ABSTRACT
The method of multiple working hypotheses (MWH) has been championed by many authors (e.g. Haines-Young and Petch, 1986; Baker, 1999). An early paper to explore this method was by Gilbert (1896) in relation to what he described as a topographic problem. Gilbert called a hypothesis a ‘scientific guess’ (1896, p. 1) and viewed the initial guess as the first step an investigator made to connect a group of facts and a cause. Rather than starting with a blank canvas Gilbert saw the hypothesis as linking existing cause and effect. The role of the hypothesis was to provide a means for extracting information that would assess the validity of the link between the two. Gilbert reasoned the link as follows:
In other words, he [sic] frames a hypothesis or invents a tentative theory. Then he proceeds to test the hypothesis, and in planning a test he reasons in this way: If the phenomenon was really produced in the hypothetic manner, then it should possess, in addition to features already observed, certain other specific features, and the discovery of these will serve to verify the hypothesis.
