it gets very dangerous.
And then it also brings me back to what I said earlier. That polarization isn't always a problem, but if it reaches a point where you don't talk to each other anymore and just talk to people who you agree with, then of course at some point society cannot function anymore. And also the problem arises. If you don't agree on facts, then of course you also can't really have functional discussions anymore.
Michael Garfield (39m 52s): You end your piece on attitudinal entropy by facing forward into the questions that remain unanswered, and then also how this model relates to other models of attitude. So I'm giving you a sort of choose your own adventure here about which of those to tackle first. But I think for the sake of feeling complete at the end of this discussion, I'd love to know how you compare this model to other models and then where you are pointed beyond what we've already discussed in this call into the questions that this work has opened for you.
Jonas Dalege (40m 28s): Let's start with the model comparison, which in psychology often it's a bit different from more formal fields because there's much less formal theories in psychology. So it's often a bit difficult to directly compare different models. So I would say the most important thing to go forward is, okay, let's see how these different predictions that we have pan out. And one thing I I'm excited about and want to understand also better is actually extreme shifts in attitudes. Our model actually predicts that generally, if a partner is very important to you, you will be very resistant to change, but at the same time, if you get enough information that you cannot hold onto your original attitudes, you will flip to another extreme.
And there's really quite a lot of anecdotal evidence on this. So for example, there's this German lawyer who used to work for the F, which is used to be a left wing terrorist organization in Germany, mostly during the sixties and seventies, but then he switched at some point to become one of the most outspoken neo-Nazis in Germany. So he's completely switched from the most extreme left to the most extreme, right? Also there's this evidence about people actually losing their faith.
So for example, I think there was this pastor within, at some point for him, it didn't make sense anymore. And then he started this group with which he helps people actually discarding their religion enters on group for becoming atheists. So there's a lot of evidence about us that people actually cannot sometimes change their attitudes quite, quite radically, but there's not that much systematic studies about this. And I think this is really something our model has a lot to say about. And I think it would be really great to test this in a systematic way.
Michael Garfield (42m 11s): Yeah. You know, to bring up the whole thing about the converted atheist, that is just such an interesting topic because it's so clear to so many people that a lot of these, Oh, I used to feel that way. And then I had a revelation kind of personal narratives are really just a shift across the Y axis into an equally extreme, but opposite position that resolves a person's cognitive dissonance that they've been developing about the creator or whatever, but it doesn't actually indicate an evolutionary transition for that person in terms of like their psychological development.
They haven't actually gained any kind of distance from the level of organization at which they were constructing their identity in the way that we were talking about earlier with James Fowler. You give another really interesting anecdotal example of this in the op-ed that you sent me on the proud boys, and I'd love to offer something concrete to people. I think talking about Trump supporters before and after January 6th and the attack on the Capitol is a really clear case that illustrates the insights of your model here.
Jonas Dalege (43m 23s): Yeah. I was really struck by this article in the New York times about how the proud boys actually shifted their attitudes to work Trump in a quite extreme fashion after the storming of the capital, because they felt betrayed by Trump because they didn't receive any parts and so on. And then really a few days after the storming of the Capitol, they started to mock Trump on their social media and really called him weak, a trader, a total failure and so on. And then, I mean, this, this is a group that used to see Trump as the Messiah, almost for them.
Trump was really a very potent, very unifying figure, but then the shift then that they, that they saw them as betraying them. That really shifted their attitudes completely from seeing him as a strong and loyal. And so on it as a great president, they really, they saw him as weak and disloyal and really as a, as a complete failure, as I, as I said, and I think this really illustrates the working of our model quite well, because it's not that they just shifted their attitude a little bit. I mean, they could have also said, okay, yeah, he might have been disloyal to us, but maybe, well, he didn't also have that much opportunities there and so on, but they really had to shift completely.
And of course it's also amplified actually by the group dynamic. Then if some person in their social media started to call them failure, of course, it's also put pressure on the other members to agree with this or shift the attitudes. But of course, as you also said, they just shifted from one extreme to the other. It's not that they're actually become any more nuisance than before. It's just from one highly idealized attitude. They shifted just to one other, very idealized and very simplistic attitude.
Michael Garfield (44m 60s): So maybe a little, the place to wrap this then would be to, you know, just invite a little bit of discussion about a question that you pose at the end of this paper, about the possibility of finding neural substrates for this. You compare what's going on here in networks of beliefs with the heavy and learning that fire together, wire together, learning of neurons in the brain. And you know, when you look at this kind of phenomenon at the social level, again, if you think about it as a human behavior in society, as a, as a form of collective computation than it is sort of like these people that are firing together, behaviorly are forming these neural motifs at the level of society and thereby determining the identity of a nation or, or whatever.
Do you think that it's going to be just a clear one to one? Are we going to find the fingerprint of belief in the brain in this way? Or do you think it's going to be a bit more of a abstract relationship between a person's connectome and their beliefs and behavior? Do you think we're really going to be able to like brain scan somebody in 20 years and tell them what they believe? Yeah,
Jonas Dalege (46m 13s): No, I don't think so. There's also quite a movement in brain size in Europe. Neuro-psychology acknowledging that trying to locate certain functions in the brain is not really the optimal way to go, but it's really much more about the complex interactions between different brain areas that can actually tell us much more about the functioning of the human mind. And I think I completely agree with this view and I think it also lends this to our work that is really at the psychological level where we also want to stress, okay. The complex interactions between different psychological factors, this matches this view that we have to study the brain in its complex form much better.
Michael Garfield (46m 51s): Awesome. Any parting thoughts before we put a pin on this?
Jonas Dalege (46m 55s): I think it's very important to think a bit more about how we can actually get back to more nuanced opinions, tolerating also ambiguity.
Michael Garfield (47m 4s): Awesome. This has been a pleasure to talk to you today. Thanks for being on the show.
Jonas Dalege (47m 9s): Sure. Thank you for having me.
Speaker 3 (47m 13s): Thank you for listening. Complexities produced by the Santa Fe Institute, a nonprofit hub for complex systems science located in the high desert of New Mexico. For more information, including transcripts research links and educational resources, or to support our science and communication efforts. Visit santafe.edu/podcast.