Mutual Transformative Change with Richard Watson

This is a way-making research conversation with Prof. Richard A. Watson of the University of Southhampton (Institute for Life Sciences/ Department of Computer Science, Agents, Interaction and Complexity group). You can watch the video here. We discuss the traditional ideas of adaptation and a Richard Dawkins style approach to Darwinism and then how this might be reimagined, as Richard Watson is doing in his work. We also discuss what it means to be an individual . Are love and logic opposites? How might we dance with a music that is both scientific and personal?

For Songs of Life: https://youtu.be/zdmY6q2ZWm8?feature=shared

Richard Watson’s personal website.

Compositional Evolution

Publications.

Conversations with Chris Fields and Michael Levin.

Master thesis that Richard kindly mentions.

Valencia Event

Video of the Conversation, in case you want to see the moment at the end where he shows us love on the wall.

Biography of Richard Watson:
"Dr Richard Watson studies evolution, learning, cognition and society and their unifying algorithmic principles. He studied Artificial Intelligence and Adaptive Systems at Sussex University, then PhD Computer Science at Brandeis in Boston. His current work deepens the unification of evolution and learning - specifically, with connectionist models of learning and cognition, familiar in neural network research – to address topics such as evolvability, ecological memory, evolutionary transitions in individuality (ETIs), phenotypic plasticity, the extended evolutionary synthesis, collective intelligence and 'design'. He has also developed new computational methods for combinatorial optimisation (deep optimisation), exploiting a unification of deep learning and ‘deep evolution’ (i.e. ETIs). He is author of "Compositional evolution" (MIT Press), was featured as "one to watch in AI” in Intelligent Systems magazine, and his paper “How Can Evolution Learn” in TREE, attracted the ISAL award 2016. He is now Associate Professor at the University of Southampton."

Google Scholar

Support on Patreon and You Tube. Listen anywhere you find podcasts.

Transcript

Richard Watson Mutual Transformative Change: Reimagining evolution

Introduction: [00:00:00] Hello everyone. This is Andrea with another research conversation. Just had to walk with my dog here in Heidelberg, where it sort of snowy and icy. I was listening to this conversation I had at the end of the summer with Richard Watson. Michael Levin actually recommended him to me, but I had also come across him at those talks too, in Valencia, where all three of us were. And found his work really inspiring. So I'd actually already reached out to him. Even before Michael recommended him. And, I revisited this conversation. I'm so glad I did because it re relates really well to something I've been thinking about a lot, which is the individual and the body. And how to define these. I gave a talk here at the colloquium recently. And this issue was one that really came up a lot is how am I going to really define the individual and the body? And so we talk about that a lot in this conversation, we talk about evolutionary units and individual units. And I kind of had this epiphany just now that may be what we're trying to do as humans on the planet is [00:01:00] become responsible enough to engage in what Richard calls here. Mutual transformative change. Whoa. I don't even know what I mean by that yet it just started came to me. But I just say it because there's so many exciting ideas in this conversation. And we get really deep by the end. Uh, though, at the beginning we talk about. Some very precise academic ideas about evolution and cognition. The work that he is doing the work, the papers that he and Michael Evan have written together. That have these very illuminating. Definitions of individual and body and agency. Which I'm finding very helpful in my work. Because of course my idea in way making is that all encountering is cognition for the body. All that encountering is the development of mind and intelligence. Which are themselves physical patterns of encountering. So. I have to be really careful here that it's understood that I'm not reducing this. I'm not reducing walking and crawling and swimming to [00:02:00] cognition. It's almost the reverse. And as you'll hear here, the definition that. Watson and 11 give. Of agent or body is very helpful in this way, because it also sort of flips that traditional script. 

And it's also a way To show, at least for me that I'm not saying that plants and animals and insects have human cognition. That, when we define cognition in this way, we also have to distinguish that each body or individual will have its own development of cognition. So this conversation was illuminating in that sense. Or is illuminating. I will continue listening to it because I'm not sure I understood it myself yet. But, uh, it's also illuminating in another way. And my other conversation with Richard since then, Have been illuminating in this way, and that is in understanding how it relates to relationships and everyday feelings, which is part of what I also am trying to do in my work. What is meant by love? It might sound shocking. To say that word. In the context of bringing up the work of Richard Watson, because he studied [00:03:00] artificial intelligence and adaptive systems and. Then he got a PhD in computer science from Brandeis and he works on now on all these con connectionist ideas of learning but of course his work actually goes way beyond that too. 

And I'll link to his website so you can see all that he's doing, but we get to this warmer. Second part of it. Uh, and the second part of our conversation. And even though it's kind of uncomfortable and challenging. To put this word love with this word science. Uh, that's what we look at here and we even look at the importance of subjectivity. And as Richard says, It might even be that subjectivity itself is the most sort of original or important thing. That we bring to the table, even in these discussions of scientific evolution. So how can we hold that weird paradox in mind? About 40 minutes in, we start to talk about these. Different trajectories. And I asked Richard if he would say his experience with these two different approaches and he bravely talks about it and shares his own story a little bit. And [00:04:00] as you'll hear some great musical. Gifts come out of it. So I'll leave it there and, and let you listen. Uh, I just want to say thank you to Richard Watson and to all of you, I hope you're having a good day, wherever you are. All right. 

Bye

Andrea Hiott: hey Richard, it's great to see you. As I was already saying before I hit record, I'm really excited about your work and about the connections, and thanks so much. I already feel a lot of gratitude to you.

Richard Watson: We said it a moment ago, but yeah, I'm really excited to speak to you. The stuff that I read in your master's thesis really resonates with me very strongly, so it looks like we have 

Andrea Hiott: a lot to talk about.

Yeah, thanks. I went back to do that master's just because of all these ideas I wanted to... Somehow have them tested academically. So, um, I've been thinking about this a long time. It seems like you have too. But to Just, you know, give people an idea of, of who you are.

I thought we'd start with this. I guess it's a good dichotomy or it's an either or of the [00:05:00] idea of natural selection, it seems like we've sort of just accepted that anything that has to do with adaptation or evolution must be this Darwinism that we've learned. Universal Darwinism, as you call it, or others call it, uh, your work is challenging that I remember when I was doing my neuroscience master, I wrote in one of the, you know, forums where you have to write about papers.

Um, I feel like people are starting to challenge traditional notions of Darwinism and I got so much pushback. Um, so, okay, that's just to kind of set it up,

Richard Watson: yeah. That's a good place to start. So, um, the idea that biological adaptation occurs through heritable variation and reproductive success, you know, the principles of Darwinian evolution. That was an answer that Darwin provided for how things appear to be so amazingly complex and adapted. And prior to that, we didn't have any explanation of how biological things could be so design [00:06:00] like.

 And that idea of random variation and selection is a very simple and very powerful idea that, you know, that Dennett refers to as the universal acid that can solve any problem. Right. So, it's a really simple process. Uh, things vary at random, and those entities which are thereby better able to survive and or reproduce, you'll have more copies of it.

It's an automatic thing, right? It can't, it can't not be true. It's logically true. Now, it's, it's also a process which you can take, uh, to be not just about biology. If you have any, um, sort of set of entities which are, subject to differential survival, that some of them have properties that enable them to survive that others don't, um, hard grains of sand and soft grains of sand.

You'll end up with [00:07:00] only the hard ones because the soft ones don't survive as well as the hard ones do. 

Andrea Hiott: Oh, I hate that, but yes

Richard Watson: so that's, so that's selection um, that's natural without it being a biological thing. Now in order for it to be, but that's, you know, you can already, you can already see that that's not very interesting, is it?

Because You're just throwing away the things which survive less well, but the things that you end up with are things which you had at the beginning, in that example, right? Just the hard grains of sand. But biological evolution is obviously much more creative and productive than that. To which we say, well, if you just do that lots of times, but with random variation thrown in, then there'll be a cumulative increase in the adaptiveness of a system towards its ability to survive. So, that idea of generalizing Darwinism to other substrates. Uh, sometimes called generalized Darwinism, it's been applied to all sorts of [00:08:00] things, you know, it's been applied to, physical organizations, it's been applied to the universe as a whole.

In fact, why does the universe have such, peculiar parameters that enable stuff to exist? Or there must have been a multiverse and this was the universe that survived. Institutions, uh, or businesses, ideas or memes that travel through cultural inheritance through, um, human populations.

Yeah, 

Andrea Hiott: so we have this idea of like, selfishness, competition, survival of the fittest, which is sort of taken out of context. Almost everyone knows these terms now, right? It's almost, we assume that's the way things work. 

Richard Watson: So, uh, when it's applied to human social systems, sometimes referred to as social Darwinism, the idea that people or organizations which are better able to survive or better able to out compete other people or organizations, while those are the ones which are most adapted, then, the position is that that's been taken out of context now, right, that no, no, no, no, Darwin and biologists didn't mean that, we were just talking about how organisms [00:09:00] survive and reproduce, and when you take that to the context of culture and society, that's a misappropriation. None that they say. Yeah. It happens anyway. But it still happens. Yes. Yeah. So, uh, nonetheless, that idea that variation and selection is a process of adaptation which can happen in any substrate is sometimes known as generalized Darwinism. But the idea is also, as you said at the beginning, taken further to mean it's not just that it could apply in lots of different cases.

It's that in any case, when you see adaptation, it is Darwinism. So that's universal Darwinism that Dawkins describes as, if you were to find life on another planet that was adapted you would know, without even looking, that it must have been produced through natural selection, because natural selection is the only possible mechanism that can ever produce adaptation, except design. And since design is ruled out, because we don't, that requires genuine design, we're required an intelligent designer. If it's a natural process, it must have been natural selection. There isn't any other [00:10:00] way in which a system can get better by itself without a designer, by itself meaning without a designer.

Right. 

Andrea Hiott: That's really very well said. I remember a friend of mine who's an artist actually, found Richard Dawkins book and was like, Oh, this is great. It describes me, the selfish gene. Because he's always, in his own family, been sort of selfish. Um, I think that's just an example of how it gets, you just, you, you get these kind of big, big statements or ideas and people just see them and take it in a certain direction, which I think you're describing there.

Yeah. 

Richard Watson: And then, you know, Dawkins will say, Oh, no, I never meant that. I always, you can do, you know, uh, um, morally valuable, ethical things. If you want to, that's your choice. One of the things that's extraordinary about us is our ability to choose to see something other, uh, something other than our biological makeup.

And it's like, That's totally not what you wrote, though, is it? Right? You described us as lumbering robots determined by our genes. 

Andrea Hiott: As you've also pointed out, he's such a good writer. I mean, I love that book, even though [00:11:00] I find it not, I mean, I don't agree with this overall approach we're talking about, but it's seductive.

Richard Watson: It's a brilliant oration, not just of his thinking about the topic, but it's a brilliant oration of the received view, the received wisdom about how things work and how things are in the scientific understanding of living systems, right? It's all controlled. Everything is reductionist, so everything is controlled bottom up, and the thing that's at the bottom of biology is genes.

So genes 

Andrea Hiott: control everything. There's something so simple about it, and we like that, I think, in a way. The other notion, which we're going to get into, can feel a little more complex.

Richard Watson: So, yeah. It is clear to us that in fact, there are other adaptive mechanisms other than natural selection.

Namely learning. Learning isn't, doesn't have to be a variation and selection process. So you can describe the way in which connections between neurons change in the brain in a way that doesn't involve. A population of [00:12:00] neurons, some of which are changing at random, some of which survive or reproduce better than other neurons.

You can just say, the connection between these two neurons gets stronger, and the connection between these two neurons gets weaker. Now, Others, Campbell and others, have suggested that oh, but actually it is (and Edelman). Actually, it is a variation and selection process going on inside the head. There really are populations of neurons, some of which are being useful and retained, and other neurons are not being so useful and being culled, and there really are connections which are useful, which are being retained, and connections which are not useful, which are being culled.

And almost as though appealing to the, um, amazingness of biological evolution, which obviously works by natural selection, that natural selection can do amazing things in biology. So maybe natural selection also explains the amazing things that brains, minds, and cognition [00:13:00] does. But, Although you could implement an adaptive process based on natural selection inside the head, you don't have to. It's possible to write a simple model of a learning process based on strengthening or weakening connections that is directional. Connections between two neurons which are excited at the same time get strengthened, and you don't need to create some connections which are stronger and some connections which are weaker and keep the ones which are strong when they're supposed to be strong and throw away the ones which are weak when they're supposed to be weak.

You don't need any variation selection, to do that kind of learning process. So that means that there is in the universe more than one kind of adaptive process. There's variation and selection processes and learning processes, which are sometimes called transformational processes. It's one system that changes its shape or organization over time, not a population of different systems and you retain some and discard others.

Right? 

Andrea Hiott: [00:14:00] Right. Maybe we could clarify a little bit the difference between evolution and adaptation.

Richard Watson: Yeah. So good. That's good. So. in biology, uh, adaptation and evolution are generally, uh, treated as synonyms, that the only, since the only mechanism that can produce adaptation is evolution by natural selection, adaptation is defined as the product of evolution by natural selection, that whatever evolution by natural selection does happened because It, uh, served as survival and reproduction of the entities to which it applied, and that's what adaptation is, increasing the survival and reproduction of the entities which are involved.

Uh, I don't find that very satisfactory. Because not least, uh, you can't ask the question I was asking if you define adaptation that way, you can't ask the [00:15:00] question, are there any other sources of adaptation if adaptation is defined as the product of a particular adaptive mechanism? So, uh, Learning is, not necessarily anything to do with survival and reproduction.

It might be in the context of a brain inside an organism, but in principle, learning is to do with, uh, building a model of a data set that enables you to make inferences or generalizations over that data set, and you would call that adaptation. But it wouldn't be evolution by natural selection. So 

Andrea Hiott: here we get into your computer science background a bit and the idea of connectionism and unsupervised learning and this too, you've kind of used as a way to understand that we, we could have other options instead of this one 

Richard Watson: that's right. So the, the process of learning that goes on inside the head can be modeled in a very simple way. I mean, what goes on in brains for sure, [00:16:00] very complicated, not simple, don't understand how it all works. Totally amazing. But you can do a really simple model of a learning process. That's the sort of an abstract, simple, um, algorithm that learns.

And that's, you know, the basis of all machine learning and modern AI stuff is at the moment is connectionism, which is just about. Building models, right? So if you take it on face value that, oh, this thing happened, oh, oh, this other thing happened, and you represent those independently, you haven't made a model of it, you've just recorded what happened.

But when you make a model of it, you say, you know what, I think this happens when this happens, or the opposite of this happens when that happens, so they're correlated or anti correlated, now you're beginning to model what happens, you're beginning to make sense of what happens, not just record what happens.

And that, simple process of reinforcing, strengthening, or weakening connections between your observations [00:17:00] is sort of one of the fundamental principles of learning systems. Uh, that's necessary to explain how they can store memories of multiple things and generate, um, uh, creative solutions to problems that they haven't been exposed to before.

In other words, generalize, and these are all phenomena which are familiar to us in the context of, organisms that have brains. But that doesn't, that doesn't provide an alternative to a different source of adaptation because only organisms with brains can learn, or machine learning systems that were designed by organisms with brains.

But If you wanted an alternative for how adaptation could happen spontaneously in the natural world, well that isn't it because brains are themselves a product of natural selection. Mm-Hmm. . So it was all natural selection. After all, it's just natural selection is [00:18:00] even more clever than you thought because it also invented other adapt mechanisms.

Nonetheless, it is enough for us to know that. The existence of learning algorithms, whether in the head or artificial, shows us that there are other principles of adaptation apart from random variation and selection. The question then becomes, what kinds of systems can exhibit learning? Is it only ones which have been evolved for the purpose of doing learning, or ones which have been designed for the purpose of doing learning, or can learning systems occur spontaneously?

And the work that I've been doing over the last 10 years has been about showing how you don't need anything special for a system to do adaptation in a way that a learning, a connectionist learning system does adaptation, that that can happen spontaneously in networks with suitable properties. 

Andrea Hiott: Right.

And I think your papers on that, especially related [00:19:00] to what you call a natural induction and, and, but also the connectionism ones start to explain that, um, these are kind of hard issues though. So I don't know, maybe I can just ask some basic questions that I find a little bit difficult. I want to get to this dichotomy of logic versus love at some point, because that's kind of an interesting way that you framed it.

But before we do, is the difference that you don't need competition? Or that you don't need this paradigm of competition when it comes to something like this connectionist or unsupervised learning model or natural induction model of, of adaptation, is that giving us a way to understand how systems evolve without this framework of competition and, you know, if you don't have the best qualities, you're going to die kind of mentality. Is that the difference? Because I sometimes have trouble understanding how adaptation and learning are different, except from maybe the agent base that's, um, assessing them.

Richard Watson: [00:20:00] Yeah. So natural Evolution by natural selection, the fundamental principles are differential survival and reproduction. Some things survive, others don't. Some things reproduce, others don't. The mechanism by which the differences between my survivability and your survivability or my reproducibility and your reproducibility play out in the natural world is competition. There's a principle of mutual exclusion. One of us is going to make it and one of us isn't. That's how the population is going to change over time. That's 

Andrea Hiott: this universal Darwinism. 

Richard Watson: Well, it's the basis of, of natural selection. And if you think that natural selection is the only adaptive mechanism there is, then it's the basis of all adaptation.

Now, one might quickly counter, but biology is full of cooperation, right? You know, biology is full of organisms like the clownfish and the sea anemone that cooperate with one another. You scratch my back, I'll scratch yours. And not just that there are these sort of... Esoteric examples of [00:21:00] cooperation, but that it's actually fundamental to all of life that in order to have a multicellular organism for all the cells of your body need to cooperate with one another, and you wouldn't have any interesting changes in the level of complexity in biology if it wasn't for cooperation.

Yeah, fine. I agree with that observation. Now explain how natural selection did it. Now explain how a principle that's intrinsically based on competition. Mm hmm. Reduce those cooperative relationships. 

Andrea Hiott: Right. That would be cells competing against each other, which is like cancer or something.

Richard Watson: Yeah. So, which can happen, but why doesn't it? Well, you'd say, well, because the organisms that get their cancer under control survive better and reproduce better than the organisms that don't get their cancer under control. Oh, okay. So you just move the competitive argument up a level and that explains why the cells have to cooperate with one another.

It's they're not cooperating with one another out of the kindness of [00:22:00] their hearts, they're cooperating with one another because if they, because they are in a very controlled circumstance created by the organism that makes them cooperate in that way. And if they don't, the organism will die, and them with it.

Right. So it's still competition which explains that cooperation competition is the prime mover, as it were, and you have this sort of story that, competition is the prime mover, but the winners are those who cooperate best,

so there is cooperation, because if you cooperate well, you'll win, but winning means you out competed somebody, right? You're actually using the competition to explain the cooperation. But that only works. After you've got a higher level of organization that can compete with another higher level of organization, if it's true the natural selection is operating at the level of multicellular organisms that are competing with one another, that would explain the cooperation of the parts therein.

But for systems that are not evolutionary units, like [00:23:00] societies or populations or the biosphere as a whole, you can't have cooperation at that level because they're not evolutionary units. They're not organisms that exhibit. or undergo differential survival or differential reproduction. So there would be, according to the natural selection story, there would be limits on the scale of organization that can exhibit cooperation.

Namely, they have to be inside units which are competing. 

Andrea Hiott: As we're talking, it's really hard not to assume this like, either or mindset of, um, it's either competition and natural selection or it's natural induction.

And this other model, which you've, you've shown, right? And, and that's not even what you're saying either. I feel like if anything, you're saying, here's another way it could happen. You're trying to open up the space and say, okay, uh, it's not, it's not about either or it's about, there's many different ways we could assess what's happening here.

Is that, is that right? Or 

Richard Watson: yeah, I think they are, uh, both [00:24:00] right at different levels of explanation. Then there's, then there's an interesting case about. You know, but which came first or which is at the bottom, which is sort of separate 

Andrea Hiott: questions, right? Yeah. And this is that mindset of we need it to be one or the other, which gets to this individuality too, I think.

And how, how can we understand? 

Richard Watson: I find, I often find myself drawn into the competition between my theory and your theory between natural induction and natural selection. And my theory is better than your theory, like, Oh no, I got sucked into 

Andrea Hiott: it. It's very hard not to, right? Like even this conversation, we already talk about it as if they're, it's one or the other.

And you, you say very specifically, this is an alternative, you know, it's, it's, it's about trying to break that, but I think that's what's so hard about having what you talk about as a history, what I might talk about as a trajectory that's kind of, um, structured, uh, in a certain way. And, and now you with your work are trying to kind of notice that structure and, and offer, open the space around it.

But it's [00:25:00] weird, right? Because you're still using that structure and within that structure. Do you ever feel like that constraint of, it's exactly what you write about too. You're trying to sort of change the system from within the system, which is, you can't help but do, but the system itself has a history.

And so you're constrained by that history. 

Richard Watson: Yeah. It cycles within cycles, right? That there are, given the way things are, The components within that system are likely to be competing with one another for limited resources. But, how did they get to be that way?

How did that get to be the rules? How did that get to be the circumstances? Explaining the origination of the organization in which those relationships was playing out wasn't explained by competition. That was explained by their mutual transformative change in the nature of the relationship between the entities.

 Some people say, you know, there's just lots of different levels of selection and that explains everything else, all the other details just follow from that, right? So, there's selection at the genic level, there might [00:26:00] be selection at the cellular level, or there used to be, there might be, there still is in single celled organisms, there might be selection at the multilevel of the multicellular organism.

And in some very special hysteric cases, there might be selection at the level of groups, but it's selection all the way up and all the way down. The interesting thing, though, is that... In order to explain how biology moved from one level of organization to another, you can't use selection to do that. Why would an organism get involved in a relationship that creates a higher level entity that then forces them to do something they don't want to do, right?

Why would my cells get involved in a social contract we call a multicellular organism that now forces them to do things they don't want to do? Now if it doesn't force them to do something they don't want to do, then the multicellular organism isn't really there. You can just think of it as an ecology of cells.

Like ecosystems out there are supposed to supposedly not organisms. I don't need a high level of selection to explain what they [00:27:00] do. Nobody's doing anything they don't want to do. They just do whatever is best for them, right? It's a jungle out there, or a tangled bank as Darwin called it. Uh, but they're, they have interactions with one another, but, um, self interest explains what you see.

So if self interest doesn't explain what you see, that must mean that you're in some sort of social contract that causes you to do something else. Why would I want to be in a social contract that causes me to do something I don't want to do? So you can't explain it from the bottom up. Now you could explain that, but if there was a higher level of selection, Then an organismic contract that makes its parts do this could survive and reproduce better than an organismic contract that makes its parts do that. But that's what you wanted to explain, right? You can't use that to explain the cooperation of the parts until after it exists. So you have this sort of catch 22 that selection from below and selection from above, neither can [00:28:00] explain the origination of the higher level of organization.

Andrea Hiott: Let's come back to this individual idea. Is it? Oh yes. Because you're, that's what you're sort of talking about here too, it's like, um. Is there a way that if you change what we understand as an individual, that, that alleviates some of this?

Isn't part of it that we are talking about an individual and assuming something that means as if it's static? Yeah, 

Richard Watson: so, um there are at least two quite different notions of individuality that are relevant to living systems. One is the idea of an evolutionary unit, or a Darwinian unit, which means a unit that's capable of self reproduction.

That exhibits heritable variation and reproductive success, a proper evolutionary unit, it can undergo, a population of them can undergo a natural selection. So they're, um, you know, genes, uh, collections of genes in a single celled [00:29:00] organism, multicellular organisms, and sometimes... Uh, spatially aggregated or compartmentalized social groups can be evolutionary units, but other things can't because they don't have heritable variation in reproductive success, but a quite different notion of individuality.

Uh, is more related to ideas like agency. What's the sort of scale of a system or scope of a system that is capable of acting on its own behalf? And there I like, uh, Mike Levin's terminology of a system that exhibits information integration and collective action. So if you had a bunch of little components, which didn't integrate information with one another and didn't do collective action, you'd say, well, that's not one individual, that's many

but if they do share information with one another, undergo a collective decision point and then take [00:30:00] collective action, you say, well, that's not a population of things. That's one thing by virtue of the fact that they do information integration and collective action. They are one individual. And that's kind of what we mean by the difference between a single celled organism and petri dish full of a population of single celled organisms. And it's also what we mean by what's the difference between a bag of single celled organisms and a multi cellular organism, right? What's the difference, right? In the bag of single celled organisms, they're not doing any information integration and collective action.

They're all acting individually. But in a multi cellular organism, they clearly are doing information integration, taking in information from all of your senses all over your body. Making collective decisions and then acting collectively on them, and that's what makes them an organism. Now, what's the relationship between these two types of individuality?

So, uh, it's been suggested that, [00:31:00] uh, that kind of, uh, agential or cognitive view of information integration and collective action, uh, that cognitive view of individuality is coextensive with, with life itself, right, there's no such thing as a living system that doesn't do information integration and collective action.

That's, that's what it means for it to be one system and not just a bag of little systems.

So then what's the relationship between that notion of living systems and the evolutionary units? Well, the usual associations that it's sort of one to one and onto that only evolutionary units can be uh, agential units. And even then only in very special cases. So, we are an evolutionary unit and we're agents, I don't know, let's give it to dogs and chimps.

Well, maybe we could give it to a few other things, but when you get down to, I'm not sure about insects or maybe not single celled organisms and, you know, like now it's just getting silly. They're not really agents. They're just evolutionary [00:32:00] units. So it would be that agency is a product of evolution by natural selection and therefore agency can only be a trait of a thing, which is an evolutionary unit.

So the only thing that can be an agent is an evolutionary unit. And even then only in special cases, probably too rare to worry about. My view and Mike's view is, Mike Levin, is that it makes more sense that it's the other way around. That acting in a way that exhibits information integration and collective action is primary to living systems.

And that happens at all sorts of biological scales, regardless of whether they are evolutionary units or not. And that ecosystems do information integration and collective action, even though they're not evolutionary units. And that it was an ecosystem of cells prior to becoming an evolutionary unit that was doing information integration and collective action that enabled them to become an evolutionary unit at the higher level of [00:33:00] organization.

When their collective action turns upon things which are vital to their survival and reproduction. It creates an evolutionary unit at a higher level of organization. But the evolutionary unit is a product of the agency, not the other way around. So there's agency everywhere, not all of which are evolutionary units.

Uh, and the evolutionary units are a special case, rather than the, um, driving factor. 

Andrea Hiott: I think it's a paper that you and Michael Levin did together where you define individuality as I wrote on here where the evolution of the whole is a non decomposable function of the evolution of the parts. I think this is one of the best. Uh, definitions I've ever seen of it, and it is sort of flipping it

Richard Watson: but yeah, I mean, so when, when people say, the ecosystem isn't an organism or the biosphere isn't an organism, what do they mean, right? They mean, well, it's just a free for all, right? It's just, it's just a jungle out there. It's just everybody acting for themselves, right?

[00:34:00] Everybody is self interested. And the difference with a, with an organism Let's leave aside whether it's an evolutionary unit for the moment is that each of the components instead of having each of the components acting for their individual short term self interest You have a system which is behaving in a manner Which is consistent with the long term collective interest instead of the short term self interest and that change in physical scale and temporal scale is the thing that makes This collection is a higher level organism, and this collection isn't.

Andrea Hiott: This really gets at something that I think is important to all three of us. You, Michael Levin, and me, and probably lots of other people, and that's this trying to understand scale, nestedness, and non linearity, Um, so I don't know how to get into that except to ask maybe does universal Darwinism , does it preclude,, complexity or dynamical systems or networks?

Richard Watson: Yeah. That's a great way to get into it. So. Uh, [00:35:00] it doesn't preclude it at all, right? So, natural selection doesn't in any way preclude the fact that organisms are complex dynamical systems. Although natural selection operates on genes, Organisms aren't just the sum of the genetic contributions of each of the genes that they contain.

There's a complex dynamical interplay between all of those parts, which is what it means to, develop from the genotype into the phenotype, the bodily form and behavior of the organism. And that's complicated and dynamical. The difference, though, is that the gene's eye view says, Uh, yeah, but all that matters to selection is all that matters to evolutionary adaptation is if I change this gene, what difference does it make?

And if I change that gene, what difference does it make? As though you can treat all the other genes for a moment as though they're just a, uh, a context or environment of this gene [00:36:00] and say, if I were to optimize this gene such that it maximizes the survival and reproduction of the whole, which allele for that gene would I choose?

And if you were to do that for all of the other genes. then that would just be what the evolution of the organism is. So, you think about genes as being the difference makers, and all of this complicated dynamical system between the genes and the phenotype as being, yeah, sure, it's complicated, but it's not part of the mechanism.

If anything, it's just as an annoyance which obfuscates the mechanism, right? The, the, the real mechanism is variation and selection acting on genes. And all of that complexity makes it more difficult. It's not, it's not an enabling actor in this process.

 The real, nub of the issue is reductionism, right? Is it, is it sensible to think of all of the systems that we're interested in [00:37:00] as being, uh, the sum of the parts that they contain? Even though they have nonlinear interactions between them, you know, their parts and their interactions, uh, in their totality explains everything that happens.

What else could be going on? Well, maybe what the organism chooses to do affects what happens here, or what the cell's behaviors are affects what happens here. And the reductionist says, but what the organism chooses to do and what the cell's behavior are is all determined by what was happening at the lower level.

That's not even a thing. I could describe all of it at the lower level and only ever talk about the lower level and that would be fine. 

Andrea Hiott: That's exactly what I mean, though. That's very constrictive because it assumes there's only one trajectory

Richard Watson: So when, um, Mike shows, for example, that you can change the number of heads that an organism has without changing its genes, what should [00:38:00] we make of that?

Right. Yeah. Well, I thought. That the morphology and behavior of an organism was determined by its genes, including things like what size it is, what shape it is, and probably how many heads it has, right? That if you wanted to change how many heads it has, you should have to change its genes, right? But I've changed the number of heads it has, and I haven't changed its genes.

By the way, weirdly, it's heritable when the worm reproduces through fissioning. It's still two headed. 

Andrea Hiott: Yeah, this is just almost such a huge difference in the way we've thought about things that it's very exciting and also, I think people almost just ignore it because yeah, 

Richard Watson: yeah, but how can you ignore it?

Right? So you can ignore it is you can say, Oh, well, it could be epigenetics, right? Because you didn't change the DNA, but everybody knows that if you change the gene expression, the epigenetic marking and the gene expression of a cell, then that changes what it does. So of course, in principle, you could create different tissues, different organs, different morphologies [00:39:00] without changing the DNA because he would change the expression of it.

And obviously Mike interfered with it. He changed the bioelectric patterns of the bioelectricity in the tissue and that's how we got them to do something else. And so they get to keep their reductionist perspective about what was happening here. It was all bottom up, except for when Mike interfered with it and then he made it do something else.

Right. 

Andrea Hiott: But I don't think that holds because as you've, I think I heard you discussing with Mike and Chris Fields, maybe the, that, that actually shows that the, the environment itself could be understood as an agent. And then we, we really can't fit it into that same mold anymore. 

Richard Watson: Yeah. There's a beautiful, um, experiment with putting a single cell on a very tiny little platform, a little pillar.

And, uh, if the pillar is round, the cell just sits nicely, roundly on the top of the [00:40:00] pillar. If the pillar is square, it tends to spread out, you know, just through the physics of sitting on the top of the square pillar. And you say, well, okay, well that's not, you know, recognizing or adapting to the shape of the pillar that it's on.

It's just physics, making the cell blob out to the shape of the thing that it's sitting on. Yes, that's right. It is, but it changes the gene expression of the cell, right? The cell's shape induced by, clearly induced by the environment, changes the gene expression patterns inside the cell. So, there's a two-way relationship, right between, you know, the traditional direction is the cells determine the shape on morphology, sorry, the genes determine the shape on morphology of the cell. But if you change the shape of the cell, it changes the genetic activity, right? That's a bidirectional relationship, not one way.

And that means that if there was a way of getting information into cells that didn't come from the genes. There is a pathway that could get that [00:41:00] information into the genes, at least into their activity, if not into their sequence. Mm-Hmm. , this is 

Andrea Hiott: closer to induction 

Richard Watson: another day. The then the problem is, but, but there isn't a way in which you could get adaptive information into the phenotypes of organisms.

So all of the. Adaptive information that's in the phenotype of organisms has to come from natural selection in the genes. So even if there was a pathway that went backwards, it would only be passing back information which would come up. Right? You'd put the information in the genes, it would go up into the morphology, and then the adaptive information would go back into the genes.

But I don't know why, because it was there anyway. Who cares? And the way in which the environment has modified it... It's not modifying it adaptively. It's just messing with it, right? It's not, that's not where there's no mechanism by which you get more adaptive information on the way back than you did on the way up, right?

If you assume the natural selection is the only possible mechanism of [00:42:00] adaptation. 

Andrea Hiott: Yeah. And if you stay in that linear, it's got to be unidirectional. Um, that, that, that same model, right? I've, I've heard you, uh, say that this ends up in a kind of science or logic versus love, um, paradigm, which is a little bit what we've been talking about, and kind of a, A different way, but I wonder if you might open that up a little bit.

Like, what do you really mean by love in that, in that, uh, formulation? 

Richard Watson: Deeply vulnerable mutual knowing is my current working definition of love. That's very nice. So, uh, Imagine that I was an intelligent system trying to learn about my environment for the purposes of survival and reproduction. Uh, I might think of this as an exercise in acquiring information about the outside world, and perhaps internalizing a model of the outside world for what?[00:43:00] 

So that I can manipulate it, control it, predict it, so that it can maintain my survival and reproduction at the expense of someone else's. That's the conventional way of thinking about what agency and intelligence is for.

But, um,

Where to begin talking about what's wrong with that view of the world, right? Well, 

Andrea Hiott: maybe we could begin with how I've seen you structure it before. So we've been talking about a really, really Western idea of evolution. And, of course, now it's sort of spread and... You know, whatever.

I mean, it's, it's a bully in a way. Um, but there was another kind of trajectory and it's still around. And I mean, of course, again, we're putting things in boxes that aren't, don't really fit in the boxes, but, 

Richard Watson: You know, I can remember watching, I think it was from Carl Sagan's Cosmos, not from David Attenborough when I was a kid, explaining how natural selection worked and getting it.[00:44:00] 

When I was like 12 and then thinking about the spider in the corner of my bedroom and it's like, oh, it's not running away because it's afraid or because it, you know, it wants to get away from me. It's running away because spiders that didn't run away died. You know, it's like the natural explanation, natural selection explanation of, of behavior right there.

And that was, you know, I was inspired by that and awed by that about isn't that amazing that such a simple mechanism could produce all of the incredible adaptive complexity that we see in the natural world. I want to study that. I want to understand that better. So it's. It's been a long road. Uh, you know, there's, there's a sort of a scientific story about how I came to realize that it doesn't explain the things we wanted to explain.

And there's also a personal story.[00:45:00] 

Now, there's an, there's an interesting, there's an interesting hesitation there, right? Because scientists aren't allowed to have personal stories, right? Because personal stories might influence your scientific trajectories. 

Andrea Hiott: Right. That's the soul logic versus 

Richard Watson: love thing. Yeah. That we should be entirely objective, devoid of any subjectivity, devoid of any values even.

We're just learning about what the reality of the world is and we don't bring any anything personal to it We don't bring agendas to it. You can't have an agenda in science. That's not a scientific, right?

but um a different way to look at it is that one's subjectivity is the only thing that you bring to the table that's new, right? The only thing that you bring to the, to the universe that is, uh, [00:46:00] stands a chance of participating in the creative process is your subjectivity. It's the thing that's unique.

It's the thing that's, um, the new thing that you bring into the table. And scientists know that their subjectivity influences what they do, but they treat that as a problem that ought to be removed because it's preventing you from being objective. 

Andrea Hiott: Or they just refuse to see it and 

Richard Watson: make sure they never let it leak into the language.

I wasn't being subjective. That's really how it is. Yeah. Um, but I think that there's a, there's a different way to look at it. But anyway, I'd, shall I tell briefly the scientific story first? Basically, so I studied evolutionary algorithms, which in computer science are the sort of. abstraction of the variation and selection process.

And for sure, if you run an evolutionary algorithm on a difficult optimization problem, it gives you better solutions over time. You say, that's it, you've cracked it. You know, John Holland showed this in the seventies, [00:47:00] the first simple computational abstraction of a Darwinian process in a computer. So you run your evolutionary algorithm and it gives you better solutions over time and you think that's kind of cool and you think I wonder what would happen if I run it longer.

So you let it run for a week instead of overnight or it doesn't really do very much. You let it run for a month, you know, seems to be stuck. Now from sort of lack of scaling of the algorithm. You could say, well, if you had a really fast computer or a massively parallel computer and 2 billion years, then maybe that would do everything, right?

Maybe that would explain all of the adaptive complexity that we see around us. And for sure, it is a massively parallel life. Living systems are a massively parallel computer and they have been running for 2 billion years. So, you know, maybe the fact that you don't get anything more in a month than you got in an hour is not a showstopper for [00:48:00] that theory.

But, with a computer science hat on, a different interpretation of this is that you haven't really captured the algorithm properly. It's not that you just need to run it longer. It's that you haven't really captured the algorithm properly. And so just, a straightforward disillusionment with what that mechanism can do when abstracted inside a computer, motivated me to sort of look outside the box, right?

So. At first, the suspicion is that the algorithm that we've implemented inside the computer must not be what Darwin said. Maybe we didn't incorporate all of the details or subtleties that mattered, but we did. There is random variation. Uh, there does produce differences in fitness. And there is heritability from one generation to the next.

We keep the solutions that work and we discard the solutions that don't. Those are the three properties necessary for natural selection to occur. It is artificial [00:49:00] selection, but it's a variation and selection process that's occurring. So we didn't miss what Darwin said. So if this algorithmic process doesn't produce, isn't sufficient to explain the adaptive complexity that we see in nature, it's not because we misinterpreted Darwin is because Darwin misinterpreted nature.

It's because the way that biological evolution works. Isn't explained by evolution by natural selection as Darwin described it. And I don't just mean as Darwin described it 150 years ago, I mean, as we still understand it now in the framing of natural selection introduced by Darwin. Um, I began to think about things like the evolution of cooperation, the transitions and individuality, which I've mentioned.

And eventually came to the conclusion that natural selection couldn't explain transitions in individuality. And that if natural selection can't explain transitions in individuality, it can't explain biological evolution. Because the [00:50:00] transitions in different levels of organization, that's where all of the action is.

 The amazing biological complexity that we see around us, it's the changes in the scale of organization. That's the biological complexity that needs explaining.

And that isn't explained by those processes, it's explained by the transformative change in the nature of relationships between things, not the competitive exclusion of one thing for another. , 

Andrea Hiott: when you say individual. Transformation or was that the word you used? Transition? Um. 

Richard Watson: An evolutionary transition in individuality. Yeah. 

Andrea Hiott: Yeah, what, like, what's, what's that really mean 

Richard Watson: to you? Okay, so Maynard Smith and Zathmurray, uh, Zathmurray, uh, wrote a book in 1995, the Major Transitions in Evolution.

And they said, look, you know, you have this idea of things changing over time through an evolutionary process, like a giraffe getting a longer neck. But, uh, there are also these. fundamental changes in the evolutionary process, which are much more interesting. They describe it [00:51:00] as entities that were capable of independent replication before the transition can reproduce only as part of a larger whole after the transition.

So before the transition, they're all individuals, little things, and after the transition, they're working together as a larger whole. And their examples include Uh, from self replicating molecules to the first chromosomes, from simple prokaryote cells to eukaryote cells, from single celled eukaryotes to multicellular organisms, from multi single, solitary multicellular organisms to eusocial insects, and then some hand waving stuff about society and language and stuff like that.

And the... The crucial characteristics are that, there's a change in the level of evolutionary unit in these transitions so that things which were separate before are part of a larger whole afterwards. And that's where the real action is in evolution. Taking something of a particular scale and refining it, that's not really where the action is.

It's like, how do we get from self replicating [00:52:00] molecules... To bacterial cells, to eukaryotic cells, to multicellular organisms. That's where the real action is. And explaining those transitions. Maynard Smith and Sathomir, I should say, were fully on board with the idea that natural selection has to explain those transitions.

Because, as we all know, natural selection is the only adaptive process there is, and it explains all biological complexity, so it must explain all biological complexity. Uh, so, um, you know, so that was, I took that as my sort of personal research project to understand how natural selection could do that, and eventually came to the conclusion that it doesn't.

So the personal story is unsurprisingly. A lot to do with, can't we just be nice? does the world have to be so harsh? Does it have to be true that everything is driven by self interest? Couldn't we be a little bit kinder? Couldn't we be a little bit nicer? Now, [00:53:00] in, in evolutionary biology, that expectation and that attitude has been driven out by natural selection.

That's just, I'm sorry, but that just isn't the way the world works. It is nature red in tooth and claw, it is competition, it is survival of the fittest. It's a fight to survive. It is a fight to survive. And you know, you don't have to take it personally, that's just the machinery of the thing, right? It's not mean, it's just how it works.

And to imagine that it's anything other than that is because you're soft in the head. It's because you don't understand how natural selection really works. Well, it isn't because I don't understand how natural selection works, it's because I don't think natural selection is the explanation for it. So it turns out that there's a whole other community of people out there that already know that love is the answer, that already know that competitive mutual exclusion is not the driver that matters, and self interest isn't the driver that matters.

And I say, hey, did you know it isn't really all [00:54:00] about competition after all? And they're like, yeah, we did know that, right? And that's why we never paid any attention to the theory of natural selection, because... We know it isn't about competition. We know life isn't about competition, and that isn't the way that it works.

And you're like, oh, you knew that? It's like, oh, so, but how do we reconcile that the, you know, the personal conviction that Love is the way, not competition and mutual exclusion, not, um, the discompassionate treatment of one another in the natural world, but the compassionate and vulnerable connection with people.

How do we reconcile that with natural selection is the only mechanism that produces adaptation? It's the only mechanism by which any system can be better by itself, except by design. It's like, well, one answer is... Well, that's because humans, like many other organisms, are social creatures, [00:55:00] and in the past, it, uh, organisms which lived in social groups that took care of one another out competed other social groups that didn't take care of one another.

So, that's why you're programmed by your genes to be social. And when you go extending those social dispositions to other people that you're not related to, Or other parts of the natural world or the biosphere as a whole. That's just your, your intellect failing to grasp the fact that, uh, you sort of overextended or overgeneralized beyond the thing that really mattered, which was the survival and reproduction of your own genes.

So, there's no such thing as being nice to one another. It's just you accidentally overgeneralized from something which was actually driven by 

Andrea Hiott: competitive selfishness. I mean, you're describing what... A reaction is and is what is that reaction? Why would why would we want to fit everything into that [00:56:00] same set of constraints?

Is it just so that we feel like we understand how the world works? Um, does it relate to this? Sense of how love can be very, uh, uncomfortable and scary and the world suddenly, and I don't only mean romantic love, I mean, trying to live in, not even trying, just living in that vulnerable place, as you described is not an easy place to sit, right?

Because it is dynamic in the way we were saying, and this other view is more static and you can fit things in. Is that, do you think that's the motivation for wanting to do that? 

Richard Watson: So, you know, the scientific view would say, well, it's not driven by anything, it's just the way the world works, right?

I see the world objectively and you don't. Uh, my current understanding though, is that it's... The subjective force that drives that conclusion is, is the fear of loss of self. If you're really [00:57:00] afraid that by being vulnerable you won't exist anymore, then you'll do anything to defend that point of view. And it's not just about, it's not just a sort of biological imperative that self interest must be the driver.

There's a logic of self interest that's really hard to get out from under. If you have an individual and they have interests, then the logical thing for the individual to do is the thing that maximizes their interests. How can you get out from under that logic, whether they are an individual produced by natural selection or not, whether they're a gene or an organism or a multicellular organism or a person of a particular kind of multicellular organism, how can you get out from under the logic of self interest? If your starting position is I'm a self and my existence matters. Then everything else has to fit with that way of viewing the world, which is that everything that I [00:58:00] do is about sustaining me, everything that I learn is about ways to manipulate and control the outside world in order to persist, uh, longer myself, or maybe to live vicariously through my descendants in future generations.

Right. But once, if your starting point is the self and the persistence of the self, then all of that other stuff, it's not just that it isn't biologically true is that it just doesn't logically make sense. It doesn't logically make sense for an individual to do something that hurts themselves for the benefit of another.

Andrea Hiott: RIght, and by hurt it can just be being uncomfortable or opening yourself up this morning I woke up and I thought, wow, I, I am, I really love my partner and I'm really like, this is wonderful.

And underneath that there was like a, but you could lose it, That can be really scary, right? For good reason and there's, there's something about all of this that isn't disconnected 

Richard Watson: so [00:59:00] when, when you come to know another person, that knowledge is not just a, a sort of repository of facts which you add on and carry around with yourself.

I think that when you come to know another person, you become, uh, a model of them, right? They, in the knowing, you are changed, , your internal makeup is modified. So that you become essentially more like them or a mirror of them. So you, I don't think you can really know someone in a way that isn't vulnerable.

Because if you're, if you haven't been changed by the coming into knowing with them, you don't really know them at all. And that's the thing that's vulnerable about it. Because to really know somebody means you have to let go of the, Uh, idea that you're going to be the same. [01:00:00] I'm not going to be the same person after I enter into this relationship or after I, uh, get to know this thing about my environment, including other agents that are in it.

So Friston's work, Carl Friston's work about free energy principle and becoming a model of your environment or a model of how you can act in the environment such that nothing is surprising about the 

Andrea Hiott: environment. The Markov blanket and all that. Yeah. 

Richard Watson: So that, that, um, blanket, the interface between that, that separates me from you.

Chris Fields has pointed out that that's really symmetric, right? That the, it's just as true to say that. I'm, I'm internalizing a model of my environment as it is to say that my environment is a model of me, right? That those, those two sides of the mirror are complimentary of one another. Now, I think that's more obvious when a person meets another person, when you meet, [01:01:00] when system A meets system B and system A is sort of physiologically like system B, you can see that you can mirror each other.

But, um, I think that we also do that with everything because the only thing that we can really understand about our outside world is the aspects of it that are like us. And that in, in knowing it, we are entering into a resonance with it, uh, that is changing us as well as, um uh, us getting a deeper understanding or deeper knowledge of the thing I 

Andrea Hiott: have to push that a little bit.

Like the only thing we can know about another is something like us strikes me as. I think that might be what draws us or what allows us to communicate that we have a certain amount of overlapping regularities. Um, but none of us have 100 percent completely overlapping regularities. I mean, I think of it in terms of like way making as a trajectory, right?

We all are coming with a different sort of trajectory of everything that we've experienced, all our regularities. And for the most part, a lot of those overlap, but [01:02:00] they also are really considerably different. So there's a way in which when two people, A and B, the systems come together, there is that overlapping of regularities, but also what you were just saying, that growth and change kind of comes with how, when those two systems are interacting, they start to share those new Thank you.

New paths, or 

Richard Watson: creative paths. That's right, so in the transformation they become, in the transformation that makes them better models of each other, they become more similar. Uh, but there is still a complementarity, right? They're a mirror image, not literally the same. That complementarity is, the reflection of one another is important for them, for that relationship to create something which is more than the sum of the parts.

Because otherwise, it's like 1 plus 1 equals 1, that you didn't, I didn't have anything new, that we're identical, right? But in the, in the dance that we enter into, when we enter into a relationship, romantic or otherwise, we are learning how to move in a way [01:03:00] which doesn't clash with the other, that doesn't crash into one another as we move.

And that requires each of us to Internalize a better model of the other so that we can anticipate how they change, uh, and understand what will stress them and how they will respond to things. And that means that as we become a better model of how they are, it means that we are also stressed by the same things.

And we also respond to the environment in the same way. We hear the music in the same way, uh, so that we're dancing together. But of course, it's always underdetermined. I can only see the way that you behave on the outside. So I'm trying to become in tune with you, by being in contact and remaining vulnerable, I necessarily become in tune with you, that I begin to resonate with you

but because, because we, because I'm not able to have an internal structure that's identical to yours just by observing what you [01:04:00] do. that's an underdetermined problem. There's always a, a difference or a mismatch between what we do in the way that we dance together, but as that becomes more harmonious, the differences also take on the form of something which is harmonious and not just a buzz, not just a clash, but, uh, a space between us, which is itself, uh, dynamic and flowing rather than, um, crashing into each other but a different way to view that is that in by in coming to know you better, I become concerned with the same things that you're concerned with.

So that we are now concerned about that thing instead of. I'm only concerned about it because it affected my survival and reproduction and it, it, it's in knowing one another, we expand the scope of we become a thing, right? We become a being. 

Andrea Hiott: You can think about it and you can open the space and think of it differently so that yes, of course, we want to have good lives as individuals and survive and thrive and all of [01:05:00] that, and that's not incompatible with the fact that everything we come into contact with is changing us.

Um, and so that actually our individual self or even this evolutionary unit isn't actually what we think it is. It's not this body. It's, there's multiple ones going on, right? So it doesn't, so those, the survival of those does also matter.

Richard Watson: From the self interested point of view, it's like, why should I be interested in the well being of. Things other than me and the way that I think about that these days if you want wants to explain it in terms of how does that look from the self interested point of view.

So, it's like the, the electron needs the proton to exist, right? They're not trying, they're not in competition with one another. They create one another through their mirroring of one another. 

Andrea Hiott: Yeah, I don't know why I want to push on this modeling and mirroring I feel like there's some messiness here and like entropy and even when you were talking, I was thinking sometimes entropy for one system is [01:06:00] the nutrients of another system, if that makes sense. So if like what we're trying to understand here is almost too big to really understand at this point.

But if we can really open the space in this way, then you do start to understand that what you were describing is those interactions where we're. reflecting one another, that is itself a different form of, um, like whatever all this is we're talking about, right? Evolution, biology. Um, this morning I thought too, I had this kind of like, what is it we are really trying to figure out, right?

When you study adaptation and evolution and computation and all, what is it, what's motivating us to do that? I mean, we just want to understand all this connection and change and, and how. There's something very, very exciting in it all, right? Do you feel that too sometimes? And love is the best word to describe that.

 It is the best word to describe it. Yes. Wonderful. I was going to ask you to share that. It's so good. There's something very practical about that, [01:07:00] too, 

Richard Watson: we're drawn to love. It's in in the nature of the universe, not just in the nature of things created by natural selection.

It's in the nature of the universe to be drawn into harmony, which is another way of Being drawn into relationships that harmonize or resonate with others and that is the fundamentally creative process that in, in coming to know one another in a way which is harmonious, we become something that is more than the sum of the parts.

I think it makes sense to think of, uh, creation as a process that derives from love rather than think of creation as the process that derives from self interest.

Andrea Hiott: That's beautiful, but it's, I think it's really hard, as you were saying, to... This comes back to the science versus love or logic versus love, like there's something very, it's almost love itself, like, uh, that makes it so vulnerable and, [01:08:00] um, uncomfortable to talk about those in connection

 There's an expression which, um, people sometimes use to describe the sort of fundamental logic that underpins. The special case logic of natural selection, which is just, uh, what persists exists, right? So the things which are successful in continuing to persist are the things that we see, and the things which are not successful in continuing to persist are not the things that we see.

Richard Watson: What persists exists. That's it. That's the principle which explains the way the universe works, the way life works, the way organisms are, and it's supposed to explain why You're deluded into loving things sometimes as well, just because that enables your persistence. And, uh, whilst there's a truth to that principle, I think it misses the point.

Which I would put as what relates, creates. It's [01:09:00] true that the things that persist will persist. You didn't explain the creation. You didn't explain where they came from. You didn't explain how you got something more than what you started with. With that principle. And it's the putting things together in new ways that is actually the creative process.

And the putting things together in new ways is about relating, not about just persisting. It's about creating a new interaction between things. 

Andrea Hiott: I really love that and that I think is a really what evolution and life is, is doing, you know? I mean, we're one experiment of it as this form of a human, I guess, but you could think of it the way, I mean, it's really sunny here today.

I don't know if it is there. It's like we have 30 degrees in the Netherlands again and I'm looking at the trees and they seem in love with the sun, you know? Yeah. I mean, there's a way in which life is, is what you just described, not, uh, [01:10:00] this persistence only. There, there's a way in which the, yeah, just persisting does sound like it's static and not changing, and life is not like that at all.

There's nothing that's not, there's nothing that's the same the next minute as it is in this minute. 

Richard Watson: The, uh, the modeling or mirroring that I mentioned before you mentioned it being the feeling a little static, right? So I it you're absolutely right that it's a dynamic thing And so Friston does get that it's a dynamic thing that is, you know Because I act on the world and then my senses see what the changes are in the world and there's a dynamical flow to it But now put that That dynamical flow and that mirroring together.

So there's a thing on both sides, system A and system B, and they are acting on one another, interacting with one another, creating a dynamics at a high level of organization that's moves beyond the dynamics that they have inside each one separately. And I like to think of it more like a dance, by [01:11:00] which I mean a sort of a combination of an organization, a dynamic, and a little bit of musical harmony, right?

That it's more like a dance that I'm doing a dance and you're doing a dance and When we don't know each other, our dances are literally going to crash into each other. And as we do know each other, our dances will become more harmonious. Which isn't to say that we'll be identical to one another, but they will harmonize.

Mm hmm. Which is complementary to one another, but that, that complementariness has an element of similarity that we have to be in the same key and an element of difference that we're not just singing in unison, but we're singing in harmony with one another. 

Andrea Hiott: Oh yeah, I love this. And also that we sort of grew up with the same scale, you know, in a way we're not.

It's not, we don't, what I hear is harmonious isn't what you learned was dissonant. Um, although those can change too, the more we interact, if we, even if we don't align at first over time.

Richard Watson: But I do think that, uh, [01:12:00] ultimately, it does come from, uh, something much deeper, right?

That we are literally made of the same stuff. And that the stuff I'm made of resonates to the stuff that you're made of and other living things. And that that extends beyond living things and the physics of our, of our whole universe is made of the same stuff. Uh, and you know, you can get a, uh, a resonance between a high pitched version of a song and a low pitched version of a song, the same song at a different octave, still sort of harmonized together, even though they are At a different, uh, frequency scale and therefore at a different physical scale, but that wouldn't be possible if the octave relation wasn't intrinsic to the stuff that they were made out of.

That, that the doubling in frequency is a natural thing

Andrea Hiott: We are the same movement in a way. And I think music is a wonderful. Way [01:13:00] to try to understand it or the dance because when you're dancing you're you are you're always a vibration but you are the dance in a sense right and it's a Vibration and it's dynamic and it's changing and when we're always dancing with others I don't know something about it made me think too of how we were talking about two systems coming together and as you were saying, if it's just those two systems interacting over time, it would probably become sort of an equilibrium, almost like one thing.

And that's kind of getting at that static model. But none of us are only interacting only with one system at any one time, which also gets to the scale and nestedness. So in a way, those two systems are always interacting with multiple other systems, some of many of which. The other, system A is not interacting with the one system B is, and system B is interacting with the one system A isn't.

And so that's also kind of co creating system A and B, which is then also contributing to this dance too, right? So you have this kind of multiple sharing and changing and [01:14:00] dynamism that's going on. Yeah, 

Richard Watson: that's right. So if A, B are in a little dance together. They've harmonized together and they meet A'B who are not in their little dance together.

They haven't harmonized together. They haven't found that resonance. Then A'B are induced. To get into the relationship, because then they would be a better model of AB, right? You know, as, as they are without that relationship, they're not, uh, modeling very accurately what's going on in AB. Which is another way of saying, that that level of organization, the relationship between AB has been reproduced into A'B That it's a...

Constructs the, um, organization that mirrors itself.

Andrea Hiott: That sort of brings it, full circle back to where I sort of started with your work and how you've opened up a different space than just universal Darwinism, which isn't saying evolution is not happening. It's saying there might be another way evolution is happening.

 And it's exactly what you're talking about here, because [01:15:00] if you have these two systems. We're always influencing each other and we're setting the vibration, we're, we're creating the music of how, the present moment, how we're dancing and also kind of setting the trajectory and the frameworks for future generations.

Richard Watson: You could take that and say, well, it's just an analogy, isn't it right? You know, the biological things or the way that two atoms or two molecules interact with another physical things. It's just an analogy for how two people interact with one another.

It's not really the same thing. I think it is really the same thing. And the reductionist way of thinking about things is that, well, but none of that is a real thing anyway, because the only real thing is the atoms. The only real thing is the molecular level.

 The conventional story is a, is, um, a meaning of eliminating story, right?

An alternative view that's based instead of what persists exists, uh, what relates creates, saying that the organisms are the prime [01:16:00] movers, the organizations of Information integration and collective action come first when, and they orchestrate and are connected with, uh, all scales of organization above and below that relating can happen at any level of organization.

And it can be a prime mover. It's, it's determined mostly by other stuff that's going on at its own scale in a way which is semi independent of what was happening above or below. But it's all connected, also connected with the scales that are above and below through those harmonic relationships, right?

So the, uh, the octave relationship is double the frequency, but it's also double the physical scale. Um, double the frequency, half the physical scale and the stuff which is going on at one level of organization, for example, the level of organization that we care about, the level of organization where we have relationships, where we have beliefs, desires and intentions, where we have goals, where things have meaning to us, [01:17:00] that is as much a part of the process.

As all of the other levels in that physical system, and it means that the, that what matters to us does matter because it's part of the process that of creation, uh, and not just a product of a mechanism which was happening at a lower level of organization. Oh, that's so beautifully said. I don't mean it, I don't mean it just in the sense of an analogy.

I mean, it's, it's all part of the same system. 

Andrea Hiott: Yes, it's continuous and it matters in a very very practical way and

yeah, 

Richard Watson: because it's, it's a choice between love and fear, right? Exactly. If you start with fear, then you're worried about yourself and you're worried about not existing and you want to protect it and you want to exclude other things. And you start getting a totally different mindset. You never get, you never get anything creative that way.

If you start with love, then you're interested in moving towards things and leaning in and being invulnerable and being open to change and transformation. And that's the creative [01:18:00] process. 

Andrea Hiott: And we do start in love. I think when you're born, I don't know, there's a way in which you're starting with that connection, but you're not, you don't know it yet.

 It's like that TS Eliot thing coming back to where you started and knowing it for the first time. We all start in this complete sense of vulnerability, especially human babies where you're just completely vulnerable, right? You absolutely have to depend on love of something or someone.

Uh, not even depend on just live in it. You have to live in it and you don't even know it. And it's not that easy though, for us to just say, okay, live in love. You know, start with love. That's the real thing. Um, there's, there is this, there really is this. System and this process and you talk about in your work to about that, uh, natural selection isn't the opposite of constraint, sometimes the vibrations of the dance aren't always harmonious, and that's, that has a role too, I guess, 

Richard Watson: right? Well, so, uh, imagine that everything was just pure love.

I almost think it's, [01:19:00] it's, um, it's a little bit like there isn't anything there, but the, but the, it's just like a warm 

Andrea Hiott: hum, you know, yeah, it's 

Richard Watson: just on, right. But the, the creation of entities is a sort of breaking of symmetry. It's like it's sort of, there's a thing and then there's, and the thing is separate from the non thing.

And that's, there wouldn't be a dance between those things if you didn't have the two things. And whilst those two things exist, we can enjoy the beauty of the dance that they're in and be grateful for their self interest, which enables them to persist long enough for the dance to happen with the other thing with its self interest, whilst recognizing that those two selves were created by the hum.

They were created by the love and the complementarity between them, and that they need each other to exist and that they won't persist forever, because there's no enduring selves. And that's all part of the dance. But there [01:20:00] wouldn't be any things to be dancing if there wasn't also the yang as well as the yin, right? It's like, what's, what's the same here? But the mistake is to get too attached to the regularities and not see the transformation.

Andrea Hiott: Great. That's wonderful. I think that's, that's exactly the point. So I'll, I'll say we leave it there for now 

Richard Watson: that's been really fun. Thank you so much for the conversation. 

Andrea Hiott: It's been great. Thank you. Thanks for your time.