University of the South
Why would you wane to read this chapter? To be sure, any science has a history; but the history of a science, while a legitimate field of scholarly inquiry, is usually quite separate from the science itself. Few chemists are concerned with Johann Bechers now-discredited phlogiston theory or Dmitri Mendeleyevs life and how he came to construct the periodic table of the elements. If psychological science is concerned with what is true now, why bother with what people used to think was true? Why, then, a history of personality psychology?
USES OF A HISTORICAL PERSPECTIVE
We suggest three important reasons for personality psychologists to know something of their history. First, origins are intrinsically interesting, even compelling. As Freud (1908/1959b) once suggested, the question "Where did I come
from?" may be the child's "first grand problem of life" (p. 212). Origins are often a critical part of adult identity (witness peoples interest in genealogy)- In fact) the history of psychology is of increasing interest to psychologists, as manifest in the new APA journal, History of Psychology. In this chapter, then, we offer an outline of the "family history" of personality psychology. •
Second, the study of history can help us avoid making the same journeys over the same terrain, repeating the mistakes of the past. If we know what has gone before, we may discover that some questions have been asked before, perhaps even answered! As the philosopher Santayana (1906) cautioned, "Progress, far from consisting in change, depends on rctcntiveness. . . . When experience is not retained .. . infancy is perpetual. Those who cannot remember the past are condemned to repeat it" (p. 284).
Further, any discipline is the way it is because its early workers formulated certain questions and framed certain key concepts. By under-
standing how our field evolved and was constructed—by particular psychologists, in particular social contexts—we gain a broader perspective on contemporary questions and issues. For example, Gordon Allport's decision to ratify "trait" as the key concept for the newly emerging field of personality (Allport, 1927) had important consequences in framing such later issues as the ways in which personality is measured, stability versus change, the roles of "personality" and "situation" in the explanation of behavior, and the relationship between trait and other personality constructs such as motive (Winter, John, Stewart, Klohnen, &c Duncan, 1998).
Finally, no science is divorced from its surrounding social conditions and values-climate. For example, any attempt to study the psychology of personality presumes the belief that "personality" is important and worth explaining. This belief, in turn, rests on a philosophical individualism and a view of the individual as cause or agent. Because Americans are particularly likely to hold such individualistic beliefs and reluctant to acknowledge the importance of collectivities (Bellah, Madsen, Sullivan, Swidler, & Tipton, 1985), it is not surprising that "personality" has been a major focus of psychological inquiry as well as popular culture in the United States (see Susman, 1979). One observer suggested that "no country in the world is so driven by personality as this one" and that Americans have a "hunger to identify with larger-than-life personalities" ("Only spectacular crimes," 1994). Our historical account, therefore, will emphasize developments in the United States. It should be taken as building upon earlier historical handbook chapters by Burnham (1968), Pervin (1990), McAdams (1997) and Runyan (1997)— and, indeed, on those entire earlier handbooks.
NINETEENTH-CENTURY INTELLECTUAL ANTECEDENTS OF PERSONALITY PSYCHOLOGY
The modern psychological study of personality, which began to flourish during the early decades of the 20th century, has roots in three 19th-century intellectual themes: a deep belief in individualism, a pervasive concern with irrationality and the unconscious, and a strong emphasis on measurement. Although these three themes are by no means fully consistent with each other and not every personality psychologist emphasized
all three, they did shape (and limit) the emerging field of personality psychology in important ways that can still be recognized.
Individualism and the "Skin-Bounded" Individual
Personality psychology evolved and flourished in the Western philosophical-political climate of individualism—the belief that individuals are important and unique. Such a discipline would be unlikely to develop in a homogeneous society in which everyone lived through the same small set of life cycles. Thus the cultural historian Burck-hardt (1860/1954) concluded that in the Middle Ages, people were conscious of themselves "only as a member of a race, people, party, family, or corporation—only through some general category" (p. 101). After the Renaissance, however, people recognized themselves as individuals.
From the individualist perspective, "persons," the very subject matter of personality psychology, are construed as single bothes, bounded by skin. Anything else (objects and environments, intimate partners, other people, family, solidary groups, communities, institutions, cultures) is construed as "outside," wholly external to the autonomous skin-bounded individual person. From this perspective, they only exist and are understood as internal representations, objects of contractual relations, or encroachments on the individual.1
The individualistic perspective, however, leads to a paradox: Often we do not seem so very different from each other. For example, the many stuthes of bogus personality feedback (Dickson & Kelly, 1985) suggest that people readily accept uniform, standardized feedback as an accurate description of their own "individual" selves. And Singer (1995), after studying thousands of college students' self-stories, was "overwhelmed by the narrative similarities they bring to the important events in their lives" (p. 452, emphasis added). Further, an excessively individualistic perspective can blind psychologists to the importance of collective aspects of personality such as groups, social identities, and cultural symbols.
In revolt against the 18th-century "age of reason," the 19th-century Romantic movement exalted the role of unconscious and irrational emotions, spontaneity, and impulsivity in literature,
Chapter 1. History of Modern Personality Theory and Research the arts, and philosophy. In personality psychology, this influence was most obvious in psychoanalysis, particularly Freud's concepts of id and primary process. The notion of an unconscious still survives today, albeit in a muted and less grandiose form, in the claim that many important processes are implicit or automatic—that is, they operate outside of conscious awareness (Kihlstrom, 1990, and Chapter 17, this volume; see also Bargh, 1982; Greenwald & Banaji, 1995; Nisbett & Wilson, 1977). Still, an exclusive focus on unconscious or implicit forces may lead to a neglect or underestimation of the importance of rational planning and directed behavior (see Cantor & Zirkel, 1990).
Emphasis on Measurement
In devoting an extraordinary amount of attention to issues of measurement and psyehomct-rics, personality psychology tried to follow in the footsteps of the prestigious "exact sciences" that had developed so rapidly in the late 19th century. The modern emphasis on precision and measurement was prefigured in the writings of British scientist and statistician Sir Francis Gal-ton (1884): "The character which shapes our conduct is a definite and durable 'something,' and therefore ... it is reasonable to attempt to measure it" (p. 179).2 In the early 20th century, American experimental psychologist £. L. Thorndike (1914) insisted that "if a thing exists, it exists in some amount; and if it exists in some amount, it can be measured" (p. 141; emphasis in original).
The prestige of Binct and the early intelligence "tests" also increased the concern with measurement in personality psychology. A rapidly developing assessment technology led to a number of paper and pencil "tests" of personality, modeled after the early intelligence tests and consisting of questionnaire "items" intended to measure the "traits" of personality. Later on, personality psychologists expanded their measurement technology to include ratings and behavior observations.
The emphasis on measurement also fit with the imperative, strongly felt by personality psychology in its beginning decades, to be useful, to furnish assistance to a corporate culture and a government suddenly confronted by dramatic changes and the need to "manage" and control an American population that had suddenly become larger, more diverse, and "difficult" (Parker, 1991). This point is nicely illustrated by
the example of Woodworths Personal Data Sheet, which was probably the first personality test based on the IQ test model of adding "scores" on discrete individual "items" to get a total. By the time the United States finally entered the war in April 1917, the experience of other armies suggested that many solthers were vulnerable to "shell shock" or "war neurosis." (Nowadays this would probably be called post-traumatic stress disorder.) The American Psychological Association quickly set up a committee charged with developing a diagnostic test of "susceptibility to shock," which was conceived as one aspect of a more general emotional instability (Camfield, 1969, pp. 126, 131-132). Acting without the benefit of much prior knowledge or technique, Robert Woodworth collected a long list of symptoms from the case histories of "neurotic subjects" and turned them into a series of 116 simple questions that could be answered Yes or No (for example, "Do you usually feel well and strong?" or "Has your family always treated you right?"). Overall scores, calculated by summing scores of the individual items, differentiated "normal" solthers from diagnosed neurotics or returned "shell shock" cases (Woodwoith, 1919, 1932, p. 374). The resulting Personal Data Sheet was the first objective, self-report "inventory" purporting to measure a personality characteristic, in this case what was later labeled Neuroticism.
A focus on measurement was obviously beneficial to any science that aspired to prestige by developing along the traditional positivistic path blazed by 19th-century physics. Conversely, it was a handicap insofar as it constrained the scope of investigation and explanation to that which was easily measured, thereby neglecting more subtle and complex personality characteristics and processes. For example, self-report questionnaires constructed on the iQ-test "item" model are sensitive only to conscious (rather than implicit) sentiments. They are usually based on the atomistic assumption that a complex whole can be broken down into a series of small component parts, each of which is equal to every other and all of which combine in additive (rather than interactive or nonlinear) ways. In this respect, the "measurement imperative" of personality psychology sometimes resembled the famous joke about the drunken man who had lost his keys on the dark side of the street but looked for them on the (other) lighted side, "because that's where the light is."
THE TWO TASKS OF PERSONALITY PSYCHOLOGY
From its earliest days, American personality psychology involved two related but contrasting endeavors: (1) the study of individual difference^ or the dimensions along which people differ from each other, and (2) the study of individual persons as unique, integrated wholes (see, e.g., Lamiell, 1997; McAdams, 1997; Murphy, 1932; Sanford, 1963). These two endeavors have been variously labeled as "analytic" versus "structural," or "quantitative" versus "qualitative," respectively. According to one early commentator (Young, 1928):
A review of the whole gamut of personality stuthes . . . reveals two essentially distinct approaches to the data. One of these, which has been developed by recent psychology, concerns itself with a ... cross-sectional treatment of personality in terms of traits, attitudes and habits. The other, which has arisen from a number of sources, especially psychiatry, treats personality from a functional, historical-genetic standpoint, (p. 431)
Much of the difference between these two approaches is also captured by the "nomothetic-idiographic" dichotomy—terms first used by Windelband (1894/1904), later adopted by Allport (1937) ,3 and still a lively topic at the end of the century (West, 1983; see also Lamiell, 1998).
The Psychometric "Analysis" Approach
Discussing the analytic or quantitative approach, Murphy (1932) characterized its view of personality as the "sum of all of an individuals traits" (p. 386). In his view, psychologists who measured and stuthed the intercorreladons of separate personality traits conceived of personality "as the answer to a complicated arithmetical problem" (p. 386). The practical goal of personality research was to predict, modify, and control behavior, with individual differences conceived as coefficients to be supplied to linear, additive prediction equations.4
Influence of Intelligence Testing
We have called this the "psychometric" approach because, according to an early reviewer, it "owes its prominence to the work of Gallon, Pearson, Cattell, Thorndike and Terman with their inves-
tigations of individual differences, particularly in the field of intelligence" (Young, 1928, p. 431). During the first two decades of the 20ih century, psychologists had developed "mental tests" for selection, diagnosis, and placement, as psychologists tried to demonstrate their usefulness in solving urgent practical problems associated with immigration, labor unions, and schools, as well as the 1917-1918 national war mobilization (Danziger, 1990; Parker, 1991; Vernon, 1933; see Schaffcr, 1991, pp. 133-139, on the effects of World War I in particular).
In the post-World War I period, however, many critics questioned the predictive utility of intelligence tests in various applied settings (Parker, 1991) and suggested that measures of personality or character traits were needed to improve the prediction of military, managerial, industrial, and educational performance (see, e.g., Fernald, 1920; PorTcnbcrger, 1922; Pressey, 1921). Still, these newer "character tests" were based on the mental test model in preference to other, less "efficient" methods (Parker, 1991). The content of trait measures also reflected practical demands: "Ascendance-Submission" and "Extraversion-Imroversion," considered relevant to selecting business managers and military officers, received the greatest attention (Danziger, 1990; Parker, 1991).
"Mental Hygiene" and Personality
Yet another important influence on the psychometric approach to personality during the 1920-1930 period was the so-called mental hygiene movement. This well-funded alliance of psychiatrists, educators, and social workers viewed individual maladjustment as the root cause of a wide variety of social and personal problems (S. Cohen, 1983; see also Parker, 1991). With its "personality" focus, the movement enlisted psychologists to supply a "scientific" basis for the therapeutic efforts of mental hygiene workers. As Danziger (1990) noted, "in practice this generally came down to the construction and application of scales diat would subject 'personality' to the rigors of measurement and so convert it from merely an object of social intervention to an object of science" (p. 164).
Chapter 1. History of Modern Personality Theory and Research ness of their work on measuring individual differences suggests the need for viewing history from a political and moral perspective. In fact, most of the "practical social problems" personality psychologists tried to solve were problems faced (and framed) by elite groups: selecting managers who could maximize profits; selecting military officers who could win wars; controlling an increasingly diverse population in the country's schools, factories, crowded cities, and prisons; and controlling "deviant" behavior, at least to the extent of promoting labels suggesting individual pathology instead of social problems. Not surprisingly, the "solutions" supplied by personality psychologists tended to leave existing power relations intact or even reinforced, because the intellectual roots of personality psychology—radical individualism, supported by the mystique of "science"—were quite compatible with the ideological stance of the ruling elites who financed the research. Thus the practically oriented psychometric approach to personality psychology runs the danger of becoming a technology that deliberately or unwittingly functions mainly to serve the interests of the powerful.5
The Psychiatric and Historical "Interpretation" Approach
Murphy (1932) contrasted the psychometric approach to personality with the psychiatric focus on "personality as a whole," observing that psychiatric conceptions "added much to the richness of the term 'personality"9 by introducing topics such as dissociation and unconscious mental processes (p. 387). Young (1928) was especially enthusiastic about this historical-interpretive approach (though he noted that at the time of his writing, it "has had very little attention from the psychologist," p. 437): "[It] has arisen from the study of literary biography, from historical biography, but especially from psychiatry and sociology. Here we find the great biographers, the psychoanalysts led by Freud, Jung and Adler, the invaluable work of Healy [a neurologist] and latterly the sociological reformulation of W. I. Thomas" (p. 431).
Mainstream psychologists ignored or criticized biographical and case study methods and were (at least initially) quite hostile toward psychoanalysis (Danziger, 1990, 1997, especially p. 125). Given their concern for scientific respectability via sophisticated quantitative techniques, it is not surprising that they viewed the
case study methods of psychiatrists and psychoanalysts as part of an "old-fashioned and unscientific" medical tradition (Hale, 1971, p. 115).
Discussing the case study literature in their review of personality, Allport and Vcrnon (1930) noted the contributions of psychiatrists and sociologists (rather than psychologists) and echoed the prevailing view that the method was "unsatisfactory." However, they went on to suggest mat "the concrete individual has eluded study by any other approach" and concluded with the hope that "in the future there will undoubtedly be attempts to standardize the case-study in some way which will reduce its dependence upon the uncontrolled artistry of the author" (p. 700).
The whole-person or interpretive approach to personality had been especially prominent in German psychology during the first three decades of the 20th century. Influenced by the philosopher Dilthey s view of psychology as a "human science" (Geisteswiuenschafi), it emphasized the ways in which personality characteristics and other psychological processes were organized or patterned within the unique individual. Writing for an American authence about the German qualitative approach to the "undivided" personality, Allport summarized this perspective: "More fundamental than differential psychology [i.e., the psychometric focus on dimensions of difference among people], by far, is the problem of the nature, the activity, and the unity of the total personality" (Allport, 1923, p. 614; emphasis in original). Allport traced the German emphasis on structured "wholes" in Gestalt psychology, the personalistic psychology of William Stern, and Eduard Spranger's method of "intuitive understanding" (Ventehen) (Allport, 1923, 1924a, 1924b, 1929).
However, Allports early attempts to introduce this approach to personality to American psychologists apparently met with little response. Vernon (1933) noted that most American psy-chometrists were either unaware of the German qualitative approach or else found it scientifically unacceptable.
The Uneasy Coalition between "Analysis" and "Interpretation" Approaches
For the past 70 years, the two approaches to personality psychology—the study of individual personality differences and the study of individual persons—have existed in an uneasy coalition or truce. As part of their efforts to broaden (and systematize) personality, Allport (1937) and
Murray (1938) tried to move the field toward studying persons as integrated wholes, but as several later commentators noted, personality psychologists generally avoided stuthes of individual lives through most of the century (see, e.g., Carlson, 1971; McAdams & West, 1997; Runyan, 1997).
Even Allport and Murray seemed ambivalent about case stuthes. As a journal editor and professor, Allport advocated studying individual persons (sec, e.g., Allport, 19245, 1929, 1937, 1942) and tried (unsuccessfully) to develop a set of systematic rules for writing life histories (Allport, 1967; see also Barenbaum, 1997a, 1997c; Garracy, 1981). Yet he himself actually published only one case study (Allport, 1963), and much of his research involved nomo-thetic traits (Allport, 1928; Vernon & Allport, 1931). Situated within Harvard's scientifically oriented academic department of psychology, Allport opted for a moderate, eclectic approach that included both analytic and interpretive methods (sec Nicholson, 1997a, but also Pandora, 1997).
Situated as a maverick in opposition to the "scientistic" Harvard department (Robinson, 1992), Murray still needed to justify his foundation support on the basis of the "scientific" status of his work (see also Triplet, 1983, 1992). Although Murray's group developed elaborate procedures for the intensive study of individual persons, their book (Murray, 1938) actually included only one case.6 Later, in establishing an assessment program for the wartime Office of Strategic Services (OSS), Murray modified his "diagnostic council" method even more in the direction of more quantitative assessments (McLeod, 1992).
From a historical perspective, it is easy to see that the psychometric tradition was already well established in psychology by the time that personality emerged as a separate field, so that this prior "triumph of the aggregate" (Danziger, 1997, p. 68) led to an early dominance of the "individual differences" tradition over the "study of individuals" tradition. For example, when the American Psychological Association took over publication of the Journal of Abnormal and Social Psychology (from the psycruatrically oriented American Psychopathological Association) in 1925, the publication pattern changed dramatically (Danziger, 1990):
What was given up was an earlier tradition of detailed study of individuals.... Whereas in 1924[,]
80 percent of the empirical papers published were based on the study of individual cases, that proportion dropped to 25 percent in the following year. Their place was taken by statistical stuthes based on group data, as was required by the prevailing Galtonian paradigm for psychological measurement, (p. 165)
The marginal status of stuthes of individual persons in personality psychology, then, really continues an early trend. Yet interest in individual cases has continued to survive—at first mainly among students and associates of Allport and Murray (e.g., Polansky, 1941; Rosenzweig, 1943; White, 1952), but in recent years among a much broader and larger array of personality psychologists (Runyan, 1997, p. 42; see the review by McAdams & West, 1997; also the collections edited by Franz & Stewart, 1994, and McAdams & Ochberg, 1988)7
1921-1946: THE CRITICAL FORMATIVE ERA OF PERSONALITY PSYCHOLOGY IN THE UNITED STATES
We suggest that while the "official" emergence of personality psychology is often dated by the appearance of the "canonical" texts by Allport (1937), Murray (1938),8 and Stagner (1937), it really began at least 15 years earlier. The 1921-1938 period was a time of intense research activity; the years between 1938 and 1946 were really the consolidation of "modern" personality concepts and methods, as well as the widespread institutional recognition of personality as a specialty within psychology departments (Danziger, 1990; 1997; Parker, 1991). We first review the period before the canonical texts, then discuss the latter period in terms of the influence of three major figures who had substantial and enduring impact on the field: Gordon Allport (1897-1967), Henry Murray (1893-1988), and Raymond Cattell (1905-1998).
Origins: The Decade of the 1920s
The first American review of the psychological literature on "personality and character" appeared in 1921 (Allport, 1921).9 Most of the various sources cited by Allport involved traits, which suggests that the trait concept had already achieved a theoretical dominance by that time (Danziger, 1990; Parker, 1991). The article
Chapter 1. History of Modern Personality Theory and Research focused primarily on the distinction between "personality" and "character," two concepts that had been used interchangeably by American psychologists up to that time. In agreement with the behaviorists, Allport suggested that the latter term, defined as "the personality evaluated according to prevailing standards of conduct" (p. 443), was not an appropriate topic for psychological study. Allport continued to advocate the use of "personality" in preference to "character" (1927; Allport & Vernon, 1930); this soon became standard practice, not only for reasons of scientific respectability but also on account of a cultural shin from moralism to pragmatism (Nicholson, 1998).
Before the 1920s, the term "personality" had been used primarily in discussions of abnormal psychology, which was considered to be the province of the medical specialty of psychiatry (Parker, 1991). Another indication that the field of personality was emerging as a separate and autonomous area of psychological research was the 1921 expansion of the Journal of Abnormal Psychology into the Journal of Abnormal Psychology ana Social Psychology™ with Floyd Allport as an additional editor "in cooperation with" Morion Prince. In a joint editorial, they noted that psychopathologists and social psychologists shared an interest in the "dynamics of human nature" and invited contributions on a number of topics, including "the foundation-study of human traits" and "the personality of the individual" (Prince &: Allport, 1921, p. 2). The lead article of the first issue of the expanded journal was the study of personality traits by the Allport brothers (Allport & Allport, 1921).
Seven years later, Prince and Moore (1928) introduced a special issue of the Journal of Abnormal and Social Psychology devoted to personality with the observation that "the rapidly increasing number of manuscripts dealing with problems of temperament, character and personal traits shows that an extraordinary pressure of investigation is being directed toward this subject which until quite recently numbered but a few brief pages in any standard textbook of psychology." They noted the "necessity of ever widening approaches and more manifold stuthes," and concluded with a prophetic reference to "the complex task that faces the author of a Psychology of Personality ten years hence" (p. 117).
By the end of the 1920s there were several institutional indicators of the widespread interest in personality as a field of psychology. During the period 1923-1928, the American Psychological Association responded to its members' increasing interest in personality by scheduling convention sessions on "character" and "personality," sometimes in the practical context of vocational guidance or selection (Parker, 1991). Similar developments occurred at other conferences and meetings. Support for personality research came from a variety of foundations. Other signs of recognition of the new discipline included the founding of the journal Character and Personality (later renamed Journal of Personality) in 1932, and the 1934 addition or personality" as a category in the Psychological Abstracts (Parker, 1991).
Allport: Defining and Systematizing the Field of Personality
Gordon Allport s early efforts to define and systematize the field of personality included what he later described as "perhaps the first American dissertation written explicitly on the question of component traits of personality" (Allport, 1967, p. 9}.n Trained in social ethics as well as psychology, Allport attempted in his 1922 dissertation to define "a sound conception of human personality" on which to base "effective social service," arguing that "sound theory must underlie application" (Allport, 1967, p. 7). Rejecting as superficial many of the existing trait measures derived from exclusively practical concerns, he adopted a behavioristic definition of traits as "systems of habits" and designed measures of those traits he saw as basic components of personality.
As a postdoctoral student in Germany (1922-1923), Allport encountered psychological approaches to personality that "converted" him from his "semifaith in behaviorism" (1967, p. 12). From that time onward, he advocated structural approaches such as Gestalt psychology, Sprangers intuitive method, and Sterns person-alistic emphasis on the uniqueness and unity of personality (see, e.g., Allport, 1923, 1924a, 1924b, 1929). Stern was especially influential in shaping Allport s views (Allport, 1967): "From Stern in particular I learned that a chasm exists between the common variety of differential psychology (which he himself had largely invented ...) and a truly personalistic psychology that focuses upon the organization, not the mere profiling, of an individuals traits" (p. 10). Allport was
particularly impressed with Sterns "repudiation" of his earlier view of "personality as a sum-total of traits" in favor of an emphasis on "the total personality" (Allport, 1924a, p. 359), and Sterns influence is especially reflected in many of Allports later (1937) views (e.g., the very definition of "trait," the concept of functional autonomy of motives, and the distinction between idiographic and nomothetic methods; see Baren-baum, 1997b).
Returning from Europe in 1924, Allport continued his efforts to define a systematic psychology of personality. Having accepted an offer to teach social ethics at Harvard, he taught what he later described as "probably the first course on the subject [of personality] offered in an American college" (Allport, 1967, p. 9). With the title "Personality and Social Amelioration," the course was first offered in 1924 in the Department of Social Ethics. The following year it was cross-listed in psychology under a new title, "Personality: Its Psychological and Social Aspects" (Nicholson, I997b, pp. 734-735). During his time at Dartmouth (from 1926 to 1930) and later at Harvard (in the psychology department), Allport refined his definition of traits (1931) and attempted to synthesize analytic and interpretive methods of studying personality. He experimented with case methods in teaching (1929) and developed the Study of Values, an instrument based on Spranger's six fold conception of value (Vernon & Allport, 1931).
Although personality research had proliferated during the 1920s, Allport and other commentators noted a continuing lack of attention to the development of personality theory. According to Allport and Vernon (1930), "experimental approaches in recent years have far outgrown an adequate philosophical and theoretical Background, especially in America" (p. 702). In a similar vein, Murphy (1932) commented that "those who have occupied themselves with the measurement of personality traits have in general been even less concerned with the theory of personality than most intelligence testers have been with theories of intelligence" (p. 386). Vernon (1933) attributed this lack of attention to American psychologists' preoccupation with applied issues: "The preliminary study of fundamental theoretical problems seems to have been neglected for the sake of practical results" (p. 166).
Allport's 1937 text, like his early reviews, attempted to define and systematize the field of personality psychology in order to provide
"co-ordinating concepts and theories" (1937, p. ix). His survey of existing definitions of personality and methods of studying it led to a renewed emphasis on trait as the fundamental unit of study for personality. Traits, Allport suggested, were neuropsychic systems with dynamic or motivational properties. At the same time Allports focus throughout was on "the manifest individuality of mind" (p. vii), which implied both idiographic and nomothetic methods. Although some critics considered this focus on the individual unscientific (Bills, 1938; Skaggs,' 1945), most reviewers immediately recognized Allports "precise, well-integrated, and thoughtful review" of the field of personality and his "coherent defense of the individual as a proper subject for scientific study" (Jenkins, 1938, p. 777). Even those who took issue with Allport's views agreed that the book would be influential (Guilford, 1938; Hollingworth, 1938; Jenkins, 1938).
Allport's influence on the emerging field of personality psychology and his anticipation of issues of enduring significance in personality theory and research are well known (see, e.g., Craik, Hogan, & Wolfe, 1993). We have discussed above his efforts to reconcile analytic and interpretive approaches and his emphasis on the structure and organization of the individual personality. Further, on account of his pioneering contribution to the lexical study of traits (Allport & Odbert, 1936), as well his insistence that "a theory of personality requires more than a descriptive taxonomy" of traits, John and Robins (1993) claim Allport as the "father and critic of the Five-Factor Model" (pp. 225, 215), (We mention below the influence of Allports concepts of motivation and the self.) Cohler (1993) has succinctly summarized Allport's importance to the field of personality:
Allport's primary contribution to the study of the person may be less a matter of theoretical notions, methodological prescriptions, or empirical work than of his uncanny ability to comprehend the major issues in the field. As earl/ as his first papers, Allport was aware of the fundamental problems confronting those who wished to study persons, such as the problem of distinguishing between text and interpretation, the advantages and drawbacks of individual difference formulations in the study of personality structure, the fact that traits as well as environments were ever-changing, and the challenges of accounting for continuity and change in lives over time. (p. 142)
Chapter 1. Hiitory of Modem Personality Theory and Research Murray: Expanding Disciplinary Boundaries
Henry Murray came to the study of personality by way of psychoanalysis and abnormal psychology. Trained originally in medicine and in biochemistry, Murray chose a career in "depth psychology" after encountering the work of Jung and Freud. Only after accepting a position as assistant director of Morton Prince's Harvard Psychological Clinic in 1926, however, did he realize that the academic psychology of the time had very little in common with psychoanalysis (Murray, 1940; Murray, 1967; Robinson, 1992; Triplet, 1992). At a time when psychologists were struggling to define and delimit the disciplinary boundaries of their field, Murray's unorthodox and divergent interests were not acceptable to proponents of a strictly scientific psychology; in fact, they almost cost him his position at Harvard (Triplet, 1983).
Murray's eclectic, multimethod approach reflected his medical training and the nonpositivist philosophy of science of Whitehead and Hen-derson (Laughlin, 1973; Robinson, 1992), as well as the theories of Freud and Jung. Yet at the same time, Murray's background in biochemistry permeated his work; for example, in the explicit analogy between his classification of the "variables of personality" and the periodic table of chemistry (1938, pp. 142-143). The difficulty of such an eclectic enterprise is perhaps indicated by Murray's numerous revisions of his system of personality variables (Murray, 1951, 1968, 1977) and by the tentative titles he used for these works (e.g., "Preparations for the Scaffold of a Comprehensive System"; Murray, 1959).
While Explorations in Personality was praised in a representative mainstream review (Elliott, 1939) as "by far the most comprehensive attempt to bring Freudian psychology and experimental psychology into line with each other" (p. 453), the tenor of the review was decidedly mixed. On the one hand, Elliott noted that as a practicing psychoanalyst, Murray was "unique among those who hold a major position in psychology in a leading American university" (p. 453). He was impressed with the new tests and procedures (especially the Thematic Apperception Test) and with Murray's analysis of the variables of personality and environment. On the other hand, Elliott suggested that psychoanalysis was unproven and unscientific and criti-
cized Murray for overlooking existing work in experimental and differential psychology as well as for insufficient use of statistics. He considered the book's single case study too speculative.
According to Triplet (1983, 1992), Murray's work had a profound effect on the expansion of the disciplinary boundaries of personality psychology. By demonstrating the use of experimental techniques to investigate psychoanalytic concepts, by teaching courses on abnormal and dynamic psychology in a prestigious academic psychology department, and by inspiring a large number of graduate students, Murray played an important part in expanding the definition of personality psychology to include psychoanalytic theories that had earlier "had pariah status in academia" (Smith, 1990, p. 537). While he advocated the interpretive study of lives, his most enduring legacy was probably the catalog of "variables ofpersonality" (the title of Chapter 3 of Exploration^ see Danziger, 1997) and the Thematic Apperception Test (Morgan & Murray, 1935).
Murray's diagnostic council method was generally overlooked, however (McLeod, 1992). Although variants of the approach were used in later assessment settings such as the Office of Strategic Services (OSS), and the Institute for Personality Assessment and Research (IPAR) at the University of California (MacKinnon, 1975), as well as in organizations such as AT&T (Bray, 1982), the practical demands of assessment even in those well-financed settings led assessors to base decisions "on statistical averages of ratings, rather than on discussion" (McLeod, 1992, p. 10). Perhaps only when executives of the very highest levels (and salaries) were to be assessed, could the full range of person-focused techniques be financially justified.
Cartel] and the Measurement Imperative
The third major figure in the formative years of American personality psychology was Raymond Cattell. Cattell was born and educated in England (with an undergraduate degree in chemistry and a doctorate in psychology). He was greatly influenced by his assisrantships at the University of London with Charles Spearman,12 who first developed the technique of factor analysis, and later at Columbia University with E. L. Thorndike. Most of his professional career was spent at the University of Illinois, where he introduced many conceptual refinements and elaborated methodological developments, espe-
daily concerning the use of correlation, factor analysis, and other mulcivariate techniques, into the field of personality. Taking seriously Thorndike's famous dictum about measurement as well as the early work on intelligence testing, Cattell prefaced his first major book on personality with the assertion that "it is on measurement that all further scientific advance depends" (1946, p. iv). He went on to argue that "the ideal of a science of personality description [is] to build its traits upon a foundation of objective test measurements, as has been done to a very large extent in the analysis of abilities" (p. 210).
Kinds of Data
In reviewing the extant literature on personality measurement, Cattell first distinguished among different types of data (1946, pp. 10, 210-213; chaps. 9-11): ratings by skilled judges (R data), self-ratings (S data), and measurements on tests or experimental situations (T data). (Later he added life-outcome or L data to the list; see Cat-tell, 1957.) He believed that in the long run, T data were the soundest; but since for the foreseeable future objective tests could not measure everything of importance, he preferred ratings (1946, p. 294).
Kinds of Personality Traits
Like Allport, Cattell adopted trait as the fundamental conceptual unit of personality; but at the same time he also, in the spirit of Murray, distinguished motivational or "dynamic traits" (also called ergic traits or ergs) from stylistic or "temperament traits," as well as "ability traits." In his view, each kind of trait had its own pattern of correlational relationships among its component variables and the external situation: Thus, dynamic trait variables "change most in response to change of incentives" and showed complex higher-order correlations, while temperament trait variables "change least in response to any change in the field" (1946, p. 167). CattelTs major contribution to personality was his analysis of temperament traits via mathematical and statistical techniques.
The Search for Personality Coherence through Factor Analysis
Given the enormous array of data available to the personality psychologist—"too many variables, and too many influences behind these
variables" (1946, p. 272)—Cartell turned to correlation and orner multivariatc techniques for clarification. Believing that the essence of a trait was covariation or correlation, he concluded that the "most potent method of attacking the tangle is to work out correlation coefficients between the inconveniently multitudinous variables abounding in the subject and to seek some smaller number of 'behind the scenes' or underlying variables, known as factors" (p. 272).
Cattelis first major investigation of personality (1945) used peer ratings of 208 men on 35 trait-word clusters derived from the exhaustive Allport and Odbert (1936) list of 4,504 traits. After carrying out a factor analysis of these data, Cattell stuthed various rotations and finally settled on an oblique rotation that yielded 12 factors, which were also consistent widi cluster stuthes carried out by other investigators. (The rotations process, which could now be done on a personal computer in seconds, took 6 months!) He concluded that these 12 factors represented "the established primary traits" (1946, Chaps. 10-12).
Cattell thus introduced or established many conceptions and techniques that are an enduring part of contemporary personality psychology. His distinction between different kinds of data (especially the difference between R and S data, on the one hand, and T data, on the other) re-emerged in the 1970s dispute about the cross-situational and temporal consistency of personality (Block, 1977; Epstein, 1979). His formulation and defense of the "lexical hypothesis," that "language covers all aspects [of personality] that are important for other human beings" (1946, pp. 215-216),13 prefigured the adjective-based trait assessments widely used over the next several decades. Finally, his discussion and examples of rotated factor analyses set the stage for an entire approach to personality, though there would be considerable controversy about his particular methods and assumptions. For example, Eysenck preferred orthogonal rather than oblique rotations, and so he argued that Cattell's 12 oblique factors were really equivalent to his own three orthogonal "superfactors" of Extraversion, Neuroticism, and Psychoticism (Eysenck & Eysenck, 1985). Yet in spite of Cattells general influence, few personality psychologists adopted his full views about personality measurement as outlined in his "specification equation" for predicting behavior (Cattell, 1957, pp. 302-306).
Chapter 1. Hiitory of Modern Personality Theory and Research Psychometric Technology
We have seen that the prestige and apparent success of intelligence testing at the beginning of the 20th century convinced many personality psychologists that personality could (and should) be measured in a similar way, by scales of "items." While factor analysis was the apotheosis of this ideal, the same general conviction guided the construction of numerous other scales, inventories, and questionnaires not based on factor analysis: omnibus instruments such as the Bernreuter Personality Inventory, Minnesota Multiphasic Personality Inventory, California Personality Inventory (CPI; Gough, 1957; 1987), Adjective Check List, and Personality Research Form (PRF; Jackson, 1974), as well as countless scales designed to measure particular personality characteristics.
As statistical methods of scale construction and refinement became increasingly sophisticated, the psychometric rule book expanded to include matters such as test-retest reliability, internal and cross-situational consistency (Cron-bach, 1951), convergent-discriminant validity (Campbell & Fiske, 1959), correction for attenuation, and the distinction between "trait" and "state." Some psychologists protested that psychometric rules unduly constrained personality theorizing and research because they did not take account of nonlinear, interactive, or functionally substitutable (but not correlated) relationships among components of a concept (see Atkinson, 1982; Fleming, 1982; Winter et al., 1998). At the end of the century, the increased popularity of chaos theory and associated mathematical concepts (e.g., Vallacher & Nowak, 1997) suggested possible alternatives to the classical psychometric rules.
THE FOURFOLD FLOWERING OF PERSONALITY PSYCHOLOGY: TRAITS, MOTIVES, COGNITIONS, AND CONTEXT
By 1946, shortly after the end of World War II, the main concepts and issues of personality psychology were established. For almost all personality psychologists, traits were viewed as a major element of personality. Many would say traits were the only personality element (e.g., A. Buss, 1989); others, however, argued for the fundamental distinctiveness and importance of mo-
tives as well (see Winter et al., 1998). Rating scales and questionnaires were firmly established as the preferred method of personality measurement, especially for traits.
Our account of the next 50 years of personality psychology is framed in terms of the four elements or classes of theory and variables and theories introduced by Winter (1996): traits, motives, cognitions, and social context. The articles in the special issue of Ac Journal of Personality on "levels and domains in personality" (Em-mons & McAdams, 1995) offer a similar scheme for organizing personality variables.
In personality psychology, the concept of trait has been used to denote consistent intcrcorre-lated patterns of behavior, especially expressive or stylistic behavior (see Winter et al., 1998, pp. 232-233). Our discussion of traits will be organized around four rather disparate topics, because these reflect the main lines of theorizing and research about traits. Perhaps the most frequently stuthed topic has been the number, nature, and organization of "basic" traits. Personality psychologists have used three different strategies. In approximate order of popularity, they are: factor analysis and related mathematical techniques; rational or a priori theorizing, often involving the construction of typologies; and the idiographic approach, which essentially ignores the question of "basic" traits. After discussing developments in each of these three strategies, we turn to another frequently stuthed topic, namely the relation between traits and biology.