WHAT IS A TRAIT OF PERSONALITY!

*
B Y GOEDON W. ATJVPOBT
HAEVABD TTSJIVEESITY

T THE heart of all investigation of personality lies the puzzling problem of the nature of the unit or element which is the carrier of the distinctive behavior of a man. Reflexes and habits are too specific in reference, and connote constancy rather than consistency in behavior; attitudes are ill-defined, and as employed by various writers refer to determining tendencies that range in inclusiveness from the Aufgabe to the Weltanschauung; dispositions and tendencies are even less definitive. But traits, although appropriated by all manner of writers for all manner of purposes, may still be salvaged, I think, and limited in their reference to a certain definite conception of a generalized response-unit in which resides the distinctive quality of behavior that reflects personality. Foes as well as friends of the doctrine of traits will gain from a more consistent use of the term. The doctrine itself has never been explicitly stated. It is my purpose with the aid of eight criteria to define trait, and to state the logic and some of the evidence for the admission of this concept to good standing in psychology.

1. A trait has more than nominal existence. A trait may be said to have the same kind of existence that a habit of a complex order has. Habits of a complex, or higher, order have long been accepted as household facts in psychology. There is no reason to believe that the mechanism which produces such habits (integration, Gestaltung, or whatever it may be) stops short of producing the more generalized habits which are here called traits of personality. 2. A trait is more generalized than a habit. Within a personality there are, of course, many independent habits; but there is also so much integration, organization, and coherence among habits that we have no choice but to recognize great systems of interdependent habits. If the habit of brushing one's teeth can be shown, statistically or genetically, to be unreiEad at the Ninth International Congress of Psychology, New Haven, 1929. 368

Gordon W. Allport

369

lated to the habit of dominating a tradesman, there can be no question of a common trait involving both these habits; but if the habit of dominating a tradesman aan be shown, statistically" or genetically, to be related to the habit of bluffing one's way pastguards, there is the presumption that a common trait of personality exists which includes these two habits. Traits may conceivably embrace anywhere from two habits to a legion of habits^ In this way, there may be said to be major, widely extensifiecf traits, and minor, less generalized traits in a given personality. 3. A trait is dynamic, or at least determinative. It is not the stimulus that is the crucial determinant in behavior that expresses personality; it is the trait itself that is decisive. Once formed a trait seems to have the capacity of directing responses to stimuli into characteristic channels. This emphasis upon the dynamic nature of traits, ascribing to them a capacity for guiding the specific response, is variously recognized by many writers. The principle is nothing more than that which has been subscribed to in various connections by Woodworth, Prince, Sherrington, Coghill, Kurt Lewin, Troland, Lloyd Morgan, Thurstone, Bentley, Stern, and others. From this general point of view traits might be called "derived drives" or "derived motives". Whatever they are called they may be regarded as playing a motivating role in each act, thus endowing the separate adjustments of the individual to specific stimuli with that adverbial quality that is the very essence of personality. Some psychologists may balk at the doctrine of the absorption of driving power into the integrated mechanism of traits. If so, it is equally possible, without violence to the other criteria of this paper, to accept the view that a trait is a generalized neural set which is activated ecphorically or redintegratively. But it seems to me that this second doctrine is only slightly less dynamic than the first. The difference is simply one between trait considered as a drive aroused through the operation of a specific stimulus, and trait conceived as powerfully directive when an effective stimulus arouses the organism to action. 4. The existence of a trait may be established empirically or statistically. In order to know that a person has a habit it is necessary to have evidence of repeated reactions of a constant type. Similarly in order to know that an individual has a trait it is necessary

-"370

What Is a Trait of Personality t

to have evidence of repeated reactions which, though not necessarily constant in type, seem none the less to be consistently a function of the same underlying determinant. If this evidence is gathered casually by mere observation of the subject or through the reading of a case-history or biography, it may be called empirical evidence. More exactly, of course, the existence of a trait may be established with the aid of statistical techniques that determine the degree of coherence among the separate responses. Although this employment of statistical aid is highly desirable, it is not necessary to wait for such evidence before speaking of traits, any more than it would be necessary to refrain from speaking of the habit of biting fingernails until the exact frequency of the occurrence is known. Statistical methods are at present better suited to intellective than to conative functions, and it is with the latter that we are chiefly concerned in our studies of personality. 5. Traits are only relatively independent of each other. The investigator desires, of course, to discover what the fundamental traits of personality are, that is to say, what broad trends in behavior do exist independently of one another. Actually with the test methods and correlational procedures in use, completely independent variation is seldom found. In one study expansion correlated with extroversion to the extent of +.39, ascendance with conservatism, +.22, and humor with insight, +.83, and so on. This overlap may be due to several factors, the most obvious being the tendency of the organism to react in an integrated fashion, so that when concrete acts are observed or tested they reflect not only the trait under examination, but also simultaneously other traits; several traits may thus converge into a final common path. It seems safe, therefore, to predict that traits can never be completely isolated for study, since they never show more -than a relative independence of one another. In the instance just cited, it it doubtful whether humor and insight (provided their close relationship is verified in subsequent studies) represent distinct traits. In the future perhaps it may be possible to agree upon a certain magnitude of correlation below which it will be acceptable to speak of separate traits, and above which one trait only will be recognized. If one trait only is indicated it will presumably represent a broadly generalized disposition. For example, if humor and insight cannot be established as

Gordon W. Allport

371

* ] -,, I

independent traits, it will be necessary to recognize a more inclusive trait, and name it perhaps "sense of proportion". 6. A trait of personality, psychologically considered, is not the same as moral quality. A trait of personality may or may not coincide with some well-defined, conventional, social concept. Extroversion, ascendance, social participation, and insight are free from preconceived moral significance, large because each is a word newly coined or adapted to fit a psychological discovery. It would be ideal if we could in this way find our traits first and then name them. But honesty, loyalty, neatness, and tact, though encrusted with social significance, may likewise represent true traits of personality. The danger is that in devising scales for their measurement we may be bound by the conventional meanings, and thus be led away from the precise integration as it exists in a given individual. Where possible it would be well for us to find our traits first, and then seek devaluated terms with which to characterize our discoveries. 7. Acts, and even habits, that are inconsistent with a trait are not proof of the non-existence of the trait. The objection most often considered fatal to the doctrine of traits has been illustrated as follows: "An individual may be habitually neat with respect to his person, and characteristically slovenly in his handwriting or the care of his desk." In the first place this observation fails to state that there are cases frequently met where a constant level of neatness is maintained in all of a person's acts, giving unmistakable empirical evidence that the trait of neatness is, in some people at least, thoroughly and permanently integrated. All people must not be expected to show the same degree of integration in respect to a given trait. What is a major trait in one personality may be a minor trait, or even non-existent in another personality. In the second place, we must concede that there may be opposed integrations, i.e., contradictory traits, in a single personality. The same individual may have a trait both of neatness and of carelessness, of ascendance and submission, although frequently of unequal strength. In the third place there are in every personality instances of acts that are unrelated to existent traits, the product of the stimulus and of the attitude of the moment. Even the characteristically neat person may become careless in his haste to catch a train.

372

What Is a Trait of Personality?

But to say that not all of a person's acts reflect some higher integration, is not to say that no such higher integrations exist. 8. A trait may be viewed either in the light of the personality which contains it, or in the light of its distribution in the population at large. Each trait has both its unique and its universal aspect. In its unique aspect, the trait takes its significance entirely from the role it plays in the personality as a whole. In its universal aspect, the trait is arbitrarily isolated for study, and a comparison is made between individuals in respect to it. Prom this second point of view traits merely extend the familiar field of the psychology of individual differences. There may be relatively few traits, a few hundred perhaps, that are universal enough to be scaled in the population at large; whereas there may be in a single personality a thousand traits distinguishable to a discerning observer. For this reason, after a scientific schedule of universal traits is compiled, there will still be the field of artistic endeavor for psychologists in apprehending correctly the subtle and unique traits peculiar to one personality alone, and in discovering the pattern which obtains betiveen these traits in the same personality.