Truth vs Reality: How we evolved to survive, not to see what’s really there | Donald Hoffman

Galileo was quite controversial, of course, in his time, because he was claiming that something that we all could see with our own eyes wasn't true. We could all see that the earth doesn't move and that the sun, moon, and stars go around the earth. And we believed that as a race for about 2,000 years. Galileo was saying that your eyes are lying to you. The earth actually moves and it's not the center of the universe. He was put under house arrest for it. We don't like to be told that our senses aren't telling us the truth.

Then Galileo took it another step. He said it's not just that our senses are lying about the movement of the earth; he thought that tastes, odors, colors, and so on reside in consciousness. Hence, if the living creature were removed, all these properties, these qualities, would be utterly annihilated. That's almost a direct quote in the translation. He was saying that our senses are also making up the tastes, odors, and colors that we experience. They're not properties of an objective reality. They're actually properties of our senses that they're fabricating.

By objective reality in this case, I'm going to use that term in a very specific way. I mean that something is objectively real if it would continue to exist even if there were no creatures to perceive it. For example, the standard story is that the moon existed before there was any life on Earth and, perhaps, before there was any life in the universe. But it still existed. Its existence does not depend on the perceptions of any creatures. That's the sense in which I'll talk about objective reality. What Galileo was saying was that colors, odors, tastes, and so on are not real in that sense of objective reality. They are real in a different sense. They are real experiences.

Your headache is a real experience, even though it could not exist without you perceiving it. It exists in a different way than the objective reality that physicists talk about. Galileo was quite brave and quite out of the box in his thinking by saying not only the earth in his movement, but even colors, tastes, and odors are our perceptual constructions. But he wouldn't go the next step. He wouldn't say that shapes, mass, velocities of objects, space, and time themselves are our constructions. He thought that those were objectively real. The shape of the moon, the position of the moon, is an objectively real thing, including its mass and its velocities.

This is a distinction that was later called the primary and secondary qualities distinction by John Locke. Primary qualities are things like position, mass, shape, and so forth. These are presumed to exist even if no creature observes them. Whereas colors, odors, and tastes are secondary qualities that are sort of mostly the contribution of our senses.

In brief, what I'm saying is we need to take the next step beyond what Galileo said. It's not just tastes, odors, and colors that are the fabrications of our senses and are not objectively real. It's, rather, that space-time itself and everything within space-time—objects, the sun, the moon, the electrons, quarks, their shapes, if objects have shapes, their masses, their velocities— all these physical properties are also our constructions. I've come to that conclusion. It was a bit of a shock to me. We always assume that our senses are telling us the truth.

It was quite a stunning shock to me when I realized that maybe we needed to take a step beyond Galileo on this. The reason I'm saying this is because of evolution by natural selection. Most of my colleagues in cognitive and neurosciences assume that our senses tell us the truths that we need to survive. That seeing reality accurately will make you more fit. That makes perfect sense. The argument is that those of our ancestors who saw reality more accurately had a competitive advantage over those who saw it less accurately in the basic activities of life, like feeding, fighting, fleeing, and mating.

Because they had a competitive advantage, they were more likely to pass on their genes which coded for the accurate perceptions. After thousands of generations of that, we can be quite confident that we see reality as it is. Of course, not all of reality. No one claims that our senses exhaustively tell us all the truths about objective reality. From an evolutionary point of view, the idea is we see those aspects of reality accurately that we need to survive.

So when we see space and time, we see physical objects with their shapes and motions, and so forth, we're seeing truths, objective truths. Truths about objects that would exist even if no creature were there to perceive them. That's the standard view. It seems intuitively plausible—the argument that I just gave is actually in the textbooks in my field. But it turns out that we don't have to just deal with plausibility here. Evolution by natural selection is a mathematically precise theory.

There is the field of evolutionary game theory that was established in the 1970s by John Maynard Smith, and it has flourished. It's now a very advanced and very interesting mathematically precise field. It unites Darwinian evolution by natural selection with the tools of game theory. It's very, very powerful. We don't have to guess or wave our hands anymore. We can actually run simulations and prove theorems about the effects of natural selection on our senses.

We can ask a technical question: Does natural selection favor organisms with sensory systems that tell them truths about reality, objective reality? It's a clean technical question. It turns out there is a clean technical answer that comes from evolution, and it is quite surprising. I first started this about 12 years ago with a couple of graduate students of mine—Justin Mark and Brian Marion. We ran hundreds of thousands of evolutionary game simulations in random worlds with resources and creatures that had to forage for these resources.

We played god. Some of the creatures got to see the truth. Others didn't. We had them just perceive the fitness payoffs. We can talk a little bit about fitness payoffs a bit later. That's a key notion in evolution. What we found was in the simulations that organisms that saw the truth never out-competed, never outperformed creatures in our simulations that saw none of the truth and were just perceiving the fitness payoffs. That gave me some confidence that maybe there was a theorem here.

So I proposed a theorem to a very talented mathematician named Chetan Prakash with whom I've worked for many years. Chetan and I discussed it and worked on it. Chetan brought it home. He proved the theorem. An organism that sees reality as it is is never more fit than an organism of equal complexity that sees none of reality and is just tuned to the fitness payoffs. Translated, that means if you see the truth, you'll go extinct.

So the question is, of course, what are our fitness payoffs? What's going on there? That's a technical term in game theory. The payoffs are what sort of drive the game. But I think an analogy can help. Think of life as like a video game. In many video games, you have to try to grab as many points as quickly as you can at the level that you're at. If you get enough points in minimal time, you might get to the next level. If you don't, you die. You have put in some more money or start over.

The idea is that life is like that. It's like a video game, but instead of points in a game, we have fitness payoffs. Getting the right kind of food, high-quality food, not eating poisonous things, breathing the right kind of air, finding the right mate, and so forth. These are all fitness payoffs that we can get. If we get more fitness payoffs than the competition—it's not like getting millions of fitness payoffs; you just have to do a little bit better than the competition—then you have a better chance of passing on your genes that code for your strategies for getting fitness payoffs.

You don't go to the next level like in a video game, but your genes and your offspring go to the next level. That's, informally, the idea of fitness payoffs. They're what drive success or failure in evolution and life. What we discovered was two things. First, that fitness payoffs themselves destroy information about the structure of the world. It's truly stunning. Fitness payoffs depend on the state of the world.

I can give you an example. What is the fitness payoff of, say—I like this example of a T-bone steak. For a hungry lion looking to eat, that T-bone steak offers a lot of fitness payoff. It will help it to stay alive and be strong. For that same lion that's well-fed and looking to mate, the T-bone steak offers no fitness payoffs. For a cow, in any state, a T-bone steak is not a good thing. The payoffs depend on objective reality, whatever that might be, whatever the state of the world might be.

Also on the organism, like lion versus cow, its state, hungry versus fed, and its action, eating versus mating, for example. Fitness payoffs, as you can see, are complicated functions. They depend on the state of the world, whatever the world might be, but also on the organism, its state, and its action. If we fix an organism, state, and action, then fitness payoffs are functions from the world, whatever the world might be, into a set of payoff values, say from 1 to M, fitness payoff values where 1 means you're dead, M means you're getting the most you could possibly get.

What we've discovered is that function, those fitness functions, almost surely destroy information about the structure of the world. I can give you a concrete example of what I mean by that. Suppose—by the way, when I said that, I don't need to know anything about the world. I don't need to propose I know anything about what reality is. These terms hold anyway. That's the nice thing about the mathematics. You might say, "Well, how could you prove such a theorem unless you know what the world is?" It turns out you can.

These theorems hold regardless of what the world is. Suppose we take, for sake of argument, a world in which there really is oxygen concentration. There is air, and there's oxygen. Oxygen concentration can go from 0 percent to 100 percent. That's what mathematicians would call a total order: 0 is less than 1, less than 2, all the way up to 100. That's a total order. It turns out—so that would be a structure in the objective world in this example.

Now the percentages of oxygen that will maintain human life are about 19.5 percent to 22.5 percent. If you get outside that range, you'll be in distress and eventually die. There's this very narrow range of oxygen concentrations useful for life. Suppose you had a creature that had only two colors that it perceived. A very simple creature. It just sees green and red.

Assume we're going to say green is greater than red; we'll just put an order on green and red. So green is greater than red. Suppose we look at two different creatures. One sees as much of the truth as possible with just two colors. In that case, you might use red for 0 percent of oxygen to 50 percent. So red is for very little oxygen to medium. Green would be from medium to 100 percent. That way if you saw red, you'd know there was less oxygen. If you saw green, you'd know there was more.

You're knowing as much about the truth of the objective reality—namely the amount of oxygen—as you could possibly know, given the limits of your sensations. That would be a truth organism. Now consider a fitness organism that again only has two colors, green and red. To encode fitness, you could do the following: Let's use red for 0 through 19.5, which will kill you. Green would be from 22.5 to 100, which will also kill you. In other words, we use red for those amounts of oxygen that will not sustain life. We'll use green for that narrow band from 19.5 to 22.5 that will sustain life.

If I see green, I know I'm good. I don't need to change anything. I'm going to live. If I see red, I know I'm in trouble; I need to do something differently. But notice if I see red, I have no idea about the truth, about how much oxygen there is. There could be 100 percent; there could be 0 percent. I have no idea.

That concrete example gives you an intuition about why seeing the truth is a very different thing than seeing fitness and why they are really at odds. Our intuitions are, of course, that if I see the truth, that will make me more fit. This example makes it very clear that seeing the truth is, in most cases, the opposite of seeing what's fit. We were actually able to prove this—this is now inside baseball language, but I'll throw it out there—the set of fitness payoffs, if the world has N states, N as in Nancy, and the fitness payoffs have M values, M as in Mark.

There are going to be M raised to the N power total fitness payoffs, very simple math, combinatorics. For any structure in the world that you want to consider—a total order, a symmetric group, a cyclic group, a measurable structure, a topology—you can ask in each case how many of those M to the N fitness payoffs will preserve that structure. Mathematicians call them homomorphisms.

The homomorphisms of a topology are what we call continuous functions. The homomorphisms of measurable structures are what we call measurable maps, and so forth. It is straightforward to show in each case that the probability—well, I wouldn't say straightforward. If you have a mathematician working with you, looking over their shoulder, it looks straightforward, but of course, it's hard work for the mathematicians. Some of the combinatorics is pretty straightforward for mathematicians.

In each case, we show that the ratio, the number of homomorphisms, the fitness payoffs that preserve the structure of the world that tell you something about the truth, to the total number of fitness payoffs, that ratio—the truth preserving payoffs versus all payoffs—we look at that ratio. That ratio goes to 0 as the number of states in the world increases and the number of fitness payoffs increases, and that means that the fitness payoffs generically destroy information about the structure of the world.

Our senses will be tuned to the fitness payoffs. Being tuned to the fitness payoffs means that you will not be tuned to the structure of the world because the fitness payoffs have lost that structure. That's how devastating this is. We're in a dilemma here. We have two things that we deeply believe. We deeply believe in evolution by natural selection. We deeply believe in physicalism—that space-time and physical objects as we perceive them are a true representation of reality as it is. Those two claims are in conflict.

Both cannot be true. I'm saying that space and time and physical objects don't exist unless they're perceived. Someone might say, "Well, look, Don, if you think that that train coming down the tracks is just some little thing that you're creating on the fly, you're making that up. It's just an icon on your desktop, or a symbol in your virtual reality. Why don't you step in front of that train? After you're dead, and your ideas with you, we'll know that that train was real, and it really can kill."

I wouldn't step in front of the train for the same reason that I wouldn't, for example, if I'm writing an email. The email icon is blue and rectangular and in the middle of the screen; that doesn't mean the email itself on the computer is blue and rectangular in the middle of the computer. I don't take the icon literally. It's not literally true about what's in reality. But I do take it seriously. I would not drag that icon to the trash can carelessly. If I dragged the icon to the trash can, I could lose all of my work.

I take my icons seriously, but not literally. That's the case with the train as well. Evolution by natural selection has shaped us with perceptions that are designed to keep us alive. If I see a snake, don't pick it up. If I see a cliff, don't jump off. If I see a train, don't step in front of it. We have to take our perceptions seriously, but that does not entitle us to take them literally.

Another objection is: I look over there and see a train, and I ask all of my friends—they'll also say that they see a train. Given that we all see the train, surely that means there is a train in objective reality. There really is a train. That seems very compelling. Of course, we all look; we see the moon. We all agree that there is a moon, so therefore the moon must really exist. But again that's a logical error.

In the visual example that we looked at earlier with the cubes that were floating in front of the disks, we all would agree that we saw a cube. But we would all agree that the reason you saw a cube was that you created the cube. The cube doesn't exist unless you create it. So the reason we agree about trains, and the moon, and cars, and apples, and so forth is that we have a similar interface. We’re constructing similar objects, and because we construct our worlds in similar ways, we tend to agree.

Although 4 percent of us have synesthesia and can view the world in very, very different ways from the rest of us. The bottom line is agreement only means agreement; it doesn't mean that we're seeing objective reality. Descartes famously said, "I think, therefore I am." That raises a really interesting question about consciousness and what we call the physical world. A physicalist would say, "Look, I know about things like umbrellas, and the moon, and rocks. I mean all of this concrete, stable stuff, I know about that reality.

But when you talk about consciousness and conscious experiences, I'm not sure that you're really talking about anything real. It might just be an illusion. It might be a figure of speech. We could be very, very deeply wrong." That’s the view of most of my hard-nosed physicalist colleagues. "I know about this physical world out there. That's really solid. It's something we can really look out and test. This stuff about consciousness is airy-fairy, wiggly. I don't know what you're talking about. It's too squishy for me."

There is a completely different point of view. It's to say, "Look, when I look over here, I'm having an experience, say, that I would call an apple. I know that I'm having an experience. When I close my eyes, my experience of the apple stops. Now I'm just experiencing a gray field. You want to tell me that, in fact, there's a red apple that still exists even when I'm perceiving a gray field? Well, that's actually more than I know.

All I know is that when I open my eyes, I'm experiencing a red apple. When I close my eyes, I'm not. It's an extra step. It's a big jump to say that experience of the red apple is actually true of a real red apple and that red apple still exists even when I'm stopping, when I don't have that experience. That's more than I know. I think it's far less problematic and less going out there on a limb to just say I have my experiences. I don't know what the objective reality is that's out there."

So physicalism is actually a stronger and more problematic claim. It's saying that there is a reality that in some way matches our experiences and continues. I'm just saying we have these experiences. The "cogito, ergo sum," that "I think, therefore I am," I think Descartes wasn't just talking about thinking in the normal sense of abstract cogitation or reasoning. I think he was talking about perceiving. In that sense, I would say yes. I'm having experiences. I'm perceiving.

My thoughts, my rationality, I'm experiencing that as well. That's the starting point. I wouldn't go all the way with Descartes. He says therefore I am. I don't know what the word "I" refers to there. Certainly, conscious experiences seem to exist from that. The "I" may be another construct, another symbol that I make. The symbol that I call Don, the I, may not be absolutely necessary for the experience.

I don't know if I'd go all the way with Descartes on that. But I would go part of the way and say, yes, saying that there is a world of experiences is going less out on a limb than saying there is a world of objective objects that resembles my world of experience, in addition to my experiences. In my own scientific investigations, I've proposed a very controversial theory. I've received lots of very pointed and sharp criticism, some in print, some published, some in person.

What I've found is I've learned a lot from each of the criticisms. In many cases, they forced me to think about an aspect of the theory that I had not thought about that way before. In some cases, they put me into days of doubt, where I was looking at that part of the theory, wondering about it, and then either revising the theory or realizing, "Oh, wow, the theory has resources that I didn't realize to deal with that problem."

That’s the power of a nice theory, a mathematically precise theory. When you write down the theory, the theory then becomes your teacher. It becomes smarter than you in a way. When Einstein wrote down the equations of general relativity, he did not know that they entailed the existence of black holes. In that sense, the equations were smarter than Einstein. Einstein didn't believe in black holes for decades. The equations were very clear that they could exist. Einstein said no. It turned out Einstein was wrong and the equations were right.

It's very interesting. We do these theories because we can learn from them. When you have criticisms, it forces you to examine parts of your theory very, very carefully and ask, "Is this correct? Do I need to revise? Or are there the resources within the theory to handle this objection?" In many cases, I discovered new strengths in the theory that I hadn't known before.

Then I would use them later on as advertisements for the theory: "You might say such and such is a problem; here’s the answer." A lot of my colleagues now, when I talk with them, I know the first 10 objections they're going to have because I've been given those objections. I've thought about them. I give them the objections. I give them the answers. It forces the discussion to a new level which is good.

All the easy objections, quote unquote "easy objections," are taken out of the picture. Let's go deeper. Give me an objection I've not heard before so I can learn something new. That's sort of the attitude. Take the objections. Learn from them. It's always a growing experience. If you have the attitude "If someone is disagreeing with me, no, I'm not going to listen to that," that's when you stop learning.