Marvin Minsky "The Emotion Machine": Chapter 4. "How We Recognize Consciousness"

4-3 How do we recognize Consciousness?

Student: You still haven't answered my question: if “consciousness” is just a polysemous word, which makes it such a particular thing.

Here is a theory explaining why it happened: Most of our mental activity proceeds, to a greater or lesser extent, “unconsciously” - in the sense that we can hardly guess its existence. But when we encounter difficulties, it launches high-level processes that have the following properties:
')

1. They use our latest memories.
2. They often work in series than in parallel.
3. They use abstract, symbolic or verbal descriptions.
4. They use models that we lay down about ourselves.

Now suppose that the brain can create a resource C that starts when all the above processes start working together:


If such a C-detector turns out to be quite useful, then this may lead us to the idea that it reveals the existence of some kind of “Conscious Stuck”! In fact, we can even assume that this entity is the reason for the existence of a set of the processes described above, and our language system could associate a C-detector with such words as "awareness", "self", "attention" or "I". To see why such a view may be useful for us, we need to consider its four components.

Recent memories: Why should consciousness use memory? We constantly perceive consciousness as the present, and not the past - as something that exists now.

In order for any mind (like any machine) to know what has been done before, it must have records of recent activity. For example, suppose I asked the question: "Do you realize that you are touching your ear?" You can answer: "Yes, I am aware that I am doing this." However, in order to make such a statement, your language resources had to respond to signals from other parts of the brain, which, in turn, responded to previous events. Thus, when you start talking (or thinking) about yourself, you need some time to collect the requested data.

Generally speaking, this means that the brain cannot reflect on what it is thinking right now; at best, he can look at some of the recent events. There is no reason that no part of the brain could process the results of the activity of other parts of the brain - but even in this case there will be a slight delay in obtaining information.

The sequential process: Why are our high-level processes for the most part consistent? Wouldn't it be more efficient for us to do many things in parallel?

Most of the time in your daily life, you do a lot of things at once; it’s easy for you to walk, talk, see and scratch your ear at the same time. But very few people are able to tolerably draw a circle and a square using both hands at the same time.

Everyman: Perhaps each of these two tasks requires so much attention from you that you cannot focus on another task.

This statement will make sense if we assume that attention is given in limited quantities - but based on this we need a theory explaining what this kind of restriction can impose, given that we can still walk, talk and look at the same time. One explanation is that such restrictions may occur when resources start to conflict. Suppose the two tasks performed are so similar that they need to use the same mental resources. In this case, if we try to do two similar things at the same time, one of them will be forced to interrupt our work - and the more conflicts like these occur in our brain, the less we can do at the same time.

Then why can we see, walk and talk at the same time? This is presumably due to the fact that our brain has different systems located in different parts of the brain for these activities, thus reducing the number of conflicts between them. However, when we are forced to solve extremely complex tasks, then we have only one way out: in some way to divide this task into several parts, each of which will require high-level planning and thinking. For example, solving each of these subtasks may require one or more “assumptions” about a given problem, and then a mental experiment will be required to confirm the correctness of the assumption.

Why can't we do it at the same time? One of the possible reasons can be quite simple - the resources that are necessary for the preparation and implementation of plans have evolved quite recently - about a million years ago - and we do not have enough many copies of these resources. In other words, our higher levels of “governance” do not have enough resources — for example, resources to keep track of the tasks to be done and resources to find solutions to the tasks with the involvement of funds with the least amount of internal conflicts. Also, the above processes most likely use symbolic descriptions that we described earlier — and these resources also have a limit. If this is the case, then we simply have to consistently focus on goals.

Such mutual exclusions can be the main reason why we perceive our thoughts as a “stream of consciousness”, or as an “inner monologue” - a process in which a sequence of thoughts can resemble a story or a story. When our resources are limited, we have nothing else but slow “sequential processing of information”, which is often called “high-level thinking”.

Symbolic description: Why are we forced to use symbols or words, instead of, say, direct contacts between brain cells?

Many researchers have developed systems based on prior experience by changing the connections between different parts of the system, called "neural networks" or "student machines through the creation of contacts." It has been proven that such systems are capable of learning to recognize different types of patterns — and it is likely that a similar low-level process that is at the heart of “neural networks” can underlie most of our brain functions. However, although these systems are extremely useful in various useful areas of human activity, they cannot meet the needs of more intelligent tasks, because they store their information in the form of numbers that are difficult to use by other resources. Some may use these figures as a measure of correlation or probability, but they will not have the slightest idea what else these figures can say. In other words, such a presentation of information does not have sufficient expressiveness. For example, a small neural network might look like this.


In comparison, the figure below shows the so-called "Semantic Web", which shows some of the connections between parts of the pyramid. For example, each link that points to the concept of supports (supports) can be used to predict the fall of the upper block, if the lower blocks are removed from their seats.


Thus, while the “network of links ” shows only the “strength” of interaction between elements, and says nothing about the elements themselves, the three-level links of the “semantic network” can be used for various considerations.

Models of ourselves: Why do we include “models of ourselves” in the necessary processes in your first diagram?

When Joan thought about what she did, she asked herself the question: "What would my friends think about me?" And the only way to answer the question would be to use descriptions or models that represent her friends and herself. Some models of Joan would describe her physical body, others her goals, and still others her attitudes toward various social and physical events. In the end, we would create a system that includes a set of stories about our past, ways of describing the state of our mind, a set of knowledge about our capabilities and visualizations of our friends. Chapter 9 will explain in more detail how we do things like this and create “models” of ourselves.

When Joan creates a data set of models, she can use them for self-reflection - and then she finds herself thinking about herself. If these reflexive models lead to any choice of behavior tactics, then Joan will feel that she is “in control of herself” - and probably uses the term “conscious” to summarize this process. Other processes that occur in the brain that she is unlikely to be aware of will be attributed to Joan to the area beyond her control and call them "unconscious" or "unintended." And as soon as we ourselves can create machines with a similar line of thinking, they may also learn to say phrases like: “I’m sure you know what I mean when I talk about“ mental experience ”.”

I do not insist that such detectors (as a C-detector ed.) Should be involved in all the processes that we call consciousness. However, without the presence of ways to recognize specific patterns of mental states, we may not be able to talk about them!

∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞

This section began with a discussion of some ideas about what we mean when we talk about consciousness, and we have suggested that consciousness can be described as the discovery of some high-level activity in the brain.


However, we also wondered what might trigger the beginning of these high-level activities. We can consider their manifestation in the following example: for example, among Joan's resources there are “Problem Detectors” or “Critics” that work when Joan’s thoughts come up against problems — for example, when she does not achieve some important goal, or does not solve any some problem. Under these conditions, Joan can describe the state of her mind with the terms "misfortune" and "frustration" and try to get out of this state with the help of a sensible activity, which can be characterized by the following words: "Now I have to force myself to concentrate." Then she can try to think about the situation, which will require the participation of a set of higher-level processes — for example, by activating a set of brain resources:


This suggests that we sometimes use “consciousness” to describe actions that initiate processes rather than recognize the beginning of the work of high-level processes.

Student: On what basis do you choose terms for your schemes, and define the words “consciousness” through them? Since “consciousness” is a polysemous word, each person can make up a list of terms that can be included in it.

In fact, since a large number of psychological words have many meanings, we will most likely switch between different sets of terms that best describe these set-valued words, for example, “consciousness”.

∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞

4.3.1 The Illusion of Immanence

“ The paradox of consciousness - the more intelligent a person is, the more layers of information processing separate him from the real world - this, like many other things in nature, is a kind of compromise. Progressive distancing from the outside world is the price paid for any knowledge about the world in general. The deeper and wider [our] our knowledge of the world becomes, the more complex layers of information processing are necessary for further knowledge. ”
- Derek Bickerton, Languages ​​and Species, 1990.

When you enter the room you have the feeling that you instantly see everything in your field of vision. However, this is an illusion, because you need time to recognize objects that are in the room, and only after this process do you get rid of the wrong first impressions. However, this process proceeds so quickly and smoothly, which requires an explanation - and it will be given later in the chapter §8.3 Pan-Analogy.

The same happens inside our mind. We usually have a constant feeling that we are “aware” of things happening around us now . But if we look at the situation from a critical point of view, we will understand that there is some kind of problem with this view - because nothing can be faster than the speed of light. This means that no part of the brain can know what is happening "now" - neither with respect to the external world, nor other parts of the brain. The maximum that the part we are considering may know - what happened in the near future.

Everyman: Then why does it seem to me that I am aware of all the signs and sounds, and also feel my body at every moment of time? Why it seems to me that all signals perceived by me are processed instantly?

In everyday life, we can assume that everything that we see and feel is “aware” here and now, and usually it does not go sideways to the assumption that we are in constant contact with the outside world. However, I will argue that this illusion stems from the features of the organization of our mental resources - and the above described phenomenon should finally be given a name:

Illusion of Immanence: Most of the questions you ask will be answered before higher levels of consciousness begin to connect to the search for answers to these questions.

In other words, if the answer to a question that interests you, you get before you realize that you needed it, you get the feeling that you knew the answer right away and it seems that no work of reason happened.

For example, before you enter a room that you know, it’s likely that you’ve already scrolled through the memory of that room and, perhaps, after you’ve entered, it will take some time to notice the changes that have occurred in the room. The idea that a person constantly perceives the present moment is irreplaceable in everyday life, but much of what we suppose we see is our stereotypical expectations.

Some argue that it would be great to be constantly aware of what is happening. But the more often your high-level processes change their views on reality, the more difficult it will be for them to find meaningful information in changing conditions. The strength of our high-level processes derives not from the continuous changes of our descriptions of reality, but because of their sufficient stability.

In other words, in order for us to feel how much of the surrounding and internal environment persists over time, you need to be able to explore and compare descriptions from the recent past. We notice the changes in spite of them, and not because they occur. Our feeling of constant contact with the world is the Illusion of Immanence: it arises when every question we ask is answered in our head before the question is asked - as if the answers were already there.

In Chapter 6, we will look at how our ability to activate knowledge before we need it can explain why we use things like “common sense” and why it seems “obvious” to us.

4.4 Reassessing Consciousness

“Our mind is so well created that we can begin to think, without any understanding of how it works. We can only be aware of the result of this work. The realm of unconscious processes is an unknown being who works and creates for us, and ultimately brings the fruits of our efforts to our knees. ”
- Wilhelm Wundt (1832-1920)

Why is "Consciousness" a mystery to us? I argue that the reason for this is our exaggeration of our own insight. For example, at a specific point in time, the lens of your eye can focus only on one object at a limited distance, while other objects out of focus will be blurred.

Philistine: It seems to me that this fact does not apply to me, because all the objects that I see are perceived by me quite clearly.

You can see that this is an illusion if you focus your eyes on your fingertip while looking at a distant object. In this case, you will see two objects instead of one, and both will be too blurred so that they can be viewed in detail. Before we did this experiment, we thought that we could see everything all at once clearly, because the lens of the eye was so quickly adjusted to the examination of surrounding objects that we didn’t have a feeling that the eye could do it. Similarly, many people think that they see all the colors in their field of view - but a simple experiment showed that we see the correct colors of things only near the object that our eyes are directed at.

Both of the above examples refer to the Illusion of Immanence because our eyes react incredibly quickly to things that attract our attention. And I argue that the same thing applies to consciousness: we make almost the same mistakes with regard to what we can see inside our mind.

Patrick Hayes: “Imagine what it is like to be aware of the processes by which we create imaginary (or real) speech. [In this case] a simple act, like, say, “inventing a name”, would turn into a sophisticated and skillful use of a complex mechanism of lexical access, which would be like playing on an internal organ. Words and phrases that are necessary for us to communicate will themselves be distant goals, for the achievement of which requires knowledge and skills that the orchestra has, playing a symphony, or a mechanic who understands an intricate mechanism. ”

Hayes goes on to say that if we knew how everything inside us works then:

“We would all be in the role of servants of the past of us; we would run inside the mind trying to understand the details of the mental machinery, which is now incredibly conveniently hidden from the eyes, leaving time to solve more important issues. Why are we in the engine room if we can be on the bridge? ”

With this paradoxical view, consciousness still seems surprising - not because it tells a lot of information about the world, but because it protects us from the above described tedious things! Here is another description of this process, which can be found in Chapter 6.1. "Society of Mind"

Think about how the driver drives the car without any knowledge of how the engine works, or why the wheels of the car turn left or right. But if we start thinking about it, we will realize that we control both the machine and the body in a rather similar way. This also applies to conscious thought - the only thing you have to worry about is the choice of direction, and everything else will work by itself. This incredible process includes a huge number of muscles, bones and ligaments, controlled by hundreds of interacting programs that even experts cannot understand. However, all you have to do is think “turn in that direction” and your wish is automatically fulfilled.

And if you think it could hardly be otherwise! What would happen if we were forced to perceive trillions of connections in our brain? Scientists, for example, have watched them for hundreds of years, but they still do not understand how our brain works. Fortunately, in modern life we ​​just need to know what needs to be done! This can be compared with our vision of a hammer as an object that can be hit on things, and a ball - as an object that can be thrown and caught. Why do we see things not as they are, but in terms of their use?

Likewise, when you play computer games you control what is happening inside the computer mainly due to the use of symbols and names. The process that we call "consciousness" works by and large in exactly the same way. It seems that the highest levels of our consciousness are sitting at mental computers, driving huge machines in our brain, not understanding how they work, but simply “pushing” on various symbols from the list that continually appears on mental displays.

Our mind developed not as a tool for observing oneself, but for resolving the practical problems associated with food, protection and reproduction.

4.5 Models of Self and Self-Consciousness

If we consider the process of forming self-awareness, we should avoid single signs of its manifestation, such as recognizing and separating individual parts of our body from the environment, using its words such as “I”, and even recognizing its own reflection in the mirror. The use of personal pronouns may be due to the fact that the child begins to repeat the words and phrases that others say about him. This repetition may begin in children at different ages, even if their intellectual development proceeds in the same way.
- Wilhelm Wundt. 1897

In § 4.2, we assumed that Joan “created and used models of herself” - but we did not explain what we mean by the word model. We use this word in several meanings, for example, “Charlie is a model administrator”, which means that you should focus on it, or, for example, “I create a model airplane” which means creating a smaller object in size. But in this text we use the phrase “model X” to denote a simplified mental representation, which allows us to answer some questions about a complex object X.

Thus, when we say "Joan has a mental model of Charlie, " we mean that Joan has some mental resources that help her answer some questions about Charlie. I singled out a few because each of Joan’s models will work well with certain types of questions — and will give incorrect answers to most other questions. Obviously, the quality of Joan’s thinking will depend not only on how good her models are, but also on how good her skills are in choosing these models in specific situations.

Some models of Joan will predict how physical actions can affect the world around them. She also has mental models that predict how mental acts can change her mental state. In chapter 9 we will talk about some models that she can use to describe herself, i.e. answer some questions about her abilities and inclinations. These models can describe:

Her various goals and ambitions.

Her professional and political views.

Her notions about their competencies.

Her ideas about her social roles.

Her various moral and ethical views.

Her faith in who she is.

For example, she can use some of these models to evaluate whether it is worth relying on herself when doing something. Moreover, they can explain some ideas about their minds. To show this, I will use the example offered by the philosopher Drew McDermott.

Joan is in some room. She has a model of all the objects in this room. And one of the objects is Joan herself.


For most objects, there will be their own sub-models, which, for example, will describe their structure and functions. Model Joan for the object "Joan" will be the structure that she calls "I", which will include at least two parts: one of them will be called the Body , the second - Mind .


Using different parts of this model, Joan can answer “ Yes ” to the question: “ Do you have a mind? ". But if you ask her: “ Where is your mind? "- this model will not be able to help answer the question as some people do:" My mind is inside my head (or inside my brain) . " However, Joan will be able to give a similar answer if I contain an internal connection between the Mind and the Body or an external connection between the Mind and another part of the body called the Brain .

More generally speaking, our answers to questions about ourselves depend on the models that we add up about ourselves. I used the word model instead of model, because, as we will see in Chapter 9, a person needs different models in different conditions.Thus, there can be many answers to the same question, depending on what goal the person wants to achieve and these answers sometimes will not be the same.

Drew McDermott: Few believe that we have such models, even fewer people know that we have them. The key feature is not that the system has a model of itself, but that it has a model of itself as a conscious being. ”- comp.ai.philosophy, February 7, 1992.

However, these descriptions of themselves may be incorrect, but they are unlikely to continue to exist if they do nothing useful for us.

And what will happen if we ask Joan: “ Did you realize what you just did and why did you do it? "?

If Joan has good models as she makes her choice, then she will feel that she has some “ control ” over her actions and uses the term “ conscious decisions ” to describe them. But for a type of activity for which she does not have good models, she can classify as independent of her and call her “ unconscious ” or “ unintended ”. Or vice versa, she may consider that she is still in complete control of the situation and makes some decisions based on “ free will ” - which, despite what she can say, will mean: “ I don’t have a good explanation of what made me do this act ".

Thus, when Joan says: “ I made a conscious choice ” - this does not mean that something magical happened. This means that she attributes her thoughts to different parts of her most useful models.

∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞

4.6 Cartesian Theater

“We can consider the mind as a theater that puts simultaneous productions. Consciousness consists of comparing them with each other, choosing the most appropriate in the given conditions and suppressing the least needed by strengthening and reducing the degree of attention. The best and most noticeable results of mental work are selected from data provided by a lower level of information processing, which is eliminated from even more simple information and so on. ”
- William James.

We sometimes compare the work of the mind with a performance staged on a theatrical stage. Because of this, Joan can sometimes present herself as a spectator in the first row of the theater, and “thoughts in her head” as playing actors. One of these actors was a pain in her knee (§3-5), which began to play a major role. Soon, Joan began to hear a voice in her head: “ I have to do something with this pain. It prevents me from doing anything. "

Now, when Joan begins to think about how she feels and what she could do, Joan herself will appear on the stage. But in order for her to hear what she is saying, she must also be in the hall. Thus, we have two copies of Joan - as an actor, and as a spectator!

If we continue to watch this performance, more copies of Joan will appear on the stage. There should be a Joan-writer for writing a script for speeches and a Joan-designer for staging scenes. There should also be other Joan backstage to control the scenes, lights and sounds. Joan-director for the production of the play and Joan-critic should appear so that he can complain: “ I can not bear this pain anymore !”

However, when we carefully consider this theatrical point of view, we will see that it puts us additional questions and does not provide the necessary answers. When does Joan the critic begin to complain of pain, how does she feel about Joan currently performing on stage? Do we need a separate theater for each of these actresses to conduct performances with only one Joan? Of course, the theater in question does not exist, and the Joan objects are not human. They are just different models of Joan herself, which she created to present herself in various situations. In some cases, these models are very similar to cartoon characters or caricatures, in others - they are completely different from the object from which they are drawn. Anyway, Joan's mind abounds with various models of Joan herself - Joan in the past, Joan in the present and Joan in the future.There are remnants of Joan's past, and Joan what she wants to become. There are also intimate and social models Joan, Joan-athletes and Joan-mathematicians, Joan-musicians and Joan-politicians, and various types of Joan-professionals - and because of their different interests, we cannot even hope that all joan will get along. We will discuss this phenomenon in more detail in Chapter 9.

Why does Joan create such models of herself? Reason is a confusion of processes that we barely understand. And whenever we are confronted with the fact that we do not understand, we try to present it in familiar forms, and there is nothing more appropriate than the various objects that are around us in space. Therefore, we can imagine the place where all the thought processes are located - and, what is most striking, many people actually create such places. For example, Daniel Dennett called this place the “Cartesian Theater”.

Why is this image very popular? Firstly, it does not explain many things, but its presence is much better than using the idea that all thinking is carried out by one I. It recognizes the existence of various parts of the mind and their ability to interact, and also serves as a kind of “place” where all processes can work and communicate. For example, if different resources offered their plans for what Joan should do, then the idea of ​​the theater stage could give an idea of ​​their common place of work. In this way, the Cartesian Theater Joan allows her to use many of the real life skills she has learned “in her mind.” And it is this place that gives her the opportunity to start thinking about how decisions are made.

Why do we find this metaphor so believable and natural? Possible ability“Modeling the world within one's mind” was one of the first devices that led our ancestors to the possibility of self-reflection. (There are also experiments showing that some animals also create in their brains similar to the map of the environment they are familiar with). In any case, the metaphors like those described above permeate our language and thoughts. Imagine how hard it would be to think without hundreds of different concepts like: “ I come to my goal .” Spatial models are so useful in our daily life, and we have such powerful skills to use them that it begins to seem that these models are used in any situation.

However, we may have gone too far, and the concept of the Cartesian Theater has already become an obstacle to further consideration of the psychology of the mind. For example, we must recognize that the theatrical scene is only a facade that hides the main action that takes place behind the scenes - what is happening there is hidden in the minds of the actors. Who or what determines what should appear on the stage, that is, who chooses exactly who will entertain us? How exactly does Joan make decisions? How can such a model present a comparison of two different possible “future outcomes of a situation” without simultaneously maintaining two theaters?

The image of the theater in itself does not help us to answer such questions, because it gives too much mind to Joan, who is watching the production from the audience. However, we have the best way to present this Global Work Area, which was suggested by Bernard Baars and James Newman, who suggested the following:

« , «». … . … , … [] . … « » , , , « ». (.. ). , .»

This last paragraph warns us not to assign too large a role to the compact self or “homunculus” - a miniature person inside the mind doing all the hard mental work, instead we have to distribute this work. For, as Daniel Dennett said,

“homunculi are boogie-names if they copy all our talents who provide our work, although they had to deal with their explanation and provision. If you assemble a team or committee from relatively ignorant, narrow-minded, blind homunculi to create rational behavior of the whole team, this will be progress. ”- in Brainstorming 1987, p. 123.

All ideas in this book support the above statement. However, there are serious questions about the extent to which our minds depend on a common workspace or a bulletin board. We conclude that the idea of ​​a “cognitive market” is a good way to start thinking about how we think, but if we look at this model in more detail, we will see the need for a much more complex representation model.

∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞

4.7 Sequential Flow of Consciousness

“The truth is that our mind is not in the present moment of time: memories and anticipations take up almost all the time the brain works. Our passions are joy and grief, love and hate, hope and fear belong to the past, for the reason that caused them should appear before the effect. ”
- Samuel Johnson.

The world of subjective experience seems perfectly continuous. It seems to us that we live here and now, moving steadily into the future. However, when we use the present tense, we are confused every time, as already noted in §4.2. We may know about what we have done recently, but we do not have the opportunity to know what we are doing right now.

Everyman: Funny. Of course, I know what I'm doing right now, and what I think right now, and how I feel right now. How does your theory explain why I feel a continuous stream of consciousness?

Although what we perceive seems to be a “real tense” for us, in reality everything is much more complicated. To build our perception, some resources must pass through our memory in succession; sometimes they have to look at our old goals and grieves to assess how far we have come to a specific goal.

«[ ] , . , , , « », . , , .»

In addition, it is safe to assume that different parts of your minds process information with significantly different speeds and different delays. So, if you try to present your recent thoughts as a consistent story, your mind will have to somehow compose it by choosing previous thoughts from various streams of consciousness. In addition, some of these processes attempt to anticipate events that attempt to predict the “predictive mechanisms” that we describe in §5.9. This means that the “content of your consciousness” is connected not only with memories, but also reflections about your future.

Therefore, the only thing you really cannot think about is what your mind is doing right now, because every brain resource can know at best what other brain resources did a few moments ago.

The Philistine: I agree that, by and large, what we think is connected with recent events. But I still feel that we have to use some other idea to describe the work of our mind.

HAL-2023: , , . , , , . , , .

: , , , , . , . ?

HAL-2023: , , « ». - , , .

: ?

HAL-2023:From time to time. Although these recordings can make me more “self-aware” than any other person, they do not contribute to improving the quality of my work, because I only use them in emergency situations. Error handling is so tedious that it makes my mind work extremely slowly, and therefore I begin to look at recent activities only when I notice my inhibited work. I constantly hear people say, "I try to connect with myself." However, believe my experience, they will not come very close to resolving the conflict if they can do it.

∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞

4.8 The Secret of "Experience"

Many thinkers argue that even if we learn everything about how our brain works, there will be another fundamental question: “ Why do we feel things ?”. Philosophers claim that the explanation of “subjective experience” may be the most difficult problem of psychology, and perhaps this problem will never be solved.

David Chalmers: “Why, when our cognitive systems start working with visual and sound processing of information, do we get a visual or audio experience, such as a deep blue feeling or medium C sound? How can we explain why there is something that can entertain a mental image or survive an emotion? Why should physical information processing engender a rich inner life? Gaining experience goes beyond the knowledge that can be obtained from a physical theory. ”

It seems to me that Chalmers believes that gaining experience is a fairly simple and clear process - and therefore should have a simple, compact explanation. However, as soon as we realize that each of our daily psychological words (such as experience , sensation, and consciousness ) are related to a large number of different phenomena, we must abandon finding the only way to explain the content of these multi-valued words. Instead, we should initially formulate theories about each multi-valued phenomenon. Then we may be able to find their common characteristics. But until we can properly divide these phenomena into parts, it would be rash to conclude that what they describe cannot be “derived” from other theories.

Physicist: Perhaps the brain works according to rules that are still unknown to us, which cannot be transferred to the machine. For example, we still do not fully understand how gravity works, and consciousness can be a similar example.

This example also suggests that there must be one source or cause of all the wonders of "consciousness." But as we saw in §4.2, consciousness has far more meanings than can be explained using one or the general method.

Essentialist: What about the fact that consciousness makes me aware of myself? It tells me what I think now, and thanks to him I know that I exist. Computers calculate without putting in any meaning, but when a person feels or thinks, the feeling of “experience” comes into play, and nothing more basic than this feeling.

In Chapter 9, we will discuss what it would be to believe that you are “aware of yourself”, with the exception of very gross daily approximations. Instead, we constantly switch between the different “models of ourselves” that you have, and each of them is based on a different, incomplete set of incomplete data. “Experience” may seem to us understandable and direct - but often it is not correctly drawn up here, because each of your different views on yourself can be based on missteps and different types of errors.

Whenever we look at someone else, we see its appearance, but not what is inside. This is the same as looking in a mirror - you see only what lies outside of your skin. Now, in the popular representation of consciousness, you also have a magic trick that allows you to look at yourself from the inside and see everything that happens in your mind. But when you think about this topic more carefully, you will see that your “privileged access” to your own thoughts may be less accurate than the “understanding” of you by your close friends.

The Philistine: This assumption is so stupid that it annoys me, and I know this thanks to some particular thing flowing from within me that tells me what I think.

Your friends can see that you are concerned too. Your mind cannot tell you the details of why you feel irritated, why you shake your head and use the word “ annoying ” instead of “ bothering ”? Indeed, we cannot see all the thoughts of a person, watching his actions from the outside, but even when we look at the thought process “ from the inside ”, it is difficult for us to be sure that we really see more, especially considering that such “insights” are often wrong . Thus, if we mean “ consciousness ” “ awareness of our internal processes ”, then this is not true.

“The most merciful thing in the world is the inability of the human mind to relate all that it contains to each other. We live on a quiet island of ignorance, in the middle of the black sea of ​​infinity, but this does not mean that we should not travel far. The sciences, each of which pulls us in its own direction, have so far been little harm to us, but someday the unification of fragmented knowledge will open up such terrifying perspectives of reality and a terrible situation in it, that we either go insane from revelations or run away from deadly light united knowledge into a world of safe new dark age. "
- G.F. Lovecraft, Call of Cthulhu.

∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞

4.9 A-brain and B-brain

Socrates: Imagine people as if they are in an underground dwelling like a cave, where a wide opening stretches across its entire length. From an early age, they have fetters on their legs and neck, so that people do not move, and they see only what is right before their eyes, because they cannot turn their heads because of these fetters. People have their backs to the light emanating from the fire that burns far above, and between the fire and the prisoners passes the upper road, fenced by a low wall like the screen behind which magicians place their assistants when the dolls are shown on top of the screen.

Glavkon: I represent.

Socrates: Behind this wall, other people carry various utensils, holding it so that it is visible over the wall; they sneak and statues, and all sorts of images of living beings, made of stone and wood. In this case, as usual, some of the carriers are talking, others are silent.

Glavkon: Strange you draw an image ...

Socrates: Like us, they see nothing but their shadows or the shadows of these various things cast by fire on the cave wall in front of them ... Then prisoners will consider reality nothing more than these shadows - Plato, the State.

Can you think about what you are thinking right now ? Well, literally, it is impossible - because every thought will change what you think. However, you can be content with something a bit smaller if you imagine that your brain (or mind) consists of two different parts: let's call them A-brain and B-brain .


Now suppose that your A-brain receives a signal that is carried from organs such as eyes, ears, nose, and skin; then he can use these signals to recognize some events in the outside world, and then he can respond to them by sending signals that cause your muscles to contract - which, in turn, can affect the state of the surrounding world. Thus, we can present this system as a separate part of our body.

Your B-brain does not have sensors like the A-brain, but it can receive signals from the A-brain. Thus, the B-brain cannot “see” real things — it can only see their description. Like the prisoner in Plato's cave, who sees only shadows on the wall, B-brain confuses descriptions given by the A-brain, real things, not knowing what they really are. All that the B-brain sees as the “outside world” are events processed by the A-brain.

Neurologist: And this also applies to all of us. For whatever you touch or see, the higher levels of your brain can never directly touch these things, but can only interpret the idea of ​​these things that other resources have added for you.

When the fingertips of two people in love touch each other, no one will argue that physical contact itself has any special meaning. Indeed, in the very similar signals there is no point: the meaning of this contact lies in the representation of this contact in the mind of people in love. Nevertheless, although the B-brain cannot directly perform a physical act, it can still influence the world around it indirectly - by sending signals to the A-brain, which will change its response to external conditions. For example, if the A brain stops in repetition of the same things, the B brain can easily interrupt this process by sending the appropriate signal to the A brain.

Student: For example, when I lose my points, I constantly start searching from a certain shelf. Then a voice begins to blame me for this, which makes me think of searching elsewhere.

In this ideal case, the B-brain can tell (or teach) the A-brain what exactly is worth doing in a similar situation. But even if the B-brain does not have any specific advice, it may not indicate anything to the A-brain, but begin to criticize its actions, as described in your example.

Student: But what would happen if, when I walked along the road, my B-brain would suddenly say: “Sir, you repeat the same actions with your foot more than a dozen times in a row. You should stop right now and some other activity.

In fact, it can be the result of a serious accident. To prevent such errors, the In-brain must have suitable ways of representing things. This accident would not have happened if the B-brain imagined “moving to a certain place” as one long act, for example: “Keep rearranging your legs until you cross the street”, or as a way to achieve the set goal: “Keep reducing the existing distance. Thus, B-brain can work as a manager who has no knowledge of how to do this or that job, but still can give “general” advice on how to do certain things, for example:

If the descriptions provided by the A-brain are too vague - the B-brain will force you to use more specifics.

If the A-brain presents things in too much detail, the B-brain will offer more abstract descriptions.

If the A-brain does something for too long - the B-brain will advise using other techniques to achieve the goal.

How could the V-brain acquire such skills? Some of them could have been built into it from the very beginning, but there must also be a way to learn new skills through learning. For this, the brain may need help from other levels of perception. Thus, when the B-brain oversees the A-brain, another object, let's call it “C-brain”, will supervise the B-brain.


Student: How many layers does a person need? Do we have dozens or hundreds of them?

In Chapter 5, we describe a mind model in which all resources are organized into 6 different levels of perception. Here is a brief description of this model: it begins with a set of instinctive reactions that we have from birth. Then we can begin to reason, imagine and plan for the future, developing ways of behavior that we call “deliberate decisions.” Later, we develop the ability to “reflexive thinking” about our own thoughts. After - we learn self-analysis, which allows us to think about how and why we can think about such things. Finally, we begin to consciously think about whether we should have done all this. Here is how this scheme can be applied to Joan’s thoughts during a crossing:

What made Joan turn on the sound? [Instinctive reaction]

How did she know that it could be a car? [Studied reactions]

What resources were used to make a decision? [Thinking]

How did she decide what to do in this situation? [Reflection]

Why was she thinking about her choice? [Self-reflection]

Did the actions comply with its principles? [Reflection of self-awareness]

Of course, this is too simplistic. These levels can never be clearly defined, because each of these levels, in later life, can use the resources of other levels. However, the established framework will help us begin to discuss the types of resources used by adults, as well as how they are organized.

Student: Why should there exist any levels at all, instead of a single large cloud of interconnected resources?

Our argument in favor of our theory is based on the idea that for the development of effective complex systems, each step of evolution must find a compromise between two alternatives:

If there are few links within the system between its parts, then the capabilities of the system will be limited.

If inside the system there will be many connections between its parts - each subsequent change of the system will limit the work of a large number of processes.

How to achieve a good balance between these extremes? The system can begin development with clearly demarcated parts (for example, with more or less separated layers), and then build links between them.

Embryologist: During embryonic development, the typical structure of the brain begins to form due to the separation of more or less demarcated layers or levels, which is reflected in your diagrams. Then, separate groups of cells begin to form bundles of fibers that extend across the borders of the brain zones for fairly long distances.

The system can also start with the establishment of a huge number of links, and later delete some of them. A similar process is happening to us: in those times when our brain evolved, our ancestors had to adapt to thousands of different environmental conditions, now many reactions that were previously “good” turned into serious “mistakes” and we need to correct them by removing unnecessary connections.

Embryologist: Indeed, during embryonic development, more than half of the cells described above die off, barely reaching their goal. It seems that this process is a series of edits, during which various types of "errors" are corrected.

This process reflects the main limitation of evolution: it is dangerous to make changes to the old parts of the body, because many parts that evolved later depend on the work of the old systems. Consequently, at each new stage of evolution, we add various “patches” to the structures that have already been developed. This process led to the appearance of an incredibly complex brain, each part of which works in accordance with certain principles, each of which has quite a few exceptions. This complexity is reflected in human psychology, where every aspect of thinking can be partially explained by the terms of clear laws and principles of work, however, every law and principle has its own exceptions.

The same limitations arise when we try to improve the performance of a large system, such as an existing computer program. For its development, we are adding more and more patches and patches, instead of rewriting old components. Each specific "error". Which we can fix, in the end, can lead to an even greater number of other errors and make the system extremely cumbersome, which perhaps is happening to our mind now.

∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞∞

This chapter began with a presentation of several widely held views on what “ consciousness ” is and what it is. We came to the conclusion that people use this word to describe a huge number of mental processes that no one fully understands. The term "conscious" is quite useful in everyday life and seems almost indispensable for talking on a social and ethical level, because it keeps us from wanting to know what is in our mind. The same can be said about most other psychological words, such as understanding , emotion and feeling .

However, if we do not recognize the ambiguity of many-valued words used, we may fall into the trap, trying to clearly define what these words "mean". Then we found ourselves in a problem situation due to the lack of clear ideas about what our mind is and how its parts work. So, if we want to understand what the human mind is doing, we need to divide all mental processes into parts that we can analyze. The next chapter will try to explain how Joan's mind can do the work that is characteristic of the human mind.

For the translation, thanks to Stanislav Sukhanitsky. If you want to join and help with translations (write in a personal or email alexey.stacenko@gmail.com)

"The Table of Contents of The Emotion Machine"
Introduction
Chapter 6. COMMON SENSE [ eng ]
Chapter 7. Thinking [ eng ]
Chapter 8. Resourcefulness [ eng ]
Chapter 9. The Self [ eng ]

Ready translations

• "The Art of Doing Science and Engineering" by Richard Hamming
• Full translation of the book about building communities: "Social Architecture"
• "Hackers and Artists." As we all 13 years all rune Paul Graham translated
• Translation of the book “Skunk Works. Personal memoirs of my work in Lockheed "
• Translation of the book Andrew Una Passion for Machine Learning
• Course lectures "Startup". Peter Thiel. Stanford 2012

Current translations to which you can connect

• The Dream Machine: The History of the Computer Revolution.
• Translation of Richard Bartle's book “Designing Virtual Worlds”