Scientific discovery is a central topic in the philosophy of science, and has been treated in many different ways. In this article, Tom Nickles, one of the initiators of a shift from a non-inferential to an inferential treatment of scientific discovery (Nickles 1980a,b), looks at how the issue has evolved. He describes the various responses to the challenge of separating the contexts in which scientific discoveries are made, and the resulting debates about whether or not rules can be formulated that would allow for mechanical derivation of novel ideas. These diverse approaches employ different terminologies and draw on a wide range of resources, including philosophical analyses of actual scientific discovery, empirical studies from cognitive science, computer simulations of reasoning processes, and theoretical work in artificial intelligence and the history of sciences.
Philosophical discussion of the nature of scientific discovery began in the 19th century, when William Whewell’s two volumes of Philosophy of the Inductive Sciences were published. This was an important contribution because it clearly separated the creative moment, or eureka thought, of having a new insight from processes of articulating, developing, and testing that insight.
The idea was that the eureka moment, narrowly construed, was an unanalyzable act of intuition or (Romantic) genius. More recently, there has been a move towards recognizing the relevance of a wide range of contextual factors to the process of scientific discovery. These include the specification of facts through systematic observation, measurement and experiment; the clarification of ideas through exposition of axioms and definitions; and generative justification processes that are a precursor to consequential justification processes used to test the newly articulated hypothesis.