In the 1930s, Jung's personality typology elicited much research focused on characterizing the multidimensionality of Extraversion and linking Extraversion with physiological processes and various perceptual and cognitive dimensions (see Carrigan, 1960; Guilford & Braley, 1930). During this period, many instruments were developed to measure both Extraversion and Introversion (see, e.g., Gilliland, 1970). For example, J.P. Guilford and R.B. Guilford (1934, 1936) used the then new technique of factor analysis to organize Extraversion-Introversion items into facets of sociability, impulsivity, and masculinity-femininity, among others. H.J. Eysenck's model of Extraversion flowed directly from the early work of the Guilfords. The Maudsley Personality Inventory (H.J. Eysenck, 1958) contained Eysenck's original Extraversion scale that was built from items earlier used by the Guilfords in their Guilford-Zimmerman Temperament Survey (1949) and featured aspects of both sociability and impulsivity. In the rest of this chapter, we concentrate on the various conceptualizations and operationalizations of Extraversion. Although there has been extensive theorizing about the psychobiological nature of Extraversion (see, e.g., H.J. Eysenck & M.W. Eysenck, 1985; Zuckerman, 1991), we focus only on phenotypic personality. First, we discuss several conceptualizations developed on the basis of studies of adult personality. Second, we discuss concepts from infancy and childhood literature with implications for the Extraversion construct. We conclude with a description of page_22 Page 23 Extraversion that is based on parental free descriptions of their children's personalities, descriptions obtained in seven different countries. Conceptions of Adult Extraversion In the 1940s and 1950s, the number of researchers investigating Extraversion (or closely related concepts) increased rapidly. It is beyond the scope of this chapter to give an exhaustive review of all these studies. Instead, we focus on three adult personality taxonomies. Each taxonomy is touted as completely covering the personality domain. Each differs, however, in the number of personality dimensions believed necessary to describe all personality traits. Cattell's Model Working separately from early theorists like Eysenck and Guilford, Cattell was one of the main proponents of the trait approach to personality. Cattell was convinced that the discovery of the major dimensions of personality embodied in the natural language would lead to a comprehensive model of personality structure. Beginning with Allport and Odbert's (1936) effort to compile a complete list of trait-related terms in the English language, Cattell applied various selection procedures that enabled him to use a pool of 35 clusters. These 35 clusters were assumed to give a comprehensive summary of the whole domain of personality differences. (For an extensive review of the selection procedures applied by Cattell, see John, Angleitner, & Ostendorf, 1988.) Using this 35-variable list, Cattell conducted a series of factorial studies that culminated in his structural model of personality (John et al., 1988). The structural model contained eight second-order factors and 16 primary factors. The first second-order factor (in Questionnaire data or Q data) mentioned by Cattell (1957) was labeled Extraversion versus Introversion, or as Cattell would probably prefer, Exvia versus Invia. Before his 1957 model, Cattell did not seem to be convinced of the need for a concept of extraversion-introversion as a unitary trait. Based on repeated experiments that showed a clear, second-order simple structure factor among the Q data, he commented: With this possibility of defining extraversion-introversion, as a unique second-order factor, instead of as a rough correlation cluster or a vaguely conceived mixture of the personality primaries surgency, cyclothymia, and parmia it is perhaps worthwhile to make a determined attempt to rescue the label "extravert-vs-introvert" from the scientific disrepute and uselesspage_23 Page 24 ness into which it has fallen through popular adoption. (Cattell, 1957, p. 267) Repeated factor analyses of Cattell's 16PF (Cattell, Eber, & Tatsuoka, 1980) have, for the most part, rather consistently
