Self-developing systems are in the center of attention of post-non-classical scientific rationality (Stepin, 2003).The paradigm "subject – self-developing reflexive-active system (environment)" (Lepskiy, 2010) becomes a key paradigm of control and cybernetics. It is important to note that the environment is considered to be the meta-subject. As a result the paradigm can be presented as "subject – meta-subject".

A self-developing reflexive-active environment is a metasubject, which possesses invariant similar to the properties of subjects: purposefulness (activity), reflexivity, communicativeness, sociality, ability to develop, etc. Such an environment has integrity that essentially distinguishes it from networks. This is an interaction of active elements, organized in a special way. Active elements can be created on the basis of natural intelligence (the personality, group, etc.), on the basis of artificial intelligence (agents) and also on the basis of integration of natural and artificial intelligence.

The organization of interaction of active elements among themselves and with the environment in general is defined by the system of values, principles, ontologies (maintenance, support, development, designing, providing innovations), criteria (efficiency, safety, development, satisfaction) and also by the specialized subject-focused information platform (Lepskiy, 2010; 2015).

The idea of self-developing reflexive-active environments was created under the influence of the following inter-disciplinary ideas and concepts. Philosophy, sociology and psychology have given us the ideas of post-non-classical scientific rationality, which integrates concepts of various scientific schools (Stepin, 2005), ideas of noosphere (Vernadsky, 2007), the concept of the society as a social system (Luhmann, 1982), principles of the Russian psychology (Leontiev, 1978; Vygotsky, 1981; Rubinshteyn, 1997), studies of the Russian methodologists (Shchedrovitsky, 2002) and etc.

Cybernetics has given us an idea of second-order cybernetics by Heinz von Foerster (Foerster, 1974), Stafford Beer‘s models (Beer, 1981), W. R. Ashby principle of complexity in control (Ashby, 1956), the reflexive models of Vladimir Lefebvre (Lefebvre, 1967, 1982), a synthesis of representations of cybernetics and its development by Stuart Umpleby (Umpleby, 2014), ideas of Valentin Turchin about metasystem transition and concepts of the future of cybernetics (Turchin, 1977), etc.

The model of organization of self-developing reflexive-active environments, described in this abstract, will allow us to solve a number of current scientific and practical problems (Lepskiy, 1998, 2010; 2015). It will:

– support of processes of identification of the society (project identification gets the leading role);

– assemble of the subjects of development into metasubjects, help to consolidate state, business and society actors on the basis of shared interests, stimulate and support development of the civil society;

– improve mechanisms of democracy on the basis of convergence of direct and representative democracy;

– overcome market egoism through transition to a harmony of subjects of development;

– create opportunities for all subjects in the field of social activity and mobility;

– stimulate and support the formation process of a new type of elite – an elite of development, and create necessary provisions to include it into the processes of strategic control;

– help to design complexity problem solution in the processes of social systems control (Eshbi principle);

– create development conditions of new socially oriented economic mechanisms of development;

– create effective mechanisms of innovative development;

– decrease social tension, prevent conflicts, increase security with technologies of the operated chaos, "orange revolutions" and other destructive influences;

– initiate transition processes from technogenic to a socio-humanistic civilization, etc.

This paradigm can be applied for the organization of active knowledge, for reflexive mechanisms of management of complexity, etc.

Formation of this paradigm is inseparably linked with formation of the subject focused approach (Lepskiy, 1998).

The necessity of complex use of natural-science and humanitarian fields of knowledge generates high methodological complexity. The solution of this problem is possible upon transition from an interdisciplinary to a transdisciplinary approach. An exit out of limits of separate disciplines and the conceptual directions with involvement of external experts is necessary.

Now formation of scientifically ensuring control and the use of cybernetics in the context of post-non-classical rationality has begun (Lepskiy, 2015). In our opinion, an issue of formation of post-non-classical third-order cybernetics is realized. Thus, the main thesis would be from “observed systems” to "observing systems" and to "self-developing reflexive-active environments". From the paradigm "subject – object" to the paradigm "subject – subject" and further to the paradigm "subject – metasubject".

Transition in control to the paradigm "subject – metasubject" led to formation of new types of control. Control through self-developing environments becomes dominating. Control of "the soft force", control of chaos, control of complexity, control via "mechanisms of functioning of the environment", control "via mechanisms of assembly of subjects" and many other types of control.

The philosophical and methodological analysis of cybernetics evolution proved its connection with the development of scientific rationality (classical, non-classical, post-non-classical). The classical scientific rationality is similar first-order cybernetics. The non-classical scientific rationality is connected with the second-order cybernetics. The cybernetics of self-developing reflexive-active environments (third-order cybernetics) corresponds to the post-non-classical scientific rationality.

The analysis of cybernetics evolution in the context of development of scientific rationality allows us to define specific traits of the second and third order cybernetics: basic philosophical approaches, basic paradigms, basic objects of control, the dominating types of activity, basic scientific approaches, basic areas of knowledge, basic types of control, basic models, basic mechanisms and technologies, basic ideas of knowledge, the dominating ethical regulators.

In post-non-classical representation cybernetics of the first, second and third order are integrated as a unity, as a uniform area of knowledge. This is achieved through the system of ontologies of the organization of self-developing reflexive-active environments.

It is important to note in this article, that the self-developing reflexive-active environment is influenced by cross-disciplinary ideas and concepts of philosophy, methodology, sociology, psychology, cybernetics, etc. Self-developing reflexive-active environment is a metasubject, which possesses invariant similar properties of subjects: purposefulness (activity), reflexivity, communicativeness, sociality, ability to develop, etc. Such environment has integrity that essentially distinguishes it from networks. This is an interaction of active elements, organized in a special way. Active elements can be created on the basis of natural intelligence (the personality, group, etc.), on the basis of artificial intelligence (agents) and also on the basis of integration of natural and artificial intelligence. The organization of interaction of active elements among themselves and with the environment in general is defined by the system of values, principles, ontologies, criteria and also by the specialized subject focused information platform (Lepskiy, 2010; 2015).