Структура Устав Основные направления деятельности Фонда Наши партнеры Для спонсоров Контакты Деятельность Фонда за период 2005 – 2009 г.г.
Чтения памяти Г.П. Щедровицкого Архив Г.П.Щедровицкого Издательские проекты Семинары Конференции Грантовый конкурс Публичные лекции Совместные проекты
Книжная витрина Где купить Список изданных книг Готовятся к изданию
Журналы Монографии, сборники Публикации Г.П. Щедровицкого Тексты участников ММК Тематический каталог Архив семинаров Архив Чтений памяти Г.П.Щедровицкого Архив грантового конкурса Съезды и конгрессы Статьи на иностранных языках Архив конференций
Биография Библиография О Г.П.Щедровицком Архив
История ММК Проблемные статьи об ММК и методологическом движении Современная ситуация Карта методологического сообщества Ссылки Персоналии
Последние новости Новости партнеров Объявления Архив новостей Архив нового на сайте

Toward system principles: General system theory and the alternative approach

In: Systems Research and Behavioral Science, 2004, 21, pp. 109-122.
Clarkson University, Potsdam, New York
Vitaly Dubrovsky


General System Theory (GST) sets its goal as unification of science, and its subject-matter as formulation of general system principles, or principles applicable to all systems. Unfortunately, no such principles have been formulated to date. This paper demonstrates that GST is incapable to formulate system principles due to its naturalistic methods and realist ontology that represents system as a real object, or thing. This paper suggests the alternative approach to development of system principles based on the assumption that the purpose of the system approach is not unification of science but dealing with complexity, and the corresponding method is analysis-synthesis. According to this approach, reality is neither systemic nor non-systemic, and only our methods of treating real objects and corresponding representations of them are either systemic or not. The paper asserts that system science cannot study real systems, because they do not exist. Instead it should study history of system reasoning as the process of development of system principles. As a preliminary step, by analyzing classical definitions of system, this paper identifies the principal system concepts and categorical oppositions that can be used as an initial framework for empirical studies of system reasoning. It also outlines methodology for such studies.

Keywords: system principles, system method, ontological picture, levels of thought.




Half a century ago, Ludwig von Bertalanffy formulated a new discipline, General System Theory, and defined its subject matter as "formulation and derivation of those principles which are valid for 'systems' in general" "whatever the nature of the component elements and the relations or 'forces' between them." (von Bertalanffy, 1968/1998, p. 32). General System Theory (GST) was a revolution. The purpose of this revolution was unification of science, and the essence of this revolution was interdisciplinarity. It produced a new type of scientific knowledge: interdisciplinary knowledge.

GST theorists believed that at least some isomorphisms, formulated on the basis of theoretical models of more than one scientific discipline, could be extended to other sciences, as well as the studies of new types of objects and, thus serve as general principles applicable to all systems. They also believed that such interdisciplinary models as machine, open system, living organism, self-referential system, and others, could be effectively applied to different areas of science, engineering, and practical interventions.

Although the success of GST in the unification of science is still under dispute (e.g. Checkland, 2000), it is clear that GST has failed to formulate general system principles, or principles applicable to all systems. Bertalanffy himself has failed to formulate a single principle that is applicable to all systems, save the principle of wholeness, or unity, that is actually a part of his definition of system. What he suggested as general system principles cannot be considered as such. For example:

 … A principle of relaxation oscillations occurs in physical systems as well as in many biological phenomena and certain models of population dynamics. A general theory of periodicities appears as a desideratum of various fields of science. Efforts therefore have to be made towards a development of principles such as those as of minimum action, conditions of stationary and periodic solutions (equilibria and rhythmic fluctuations), the existence of steady states and similar problems in a form of generalized with respect to physics and valid for systems in general (von Bertalanffy, 1969/1998, p. 80).

Ackoff (1964) criticized such principles for not being applicable to conceptual systems. He suggested that GST cannot claim generality if its principles are not applicable to conceptual systems and insisted that development of the principles applicable to both material and conceptual systems is the most promising direction, since it would fill the gap between natural and social sciences.

In early 2000, summarizing the on-line discussion on system principles, Tom Mandel (E-mail to SIG Chairs of February 10, 2000) wrote: "I think it is amazing that no one has come forward and said, 'Here, these are our principles!' (I thought that if I claimed there were no principles, someone would prove me wrong.) Are we really just now getting around to determining what our principles are?" If after more than fifty years of existence of our discipline, we still cannot address its subject matter, what is wrong: the definition of the subject matter, the methods we use, or both? Below, I will demonstrate that the latter is the case.




Following von Bertalanffy (1969/1998), I will unite under the umbrella of GST three different methods for achieving interdisciplinarity: (1) identification of isomorphisms common to systems studied by different sciences, (2) formal construction of theoretical models, and (3) systems research. The section below demonstrates that none of these methods, in principle, cannot produce system principles applicable to all systems (from now on in this paper, just "system principles").


The Identification of Isomorphisms Cannot Produce System Principles

According to von Bertalanffy (1969/1998) the purpose of the identification and derivation of isomorphisms, or structural similarities common to theoretical models of different scientific disciplines, is to develop system principles. It seems that the claims that system principles can be obtained by derivation of isomorphism can be explained by a confusion of what principle is. For example, in works of von Bertalanffy, the term “general system principle” has at least three different meanings. The first meaning is a derived isomorphism applicable to a certain type of systems:

In many cases, isomorphic laws hold for certain classes or subclasses of "systems," irrespective of the nature of the entities involved. There appear to exist general system laws which apply to any system of a certain type, irrespective of the particular properties of the system and of the elements involved (von Bertalanffy, 1969/1998, p. 37). 

The second meaning of the term is a derived isomorphism applicable to all systems:

Rather we ask for principles applying to systems in general, irrespective of whether they are of physical, biological or sociological nature. If we pose this question and conveniently define the concept of system, we find that models, principles, and laws exist which applied to generalized systems irrespective of their particular kind, elements, and "forces" involved. (von Bertalanffy, 1969/1998, p. 33). … the principles holding for any type of systems would have to be further developed (von Bertalanffy, 1969/1998, p. 80).

The third meaning of the term "general system principle" is the basis for derivation of isomorphisms: "The isomorphism found in different realms is based on general system principles of more or less well-developed 'general system theory'" (von Bertalanffy, 1969/1998, p. 84). Von Bertalanffy also calls such basis "general system properties": "A consequence of the existence of general system properties is the appearance of structural similarities or isomorphisms in different fields" (von Bertalanffy, 1955/1999, p. 33).

Only this third meaning corresponds to the common acceptation of the term “principle”. ”Webster Dictionary defines "principle" as "a fundamental, primary, or general law or truth from which others are derived" (similar definitions can be found in dictionaries of philosophy, e.g., Honderich, 1995 and Angeles, 1981). According to this definition, only the primary, non-derived general statements, that are applicable to all systems and serve as the basis for all derived statements, such as isomorphisms, could be properly called "system principles". This means that, as being derived, even isomorphisms that were applicable to all systems cannot be properly called "principles". In other words, being based on the assumed "system principles", or "system properties," derivation of isomorphisms in principle is incapable of producing non-derived system principles.

Formal Construction of Theoretical Models Cannot Produce System Principles

In opposition to the method of identification of isomorphisms, Ross Ashby (1958) suggested a "non-empirical" method that starts "at the other end". "Instead of studying first one system, then a second, then a third, and so on, it goes to the other extreme, considers the set of all conceivable systems and then reduces the set to a more reasonable size" (Ashby, 1958). Under a "conceivable system" Ashby meant any arbitrary set of variables, such as temperature and humidity in a room and a course of the dollar in Singapore (Ashby, 1958). According to Ashby, scientists always select their systems by discarding a large number of possible combinations of variables, mostly on the basis of their intuition. This description misrepresents what scientists actually do as well as Ashby’s own method. Actually Ashby, first, formally defined machine as anything that has "machine-like” behavior; second, referring to a wide interdisciplinary empirical material, he constructed a conceptual model of machine, with inputs, outputs, internal states, etc.; and third, he expressed relationships among these machine constituents using mathematical formalism. In exactly the same way, von Bertalanffy introduced his organism as an open system: "The organism is not a closed, but an open system. We term a system 'closed' if no material enters or leaves it; it is called 'open' if there is import and export of material" (1969/1998, p.121). Then, referring to interdisciplinary empirical knowledge in physics, biochemistry, and biology, he constructed his conceptual model of open system with "steady states", "dynamic equilibrium", "equifinality", etc. and, finally, expressed their relationships in a set of equations. This three-step method is typical for formal construction of scientific theoretical models. Being constructed such models are used as “cookie-cutters” to identify machines and open systems in the empirical world, thus determining appropriate variables and their relationships.

It should be noted that machine, open system, and similar concepts are considered general systems concepts because, unlike the traditional disciplinary science, they are aimed at the unification of science and constructed upon wide interdisciplinary empirical material. Although this approach does produce non-derived theoretical models that can be applied across scientific and engineering disciplines, these concepts are not applicable to all systems. For example, living systems cannot be adequately represented as machines (Bertalanffy, 1969; Rosen, 1991) and the concepts of inflow or outflow of material are not applicable to "immaterial" symbolic and conceptual systems that are, therefore, neither open nor closed.


Systems Research Cannot Produce System Principles 

Ackoff (1964) suggested that the method of identification of isomorphisms is ineffective due to GST's implicit assumption that the structure of reality is isomorphic to the structure of science, so reality can be divided into physical, biological, sociological, etc. He suggested an alternative "systems research" approach (at that time represented by operations research and system engineering) that deals with real systems "as they are" and that produces knowledge that cannot be separated according to the existing division of science. According to Ackoff, unification of science, or interdisciplinarity can be achieved by development of integral theoretical models in the process of system research conducted by interdisciplinary teams of specialists. These theoretical models go beyond generalization of the existing theories. Examples of such models are the theory of inventory control and queuing theory.

Responding to Ackoff's criticism of GST's approach to unification of science, Rapoport (1964) noticed that system phenomena cannot be studied in a "conceptual vacuum" and, therefore, some conceptual theoretical assumptions should be made to serve as the basis for formal models of systems research. In fact, discussing the theory of inventory control, Ackoff (1964) did employed “disciplinary” concepts of inventory management as underlying the theoretical constructs and corresponding mathematical formalism. This is reflected in his suggestion that a metabolic process in a living organism, hitting system, water supply of a certain geographical region, or any other object can be studied as an inventory control system as long as its variables behave “inventory-like”.

Anticipating the argument that the theories produced by systems research are limited in applicability to specific types of situations, i.e. are not general, Ackoff noticed that some isomorphic aspects of these theories are already identified and that higher-level generalizations are forthcoming.

Thus from the system principles perspective, systems research can be viewed as a combination of the two methods discussed above: (1) formal construction of theoretical models upon empirical material that does not belong to a traditional scientific discipline followed by (2) identification of the isomorphisms common to these models. This combination has nothing to offer in addition to these methods taken separately: the isomorphisms are still identified by generalization of theoretical models with limited application, the fact that the models do not belong to the traditional scientific disciplines does not make any difference. Therefore, none of the three OTC methods, identification of isomorphisms, formal construction of theoretical models, and system research, in principle, can produce system principles applicable to all systems.

Finally, GST cannot formulate general system principles using a "convenient definition of system", because it does not have one. Our widely circulated definitions of system as "unity of opposites", "organized wholes", "complex of elements standing in interrelation (or interaction)", "the relationship of elements taken as a whole", etc. proved to be as fruitless as the GST methods discussed above. The problem of method means that GST does not have a method that addresses its subject matter: formulation of principles applicable to all systems.





The Realist View of Systems in GST

Scientific disciplines represent their subject matter in ontological pictures, or, as von Bertalanffy called them, "schematized pictures of reality". Being realists, natural scientists believe that ontological picture of their discipline depicts the reality as it is, independent of their biases, research methods, etc. For example, when physics was a dominant discipline, the belief that the ontological picture of physics represented the reality led to the ideology of reductionism. GST emerged in opposition to reductionism, acknowledging that biological, behavioral, and sociological ontological pictures are irreducible to the ontology of physics and to one another, because they correspond to the different levels of reality itself. GST states that the unifying aspect of reality is system (von Bertalanffy 1969/1998, pp. 49, 55), and its ontological picture is "the model of the world as a great organization" (von Bertalanffy, 1969/1998, pp. 48-49).

All general system theorists share the realist view of systems. Some of them believe that the world is filled with real physical, chemical, biological, and sociological systems and the system isomorphisms are real features of these systems. Others believe that open and closed systems, machines, inventory control, and queuing systems are the models of real “systems as they are”. But all of them believe that the reality itself is a hierarchy of systems with the universe at the top and the elementary particles at the bottom. This realist bias determines the ontological picture of GST.


Concept of System in GST: Unity, Parts, and Relationship

The most cited von Bertalanffy’s definition of system as "set of elements standing in interrelationship" (1968/1998, p.55; also p. 38) should not be taken in isolation from another part of his definition: "If we are speaking of 'systems', we mean 'wholes' or 'unities'" (p. 66) that are more that the sum of their parts (p.55). Arguing against Russell's (1948) rejection of this "organic" view of system, von Bertalanffy explains the relationship between the two parts of the definition of system: "the laws governing behavior of the parts can be stated only by considering the place of the parts in the whole" (von Bertalanffy, 1968/1998, p. 67).

Thus out of three principal constituents of definition of system, Unity (also called "consistent whole," "complex whole," "wholeness", "synergy", etc), Parts (also called "elements," "constituents," "components," etc.) and "Relationship" (also called “interrelationship”, "interactions," "structure," and "organization") GST emphasizes Unity. It views Unity as a new ontological addition to the traditional scientific analysis: “Systems, of course, have been studied for centuries, but something new has been added. … The tendency to study system as an entity rather than as a conglomeration of parts…" (Ackoff, 1959).

So, in addition to being a "set of elements standing in interrelationship", system is also a separate ontological entity--Unity. In GST, the relationship between Unity and Parts in Relationship is explained by the doctrine of emergence of Unity out of Parts and their Relationship:

The characteristics of the complex, therefore, appear as "new" or "emergent"...We can also say: While we can conceive of a sum being composed gradually, a system as total of parts with its interrelations has to be conceived of as being composed instantly" (Von Bertalanffy, 1969/1998, p.55).

According to the doctrine of such intrinsic emergence of Unity from Parts and their Relationship, once emerged, Unity coexists with Parts and Relationship as a separate ontological entity. Corresponding traditional ontological picture of system (Figure 1) depicts Unity as a large circle, drawn around the Parts (small circles) and their Relationship (lines connecting small circles). This ontological addition of Unity, makes a system more than the sum of its Parts and Relationship and distinguishes it from Russel’s "physical system" that is equal to the sum of its parts (and relationship).


(Figure 1 is about here)


Problems with the Concept of System in GST


The paradox of emergence

A goal of the doctrine of emergence is to explain how Unity (whole) can be more than the sum of its parts. This doctrine would be banal and trivial if one would interpret it as the statement that Unity is more than the sum of its Parts because of Relationship. Such interpretation would make the concept of emergence of Unity needless, and, obviously, it was not the intent of its authors. The doctrine of emergence makes a much stronger statement that Unity emerging out of Relationship of Parts is something ontologically “new”, “additional to”, and different from Parts and their Relationship; it can be neither predicted from the examination of the constituent Parts and Relationship (Angeles, 1981) nor reduced to them (Honderich, 1995).

The epistemological interpretation of paradox of the emergence can be formulated in the following way: Because the emergent Unity is always new and unpredictable from Parts and Relationship, the only way to determine it is to have it already emerged and known, and, therefore, in actual investigation it is never new or unpredictable. It is not surprising, that we usually introduce systems by referring to its Unity.

In the ontological interpretation, the emergence paradox can be formulated as a contradictory status of existence of Unity in the diachronic and synchronic dimensions. This contradiction can be best explained using the “child analogy”. On one hand, in the diachrony, Unity emerges from Parts and Relationship similar to a child that “emerges” from parents and their relationship. It is a new additional entity, unpredictable and irreducible to its “parental” Parts and Relationship, i.e. independently existing along with the “parents” (the large circle in Figure 1). On the other hand, in the synchrony, while disappearance of the parents would not necessarily make child disappear, Unity would cease to exist with disappearance of the Parts or Relationship. Thus according to the emergence theory, Unity emerges as an entity existing independently of Parts and Relationship, but then it exists only as their epiphenomenon, dependent on the existence of Parts and Relationship. Later we will see Emanuel Kant’s interpretation of Unity that avoids the emergence paradox.

The paradox of system environment

The paradox of emergence leads, to false interpretations of system. One such interpretation is that system has Environment (von Bertalanffy, 1969/1998). Since the emerged Unity is usually represented as a large circle that encircles “parental” Parts and Relationship, they are interpreted as being inside the System. Since "internal-external" is a relative opposition, i.e. one opposite cannot be thought without the other (Aristotle Categories, Metaphysics; Hegel Encyclopedia of Philosophical Sciences, Vol. 1), what is outside of the System also requires interpretation. The usual interpretation is that outside of the System is its Environment; and the relationship between System and its Environment is usually characterized as Interaction. Because "interaction" means reciprocal action or influence of two or more entities (e.g. Webster Dictionary), Interaction of System and its Environment must also be ontologically represented as in Figure 2 below. In other words, the statement that system interacts with its environment assumes a combination of two incompatible ontological pictures. The first ontology depicts system as being inside of its environment (Figure 1); the second depicts it as being outside of Environment (Figure 2). In other word, at the same time, system is inside and outside of its environment.


(Figure 2 is about here)


Resolution of the above paradox requires reconsideration of many system concepts, including Unity, emergence, and environment, which will lead to reinterpretation of many other system concepts, such as subsystem, supra-system, open and closed systems, and the like. The concept of the interaction of organism with its environment has exactly the same problems. B.F. Skinner (1974), who was very sensitive to philosophical and logical problems, avoided the paradox of environment by using only the ontology of interaction (Figure2). He declared that organism is not a system, but a simple "locus of behavior," and refused to consider any internal mechanisms of behavior, including neural and cognitive information processes.



The problem of subject matter of GST is reflected in the paradoxical ontological picture of system. It demonstrates deficiency of the realist view of systems as real. This view leads to reification of the concepts of Unity and Parts and interpretation of their logical Relationship as natural processes of interaction, as well as mystical misinterpretation of logical status of Unity as an emerging real thing. The latter, in turn, leads to impotent methods of formulating system principles. If systems are real, then the way to derive generalized knowledge of systems is to empirically study different real systems, construct theories upon the empirical knowledge, and try to derive isomorphisms common to all real systems. At the same time, as we have seen above, such derivation cannot produce system principles and itself should be based on already existing non-derived system principles.

But if systems are not real, then what is their ontological status? Ackoff (1960) suggested that real physical or conceptual systems could be treated either as systems or as non-systems. If the analyst represents the behavior of a system as a product of the interaction of its parts, then she treats the system as a system. Otherwise, she treats the system as non-system, as a simple object. According to Ackoff, systems are still real, but their treatment and corresponding representation can be systemic or not. But since we can treat and represent "systems" as systems, but also as non-systems, then why not assume that reality is neither systemic nor non-systemic, and only our methods of treating real objects and corresponding representations of them are either systemic or not (Shchedrovitsky, 1966; 1975). For example, this assumption became the basis for the development of the Checkland’s Soft System Methodology (Checkland, 2000). And this assumption I will use in the rest of this paper.



We cannot point our finger and say "this is a system”, or “this is an element", like we say "this is a teacup”. Systems are not sensible objects. No one sees parts, relationship, or unity, as no one sees causes and effects, or forces and interactions. All we can see are particular things and events.

I asked my students to give me an example of system, and they pointed to a PC on my desk. I responded that what I see is things that we call keyboard, mouse, monitor, printer, computer tower, and wires, assembled together, but I do not see system, elements, relationship, or structure. They explained that if I switch my PC on, then keyboard, mouse, and microphone would function as input devices, communicating input signals to the CPU. The CPU would process the data and communicate the result to the output devices, such as the monitor, printer, and speakerphones. All these devices are components of the computer system. I insisted that, although I saw the devices, I still did not see inputs, outputs, data, processing, etc. They explained that, although I cannot see these entities, I have to know them. They are explained in the theories of electrical engineering and computer science. I accepted that, but still asked them about the relation of inputs, outputs, processor, and storage to PC as a system with its elements and their relationship. My students responded that the inputs, outputs, processor, and storage are elements of a computer system. If any of these elements were missing, we would not have a computer. Only all these elements, linked in a certain way and working together, constitute what we call a "computer system”.

In general, when we characterize a sensible object, a thing, e.g., a cat, as a system, we actually mean that, according to what we have learned in biology classes, a cat is a biological organism that consists of organs interacting in a certain coherent way that keep the organism alive, thus constituting a system. The problem of applying the concept of system to sensible objects and its relation to the ontological status of systems was first recognized and discussed by Emmanuel Kant in his Critique of Pure Reason (1781).

Kant was the first to distinguish three levels of thought: "All our knowledge begins with sense, proceeds thence to understanding, and ends with reason…" (Kant, 1781/1943, p. 189). Although later, other levels of thought were introduced (for example, in Phenomenology of Spirit Hegel split reason into negative reason--dialectics, and positive reason--speculation), the three Kantian levels of thought, perception, understanding, and reason, are sufficient for the current discussion.

According to Kant, system is an integral principle of reason: principle of unity of understanding (knowledge): "Reason cannot permit our knowledge to remain in an unconnected and rhapsodistic state, but requires that the sum of our cognitions should constitute a system" (Kant, 1781/1943, p. 446). At the same time, system is a set of principles that determine how reason unifies understanding: “The understanding may be a faculty for the production of unity of phenomena by virtue of rules; the reason is a faculty for the production of unity of rules (of the understanding) under principles” (Kant, 1781/1943, p. 191). These to views on system are combined in self-reflection of reason as being itself a system: “For our reason is, subjectively considered, itself a system, and in the sphere of mere conceptions, a system of investigation according to principles of unity…” (Kant, 1781/1943, p. 414). In other words, system is a conceptual construct of “pure”, or self-reflecting, reason: the system of system principles.

As belonging to reason, system principles cannot be directly applied to empirical experience, or reality of things and events: “Reason therefore never applies directly to experience, or to any sensuous object…” (Kant, 1781/1943, p. 191). System concepts can be directly applied to conceptual constructs of understanding, and through them, “projected” onto reality:

If pure reason does apply to objects and the intuition of them, it does so not immediately, but mediately--through the understanding and its judgments, which have a direct relation to the senses and their intuition, for the purpose of determining their objects (Kant, 1781/1943, p. 194).

Kant asserts that a concrete systemic object cannot be determined as a result of empirical observations of real objects. On the contrary, a theoretical model (“schema”, or “the arrangement of parts”) should be determined a priory according to the formal system principles ("the idea of system"), on one hand, and the subject matter ("the idea of the science"), on the other. This model then would prescribe which empirical object should be identified as a whole and how it should be divided into parts:

We require, for the execution of the idea of system, a schema, that is, a content and the arrangement of parts determined a priori by the principle which the aim of the system prescribes… A science, in the proper acceptation of that term, cannot be formed technically, that is, from observation of the similarity existing between different objects, and the purely contingent use we make our knowledge in concreto with reference to all kinds of arbitrary external aims… The schema of a science must give a priori the plan of it (monogramma), and the division of the whole into parts, in conformity with the idea of the science; and it must also distinguish this whole from all others, according to certain understood principles (Kant, 1781/1943, p.467).

Thus Kant ascribes to system the ontological status of conceptual construct of pure reason—system of system principles for designing scientific theoretical models (“schemata”). Only through the systemic theoretical models, system principles are applied to the sensible objects. This means that system is not a real, but actual object that belongs to the actuality of reasoning. Studying “real” systems or their models cannot produce system principles. Instead, I suggest reconstructing system principles by studying historical cases of system reasoning. The latter requires at least initial conceptual apparatus for description of system reasoning.






The classical definitions of system


In his Treatise on Systems, historically first discussion of systems, De Condillac (1749) gave the following definition of system:

Every system is nothing else but an arrangement of different parts of some art or science in a certain order in which they mutually support each other and in which preceding parts explain the following ones. The parts, which explain the other parts, are called principles, and the fewer principles, the better the system, so the perfect system should have only one principle (de Condillac, 1749/1938, p.3).


(Figure 3 is about here) 

Forty years later, Emmanuel Kant gave the most comprehensive definition of system:

By a system I mean the unity of various cognitions under one idea. This idea is the conception--given by reason--of the form of a whole, in so far as the conception determines a priori not only the limits of its content, but the place which each of its parts is to occupy. The scientific idea contains, therefore, the end, and the form of the whole which is in accordance with that end. The unity of the end, to which all the parts of the system relate, and through which all have a relation to each other, communicates unity to the whole system, so that the absence of any part can be immediately detected from our knowledge of the rest; and it determines a priori the limits of the system, thus excluding all contingent or arbitrary additions. The whole is thus an organism (articulatio), and not an aggregate (coacervatio); it may grow from within (per intussuscetionem), but it cannot increase by external additions (per oppositionem) (Kant, 1787/1943, pp. 466-467).

Principal System Constituents: Unity, Parts, and Relationship

As the contemporary definitions of system, the above classical definitions also use Unity, Parts, and Relationship as their principal constituents, but interpret them differently. All the constituents are interpreted as conceptual entities, not real things. De Condillac and Kant interpret relation between Unity and Relationship of Parts differently from each other and from the natural emergence. According to de Condillac, unity is not a special entity, but a special arrangement of parts with one part, or principle, uniting all other parts as a primary single basis for their explanation.

Unlike de Condillac and similarly to GST, Kant emphasized Unity as a separate entity: “I understand that contingent unity of the manifold which is given as perfectly isolated (at least in thought), placed in reciprocal connection, and thus constituted a unity (Kant, 1781/1943, p. 248). But unlike GST, Kant emphasized the priority of unity over the relationship of parts. According to Kant, unity is not something "new" emerging from relationship of parts. To the contrary, it is prior to relationship of parts (“conjunction”). 

Conjunction is the representation of the synthetic unity of the manifold. This idea of unity, therefore, cannot arise out of that of conjunction; much rather does that idea, by combining itself with the representation of the manifold, renders the conception of conjunction possible… It is therefore evident that category of unity presupposes conjunction. (Kant, 1787/1943, p.76).

G. P. Shchedrovitsky (1964/1966) specifies the relation between Unity and Relationship by their roles in the process of analysis-synthesis. He does not interpret Relationship (“a network of links”) as a “real” constituent of system, e.g. physical interaction of parts; instead, he interprets Relationship as the representation of synthetic operations applied to Parts in the process of conceptual restoration of Unity. In other words, Relationship is not present in the original Unity, but is added in the course of synthetic restoration of Unity. The following is a physical metaphor illustrating the idea. Suppose, one drops a teacup (unity), so it breaks (“analysis”) into pieces (parts). She then glues the pieces together (“synthesis”) in such a way that she can drink tea from it again (restored Unity). In this metaphor, the glue symbolizes a new addition (Relationship) that was not present in the teacup before it was broken, but had to be added in order to restore the cup.

Unlike the GST’s realist interpretation of Unity as emergent from Relationship of Parts, Kant’s interpretation of Unity avoids the paradoxes of the GST’s subject matter. In the epistemological dimension, concept of Unity exists prior to analysis of a complex whole into its parts, so nothing “new” or “unpredictible” emerges. The challenge is to analyze a whole in such simple parts and determine such relationship, that the corresponding synthetic procedure would restore (explain, reveal design, or fulfill) the initial Unity (Figure 4). In the ontological dimension, in the actuality of system reasoning, Unity and Parts with their Relationship are two different complementary representations of system linked by the method of analysis-synthesis.


(Figure 4 is about here)


Main System Categorical Oppositions: Form--Content, Complex--Simple, and External--Internal

The Opposition of Form and Content

Kant defines system as "the unity of various cognitions under one idea," that is the reason conception "of the form of a whole" that "determines a priori not only the limits of its content, but the place which each of its parts is to occupy". In other words, system, as a form, determines the unifying organization of its content: theoretical model, or “schema”. Thus system is opposed to the conceptual constructs of understanding as form to content.


The Opposition of Complex and Simple

We are accustomed to think of wholes as complex and parts as simple, so usage of the terms "complex whole" and "simple parts" is common and has a long tradition that can be traced down to Empedocles. Kant speaks "of a whole, which necessarily consists of simple parts" (Kant, 1781/1943, p. 248) and for us it is a habit to explain complex wholes in terms of simple parts. The identification of whole with complex and part with simple became a source of many paradoxes because these categories belong to different levels of thought. While “whole-parts” is a category of understanding, “complex-simple” is a category of reason, and these two levels are not always parallel.

In the Aristotelian (Categories, Metaphysics) terms, whole-parts is the relative type of opposition. Whole and parts unthinkable without each other: when we say "whole", we mean that it consists of the corresponding parts; and when we say "part", we think of the whole of which it is a part. At the same time, the opposition of simple-complex is of the possession-privation type: complex means privation of simplicity. We say that an object is simple because it is "easy to understand, deal with, use, etc." (Webster Dictionary). This can be thought of independently of any complexity. On the other hand, "complex" always means lack of simplicity, which is traditionally associated with composite nature and calls for analysis of the complex into its simple constituents.

If we compose two equal isosceles right triangles (parts) into a large isosceles right triangle (whole), few people, if any, would interpret the large triangle as something more complex than the initial small ones. Organs are not simple parts, if one tries to understand how organism works. Therefore, wholes are not necessarily complex and parts are not necessarily simple. However, wholes indeed can be complex and parts can be simple; and there is nothing wrong in the expression "complex whole" or "simple parts", if we understand that we are combining categorizations from two different levels of thought, understanding and reasoning, respectively. The purpose of decomposition is not always to obtain simple parts and the purpose of composition is not always to obtain a complex whole. At the same time, the purpose of analysis is always representation of complex entity in simple terms, and the purpose of synthesis is always representation of an entity as a complex unity. For these reasons, representations of objects as systems should deal with complex wholes and simple parts. While opposition of Complex-Simple characterizes system as a form, the opposition of Whole-Parts characterizes the systemically organized content.

The Opposition of External and Internal

The opposition of external and internal is an important constituent of a typical definition of system. Applied to system, this opposition can have several different interpretations. The first and the most popular is a physical, or spatial interpretation, corresponds to the level of perception of sensible objects. According to this interpretation all parts and their relationship are spatially internal to the system (Angeles, 1981), while the environment is external to it. For example, the concept of open/closed systems employs this interpretation of external-internal.

To the above physical interpretation of external-internal, Kant opposed a logical one, corresponding to the level of understanding (Kant, 1787/1943, p. 170). According to this interpretation, everything what belongs to a system’s conceptual content is internal, what does not belong to it is external:

By a system I mean the unity of various cognitions under one idea. This idea is the conception--given by reason--of the form of a whole, in so far as the conception determines a priori not only the limits of its content, but the place which each of its parts is to occupy. (Kant, 1787/1943, pp. 466-467).

Correspondingly, "the place” that each part occupies is not a location in physical space, but the place in the logical organization of the systemic theoretical model (schema). In the same way, I interpret von Bertalanffy's statement, "that the laws governing behavior of the parts can be stated only by considering the place of the parts in the whole" (von Bertalanffy, 1968/1998, p. 67). According to this interpretation, G. Shchedrovitsky (1975) identified the corresponding group of systemic operations: insertion of an element into its place in the object's structure and reverse operation of extraction of an element out of the structure.

At the level of reason, the opposition of external and internal can be viewed as the opposition of the two system representations: Unity as external and Parts and their Relationship as internal. In the process of analysis-synthesis, system reasoning shifts between the two representations, starting with one that is overt, or directly accessible, and shifting to one that is covert and must be discovered by means of appropriate analytical or synthetic procedure. In the case of sensibly single objects, external characteristics usually are directly accessible for observation and measurement. At the same time, determining their internal structure, elements, and properties requires special analytical efforts. For example, if water is given directly, electrolyze is required to break it into hydrogen and oxygen. On the contrary, in the case of a sensibly multiple objects, parts and their properties are directly accessible, but determining the integral properties of the composite whole requires special synthetic efforts. For example, in anthropological investigation of family in a given culture, behaviors of men, women, and children of different age are directly observable. At the same time, integral features of a family, role composition, boundaries, and external social functions are covert and require special synthetic effort. For its relevance to the method of analysis-synthesis, this paper favors this interpretation of external-internal.





The New Meaning of Empirical Study of Systems


As it was demonstrated in the first part of this paper, the realist approach that treats systems as real objects or things, has failed to produce system principles. This paper suggests the alternative approach that is based on Kant’s teaching that system is not a real object, but a conception of reason: system of system principles. According to this approach, system science cannot study real systems because they do not exist. Instead, system science should be the empirical investigation of system reasoning in philosophy, science, engineering, and practical applications (Shchedrovitsky, 1965; 1975). Applying the idea of Wilhelm Windelband (1900) that history of philosophy can be represented as development of principles of thought, I suggest that the study of system reasoning should be a historical study of development of system principles. Among other things, this means that empirical analysis of every case of system reasoning should have a two-fold purpose: (1) reconstruction of the underlying system principles and (2) determining place of the case in the chain of historical development of system principles. 

The Method for Empirical Study of System Reasoning  

According to the Activity Approach (Shchedrovitsky, 1965; 1974; 1975), the ontological picture of system is determined by analytic-synthetic procedures: 

When one characterizes “system” (either as a concept or an object), one asserts that system is a complex unity in which one can distinguish constituting parts - elements as well as structure, a network of links, or relationships among the elements. It is as if through this definition, we directly see the object comprised of elements and links among them; what we see is an ontological picture of system approach. But an ontological picture, as we have discussed before, incorporates, or “convolutes in itself” all the procedures and methods, which we apply to various symbolic constituents of scientific disciplines, representing their subjects as systems. And these procedures and methods should be discovered, in order to determine categories of the systems approach (G. Shchedrovitsky, 1975/1995, p. 249). 

Shchedrovitsky (1975) identified three groups of system procedures that underlie the ontological picture of system. The first group includes decomposition of an object into its parts and the opposite procedure of object’s reconstruction by composing its parts into the whole. The second group of system procedures includes measurement of different heterogeneous characteristics, or "aspects" of an object as a whole as well as measurement of its parts and the opposite procedure of configuration (Lefebvre, 1967), or explanation of the characteristics of the whole by relationships of the characteristics of its parts. The third group of system procedures includes insertion of an element into its place in the object's structure and the reverse operation of extraction of an element out of the structure.

It should be noted, that in most cases, system reasoning is applied not to system per se (e.g., as in self-reflection of reason), but to specific systems that are usually viewed as real objects: physical, chemical, biological, etc. Since system concepts can be directly applied only to conceptual constructs of understanding, every case usually deals with well established empirical knowledge or practical experience: “we can compose true systems only if we have enough observations to understand relationships among the phenomena” (de Condillac, 1749/1938, p.168). This is why, in concrete cases of system reasoning, philosophers, scientists, or practitioners, use specific terms of the respective discipline (system content) often mixed with the formal system terms. Therefore, in case studies of system reasoning, it is frequently necessary to translate the terms of specific disciplines into the purely system terms. I believe that the concepts of Unity, Parts, and Relationship, and categories of Form-Content, Complex-Simple, and External-Internal is a sufficient preliminary “dictionary” to start with. 

In order to identify cases of system reasoning and relate them to each other and to the history of development of system principles, we need to define a characteristic common to all the cases. I believe that this characteristic is dealing with complexity by means of analysis-synthesis. At the first step, each selected case identified by this criterion should be presented in the terms of the discussed above system concepts, categories, and procedures of analysis synthesis.

At the next step, the underlying system principles should be reconstructed and formulated. If we accept that principle is a non-derived “fundamental, primary, or general law or truth from which others are derived”, then we should conform to the Aristotelian standard of introducing principles by means of opposition (Physics, 189 a 10). Such formulation should produce two types of principles identified by Kant (1787): "regulative" principles of method, and "constitutive" principles of representation.

It is our hope that we will be able to order the analyzed cases of system reasoning along the line(s) of development, where every subsequent case would have more conceptual and categorical distinctions, more developed system of system principles, and less logical problems than preceding cases.




Ackoff, R.L. 1960. Systems, organizations and interdisciplinary research, General Systems 5: 1-8.

Ackoff, R.L. 1964. General system theory and systems research: Contrasting conceptions of system science. In Proceedings of the Second Systems Symposium at Case Institute of Technology. 

Angeles, P.A. 1981. Dictionary of Philosophy. Harper & Row: New York. 

Ashby, W.R. 1958. General Systems Theory as a New Discipline. General Systems 3: 1-6. 

von Bertalanffy, L. 1969/1998. General System Theory. George Braziller: New York. 

Checkland, P.B. 2000. Soft Systems methodology: A thirty year retrospective. Systems Research and Behavior Science 17: S11-S58. 

de Condillac. 1749/1938. Treatise of Systems that Uncovers Their Shortcomings and Merits (Traktat o Systemah). State Social-Economics Publishing: Moscow.

Honderich, T. (ed). 1995. Oxford Companion to Philosophy. Oxford University Press: New York.

Kant, Emmanuel. 1781/1943. Critique of Pure Reason. Willey: New York.

Lefebvre, V.A. 1967. Conflicting Structures (Confliktuyushchiye struktury).: Higher School Publishing: Moscow, Russia.

Rapoport A. 1964. Remarks on general systems theory. In Proceedings of the Second Systems Symposium at Case Institute of Technology.

Rosen, R. 1991. Life Itself: A Comprehensive Inquiry into the Nature, Origin, and Fabrication of Life. Columbia University Press: New York.

Russell, B. 1948 Human Knowledge, Its Scope and Limits. London.

Shchedrovitsky, G.P. 1966. Methodological problems of system research. General Systems 11. (Metodologicheskiye Problemy Systemnogo Issledovaniya. Moscow, 1964).

Shchedrovitsky, G.P. 1975/1995. Fundamental concepts and categories of activity theory. In Selected Works of G.P. Shchedrovitsky, A.A. Piskoppel and L.P. Shchedrovitsky (eds). School of Cultural Policy: Moscow, Russia.

Skinner, B.F. 1974. About Behaviorism. Knopf: New York.

Windelband, W. 1900/1958. A History of Philosophy. Vol 1. Harper Torchbook: New York.

© 2005-2012, Некоммерческий научный Фонд "Институт развития им. Г.П. Щедровицкого"
115419, г. Москва, ул. Орджоникидзе, 9, корп.2, под.5, оф.2. +7 (495) 775-07-33, +7(495) 902-02-17