Printer Friendly

A life full of learning.

I began my career with a dedication to reducing injustices. The injustices that intrigued me were those that inhibited the expansion of liberating alternatives. Soon I narrowed my focus even further to those injustices created by human beings when they were acting to reduce the injustices. The more that I studied these phenomena, the more I found myself studying processes that were self-sealing, compulsively repetitive, and non-interruptable and changeable by the very people who created them. Kubie (1958) described such processes as neurotic and undermining creative change. Fromm (1955) made a similar point, emphasizing societal variables.

It became increasingly clear that human beings were skillful at maintaining vigilantly these corrosive and non-learning features. These features continue to exist today in organizations. Yet there are no courses at universities or other executive programs that teach human beings, groups, intergroups, and organizations the skills of producing these counterproductive, compulsively repetitive processes. How do we explain their omnipresence and perseverance?

I chose to become a scholar because I felt a deep sense of responsibility to test the claims that I would make in the most robust manner possible. I sought to assure myself that I was not unwittingly kidding myself or others in the claims that I was making.

At the outset of my career, I assumed that justice and seeking truth represented unbounded good. As I will illustrate below, I learned that both features contain inner contradictions. As justice is strengthened, it can also produce knowledge that violates it. As truth is strengthened by the use of the ideas in good currency in the scholarly community, it can also create conditions for limiting truth.


A key feature of the above is the inability to learn; to detect and correct error. How come human beings are skillful at non-learning? In order to find valid answers, it became necessary to distinguish between single-loop and double-loop learning.

Single-loop learning occurs when a mismatch is detected and corrected without changing the underlying values and status quo that govern the behaviors. Double-loop learning occurs when a mismatch is detected and corrected by first changing the underlying values and other features of the status quo. Single-loop learning remains within the accepted routines. Double-loop learning requires that new routines be created that were based on a different conception of the universe (Argyris 1999).

Sometimes double-loop learning is equated with Bateson's 'deutero-learning'. Schon and I have expressed our intellectual indebtedness to Bateson (Argyris and Schon 1974). However, we also made a distinction between double-loop learning and deutero-learning. We understood deutero-learning to mean second-order learning, reflecting on the first-order actions. Deutero-learning can occur by going meta on single or double-loop learning. The distinction is important because the knowledge and skills required to produce double-loop learning are significantly greater and more complicated than those required for deutero-learning on single-loop issues.

A second fundamental feature embedded in this view of learning is that learning occurs when understanding, insight and explanations are connected with action. Errors are not simply detected; they are corrected and correction implies action. How do you know when you have learned something? When you can produce in the form of action whatever you claim that you know. This requirement is much easier to meet when the learning is single-loop than when it is double-loop.

A third fundamental feature is that learning always implies effectiveness. Was the action intended actually produced? Did the action produced persevere without harming the present level of effectiveness? Theories of effective learning are normative. Implementing them requires attention to prescriptive knowledge. Little did I know when I started that my concept of effective learning would require normative and prescriptive knowledge: the very two types of knowledge that were, at that time, in disrepute among most scholars studying organizations.

If people create injustices when they are trying to prevent them or to reduce them, then people must be skilled at producing the very conditions that they abhor. If they are skilled at reducing their liberating alternatives, then they are skilled at producing inner-contradictions. Inner-contradictions are produced when the actions lead to positive consequences (the strengthening of liberating alternatives) and negative ones (the weakening of liberating alternatives). For example, the history of employee participation indicates that employees used a theory of interpersonal effectiveness that assured success with routine problems and, at the same time, set the stage for failure around the non-routine, more difficult problems (Argyris 2000, in press).

These self-defeating actions would not persevere unless the contexts in which they exist, for example, organizations, do not sanction and reinforce them. Human beings therefore create systems that support self-defeating actions.

Reflecting, I used to believe that it was valid to assume there were bad guys and helpless guys. The bad guys were those who had power over others. The helpless guys were those 'below' the bad guys. I assumed that it made sense to hold those with power responsible to take the initiative to strengthen liberating alternatives. I also assumed that the helpless victims would respond positively to enhancements in their power and influence to have more control over their destiny.

Both of these assumptions were flawed. When it came to making non-trivial changes in organizations, people with power could not produce them successfully without the early involvement of those who had less power. The expectation that those with lesser power, if empowered, would have the desire to develop skills to use their empowerment to create more effective organizations and liberating alternatives was also flawed. As I learned over the years, those with more and those with less power were equally incompetent to produce liberating alternatives, a conclusion illustrated by Friere (1972).

Those with more power, however, should be expected to take the first initiatives. It also meant that they should learn a new mindset and skills to design and implement such non-trivial changes. It also meant that those with less power should get the same education.

My early research strategies, I hope to show, did lead to organizational changes. However, what I did not predict correctly was how limited these changes would be in scope and perseverance. The weakness of the early research was that it was guided by concepts such as hierarchical power and control and not by a theory of learning. A theory of learning that included those at the top and bottom took a while for me to formulate. Once I did so, some of the early research findings were still valid and some were not. I will illustrate the point below.

There was another blindness on my part. We, as scholars, espoused liberating alternatives. It took me a while to realize that we did so in ways that limited the study and implementation of what we espoused. Management was shooting itself in the foot by using a flawed theory of effective management. We, as scholars, were shooting ourselves in the foot by using flawed modes of theory building and flawed methodologies for research on double-loop issues. Managers exhibited defensive reasoning mindsets to defend their actions and policies. So did scholars. Left to their own respective devices, both groups created conditions, in the case of double-loop learning, where the blind were leading the blind.

The Individual and the Organization

I am getting a bit ahead of my story. When I began my career, my aspiration was to conduct interdisciplinary research. This seemed obvious to me because even a cursory observation of human beings in organizations would lead one to conclude that human action was not organized, according to academic disciplines. Yet, I learned that this 'obvious' observation was not popular among many scholars.

For example, I was asked by the Social Science Research Council to conduct a study to suggest what would lead scholars to recognize the relevance of interdisciplinary research, thereby expanding their attention to go beyond their traditional disciplines. Once the study was completed (Argyris 1954), it was distributed to scholars from several dozen research universities. A meeting was held, chaired by Professor Douglas McGregor. He asked me to begin by presenting what I thought was the most important and surprising finding from my research. I responded that I was impressed by how each of the 'senior' professors valued their own perspectives. However, most of them appeared to be unaware of the work being conducted at the other schools, especially where potential interconnections could be made. The conference was called to enhance the conduct of interdisciplinary research. I wondered how that hope would be realized given the 'stove-pipe' mentality of those present in the room.

I expected to be attacked for what Douglas McGregor called my 'impertinent response'--said in a way to encourage the dialogue. This did not happen. Each senior professor began with some positive comment about the thoroughness of my descriptive research and soon segued into talking about their research. Years later, these defensive routines continue to flourish. For example, after the appearance of Kuhn's (1970) work on scientific revolutions, scholars quote the book as evidence that scholarly community defensive routines were 'natural' in the practice of research. What impressed me about the argument was that knowledge that described the existence of defensive actions was interpreted as natural and not to be changed. Yet, the same scholars identified defensive routines in organizations plus their counterproductive consequences and implied that changing them was necessary to enhance learning and performance.

Nevertheless, I decided to give it a try. The first step was to immerse myself in reading and re-reading my nearly 400-page report. The more I did so, the more challenging the task seemed to become. What kept me going was the realization that the human beings and the organizations that were being studied were faced with the same degree of complexity. Somehow, they were trying to put their arms around the complexity that they were creating. How might I, as a scholar, begin?

I remember having read the work of some physical scientists who were striving to provide an explanation for how the universe evolved. They did so by beginning with hypothesizing a few features that probably existed at the outset. By knowing the properties of each, they were able to conjecture the likely consequences if these features interacted. They repeated the analysis by asking the same question of what would happen if these second-order features interacted.

I decided to experiment with using the same logic (Argyris 1957). I began with the hypothesis that, at the outset, there were human beings and hierarchical organizations. Defining the likely properties of human beings was not easy because there were many different views of the nature of human personality, most of which were developed by clinicians and personality theorists. I finally chose the work done by scholars who studied the development of human beings (social psychologists, anthropologists, and a few sociologists). I made this choice because the properties that they talked about were relatively directly observable and they existed regardless of which personality theory one preferred.

For example, human beings were seen as developing from infant-like stages toward adulthood. The former were characterized by such features as dependency on adults such as parents. Infants also had few abilities, most of which were skin-surfaced or superficial. Adults, on the other hand, strove to become independent of their parents, to take on responsibility for their lives. Through acculturation and education, they developed many abilities, a few in depth, so that they could become more generative.

I did the same type of analysis about organizations. I specified the initial features to be hierarchies and managerial controls. I showed that if hierarchies had their way, they would create work worlds for human beings that were consistent with the features of infancy.

I expanded the conjecture. Those workers who valued adult-like work settings would likely experience a conflict and would likely be frustrated. They would also be in a double-bind. On the one hand, they came to work to earn an income. On the other hand, the price was to accept those infant-like work conditions.

I also conjectured that there would be workers who did not care about jobs with adult-like features. They would not experience the conflict and frustration as long as they earned an acceptable income and a reasonable sense of job security. These predictions were not disconfirmed in a study where employees reported that their biggest concern was money and security (Argyris 1965). They believed that management was meeting these conditions. Also, it is fair to say that as a result of paternalism, management was sensitive not to ignore the 'hygienic' human factors. Not surprisingly, the employees and management reported good relationships. One unexpected finding was that first line managers reported that their jobs were 'simplified'. They faced few leadership challenges, and they had few leadership responsibilities because of the employee--management relationship. The second-order consequences of having simplified non-challenging jobs was accepting the simplification, an act consistent with simplified roles.

The second-order consequences of hierarchical work settings and simplified jobs were that the employees acted to reduce the frustration and dependence on management. They did so by establishing informal systems (e.g. restriction of output, rate-setting, goldbricking) as well as fighting the unfair features of controls (ranging from job designs, to production budgets, to quality control).

A third-order consequence was to activate management fears that the workers could take over. They reacted by reinforcing hierarchy and controls. This, in turn, activated the fears of the workers, who reacted by strengthening their informal work activities. They also created trade unions and fought hard for laws to protect both the inside informal activities and the trade union activities. The fourth-order consequence was that these resulted in self-fueling, self-sealing consequences that created ultra-stable states that inhibited genuine corrective actions.

It was possible to derive strategies for interrupting and reducing these counterproductive consequences. For example, create jobs that had more adult-like features (job enlargement, job enrichment) as well as provide workers with greater control over the work flow requirements through the use of participatory strategies, especially empowered work groups.

The early documented results were positive. The number of organizations (private and public) re-designing their work conditions increased (Argyris 1964). However, as time went by, negative consequences began to appear. For example, management espoused and encouraged employee participation. They could enrich and enlarge their jobs. They could participate in reducing symbols of dependence, for example, the time clocks. However, if the employees attempted to obtain more controls over the power relationships that would reduce dependence, the management often resisted. The resistance was in the form of manipulation in the sense that the management denied the resistance, made it undiscussable, and made the undiscussability undiscussable (Argyris 1964, 2000, in press). The early versions of my theory did not predict these consequences.

I decided that the next step was to engage top management around their basic values. If they did not alter their values, the structural changes would have limited results. Also, there would be an increasing lack of trust between workers and managers. In my mind, this made it necessary to focus on leadership behavior. I had published a study on executive leadership, that showed how managers and their subordinates created counterproductive consequences for each other. They also created consequences that made it difficult for them to learn how to get out of these conditions (Argyris 1954).

Toward a Theory of Effective Action

In other words, managers were not effective in learning how to reduce the very conditions that they decried. Later I found this was also true for the subordinates.

I had to face up to the following doubts about my work. I conducted empirical research that connected the apparent managerial ineffectiveness to the top-down, hierarchical values. However, I also found that subordinates and workers, when given an opportunity to 'take charge' of their own efforts, produced the same counterproductive consequences, yet they did not espouse the hierarchical values of those on top (Argyris 2000, Barker 1999).

This led to the hypothesis that there were more fundamental causes than hierarchical values that remained to be discovered. Donald Schon and I eventually realized that the more fundamental causes were the theories of action that human beings held about the effective management of human relationships. Our research led us to the conclusion that all human beings used the same theory to act regardless of their position on the hierarchy, the type, size, and age of the organizations, their gender, race, education, wealth and the culture in which the organizations existed (Argyris 1982, 1985, 1990, 1993, 2000; Argyris et al. 1985; Argyris and Schon 1996). It took us a while to begin to trust our conclusions. We had been educated to believe that if there was no observed variance, one probably had a flawed theory and/or used flawed empirical research.

As we immersed ourselves in our observations, we differentiated between espoused theories of effective action and the theories-in-use. It was the espoused theories that varied widely. But these theories did not predict accurately what we observed. We proposed that it was the theory-in-use that predicted correctly our observations. We also had overwhelming observations that these theories-in-use did not vary.

Some of the most relevant features of theories-in-use include:

1 Most human beings hold micro theories of effective action that are relatively effective in changing routines (single-loop learning) and relatively ineffective in making more fundamental changes in the status quo (double-loop learning). We have called this unilaterally controlling, emotion-suppressing theory-in-use Model I.

2 Model I necessarily leads to skilled incompetence and skilled unawareness. The skilled unawareness and the skilled incompetence are protected by the use of defensive reasoning, which is at the core of Model I sense-making.

3 Many individuals and organizations espouse a wide range of theories of action. When it comes to theory-in-use, the overwhelming number of them use Model I.

4 Although in organizational research Model I is omnipresent, it would be incorrect to conclude that Model I actions are caused only by the top-down, unilateral hierarchies and control systems. The most fundamental causes are the theories-in-use programmed into human beings early in their lives. Thus subordinates and employees can cause counterproductive, limited learning activities.

Model I theories-in-use depend on defensive reasoning mindsets. The characteristics of defensive reasoning mindsets include:

1 The objective is to protect and defend the actor(s) or supra individual unities such as groups, inter-groups, and organizations.

2 The primary reasoning processes include making the premises tacit, assuming that they are valid, and testing the assumption by the use of self-referential logic. Self-referential logic means using the logic to test a claim that is the same as the logic used to create it in the first place. For example, the claim 'trust me, I know this organization' means that the other should use the same reasoning as the one making the claim.

3 Transparency is avoided in the service of protecting self-denial and self-deception.

4 Self-deception is denied by cover-up. In order for the cover-up to work, it too must be covered up. The strategy used to cover up the cover-up is to make both undiscussable. Undiscussability is protected by making the undiscussability undiscussable.

Human beings who use the defensive reasoning mindset create behavioral systems in organizations that are consistent with and strengthen the above, such as organizational defensive routines. This results in ultra-stable systems that are self-sealing and anti-double-loop learning. Human beings report that they are helpless to make any changes because they do not know what to do and because, as victims, they could not act in ways to reverse the ultra-stable, anti-learning state of their universe.

Please note that these are the same counterproductive conditions reported in the research on the individual and the organization cited previously. We now have an explanation of why the structural changes described previously resulted early on in positive consequences, yet their persistence was limited, and they led to persistent counterproductive, anti-double-loop learning activities. The explanation is that all human beings hold Model I theories-in-use that are consistent with hierarchical values. Give human beings a chance to re-design their work lives to encourage liberating alternatives and we will find that the re-designs are likely to be effective until they are confronted with double-loop issues. The moment double-loop issues are activated, such as fear by management of loss of control or fear by workers that they are being manipulated and victimized, the only mindset related to effective actions that they can use is Model I, whose effectiveness depends upon the use of a defensive reasoning mindset. The same explanation seemed valid for the demise of other organizations such as alternative schools that were intended to create liberating alternatives (Argyris 1974).

Some of the most powerful and somewhat painful proofs of these claims became evident during face-to-face encounters with executives who were attending seminars that we had designed to enhance their leadership effectiveness. The early sessions assumed that if we could diagnose the hierarchical values of executives, and if we could show them convincingly that these values 'caused' their ineffective leadership behavior, and if they then chose to change their hierarchical values, the effectiveness of their actions could be increased. Valid knowledge would make them more competent.

The flaw in this assumption became evident during face-to-face dialogues with executives who attended our seminars in order to enhance their leadership effectiveness. During the seminar, I would feed back to the participants our diagnosis of their leadership effectiveness. I remember how secure and a bit proud we were because the diagnosis was based on a quantitative scheme that had high inter-observer reliability, as well as validity when compared with other measures. I remember how pleased I was that we could produce quantitative results that rank-ordered their actions and predicted the consequences correctly (Argyris 1965).

I also remember that the executives were glad that I was glad, but they were not impressed. First, they could not see how they could move from thequantitative data to prescriptions that were implementable. Second, after we explained how this might be possible, they responded that if they changed their behavior, they doubted that the changes would persist. One executive said, in effect, he would cooperate with the research if it helped my [the author's] promotion. I will never forget my reaction that I kept private. Is this what I meant by creating liberating alternatives? Is this what I meant by strengthening justice?

These doubts were expressed by the 'subjects' in a seminar where they expected to learn how to lead more effectively. Seminars like these required more than the type of debriefing 'subjects' typically received after participating in research intended to enhance knowledge about leadership. We, as scholars, and they, as executives, had created a psychological contract. We, the scholars, were not meeting our promises. They, as 'subjects', did not mind saying so. This pushback by the subjects helped us to seek new ideas and to develop new theories. It was experiences such as these that led us to seek to identify the mindsets in the executives' heads that led them to predict their actions more accurately than we did. This led to the development of the theory of action approach described above.

The Defensive Reasoning Mindset of Scholars Who Document the Defensive Reasoning Mindset of Practitioners

I believe that we, as scholars, hold theories-in-use about effective theory building and effective empirical research that are consistent with Model I (Argyris 1980, 1993; Argyris and Schon 1996). We have our own defensive reasoning mindset to make us skillfully unaware. We also have become skillful at protecting ourselves from facing up to our defensive reasoning mindset. We do so by claiming that all we are doing is acting consistently with the best ideas and norms in the scholarly community to which we belong.

For example, scholars have for years documented the defensive reasoning mindset and organizational defensive routines (Argyris 1980, 1990). Three recent examples come to mind:

1 The defensive reasoning mindset and the organizational defensive routines reported to exist in Intel in 1993 (Burgelman 1993) continued to exist ten years later (Burgelman 2002).

2 The defensive reasoning mindset and organizational defensive routines reported by Van de Ven and Polley (1992) were not only responsible for the failure of the introduction of a new product, but they persisted for several years (the duration of their study).

3 Weick and Sutcliffe (2002) document that many organizations operate over lengthy periods of time by using a mindset that may be described as mindlessness, especially when they are dealing with important and often taken for granted defensive explanations that cause the mindlessness to persist.

One reason why practitioners may not be able to reduce the compulsively repeatable anti-learning behavior required to solve the double-loop problems is that the scholars who document these problems do not conduct research on how to reduce them. Why don't the scholars conduct the relevant research to reduce the ultra-stable anti-learning self-sealing processes that they document? Does this not violate one of the most sacred rules of scientific inquiry, namely the unfettered expansion of valid knowledge? This inconsistency has persisted for decades and shows no sign of waning, even though it violates the basic canons of scientific inquiry.

A defensive mindset appears to be operating when scholars choose not to conduct research on how to reduce the practitioners' defensive reasoning mindset. If true, what are the causes of using such defensive reasoning mindset? Also, what are the causes of its persistence? I suggest three possible explanations. They are (1) the scholarly commitment to describe the universe as-is, (2) the scholarly commitment to internal and external validity at the expense of implementable validity, and (3) the methodologies in good currency of normal science and of interpretive, humanistic research.

First, the fundamental norm of scientific inquiry is to describe the chosen universe as completely and accurately as possible. The test of the validity of the knowledge produced is that predictions are made and tested robustly. There are two features of defensive reasoning embedded in this norm. The first is that the focus upon the universe as-is tells us little about how to create a universe where defensive reasoning is not rewarded. If the universe is dominated by Model I theory-in-use, and if Model I requires that effective action be causally connected with defensive reasoning; and if defensive reasoning produces defensive norms that result in the ultra-stable state that prevents new and different universes (e.g. those that reward double-loop learning), then there will be little attention paid to creating new, and up to this point rare, universes.

The Behavioral Theory of the Firm comes to mind. The core of the theory is its relational variables. For example, quasi-resolution of conflict, uncertainty avoidance, problematic search and limited organizational learning (Cyert and March 1963). To my knowledge, scholars of that persuasion have not conducted research to see how defensive reasoning mindset and organizational defensive routines contribute to these variables. Nor have they conducted research on how to reduce the counterproductive consequences on these variables, for example, on limited learning (Argyris 1996).

The second feature is that if research is not conducted on how to overcome the present status quo, then we will never know how the status quo feature of the universe as-is will react to serious attempts to change it. Yet, this knowledge is part of a complete description of the world as-is. This means that we will have not only lost an opportunity to create new universes, but we will also have lost the opportunity to describe important features of the as-is universe when the as-is features are genuinely threatened.

The second explanation is related to the assumptions that all scholars should produce knowledge that exhibits internal validity and external validity. The former means that scholars have done their best to minimize the confounding influence of unrecognized variables upon the causal claims made and the results produced. External validity means that the results obtained in the context of inquiry are relevant beyond that context.

The defensive reasoning embedded in this rule is illustrated by the many empirical studies of variables such as 'concern', 'trust', and 'openness' that meet the requirements of internal and external validity, yet the knowledge produced exhibits low implementable validity. Propositions that exhibit implementable validity specify the behaviors and their sequences required to achieve the intended consequences. It is one thing to conduct research that shows that trust is important in decision-making. It is quite another to produce propositions that tell the mind how to create trust. The former is about knowledge that indicates understanding and explanation. The latter is about how to produce trust (procedural knowledge). It is implementable validity that provides the most robust test of the actionability of the propositions (Argyris, in press).

There are two currently accepted views on conducting research. The most widely held view is normal science research. The defensive reasoning embedded in this view is surfaced when we realize that the theory-in-use of rigorous research (experimental or correlational) is consistent with Model I. It is the researchers who are in unilateral control over the methods used; who define the criteria of victory; and who design research conditions that minimize the expression of feelings, especially of doubt, bewilderment, and at times anger at being dominated by researchers in the name of validity (Argyris 1980, 1993).

The expression of these feelings is rarely encouraged by researchers who remain within the status quo of normal science and who do not take implementable validity seriously, especially around double-loop problems. Imagine trying to conduct research where the subjects are asked to take seriously new and rare universes; to learn the new skills that will be required; to reflect critically upon their skilled incompetence and skilled unawareness. In our experience human beings do not find it easy to achieve such learning. Lammers (1974) suggests that few scholars are taught the interpersonal skills required to engage the subjects' defenses that arise under these conditions. I believe that he is correct. I also believe that scholars have the responsibility to learn these skills if they are not going to inhibit the study and implementation of double-loop change and implementable validity.

The second mode of research is often described as the interpretive, humanistic, where telling stories is a key methodology. I analyzed many such stories and concluded that when they are told consistently with high standards of story telling, they are likely to be valid stories about the universe as is. These stories are in the form of espoused theories; they have little to do with theory-in-use. The interpretive mode has its own criteria for internal and external validity. However, I could not find examples where such research exhibited implementable validity, especially about double-loop issues such as reducing the defensive reasoning mindset and the defensive routines in organizations and scholarly communities.

I was able to find examples of defensive reasoning mindset operating (Argyris, in press). Martin (1990) tells a story of an officer of a company exhibiting sexism in the way he attempted to honor the efforts of a loyal female employee in the midst of her delivering her child. Martin documents the probable sexism in several ways. The defensive reasoning surfaces when Martin suggests that the way to correct or prevent such errors is to define new human resources policies. Such advice permits the officer to distance himself from his personal causal responsibility for acts that are sexist. It also makes it less likely that the loyal female employee will face up to her discomfort, if any, when being treated unfairly by the company for behavior that the officer did not recognize as sexism. Martin's case illustrates the officer's skilled incompetence and skilled unawareness. Yet, she appears to bypass advice that may help the officer reduce both of these features.

Closing Comments

My research endeavors have had two major themes. One is the focus on double-loop learning and implementable validity. Two is questioning the ideas and norms in the scholarly community about theory building and empirical research. The challenge has not been about getting rid of the ideas in good currency but to build upon them.
Some of the most important movements are:

From an emphasis on:                  Toward an emphasis to include:
single-loop learning                  double-loop learning
internal and external validity        implementable validity
normal science causality              design causality
understanding and explaining          producing action
describing the universe as is         creating new and rare universes
propositions that are descriptive     propositions that are normative
                                        and prescriptive
generalizable propositions            propositions that are applicable
                                        in the individual case
seeking knowledge that is as          seeking knowledge that is
  complete as possible                  designedly incomplete,
                                        supported by theory and
                                        methods to fill in the gaps

This perspective made it less necessary to take an either/or stance on these features. For example, in my work, I adhered to the positivistic rules of testing claims through the process of disconfirmation. I also was critical of positivistic statistical procedures when they were not actionable in everyday life. For example, it is not likely that human beings can produce (i.e. create) in everyday life curvilinear relationships.

In order for human beings to create the required procedural knowledge to produce curvilinear phenomena, they would have to revert to the procedural knowledge the scholars used to generate the curvilinear relationships in the first place. Such procedures require 'conditions' that are not restricted by everyday events. Moreover, even if they could be produced, the propositions would be consistent with Model I theory-in-use that is embedded in the methodologies used. This would exclude double-loop learning. Elsewhere, I have presented illustrations of humanistic-interpretive research to show that the propositions are not supportive of the human objectives that they espoused (Argyris 1993).

I have had many experiences where the positivistic saw me as being too humanistic and where humanists saw me as a closet positivist. This was frustrating. However, it rarely led me to feel personally attacked. It did lead me to continue my efforts to show how features of these positions could be integrated. This appeared to frustrate representatives of each side because the integration would produce positive results where conducting research on double-loop issues and when implementable validity was valued. Neither side was conducting such research (Argyris, in press).

Emphasizing implementable validity led to the distinction between internal and external implementable validity (Argyris, in press). Internal validity occurs when scholars implement their ideas consistently with their theories. External implementable validity occurs when scholars implement their ideas with practitioners.

In my opinion, the history of Tavistock Institute produced knowledge with high internal and external implementable validity. For example, the pioneering research of Frank Hellers, Elliott Jaques, Ken Rice and Erick Trist are a few names that come to mind. Ironically, Tavistock also provides illustrations of how theory can be used to distance the practitioners from the scholar, the very conditions that the founders of Tavistock abhorred. I remember participating in one of their inter-group conferences. At one point, I asked our leader for the reasoning behind a claim that he made. He responded that my behavior illustrated hostility towards the leader. I then said that I was bewildered by his inference. He responded that this comment illustrated an attempt to remove the leader. Many of my fellow participants, all of whom attended because they, like myself, felt that the inter-group ideas were fresh, concluded that even if the ideas were valid, they could not be used in organizations. The message of this story is that these clinical interventions might be useful in a clinical setting where there are clients who are mentally disturbed and require therapeutic help. The participants did not attend the workshops because they felt that they required such help.

In closing, I have had good fortune assisting me throughout my endeavors. The first was by my design. I sought to emphasize the importance of theory and to connect the theory with my deepest feelings. This led me to an ongoing dialogue with my ideas, especially unrecognized gaps and inconsistencies. This also kept me attentive to the theories and research of others that could point out possible gaps and inconsistencies, as well as ways to correct them. May I say that my attention went beyond the philosophy of science in my field (see, for example, Newton-Smith 2000).

The second bit of good fortune was not by my design. I refer to the information science revolution including such components as cognitive psychology and sociology, as well as cognitive neuroscience. A theory of action perspective is consistent with many perspectives in the information sciences. For example, that action is produced through the use of knowledge stored in the mind/brain; that this knowledge is in the form of designs-in-use; that the fundamental feature of designs-in-use is design; and that the propositions produced should be generalizable and, at the same time, applicable to the individual case.

In sum, the commitment to the development of actionable knowledge, in the service of one's core values and feelings, with the intention to help human beings and organizations to act more effectively in the 'real world', and with the ultimate concern for justice, has helped to make life full of learning and fun.


Argyris, C. 1954 The present state of human relations research. New Haven, CT: Labor and Management Center, Yale University.

Argyris, C. 1957 Personality and organization. New York: Harper Brothers.

Argyris, C. 1964 Integrating the individual and the organization. New York: Wiley.

Argyris, C. 1965 Organization and innovation. Homewood, IL: Irwin.

Argyris, C. 1974 'Alternative schools: A behavioral analysis', Teachers College Record 75(4): 429-452.

Argyris, C. 1980 Inner contradictions of rigorous research. San Diego, CA: Academic Press.

Argyris, C. 1982 Reasoning, learning and action: Individual and organizational. San Francisco, CA: Jossey-Bass.

Argyris, C. 1985 Strategy, change and defensive routines. New York: Harper Business.

Argyris, C. 1990 Overcoming organizational defenses: Facilitating organizational learning. Needham, MA: Allyn and Bacon.

Argyris, C. 1993 Knowledge for action. San Francisco, CA: Jossey-Bass.

Argyris, C. 1996 'Unrecognized defense of scholars: Impact on theory and research', Organization Science 7(1): 79-87.

Argyris, C. 1999 Organization learning, 2nd edn. Oxford: Blackwell.

Argyris, C. 2000 Flawed advice. Oxford: Oxford University Press.

Argyris, C. in press 'Actionable knowledge', in Organization theory as science: Prospects and limitations. C. Knudson and H. Tsoukas (eds). Oxford: Oxford University Press.

Argyris, C., and D. Schon 1974 Theory in practice. San Francisco, CA: Jossey-Bass.

Argyris, C., and D. Sch6n 1996 Organizational learning IL Reading, MA: Addison-Wesley.

Argyris, C., R. Putnam, and D. Smith 1985 Action science. San Francisco, CA: Jossey-Bass.

Barker, J. R. 1999 The discipline of teamwork. Thousand Oaks, CA: Sage.

Burgelman, R. A. 1993 'A process model of internal corporate venturing in a major diversified firm'. Administrative Science Quarterly 28: 223-344.

Burgelman, R. A. 2002 Strategy is destiny. New York: The Free Press.

Cyert, R. M., and J. G. March 1963 A behavioral theory of the firm. Englewood Cliffs, NJ: Prentice Hall.

Friere, P. 1972 Pedagogy of the oppressed. Harmondsworth: Penguin Books.

Fromm, E. 1955 The sane society. New York: Holt, Rinehart & Winston.

Kubie, L. S. 1958 Neurotic distortion of the creative process. Lawrence, KS: University of Kansas Press.

Kuhn, T. S. 1970 The structure of scientific revolution, 2rid edn. Chicago, IL: University of Chicago Press.

Lammers, C. J. 1974 'Self management and participation: Two concepts of decentralization in organizations'. Organization and Administrative Science 5(4): 17-33.

Martin, J. 1990 'Deconstructing organizational taboos: The suppression of gender conflict in organization'. Organization Science 1(4): 339-359.

Newton-Smith, W. H. 2000 A companion to the philosophy of science. Malden, MA: Blackwell.

Van de Ven, A., and D. Polley 1992 'Learning while innovating'. Organization Science 3(1): 92-115.

Weick, K. E., and K. M. Sutcliffe 2002 Managing the unexpected. San Francisco, CA: Jossey-Bass.

Cris Argyris is the James Bryant Conant Professor Emeritus of Education and Organizational Behavior at Harvard University. He received his PhD from Cornell University in 1951, and has received 11 honorary doctorates. The Chris Argyris Chair was recently established at Yale University. His current research interests are learning, leadership, theory methods of intervention and organizational change.

Address: Monitor Group, Two Canal Park, Cambridge, MA 02141, USA. E-mail:

Chris Argyris


University, USA

and Monitor, PLC
COPYRIGHT 2003 Sage Publications, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2003 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Argyris, Chris
Publication:Organization Studies
Date:Sep 1, 2003
Previous Article:Vita Contemplativa.
Next Article:John Hassard, Ruth Holiday and Hugh Willmott (eds): Body and Organization.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters