I. Introduction

§ 1. Logic and language

Logic has often been defined as the science that deals with the laws of thought. This is aa ambiguous characterization unless we distinguish between psychological and logical laws of thought. The actual process of thinking evades distinct analysis; it is in part logically determined, in part automatic, in part erratic; and what we observe as its constituents are isolated crystallizations of largely subconscious currents hidden below a haze of emotional processes. As far as there are any laws observable they are formulated in psychology; they include laws both of correct and of incorrect thinking, since the tendency to commit certain fallacies must be considered a psychological law in the same sense as the more fortunate habits of correct thinking. This distinction itself, the distinction between valid and invalid thinking, cannot be made within psychological analysis.

If we want to say that logic deals with thinking, we had better say that logic teaches us how thinking should proceed and not how it does proceed. This formulation, however, is susceptible of another misunderstanding. It would be very unreasonable to believe that we could improve our thinking by forcing it into the straitjacket of logically ordered operations. We know very well that productive thinking is bound to follow its own dark ways, and that efficiency cannot be secured by prescriptions controlling the paths from the known to the unknown. It is rather the results of thinking, not the thinking processes themselves, that are controlled by logic. Logic is the touchstone of thinking, not its propelling force, a regulative of thought more than a motive; it formulates the laws by which we judge thought products to be correct, not laws that we want to impose upon thinking. Creative thought processes, even of the trained mind, do not move along prepared paths but follow a method of trial and error, in which logic separates the right from the wrong results. If it is true that to a certain extent we can improve our thinking by studying logic, the fact is to be explained as a conditioning of our thought operations in such a way that the relative number of right results is increased.

When we call logic analysis of thought the expression should be interpreted so as to leave no doubt that it is not actual thought which we pretend to analyze. It is rather a substitute for thinking processes, their rational reconstruction,1 which constitutes the basis of logical analysis. Once a result of thinking is obtained, we can reorder our thoughts in a cogent way, constructing a chain of thoughts between point of departure and point of arrival; it is this rational reconstruction of thinking that is controlled by logic, and whose analysis reveals those rules which we call logical laws. The two realms of analysis to be distinguished may be called context of discovery, and context of justification. The context of discovery is left to psychological analysis, whereas logic is concerned with the context of justification, i.e., with the analysis of ordered series of thought operations so constructed that they make the results of thought justifiable. We speak of a justification when we possess a proof which shows that we have good grounds to rely upon those results.

It has been questioned whether all thinking processes are accompanied by linguistic utterances, and behavioristic theories stating that thinking consists in linguistic utterances have been attacked by other psychologists. We need not enter into this controversy here for the very reason that we connect logical analysis, not with actual thinking, but with thinking in the form of its rational reconstruction. There can be no doubt that this reconstruction is bound to linguistic form; this is the reason that logic is so closely connected with language. Only after thinking processes have been cast into linguistic form do they attain the precision that makes them accessible to logical tests; logical validity is therefore a predicate of linguistic forms. Considerations of this kind have led to the contention that logic is analysis of language, and that the term 'logical laws' should be replaced by the term 'rules of language'. Thus in the theory of deduction we study the rules leading from true linguistic utterances to other true linguistic utterances. This terminology appears admissible when it is made clear that the term 'rules of language' is not synonymous with 'arbitrary rules'. Not all rules of language are arbitrary; for instance, the rules of deduction are not, but are determined by the postulate that they must lead from true sentences to true sentences.

It is the value of such an analysis of language that it makes thought processes clear, that it distinguishes meanings and the relations between meanings from the blurred background of psychological motives and intentions. The student of logic will find that an essential instrument for such clarification is supplied by the method of symbolization, which has given its name to the modern form of logic. It is true that simple logical operations can be performed without the help of symbolic representation; but the structure of complicated relations cannot be seen without the aid of symbolism. The reason is that the symbolism eliminates the specific meanings of words and expresses the general structure which controls these words, allotting to them their places within comprehensive relations. The great advantage of modern logic over the older forms of the science results from the fact that this logic is able to analyze structures that traditional logic never has understood, and that it is able to solve problems of whose existence the older logic has never been aware.

We said that logic cannot claim to replace creative thought. This limitation includes symbolic logic; we do not wish to say that the methods of symbolic logic will make unnecessary the imaginative forms of thought used in all domains of life, and it certainly would be a misunderstanding to believe that symbolic logic represents a sort of slide-rule technique by which all problems can be solved. The practical value of a new scientific technique is always a secondary question. Logic is primarily a theoretical science; and it proceeds by giving a determinate form to notions that until then had been employed without a clear understanding of their nature. Whoever has had such an insight into the structure of thought, whoever has experienced in his own mind the great clarification process which logical analysis accomplishes, will know what logic can achieve.