![]()
CHAPTER I
THEORIES OF PERSONALITY ORGANIZATION
“EXPERIMENT WITHOUT theory is blind; theory without experiment is lame.” There is perhaps no field in psychology where this saying of Kant’s applies with greater force than in the study of the structure of personality. Observers have been struck again and again by the fact that what should be a unitary field of study is cleft in two; that instead of an harmonious co-operation between theory and experiment, we have, on the one hand, an experimental school which investigates in the minutest detail processes having only the most tangential relevance to personality or to any plausible theoretical orientation, and, on the other, theoretical schools of the “dynamic” type whose theorizing proceeds without any proper basis in ascertained fact and without any consciousness of the need for verification. Most psychologists would agree that this division of labour has been carried to such extremes that it is threatening the very conception of “personality” as a legitimate field of scientific study.
Corresponding to this division into “experimentalists” and “theoreticians”, there are a number of other divisions among students of personality hardly less deep and hardly less acrimoniously debated. Yet to the onlooker it often appears that while both sides are right in their positive claims, they are wrong and one-sided in their condemnation of what other schools and other points of view have to contribute. Few would seriously argue that experiment could fruitfully be carried on without theory or theory lead to important advances without the check of experimentation. Similarly, most of the other disputes which appear so formidable in cold print seem amenable to compromise when each side’s arguments are carried to their logical conclusion.
As an example, we may take the very definition of the term “personality” itself. Here we find immediately an apparently irreconcilable opposition between those who lay stress on behavioural acts and those who lay stress instead on dynamic concepts. As an example of the behavioural type of definition, we may quote Watson (1930), according to whom personality is “the sum of activities that can be discovered by actual observation over a long enough period of time to give reliable information”. As an example of the dynamic type of definition we may quote Prince (1924), according to whom “personality is the sum-total of all the biological innate dispositions, impulses, tendencies, appetites, and instincts of the individual, and the acquired dispositions and tendencies”.
It is obvious that the concepts which enter into one kind of definition—observable behavioural acts—play no part in the other, which deals entirely with dynamic concepts—impulses, dispositions, instincts, and the like. Yet the opposition clearly cannot be as complete as it appears. We have no direct knowledge of instincts, dispositions, and impulses; they are abstract conceptions created to unify and make intelligible the observable behavioural acts from which they are abstracted. Without these behavioural acts the concepts would have no assignable meaning: all we can know about human behaviour must ultimately derive from observations of behaviour. Yet such observation of behaviour by itself is not enough. We must have concepts which denote aspects of behaviour common to a number of situations; science cannot exist without abstractions based on common properties. Both definitions therefore are one-sided; a proper definition must stress both the empirical source of our data and the theoretical nature of our unifying concepts.
For the purposes of this book, we shall adopt the following definitions: Personality is the more or less stable and enduring organization of a person’s character, temperament, intellect, and physique, which determines his unique adjustment to the environment. Character denotes a person’s more or less stable and enduring system of conative behaviour (“will”); Temperament, his more or less stable and enduring system of affective behaviour (“emotion”); Intellect, his more or less stable and enduring system of cognitive behaviour (“intelligence”); Physique, his more or less stable and enduring system of bodily configuration and neuro-endocrine endowment. It will be noted that this definition, which owes a great deal to Roback (1927), Allport (1937), and McKinnon (1944), stresses very much the concept of system, structure, or organization; in this it goes counter to the doctrine of specificity of behaviour, which held almost complete sway in American research from the early nineteen-twenties until quite recently. A few words may therefore be said regarding this issue of specificity versus generality, particularly as from one point of view all the experimental work reviewed in this book is intimately related to this problem.
Common-sense psychology unhesitatingly describes and explains behaviour in terms of traits, such as persistence, suggestibility, courage, punctuality, absent-mindedness, stage-struckness, “being one for the girls”, stuck-upness, and queerness, or posits the existence of types, such as the dandy, the intellectual, the quiet, the sporty, or the sociable type. For the greater part, orthodox psychology has taken over these concepts, and has presented us with traits such as ascendance-submission, perseveration, security-insecurity, and with types such as extraversion-introversion, schizothymia-cyclothymia, or Spranger’s Lebenstypen. This easy acceptance of these concepts has been challenged, however, by a number of critics, who hold that “there are no broad, general traits of personality, no general and consistent forms of conduct which, if they existed, would make for consistency of behaviour and stability of personality, but only independent and specific stimulus-response bonds or habits”.
This theory of specificity has its roots deep in the experimental tradition, and its à priori improbability should not prevent us from glancing at the main sources from which it draws its strength. The first of these sources is the Thorndikian type of learning theory prevalent around the first decades of this century. Learning is conceived in terms of S-R (stimulus-response) bonds after the manner of the reflex or the conditioned reflex, and these bonds are, of course, conceived to be entirely specific. If the organization of personality is largely a matter of learning—and here the great majority of writers have favoured an anti-hereditarian view, without however basing themselves on any convincing experimental evidence—then the specificity of the learning process should be mirrored in the final product of learning, i.e. the adult personality. And while S-R theories in the field of learning have been challenged by S-S (sign-significate) theories which maintain that learning is part of a larger problem of organization, particularly perceptual organization, these non-specific theories came into the field more recently, have been somewhat less influential historically, and have not carried over into the field of personality description to the same extent as the specificity theories.1
A second source, not unrelated to the first, has been the vast volume of work done on the problem of “transfer of training”. It used to be assumed that certain specific acts (learning verses by heart, or doing problems in arithmetic, or writing out French irregular verbs) would in the course of time lead to improvement in general abilities or faculties (memory, will-power, logical ability, and so on). James and Thorndike showed in a number of investigations that this easy assumption had little empirical foundation. When two groups of subjects were equated for their ability in a given task, such as learning poetry by heart, for instance, and one group subjected to a period of training in memorizing material which might even be closely similar to that on which they had been tested, while the other group was not given any training, then the predicted superiority of the former group over the latter on a repetition of the original task was not observed. Learning, apparently, is relatively specific: there is no general effect on the hypothetical faculties which such training was supposed to improve. Any transfer effects which might be observed were considered to be due, not to the action of broad mental “faculties”, but to the fact that the original and the practised activities had certain elements in common. This theory is known as the “theory of identical elements”; in Thorndike’s (1903) own words, “a change in one function alters any other only in so far as the two functions have as factors common elements.… To take a concrete example, improvement in addition will alter one’s ability in multiplication because addition is absolutely identical with a part of multiplication, and because certain other processes—e.g. eye movements and the inhibition of all save arithmetical impulses—are in part common to the two functions.” Development of personality, no less than of linguistic or numerical skills, is therefore seen as specific training of individual association, never as generalized improvement of larger mental units or “faculties”.1
A third source of the specificity theory of personality organization, equally influential as the other two, has been the direct experimental attack on the problem by Hartshorne and May (1928, 1929, 1930). These writers carried out a large-scale project, described in some detail on a later page, in which many hundreds of children were given the opportunity to behave in a dishonest, deceitful manner under conditions which apparently made discovery impossible, but which in reality were completely under experimental control. Other types of behaviour (persistent, moral, charitable, impulsive, and self-controlled behaviour, for instance) were also investigated by means of ingenious and largely novel techniques. The statistical treatment of the data was beyond cavil, and in view of the brilliance of the design and the technical excellence of the execution, this study has rightly been regarded as crucial in respect to the theory of specificity. When therefore Hartshorne and May found very low intercorrelations between their tests, and discovered that children who were honest, or persistent, or co-operative, or charitable in one test-situation were not always honest, or persistent, or co-operative, or charitable in another, their conclusion that these alleged qualities were “groups of specific habits rather than general traits” was very widely accepted as finally settling the issue in favour of the theory of specificity.
This powerful and imposing theoretical structure was subject to a variety of damaging criticisms, however, and none of the three sources on which it bases itself has remained unscathed. We have already mentioned that S-R theories were opposed by writers whose outlook was formed or at least influenced by Gestalt notions; Köhler, Koffka, Tolman, Adams, Zener, and others have developed theories which account for the observed facts without invoking the specific connections posited by the followers of Thorndike, and, indeed, Thorndike himself has admitted concepts into his system which are incompatible with a completely specifist point of view. There is no sign of any decision in this battle of learning theories, but it is already clear that if one’s theory of personality organization must be determined by one’s learning theory, then there is still freedom of choice between a “specific” and a “general” type of learning theory. It would seem to follow that a direct attack on the problem of specificity in the field of personality itself would be more promising than a somewhat lengthy wait for a decision in the field of learning theory.
Much the same must be said about the conclusion to be drawn from investigations into the problem of “transfer of training” and of “identical elements”. Allport’s (1937) brilliant criticism of the specifist contention is probably too well known to need repetition. By showing that the very notion of an “element” is completely ambiguous in the writings of those who support the Thorndikian view, and that the alleged “identity” of these elements is merely an a posteriori justification of the observed phenomena, without any value in predicting and without any possibility of verification, he has succeeded in throwing great doubt on the tenability of this whole view. When his criticisms are seen in the light of experimental work, which fails to show the theoretically predicted correspondence between improvement after practice, and the similarity between original task and practised task, we can only conclude that regardless of the eventual outcome of the argument regarding “transfer of training” and the theory of identical elements, our decision with regard to the question of specificity in the field of personality must rest on direct evidence from that field, rather than in deductions from principles of such uncertain validity.
We are thus led to a re-examination of the results of the Hartshorne-May study. While the detailed results are presented in a later chapter, we may here note certain doubts regarding the interpretation of their perfectly valid results made by these two authors. Let us examine first of all their finding that a child who behaves in a dishonest manner in one situation does not necessarily behave in a dishonest manner in another situation; their conclusion is that honesty is not a general trait but specific to the two situations. But this would assume that the two situations made equal demands on the hypothetical honesty of the child, a view for which there is no evidence at all. A child may fail a difficult item in an intelligence test and pass an easy one; because he passes one and fails on another, we do not argue that he is not behaving in a consistent manner! A child may tell what he considers a white lie, but balk at cheating; or he may cheat, but balk at stealing. To imagine that an advocate of the view that a general trait of honesty existed would necessarily deny the existence of degrees of temptation, or of degrees of immorality as between one act and another, is quite unrealistic, and there is no such implication in the “generality” theory. Related to the first point is a second, made by Hartshorne and May, and by many other writers since, namely that while some children do show the postulated trait, i.e. are always honest or persistent, and while others are consistent in never showing it, i.e. being always dishonest or lacking in persistence, the majority sometimes show the trait and sometimes not. Thus the trait is supposedly applicable only to a few cases, i.e. those who demonstrate it consistently, and not to others. By a similar argument it might be maintained that the concept of intelligence is applicable only to those who never fail an item or to those who fail every item! If we conceive of honesty as constituting a continuum, then the most honest should indeed never cheat and the least honest always; intermediate grades of honesty should be reflected in action by cheating when temptation is strong or when the immorality involved is rather slight, and by not cheating when temptation is weak or the immorality involved strong. For a given degree of temptation and immorality of the act, we would then be able to predict with as much accuracy for the intermediate child as for the extreme, just as we can predict for the child of average intelligence as easily as for the genius or the dunce whether he will succeed or fail with any given problem.
As a third argument, Hartshorne and May advance the view that the very low intercorrelations between the different tests for each one of the various personality qualities me...
