Science (from the Latin scientia, knowledge) is too diverse an undertaking to follow any single method. Yet from the time of Francis Bacon, well into the twentieth century, the myth has persisted that true science follows a particular method – Bacon's "inductive method". This allegedly involved collecting a vast number of individual facts about a phenomenon, and then working out what general statements fitted those facts. After the 17th century nobody attempted to follow that program. In the 19th century, philosophers came to recognize the possible existence of the "hypothetico-deductive method". According to this model, the scientist studied the phenomena, dreamed up a hypothetical explanation, deduced some additional consequences of his explanation, and then devised experiments to see if these consequences were reflected in nature. If they were, he considered his theory (hypothesis) confirmed. But Karl Popper pointed to the logical fallacy in this last step – the theory had not been confirmed, but merely not falsified; it could, however, be worked with provisionally, so long as new tests did not discredit it.

Philosophers of science now recognize that they cannot justly generalize about the psychology of scientific discovery; their role must be confined to the criticism of theories once they have been devised. Historians of science, meanwhile,have pointed to the importance in scientific discovery of "external factors" such as the contemporary intellectual context and the structures of the institutions of science. Once distinct terms – "theory," "hypothesis," "explanation," "description," and "law" – are all now seen to represent different ways of looking at the same thing – the units in what constitutes scientific knowledge at any given time. Indeed there is still no general understanding of how scientists become dissatisfied with a once deeply-entrenched theory and come to replace it with what, for the moment, seems a better version.