There is a question: “What is the self?” This question is multi-faceted, and depending on the perspective, numerous answers can be contemplated.
From the perspective of what kind of logic would make it easy for intelligence to recognize self and non-self, I embarked on finding an answer to this question.
From this viewpoint, I concluded that the part one can control is the foundation of self-recognition. In previous articles, I have applied this perspective to the process of self-recognition in toddlers and considered how it can be implemented as an algorithm (refer to articles 1,2 and 3).
In this article, I will briefly review the idea that the controllable part is the self. Then, I will explain how the range of self-recognition is expanded using the concepts of control extension, protection, and ownership. Furthermore, I will touch upon how the concept of self can be applied and understood not only in intelligence but also in living organisms.
Self as Predictable Chaos
Intelligence observes the world and learns the patterns, laws, and mechanisms present. It then uses pattern recognition, application of laws, and simulation of mechanisms to predict the unknown parts of the world.
At this time, there remain parts that are difficult to predict due to lack of learning or inability to grasp patterns and laws. If we call the predictable parts “order”, the unpredictable parts become chaos.
We have no choice but to fill in that chaos with assumptions and imaginations to predict the world.
Intelligence then recognizes parts of the world that, despite not being able to grasp patterns or laws, somehow always align predictions with reality.
This part is the self. In other words, it’s like one’s hands and feet, the controllable part. This part is unknown and does not follow patterns or laws, making it chaotic, but on the other hand, its future can always be perfectly predicted.
In other words, it’s predictable chaos. I believe that this is the foundation of self-recognition.
This mechanism of self-recognition can be easily realized if intelligence has the ability to learn and predict the world. Since all that needs to be done is to identify the part that is chaotic but always predictably accurate, it can be achieved with a simple algorithm.
Extension of Control
Once intelligence recognizes the part it can control as the self, it seeks more controllable parts.
It moves the shoulder, then the elbow, and the five fingers. By combining these, it learns to move the hand as desired. It learns to move the thigh, knee, ankle, and then to walk.
Furthermore, it will learn to use tools. By mastering tools like sticks, scissors, bicycles, and smartphones, it can increase what can be controlled and its range.
Moreover, the skill and ability to control these tools is recognized as the self, and while in control, the tools are recognized as the self, just like hands and feet.
When playing computer games, the character being operated is recognized as the self. That’s why we feel anxious when attacked by enemies and unintentionally shout when falling into a pit.
What is Needed for Control
Recognizing something that can be controlled as oneself is thought to be the origin of self-awareness. From there, one’s perspective of oneself is believed to expand.
First, if our arms and legs are injured or lost, there’s a risk that we can’t control them. Therefore, we have a desire to protect what we can control.
Feeling pain is a function innate to us through evolution. Even if there’s a part that doesn’t feel pain, if it’s controllable, we will want to protect it as if it feels pain. This is because if we don’t, a part of ourselves might be lost.
Hence, intelligence will also recognize itself in terms of what to protect, aside from what to control.
Moreover, what needs protection is not limited to directly controllable parts. Resources like the energy needed for those controlled parts to operate, and freedoms and rights, are also things we aim to protect. If they are lost, we might lose control just as if something controllable was damaged.
This is the concept of ownership. We own resources like energy and money, and rights including freedom, and we aim to retain or even increase them. This is because ownership affects our ability and scope of control.
Therefore, intelligence also has a perspective of recognizing owned objects as oneself, separate from control and protection.
While control, protection, and ownership often overlap, they don’t align perfectly.
Each perspective has different subjects, but generally, we recognize ourselves by integrating these views.
Maintaining and Expanding Control Capability
Intelligence recognizes itself to expand its control capability. We increase our possessions, protect them, and use them to expand the range, precision, and power of our control.
The means might be rational, but the purpose itself is instinctual. Higher control capability means higher survival capability and a better chance to protect our offspring.
If traits that can enhance control capability have survived, it’s not surprising that our instincts incorporate the desire to expand control capability, including the perspectives of protection and ownership.
However, just because it’s instinctual doesn’t mean it’s the right thing to do. In many situations, we need to rationally suppress or control our instincts.
The same goes for the desire to expand control capability. We should leverage this desire up to a point, but overreaching shouldn’t be condoned.
It’s crucial to understand the concept of “knowing when it’s enough” rationally.
Self in Organisms
Like intelligence, organisms distinguish between self and external world. Even primitive single-celled organisms physically distinguish themselves from the external world with a cell membrane.
Assuming the same model applies, even organisms might have controllable parts at the core of their identity.
Maintaining or possibly expanding that control capability is crucial for organisms. Therefore, one could argue that a cell membrane exists in single-celled organisms to protect their controllable capabilities. Not only controllable parts, but also resources like energy and chemicals can be seen as protected by the cell membrane.
And organisms are believed to have evolved towards expanding their control capabilities. Multicellular organisms likely have a larger scope, precision, and power of control than single-celled ones. Similarly, vertebrates probably have more control than invertebrates.
Also, from a protective perspective, organisms have evolved significantly. It’s not just about defending with a cell membrane. Organisms need to intake nutrients and oxygen.
They have mechanisms to ensure they don’t intake harmful substances or can expel them if they do.
The mechanisms of detecting abnormalities through taste or smell, or repelling intruders through immunity, are classic examples. Especially in immunity, there’s the challenge of distinguishing self from foreign substances. This defensive evolution is a critical aspect of organisms’ evolution.
In Conclusion
This article explained the idea that the origin of self-recognition in intelligence lies in controllable parts. It also touched upon the idea that controllable parts can be extended, and parts encapsulated by concepts like body protection and resource ownership are also recognized as the self.
Furthermore, we looked at the distinction between the external world and self from a biological viewpoint. By evolving protective functions like the cell membrane to differentiate from the external world, it’s believed that organisms have aimed to enhance their self — the range, precision, and power of control.
Presumably, when considering making artificial intelligence recognize itself, starting from the perspective of controllable parts might lead to a rational self-recognition mechanism.
References
article1
article2
article3
