Abstract
Carroll Izard completed his dissertation in 1952, beginning a career spanning more than six decades that coincided with clinical psychology maturing as a profession, and the birth of clinical science and cognitive neuroscience. Izard’s focus on discrete emotions as evolved systems that organize information, prepare responses, and shape the development of personality and relationships persisted through his career, despite “emotions” often being overshadowed by psychodynamic, behavioral, or cognitive perspectives. His theoretical work anticipated and now integrates contemporary neuroscience and relational perspectives. Exploration of discrete emotions has kept lines of inquiry open that enrich our understanding of psychopathology. Izard also embraced clinical science, combining basic research with effectiveness studies addressing the unmet need for mental health services via emotion-focused interventions.
In 1952, the global community fumbled towards a new order in the aftermath of the Second World War. Hopes for peace were confounded by events in Korea, which embroiled the United States, the fledgling United Nations, Soviet Union, and mainland China in a lurching escalation that threatened to reignite still larger conflict. Technology, always spurred by war, brought major advances in computing and communication, but also capacity for destruction: 1952 saw the first explosion of a hydrogen bomb. Eisenhower became president. Elizabeth II started her historic reign. The American Psychiatric Association published its first effort at codifying and formalizing the definitions of psychopathology, the Diagnostic and Statistical Manual of Mental Disorders (DSM; American Psychiatric Association, 1952). And Carroll Izard defended his dissertation—an investigation of how patients with schizophrenia perceived emotional facial expressions—launching a career of curiosity and scholarship that has spanned more than six decades of continuous productivity.
The field of psychology was going through seismic shifts of its own as he joined it. Hundreds of thousands of soldiers returning with “combat fatigue” and “shell shock,” as well as millions of affected civilians, created a social need for better understanding of the causes and treatments for psychopathology that transformed the field of clinical psychology, especially in the United States. The influx of college graduates and graduate students, the birth of the Veterans Administration, the formulation of the Boulder and Vail models of clinical psychological training, the invention of the predoctoral clinical internship, and the more recent emergence of postdoctoral specialization as an advance component of research training… all of these defining events shaping the field of clinical psychology in the United States have occurred within the course of his career.
Though no mere observer of events, he also was not primarily interested in professional politics. He loves ideas, and he has always vigorously engaged in pursuing them. Alert to changes in technology and in priorities that would help his quest for knowledge, he engaged with the field’s trends while focusing with remarkable concentration on his own vision and interests. Undoubtedly a pioneer in the study of emotion, he has been more of an alchemist than a cowboy, intent on integrating different ideas and perspectives in pursuit of transformative knowledge rather than wandering newly opened frontiers and asserting his claims.
His contributions to our understanding of developmental psychopathology are remarkable. I will focus on four: how his work on emotion fits in the broader arc of conceptual models of clinical psychology, the developmental nature of his work on emotion, his work on discrete emotions compared to other models of emotion, and his embodied stance on the dynamic tension between research and practice.
Emotions as a Model of Understanding Psychopathology
When Carroll was completing his doctoral work, Freud and Jung were publishing actively. Psychodynamic models were ascendant in Europe, and dominated the first versions of the nascent DSM (American Psychiatric Association, 1952). Many departments focused on psychodynamic models in their training, and Exner (1999) was beginning to gather data for his system of scoring the Rorschach. Carroll had completed a master’s of divinity earlier, and psychodynamic models had a lively dialog with religion and with the arts. Carroll appreciated the value of a model that could connect basic psychological processes with the highest expressions of human creative achievement (e.g., Izard, 1972, 1985), but he also was drawn to the evolutionary theories of development, weaving in a skein of biology and of ethology. He read Darwin’s (1872/1965) work on emotional expression in animals, and began working with Sylvan Tomkins (1963) to understand affect and consciousness and cognition in humans. When behaviorism mounted a serious challenge to psychodynamic approaches for conceptualization of psychopathology (Skinner, 1971) and for the training of clinical psychology (Cummings, 2000), Carroll continued to focus on emotions, musing about their adaptive function. As the cognitive revolution overturned narrow behaviorism as the centerpiece of American clinical psychology (Beck, 1991; Ellis, 1973), Carroll persevered in elaborating models of emotion. Through these major sea changes in orientation, Carroll steadily, thoroughly explored emotion, regardless of the prevailing zeitgeist. He was convinced of a fundamental and unifying role for emotions in phylogenetic evolution, as well as in individual development (Izard, 1993; Izard, Youngstrom, Fine, Mostow, & Trentacosta, 2006).
The blossoming of affective neuroscience in the last two decades provides a valediction of his confident commitment to the core role of emotion. Recent developments with which the field is grappling, such as the National Institutes of Mental Health (NIMH) emphasis on research domain criteria as the organizing principle for new research instead of DSM diagnoses (Cuthbert & Insel, 2010), fit more naturally with his thinking. He has trusted his vision of where to find the most productive wellsprings of inquiry, even when research initiatives have invested heavily in other enthusiasms.
Starting with the human face as a point of entry, Izard and Ekman focused on distinct emotional expressions that could be reliably distinguished, generalizing across raters, across cultures (Ekman, 1989; Ekman & Friesen, 1971), and across ages (Izard, 1982), with Carroll creating coding systems that captured expressive information down to the first months of infancy. A good theory is a powerful simplifier, distilling complex phenomena into a few essential parameters without distorting the representation or omitting key facts. Phases of Carroll’s work explored the potential explanatory value of reductionist models of emotion: that certain emotions represented human universals that did not depend on culture or learning history to shape their formation (Izard, Hembree, Dougherty, & Spizzirri, 1983), that all emotions are tied to the face (Izard, 1971), that discrete emotions recognizable on the face represented affective “atoms” that were not further divisible (Izard, 1971).
To his credit, he always acknowledged when data did not assimilate neatly into a particular model, and he was always willing to make accommodations to integrate new findings. These syntheses generated some of his most significant writing, measured in terms of citation, such as his formulation of cognitive-affective structures as formative elements of cognitive schema and personality development (Izard, 1992), or the four systems of emotion activation (Izard, 1993) as a hierarchically arranged set of evolved mechanisms for eliciting emotion—with the cognitive appraisal system being the most recently evolved (Izard, 2007). These models created the framework for integrating cognitive and learning models of psychopathology with emotion, while also accommodating ethologically informed developmental models, such as attachment theory. As Beck’s initially cognitive models of depression and anxiety became more elaborate, including cognitive schema as overvalued, affectively charged core beliefs (Beck & Clark, 1988), Carroll’s models of affective–cognitive interactions (Izard, 1989; Izard, Krauthamer-Ewing, Woodburn, Finlon, & Rosen, 2009) bridged the span from the emotions side of the theoretical divide.
Carroll continued to read voraciously and digest the ideas, yet he was not an abstract theoretician. Eager to find new information, he bought or built equipment at the forefront of contemporary technology. Of course there were one way mirrors, microphones dangling from the ceiling, and video cameras (when I was a student, they were recording on VHS—there was a walk-in library full of shelf after shelf of carefully indexed and labeled tapes). More strikingly, Carroll was on a first name basis with the experts at the University of Delaware machine shop, and they would build him custom gear when there was not an “off the shelf” version available. Carroll was scrutinizing emotions before there was
Thinking Developmentally About Emotion
As Carroll constructed differential emotions theory, he built an intrinsically developmental focus. This did not mean simply enrolling children as research participants. When he was coding facial expressions in infants using MAX, he also was videotaping the mother’s facial expressions and coding them. He looked for significant behaviors in dyadic interactions and refined novel coding schemes to quantify the frequency, intensity, or duration of these exchanges. He continued to look for new ways of measuring emotions and then tweaking the assessment to make it developmentally and culturally appropriate. As he grew into thinking more developmentally about emotions and about psychopathology, he catalyzed the crystallization of developmental psychopathology as a discipline and formal perspective.
The Differential Emotions Scale (DES) is an excellent “case study” illustrating this cycle of development and extension in Carroll’s work. A component of the validation of the facial expression coding systems involved checking the association between a person’s expression and their subjective mood state. The DES grew out of the need for a consistent way of evaluating state emotion, with triplets of items that used a Likert-type scale focusing on each discrete facial expression. These emotion facets or subscales each had a name (and a two-letter code) corresponding to the facial code. Happiness and joy were two items on “EJ”—“Enjoyment/Joy.” The number of triplets expanded over time as Carroll probed emotions that might or might not be “discrete” or “universal” or have a distinct facial expression, such that revised versions of the DES included self-directed hostility—important for testing some models of the development of depression, and anticipating Gross’s later findings on “emotion suppression” (Gross & Levenson, 1993)—and contempt, which, regardless of its status vis-à-vis facial expression, has proved to be a distinctly powerful and toxic emotion in dyadic relationships (Gottman, 1994). By the time that the DES was formally published, it was on its fourth revision, with validity data for both state and trait versions (Izard, Libero, Putnam, & Haynes, 1993). But still it grew. As the lab started working with families enrolled in Head Start, funded by the William T. Grant Foundation to investigate associations between emotion knowledge, emotion regulation, and behavior, the DES morphed into a collateral informant version, where the caregiver reported about the typical emotion traits of her child (DES-V), and a simplified version for children to report about their own emotions (DES-VI)—often conducted as a structured interview to bypass issues of reading ability. My doctoral dissertation also included a simple, global rating of each emotion from the DES-IV on a visual analog scale, coding the overall level of expression of that emotion by the child during a 2-minute video clip of the dyad performing a standardized “cooperative construction” challenging task (Youngstrom, Ackerman, & Izard, 1999). The version that most of the research community uses (DES-IV) had been through multiple iterations before publication, and it expanded into a suite of related assessment strategies, including developmental modifications, cross-informant and observational methods—all with the goal of better characterizing the role of emotions in the development of the child and family (Youngstrom & Green, 2003). The creation of assessment strategies focused on discrete emotions instead of broad dimensions has helped reveal subtle but important gender differences in emotion expression (Chaplin & Aldao, 2013), among other areas.
The measurement of emotion also extended across methods. Carroll routinely used standardized performance tasks developed and validated in other areas of developmental psychology. These included tasks from Block and Block’s (1980) longitudinal work, as well as some of the first ambulatory pediatric measures of heart rate (kids wore backpacks in the lab carrying all the sensor paraphernalia while they did the tasks; Izard et al., 1991), and a variety of other innovations. Multiple projects had longitudinal components, and Carroll committed to pursue the goals of internal and external validity even when they sometimes pulled in opposite directions. The longitudinal work with families from Head Start embodied the quest for external validity, where the research literally moved from the lab into the community centers. Carroll bought portable TVs with built-in VCRs that the team could lug to each Head Start center for the family visit, so that we could play back the standardized video clips for the caregivers to rate. That investment in bringing internal rigor to a generalizable context improved understanding of the role of caregiver dysphoria in coloring descriptions of child emotions and behavior (Youngstrom et al., 1999).
Discrete Emotions and Psychopathology
Another major contribution has been continued emphasis on the role of discrete emotions, as opposed to major dimensions. Analyses of self-reported emotions tend to find two or three major dimensions, which many have labeled positive affect, negative affect, and dominance (Mehrabian, 1996; Russell & Mehrabian, 1977; Tellegen, Watson, & Clark, 1999). These have tremendous explanatory power in terms of relationships to personality and psychopathology (Watson et al., 1995), along with a well elaborated nomothetic network of relationships with psychophysiological measures (Cacioppo, Berntson, Klein, & Poehlmann, 1998; Izard et al., 1991; Levenson, Carstensen, & Gottman, 1994). There is also strong correspondence between these major dimensions and the organization of the research domain criteria, which include positive valenced, negative valenced, and social dominance as major categories in the matrix of constructs (Cuthbert & Insel, 2010).
Carroll persisted in testing whether discrete emotions showed any incremental value. He followed the development of dimensional approaches assiduously, but continued to examine discrete emotions measured by facial expressions, more global observational ratings, self-report, and collateral report. He found that these helped produce a more textured description of depression in childhood (Blumberg & Izard, 1985, 1986; Izard, 1972) at a time when the idea of depression in youth itself was iconoclastic (Kovacs, 1989). Thinking of sadness as a discrete emotion helps interpret its position on the affective circumplex, opposite happy and off the main axis of “negative affect,” consistent with tripartite models of low positive affect being the distinctive component of depression (Watson et al., 1995). Similarly, anger and fear show a variety of correlates, including cortical activation as well as major associations with externalizing versus internalizing behavior problems, that belie their juxtaposition in two-dimensional models of affect (Carver & Harmon-Jones, 2009). More differentiated emotion models afforded connections to peripheral and central nervous system psychophysiology (LeDoux & Phelps, 2000), and later to functional imaging (Vytal & Hamann, 2010).
The Practice of Science, and Science to Practice
One of the major creative tensions in clinical psychology has been between practice and basic science. The need for treatment outstrips the evidentiary basis for effective intervention, and science and practice often act as lodestones that polarize as well as magnetize. In the span of Carroll’s career, the Psychonomic Society (in 1959) and the Association for Psychological Science (in 1988) were major secessions of scientists from a practice dominated by the American Psychological Association. During my graduate training, the University of Delaware chose to become one of the charter members of the Academy for a Science of Clinical Psychology (McFall, 1991).
Where others perceived competing aims or irreconcilable values, Carroll avoided polemic and demonstrated the possibility of alloying science and practice into a stronger material. On the one hand, his scientific credentials could not be impugned, and his commitment to research was clear. On the other, the overarching aim of his work was to produce evidence-based tools that would alleviate or prevent problems. The body of work continued by alumni from the Human Emotions Laboratory not only strives to generate techniques for improving emotion knowledge, emotion regulation, and adaptive interpersonal skills, but it also moves quickly to effectiveness work—implementing these in Head Start and other real-world settings (e.g., Izard et al., 2008; Schultz, Izard, Ackerman, & Youngstrom, 2001; Slatcher & Trentacosta, 2011). Through a shift in perspective, the science–practice gap can be reframed as a false dichotomy. Carroll modeled a scientific approach to clinical issues. Some of his former students have become scientifically oriented clinicians, applying research findings in their work and in supervision and training of the next generations. Others are continuing the laboratory tradition of clinically relevant research. All are continuing to extend Carroll’s legacy of focusing on emotion as an organizing principle for clinical science and practice.