Induction and Deduction Revisited

Abstract

What is the distinction between an inductive argument and a deductive one? This seems a straightforward question, but there has been very little discussion of it lately in the literature. Perhaps then it is a question with a straightforward, settled answer, not worth further talk. If that is so, it ought to be possible to state the answer. Suppose we try. We can begin by looking briefly at the jobs which the distinction is supposed to do. Gilbert Harman has claimed whenever one characterizes a set of statements as premisses and conclusion one is thereby thinking of the statements as constituting a deductive argument. This leaves no room for a separate class of inductive arguments, and according to Harman it follows that there is no 'subject of inductive logic'.' Wilfrid Sellars, having concluded that all 'probability' arguments are best construed as camouflaged deductive arguments for conclusions which explicitly state that it is reasonable for the arguer to accept certain statements, leaves us with a negative prospect for a distinct set of inductive arguments, and hence for a distinct inductive logic.2 Sellars and Harman thus accept the well-entrenched view that if there were a legitimate distinction between inductive and deductive logic, that distinction would be drawn by appeal to a more fundamental distinction between arguments-a distinction between inductive and deductive arguments. Neither Harman nor Sellars admit the existence of any members in the class of inductive arguments. Other philosophers, such as Arthur Burks,3 viewing the distinction between deductive and inductive logic in much the same way, insist there are inductive arguments to serve as subject matter for inductive logic. In fact it has become commonplace to hold that inductive logic, whatever it may be, is the logic of inductive arguments, whatever they may be, and similarly deductive logic is the logic of deductive arguments.