A few years after Charles Darwin published On the Origin of Species, upsetting centuries of certainty about the history of life, he wrote a now-famous letter to Joseph Dalton Hooker, British botanist and advocate of evolutionary theory. "But if (and oh what a big if),” Darwin’s letter reads, “we could conceive in some warm little pond with all sorts of ammonia and phosphoric salts, light, heat, electricity etcetera present, that a protein compound was chemically formed, ready to undergo still more complex changes.”
That was 1871. Nearly 150 years hence, humankind has worked out the details of the evolutionary process to exquisite depth and resolution, but abiogenesis - the origins of life - remains one of the greatest mysteries of our world. Fierce theoretical debates rage on between those who think life got its start in deep sea hydrothermal vents and those who think it started in “some warm little pond” – not to mention more heterodox hypotheses. The consequences are enormous – shaping plans for interplanetary exploration, changing our approach to medicine, and maybe foremost, settling the existential question of what life is in the first place.
This week’s episode was recorded live at the Santa Fe Institute’s InterPlanetary Festival in June 2019. The panel features evolutionary theorist David Krakauer, President of SFI; biochemist Sarah Maurer, Assistant Professor at Central Connecticut State University; and SFI Professor Chris Kempes, who works on biological scaling laws. In this discussion, we present a spectrum of perspectives on the origins of life debate, and speak to the importance of presenting this unsettled science as itself an evolutionary object...
Visit our website for more information or to support our science and communication efforts.
Join our Facebook discussion group to meet like minds and talk about each episode.
David Krakauer's Webpage & Google Scholar Citations.
Sarah Maurer's Website.
Chris Kempes's Website.
InterPlanetary Festival's Website.
Complexity Explorer's Origins of Life Online Course.
Follow us on social media:
Twitter • YouTube • Facebook • Instagram • LinkedIn
Michael: Welcome everybody. This is exciting. I know that Jenna just named everyone, but we'll go down and everybody can have an opportunity to introduce themselves real quickly, and then we will get into the meat of this especially tricky issue. So I'm Michael Garfield and I run Santa Fe Institute’s Social Media.
David: Yeah. I am David Krakauer. I'm the president of the Institute. I work on the evolution of intelligence and stupidity on Earth.
Chris: I'm Chris Kempes. I'm a Professor at the Santa Fe Institute. I think a lot about how the physical laws interact with biology, how we find systematic patterns in laws across all of life.
Sarah: I'm Sarah Maurer. I'm an Associate Professor of Chemistry and Biochemistry. I'm from Central Connecticut State University, and I do a lot of wet lab chemistry, so I build simple models of cells, and look at what kind of lifelike properties they have.
Michael: So what we have here on the panel is a rather diverse group of perspectives on all of the different aspects of this problem of biology, which is: Where does it even come from? What are we even describing when we describe this question? As Schrödinger puts it, "What is life?" Right? This is still a question that's contested. So I think it would be interesting to hear from each of you. Why do you think that this particular question of the origins of life has been so resistant to a clean answer? Why is it that we're still working on this after 60 years of really pushing on this question?
Chris: Yeah, I mean, I think... Well, so there's many reasons it's been a hard question. I think in general, life has been a hard question. So even until very recently, we often thought about life as very different forms, lots of individual species. Before we had an understanding of DNA and genetics, and the central dogma of molecular biology, thinking about life was much more confused. Before Darwin even thinking about life was much more confused. So I'd say in general we've struggled with trying to define or think about what life actually is and what are the right lenses to use for it, and then thinking about an origin. I think what becomes very challenging is how do you deal with the huge expanses of time?
If we want to think about the specific origin of life to our current form, and what that trajectory has looked like, and then there are very hard problems about how do you go from a completely abiotic world to something simple and evolving, to all the way up to the very complicated molecular machines and machines that we have as single cells. So I think it's huge expanses of time, and then certain types of transitions that we don't have good ways of talking about, thinking about defining.
Sarah: I also think that it's really challenging because we only have this one sample of life, and so it's really hard to understand how you could do anything from one example. So we're really struggling with even knowing what parts of life are a requirement for early life or for like the first living things, and what parts of life that we see today are from a long evolutionary trajectory that we develop over time, and really aren't necessary for something to actually be alive, per se. So I think finding a second instance of life will be the best way that we can actually develop the theory of life.
David: Yeah. I want to come back a little. So the first point to make is that some of the terms that we use most frequently that we deploy in everyday life are actually the hardest to understand. So love, for instance or hate, or consciousness, or life, and so it turns out, somewhat paradoxically, that some of the more technical terms that we use are easy in the terms that we use that represent averages of opinions like life, are difficult. So that's one point to make. So it's a little bit counterintuitive, but broadly speaking, there are two schools here. One you could call the biological naturalists. The biological naturalists, and Sarah might be one, tend to think that life evolved only once, and that fundamentally understanding life means understanding the particular contingent chemistry of life that's metabolic replicating, open-ended evolving.
The other school of which I'm a member, and Chris might be somewhere in the middle, are the functionalists, and we believe that life is a self-propagating computational element that has much more to do with information than it has to do with chemistry, and that we've created life countless numbers of times. Every time you open your computer, you generate life de novo when you run code, and the fact that it depends on the energy we supply is no different from the fact that life on Earth depends on the energy the sun supplies. So it's worth bearing that in mind, if you like computational theories of life and chemical theories of life.
I think one of the problems of the field is it's been dominated by the chemical perspective that's trying to look for the unique conditions of the early Earth, rather than be a little bit more expansive, and ask what the general principles are that you can observe anywhere, and the conditions that even prevail today in modern Earth.
Sarah: I guess the only thing that I would append is that there has never been a computer that was not made by a living thing, and so we might consider that computers are just a next step of evolution of the only life that we know, and maybe it has a different type of information, but it is really just an extension of Earthlings, and so is it really making life every time you turn on a computer or is it just like giving birth?
Michael: That's an interesting point that leads into this question that I had in high school biology when my teacher said, "Okay, viruses are not alive. This doesn't count because they require a host in order to reproduce," right? "They require a context," and I was like, "Wait." At 14 I was like, "Wait a minute. I require a space suit. If I leave the atmosphere, I require this complex supply chain that produces all of this food, and even if I'm growing all my own food, I still require a microbiome to digest it."
So this question of context I think is an interesting one, and given that all your work is being done in, like you said, a wet lab, I'm curious how each of you think about the role of context, closure, constraint, and a distributed, I don't know, support or intelligence of the environment in this question. Especially given that so much of the investigation in this particular topic is treating this issue as though it's like a carry over of the spontaneous generation of the ancient world, this notion you can just leave meat in a jar and it'll grow flies. So like where does the context fit into this investigation?
Chris: Great. So I'll start. I think for me there's no problem with viruses, and I think David would agree with this, but for me a virus is a living thing that has a very complicated ecology, has a very complicated environment that it needs for its own replication and so forth. That actually turns out to be true for most organisms that we know. So one of the great surprises of microbial ecology has been if you take almost any bacterium, say from an environment, the likelihood that you can figure out how to culture that in the lab is very low, and so the types of things that we regularly grow in laboratory conditions are things that have very simple environments and we understand what to do, but that's not true for most organisms.
We can't figure out the set of contingencies they need from their environment to exist, and so I think that's one of the fundamental pieces that we need to think about for life at any scale is: What is the connection between the living thing and it's contingencies with some environment? That really doesn't matter if you're talking about a molecular virus or some sort of other living process like cultural evolution or something like that. It really is about the thing that is replicating, and existing, and evolving, and the types of environmental conditions it needs for that to occur.
David: Yeah, I mean, so something Chris and I worked on several years ago, we developed a mathematical theory for this, and one of the ways you make progress in science, or I guess in thinking, is you take a concept like life, and you get rid of it, and you replace it with its constituent parts, and so we choose not to talk about life, and you can talk about things like autonomy or minimality, and that's the spectrum that Chris is describing. I think this is extremely dangerous to think of life as autonomous and something you can build because it eradicates the need to understand the ecological dependencies exactly because it's talking about...
One of the ethical implications of which is to imagine that you could live on a planet with just one species, namely us. So life is an ecological phenomenon, and if there's any ever going to be a biological definition of life, it's going to have to be whole planet, and it's quite interesting if you look at a textbook on life, it never considers the ecological network. It will always say replication, metabolism, and so forth, but it's to miss the point. So the virus makes very clear that the completion of the life cycle requires a huge number of host factors, and that's no different from the huge number of ecological services and factors that we require to compete our life cycle. So actually to think of ourselves as a virus would be extremely useful. I don't mean that in a parasitical sense, but just in terms of the context that Michael's alluding to.
Sarah: I think that's true that when we talk about life, we have to be talking about life from the context of the entire planet, and you can't. I studied honey bees for a short period of time, and a single worker honeybee will never be able to reproduce. Right? It lives for a very short time, and it gathers resources for the hive, and I had a lot of struggles when I was 21 and murdering honeybees every day about whether or not I was killing a living thing because it didn't really feel that way in a lot of the really definitive biological terms, but I do think that when we look at something and say, "Is this living or not?" Like an individual single instance of whether something is alive or not, that we as humans have like this gut reaction to whether or not that's true.
That shapes a lot of our interaction culturally with each other and, also, the dialogues that are occurring between biologists in between different fields, right? So physicists talk about life in a very different way just because of the way that they look at something and have that gut reaction, "Is this alive or not?" So I think that it is challenging in a lot of ways to do interdisciplinary work and to really contextualize what your experiment will be addressing or what your work is really trying to achieve just because there is such a different feeling about it or a different sociology behind it.
David: Can I just add?
Michael: Yeah.
David: So another way to say this is, and this may be a little bit gobbledygooky here, but transient forms of matter that exist in this ecological network that manifest agency, that is have purpose and function, we tend to confuse with life, whereas we should be conferring the life property on the entire graph, but what we're really doing in all of these experiments is doing experiments on the nature of agency, and it's an important element of nodes in that living draft, but it's not life itself.
Michael: Sarah, do you understand the work that you're doing in terms of understanding the chemical basis for agency or, I mean, what is your... As you said, you're kind of on the side that life is append to a particular chemical instantiation, right?
Sarah: I think that we're all saying that. Right?
Michael: Yeah.
Sarah: That it is really dependent on the exact conditions of Earth right now. I think agency is very far from where my experiments are. So my experiments are maybe building that the lines on the graph that you would then plot, right? So we're just really building the chemical context for life to evolve in, and to actually generate that next step. So I don't think that I would say I have any agents in my system, and that I can get things that will have transient properties that could be considered life or individuals that have those things that we're looking for, but we are still very far from having a single agent that could be made in any context.
Michael: Something I haven't heard raised in this discussion so far is the thermodynamics of life, which I think bridges a number of disciplines. There's this growing, I don't know, momentum around this idea that life is a sort of thermodynamic inevitability because of the way that metabolisms come together and dissipate energy more effectively in a system. I'm curious what your thoughts on that position, whether you see... Like for example, is there such a thing as anti-entropy? Is that what life is doing? Is life cutting against that current of the second law or not, and in what ways might it not be?
Chris: I mean, I think one way to break that question down is to ask under what conditions do you get subsets of the system that have, surprisingly, more structure, able to then harness energy from other parts of the system. Right? So that's that the thermodynamic answer. I'm not sure negative entropy is the most useful concept because I think part of this is just very simply trying to count how many states you are, and you might have, and think about the likelihood of getting a very rare, complicated state in some sub portion of the system that then can operate on other parts of that system. So that requires certain types of energy flux for that to be true.
So there's not so much that the states are destroyed quickly, but also enough energy that you can maintain order, and so I think there are sweet spots for energy, but I think most of the standard thermodynamics we would think of are sufficient for thinking about living structure.
Sarah: I think one of the things that you said was that it's an inevitability, and if that was true, then we would see it on every body in the solar system, right? Every place that has a dynamic, a space, we would see that there were signs of life, and I don't think that we see that, and so the inevitability of this process seems to not be true to me. I don't know a lot about the theoretical energetics, but I do think that life works with entropy in a lot of ways to have something that we perceive as more ordered. For example, oil separating from water is driven by entropy, but you wouldn't look at a salad dressing that is phase separated and think, "Oh, that's more disordered than if the oil was mixed into the water." Right?
Again, this is one of those gut reaction things where we're like, "Oh, this is entropically unfavorable." So if you're not familiar, entropy is just how, how much disorder there is or how many possible states a system can have, and how many of those states it is trying to occupy. So if there's one state and that one state is occupied, that would be low entropy. If there are a hundred states and it is switching between all 100, that'd be high entrppy, and so you would think that life having many possible states, it sometimes is working with entropy. So even protein folding is, in part, entropically driven.
David: Yeah. Just to build on this a little just to clarify, and that slightly paradoxical insight. I think Shawn's going to talk later in this podcast, and he has a nice example I like, but the thing that's shocking, I think... So the question Michael is asking is: Do potential differences that allow for the possibility of current flow generate as they move towards equilibrium generating entropy, complexity?" One form of transience we stabilize complexity with agency is life, and the example that Sean [Carroll] gives that I always like is you go into a coffee shop, and you get a fancy latte-like drink that has cream on the top, and coffee underneath, right? So that's an ordered state.
It's a boring one, but it's an ordered state, and it's easy to describe. Now if you leave that in a room at room temperature, what does it do? If you come back an hour later, what will that look like?
Audience Member: It's mixed up.
David: It's mixed up. Right? It's kind of boring again, right? It's just all brown, and so you go from one boring state, which has been separated, to another boring state, but between those two states, you produce vortexes, and because you're not at zero degree Kelvin, you produce currents, and the description length of that coffee is vastly greater in the transient than it is between the initial ordered state and the disordered state, and some people would argue that the coffee cup is maybe alive just for a short while, while it's cooling down. I might be one of those people. So there's this law of the universe that's as fundamental as the second law of thermodynamics, which is the law of complexity production, which is between order and disorder.
There is complexity and somehow in that state, something maybe special has to happen to the chemistry to create agency, but maybe it's already there. I don't think we know.
Chris: Well, and I'd also say another nice thing to recognize here sort of building on what David is saying is using very simple chemistry and very simple thermodynamics, it's possible to get many of the features that we're interested in for life. So if you have just two chemicals interacting with a flux of one of them, so that's effectively an energy source, and you have the right decay in the system, then you will get regions of high concentration in one of those chemicals that then self replicates itself, and you get these repeating cell division and death, and you from a lot of mathematical models would say that looks like what you want to have for cell dynamics. Incredibly simple system, right? Sort of boring, susceptible to perturbations will go away easily. That doesn't take much more than very simple chemistry and a little bit of thermodynamics.
Sarah: But in those systems, the number of boring solutions is much greater than the number of interesting solutions.
Michael: So to this point, the question of where do we look for an environment like the environment that we imagine life may have emerged on Earth is this highly contested topic. You've got people who are really, they're on the deep sea hydrothermal vent camp. You've got the hot springs camp, and these different perspectives. They have consequences for our space programs, for the energy, and the attention that we invest in different strategies for exploring the possibility of life beyond Earth, even if we're just using spectroscopy to analyze chemical signatures of other worlds out there. So you can get armchair speculative on this, but I'd love from where you stand on this, each of you, kind of where you imagine life may have emerged and why. Like why would those conditions have been favorable compared to someone else's version of this?
Sarah: In the right audiences, this would be the question most likely to start a fist fight. So yeah, some people feel…
Chris: Right.
Michael: You're safe here.
Sarah: ...very strongly about like their warm little pond, if you will, and so I think that there are a lot of people who are pushing something that works. Their unique chemical system works in this one location, and this has to be the one. I don't feel that strongly. I think that we can make unique systems in pretty much any environment. So any temperature range, any pH, any salt conditions, even if we're talking about different liquids, so not water. If we're talking about not liquids, so atmospheric kind of systems, you can find interesting patterns and chemistries evolving in those spaces. So I'm very unbiased. I have very little in the game there. If I can build something that's cool in a new place, I would like to do that, and let's see what we can get.
Michael: And that's how you avoid fistfights, folks.
Chris: Yeah. I mean I think my answer to that would be, I don't think we have the right theories yet to say, and I think in the history of thinking about life, the most dangerous and often wrong statement is, "Life can only do X or life can never do Y," and those typically turn out to be very wrong. So we'll never find an organism that lives at this pH or this temperature or whatever it might be, and so I think that that highlights that we don't have the right sort of theories for thinking about life, and I think those predictions get even more dangerous if you go to thinking about where life might first evolve or how many different ways or how many different environments, and how big or small that space is. I think we simply don't have the right frameworks, yet.
David: Yeah. One of my favorite remarks at this conference thus far came from the extremophile cities panel where I don't remember, it might've been Nicholas [de Monchaux] that made the point. He said, "If an alien species or group of species, maybe an alien phylum visits you at the Earth, they would consider the city as the dominant form of life, and humans their microbiome," and I thought that was kind of beautiful, and I completely believe it. I'm one of those people who believes that Hamlet is alive. I think that in the environment called the brain that is capable of speaking English or middle English, Hamlet is transiently alive, and propagates itself by virtue of being such a beautiful, tragic player.
So I think once you mathematize the concept of a system capable of propagating information into the future that modifies the environment in which it lives, so as to benefit itself, there are a range of different things that could be living, and I just think we have to be pluralistic. I think a very interesting one is the chemical one. Another interesting one is the pneumatic cultural one or the computer virus. If you have work on intelligence, I feel the same way. The monolithic perspective that there is only one of something is a disaster, and I'd much rather be thought about forms of life than one.
Sarah: Can I actually ask you a question, David?
David: Yeah, please.
Sarah: So what do you think is the biggest barrier to us actually developing this model? It sounds like you think that we have everything we need to do this and that we just haven't come up with it yet. So what would be an example of something that would really kickstart this theory, this like hardcore scientific theory?
David: Yeah. I think we have, so I would say that we have the mathematical theory to explain what life-like phenomena are. We have failed at creating one special kind, which is the chemical kind. So there's probably a laboratory experimental challenge that I don't think has been solved, but I think the theoretical challenge was solved a long time ago, and I think part of the problem is hegemonic. I think it's that the people who dominate the field of the origin of life are prebiotic chemists, but I actually don't think it's a problem. I think actually the theory of living systems is kind of come quite a long way, but the experimental realization is so hard. I think it's a technical problem.
Michael: So there's a... Oh, go ahead.
Chris: Well, I was going to say, and I think somewhere that sits in between the pure theory and the full chemical realization of a new form of life is lots and lots of interesting in silico evolution. We have learned a lot about evolutionary theory. We've built lots of interesting evolutionary models. We've come to understand a lot about what's required for something to adapt and persist simply by building simple organisms in computers, and putting those in simple or complex ecosystems, and understanding a lot about what happens there, and the theory typically holds up. I think we do have a middle regime where that has worked very well.
Michael: So another way of posing this similar kind of question, and let's limit it to just the chemistry of abiogenesis. The question is around the enormous search space of chemical possibilities, and the fact that we don't actually even really know how big that space is. We were talking about this a little bit earlier. What are some of the ways that people researching in this domain are trying to winnow down the possibilities, what we are actually looking for, at least in our own life history, if not possible alternate biochemistries?
Sarah: I think really strangely, instead of it being winnowed down, it has expanded and grown in maybe the last 30 years. When we first started looking at chemical mixtures, we were looking at ambient temperatures, and really mild neutral conditions that you find most bacteria in. Maybe now people are trying even more things like crazy pressures and crazy temperatures. If anything, because we didn't find an easy way in a mild solution, now we're trying easy ways in harder solutions or in weirder solutions. I think that because this didn't happen in the last 50 years, this is not easy, right? It's not that you're going to find the right environment and then suddenly it's going to happen for you unless you're really, really lucky. I don't think that thinking about it really hard is going to improve our... We've been thinking about it really hard for 50 years. Right?
That clearly is not yielding fruit, so maybe we need to try some other discussion. I think that this idea of using computer modeling or even using some of the robotic technology that we have to explore to not thoughtfully, but just like spam the space, and try everything as best we can, is maybe the next mechanism. We saw big jumps with computing by doing this, right? A lot of our computational models now just kind of randomly try different things, and see how we get to the best fitness. We just need to do that with chemical complexity.
Michael: That kind of parallel approach actually almost sketches out a narrative, right? An origin story, right?
Sarah: Yeah.
Michael: The Earth itself may have been running that kind of massive parallel experiment.
Sarah: Yeah, exactly. If we don't do it on the scale of Earth, maybe we will never be able to recreate it. Let's terraform.
Michael: So that's like that's like a [Jose Luis] Borges kind of issue, right? A map the size of the territory.
Sarah: Yeah.
David: Yeah. I object to the question. I mean, the question that has to be answered is what's special about chemistry, and a useful analogy to me is digital versus analog coding. The way to think about chemistry maybe in this context is that the Fabian statistics or the Pauli Exclusion Principle, or whatever it is that gives you discrete entities. This is something one of our faculty actually works on a lot, Eric Smith, what the periodic table does. It gives you Lego, and it gives you discreet forms of stability that can be combined into relatively stable aggregate discreet forms. What chemistry does is give you digital logic, right? And so with a very interesting combinatorial constructive aspect.
That's the more fundamental principle, not chemistry. I think it's so if you can describe that concept in mathematical terms or however you want to do that. You have category theory, that's what some people are doing, then you can get at the deeper constraints about what would allow for sort of an open ended, self-sustaining process. I think, again, we should just turn things around. It's not the contingency of the particular universe in which we live. It's something more profound about the discreet fault-tolerant building blocks, and once we can realize them, however we realize them, we'll be able to create. Maybe that's the technical challenge of life.
Chris: Yeah. I think building on that last idea, I mean the interesting thing to do there would... I mean what we ultimately want is a theory that says, "If I have this size of a combinatorial space with this types of possibilities, what sorts of lifelike things can I get in that? So if I design a computer system where I've specified a much smaller combinatorial space, then say all of chemistry, what then do I expect about the range of lifelike entities I could get in that system opposed to chemistry?" We want something that covers that full range, I think.
Michael: Not to get super meta, but this all suggests that the study of the origins of life itself is speciating, that it may be an instance like you were suggesting earlier. These technologies are an instance of this process.
Sarah: Yeah, I mean, I think there is, and all sciences do experience evolution of thought and evolution of ideas, right? We have continuously been updating our ideas on all forms of science, right? So this is a natural process. One of the interesting, more recent, I don't know, trends I guess that I've noticed in origins of life studies is that everyone used to just talk about RNA, and in my very first conference, someone, a graduate student, got yelled at because he was talking about proteins in an origins of life conference. That is not happening anymore. It has really changed, and now we talk about complex mixtures, and we are talking about RNA interacting with other things or maybe not RNA at all.
I think that we have pounded on the RNA door for very long time. That hasn't yielded fruit, and so now scientists are like, "Okay. Well, maybe that isn't the right answer," kind of growth, and I think that is promising because it's unlikely that there was only one chemical on Earth. Right? Like it was definitely a messy, messy system four billion years ago.
Michael: What do each of you consider to be some of the most exciting promising research in this area over the last few years?
Sarah: Mine.
Chris: Yeah.
Sarah: No.
Michael: Well, then you have to talk about your work in more detail.
Chris: Yeah.
Sarah: I do think that there are a lot of people who are trying to make not just one thing, but make a network of things come together from a more abstract point of view. So a lot of the artificial life work, a lot of this synthetic biology work, and we say that we're trying to create life from like modeling the early Earth, but the reality is that even using really sophisticated biology, we don't know how to make chemicals alive, right? Even using like if you destroy a bunch of cells, you can't put them back together. Right? We don't know how to do that, and so saying that we could model early Earth and do this, I mean it's very unlikely that that will occur, and so the origins of life focuses that I think are really interesting are looking at how small networks of molecules that could have been around on early Earth can evolve chemically.
So the chemical evolution processes that we're seeing are going to inform a lot of the research in the next few years.
Chris: I think, for me, one of the most exciting things is moving away from trying to design a very specific reaction that you think is the one that leads to an origin of life or will eventually, running that in the lab, and then being disappointed, and going more towards what Sarah was just saying about bringing in more evolutionary thinking. So I think what we've learned from evolution in the last hundred years is that it finds surprising and interesting solutions. It's good at searching spaces and so forth.
A lot of that's been translated over into computer algorithms like genetic algorithms, which are largely informed by evolutionary theory, and so I think bringing that perspective to the chemical space is a very interesting one by saying, "Let's build complicated chemical environments, but let's add to that selection and variation in an evolutionary process with the hope that that helps us understand the space that that searches." So on the explicit side, I think the actual chemical reactions that people are doing that's much more interesting, and then I think there's a lot of really interesting theoretical avenues, which David has alluded a lot to.
David: Yeah, just building on both of those. So I'll give you an analogy. So I'm going to claim, and I think Sara said this and Chris made this point yesterday, that we're about to experience a renaissance in artificial life. So if you remember neural networks, when I was in grad school they were laughed at, and I actually worked on neural nets. It was a joke, and then GPUs came along, and Moore's Law helped, and Big Data helped, and all of a sudden a technology that was considered a little bit naïve, these very simple computational models of nervous systems, became immensely powerful, and modern machine learning, as we all know and sometimes called inappropriately called AI, has reached a maturity, and I think the same kind of argument would apply to artificial life. I think we have interesting constraints on the simulation of living systems.
I'm not sure exactly what they are. Will we be able to mobilize GPUs the way that folks in machine learning did just to model chemical-like events? I think once we do that, there'll be this huge new field of artificial life, and it'll lead to a new faction of people who say, "That's not life at all!" In the same way that folks say machine learning is stupid and naïve, and it's not really like human intelligence. Thank God, by the way. I think that that would be my hope that there's going to be a renaissance in artificial life.
Michael: Any research in particular, like specific publications worth discussing in this area? You can talk about your own. I'm putting you there.
Sarah: So I think what I try to do is make these small cells, and the really cool part about small cells is that each one can have a unique chemistry, but you can use different unnatural techniques to select for ones that are better, and so maybe you're selecting for ones that are persistent. Maybe you're selecting for ones that are fluorescent, and that's a real biochemical tool. We just tend to make things fluoresce when they do a specific task, and then we can pull them out of the rest of the population, and so I think that one of the things that I'm really excited about that I'm doing is working on how you can get these really simple collections of chemicals that aggregate. They aggregate into really like when you look under the microscope, you think, "Oh, that kind of looks like a cell," right?
And so you can actually put these through rounds of selection, and select for different fitnesses, and I think that what we're going to see is that as we get better at making these selections... So it's a chemically challenging problem to pool the ones out of the population that you like. Once we get better and better at that, we're going to see that we can actually evolve specific features without actually having something that you would think of really as alive. Right? It's never going to escape my lab and evolve, but I can like kind of force it to evolve, and you can do this not just with my cells, but lots of different types of small individuals. So maybe like grains of sand or little oil droplets.
So instead of forcing some idea of what the chemistry should be, you can kind of let it happen and see what comes out of it, and I think that that's a really interesting tool that we're going to be able to harness with the increase in our technological abilities. So as things become cheaper and as our tools become stronger.
Chris: Yeah, I'd say there's a huge amount of interesting research, and so rather than highlight that, I'll just talk about our own so we highlight our own work.
Michael: That's good. Yes.
Chris: Yeah, along these lines, David and I are working on a paper trying to take a lot of these very abstract mathematical theories, and apply them to particular case studies, and hopefully that will help clarify this connection between a lot of the tools and framework that has been worked out, and has a lot of power, and how you start putting that into practice in very specific examples. So stay tuned for that.
Michael: Okay. So all three of you have contributed to this new course, that Complexity Explorer SFI's online classes. They've put together an origins of life course that's a little different than your normal kind of course in that it is this unsettled area.
Sarah: Yeah. So if you are finding this discussion really interesting and you want to learn more about origins of life, I'm teaching a course online. It's free, complexityexplorer.org, and the origins of life course will run through September. You get a certificate at the end. If you decide that you don't have time or it's too much, you're not going to be pressured to finish it. So don't stress out about it, but if you think this is interesting, go check out the course because I think that citizen scientists can play a big role in shaping how we talk about things, and how we bring ideas to the public, and it's better to have an educated public, and so the Complexity Explorer's idea of having these free open courses, it's a really good thing for everyone to be engaged in, and to at least try out once, and this would be a good one, if you're enjoying this panel.
Michael: Do any of you have thoughts on like what it's like to work on a course that is so up in the air, so contentious?
Sarah: Scary.
Chris: Yeah. I'd say very much in the spirit of the conversation that we've just been having, in this course we've really tried not to commit ourselves to any particular historical story, and tried to really present people with the broad range of tools and concepts that are needed to make progress in this world, and so that spans, what we know about modern biology, evolutionary theory, certain mathematical concepts, some detailed chemistry, but we have tried not to tell a very detailed history.
David: No, no, I love it. Just to get to your point, the only course that you want to take is the course that's up in the air. The only kind of knowledge that's interesting is the unstable form because you could actually contribute to it. If it was a little rigid and fixed, it would be totally dull and so you ignore it.
Michael: So the final is to build your own cell, right?
Sarah: Or write a set of equations that can be then used to predict the chemistry that is coffee mixing in a cup.
Michael: No obligation. We have time for a couple of questions, if anybody has a burning question. Yeah?
Audience Member: There's got to be an energy we don't understand here, we haven't explained yet, that exists. Tell me more about it.
Michael: So the question for the record is from like a vitalist position, right? This question, "Why is the assumption that life has an élan vital, right? Its own life force. Why might that be a mistake according to the people on this panel?"
Chris: Well, I think part of the answer is David's Hamlet example, right? So Hamlet is a certain living thing with a different transience, and in the example that you gave, the ways in which you're dead are progressing in time, right? So at first many of your cells are still metabolizing. They're still active. If you measured those, you wouldn't necessarily able to tell that the body had died, yet. If you measure circulation, you can tell that something's died. If you measure mental activity, you can tell even quicker that something's died. So I think it is a question about this transience, and what environment is maintaining a particular type of set of states or process? I think that can be explained very physically.
David: Yeah. So I do disagree with you. I think Chris is right. It does require new thinking and new theory, but it's not about energy. It's about information, and that's a very important distinction to make. So what's happened at different scales of your organization is that information is ceased to be propagated forward in time, and as Chris pointed out at the cellular level, as far as a cell in your toe is concerned, nothing happened perhaps for an hour or so, but in terms of the aggregate computation that you think as self awareness, that has been significantly compromised. There's no energy issue here. It's a computation informational one, and it's one that exists uniquely at a level of aggregation that you're aware of. So we need theories for that, but it's true. It's a challenge, but it's not about energy.
Sarah: I think this is another problem with how we talk about science versus how we think about our own lives, and so you think that if you were shot and buried that you are dead, but maybe if you had children you would actually still be alive. So if there's a single cell and it divides into two daughter cells, is the single cell still alive or is one of the daughter cells the single cell or are both the daughter cells, the single cell? So I think how we define ourselves being alive is only a unique production of our consciousness, and maybe if you had children and your information is still out there, and the way you raised your children and that is going through time, then maybe you are still considered alive from some perspective.
Michael: Yeah. To that point, also, David Eagleman wrote this fantastic book, Sum, which is what he calls “possibilian” philosophy, which is, “Well, we don't know, so let's investigate all of the different options.” And he writes something 40-something different versions of the afterlife, and one of them draws on this kind of pre-modern understanding that you die when people stop remembering you, when people stop telling your stories, and that there is... I think that speaks to this postmodern construction of life as an informational process, but in a way that doesn't limit the investigation to a particular contemporary scientific thing, but is able to draw on the cultural diversity of these perspectives.
Yeah. So this guy just died and fell in your lap. I'm sure you have a question.
Audience Member: I'm a little traumatized, but I’m okay. What do you think about the Great Filter hypothesis and how that’s why we aren’t seeing any extraterrestrial cultures?
Michael: So again, for the record, the question is: If life is likely, why don't we see it? And then there's been this conversation around the notion of a great filter that there may be these crucial moments in the evolution of a biosphere at which point things are more likely to collapse, and that we just haven't gotten there yet, or somehow, we've dodged all of them so far, and so what are your thoughts on that?
Chris: Yeah, I'm going to fall back on I think there's lots of probabilities in the Drake equation, which is really just a nice way of formalizing the overall probabilistic thinking of getting complex life or communicating complex life. I think many of those probabilities we don't have a good sense on, and I've seen convincing calculations that bound many, many orders of magnitude in that probability by pretty careful reasoning about individual terms, and what the lower and upper bound on that could be. So, one place where we were very wrong, for example, was the number of stars that have planets, right? So it wasn't so long ago that we discovered the first star, I mean the first planet around a star that wasn't our own, and at that time the thinking was, "Okay, great, we verified it, that this happens."
"Maybe one or two percent of stars have planets," and then that probability has crept up all the way to effectively one. Every star that we see you can make a safe guess has a planet around it, and so that radically shifts the probabilities from something that's maybe close to zero to something that's one. I think how easy it is to go from a single cell organism to something with a slightly more complicated architecture, like a single cell eukaryote to multicellular life. That overall process, again, we don't have we have a... We have one historical perspective on that. We don't have a really good general theory for that sort of thing. So I think all of those probabilities or many of those probabilities are begging for more general thinking that we just don't have yet.
Sarah: I think there's a really expensive way to figure this out, which is doing missions to other places, and so if life evolved on Mars separately from the life on Earth, we might find remnants of it as bacterial fossils or something like that. Right? So, we are plan... NASA, not me, I'm not NASA, but NASA is planning a number of missions to moons and to Mars, and so we may actually end up seeing. Without having to account for a great filter, we might actually end up seeing some of these life forms that are no longer alive. So remnants of life, and that might inform a lot of what we know about these probabilities.
David: I think Chris, to sort of build on this, just to hammer it home. So what was it like two decades ago?
Chris: Yeah.
David: We didn't know that there were exoplanets. That's 20 years. Now what have we got, 20,000? I don't know how many, tens of thousands of them. That's 20 years, and now we have to work out a means of determining whether there are spectral signatures on the surface of those planets that correspond to what we think the chemistry of living systems is. I think it's way premature about great filters. I think that I'm going to put money on the table with you as a witness, now. I'd say within a decade, write it down, it's a sport person's bet, within a decade…
Michael: June 16th.
David: ...we'll find out-of-equilibrium spectral signatures on exoplanets that are consistent with life even as we know it, and so I don't. I think it's premature.
Audience Member: Where does the self-organizing system shift from, well, shift to life? Is it possible that we just want to call life something that we can relate to?
Michael: So if I'm to get your question right, as a philosopher, the question is: Is life emergent or is it a quality of self-organizing processes wherever we find them? Which, again, I think we run the danger of retreading some stuff here, but I think you got it in a little late.
Audience Member: Yeah, totally. Guys, I missed the first half, so…
Michael: Yeah.
David: Yeah. We answered that right at the beginning in depth.
Michael: But did life emerge more than once? In which case, we might have a problem of competitive exclusion of your question here. So, I don't know if you want to try to take a different angle on it.
David: I mean, I will say one thing. So I'm very sympathetic to that perspective. So we develop mathematical lenses, if you like, and what that mathematic... It's a little bit like Giulio Tononi’s work in consciousness, if you're familiar with that, and so you develop a mathematical expression, and think of it as a lens, and you look at the world through that mathematical expression, and if you see nothing through that expression, you're not seeing this life form, and we develop a variety of such lenses, and by holding them up to exactly what Chris is saying, by holding them up to reality, by looking through them, you find a certain kind of life according to the axioms and definitions that go into that mathematical expression.
David: So we have definitions of life that are captured by these quantities that we think we will observe, for example, in computer code, and so then that would be by a rigorous operational definition and legitimate form of life. Is it the fundamental form of life? Maybe not, but it's a form of life. So I think one can actually answer that question very mathematically and quite rigorously.
Michael: Yeah.
David: Yeah.
Chris: Yeah, and I think a lot of the way this may shift in time will perhaps happen in the same way that chemical thinking shifted throughout human history. So, originally people thought water was an element amongst other elements, right? It turns out water is not an element. There's a more fundamental chemistry. There are fundamental building blocks, but water does capture a certain phase of matter. It captures a very complicated physical idea, and so there was something to that, and so I think it may be that as things become more formal, we understand where something like life sits in some spectrum like that where there may be more fundamental things underneath, and there may be more complicated higher order things we use to explain it, but I think the phases of water is a nice way to think about how we might eventually come to understand life. Yeah.
Sarah: This is really outside my realm, but I think the one thing that you did capture is that when we talk about emergence of life, we often correlate it with or compare it to the emergence of consciousness, and so these two concepts are two really good examples of how really simple mixtures have properties that you would never predict. So that's what we mean by emergence, right? The chemicals themselves aren't very interesting, but they make these really interesting structures. Whatever terminology used philosophically to describe the emergence of consciousness, we would probably also apply those to the emergence of life.
Michael: Well folks, this has been fun. If you feel like you're unsatisfied, again, where we've got the origins of life course at complexityexplorer.org, and then if you want to listen to this again, we will be starting up the official Santa Fe Institute podcast later this summer. So just stay tuned, follow SFI on Twitter or whatever it is that you do, and we'll be sure to let you know about that, and thanks everybody for participating.