ABSTRACT
The support-function invoked in normal judgement applies standards of evaluation determined in some way by the variables that are relevant to the particular set of materially similar hypotheses to which the hypothesis judged belongs. Certain very serious difficulties stand in the way of contriving one, or of contriving even an informal measure of inductive support that relies at some point on a practicable measure of inductive relevance. Thus a support-function can be readily constructed, for any particular field that is adequate to the task of ranking or grading, though not of measuring, the support given to a hypothesis by the results of any prudent experimenter’s tests on it. In order to escape falsification by trials that manipulate the variable and to be able to go on and achieve a high level of support, a hypothesis about the alleviative properties of barbiturates, or of whisky, must be suitably qualified in its formulation or suitably restricted in its domain of discourse.
