Abstract

Over the past two or three decades there has arisen a fairly widespread dissatisfaction with the usefulness of the formal logical techniques of Frege and Russell for teaching the evaluation of reasoning in natural language. The problem with these formal methods, I argue, lies in certain questionable presuppositions that underlie these methods. ;My own case will rest mainly on the following three claims: first, that the millennia old belief that there are two fundamentally different types of argument, viz., induction and deduction, is crucially mistaken and hampers rather than helps our understanding and evaluating of reasoning; second, that the belief that only formally valid arguments are deductively valid is a false dogma, one that is pernicious not only in its own right, but also because it has lent support to the other dogmas we must set aside if we are to have any radical progress in our understanding of and instruction in reasoning in natural language; and third, that logic is better seen not as a body of knowledge, but as a practice or art; and, as with any art, we come to perform it better not through the learning of a system of propositional knowledge but through appropriate training. In addition, I shall outline what I take the appropriate training to be, one that makes use of the imaginative search for counterexamples, and which makes logical training a more practical enterprise