Making Sense with Sam Harris - #178 — The Reality Illusion

Episode Date: December 11, 2019

Sam and Annaka Harris speak with Donald Hoffman about his book “The Case Against Reality.” They discuss how evolution has failed to select for true perceptions of the world, his “interface theory” of perception, the primacy of math and logic, how space and time cannot be fundamental, the threat of epistemological skepticism, causality as a useful fiction, the hard problem of consciousness, agency, free will, panpsychism, a mathematics of conscious agents, philosophical idealism, death, psychedelics, the relationship between consciousness and mathematics, and many other topics. If the Making Sense podcast logo in your player is BLACK, you can SUBSCRIBE to gain access to all full-length episodes at samharris.org/subscribe.

Welcome to the Making Sense Podcast. This is Sam Harris. Okay, housekeeping. Well, the last housekeeping was intense. Got some new music all of you are dealing with emotionally. Got some grief over the new music. Let's just hang out with it for a while. See how we feel in the new year. Also dropped a paywall on the podcast. For those who need my rationale around all that, you can listen to the last housekeeping in the public feed. Those of you who are subscribers never even heard it. Anyway, to make a long story short, unless you subscribe to the podcast through SamHarris.org, you will only be getting partial episodes now. For instance, today's podcast is around three hours long, but if you're listening on the public feed, you'll get

the first hour, merely. So if you care about the conversations I'm having here and want to hear them in their entirety, subscribing through samharris.org is the only option. I'm having here and want to hear them in their entirety, subscribing through SamHarris.org is the only option. I'm clearly at odds with the trend here of all podcasts being free and ad-supported, but all I can say is that the response has been fantastic and the podcast is on much better footing even after only a week. So thank you for that. As always, if you actually can't afford a subscription, I don't want money to be the reason why you don't get access to my digital content, whether that's the Making Sense podcast or the Waking Up app or anything else that I might produce in this space. And the solution for that

is, again, if you can't afford it, simply send an email to support at samharris.org for the podcast and support at wakingup.com for the app, and you'll get a free year. And you can do that as many times as you need. We don't means test these things. There are no follow-up questions. This is based on your definition of whether you need this for free. And that's as it should be. So anyway, this is the business model. The podcast is now a subscription, just like the app.

And if you can't afford it, you can have it for free. Okay, so today I'm speaking with Donald Hoffman, and I'm joined by my wife Annika. This is the first time we have jointly interviewed a guest, and I'm sure it won't be the last. we have jointly interviewed a guest, and I'm sure it won't be the last. Annika's interest in this topic definitely helped us get deeper into it. Donald Hoffman is a professor of cognitive science at the University of California, Irvine. His writing has appeared in Scientific American and on Edge.org, and his work has been featured in The Atlantic, Wired, and his work has been featured in The Atlantic, Wired, and Quanta. And his new book is The Case Against Reality, Why Evolution Hid the Truth from Our Eyes. And there was an article in The Atlantic profiling him that made the rounds. He also had a TED Talk that many found bewildering.

As you'll hear, he has what he calls a user interface theory of perception, and many people find this totally confounding, and it can seem crazy at first glance, and even at second glance. And I must say, when I first read the Atlantic article and watched his TED talk, I wasn't entirely sure what Hoffman was claiming. As you'll hear, Anika got very interested in his work and had several meetings with him, and then we finally decided to do this podcast. And it is a fairly steep conversation. I do my best to define terms as we go along, but for those of you for whom this is your sort of thing, I think you'll love it. Over the course of three hours, we really leave virtually no stone unturned

in this area. We talk about how evolution has failed to select for true perceptions of reality. has failed to select for true perceptions of reality. We talk about Hoffman's interface theory of perception. We talk about the primacy of math and logic and what justifies our conviction there. We talk about how space and time cannot be fundamental to our framework. We talk about the threat of epistemological skepticism, causality as a useful fiction, the hard problem of consciousness, agency, free will, panpsychism, what Hoffman calls the mathematics of conscious agents, philosophical idealism, death, psychedelics, the relationship between consciousness and mathematics, and many other topics. And now, Annika and I bring you Donald Hoffman.

We are here with Donald Hoffman. Donald, thanks for joining us. Thank you, Sam. It's a great pleasure. So this is unusual. This is the first time that Annika, my wife, who's only been on the podcast once, many of our listeners will remember that podcast. It's the first time anyone has heard me laugh out loud in a decade. So you came to my attention on the basis of an Atlantic article, I think, that was making the rounds. And you also had a TED Talk.

I don't know which preceded the other. But then Annika just got completely obsessed with what you were doing. And maybe once a month or so, I would hear that there was some export from a conversation she was having with you. So it just seemed like it would be professional malfeasance for her not to really anchor this conversation. So. So, Annika. That was all in the context of writing my book. I was doing research for my book, and Don was working on a book on a similar topic, or really on the same topic,

with a different perspective. And so, yeah, so I had wanted his input on my manuscript and was honored that he trusted me with his manuscript, and we actually gave each other, And we kind of, we actually gave each other, we were kind of in the writing process together, so gave each other notes. And then Don was extremely generous with his time and continued to meet with me as I had many follow-up questions. And yeah, yeah, put up with my curiosity, even though

I'm not sure any of it was helpful to you, but it was great for me. It was very much fun for me and very, very helpful because you also gave me feedback on my book and really helped bring my book to a broader audience as well. So I was grateful. And I was really grateful that you did all the driving. Yeah, right. So before we jump into your thesis, which has the virtue of being on what I think is perhaps the most interesting topic of all, and some of the points you make are so counterintuitive, it's just seem crazy on their face. So it's going to be fantastic to wade into this with you. But how do you summarize your academic and intellectual background before we get started? How do you summarize your academic and intellectual background before we get started?

Well, so I did my undergraduate bachelor's at UCLA in what was called quantitative psychology. It was like a major in psychology and a minor that had computer science and math courses in it. And while I was doing that, I took a graduate class with Professor Ed Carter-Writton, which we were looking at artificial intelligence and ran across the papers of David Marr. This is like around 77, 78. And his papers just really grabbed my attention. Here was a guy that was trying to build visual systems that worked with mathematical precision, not just waving your hands, but actually writing down mathematics and something that you could actually build eventually into a robotic vision system. So I found out he was at MIT in the AI lab and what's now the brain and cognitive sciences department. And I was lucky enough to get to go there and work with him. He died a little over a year after I was there. So I only got to work

with him for 14 or 15 months. Yeah, very young. He was like 35, right? was there. So I only got to work with him for 14 or 15 months. Yeah, very young. He was like 35, right? 35. He had leukemia. But I did get to work with him and see how his mind works. It was revolutionary. It was a wonderful time there at MIT. And then my other advisor was Whitman Richards. David Marr and Whitman Richards were my joint advisors. and then Whitman was my sole advisor after Mar died. And so I was very interested in going there and the problem of, you know, are we machines? And I figured what better way to get at that question than doing something in an artificial intelligence lab where we try to build machines and understand the scope and limits of what machines could do.

So I was always very interested in human nature and how artificial intelligence is related to humans. Are we just artificial intelligences ourselves, just machines, or is there something more? And I didn't want a hand wave. I really wanted to understand what it means to be a machine and what might be different or not about humans. And so that's sort of my intellectual background. And what I focused on because of Marr was perception, visual perception. Yeah. So he wrote a book that was quite celebrated, a very early, detailed look at visual perception, which it's amazing what a contribution he made in such a short time. Decades after his death, his book is still recommended as a must-read book in cognitive science and neuroscience. Absolutely. It was brilliant, and he was brilliant in person.

The lab meetings were electric. He had assembled this world-class group of scientists around him. They congregated around him. And I just was so lucky to be watching this new science being revolutionized by this young man. Yeah, at 35, he did all this and died. It was truly stunning. Yeah. You're now at Irvine as a professor, right? That's right. University of California at Irvine.

Yeah. And you have been meeting over the years with some of the great lights in consciousness studies, for lack of a better word. There was these meetings of the Helmholtz Society. Isn't that what you were called? Yeah, the Helmholtz Club. Helmholtz Club, yeah. And that had Francis Crick in it. And I never met Francis, but Joe Bogan, who you write about in your book, is somebody who I did meet.

And he was quite a character. He was quite a character. He was fun at dinner. Yeah. He was the neurosurgeon who did the bulk of the split brain procedures for which Roger Sperry won the Nobel Prize. That's right. And Aran Zidell at UCLA was involved in that work, and Michael Gazzaniga. Yes. Yeah. Before we jump in, I want our listeners to be sensitized to how seemingly

preposterous some of your initial claims will be. And I can guarantee you that on certain of these points, the sense of their counterintuitiveness will wear off. And there's something thrilling about this. The thrill that was exemplified by Annika's obsession with your work, I know has spread to other people. We have a friend who, perhaps I shouldn't name, who claimed that she accosted you at some function and just completely fangirled you as a groupie. So we know that I think once you start wearing sunglasses indoors, you will have started a cult, and then we will put the word out against you. But in the meantime, perhaps the best place to start, I mean, I would imagine we should just track through it the way you do it in your book, starting with the interface theory of perception. But you can start wherever you want, and we just want to go through it all, and we'll have questions throughout.

one. We just want to go through it all, and we'll have questions throughout. Right. So most of my colleagues who study perception assume that evolution by natural selection has shaped us to see truths about the world. None of my colleagues think that we see all of reality as it is, but most of my colleagues would argue that accurate perceptions, what we call veridical perceptions, perceptions that tell us truths about the world, will make us more fit. So accurate perceptions, veridical perceptions are fitter perceptions. And the argument that's classically given is actually quite intuitive. So the idea is that those of our ancestors who actually were better at feeding, fighting, fleeing, and mating because they could see reality as it is, were more likely

to pass on their genes, which coded for the more accurate perceptions. And so after thousands of generations of this process, we can be quite secure that our perceptions are telling us truths about the world. Of course, not exhaustive truths, but the truths that we need. We see those aspects of reality that we need to stay alive and reproduce. And that seems like a really compelling argument. It seems very, very intuitive. How could it go wrong? So at first glance, it seems some measure of veridicality, some measure of being in touch with reality as it is, would increase an organism's fitness. There must be a fit between tracking reality as it is and adaptive advantage. Exactly. That's the standard intuition for most of my colleagues. Stephen Pinker has actually

published papers where he points out some contradictions to that idea. But most of my colleagues would go with the idea that, yeah, it's better, it's more fit to see reality as it is, at least part of reality. initial intuition was that maybe it would just take too much time and too much energy to see reality as it is. So evolution tries to do things on the cheap. So maybe the pressures to do things quickly and cheaply would maybe compromise our ability to see the truth. And so I began to work with my graduate students, Justin Mark and Brian Marion, around 2008 or so, 2009. And I had them write some simulations where we would simulate foraging games where we could create worlds with resources and put creatures in those worlds that could roam around and compete for resources. And some of the creatures we let see all the truth, so they were the vertical creatures. And others I didn't let see the truth at all. We had them

only see the fitness payoffs. And had them only see the fitness payoffs. And we can talk about what fitness payoffs mean. That's an important concept. But what we found was in these simulations that the creatures that saw reality as it is couldn't out-compete the creatures of equal complexity that saw none of reality and were just tuned to the fitness payoffs.

And so that began to make me think there was something real here. So now I should say what fitness payoffs are. So in evolution, you can think of evolution by natural selection much like a video game. So in a video game, your focus is to collect points as quickly as you can without being distracted by other things. And if you get enough points in a short enough time, you then might get to go to the next level. Otherwise, you die. And in evolution by natural selection, it's very, very similar.

Instead of the game points, you have fitness payoffs. And you go around collecting them as quickly as you can. And if you get enough, you don't go to the next generation, but your genes get passed to the next generation. And so to be a little bit more specific, think about the fitness payoff that, say, a T-bone steak might offer. So that if you're a hungry lion looking to eat, that T-bone steak offers lots of fitness payoffs. But if you're that same lion and you're full and you're looking to mate, all of a sudden that T-bone steak offers you no fitness payoffs whatsoever. And if you're a cow in any state and for any

activity, that T-bone steak is not a fit thing for you whatsoever. And so that gives you an intuition about what we mean by fitness payoffs in evolutionary theory. Fitness payoffs do depend on the state of the world, whatever the objective reality might be. They do depend on the state of that world, but also, and importantly, on the organism, its state, and the action. And so fitness payoff functions are very complicated functions, and the state of the world is only one of the parts of the domain of that function. There's lots of other aspects to it. And so they're really, really complicated functions of the state of the world and the organism, its state, and its action. Right. Well, so now I think you should introduce the desktop analogy, because again, what you just said can sound suspiciously similar to more or less what every life scientist and certainly neuroscientist would agree is true, which is whatever reality is, we see some simulacrum of it that is broadcast to

us by the way our nervous system sections up the world. So we see within a certain bandwidth of light, bees detect another bandwidth, and we, by the very nature of this, don't get all the information that's available to be gotten. So we don't have a complete picture of the thing in itself or the reality that's available to be gotten, so we don't have a complete picture of the thing in itself or the reality that's behind appearances. But implicit in that kind of status quo assumption is that the things we do see really exist out there in the real world in some basic sense in space and time. Again, it's not clear how much gets lost in translation, but there is some conformity between what we see as a glass of

water on the table and a real object in the world in third-person space. How is your vision of things departing from what is now scientific common sense? Yeah, it does depart dramatically from that standard view. The standard view, as you said, is that we may not see all of the truth, but we do see some aspects of reality accurately. And what the evolutionary simulations and then later theorems that my colleague Chetan Prakash proved indicate is that our perceptions were shaped by natural selection not to show us just the little bits of truth we need to see, but rather to hide truth altogether and to give us instead a user

interface. So a metaphor I like to use is if you're writing a book and the icon for the book is blue and rectangular in the middle of your screen, does that mean that the book itself in your computer is blue and rectangular in the middle of the computer? Well, of course not. Anybody who thought that really misunderstands the point of the user interface. It's not there to show you the truth, which in this metaphor would be the circuits and software and voltages in the computer. The interface is there explicitly to hide the truth. If you had to toggle voltages to write a book, you'd never get done. And if you had to toggle voltages to send an email, people would never hear from you. So the point of a user interface is to completely hide the reality and to give you very, very simplified user interface to let you

control the reality as much as you need to control it while being utterly ignorant about the nature of that reality. And that's what the simulations that I've done with my students and the theorems that I've done with Chetan Prakash indicate, is that natural selection will favor organisms that see none of the truth and just have this simplified user interface. So to be very explicit, three-dimensional space, as we perceive it, is just a three-dimensional desktop. It's not an objective reality independent of us. It's just a data structure that our sensory systems use to represent fitness payoffs and how to get them. And three-dimensional objects like tables and chairs, even the moon, are just three-dimensional icons in that interface. So once again, they're not

our species representations of a true glass that's really out there or a true table that's out there? They are merely data structures that we're using to represent fitness payoffs and how to get them. and also of how evolution gives us this false picture of what the deeper reality actually is. I have a few questions here. I'm going to start. I'm not quite sure where it will go, but there are at least three things that have been brought up so far that I feel like it's important for us to get clear on terminology and framework before I start really disagreeing. And I should say that you and I have now spent many meetings together. I spend a lot of time challenging you, mostly because I actually think there's something very interesting that you're doing, and I think you're onto something. And so in the same way that in my editing work, I give the most notes to the books I'm most passionate about. It's in that spirit.

So beginning with evolution, I've actually said to you many times that I don't actually think you need the evolution argument to make your case for your theory. So some of this pushback is actually moot, but I still think it's interesting. And I think I agree with this evolution argument up to a point. So my first question is really to just get us on the same page or see if we are on the same page as a a point. So my first question is really to just get us on the same page or see if we are on the same page as a starting point. I know that you believe that, or you're hopeful, you're optimistic about the fact that we can ultimately understand what that deeper reality is. And so there must be boundaries to the systems that we're using our brains which have

evolved where we can actually get access to the truth so so up until to a point our brains are giving us all of this false information but there's some sense in which we can actually get access to things that are true about the nature of reality. So my question is, where do you draw the boundary of an evolved system that by definition gives us false information about the nature of reality so that outside that boundary is where we might be able to gain access to information that delivers us the truth? And there's kind of a second part to that, which is where we might disagree. I believe we've already begun to cross that boundary with science. And so the way I follow your evolution argument is simply about direct perceptual information that we get rather than ideas, scientific experiments.

So if you just take light, light I think is always the simplest example. We have not evolved perceptual systems to really understand what light is, right? Like everything we've learned about light through the sciences up to quantum mechanics where it gets completely mysterious and we really don't actually know what light is. So we can kind of all agree, and not just the three of us in this room, but all of us, you know, most scientists would agree that ultimately, we still don't have this information about what the fundamental nature of reality is. We're still stuck there. But I would say that we have learned, we've gotten much closer to that

by these processes that I think are outside the boundary of this evolved system that is, by definition, delivering us false information. Right. Great question. And there's a couple points about it. First, the arguments that I've given from evolution by natural selection against veridical perceptions do not hold against math and logic. So that's very, very different than some other Christian apologists like Alvin Plantinga, who have made an argument that sounds very similar to mine, that they say that if our senses, if our cognitive capacities evolved, they would be unreliable. That includes our theory-building capacity, and therefore the theory of evolution is unreliable, and therefore evolution is false. I'm making no such argument. It's the furthest thing from my

mind. I'm focused only on the senses. And the reason why the argument that says our senses are not veridical doesn't hold for math and logic is that there are evolutionary pressures for us to reason about fitness payoffs. Two bites of an apple give you roughly twice the fitness payoff of one bite of an apple. Whatever objective reality might be, we need to be able to reason about fitness payoffs. And so, whereas the selection pressures are uniformly against vertical perceptions, they're not uniformly against some elementary competence in math and logic. Now, I'm not, of course, arguing that natural selection is shaping us to be geniuses at math and logic. Far from it. It's just that the selection pressures are not uniformly against ability. And every once in a while you get a genius coming up. But don't we think the math and logic are giving us space-time? I mean,

this can get into a deeper question because, of course, we now have quantum mechanics, which is putting all of this into question. And many physicists, if not most, are talking about space-time being something that emerges out of something more fundamental. But they would still say that it emerges. And so it seems that it's hard to take. So I guess my argument with where you take this evolution argument is as far as space-time itself. Because it seems that we don't yet know whether space-time is a true illusion in some sense. But I would say our math and logic has taken us that far, not simply our perceptual systems. Actually, let me see if I can add to this

point, because this is something that came up for me as well. So if we confine this to perception, for me, it's no longer counterintuitive. But again, this will be counterintuitive for many, many people. So the claim is that fitness trumps truth so fully that apprehending the truth perceptually is just not an evolutionarily stable strategy. You're going to be driven to extinction among creatures that are optimized for fitness. And that sounds a little crazy, but when you think of what fitness means, fitness means simply being optimized for survival and procreation, right? So as long as you're optimizing for that, it's easy to see that you successfully out-compete

anything that isn't optimized for that. And there's also this additional piece, which you mentioned, which is there's clearly fitness value, i.e. survival value in throwing away information that isn't related to fitness, right? So that every organism is going to have some bandwidth limits and metabolic limits and tracking every fact that's out there to track can't be a priority. And then there's this additional component, which is if the inability to make certain distinctions doesn't relate to increased fitness, evolution would not have selected for that ability to make those distinctions, right? So you'll expect organisms to be blind to certain features of reality just in principle.

But there is a sense in which your thesis does bite its own tail and seems to at least potentially subvert itself in that the moment you start to say that, okay, space and time, they don't exist, they're data structures, therefore our notion of objects is a pure interface issue. It's like a trash can on the desktop. It doesn't really map onto reality as it is. You just bracketed logic and rationality, which may be defensible, but evolution itself, the very notion of natural selection, is more than just rationality. It is a causal picture, and we might say that causes and the notion of cause and effect, right, or the notion that causes precede their effects rather than some notion of teleology, these things are also just data structures. So that every piece you want to put on the board to give a Darwinian account of anything

does sort of fall in the bin of more space and time, more objects. And so how doesn't this thing completely subvert itself and land you in something like just a global skepticism, which says, you know, we're in touch with some seeming reality, which we really can't ever know anything fundamental about. Yeah, great question, both of you. So the idea first that evolution by natural selection, So the idea first that evolution by natural selection, as we all know and love it, involves things like DNA and organisms and space and time and so forth. So how could I ever use the theory of evolution to show and claim to show that things like DNA are just data structures? They're just interface symbols.

And the reason I can do that is because John Maynard Smith actually took the theory of evolution by natural selection and mathematized it. He realized that we could abstract away from all of the sort of the extraneous empirical assumptions of space and time and DNA and so forth, and we could look at what he calls just evolutionary game theory, and so that the logic of natural selection itself can be reduced to competing strategy, where you make no ontological assumptions whatsoever about the world in which those strategies are playing. So it allows one, when someone says natural selection favors true perceptions, evolutionary game theory provides you precisely the tool you need to ask how to assess that

question independent of all these other empirical assumptions that are standard in biological evolutionary theories. And so that allowed me to do this. Now, there's another aspect to the argumentative strategy that I'm taking here, and that is that one reason that I went after the evolutionary argument was I actually announced the interface theory in my book in 1998, Visual Intelligence. And people liked the book except for the chapter on the interface theory, and they thought that was nuts. And I realized I wasn't going to get my colleagues to pay attention to that idea unless I talked to them in a language that they really understood. And it was that that motivated me to go after the evolutionary argument

a few years later. So the reason I use evolution is not because maybe it's the best argument. It's because it's the argument that I knew my colleagues would listen to. So first, I'm abstracting away from the whole apparatus of biological evolution to just the nuts and bolts

of evolutionary game theory, which doesn't bring the ontological assumptions. And second, my attitude as a scientist toward any scientific theory is, they're just the best tools we have so far. I don't believe any scientific theories, including my own. I think belief is not a helpful attitude. This is the best tool we have so far. Let's look at what this tool says about the claim that natural selection favors vertical perceptions. And whatever deeper, so what that tool is saying to me is there's just no grounds for thinking that any of our perceptions of space and time and objects in any way capture the structure of whatever objective reality might be. And one thing that's nice about this mathematics as well is you might say, well, how in the world could you possibly show that the structure of our perceptions doesn't capture the

structure of the world unless you knew already what the structure of the world is? I mean, aren't you shooting yourself in the foot there? And it turns out you don't have to. It's really that wonderful in the mathematics that you can show that whatever the structure of the world might be, the probability is zero, that that's what we're seeing. Right. And that makes sense to me too. I'm still stuck on how it extends all the way to space and time. And I think we shouldn't spend too much time on the evolution piece, mostly because I actually think you don't need it. But just from a philosophical perspective, I think it's very interesting. And I'm still curious myself kind of how far this goes, because it's clearly true, up to a point at least. So if Darwinian evolution

by natural selection is a theory about objects and space and time. I mean, this is just a question for you about how you view this. Where can you stand outside of space, time, and matter to talk about evolved perceptual systems? But more specifically, what does evolution look like, or how do you even talk about evolution outside of space-time? So what are we saying is evolving? What are we saying is surviving? What do evolution and survival even mean in a context outside of space and time? Or is that just an abstract idea that you haven't... No, that's the right question. And that's the power of evolutionary game theory. What John

Maynard Smith was able to do was to show we could talk about abstract strategies competing. Not in any particular assumption about space and time. He was able to abstract away from all the details of biological evolution in space and time and organisms and say the essence of Darwin's idea are these abstract strategies, and we can look at how these strategies compete in an abstract space. So what is it that's surviving? It's an idea? It's a meme? What survives? So what you do is you imagine that there's a population of entities that are competing using these strategies. So they're

abstract entities in an abstract space with these strategies. And what you do is you just, there's something called the replicator equation. And what you find in the replicator equation is that the number of entities that have a good fitness strategy will start to increase. Their proportion will increase. The strategies that have a bad fitness strategy will start to increase. Their proportion will increase. The strategies that have a bad fitness strategy or a lesser strategy. And so what you have is the proportion of the population that has various strategies goes up and down. Well, then I guess my question goes back to what do you mean by entity? So these are just abstract entities that in evolutionary game theory,

you don't need to know what the entities are. They're just place markers. You're imagining they're entities outside of space and time. And that's what the mathematics allows you to do. Well, let me just piggyback on this. You're getting tag-teamed. Oh, that's what I was hoping for. I apologize in advance. I apologize in advance. But isn't the very notion of competition and differential success parasitic on the notion of time, parasitic on the notion of causes preceding their effects?

And entities is, you know, I think what Annika's fishing for there is entities seem somehow derivative of objects, at least the concept of an object. I mean, we're talking about something that's discrete, that's not merely a continuous reality, right? Things can be differentiated. So how are we not using the same cognitive tools that have got hammered into us by evolution, whose process has only selected for fitness and therefore left us epistemologically closed to the nature of reality. Absolutely. So you're right that the evolutionary, the replicator equation itself does have a time parameter, right? Or at least a sequence parameter. It depends on whether you do it discreetly or continuously. And so that's going to be built into it. Absolutely. So by the way, as I said, I'm not committed to the truth of evolution by natural selection.

Absolutely. So, by the way, as I said, I'm not committed to the truth of evolution by natural selection. I'm just using that theory itself to say that whatever the structure of the world is, that theory says the chance is zero reality that will give back evolution by natural selection as a special case within what I call our space-time interface? And that's actually what I'm hoping for is to have a deeper theory that will go beyond space and time. It'll go beyond time in the sense that there will be sequence and there will be perhaps a notion of cause following effect, but not in a global space-time temporal framework. It'll be completely asynchronous and so forth, and we'll get what we call causality in like a Minkowski space, Einstein's Minkowski space, or a general relativistic curved space-time, as a projection of a much more deep theory of reality in which the very notion of dimension doesn't hold,

in which time doesn't hold, but we can show that though, that so far I'm thinking about dynamics on abstract graphs and asynchronous dynamics, but that can be projected and simplified into what we call space-time and its causality, say in Minkowski's space. Mm-hmm. I think it's just useful as a launching off point to every place we'll go from here to just say that at the very least, I think this evolution argument is very useful in terms of opening our eyes to something that I actually think we, in some sense, we already know. And again, you know, looking at something like light is a good example where we clearly, we have not been given any tools, perceptual tools, to understand how electrons operate, what is

actually happening at a fundamental level. And of course, there are all these theories now from everything in string theory to many worlds trying to sort out all of these things that we see through our science that we have absolutely no intuitions for, we have no insight into, we're just getting at through math and logic. And so clearly we haven't evolved systems that help us here. And so I feel like we can agree to two points that we can move from here onward. And the first one is that we can all agree, and scientists in general, we don't know what's fundamental, nor do we perceive the truth about the fundamental building blocks of reality. And two, and this is where I'd like to set this up

for where consciousness is going to, it was about to come in, we can agree that physical science has not given us an explanation for consciousness. We have no understanding of how consciousness arises out of physical processes. And so it seems that we can at least agree that it's a legitimate question, or it's a legitimate project to wonder if consciousness is something that's more fundamental and that we're missing that piece and that we've thought about it backwards all this time. And that's one of the things that I think is so great about your work. And I think this is a very important project. Okay. So before we get to consciousness, which is central to our interest and where there's more controversy, at least in my mind, I want to anchor what you've said to a very straightforward perception so that our listeners

can get in touch with how counterintuitive your thesis is. So when the three of us are in a room together, apparently there are objects we can see, what is the status of those objects like a glass of water when none of us are looking at it? And what is its status, given the fact that it apparently is always there for any one of us to look at? We have some kind of consensus, intersubjective language game we can play here that can reference the glass of water, you know, at will, how does that map onto your theory of non-veridical perception? Right. So I think a good way to see what I'm saying and how counterintuitive it is, is to think about, say, playing a game like Grand Theft Auto, but with a virtual reality add-on. So you have a headset and you're seeing a three-dimensional

world of cars and your own steering wheel and so forth. And you have a headset and you're seeing a three-dimensional world of cars and your own steering wheel and so forth. And it's a multiplayer game, so there are people around the world that see the same car that you're driving and see all the other cars that you see. And in that case, there of course is no real car that anybody's seeing. There's just some, in this metaphor, a bunch of circuits and software and so forth. That's the objective reality in this metaphor. But all the players will agree that they see a red Corvette chasing a green Mustang down the highway at 70 miles an hour. They all agree, not because there's literally a red Corvette chasing a green Mustang. There is some objective

reality, but it's not Corvettes and Mustangs. That's what we each see. And each person with their own headset is getting, in this example, photons thrown to their eyes, and they're rendering in their own mind the Corvette chasing the Mustang. So there are as many Corvettes and Mustangs as there are people playing the game because they each see the one that they render. And I might be looking at the Corvette and I look away and I'm now looking at my steering wheel. I no longer see the Corvette. I have garbage collected the Corvette. I'm not making that data structure anymore. Now I'm rendering a steering wheel. And now I look back over at the Corvette, now I'm re-rendering the Corvette. So it looks like the Corvette was always there because when I look away and look back, it's right where I expect it to be. But in fact, there is a reality. It's not Corvettes, it's not Mustangs, it's not steering wheels. So here's the counterintuitive claim. I'm claiming

we all have a headset on, all of us. And we all have this space, time, physical objects, the glass of water. Those are all things that I render on the fly when I look at them, and then I garbage collect them. And that's part of the evolutionary argument. I garbage collect them because I'm trying to save energy and time and memory. So I render it only as I need it, And it's really just the glass I'm seeing is a representation of fitness payoffs. Those are the fitness payoffs I need to pay attention to now. Now I'm throwing that fitness payoff description away. Now I'm looking at fitness payoffs over here. So it's a rapid rendering of fitness payoffs in real time.

So here's one of the areas where I worry that the language that you're using, the terminology you're using, may actually give a false impression of what you're saying. This is where some of my notes came in. I don't know how many of these notes you have taken or will take, but I worry that, I actually think I agree with you there, but there's something about the way you're saying it that I think gives a false impression of what you're saying. something about the way you're saying it that I think gives a false impression of what you're saying. So if you say, you know, the race car isn't there, you know, the moon is an example you give often. I mean, you also will say, which I think is more accurate and closer to what you're saying, is something exists. Something is there in reality that my perceptual systems are kind of turning into this sight of a moon. And I think it's confusing to readers and listeners when you say it doesn't exist, as if the fundamental nature of reality behind whatever that moon is doesn't exist, that there's nothing there. So it seems more accurate to say we simply

don't understand the deeper reality behind the moon and behind apples. And that this is something in a way like it's less controversial. This is something we can all admit, given our current understanding of the physics. And so part of my gripe there, I think, is just with the language that you're using. And there's something incredibly interesting about that, that something is there. There's something I'm interacting with. The example I often like to use with you when we meet is a tree. If we plant a tree and leave it, it is out of our conscious experience. There are all these processes that will be taking place in what we call them, how we view them as water and nutrients being sucked up from the earth, and it will grow and will come

back in a year. And all of those processes would have taken place whatever they are at bottom we we may not understand but something is going on in the universe that we have our access to however far from the truth it is there's something taking place there and so to explain it as when I leave, there's absolutely nothing there and there's no tree. And then I come back and somehow I create this as if it's... Yeah, I think that's a very important clarification. So I agree with you completely that I'm not saying that there isn't an objective reality that would exist even if I don't look at it. There is an objective reality. It's just that what I see is utterly unlike that objective reality. And in the metaphor that I was giving of virtual reality, I might see a red Corvette.

The reality in that metaphor would be circuits and software that aren't red, that don't have the shape of a Corvette. They're utterly unlike a Corvette. But when I interact with that objective reality that's there, even if I don't see the corvette i then will see the corvette so that's how different i think it's potentially confusing as an analogy only because as a user of video games you can turn you can turn the video game off it's not a self-sufficient world it's not reality that that continues on and does it does its thing i I agree with you on that. Yeah, it gives a slightly false impression. Right. I agree that the reality is continuing on regardless of what... I have life insurance.

Right, right. And the reason I have life insurance is because I agree with you that there is some reality that will continue to go on even if I'm not here. Right, right. Okay, so let me make that point with a slightly different topspin, because those concessions seem to bring us back to the standard consensus view of science in some way. So there's this appearance-reality distinction. There's our sensory experience, which is our interface, which everyone agrees does not put us in direct contact with the thing in itself or underlying reality. But you're conceding that there is an underlying reality and there must be some lawful mapping between what we see on the interface and that underlying reality which

actually renders our mutual perceptions of things like trees and glasses and cars predictable, where we can both agree that if we go to look for the same object, each one of us is likely to independently find it, whatever the relationship is between that interface data structure and reality itself. So there has to be some kind of isomorphism between our virtual reality experience and reality itself, even though we don't have, by virtue of evolution, all of the right conceptual tools so as to say what it is. There is going to be a mapping between objective reality and our perceptions, and that mapping will be as complicated or more complicated as the mapping

between all the circuits and software in the virtual reality machine and the actual like Grand Theft Auto world that I perceive. And if you think about it, there's going to be hundreds of megabytes of software, all these complicated circuits. All I'm seeing is simple cars and so

forth. So there's going to be, in computer science, there are all these virtual machines that you create, many, many levels of virtual machines between what you see in the Grand Theft Auto game and the actual objective reality in this metaphor that's going on there. And so I'm saying that the idea that the reality is going to be isomorphic to space-time is too simplistic. I agree that there's going to be some systematic mapping that is going to be quite complicated. Another way to put it is this. If I said to you, I want you to use the language of what you can see in your interface in the virtual reality, so the pixels that you can see, the colors and pixels, that's the only language you can use. I want you to tell me how this virtual world works.

You can't do it because the language of pixels is an inadequate set of predicates to actually describe that world. And I'm making the very strong claim that whatever objective reality is, the language of space and time and physical objects in space and time is simply the wrong The language of space and time and physical objects in space and time is simply the wrong language. There is a systematic mapping, but the language of objects in space and time could not possibly an aphorism that almost contains this thesis in seed form, which is, not only is reality stranger than we suppose, it's stranger than we can suppose. By giving a deflationary account of our notion of space and time, you are saying, whatever this mapping is between appearance and reality, we are so ill-equipped to talk about it based on this interface analogy that it is, on some level, far stranger and far more foreign to the way in which we're thinking about things than anyone has. Absolutely. Yeah, so your claim isn't actually, so I'm just trying to get at what is truly novel about your claim.

One thing that's novel is the expectation that evolution has selected for some approximation to what is true seems false. Right. So fitness trumps truth. Trump's truth. And as a result, whatever this mapping is to underlying reality, we are in a far greater state of ignorance about it than most people expect. That's right. Absolutely, you've nailed it on the head. And I would say this, that it's the relationship between a visualization tool and whatever it is that we're visualizing, right? So there's going to be this objective reality that's out there. And evolution just gave us this very, very dumbed down species-specific visualization tool. The very language of that tool is probably, I mean,

the whole point of a visualization tool is to hide the complexity of the objective reality and just give you, you know, a dumbed down tool that you can use. And so the very language of space and time and objects is just the wrong language for whatever the thing is. Just like... I would say though that, you know, as far as I understand, most up to this point, I know we're going to talk about consciousness soon and then we'll get into a different realm, but up to this point, everything that you've just said, clearly not being the final answer to what is fundamental. And everything we see out of quantum mechanics gives us a real philosophical problem similar to the one you're describing, which is, it seems that the fundamental nature of the universe, what the universe is actually made of, is not anything like what we experience it all the way to the point of space and time. That's right. And so it's really interesting because if you look at our biggest scientific theories in physics, general relativity and also special relativity, are about space-time.

Space-time is assumed to be an objective reality and a fundamental one. In quantum field theory as well, the fields are defined over space-time. And so physics, as Nima Arkani-Hamed has put it, and he's a professor at the Institute for Advanced Study at Princeton, he's pointed out that for the last few centuries, physics has been about what happens in spacetime. But now they're realizing that to get general relativity and the standard model of physics to play well together, they're going to have to let go of space-time. It cannot be fundamental. And he's not worried about it. And in fact, he says most of his colleagues agree that space-time is doomed, and there's going to be something deeper. And that's wonderful, because we're about to learn something new. There's a deeper framework for us to be thinking about physics, and space-time will have to be emergent from that deeper framework. I actually watched a lecture of his recently, and I wrote down this short quote.

He says, all these things are converging on some completely new formulation of standard physics, where space-time and quantum mechanics are not our inputs, but our outputs. And I thought that was very well said. So as far as I understand where physics is at at this point, I think all of these physicists would agree with you up until this point. And I think now we can probably cross over. Although I would just point out that they might agree for different reasons, right? They're not using the same evolutionary logic. But that there's nothing intrinsic in what Don is saying about

how false our view of the fundamental nature of reality is. That it is that. That you can actually take it all the way to space-time and that we're probably wrong in all of those assumptions about what we think. I agree, and I think it's really interesting

that the pillars of science are all saying the same thing. Evolution by natural selection is saying you need to let go of space-time. And then the physicists trying to get general relativity and quantum field theory to play, they're saying you have to let go of space-time. When our best science is saying that, it's time for an interesting revolution. That's going to be fun. I mean, it's going to be very exciting to see what happens when we go behind space-time. It's so counterintuitive though, right? We've just assumed that, I mean, our story is space-time came into existence 13.8 billion years ago at the Big Bang. It was the fundamental reality. We're saying there's a deeper story. That story is only true up to a point there's a much much

deeper story and that's more like an interface story that's the projection of a much deeper story we're gonna have to find and that is tremendously fun yeah well so now we're now going to move on to consciousness which will be interesting i just guess I want to flag my lingering concern that your rationale, if taken in deadly earnest, may still kick open the door to epistemological skepticism, for me at least. Because I think if one's space and time are dispensed with causality and kind of an evolutionary rationale does, I mean, this is kind of the Plantinga argument you referred to. It's just once you start pulling hard at those threads, I'm not sure how much the fabric of epistemology can be defended.

I agree with you, Sam, in the following sense. I think that it might actually go that way just on the evolutionary arguments alone. So what I'm going to want to do is to, whatever the deeper theory of reality that I propose, it needs to be such that it will not fall into the epistemological problems that you're raising. So the deeper theory needs to avoid those epistemological problems and show why that deeper theory looks like evolution by natural selection when we project it into our space-time interface. In other words, so that these kinds of problems might arise because evolution by natural selection itself is not the deepest theory. It's just an interface version of a deeper theory. Right.

Yeah, so on this topic of causality and time and whether this project even makes sense, which I know is a place you and I have gotten to before in our conversations. When you say things like the brain and neurons are not the source of causal powers and that we need to find another source, my question is why would you assume that there are causal powers at all in the fundamental nature of reality. So it's not clear to me why we include causal powers as part of a fundamental reality if space-time doesn't exist. I don't quite see how there is causality without time, at least in the way that we typically think about it. I mean, just to take an example, which is kind of standard physics, although often neglected, the notion of a block universe, right? The notion that, you know, the future exists just as much as the present as as the past. And so that there really are no events. There's just a single datum, which is the entire cosmos, right?

And its connections. So causality under that construal is really an illusion. That's right. And without endorsing the block universe view, I would say that causality in space and time is a fiction. It's a useful fiction that we've evolved in our interface. we've evolved in our interface. But that strictly speaking, causality in space and time is not, because space and time is not the fundamental reality, the appearance of causality, like my hand pushing this glass and moving it, it gives the appearance that my hand has causal powers and is causing the glass to move. But in fact, that's just a useful fiction. It's like if I drag

an icon on my desktop to the trash can and delete the file. It looks like the movement of the icon on the desktop to the trash can caused the file to be deleted. Well, for the casual user, that's a perfectly harmless fiction to believe. If you move the icon to the trash can, it causes the file to be deleted. It's perfectly harmless. But for the user, for the guy who actually wants to build the software interface for this, go under the hood, that fiction has to be let go. So I'm claiming that within space and time, all the appearance of causality is a fiction. Now, in terms of a deeper theory, because you're asking a deeper theory, what about causality? Well, my argument is that causality is part of the illusion of time.

Assuming time is some sort of illusion and time is not fundamental, at least as far as we usually talk about. I mean, I can think of, this is another conversation of how we can almost redefine causality, which in my view I have. I think there's a way to talk about different things being connected. view I have I think there's a way to talk about different things being connected but in terms of the way we our definition of causality and how we use it it is dependent on time it is a part of things that play out in time you need something you need something to happen in the past to cause something to happen in the future it's this it is this direct relationship in time. And so I don't even know how you would talk about causality without time. It needs time for its own definition.

So I think if we're redefining causality, which I think is kosher, actually, I think that's something we can talk about. I've never been clear whether that is what you mean. Are we kind of redefining what causality is? what you mean? Are we kind of redefining what causality is? And is it more like connections between things rather than one thing happens and then another thing happens in response? Yeah. I would also add another aspect here, which is the notion of possibility may be spurious, right? So that it may in fact be that nothing is ever possible. There's only what is actual, right? There's only what happens. And our sense that something else

might have happened in any circumstance, that just might be a, again, part of this user interface that has seemed useful because it is useful to try, like when we're apparently making decisions between two possibilities and we need to model counterfactuals, right? Counterfactual thinking is incredibly useful. And yet what if it is simply the case as it, you know,

as it would be in a block universe that there's just, you know, the novel is already written and you're on page 75, but page 168 exists already in some sense. I don't think you need the block universe, though, because I think there's... No, that's just one way of getting at the point. Yeah, I mean, it's a good visualization, but I think most physicists will have some argument about it being described that way. But I think the analogy holds. And I was just reading Carlo Rovelli's book on time, and he makes this point as well, that at a certain level, there is no difference between past and future. And essentially, I mean, his thesis in the book is's outside of spacetime, that is not going to be dependent on time. It'll be more like relationship, as you talked about. about probability and how we interpret probabilities in scientific theories.

So there are probabilities that are epistemic in the sense that maybe there's a deterministic reality out there, and I just don't know enough about it. So the probabilities are subjective. It's my lack of knowledge. There's frequency, but our sense of probability may be spurious. but our sense of probability may be spurious. That's right. But then if there are probabilities in which no matter how much my knowledge increases, the probability will not disappear.

And so we often call those in science objective chance. And I think we'll want to have a conversation about how we think about probabilities and objective chance. It will actually take us into the question about free will and so forth, my version of notions of free will versus determinism. So I think that that's going to be an interesting conversation. So I agree that we need a notion of causality that transcends time, and I'm proposing one. By the way, it's interesting, I know you talk with Judea Pearl,

and he's got, of course, these directed acyclic graphs, models of causal reasoning, which are brilliant, and they've actually given us a mathematical science for the first time of causal reasoning. But in his book, Pearl doesn't define causality. He refuses to define the notion of causality. In some sense, what we're facing here is that every scientific theory, and this is a really important idea, I think. No scientific theory is a theory of everything. There's no such thing. Every scientific theory makes certain assumptions.

We call them the premises or the assumptions of the theory. And only if you grant the theory those assumptions can it go and explain everything else. And so we're going to have in every scientific theory certain primitives that are unexplained. They are the miracles vis-a-vis that theory. Now you may say, well, I can get you a deeper theory for which those assumptions come out as consequences, but you will have a deeper set of assumptions. There's going to be an axiom somewhere at the bottom. Absolutely.

And that's a humbling recognition for a scientist to realize that we will never have a theory of everything. We will always have a miracle or a few miracles. We want to keep them as few as possible i don't like that you call them miracles i would like like to have the record show i understand that but i we call them assumptions why not call them axioms yeah well because i want to really make it's another place where i think people might actually be confused about what you mean which is why i'm sure i'm glad that you're pushing trying to protect you. So I'll just say that there are things that the theory cannot explain.

And there will always be things that every scientific theory cannot explain, and it's a principled problem. So the interesting thing will be, in a deeper theory, will we have something that's like a causal notion that will be a primitive of the theory? And it may not be dependent on time, but there will be primitives and explanation will stop. I guess, so my question, my issue really is why use the word causality when you're speaking in more fundamental terms? So why not say something like connections, relationships to me seem much

more, much closer analogies. And so to say what we view as causality is in fact something more like a connection or a relationship. I'm completely on board. I agree with you completely. I think a deeper theory, we may think that the term causality is just not a very useful term anymore. It was useful in space and time and connection or influence is a better term at a deeper level. Okay, so on to consciousness and free will and other dangerous topics. Yes. What, in your view, is the connection between consciousness and the base layers?

If you'd like to continue listening to this podcast, you'll need to subscribe at SamHarris.org. You'll get access to all full-length episodes of the Making Sense Podcast and to other subscriber-only content, including bonus episodes and AMAs and the conversations I've been having on the Waking Up app. The Making Sense Podcast is ad-free and relies entirely on listener support. And you can subscribe now at SamHarris.org.