The description of an image with pixels allows e.g. the exact reproduction of each single point whose size is determined only by the number of pixels. However, a much higher number of pixels or coefficients are needed than for a fractal description which results from improving an initial assumption stepwise from generation to generation, but is unable to accurately render individual pixels.
States can be determined rationally. In philosophical terms, it means the investigation of being, which can be done with the utmost accuracy in the framework of statics while renouncing dynamics. The addition of dynamics inevitably leads to a renunciation of the highest accuracy, but allows to capture the development of processes what constitutes the very essence of dynamism, modernity and learning. Determining and learning are thus fundamentally different categories, as humanists say, or different dimensions in scientific language.
Clearly conceiving this difference is of utmost importance in-between theory and practice. Pixel images represent states, and dynamics can only be created as an illusion by rapid succession of such images, which is the basis of television. Fractal images, on the other hand, can continue to develop at any time, which corresponds to learning processes. In principle, every single image can be derived by changing or specifying a single coefficient from the previous image with minimal loss of time. This consideration is not just valid for images, but for all dynamic processes and their description. Learning is generally the addition of a "little bit" in a subsequent step and thus should theoretically be more effective, the faster these steps take place. However, this can only partially be realized, because each step also means an expenditure of energy, which is not available indefinitely.
Life derives from automatons or immovable unicellulars. These first jump from one state to the next under energy change, which is described by quantum numbers or at low energies by genetic code. These states can in principle be grasped rationally. The transition from static to dynamic behavior is accomplished as a crucial step in biological evolution in nature through the transition from immobile crystals to mobile enzymes, enabling learning and effective development. This is the basis of the evolution of extremities, in the simplest case of flagella in bacteria.
The lower abdomen of higher living beings derives from unicellulars. So the first step following is the evolution of extremities, then of a head and finally the clear distinction of an upper body. These further developments of biological evolution can be understood as consequences of learning processes in the sense of Darwinism. The assumption of a statistical emergence of new states only by mutations could not explain the actual speed of evolution. Faster learning processes through transition from generation to generation are therefore crucial.
The trick of learning in the modern dynamic sense now consists in reducing the generation time, which must no longer be identical with the lifetime, but only will depend on the energy being available. Each individual learning step is thus "simply" an improvement on the previous "generation" in learning, which no longer necessarily has to do with birth and death and thus avoids the problem of destructive singularities. Learning becomes a practically continuous process, even though it is gradual. The individual steps need only go as fast as possible and with the lowest possible energy consumption, which is the basis of all competition.
Every single step of learning represents an initially infinitesimal small emergence. The higher the density of the involved components, the more this comes close to actual emergence. This is generally true, so equally for elemental particles, human creativity and astronomical super- or kilonovae.