A preliminary typology of organizational learning: synthesizing the literature

Journal of Management, Fall, 1996 by Danny Miller

Two organizational phenomena best illustrate synthesis: configuration and systems thinking. Synthetic learning can pull together information in a way that gets rid of extraneous details and converges on what is most important. It identifies a "configuration" - a theme that reveals how the parts of a problem, strategy, or even organization interrelate (Miller, 1987, 1990b). This may help managers discover a core competence or critical resource, or identify a competitive advantage (Peteraf, 1993; Prahalad & Hamel, 1990). Synthetic learning may also find synergies among organizational departments or skills, or it may discover a desirable nexus between organizational competences and challenges (Hambrick, 1994; Miller, 1987; Mintzberg, 1973, 1989).

Another application of synthetic learning is systems thinking (Beer, 1981). Some executives are able to view their organizations as dynamic, evolving configurations: they not only see how the parts fit together, they also develop a feel for the dynamics of the system. They begin to understand networks of causality, to discern complex problematic relationships, vicious circles, emerging opportunities, and shifts that signal special challenges (Senge, 1990; Stata, 1989; Steinbrunner, 1974).

There are many strengths associated with synthesis. It can discover powerful synergies and great opportunities, and it may give managers a more profound understanding of the underlying dynamics of their organizations (Mintzberg, 1989; Senge, 1990). But synthesis may create a vision that is so integrated and compelling that it defies necessary alteration (Miller, 1990b, 1993). Also, effective synthesis is beyond the intellectual capacity of many managers.

Experimental Learning

Scholars of organizations have noted that the analytical model ignores limitations on decision makers (March & Simon, 1958). They believe that voluntarism is bounded by intellectual, temporal and economic constraints (Grandori, 1984). They also suggest that overcoming these constraints requires simpler, more incremental approaches to learning - for example, performing small experiments and monitoring the results (Quinn, 1980). Managers can explore their complex environments in a gradual, piecemeal way instead of making grand long-term plans (Burgelman, 1990; Staw, 1977).

Like analysis, experimentation is an intendedly rational, methodical approach to learning. There is a deliberate attempt to systematically gather and interpret information in the hope of improving the behavioral repertoire of the organization. But now action sometimes precedes analysis in the learning cycle (Weick, 1979). Experimental learning is also more spontaneous than analytical learning as it is not so governed by detailed plans. Search occurs by conducting experiments remedially, opportunistically, and in many areas (March, 1989, 1991). Choice - the decision of how far to go with experiments - is generally determined by feedback about results (Lant & Mezias, 1992).

Experimentation reduces the burden on top managers by occurring at many levels and places of an organization (Hart, 1992). In some cases, "atomistic" experimentation takes place - small projects and adaptations that are done almost independently by operating units (Garud & Van de Ven, 1992). Elsewhere, experiments are performed higher up and are more coordinated and interrelated (Pascale, 1989). Although participation in experimental learning is sometimes quite broad, what is learned may be used only locally. Thus, experiments are more likely to influence situational tactics and practices than global strategy. Occasionally, however, experiments will produce dramatic local victories or blunders that will be noticed throughout the system and serve as seminal strategic lessons (Lant & Mezias, 1992).