Methodological Function of Hypotheses in Science: Old Ideas in New Cloth

The August 1999 issue of Genome Research presented a very interesting editorial, “Hypothesis-Limited Research,” on the status and discovery function of scientific hypotheses in disciplines generating large amounts of data (Goodman 1999). Dr. Goodman points to two types of limitations generated by hypotheses—one theoretical and one pragmatic. She argues that generating at least some type of data, for example, largescale sequencing, does not require hypotheses. From a practical point of view, a hypothesis often complicates the scientific discovery flow, generating a hypothesis is time consuming and biases data interpretation. Consequently, Goodman urges giving up on proposing theory first and collecting data afterward. She supports her suggestion with historical examples, presenting great scientific discoveries that were not based on a hypothesis, although contrary interpretation has been presented by others (Łastowski 1996). Let us have a look at how heretical Goodman’s proposal is, if at all.