We present several methods of determining, not guessing, complex chemical reaction mechanisms and their functions. One method is based on the theory of correlation functions of measured time series of concentrations of chemical species; another is on measurements of temporal responses of concentrations to various perturbations of arbitrary magnitude; a third deals with the analysis of oscillatory systems; a fourth is on the use of genetic algorithms to determine functions of chemical reaction networks. All methods are applicable to chemical, biochemical, and biological reaction systems and to genetic networks and systems biology. The methods depend on the design of appropriate experiments on the whole system and corresponding theories for interpretation that lead to information on the causal chemical connectivity of species, on reaction pathways, on reaction mechanisms, on control centers in the system, and on functions of the system. The first three methods require no assumption of a model or hypothesis, nor extensive calculations, unlike the interpretation of measurements made on a gene network at only one time.