The Generation of Semantics in Natural Language and the Formation of Brain Intelligence

The foundation of life phenomenon is the abilities of representation, memory and behavior of a life form originated from complex motion and interaction among a great number of microscopic matters. Human natural language can describe and interpret this matter world, and also support naturally the subjective initiative of humans. Not as a computer system, the brain neurons can link more effectively with each other in 3D physical space such as in which no word is stored repeatedly, and the neurons form the matter base of consciousness activity. It is known that the radial basis function representation principle is widely used to represent color, sound, smell and gustation, this paper attempts to use a quantitative analysis method to estimate with an unbiased way the number of single wavelength color attributes that can be represented by our brain. It is also known that some simple neurons in the primary cortex are sensitive to the direction of short line, this paper introduces the representation principle of single basis inclination which can represent the inclination attribute of shorter line, and our brain can further extract more topology attributes of a shape through introducing the concept of vision event which is the evolution foundation of cognition ability. By the mechanism of semantics expansion our brain can construct more complex semantics. This paper emphasizes that both the semantics in natural language and the original semantics only perceived by our brain exist in the neural fiber plexus nodes of brain, they are twins. We humans always attempt to consider this world as a whole, just as at a certain moment the neurons activated by the brain attention form an expanded semantics, and the shifting of attention forms a consciousness stream, so the semantics is furtherly expanded, and especially it is our behavior to have shaped the complex consciousness stream structure. This paper also elucidates the principle of matter primacy, and through analyzing the development process of human intelligence, we have found that the subjective initiative in human behavior makes us distinguishing from animals. In this paper, we also explain that why the brain consumes less power and the limitations of human intelligence.  Here, following the essence of intelligence, through developing a kind of computer language and the corresponding language parser(“analysisformer”) to support a Digital-Twin system, we can dynamically establish a large vector database with practical training and simulation, then through these numerous examples of behavior and reinforcement learning method we can map the intelligence principle into an intelligence automation system such as a humanoid robot and some kind of dynamic system that requires real-time optimization.