1. INTRODUCTION

“The whole is more than the sum of the parts”--Aristotle

This synergistic phrase, long a central tenet of systems theory, is often cited in the recent phenomenal growth of many information systems. It is evoked to explain the growth of the Internet itself, networks of all sizes, P2P applications such as Napster, supply chain management, consumer trust, E commerce, and globalization.  Though central and common to these current IS phenomena, this principle of systems theory is not, at times, well understood (Klir, 1991; Porra, 1999, Senge, 1990).  Moreover, this limited understanding is not isolated to this particular tenet, but suggests a number of the principles of systems theory could be better understood. To address this problem, the purpose here is to systematically clarify the tenets of systems theory in order to improve its use. 

Systems theory is better conceived as a collection of disparate theories.  By highlighting both the variety and commonality of systems theories, this paper suggests a classification scheme of four systems theories, and posits that systems theory is better conceived as a system or taxonomy of these four specific theories. 

The systems view, long established in some circles (Ackoff 1971), is gaining support elsewhere in the social sciences:  in human factors (Meister, 1989), in sociology (Kling, 2001), in organizational communication (Miller, 1995), in teamwork (McGrath, 1984), and in management, where Senge’s application of systems has recently been recognized as a seminal work in 75 years of management research (Harvard Business Review, 2000).  Elsewhere, with advances in computational power, systems theory is being reborn in the natural sciences where aspects of it are labeled chaos theory, complex systems, and non-linear systems (Bohm and Hiley, 1993; Campbell, 1989; Gell-Mann, 1994; Prigogine, 1980; Waldrop, 1992). 

Within information systems, applications of systems theory are growing as ambitious information systems become increasingly complex (Xu, 2000; Zeigler, 1996). For example, IS research relies on systems principles to investigate modularization and bundling (Shilling, 2000), systems design (McLeod, 1995; Xu, 1995; Zhu, 1998, 2000), decision support systems (Eom, 1998; Takahara and Shiba, 1996, 2000; Zhu 1998, 2000), and networks (Xu, 2000).  Moreover, the growing use of systems theory is also evident in introductory textbooks (Alter, 1999; O’Brien, 1999; Oz, 1998; Zwass, 1998), as well as highly regarded academic journals (Lewin, 1999; Porra, 1999).  Further, a new discipline of Social Informatics invokes systems theory to highlight the pernicious and lasting social impacts of information systems (Kling, 2001).  Unfortunately, this growing impact of systems theory on IS has not been well demonstrated until recently (Eom, 1998, 2000; Kirby, 1993; Xu, 2000).  This inattention may contribute to IS graduates’ lack of ability to think systemically, a key shortcoming identified by employers (ISCC ’99, 1999).

Systems theory is particularly valuable for IS during this current period of growth and change.  First, systems theory provides the language and perspective to understand adaptation, the key to understanding MIS use in organizations (Anderson, 1999). Second, the explosive growth in technology has fueled concomitant growth in the complexity of IS applications.  Systems theory is well suited to finding the hidden order at any level and structure in large complex systems.  Third, technology growth also has spurred new organizational structures of networks and alliances.  Systems theory, as evidenced by Senge’s work, is a robust language for these new structures and the IS that support them.  Fourth, the field of IS has shifted from production to identification and delivery of information.  Systems theory, with its fundamental input, process, output framework is well suited to model this new role of IS.  Finally, information economics is different than the physical good economics where single physical assets have singular values. Systems theory, based on non linear combinations of value, is better able to model information economics than reductionism.  In sum, a detailed understanding of systems theories is essential for IS professionals and scholars confronting today’s systems.

As the systems view gains popularity, it is necessary to reconceive the idea of systems as a collection of systems theories.  Some 30 years ago, Ackoff (1971) proposed an initial system of systems concepts, an integrated set of ideas or concepts about systems theory.  He organized over 30 principles of systems, definitions, and concepts into one comprehensive framework in order to sharpen definitions that had become dull with use and to aid understanding.  Suggested here is a system of theories rather than a system of concepts, but for similar reasons. 

The need to once again clarify is evident.  No mainstream introductory MIS text acknowledges more than one systems theory exists, well know popular business books based on systems theory (Wheatley, 1992; Senge, 1990) also fail to distinguish the variety of systems theories.  Recognizing the need to qualify the use of systems theory, Porra (1999) is an example against this trend.  Behind this indiscriminate use is the impression that the underlying principles are well accepted, that only one systems theory exists. However, as argued here, while systems theory in general has several well-accepted principles, systems theory has at least four variations whose underlying assumptions about knowledge and philosophy are in conflict.

3. DIFFERENCES AMONG SYSTEMS THEORIES

“Systems thinking, if anything, should be carried out systematically.” (Ackoff, 1971)

Taxonomy, the science of classification, dates to Plato (Mayr, 1969).  The purpose of a classification is to describe the structure and relationships of the constituent objects, to achieve economy of memory and facilitate communication, and to ease observations and retrieval of information (Fleishman and Quaintance, 1984; Jarvenpaa, 1988). A classification is a first step in organizing knowledge; it can facilitate comparison, highlight difference, limit generalization, expose gaps in knowledge, assist theory development, provide comparability, assist in extrapolating results to practice, help form theory, lead to new research questions, point out where extensive research has been done or where more research is required, and identify a critical set of characteristics and definitions for these characteristics. (Fleishman and Quaintance, 1984; Melton, 1964; Sokal, 1974).  For these reasons, systems theory needs a classification.  Further, classifications typically arise in heavily researched fields in which impetus has come from an overwhelming cascade of unorganized facts where a metalanguage to describe concepts in a field is needed.  In sum, a classification is a system that identifies regularities and order at a high level of organization.

Hard Systems

Hard systems theory employs quantitative techniques from a positivist epistemology similar to the traditional sciences.  Epistemology, or the grounds of knowledge, examines how we understand and communicate knowledge about the world (Burrell and Morgan, 1979).  Positivism is based on formal propositions, quantifiable measures of variables, hypothesis testing, and drawing of inferences; it creates a cumulative growth of knowledge as in the natural sciences (Klein & Myers, 1999).  What makes it different from traditional science is that its level of analysis is more holistic; the object of inquiry is typically large-scale systems in operation.  Labeled systems management, management science, and operations research, it is teleological:  assuming the existence of purposeful global goal seeking functions that it seeks to optimize (Forrester, 1971).  The aim is to predict the behavior of the system within a framework of self-control, optimization and objectivity.  The method of research is nomothetic (the study of cases or events as universals, an emphasis on measurement and identification, and a view to formulating general laws), and is rapidly becoming entirely quantitative.  It is important to note the evolution of this field.  Operations research and management science have evolved from roots in hard systems and now are almost entirely reductionistic; in a sense, hard systems is no longer practiced.  Systems dynamics is a label ascribed to the most systemic versions of current operations research and management science.  Hard systems was developed to solve the complex problems of logistics and resource management in World War II (Banathy, 1996).

Complex Systems

Within the past 15 years, this school of thought has emerged sharing many of the same tenets with hard systems, but extending their common mechanistic, positivist-nomothetic, predictive, regulative approach to non-purposeful domains in the natural, and artificial sciences. A key descriptor is the non-linear relationship between the whole and key “subroutines”.  These subsystems are complex functions, which tend to be performed in few locations and result in emergent behavior, or indirect effects, in the overall system.  These sublocation processes explain and predict emergent properties at the level of the whole.  For example properties of the immune system such as disease response and memory emerge from changes in a few specific cellular process; the emergent property of memory can be explained by subprocesses in spin glasses (magnetically charged glasses) (Campbell, 1989).  Increasingly the term adaptive is used to describe complex systems. 

This group has proposed emergent property models for the immune system, evolutionary biology, spin glasses, computational physics, dynamical functions, ecosystem dynamics and chaos theory (Devaney, 1990; Kauffman, 1993; McNaughton, 1989; Mitchell, 1995; Stein, 1989b; Zurek, 1990).  This school is concerned with explanation and prediction via pattern recognition, modeling agent interaction, and understanding local goal seeking ("niching") rather than teleological global optima.  Further, complex systems behavior is thought to be highly dependent on initial conditions; small variations in these conditions have significant non-linear impacts on system performance.  The complex school is critical of the hard systems approach as inadequately addressing complexity or emergent phenomena, overly relying on simplifying linear approximations, and unsuited to the inherently dynamic, iterative, interactive nature of complex systems that produces the emergent phenomena (Santa Fe Institute, 2001).  It attempts to quantitatively predict system-wide behavior by building mathematical, but non-linear models of the system's components.  (Linear functions by contrast, predict model behavior based on weighted sums of input values.)  One of the key differentiators of the complex framework is methodology, its specification and use of its own distinct computational tools. The dominant method is simulation.  Models of the system's components and their interaction are programmed.  Initial conditions, input from random number generators, are varied, and the quantitative patterns or symmetries developed over the multiple iterated runs are evaluated. 

The development of this framework was concomitant with growth in network computing technology.  These technological advances in the 1990s enabled the type of distributed computing and other tools that characterize the complex systems approach.

Soft Systems

Soft systems was developed to complement the hard systems approach, differing in epistemology, key principle, purpose, and method.  It arose from a need to better address complex contemporary social issues (Flood and Jackson, 1991).  Its interpretivist epistemology holds that subjects or groups construct knowledge because of selection pressures from the environment (Heylighten, 2001).  The interpretivist suggests that our knowledge of reality is gained only through social constructions such as language, consciousness, shared meanings, documents, tools, and other artifacts.  Moreover, various stakeholders have unique and valid views of the problem space.  Further, problem identification and selection are largely idiosyncratic:

The social world is perceived (and constructed) by men according to the particular world-views.  This is a cultural mechanism which maintains desired relationships and eludes undesired ones.  The process is cyclic and operates like this: our previous experiences have created for us certain standards or norms, usually tacit; the standards, norms and values lead to readiness to notice only certain features of our situations; they determine what facts are relevant; the facts noticed are evaluated against the norms or standards, so that the future experiences will be evaluated differently. (Vickers, 1983; p. 17)

Another fundamental difference of soft systems is the idea that goals may be ambiguous, conflicting, non-quantifiable, and indeterminate.  That is, ambiguity of problems is not a result of underdeveloped analysis tools; it is how things are.  Thus, problems involve judgment, weighing moral issues and creation of form (Checkland, 1981).  As a result, solutions do not emerge from one decision, but over time where action and refinement has a better chance of success.  Direct cause and effect is rejected, a more indeterminate problem space is considered more realistic, and as a result, this approach is often described as organic.  Therefore, social problems rich in complexity and change need to be managed rather than decided or solved, the predict and control framework of complex and hard systems yields to design and invention (Flood and Jackson, 1991).

The ideographic method of soft systems is founded on the premise that the world is understood only by first hand knowledge of the subject (Burrell and Morgan, 1979).  It encourages participants to accept multiple realities, multiple worldviews of a problem.  That is, participants are shown the idiosyncratic nature of their own worldview and how this affects problem identification and solution.  As a result, theory and practice are inseparable; practitioners attempt to help participants in social problems see themselves within the higher-level system or context (Flood and Jackson, 1991).

Finally, validation in soft systems methodology is difficult if not impossible.  External validity in an interpretivist epistemology depends on improved behavior of participants.  However, this opportunity for improvement assumes stakeholders are guaranteed free and open discussion about changes to be made.  That may be unrealistic to assume.  In reality, powerful participants in the process are unlikely to risk their dominant position and submit their privileges to the vagaries of others' ideal demands (Jackson, 1991).  This critique leads to the critical systems position.

Soft systems emerged in response to the failure of hard and quantitative tools to model messy social problems in the 1960s.  It represented a shift away from mathematical modeling to understanding the process of building consensus viewpoints.

Critical Systems

The critical approach takes it name from the radical humanist paradigm of sociology.  It is committed to the moral concepts of individual progress and emancipation from constraining paradigms and traditions.  Sharing foundations of interpretivism, ideographic methodology and purpose with the soft approach, it views soft and hard systems as regulative approaches, unaware of their own conservativeness, and more generally the role of power in shaping social action and meaning.  The critical approach shares with the radical humanists the view that consciousness is dominated by social infrastructures in which an individual operates (Burrell and Morgan, 1979).  Hard systems explicitly, and soft implicitly--although it claims to be political and ideological neutral (Flood and Jackson, 199l)--take as a given organizational mission and needs.  Problems are resolved to return the system to equilibrium.

 According to Jackson (1991), Ulrich (1991), and Schecter (1991), the critical approach is founded on critique, emancipation, and plurality.

Critique is a commitment to questioning the methods, practice, theory, non-native context and limits of rationality of all schools of thought.  It requires a never-ending attempt to uncover hidden assumptions and conceptual traps.  The commitment to emancipation is a commitment to human being and their potential for full development via free and equal participation in community with others.  The commitment to pluralism insists that all systems approaches have a contribution to make and that no single approach is adequate to address the full range of problematic situations. (Schecter, 1991; p. 211)

One example of the critical school's methodology is presented in Ulrich (1991).  He argues that problem selection and identification requires numerous boundary judgments of what is relevant beyond the control of logic or reason.  Specifically, participant consensus on issues relevant to the problem should be motivated by considering "what should be" rather than "what is" to avoid overlooking hidden boundary judgments.  Four general issues should be discussed, what should be sources of motivation, what should be sources of control, what should be sources of expertise, and what should be sources of legitimization in the domain of the problem space. 

In general, critical theory attempts to question objectives toward which discussions are offered.  It disagrees with the sociology of the soft approach that free and open debates are ever possible.  It points to the weakness of soft systems theory's attempt to resolve plurality of ideas via exchange.

In the end, validation is possible,

... only via the social actors involved in the process.  The analyst's success is measured by the extent to which the patient recognizes himself in the explanations offered and becomes an equal partner in the dialogue with the analyst.  The actor in the social world very often suffers false-consciousness and does not truly comprehend his situation in that social world.  It is incumbent, therefore, on the critical theorist to employ a social theory capable of explaining the alienated words and actions of oppressed groups in society. (Jackson, 1991; p. 133)

Critical systems theory is a manifestation of the movement from confidence to skepticism in philosophy and sociology.  This growing trend in philosophy in the 1960s created the conditions for the emergence of radical sociology which provides the philosophical support for the critical school in the 1990s.

Summary of Major Differences

Each theory differs in unique ways from the others.  However, three main differences between the top two theories in Table 1 (hard and complex) and the bottom two (soft and critical) are instructive of how significantly these frameworks vary. 

It is evident that to the soft and critical frameworks, systems is a philosophic choice, to hard and complex, it is distinguished as a methodology.  That is, for the critical and soft perspectives, systems theory explains how individuals perceive and conceive their world.  This epistemological role for systems contrasts with the hard and complex perspectives that assume a positivist epistemological perspective of traditional science, and view systems as a method to investigate objective phenomena using a non reductionistic process within that accepted epistemology. 

This methodological difference is a manifestation of a key epistemological one.  The interpretative systems theories (soft and critical) insist that an objective viewpoint outside of a system is not available, the observer or interpreter is always a part of the whole system under scrutiny.  The positivists (hard and complex) separate subject and object and hold that with appropriate methodology the object system can be understood.

5. SUMMARY

Systems theory often means different things to different people, a malady that precipitated development of classification schemes in psychology and biology (Fleishman and Quaintance, 1984; Klir, 1991).  To help preserve the utility of systems theory concepts for MIS, this classification starts ontologically, and develops epistemological, methodological, sociological, and other key descriptors. Again, the goal here was to clarify the variety of systems theories to make each of them more useful, and to suggest that a meta theory of systems theories is more appropriate conception.

To be sure, systems theory is not a panacea; a useful understanding demands consideration of its limits.  In addition to the main philosophical criticisms mentioned earlier, it is difficult to test hypotheses in a conventional sense using systems theory.  It cannot compete with what traditional reductionistic science has become over the centuries—a consistent, reliable, language and valid way of knowing.  The questions systems theory poses often do not translate into the form of specific testable hypotheses common in traditional science.  As a result, its language may appear fuzzy or vague; its appeal is as a different way to think about the world, not as a clear way to test it.  A nice concept, but it can not do anything useful. With soft and complex systems theories, each user determines validity, making standardization, training, and regulation difficult.  Systems theory, a language without a sentence, pays for its generality with validity.  It's maddeningly broad and vague, generating curious and intriguing insights that disappear like a Cheshire cat when examined in the light of traditional science. 

These limits are not taken lightly or dismissed.  What then should be done with systems theory?  This paper, using a soft systems perspective, suggests reestablishing a theoretical foundation by identifying systems theory as four specific variations, while also recognizing that taken as a whole, these four comprise a classification or a system of systems theories. 

In addition to this classification, this review of the variations of systems theory aims to unhinge long-held and unquestioned epistemological views.  Ideally this will can lead to closer scrutiny of systems theories, and to stimulating communication that establishes better understanding and research. 

In closing, systems theory has evolved as it suggests all systems do.  Systems theory, in response to the pressures of use, has adapted from a singular general theory into four separate and more coherent independent theories.  Now, treating these four as a system, systems theory again adapts as the use of the four independent theories reveal a hidden order, a pattern resolved only at the taxonomic level.  

What is valuable to MIS is that these four schools of systems thinking are becoming established in their own right, with important implications for MIS.  As MIS continues to seek philosophical support  (Courtney, 2000) it would do well to build on the foundation these four pillars provide. 

"The more science becomes divided into specialized disciplines, the more important it becomes to find unifying principles.”  (Haken, 1988)

6. SUGGESTED READINGS

            General Systems Theory