If a scientific hypothesis were a self-generated bright idea which owed nothing to scientific activity, then empiricism governed by hypothesis could be considered to compose a self-contained manner of activity; but this certainly is not its character. The truth is that only a man who is already a scientist can formulate a scientific hypothesis; that is, an hypothesis is not an independent invention capable of guiding scientific inquiry, but a dependent supposition which arises as an abstraction from within already existing scientific activity. Moreover, even when the specific hypothesis has in this manner been formulated, it is inoperative as a guide to research without constant reference to the traditions of scientific inquiry from which it was abstracted… Or consider the example of cookery. It might be supposed that an ignorant man, some edible materials, and a cookery book compose together the necessities of a self-moved (or concrete) activity called cooking. But nothing is further from the truth. The cookery book is not an independently generated beginning from which cooking can spring; it is nothing more than an abstract of somebody’s knowledge of how to cook: it is the stepchild, not the parent of the activity. The book… speaks only to those who know already the kind of thing to expect from it and consequently how to interpret it.
Michael Oakeshott, Political Education.