The Earlier History of Biological Theories of Personality
Hans Eysenck, Jeffrey Gray, and Robert Cloninger
We are what we feel and perceive.
If we are angry, we are the anger.
If we are in love, we are the love.
If we look at a snowy mountain peak, we are the mountain.
While dreaming, we are the dream.
—Thich Nhat Hanh, Silence: The Power of Quiet in a World Full of Noise
FEW OF THE PERSONALITY THEORIES we have reviewed so far and compared to our affective neuroscience theory of personality have delved deeply into brain sources of personality variability. Some have offered speculations, but few have attempted to integrate neuroscientific thinking into their theories. In this chapter we briefly review the work of three eminent scientists who made serious efforts to not just name and describe personality traits but to consider the possible underlying neural substrates of human personality differences. We highlight the work of Hans Eysenck, who spent most of his career as professor of psychology at the Institute of Psychiatry, King’s College London, and his most eminent colleague and student, Jeffrey Gray, who added the first modern neuroscientific dimensions to emotion theorizing, and conclude with the seminal work of Robert Cloninger at Washington University in St. Louis, who focused on possible neurochemical substrates of human personality. As we share the revolutionary flavor of their work, we also contrast their seminal contributions to our affective neuroscience approach.
HANS EYSENCK
Hans Eysenck (1916–1997) was born in Germany during the First World War and emigrated to England after refusing to join the German military during the tragic events leading to the Second World War. He received his Ph.D. in psychology in 1940 from the University College London and, as fate would have it, began his career in a WWII British emergency hospital treating psychiatric casualties. He was an early advocate of diagnosing psychiatric patients using psychological dimensions rather than distinct categories based on symptom states—a battle he fought with the psychiatric powers on the European side of the Atlantic. So, it is not surprising that much of his career focused on delineating psychological characteristics, which he surmised were foundational for human personality.
His approach was similar to Raymond Cattell’s (see Chapter 12) in that he utilized factor analysis, which was increasingly popular at the time. However, he deviated from Cattell by shifting his emphasis from factor-analytically identified “traits” (Cattell’s source traits) to higher-order superfactors derived from the lower-order traits (Eysenck, 1990). Eysenck called these higher-order factors “types” and over his career identified three, Extraversion, Neuroticism, and Psychoticism, which had considerable impact on psychiatric theorizing.
Eysenck also differed from Cattell in his source of theoretical ideas. While Cattell used many Freudian concepts, Eysenck was influenced by Pavlov. In his 1957 book The Dynamics of Anxiety and Hysteria, Eysenck offered Pavlov’s idea of “excitation versus inhibition” in the nervous system as the basis of Introversion versus Extraversion, with introverts being characterized by more excitatory neural processes and more sensitive to sensory overload than extraverts. As the basis of his Neuroticism, Eysenck adopted Pavlov’s idea of nervous system “mobility,” which was the ability of the nervous system to give one impulse priority over another. In other words, Eysenck was the first to robustly link personality theorizing to the emerging neuroscientific revolution that embraced the systematic experimental study of behavior.
In 1967 Eysenck amended his theory, adding his third personality dimension, Psychoticism-Socialization, reporting that higher levels of testosterone were associated with lower levels of socialization and higher psychoticism. He also slightly adjusted his definitions of Extraversion and Neuroticism by suggesting that the ascending reticular activation system controlled the cortical arousal characteristic of Introversion-Extroversion differences, and that the emerging limbic system concept along with the sympathetic nervous system accounted for how human personalities were influenced by an underlying Neuroticism-Stability dimension. He also reported twin studies that further convinced him that biological causes played an important part in personality.
Eysenck may be most easily summarized as an early variant of the Big Five model discussed in Chapter 12. At his higher-order level, he accepted the Big Five Extraversion and Neuroticism dimensions. However, he argued that his Psychoticism dimension was a combination of the Big Five Agreeableness and Conscientiousness factors. He did not include the fifth Big Five factor, Openness to Experience, because he regarded it as a cognitive rather than a personality dimension. He later published the Eysenck Personality Questionnaire (Eysenck & Eysenck, 1975), which focused on empirically measuring his three psychiatrically focused personality dimensions: Extraversion, Neuroticism, and Psychoticism.
Although Eysenck’s reliance on factor analysis places his thinking in more of a top-down than a bottom-up category, he consistently emphasized the importance of psychological science linking his thinking to neuroscientific foundations of personality. In a paper debating how many basic human personality factors actually existed, Eysenck cited the “need for a nomological or theoretical network to accompany and be part of any model. Only thus can we avoid the problems of subjectivity and the possibilities of misinterpretations attaching to the conceptualization and naming of factors” (1992, p. 670). Further on he wrote, “We need to anchor our dimensions of personality in something more concrete than the morass of factor analysis, and biology supplies us with the necessary tools” (p. 672). Had he considered affective neuroscience perspectives and the dramatic emotional consequences of deep brain stimulation, he likely would have further modified his ideas about the biological bases of personality, a step that was taken by his most illustrious student.
JEFFREY GRAY
Jeffrey Alan Gray (1934–2004) was born in London. He did not begin his studies in psychology, receiving his first university degree in modern languages before beginning graduate studies. He was also trained in Russian during his British national military service and translated the work of Russian psychologist Boris Teplov into English for Eysenck while working as his graduate student.
In contrast to Eysenck, Gray began with a more bottom-up perspective. He completed his dissertation on emotional behavior in animals, and much of his life’s research dealt with diverse physiological influences on learning behavior in rats. He believed that the “reinforcement contingencies” encountered by animals created emotions (see Gray, 1990), a view that was countered by the suggestion that shifts in emotional and other affective feelings constituted the construct of “reinforcement” (Panksepp, 1990b). From his perspective, and building upon Eysenck’s seminal thoughts, he postulated that the brain has a basic Behavioral Activation System and a countervailing Behavioral Inhibition Systems.
Figure 14.1. Geometrical factor alignment of Gray’s behavioral activation system (BAS) and behavioral inhibition system (BIS).
While he built on Eysenck’s factorial studies of neuroticism and extraversion, Gray argued that a geometric shift in Eysenck’s model was needed to properly account for the two major personality dimensions Gray proposed: a behavioral inhibition system (BIS) that gave rise to anxiety and a behavioral approach system (BAS) that engendered impulsivity (see Figure 14.1). Gray contended the BIS and BAS constructs did not line up directly with Eysenck’s extraversion and neuroticism dimensions. Gray’s anxiety dimension fell between high neuroticism and low extraversion but inclined more toward neuroticism. Likewise, impulsivity aligned between high extraversion and high neuroticism, inclining more toward extraversion, as illustrated in Figure 14.1 (Gray, 1970; Corr, Pickering, & Gray, 1997). Gray argued that extraversion and neuroticism were derived from impulsivity and anxiety, which were their fundamental neurophysiological bases. In 1994, Carver and White gave a boost to research into Gray’s theory by constructing personality scales to measure his hypothesized BAS and BIS systems.
In laying out his theory, Gray also deviated from classical learning theory by asserting that there was not just a single reinforcement mechanism but two: BAS reflected sensitivity to rewards, and BIS, sensitivity to punishments. However, from an affective neuroscience perspective there are many distinct reward and punishment systems: positively and negatively valenced affects manifest by the primary emotions that are closely related to personality, but there are also the homeostatic affects such as HUNGER that have been so widely utilized in rat learning research. Indeed, these powerful brain-based emotional command systems can also be thought of as evolved learning systems that facilitate ongoing adaptation to our specific environments. In other words, the shifting tides of many positive and negative affective systems constituted what behaviorists had become accustomed to calling “reinforcement” processes in the brain (Panksepp, 1990b), and to understand the affective foundations of personality one has to focus on more affective systems than just the BAS and BIS.
From Gray’s learning theory perspective, the reward and punishment systems were each unidimensional. Yet, each of seven primary emotional-adaptation systems of affective neuroscience may possess different conditioning parameters in the brain, with additional, distinct homeostatic affects (e.g., HUNGER or THIRST, etc.) and sensory affects (various pains and pleasures). However, only the emotional affects are major contributors to personality development, even as they share many of the general brain chemistries for learning, such as glutamate and gamma-aminobutyric acid (GABA). For example, the SEEKING system is more of a general anticipatory reward system, whereas the CARE system may be more of a maternal-social reward system. The PLAY system may be a social emotional system that is especially robust during childhood, with some neurochemistries that are distinct from CARE and LUST. In addition, the PANIC/Sadness system may have acquired addictive-like qualities (being very strongly modulated by brain opioids), which would make its extinction patterns different from those of the FEAR system. As each of these primary emotions are thought to have evolved for distinct survival issues, serving different purposes at different times in our ancestral past, it is likely their learning parameters were evolutionarily adjusted to meet unique survival needs as well. Moreover, the development of the Affective Neuroscience Personality Scales was motivated by the need for personality psychologists to measure each of the distinct primary emotions that are most closely related to personality and psychopathology and to provide an assessment of their relative strengths and weaknesses in an individual’s life. Although both Eysenck and Gray emphasized the need for personality studies to focus on brain issues, a sufficient analysis of core mammalian affects was not evident in either Eysenck’s or Gray’s theorizing.
In 2000, Gray and McNaughton published a revision of Gray’s theory. In the revised reward sensitivity theory, the BAS remained largely unchanged. However, the BIS assumed responsibility for goal conflict resolution and became a “risk assessment” system: “the ‘increased attention’ output of the behavioural inhibition system” (p. 20). While the BIS once accounted for all punishment, in the revised theory it retained only its anxiety function, which was distinct from fear. Yet, BIS anxiety also included “separation anxiety” (p. 91), which was considered a childhood phenomenon and was not given any special status as in the affective neuroscience PANIC/Sadness system. Specifically, in the modified BIS, anxiety was generated by neural processing not only in the septal-hippocampal circuitry but also in the cingulate and prefrontal cortices. A proposed new brain system, the fight-flight-fear system, dealt with all defensive behavior, including fear. So, one of the more controversial positions taken by Gray and McNaughton (2000) was that fear and anxiety represented different brain systems and what we would call PANIC/Sadness was included with BIS anxiety. From our perspective, the natural emotional systems of mammalian brains were not adequately integrated into the revisions of the BAS and BIS.
Joseph LeDoux (1994) has shown that the amygdala is clearly linked to both fear and anxiety. Further, while the hippocampus provides cognitive-spatial information related to fear conditioning, it is not needed for the threat-evoked affective experiences of anxiety or fear. Others have also found that deep brain stimulation in the periaqueductal gray can generate alertness (corresponding to risk assessment and increased anxious attention), freezing, or fleeing in animal models, depending on the strength of the electrical current (Vianna, Landeira-Fernandez, & Brandão, 2001). This also applies to humans. For instance, Dean Mobbs and colleagues published a brain imaging study in humans that simulated a predator attack (Mobbs et al., 2007). They described a single “forebrain-midbrain network, which includes the vmPFC (ventromedial prefrontal cortex) at the lowest level of threat and interacts with the midbrain PAG (periaqueductal gray) as the threat level increases” (p. 1082)—meaning as predator threat is imminent, as opposed to far away, and dread increases (discussed in more detail in Chapter 16). All this is consistent with Panksepp’s observations that anticipatory anxiety and raw fear are all part of the same ancient FEAR circuit. Thus, in contrast to Gray’s theory, it seems likely that “anxiety” and “risk assessment” in potentially dangerous situations are tapping into the same emotional FEAR circuit that is activated when survival is imminently in jeopardy.
Perhaps when one has worked with deep brain stimulation à la Walter Hess and Jaak Panksepp, it is easier to conceptualize the foundational emotional systems embedded in the subcortical brain than when starting from a learning theory perspective. Furthermore, why not study the behavioral, biological, and psychological mechanisms of each affective brain system separately, rather than lumping all “rewards” together into a single BAS and all “punishments” together into a single BIS or FFFS? That would better allow us to consider specific and general “sensitivities” in each emotional or homeostatic affect system and further refine our definition of mammalian affective space—our genetically endowed sensitivities as well as life-span changes in sensitivities derived from addictive or traumatic experiences.
CLONINGER’S CONTRIBUTIONS: BRIDGING TO NEUROSCIENTIFIC FOUNDATIONS OF PERSONALITY
C. Robert Cloninger (born 1944) is an American psychiatrist and geneticist. He attended medical school at Washington University in St. Louis and has remained with that institution throughout his career. His personality interests and theories may fit more in the clinical tradition of personality theory. With his unified biosocial theory of personality (Cloninger, 1986), Cloninger connected clinical approaches focusing on abnormal human behaviors to the temperamental characteristics found in all humans and to emerging potential linkages to animal research. In other words, like McDougall and Eysenck before him, Cloninger advanced the view that “abnormal” disorders represented extreme cases of the same personality dimensions evident in normal people at more moderate levels (Cloninger, 1987). Thus, in line with Jeffrey Gray’s work, Cloninger promoted a view that could be linked to emerging neuropsychological work in animals.
Cloninger’s biosocial theory of personality hypothesized that there were three genetically inherited personality traits, which he would later call temperaments: Novelty Seeking, Harm Avoidance, and Reward Dependence, which could be measured with his Tridimensional Personality Questionnaire (TPQ; Cloninger, 1987). With the release of the TPQ, Cloninger included a discussion of how eight clinically diagnosed personality disorders from the DSM-III (American Psychiatric Association, 1980) could be described by his three biosocial personality scales. That is, there were eight possible combinations of high or low scores on his three TPQ scales, and he illustrated how combinations of the three biosocial personality traits potentially defined the clinical personality disorders. For example, the Antisocial personality disorder was characterized by high Novelty Seeking, low Harm Avoidance, and low Reward Dependence scores whereas the Histrionic personality disorder was described by high Novelty Seeking, low Harm Avoidance, but high Reward Dependence scores. This underscored Cloninger’s belief that the underlying structure of normal personality traits was the same as that for clinically defined behavior disorders. In other words, maladaptive behavior could be placed on the same continua as behavior describing the general population.
Cloninger’s Brain Systems
Around the same time, W. John Livesley (1991) was also working on a common model to bridge the divide between the clinical and normal personality worlds (for a fuller treatment, see Chapter 18), but Cloninger went further by hypothesizing that each of his three biosocial personality dimensions reflected variations in different brain systems, which were largely based on the major monoamine neural modulators (Cloninger, 1987) dopamine, serotonin, and norepinephrine (collectively called “biogenic amines”) that were receiving massive neuroscience and behavioral/psychological experimental attention in the 1970s and 1980s. Figure 14.2 depicts the chemical structures of dopamine, norepinephrine, and serotonin, all of which are similar small molecules. All monoamines contain one amine group, which for dopamine and norepinephrine is connected to an aromatic benzene ring; serotonin has a more complex “indole” ring structure. We highlight his monoamine theory for historical reasons as current neuroscientific thinking has moved toward more complex/sophisticated views in the last three decades.
Cloninger argued that Novelty Seeking was primarily associated with behavior-activating dopaminergic neurons, with dopaminergic neural projections from the midbrain ventral tegmental area and substantia nigra to the forebrain. The major neural modulator in the Harm Avoidance system was postulated to be serotonin. That system had its roots in the behavior-inhibiting serotonin neurons largely located in raphe nuclei, with projections to the limbic system and the prefrontal cortex. However, Cloninger also noted that benzodiazepines, with actions similar to alcohol, block such learned behavioral inhibition via gamma aminobutyric acid GABA inhibition of serotonergic neurons. Norepinephrine was the major neuromodulator that Cloninger linked with his Reward Dependence system, with major projections from midbrain locus coeruleus to much of the brain, including the neocortex. Reward Dependence functioned as a behavioral maintenance system, with norepinephrine increasing the “signal-to-noise ratio” or salience of relevant stimuli, helping them stand out from those that are less relevant (Cloninger, 1987, pp. 576–577).
Figure 14.2. Molecular structures of the catecholamines, dopamine and norepinephrine, and the indoleamine, serotonin.
Novelty Seeking
Novelty Seeking represented the dopamine-promoted “heritable tendency toward intense exhilaration or excitement in response to novel stimuli or cues for potential rewards or potential relief of punishment, which leads to frequent exploratory activity in pursuit of potential rewards as well as active avoidance of monotony and potential punishment” (Cloninger, 1987, p. 574). Thus, Cloninger envisioned people characterized by high Novelty Seeking with words like impulsive, exploratory, fickle, excitable, quick-tempered, extravagant, and disorderly, versus those low in Novelty Seeking being described as reflective, rigid, loyal, stoic, slow-tempered, frugal, orderly, and persistent (p. 575). Compared to the Affective Neuroscience Personality Scales (ANPS), Novelty Seeking seems most similar to the SEEKING scale.
Harm Avoidance
Cloninger envisioned Harm Avoidance as the serotonin-promoted “heritable tendency to respond intensely to signals of aversive stimuli, thereby learning to inhibit behavior to avoid punishment, novelty, and frustrative nonreward” (Cloninger, 1987, p. 575). Those characterized by high Harm Avoidance were cautious, tense, apprehensive, fearful, inhibited, shy, easily fatigable, and apprehensive worriers, with low Harm Avoidance types being more confident, relaxed, optimistic, carefree, uninhibited, outgoing, and energetic (p. 576). The Harm Avoidance dimension has many similarities to the ANPS FEAR scale.
Reward Dependence
Cloninger’s third biosocial dimension, norepinephrine-mediated Reward Dependence, was the “heritable tendency to respond intensely to signals of reward (particularly verbal signals of social approval, sentiment, and succor), and to maintain or resist extinction of behavior that has previously been associated with rewards or relief from punishment” (Cloninger, 1987, p. 575). He portrayed people who showed high Reward Dependence as “eager to help and please others, persistent, industrious, warmly sympathetic, sentimental, and sensitive to social cues and personal succor but able to delay gratification with the expectation of eventually being rewarded. In contrast, those who are lower than average in Reward Dependence are socially detached, emotionally cool, practical, tough-minded, and emotionally independent in what they choose to do” (576–577). Given a consistent association between the Big Five Extraversion and Reward Dependence (De Fruyt, Van De Wiele, & Van Heeringen, 2000; Zuckerman & Cloninger, 1996), it might be positively correlated with the ANPS PLAY scale. Overall, however, the vast diversity and complexity of brain neurochemical systems (e.g., Panksepp, 2011a) made such “aminergic” simplifications of personality open to debate even at the time they were proposed. Brain biogenic amine systems are surely critical brain arousal foundations of personality, but many other neurochemistries need to be considered in the present era, as well as various learning processes, as Cloninger well recognized.
Neuromodulator Commentary
While Cloninger argued for the dynamic interaction of the three temperaments, his attempts to link each to a major neuromodular—dopamine, serotonin, and norepinephrine—was probably painting with too broad a brush. For example, Ritalin (methylphenidate) as a dopamine stimulant should have a behavioral activation effect. However, counter-intuitively, it is widely prescribed to treat hyperactive children and those diagnosed with attention-deficit hyperactive disorder making them less impulsive, with many children reporting feeling less playful and having less fun when taking Ritalin. Panksepp et al. (1984) have shown that, while moderate levels of dopamine must be present for play to occur, high levels of dopamine decrease mammalian play, illustrating the complex role dopamine can have on behavior.
Similarly, serotonin-enhancing drugs are often used as antidepressants. Many selective serotonin reuptake inhibitors (SSRIs), such as Prozac (fluoxetine), Paxil (paroxetine), and Zoloft (sertraline), are used to treat depression and act by retaining serotonin in the neuronal synapse, where it can remain active, which is consistent with Cloninger’s predictions. However, the linkage of serotonin levels to Harm Avoidance has received mixed results (Paris, 2005). Further, recent reports suggest that SSRIs are only marginally effective at treating depression (Pigott, Leventhal, Alter, & Boren, 2010). In Chapter 18, we will show how opioids and other pharmaceutical treatments unrelated to the broad neuromodulator serotonin offer the promise of more effective, affectively targeted treatments for depression.
Cloninger Expands His Theory
In good novelty seeking fashion, in 1993 Cloninger, Svrakic, and Przybeck explored new personality dimensions to improve Cloninger’s biosocial theory of personality and increased his eighty-item TPQ to include many more items, scales, and scale facets. Cloninger expanded his personality model beyond the biosocial neurotransmitter-based brain systems to a new psychobiological model that included a total of seven personality dimensions. With his new Temperament and Character Inventory (TCI), he also added a fourth “temperament” dimension called Persistence, although he did not provide pharmacological or other biological mechanisms for it. In addition to the temperament traits, he also introduced the concept of “character” dimensions, which he hypothesized to be (1) less genetically heritable, (2) maturing in adulthood rather than in childhood, and (3) influencing personal effectiveness through learning around self-concepts. He called his three new character dimensions Self-Directedness, Cooperativeness, and Self-Transcendence, which were thought of as being more cortical, with his original temperament dimensions being more emotional, controlled by subcortical neural systems. In other words, with the character dimensions, his personality system became considerably more cognitive.
Cloninger conceived of the three character dimensions as corresponding to one of three self-concepts, or “the extent to which a person identifies the self as (1) an autonomous individual, (2) an integral part of humanity or society, or (3) an integral part of the unity of all things” (Cloninger et al., 1993, p. 978). The Self-Directedness scale corresponded to the autonomous individual level. Cooperativeness reflected identifying with humanity. Self-transcendence was related to identifying with the universe as an interdependent whole.
Cloninger also attempted to differentiate between temperament and character dimensions based on the learning and memory systems involved. In addition to expressing inherited behavior biases, temperament personality dimensions relied on presemantic perceptual learning that basically reflected associative learning or classical and operant conditioning. This elementary level of learning required no conscious awareness and was closely related to the concepts of procedural or implicit memory (Bachevalier, 1990).
The character personality dimensions theoretically reflected conscious experiences and learning abstract concepts. The character dimensions drew on retrievable memories of facts and events. This level of learning is also referred to as declarative or explicit memory (Smith & Squire, 2005; Squire & Zola-Morgan, 1991).
In any case, the temperament scales became Cloninger’s elementary components of personality. The new character dimensions added higher-level humanistic concepts to Cloninger’s personality model. Beyond more concrete social bonding and nurturing that would reflect more of the primary PANIC/Sadness and CARE systems, he introduced elements to describe how we relate to our world at an abstract, conceptual level, which would seem to be positioned at a more tertiary level of BrainMind organization than the basic inherited foundations of personality. What was missing in Cloninger’s early views was a clear vision of the inherited emotional proclivities in the development of personality, and the recognition that the biogenic amines modulated all of the primal emotions that the ANPS seeks to illuminate. In a sense, Cloninger’s expansive additions, attempting to deal with the fuller complexities of human life (in a sense becoming a “theory of everything”) diluted his initial vision of the fundamental instinctual-neural foundation of personality.
A CRITIQUE OF CLONINGER’S EXPANSIVE THEORY OF PSYCHOLOGY
When one probes the origins of Cloninger’s revised and expanded personality test, the TCI, one wonders how far Cloninger has truly extended the science of personality beyond other existing personality assessments and the factor-analytic methods he has criticized. Cloninger openly noted that the temperament scales from his original personality test, the TPQ, “were largely uncorrelated with some measures of social cooperation, such as the Agreeableness scale of the Neuroticism-Extraversion-Openness personality inventory, the Aggression scale of the Multidimensional Personality Questionnaire (MPQ), and the Hostility scale of the Profile of Mood States” (Cloninger et al., 1993, pp. 978–979). These scales were largely derived from factor-analytic analyses and must be seen as the psychometric basis for his character dimension called Cooperativeness.
He also pointed out that individual self-acceptance and self-esteem, key elements of the Self-Directedness character dimension, “were not well explained by the TPQ” (Cloninger et al., 1993, p. 979). He cited the Alienation scale of the MPQ and the Repression factor of the Minnesota Multiphasic Psychological Inventory as scales predicting positive self-regard and the acceptance of individual responsibility, including the capacity to admit to unflattering statements about oneself. These scales form the foundation of Cloninger’s Self-Directedness character dimension, along with Julian Rotter’s Locus of Control scale, which Cloninger cited in his discussion of Self-Directedness (Cloninger et al., 1993).
The story is the same with Auke Tellegen’s MPQ Absorption scale, which Cloninger reported was not correlated with any of his four temperament traits (Cloninger et al., 1993). In a description of Self-Transcendence, Cloninger described the “absorption” that leads to “identification with things outside the individual self” (Cloninger et al., 1993, p. 982). Again, the MPQ Absorption scale may largely form the basis of Cloninger’s Self-Transcendence scale.
Cloninger was able to report a factor analysis of the twenty-five facets that made up the seven scales of the TCI, in which the facet loadings were mostly consistent with how he “rationally” defined the facets (Cloninger et al., 1993, p. 986). However, this is not really surprising, because the material for the TCI character dimensions seems to have been drawn from sources that derived their scales using linear factor-analytic techniques, such as Tellegen’s MPQ and Costa and McCrae’s NEO.
CLONINGER ON FACTOR ANALYSIS
Cloninger has argued strongly that factor analysis is not capable of deriving adequate personality scales. He sees factor analysis as a linear statistical procedure that limits attempts to describe the human personality, which is greatly influenced by dynamic, nonlinear factors (Cloninger, 2004). Consequently, he prefers to rationally define personality scales based upon his research into the underlying causal structure of personality. He writes, “Factor analysis of individual case descriptions can only determine the minimum number of measured dimensions, and cannot decompose their underlying causal structure. Extrastatistical information is needed to specify the structure of the underlying biologic and social variability in personality traits. In other words, descriptive data about individual behavior are not sufficient to permit any strong preference among alternative ways of summarizing personality traits” (Cloninger et al., 1993, p. 976, italics in original).
We are in full agreement with the idea that factor analysis by itself is not adequate to specify the biological systems inherent in the BrainMind and manifested as personality. Panksepp did not derive his blue ribbon affective neuroscience emotions from a statistical analysis of descriptive data. Each of his foundational subcortical emotions reported in Affective Neuroscience (Panksepp, 1998a) was solidly supported with experimental brain research, such as deep brain stimulation, in which a simple electrical stimulus to a particular region of the subcortical brain produced clear, unambiguous emotional action patterns that were accompanied by diverse affective feelings, which may be foundational for the developmental complexity of the higher-order MindBrain organizations that Cloninger moved on to, without ever conceptualizing the fundamental emotional infrastructure of higher mental tendencies.
CRITIQUE BY FARMER AND GOLDBERG
Unfortunately, in the current personality research climate, psychometric criticism is likely to emerge for assessments that do not meet the requirements of the factor-analytic community. And, it is along these lines that Richard Farmer and Lewis Goldberg, two Oregon personality theorists, carefully analyzed Cloninger’s TCI using their own large independent data set of TCI data (Farmer & Goldberg, 2008b). In their factor analysis of TCI facets, they were able to confirm much of the same TCI structure that Cloninger had reported earlier (Cloninger et al., 1993, p. 986), but they were able to derive only six interpretable factors. As with Cloninger’s data, in their statistical analysis not all facets fell on the predicted factors. Unlike Cloninger, their Harm Avoidance and Self-Directedness facets combined to form a single factor. They were largely able to verify the Persistence, Novelty Seeking, Self-Transcendence factors and all but one of the Cooperativeness facets loaded together on the same factor. Like Cloninger, Farmer and Goldberg found the Reward Dependence facets did not load together.
When analyzing the TCI items rather than facet scores, Farmer and Goldberg were again unable to interpret seven factors or to separate Harm Avoidance and Self-Directedness into separate factors. Also, Cloninger’s Novelty Seeking items split up and did not form a clear factor. However, most of the Reward Dependence items sorted onto the same factor, with similar clarity for the Cooperativeness, Persistence, and Self-Transcendence items.
Perhaps the biggest problem in the Farmer and Goldberg analyses was that Harm Avoidance and Self-Directedness loaded together on the same factor whether analyzing TCI facets or items. In other words, Farmer and Goldberg could not distinguish between Harm Avoidance as a temperament trait and Self-Directedness as a character trait. Statistically, these two Cloninger dimensions looked like they were measuring the same thing and that Farmer and Goldberg, at best, could only account for six personality dimensions when examining Cloninger’s TCI.
While a Big Five factor analysis also lumped together all three of the ANPS negative emotions, as we described in Chapter 2, Cloninger had claimed that Self-Directedness and the three character measures did not measure emotional traits and were more cortically dependent. Thus, Farmer and Goldberg’s finding that Harm Avoidance and Self-Directedness formed a single factor suggested that temperament and character dimensions shared a common underlying basis. Accordingly, perhaps there weren’t distinct temperament and character domains in the human personality. In our view, it is likely that both emerge from the way basic emotional strengths and weaknesses interact with developmental and life time learning experiences.
Farmer and Goldberg (2008a) further pointed out that Cloninger’s claim of serotonin being specifically related to Harm Avoidance and not to character dimensions has often not been supported by outside research. Both Harm Avoidance and Self-Directedness were found to be strongly related to serotonin activity (Peirson et al., 1999), and Self-Directedness and Cooperativeness but not Harm Avoidance were strongly associated with a genetic polymorphism that modulates serotonin (Hamer, Greenberg, Sabol, & Murphy, 1999). Similarly, cognitive-behavior therapy outcomes (Dalle et al., 2007) and recovery from drug dependencies (Borman et al., 2006) have been associated with changes in both temperament and character dimensions. These findings contradicted Cloninger’s prediction that pharmacological treatments would modify temperament traits with psychotherapy influencing character traits and specifically that “cognitive-behavioral therapy techniques may facilitate learning self-directed behavior” (Cloninger et al., 1993, p. 988).
Indeed, more recently Cloninger himself has softened his claim that, in contrast to the temperament dimensions, character dimensions were not genetically determined and only influenced by environment. Cloninger has conceded that data from twin studies show that “Each of the character dimensions shows moderate heritability and associations with multiple candidate genes . . . and little or no effects of environmental influences shared by siblings reared together” (2004, p. 46).8 Cloninger later elaborated: “To my surprise, we found that the heritability of character was as great as that of temperament. This important finding suggests that the popular dichotomy in science between the neurobiological and psychosocial paradigms of human personality is not well justified” (Cloninger, 2008, p. 297). Thus, another of the three hypotheses Cloninger initially claimed distinguished his character dimensions from his temperament dimensions was not supported.
Cloninger has also backed away from his claim that character dimensions would mature in adulthood rather than childhood. He conceded that “the amounts of increase in character with age are small on average and negligible after middle age” (2004, p. 47). With that finding, all three of the pillars Cloninger hypothesized would separate character traits from temperament traits have failed to be confirmed.
CONCLUSIONS ABOUT THE CLONINGER PSYCHOBIOLOGICAL THEORY
There is substantial evidence that Cloninger’s temperament and character dimensions do not represent distinct personality domains. Perhaps most telling from an affective neuroscience perspective is that both sets of dimensions exhibit significant genetic heritabilities and seem to correlate with emotions. Because we believe that there are no inherited functional “modules” in the neocortex (e.g., it is well established that even cortical vision is learned) and that all known personality dimensions have a basis in subcortical brain affective systems (or the regulation of subcortical brain emotion systems), Cloninger’s lack of evidence for additional neocortical personality characteristics beyond conscientiousness strongly supports the affective neuroscience theory of personality. Although we should not exclude the possibility of the future identification of novel neocortical personality traits in addition to those regulating subcortical primary emotions, multivariate genetic analysis suggests that Cloninger’s seven personality dimensions were most likely linked to the affective dynamics of the various evolved subcortical brain emotional systems (Ando et al., 2004).
With the expansion of his theory and the TCI psychological assessment in 1993, Cloninger veered significantly toward a more purely philosophical direction, advancing claims that were difficult to substantiate empirically. In his 2004 book Feeling Good: The Science of Well-Being, Cloninger presents ideas that end up being more humanitarian-positive psychology, with insufficient substantive brain science. His prescriptions become more sensible guides for living than evidence-based psychobiological inferences.
To his credit, Cloninger has assembled a fine personality instrument. However, while it may be one of the better instruments for diagnosing personality disorders (Farmer & Goldberg, 2008b), it is by no means the only good instrument available in this arena. One might have wished that Cloninger had attempted to refine the bottom-up neurobiological aspects of his original theory. Instead, he attempted to expand toward a more top-down approach to understanding human well-being.
Altogether, Cloninger is a creative scholar who has enlivened the field with novel perspectives, reflecting the consideration of an incredible range of information, including recent participation in a far-reaching genetic analysis of schizophrenia (Arnedo et al., 2015). However, he has yet to provide a biological theory of personality that adequately accounts for the fundamental affective infrastructure of mammalian brains.
At this point, Darwin’s famous theoretical dictum about mammalian mental differences being “of degree not of kind” seems to be holding true in the world of personality theory. Along these lines, it is interesting that Cloninger has not focused on the Big Five Conscientiousness dimension as a possible cortically based character trait that may acquire the capacity to regulate subcortical primary emotions. Among the great affective neuroscience puzzles that needs more experimental attention is the developmental construction of this Conscientiousness regulatory trait, which, we assume, appeared on the neuroevolutionary scene long before the emergence of Homo sapiens.
The cortex, with its massive capacity to learn, provides vast diversities of mental life that are truly unique to each individual. Yet, such differences emerge from the shared emotional and other affective features of our personalities, based robustly on survival systems that evolved long before humans walked the face of the earth. That the intrinsic genetic and developmentally guided epigenetic strengths and weakness of these systems are critical for the emergence of our higher-order characterological traits seems more likely than ever, which leads us to the subject of the next chapter.