6 Landscapes of Computational Creativity

Novelty, Search, Imagination, and Scientific Discovery

Chapter 5 fixed the book’s principal creative artifact: a presented theory, its models, and an explicit relation to what was known before. Before Chapters 7–9 develop Boden’s three modes in detail, it is important to locate that target among neighboring accounts of creativity. Otherwise a recovered equation, a surprising image, a novel search path, and a theory extension can appear to be interchangeable successes even though they produce different objects and require different evidence.

Creativity has been studied as a psychological capacity, a philosophical problem, a property of artifacts, a search process, an information-theoretic event, and a program of automated scientific discovery. No single literature owns the question. The Association for Computational Creativity and its International Conference on Computational Creativity have made this plurality into an interdisciplinary research community spanning artificial intelligence, cognitive science, design, engineering, and the arts [ Association for Computational Creativity , 2026a , 2026c ] . The Association also maintains an extensive online bibliography of books, theses, proceedings, conference papers, and other publications. Because that resource can grow with the field, it is the preferred current pointer beyond the selective map developed here [ Association for Computational Creativity , 2026b ] .

This chapter organizes that broad landscape by asking a concrete question: what is the object of the creative claim? It may be an artifact, an observation judged surprising, a path through a conceptual space, a recovered equation, an experimental program, or an extended theory. These are not rival definitions so much as different units of analysis. Once the creative object is fixed, one can ask how it is generated and what evidence warrants calling it creative.

The argument proceeds in four stages. We first compare conceptual-space, developmental, and neurocognitive accounts of what changes. We then separate novelty and surprise from explanation and theory construction. Next we examine search, imagination, cellular automata, and evolution as mechanisms that generate candidates at different levels of description. Finally, symbolic regression, automated laboratories, and contemporary scientific agents show how those distinctions become experimental. The point is not to force these traditions into DIAL, but to identify what each contributes and what remains missing when the desired artifact is a persistent, revisable theory.

¹1. The chapter is a map, not an exhaustive history. Its purpose is to locate the book’s theory-construction objective among complementary research programs and to identify reusable evaluation methods. ↩

. A creativity claim needs both reference and evidence coordinates. Its reference coordinates identify the object or process being credited, the space relative to which the change is novel, whether the process remained assimilatory or proposed a package change, and which package component changed. Its evidence coordinates record the less-disruptive controls, the proposal’s epistemic status, and the certificate warranting admission. “Creative output” without these coordinates is too underspecified for cumulative experimental science.