UC Berkeley

In this dissertation I analyze the process of crafting definitions whose purpose is classification. The context I examine is undergraduate upper-division physical science majors defining and naming sub-categories of a physical phenomenon in the context of a design task over an extended period of time. The goal of the design task is one of classification: help people better identify the phenomenon out in the world.

I first develop an analytic framework for the process of improving a definition. This framework involves an interplay between four main elements: the current state of the definition, criteria for what makes a good definition, examples, and definitionally unarticulated knowledge (DUK). By DUK, I mean implicit judgments that definers make about categorization that have not been incorporated into the definition, per se. I show how criteria and the practices associated with meeting them guide the refinement work.

Once participants craft their definition a question emerges: to what extent does it serve them as a tool when they make sense of new examples? I zoom in on moments where new examples are introduced by participants or the facilitator and describe the extent to which students rely on their own crafted definition. The consideration of a new example can sometimes lead to participants naming a new sub-category and I analyze this process as well.

The detailed empirical analysis is bookended by a set of workshop design principles that made the analysis possible and some educational implications of the work. I lay out a set of design principles for creating activities wherein people engage productively in collaboratively defining. I conclude by exploring the educational implications for college physics teaching as well as more general instances where one is interested in crafting a definition for the purpose of classification.