Dialectron : a prototypical dialectic engine for the support of strategic planning and strategic decision making

Abstract

The use of dialectic in strategic management has been extensively debated. It was first proposed by Mason (1968), who soon after (Mason, 1969) argued its superiority over devil's advocacy and expert advice. Cosier first questioned Mason's claim in 1978, and initiated a controversy which remains unresolved (Cosier and Schwenk, 1990; Schwenk, 1984, 1988, 1989, 1990; Schweiger and Sandberg, 1989). Chanin and Shapiro (1985) have pointed out weaknesses on both sides of the debate and have made recommendations for future research. MIS, and more specifically DSS, can contribute to this research issue with the development of computer-based tools designed to support participation in simulated strategic planning. The research to be presented constitutes an attempt to apply Brinberg and McGrath's (1985) Validity Network Schema (VNS) (itself the result of a search for a better paradigm for scientific research) to a) the development of an improved conceptualization of dialectic and to b) the design and implementation of a DSS prototype which shows some of the possibilities for decision support available. The first part of the study, whose research design resembles the dialectic method used in Adler (1958, 1961), established points of correspondence between uses of the term in various disciplines. The second part of the study resulted in a DSS prototype named DIALECTRON, whose design architecture is the result of combining 1) Rescher's (1977) characterization of debate (Mason's dialectic) as two parallel disputations (Mason's devil's advocacy), 2) Rescher's inventory of dialectic moves and countermoves, 3) Ackoff's (1981) concept of interactive planning, 4) cognitive mapping, Axelrod's (1976) modelling technique, and 5) Mitroff et al.'s (Mitroff et al., 1970, Mitroff, 1971) mathematical model of the dialectic process, all within the spirit of the concept of dialectic developed in the first part. Through the integration of these paradigms within a state-of-the-art artificial intelligence application development environment, it was possible to construct a prototype capable of concurrently representing and synthesizing two conflicting cause-and-effect models of a typical strategic planning situation, and to show how one such tool could be used as an aid in various knowledge manipulation procedures involving conflicting problem models.