Development as an open-systemic phenomenon involves feed-forward processes that guide the organism to face always uncertain future states of its relations with the environment. This creates a major conceptual problem for science--and a practical one for education--its theoretical terms need to capture both the certain and uncertain aspects of the developing system. Psychological theory has failed to model developmental processes since it has been built on static ontology of being, rather than on the epistemology of becoming. This contrast is accentuated by the use of transitivity relations in psychological models which are axiomatically closed to the open-endedness of the future in conjunction with the uniqueness of the past-to-present trajectory. Flexibility of developing systems is guaranteed by intransitivity cycles that set up conditions for their own rupture. Development as multi-level interdependent process requires breaking of intransitivity cycles and is organized as an autocatalytic process. Novelty is the key feature in all developing systems, and our formal models in psychology need to open the process of emergence for empirical and theoretical inquiry. Implications of such theoretical reconstruction for educational practices will be discussed--demonstrating that (a) education is necessarily open-ended and constructive process, and (b) outcomes of educational efforts cannot be predicted--while setting the direction towards educational goals is the key to any outcomes. (Contains 7 figures and 2 notes.) [Abstract presented in both English and French.]