When in the last two decades of the 20th century new ideas on complexity had started to diffuse in social sciences, the economic thought became a typical area of interdisciplinary research in which the concepts of chaos, nonlinear movement, turbulence, and instability became best to apply. The article looks at how the approach of complexity has been implemented in the Santa Fe interdisciplinary project almost in it's first ten years between two major workshops on the economy as an evolving complex system. At first it concentrates on general issues of complexity discussed by natural and social scientists in Santa Fe Institute accepting all of them as certain bias for the epistemological breach. Then it turns to the several topics, which problematized the Santa Fe thought in particular and made it unique among the scholars who have adopted the new concepts proposed by complexity science. It also argues that the term of so called "restricted complexity" applied by some epistemologists to characterize a limited score of research topics in Santa Fe is far from being adequate. Although all developments were not equally successful, they did turn out to have enormous resonance in terms of extensive theoretical and practical pertinence and broad areas of application in social sciences even then the complexity approach was challenged by widely spread revision inspired by Niklas Luhmann legacy. Based on a body of work of Brian Arthur and John Holland and their colleagues from Stanford and Massachusetts where complexity has been understood in ways of system behavior, the article presents a complex adaptive system not as the opposite but as a source and confederate of certain order.