Dynamic Systemic Learning

The concept of system dynamic learning assumes that there is no single solution for a task, but that there are always several possible solutions available in a kind of "solution space". The more information can be stored during learning, the more information can be accessed later when searching for a respective solution. Therefore, as many different estimated solution approaches as possible should be practised, and not only those that seem suitable from the outset.

Content

The concept of dynamic systemic learning

The concept of dynamic system learning assumes that there is no unique solution to a task, but that there are always several possible solutions available in a kind of "solution space". For example, there are many different ways to grab a glass and drink from it. Individual skills are stored in multidimensional "maps" of information that can be used to control various aspects of such movement. The more information stored in such a "map", the more information can be accessed when searching for a particular solution. In order to fill these maps as richly as possible, not only the solutions deemed appropriate should be put into practice. It can also be useful to try several different and even mismatched implementations. In this way, the dynamic system learning model differs significantly from the model-based approach. However , by intentionally trying even inappropriate execution forms, dynamic systemic learning also outperforms variable learning. These different approaches to learning are explained below, using skills sheets as an example.

Model-based learning

In model-based learning, an optimal solution (the green dot on the map) is always given, which should be mostly practiced. This model is mainly oriented towards the perfect result of the movement, so towards the correct form. Deviations from this pattern should be avoided (yellow), and larger deviations are considered mistakes during the learning process (red). Individual parts or creativity are thus considered as disturbance variables. At school, deviations from the prescribed letter form are often sanctioned and negatively evaluated. This naturally leads to a certain learning behavior, in which mistakes are avoided and a high degree of conscious control of movements is used.

Variable learning

During learning, the tasks are varied to some extent to provide more information to the learning system. Learning is seen as a repetitive problem-solving process where variation not only enhances the learning process itself but also adaptability. Inevitably, there are also more failures, but these deviations from the default values (yellow) are no longer necessarily considered negative or mistakes, but useful for learning. The degree of variation also depends on the student's proficiency level, which should not be under-challenged, but neither should it be over-challenged. The learning result of variable learning is significantly better, especially in terms of transfer, than in the case of learning by repetition.

Dynamic (or Differential) Systemic Learning

A rich learning environment with a permanent variation of tasks and task context tries to fill the maps with stored information as completely as possible (green dots). This could be, for example, writing fast, accurate, large, cursive (task context) different letters (variation). In principle, it does not matter whether or not these experiences are useful for solving the task. Of course, in order to be able to categorize information properly, accurate feedback on the success of that movement is required. If fluctuations and inaccuracies occur during learning, they are even intentionally amplified and replayed as a new task to promote the self-organization process in finding an individual solution. For example, if the pressure is too firm, the student must press harder until they learn what too firm pressure feels like.

The decision-making process in such a complex information system is explained on the basis of a model. If the maps are well populated, and later the system needs a suitable solution for a particular task, then some sort of interpolation of ALL the information stored in these maps calculates an intersection or centroid that represents the best solution at that time. For example, when you run in the forest and jump over a tree trunk, the motor program that corresponds to the size, distance, surface, etc. is selected. The better these maps are filled in general, the better and more solid the solution will be.

Thus, dynamic systemic learning differs fundamentally from a classical cognitive approach to learning. While the cognitive approach to learning largely predetermines the content and method of learning and provides only appropriate solutions, dynamic system learning is more about designing a rich learning environment where the learner can accumulate as many experiences as possible.

Numerous scientific studies have been able to impressively demonstrate that dynamic variable and systemic learning is more demanding in the learning phase, but achieves a significantly better result when transferred to everyday life. In principle, children's self-organized learning (for example, learning to walk) can also be considered as variable learning. Cognitive learning, on the other hand, is subject to the illusion of learning: during the learning phase, movement appears to succeed because it is performed under controlled conditions, slowly, and with a high degree of movement control. However, transfer to everyday life is unsatisfactory when the learner has to face new situations or when the movement has to be performed automatically. In the case of variable learning and dynamic system learning, many more errors are committed during the learning phase. However, these errors serve as an additional source of information and allow the automation of movements at an early stage, as less conscious control of movements is required. In transfer, variably learned movements are more adaptable and require much less working memory due to movement automation.of variable learning and dynamic system learning, many more errors are committed during the learning phase. However, these errors serve as an additional source of information and allow the automation of movements at an early stage, as less conscious control of movements is required. In transfer, variably learned movements are more adaptable and require much less working memory due to movement automation.

What does this mean for my teaching practice?

Even though constant practice during the practice phase seems to bring a higher success rate, the transfer result to other tasks is very limited. The learning system does not react to constant repetition, but rather to frequent new requests (see → Hippocampus). The task of the teacher in dynamic systemic practice is to design a rich learning environment that stimulates different and individual ways of solving problems.

Reflection question

Why does cognitive learning (learning by understanding) only lead to limited learning success?

Quiz

1) One principle of system dynamic learning is

A) the avoidance of mistakes
B) constant practice with repetition
C) repeated implicit experience

2) Why is variable practice more efficient than constant practice?

A) You have a better understanding of the task
B) You get more information through constant adaptation.
C) There is immediate learning success.