Provisos and the magnitude of scientific change

The failure of a prediction forces some modification of the theory which in many cases means locating some factor that interfered in the experiment. A factor may be found of a type already recognized by the theory. Or it may be the case that it becomes necessary to propose an as yet unknown factor. Here we suggest an ordering of theory changes according to the type of factor proposed. The ordering may be used to distinguish normal change from revolutionary change. Our proposal is based on the thesis that a theory admits only a finite number of types of factors.