The central issue of this study is knowledge constructions. A double perspective penetrates investigations: constructions in man and in science.

The human perspective focuses on a psychological question of learning: the spontaneous acquisition of new knowledge in the individual. The scientific perspective focuses on philosophical premises for the formation of scientific knowledge. The guiding principle here is that the psychological and philosophical themes condition each other.

The three central chapters are organised in response to claims made by Kant against a science of the psyche, concerning its suggested lack of objective observation, measurement, mathematisation. I suggest a method, perform an empirical study accordingly, and supply a formal account of psychic processes.

The empirical investigation concerns subjects' inadequate constructions of a text editor, and their consequent reconstruction attempts. Verbal reports are taken and analysed within a problem solving paradigm. Results suggest that implicit reconstruction processes have to be introduced to account for verbal and key-stroke events. Ontological implications are discussed. Information processing accounts for insight are criticised. A constructivist approach to insight is offered.