In everyday life, we are used to thinking that our senses, our perception - sight, sounds, textures, tastes - give us an accurate picture of the real world. Of course, when we think about it for a second - or succumb to the trick of our senses - we realize that we will never be able to perceive this world exactly. Our brain rather makes assumptions about what this world is like, as if imitating external reality. Still, this imitation should be pretty good. If it were not so, would we not be left on the sidelines of evolution? True reality can always remain beyond our reach, but our senses should at least generally outline what this reality can be.
Cognitive scientist Donald Hoffman uses evolutionary play theory to show that our perception of independent reality must be an illusion. He believes that our senses do not owe us anything. Hoffman is Professor of Cognitive Sciences at the University of California, Irvine. For the past thirty years, he has studied perception, artificial intelligence, evolutionary game theory and the brain and made a very dramatic conclusion: the world presented to our perception has little to do with reality. Moreover, he says, we must thank evolution itself for this magical illusion, since the need for evolution grows with the belittling of truth.
An attempt to understand the nature of reality and to separate the grain from the chaff, the observer from the observed, is made at the border of neurobiology and fundamental physics. On one side, you'll find scientists trying to understand how a kilogram of gray matter, obeying only the usual laws of physics, leads to a conscious first-person experience. They call it the "hard task."
Donald Hoffman
On the other side, there is quantum physics, surprising everyone with the strange fact that quantum systems do not seem to be separate objects localized in space until we begin to observe them. Experiment after experiment, scientists have shown - contrary to common sense - by assuming that the particles that make up ordinary objects exist independently of the observer, we get the wrong answers. The main lesson of quantum physics is quite clear: there are no public objects present in some pre-existing space. As physicist John Wheeler put it, "The view that the world exists 'out there', independently of us, is no longer valid."
So, while neuroscientists are struggling to understand how something like first-person reality can exist, quantum physicists are dealing with the mystery of how something other than first-person reality can exist. And here lies the area of Hoffman's work - stretching the boundaries in an attempt to create a mathematical model of the observer, to get to reality on the other side of the illusion. Quanta Magazine interviewed the scientist about his work and results.
Humans often use Darwinian evolution as an argument that our perceptions accurately reflect reality. They say: “Obviously, we must be well versed in this reality, otherwise we would have been erased long ago. If I think I see a palm tree, but in fact there is a tiger, I am in trouble."
Promotional video:
Right. The classic argument is that those of our ancestors who saw more had a competitive advantage over those who saw less, and therefore most likely passed on their genes that determine more accurate perception. This means that after thousands of generations we can be quite sure that we are the descendants of those who saw more accurately, and we see more accurately. Sounds logical. But I believe that this is fundamentally wrong. This argument does not reflect the fundamental fact about evolution, which is its fitness function (fitness function) - mathematical functions that describe how well a defined strategy achieves its goals of survival and reproduction. The mathematician and physicist Chetan Prakash proved the theorem I mentioned, and it says: according to evolution by natural selection, an organism that sees reality as it is,will never be more adaptable than an organism of equal complexity, which does not see reality at all, but is able to adapt. Never.
You have done computer simulations to show this. Can you give us an example?
Suppose in reality there is a resource, water, for example, and you can quantify it objectively - very little water, average amount of water, a lot of water. Now, suppose your fitness function is linear, so a little water will give you a little fitness, medium water will give you average fitness, and plenty of water will give you the most flexibility - in which case an organism that sees the truth about water around the world can to win, but only if the fitness function is built in accordance with the present structure in reality. But truth be told, this will never happen in the real world. A bell-shaped curve is more likely: say, too little water - you will die of thirst, too much water - you will drown,and somewhere in between it will be good for survival. Now the fitness function no longer matches the structure of the real world. And that's enough to nullify the truth. For example, an adaptive organism might consider low and excessive water levels to be, say, a "red signal" indicating poor fitness, and intermediate values as green indicating high fitness. His perception will be attuned to fit, not truth. He will not see any difference between small and large - only red - even if it exists in reality.indicating low fitness, and intermediate values are green, meaning high fitness. His perception will be attuned to fit, not truth. He will not see any difference between small and large - only red - even if it exists in reality.indicating low fitness, and intermediate values are green, meaning high fitness. His perception will be attuned to fit, not truth. He will not see any difference between small and large - only red - even if it exists in reality.
But how will seeing false reality be beneficial for the survival of the organism?
There is a metaphor that has become available to us in the last 30-40 years, and this is the desktop interface. Suppose there is a blue rectangular icon in the lower right corner of your computer desktop - does this mean that the file itself is blue, rectangular and lives in the lower right corner of your computer? Of course not. It's just the way things are arranged on your desktop - it has color, position, and shape. These categories are simply available to you, and none of them will tell the truth about the computer itself. And this is interesting. You couldn't form a true description of the internals of a computer if your whole vision of reality was limited to the desktop. Still, the desktop is quite useful. The blue rectangular icon guides my behavior and hides in a complex reality that I don't want to know. This is the key idea. Evolution has endowed us with the organs of perception that allow us to survive. They guide the mechanisms of adaptation. But some of this is hidden in mechanisms that we don't need to know. And this, however, is a huge part of reality, whatever this reality is in reality. If you spend too much time analyzing all this, the tiger will devour you.
Does this mean that everything we see is one big illusion?
We are equipped with senses that allow us to live, and therefore we have to take them seriously. If I see something that looks like a snake, I will not take it. If I see a train, I will not stand in front of it. These symbols keep me alive, so I am serious with them. But it is wrong to believe that if we are to take them seriously, we must also take them literally.
If snakes are not snakes and trains are not trains, what are they?
Snakes and trains, like particles in physics, do not have any objective functions independent of the observer. The snake I see is a description created by my sense system, telling me a sequence of actions conditioned by my fitness. Evolution creates acceptable solutions, not optimal ones. The snake is an acceptable solution to a problem that tells me how to act in such a situation. My snakes and trains are my mental representations; your kites and trains are your ideas.
When did you start thinking about it?
As a teenager, I was interested in this question: are we machines? My reading of science has shown that yes. But my grandfather was a priest, and in the church they said no. So I figured I needed to figure it out on my own. It's kind of an important personal question - if I'm a machine, I need to find out. If not, you also need to find out what this special magic is, that I am not a machine. As a result, in the 1980s, I ended up in the MIT artificial intelligence lab, where I worked on machine perception. In the field of vision, there has been an unexpected success in developing mathematical models for specific visual abilities. I noticed that they have a common mathematical structure, and I thought that it would be possible to write out a formal structure that, possibly, covers all possible observation modes. I was inspired in part by Alan Turing. When he invented the Turing machine, he tried to create an abstract computing machine. And instead of putting a bunch of unnecessary things on it, he said: let's take the simplest mathematical description that can work. And this simple formalism formed the basis of computer science, the science of computing. And I wondered if such a simple formalism could be used as the basis for the science of observation?
Mathematical model of consciousness
Exactly. Intuition told me that there is a conscious experience. I feel pain, tastes, smells, I can see, experience, feel emotions, and so on. One part of this structure of consciousness is the collection of all kinds of experiences. When I have this experience, based on the experience I have, I may want to change what I do. Therefore, I must have a collection of possible actions that I can take and a decision-making strategy that, based on my experience, allows me to change my actions. This is the main idea. I have an X space for experience, a G space for actions, and an algorithm D that allows me to choose new actions based on my experience. I also add the space W for the world, which is also a space of possibilities. This world somehow affects my perceptionstherefore there is a map P from the world to my experience, and when I act, I change the world, therefore there is a map A from the action space to this world. Here is the whole structure. Six elements. And I think this is the structure of consciousness.
But if there is a W, are you implying the existence of the outside world?
This is the most interesting thing. I can extract W from the model and put a conscious agent in its place, thus obtaining a chain of conscious agents. Basically, you can end up with entire networks of arbitrary complexity. And this is the world.
Is the world just other conscious agents?
I call this conscious realism: objective reality is only conscious agents, points of view. I can take two conscious agents and make them interact, and the mathematical structure of this interaction will also satisfy the definition of a conscious agent. And math tells me something. I can take two consciousnesses, and they can give rise to a new, united, unified consciousness. Here's a concrete example. Our brain has two hemispheres. But when you perform a brain severing operation that completely severs the corpus callosum, you get clear evidence of two separate consciousnesses. Before this separation, consciousness was one. So it cannot be said that there is a single agent of consciousness. I didn't expect mathematics to lead me to admit it. It follows from this that I can take individual observers,combine them and create new observers, and do it ad infinitum. There are some conscientious agents.
If it's all about conscious agents, first-person perspectives, what about science? Science has always been a third-person description of the world
If what we are doing is measuring public objects, and if the objectivity of the results is that you and I can measure the same object in the same situation and get the same result - from quantum mechanics it becomes apparent that this is not working. Physics tells us that there are no publicly accessible physical objects. What to do? I can tell you that I have a headache, and even believe that I will tell you well, because you too have ever had headaches. The same goes for apples, the moon, the sun, and the universe. Just as you have your own headache, so you have your own moon. But I suppose she will be the same as mine. This assumption may be wrong, but it underlies my post and it is the best,what can we do in relation to publicly available physical objects and objective science.
It doesn't seem that many people in the field of neuroscience or philosophy of mind are thinking about fundamental physics. Do you think this is a stumbling block for those trying to understand consciousness?
I think it was. Not only do they ignore progress in fundamental physics, they often do it on purpose. They openly state that quantum physics is not concerned with aspects of brain function that are a partial cause of consciousness. They are sure that the point is in the classical properties of neural activity, which exists independently of observers - the strength of synaptic connections, dynamic properties, and so on. These are very classical concepts of Newtonian physics, in which time is absolute and objects exist absolutely. And then neuroscientists don't understand why they don't have breakthroughs. They move away from incredible breakthroughs and insights made by physics. "We will be with Newton even after 300 years."
I suspect they respond to things like Roger Penrose and Stuart Hameroff's model where you still have a physical brain in space, but presumably doing quantum work. Instead, you say, "Look, quantum mechanics tells us that we should question the very notions of 'physical things' being in 'space'
Neuroscientists say: "We don't need to involve quantum processes, we don't need a quantum wave function collapsing in neurons, we can just use classical physics to describe processes in the brain." I will emphasize once again the big lesson of quantum mechanics: neurons, brains, space … these are all just symbols that we use. They are not real. There is no classic brain doing some kind of quantum magic. There is no brain! Quantum mechanics says that classical objects, including the brain, do not exist. This is a very radical statement about the nature of reality, and does not include the brain doing sophisticated quantum calculations. So even Penrose didn't go far enough. But most of us, as you know, are born realists. We are born physicalists. It's very, very difficult to refuse.
Coming back to the question you started with: are we machines?
The formal theory of conscious agents that I am developing is computationally universal - in a sense, it is a machine theory. And since this theory is computationally universal, I can extract cognitive science and neural networks from it. However, at this point in time, I don't think we are machines, in part because I can distinguish between a mathematical representation and the thing that is represented. As a conscious realist, I postulate conscious experience as ontological primitives, the basic ingredients of the world. I affirm that my experience is above all. The experience of everyday life - my true sense of headache, my true taste of chocolate - this is the ultimate nature of reality.
Based on materials from Quanta Magazine
ILYA KHEL