 The dominant paradigm in Western thought is that the world is composed of physical stuff, that space and time are fundamental. But could that be wrong? Might consciousness be truly fundamental? Do we see reality the way that it actually is? Or is the world that we see entirely in our minds? These are the questions I'm discussing with Professor Donald Hoffman on the 75th episode of Patterson in Pursuit. Hello my friends, I've got a fantastic episode for you today. We're talking about the fundamentals of metaphysics. Most people in the West are biased and we assume without really examining that the physical world is real, it's external to us, that our minds somehow grow out of the physical world. But there are a few modern thinkers out there that are rejecting this idea. They think that actually the physical world, the extent it exists, grows out of our minds. That consciousness and ideas and experience is fundamentally what the universe is made of. That's the real stuff. Now if you've never heard that idea before, it might sound pretty wild and out there. But that is the position of my guest this week, Dr. Donald Hoffman, who is the Professor of Cognitive Science at the University of California, Irvine. He's also the author of Visual Intelligence, How We Create What We See, and he's won a number of prestigious academic awards as well. One of the unique things in Dr. Hoffman's argument is his use of mathematics, which if you've been listening to the show for a while, you know I have very passionate beliefs about mathematics, and so it's interesting to hear how he incorporates mathematical arguments into his theories about metaphysics. In order to continue to bring super-high quality interviews like the one you're about to listen to, this podcast needs to continue growing, and leaving a review on iTunes and telling your friends about it is a great way to grow it. Alright, so I hope you enjoy my conversation with Dr. Donald Hoffman. Dr. Donald Hoffman, thanks very much for coming on Paterson in pursuit. It's a pleasure to have you on the show. My pleasure. Thank you for having me. You've got some very interesting ideas, and I want to talk to you about them. It's right up my alley for the topics of this show. You've got some ideas in the philosophy of mind. You've got some ideas in metaphysics and how they interact. You gave a TED Talk, which I think is a perfect summarization for what I want to talk to you about today, which is, do we see reality as it is? I want to start with a very basic framing of the conversation, and then that's going to take us right into some more advanced topics. There's a very normal notion that when you look out, we see the world. There's all kinds of objects out there in the world, and they're colored objects. In front of me, for example, I've got some flash drives. They're colored green. I've got a blue flash drive in front of me. Just everyday objects. We don't think twice about their existence, but there are a few lines of philosophic reasoning where people have made the case that reality actually isn't made up of ordinary objects as you think of them. In fact, reality might be totally different than that. What you're looking at is a representation of reality, which may or may not correspond to the external world. When people hear that, usually they're filled with skepticism and they go, oh geez, that's why I don't like listening to philosophers talk, because that seems totally count intuitive and crazy, but in a nutshell, do you think that there is meat to the kind of skepticism about ordinary objects? Is there more going on than what our common sense tells us about us inhabiting a world of, let's just say, colored objects? I do think that there's something more than what we think is going on. We see when we open our eyes a three-dimensional world full of objects with their shapes and colors and motions, and we assume that our perceptual systems are giving us a true description of what really exists, even if we weren't looking or if no one was around to look. Not exhaustive, no one thinks we see all of reality, but we tend to think that we see the reality that we need, and I think that that's absolutely false as it turns out. Okay, so what do you mean absolutely false, as in the world is totally different or there's no such thing as the external world? What do you mean? Okay, I am a realist in the sense that I think that something exists even if there were no living beings. I have life insurance, for example, so I'm getting hard earned money that even after I'm gone there's going to be something around. So I'm a realist, but I think that the forms of our perceptions of space and time and physical objects and their colors and motions and so forth, those are simply a species-specific way of getting information that we need to stay alive. And they're not in any way an insight to objective reality. They're just an insight into how we need to act in order to stay alive long enough to reproduce. Okay, so I'm sure you've heard this argument a million times, but it goes like this. There is an evolutionary case to be made that reality is roughly the way that we perceive it because if you, let's say you're some critter walking through the grass and you hear a rustle in the grass, if your senses tell you that there's a snake that's about to gobble you up and you're correct in that assessment, then you have an evolutionary advantage. But if your senses represent the world wrong to you and you don't think it's a snake, you have an inaccurate perception of reality but it actually is a snake then you get eaten and then you get kind of weeded out of the gene pool. Do you think that that's a flawed argument that evolutionary bias towards our senses being reliable? That is the standard argument and it's made by very, very smart people. It's in the textbooks and I think that it completely misunderstands evolutionary theory. But it certainly has a prima facie plausibility. It does seem reasonable to say that those of our ancestors who saw more truly had a competitive advantage over those who saw less truly and therefore were more likely to pass on their genes that coded for the truer perceptions and so after thousands of generations of that we can be pretty confident that we are the offspring of those who saw most more truly and so we see more truly most of the time in normal situations. That seems quite plausible but I began to look at this a little bit more carefully. It turns out evolution by natural selection is not just a hand wave of the theory, it's actually mathematically precise. John Maynard Smith and others have created what's called evolutionary game theory. There's evolutionary graph theory and so we can now actually begin to trace out what are the in fact logical implications of natural selection. And so I did this with some graduate students, Justin Mark and Brian Marion. We did hundreds of thousands of simulations of randomly chosen worlds in which we could put resources down wherever we wanted and we could control the fitness functions which we can probably talk about later what fitness functions are. We put organisms in those worlds that could see all of the reality or none of the reality and some that were just tuned to fitness and saw none of the truth at all and we could let them compete. And what we found was that organism, if you have an organism that sees all of reality, exactly as it is, competing with an organism that sees none of reality and is just tuned to fitness and the organisms are say of equal complexity, then the organism that sees all of reality is never more fit. And that means that natural selection, if there were a mutation that happened to grant an organism true perceptions, natural selection would then go about systematically stamping it out of the gene pool. So the result is, then it turns out I worked with a mathematician, Chaitan Prakash, I conjectured a theorem and Chaitan has proven it and the theorem basically says this. If our senses evolved and were shaped by natural selection, then the probability that our senses were shaped to see reality as it is is zero. So we have a choice. I and most scientists do accept that we've evolved by natural selection. Well, we've evolved and that natural selection plays a role. I mean, it's not the only thing that goes on. There's genetic drift, pleiotropy and a bunch of different things that go on in evolution. But the argument that you gave early on and that I gave as well is an argument from natural selection that true perceptions are fitter perceptions. That's a natural selection argument. And so what we proved was that if we evolved and there is any effect of natural selection, then the effect of natural selection is to stamp out true perceptions completely. Okay, so let me let's go a little bit deeper into that because it seems very counterintuitive that in this particular model, the creature that has true perceptions is at a disadvantage. So why would it be the case that that would be disadvantageous? Well, intuitively, I think the way to think about it is suppose you're playing a video game and to stay alive and get to the next level, you need to be hunting for places where you can grab points, killing stuff or running over stuff or grabbing stuff to get points. But suppose that instead of doing that, you spend your time just trying to see stuff and enjoy the scenery and so forth. Well, you're not actually playing the right game and you'll never make it to the next level. Unless you're actually going after whatever gives you points, you'll die or you won't go to the next level. And that's the way it is with evolution by natural selection. Evolutionary theory makes a clean distinction between objective reality, whatever objective reality is, and fitness functions, which are essentially the point functions that tell you how to get points. And the distinction can be seen quite clearly. So suppose, take for example, I have a nice juicy T-bone steak and I ask what are the fitness consequences of this steak say for a hungry lion? Well, for a hungry lion, that T-bone steak could offer a lot of fitness points. But now if that same lion was hungry and it was looking to mate, that T-bone steak offers it nothing. And for a cow in any state and for any action, that T-bone steak offers it no fitness points. So the way that evolution is typically thought about, the fitness points that you can get depend, yes, on the objective world, whatever that world might be, but also on the organism like a lion versus a cow. Its state, is it hungry or thirsty or looking to mate? And then the action, does it want to eat, is it going to eat, or is it going to mate? So a fitness function is an extremely complicated function and the state of the objective world, whatever that world might be, is only one aspect of the domain of that function. And so if you aren't going around, if your sensory systems are not all about telling you how to get fitness points, you'll die or at least you'll be less fit. And so what our theorem shows is that an organism that's spending any of its time trying to see the objective truth, that's time wasted because that's not time trying to hunt down the fitness points. And so an organism that spends all of its perceptual resources hunting down fitness points will always be more fit and therefore will out-compete an organism that sees reality as it is. And this means, to be very, very clear, it means that not only are physical objects like apples and tables and chairs and so forth, not objective reality, neither are space and time. Space and time and physical objects are simply a species-specific representation of fitness points on offer and how we can act to get them, but they're not the objective truth. In fact, they're there to hide the truth as we can talk about. So we'll definitely have to jump into space and time a little bit later because that's pretty, and that's very heavy stuff. I want to keep exploring this notion of the mathematical formula you guys are dealing with in this model that you've created. So it sounds like, so could I roughly summarize it like this, that organisms that are spending finite resources on perceiving reality as it is are inefficient and those which are spending their resources on reproduction and fitness are efficient. That's right. It's a waste of your time to be seeing the truth when you should be hunting down fitness points. The question that comes to mind would be doesn't that imply though that there's like a separation between the hunting and the fitness point and reality as if they're completely separate domains? Wouldn't they be actually completely overlapping because in order to get the food in order to reproduce, you still have to be tied to reality in some respect, right? Right. So I think the good analogy for what I think is going on here is the desktop interface on your computer. So suppose, and this gives you the idea of the relationship between perception and reality that I have in mind here, that evolution is done for us. So suppose you're writing an email to a friend and the icon on your desktop for that email is blue and rectangular and in the middle of your screen. Does that mean that the file, the email file itself in the computer is blue and rectangular and in the middle of the computer? Well, of course not. Anybody who thought that is misunderstanding the whole point of the desktop interface. It's not there to show you the truth of the computer, the diodes and resistors and voltages and magnetic fields. I mean, that's nasty, nasty stuff. It's in fact there to hide that truth and to give you the ability to control that reality, to do what you need to do while you're totally ignorant about that reality. Most of us have no idea how computers are put together. But if you had to know the reality, if you had to toggle voltages to craft an email, your friends would never hear from you. So we pay good money for interfaces because they hide the truth and they give us knobs and wheels to control the truth while we're totally ignorant of the truth. And so we have this simplified eye candy that lets us do what we need to do to send emails, edit photographs and so forth. And that's what evolution has done for us. So space time, as you perceive it, 3D space time, space and time is your desktop and physical objects are just icons on the desktop. And it's there not to show you the truth, just like the blue rectangular icon on the desktop doesn't show you the truth about the email. It hides it. It just gives you eye candy. So, you know, apples and cars, the sun and the moon, this is all eye candy that are part of the evolved user interface of Homo sapiens. And it's there to keep us alive. It's there explicitly, I would say, to hide the truth. Space and time has no resemblance to whatever objective reality is. And therefore nothing inside space and time resembles any aspect of objective reality. It's all a dumbed down user interface that we've evolved to stay alive because we don't need to know the truth to do what we need to do, which is to have kids. So to play the analogy a little bit farther with the computer interface, we pay a lot of money for engineers to have access to the underlying mechanics to structure things in such a way so that you and I have very high level illusions and can function through our desktop interface. But would the analogy be that nature is kind of the engineer? Nature is the one that has access to how the system operates and over time has structured things to an extremely high level of illusion and abstraction that may completely not correlate at all to what's actually going on. Exactly right. So the standard Darwinian selection story here would be that, of course, from the Darwinian point of view, there's no intelligent guidance going on in this. So what happens is you have random mutations which throw up random perceptual systems and most of them don't work. The ones that work are the ones that hide the truth and just happen to tune you to the fitness. And so you don't expect, so I'm definitely not claiming that our perceptual systems have evolved to give us the absolute best possible information about fitness. They're also just what economists call satisfying solutions. They're just good enough to get you by and that's what evolution has really given us. Our user interface is just good enough to give us the information we need to stay alive but it's not trying to give us exhaustive information even about fitness, but certainly not about truth. I definitely like some of these ideas specifically when we're talking about the visual field and the color representations, let's say, in the visual field. The idea, for me, it comes from a totally different perspective but the idea that there's redness in the world, like platonic redness, I don't buy that argument. I think redness is a way that our minds represent information to our consciousness. But I want to know in your theory how far this goes to other systems. If we're talking about the visual system, let's say that the world does not look the way that it appears to us but what about the conceptual system? When we're talking about reproduction as a concept, we're talking about fitness as a concept, is that also totally illusory and does that not correspond at all to something that's actually happening in the world? Great question. First on the redness, I certainly agree with you that redness and so forth is not an insight into the objective route. I would just point out that Galileo in 1623 said exactly the same kind of thing. This is actually a quote in English translation. He said, I think that tastes, odors, colors, and so on reside in consciousness. Hence, if the living creature were removed, all these qualities would be wiped away and annihilated. Galileo still thought that physical objects, their shapes and positions in space and so forth were part of an objective reality. I'm taking one step beyond what Galileo was saying. I'm saying that even physical objects and their shapes would be annihilated but even space and time themselves would be wiped away and annihilated. All of these things only reside in consciousness but now you raise another question which is at a couple levels, one, does this apply to our conceptual systems more generally, not just our perceptual systems? And I would also take a question this other way which is to say, haven't I really refuted myself? I used the theory of evolution by natural selection which assumes that there are physical objects such as DNA and organisms and resources that exist in a space-time world and I've used the theory of evolution to then propose a theorem which concludes that space and time and physical objects like DNA do not exist when they're not perceived. So haven't I really shot myself in the foot logically? And the reason is something that has been brought up by thinkers like Dawkins and Dennett and John Maynard Smith and that is that there's this notion of what they call universal Darwinism that at the core of evolutionary theory is an abstract algorithm that involves variation, selection, and retention. Anytime you have any system in which there is random variation, some mechanism of selection and some way of retention, then you can have evolution. So that is what is captured in the mathematics of evolutionary game theory, that abstract evolutionary core, the abstract Darwinism, the universal Darwinism. And so that's what I actually used in my theorem. I used universal Darwinism as it's embodied in evolutionary game theory. And the nice thing about that is as Dan Dennett has pointed out, universal Darwinism is like this universal acid. It tends to go through and wipe away a lot of our deeply held beliefs. And it's a lot more acidic than we might have thought. It's actually wiping away some of the extraneous assumptions of standard biological evolutionary theory, namely that space and time and matter exist unperceived. So the algorithmic core of evolutionary theory is what I'm using and that algorithmic core turns out to be incompatible with other peripheral assumptions of evolutionary theory. So I haven't shot myself in the foot. What I've shown is that the logical core of evolutionary theory is logically incompatible with some extraneous assumptions like space and time and physical objects like DNA that are assumed to exist unperceived. Okay, a couple questions. So are you saying that it is kind of a logical, as we just say it's a necessary conclusion that if this universal Darwinism idea is correct, that it must be the case there is no such thing as space and time and so on in an objective sense? What our theorem shows is that, I can't put the word must on it, but I can say the probability is zero. What would the difference be? Well, must, if you were thinking about it in terms of it, it was a logical impossibility. It's not a logical impossibility. Just like it's not logically impossible that I might win the lottery, but I wouldn't waste my time on it. But wouldn't that mean the probability is ever so slightly greater than zero? So if we were to say the probability is definitely zero, it seems like, at least the way that I'm thinking about things, that would be a claim that it must be the case that reality is not this way. Actually it's not. This is a slight technical point about probability theory that most people perhaps don't know about. That is that things that are probability zero can happen. And in fact, they can happen infinitely often. So that's why when I say something is probability zero, I cannot say that that means that it's logically impossible. It's logically possible for things that are probability zero to happen. And in fact, they can happen not only countably often, but uncountably often. So it's in principle. So that's why I'm sticking with the probability zero. I think the right intuition is not to say that there's a logical impossibility that we see reality as it is, but rather that it's the same kind of impossibility that holds when I say it's pretty much impossible for me to win the lottery. Okay. So in regards then to the concepts, to our conceptual screen, when we say something like universal Darwinism, that's a concept in our mind. Does that correlate to the world in some way? Yes. Now that gets to a really interesting point. As a scientist, what I try to do is take our best current theories and push them toward their limits. And so the logical structure of my argument is to say, let's assume, for sake of argument, that evolution by natural selection is true, as embodied by evolutionary game theory and evolutionary graph theory. So just suppose for sake of argument that it's true and the reason I'm doing that is because it is one of the pillars of modern science. It doesn't mean it's right, but it surely is, there's no other theory of biological species and their properties that comes anywhere near the theory of evolution in terms of explanatory power. So even if evolutionary theory is not correct, there's nothing else to compete with it. So why not start there? And so that's why I started there. And so the structure of my argument is really a conditional one. If we assume evolution occurs and that natural selection plays a role, then we find out that we get a conflict with physicalism. Physicalism, you cannot endorse physicalism and evolution by natural selection. They're incompatible, not logically incompatible, but the one, if we evolve a natural selection, the probability is zero, that physicalism is true. So the point that, and this is the way science works, belief in the theory is really not a good scientific attitude. I admire our theories. I think that they're fantastic tools. And the point of a scientist, a state-of-the-art scientist is say, okay, these are the best theories we've got right now. Let's push them to their limits. Let's see if we can break them. If we can break them, we shouldn't be upset about it. We should be overjoyed because now it's a chance to learn. And so that's what I'm doing here. I'm taking evolutionary theory, pushing it as far as I can. And for most people, this is a point of breaking. If evolution by natural selection entails that physicalism has a probability of zero, of being correct. For most people, that means we need to step back and look at some of our assumptions. Interesting. Well, so within that theory though, is there any distinction between the perceptual system and the conceptual system? There is. So I'll give you a concrete example. One aspect of our conceptual system is our conceptual ability to do logic and mathematics. Yeah. And the argument that I've given in my papers and then the theorem that we just proved that says evolution by natural selection entails that the probability is zero, that our perceptual systems are tuned to reality as it is. That theorem and my simulations do not bear at all on our logic and mathematical abilities. Those are utterly separate. And I can give you an intuition about why those conceptual abilities are not subject to my theorem about perception. Okay. And the reason is this. We have to be able to reason at least with elementary facility about fitness. For example, I need to understand, at least in a practical way, that two bites of an apple offer me roughly twice the fitness payoff of one bite. If I got that wrong, if I thought that two bites would actually be less fit than one bite and I acted accordingly, I would go extinct. So there are selection pressures for us to have elementary facility with logic and mathematics. I'm not saying that there's selection pressures to make us geniuses. Selection pressures are never there to make us geniuses. They're there just to keep us alive long enough to reproduce. So there are selection pressures from the standard evolutionary theory, universal Darwinism, to keep an elementary ability with math and logic. And then every once in a while, all the things aligning you get of von Neumann or David Hilbert or some mathematical genius that pops up. But most of us struggle. Most of us have some elementary ability and we really have to focus very, very hard not to make mistakes. And that's what you would expect from an evolutionary point of view. So I'll say this. There are some from a religious point of view who've tried to make a general argument from Darwinian theory, evolution by natural selection, to say that all of our cognitive abilities are suspect and are unreliable if we evolve by natural selection and then they take it as a refutation of Darwinism. And that's not at all what I'm doing here. I'm saying we need to look at each of our cognitive and perceptual abilities one by one and don't go broad brush. Look at them each one at a time and do mathematical theorems and proofs about what natural selection entails. Wouldn't that give the conceptual mathematical system a kind of empirical grounding? So when we're learning about the utility that might be broad or the fitness that might be brought by eating an apple, within that model, wouldn't we only grasp mathematical truths through empirical experience? And if that's the case, wouldn't that also put mathematics on the same kind of ground as physics where it could be possibly wrong? Well, so to answer that question, I now have to step outside of evolutionary theory. And I have to now propose an ontology. I have to propose a theory of what reality is. So the structure of my argument so far is to say if we take one of our best confirmed theories, namely Darwinian evolution by natural selection and we push it, we find out that physicalism has to go out the door. Space and time aren't fundamental. We find out that mathematics doesn't necessarily have to go out the door. I'm not saying that mathematics is fundamental or logic is fundamental yet. I'm just saying that I don't get there by evolutionary arguments, whereas I do get rid of space and time and physical objects by evolutionary arguments. I don't get rid of math and logic because I still have to use math and logic to do elementary reasoning with fitness functions properly. So to really answer your question, I have to now go beyond the... So right now everything I've done is sort of destructive. I've said let's use universal Darwinism to see what we can destroy. But now if I want to give a positive theory, I don't have space and time and physical objects, what theory do I want to propose? See now it's open to me to propose various kinds of theories, and I can propose theories in which perhaps mathematics and logic, according to my assumptions, are part of the objective reality. And then the burden on me at that point is then to construct a consistent theory and a consistent story about how I perceive that reality. So I need a theory about reality and a theory about how my perceptions relate to that reality such that everything works. If everything isn't consistent, then I've messed up. Okay, so would you say that if it's the case that mathematics and logic are not understood through the evolutionary process or that they have a more fundamental relationship with objective reality, would you say that our minds grasping of those concepts could not have come about through the Darwinian process? Well, I don't think that it's a logical claim that follows from Darwinian theory that we couldn't have evolved elementary logic and mathematical abilities even within the Darwinian framework. I don't think that that follows. Yeah, I don't think that follows at all. But I do think that if I'm going to propose a non-physicalist theory, and I do, I propose, because I'm interested in solving the mind-body problem, how is consciousness related to our perceptions of a physical world and to our perceptions of our brain. So I'm interested in solving that problem. So I have been working on a theory of reality that I call conscious realism in which I take consciousness and what I call conscious agents to be the ultimate nature of reality. And in my theory, mathematical structures and logic are an integral part of the nature of that reality. And so I do propose a theory of reality that allows that our perceptions of mathematical theorems and properties could be an insight into the nature of objective reality. Now, again, I could be wrong, but at least I'm not proposing something inconsistent. I'm proposing a theory in which mathematical properties and logic are part of the objective reality and in which our perception of mathematical theorems and properties could be true. But I leave open the possibility that my ontology is utterly false. You're proposing there are two different mental systems taking place that in mental system one where you have the perceptual faculty just through Darwinian evolution, through natural selection, you cannot get a true perception of reality or true representation of reality, but through the exact same process in a different system, you can actually get the truth about the mathematical structure of the universe or something like that. So there's two structures at play here. Is that right? Well, it's close. The idea is that even in our perceptions of space and time, the physicalist perceptions, there's mathematics all throughout that. And if you take a class in physics, you see mathematics everywhere in that. And so it's not like the physicalist theory of reality, is devoid of mathematics. The physicalist theory of reality also says that mathematical structures in some sense intrinsic to that reality. And when physicists talk about writing down the theory of everything, they're thinking about writing down some mathematical equation which captures everything, maybe hopefully on the back of a napkin. And so when I say that we have to let go of physicalism, it's not that math is wrong, because the physicalists also use math, is that the particular mathematical structures, for example, a three-dimensional space, one-dimensional time, Lorentz transformations and all these other things, those particular structures are simply the wrong structures to describe objective reality. They're more the way of describing our user interface. And so mathematics is being used by the physicalists, and I'm also using it as part of the nature of the objective reality. But what I'm saying is that the particular structures of space and time that we've described are, in fact, the wrong structures to posit for objective reality. Just like the desktop on my screen here is the wrong structure to posit for the diodes and resistors inside my computer. I can use mathematics to describe my screen. It's just the wrong structures for the true objective reality in this example. Interesting. Maybe you mentioned it before, so I might be just asking a question I already asked. Matt, the way you put that made sense to me, but what is the argument then that mathematics is that we do have true insight to the nature of objective reality and it is mathematics, if that's not just gained through evolution by natural selection, that that's our best concept about how we think the world works? Well, so I leave, certainly open the possibility that math and logic, as we experience them as humans, are not an insight into objective reality. I think that if you ask me, do I know that they're insight into objective truth, I don't know that. What I'm trying to do is build a scientific theory in which I go ahead and assume that they are and see if I can make a consistent system. I see. And this is the way science always proceeds. You never know. In science, you lay down assumptions and you don't know if your assumptions are true, but you say grant me for sake of argument, grant me these assumptions and then I'll show you what I can build and then we'll see if what I build is consistent or not and if it's not consistent, then we'll have to change things up. And so that's what I'm trying to do here, but you're asking the deeper question, can I know for sure where their math and logic are part of objective reality and I can't know that for sure. What I can say is given my current limitations as a human being, trying to do science, if you take away the tools of math and logic, I literally don't know where to go and I don't know what to do. So I find myself forced to use those tools. But again, and I don't have a theorem that says that my perceptions of math and logic aren't inside into the truth. So since I don't have such a theorem, I might believe that that's the case and I might assume that it's the case, but I can't know that it's the case. So it's something like an inescapable assumption that if it's true that our conceptual scheme is fundamentally flawed in such a way where we're screwed. That's right. And from the point of view of evolution by natural selection, I can put it this way. Let's say that all for the moment assume a physicalist ontology and the full Darwinian evolution apparatus, assume that theoretical apparatus. In that case, you can actually show that if you don't, for example, use all the laws of probability, there's a set of axioms for probability theory called the komogrovaxiums. And you can prove that if someone reasons in probabilistic situations in any way other than the komogrovaxiums and also Bayesian updating, then you can make what's called a Dutch book against them. You can set up a bet in which no matter what the outcome is, they always lose and you always win. So that's the coherence constraint on the mathematics of probability theory. And from an evolutionary point of view, I would say that that would cash out saying, look, if anybody was doing, and we all reason under uncertainty, so we always have to deal with probability. That's the way the world is. And so an organism that doesn't approximate proper probabilistic reasoning can have some other organism that's just a little bit better than it make a Dutch book against it, in a fitness sense, and make it go extinct. So you can make arguments within that framework to say that there could be selection pressures that would move us toward not exhaustively correct and brilliant probabilistic reasoning, but that would get rid of the any reasoning that's worse than your neighbors. You have to be better than your neighbor. Okay, that's a good way of putting things. And I think that would be a good segue to the next section, in which you mentioned ontology a few times and you're saying you're trying to construct an ontology kind of around this and see what follows and try to make it consistent. So in the ontology that you're constructing, what are the things that exist? If I were to try to take a non-physical standpoint and say, okay, what kind of objects exist? Well, there are colors, there are sensations, there are feelings, maybe I could say there are beliefs, there are ideas, there are concepts, but that doesn't get me anything outside my mind. Is that the ontology that you're building? Is it all within the contents of our own consciousness? Yeah, roughly yes, it's going to be, the ontology is that conscious experiences, the ability to make decisions based on those experiences and the ability to somehow act based on those experiences is, so that whole structure of being able to experience, decide, and act, something I call a conscious agent, and I have a mathematical definition of a conscious agent. So it's a completely rigorous definition and the claim is that objective reality consists entirely of conscious agents and their interactions. And so that's the precise claim, but intuitively it's exactly what you were talking about, you know, ideas, experiences, and so forth. I think you can get a feel for it by a very simple example. If you look in the mirror, all you see immediately is your skin, your hair, your eyes, and so forth. But you know firsthand that what you can't see is a very rich world, your beliefs, your aspirations, your fears, your love of music, your headache. It's a very rich world of conscious experiences that's completely hidden by the face in the mirror. No one else can see it directly, but you know that that face in the mirror, behind that, so to speak, behind that icon in the mirror, behind that face, is this infinitely, well, it seems infinitely rich world of conscious experiences. And I'm saying that's what I'm going to take as my ontology, is a social network, an infinite social network of conscious agents that have conscious experiences, fears, beliefs, hopes, desires, and they're interacting. And that will be one big social network. That's the idea. When you say a conscious agent, you're essentially just saying, I think people or people in their experiences and then the interconnection therein. Is that correct? Oh, definitely not. No, okay. Conscious agents, people could be conscious agents, but there's an infinite variety of conscious agents. Most conscious agents will have qualities of experiences that I can't even imagine. So anytime I'm interacting with anything, so for example, when I interact with you, I have some idea about your mental insight, who you are as a conscious agent. When I interact with a macaque monkey, I have quite a bit less insight into their conscious world. When I interact with a rat, quite a bit less an ant, even less. By the time I get down to things that I call rocks and so forth, I have no insight into any world of consciousness that might be hidden behind that rock. But I'm claiming that nevertheless that rock, even though that rock is just an icon of my interface and I've effectively given up on trying to understand the conscious nature of the world in conscious terms, nevertheless that's what I'm doing, interacting with conscious agents that might be utterly unlike me in every respect in their experiences. Is there a positive belief that those icons, so say, of the rock that there is an internal experience going on? Is there some kind of conscious agent there? Or is it that maybe there is and maybe there isn't? Well, I'm not a panpsychist, so I'm definitely not saying that the rock is conscious. In the same way that I, you know, for example, when I say that, to be really clear, suppose we go back to my blue icon on the desktop for my email and I say, well, there's this blue icon on the desktop, but the email itself in the computer isn't really blue and rectangular. What's really going on in this example is there's voltages in all sorts of circuits inside there. But then you might say, well, so are you saying that the blue icon on the desktop has voltages and circuits inside of it? No, no, no, no, no. The blue icon is just an icon. The reality that I'm interacting with is utterly, utterly unlike that blue icon. And the blue icon is pretty much given up. It's not telling me anything about those voltages and circuits. And that's what I'm saying about the rock. The rock is just my stupid symbol. And it's really given up on telling me about the conscious agents that I'm interacting with. It just doesn't have the power to do that. But I am interacting with conscious agents just like with the blue icon. I am interacting with voltages inside circuits, but there's nothing inside that blue icon that's voltages or circuits, you know, and nothing that resembles them. That's the same thing with the rock. So I'm definitely not a panpsychist. I'm saying something very, very different. The rock is just my symbol. It's a symbol, a very limited symbol, because I'm a very, very limited member of a humble species. I have very limited resources. And so I have to give up. The universe is infinitely complicated. So I literally ignore essentially all of it. And the part that I don't ignore, I use my very feeble tools that are there just to keep me alive. And so I end up with stupid things like rocks. And then what I do, what physicalism is, is I misinterpret the limits of my interface as an insight into objective reality. So that's what we do. We assume that the limits of our, you know, what are literally limitations of our perceptual interface, we mistake them for insights into objective reality. And that's, I think, is the real psychological source and genesis of physicalism. Yes, I, that last part I definitely strongly agree with that people mistake the contents of their perception for the way that the world is, the redness in the world. Even labeling objects as objects. When you call it a rock in the world, it might be the case that you're just putting boundaries around something, artificial boundaries, that if you didn't place those boundaries, the actual object wouldn't be there. The object is something like a concept. But you use, so this... I agree with you. I'm very interested in this term, the conscious agents. So is the, is there a belief that there is still a kind of uniqueness of perspective amongst these conscious agents? Because the way that I, the phrase conscious agent makes me think of like the you know, your person and my person and there's a difference between our experiences. I'm talking to you and I have the experience of talking to this microphone. You're talking to me, you have the experience of talking from another perspective into that microphone. Is that distinction still there? Are you saying that is the world populated by many of these perspectives? Or is it one layer more like, can we not even make it that kind of claim? It's both in the following sense that I see this as an infinite collection of conscious agents, but then it turns out the mathematical structure for conscious agents that I've written down when I analyze it, I can take two conscious agents, have them interact and it turns out that the dynamical system of two conscious agents satisfies the definition of a conscious agent. So it becomes a single conscious agent. This is a solution to what's known as the combination problem. Panpsychists have this problem and this gives a mathematical solution to that. Even though I'm not a panpsychist, I do have a solution to the problem of how do you combine not only conscious experiences but actual conscious agents points of view together. And so you could imagine and this is an interesting direction that I want to go with the theory is as I take more and more agents and combine them into ever and bigger and bigger, bigger agents is there an ultimate big agent and I don't know the answer to that. It's going to depend on the mathematics. There may not be one big infinite agent, there might be an infinite number of biggest agents. Mathematically there may not be a single top point of this whole lattice. It may have many top points to this graph and so I just don't know what the answer is going to be there and that's the nice thing by the way about doing a mathematically precise theory. And the reason why I love science is that when you write down a mathematical theory the theory then is smarter than you. You become a student of the theory. You don't know where it's going to take you and it may take you places you don't like. When Einstein wrote down the theory of the equation of general relativity, he had no idea that it implied black holes and he didn't like it and that's the power of these scientific theories. Once you write them down then the mathematics is there and you have to follow the math wherever it goes and so I'm excited to see is there a top agent or not. I think there's a lot of people who are of the religious persuasion that would be very interested in those kind of questions too because I know that's where I'm sure somebody would say well what we mean by the infinite agent is God. I don't think that's a difficult jump I think for people to make but I do want to ask you. There's a lot of this comes down to mathematics and you say you got to follow the mathematics wherever it leads. How do you come up with the mathematics in the first place? So what's the actual process of starting with the conscious experiences that you're having and saying I'm going to try to quantify them and symbolically represent them in this particular way. How do you even begin to do something like that? Well it's a fun and creative process. It's some of the most fun part of science is where you have to take your intuitions and sort through them and say what are the key intuitions that I have about consciousness and you don't want to you don't there's a matter of taster. You have to have just the minimal you want the minimal structure necessary. You don't want to have too much stuff. You don't want any redundancy. So the kind of thinking I did here is I have conscious experiences. That seems to be a necessary part of consciousness like as you talked about my experience of red the taste of chocolate and so forth. Now I wanted to say the minimal that I could possibly say about them so I just want to say that I need to be able to talk about the probability that I might have the experience of red or the or the taste of chocolate and nothing more. So mathematically it turns out one way to do that is something called a measurable structure. You can do it even more weekly. You can have what's called a sigma additive class but the point is that you get the I'm not using things like I'm not assuming there's a topology or a metric or any you know you know Euclidean space none of that stuff. I'm assuming the minimal that I could assume that's compatible with doing science and that is I'm simply assuming that I can speak probabilistically about whether I have an experience or not. Then the same thing about actions. I assume that I have a set of actions and again I just assume that I can speak probabilistically. I don't assume that that structure that set of actions is structured with a Euclidean space or topology or anything. I assume that has the minimal structure possible and then my decisions I simply model them with the minimal structure I could have there which is what's called a Markovian kernel. It's again the minimal assumption I could make about how decisions are made and I just start with that structure. I don't even have a notion of self consciousness. This is just raw consciousness but it turns out that the mathematical structure I wrote down just literally has a space of perceptions I call X for experiences. A space of actions G and a decision kernel D and then an action map A and a perception map P and a counter. That's it. It's just a really really simple structure and even the notion of memory is not built into it explicitly. The notion of self awareness is not built into it but using this formal structure turns out I can show how to build a self awareness. I can build a memory. So I'm trying to show that I've got this minimal architecture that will allow me to erect an entire theory of cognitive science and consciousness out of it. It turns out it's provable that the mathematical structure I have is what's called computational universal. Anything that's computable can be modeled in this formalism. So I know that I can model all of what we know in cognitive science in this formalism and I can build and I'm working with collaborators to build a theory of self awareness and memory and so forth out of those things. So when you are trying to build those more advanced features into your model like memory, how do you take it from probability that an experience will happen into something like memory? How do you do that? Well you're going to have circuits of conscious agents in which you can have reiterations and loops that will effectively instantiate what will be a memory for a higher level agent. The lower level agents that are involved in the memory themselves don't know that they're doing a memory but they essentially function to create a memory for a higher level agent and also to give it a sense of self. So I have to construct my sense of a body. My hand and my body are all sensory constructions on my interface. I literally don't know my body is not what I am. It's just an icon on my desktop like other icons and it's there to hide the truth. It's not there to show me the truth. So I am a mystery to myself. When I look at my hands I think I'm seeing the truth. I'm seeing a symbol that's there to hide the truth. What am I really? Why? I'm eager to find out. I hope this theory will help me to understand what I really am but my assumption has been to glib all my life that I assume that I know who I am. Okay, so I'm going to try to rephrase that and I'll probably do a poor job because I'm just running with some of those base concepts but if I make an error, please correct me. It sounds like the claim is the fundamental units of existence that we can know to some extent are pure experience, the contents of experience, taste, smells, feelings, colors, those types of things. If you codify or if you mathmatize the probability of having any type of experience you can then build all of reality based on the probability that an experience will happen. Is that the idea that literally the model of reality is identical to experiences or possible experiences? You've captured a huge part of the model there's one other aspect and that is the notion of agency so that I not only have experiences but I in my model I'm assuming that I can also choose how to act based on the and that I can act and so again these are assumptions that I'm making I could possibly be wrong and in fact the chances of me being right on the first go out are very very slow but the goal in science is to be precise so that you can find out precisely why you're wrong and that's what I'm trying to do here. So I'm going to be bold and make bold claims that agency and the ability to make decisions are on equal footing with the conscious experiences that you described as part of the fundamental structure and then I will see where that takes me and if I'm wrong then we'll try to fix it. So would you say this is like free will you're saying you're building the existence of a real type of free will into your model of what exists? Right so that's the way I interpret the what I call the Markovian Kernel Formalism for Decisions I myself interpret it as free will but I think the Formalism would allow someone who didn't like that interpretation to try to interpret it in another way I myself view it as a free will and I'm thinking about free will as a primitive but in an interesting sense it turns out that if I I told you I could build up conscious agents I take two agents put them together in a dynamical system they can create a higher level agent and I keep doing that so I can get really complicated agents well I can take one of these really high level agents and look at all the agents that are in what I call its instantiation and there could be you know countless agents that are part of its instantiation and so the agent at the top level has its own free will decision ability but then I can recursively unpack that into the free will decisions of agents at various levels of its instantiation and so what I'm going to get is not this trivial theory of free will I'm going to get a mathematically rich recursive unpacking of the notion of free will throughout an entire instantiation and the whole thing will have to fit together or I'm wrong so it's going to be a very interesting so the math that's the nice thing about doing mathematics in a science is that if it doesn't work the mathematics is going to put that right in your face and say it doesn't work okay so I've got two more questions for you one if you had a way to predict or to accurately put into some kind of formula the probability of any given experience happening and if you had a way to verify that that function that mathematical structure that you've built is accurate would you say that is a satisfactory theory so in other words if your theory can explain all of the experiences and their probabilities of happening that's kind of the end game that's what you're trying to do that would be just the start of the game what I have to do to be taken seriously by myself or my colleagues in the sciences I have to show that starting with an ontology in which conscious agents and their interactions is all that exists I have to show that when I project conscious agents back into the space-time interface of homo sapiens that I get back all of general relativity, quantum mechanics quantum field theory and also that when I look at the subset of my interface that I call organisms and their interactions that I get back Darwinian evolution if I don't get back all of that from my theory then I'm wrong so this theory is going to have some very very serious empirical constraints on it and in fact even that's not enough I really need to make new predictions I need to actually for example end up coming up with some modifications to general relativity maybe a quantum gravity and make new physical predictions that you cannot make with current modern physics or again I failed so the bar is high that's why I'm going after a mathematically rigorous theory here and I've got some physicists and mathematicians working with me we're headed toward it but I don't think we'll get there before dinner so that's a perfect segue to the last question which is about this relationship between the mathematics and reality you've got a way to mathematically model the conscious agent and I think that could be understood as the probability of a certain experience taking place is there a distinction in your theory are you making a kind of ontological distinction between the mathematics and the actual reality so are you just describing the actual agents in the world or are you saying that the real truth is the mathematics that you're actually getting at the fundamental which is the mathematics itself that's a great question and my attitude is that the mathematics is just a model and the conscious experiences are the reality in the same way that suppose I was building a simulation of the weather and I brought you said I want you to come into my air cooled supercomputer room but you don't need an umbrella when you go to see the supercomputer because the mathematics won't create the weather similarly the mathematics won't create consciousness nevertheless consciousness is not a blob it has structure consciousness is structured it's like the bones but if you just write down the bones then you've got a desiccated theory you've taken the life out of it so the mathematics is like the bones but you need to put the living flesh onto it and that's what the consciousness is so I don't think that the math is divorced from the consciousness you couldn't in some sense consciousness without mathematics is like a chicken without its bones it's not going to happen but a bones by itself is not the chicken either they have to be brought together and so I have that very dynamic feeling of a relationship between the mathematics and the consciousness but the math by itself is definitely just a theory it's not conscious experience I would love to hear and see exactly the argument for the reliability of our mathematical intuitions because I think you can make a pretty strong argument against the accuracy of our sense perceptions but that that one layer deeper which is about the math and the logic I think that that seems like that is such an absolutely fundamental assumption that needs at least some kind of ontological answer if we're conscious agents and we're conscious creatures how is it that we have this remarkable ability to build these mathematical models that say this actually describes that seems very remarkable I'm not saying it's wrong, I think mathematics probably is true it can tell you true things I just don't have a great explanation for how that's possible right now I agree so in my theory I will have to do that I think from the evolutionary point of view I think that the argument I gave about you need to have elementary ability with fitness reasoning two bites of an apple give you roughly twice the fitness path of one from within the framework of evolutionary theory you can put together a pretty compelling argument that if logic and math are parts of objective reality then there are selection pressures for us to the selection pressures are not universally against us having some elementary competence in logic and math just enough to stay alive but I agree if you're after the deeper metaphysical question about you know why should it be that math and logic are parts of objective reality that's a very deep question well that's a great note to end it on kind of we'll end it on another mystery this has been awesome I really appreciate the conversation yes it's been a great pleasure thank you very much okay that was my conversation with Dr. Donald Hoffman I hope you guys enjoyed it and found it intellectually stimulating I certainly did there's a whole lot more to say on this topic and one of the many things that I loved about this conversation is that it demonstrates just how radical a different world view will be generated based on different foundational premises I say it all the time and you guys are probably sick of hearing it but the foundations and the fundamentals are the most important ideas in your world view so difficult questions about metaphysics about mathematics the relationship between mathematics and the world relationship between our experiences and our knowledge all of these topics are incredibly interesting and incredibly important and there's so much work to be done so I hope you enjoyed that conversation and I'll talk to you next week