We intend to uncover generative principles for complex, biological systems, looking the reflections as well as the analogs of decision making property in quantum physics: measurement, self-interaction of the electron, Berry phase and quantum anomalies. We assume that classical analogs of the mentioned phenomena could be related to the evolvability, growing of complexity and decision making in biological systems. The reflection is a map (coarse graining) from microscopic motions to a macroscopic scale that relates with a free-energy cost and is often accompanied by the emergence of order-parameters. In this context we identify the self-reflection phenomenon, which is exemplified by cognition, information transfer near the error threshold, and tightly related evolution-ecology phenomena. We propose that complex systems that have similar reflection structure are to be described by similar mathematical tools including stochastic (information) thermodynamics and the large deviation theory. We introduce the concept of complexity engine: the group of two (or more) autonomous features of complex systems that are in a partial conflict with each other. Analogues of this are wave-particle duality in quantum mechanics and data-program duality in digital life. We formulate a fundamental problem: does the three-dimensional space provide a complexity engine for the emergence of life?