The method of multiple working hypotheses
Thomas Chrowder Chamberlin 1890 Science (old series) v15 p92
Chamberlin, writing near the turn of the nineteenth century, advises naturalists to invent and/or test several testable hypotheses for each question that they investigate. This method helps avoid the usual &quot;parental affection&quot; theorists develop when testing only one idea at a time. Moreover, he suggests, a good interpretation of a complex phenomenon may result in the retention of more than one hypothesis. For example, the formation of the Great Lakes probably resulted from a combination of preglacial stream erosion, glacial ice erosion, and crustal deformation, not any one of these processes alone. The advantages of the multiple-working-hypothesis method include increased objectivity, flexibility in response, and improved ability to recognize one's own errors and ignorance. Drawbacks of the method are difficulty in explanation (there's so much more to explain) and an increased delay in settling on and reporting findings.
It was reprinted in Science in 1965 (v. 148, p. 754-759) and this version includes a bibliographical note that clears up the publication dates and versions.
The Related Website below links to an online version of the paper.