Kalle Lyytinen

Information systems development is a high‐risk undertaking, and failures remain common despite advances in development tools and technologies. In this paper, we argue that one reason for this is the collapse of organizational intelligence required to deal with the complexities of systems development. Organizations fail to learn from their experience in systems development because of limits of organizational intelligence, disincentives for learning, organizational designs and educational barriers. Not only have many organizations failed to learn, but they have also learned to fail. Over time they accept and expect poor performance while creating organizational myths that perpetuate short‐term optimization. This paper illustrates learning failure in systems development and recommends tactics for overcoming it.

Technologies that support communication and models used in the development of communications need good underlying theories. One theory suggested as a base for design is speech act theory. Both communication support tools and modelling notations informed by speech act theory have been proposed. Speech act theory forms no unified, single theory, but actually houses several variants for dealing with semantics, pragmatics, and social context of communications. They all have one common feature: they assume that language is not merely a means of describing but also a means for doing things. In this paper we present an overview of speech act theories and their uses in information systems research. Our focus is on examining success of speech acts and commitment negotiations. We suggest a new model of commitment making that combines different speech act theories and their criticisms. It can be used in developing more adequate models of information systems and in highlighting requirements for flexible coordination support tools. The creation of shared understanding and agreement as well as the representation of context are emphasised.