Positioning knowledge: schools of thought and new knowledge creation – Part 2

By Phin Upham


Hotelling (1929) described a competitive game that was later adapted to explore strategic positioning in ideological space. In the game, two hypothetical newspaper sellers, competing for readers who are distributed evenly along ‘‘Main Street,’’ can set up their stand anywhere in town. Assuming that, for the same price, customers will buy the closest newspaper if one newspaper seller were to position himself anywhere but the center of Main Street, the other would position himself a little closer to the center point and gain the majority of the customers. Thus both sellers end up converging at the midpoint of Main Street. In this model each player explicitly takes into consideration the moves of other players when acting. A political variant of this principle was applied to the ideological landscape of voters to explain the middle-of-the-road views generally espoused by candidates of major political parties (Downs 1957). Given an even distribution of voters, politicians in a two-party race, in order to appeal to the greatest number of voters, will converge to mainstream positions where they maximize their access to voters in a Nash equilibrium. More generally, Downs’ work finds that in an intellectual landscape, given Hotelling’s assumptions, a central position closest to the greatest number of consumers is optimal.

Further extending Hotelling’s theory, Downs (1957) also challenged his assumption of a ‘‘normal’’ distribution of consumers, and other researchers have further extended Hotelling’s game by including multiple players or additional consumer (voter) or competitor (candidate) entry (Krishna 2001; McGann 2002). The Hotelling–Downs framework has been usefully applied broadly to such areas as marketing and brand positioning (Choi and Coughlan 2004), news coverage (Gasper 2005) and simulations (Marks and Albers 2001).

The principles of acting to maximize intellectual proximity to the greatest number of consumers and of employing a dynamic strategy that takes others’ moves into consideration provide a powerful framework for analyzing new knowledge development. The strategic positioning of new knowledge developers explored in this paper resembles a very complex multi-player version of the Hotelling problem—one, however, that differs along two key dimensions. First, we attempt to include the dimension of schools of thought, which makes the landscape ‘‘clustered’’ and has profound consequences for the application of Hotelling’s strategic principles. Second, in our framework the candidates and the voters are flip sides of the same coin (each producer of new knowledge is also a consumer of new knowledge in our knowledge landscape), simultaneously competing with and supporting each other. In the context of new knowledge development, the positioning theory model significantly underestimates the importance of schools of thought in the knowledge landscape.

Search—near and distant

The tension between local and distant search has been explored by juxtaposing the Strategies of exploration and exploitation (March 1991). Exploration (involving distant search) is an attempt to add value by finding a new opportunity, while exploitation (involving local search) involves building on existing resources or knowledge in an attempt to extract value. Search strategies have significant effects on the development and structure of their landscape (Levinthal 1991, 1997, 1998). In the long term, exploration does produce benefits, but it must be ‘‘paid for’’ by exploitation (Barnett and Sorenson 2002). In the shorter term and from a research perspective, interdisciplinary research—explorative by its very nature— increases the difficulty of publishing papers, training graduate students, or receiving funding for a subject (Birnbaum 1981b).

However, since the dichotomy between exploration and exploitation is always Operationally dependent on the choice of boundary, the way in which we delineate boundaries will determine whether we define a search as near or distant. A number of recent studies examine the tradeoffs of different search strategies, but they define their relevant boundaries of analysis in different ways. Katila and Ahuja (2002) argue that firms can differentiate themselves by creatively and meaningfully reusing old technology to create new knowledge as well as by finding new technologies to achieve breakthroughs. Nerkar (2003) sees firms successfully choosing between recent, cutting-edge knowledge and knowledge that integrates understandings developed across time spans. They thus differentiate between middle-level and radical exploration. Herbert Simon famously looked at satisfying to explain the tension between usefulness and truthfulness—that at a certain point one stops looking for a better answer if the one has an adequate one (March and Simon 1958; March and Shapira 1987; March 1991). Over all, these researchers find that search patterns have profound effects on knowledge creation. This previous research is tied together by the core idea that when new knowledge is developed, some boundary, internal or external, is extended or challenged.

Although theories of local search often take for granted that there is local and distant knowledge, what makes knowledge accessible and close or inaccessible and far is left unexplored from an intellectual, psychological, and resource point of view. Delineating these boundaries is indeed a complex process. We argue that a key part of this intellectual boundary-shaping in knowledge development can be found in socially constructed schools of thought. Schools of thought are a key factor for new knowledge developers in perceiving information as near or far, and we believe that an appropriate intermediate spanning of boundaries between schools of thought is an important potential driver of knowledge creation. By providing an explanation for why new knowledge is close or far, we believe we move towards generalizing and integrating previous research on boundary spanning in knowledge creation.


In this section we extend three hypotheses about the relationship between the position of new knowledge within a school of thought and the subsequent impact of that knowledge. First, we argue that being part of a school of thought increases intellectual impact. Indeed, most new knowledge is not in any school of thought at all. Second, we argue that a position near the semi-periphery of a school of thought (but still within the school) leads to greater overall impact of new knowledge. We also find that two strong and competing hypotheses are supported by existing theory. In our third hypothesis, we argue that over time an author’s tendencies to explore between schools of thought, i.e., whether a knowledge developer tends to explore many different paradigms or specialize within a paradigm, will affect the impact of his or her ideas (Fig. 1).

Hypothesis 1 New knowledge has more impact if it is within a school of thought than if it is not.

Given that new knowledge is part of a school of thought, within that school of thought it can also be at the ‘‘core’’ or ‘‘periphery.’’ New knowledge that is core to its school is consistent with the rest of the school in its sources of ideas; new knowledge that is peripheral draws on knowledge that differs in some significant way, usually uncommon knowledge or knowledge from outside the school. The level of compliance of new knowledge with a particular school of thought depends on the number of criteria it satisfies in order to identify with that school. Its association with the school weakens as it draws more and more from notions that may not be firmly help by members of the school or our the central focus of other schools. In this way, for the purpose of research and exploration of knowledge creation, a theoretical distinct line can be created to separate schools from one another.

We believe that while new knowledge creators receive benefits from being in a school, membership can also be constraining if they blind themselves to good ideas outside that school. Specifically, we believe successful research usually draws from its own school and also a few core ideas from one or perhaps two other schools, synthesizing knowledge that is near and distant. In the area of technological innovation, studies have shown that patents which combine technologies from different patent classes tend to have more diverse, and potentially greater, impact (Trajtenberg 1990).