Category Structure and Generative Thought

Texas A&M Research Foundation, College Station TX

Investigators

Abstract

The ability to generate useful new ideas underlies human accomplishment in technological, scientific, economic and artistic realms, and understanding how new ideas are formed has importance for all of those areas. The major objective of this research is to develop a theoretical account of the processes involved in novel idea generation, and the ways in which people's existing knowledge influences the form of the new ideas they generate. Previous research has shown that people often begin the process of developing new ideas by retrieving from memory fairly specific instances of known concepts to serve as starting points, and then patterning the novel idea after those instances, but much about this process remains unclear. Building on that earlier work and other research on the nature of human concepts, this new research seeks to provide a more coherent view of the process. It goes beyond the obvious point that new ideas are derived from existing ones to assess how the structure of people's existing concepts influences the structure of novel ideas they produce. Studies are designed to determine why some instances are more likely than others to be retrieved and used as starting points, the extent to which reliance on those instances constrains the originality of the newly formed ideas, the extent to which reliance on those instances is rigid versus flexible, and how situational factors influence what the person retrieves from memory as a starting point for the new idea. Predictions are based on the path-of-least-resistance model, which states that people will tend to rely on the specific instances that are most highly representative of a given category. Some studies will require participants to imagine novel entities from particular categories (e.g., fruit that might grow on another planet, toys that might be used by blind children). Base categories will include a variety of domains, including tools, toys, animals, fruit, vegetables, snack foods, rituals, and several others. Companion studies will use multiple measures to determine how representative different instances are within the base categories (e.g., how typical, frequent, and familiar specific instances of real Earth fruit are judged to be). The measure hypothesized to be most influential is how readily specific instances come to mind. The idea is that the category instances that come to mind most readily are the ones that will be most frequently retrieved as candidate starting points in formulating novel ideas. The rationale is that generating new ideas is cognitively demanding, and people tend to simplify the task by pursuing ideas that come to mind readily. In these studies, the originality and other properties of the imagined ideas will be assessed to determine the degree to which relying on such specific instances constrains the form of new ideas. Other studies will test the hypothesis that representativeness is flexible rather than fixed and that recent experiences will influence the particular instances that are retrieved and used in imagination. They will attempt to influence idea generation by giving people prior exposure to certain instances or by presenting key attributes of those instances. In other words, rather than just using the static structure of existing concepts to predict the form of novel ideas, the studies will attempt to manipulate the form of the novel ideas by way of the other factors present in the situation.