Individual learning as evolutionary construction of behavior 1 Mikhail S. Burtsev Keldysh Institute of Applied Mathematics, 4 Miusskaya sq., Moscow RU-125047, Russia mbur@narod.ru http://www.keldysh.ru/pages/mrbur-web/ Abstract. Deterministic causation in individual learning is commonly accepted among neuroscientists and adaptive behavior scholars. At the same time there are natural and artificial evolutions which are based on alternative style of causation of adaptive change. This paper provides an attempt to formalize individual learning based on evolutionary causation and discuss some of its consequences on the modeling of individual learning. Introduction A mainstream of modern study of animal learning in the fields of neuroscience [1-4] and adaptive behavior [5-8] adopts explanatory scheme of Newtonian physics. The rules of synaptic plasticity in neuroscience and algorithms of ANN weights adjustment are deterministic like laws of motion in classical physics. In this framework process of individual learning is guided by some external or internal cause. A deterministic explanatory scheme is not unique in science. We have examples of theoretical approaches which are based on other types of causality such as quantum physics and evolutionary theory. Hence some questions concerned with causality in the theories of learning arise: • Is it possible to use any other type of causality than deterministic in the theories of learning? • What consequences will follow the acceptance of this other causality? • Can we build an animat based on nondeterministic causality which will demonstrate adaptive behavior? As it was mentioned above an evolutionary theory is based on non-deterministic causality. Evolutionary change is not determined by any given cause, but it is result of interaction of two processes – variation and selection. Such type of causation of change will be called below evolutionary causation. There are a number of attempts to use principles of evolution and thus evolutionary causation in the theories of learning. These are an evolutionary epistemology [9,10] in philosophy and theories of neuronal selection in neuroscience [11-13]. Up to date all these theories gain little acceptance as in the field of neuroscience as in the field of 1 draft, any critics and suggestions are welcome.