Ep 100: A retrospective on agency

Big Biology celebrates its 100th episode!

You’re only 100 once, and though we hope there will be many more episodes to come, we wanted to celebrate this milestone with something special. A recurring theme from many of our conversations with guests on the show has been agency. This is a BIG and oftentimes controversial topic in biology, and in this episode we try to pull together the different threads from our past guests’ approaches to agency and discuss what agency means and why we think it is so important for us to better understand it. 

We also hear from our recent addition to the podcast team, Cameron Ghalambor, who has very different thoughts on agency to Art and Marty, making for a fascinating discussion. 

Of course, we don’t have all the answers, but that’s never been what this show is about. It’s an invitation to think and converse about the biggest questions in biology. And that’s exactly what we do here.

Cover art by Keating Shahmehri

Donate to our spring fund drive! Or, become a patron! Or both – hey, we’re not picky!

  • Marty Martin 1:00

    Light the fireworks and blow the trumpets, it's the 100th episode of Big biology.

    Art Woods 1:05

    That's a hecto-episode for those of you who tend to invoke Greek prefixes unnecessarily.

    Marty Martin 1:10

    Okay, it's gonna be that kind of show, huh? Even episode 100? Have you no respect man?

    Art Woods 1:15

    We considered many options for our 100th episode, but in the end, we decided that it would be fun to pick out just one theme and piece together how different guests have approached it.

    Marty Martin 1:24

    We picked agency, a totally non controversial topic right, Cam? Cam? Cam?

    Art Woods 1:31

    Yeah, we'll hear from Cam later in the show. Let's just say he has different opinions on agency than me and Marty.

    Marty Martin 1:36

    But why did we choose agency?

    Art Woods 1:38

    We chose it because many past guests have convinced us that we really should be paying it a lot more attention in our own research.

    Marty Martin 1:44

    And all of that unintended cajoling worked. We've come to think a lot more about agency and its implications for the work we do. For instance, it's really helped me to think about how epigenetic marks like DNA methylation might enable organisms to respond plastically and invade new areas,

    Art Woods 1:58

    And it's helped me design better experiments on niche construction and thermoregulation in insects.

    Marty Martin 2:03

    Also, agency or at least one of its core elements information has so resonated with us that we've profoundly changed the way we think about stress in animals. Art and I have been interested in the topic for years and after talking about information and agency to many guests on Big Biology, we've changed our approach to this topic, big time.

    Art Woods 2:20

    But in the broadest sense, we chose to focus on agency in the 100th big biology episode, because it illustrates why so many biologists have called for updates to the modern synthesis, an "extended evolutionary synthesis," some have called it. They want a more complete theory of life.

    Marty Martin 2:34

    Unfortunately, the message of the extended synthesis crowd often came across as the original modern synthesis was incomplete or wrong, which is not just off putting it misses the point.

    Art Woods 2:44

    The major innovation of the modern synthesis was to take statistical shortcuts and thus make tractable that which until then had been out of reach. It made the complex simple.

    Marty Martin 2:53

    Darwin's ideas about population change over time were first merged mathematically with the principles of inheritance that Mendel gleaned from his pea plants.

    Art Woods 3:00

    Then Fisher, Haldane, Wright and others invented concepts like genotypes, populations, and relative fitness to predict genetic evolution. And voila, a lot of biology became intelligible-

    Marty Martin 3:12

    and predictable-

    Art Woods 3:13

    based on a few simplifying assumptions.

    Marty Martin 3:15

    Here's Dennis Walsh, our guests from Episode 62, explaining the modern synthesis:

    Dennis Walsh 3:19

    The modern synthesis is a really abstract theory. You take these abstract entities- genotypes, and you construct an abstract entity, a population assemblage of genotypes. And then you apply to the genotypes is very abstract parent parameter-growth rate, relative growth rate. And you can track the dynamics of these abstract populations in this very, very powerful way. And the-

    Art Woods 3:49

    While skipping over the organism.

    Dennis Walsh 3:50

    Exactly, exactly. Leaving them out, right. And, and it's absolutely right, that this parameter fitness or growth rate that has all those biological things, packed into them, it can accommodate any amount of biology. But I think what the defenders of the modern synthesis don't do is they don't acknowledge or realize the level of abstraction at which this theory actually works. It's not about the nuts and bolts of biology, it's basically identifying a universality phenomenon that we find in thermodynamics. And any time you have an assemblage of differentially growing or changing entities, you're going to have this change in the population of this ensemble, that's expressible at a really high level of abstraction. And I think that's the core of modern synthesis. That's what makes it so powerful. But it's not a theory of the metaphysics of evolution or or as I say, in the book, what happens when evolution happens. It's a very poor account of what happens when evolution happens.

    Art Woods 5:01

    In this light, it's strange to say that the modern synthesis needs an update because it always rested on what William James called "vicious abstraction." By intentionally leaving out many of the details that make life, life. We learn things we never knew and made all sorts of progress in animal husbandry, improvements in crop productivity and even in human medicine.

    Marty Martin 5:21

    That said, we empathize with calls by some for an extended synthesis. We agree that biology is more complex than the modern synthesizers imagined. And we think it's time to strive for a more inclusive and cohesive theory of life- one that captures its origins, its adaptive operations, and its evolution.

    Art Woods 5:38

    To channel Sarah Walker guest on episodes 9 and 93, we need a model of life that focuses more on the alive parts of life.

    Sara Walker 5:46

    In my mind, life is a dynamical process. And it's one where you have particular informational patterns that are like structuring physical systems across space and time. People get mad that I think a computer is life or a screwdriver is life, but those things literally would not be created without information.

    Art Woods 6:02

    Yeah well that's consistent.

    Sara Walker 6:03

    Yeah. And then alive are the systems that are actively constructing things. They're the ones doing the information processing to actually build those things and like and use internalized information to actually do the construction.

    Marty Martin 6:15

    The modern synthesis took shape at a time in history when inheritance was just coming to be understood. And the molecular revolution, including the discovery of DNA, was just getting underway.

    Art Woods 6:24

    Progress via synthesis thinking was great, but it also led many subfields of biology to forget about the original "vicious abstraction." Most egregiously, genes came to take on causal powers they simply can't have including the combination of this forgetfulness, The Selfish Gene concept.

    Marty Martin 6:40

    To Richard Dawkins, George Williams, Arvind Agren, a guest on Episode 73, and many others, willow catkins, blowing from trees are literally DNA rain.

    Art Woods 6:49

    To their way of thinking, one need only understand how the DNA from one willow generation gets into the next. That's it. That's all. Biology done.

    Marty Martin 6:57

    No, can't be the modern synthesis can't be the main paradigm for understanding life, if it doesn't explain most of how life is alive, how complex systems emerge and come to be resilient, how homeostasis works, and how collections of subentities are integrated into working wholes.

    Art Woods 7:14

    One of our very first guests, Massimo Pigliucci, from Episode 7 beautifully laid out the limits of the modern synthesis. But to be fair, he was using Dick Lewontin's analogy.

    Massimo Pigliucci 7:23

    So he said, "Look, imagine you're building a house. And instead of in the United States, where most houses are built of wood, which is why they don't last, you build it, you know, the old fashioned way, the European way with bricks and lime. Okay, so you say okay, so you start putting the first layer of bricks and then lime on it, and then bricks and lime and bricks, and lime. Now, once you get the final house, you could, if you wanted, ask the very quantitative question, well, what is the weight of the house in bricks? And what the weight of the house in lime? And there is an answer to that question, and I'm sure it will be something like 98% bricks and 2% lime. That tells you precisely nothing about how to build a house. Because it isn't about, you know, it's you're not going to come up with 98 bricks, and then two little pieces of pieces of lime and then say, 'Oh, I got the house.' No, you get the house by the specific patterning of the bricks and the lime, right? And so the idea there is that even if you could show that, let's say 90% of genetic variation in phenotypes in a particular human trait is the result of genetic influences, that still doesn't mean the way in which is usually interpreted. Oh, so genes do all the work and the environment is not important. You take out that 10% in that specific pattern, and you get nothing, absolutely nothing, because genes by themselves don't do crap."

    Marty Martin 8:50

    And yes, you can buy that on a bumper sticker in the Big Biology store.

    Art Woods 8:54

    Massimo was talking here about the importance of phenotypic plasticity, but his point applies more broadly. The modern synthesis focuses on changes in the fraction of variation that things, either genes or environments, explain.

    Marty Martin 9:06

    But that kind of model doesn't capture much of what we wanted to understand in the first place: how the house comes to take the form it does, and how the house doesn't fall apart soon after being built.

    Art Woods 9:16

    Can we come up with another theory of life, a simple one, but also one that better captures the key things that distinguish life from non-life?

    Marty Martin 9:23

    We think we can and we think at its core will be the concept of agency, a system's propensity to maintain its integrity by either changing the disruptive external forces it experiences, or adjusting its internal makeup to better suit those challenges and opportunities.

    Art Woods 9:38

    Or to blend the words and ideas of a bunch of past guests, agency is the intentionally cognitive set of activities that a living system uses to achieve dynamic stability.

    Marty Martin 9:48

    The architecture and updates of the Bayesian priors and its Markov blanket model of the world.

    Art Woods 9:53

    In this 100th episode of Big Biology, we focus on agency with input from a subset of our past guests to try to forecast where biology might be going and what values lie in the conceptual transformation.

    Marty Martin 10:04

    And in the last third of the show, Art and I will talk with the newest host of Big Biology, Cam Ghalambor, on his thoughts and skepticism about agency.

    Art Woods 10:11

    I'm Art Woods.

    Marty Martin 10:11

    And I'm Marty Martin.

    Art Woods 10:12

    And this is Big Biology, Episode 100.

    Marty Martin 10:29

    To make a case for agency, let's start at the literal beginning, life's origins. In Episode 49, we talk with Nick Lane about this topic and his book, The Vital Question.

    Art Woods 10:38

    Most origins of life scientists at least historically identify with one of two camps- the RNA first world or the metabolism first world. Nick has feet in both worlds, with heavy doses of systems and thermodynamics thinking.The

    Nick Lane 10:51

    Yeah, I mean, the old idea of a primordial soup, it's not gone away, you know, Darwin's warm pond is probably still the dominant idea in the origin of life field, except when we now call it terrestrial geothermal systems. And a lot of that chemistry works quite well. You start with cyanide or cyanoacetylene, you use UV radiation, and you're able to make all the building blocks of life. And the problem for me with all of that. It's beautiful chemistry that works well, but it doesn't look anything like biochemistry. And so you're still left with this question. Okay. So we've got all these monomers floating around in solution. What happens next? How do they invent life from there?

    Marty Martin 11:31

    As you can hear, Nick wonders whether particular conditions on young Earth led to the appearance of simple, but self-sustaining complex systems.

    Art Woods 11:38

    Data from his and other labs show that proto-life processes were originally possible because of the availability of particular forms of matter and energy in particular places.

    Marty Martin 11:47

    Many sites on early Earth would not have supported life-like systems, but some might have been just rich and resource dense enough for those systems to maintain their integrity and eventually reproduce.

    Art Woods 11:57

    One way, perhaps the way that these systems were able to maintain integrity, was by instantiating information into their structural elements, which include the nucleic acids that became today's RNA and DNA.

    Marty Martin 12:08

    Now IT remains a mystery how lifelike systems came alive from this point of being just a dynamically stable system to the complex homeostatic modular systems that resemble modern microbes.

    Art Woods 12:19

    Probably there was no bright line between these stages anyway, an early step in this process must have entailed a separation of internal and external states, some sort of physical boundary, like what would become a cell membrane.

    Marty Martin 12:30

    Exactly what that first membrane would have been, much less how the inside came to take on such diversity, is to put it lightly, a work still in progress.

    Art Woods 12:39

    Nevertheless, according to Karl Friston our guest in Episode 70. All viable complex systems have what mathematicians call Markov blankets. The first barrier must have provided a clear demarcation between outside and in.

    Marty Martin 12:51

    And to be clear, Karl isn't referring to Markov's blanket as the thing he slept under, during particularly cold Russian winters. Rather, Markov blankets are relatively simple, interconnected sets of states whereby complex systems shield themselves from entropy.

    Karl Friston 13:03

    At its simplest, a Markov blanket is just a way of partitioning the states of a universe into a system of interest, say you or me or the virus, or the vegan, and everything else. More specifically, it's a partitioning, a dividing into three sets of states, those states that are internal to a system, those states that are external to the system, and then some intervening states that mediate the exchange between the inside and the outside. So if you were a physicist, or if you go back to your sort of school by physics, the Markov blanket enshrouds the internal states, and if you're a biologist, then you can think of this as the cell surface. It's the thing that sort of, together with the internal states constitutes the unit of description, of discussion, and is responsible for mediating the reciprocal, the two way causal exchange, between the inside and the outside. Mathematically it inherits from the work of Pearl in Bayesian networks, and is defined operationally in terms of what's called conditional independence. Which means that quite simply, if I wanted to know how my internal states are going to change in the future, and the immediate future, then I only need to know the Markov blanket states, the surrounding states, I don't need to know the rest of the universe.

    Art Woods 14:43

    Karl says that Markov blankets are the key elements in his free energy principle. The idea that all enduring complex systems including living ones must resist entropy by minimizing surprise.

    Marty Martin 14:55

    Karl's surprise here is mathematically defined. He means how unexpected information gleaned from the environment, internal or external, is related to a set of Bayesian priors. These are, in effect, parameters that comprise the Markov blanket. Here's Karl on surprise and free energy minimization ideas, he attributed to Richard Feynman.

    Karl Friston 15:12

    So he was dealing with a problem of trying to characterize the behavior of small particles in quantum electrodynamics. Trying to understand the probability distributions or the beliefs about different paths that particles could take, realized that to describe it properly, he had to turn what was an impossible integration problem, marginalization problem into an optimization problem, that he could then solve using standard techniques. That's a key move. What that does is it creates, it takes a system that can be described probabilistically in this instance, quantum mechanics, and converts it into an object that can be understood in terms of optimization. And that means you've now got a normative teleological gloss on describing how the system works, because it looks as if something is being minimized or maximized. So what is that thing? Well, the thing is the variational free energy. So it is exactly the same construct that is used in machine learning and high-end deep learning like variational autoencoders, where the negative physics, Feynman variational, free energy is known as an evidence lower bound or the elbow.

    Karl Friston 16:35

    So in that name, in that acronym, you have the key thing, which is evidence. So what we are talking about now is a generic mathematical way of describing the probabilistic dynamics or evolution of any system in a normative sense in a teleological sense, as trying to optimize a bound on evidence. So what is evidence? It's just the probability of some outcomes, given some model or hypothesis or construct explanation that you consider generated those outcomes.

    Art Woods 17:19

    If Karl's hypothesis that all complex systems have Markov blankets-

    Marty Martin 17:23

    all the way down, as he likes to say-

    Art Woods 17:25

    Lifelike systems have to make choices to endure, they must either push back on the world to make it what they need it to be, or they must update their models of the world. If their models become sophisticated enough that they can come to plan about how to push back, they by definition, have agency.

    Marty Martin 17:40

    Given enough time, and the right contexts, one should expect such systems to become common. One more time, here's Karl.

    Karl Friston 17:47

    If you believe that the universe is essentially a random dynamical system, what you are saying is that the variables and state of the universe evolve, which means that they have a trajectory. If they have a trajectory, then if you want to understand that trajectory in a normative sense, it's basically doing some form of gradient descent, some sort of either hill climbing or hill descending, which gives you this sort of teleological optimization perspective. But in so doing, even an elementary particle is in effect, selecting a path to pursue. So in that elemental sense, there has to be agency because there's time you can't have dynamics without time. And if there's time, then there are trajectories, if there are trajectories, then I go over there and I don't go over there. So you know, I agree entirely that you can't, you can't move away from or deny, an agential aspect, even to elementary particles.

    Karl Friston 17:47

    Having said that, I think there is a fundamental difference between agency that involves planning and agency that is just an expression of density dynamics. So my favorite example is the difference between a virus and a vegan, the virus, you're certainly has, has attained a non equilibrium steady state, it's a beautiful little model of its eco niche, its milieu in which it survives, it does all the right things entailed in its sort of molecular structure and kinetics are all the right substrates to be interpreted as a model of the kind of inputs and outputs in the world, the intracellular world, usually that inhabits. And one could say the same as a vegan. But a vegan, of course, can do a lot more than a virus. And if you can plan, then, by definition, you must have a generative model of the consequences of your action in the future. And because in the future, now says well, you've got a generative model with a temporal depth with the horizon. So then you can ask well, how far into the future can I see? Well, in a sense a virus could probably see a few nanoseconds or milliseconds into the future simply by committing to a particular trajectory. But of course, a vegan can not only see a few milliseconds into the future, she can see, minutes, hours, days, months in terms of

    Art Woods 19:03

    Yeah half a century. Yeah.

    Karl Friston 20:18

    Half a century, exactly

    Art Woods 20:26

    Okay, so combining Nick's and Karl's thinking so far, the first forms of life on Earth must have been self-sustaining systems,

    Marty Martin 20:32

    or "autocatalytic sets" in the words of Stuart Kaufman-

    Art Woods 20:36

    appearing in one or more places conducive to persisting.

    Marty Martin 20:39

    To avoid succumbing to entropy, the Markov blankets of successful systems would have become more and more complex. If they couldn't change their external states, they'd instead update the parameters of their internal states, eventually instantiating information about past successes into nucleic acids, and probably many other structures.

    Art Woods 20:57

    Paul Davies, guest on Episode 33,

    Marty Martin 20:59

    and author of many books, including the focus of our chat, Demon in the Machine, said

    Art Woods 21:03

    that this use of information, in the sense Karl talked about, is one of the things that distinguishes life from non-life.

    Karl Friston 21:09

    I think everybody thinks about the world about them will realize that life stands out. Living things are in a class apart, they perform the most amazing feats, they seem to be different, in a very fundamental way. And not just difference in degree but difference in principle from non-living systems. They are very, very odd. Now, biologists tend to not find that subject matter quite so odd, because they're dealing with it every day, life's what they study. So there's like, of course,

    Art Woods 21:40

    Of course we know what it is.

    Karl Friston 21:41

    But to a physicist, it looks like magic. It really does. And I remember thinking, well I suppose if you take a living organism, it's made up of normal atoms doing normal physics things. How is it that a collection of these stupid atoms blundering around just following the rules of physics, can collectively combine to produce some-what looks like magic, some form of magic? How can that happen? And it is a very, very profound mystery.

    Marty Martin 22:11

    To Paul, whereas we're taught from our first biology classes that DNA encodes information. What that really means is incredibly important for understanding life.

    Art Woods 22:21

    Without doubt, some kind of information is in DNA. But information is everywhere in life, what's in DNA is not all of it. And even in DNA, the information that resides in the sequence isn't necessarily even the most important bit. Here's Paul, again.

    Karl Friston 22:35

    You're absolutely right, that the information contained in DNA, the genetic information is something that people are familiar with. But it doesn't stop there. So genes rarely act in isolation, they can switch each other on and off. And they can form networks, sometimes of great complexity, and information swirls around these networks. And sometimes they're very much like components in some electronic system. They form modules. And these modules in turn, coupled to each other, form bigger networks. And we're talking here, unlike in electronics, about components being wired together chemically, not electrically, but the same principles apply, that these are logical operations that these components can carry out. And they can compute and regulate and fulfill many of the functions of modern electronics and computing, but they're doing it with a chemical basis. And so genes form networks, but it doesn't stop there, because cells themselves can form communities, they can signal each other chemically. So we're using this information language all the time, when we talk about cell-cell signaling, cooperation among colonies of cells, even bacteria can form communities that can carry out coherent tasks. And then when we come up to multicellular organisms, take social insects, for example, one of the really fascinating areas of study here at Arizona State University is with ants and ant communication and they form colonies and they engage in collective decision making. You can see these pictures where ants are sort of clustering around, you know, having a little conference and you're wondering, you know, what are they talking about? And we're beginning to understand now, there's no sort of chief ant that says, you know, okay, lads, and it's not lads, because they're all female, gals, you know, we're off to a new nest. It's one of these things that is done collectively, it's distributed across them. And it's done through information exchanged through all sorts of chemical and physical cues. And it goes on all the way up. We've talked about the brain. This is the biggest information processing system that we know, but it again, it doesn't stop there. It really encompasses the entire planet. When we look at ecosystems there's a lot of information flow, there are mobile genetic elements, things like viruses that get around the environment, couple widely separated systems together. So I like to say that the biosphere is the original world wide web.

    Marty Martin 25:21

    The problem with adding information to our theory of life is that it's a very abstract concept, so we've had a hard time measuring it.

    Art Woods 25:28

    In the 1950s, for instance, cybernetics was all the rage and many scientists argued that by studying information, the mysteries of life would soon be solved

    Marty Martin 25:36

    Wrong.

    Art Woods 25:37

    Although in some contexts, information theory was really fruitful.

    Marty Martin 25:40

    We're looking at you neuroscience.

    Art Woods 25:41

    In most biological sub-disciplines, it just never delivered. The information was truly a bad idea, or was too diffusely defined leading people to talk past each other, or it was just never really measured well in any of its possible forms.

    Marty Martin 25:54

    Several guests now feel that the tides have turned and then information will in fact, be integral to biology's future. For instance, Karl Friston thinks that the information geometry of a system will be what distinguishes a stone from a virus.

    Art Woods 26:07

    Sarah Walker, on her second visit, in Episode 93, seemed to agree. She and collaborator Lee Cronin proposed that something called "the assembly index," will be useful to finding extraterrestrial life. These indices can be calculated in principle for anything and inherently they represent the kind of thing Karl meant about information geometry.

    Marty Martin 26:26

    Sara used Harry Potter's Hogwarts Castle, one made of LEGO that is, as an example. Sara thinks that the higher the assembly index is of a thing, the more likely it was produced by non-random processes, namely life. Only information-rich things like life could build high assembly index structures.

    Sara Walker 26:43

    So with Hogwarts Castle, for example, imagine I'd just put the Legos on the table. And I didn't give you an instruction. And maybe you're a child that never read Harry Potter and I said, "Make Hogwarts." Like, what is the likelihood of you even being able to build that object, right? So the fact that you could even imagine the experiment probably means that you have some cultural association with my cultural background, right. But like, there was a goal in mind, and you can imagine building toward that goal, and probably you were assuming you had, you know, the Lego instructions in front of you to build it. So Hogwarts has a very high assembly index. The minimal path to make Hogwarts by randomly constructing it, just by joining operations, is quite large. You know, if I had said, "Let's just stick three blocks together, red, blue, red," you know, that would be pretty easy for you to randomly assemble. And so the idea is that everything can be tiered by this minimal path, which we call the assembly index. And the things that have a larger depth and time require more minimal steps, more memory to produce them are more evolved objects, they require more evolution to get to them more knowledge, more learning.

    Marty Martin 27:44

    If you heard that episode, you could probably tell that I was skeptical about Sarah's assembly index as panacea, or I'm too dense to understand her.

    Art Woods 27:51

    Yeah, she's way smarter than you.

    Marty Martin 27:53

    Agreed. But it felt like I could follow how assembly indices could be calculated for anything. It just seemed that once again, a vicious abstraction, extinguished the fire of life, its dynamism. I understood that even processes like metabolism could be captured by the assembly index. But I wasn't convinced that an idea developed to identify alien life was sufficient to explain living life. I mentioned this to Sarah and she said assembly theory accounted for it, but I just couldn't quite accept her explanation.

    Art Woods 28:19

    Well, for once, you're in good company Marty is this very thing that motivated Dan Nicholson, guest on episode 82. To co edit the book, Everything Flows. To Dan, a key trait of life is its dynamic equilibrium. It's for this reason that we titled his episode, Organisms are Not Machines. In other words, life is more a river than a river boat.

    Dan Nicholson 28:38

    Organisms are from a physical perspective, these systems, okay, they are systems that have to maintain their organization by constantly bringing in matter and energy from the environment, so that they attain this steady state, if they lose the steady state they die. It's an irreversible process, and that you can think of everything that organism does ultimately as being reducible to that, even if that, in practice, may not necessarily be helpful, but you know, at least right that whatever else organisms are, what can't be denied, is that the self-organizing systems. I mean, it's puzzling to me because it isn't something that often comes up in biology. If you really press a biologist, they'll say, "Well, of course." And yet, it doesn't usually feature in the way biologists explain certain phenomena, right? So it's a reminder, okay, that you can provide a physical explanation for certain capacities that organisms have, okay, in a way that is not shouldn't be controversial, shouldn't be problematic. And yet, it's a way of thinking about them physically, that is very different from the traditional, mechanistic, reductionistic, deterministic view that has dominated the biological discourse since the 17th century, right. So I say, okay, I'm giving you an alternative. And I'm going to anticipate your protest that is not scientific by saying I'm grounding it in physics, and I'm going to show to you that this grounding leads to really interesting implications for how you should think about biology.

    Marty Martin 28:39

    Dan pushed so hard against that life has "things," namely machines, because this metaphor has so biased the field. Machines like cars don't get tired, they don't rebuild their own tires when the old ones were down, and they don't make baby cars.

    Art Woods 29:59

    Organisms do all of these things. And when we abstract organisms into simpler systems like machines, or worse genetic blueprints, we miss the main points about life.

    Marty Martin 30:21

    We want a simple model of life, because a too complex model defeats the purpose.

    Art Woods 30:25

    Okay, first, before we dive into agency, we have to say that yeah, we know we get it that agency is a taboo topic to many biologists. It smacks of some "God of the Gaps," some form of soul, or a spooky force galvanizing our cells.

    Marty Martin 30:41

    Historically, though, agency and other vitalistic ideas didn't leave the bad taste that they now do. Claude Bernard, Louis Pasteur and many others had much more sophisticated and nuanced ideas of agency than the simplistic negative one that's so common now.

    Art Woods 30:56

    Hopefully, by this point, in the episode, you can tell that the agency we mean is not a mystical spiritual one. We see no room for such things in life, and we're both fairly staunch atheists.

    Marty Martin 31:05

    However, we and the many guests we quoted earlier do think that agency, or something like it, promises the chance to truly and fully integrate biology, maybe not all the details, but all the major elements.

    Art Woods 31:18

    So let's get after it, with help from past guests, starting with Dennis Walsh.

    Marty Martin 31:22

    Dennis's book, Organisms, Agency, and Evolution had a huge effect on us. To Dennis, for life to sustain itself and evolution to occur, agents must seek out resources, avoid danger, and generally expose themselves, or not, to selective pressures.

    Art Woods 31:36

    Some things are inherited memory tokens as Scott Turner called them. And these factors plus DNA affect change over generations.

    Marty Martin 31:44

    But what happens within generations, organisms' struggle for existence, is what Dennis thinks needs a lot more attention than it gets.

    Art Woods 31:51

    In other words, if organisms didn't have agency, they'd never succeed in the struggle. Organisms as systems, always at risk of breaking down, must have agency to exploit what Dennis calls affordances.

    Dennis Walsh 32:03

    So I think the affordance concept is really important. It helps us to explain well, the external dynamics of organisms, how they move through their environments and why, but their internal dynamics too, why the parts are integrated in the way they are, why organisms synthesize these very materials out of which they're made, because they're conducive to the pursuit of the organisms' goals, and the exploitation of their affordance. And they also like, create affordances that are structures or capacities, confer on organisms capacities to, capabilities to pursue their lives in this particular way, right. So we should understand the integration of organisms and their movement through their environment, in terms of the creation and exploitation of affordances. And affordances are dynamic, as you respond to an affordance, other affordances open up right, so there's this constant creation and exploitation of affordances going along. So my thought about agency was we should start there. We should acknowledge that this is what organisms are like, this is the kind of defining feature of life and see how working from the taking the affordance notion as basic, transforms our understanding of the dynamics of evolution.

    Marty Martin 33:24

    One of our other repeat guests on Big Biology, Episodes 39 and 65, is Mike Levin. In a first chat with Mike, we talked about the inheritance of body form in multicellular organisms.

    Art Woods 33:33

    Convention, of course, has it that a developmental program somehow resides in the genes, one that unfolds over some period of time to produce a mature body plan.

    Marty Martin 33:42

    But not always true. At least not in flatworms, and frogs.

    Art Woods 33:45

    Inheritance of their body plans has to do with, believe it or not, electrical fields,

    Marty Martin 33:50

    But that's for you to go hear in Episode 39, if you'd like.

    Art Woods 33:52

    Our chat with Mike on Episode 65, focused on agency, or more accurately, something he called cognition and agendas, and it was largely based on a 2020 article that he co authored with Dan Dennett and Aeon magazine called "Cognition All the Way Down."

    Marty Martin 34:07

    The main point of the article was that only agents can have agendas. Put a mouse on a ball on the top of a very steep hill, and both might initially resist rolling down the slopes, but shake that hill enough and only one system stays in place. Or climbs back to the top,

    Art Woods 34:21

    Mike and Dan, realized that any system with an agenda also requires a form of cognition, some way to identify options and choose among them to remain in the same state or to exploit another one.

    Marty Martin 34:31

    This choosing of options is what Mike means when he says cognition. And cognition is core to what agents do. Here's Mike talking about a molecular reaction arguing that, in some sense, even molecules have simple forms of cognition.

    Mike Levin 34:44

    People sometimes say to me, especially let's say molecular biologists, will sometimes say "Look you're talking as if this thing made decisions, whatever but we but but that's just the metaphor, right? You don't really mean that, you know, I make decisions, this thing is just chemistry." And I think it's very important. I think that's something major mistake. And I think it's really important to get from the get go that I think agency and cognition are a continuum, or a spectrum

    Art Woods 35:09

    To Mike, cognition can and should be broadly defined. The most conspicuous form involves the brain, but really any system that has an information geometry can be understood as cognitive.

    Marty Martin 35:19

    So subsequently, even simple systems like molecules and rocks can be cognitive, in that their interactions with other entities can lead to predictable changes in their form.

    Mike Levin 35:27

    For me, agency is a kind of center of gravity for things like decisions, preferences, memories, and in the more advanced implementations, sometimes things like blame and credit and other, you know, other things like that. Agents are things that can make mistakes, right chemistry and physics doesn't make mistakes, it just sort of does what it does. But agents are a, and I think this is a point that Dan has made before, that agents are capable of making mistakes. And so there is a kind of a whole spectrum of different levels of sophistication that different agents can achieve. So with the thing with the ball and the mouse on top of a hill is just an example of this. You know if you got a ball at the top of the hill, you can use equations that will tell you what it's going to do, and those equations have almost no reference whatsoever to information processing, to memory, to learning, to preferences, you don't need any of that you have a much simpler model that does pretty much everything you want to do to predict what that system is going to do. If you've got a mouse at the top of the hill, Newton's equations about what it's going to do if it were rolled down the Gravity Well are, unless the mouse is dead, are almost useless. Because now if you really want to understand what that system is going to do, or modify it and make the mouse go somewhere else, you have no hope, other than through a model that takes seriously what that system actually is. It's a system with preferences with memories with all kinds of internal states that are going to determine what happens later on.

    Art Woods 36:58

    Okay, so now that we've laid out what agency is, let's confront the elephant in the room. Does it matter? Do we have any evidence that agential thinking will contribute new and important things to biology?

    Marty Martin 37:08

    Because we started the show today by highlighting the vicious abstraction at the heart of the modern synthesis, let's start there, with the evolutionary implications of agency.

    Art Woods 37:16

    Specifically, let's flesh out the roles that agency plays and what's become a hot topic in evolutionary physiology, thermoregulation.

    Marty Martin 37:24

    Boo, how do your boys put up with those terrible jokes?

    Art Woods 37:26

    They love me. Ectotherms often use behavior to get body temperatures they want, and recent thinking on this topic has focused on something called the Bogert effect, which is named after the American herpetologist, Charles Bogert. In the 1940s and 50s, Bogert studied lizard thermoregulation and found something surprising.

    Marty Martin 37:29

    Lizards in many different geographic localities, living under many different prevailing thermal regimes, had body temperatures that vary little. Primarily because many of those lizards were such good behavioral thermal regulators.

    Art Woods 37:53

    Bogert wondered whether behavioral thermoregulation could blunt the effects of selection on other aspects of their thermal biology, specifically their upper thermal tolerances.

    Marty Martin 37:53

    And indeed, it can.

    Art Woods 37:57

    Here's Martha Munnn1oz, in Episode 81, explaining her work on the Bogert effect

    Martha Muñoz 38:10

    Ray Hughey, in collaboration with Paul Hertz and Barry Sinervo, wrote a really impactful conceptual paper, taking Bogert's qualitative ideas and creating a quantitative hypothesis testing framework with which to put them to the test. If we take Bogert's argument at the broadest possible level, what the argument says is that when any kind of regulatory homeostatic behavior is at play, that has the capacity to reduce environmental variation across some environmental gradients, and that that buffering should limit physiological divergence and/or slow the rate of evolution. And so Ray and colleagues effectively just gave this a new name the Bogert effect and devised a series of approaches for testing it. And what they did was basically developed two premises that should be true under the Bogert effect.

    Martha Muñoz 39:08

    The first is that regulatory behavior is occurring. That while it seems quite obvious is very complicated to actually demonstrate in the field. It requires understanding the environment that's available to organisms, so you need a null distribution of temperatures that organisms, if you're doing thermoregulation, that organisms could theoretically play with. Then you need to demonstrate that body temperatures are actually in a mean and range that is so separate from what's available in the environment that our metrics indicate that they're regulating. And so the null distribution tells you what the environment is, and then you can compare that to observed body temperatures and, through a series of metrics, basically quantify the degree of thermoregulation that organisms are engaging in and this is step one. So the second premise that should be true under the Bogert effect, is that thermoregulation should be associated with limited physiological divergence or weaker selection on physiology.

    Marty Martin 40:05

    The short version of Martha's work, to put a fine point on it.

    Martha Muñoz 40:08

    I've discovered that lizards from tropical islands tend to thermoregulate more than lizards from the Latin American mainland. And we discovered that the rate of heat tolerance evolution is about three and a half times slower on islands than on the mainland.

    Art Woods 40:24

    Get that? 3.5 times slower evolution on islands because island lizards do so much more thermoregulation, their physiology is shielded from selection.

    Marty Martin 40:33

    Although Martha didn't explicitly use the word agency, when she talked to us, we think it fits well into the arguments we're making in the episode.

    Art Woods 40:39

    To recast Martha's work explicitly in this language, we would say that lizards are agents moving around in their environments and exploiting affordances by which I mean choosing microclimates to give high body temperatures when it's otherwise cool and microclimates that give low body temperature when it's otherwise hot. The outcome of this agency is remarkable differences in the macroevolutionary trajectories of lizard physiology in different lineages.

    Marty Martin 41:04

    Martha's not nearly the only biologist that feels this way. And perhaps not surprisingly, the biggest proponents of agency as an important biological force tend to be the physiologists, especially those focused on homeostasis, how a system maintains stability, think thermostats.

    Art Woods 41:17

    Scott Turner, from Episode 36, said that adaptations generally don't really make sense except in the light of agency.

    Marty Martin 41:24

    What we so liked about Scott's approach is that he weaves together adaptation and physics with homeostasis as a thread.

    Art Woods 41:31

    To Scott and to Claude Bernard, a contemporary of Darwin and the father of modern experimental medicine. Stasis is the process that distinguishes life from non-life.

    Scott Turner 41:40

    The very persistence of an organisms form is itself a form of homeostasis. And that, of course, is maintained by this enormous complex of adaptive barriers that separates us from the environment, the linings of the lung, the linings of the intestine, the sensory interfaces, and those kinds of things, all of which are mediated by epithelium-like structures. And you can take some fairly simple aspects of conservation of mass and thermodynamics to be able to extend adaptive boundaries outward from the organisms. And in the case of the termites, of course, they build these, these are the African termites that build these massive mounds, as infrastructure for their subterranean colonies. What these mounds are is they are a big, massive adaptive boundary that has been constructed between the termites themselves and the environment, which they are, of course, totally unsuited to be living in on their own. And the more I studied them, the more I came away, impressed with just how extensive this reach was, you know, so it extends not only to managing the atmospheric composition within the nest but it also co-ops, the physical environment, the entire hydrology of the environment over a fairly extensive range, to be able to enable termites to live in a dry environment. But because they reconstruct their environment to manage water flow through it, they can survive in those kinds of environments.

    Marty Martin 43:23

    Core to Scott's thesis is that organisms and perhaps other levels of biological organization have to be intentional. Here, he talks about how mole crickets actively modify their burrows to get just the sound they want, so they can attract mates.

    Art Woods 43:36

    In some cases, like in humans, consciousness can come into play and be an aspect of intentionality. But intentional behavior need not be conscious, it just needs to be directed at something in the environment based on some pre-existing model in the brain or elsewhere in the body about expected outcomes and current needs. Again, states of a Markov blanket.

    Scott Turner 43:56

    So what is it that we do when we have an intention? You know, well, we there's a conscious part of it definitely kind of want to stay away from that little bit. But you know, this intentionality can be framed in a way that links the cognitive interpretation of the environment with the connection to the engines, if you will, that can modify the environment. And so, when you look at the borrows, the tuned boroughs of mole crickets, for example, you know, these creatures build a burrow, it ends up in the shape of an exponential horn, this helps project the sound of the call much further than it would otherwise. And if you look at what's happening during the construction of that burrow, the cricket borrows a little bit, emits a chirp, listens to it, and if it's not quite right, it continues to modify its burrow until it gets the chirp that it wants, once again, I'm putting up scare quotes here, that it "wants." Yeah. Right. And that's kind of an intentionality isn't it? And so if we want to try to develop a concept of what intentionality is that can be kept independent from the kind of mysticism that tends to trip this up, then, to me, the simplest definition is coupling modification of the environment with the cognitive interpretation of the environment.

    Marty Martin 45:31

    Before wrapping our story, we think it's important to point out the practical reasons for understanding agency one more time. Here's Mike Levin on just two such reasons.

    Art Woods 45:48

    First, our health.

    Mike Levin 45:50

    So imagine in the next 10 years, we solve the two things we're going to solve genome editing, so somebody will have come up with a nice clean way of making genetic edits exactly where you want it and nowhere else

    Art Woods 46:01

    Perfect editing.

    Mike Levin 46:02

    Yeah, forget all that, let's say you get totally perfect editing. And let's say, and by the way, let's also say that stem cell biology gets solved, so that from a stem cell, you can get any other single cell type that you want. Okay. So now you have all this. And so now that means that you're going to solve some really nice low hanging fruit. So single gene diseases, of which there are some,and single cell diseases, you know, Parkinson's maybe things like that. But then you're going to reach the much, much deeper question, well okay, somebody's missing their hand, let's say there was an accident or birth defect or, you know, whatever, or an eye, and now what? Because the point isn't to be able to edit the genome cleanly. The point is, what in the world would you edit? So if you don't, you know, if we don't have a good account of morphogenetic agency, and competency, if we don't have a reverse engineering of this kind of software of life that enables it, to have modularity and so on, you're talking about micromanaging at the molecular level, all of the steps that go on to making a complex organ. That is not going to happen, nevermind our lifetime, you know, I don't know how many years it's going to be before that's even feasible, if it even is at all feasible, that kind of micromanagement.

    Marty Martin 47:13

    Second, here's Mike on rapid advances in our technology.

    Mike Levin 47:16

    We are going to see cyborgs and hybrids, and you know, every kind of combination of biology, technology, artificial intelligence, and software and hardware, merges of living tissue with engineered, you know, all kinds of things. What that means is the older categories, things like, what is a robot? What is a machine? How do we recognize agency? What is something that was evolved versus designed? Does it matter for these things? We have to start wrestling with this now.

    Massimo Pigliucci 47:49

    Because in the olden days, it was very easy to tell. And even then, of course, we made all kinds of mistakes with various kinds of humans and animals. We made all sorts of terrible mistakes. But generally speaking, you could do this,would sort of knock on something. And if you hear a metallic sound, you would say, oh, yeah, you can do whatever you want with this. And if it was squishy, and you know, sort of warm and furry, you would say you if you do certain things with this, you're going to jail, right? You have to be nice to this one. It's a horse or a dog, whatever. And that was easy, because you could rely on two things. You could rely on the thing it's made, what it's made of, and you could rely on an origin story. You could say, well, this thing evolved. And this thing was created in the lab and that makes all the difference. Those categories are gone. I think even now, these categories are no good and they're absolutely going to be no good going forward.

    Marty Martin 49:05

    Let's suppose that Darwin had remained in the clergy, that Fisher failed out of math, and Mendel flew kites instead of growing peas.

    Art Woods 49:11

    Suppose that Lamarck, Shannon, Cannon, Waddington and McClintock got more positive attention or just weren't excluded by others in their fields.

    Marty Martin 49:19

    Suppose Lysenko was honest, or was just ignored.

    Art Woods 49:21

    Suppose that Williams, Franklin, Watson and Crick had access to the computing power of the average smartphone,

    Marty Martin 49:27

    Or suppose that Dennis Noble, Urin Alon, or other systems biologists had been working in the early 1900s.

    Art Woods 49:32

    Suppose that Biomedicine hadn't become a marauding capitalistic behemoth.

    Marty Martin 49:36

    If these counterfactuals were true, biology today would probably be a discipline much less focused on genes and much more focused on life, is what Stuart Newman calls Active Matter.

    Art Woods 49:46

    Let's make it so.

    Art Woods 49:48

    Oh and before we go as promised, here's a quick exchange between Cam, Marty, and me on agency. Currently Cam views agency differently than we do.

    Art Woods 50:19

    Okay, we got the three of us here in a room. It's me, Marty, and Cam. And we're going to talk for a little while about agency. We have a divergent set of points of view about the importance and utility of agency and sort of bigger broader issues in biology, like, you know, how much does the modern synthesis need updating? So, Cam, why don’t you go first and have at it.

    Cameron Ghalambor 50:41

    So I want to start off by saying that I don't deny that organisms are complex systems, and that agency exists. I absolutely agree that living systems have evolved to be robust, to exhibit homeostasis, to be flexible, to be plastic, to be self-regulating, that they interact with their environments in complex ways. And you can see this at different sorts of levels of biological organization. But what I'm struggling with is, I guess, like, my first question for you is, can you have agency without natural selection?

    Marty Martin 51:22

    Yeah, sure.

    Cameron Ghalambor 51:24

    Can you give me an example of how that?

    Marty Martin 51:26

    Okay, so this is where it gets interesting, and yet, maybe complicated and off page, first question.

    Art Woods 51:31

    Immediately

    Marty Martin 51:32

    Yeah first question, it's already off the rails. by natural selection, I'm assuming that you mean the one that Darwin pointed to the one that Darwin popularized? Or do you mean something broader?

    Cameron Ghalambor 51:44

    Is there another version of natural selection, aside from the one that sort of Darwin coined or?

    Marty Martin 51:50

    So Natural Selection, what I'm saying is that evolution can happen independent of life right now, no one would argue that I think you don't have to have living systems to evolve, you just have to have complex systems that change through time, one of which is life. Right. So Darwin didn't bother to talk about any other kind of complex system, he was only interested in the ones with fins and feathers and such.

    Art Woods 52:10

    But Marty, I think I may disagree with you on this point.

    Marty Martin 52:13

    Good

    Art Woods 52:13

    So do you think you can have agency without selection, no matter how you define the selection, like, to me, it doesn't actually really matter how we define selection, let's just take it as Darwin's kind of natural selection. And then try to use that to answer Cam's question. Can you get agency without that form of selection operating?

    Marty Martin 52:33

    I don't know that I can answer the question that way, because I have it in my head in such a different way. Let me just try really briefly to articulate what it is that I mean, and why I'm pushing back on the natural selection is not Darwinian. And you're inevitably anytime you get at the existence of a complex system, where you get a system that comes to sustain itself through time, the process by which that happens, as long as sustaining happens for long enough, if you ever get to a point of replication, you're going to have the instantiation of information such that the future generations of that system are different than the last ones. That's evolution by natural selection of any system, living and nonliving. That's what I'm talking about. Right? So Darwin happened to pick on living systems. But that process should apply to any kind of system that persists in time.

    Art Woods 53:18

    So you would say, for example, that like, we talked to Tim Lenton, about this, and he had this idea of, say, grasslands being a complex system, right, that have a lot of different components that are interacting in the grasslands, because of the way grasses affect fire dynamics and grazing dynamics. And, you know, the interaction between grasses and trees. Those systems are complex, and they're self-sustaining over long periods of time. So would you say that that ecosystem has agency?

    Marty Martin 53:47

    Yeah, I have a hard time with that. I think just because of my bias about organisms. Based on everything that I've said most of the people that we talked to, including Tim, I think they would say yes, that has agency. Because agency in the most generic sense, is the Karl Friston one of an updatable set of Markov blanket states. Right. And so as long as you have one, that reifies itself, because of the way that it's updated its state, that should mean that it has agency.

    Art Woods 54:15

    I suppose I could accept that. Although, it seems a lot more likely to me that it's, you know, individual organisms and parts within those organisms that are going to actually develop sophisticated forms of agency, because that agency itself is going to be shaped by natural selection.

    Marty Martin 54:32

    Well, yeah. Now natural selection, you mean the Darwinian one?

    Art Woods 54:35

    I do.

    Marty Martin 54:36

    Yeah. So I mean, I agree with you, but I'm uncomfortable in fully agreeing with you only because I can't get my head around why that should fundamentally be different, organismally and suborganismally, than at the level of what Tim was talking about/

    Cameron Ghalambor 54:50

    Well I'm still a little bit confused by non-Darwinian selection. I mean, I think, in the general sense when we talk about evolution by natural selection, we're referring to a set of conditions that when those conditions are met, then there's some predictable outcome. And those conditions are simply just that there's variation, some of that variation is heritable, and if some of that variation is associated with differential survival, or reproduction, then those individuals that you know, have higher fitness become more represented in a population.

    Cameron Ghalambor 55:33

    And so, in the most general sense, that can apply, you know, if we think about it as a levels of selection problem, something like a transposable element is a selfish bit of DNA, it doesn't have like a heritable component, in the same way that a, you know, multicellular organism would have. But if there is an element that tries to propagate and make more copies of itself, within the environment that it lives in, inside that genome, it has higher fitness, and it will propagate and increase. That's what happens.

    Marty Martin 56:13

    That's true. But I think what we're talking about is a different kind of thing. It's a much more inclusive thing, than, you know, the sort of temporal changes in lineages like transposable elements.

    Cameron Ghalambor 56:25

    Well, but you know, I guess that still confuses me, because you can think of a computer program as also something that could evolve. It has some sort of information, it's not living, but it still conforms to the same principles of, you know, Darwinian natural selection, even though it's not a living organism, those programs that have higher fitness increase, those that have lower fitness decrease. And so it's the same general concept, so that's why I guess I'm struggling with why that's different from, you know, any other kind of system that might evolve.

    Marty Martin 57:09

    So there's two big things in, in the example that you're using that stick out for me, one of them is that when we say that evolution by natural selection from Darwin is those big three, heritability variation and differential survival and reproduction. That's all true. But it doesn't drive home what Scott Turner really emphasized, where Darwin was really coming from and a lot of people around his time, it doesn't capture the struggle for existence component nearly well enough, meaning that it oversimplifies, how hard it is to be alive. Before you even get to the heritability of variation, it's just plain old, what is it to be alive that will allow any of those other three things to make a difference? And so the second piece that I think is really important in this space is why people traditionally pick on traditional thinking evolutionary biology, we don't have ideas about the origins of variation. We just say that there's variation, but coming from a general homeostatic complex systems mindset, you immediately for free, when you start talking about minimizing surprise and entry, reproduction or entropy exporting, you end up with an explanation for not just how much variation, but the sources and kinds of variety of information that so many people want to know. So it's just a richer landscape. I mean, it just does work that isn't offered by modern theory. But I'd be interested to hear what Art says about it. I think I agree with what you just said. I was hung up on thinking about the computer software program that Cam was just mentioning. And I was thinking maybe have a similar sort of analogy last night when I was anticipating this conversation. And that was I was trying to think about, like, you know, what would be a fair way to characterize the difference between maybe a modern synthesis view, and something that took a broader a broader stance/ And this comes from, you know, I think the comments that you've made Cam about this conversation in our writings to each other and what I've heard from other people. And that is, I think you could say, well, this agency stuff is all really interesting, and yeah, physiology interesting, but that's not what the modern synthesis was designed to do. It was designed to provide a very simplified quantitative statistical way of understanding how, you know, variation and filters on variation has translated into micro and macro evolutionary change over time. And it does that quite well.

    Art Woods 59:38

    But okay, so here's this analogy that I had in mind last night thinking about this. So imagine we're trying to explain the evolution of flying machines from say, the Wright Brothers plane up to modern fancy jets. What would the modern synthesis say about that sort of evolution and diversification of flying machines? It would be something about you know, the plans, the blueprints, the electronics that have diversified and become more complex over time. And we could even draw a phylogeny of airplanes that were based on what we know about how that information was transmitted among individuals and among companies. But that knowing that somehow still doesn't explain that much about how airplanes operate, you know, about where the Bernoulli effect comes from, about why the wings and the tails are, where they are, and how you steer the plane in the air. Like all this interesting stuff about what makes a plane a plane isn't captured by that sort of theory of transmission of plans. And so, I mean, maybe this is not a good example, because, like, what

    Marty Martin 1:00:46

    It's a great example.

    Art Woods 1:00:47

    But what I'm invoking here is not like different parts of an airplane that have agency, so much as saying, you know, there's this sort of narrow path that describes the evolution, but we want to know a lot more than that, because that's where a lot of the interest lies.

    Cameron Ghalambor 1:01:02

    So I agree with you. And I think this is a little bit reminiscent of some debates that happened, sort of at the interface of ecology and evolution between sort of the evolutionary explanations for patterns versus sort of more proximate, mechanistic explanations. And, you know, during the 50s, and 60s, I think there was a lot of debate about, just among ecologists, to try to explain the phenomenon that they saw. It was very strictly a mechanistic interpretation of what regulated populations, you know, through, for example, rainfall. And they viewed those explanations as very distinct from any kind of evolutionary explanation that. Liike ecology, and especially sort of like functional ecology was its own separate discipline and provide its its own set of explanations for how things worked. And you could think about it in the same sense of like a more mechanistic explanation for behavior, you know, an organism exhibits a certain behavior, because there's a stimulus from the environment that causes a hormonal response, which triggers some sort of neural response, which then, you know, eventually, through various complex pathways, results in a certain kind of behavior. Those are not evolutionary explanations, the evolutionary explanation for that would be the, you know, animal does that behavior because it gives it a fitness advantage. It doesn't say anything about the nuts and bolts about that.

    Cameron Ghalambor 1:02:37

    And I think there was a lot of debate about this through the 50s. And 60s, and Ernst Mayr wrote about this kind of cause and effect problem in biology, and in the end, you know, the consensus was that, look, we're studying the same things. And these approaches are not antagonistic to one another, they're complementary to one another. And we shouldn't think of them as being opposed, we should think of them, as you know, different ways of understanding. And so when reading Ernst Mayr's descriptions of the big meeting that sort of led to the coining of the term modern synthesis, he makes some comment about how they invited developmental biologists and physiologists to attend the meeting. And nobody was interested. It was because they really didn't see how-

    Art Woods 1:03:25

    They thought they were unrelated.

    Cameron Ghalambor 1:03:26

    Yeah, it just, it didn't cross their minds that you know, what you study at an organismal level, the mechanisms of physiology and development and behavior would have anything to do with these sort of population level evolutionary responses. And so I wonder if some of the debate and disagreement over the importance of agency and some of these other mechanisms like understanding the nuts and bolts of like an airplane, the physics, and mechanics, and the electronics are like, those are also very important, and those are very complementary. But those aren't going to be explained by a general theory of population genetics, that is thinking about the processes of genetic drift and mutation. And those all become, I think, what Marty calls the vicious abstraction. You can't see those mechanisms, but that you have to sacrifice.

    Cameron Ghalambor 1:04:25

    And I think that, you know, one thing that I know, both you and Marty have talked a lot about as kind of a different perspective is the systems level perspective. And I think systems biology and systems level thinking, that does offer a different perspective on the sort of traditional single locus to allele population genetic kind of simplification model because it does bring in a lot of complexity that isn't necessarily in those population genetic models. But people are studying that, there are evolutionary biologists, you know that that think about, for example, the concept of robustness. Robustness is not a concept that is easily pulled out of quantitative genetics or population genetics. But it's still a type of problem and phenomenon that's studied within evolutionary biology.

    Art Woods 1:05:21

    I just first wanted to make a comment about this idea of a vicious abstraction, which we talked a lot about in the script that we just read, and that you just brought up. And to be clear, there's nothing wrong with simplification and models. And in fact, that gives them enormous amounts of power. Right? I mean-

    Marty Martin 1:05:38

    That's the whole reason, yeah

    Art Woods 1:05:39

    That's essentially the philosophy of doing a model is, you know, what's the minimal amount of stuff you can write down to capture something that's important and of essence, in a system. And like, you know, another sort of non evolutionary example of that would be thinking about, well, what are good mathematical models of how populations fluctuate over time. And, you know, there's the exponential growth model, which has just a very simple equation that underlies it. There's the logistic growth equation, which adds in just another couple of terms that allow you to incorporate density dependence into an exponential growth model. And those have been vastly powerful, precisely because they're really vicious abstractions. Right.

    Art Woods 1:06:18

    And so, and I think the way to relate that back to what you're just talking about, about evolutionary theory is that, you know, the modern synthesis is based on these abstractions that are super powerful, because they boil it down to some kind of essence, that really matters. And I think one path would be to say, you know, this stuff about agency and physiology and homeostasis is just something else. And, you know, it's not what the modern synthesis was designed to explain. And the problem with that, I think, is that what I can see glimpses of is that this complex system stuff and agency and homeostasis feel to me like they feed back on that evolutionary process in a really important way. In other words, by doing the vicious abstraction, you've actually lost sight of something that's not tangential, but something that's really central to the way lineages evolve and diversify. And so it feels like we need to kind of bring that back into the sort of the main thread.

    Marty Martin 1:07:14

    Yeah, that's where I would end because I mean, to first to be fair and clear, that's not my word, I'm not that creative. That was William James, particular word, but it wasn't him that claim that necessarily

    Art Woods 1:07:23

    Vicious abstraction?

    Marty Martin 1:07:25

    Yeah, he was making the general point about, you know, when you're trying to model something, by its nature, you want to simplify it, but I think it's, it's at core, it's been really good, you know, really generous of you to include all of this text in the script and the other emails and things Cam, because I think at the bottom, we're interested in different things. Like, it's not surprising to me for you to say that you believe that agency is real, and all those sorts of things. And I think most evolutionary biologists, not everyone, but most of them would go along with agency, if agency is defined in a really specific way as to be, you know, the kinds of things that help organisms not fall apart or help systems not fall apart.

    Marty Martin 1:08:04

    But I think that the piece that's a little bit weird, what Art is alluding to, it's not so much that the modern synthesis was wrong. The question is whether the original vicious abstraction is the best, most effective abstraction. And I just think that it can't be. Exactly what replaces it, now, that's a totally fair question. And I think, you know, when the people are kicking the modern synthesis and people saying, what else do you want, just listing terms, niche construction or epigenetics, that's not enough, because that's just a laundry list is something else that isn't necessarily incompatible. So it is the responsibility of those that wants something different to articulate what's different. And I think that's, you know, that's why we did this thing with agency. It seems to be because most of the stuff are systems, like they're all systems. It is really interesting to me to hear you say that you are cool with robustness, and systems thinking. But since some of this other stuff gets a little bit wonky, you're drawing the line in a different place than I can understand, because I don't think it's really all that different. But at the end of the day, wouldn't you say that we all want a vicious abstraction, we just want a different one? Or is there value in a different one?

    Cameron Ghalambor 1:09:10

    Well, I think that any kind of model that improves our understanding of how things work is something that, you know, who wouldn't want that? Everybody wants that. And I think that you know, my problem with a lot of the people who you know, beat up the standard evolutionary theory and say like, well, it doesn't include epigenetics, or it doesn't include niche construction, it doesn't include plasticity, even though I study plasticity and I find plasticity perfectly compatible within the context of, you know, the standard evolutionary theory. What I haven't seen articulated is what changes. So at the core of like, what what we would call standard evolutionary theory are certain processes, you know, those include mutation, genetic drift, gene flow, recombination, selection, and then you can study and you can model and you can empirically measure interplay between gene flow and selection.

    Marty Martin 1:10:21

    Well, hold on, let me add something else in here, because I guess I think I didn't finish the thought before. It's not just that we want a model that includes a vicious abstraction, because useful models have to have vicious abstractions. What is it that we're trying to model? What I'm talking about is not the same thing, because modeling evolutionary change, especially when the history of those kinds of models have been about genetic change, right? This kind of mentality, whatever this vicious abstraction is going to be will be something about the viability, the sort of sustainability of a system. Now that can include reproduction, right, and evolution, subsequently, but it doesn't have to. So we're talking, we're coming up with a model that explains the existence, like the origins of life, the persistence of life, and variations into the future, on life, based on these Markov updates, and that kind of thing. But it's not quite the same thing.

    Marty Martin 1:11:17

    So you know, gene flow and drift, and those kinds of things aren't necessarily even included and why I'm picking on this. And I want to hear what you think about it, knowing that this is the like, to what end are we modeling, I just really strongly feel that even though the modern synthesis was not intended to have the impact on biology that it has had, it's time to confront the fact that it has, that biomedicine, for example, spends an enormous amount of money, so much more money on things genetic than things other, right? And so practically, it's time to stop living in the modern synthesis world, or tacitly representing it as our backbone, because people's lives are at stake.

    Cameron Ghalambor 1:12:15

    So I think, part of my disagreement then is holding up this thing, that people keep referring to "the modern synthesis," which is, you know, it's like evolutionary biology had the synthesis 50 years ago, and nothing's happened ever since, which is how it's often depicted.

    Marty Martin 1:12:36

    Yeah, that's unfair. That's you're right.

    Cameron Ghalambor 1:12:38

    So I think a more productive, maybe way is like, within kind of what we would think of as standard evolutionary theory, what are the explanations for what I think your first step that you're very interested in is the origin of life. Like, I know, for example, you've interviewed people like Sarah Walker and Nick Lane and, you know, does RNA come first? Or does metabolism come first? The population genetic based theory doesn't say anything about that. There may be some other theory out there that maybe invokes a role for natural selection, in favoring one variant over another variant of, you know, which one may be more effective. But I think that's more kind of a discussion for biochemists and chemists, and biologists

    Marty Martin 1:13:36

    I disagree

    Cameron Ghalambor 1:13:36

    Well you know, it's certainly more at the interface. And so then , if we want to then move beyond that, to say that, you know, successful systems are those that do a good job of replicating themselves, that can keep themselves alive and persist entropy and maybe modify their environments in ways that make them survive better and are more suitable for themselves? I think that's all fine. I think articulating a model or a theory for that kind of process of life that's very, very general, that is totally fine. And I think standard evolutionary theory maybe can be incorporated into that, to talk about, you know, how things go in one direction versus the other or, but I don't, I see that as a, again, more of a complementary type of theory. It's maybe most evolutionary biologists are focused on microevolutionary change. But you know, I can say going to like the evolution meetings, you don't see people giving talks, for example, or presenting posters on, at least it's not very common, unlike, Did RNA come first? Pr did metabolism come first? Like certainly that's a big evolutionary question, especially, you know, in the history of life.

    Marty Martin 1:15:06

    But that, at bottom, is the thing. There's a bunch of people, I think Art and I are really physiologically inclined. I think we get jazzed about this, because the number one thing for us when you talk about any living system, is that it's not dead, that there has to be homeostasis. I mean, it's silly

    Art Woods 1:15:23

    I'm not dead yet

    Marty Martin 1:15:25

    That's the core- I'm not dead yet. But that's the core and evolutionary biologists, to me, bizarrely, and I am one because my PhD says it, but they don't care, necessarily, that life resides in those equations. And that's just perplexing, and I don't think even viable in 2023, but aren't what do you think you've been quiet?

    Art Woods 1:15:46

    Well, I want to go a different direction. So what I hear you guys arguing about is, you know, whether we should localize this argument onto a sort of more standard view of how we view evolution right now, among living organisms using the mechanisms that we understand versus taking this approach that Marty is advocating that sort of goes beyond evolution of individual populations, and, you know, genes, and differential survival. So to sort of thinking about complex systems and their origin more, more broadly.

    Art Woods 1:16:18

    I want to circle back to sort of how I feel like some of these ideas affect my thinking about more standard use of evolution. And this comes out of just a couple days ago. So Alisha Shah invited me to Kellogg Biological Station, I gave a talk there a couple of days ago and did a lot of thinking about all of these ideas interface with my own work on thermal ecology and thermal physiology of insects. And I was struck again by, I guess, this idea that maybe was articulated most strongly by Scott Turner, that what's interesting about agential thinking is that if you think about how populations evolve, you know, what is there? There's some kind of variation, there's some kind of filter, and then some subset of that original variation makes it through that filter and becomes the subsequent population that's reproducing. And that's a very kind of microevolutionary view. It's a standard thing that we all say, in our basic biology classes. And what agency adds to that is that it means that the organisms that have the variation are also, by their actions, you know, creating and modifying the very filter that is selecting on them. And that to me feels like, I mean you could say that's a small thing. And we could explain that as you know, the fact we've known forever that, yeah, organisms interact with their environment, organisms have behavior. But I think that underplays this sort of fundamental importance of organisms, both, you know, at some level, creating the variation, and modifying and creating the very filter that's doing that selection. And that feels to me like a sort of profound shift in thinking about, about where variation and, you know, potential causes of change come from? Discuss.

    Cameron Ghalambor 1:18:03

    Yeah, so I think that is a, I think that is a more sort of rich and productive way of thinking about things. But again, that kind of thinking is also captured, for example, within the, what people call eco-evolutionary dynamics. So, like the guppies that I studied, worked by Dave Reznick, John Endler, and others, but Ron Bassar and Joe Travis recently, have shown, for example, that when you move guppies from a river that is full of predators, that keeps the population density of the guppies down, you get a certain life history that evolves sort of a live fast, die young kind of strategy, and whether guppies naturally colonized or experimentally are put into these streams, where these predators are absent, they modify their environment, but not always in a good way. In the absence of predators, the populations grow, they become, you know, very high density kinds of populations. And you would think that maybe in the absence of predators, this is sort of a nice, happy paradise. But in fact, it's kind of a nasty place to live. You're in these very small streams under a closed canopy with very little productivity, and the population densities are high, so it's a very competitive, very nasty kind of environment.

    Cameron Ghalambor 1:18:14

    And the sort of previous thinking was that the evolution of the sort of low, the more slow life history was simply just due to the absence of mortality from predators. But now it's appreciated that it's much more driven by the density dependent kind of competition for food and the sort of nastiness that occurs in these kinds of streams. So, and of course, that then changes how the guppies evolve, right, they go from one situation to the other. And I think that's where the concept of agency comes in. I worry about that then becoming a bit circular, because if guppies were not to, weren't able to adapt and evolve to these new sets of conditions, they would go extinct. And the fact that they don't go extinct means that there's been some sort of compensatory evolutionary change in the life history, in the physiology, in the behavior that allows them to persist. And so the system, in this case, the population level, persists in this new environment, and it's been shown that this can evolve very quickly. And it's been replicated many, many times. So is that kind of interaction between the guppy and its environment, agential? I would refer to that as either density dependent evolution or eco evolutionary dynamics. We could also call it agenital behavior, but I think all three are capturing the same kind of interaction between the organism, the environment, changing the selection pressure, acting on that variation, and this more kind of feedback loop.

    Art Woods 1:19:35

    Yeah, yeah.

    Cameron Ghalambor 1:20:06

    Rather, maybe a more linear sort of response?

    Art Woods 1:21:18

    Sure, sure. I mean, I think honestly, that's a super beautiful example. And you know, to me that that's a great example of organisms sort of creating the conditions themselves that are then selecting back on them. And density dependence is clearly an example of that, that happens not only in guppies, but in many other groups. But like, to me, what's important about this idea of agency is that it allows me to see that that kind of density dependent selection that you just described, is a subset of a much broader set of things that organisms are doing, you know, exploiting affordances in their environments, and taking advantage of opportunities and avoiding threats. And you know, it's not just guppies creating dense, nasty, competitive streams, it feels like this is a characteristic of life everywhere, and that that agential thinking somehow sort of unifies those various ways by which organisms are interacting with their environments.

    Marty Martin 1:22:14

    Yeah, I mean, it becomes very specific, right? The amazing thing about agency is that agency is directed at something, it's for something, I know teleology we're not supposed to simplify such things, but it is for something, which means that the kinds of things that you could expect to find in populations and subsequently how they're going to evolve, it's not just anything, right? There's going to be particular paths that lineages can take, because of the kinds of challenges that they have this, again, this variation, it doesn't just become any old thing, it becomes that thing, that physiological problem that agency is for.

    Cameron Ghalambor 1:22:48

    Yeah, and again, I mean, so like, at least in the literature that I read, you know, I see people talk about how contingent is evolutionary change based on the past like, that is a topic that is thought about, you know, what are the constraints? What are the trade offs? How do those dictate the direction that evolution will go, you know, go toward. Is there a bias because of some kind of historic sort of constraint like that? And so, you know, if the concept of agency can be brought into that type of thinking, and say, like, you know, given what we know about this particular lineage, for example, and agency is one of the other kinds of baggage that they bring with them, you know, sometimes it's that baggage can be not so good, you know, because it really prevents you from exploiting certain kinds of environments and evolving in certain directions. But then agency could certainly be thought of as a way of, for example, facilitating evolutionary change into, you know, for example, colonizing new environments because of something that's there. But you mentioned the word teleology, and teleonomy, like, I think that's where a lot of people get uncomfortable within the evolutionary community, because then it implies that there is some, you know, known outcome that will happen. And I think, repeatedly, the evidence suggests that, you know, there isn't necessarily going to be evolutionary progression towards, for example, greater complexity or greater than a gentle sort of behavior. I mean, we see, for example, that there are some lineages that have more or less been unchanged for, you know, millions and millions of years, living in very constant, usually, kinds of environments with no change. And then, you know, we've seen other lineages that have like, diversified and expanded. And so-

    Marty Martin 1:24:57

    I don't think agency would necessarily predict progress to complexity. As long as the system remains viable, it doesn't need to become anything different. If the world is predictable, it maintains the same old model, and I think that's totally fine. But if and when the world does change it has to act back or it has to update its model. I don't think that those ideas are incompatible. But bringing up this idea, it reminds me to ask you about something that you mentioned a couple of different times. And I think it's a part that I feel a little bit responsible to touch. Do you think that talk? Or where do you think in 2023, the field stands with agency as some kind of spooky force? Like, do you feel that biologists have to be extra careful about using words like this, because of the potential influence it has on folks that are against or have different agendas, like intelligent design, that kind of thing?

    Cameron Ghalambor 1:25:44

    Well, I'm not an expert on these kinds of ideas. And I'm sure there are other people who've thought about this a lot more. I can only say, from my perspective, one thing that makes me very uncomfortable is that most of the papers that I've read that deal with agency, are funded by the Templeton Foundation. And so within that group within that set of researchers and philosophers and biologists who are funded through that agents

    Art Woods 1:26:15

    Through that agency?

    Cameron Ghalambor 1:26:17

    Through that agency.

    Marty Martin 1:26:18

    Agency agency?

    Cameron Ghalambor 1:26:19

    You know, so does that mean that it's just a little bubble of people that have similar thinking and are all in agreement with one another, and there's nothing more than? That could be the case. You know, what makes me uncomfortable, and I think makes a lot of other people uncomfortable is this way of depicting evolutionary biology as being somehow in crisis and that there's a real problem here. And there's some fundamental problem that needs to be corrected. And whether it's, you know, what people are calling the extended evolutionary synthesis or using new terms like agency, like so within biology, I don't think this is a big deal. But does this open the door for the general public to then say, look, biologists, evolutionary biology, is all wrong, or has been wrong and is in this crisis mode, and that somehow gives the perception that it gets discredited, it shouldn't be trusted, because those guys still haven't figured things out, that gives me strength to push maybe a creationist or intelligent design kind of agenda. Those are concerns. I mean, those are there are people out there who would exploit these kinds of disagreements in that kind of way. So I, you know, I think it's extremely healthy to we don't want everybody to have the exact same views. We want people to have a diversity of views.

    Cameron Ghalambor 1:27:51

    Wouldn't be science, would it?

    Cameron Ghalambor 1:27:52

    Exactly. You know, and, and we want to encourage that. And so, you know, I think the challenge here is articulating and incorporating a view of agency that is compatible with and appealing to those people who are actually doing evolutionary biology research, and how would you incorporate it into, you know, you're studying guppies or insects or birds? Would it fundamentally change the kinds of experiments you design, you know, the assumptions that you're making the interpretation of your results in ways that the standard theory wouldn't do?

    Art Woods 1:28:31

    I do hear your concerns. And I mean, it's unfortunate that you know, a lot of the people talking about agency are also funded by, you know, this agency that could raise some doubts. I mean, I agree with you, there Cam.

    Cameron Ghalambor 1:28:56

    Earlier, you said that a grassland ecosystem could have agency, it's hard for me to see how you use the term without cognition.

    Marty Martin 1:29:07

    Cognition, for sure, because any complex system with the ability to update is learning. So that's just that's cognition.

    Art Woods 1:29:14

    But I think you guys may differ on this meaning of cognition, too, right?

    Marty Martin 1:29:18

    Oh, probably. Yeah, I mean, the last, the last part of the conversation we focused on the thing we just recorded was with Mike Levin, that was it called cognition all the way down, Art?

    Art Woods 1:29:27

    I thought it was agency all the way down.

    Marty Martin 1:29:29

    Agency all the way down. But cognition is the main word that he uses in it.

    Art Woods 1:29:33

    And the people that are invoking cognition this way are generally careful, while they're rare, probably for good reason. And they're very careful to distinguish it from consciousness, right? So there's nothing about cognition that requires consciousness awareness

    Marty Martin 1:29:48

    Awareness is secondary, correct? Yeah, yeah. Well, guys, we've been going on for a little while, and I've gotta go in a few minutes. What are the like main landing points that we want to hit? Cam, is there something burning that you want to say?

    Cameron Ghalambor 1:30:01

    I think one thing that I wanted to say was that one thing that I've really appreciated about Big Biology as a listener, and now joining as a co-host, is interviewing people with diverse opinions, for sure, even if I don't agree with them, but also interviewing people who are also interested in the philosophy of biology. And I think, you know, most working biologists don't read philosophy of science as much as they probably should. And I think that's a really great thing that big biology has done to introduce working biologists to this kind of philosophical approach. But, having said that, I think, and I know I've heard you and Art ask these questions in the past is, it's one thing to talk about philosophical ideas, as they relate to evolution and agency and everything. It's another much more difficult task to then bring those ideas into sort of the working nuts and bolts of what people actually do

    Marty Martin 1:31:12

    - Yeah absolutely

    Cameron Ghalambor 1:31:12

    -in terms of their research. And I know you've, you've asked that question in the past of some of the guests of like, how do I incorporate that into my own research?

    Art Woods 1:31:19

    So now you're pinning that on us? I see.

    Marty Martin 1:31:22

    Yeah.

    Cameron Ghalambor 1:31:22

    And so now, I would just say that I look forward to another hundred episodes engaging with both of you in conversations about agency and the complexities of life and how life evolves. Yeah, there's a lot of room for different ideas and a lot of work that needs to be done.

    Art Woods 1:31:46

    It's a great place to wrap it up. I think, honestly.

    Marty Martin 1:31:47

    Great place to wrap it, yeah. Did you get everything that you needed to, Art?

    Art Woods 1:31:51

    Yeah, I think I'm fine.

    Marty Martin 1:31:52

    Okay, good.

    Art Woods 1:31:53

    That was fun.

    Marty Martin 1:31:54

    It was fun.

    Marty Martin 1:32:03

    Thanks for listening. And if you like what you hear, please tell a friend, mention us on social media,

    Art Woods 1:32:06

    or if you're feeling really generous, remember, we're a nonprofit. We always welcome donations to help support our production team, and especially our student interns,

    Marty Martin 1:32:14

    and a special thanks to our dedicated fans for helping us reach this 100th episode. Without your support and enthusiasm. We'd never made it here.

    Art Woods 1:32:21

    Thanks to Steve Lane who manages the website and Ruth Demree for producing the episode.

    Marty Martin 1:32:25

    Thank you as well two interns Dayna De La Cruz and Kyle Smith, who helped produce the episode. Katie Shahmehri does our awesome cover art.

    Art Woods 1:32:32

    Thanks to the College of Public Health at the University of South Florida, the College of Humanities and Sciences at the University of Montana and the National Science Foundation for support.

    Marty Martin 1:32:42

    Music on the episode is from Podington Bear and Tieren Costello

Big Biology