Philosophy 100 Essential Thinkers

by Philip Stokes

Bacon was less interested in the problem of justifying inductive generalisations, than in how to generate good inductive hypotheses out of the masses of data collected by observation. Bacon devised a new method. To illustrate it, Bacon shows how one might generate an hypothesis on the nature of heat. One should, Bacon tells us, list all those things in which the property under investigation, in this case heat, is present, then all those things in which the property is absent and finally all those cases which admit of varying degrees of the property in question. From such a list, Bacon believes the natural hypothesis will present itself, which in this case, as he well knew at the time, is that heat is produced according to the movement or excitation of molecules within a body. Although Bacon’s method is undoubtedly one way of applying order to a body of data, and even perhaps a useful way in some cases, it nevertheless seems unlikely to fulfil his ambition, which was to find a systematic way of deriving scientific hypotheses from the arrangement of data. It is unlikely that there ever could be such a system. Bacon failed to take into account the creativity and imaginative aspect of scientific theory building. No matter how systematically one organises data, inductive hypotheses cannot be guaranteed to appear out of them. One may find that some facts deductively follow from a certain ordering of data, but that is not what Bacon was after.

Consider data science ...