ABSTRACT
Economics, for instance, does not deal with human behaviour directly, but is forced to make assumptions about the reactions of human beings to economic facts; it uses a (largely unrealistic and unsound) psychology of its own as an intermediary between economic fact and social behaviour. The psychologist has no need of such intermediaries; he is in direct contact with the central link in the chain of causation between antecedent condition and resultant action. This central link, in a very general sort of way, is, of course, the human being, but this is far too vague and general a statement to be of any scientific usefulness, and we must look for something more definite and more specific in terms of which we can frame our theories.Psychology for a long time was handicapped by the absence of a concept which could be used in this connection. Taking over from Tetens and Kant the general division of mental life into willing, feeling, and thinking, or conation, affection, and cognition in more technical language, it failed to account for the facts of social life which inevitably demand an integration of all three elements. As-sociationist philosophers attempted to solve the problem by concentrating exclusively on cognition and thus rendered their accounts almost entirely in terms of a wraith-like ‘rational man5 whose actions were determined by reason alone. ‘Economic man5 is a direct successor o f ‘rational man5, equally wraith-like and equally absurd. Psychoanalysts, on the other hand, attempted to dethrone reason completely and posited ‘irrational man5, a creature wildly driven by impulses and emotions which he did not understand, and using reason merely to rationalize his actions ex post facto.Clearly, these are both simplifications which are of little use in a scientific account of political behaviour. Human beings do not always act in a completely rational, philosophical manner, debating the ultimate causes and consequences of actions and deciding on the basis of pure ratiocination; neither are they merely the playthings of emotional surges dating back to and deriving their strength from events in their far-distant childhood. What is needed is a concept which will serve to integrate all these divergent ingredients. This concept was introduced into science by that great trio of British psychologists, G. F. Stout, A. F. Shand, and W. McDougall. Using the term ‘sentiment5 which had earlier been proposed by Shaftesbury, Adam Smith, Herbert Spencer and Alexander Bain, Shand defined this term as denoting systems of
character which organize and direct the various primary emotions and impulses. These systems, as Stout had pointed out earlier, are not actual feelings but dispositions; they dispose a person to have certain feelings when presented with the object around which the sentiment has grown. Such sentiments are not innate but acquired; as Stout points out, ‘an object which has been connected with agreeable or disagreeable activities, which has given rise to manifold emotions, which has been the source of various satisfactions or dissatisfactions, becomes valued or the opposite in and for itself’. A sentiment so acquired has dynamic properties because ‘although a sentiment is only an organization of a part of the character, it is in a dynamical relation to the rest and gives a peculiar orientation to the whole’.These adumbrations of a consistent theory were transformed into the basis of a systematic social psychology by McDougall, whose Introduction to Social Psychology first appeared in 1908 and may be said to mark the beginning of that branch of study. Defining sentiment as ‘an organized system of emotional dispositions centred about the idea of some object’, McDougall also stresses the importance of learning in the development of sentiments. ‘Each sentiment has a life history like every other vital organization. It is gradually built up, increasing in complexity and strength, and may continue to grow indefinitely, or may enter upon a period of decline and may decay slowly or rapidly, partially or completely.’McDougall adds the idea that in the course of development all the sentiments of an individual will tend to build themselves into a hierarchical system at the apex of which is usually placed what McDougall calls the sentiment of self-regard. To him, sentiments and their organization are the building stones of character; . . . ‘the development of integrated character consists in the growth of a harmonious system of the sentiments, a hierarchical system in which the working of the sentiments for the more concrete objects is regulated and controlled by the sentiments for general and more abstract and ideal objects, such as devotion to the family, the clan, the occupational or civic group, the nation, or mankind, the love of justice, humanity, liberty, equality, fraternity; and by hatred for cruelty, for injustice, for oppression, for slavery. And volition in the full and higher sense implies that this hierarchy of sentiments culminates in and is presided over by a sentiment of self-regard which, by incorporating in its system these higher abstract senti-
merits has become an ideal of self, an ideal of character and of conduct to which our daily actions must conform and with which our long range motivations, our ambition and personal loyalties, must harmonize.5We may briefly summarize now the nature of a sentiment. In the first place it is an organization of conative, affective, and cognitive parts of the mind; in the second place, it is dynamic in the sense that it determines to some degree the behaviour of the organism; in the third place, it is a disposition, a set to react in a certain way once it is aroused; in the fourth place, it is learned rather than innate; and in the fifth place it combines with other sentiments to form a larger structure.In spite of the great usefulness of this theory, and in spite of the fact that it fulfils an obvious need, it did not live up to its promise. For one thing, it remained theoretical and failed to seek substantiation in large-scale experimental studies. In the second place, it was linked too closely with McDougall’s doctrine of instincts, which proved unacceptable to later workers. It was too necessary’ a concept, however, to be completely forgotten, and consequently it suffered a sea-change; while sentiment as a concept was retained, the term itself was dropped and a large variety of others substituted. The most widely accepted of these terms was that of attitude, but in psychoanalytic circles the term ‘complex’ also has frequently been used. It would not be profitable to enter into the long history of discussion of definitions beginning with the symposium on the relations between complex and sentiment held shortly after World War I by the British Psychological Society. It is sufficient to say that there is little agreement between psychologists on the differential use or meaning of these terms, except that complexes tend to be regarded as morbid, symptom-producing sentiments, and that the main distinctions between sentiments and attitudes appear to be that sentiments are more lasting and more highly organized than attitudes, and that the objects of attitudes are usually more abstract than the objects of sentiments. This would agree with popular usage; we tend to refer to personal sentiments but to social attitudes, and there is little doubt that our personal feelings on the whole are more highly organized and less abstract than our social ones.1Be that as it may, we shall here use the term ‘attitude5 very much in the way in which it has been defined by G. W. Allport:
‘An attitude is a mental and neural state of readiness organized through experience, exerting a directive or dynamic influence upon the individual’s response to all objects and situations with which it is related.’ As so defined, attitude is clearly a hypothetical construct, or an intervening variable; it cannot be directly observed, but has to be deduced from other events which are directly observable. It is, therefore, in the same position as such concepts as electrons, protons, positrons, etc., in the physical field, or as drive and habit in the psychological field. Such concepts can be extremely dangerous unless firm limits are set to speculation, and it is important to realize what the conditions are under which such concepts are acceptable to science.Hull has set down the rules for acceptance in admirable brevity: ‘Despite the great value of logical constructs or intervening variables in scientific theory, their use is attended with certain difficulties and even hazards. At bottom this is because the presence
F I G U R E i Diagrammatic representation of a relatively simple case of an intervening variable (X ) not directly observable but functionally related ( /) to the antecedent event {A) and to the consequent event (B), both A and B being directly observable. When an intervening variable is thus securely anchored to observables on both sides it can be safely employed in scientific theory.
