Routines in theory and in practice, a criticism of the cognitive perspective

The aim of this article is to provide a better understanding of the concept of routines. We will start by showing how Schumpeter introduced this concept and how Nelson and Winter have disseminated it so as to offer a deeper understanding of what a firm is. Through the notion of routines evolutionary theory opens up the organizational "black box" and offers a deeper insight into the firm's behavior. If routines are regarded as the loci of organizational memory, the articulation of knowledge (i. e. rendering it explicit by means of a natural, symbolic, or other language) is far from perfect, owing to the presence of tacit knowledge and the costs of articulation. We will attempt to show the evolution of this notion in some recent research (Cohen et al., 1995). These works, based on an articicial intelligence's approach developed by Simon, describe in a computational dimension which separates the body from the spirit. We will criticize this perspective by showing that tacit knowledge is not always included in cognitive maps but rather in technical artefacts and in operators' bodies, as underlined by Rosenberg. This will bring us to discussion of other dimensions of organizational memory that are able to capture the diversity of tacit knowledge without excluding the various reservoirs of knowledge (interference of technical artefacts, social interactions...). This last point will be discussed through a review of recent empirical studies which underline the articulation of knowledge by considering initial events, such as social interactions, bodily and mental skills, and technical artifacts, as an important dimension in building up routines.