Watch Episode Here
Listen to Episode Here
Show Notes
This is a ~2-hour Q&A with Mark Solms (https://en.wikipedia.org/wiki/Mark_Solms) at the "Public conversation about their work and its implications for psychoanalysis" series of the Annual meeting of the American Psychoanalytic Association.
CHAPTERS:
(00:01) Intro and morphogenesis
(08:17) Embryos as collective minds
(19:38) Regeneration and cancer minds
(33:24) What counts as mind
(44:37) Emergence, TAME, and tools
(56:12) Levin's curiosity and origins
(01:06:52) Evolution and collective intelligence
(01:20:58) Panpsychism, language, psychoanalysis
(01:29:22) Q&A: self, feelings, aging
(01:40:49) Q&A: free will, immortality
PRODUCED BY:
SOCIAL LINKS:
Podcast Website: https://thoughtforms-life.aipodcast.ing
YouTube: https://www.youtube.com/channel/UC3pVafx6EZqXVI2V_Efu2uw
Apple Podcasts: https://podcasts.apple.com/us/podcast/thoughtforms-life/id1805908099
Spotify: https://open.spotify.com/show/7JCmtoeH53neYyZeOZ6ym5
Twitter: https://x.com/drmichaellevin
Blog: https://thoughtforms.life
The Levin Lab: https://drmichaellevin.org
Transcript
This transcript is automatically generated; we strive for accuracy, but errors in wording or speaker identification may occur. Please verify key details when needed.
[00:01] Mark Solms: This event is a new thing on the APSA program, called Interview with a Major Scientist from an Allied Discipline. Since it's the inaugural event, I had carte blanche to choose my most favorite major scientist from an allied discipline. There you see him on the screen. He is, in my unbiased opinion, the most interesting scientist working in an allied discipline to ours in the world today. So you're in for a treat, as you're going to see. Michael Levin is described as a developmental and synthetic biologist. He started out in computer science. Your first degree at Harvard was in computer science and biology. Then you shifted over decisively to biology for your doctoral degree at Harvard. Currently, Mike Levin is Vannevar Bush Distinguished Professor at Tufts University in Boston. He's the director of the Allen Discovery Center, and also the director of the Center for Regenerative and Developmental Biology. He's also co-director at Harvard of the Institute for Computationally Designed Organisms. Is that all correct, Mike?
[01:56] Michael Levin: It's Tufts and UVM, with Josh Bongard's group at UVM.
[02:02] Mark Solms: So all of that is just formality. Those are the titles that Mike has. Rather than tell you about all of those things, I'll just introduce you to the mind of Mike Levin and you will privately award him with prizes and honors yourself, I'm sure. If there's one word that I could use to describe the essence of what Mike Levin studies, that word is morphogenesis. And in order to explain what that word refers to, I'm going to read you something that's written on one of the websites of one of the centers that he directs. I'll just read you the opening sentence. It says the following: "The capacity to generate a complex behaving organism from the single cell of a fertilized egg is one of the most amazing aspects of biology." And that center seeks to understand that amazing aspect of biology. How does a complex behaving organism arise from a single cell that has been fertilized? That's what's meant by morphogenesis. But a wonderful and miraculous and complex problem as that is, it doesn't begin to describe the totality of what Mike Levin does. But that's the essence of it. As I said, if I have to reduce what it does to one word, it's morphogenesis. And that's what morphogenesis is about. How does one cell become a complex organism or organ? When I said at the beginning that Mike Levin is why I selected him for this inaugural event, which is now going to become an annual part of our calendar, an interview with a major scientist from an allied discipline, I said that he's the most interesting scientist that I could think of for this purpose. I think one of the reasons he's so interesting, if not the major reason, is because he never lost his childlike sense of wonder about questions like that one. How does a single fertilized cell become you and me? It is miraculous. It's amazing. I'm not sure it applies in biology as much as it does in neuroscience. I was inspired to study the neurosciences for the same reasons of childlike wonder. What is a mind? How does it arise in the physical universe? And how does it relate to the organ in our heads? I thought neuroscience was about that. And I quickly got that childish sense of wonder and amazement beaten out of me by my professors. I was taught you shouldn't ask big questions like that. It's bad for your career. You must ask much smaller, much more boring questions. And you either gradually identify with those smaller, more boring questions and get on with your career, or you do what I did, which is to jump ship and train in psychoanalysis. So that I could learn about the thing that I thought I would learn about in neuroscience. I'm not sure it does.
[06:16] Michael Levin: It applies. Oh, yeah, it applies.
[06:19] Mark Solms: Well, then my admiration for you is entirely deserved because you most certainly have not lost that sense of wonder. So with that introduction, let's begin. I'm going to ask you a series of questions. In some places, I'm going to be more in conversation with you than others. Mainly, it's an interview of a major scientist rather than a public conversation with one. But there are times when I will converse with you. Mostly, I would just like you to speak to our audience in this grand ballroom. My questions are prompts. Feel free to meander from the question because sometimes people like me ask people like you the wrong questions. So if you feel you've been asked the wrong question, give the answer to the question that you think you should have been asked. So let me start with that first question, the one on your website, the one about how does a single fertilized cell become a complex behaving organism like us? And just to convey the limits of my knowledge. I think this probably applies to many of us because most of us in this audience are psychoanalysts. We are not biologists. So when I say how does it happen, I have such naive questions in mind as how does one cell know that it's meant to be part of the heart and another one know it's meant to be part of the skin and another one know it's meant to be a bone. And how do they know when to stop? How do they know what they're supposed to be? How do they know when they've reached their destination? I now have become a FEMA. This is what I was always trying to be. Please, Mike, tell us.
[08:17] Michael Levin: I want to say right at the beginning that this is an incredible honor. Thank you so much for having me. Mark, ever since, I think it was about 2002 when I first read "The Brain and the Inner World." At that time, I had no inkling that at some point I would be able to talk to you about these things. And of course, "The Hidden Spring" after that and so on. I'm super pleased. Thank you for that amazingly kind introduction. I've registered for your Mark Solms in lockdown course, but I know very little about psychoanalysis. I'm going to tell you the things that I know. I do think that there's a lot of interesting overlap. I'll start here. Alan Turing, who is, of course, the father of computer science, very interested in artificial intelligence — in fact, intelligence in its general form, in different embodiments — wrote a paper towards the end of his life. He wrote a paper on morphogenesis, on the appearance of order from disordered chemicals and embryonic development. You might wonder why somebody who's interested in intelligence and computation and reprogrammability would be interested in chemicals in the early embryo. I think Turing saw a very deep truth that we have been pursuing since then, which is that the self-assembly of the body and the self-construction of a mind have very, very fundamental invariances. They have symmetries. It's fundamentally the same process, and I'm going to argue that this is the case. The things I'm about to tell you are grounded in perfectly mainstream science that we publish, but my views on these things are in many ways not the mainstream view. The things I'm going to tell you are not what you're going to see in genetics textbooks. The first and most important thing that is of relevance to all of you that you encounter in embryonic development is that the story you've been told and probably have in mind about how the body assembles — as a set of genes having specific effects, mechanical rules, they interact, and eventually, boom, there's the body — is a very fundamentally incomplete story. Details to be figured out, but more or less solved. It's this mechanical process that rolls forward where the biochemicals have their say. This is not correct. What is really going on is a process of collective intelligence from the very first cellular division. Intelligence did not wait until brains showed up on the scene. Intelligence is extremely ancient. Development is not a process of mechanical workings out of biochemistry. It is a process of creative problem-solving. What happens here is that you have a single cell. The single cell starts to divide. Let's imagine an early stage where you have a nice flat blastodisc. There's maybe 100,000 cells or something. We look at this and we say, well, there's an embryo. There's one embryo.
[11:32] Michael Levin: What are we counting when we say there's one embryo? There's 100,000 cells. What makes that one embryo? What makes it an embryo is that all of those cells are committed to a particular worldview, they all have a goal that they're going to cooperate towards, and that goal is to reach a particular anatomical configuration. This is a problem-solving process because we know that we can introduce all sorts of barriers between them and their goal. It isn't mechanical. It isn't just rolling forward. You can intervene, and they will solve a wide variety of problems in order to get that goal met. They have a literal memory of their goal. That memory is bioelectrical. We now have the ability to see it. It uses exactly the same mechanisms as the brain to hold its memories of goals and past experience. It involves ion channels, it involves neurotransmitters, it involves electrical synapses, all the same stuff. In fact, the brain got its mechanistic tricks from this much more ancient system. And whereas our brain, at least in its simpler versions, thinks about moving us in three-dimensional space during behavior, that system — all of your body cells have all of that same machinery. And what it thinks about is shape. It thinks about how to get from where it is now. So morphogenesis, the reason we study it is because it is a kind of behavior by an unconventional collective intelligence. We are a collective intelligence made of neurons and some other stuff. Embryos are a collective intelligence made of all kinds of cells that navigate this anatomical space of possibilities. They have to have memories, they learn from experience, they make decisions, they have valence and preferences about what happens. They can make mistakes, they have perceptions that can be more or less correct about what's going on. All of that happens from the very earliest moment of development. And one kind of interesting thing is that when we say there is one embryo and one individual or one human, let's say in the case of a human embryo, is going to result from this process, that is not written in stone. What you can actually do is — and this is how twins and triplets and so on are formed — I used to do this in duck embryos as a grad student. You take a little needle and you put some scratches into that blastoderm; every little island, not being aware of the others, will build its own embryo. You will have multiple — you can have any number. The number of individuals, the number of minds or selves in that pool of potentiality is anywhere from zero to probably half a dozen or more. And so how many individuals are in that embryo is not fixed. It will determine during the process; it assembles itself. And so what's happening here is that all of these cells link together into an electrical network that processes information. It has memories of what the shape is that it's supposed to form. And it will do its best to reduce the error between the current state and the final state. Therefore, it will keep rolling. There's a process of stress minimization. There's a perception and so on. All of this is going on very early.
[14:50] Mark Solms: Amazing. Thank you. When you say your views are not mainstream, the talk on the street is that the mainstream is going to have to catch up with your views because the level to which Mike Levin has studied these processes is the level to which he understands mechanistically the processes that he's talking about, which is illustrated by something I saw in his lab. They're these little worms, I don't even know what their species name is, tiny little worms with big eyes, and their natural way of reproducing is they just divide. So they're immortal worms. They don't reproduce sexually; they reproduce by division. Part of this process: if you cut the worm in half, the tail grows a head and the head grows a tail. And then you've got two worms. So it's an accidental version of the same thing that happens in its normal reproductive mechanisms. Mike studied where the cut is made: how do the cells at the severed part of the organism know "I must become a head"? And conversely, on the head half, how do those cells know "I must become a tail"? So he studied, mechanistically at the finest level, what exactly is going on there. And this is when you made your discoveries about the electrical properties of these cells, which I wouldn't begin to be able to describe because I don't understand it. Why I know that Mike understands it is because once he had discerned what those mechanisms are, he then tinkered with them. Because he had understood how the tail produces a head and how the head produces a tail, he tinkered with the mechanism and made the head produce another head. So Mike Frankenstein—Mike Levin—produced a new species, a new worm with two heads. And now you have a species with a novel problem, never encountered in natural selection. What do you do when you've got two heads? So this is the sort of intelligence he's talking about: a creative problem-solving process. I wanted to tell our audience that so that they get some sense of the extraordinary work you're doing. Feel free to correct or supplement what I said about your worms. I want to go on to another question which flows from what you've just told us. Some of the things you've told us I'm going to come back to, collective intelligence. But for now, sticking to more basic questions. Given what you've described, clearly you're speaking about problem solving, but what about problems not solved? Presumably the things you're studying are terribly important to understand birth defects and neoplastic disease. Why do these cells produce a clump of tissue that is unwanted and pathological? I presume your understanding of processes like birth defects and neoplastic diseases will also give us invaluable knowledge for the correction of birth defects and the treatment of cancers. What about regeneration of lost organs? Not only in the sense of regeneration of a severed limb along the lines of what you've done with your worms, but also, rather than transplanting an organ, you just grow a new one. Please speak to those issues: birth defects, neoplastic disease, the correction of these things, regeneration of missing organs. Over to you.
[19:38] Michael Levin: Thank you. Let's start here. I'm going to say a lot of things about the mind of these cells and so on. A lot of people hear that and think this is a metaphor. I have a very clear criteria for everything that I say, and I try not to say anything that doesn't meet that criteria. The criterion is these ideas are of practical utility today in driving forward biomedicine and bioengineering. In other words, we live in an amazing time where the state-of-the-art is such that we can take some very deep philosophical ideas about the nature of the self, the boundaries between self and world, collective intelligence, all these things, and we are now at the place where they become empirically testable. I know the ideas I'm going to tell you about today are useful because they have already driven new scientific discoveries in our lab and in other labs. My criterion for all the things that I'm going to tell you are these are things that actually lead to new discoveries of things that were not observed using the prior standard paradigm. The biggest idea in this domain is that whereas current molecular medicine deals with the body as basically a mechanical machine and tries to micromanage all the molecular states, you can imagine that the neuroscience version of this would be if you wanted to control an animal's behavior or get an idea across to another human, you would have to reach in there and start rearranging molecules in the synaptic clefts to get it to happen. Obviously, we don't do that. The brain and the sensory organs provide an amazing interface by which we can exchange fairly minimal amounts of information and let the system itself rearrange its own molecular states however it needs to be. So when you hear me speaking, I don't need to worry about what happens in your synapses. You're going to take care of that. You've got this interface. What we've basically discovered is that the exact same thing works in the body. When our lab creates a second head on a flatworm, an ectopic eye on the tail of a tadpole, induces an animal to regenerate a limb where they normally don't, or repairs birth defects, what we are not doing is touching the genetics. We almost never touch the genetics. We are not micromanaging the stem cells or the gene expression or where the molecules go or what happens to the proteins. We don't touch any of that. What we are doing is communicating, literally communicating novel goals to the collective intelligence of the cells because we have figured out the interface, and, no surprise, it's a bioelectrical interface just like the one that we use with our nervous system. And what you can do is take advantage of the intelligence of that system by triggering specific memories with specific behavior-shaping protocols. We are not micromanaging the material. We are communicating, collaborating with, and cooperating with the competencies that are already there. Some examples: we've shown in model systems; we're not in humans yet, but we have a few efforts heading towards preclinical testing and clinical testing. We have shown repair of birth defects of the face, brain, heart, and gut. Birth defects induced both by genetic malformations and chemical teratogens, by reinforcing the correct bioelectrical memory of what the correct structures are supposed to be, can be overcome. We've done some work on regeneration, inducing regeneration of limbs in animals that normally don't. That whole story I told you about development is: I think that there really is no development. It's all regeneration. Development is regenerating the entire body from one cell. The planaria that Mark talked about are amazing because from a small piece of the worm, you can get the entire worm. Half of the human population can do an even more amazing trick, which is to take a single cell and regenerate the whole body from that. Regeneration, remodeling, repair, development are an error minimization process. It's a system that has goals, it has anatomical goals, and it works. I could tell you some examples, examples of creative problem solving using new and different approaches to get your goals met despite all sorts of circumstances. One of the most amazing things about this process is that right in front of your eyes you see the journey from physics to mind. This is why I moved from computer science and the AI that I wanted to do when I was young to developmental biology. You start with a blob of an unfertilized oocyte, and people look at this and say it's amenable to physics and chemistry, but there's no mind there, it's just a chemical machine. Slowly and gradually, you get from there to whatever it is that we are. That's a slow and gradual process. What's amazing about developmental biology is that it can make mistakes. That's fundamental because chemistry can't make mistakes.
[24:13] Michael Levin: Chemistry just does what chemistry does. There's no notion of a mistake in chemistry. But developmental biology can make mistakes because it can fail to meet the goals that it works very hard to achieve. This is very important and we can now develop techniques to literally see those memories. We do similar things to what neuroscientists do when they try to image the brain and do neural decoding to extract memories and so on. We can do that. We can rewrite some of those memories and thus, as Mark said, convince the cells of a planarian that they should actually have two heads instead of one. The reason that he said that it's a new species is because that memory holds. Once you've convinced them that a proper worm has two heads, if you then cut that worm into pieces, those pieces will continue to make two heads. It's literally a new lineage with a new anatomy and a new sensorimotor architecture. The last thing you mentioned was cancer. This is of a lot of interest to this group. The conventional story of cancer is that it is fundamentally a hardware disorder. That is, you've got some genetic damage. As a result, certain regulatory systems go haywire, the cells overproliferate, they form tumors and so on. I'm going to give you a different story about cancer. What I think is going on, and we've used this, is an example of where deep philosophical ideas about what selves are and how you set the boundary between self and world become therapeutics that we have now used to normalize and suppress cancer in vivo, again not in humans but in animal models. You can start with two questions. The first is not why do we get cancer; the first is why isn't everything cancer all the time? Instead of a chaotic group of cells pursuing single-cell agendas like overproliferation, why do they ever cooperate to build a specific structure? A related question is why in robotics and computer science we don't see the cancer problem. Why is there no robot cancer yet? There will be; right now there isn't. Why is that? It's because most of our artificial devices are built with a very flat architecture. The whole, you hope, has some intelligence and is smart, but the parts tend to be very passive and stupid. The parts tend not to be smart. That is not our architecture. We have a multi-scale competency where the molecular networks inside your cells have learning capacity. The cells themselves have learning, decision-making, memory, predictive, and other capacities. The tissues, the organs, every subsystem has its own agenda. It has its own competencies. Every system is constantly trying to hack its own parts and its neighbors in order to get certain goals met. That means the collective, the network of cells, can remember very large, grandiose goals. Individual cells can only remember tiny goals, like keeping their pH at a certain level or certain hunger bands. But the collective can remember these grandiose goals, like a limb with five fingers. Individual cells don't have any idea what a finger is, but the collective absolutely does.
[28:47] Michael Levin: In the salamander, if you start cutting off the fingers, the cells will build exactly what they need to build, and then they stop. They know when the goal has been achieved. What keeps all of these cells working together towards these large-scale system-level goals is an electrical network, just like in the brain, just like what enables our collective intelligence. The cells assemble into a network that stores these memories. If, due to stress, chemical carcinogens, or certain oncogenic mutations, these cells become disconnected from that network, their cognitive light cone shrinks. When I say cognitive light cone, I mean the scale of the largest goal they can pursue. Every being, whether biological or not, whether a cell or an animal or a human, has a cognitive light cone in the sense of the largest, in space and time, goal they can possibly represent and work towards. The collective has these large-scale goals about fingers and hearts and livers and so on. The individual cells, once they disconnect from that network, that cognitive cone shrinks down. The only goals they can now represent are local amounts of nutrition and pH. The boundary between self and outside world shrinks. Cancer cells are not more selfish than normal cells. They just have smaller cells. That boundary is plastic. During embryogenesis, the boundary grows because these cells become part of a collective that dominates their activity. When they disconnect and become cancerous, it shrinks and they have tiny cells and they go on. As far as they're concerned, the rest of the body is the external world. It's not them any longer. What do you do in the outside world? You do what's best for you: go where life is good, you consume. This is metastasis. The idea that cancer is fundamentally a dissociative identity disorder of the somatic intelligence. You have a somatic intelligence whose goal it is to achieve certain things in anatomical space, and it becomes fragmented and dissociated. I think there is a lot here that parallels this notion of a shrinking boundary that fragments the whole into multiple cells. All of this would be a very strange fantasy, except that the implications are that if you take those cells and instead of trying to kill them, as we do currently with toxic chemotherapy, or instead of trying to fix the genetic mutations, what if all we do is forcibly reconnect them to the rest of the collective? We're able to do that using electrical synapses known as gap junctions. We can control the voltage of these cells. Here's what happens. You have these animals; you inject a very nasty human oncogene, like KRAS mutations. Normally they make tumors that metastasize. If you take those cells and forcibly reconnect them to the rest of their cellular environment, the collective takes over. They continue working towards making skin and muscle. There is no tumor, even though the genetic defect is still there. One of the things I'd really love to talk about is the issue of organic versus psychological disease. What this is showing is that certain kinds of informational states repair and dominate hardware defects. These cells have a hardware defect. They have a terrible oncogene. If you sequence the genome, you would make the wrong prediction. You would say that there is going to be a tumor. I see blazing expression of oncogenes. But that's not what happens. That's because in this system, very much like in the cognitive system, you can learn and think and have states that are not determined by the hardware and that override the default functionality of that hardware.
[33:24] Mark Solms: That's an incredibly novel way of conceptualizing the problem. I want to reiterate for our audience that it's not just a concept, that you've actually demonstrated proof of concept on animal models where you've actually done these things. It's fantastically important and we are all in the long run going to be enormously indebted to you for these novel concepts of such enormous practical consequence. One of the problems with the setup we have here today is you can't see or hear the audience. When you said that these cells that give rise to malignancies have a dissociative identity disorder, they laughed, which shows they understood exactly what you're talking about. I think I take it to mean that they think it's a very apt analogy. Taking it from there, this analogy, I want to drill down now, picking up on something that you've said. Using different words, you've said it several times. I want to be clear, I want the audience to know that you mean it literally, that when you speak of the cell having an intelligence, of the cell or cell groups solving problems, that they are part of a collective, that they identify with, that they have minds. I want to be clear that this is literally what Mike Levin is saying to us. He's not speaking metaphorically as far as I understand. That leads me to be obliged to ask you, Mike, because remember, these are mental scientists, mostly psychoanalysts here. What do you mean when you use that word, mind? When you say a single cell or a flatworm has a mind, what do you mean? And specifically, do you mean that it is a conscious being? That's the nub of what's meant by mind. Please speak to that.
[35:37] Michael Levin: I'll just make a note to make sure I get to the consciousness issue after that. The first thing I have to say is that if we take the lessons of developmental biology and evolutionary biology seriously, then we are committed to a continuity thesis in the following sense. We know that we adult humans have minds, whatever mind is, we have it. Now we have to ask the question, how did we get here? If you track the progress backwards, both on an evolutionary time scale and on a developmental time scale, eventually you will end up as one cell. Now you have two choices. There are many people who think that terms like mind, machine, living organism, inanimate object are sharp categories. If you think that these things are sharp categories, if you are used to asking questions like, does it or doesn't it have a mind? Is it or isn't it conscious? then you have a particular task. You have to figure out where during this very slow and gradual process from a single cell to whatever we are this sharp category snaps into being. The problem is that neither developmental biology nor evolutionary biology offers any such place. There is no bright line at which you go from physics to suddenly you have a mind. My analogy with my students is the term "adult." If you want to run a court system, you need a definition: on the night of your 18th birthday you're an adult, and that makes things nice and easy and you don't have to worry about these grandiose questions. As clinicians, as philosophers, as scientists, we all have to understand that nothing happens on the night of your 18th birthday. Instead of some magical click over into a sharp category, what we really have to tell are stories of transformation. We really have to understand what's happened slowly and gradually to scale up the primitive competencies of single cells into the kind of problem solving, a first person perspective and everything else that we have. We have to be able to go backwards and ask that question. We'll talk about machines and AI later. We have to tell a story of how it is that we got here. That's the first thing: continuity thesis and an explanation of the scale-up. The second thing is that all of the mechanisms that we normally associate with underlying intelligence, behavior, mental activity, consciousness — all of these things are ancient. They're ancient evolutionarily and they're ancient developmentally. From the first moment of fertilization, even before that, the ion channels and neurotransmitters are present; there is serotonin around early on and other neurotransmitters — all of this is there and operational. All that's happening from the earliest stages when you were in fact a single cell, either a microbial cell or an egg, to the stage that we are now is two things. One is a rapid speed-up of information processing. Neurons are very good at processing information quickly, which was necessary once muscles came on the scene so we could chase each other and eat each other. Fundamentally, most of the computations that we do — with the possible exception of language and some other things — the vast majority of what animals do as part of their intelligence was already being done in another space. What evolution did was pivot these competencies from anatomical problems to 3D problems of motility and behavior.
[40:06] Michael Levin: Now we have to decide, what is it that we mean by having a mind? What criteria do you use when you look at something and you say it has a mind? This is very important because, as we'll get into, it's not enough to say you look like a human, so I know what to do here. We are going to have cyborgs and hybrids and various kinds of extended, augmented, and altered humans. We are going to have, which we already have, different kinds of novel animal creatures that are nowhere on the tree of life with us because they're either engineered or hybrids or chimeras. AIs and robotics and possibly aliens someday — all of these things — we have to solve the problem of other minds. We have to ask ourselves, what criteria do you use for deciding that something has a mind? Here are some options. One option is composition. If you're made of the same stuff, then you probably have minds like us. The individual cells and embryos and all kinds of weird stuff have all the same components that our minds have. In fact, we could argue for a long time on what a neuron is. It's not simple. I've been at neuroscience conferences where they talk about neurons and ask, "What's a neuron?" They'll write down things that describe neurons, and I point out that every cell in the body does this. If you're interested in the materials, there's massive conservation. Another way to go is behaviorally. You could say you have a mind if you behave as if you had a mind, meaning we can observe behavior associated with feelings and with solving problems to reach goals by different means. This is William James's definition of intelligence, which I really like. And counterfactual thinking, memory, learning, and all of these things. As a matter of experiment, not metaphor, if you use the tools of behavioral science and neuroscience in these other contexts, you make new discoveries and capabilities. In other words, neuroscience is generally not about neurons at all. It's about the multi-scale assemblage of selves by aligning parts into a novel emergent being. This does not require neurons. Many cells can do it. You have to be more imaginative in asking what problems they are solving, what goals they are trying to meet, what capabilities they might have, and what space they are working in. I mean that completely literally: if we are willing to test those things, and I can give you many examples of what we've done and what other people have done in a diverse intelligence field, if you're serious about specifying what criteria you use to recognize mind, you are going to have to find the same thing in all of these other systems. For that reason, I'll say this about consciousness. I don't have a new theory of consciousness to give you here; I'm working on it. But for the same reasons that you attribute consciousness to other humans — meaning composition, evolutionary continuity, and behavior — you should take very seriously, at the very least, the existence of consciousness in all other body organs. All of the same stuff, with the probable exception of formal language happening in the brain, is happening in the other systems in your body. The fact that you don't feel like your liver is conscious is okay because you don't feel how I'm conscious either. The liver can't speak for itself, although we might change that; we're working on letting you talk to your organs. For those reasons, we should take this very seriously. I am completely being literal when I say that these things have intelligence and problem solving, have minds, and almost certainly have consciousness.
[44:37] Mark Solms: Thanks, Mike. To contextualize this for our audience, my own work, Mike referred to a book I wrote called "The Hidden Spring," where I was trying to identify what the fundamental mechanism of consciousness is. I reduced it to a very basic capacity just for raw feeling. I tried to identify the brain structures that generate raw feeling. They, in my best estimation, boil down to upper brain stem structures, which exist in all vertebrates. This leads me to claim that any vertebrate must have feelings, any vertebrate must be conscious. There are testable predictions. You can say, if that is the case, then I predict that if I do this, the animal will behave in that way, including manipulation of purely hedonic variables, things that only have to do with feeling. I had a very hard time, I still do, convincing many of my colleagues that a snake is conscious, that a lizard is conscious. When Mike makes arguments of the kind that he just has, which I have to say have a certain reasonableness to them, a certain clarity, a certain logic, my heart thinks, if I can't persuade my colleagues that all vertebrates are conscious, how am I ever going to get them to take seriously what Mike is saying? I have to confess, the way I've dealt with that is I resist myself being fully persuaded by what Mike is saying. That's the rationale for my next question, Mike. Surely you're right, there's a logic in what you're saying. It seems an entirely reasonable, seamless argument. To put it in this crudest of terms, obviously, if we are, in all of our mental and behavioral complexity as human beings, if we all of us were once a single cell that was fertilized, then there's some basal form of everything that we become that is latent, is imminent in that cell. This is reducing to the absolute minimum what you're saying: there are minimal versions of those processes, which then just get scaled up. I get that. But here's the question. Is there not a point at which it's no longer the same thing? You could say waves of the ocean are made of H2O molecules. The waves of the ocean somehow are imminent in those molecules. But it's meaningless to speak about an individual water molecule as having the properties of water under various conditions. In other words, are they not emergent properties, which it's not really reasonable to speak of until you get to certain levels of scaling up the processes you're talking about? Speak to us about emergence and what would your answer be to a critique like that: yes, of course, everything has some precursor, but that doesn't mean that nothing has a beginning, that there's no transition at which point we can now say it becomes reasonable to speak of a mind, whereas prior to that, it's just the equivalent of molecules not yet being water.
[48:27] Michael Levin: In my framework that I'm developing for this, which is called TAME, Technological Approach to Mind Everywhere, I've used the following metaphysical stance. I want to commit to two things. First, anything that I say on this matter has to be empirically useful. It cannot be just an opinion such that someone else can have a different opinion, and we agree to disagree. All of these things have to have a strict connection to empirical experiment so that we can follow the consequences of our assumptions and see who discovers more new biology, computer science, and new capabilities in biomedicine. All of these things have to be decidable by contact with the experimental world. For the purposes of this question, I take all of these mentalistic claims, all these cognitive terms, to be interaction protocol claims. In other words, we are not saying what something is. I'm not sure we ever can say what anything actually is. What we can say is, here's the way I plan to relate to this system. I'm going to use this set of tools and concepts to relate to the system, and we're going to see how that works out for me. I'll give you an example. All of this is going to get back to the point of whether it is silly to talk about some of these very primitive kinds of things as kinds of minds. Imagine a spectrum of persuadability. It stretches from left to right. I'm going to give you four waypoints on the spectrum. The first is a mechanical clock. The second is a thermostat, the third is a dog, and the fourth is a human. There are a few interesting things that happen as you move from left to right across this spectrum. The first thing is that in order to have a fruitful, mutually enriching, effective interaction with each of these systems, you're going to use a completely different set of tools. With a mechanical clock, your main tool is rewiring. You have to micromanage all the details. You cannot convince it of anything. You cannot reward it or punish it. You will not get any traction that way. Once you're dealing with a thermostat, something interesting happens. You can use the tools of cybernetics and control theory to say I don't even necessarily know how the whole thing works. What I do know is that it's a goal-directed system, and I know how to rewrite the set point. Every homeostatic system, like your thermostat, has a set point. If you rewrite the goal, the system will take care of holding to that goal. In other words, it has a degree of autonomy. It's minimal. It doesn't have a lot of opinions about what you've set the set point to, but it's enough to have some autonomy.
[52:19] Michael Levin: Now that you have options besides physical rewiring. If you get an HVAC technician into your home to fix it, if they don't know that thermostats have goals, you've hired the wrong person, because they're not going to be able to have this interaction. Now you move on to something like a dog. You need to know even less about how this thing works. Humans trained dogs and horses for thousands of years knowing no neuroscience, but what you do need to know is that there's an interface where you can provide rewards and punishments and signals and communications, and the system will have lots of autonomy in complying with whatever you're trying to teach it to do. Then you go into behavioral science and that bag of tools, and then you move to the right and deal with humans, where it's on steroids and using the language and tools of psychoanalysis, psychology, and psychiatry; you can do even more. What we're getting at here is that when you say this system has or doesn't have a particular property, what you're telling me is this is the bag of tools I'm going to use with it. Sometimes people also use to try and argue against my view: they will use the paradox of the heap. They will say when you have a pile of sand and you remove grains, adding or subtracting one grain doesn't do anything, but we know there are piles and there are not piles. My answer to that is very instrumentalist. I'm not interested in arguing about what a pile is. What I want to know is if you call me and tell me that we need to move a pile, here's what I need to know. Am I bringing tweezers, a spoon, a bulldozer, a shovel? What is the relationship we're going to have with this thing. That is what we need to know. There will be overlap and these are not sharp categories. The reason I think these things are not silly and provide very effective perches on the physical world is that we now know experimentally that when we appropriate concepts and laboratory tools from behavioral neuroscience and apply them to cells, tissues, and even subcellular networks, we've shown Pavlovian conditioning in subcellular molecular networks. When we apply these things, we make new discoveries and reach new capabilities and we're able to do things we couldn't do before. When you do this, it works. Now we have two views. You can have someone who says this is ridiculous: these things are only usable for brains and brainy organisms, and I'm only going to use these tools for those organisms; here are the limits of your world and the things you can discover. Or you can say those categories are holdovers from prior times when we had limited knowledge and limited imagination. It turns out that all of those things are not about brains at all. They have much broader utility. A few things we've used: various SSRIs, compound anxiolytics, hallucinogens, serotonergic modulators, training of various types including Pavlovian conditioning, instrumental learning, sensitization, habituation, and perceptual bistability. We've made planaria where the collective intelligence sees the rabbit-duck illusion. They have a bioelectric pattern that can't quite decide whether it is one head or two. When you cut them into pieces, each piece will settle on one and you get one or the other. All of these things work well in these other contexts. The tools do not distinguish. If the tools don't distinguish and if porting those tools leads us to new discoveries, I rest my case that that's how we decide whether these things are appropriate or fanciful.
[56:12] Mark Solms: I think it's a good point for you to rest your case. It is frighteningly compelling. I used the word "confession" a few minutes ago when I said I spoke of my own resistance to allowing myself to accept the compelling logic of what you say. I realize that in that resistance is an anxiety about not going along with the compact majority starting to hold out-there views and all the prejudices that come with that. So I'm going to veer off script, bearing in mind that I'm at a psychoanalytic conference, and ask you perhaps an inappropriate question, which is this. Maybe you've never thought about this, but I would be surprised if you haven't. Why is it that you do not start with the same constraints and conventions and prejudices as all of the rest of us do in science? The striking thing about you as a person is that you allow yourself to take things to their logical conclusion, even if it leads you to places where angels fear to tread. So what is it about you? How did you become like this?
[57:40] Michael Levin: I can tell you what I think. No one really knows. First, all of the things I see here are very disturbing to two sets of people, which pretty much covers almost everybody. There are the set of people who are mostly from a mechanistic reductionist perspective — my molecular biology colleagues, genetics, things like that — who think that it is completely inappropriate to use mentalistic concepts. Maybe some of them deny consciousness outright, but most of them will just say that something's going on in adult human brains, but otherwise it's a chemical machine and that's all there is to it. And you shouldn't port these other things. That worldview is completely incompatible with what I'm telling you now. There's another set of people who find it equally disturbing. That set of people are the exact opposite, and those are the organicists. The organicists have been rejecting and resisting this machine metaphor for probably several hundred years. Their point is different. They say, look, all of life has this ineffable magical quality. They're not quite vitalists. They have stories to tell about what it is that's magic, but it's basically living things have this amazing property. Machines don't. They like this sharp category. There are the living things that are living and cognitive. And then there are the inanimate machines, and the machines will never have this. They are not real minds; they are faking. In answering your question, I have an even harder bar because I have to explain why I'm weird on both counts. On my continuum, there are no sharp categories. There is no sharp category of life. There are no machines in the sense of dumb mechanical systems that only do the obvious things that the materials or the algorithms force them to do. I don't believe that exists at all. I think these cognitive terms can potentially apply to some extremely minimal things like active matter and some even much weirder things that I could talk about later. You haven't heard this stuff yet; it's even weirder than the things we have talked about. So why is that? I don't know. What I can tell you is this. I'll tell you about my childhood. The two parts I remember are these. One is that as a kid I had asthma. I was born in the USSR, and we had no medications. I was probably five or six and I had terrible asthma. The one thing that happens is your airway closes up, you get nervous about it, and that makes it close up more, and then you get more nervous, and it's a cycle. What my dad used to do to break that is we had this TV — you can imagine it's 1974 or something. It was this gigantic wooden thing with glowing vacuum tubes, and the screen was like this big. What he would do is take the back off the TV and we would sit there and look at the back. There was nothing interesting going on in the front of it. They never showed anything about it except collective farming and things like that.
[1:00:38] Michael Levin: So you would look at the back, we would sit there and look at the back and there were these vacuum tubes and I distinctly remember sitting there thinking, my God, somebody knew how to put all these things together in the right way. I would ask him, "Who knew how to do this?" He said, "There are engineers." I asked, "You can learn to do this?" He said, "Yeah, you can learn to do this." "OK, that's it." So then that was it. I decided I would be an electrical engineer. I did a lot of playing around with circuits. I had a slightly older friend; we would go outside, and he was into bugs and insects. He would lift up rocks and show me that these are the eggs and they're going to become the beetles and the caterpillars. Then the caterpillar is going to become a butterfly and fly off. This was my entire childhood: looking at these things together and trying to say, what's going on here? Clearly these things were put together as well, by whom and how, and they seem to put themselves together. Television doesn't much care what it shows. These things do care. It appears that way. What is the same? What is different? How did they get here? I read a ridiculous amount of science fiction as a teenager once we emigrated to Boston. I had access to all these things. I read a lot of science fiction where that community, since the 40s, and of course, in some cases, well before that, has been dealing with all of these problems. This fundamental problem of a spaceship landing on your front lawn, this thing trundles out. It's shiny and metallic looking, but it's holding a poem that it wrote for you on the way over here about how happy it is to meet you. Now you've got to decide. What is this? What kind of relationship are we going to have? How do I decide? That fundamental thing, which is the basis of a lot of science fiction — all the problems are already here. Once you've seen that, you can no longer think the traditional way. When I argue with people, they often don't see it. To me, this is the fundamental problem of other minds and unfamiliar embodiments.
[1:03:38] Mark Solms: Thank you for allowing me to veer off script and ask you that question, because what's astonishing is your answer restates the way I introduced you. I said that what distinguishes you as a very distinguished scientist is that you've never lost your childlike sense of wonder. And you've just told us, yes, exactly. That's where it started. You are exactly those questions.
[1:04:10] Michael Levin: I'll just say, being in the lab, I see things on a weekly basis that completely blow my mind and fill me with this incredible wonder about what we're seeing, that shatter all kinds of preconceptions. It's completely obvious to me that we know almost nothing about these big questions. The certainty that a lot of people bring to this — I know what this AI can do. It's linear algebra. I built it, so I know what it can do. The amount of things I see on a weekly basis in the laboratory that blow up all of our conceptions, I think it's very clear that kind of wonder where we don't come in thinking that we actually know what's going on is pretty critical.
[1:05:07] Mark Solms: I'm so fortunate to have visited your lab and to have seen with my own eyes these wonders that you do not shy away from and that you invite us to explore with you. Speaking of scripts, I must tell my audience that I intend to give them a chance to ask you questions themselves; at least the last 20 minutes I'll hand him over to you. I still have my script here, which fortunately is a little shortened because the last things you told us before I veered off into your childhood answered what my next question was going to be, which was whether there are fundamental distinctions between biological and non-biological intelligences. You've answered that without me even asking the question. What you've said now makes me want to insert a question I didn't intend to ask because of the way you put things a moment ago. What is your position on the fundamental question of natural selection and how teleological, goal-directed purpose arose? Do you have a different take on all of this? To take it to the extreme, do you have a different take on the question of intelligent design?
[1:06:52] Michael Levin: I'm not going to support intelligent design in the way it's formulated normally. But I will say this, the typical way that people look at this, which is completely consistent with how they look at all this other stuff as binary categories, yes and no. So we have two options that we've been given. The evolutionary process is completely random, blind, and maximally short-sighted. Completely blind and stupid. Or it is long-term guided by a massive superhuman intelligence. Those are the two options we've gotten. My whole commitment to the gradualism thesis is that those are not the only two options. Processes can have some degree of not being blind and stupid, but it doesn't mean they have long-term planning forward to chain reasoning or any of that stuff. So my suspicion is that evolution, I don't see any reason to think that it has long-term goals or anything like that. However, I also don't think it's completely blind and unintelligent. And the reason I say that is this. We've done computer simulations of this process. Evolution goes completely differently when it works on an intelligent substrate. If you are running an evolutionary search with random mutations and so on, a substrate that is mechanical. You have the genotype, the genes tell you what the characters are going to be, the traits, the behaviors, and that's it. There's a direct mapping from here. It might be very complicated, but it's just a direct mapping. Then you get the conventional story of evolution, and this is what everybody learns in evolutionary biology class. However, that isn't what we have. What we have is you've got the genes, you've got the phenotype, but in between, that process of morphogenesis is not a mechanical process. It is an intelligent problem-solving process where the genes are not determinants, they're actually prompts or incoming data that the process of morphogenesis has to utilize to get where it's going. When you do this with a process like that, one that is intelligent, the whole search process works completely differently. And this also ties into something you were saying, the distinction between biological architecture and current technologies. There is a major difference. What happens there is that just imagine that if you have a tadpole, and tadpoles have to become frogs, in order to become frogs, they have to rearrange their face. The eyes have to move, the mouth has to move, everything. You might think that this is a mechanical process. The genes somehow tell every organ which way to go and how far, and then you get your frog. Well, what we did was we made so-called Picasso tadpoles. You put the eyes on the top of the head, you put the mouth off to the side. It's like a Mr. Potato Head doll. Everything is mixed up. And what you get are completely normal frogs out of this, because all of these things will undergo unnatural novel motions to get you to a frog. So now think of what this means evolutionarily. If you have a material that is able to do new things to reach its anatomical goal, and there are some amazing examples of that, then imagine you have an animal with a correct face. It comes up for selection. Selection doesn't know, do you have a correct face because your structural genome was excellent or because your structural genome wasn't actually any good, but you fixed it? Selection can't tell. And so that competency, the ability to fix these defects, hides information from selection. It hides information about the genome. What happens then is that the genome has a hard time picking out the best genes, but what it can do is pick out the systems with the highest competency. But that makes it even worse because now it's even harder to see the genome. You start having this intelligence ratchet that once you get on this, the pressure to have an excellent genome comes off. The planarian genomes.
[1:11:24] Michael Levin: Planaria are immortal, cancer-resistant, highly regenerative. Their genome is incredibly noisy and dirty. That's because the pressure on the genome has fallen off because the algorithm is so amazing at making a perfect worm no matter what your hardware looks like. That kind of evolution that has intelligence baked in, where there's a feedback loop between the evolution and the intelligence, once it gets going, that's it. You have this incredible ramp up. That's what leads to the appearance of intelligent creatures, and because it's baked in at the very beginning, that whole process. Random mutations. The mutations aren't there to take you to some specific place, but because you're making it in a competent material, the process as a whole is not at all stupid and blind, and it is in fact doing much better than the standard story of evolution would lead you to believe. The final thing that I'll say is here's the difference between current technology and biological architectures. In the technology that we make, standard computers have two things. They have abstraction layers, which means that when you are working in a high-level programming language, you don't need to worry that your copper is getting hot or anything like that. You don't think about that at all, because if the computer's made properly, the higher levels don't need to know anything about the lower levels. You can trust them that they're reliable. That's the first thing. That's not biology. In biology, everything is unreliable. You have no idea how many copies of any protein you have. Things are getting degraded. You can't trust any of that. The other thing about computers is that they operate on the fidelity of information. Once you store information somewhere, there's redundancy, there are error-correcting codes, there's all this stuff designed to make sure that data stay there exactly how you put it there. Biology is exactly the opposite. Biology leans into the paradox of change. This idea that you can't remain the same, you'll die. And if you change, then you're also kind of gone. You have to change constantly because everything is unreliable. The environment is unreliable, your own parts are unreliable. You are going to be mutated on the evolutionary scale. You can't trust any of that stuff. What that means is that biology, from the very earliest moments, both evolutionarily and developmentally, commits to an active reinterpretation of the information it has, no matter what it meant before. It's a beginner's mind idea where I have this genome, I have no idea what this genome used to do or how it fit in previous environments. All I'm going to do is make the best use of it that I can now. I have many examples of that. That commitment to saliency and to solving problems as opposed to maintaining whatever the correct meaning of the information was is this creative urge at the very basis of life where you don't know. Now I'm going to tell you the cognitive version, at least what I think is the cognitive version of that. At any given moment in time, you don't have access to the past. What you have access to is the memory traces, the engrams that the past events have left in your brain and body. I think that biological cognitive systems, unlike current computer systems where the meaning of information is always the same, have to constantly reinvent the story that we're telling about ourselves. What do my memories mean? It's a story that is not trying to necessarily keep fidelity with the past; it's trying to be adaptive into the future. This is a story that I can concoct together to take me forward and decide what I'm going to do next. Biology, because of that—the same thing with the genome, the same thing with the information that embryos have—is constantly trying to; it's storytelling on a molecular, cellular, and behavioral level. You are constantly trying to pull together a new story of what you are, what the outside world is. You're making models, you're minimizing uncertainty and stress, but you don't know what any of it means. You have to create, you have to creatively come up with meanings all the time on the fly. That makes the evolutionary process completely different than the way we think of it on a passive material.
[1:16:00] Mark Solms: What you've just said, I think, demonstrates more clearly than anything else you've said so far why what you're doing is of such relevance to psychoanalysts. Because what you've just said about living organisms in general is our take on; that's what we deal with all the time. I like to say to my students, memories are about the past, but they're for the future. And I think you've just elaborated that point brilliantly. I want to go back to one of the fundamental ideas that have been a red thread through everything that you've said to us. And that is this question of collective intelligence. You spoke of how each cell has its own intelligence, its own cognition, its own problem-solving abilities. But you also said that it has a very small problem-solving horizon. Once it becomes part of a collective, the problem-solving horizon expands, and it's now contributing to the solution of bigger problems. That makes very good sense to me. But I want to have your take on what happens once you go beyond the individual organism. Does not the same process apply? Is there a fundamental difference to the collective intelligence you're speaking of when you say each one of the cells and organs of our bodies contribute to a bigger and bigger problem-solving entity? Is there a fundamental difference between that and, say, swarm or herd intelligence? If so, in what way are they different from the collective intelligence of the constituent parts of an individual body?
[1:18:14] Michael Levin: In principle, I think they're not different. The fundamental through line here is mechanisms that I call cognitive glue. These are just policies that enable individual parts, and those can be molecules, cells, individual organs, animals, groups. They enable individual parts to become aligned towards common purpose where the common purpose doesn't belong to any individual part, it belongs to the collective. I think that can happen on any scale, and Chris Fields and I have a paper on this scale-free biology idea where it's actually the same policies apply at all scales. However, especially with my story of cancer as this dissociative disorder, people very naturally map that onto certain social issues and they will say, this sounds exactly like selfishness in society. We should all just gap junction ourselves together into this giant syncytium and then life will be good. I want to point out two things. First, I don't think there's any guarantee that larger scale collectives are actually smarter than their partners. It can be, and in biology that tends to be true, but there's actually no guarantee of that. The group or the herd might actually have an intelligence that's a much lower grade of intelligence than its individual parts. That's certainly possible. Second, we're still, as a community, very bad at predicting when collective intelligences form and what goals they will have. It is entirely possible for large-scale systems to have goals that are quite detrimental to the goals of their components. So it's quite dangerous to form these large-scale collectives. You can't just assume that it's going to be good for everybody. The dumb example that I always use is, if you go rock climbing and you have a wonderful day, you've met some social goals, you've met some personal goals, great, you're happy, you left a whole bunch of skin cells on that boulder face. And they weren't asked whether this was a good trade-off for them. You don't think about it twice. That's common; collectives are quite tolerant of terrible things happening to their members. There's no guarantee. In general, it's exactly the same dynamics. Perhaps the optimistic version of this is that someday when we know what we're doing, we should be able to design policies that allow us to level up to a larger scale collective intelligence while keeping the kinds of things to which we are committed for the welfare of the individuals.
[1:20:58] Mark Solms: Freud was of the view that group psychology is regressive. In other words, a group does not behave better than its constituent parts. In fact, it behaves in a rather more primitive way. I have three more questions for you, and some of them are rather small. So I'm heading toward handing you over to Q&A. The small question, in the sense that I'm putting it to you very bluntly, but it's a very big question in terms of its implications, but I suspect you'll give it a fairly brief answer. Are you a panpsychist? In other words, is there something it is like to be absolutely everything? At least in biology? Does every biological entity have a subjective existence?
[1:21:57] Michael Levin: The short answer is yes. Slightly longer. You would be shocked at the set of things that I'm now considering being in that, because it's wider than biologicals, it's wider than embodied robotics and AIs and things like that. I've recently realized that I think it's much, much wider than that even. But I just want to say one thing, which is that I resist the framing of binary yes or no. So when we say, "is there something that it's like," I would like to maintain this idea of what kind and how much. And so I think that almost everything that you would name is somewhere on the continuum, but a great many things are so far off the continuum that very little, not zero, but very little of the tools of cognitive and behavioral science apply. I think in this world there probably is no zero in this universe. But yes, if we say that it's a continuum, then yes, I would say absolutely.
[1:23:04] Mark Solms: One thing in this conversation that you said really surprised me. Most of the rest of what you said, because I'm such a student of your work, did not surprise me. But you said twice, with the exception of language, what is so exceptional about human language?
[1:23:24] Michael Levin: I don't know. I'm being very careful. The things that I say sound wild and crazy, but I'm being extremely careful when I say all these things. The only reason I leave language out of it is because I have no evidence that these things have formal language in the sense of a limitless generative grammar. They might. I'm not saying they don't. They could well, but I have no evidence for it. We do have evidence for all kinds of other things, learning and decision-making and perceptual errors and all these. We have all this other stuff. I have no evidence of actual formal grammar use. So if someday we find it, I will not be terribly surprised, but I don't claim it because we have no evidence of it.
[1:24:08] Mark Solms: That leaves room for the possibility that reflective self-awareness is uniquely human.
[1:24:18] Michael Levin: It may be, except then the fly in that ointment is going to be that I don't believe that human is a binary category either, and we're going to have a problem with that on 2 fronts. One is the evolutionary ladder, where at some point you're going to have to tell me which of our ancient ancestors had it or didn't. I think that's almost impossible to do. So there's going to be some scaling, and this one has more than that. The other issue is that as we talk about humans, and this gets into the whole human and machine thing, we already have, and we increasingly will have, walking around combinations of standard human hardware with some sort of augmentation replacement, extra hemispheres grafted on, connections to other humans or the global internet. So what we say about those creatures is going to be, first, very important from an ethical point of view. Second, I think we're going to have to drop this yes human, no human thing. It's not going to serve us in the coming decades at all.
[1:25:33] Mark Solms: My very last question for you, Mike: in the interdisciplinary conversations that you've involved me in, which have always been absolutely fantastic, you've introduced me to people, incredible minds in various related fields, and I've always gained an enormous amount from taking part in those conversations. But in those conversations you regularly turn to me and say, "What would the psychoanalytic perspective on that be?" I want to put that question back to you and ask: What might the special contribution of psychoanalysis be to science in general and to the sorts of questions you and I have been talking about?
[1:26:28] Michael Levin: My amateur understanding of psychoanalysis, and it's mostly driven by conversations with you, is that it helps to understand the underlying reasons and causes for patterns of thought and behavior. To the extent that that's true, I think this is a critical, critical need, because as we continuously confront unfamiliar other beings, different kinds of embodied minds. This is all the products of synthetic bioengineering and humans modified in various ways. Every combination of 30% human brain cells, 70% something else, every possible combination of things is going to be part of our world. Developing an understanding of and helping these beings. I think in the future, your patients, the standard human architecture that we're looking at is going to be a minority of your patients at some point. You're going to be dealing with beings in this world that did not get here in exactly the same way that we did, but they have many of the same questions. They have stressors, they have existential concerns, they have problems about their life, they have questions about their own drives and behaviors that they don't understand. Why do I keep doing this? All of these things are going to be here, and the current tools that are being applied, which are basically the tools of biochemistry and physics, are not going to cut it any more than they cut it for your human patients. Nobody wants to hear a story of which protein went where when they're having deep, meaningful crises in their life. I really don't know what is needed from our end, meaning from your community, to develop effective psychoanalytic tools for beings that have never been here before. You have a rich tradition developing your craft for standard humans. I don't know what that would be like if the other kinds of humans we used to have had made it into today, and we had, whatever, four or however many there were. But that's nothing compared to what's coming. What's coming is an incredibly rich diversity of embodied minds and finding approaches to help those beings. A lot of the stuff I write, especially going forward that I plan to write, is not meant for these humans at all. It's meant for new beings that come into the world and really need to be able to understand themselves and us for us to all live together in productive ways. I think you guys are going to be the tip of the spear for that, for helping humans adjust to it and helping these other beings adjust to the world that they're in.
[1:29:22] Mark Solms: Thank you. Thank you, Mike Levin. I'm now going to hand you over to the audience. The first person is at the mic. Say one, two, three so I can see that Mike can hear you.
[1:29:37] Unknown: One, two, three. Can you hear me?
[1:29:38] Michael Levin: Perfect. Yeah, great. I can hear. Yeah.
[1:29:40] Unknown: Thank you so much for this. This is one of the most exciting things I've heard in a long time. I really appreciate it. It reminds me of when general relativity changed our notion of how the universe can communicate with itself as opposed to Newtonian mechanics. That's very exciting. And it sounds like you have some deeply subversive ideas, which reminds me of Freud, who is incredibly subversive. One of the most subversive ideas he had is that one of the main functions of the mind was self-deception. Indeed, from a psychoanalytic point of view, we're all lying to ourselves, constantly, a quintillion times a minute about why we think what we think, what we're experiencing, and so on. One of the most important ways that's coming up is in implicit bias research, which is "I'm not a racist," and then you find ways to befuddle, confuse, stun the conscious mind, and then you are a racist. So my question is, have you ever, or could you ever imagine an experimental paradigm to, again, reject binaries: could we see ways in which even subcellular structures intentionally omit awareness within themselves of other states, traits, functions, et cetera, in order to function more optimally? Could we imagine a continuum of that, or would we have to say, I just need a bulldozer to relate to it, that's what you psychoanalysts are for, it's part of the human mind, et cetera. I would love to hear your thoughts about that. Thank you.
[1:31:25] Michael Levin: Thanks. I appreciate that. That's a very interesting point. I absolutely think that this goes all the way down in cell behavior. It happens in two dimensions. One is laterally. What exists in the biological architecture are a massive number of sub-modules which are constantly trying to hack each other. Every subsystem is constantly trying to get all the stuff around it to do whatever it wants. They sometimes cooperate, sometimes they resist. I assume that happens too. You have different modules with different goals, and one module fooling another module by various things that it does, both locally and globally, is absolutely par for the course in biology. Then you have this other thing that goes through time, which I said before: you don't know what your memories really mean. One thing to think about is that memories are messages from your past self, and your actions are messages towards your future self. You're doing what the ecologists would call niche construction. You are changing your environment to make certain other options easier or harder, to deform the energy landscape for your future self. This is a dumb example. If you know that around midnight you're really going to want to go out and smoke, you might hide your car keys somewhere where you know that at midnight you can be too tired to go look for them. It's bending your landscape so that it's midnight and you're like, damn it, who are you mad at? Who puts your keys away? It's just your past self. So these kinds of things that shift the available signals, the available options, the perceptions and the actions of your future self are something that biological systems do all the time. Both laterally and vertically in time, there is a lot of that. We can say that we're lying to each other, but I probably prefer a more positive outlook on it: what you're doing is storytelling, and the point of the story isn't to be veridical to some kind of ground truth, which you're never going to have access to anyway. The point of the story is to optimize your life going forward. Looking at it, you say, I see how you did that. You completely neglected this thing that's actually true, and you made sure that you don't see that by shutting off certain sensors. I think that's what fundamentally is the drive of life: its sense-making. It's some version of optimally moving into the future, which may or may not correspond to what looks like truth from another perspective.
[1:34:21] Mark Solms: Thanks, Mike. I want to ask a follow-up to the last question. People are moving to the mic, so I'll be very brief. Because it cuts back to your earlier comment or your response to my question about "is there something it is like to be even the simplest?" This "something it is like," to me, the simplest form of that is just to have a raw feeling, that the feeling state. We feel things; this adds something — it's something new in the universe to feel something. And this is the essence of what we deal with in psychoanalysis. We're motivated by the feelings rather than what the feelings are there for. If I can take a typically psychoanalytic example: sexuality. Clearly the reason why sexuality is rewarded with such pleasures is because it's what keeps our species going. But what motivates each one of us to have sex is not to do our biological duty and reproduce. It's because it feels good. And so there's a kind of a gap between the feeling and what it motivates us to do and the underlying biological reason as to why such a feeling evolved. And that gap seems to me to be what the previous question touches on. Do you have a thought about that? I'm feeling guilty about having asked this because we now have two more people queued up. Don't give too long an answer.
[1:35:55] Michael Levin: I'll say two things. One is I completely agree with your emphasis on feeling, but I think there's a flip side, which is the need to act. What does it feel like? I don't have a good vocabulary. I'm still hunting for the right words. The flip side is critically, if you're an active agent, you are responsible for choosing the next action. Both of those things drive the behavior. In case I haven't said enough crazy things today, I'll say one more thing, which is that feeling goes all the way down to very minimal systems. The notion of geometric frustration — the kind of thing that if you're tiling the plane and there are certain shapes that don't fit right, or other Ising models and other mathematical systems where you can't get everything aligned. I think geometric frustration is actual bona fide frustration. It scales up, but I think it literally is a minimal version of frustration because alignment of your parts towards some goal is what beings are made of.
[1:36:57] Mark Solms: Great, thank you. Another question from the audience.
[1:37:02] Michael Levin: You spoke with great confidence about the future of a new kind of human, a mixture of mechanical. And I felt my personal defenses and anxiety rising quickly and then just told myself, that's okay, I won't be alive by then. Could you give us some science fiction thoughts about the time scale of that? I can't see you, so I can't even estimate, but I'm 55, and I think I will see this. We already have this. I don't think you can put this off onto our children or future generations. This is not sci-fi. We already have people walking around with implants in their brains that alter their cognition in certain ways. This is going to take off exponentially. In the next decade or two, we are going to have people with radically different sensory-motor capabilities. There are going to be people, the way that we see and hear the things around us, that have gotten sensors. I've seen people who get stuff implanted under their skin so that they can feel magnetic fields. They have little particles so that they can feel the magnetic fields around things. There are going to be people who perceive — what do you want to perceive? The solar, weather, the stock markets, political movements across the — whatever it is, there are going to be people who have those primary perceptions, the way that we have our five senses. There are going to be people connected to others. The sci-fi has already talked about all of this. I don't have to retread it. But this is not some far-off future. The technologies are here. It all works. Biology is incredibly interoperable for exactly the reason I said: every embryo is solving problems from scratch. And that means that if you find yourself next to some weird nanomaterial that's computationally useful to you, cells have no problem using it. I think we will all see this. I don't think this is far off.
[1:39:13] Mark Solms: Thanks, Mike. Next question.
[1:39:15] Michael Levin: I'm David Nickel from Denver, Colorado. I have to do a disclosure. We work on aging in my lab. There's a spin-off company that's funding some of this work. Take everything with a grain of salt. I think that aging is fundamentally a cognitive problem, not of the brain. It's a cognitive problem of the morphogenetic intelligence. Our latest paper looks at aging as basically a system that has met its goal, as in completion of embryogenesis, and somewhere after 18 it realizes that its goal has been completed, it has no more goals, and everything goes downhill from there. I think it's fixable. I think we are working on tools to remind the collective intelligence of the body of goals, or possibly the answer is you're going to need new goals. And so that means change and growth in ways that we've never seen before.
[1:40:42] Mark Solms: Thank you for that wonderfully optimistic answer. Here comes the next question, Mike.
[1:40:49] Mark Solms: Hello, thanks Mark and Michael for this fascinating discussion. I'm a faculty member at the University of Arizona. Your talk was intriguing — how you explained forming a complex organism from a single cell. There was also discussion on consciousness. That makes me think about a question that comes up repeatedly. It's a very simple question regarding free will. What are your thoughts? Does free will exist or not?
[1:41:26] Michael Levin: I don't have a quick answer on free will, but I can say a few things. I think it's critical to define terms. If you would like to have free will, you have to say what you are free from. You're not going to be free from physics. You're not going to be free from your own past experiences. You're not going to be free from certain aspects of your own cognitive architecture. I do think we have free will in a useful sense. Here's the sense of free will I think we have. I think we have free will on a temporally extended scale. I think we have very limited free will over what we do in the short term, but by consistent effort and by stacking certain actions in a very long chain, you have the ability to alter the behaviors that you're going to have in the future. That might be commitment to education, meditation, anger management, psychoanalysis. If you show up and continuously apply effort to bend your own cognitive system in the direction of your values, eventually what you will see is a shift where you now are doing more of the things that you find of value and doing less of the things that you think are not good to do. Therefore, you have exerted your free will to mold yourself in a way that is in accordance with your values. That, I think, is the sense of free will that we have. It's the continual application of effort towards large-scale goals. What we have free will over is our attitude and our interpretation of events. Using that and managing that towards modifying ourselves and our future in the way that we want. So I think that's a useful version of free will. I visualize it like calculating the area under the curve in calculus, where your free will at any given moment is infinitesimally small, but over time it actually adds up to a non-zero quantity. That's my simple version of free will.
[1:44:01] Mark Solms: Great. Here comes another question.
[1:44:04] Unknown: Greetings. Carlos Lopez, a clinical psychologist. Thank you so much. It's been an extraordinarily stimulating discussion. Is there a difference? What is the difference between humans and other creatures or things? This is really called into question. We talk about language. One thing that so many people raise is that humans have this conception of language, and we're also storytelling creatures. We're narrative creatures, and it's a constant narrative-making process. I would love to have your response to the thesis that is: is there something about narrative making which may just transcend our species and occurs in all aspects of potentially life or existence of consciousness? I would love to have your response or reaction to that idea. Thank you.
[1:45:05] Michael Levin: I think that all active agents, which certainly includes all life and maybe broader than that, are storytelling narrative creatures. Most of their stories are not verbal. Most of their stories are hard to recognize. We have a lot of mind blindness. We evolved with a very specific firmware that lets us recognize creatures in a very narrow range of environments, scales of time and size. But I think all of life, and probably much more than that, is fundamentally narrative-making, sense-making, storytelling creatures that are continuously trying to make and update a story about themselves and about their world that enables them to take actions in the future. Some of these are tiny, very minimal stories, and some of them are incredibly rich and complex self-models and world models, but they're all on the same spectrum. We're all in the same boat from the tiniest active matter system trying to persist in the world all the way through humans. We have to figure out what our memories mean, what sense we make of our world, and most importantly, what next. That's the biggest question I've always had: what do I do next? I think stories and narratives enable you to solve that problem and do something adaptive.
[1:46:41] Mark Solms: While I'm waiting to see if somebody else comes up to the mic, I'll make an observation and then if still nobody comes up, I'll ask a question. I think it must be because I'm in the setting that I'm in, Mike, that I'm looking at you with psychoanalytical eyes. It just strikes me as I listen to your responses to these questions that another thing that stands out about your mind is that everything you say is deeply consistent with everything else you say. As you answer these very different questions, there's a deep consistency in the underlying assumptions that you apply to each of them. And I think that that must be part of the courage that enables you to take your line of thinking to these unpredicted and unlikely places: you are following a line of evidence and a line of reasoning wherever it takes you. But still, nobody's coming up to the mic. So I'm going to go back to ask you a question that arose from one of the other questions about aging and mortality. One of the things that struck me when I was at your lab about these flatworms that you study is the fact that they, because of the way they reproduce, are genetically immortal. They never die. I know I asked you about it in your lab and you told me if you absolutely crush them to smithereens, they will die. But you can chop them up into many, many, many bits and they just all regenerate. Each bit makes a whole new worm with the same— and here's the nub of the question— with the same DNA. So because they don't have sexual reproduction, there's just an accumulation of errors in the DNA. And so I said to you, so malignancies must be a huge problem with these little creatures. And you said, no, there's no cancer in them at all. While we're waiting to see if anyone else has the courage to ask you a question, tell us about that. How the hell does that work? These immortal worms, why do we have sexual reproduction at all?
[1:49:29] Michael Levin: This drove me crazy for decades, and I finally, within the last year and a half, have what I think is an understanding of what's going on. The thing that drove me crazy about this is that these animals, exactly like you said, in sexual reproduction, the beauty of it is that if you get a bunch of mutations in your body during your lifespan, your children don't inherit those mutations. While going through the egg phase, that center of that bow tie going from animal to egg back to animal, it purifies all those somatic mutations. So you get this Weismann barrier. Planaria don't do that. They rip themselves in half. They regenerate. Every mutation that didn't kill the stem cell that got it is going to be propagated into the rest of the body. They are so genetically messy, they're mixoploid. Every cell could have a different number of chromosomes. They look like a tumor at that level. So that animal then has immortality, perfect regeneration, cancer resistance. What have you ever heard in any biology class that would lead you to believe that the animal with the dirtiest genome would have those properties? We are told exactly the opposite. The genome is critical. The genome sets what you're going to do — keep the genome clean. This is exactly the opposite. It made no sense to me until I finally understood the story that I told a few minutes ago, where what happens is that if you have a material that is competent in fixing certain errors, it makes it hard for evolution to select for the best genomes. Evolution ends up spending all of its time then augmenting the competency, which makes it even harder to fix the genome. So eventually there's this positive feedback loop that takes the pressure off of the genome and everything goes into the competency. What I think has happened is that planaria are a lineage where that process went all the way to the end, that basically it's completely maximized. Current planaria assume that the hardware is going to be junk, that there's going to be mutations and problems everywhere, and it does not rely on any of that. What it does rely on is an amazing algorithm, which we're just now starting to figure out, that enables you to build a proper worm despite all this garbage. That has two interesting implications that only this framework explains. One is that for 40 years, people have been trying to make transgenic planaria. It doesn't work. No one has been able to knock in genes into planaria. I think the answer is because they ignore their genome to some extent anyway. That's the first thing. The second thing is every other animal, you can call a stock center and get mutants. You can get fruit flies with extra wings and you can get rats with crooked tails and you can get all these mutants. Planaria: there are no genetic mutants. The only permanent lines of abnormal planaria are ones that we had made, and they're not genetic. It's the two-head line, and then the confused one that makes one or two. There are no genetic lines. This is not published, so take everything with a grain of salt. We've kept planaria growing in mutagen — the kind of stuff that you dip a fruit fly in for 30 minutes, and then you get all these mutants. We've kept them in mutagen for over a year. Nothing. They just — it's gone all the way so that the assumption is that your hardware is going to be junk and that you're going to need to do what you can to build an organism. Other organisms went partial along that spectrum, and I can't tell you why. There's probably ecological reasons. Salamanders go to some extent in that direction, and they're tolerant to a lot of things, including amputation of organs and limbs. Mammals are not quite along. You have other animals like C. elegans, the nematode, which is all the way on the opposite end of the spectrum. Every nematode is identical. Their cells are numbered. They all have the same number and position of cells. They're cookie-cutter. They're completely the opposite. So every lineage is somewhere, but I think planaria went all the way to the right. Everything went into the algorithm of being able to do the job even when the materials are not reliable.
[1:54:03] Mark Solms: Mike Levin, I'm very glad that you have demonstrated so fully to my audience why I introduced you as the most interesting scientist in the world today, working in an allied field to our own. Thank you for giving us two hours late in the evening for you on the East Coast, of your precious time, expanding our minds and perhaps blowing one or two of them.
[1:54:38] Michael Levin: Thank you so much, Mark. Really appreciate it. Fantastic. Good to see you as always. Great discussion. Thank you.
[1:54:50] Mark Solms: Until next week when we meet privately.
[1:54:52] Michael Levin: See you soon. Thanks. Thank you, everybody. Thanks, Mark.
