George Soros makes an important analytical contribution to understanding the concept of reflexivity in social science by explaining reflexivity in terms of how his cognitive and manipulative causal functions are connected to one another by a pair of feedback loops (Soros, 2013). Fallibility, reflexivity and the human uncertainty principle. Here I put aside the issue of how the natural sciences and social sciences are related, an issue he discusses, and focus on how his thinking applies in economics. I argue that (...) standard economics assumes a ‘classical’ view of the world in which knowledge and action are independent, but that we live in a complex reflexive world in which knowledge and action are interdependent. I argue that Soros's view provides a reflexivity critique of the efficient market hypothesis seen as depending on untenable claims about the nature of random phenomena and the nature of economic agents. Regarding the former, I develop this critique in terms of Cauchy distributions; regarding the latter I develop it in terms of rational expectations and rational addiction reasoning. (shrink)

In 1987, George Soros introduced his concepts of reflexivity and fallibility and has further developed and applied these concepts over subsequent decades. This paper attempts to build on Soros's framework, provide his concepts with a more precise definition, and put them in the context of recent thinking on complex adaptive systems. The paper proposes that systems can be classified along a ‘spectrum of complexity’ and that under specific conditions not only social systems but also natural and artificial systems can be (...) considered ‘complex reflexive.’ The epistemological challenges associated with scientifically understanding a phenomenon stem not from whether its domain is social, natural, or artificial, but where it falls along this spectrum. Reflexive systems present particular challenges; however, evolutionary model-dependent realism provides a bridge between Soros and Popper and a potential path forward for economics. (shrink)