Modularity, near-decomposability and interdependence: On conceptual parsimony again

The last twenty years have witnessed an increasing use of the concept of modularity. Initially used to refer to product design and its underlying technologies (e.g. Baldwin & Clark, 2000; Henderson &  Clark, 1990; Langlois & Robertson, 1992; Sanchez, 1995; Schilling, 2000), it is also increasingly used to refer to organizational design, i.e. to the links across organizational units (e.g. Ethiraj & Levinthal, 2004; Langlois, 2002; Sanchez & Mahoney, 1996; Schilling, 2001).

A modular system is a system where clear, separable modules or components can be identified. Thus, modularity means full decomposability or almost no interdependence between modules. Herbert Simon (1962)  introduced the term near-decomposability to refer to complex systems where some interdependence always remains, despite the designer’s effort to fully decompose the system into independent modules. Interdependence generates complexity and thus the need for coordination and collaboration (see also Sanchez, 2002).

Therefore, the underlying, essential design feature of modularity is interdependence, a concept for which one of the pioneering scholars of organizational designers (Thompson, 1969) provided a typology (pooled, sequential and reciprocal) that gives us a more sophisticated, nuanced lens to address the challenge of designing and managing interdependent, complex systems. Thus, do we need the concept of modularity? Isn’t it a synonymous of full decomposability or non-interdependence (Sanchez, 2002)? If parsimony is a hallmark of science, should we then dispense with modularity and keep decomposability and interdependence?