Reductionism presents itself with various aspects, such as the purpose of explaining properties of the macroscopic with those of the microscopic; to consider the existence of a true (possibly unique) level of description; the completeness of models and descriptions; the existence of objective, context-independent laws; the role of the observer as a generator of relative points of view; reducibility to the optimal scale where everything can be explained; and deductibility from principles. However, it is considered here that reductionism is not always clearly identifiable, but can also be delineated, in our case, in the disciplinary, technical, compositional, generic, and metaphorical use of the concept of systems. In this case the local disciplinary effectiveness of the concept of systems is considered sufficient, not necessarily implying interdisciplinary extensions and coherences.
In this regard, we recall the concepts of multidisciplinary, interdisciplinary, and trans-disciplinary levels, when even simultaneous different gradations between them may occur for different multiple approaches to real cases.
In a nutshell, the multidisciplinary aspect can be understood as when problems are faced side by side with different, disciplinary, specialised, and well-distinguishable approaches that must, however, be coordinated. An example is given by a company that requires engineering, design, accountancy, marketing, and human resources skills, all to be coordinated by the management. Still, everyone does his/her job. It is matter of separated competences and specialisations.
The interdisciplinary aspect can be understood as when problems and solutions in one discipline can be used in another, for example through the different meanings given to the variables of a model. For instance, the well-known Lotka-Volterra model was originally introduced in 1920 by Alfred Lotka as a model for oscillating chemical reactions. Later it was applied by Vito Volterra to predator-prey interactions. Other ulterior applications relate to populations dynamics and economy.
The trans-disciplinary aspect can be understood as the study of systemic properties per se, without reference to specific phenomena. Examples include openness, emergence, coherence, and resilience. The applications of these properties take place in their turn in disciplinary contexts, e.g., resilience in social systems and emergence in biology, with eventual outcomes and other disciplinary applications. A further example is how to keep, induce, or even destroy coherence within a population of interacting agents, e.g., Brownian motion.
In this context local, interacting systemic solutions may transform into new problems (often hastily considered as side effects).
3.1. Systems and Non-Systems
An interesting case considering the problem of using (not confusing) multiple systemic and non-systemic approaches is given by the DYnamic uSAge of Models (DYSAM) when the same problem or phenomenon can be represented in different ways and modelled using different approaches ([
3], pp. 201–204; [
5], pp. 64–85) that is, in multi- and interdisciplinary ways. This is the situation when the system to be studied is so complex, e.g., constituted by processes of emergence, that it is impossible, in principle, to fully describe it by using either a single model or a fixed sequence of models.
The conceptual background of DYSAM ([
5], pp. 64–70) includes the well-known Bayesian method, statistical approaches based on a “continuous exploration” of the events occurring within the environment under study. The method is based on the well-known Bayes’ theorem [
17,
18], named after the reverend Thomas Bayes (1702–1761). The background in physics includes the well-known Uncertainty Principles [
19]. One example is given by the approach introduced in 1927 by Werner Heisenberg (1901–1976) [
20], in which the measurement of homologous components, such as position and momentum (the product of the mass of an object and its speed), the search for increasing accuracy in knowing the value of one variable correspondingly involves reduction in knowing the value of the other variable. Another example is given by the Complementarity Principle introduced by Neils Bohr (1885–1962) in 1928, in which the corpuscular and wave aspects of a physical phenomenon never occur simultaneously. Moreover, so-called Ensemble Learning using machine learning techniques and algorithms [
21] and so-called Evolutionary Game Theory are parts of this background [
22,
23,
24].
When dealing with complex systems such as collective behaviours, we may simultaneously use models considering, for instance, metrical, topological, and energetic aspects. In the same way, problems may have different non-equivalent aspects to be represented and approached in different ways. For instance, a social problem may simultaneously be political, economic, sociological, and related to defence and security; a business problem may simultaneously be financial, organisational, managerial, and related to marketing, delivery, production, and warehousing processes. Moreover, a problem may have geometric or analytical representations.
DYSAM helps to consider how we may use approaches having different systemic natures, for instance, based on first or second Systemics or having a reductionist, non-systemic nature.
We consider here the issue of the coexistence between classical reductionism and its new form as the non-systemic usage of systems, local non-interacting systems, and—as we may say—systems without Systemics, as for cases summarised in
Table 1.
Systemics has a general, global semantic nature and trans-disciplinary, non-prescriptive, perhaps methodological implications, yet it is contextual, conceptually irreducible to the local, partial, and separated. It is Systemics that allows overall coherences and avoids misalignments with the consequent conceptual and phenomenological negative or unwanted effects.
This is a matter of recognising and overcoming possible partial and temporary local effectiveness. It is a matter of different modalities and approaches used simultaneously, such as in DYSAM. It should be discussed how such coexistence cannot suffer from contradiction, in a dialectical way [
25], similarly to when we accept the classic and non-classic nature of the matter that is revealed during some phase transitions.
DYSAM introduced the need to use different models of different natures to deal with the impossibility of zipping all the features of complex systems into a single model, while also considering the issue of theoretical incompleteness [
26,
27].
We consider the case of phenomena with multiple natures such as partial, not long-range, and local systemic natures, as seen in the quasi-systems mentioned below.
The structural dynamics of complexity should be approached not only in pragmatic way that is tolerant of multiple differences, such as with DYSAM, but also by using a more structured methodology and model suitable for the scenario underlined above. In this regard, we mention:
- (a)
A methodological approach named Logical Openness when a zipped, complete, and explicit model of the system and its interaction with the environment is conceptually not possible ([
3], pp. 47–51), [
28] because of the theoretical incompleteness [
29], varieties of interactions, and structures involved. For clarity, we specify that a model is considered as logical closed when a formal description of the relationships between all state variables is available in the model’s equations and a complete and explicit description of system-environment interactions is possible and available. This is not the case, for instance, when dealing with collective behaviours ecosystems and processes of learning;
- (b)
The more formal Meta-Structure approach, where the meta-structure project considers mesoscopic representations [
30,
31] and related properties of large varieties of structurally different interactions which are analytically intractable and impossible to represent in explicit ways [
32,
33,
34];
- (c)
The concept of quasi-systems mentioned in
Section 3. In short, a quasi-system is a system that possesses properties in partial, non-regularly changing ways; having partial or temporal inhomogeneities in its system status; and inhomogeneous possession or emergent acquisition of systemic properties, such as in the case of meta-stability, which is the ability of the system to maintain or switch between states in response to small fluctuations. In short, a system is not always a system and structurally not always the same system. The quasiness of complexity occurs when systemic aspects are predominant only and there is the need to methodologically deal with such aspects.
As mentioned above, we distinguish between non-systemic usages of systems, for instance in non-systemic, non-interacting environments or neglecting their natures of sub-systems or part of coherent multiple systems in the process of emergence, and simultaneous usages of systems and non-systems, as in DYSAM. In the first case it is a matter of misunderstanding the natures of systems, while in the second case it is a matter of multiple usages. As considered by quasi-systems, the coherence of multiple systems in processes of emergence may not be absolute. However, quasi-systems tolerate dynamical changes of states within sequences of supposed multiple systems, when there is temporary, partial, local loss of coherence, and the status of the system is then recovered by processes of resilience and compensation, the interchangeability of entities playing the same roles at different times, the equivalences among roles, and balancing. This is the case when, for instance, long-range correlation is not always 100% valid, ergodicity is partial, and so on. Depending on the nature of the phenomena, this tolerance has a threshold of admissibility, beyond which there are increasing non-restorable desegregations.
The aim here is to consider explicit and implicit aspects of coexistence, transformation, and use side by side with combinations of reductionism as local, disciplinary usages of systems, and non-complex and complex system representations. The interest is to at least recognise this situation.
This coexistence should certainly not only be recognised with the aim of developing processes considered to overcome reductionism, but with the aim of reformulating of problems, allowing new approaches and theorisations.
Is this a bionomic situation destined to be maintained, without having to necessarily or ideologically resolve the situation in the final adoption of Systemics?
It recalls the combination of constructivism and objectivism, of which gradualness is often used, rather than definitive overshoots. Pragmatically speaking, from time to time some objectivism is necessary to simplify, fix the ideas, and establish a starting point, while constructivism helps researchers to avoid getting wrapped up in the same simplified, initial approaches.
3.2. Cases and Examples
We will now consider cases as examples.
First of all, we need to distinguish between the properties possessed, acquired as consequences, and systemic properties continuously acquired.
Examples of properties possessed include age, spatial position, and weight.
Examples of properties acquired as consequences include properties of food after cooking, properties of buildings after renovations, new colours obtained from mixing primary colours, e.g., red-green-blue, and properties of objects after their forging such as knife sharpening, welding, and casting. In the latter cases we speak of results.
In the case of systemic properties, as introduced in the literature [
1], they occur due to the interaction (one’s behaviour depends on another’s behaviour) among components. There are two main kinds of systemic properties (related to the first or second Systemics, see
Section 1) continuously acquired through the interaction among components.
Examples of non-systemic usages of systems include non-comprehensive health policies, not considering, for example, side effects, unrelated to other aspects such as waste treatment, pollution, and natural-artificial food quality; or the systemic properties of cars without considering the emergence of traffic and urbanistic plans. Furthermore, in network representations [
35,
36,
37] the nodes may be systems of different natures and even non-systems, such as detectors, corporate departments, stores, and financial positions.
In
Table 1 we list three main different cases of reductionist usages of systems, where mixed and graduated cases are possible when occurring as a methodology and not as confusions and misunderstandings.
As an example of reductionist, simplified usage of concepts of pre-complexity Systemics for phenomena with complex natures, we mention how the problems of the current post-industrial society are complex in nature and require appropriate approaches. Often these issues are still addressed with pre-complexity approaches [
38].