The epistemological basis of the participatory action-research (PAR) methodology in the so-called 'paradigm of complexity' is plagued with misconceptions, imprecision and significant omissions. The appropriate and contextualised translation of concepts from the natural sciences for use in the social sciences is particularly necessary in qualitative-structural and participatory-dialectic methodological trends. This paper focuses on the concepts of 'entropy', 'complexity' and 'strategic action'. Based on a general notion of complexity relating to the consideration of uncertainty, innovation and the contextualisation of systems, performance-related aspects indicative of actions implied by this paradigm may be highlighted. Thus, it is argued that the conceptualisation of operations such as 'acting in order to know' and 'act knowing/know by acting', encompassed in the notion of 'strategic invention', complement and exceed the scope of typical planning operations and even self-planning ('knowing in order to act'). In short, these definitions are coherent with an ecosystemic perspective of social and natural reality, in which it is necessary to contextualise what is more or less complex in the world, in our knowledge of the world and in our practical actions when acting in the world. Consequently, participatory action-research methodologies should adopt this complex ecosystemic epistemological perspective, and be designed through a 'strategic invention' approach, in order to clarify the concepts imported from other scientific disciplines.
2008. Vol. 33, no 2, 153-177 p.