106 | Stuart Bartlett on What “Life” Means

Someday, most likely, we will encounter life that is not as we know it. We might find it elsewhere in the universe, we might find it right here on Earth, or we might make it ourselves in a lab. Will we know it when we see it? “Life” isn’t a simple unified concept, but rather a collection of a number of life-like properties. I talk with astrobiologist Stuart Bartlett, who (in collaboration with Michael Wong) has proposed a new way of thinking about life based on four pillars: dissipation, autocatalysis, homeostasis, and learning. Their framework may or may not become the standard picture, but it provides a useful way of thinking about what we expect life to be.

Support Mindscape on Patreon.

Stuart Bartlett received his Ph.D. in complex systems from the University of Southampton. He is currently a postdoctoral researcher in the Division of Geological and Planetary Sciences at Caltech, and was formerly a postdoc at the Earth Life Science Institute at the Tokyo Institute of Technology.

• Web site/Blog
• Caltech web page
• Google Scholar publications
• ResearchGate page
• “Defining Lyfe in the Universe: From Three Privileged Functions to Four Pillars,” Bartlett and Wong (2020).

[accordion clicktoclose=”true”][accordion-item tag=”p” state=closed title=”Click to Show Episode Transcript”]Click above to close.

0:00:00 Sean Carroll: Hello everyone, welcome to the Mindscape Podcast. I’m your host, Sean Carroll. You might have noticed that we’ve talked a lot in the podcast about life, about the scientific concept of life, especially in the context of possible kinds of life elsewhere, not here, right on earth, or at least not naturally occurring right here on earth. We talked to Kate Adamala about synthetic life, about actually building living organisms in the laboratory. We talked to Sara Imari Walker about information and what that has to do with life, how living beings use information but also how information helps us define what it means to be a living being. And we talked to Kevin Hand about actually looking for life elsewhere here in the solar system. In all of these discussions, we’ve mentioned the fact that we don’t have a once-and-for-all definition of what life actually is, but we sort of mentioned that, mumbled about it, and then moved on to something else.

0:00:55 SC: So let’s take the opportunity to actually dig in. Today’s guest is Stuart Bartlett, who’s a researcher here at Caltech at my own university. And he’s an expert in what life actually means, what the word life means. He and his collaborator, Michael Wong recently published a paper that actually goes through all of the different features of what come into being a living being, and make the point that some of these features might not exist in every single thing you wanna call life. So they actually define some new words, and they hypothesize about which of these features such as learning and homeostasis would come first, which are more important, and then which would come later. And the answer is we don’t know. It’s a wonderful little organizational tool to remind us of how much work there is to be done, to really understand what it means to undergo a transition from simply some complex chemical reaction to something we would actually call life.

0:01:53 SC: And it harkens back to the various talks that we’ve had. Information plays a huge role here. The possibilities of building life in the laboratory are mentioned. We go pretty deep, Stuart and I in this conversation. We talk about life on computers, surprises in evolution, a whole bunch of good things. So one of those mind is stretchy episodes and possibly one that’s gonna be really, really relevant to science in the near future, as our ability to look for life elsewhere, both in the solar system and on other planets, on other stars, becomes better and better. So there’s a good one to make your brain grow and also prepare us for the future. What else do you look for in a good podcast episode? So with that, let’s go.

[music]

0:02:50 SC: Stuart Bartlett, welcome to Mindscape Podcast.

0:02:52 Stuart Bartlett: Thank you for inviting me. It’s a great honor to be with you today. Thanks.

0:02:56 SC: I wish more people would recognize what an honor it is, but I’m glad that you put it that way. We are honored to have you here, especially… I’ve talked a few times on the podcast about life, the definition of life, and looking for life in other planets, the origin of life. But it’s always been in the service of something else. We talked with Kate Adamala about synthetic life and Sara Walker about the relationship of information and life. And Kevin Hand we talked with about looking for life here in the solar system. But so, lurking in the background here something we need to bring up every single time is, “What do you mean by life? What is the definition of this?” So I’m glad we’re finally going to answer this once and for all. But maybe I can start by just asking, “Do we need a definition of life? Is this something that is just arguing over words or is it really crucial to scientific understanding?”

0:03:50 SB: Yeah, that’s a great question. And you can really argue this either way. Of course, the most conventional affirmative answer is that we need a definition of life so that we can design, plan, strategize and aim towards a common objective. So much of modern science is objective-driven. When you write proposals and you design missions, you have to have a very, very clear goal in mind, and a lot of what we do is taking incremental steps towards those goals. And Astrobiology, even though it is, let’s say at least 50 years old, it’s a field which has a broad scattering of opinions and a complete lack of consensus on almost anything.

[laughter]

0:04:43 SC: There are a lot of opinions, I’ve noticed that out there. There are definitely people with strong opinions about things.

0:04:47 SB: Exactly, it’s one of those situations where the less something is constrained, the more sure people feel about their opinions on it. [laughter] And so one could say that we should have a definition of life so that we’re… At least to a certain extent agreed on what we’re looking for, especially when we design missions, looking for extraterrestrial life. And also in the context of the origins of life, a very difficult problematic question is, “What would qualify as a successful demonstration?” So for example, if someone did an experiment in the lab and they said, “I have proven that life began in this way.” We have no… In the Astrobiology field, we have no agreed-upon standard of what constitutes absolute proof of an abiogenesis number two. I mean, NASA has its definition, which is probably the most commonly adhered to definition, but there is still…

0:05:57 SC: And so tell us what that is.

0:06:00 SB: Yes, the NASA definition of life is a self-sustaining chemical system capable of Darwinian evolution. An evolution or Darwinian evolution in particular, features heavily in most definitions of life, and… In the early days of origins of life research and still to a stronger extent today, there’s this almost tacit assumption that whatever your prebiotic system is, once it shows definitive signs of Darwinian evolution, then you can probably leave the rest to be discovered and constructed by that evolutionary process. So, it’s sort of tacitly hoped that evolution is the explanation for the complexity of life and you just have to get to that starting point and then evolution will take over for you. I don’t share that opinion. I don’t know whether… I’m unsure as to whether just Darwinian evolution is sufficient to explain the diversity of life on earth, so I’m not confident that we can rely on Darwinian evolution to explain the complexity that we see after that origins starting point.

0:07:20 SC: Okay, wait a minute, so these are fighting words, obviously, I can’t just let you get away [chuckle] with saying that. [chuckle] So, maybe you can define exactly what you mean by Darwinian evolution, ’cause I suspect a lot of work is being done by that phrase? And then, in what ways do you think it’s insufficient?

0:07:36 SB: Yeah, so the first course that I took on evolution was taught by Professor Richard Watson at the University of South Hampton. And I’ll never forget, he would summarize it in a very compact way in the following phrase, “heritable variation leading to reproductive success.” So, there’s basically three components to Darwinian evolution, there’s heredity, so the ability for information to be passed from one generation to the next. It has to be a variation. So, when an organism reproduces, the next generation has to consist of a finite number of individuals, and they cannot all be identical. And we tend to say that if they’re identical then that’s replication. If there’s variation then it’s reproduction.

0:08:28 SC: Okay.

0:08:28 SB: So there’s variation among the offspring, and then there are selective forces which will tend to favor those members of the new population which are more fit to the… Better suited to the environment and it will disfavor those which are less suited to the environment. And of course, this is the theory of biology. And Darwinian evolution has been studied in all kinds of different ways beyond just phylogenetics and studying the modern biosphere. In the field of ALife, people study the process of evolution through digital evolution experiments. So there are some…

0:09:28 SC: You should probably define what ALife is.

0:09:32 SB: Yes. So ALife is a very interesting discipline. ALife stands for artificial life, and by artificial life we mean alternative forms of life in some sense. So, the research in artificial life consists of things like artificial intelligence, and robotics, evolutionary robotics. So, people… When people design robots, maybe they can design it and hand program it, and it works, but often when you’re trying to get a robot to solve difficult tasks, you have to basically evolve the software, and often these robots become “intelligent” in their own right, and you basically evolved their programming. So maybe you have several different sets of programming for your robot, and you see how well it achieves the desired task. And then, you rank their performance, take the best ones, and then maybe re-combine the features of the best ones and then reiterate that process. And you basically mimic the evolutionary process in the programming of your robot and eventually the performance should increase.

0:10:47 SC: Right. Okay, so even in the artificial life context, people are using Darwinian evolution, but you wanna say it’s not sufficient to do something. Tell me what it’s not sufficient to do.

0:10:58 SB: Researchers in the field of ALife have for a long time been seeking this grand goal of what they call Open-Ended Evolution. An open-ended evolution means evolution of a population of individuals, which just keep changing and keep becoming more complex over time, and show no sign of sort of settling and saturating. And when we look at the biosphere on earth, it looks like evolution of earth-based life is open-ended, because there’s no sign that evolution ever got bored and stopped innovating and generating new levels of complexity. So, if we understand the evolution of life on earth, we should be able to… As Feynman said, “When you understand something, you should be able to reconstruct it.”

0:12:02 SB: And so, if we understand evolution, we should be able to reiterate it in a way that shows this open-ended generation of novelty. And so, researchers in the field of ALife have built a large range of models where you have artificial organisms which reproduce with heredity and variation, and then they impose selective forces on those populations. In many of the examples, these are basically digital organisms, they’re just sets of code and they’re actually competing for CPU time and memory access. And the idea is that if we let these populations evolve, if the process is similar to evolution of life on earth, then we should see emerged and novel features. So maybe, for example, they form groups and the groups start competing or they form new… They invent new ways of existing in their digital environment.

0:13:07 SB: And so that’s what would potentially count as open-ended evolution. But in reality, what tends to happen is that those systems do evolve to a certain extent. So there are changes in the code and they do develop certain functionalities, but what tends to happen is, you have one set of those entities, which are able to reproduce themselves quite well. But then you have a set of parasites which come along with them, which emerged in the system, and they will essentially piggyback on the more stable replicators and those more stable replicators then find it difficult to generate new levels of novelty because their resources are being expended, not only on their own replication but on the replication of just those small bits of code which are a bit like viruses in some ways. And what you find is that the complexity of the system basically just saturates out at that point, and it is not able to develop further.

0:14:13 SB: And there was a similar experiment by Spiegelman on small sections of RNA. And you can ask the same question. If you have RNA molecules, which are able to perhaps replicate themselves in a collective sense. Again, you have a replication process or a reproduction process. And there is also heredity because there is information in the sequence of the RNA molecules. And there’s also a selection for reproductive efficiency. And again, one would hope that if you let the experiment run, you would see emergent, new levels of function or complexity. But it turns out that actually, those RNA molecules evolve towards simplicity, not towards complexity. And actually, they end up evolving towards the minimal sequence capable of replication. And then they stop there.

0:15:10 SC: This is fascinating. I’d actually never heard of this idea that in the artificial life context, complexity saturates, and furthermore, it’s because of these parasites. So, I’m not quite sure what the causality is there. Why does the existence of these parasites cause the main artificial virtual life forms to stop innovating?

0:15:30 SB: Yes, so the in general, the answer seems to be something called the cooperation barrier, which is that when a replicating system is sort of headed towards, let’s say, a higher level of organization. For example, if there is a set of simple replicators and we believe that there’s a way for those simple replicators to cooperate, and perhaps share resources. Maybe some of the replicators are more specialized at some functions than others. And we know that if they cooperated and shared those functions, then the collective whole would actually be able to replicate at a more efficient rate. We can ask the question if there is some kind of collective favorability for that cooperation, what does it take for that cooperation to emerge, if we know that it’s favorable? And it turns out that this cooperation barrier is the fundamental issue. So, let’s imagine that by some fluctuation, these originally selfish individuals begin cooperating and they discover this higher level of organization which is beneficial.

0:16:56 SB: If one of those individuals in the cooperating group reverts back to a more selfish way of living where it takes its share of the resources but doesn’t contribute a similar level of resources to the group, then in the short term, that individual will be favored. Because, it can basically steal and reproduce itself and without having to wait for the whole collective to reproduce itself. And what you find in these experiments is that those… We sometimes call them cheaters or selfish individuals. Those cheaters, in that short term favorable situation where they can basically steal resources from the group, they will tend to out-compete the cooperators and take over the group, and then you go back to the previous level of organization without the cooperation. And so this is something that’s also studied in evolutionary game theory.

0:18:00 SC: Yeah.

0:18:01 SB: And, it seems to apply to those digital evolution experiments and perhaps even to some of these RNA replication experiments. And in general, on the open ended-ness side, people have looked at the history of life on earth and tried to delineate what they call major transitions. That idea is due to Erin Seth Mary and John Maynard Smith, where we see changes in the organizational structure of life. So the transition from single-celled organisms to multi-celled organisms is one example. And so, in that transition, it’s similarly an issue of why would those previously individual selfish organisms decide to give up some of their own resources and cooperate for that collective favorability because often it takes some time for the advantages of being in a group to be reflected in the selective forces and…

0:19:11 SB: So this is sort of a generic problem in evolutionary game theory and artificial life and there are different theories out there for how you combat this cooperation barrier, so some people argue that there has to be some kind of regulation mechanism or a suppression mechanism to try and suppress the cheating behavior. And we can also draw analogies with society itself. With society itself there’s a also a short-term selection pressure for cheating and stealing. In theory there are ways that you can cheat and steal from the group and in the short term you will be better off. But the group in the long term will be worse off, and we have to try and suppress that uncooperative behavior with dedicated mechanisms for that.

0:20:03 SC: Okay, so fine, so what you’re saying is that at least in our artificial experiments, Darwinian evolution all by itself or the analog thereof seems to reach this bottleneck or blockade or fixed point equilibrium that doesn’t look like what life looks like here on earth. So how do you fix it? Are you gonna say that God comes in and stirs things up and makes us more complex?

0:20:26 SB: Yes, well that’s the question that everyone is trying to answer and so, again, people suggest that you need some auxiliary mechanism which is able to… Which is able to suppress the cheating behavior or you have to have selective forces at the group level which are strong enough. And in Biology there’s still a lot of debate about whether group selection really exists, ever since Richard Dawkins I guess, we tend to focus on the selection of fairly low level entities like the genes or the individuals, and so many biologists are skeptical about the possibility that there can be true selective forces acting on the group. But if we look at it at that level, let’s say we have several groups of individuals which are replicating, we could make an argument that out of those groups, the ones which are best able to suppress the cheating and best able to cooperate should probably out-compete the groups which are being dragged down more significantly by the cheating. So it has been a sort of intuitive level group selection, it feels like group selection could be acting and…

0:21:49 SC: But is it fair to say that that’s not Darwinian evolution? Isn’t that something that is just part and parcel of Darwinian evolution?

0:21:57 SB: Yeah, that is a good question, is whether it’s… You could say that it’s maybe a higher manifestation of Darwinian evolution or you could say well, it’s something different. And in fact, going back to my lecturer Richard Watson, he has proposed that actually evolution is more like a large associative learning system and that can potentially explain the formation of these groups and that basically the whole system, whole ecosystems for example, are essentially learning about their environments and learning about one another and they’re learning by observing correlations in the environment and amongst each other. And gradually the populations are becoming more knowledgeable and have a greater understanding of their environment, and that is a more general explanation of evolution, that it’s basically this giant learning algorithm. There’s also another author called Leslie Valiant who’s discussed these ideas in his book “Probably Approximately Correct”.

0:23:20 SB: And of course, it’s not trivial to figure out, if you say, “Well, ecosystems or the biosphere as a whole is a giant learning system, it’s not clear how exactly it’s doing the learning.” Again, Richard has argued that it’s associative or heavy in learning, correlation learning so, which is where the system is observing correlations in external variables. For example, I don’t know, if the sky becomes dark and cloudy it’s quite likely that then it’s gonna rain, and if you’re an organism that really doesn’t like rain, then you probably wanna get out of the way before the rain arrives and so it’s advantageous if you correlate those two things and you can get out of harm’s way before the danger actually arrives, then you will be better off than if you didn’t learn that association.

0:24:18 SC: It’s very humian in some sense.

[chuckle]

0:24:21 SB: Yeah, yeah.

0:24:23 SC: But wait, and so I think… I actually, I feel bad because we’ve gotten off of the question of the definition of life and I do wanna go back there, but this is just so fascinating that, you know, that’s okay, we should talk about fascinating things. But nothing you’re saying makes me think that there isn’t some purely reductionistic level at which we talk about genes being passed on to other future generations and then some of them reproduce and some of them don’t. And I’m very happy with there also being interesting emergent higher level descriptions that give us true causal handles on what’s going on, but they’re not overriding the reductionistic level in any way. So that’s why I think that I’m perfectly happy saying all of this is 100% compatible with traditional Darwinian evolution.

0:25:11 SB: So yeah, that’s a good point, and it is possible that all the causation is bottom up, that the Darwinian Evolution is the driving force, and that these higher level phenomena are just emergent downstream effects of the Darwinian evolution at the bottom. One of the primary counter-examples is the one of the most interesting major transitions, which again goes back to group selection, and that’s the emerging of social behavior in insects. So if we look at the ants, for example, and if we just go and look at a given ant colony, and we pluck out an individual at random, we will find maybe a worker ant for example. Now if we look at that worker ant, we could look at its cells and sequence its genes, and we could ask, “Okay, how fit is this individual?” Well, it’s pretty fit because those ants are extremely sophisticated. They have a complex social structure and they farm other organisms, very, very sophisticated behavior. So that ant is clearly doing well. And if we look at its behavior, we would observe that it’s gonna work its entire life tirelessly for the benefit of the colony. But then if we ask, “Well, it seems like it’s part of a very resilient community and it’s dedicated to its life of labor, but is it actually gonna reproduce?” And the answer is no. The vast majority of the ants in the colony are not gonna pass their genes on to the next generation.

0:26:52 SB: And so the question is, under the Darwinian Paradigm we would say that, “Well, every individual is gonna be filtered out by natural selection, and the most fit ones are gonna pass on their genes to the next generation.” But that individual worker ant, which is apparently doing really well is not gonna pass its genes on to the next generation, and it’s cooperating in this collective, but its genes are not getting anything out of that cooperation, because its genes are not gonna be passed on to the next generation. So that’s kind of one of the conundrums that is put to those who think that Darwinian Evolution alone can explain everything, is how do we explain the emergence of social behavior in insects?

0:27:42 SC: Okay, I think that this could be a whole podcast all by itself. It would be a great one. But I just wanna very quickly put a cap on it by saying what you’re suggesting is that some kind of downward causation that is not part of the traditional Darwinian recipe is helpful to understand the evolution of life here on earth, not that it’s like Creationism or something like that. [chuckle]

0:28:03 SB: Absolutely, absolutely. Yeah.

0:28:05 SC: Just making sure, ’cause there are people who will get worried about that. [chuckle] Okay, good. So, in the definition then, the reason why we got on to this is because NASA puts Darwinian evolution right there in the definition of life, and you don’t think that’s a good idea.

0:28:20 SB: Well, my concern that I share with some of my collaborators, in particular Mike Wong, where we just sort of published these ideas recently, so what we’re concerned about is whether the definition might be too acentric essentially. Because we only have an example of life to study, that makes it very, very difficult to generalize. And the question is, when we look for life elsewhere, what features do we expect extra-terrestrial life to share with life on earth, and what features might we expect to be different? And of course NASA expects that life elsewhere will be chemically based and it will be evolving. And again, me and Mike are great fans of Star Trek and in Star Trek [chuckle] you see all kinds of different forms of life, and some of which are not chemically based. And so when you try to open your mind and take inspiration from things like Star Trek, you wonder whether chemical life and evolution by natural selection is the only way of doing it. And what we fear is that, what if there’s life out there, which does not evolve by Darwinian Evolution, or maybe it’s not chemically based, but it’s there, what if we miss it because our definition is wrong or our definition is off the mark?

0:29:53 SC: And so I think it’s always been useful to me to sort of keep in mind examples of systems that we know of here on earth, that have some features of life, but don’t count as life. And then the idea of the definition of life would be to distinguish something like a hurricane is not alive, but an anteater is. Do you have your favorite examples of sort of warning counter-examples to keep in mind?

0:30:18 SB: Yes, I’m a big fan of these so-called dissipative structures, and have worked with a few of them during my career. So my favorites are convection cells with very… I mean a hurricane is a very, very elaborate complex convection cell.

0:30:35 SC: Yeah.

0:30:36 SB: A convection cell being an ordered structure that arises in a differentially heated fluid. So if you heat a fluid from the bottom and cool it from the top, instead of the heat moving homogeneously through the fluid it forms these beautiful overturning rolls. And you can do the experiment in a sauce pan, maybe with some noodles or something, the noodles will kind of align themselves with the flow, and you see this lovely hexagonal pattern of convection cells. And my other favorite example is pattern forming chemical systems. And arguably the first pattern forming chemical system was the Belousov-Zhabotinsky reaction, and famously they couldn’t get their experiments published for a long time because it formed these interesting patterns, but no one believed them that it was actually real, because it seemed to violate the second law of thermodynamics.

[chuckle]

0:31:32 SC: Wait, can you say a little bit more about what this reaction is? This is something in a petri dish or what?

0:31:36 SB: Yeah, so it’s a… You basically mix, I think it’s malonic acid… It’s not actually that simple, you do have to mix several different things together. And normally when you mix reactants together, you would expect the reaction to proceed in a particular direction, and eventually the system would come to equilibrium and the reaction would be done. In the BZ reaction in fact, you see oscillations and you see waves in the concentrations of different chemical species. So if you put some dyes in the system, then these patterns are very nicely illustrated. And so it involves things like potassium bromate, cerium sulfate and malonic acid, sulphuric acid. So you have to sort of mix these things in the right way, and then if you wait for a little while, you start to see oscillations and spiral waves in the concentrations of different chemical species. And…

0:32:44 SC: And the point is that they look organized, and they’re in some sense, processing energy from their environment, but no one is gonna say that they’re alive.

0:32:53 SB: Exactly. So there are lots of variations you can do of these kind of reactions. So there are model systems such as, for example, the Selkov reaction, which is an artificial chemical system, but it kind of encompasses these effects. In that reaction, you just react one molecule of A species with two molecules of B species, and you get three molecules of the species B. And it turns out that just that reaction, if you tune the diffusion coefficients of these chemical species and you let that system run under different sets of parameters, you can get very elaborate patterns including these spot patterns which will grow and then divide, and it looks exactly like replicating bacteria in a petri dish. So that system, which I studied during my PhD is very interesting, because again, there’s this superficial similarity to simple life, because those chemical structures will grow and divide and replicate. And so, as you said previously, we can ask, “If it looks like life, what has it got that life has, and what does it not have that life has?” And in this case, it definitely has replication, because you can see that replication happening before your own eyes. Of course, it doesn’t evolve.

0:34:18 SB: And if we try and think about heredity, well, each time there’s the replication of those spot patterns, the offspring basically carry the chemical species of the parents, and that’s about it. So the heredity is very limited. However, something that I found during my PhD is that actually, if you run, if you simulate systems with several populations of these things, you can observe cooperative behavior between the different populations. And in particular, something that I experimented with is, if one set of these spot patterns produces heat and the other population absorbs heat, they can actually regulate their temperature. And if you try and sort of kill them by heating up the system or cooling down the system, they will respond by changing their relative populations, and hence they will basically protect themselves from thermal damage. And again, those systems are… It’s just chemical reactions, very simple chemical reaction. And yet they show self-replication, they show homeostasis, they can expand into regions of their space which were previously inhospitable. And a friend of mine, Martin Hanczyc has also done similar experiments with oil droplets. And oil droplets will show chemotaxis. So they will basically follow a gradient of a chemical species which powers them, and you can use that effect to get them to solve mazes, for example.

[chuckle]

0:35:54 SC: Oh, really? Even though these are clearly just little blobs of chemicals, they’re not, no one’s gonna mistake them for being alive, but they can learn enough to solve a maze?

0:36:02 SB: Exactly, exactly. And they can… Yeah, and if you lay out a path of their food in front of them, they will follow it in the same way that bacteria do.

0:36:11 SC: Okay, so this is the challenge that we’re faced with, right? That there are all these things that are kind of… Have some properties in common with life, but are not. And you and Michael Wong have faced this challenge by inventing a whole new word. And why don’t you tell us what the new word is, and what are the ingredients that go into defining what it is?

0:36:30 SB: Yeah, so we wanted to try and clarify this question of what is life by introducing a new terminology to encompass… So we leave life with an I to refer to life that we’re familiar with, I.e. Nucleic acid-based information transfer and protein-based metabolisms, and the lipid membrane compartments that we’re familiar with. So we ascribe that to life with an I.

0:37:07 SC: So in other words, L-I-F-E.

0:37:11 SB: L-I-F-E. And then we introduced a new term, L-Y-F-E. Which at the moment we pronounce loif.

0:37:16 SC: Loif? [chuckle]

0:37:17 SB: Loif.

0:37:18 SC: I didn’t know that, I thought it was pronounced the same, okay.

0:37:22 SB: We introduced lyfe as a larger set of entities which encompass what we have introduced as a more general definition of lyfe. And so that larger set of entities satisfies four criteria which we’ve introduced, as a sort of larger overarching definition of lyfe. And life is a subset of that larger set, which applies to life on Earth after it had evolved from simpler systems. And so the four criteria to be satisfied for a system to be lyfe is dissipation, which just means that the system cannot be in thermodynamic equilibrium, it can’t be an isolated closed system because we observed that life always uses sources of free energy and it always dissipates free energy.

0:38:22 SB: So it takes concentrated energy, uses it for its own functions and then expels less useful energy and ultimately at the scale of the planet, we have concentrated UV light coming in and less concentrated infrared light going out and there’s a loss of free energy involved. The second characteristic is autocatalysis, which is another way of saying exponential growth. So we would always expect lyfe to exhibit exponential growth when it has access to abundant resources. So when you look at a bacteria in a Petri dish when it has abundant resources, it will grow exponentially. The population will increase exponentially as it replicates. And we introduce this characteristic because we don’t… We expect non-linear behavior in living systems because whether it’s replication or the enhanced exploitation of new niches, the activity of life under abundant conditions always leads to an enhancement of that activity and there’s… So we expect there’s always gonna be the potential for that positive feedback. And then third characteristic is homeostasis, which is quite common to definitions of life, because we observe that life is very good at regulating certain important variables…

0:39:58 SC: Yeah, I mean, maybe give us a definition of what homeostasis means. We’ve actually talked about it a lot. We had a podcast with Antonio Damasio, who’s a big fan of the concept, but in this context what do we mean?

0:40:09 SB: Yeah, so in this context we basically mean negative feedback mechanisms that are able to rein in key physical or chemical variables, such that the entity is stable. I mean for us humans, the characteristic example is often blood glucose level. If blood glucose level is too high or too low then we suffer fairly catastrophic effects. Temperature is another one, you know we can tolerate certain changes of temperature but we have to do a certain amount of self-regulation, otherwise we suffer a lot of negative metabolic impacts and you can also look at it at collective levels. I mean again, the Gaia hypothesis would be another, would be a whole another podcast and I’m not saying… I think the Strong Gaia hypothesis is true, but the…

0:41:02 SC: Sorry, tell us what the Strong Gaia hypothesis is. [chuckle]

0:41:05 SB: The Strong Gaia hypothesis is that the whole of planet Earth can be considered an organism in itself, which by nature of being an organism, regulates itself such that it is more resilient and stays within environ… Within a good range, a supportive range of environmental parameters. So for example, it doesn’t… The Earth doesn’t allow itself to lose its atmosphere in the way that Mars did and so it doesn’t allow itself to experience like a runaway greenhouse on the other hand and boil off the oceans. It doesn’t allow itself to plunge into a permanent snowball Earth kind of deep freeze scenario. But instead, it activates mechanisms to keep itself in a comfortable state.

0:41:56 SC: Right, okay. That’s homeostasis, we regulate ourselves to be comfy.

0:42:01 SB: Exactly, and of course you need homeostasis to keep the autocatalytic property in check. Again, going back to the corporators versus cheaters, we are multi-cellular organisms and so we have to regulate the replication rate of our own cells. And our cells are basically signing a contract with one another in that sense. And if we have a stray cell which breaks that contract and replicates faster than all the others, then you end up with cancer and so it’s a key property.

0:42:39 SC: So the… I mean maybe I didn’t quite understand the autocatalysis therefore well enough. I mean the auto… I should associate the autocatalytic property with the idea of growth, exponential growth, but should I… Is that in the number of organisms or is it in the number of cells per organism, individual organisms or is even thinking of organisms too benighted at this level of generality?

0:43:04 SB: Yeah, it’s a good question. So in general for life on Earth, we tend to think of it as the cell count, but in principle for exotic life elsewhere, maybe that life doesn’t… It’s not comprised of cells in the same way and so perhaps it would apply to a different variable, for example, free energy consumption. And we can, if we were to… If we look at graphs of the human species, changes in the human species in the last several hundred years, the population of humans has increased exponentially and also the consumption of mineral resources, the consumption of free energy, these things have also followed exponential-like curves. And so it may not necessarily apply to individual organism count but it should be possible to find some measure of that living system, which when under ideal conditions grows exponentially.

0:44:10 SC: I guess I mean maybe this… I know you’re only done with three of your four pillars here, but maybe to help us focus what these are supposed to apply to, I mean maybe this raises an interesting question of do we think of life on an organism by organism basis, or do we think of it as the entire biosphere all at once. I had David Baltimore on very recently talking about viruses, and common question about viruses is, are they alive? And Baltimore’s answer was, “Well, of course, they’re alive because they’re part of life. They can’t survive on their own, they don’t fulfill all these criteria by themselves, but they’re part of this larger process,” and is that a paradigm shift we should try to make from thinking about the process of life as a whole rather than whether this or not individual organism is alive?

0:45:00 SB: Yes, exactly, and this is a great point. And I do agree with this quote from Baltimore in that I do think it’s wrong to just say that viruses are not alive, because as time goes by, we see more and more evidence that they are, even though they have these terrible destructive effects, they are still an integral part of the biosphere. They massively out number other organisms, and so they are contributing somehow. And we probably don’t understand the extent of their contribution yet. And so we definitely apply our definition to the system level, maybe even the planetary level or the ecosystem level.

0:45:43 SC: Right.

0:45:44 SB: Yeah, if you just pluck out individual entities from the system, it’s quite likely that those individual objects will not satisfy the criteria, but the system as a whole will. And along with several other authors in Astrobiology, I would definitely, I would lean towards life being a planetary phenomenon rather than something that happens at the individual level.

0:46:14 SC: Yeah, okay, good. And so, let’s contextualize your pillars in terms of that. You mentioned dissipation. So life as a whole, the system level, clearly we’re using up the free energy we get from the sun and dissipating it. Autocatalysis, at least in some sense… This is where I’m struggling, so numbers two and number three, where autocatalysis, which is growth and homeostasis, which is stasis. [chuckle] So if you… I would guess that autocatalysis applies to the system and homeostasis applies to the individuals, or at least little sub-environments thereof.

0:46:52 SB: Yes, again, another good question, because there are different time scales at work here and different… And fluctuations, as well as equilibrium. So as you point out, if something is perfectly homeostatic, then it’s not gonna expand. And how does it innovate? Right now, we’re just suggesting that those mechanisms be represented in the system and be identifiable in the system.

0:47:40 SC: Right, okay.

0:47:41 SB: And also the interaction between the fourth pillar, which is probably the trickiest one, of learning is another interesting area, because, yeah, autocatalysis implies significant change, whereas homeostasis implies a return to equilibrium. But of course, learning implies change as well, and so the feedbacks between these pillars is something that we will elaborate on in future work. So under what circumstances can autocatalysis break homeostasis in such a way that the system, it maybe expands and innovates? And there might be other circumstances where the system is trying to be autocatalytic and expand, but its homeostatic mechanisms are bringing it back. So, yeah, this is a big question that we’re gonna explore further in future, I would say.

0:48:44 SC: But maybe… Okay, so your fourth pillar is learning, and I think maybe it would be helpful to be more specific about how that arises. Do fungi learn? Do trees learn? Do bacteria learn?

0:48:56 SB: Oh, definitely. All of the above, yeah.

[chuckle]

0:48:58 SC: Good, so explain what that means, ’cause they don’t go to school, they don’t get grades.

0:49:03 SB: Yeah, so this is… Again, quite an open aspect, and probably the most interesting and also probably the most important, because we talked about non-living systems which exhibit some of these characteristics. And crucially, those non-living systems did not exhibit any learning capability. There are non-living systems capable of learning. Neural networks is the classic example, of course. And there are even just simple physical systems like shared fluids which can show hysteresis in memory. In fact, right now, I’m working on convection systems that are capable of computation, so that…

0:49:48 SC: Well, maybe I don’t wanna give you too much of a hard time here, but I want a definition of what learning means. Like if I walk down the beach, I leave footprints on it, has the beach learned about the shape of my foot?

0:50:01 SB: Yeah, good question. So learning, it is somewhat of an umbrella term that encompasses several different aspects, and memory is one of those aspects. So, yeah, when you’ve walked down the beach and you leave a footprint in it, has the beach learned about you? Learned about your presence? So yeah, the beach has exhibited a memory and it’s interacted with you somehow. The question is, will that memory, the information in that imprint that you left, will it be processed? Will it be used? Will it feedback on anything else? So if the beach, for example, let’s say your footprint was left in the beach and then the sand sort of rearranged itself into a different structure. Maybe it did like, made a copy of the inverse of your foot and then copied that 100 times into a tall pillar and that had an effect on the way that the waves washed the sand along the sea shore and maybe made the beach more resilient to the erosive effects of the waves. Well then we would say, “Crikey, that beach interacted with Sean’s foot, and it learnt something.

0:51:10 SB: It somehow processed that information and used the information to protect itself from destruction.” If it did that, then we would say, “Well that seems like a learning behavior.” It encoded information, processed the information, and it used the information. And likewise there’s often analogies made between life and fire, because fire kinda has a metabolic process because it’s converting free energy and the fuel to lower free energy forms and it replicates itself. It grows. It shows autocatalysis. And does it exhibit homeostasis? Well, in certain things, it doesn’t explode and completely encompass everything in a huge area surrounding it instantly. So it’s autocatalytic to a certain level. But it’s not homeostatic in the general case, because once it burns itself out, it dies and it also doesn’t learn. We can imagine a different form of fire, which somehow interacts with the fuel environment around it and adjusts its burning rate adaptively such that it never burns out.

0:52:27 SB: But of course, in fact, fire doesn’t do that. So fire is not capable of learning, and hence, we don’t consider it to be alive. And so in general, we define learning as three abilities. The ability to interact with the environment and record information. The ability to process that information, so change its form somehow. And then after processing, the use of that information such that the organism, it has somehow increased its utility, its benefit as a result of interacting with that information. Because of course, there are lots of ways that you could record and process information but not actually make good use of it. And so natural selection is kind of like a feedback effect of that nature in the sense that it chooses organisms which are making functional use of information that they receive.

0:53:31 SC: Okay, so the learning is not just taking information, recording it, but using it, processing it somehow, and so just to be super-duper clear here, in what sense does a tree do that?

0:53:44 SB: Right, so trees are a great example, and I wish I knew more about trees because I know that they communicate with one another. They are very sensitive to the environment. They process information, and they make use of that information. So, there have been some experiments, I think they were done in British Columbia, where they actually discovered that the trees were communicating with each other via chemical signals and sharing resources based on which resources were abundant for one individual, and which were scarce for another individual. But one of the great examples of long-term learning by trees is right here in California, and it’s the Bristlecone pine, which is, I believe the oldest… There’s a Bristlecone pine in the White Mountains of California, which is the oldest known organism in the world. I believe it’s something of the order of 4 to 5000 years old… I think 5000 years old.

0:54:49 SB: And one of the interesting adaptations of the Bristlecone pine is that it’s cones, containing the seeds, are activated by fire. So we tend to think of forest fires as being destructive. I mean they cause huge amount of damage to our human infrastructure and cause loss of life, so they can be terrible. And so we associate fire with destruction and damage. And what’s interesting with the Bristlecone pines, they are fire-resistant to a certain extent. That’s not unique to them, there are other trees which are fire-resistant, but the fact that their cones activate and release the seeds because of fire, so the cones are basically activated by the fire. The selective advantage of that is that a lot of their competitors are not fire-resistant, so once the fire has blown over, it’s quite likely that there will be an abundance of certain resources because other plants have been destroyed.

0:55:57 SB: And so this is a great time for the Bristlecone pine to lay down it’s seeds and try and grow. And so exactly how the Bristlecone pine learnt that characteristic is a tricky question. Of course, the first order answer is, well it was again evolution by natural selection. There were some mutations which produced the cones that were activated by the fire, and over time that was selected for. So there’s the Darwinian answer, which is that it is favorable. Of course it would take a very long time for that selection process to act because you would have to have a large number of fires for the Bristlecone to out-compete it’s competitors. So whether it’s by evolution through natural selection or some other more sophisticated learning system like an associative learning mechanism, like a Lamarckian’s within lifetime, and these trees live a long time. Somehow or other, this tree has learnt that fire can be used to its advantage, even though for the majority of organisms, fire is destructive.

0:57:12 SC: So I think, yeah, I think that does clarify in my mind what’s going on, so… In perfect accord with what we said before about the systems level view rather than the organism level view, you’re counting learning as a species figuring out some kind of strategy for surviving and flourishing even… Not necessarily just an individual organism.

0:57:32 SB: Yes, yes. And whole eco systems, ’cause again, in the competition, for example, for the oldest organism, I believe there’s another collection of trees perhaps also in California or Nevada, which is older than 5000, but it’s not any one particular tree which has that record age. It’s the germ line and the sort of community of trees which have that…

0:57:57 SC: Right.

0:57:58 SB: Have that age. And of course, there are also immortal organisms like the Hydra, which don’t have any programmed cell death, so they could be… You know, in the deep ocean, there might be organisms that have been around for tens of thousands of years, maybe.

0:58:12 SC: I thought that the Hydra was a mythical beast with multiple heads.

[chuckle]

0:58:17 SB: That also. Yeah, so it’s the character in Greek mythology, but… Well, it’s also a genus. They’re small fresh water organisms.

0:58:28 SC: Okay.

0:58:29 SB: Yeah.

0:58:30 SC: I don’t know about Hydras in the real world, so that’s a new one to me.

0:58:33 SB: Yeah, there’s…

0:58:35 SC: And they’re immortal?

0:58:36 SB: There’s been some studies… Basically, there are Hydra which, the technical term is they don’t senesce. So most organisms undergo programmed senescence, which means programmed cell death. So at a certain point in an organism’s lifespan, the cells start to degrade. You know, we just call it aging. And gradually… Gradually this degradation increases and eventually we die. And basically… So I think the first article was in 1998 by Daniel Martinez, and it was… Where it was claimed that Hydra are immortal. But I mean it’s a controversial piece of work, so…

0:59:22 SC: Okay.

0:59:23 SB: I don’t know whether it’s been proven outright, but…

0:59:26 SC: Okay, but to… So just to… I think that we can get this in everyone’s brains here, so you’re saying that there are these four aspects that together would qualify you as being lyfe and maybe even life. But presumably life is a subset of lyfe, right?

0:59:44 SB: Yeah, so life is a subset of lyfe, and specifically it’s that subset which uses… The way that we focus… There are different ways that you could identify the characteristics. The ones which we focused on are the sort of core functions, which are the way that the organism replicates and carries information, which for life is using nucleic acids. And it’s a nucleic acid sequence, a polymer which has like a sugar backbone and phosphate groups and a base sequence. But of course, we could imagine another form of lyfe which doesn’t use nucleic acids, but maybe it has a sequence polymer where information is encoded in a particular… In a sequence of one of the side molecules in this chain, and it has a backbone in the same way that nucleic acid polymers have a backbone. But the backbone might be different and the bases might be different. So that would not be life. That would be lyfe.

1:00:58 SB: And the mode of metabolism, we also make a distinction there. So for it to be life we expect it to be a metabolism based on amino acid enzymes and Chemiosmosis. So Nick Lane has written a lot about the Chemiosmotic property of life. And Chemiosmosis refers to the pumping of ions across a membrane and the reverse process of allowing those ions back through a membrane and converting that energy into a different form. So for example, a lot of organisms will use a source of energy, maybe light energy or chemical energy to pump protons, for example, across the membrane. And that is a costly exercise. And then later on, those protons will be allowed to come back through the membrane and they will go through a molecular machine, which sort of turns as they do it. And as it turns, it will take ADP and phosphate molecules and form ATP which is the energy currency of life. So life on Earth, we identify it as using a metabolism based on Chemiosmosis. And the most important metabolic cycle is the rTCA cycle. So…

1:02:28 SC: Okay.

1:02:29 SB: That’s a key aspect of life’s metabolism. And then if we look at the compartments of life on Earth, we see that it’s a complex of lipid molecules and proteins. And so again, if we were to find… Let’s say we found extraterrestrial life, which was using nucleic acids for information storage and replication, and it was using the rTCA cycle as its metabolism and Chemiosmosis, but it was not using lipids or proteins for it’s membrane, it was using something completely different, that would not be life, that would be lyfe. That would be a form of lyfe.

1:03:07 SC: But you would still… You’d be pretty excited.

[chuckle]

1:03:09 SB: You’d be excited, yeah, yeah. You would, you would. You’d jump out of your chair when you sequenced it, yeah.

1:03:15 SC: I mean, are there… The other thing you talk about in your paper is there are ways that you can have subsets of all the four pillars, you have the pillars of dissipation, autocatalysis, homeostasis and learning. And you mentioned that you could have two or three at a time, and it wouldn’t count as life or lyfe, but it would still be maybe a step along the way.

1:03:37 SB: Yeah, exactly. So this goes back to the dissipative structures that we were talking about before. And one of our inspirations for this paper was to try and distinguish these different things, what do they have, which we ascribe to life and what are they missing? And so we can take all of our four pillars and look at the different subsets and of course there are eight different subsets which have less than all four of the pillars. So we talked a little bit about convection cells and hurricanes, so of course, those are systems being driven by large energy gradients, so they’re definitely dissipating energy, but we would say they’re not capable of learning.

1:04:26 SC: Yeah.

1:04:26 SB: We could ask, are they autocatalytic? Well, there are some situations where structures in fluid flows will sort of replicate. So, you know, vortices in turbulent flows, sometimes it looks like they’re replicating like a Kármán vortex street. So yeah, potentially under some circumstances, fluid flows, maybe they can be autocatalytic. Homeostasis, well, Jupiter’s red spot has been around for a long time.

1:04:54 SC: Yeah.

1:04:55 SB: So, but of course, these systems are not capable of learning because there’s no identifiable structure which is encoding information. I mean, they are… Every dynamical system interacts with its environment and processes information in some way, but if that information is passing in, being processed, and just washing out, and having no other effect, then that learning… Then there’s no real learning.

1:05:22 SC: Right.

1:05:24 SB: It’s not feeding back on the existence of the…

1:05:26 SC: So there’s always this tight coupling in these discussions between looking for life elsewhere in the universe, and I think that this is a very nice framework for thinking about what you might find along the way and how it could be different, but… And the other question of how life actually arose here on Earth. Do you think that this framework of understanding the four pillars of life in their separate ways helps you with either one of those, either it helps us understand the origin of life here on Earth, or strategies for looking for it elsewhere or what we might expect to find elsewhere?

1:06:00 SB: Yeah, so we hope that it can maybe serve as… Either as a guide or perhaps as inspiration for ways that life could arise. Because again, when we conduct experiments where we’re trying to understand the origins of life, as I mentioned before, we don’t have any agreed upon standard for what would count as a successful demonstration, what would count as a successful experiment. And so potentially, these four pillars, if they were… When we refine them over time, formalize them a little bit, maybe they could serve as a criteria for what is successful and what is a failure. And basically, in the past, a lot of experiments focused again on trying to demonstrate the emergence of a particular function which was seen to be essential for life, and then assuming that the other functions of life would then follow.

1:07:10 SB: So an example would be genetic… The replication of genetic information in the RNA world theory. So for the people working in the RNA world, for them, the replication of information in a genetic polymer is so essential to life that it has to be functioning before any other aspects of life can come along. And in general, it comes back to the assumed importance of evolution by natural selection. The opponents of the RNA world would say, “Well, there’s no physical reason for strings of RNA to just appear in a system. It’s a very difficult molecule to synthesize, and why would a physical system want to produce strings of RNA?” So the opposite… So a counterpoint to that is that, there’s no way that a system would ever want to produce genetic molecules. So instead, why don’t we focus on a process which is physically favored and see whether the functions of life can emerge out of that.

1:08:19 SB: So others focus on the metabolic aspect. Where you have a gradient of free energy, so the dissipation pillar is satisfied. So the system wants to dissipate free energy, it wants to try and equilibrate and in the process of equilibrating, perhaps it can carry out some lifelike functions, maybe synthesize amino acids, for example. And then, so for those authors, once that process is running, maybe the genetic molecules might be produced as a side effect of the metabolism, and then the use of those genetic molecules to encode sequence information would be accidentally discovered later on. And so, because we’re… As humans, when we engineer things, we tend to first manufacture the components and then put the components together to make the whole, if you think about like a production line in a manufacturing plant.

1:09:22 SC: Mm-hmm.

1:09:23 SB: So some of the approaches in the origins of life have pursued ways to produce the components of life. For some people, the most important component is amino acids leading to proteins, and for others, it’s nucleotides, which, as I mentioned they are notoriously difficult to synthesize, but there has been progress made. And so I often wonder, if those experiments succeed, and we have a chemical scheme which synthesizes those components, we will then stumble across the problem of, what is the physical driving force for those components to come together and cooperate in a way that’s recognizably alive? And so hence, that was part of our motivation for the four pillars because the four pillars are processes, they’re not objects or components.

1:10:14 SC: Right, they could be manifested in different chemical ways.

1:10:18 SB: Exactly, and so… And… I mean… I’m actually… I go back and forth… In my own mind, I go back and forth about whether it’s more likely that the four pillars emerged sequentially, or whether they all were there from the very beginning. Because, to be honest, I find it unlikely that the various functions of life emerged sequentially that… Nature always surprises us, especially the living world. And given the dramatic extent to which the living world surprises us, I find it almost impossible to imagine that the most mysterious aspect of life would be that predictable. That first came this, and then came this, and then somehow it magically came together and then we got life. So, instead, I think it’s more likely that the origins of life was this very messy, chaotic process. And maybe life began several times, and it went extinct again, almost as much as it originated. And then, maybe there was a handful of survivors. And probably, those survivors would’ve interacted and perhaps shared material, shared genetic information, undergone all kinds of exchanges, and then what emerged was a kind of amalgam of different things that had been discovered at the beginning.

1:11:51 SB: So, even the idea of life coming from a single universal ancestor, I’m not sure that we can definitively say that that was the case. There are very strong limits on how far we can reach back in the history of life based purely on paleogenetics. So I think it’s very much an open question. I could come up with scenarios where you first have dissipation, so you have a system that was driven out of equilibrium. That’s easy for the early Earth, because it was a very turbulent [1:12:27] ____ And then, I can imagine, maybe a chemical network that was autocatalytic and exponentially grew in the rate at which it consumed external resources, and then, some negative feedbacks could’ve set in to regulate itself against external perturbations or rather, if there were several versions of this metabolic system, those which stumbled upon homeostatic mechanisms would’ve out-competed the non-homeostatic ones. And then, perhaps, at some stage, either by means of genetic molecules or even non-genetic molecules, there was perhaps some learning process. Alternatively, you could come up with scenarios where all four pillars were there from the very beginning. So it’s still an open question, I think. And I wouldn’t rule out the possibility that all four characteristics were there from the beginning.

1:13:29 SC: You’ve mentioned… So there’s the four pillars which are sort of, I don’t know what to call them, “Aspects of life” the dissipation, autocatalysis, homeostasis and learning. But then, the more traditional discussion of these issues that I’ve heard, and then you’ve also alluded to, is the functions of different things that go into making life: The metabolism, the reproduction, the compartmentalization, the little membranes that surround us. And there’s a debate going on between which of these comes first. Is it too much to say, or is it what you’re aiming at to say that, shifting focus from the functions to the pillars helps us be a little more free in our imagination about how the different functions might have arisen, and the role… And the order in which they’ve arisen. Maybe it’s not, the question we should be asking is not which came first, but how the different pillars sort of came to be and hooked up to get together.

1:14:31 SB: Exactly. I think that would be a great step forward is if our thinking… Because those discussions consume a huge amount of time and energy: “Which came first?” “Well I think this thing came first,” “No, I think this thing came first.” And of course it’s possible that mother nature is just laughing at us, saying “Well, you’re all wrong.”

[laughter]

1:14:52 SC: Right.

1:14:53 SB: Because it was neither. Neither came first. There was a bit of this, and a bit of that, and interactions among everything. And so, yeah, I do think that the “X first” discussions, I fear they may be somewhat artificial and they might be consuming our resources without actually teaching us anything. And, in general, to me, the biggest enigma in the origin of life is the emergence of the information processing mechanisms. And some of the most exciting research going on in origins of life right now is reconstructing the history of the ribosomes. So the ribosome is an organelle common to all life, except viruses of course. So the ribosome is a RNA Protein Complex, which translates genetic information into protein sequences. So the information in DNA is translated to Messenger RNA, which is then translated to Transfer RNA, and the Transfer RNAs are charged with amino acids and basically the RNA sequence is passed through the ribosome and a sub-unit of the ribosome will take the tRNAs and combine the amino acids into a growing peptide chain which has the sequence that was encoded in the RNA encoded in the DNA. So the ribosome is kind of the nexus. It’s the place where this chemical information is transformed into peptide sequence.

1:16:37 SC: Right.

1:16:37 SB: And the origin of the genetic code is a deep enigma in itself. And the question is, why would a physical system invent a coding mechanism? We can always look at it after the fact and say, “Well, you have to have the coding system to translate genetic information to protein sequence information. And a protein that’s assembled without the right sequence, for the most part, is not functional. It doesn’t do anything useful. That sequence is essential.” And so…

1:17:14 SB: It’s not well-understood, the connection between well-known physical forces, like for example, the second law of thermodynamics, or the principle of least action. How does that link to the emergence of this highly sophisticated coding system? And so… So, some researches in the field have been looking at the structure of the ribosome and trying… And looking at where different sub-parts of it were added on in the past and basically retrodicting what it looked like at the very beginning. And they have been able to sort of reconstruct this very, very primitive, early, simple ribosome. And…

1:17:58 SC: So… Sorry, in other words, given the current ribosome, what’s the minimal sub-part of it that would be functional?

1:18:05 SB: Exactly. So it turns out that as the ribosome evolved, it kind of evolved by accretion, by basically adding extra sections of RNA, like little loops, for example. And it turns out that it is possible to identify these characteristic points where one of these loops or one of these RNA sections was basically tacked on to an existing version of the ribosome. And so what these researchers were able to do is identify those points where a new part was added on, and say, “Right, well, let’s take that part off. And when we take that part of, we revert the ribosome to a slightly earlier point in its history.” And it turns out that you can keep doing that until you’ve actually stripped away almost all of the additions, and at that point, you get back to a very, very minimal primitive ribosome, which still has the ability to translate genetic information to protein sequence information, but it doesn’t have all these more recent additions that have come along in the intervening years. And so it’s possible that in the near future, we will have good information about the first ribosome, and maybe we can… We can sort of reconstruct the most primitive possible cell. And then we can ask… And then… We still have a gap that we need to understand regarding the origins of life.

1:19:44 SB: But the gap would have been made a little bit smaller at that point. And so we come back to this fundamental question, which to me is a question of thermodynamics and physics and information theory, which is given the tendency of the second law of thermodynamics, which is, for systems to try and come to equilibrium, for gradients to dissipate, for things to spread out. Given that tendency, and given, let’s say, the presence of certain chemical or physical structures, how would that tendency be linked to the emergence of an information processing system? And in fact, it doesn’t need to be as complicated as a genetic code or a ribosome, because there are also individual molecules which are capable of processing information. It’s been well-known for a long time that protein molecules can process information and switch their behavior depending on their surroundings.

1:20:51 SC: Yeah.

1:20:53 SB: And there’s another class of molecules called information engines, which are molecules that basically… They can make a measurement of their immediate vicinity and then change their behavior based on the measurement that they’ve made, and then… And that behavior can actually be used to make an energy conversion. And so it’s quite similar to Maxwell’s demon, which makes a measurement and is able to create free energy where there was none before. It can’t do it for free. There’s a free energy cost involved in erasing its memory. And likewise, information engines in general will consume a source of energy, make a measurement, and create a different source of energy. And so one example is molecular walker molecules in modern life. So those walking molecules will actually… Will actually wait for a kick from their external environment, and when a useful kick, when they get pushed forward in the direction that they want to walk, they will actually consume another source of energy, ATP, and lock themselves in the forward position. And so that is how they actually walk forward. They’re not… Their motion is not powered by the ATP.

1:22:25 SB: Their motion is powered by random fluctuations in the background, but the sequence of steps that they go through is selective in the sense that if the kick is good, they will move forward and lock that position in by burning the ATP. And if the kick is bad, they will do nothing. So this is another… Is a fundamental aspect of the origins problem, which is why would a physical system, given a gradient plus perhaps some molecules or physical systems capable of processing information… Why would it become favorable for free energy dissipation to be linked to the creation of another gradient or the processing of information?

1:23:15 SC: Well, I think that’s a… It’s a good place to end. It’s a little bit of food for thought for people because I completely agree. I’ve thought about this myself, how we all know that information is processed in our own brains, and there’s a sense in which all of life processes information in some sense, but getting that sense exactly right is gonna be important to figuring out how it started, how the origin… There’s probably more than one way in which life processes information and makes use of it, and so the project of specifying what that means and how it came to be is gonna, I think, lead to a lot of breakthroughs in the future.

1:23:53 SB: I agree, yeah, definitely.

1:23:54 SC: Alright, well, Stuart Bartlett, thanks so much for being on the podcast, you’ve given us a lot to think about.

1:23:57 SB: Well, thank you. It was a great privilege.

[music][/accordion-item][/accordion]