One in eight applications papers (12 percent) used the full TP analysis, whereas over 40 percent (43/106) involved only one TP tool. Even though TP have been developed to make mutual and complementary contributions as a suite of integrated logic tools, the literature suggests that individual use of single TP tools, or tools in pairs or trios, is not only possible but has been to be found very valuable in dealing with problematic situations. See Table 23-2.

Methodological Developments and Enhancements

Kim et al.'s (2008) review also identified many methodological developments and variations that have emerged, including alternative approaches to building a tree, using specific TP tools to serve a different purpose from that originally intended, using TP tools to complement the use of other tools in addressing problem situations, and the development of new TP tools, such as the GEC, the CCRT, and IO maps.

Developments Pertaining to the Building and Presentation of the CRT The CCRT is a stripped-down CRT that facilitates communication with managers. It also serves to enhance buy-in by starting from a positive proposition (the desired objective, box A in the EC) rather than just a negative one (the core problem), and shows the relationship between these and the observed UDEs (Scheinkopf, 1999, Chapter 12; Houle and Burton-Houle, 1998; Button, 1999). Button also presents Goldratt's three-cloud approach for building a CRT, which was developed to reduce the time and difficulty of building a CRT. The traditional approach incorporates a 10-step procedure seeking likely causes for observed UDEs. The 3-UDE EC approach uses four steps to construct a CRT: (1) identify a list of UDEs; (2) generate three ECs (from seemingly unrelated problems) from the list of UDEs; (3) construct a GEC from the three ECs, thus identifying the likely core conflict; and (4) build a CRT that starts with the core conflict and harnesses the logic and pictorial representation of the GEC. While Dettmer (2007) decries such an approach, it is recommended in other texts (Cox et al., 2003), and the Kim et al. (2008) review reveals that both approaches have been much in use. Dettmer instead prefers the IO map as the starting point for TP analysis, arguing that the core conflict is more likely to be identified by the IO map's more strategic approach to identification of UDEs. There may also be occasions when three clouds may not lend themselves to a GEC when one EC is nested or embedded in the other EC (Davies and Mabin, 2009). The Strategy and Tactics (S&T) trees-covered in Chapter 25-were not evident in the review of the peer-reviewed literature, but are being used increasingly by TOC developers and practitioners.

TABLE 23-1 TP Tools-Reported Usage-1994 to 2009 TABLE 23-2 Classification of the Literature by TP Tools-in-Use-1994 to 2009 Methods Being Used Singly or in Sequenced Use Once TOC practitioners have identified What to Change by using the CRT, the second step in the traditional TP approach deals with the search for a plausible solution to the root cause; that is, to What to Change. This task can be accomplished with the aid of the EC and the FRT (see, for example, Taylor and Thomas, 2008; Taylor and Poyner, 2008). As evident from Table 23-2, many authors have seen the advantage of the EC as a standalone tool or method, and how it can lead to a win-win solution by surfacing and breaking the assumptions underlying the supposed conflict. The papers reviewed by Kim et al. (2008) described use of the EC method in conflict situations as varied as interpersonal conflict between sales manager and salesperson, writing MIS mini-cases, the creative design process, SCM, resource allocation in schools, TOC education, forest harvesting, Lean manufacturing and TOC implementation, managerial dilemmas, and "traditional" measurement. (See Kim et al., 2008 for further details.) Nevertheless, it has also been suggested that the use of the EC following development of the traditional CRT provides potentially far more diagnostic and solution generation power than individual use of the EC or CRT. One reason is that once a core problem has been identified using a CRT, it is more likely that a solution can be developed using the EC. Several papers (see Moura, 1999; Smith and Pretorius, 2003; Choe and Herman, 2004; Umble et al., 2006) describe and explain the combined use of both TP tools to identify the system's core problem and possible solution.

CRT-EC-FRT Method and EC-CRT(B)-FRT(B)-NBR Method Other variants on the "traditional" approach (CRT-EC-FRT) include the GEC-CRB-FRB-NBR multi-method approach, a refinement using the Current Reality Branch (CRB) and Future Reality Branch (FRB) (Cox et al., 2003). Cox et al. (2005) suggest using the EC rotated clockwise to provide a skeletal structure for the middle/upper part of the CRT. This and related papers (e.g., Davies and Mabin, 2007; 2009) also combine an EC portrayal of the conflict with a causal loop diagram (CLD; Senge, 1990) from System Dynamics (SD), which reveals the nature of systems relationship as well as identifies the underlying systemic structure of the problem situation. They have found that an important aspect of the EC process is that it ensures that the system goal is reflected appropriately in a second modified CLD, when it may have remained implicit or been overlooked in an initial CLD (or CRT) representation.

Validation Using CLR TOC's CLR provide guidelines for communicating any doubts or concerns about the validity of the entities and their connections within TP trees (see Dettmer, 1997). Balderstone (1999) suggested using the CLR for validating System Dynamics (SD) models, while Koljonen and Reid (1999) demonstrate use of SD models to validate TOC logic trees.

Full Thinking Processes Analysis (FTPA) While the TP logic tools or trees and the EC were developed as a suite, the review of the literature conducted by Kim et al. (2008) highlighted the frequent reported use of individual TP and the application of TP for a different purpose from that for which they had been originally designed. The latter uses, however, in no way deny the effectiveness of the TP tools as a suite that contributes to a Full Thinking Processes Analysis (FTPA). In a later section, we suggest why the FTPA may be regarded as a complete "package" or comprehensive methodology. As such, as designed, the FTPA would use all five original TP tools to examine a system in order to identify the core problem, develop solutions, and determine the implementation steps.

Nevertheless, the literature shows that the FTPA is often used and has value in seeking to overcome resistance to change by creating a logic path that can be followed by all stakeholders and participants. Houle and Burton-Houle (1998) lay out five layers of resistance, and correspondingly five phases of buy-in. Foster (2001) discussed nine layers of resistance to change and suggested that TP tools can be used to overcome each layer of resistance. Mabin et al. (2001) relate the layers of resistance to the sources of resistance identified in the change management literature and link the TP tools accordingly. Table 23-2 shows that only 13 papers in the published literature surveyed have contained complete descriptions of use of the FTPA-perhaps because the length of these analyses may prohibit the acceptance of such research in most journals. These papers detail how the FTPA can be applied to specific business situations (Klein and DeBruine, 1995; Boyd et al., 2001; Mabin et al., 2001; Reid et al., 2002; Gupta et al., 2004; Ritson and Waterfield, 2005; Reid and Shoemaker, 2006; Shoemaker and Reid, 2006), with other authors discussing the possibility of multi-methodology in detail (Thompson, 2003; Davies et al., 2005; Schragenheim and Passal, 2005). However, the reports of FTPA use demonstrate its versatility and applicability in relation to different functionality and settings, including establishing management policies, strategic planning, executing a bank merger, and in industry settings as different as the manufacturing industry, the motion picture industry, and the healthcare service sector.

The literature appears to support the views of TP developers, including Goldratt (1994), Scheinkopf (1999), and Dettmer (1997), that each TP tool in the TP set is a potentially valuable tool in its own right, without regard to its contribution in a suite or sequenced use of tools.

Summary of the Literature Review

The development of the TOC body of knowledge has been largely practice-led, manifested not only in the diverse nature of applications areas and in the diverse use of TOC tools, but also in the broader evolution of TOC methodology, methods, and tools. While the TOC TP had their origin and arose from concepts developed primarily in operations management, we note how their contribution to the development of the TOC body of knowledge has since generated impact well beyond the particular domain of operations management not just to wider business, but also organizations in general.

Earlier reviews of the literature (Rahman, 1998; Mabin and Balderstone, 2003) preceded many of the developments documented here, which have evolved since 2000. This overview has drawn primarily on the work of Kim et al. (2008) to present a review of the TP literature, as published in refereed journals and conference proceedings over a 16-year period from 1994 to late 2009, and to portray the development of TP concepts and tools since first applied in the POM (production/operations management) and OR/MS domains. The review of Kim et al. (2008) revealed specific publication and research gaps, and some common future research topics and approaches have also been identified. These will be discussed in the final section.

The review of TP tools-in-use has found that a combination of tools is often applied pragmatically according to the problem situation. Indeed, the overview has positioned the many TOC tools in multi-methodological use and in relation to each other, as well as capturing developments in multi-methodological usage across several domains. Consequently, a later section will examine the design-for-purpose and philosophical basis of the TP tools, as a means of understanding whether use of a TP tool for an alternative purpose is appropriate, and whether and how the TP tools in combination or as a suite comprise a comprehensive multi-methodological set. As a corollary, we will develop alternative perspectives on the nature of TOC methods and the TOC TP tools, their philosophical basis, and their use in problem-solving activities, that will facilitate comparison with other problem structuring and problem-solving methodologies and provide insight about the communality and complementarity of such approaches and methodologies.

It is apparent that TP have become a problem-solving method of choice for many, on their own, and sometimes in combination with other methods. Before we investigate TOC's philosophical roots, we will briefly discuss other managerial problem-solving methods in order to make a comparison.

The Nature of Other Approaches to Problem-Solving and Decision Making

The purpose of this section is to establish what other methods are being used for problem solving, in what ways they are being used, and in what ways they may be limited, thus providing a partial justification for TOC TP as an alternative or complementary approach.

As a facilitating framework for this discussion, we draw on the work of Mingers and Brocklesby (M-B) (1997) to clarify the role, function, and purpose of different problem-solving methods or tools, and for relating those methods or tools to problem content and problem-solving activity. In doing so, we seek to provide a basis for some selective comparison of traditional methods and TOC methodology.

The Relationship of Problem-Solving Methods to Problem-Solving Activity

M-B developed a two-dimensional mapping grid (see Table 23-3) with the purpose of alerting practitioners to the appropriateness of using different methodologies in different contexts, and to the possible use of multi-methodology. One dimension relates to the problem domain, specifically the nature of the world-social, personal, or material-being investigated, and a second relates to aspects of methodology, particularly the conceptually distinct but related phases of "intervention."

These phases are described within the M-B framework, for example, as building an appreciation of the social, personal, or material world that provides a necessary base for analysis of that world and relationships between key entities, before developing and assessing alternative futures and options to bring them about, and then finally being able to choose and implement alternative courses of action that bring about the desired future.

Despite Mingers' (2003, 560) later reservations about the limitations of the two-dimensional M-B framework in seeking to link methodology and method to problem content and problem-solving activity, we see value in using the M-B framework, both on its own, and in tandem with Mingers' later classificatory framework (2003).

Unstructured Approaches-Management on the Hoof

Management books paint a gloomy picture of the problem-solving and decision-making abilities of managers and organizational decision makers (Simon et al., 1987), highlighting the decision traps faced by managers (Russo and Schoemaker, 1989) and the common failings of managers (Nutt, 2002). These include, for example, weaknesses in the appreciation, analysis, assessment, and action phases of problem intervention-a failure to appropriately frame decision problems or problem situations; a failure in direction setting-that is, to determine inclusive, acceptable strategic goals and values; a tendency to jump in and act precipitously; a failure to understand or accommodate stakeholder influences and needs; a tendency toward overconfidence and to overestimate one's predictive ability, sphere of influence, or influence on past successes and future outcomes; a failure to learn from prior actions; a failure to recognize or address ethical dilemmas or the importance of ethical values, etc. (Russo and Schoemaker, 1989; Senge, 1990; Bazerman, 1996; Nutt, 2002). Some consequences are what many perceive to be the predominance of a firefighting mentality, and the preponderant use of managerial fads and fashions, such as quality circles, JIT, BPR, Six Sigma, etc.-with managers having the expectation that the use of these tools or processes, even in isolation, will help address their wider problems and deliver riches, now and in the future.

TABLE 23-3 Framework for Mapping Methodologies However, if and when "managing becomes a constant juggling act of deciding where to allocate overworked people and which incipient crisis to ignore for the moment" (Bohn, 2000, 83), it is usually deemed to be more expedient to attend to the squeaky wheel, to search for local solutions, as the big picture drifts out of sight. Managers then face the issues of tackling problems as they resurface or have adverse impact elsewhere. They may have framed their problems inappropriately, tackled the wrong problems, attacked problems at the wrong levels, or just addressed them in poor fashion.

Problems poorly addressed create more problems and take longer to fix in the long-term. Senge (1990) describes this common behavior in his Fixes that Fail and Shifting the Burden archetypes. In Fixes that Fail, an inappropriate fix might work in the short term but make the problem worse in the long term; for example, smoking may bring short-term relief but leads to long-term addiction and health problems. In the Shifting the Burden archetype, the quick fix not only makes the problem worse in the long term, but also undermines the effectiveness of any other alternative fix that could be used. For example, employing consultants may assist in the short term, but may consume resources required to develop expertise in-house.

In all of these situations, several features usually stand out. They include the lack of an overall perspective, the systems or holistic view; and a related inability to think about the wider stakeholder community, their values and views, about the wider systemic consequences over time, that is, within and without the system. More specifically, they include the related inability to think about the time-related dynamic nature of cause-effect relations and feedback. They include behavior of seeming irrationality and behavior suggesting lack of awareness of the values and perspectives of others. In essence, these are features suggesting that some formal processes may be needed-in particular, processes adopting a systems perspective. We provide a brief overview of such processes or approaches in the following subsections.

Formal or Structured Approaches

There are many structured approaches, and we have chosen to review a few of the approaches that have been compared with TOC. In the next section, we briefly outline some of the major traditional or "hard" OR/MS approaches before providing an overview of "soft" approaches, in order to provide a comparative critique of hard, soft, and TOC methods.

OR/MS Structured Approaches

OR/MS has adopted the phrase, The Science of Better, describing itself as the scientific approach to solving business problems. Similar terms have been used to describe the TP. Hence, a comparison between OR/MS and TOC would appear to be appropriate. Despite its origins as a problem-focused, multidisciplinary activity employing top scientists to attack operational problems, OR/MS has become very focused on techniques. In the United States, these techniques are almost exclusively quantitative in nature, and most modern-day OR/MS textbooks are rooted in the language of mathematics: mathematical modeling in its various forms such as mathematical programming (including linear and integer programming), simulation, heuristics, scheduling, decision analysis, data envelopment analysis, inventory control, and project scheduling. In these areas, OR/MS has achieved notable successes, largely through the use of powerful mathematical and computer modeling techniques to crack large problems. As such, OR/MS tools and techniques have predominantly contributed to the analysis and assessment phases of the problem intervention process, set out in the M-B framework. Indeed, such emphasis on mathematics and its use in this way is well recognized and even reinforced by the publication regimes of the top American OR/MS journals, which restrict their scope to those papers containing mathematically rigorous treatment (Simchi-Levi, 2009).

Some leading OR/MS authors, however, view this narrow definition of OR/MS-a collection of powerful mathematical tools-as unhelpful, even detrimental to achieving the full potential of OR/MS. As Daellenbach (1994, 112) puts it: When reading about how to do a problem formulation, the tyro management scientist is often somewhat impatient: "This seems to be all obvious-let's get down to the really interesting mathematical modelling phase! That is real OR/MS!" Unfortunately unless the groundwork for the modelling phase is properly done in the formulation, the risk is great that, although challenging, the modelling may address the wrong problem. Not only can this have serious consequences for the analyst, it also puts OR/MS into disrepute.

Refreshingly, perhaps pointedly, Daellenbach devotes the early chapters of his text to systems thinking, systems concepts, systems modeling, and problem formulation before introducing mathematical modeling.

Many OR/MS writers have made similar points about the tendency to solve the wrong problem; for example, Gass (1989), Zeleny (1981), Rosenhead (1989), and Mabin and Gibson (1998), and offered alternatives (e.g. Pidd, 1996). The debate that raged in OR/MS circles in the 1970s, led by Ackoff (1977; 1978; 1979), was largely due to this concern that the obsession OR/MS had with mathematical modeling led the OR profession astray. TOC writers have added their voices: Jackson et al. (1994) provided a powerful case comparing the standard OR-derived Economic Order Quantity (EOQ) with the EC approach for inventory control, following Goldratt's own treatment of batch sizing decisions (Goldratt, 1990b, 43); Mabin et al. (2009) compared OR's math programming approach with an EC approach to a warehouse/distribution problem. The concern of these authors over problem definition-rather than merely problem solution-is shared by the developers of the various soft OR methods, also known as Problem-Structuring Methods (PSMs), which are discussed in the next section.

Present-day OR/MS tools and methods have much to offer in addressing complexities relating to scale, time, and computation. They have much to offer when the problem is well defined and when goals, local or global, are known, understood, and accepted by stakeholders with common perspectives; when desired outcomes can be guaranteed by action; when the successful accommodation of multiple objectives is unambiguous; and objectives can be quantified. However, even when these conditions are not met, the sophistication of analysis and the scale of computer power can generate a sense of false security especially where the relationships of local goals to systems goals are not understood or accounted for, or when local goals surface as "numerical" or binding constraints in the hard mathematical formulations-and do so without being questioned. Indeed, even when problems are not well defined, assumptions are often made to make the problem tractable or amenable to mathematical formulation, often without sufficient questioning of the appropriateness of such assumptions.

The value of the TOC TP as a comprehensive methodology is that they bring such issues to the fore, forcing a consideration of the broader problem situation, global and local goals, challenging the assumptions that underpin them, oftentimes setting the solution path toward a very different goal.

While it is often claimed that effective OR/MS practitioners do seek to achieve the global and systems goals, the reality is often a suboptimization of a technical subsystem (in the material world of the M-B grid) that can be modeled, undertaken without any mandate or ability to place the problem in a wider context or to consider broader issues and ramifications.

In brief, most OR/MS methods have strengths in evaluating the relative effectiveness of alternative choices and decisions, and identifying the best among them, according to prescribed quantitative criteria and objectives, or in some cases, from a prescribed or readily imputed list of choices (for example, optimization by linear programming (LP). These methods stop short of guiding decisions on value systems, strategic direction, or other matters of identifying strategic choice for a variety of stakeholders. Furthermore, if methods of constrained or numerical optimization are examples of such methods, then by contrast, it is the methods of soft OR, along with TOC, that have been designed to grapple with the wicked problems or messes that are beyond the scope of the traditional mathematical modeling methods of OR/MS (Mingers, 2009a). We explore these matters further in the next section.

Soft OR

The concern of some OR writers over problem identification and problem definition-rather than merely problem solution-is shared by the developers of various soft OR methods or PSMs. These were showcased in the book Rational Analysis for a Problematic World (Rosenhead, 1989), which became the most referenced book in the Journal of the Operational Research Society in the next decade (Rosenhead, 2009). Soft OR is suited to messy situations, where the first issue is that of not knowing what the problem is. Soft OR methods have been designed and developed to grapple with wicked problems or messes by seeking to gain understanding of what would be desirable and appropriate goals of the organizational system and subsystems, and by seeking a broader, often predominantly qualitative, understanding of the problem domain or wider system within which it sits.

Soft OR or PSMs aim to: structure complexity of content and represent it in a transparent manner be deployed in a facilitated group environment develop model structure interactively incorporate tools to encourage participation and generate commitment to action.

However, they do so in a manner that bears the scrutiny of rigor expected of any science-based approach. Virtually none of the attributes of soft OR apply to traditional OR/MS approaches, the latter being increasingly known as hard OR (Rosenhead, 2009, S10.) The field of soft OR now includes a wide variety of approaches, developed for a range of purposes and applications, some of which could be considered systems approaches. They include: Strategic Choice Approach (SCA) Strategic Assumption Surfacing and Testing (SAST) Soft Systems Methodology (SSM) Critical Systems Heuristics (CSH) Cognitive Mapping (CM) Strategic Options Development and Analysis (SODA) Robustness Analysis Interactive Planning Soft Game Theory including Hyper and Metagames; Drama Theory The development of such soft approaches to address the design limitations of hard OR/MS approaches and methods paralleled the angst of early OR/MS pioneers such as Churchman (1967) and Ackoff (1977; 1979) about the sterility and inappropriateness of overly mathematical approaches for tackling complex social and business problems. Friend and Jessop (1969) developed SCA in the 1960s, as did Mason and Mitroff (1981) with SAST. Checkland and Scholes' (1990) major developments of SSM took place in the 1970s, as did major developments of Soft Game Theory by Howard (1971) and Bennett (1977), and Robustness Analysis by Rosenhead et al. (1972); CM was developed by Eden et al. (1983) in the 1980s; and the major drive to explore hard and soft methods in multi-methodology began to flourish in the early 1990s. However, the work of Munro and Mingers (2002) some 10 years later showed that up to that point, almost all claimed examples of multi-method intervention comprised either all hard or all soft methods, not both in combination.

Such soft approaches may provide the opportunity, in Ackoff's terms (1978), to dissolve the problem altogether, to resolve the problem satisfactorily, rather than just optimize or solve a technical problem that is an incomplete or inappropriate representation of the wider, relevant system domain.

Despite the growing evidence that soft OR is able to reach problems that traditional or hard OR cannot handle, such as organizational and individual behaviors and inconsistencies, soft OR is still not well accepted universally. While it is appreciated in the UK and elsewhere, it receives scant coverage in the United States. The hostile reception from journals like Operations Research and Management Science, which refuse to accept any papers that are "not based on rigorous mathematical models" (Simchi-Levi, 2009, 21), is the topic of current debate (Mingers, 2009a; 2009b).

Much of the soft OR story applies equally well to TOC. Indeed, most users of TOC TP would acknowledge their real benefit is in probing the very notion of what the problem is, why it exists, and what might be the outcomes if it did not exist, before diving into mathematical detail. However, there is also skepticism in traditional circles about whether TOC is a bona fide methodology, worthy of publication (Ronen, 2005).

Given the similarity in standing and appreciation offered to nontraditional approaches, one might argue that the time is right for TOC academics and practitioners to unite with academics from across a number of related disciplines, including soft OR, to persuade more editors that applications and theoretical developments that do not necessitate a mathematical approach are nevertheless worthy of publication and dissemination. In terms of rigor, TOC does have one clear advantage-the CLR governing the use of the TP provide strict logic protocols that lend rigor to the endeavors of TOC analysts.

Soft OR Methods-Theoretical Underpinnings

In this section, we draw together and reinterpret the prior discussion of soft OR methods in the context of the M-B framework, described earlier. We note, in particular, that soft approaches have been designed and developed to assist with all phases of problem intervention-appreciation, analysis, assessment, action-but especially as they relate to matters in the social and personal domains of the M-B classificatory system.

Table 23-4 provides examples illustrating how two soft OR methods, namely SSM and CM (Cognitive Mapping), map to the M-B framework. The "+" symbols indicate the relative extent to which each tool is purposively designed to attend to each phase of problem intervention in each of the problem dimensions. We note, for example, that SSM and CM have not been expressly designed to contribute to the analysis and understanding of underlying causal relationships in the material world-although their use may well contribute to doing so.

As may be inferred from commentary on the characteristics of hard OR/MS approaches, soft approaches have been developed to assist with situations where problems are not well-defined and where goals, local or global, are not necessarily understood, known, or accepted; where multiple stakeholders are involved; where desired outcomes cannot be guaranteed by action; and where success is ambiguous and the definition of success may need to be negotiated. As a consequence, soft approaches meet a need to facilitate learning about a problem, its constituency and constituents, their customs, practices, and ways of thinking; that is, they also seek to explore and accommodate a range of views, worldviews, values, and objectives without reducing them to a single measure, and they seek to encourage the active involvement, engagement, and commitment of stakeholders (Mingers, 2009b).

These latter features of the soft approaches are reflected by multiple "+" signs emphasizing the level of their contribution to the personal and social domains of problem context, across the whole spectrum of phases of problem intervention in the M-B classificatory framework. By contrast, hard approaches tend to be situated to provide analysis and assessment within the material world. Furthermore, the notion that soft approaches meet a need to facilitate learning about a problem aligns with Checkland and Scholes' view that soft systems approaches can or should also be conceptualized as "learning systems" (1990, A8).

TABLE 23-4 Mapping of SSM and CM For instance, and for Table 23-4, we note that within the action and implementation phase that both SSM and CM are designed to seek out and seek accommodation of disparate views/or consensus, and to seek enlightenment and empowerment for problem constituents as well as the problem owner and analysts. In stark contrast to hard OR/MS methods, we further note that the primary purpose of such soft methods is to understand better, not necessarily to identify best alternatives.

CM, similarly to SSM, seeks to effect a representation of how individuals view a problem, what it means to them, and how they make sense of it. The axiology or purpose of CM is to surface and understand these beliefs in order to generate consensus about possible strategic action. In contrast, we reaffirm that hard OR/MS relates mostly to the material world and focuses on analysis and assessment, leading to action in that domain. However, both hard and soft methods and methodologies may also be described as systems approaches. The nature of systems approaches and methodologies will be explored further in the next section.

Systems Approaches

Systems approaches to problem-solving typically conceptualize "problems" as existing within a notional whole or synthetic system, where a system can be defined as any grouping of people, events, activities, things, or ideas, connected by some common reason or purpose (Senge, 1990). As such, many systems can be best described as notional. In general, we can describe systems as being natural, for example, ecological systems; as being designed, for example, a car or an organization; or as being a human activity system, for example, a sports team or an ad hoc work group. Systems thinking attempts to reflect and illustrate the importance of holism, of boundaries, of feedback, of reciprocal relationships, and the notion that, say, activities or events, while perhaps separated by distance and time, cannot be understood in isolation, but instead need to be understood in terms of the patterns of relationships that create them and the patterns of behaviors that emerge from those relationships.

Systems thinking entails, above all, a sensibility about matters systemic (Espejo, 2006); that is, consideration of the big picture, the need to think holistically, to consider the whole as a network of relationships of interconnected parts or subsystems, and the need to understand feedback. Systems thinking, according to Senge (1990), involves learning to recognize structures that occur repeatedly-a notion that accords with TOC practice.

We often seek to understand a problem or problem situation by taking a Descartesian reductionist approach to analyze and understand its identifiable "component parts." However, systems thinking reflects a recognition that to understand a problematic situation fully, or why a problem exists, and persists, requires the problem to be situated within a wider context, a notional whole or system, and then to understand how the parts of that system relate or contribute to the whole-which, in itself, is an act of synthesis or synthetic systems thinking. Indeed, such conceptualization of problem situations as systems is an act of systems thinking.

In general, we may make a useful distinction between representing reality as systems, using systems language and protocols, and inquiring of what we regard as reality using systems approaches (Senge, 1990; Checkland and Scholes, 1990). The latter notion is that by examining situations using systems frameworks as learning frameworks, and by using the systems concepts of holism, boundary, feedback, etc., one can gain an understanding of complex situations where seemingly insignificant events can catalyze the playing out of complex relationships that generate unpredictable, unanticipated emergent behaviors and outcomes, which cannot be attributed to any single causal event.

Senge (1990) regards systems thinking as a discipline for seeing wholes; as a framework for seeing inter-relationships rather than events, and for seeing patterns; as a set of principles; and as a sensibility for "the interconnectedness that gives living systems their unique character" (69). In adopting a systems approach, one is therefore less likely to be reactive or over-reactive to current or local events or outcomes, where such over-reaction may potentially exacerbate undesired problems elsewhere. As a corollary, we may become more sensitized to patterns of change, to the impact of change, and of the systemic influences whereby even a positive change in one area of a system may lead to adverse effects elsewhere in another part of the system. As such, systemic sensibility to such possibilities may likely reduce the tendency to act and think suboptimally and, as Senge (1990) suggests, to be generative in terms of creating systemic structure that leads to sustainable and desirable outcomes.