Building the Learning Consortium
Edgar H. Schein
MIT Sloan School of Management
March 30, 1995
Good Evening Friends and Colleagues. I am really sorry I can't be with you in person, but I find myself in the midst of a learning experience of my own which requires me to stay in San Francisco.
Now on to tonight's topic--why build a learning consortium?. I would like to begin this talk by having you consider with me the question of why transformational learning is so difficult?
By transformational learning I mean a learning process that examines and evolves the present assumptions on which the organization operates, and develops new organizational practices based on new assumptions.
It is important to distinguish knowledge from practice. Knowledge can usually be articulated; by practice I mean tacit skills that can usually not be articulated. It is the difference between the theory of how to ride a bicycle and the bodily know-how of actually being able to ride.
For transformational learning to have occurred, we must have more than new knowledge and new insights. The new knowledge must become embedded in tacit organizational practices that become routine.
One could contrast transformational learning with adaptive coping, where we try to do better in implementing the present set of assumptions, but do not really examine or challenge them. That is, we improve and perfect the practices that are built on the old assumptions, but continue to take those old assumptions for granted as the correct way to think about things.
So we start with the question of why it is that organization after organization first exposes itself to new revolutionary ideas like employee empowerment, total quality, visionary leadership, networking, re-engineering, information technology, and all the other concepts that academics and consultants have been touting as the way of the future, then launch crash programs to implement them, yet conclude a year or so later that the ideas did not work out.
Why do so many programs of total quality and re-engineering have such high reported failure rates?
Why is it that the idea of empowerment, of involving people in the changes that will affect them, an idea which has been around at least since the Hawthorne studies of the 1920's, has to resurface about every two or three decades under new kinds of labels and still only takes hold here and there? Or, worse, why do many organizations claim to empower their employees, yet their day to day managerial practices change hardly at all?
Why is it that McGregor's Theory Y, the assumption that people are willing and able to work and to make contributions to organizations, an idea he espoused in 1960 in his classic The Human Side of Enterprise is still overridden in most organizations by the cynical assumptions of Theory X that people are basically lazy and have to be motivated and controlled by management?
There are basically two answers to these questions.
First, the more cynically based, conservative model of human nature and organizations, what many have called the "command and control" model, has worked. Organizations have been able to function effectively with cynical views of human nature and have thrived on the assumption that hierarchy is a necessary and sufficient mechanism of control and coordination.
And one reason, they have been able to function with those assumptions is that they have existed in fairly predictable and stable environments, in containable markets that changed slowly, with technologies that evolved slowly, and in economic-political environments that allowed success even with relatively low levels of efficiency and effectiveness.
But, as we all know, environments are today changing at an ever accelerating pace, technologies and markets are changing more and more rapidly, and globalism is making it necessary to compete with other organizations that are far more productive and effective. The need to change, the need for organizational transformation is demonstrably much higher today than it was ten, twenty, thirty years ago. Yet when we look at the results of todays' efforts at organizational transformation, we still find relatively few organizations that are successfully managing such transformations. So there must be additional reasons for the slow rate of this kind of learning.
This second reason has to do with organizational culture and the national cultures within which organizations are embedded. Those of us who have studied organizational cultures in some detail have found over and over again that, at best, cultures evolve very slowly, and, short of the massive destruction and rebirth of entire organizations, one sees very little fundamental culture change.
I am now talking about real culture change, not what so many espouse by casually labelling every remedial program in their organizations as the "creation of a new culture."
When one examines those claims closely, one usually finds either that the changes are minor adaptations within a given culture or, if the espoused changes really challenge an organization's basic assumptions, they are resisted, subverted, or misunderstood. In other words, the second reason why we see so little organizational transformation is that such transformation usually requires us to give up some deep cultural assumptions and build organizations on new assumptions. And that kind of process of unlearning and relearning is inevitably painful and slow.
How does such learning occur? What are the necessary and sufficient conditions for unlearning and relearning? To understand this we need to examine some of the psychological dynamics of anxiety. When our old models no longer work, when we experience dis- confirmation or lack of confirmation, we become anxious, call this Survival Anxiety. It is the anxiety that unless we change, we will not meet our goals, achieve our ideals, or, at the extreme, survive at all.
However, the prospect of giving up some of our assumptions and tacit practices also causes anxiety, call this Learning Anxiety. This is the anxiety that if I let myself become a learner I will become incompetent and, at worst, lose my identity. The more I experience learning anxiety, the more I will defensively deny and disregard the disconfirming data and cling to my old ways.
This process is what we call resistance to change, and it is entirely normal and to be expected. Now if we are logical we will realize that for a learning process to begin Survival Anxiety has to be greater than Learning Anxiety.
Our usual approach to making this happen is to escalate the Survival Anxiety to such high levels that the potential learners become even more defensively paralyzed, deny or rationalize away the disconfirming data, or go through the motions of a learning process just to get management pressure off their back.
An alternative and better method of launching the learning process is to notice that Survival Anxiety will be greater than Learning Anxiety if we reduce the Learning Anxiety. This process I have called the creation of Psychological Safety--to make the learner feel that it is possible to learn without loss of self.
The problem is that the creation of enough psychological safety to overcome our Learning Anxiety is usually very difficult, especially when we demand increased productivity from the learner at the same time. Psychological safety is dramatically absent when our organizations are downsizing, reorganizing into flatter networks, and trying to become "lean and mean."
To feel psychologically safe we need some time and space to become reflective to let the disconfirming data into consciousness, we need training in how to learn, supportive coaches, practice fields, and positive visions of the future.
What most organizations have figured out as they face this set of constraining conditions, is that they must create a "parallel system." They must empower some group or steering committee or task force to undergo some transformational learning of its own, and, to do that, they must give the group some time and room. Such a group then must make itself partially marginal in order to set up the conditions for learning. I will call this Stage 1 of the transformational learning process.
The role of the parallel system is 1) to provide psychological safety, 2) to create practice fields in which errors can be made, embraced, and learned from, 3) to expose the learners to the truly transformational concepts by bringing in academics and consultants to begin the process of cognitive re-definition, and 4) to create the conditions for testing the new concepts and develop the practices that flow from them in the context of their own company cultures.
By cognitive re-definition I mean the process of gradually transforming one's mental model by first of all recognizing that one's concepts are to some degree one's own learned constructions, not some arbitrary external reality. What I mean by systems thinking or Theory Y is my own learned definition which will differ in unknown ways from your learned definitions. It is here that Dialogue becomes so important as a vehicle for creating more common understanding.
When we then encounter a new academic theory or concept we have to broaden our own definitions and sometimes even replace them. I might think of Theory Y as participative management, only to realize that what McGregor meant was something quite different, a basic assumption that people want to work and want to link their own aspirations to those of their organizations when this is possible. I might be an occasional autocrat if the task and circumstances demanded it, and would realize that autocratic behavior does not necessarily imply mistrust of my subordinates. My concept of Theory Y has now been broadened.
Another aspect of cognitive re-definition is that we change our anchors and standards of judgment. I might believe that the best way to succeed is to compete with my peers for a promotion--the ultimately best way to be individualistic. A theory of teamwork might note that winning out over others is a very limited strategy and a better strategy might be to build an effective team with my peers so that we would collectively win out over our competitors. I might then shift my evaluation of my own subordinates from rewarding competitive behavior to rewarding team building behavior, which might be a considerable shift.
But unfortunately we have learned that even in the parallel system with more psychological safety, the new ideas, models, concepts, and assumptions will be hard to understand, and the skills needed to implement them will be even harder to learn.
Think about the cognitive shift that is involved in moving from linear cause and effect thinking to systems thinking, the shift that is involved in moving from a Theory X, Command and Control model to an organization in which fluid groups set their own timetables, distribute work according to their own knowledge of who can do what, and are collectively accountable for results.
Think about the new attitudes that managers of such organizations have to develop and the skills they need, to manage such loose and shifting boundaries. Think of the career system and reward system that would have to be invented to work effectively with such shifting groups and boundaryless jobs.
Under these conditions what is likely to happen is that the members of the parallel system only partially learn what the new concepts really imply and consequently develop practices that do not really work.
The best example of this was the abortive effort to develop quality circles in U. S. companies based on an incomplete understanding of what the Japanese meant by quality and what role groups played in the measurement of it.
One reason for such lack of understanding is that even the parallel system is still part of the broader managerial and national culture, and filters the new concepts through those cultural assumptions. "You say we have to have teamwork. OK we will train ourselves in teamwork and exhort everyone to become teamwork oriented, but don't ask me to give up the individually oriented performance appraisal system, or the individually based bonus system, or the concept of individual accountability. We still have to hold the chair of the team accountable." The irony is that often the learners themselves do not see the inconsistency.
Or alternatively, if the members of the parallel system really do get it and undergo some transformative learning of their own, they often discover that they have become aliens in their own culture and cannot figure out how to get the new practices instituted in the home culture. The parallel system then is either disbanded or gradually re-learns the old cultural assumptions.
The lesson is that a parallel system is necessary, but not yet sufficient for real transformational learning to occur. This set of circumstances brings us to the need to create the Learning Consortium.
We need a Second Stage learning process. The members of the parallel system who are now on the road to transformative learning need, in effect, their own parallel system.
They need to be able to talk with others from other organizations who are undergoing similar transformative experiences, compare notes, get emotional support and encouragement, get advice, and, in the extreme even get training. You may well ask why the academic coach or consultant cannot provide this training and support to the original parallel system. I believe the answer lies in the fact that the cultural gap between the coach or consultant and the members of the organization trying to learn is too great.
Even if the ideas get across, the academic coach or consultant rarely understands organizational culture well enough to help the learners develop new practices that can fit into the culture and can slowly evolve it to new assumptions.
On the other hand, fellow learners who are facing similar problems in similar but different organizations often have the empathy and know- how to teach each other. The academic coach or consultant is still needed to set up the conditions for this kind of mutual learning in the consortium, but it is the peer group that actually functions as both learner and teacher/coach.
The learning consortium then can be thought of as fulfilling at least three critical learning functions that are unique to the consortium:
1) Translation of the academic ideas into the language of management, thereby further facilitating cognitive redefinition;
2) Providing long range emotional support and empathy for its members to deal with the frustrations inherent in implementing the ideas within any given organization;
3) Providing mutual education and training on how the new ideas can be implemented in the different organizational cultures that are represented in the organization set.
Notice what this two-stage model implies. First, it implies that no organization can go it alone. The problems of transformation are sufficiently complex that organizational sets have to work together to create a learning field for themselves. But each organization still has to function, so it has to work through creating stage 1 parallel systems to launch the learning process.
And then the organizations have to work together to create a viable stage 2 process by launching and supporting a consortium in which their members can begin to learn together. The existence of rudimentary learning consortia is certainly not new. Every university based executive development program that is run for multiple companies is a kind of learning consortium.
International conferences, trade association meetings, management institutes, assorted round tables that have a set of member sponsoring companies, and so on are all potential learning consortia.
But the member companies who send delegates rarely think through the possibilities in having their delegates first become learning teams and then creating opportunities for those teams to work with each other. Typically one or two people go off to learn something and then come home with no mandate beyond "reporting" to their boss what they have learned.
The ideas may be written down and even circulated so others become informed on what new stuff is going on. But this process does not even meet the minimum conditions for stage 1 learning because the "back home" members are rarely feeling disconfirmed or anxious enough to take what the learner brought back seriously.
For a learning consortium to work, there first needs to be an organizational set, a group of organizations that jointly commit themselves to some transformative learning. One of the central assumptions underlying the formation of MIT's OLC is that the partner companies begin to see themselves as a set working together to help each other along in the learning process.
The creation of the liaison officers group that meets quarterly was premised on this assumption. This group was originally formed to provide an ongoing interface with MIT, to provide a structure for companies to get acquainted with each others' projects, and, hopefully to provide help to each other by collaborating on projects, training and coaching each other, and providing a base of support for the whole enterprise.
Peter Senge's original model called for the creation of partnerships not only with MIT but among the learning companies themselves so that eventually they could carry the projects without direct MIT intervention. The members of the group have all gone through the basic competency course and were individually on a personal learning track. They were therefore capable of contributing significantly to questions of how to design meetings, learning projects, and learning tools.
We found that by sharing our problems such as how much to charge companies, how to describe the deliverables, how often to meet, how to design the semi-annual meetings and their own meetings we not only got significant help, but stimulated the group to become effective.
As the group matured, the unique functions of the consortium kicked in with increasing incidents of members addressing fundamental learning questions not to the MIT faculty but to fellow group members.
Sharing of resources and stories is becoming more and more common in this group. The kind of learning that occurs here can be thought of as "lateral diffusion" in that the members represent middle levels of their organizations. They can discuss how to penetrate other hierarchical boundaries and to design processes for those other levels to become learners.
We learned that such a group could indeed become each others' coaches and gradually take on responsibility for their own learning and begin to design their own learning experiences. This process has been most evident in the last year or so when this group took on the task of designing the semi-annual meeting at which all of the partner companies, the extended faculty and research staff, and selected visitors and guests come together for two to three days of sharing project experiences, mutual exploration of what is going on in each others' companies, focused learning on selected topics, and an in-depth visit to the host company facility.
The success of these meetings has been closely connected to the design efforts of the liaison officers group functioning as a learning consortium.
The semi-annual meeting can also be viewed as a learning consortium. Not only do the attendees hear more formal reports of progress, problems encountered, and ideas generated, but during the informal time networks get built, advice is sought, plans are made for visits to each others' companies, and war stories are told to bring home important lessons.
At the semi-annual meeting the learning community consists of faculty, consultants, research staff, graduate students, and company representatives. At this meeting the companies can send people who are not yet launched as learners but who want to find out more about the whole enterprise and "test the waters" so to speak.
A third kind of learning consortium is the "Leaders Group," a set of five CEO's from partner companies who meet quarterly for two and a half days with one or two faculty and staff. They operate with a fairly open agenda built around periods of dialogue and some focused learning exercises or topical discussions.
For example, the group spent an afternoon with Karl Henrik Robert, the Swedish doctor and environmentalist and the U. S. president of IKEA, the furniture company that is using many of his ideas. The group uses the meetings to reflect, to reframe, and to regenerate their own commitment to learning.
One salient thing that was discovered early by the group is that formal power is not enough to start a learning process in their organizations. Everyone of them expressed great frustration about their inability to get others to buy into their own strongly held values and the amount of resistance to change they encountered.
But by sharing these frustrations they were also building a common frame of reference and a language that permitted them to begin to truly understand and help each other. It is very noticeable at the meetings how much business gets conducted during the informal time and how meaningful it is to the members to be able to let their hair down, seek help, and admit their frustrations.
As their own understanding of the learning process increases they will not only be able to help each other but become more supportive of the middle management layers and encouraging to leaders in other organizations. Though this consortium is small, it is clearly fulfilling its key functions.
Recapitulation
The basic argument is simple. A stage 1 process of creating a parallel system creates some of the conditions for learning. But the cognitive transformation that is required for new concepts to be truly absorbed and developed into new practices requires learning from a peer group and coaching from others who are culturally more similar to the learners. For this condition to be met a learning consortium must be created and nurtured. And out of such a consortium, truly transformational learning will become possible.
Thanks very much for your attention and good luck in your own learning efforts.