According to Tsoukas (1998: 293), the sciences have historically set the tone in intellectual inquiry. Furthermore, there seems to be a fundamental human urge to want to understand both nature and society as a unified entity. Tsoukas (1998: 293) justifies the appearance of a new scientific approach, complexity theory: “If nature turns out to be much less deterministic than we hitherto thought …then perhaps our hitherto mechanistic approach to understanding the messiness we normally associate with the social world may need revising.” Tsoukas (1998: 291) states that the Newtonian, traditional or mechanistic style is gradually receding in favor of the complex, holistic or emergent style, characterized by the ability to notice instability, disorder, novelty, emergence, and self-organization. Indeed, an increasing number of academics have started to use complexity theory to better understand organizations.
Complexity theories, generally referring to ideas and concepts at a distance from the mechanistic view, represent a research approach that makes philosophical assumptions about the emerging world view, which include wholeness, perspective observation, non-linearity, synchronicity, mutual causation, relationship as unit of analysis, etc. (Dent, 1999). The word “complexity” originates from the Latin word complexus, meaning comprehension and wholeness; complexity theories explore the totality (the wholeness) of dynamics – forces, energies, substances and forms – permeating the whole universe and connecting everything that exists in a whirling web of dynamic interrelationships and interactions (Dimitrov, 2003).
Complexity theories are increasingly being seen by academics and practitioners as a way of understanding organizations and promoting organizational change (Burnes, 2005: 74). This is so because complexity theories deal with the nature of emergence, innovation, learning and adaptation (Houchin and MacLean, 2005). In spite of the potential importance of complexity theories for organizational learning, only a few attempts have been made to improve our understanding of organizational learning based on these ideas (e.g. Stacey, 1996; Eijnatten and Putnik, 2004; Antonacopoulou and Chiva, 2007). However, none of these papers analyze or improve our understanding of adaptive and generative learning within organizations. In this paper, we put forward a new understanding of the two types of learning grounded in some ideas from complexity theories.
Complexity theories serve as an umbrella term for a number of ideas, theories and research programs that are derived from a range of scientific disciplines (Burnes, 2005: 73). Consequently, and according to this author, there is not one theory, but a number of theories (Chaos theory, wholeness theory, dissipative structures, fractals, complex adaptive systems, etc.) developed by different scientific disciplines, which are gathered under the general heading of complexity research. In fact, most of the papers that use complexity theories to improve our understanding of organizations select a few terms, concepts or ideas, which are assumed to be essential in that analysis (e.g. Houchin and MacLean, 2005). In this paper we mainly focus on two concepts: self-organization (Kauffman, 1993; Gell-Mann, 1994); and implicate order (Bohm, 1980; Bohm and Peat, 2000). These two concepts were chosen because they are essential in learning processes: complex adaptive systems learn through a self-organizing process (Kauffman, 1993; Gell-Mann, 1994); on the other hand, Bohm (1980) considers learning and creativity as the search for and representation of a new order.
Based on these concepts, we propose and explain some characteristics that describe both adaptive and generative learning. Through these characteristics we explain the process of generative and adaptive learning and make certain conceptual suggestions to better understand and foster these processes. Finally, we include both types of learning processes within the OL framework. With the aim of obtaining new insights from complexity theories, we will follow a metaphorical approach (Tsoukas, 1998; Tsoukas and Hatch, 2001; Houchin and MacLean; 2005), which avoids searching for common principles across a variety of very different systems (physical, social etc.).
Generative learning is a process that involves searching for (implicit) order, which is a holistic understanding of anything or anyone we interact with (holo-organization). When enacted or interpreted (unfolded), this implicate order becomes a new explicate order, or the manifested world, which is represented through mental models, paradigms etc. Adaptive learning involves any improvement or development of the explicate order through a process of self-organization. Generative learning is developed individually or socially at the edge of chaos, through intuition, attention, dialogue, and inquiry.
Based on these two conceptualizations, we consider learning as any change (incremental or radical) in the explicate order (individual or social). OL implies that a new or improved organizational explicate order has been developed.
In pursuing this analysis, we first provide an overview of the adaptive and generative learning typology in the OL literature. We selected the main works that explain their importance, describe them, analyze their facilitators, and incorporate them in the OL process. Secondly, we analyze the main works that explain the concepts selected from complexity theories: self-organization and implicate order. Although we mainly focus on the complexity literature, we also take into account organizational literature that has applied complexity ideas. Based on these ideas selected from complexity theory, we then present the process of generative and adaptive learning within organizations, their essential catalyzers, and a model of OL that incorporates both types of learning. Finally, we discuss the main implications of the two types of learning for OL.^
As Shipton (2006: 233) affirms, the study of OL is no longer in its infancy. Since the first works in the sixties (Cyert and March, 1963; Cangelosi and Dill, 1965), researchers have focused on different aspects of learning in organizations, in an attempt to find answers to questions such as: What does OL mean? How does OL take place? Who is learning? What is being learnt? What factors facilitate or inhibit OL? or Are there different types of OL? In order to better understand our thinking of learning in organizations, different typologies and classifications of OL research have been put forward (e.g. Miner and Mezias, 1996; Örtenblad, 2002; Elkjaer, 2004; Shipton, 2006). Recently, Shipton (2006) analyzed the whole body of OL literature through two typologies: prescriptive vs. explanatory and individual vs. organizational. The first typology differentiates between a more prescriptive, normative and practically orientated literature; and a more explanatory, descriptive, skeptical literature, centered on understanding the nature and processes of learning (Tsang, 1997). The second typology examines the level of analysis: either individual or organizational. The former considers OL to be mainly an individual activity taking place within organizations and that it emerges naturally from day-to-day practices (Simon, 1991). The latter perspective considers OL to be more than the learning of its individual members, and focuses on systematic and planned efforts to capture, share and apply the insights of the individuals and the groups to which they belong (Zollo and Winter, 2002).
However, one of the most recurring classifications used by researchers is the distinction between adaptive and generative learning (Senge, 1990). Miner and Mezias (1996: 88) explain that in the OL literature there are two streams of work: incremental and radical learning. The former, described by Cyert and March (1963), considers firms as incremental or adaptive learning systems in which routines and the firm’s adapting behavior are essential for learning (Miner and Mezias, 1996). The second stream, based on Argyris and Schön’s (1974; 1978) distinction between single and double loop learning, stresses the importance of the latter for organizations. Single loop learning implies the ability to detect and correct errors in certain operating procedures, whereas double loop learning implies being able to see beyond the situation and questioning operating norms. Single loop learning is like a thermostat that learns when it is too hot or too cold and turns the heat on or off (Smith, 2001). Single loop learning seems to be present when goals, values, frameworks or strategies are taken for granted. It is about efficiency. Double loop learning occurs when error is detected and corrected in ways that involve the modification of an organization’s underlying norms, policies and objectives (Smith, 2001). Miner and Mezias (1996: 89) point out that most papers support the importance of both learning streams in organizations.
Argyris and Schön’s (1974, 1978) distinction was probably based on Ashby (1952) and Bateson (1972), as they proposed similar concepts of learning. Practically at the same time as Argyris and Schön, very similar typologies were suggested by authors like Piaget (1969), Kuhn (1970) or Watzlawick et al. (1974), among others. Piaget (1969) discovered that children learn in two different ways. First, they can learn through “assimilation”, when a new fact is understood through a previous model. A different type of learning is needed when a new fact cannot be assimilated through a previous model. In this circumstance, children need to “accommodate” or change their model to a new reality. These two kinds of learning could be related to single and double loop learning, respectively. Similarly, Kuhn (1970) describes the evolution of science as a succession of “paradigm shifts”, each one completely reorganizing the mental models of the community of practitioners of a certain scientific field. Kuhn (1970) makes a clear distinction between what he calls “normal science”, where scientists only “solve problems” by expanding the old theory to apply it to new facts, and what he calls “scientific revolutions” where a scientist creates a completely new model to explain reality. In the same way, Watzlawick et al. (1974) distinguishes between two types of change. First order changes are incremental changes made within the system, the rules of which are not changed. In contrast, second order changes imply that the rules of the system are challenged and changed. They are no longer changes within the system, but changes of the system itself. In summary, all the divisions these authors propose show that this distinction is generally accepted, not only in the OL literature.
Argyris and Schön (1974, 1978) appear to have introduced the distinction between adaptive and generative learning into the OL literature; however they are not the only authors to consider these types of learning. Senge (1990), Lant and Mezias (1992), Virany et al. (1992), Sitkin (1996) or Fiol and Lyles (1985) mention and analyze the existence of these two types of learning in organizations.
Fiol and Lyles (1985: 807) differentiate between lower and higher level learning. The former is a focused learning that may be mere repetition of past behaviors, adjustments in part of what the organization does. Higher level learning is related to the development of complex rules and associations regarding new actions.
Senge (1990) distinguishes between adaptive and generative learning. He affirms that generative learning, unlike adaptive learning, requires new ways of looking at the world, whether in understanding customers or understanding how to better manage a business. In order to look more deeply into generative learning, he introduces the concept of “metanoia”, a Greek word meaning a profound shift of mind, which he considers to be synonymous with generative learning. He explains that for the Greeks it meant a fundamental change, transcendence (“meta”) mind (“noia”). Senge (1990) affirms that to grasp the meaning of “metanoia” is to grasp the deeper meaning of learning, as learning also implies a fundamental shift of mind. He compares the everyday use of learning, such as taking information or adapting behaviors, with generative learning and claims that real learning gets to the heart of what it means to be human. Through learning, we recreate ourselves and perceive the world and our relationship to it differently. Generative learning or “metanoia” refers to a change of the mental model, paradigm or knowledge through which we see reality. Recently, Senge et al. (2005) suggest that generative learning occurs through a process (the “U” process) that entails three major stages or elements: sensing, presencing and realizing. Sensing means becoming one with the world, mainly by observing. Presencing implies a state of becoming totally present to the larger space or field around us, to an expanded sense of self, and, ultimately, to what is emerging through us. Realizing involves bringing something new into reality.
However, OL literature has also described what structural or cultural arrangements are likely to foster both adaptive and generative learning (Argyris et al., 1985; Anderson, 1997; Senge, 1990). Adaptive learning is related to rationality, defensive relationships, low freedom of choice, and discouragement of inquiry (Argyris et al., 1985). In contrast, double loop learning is encouraged through commitment, minimally defensive relationships, high freedom of choice and inquiry.
In Senge's (1990) view, generative learning requires five disciplines: personal mastery, mental models, shared vision, team learning and systemic thinking. The first, personal mastery, is the term Senge uses to refer to institutionalized conditions for personal learning within an organization. It is related to issues of staff empowerment and the development of staff potentials. Senge explains that people in an organization have different internal pictures of the world or mental models, the second discipline, which should be made explicit so that they can be discussed openly and modified. The third discipline, shared vision, concerns the need for a certain degree of consensus within an organization, and at the same time the need for inspiration and motivation. Concerning the fourth discipline, team learning, Senge explains that teams, not individuals, are the fundamental learning unit in modern organizations; unless the team can learn, the organization cannot learn. This requires improved interpersonal communication between team members, a reduction in defensive behavior, and openness to creative thinking. The fifth discipline, systemic thinking, is crucial to examine the interrelationships between parts of an organization rather than the parts in themselves. While a focus on individual parts would only obscure the need for larger change, a focus on the whole system makes it possible to identify how organizational change might be brought about.
Adaptive and generative learning have not been extensively incorporated in frameworks or models for the process of OL. Kim (1993) develops a model of OL that links individual and organizational levels and also single and double loop learning through mental models. However, he recognizes that further work is needed for a better understanding of the role of mental models in individual and organizational learning, or the types of mental models that are appropriate for representing OL dynamic complexity.
Most of the well-known models (e.g. Huber, 1991; Crossan et al. 1999) obviate this typology. Huber (1991) describes four processes or constructs that contribute to organizational learning: Knowledge acquisition, information distribution, information interpretation and organizational memory. Crossan et al. (1999) developed a framework for the process of OL that identified the role of individuals, groups and the organization in feed-forward and feedback information flows (Crossan et al., 1999). This framework contains four related (sub) processes: intuiting, interpreting, integrating and institutionalizing, which occur over the three levels. Intuiting and interpreting occur at the individual level; interpreting and integrating at the group level; and integrating and institutionalizing at the organizational level. Crossan et al. (1999) consider that OL is multilevel, and also that OL consists not only of exploring or assimilating new learning but also of exploiting it or using what has already been learned (March, 1991; Cegarra-Navarro and Dewhurst, 2007).
In sum, mention has been made of adaptive and generative learning in the literature of OL since its first introduction in the field. However, few works (e.g. Argyris et al., 1985; Anderson, 1997; Senge, 1990; Kim, 1993) have attempted to analyze what factors are likely to enable these activities, have tried to inquire into the process in which they take place or have incorporated these processes into the OL process. In fact, this is what Visser (2007) recently termed meta-learning. The aim of this paper is to accomplish this, essentially through two concepts from complexity theory: self-organization and implicate order.