Understanding Exploratory Creativity in a Visual Domain Kyle E. Jennings, 1,2 Dean Keith Simonton, 2 and Stephen E. Palmer 1 Departments of Psychology University of California, Berkeley (1) University of California, Davis (2) kejennings@ucdavis.edu, dksimonton@ucdavis.edu, palmer@cogsci.berkeley.edu ABSTRACT This paper describes a computerized aesthetic composition task that is based on a “creativity as search” metaphor. The technique collects detailed, moment-to-moment data about people’s search behavior, which can help open the “black box” that separates independent variables that influence cre- ativity from their outcomes. We first describe the technique and provide a detailed theoretical framework. Then, we dis- cuss how the technique is typically applied, describe several in-progress studies, and present some preliminary results. Finally, we discuss relations to other work, limitations, and future directions. We argue that this technique and the re- search that it enables will facilitate a deeper understanding of the creative process, become a valued tool for creativity researchers, and contribute to methodological and theoreti- cal advances in how creativity is studied and understood. Author Keywords creativity, search, optimization, experimental methods ACM Classification Keywords J.4 Computer Applications: Social and Behavioral Sci- ences—Psychology General Terms Experimentation, Measurement, Theory BACKGROUND The creative process can be thought of as the search for an ideal solution to a problem. One way to understand creativ- ity is to understand this search process. This paper describes an under-development research technique that is based on the “creativity as search” metaphor. By capturing detailed information about the moment-to-moment progress of a per- son’s creative search, this technique can both test models of creativity as search, and enable more detailed studies of common phenomena in the creativity literature. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. C&C’11, November 3–6, 2011, Atlanta, Georgia, USA. Copyright 2011 ACM 978-1-4503-0820-5/11/11...$10.00. Search can either be seen as finding a path from a starting state to a valid end state, or as finding the best (or nearly best) solution from among many others. The former case, which we call path search, is relevant when the desired outcome is known but the way to reach it (the path) is not (e.g., proving a mathematical theorem), and has been studied extensively [19, 20, 22]. The latter case, which we call place search, ap- plies when the desired outcome (the place) is unclear but the means for achieving it are relatively straightforward, such as when an artist arranges a still life or composes a photograph. Particularly given that the creativity of artists depends more on these early decisions than on the execution of these de- cisions [10], this research focuses on how people select the desired outcome from among competing alternatives. In open-ended domains like the arts, choosing what outcome (or solution) to pursue is seldom a simple matter of deciding among a few choices. Instead, the set of possibilities is too vast to be considered simultaneously, meaning that searchers must iteratively consider subsets of the space. How people control this iterative process can be called a search strat- egy, and includes considerations such as how people move from one subset to another, and how people evaluate each solution. Though search strategies might be an important determinant of how creative the final solution is, they are not directly observable. However, if the options under consider- ation at each stage can at least partially be observed, it be- comes possible to trace how people move through the space of possibilities over time. This search trajectory offers clues as to what kind of search strategy people use. We have developed an experimental paradigm that can ef- ficiently capture detailed information about people’s search trajectories in a visual domain. Using this technique, we aim to determine: • what can be learned about people’s search strategies from their search trajectories; and, • whether differences in solution creativity can be traced to differences in search strategies. We also aim to apply our technique to better understand broader questions, such as how personality and situations affect people’s creative processes. Our technique should help open the creative process “black box” that sits between the independent and dependent variables in much of this re- search.