Robert W. Levenson

Introduction

Robert W. Levenson is a professor of psychology at the University of California, Berkeley, and the director of the Berkeley Psychophysiology Laboratory. He works on human emotion: what the body does during it, how that changes across the lifespan, and how it shapes and is shaped by the people we live with. He is a past president both of the Society for Psychophysiological Research and of the American Psychological Society, and a recipient of the latter's Mentor Award and William James Fellow Award.

Three lines of work carry his reputation. The first, conducted with Paul Ekman in the early 1980s, revived the claim that discrete emotions have distinguishable autonomic signatures. The second, conducted with John Gottman over several decades, used physiological measurement during couples' conversations to study why marriages succeed or fail. The third examines how emotion changes in normal ageing and in neurodegenerative disease, and it is responsible for much of what is known about the emotional deficits that accompany frontotemporal dementia.

Early Life and Education

Levenson received his bachelor's degree in psychology from Georgetown University. As an undergraduate he was drawn to the philosophy of the mind–body problem, and he has said that discovering psychophysiology — the existence of a research field devoted to exactly that problem — settled his direction.

He went on to Vanderbilt University for his master's and doctoral work in clinical psychology, finishing in 1974. At the time, psychophysiology was dominated by stress research: the assumption was that difficult situations activate the body in a broadly common way. Levenson's early experience in his own laboratory pushed him away from that view. He ran a study in which participants watched a film intended to be uniformly stressful and found that some looked afraid, some laughed, and some looked disgusted. What was interesting, he concluded, was not the shared stress response but the differences in how individuals responded — a conclusion that set the agenda for the rest of his career.

Academic and Professional Career

Levenson joined Indiana University in 1974 and stayed twelve years, moving to the University of California, Berkeley in 1986, where he has remained. At Berkeley he founded the Berkeley Psychophysiology Laboratory and served as director of the Institute of Personality and Social Research.

His research has been supported continuously by the National Institute on Aging, which has funded two long-running programmes: a longitudinal study of emotion in long-term first marriages, and a study of emotional functioning in neurodegenerative disease. The dementia work is conducted with neurologists at UCSF and Berkeley, and focuses on patients with focal neural loss in regions thought to be critical for emotion — a design that uses the specificity of the lesion to test claims generated from the study of healthy people.

He was a founding member of the American Psychological Society and later served as its president (APS Observer, 2013), and he has served as president of the Society for Psychophysiological Research. In 2013 he received the APS Mentor Award, given for lifetime achievement in research and education; in 2014 he received the APS William James Fellow Award. Vanderbilt named him its Psychological Sciences Distinguished Alumnus for 2009–2010. He has published more than 200 journal articles and has mentored several generations of emotion researchers, including a substantial share of the field's current senior figures.

Major Contributions to Psychology

Autonomic specificity of emotion

The question of whether discrete emotions have distinct physiological signatures is one of the oldest in the field, and by the 1970s the consensus was that they do not. In 1983, Ekman, Levenson, and Friesen published a paper in Science arguing that the earlier negative results were methodological artefacts: prior studies had failed to equate the intensity of the emotions being compared, and had not synchronised physiological recording with the actual onset of the emotional response.

Their design addressed both problems. Participants relived past emotional episodes and separately posed facial expressions muscle by muscle, while heart rate, finger temperature, skin resistance, and forearm muscle tension were recorded. The result was that heart rate increased substantially for anger, fear, and sadness but not for happiness or disgust; finger temperature rose most during anger; skin conductance rose most for the negative emotions. Different negative emotions could be distinguished from one another using combinations of measures. The paper reopened a debate that had been considered closed.

Emotion in marriage

Beginning in the early 1980s, Levenson collaborated with John Gottman on studies in which couples came to the laboratory and discussed an area of continuing disagreement while their physiology, facial expression, and speech were recorded. The work produced several findings that became standard. Married partners' physiology becomes linked during conflict, and the degree of linkage tracks relationship distress. Specific fleeting negative expressions — contempt in particular, along with defensiveness, criticism, and stonewalling — discriminate couples who stay together from those who separate. And the balance of positive to negative affect during conflict, rather than the intensity of the conflict itself, predicts the trajectory of marital satisfaction over years.

The programme eventually extended to a longitudinal study following long-term first marriages from middle age into old age, watching couples pass through the transitions that define later life — children leaving home, retirement, health decline — and asking which kinds of couples navigate them well.

Emotion and ageing

A recurring finding in Levenson's work is that emotional functioning is relatively spared by comparison with other domains. Memory and psychomotor performance decline measurably with age; emotional experience, regulation, and knowledge largely do not, and some aspects improve. In a widely cited 1991 study with Laura Carstensen, Paul Ekman, and Wallace Friesen, older adults reported emotional experiences as intense as those of younger adults while showing attenuated physiological reactivity — a dissociation between how emotion feels and what the body does that has proved fertile. His more recent work has asked where individual differences in that pattern come from, examining genetic polymorphisms and changes in cognitive ability as sources.

Emotion in neurodegenerative disease

The third programme studies patients with frontotemporal dementia, Alzheimer's disease, amyotrophic lateral sclerosis, and orbitofrontal brain injury. It has established that these conditions produce emotion deficits that are selective rather than global: patients with frontotemporal dementia show marked losses in embarrassment, empathy, and self-conscious emotion while other emotional capacities are comparatively preserved. Because the atrophy in frontotemporal dementia targets specific regions, the pattern provides evidence about which structures support which emotional functions.

Major Publications

Ekman, P., Levenson, R. W., & Friesen, W. V. (1983). Autonomic nervous system activity distinguishes among emotions. Science, 221(4616), 1208–1210.
Levenson, R. W., & Gottman, J. M. (1985). Physiological and affective predictors of change in relationship satisfaction. Journal of Personality and Social Psychology, 49(1), 85–94.
Levenson, R. W., Carstensen, L. L., Friesen, W. V., & Ekman, P. (1991). Emotion, physiology, and expression in old age. Journal of Personality and Social Psychology, 61(6), 987–996.
Levenson, R. W., Ekman, P., Heider, K., & Friesen, W. V. (1992). Emotion and autonomic nervous system activity in the Minangkabau of West Sumatra. Journal of Personality and Social Psychology, 62(1), 55–68.
Levenson, R. W. (2014). The autonomic nervous system and emotion. Emotion Review, 6(2), 100–112.

Impact and Legacy

Levenson's influence is easiest to see in the fact that questions once considered settled are now open. Autonomic specificity was written off by 1980 and is now a live research programme with its own methods, its own meta-analyses, and its own sceptics. Emotion and ageing was a marginal topic and is now a substantial field with a well-supported funding stream. The study of marriage moved from self-report questionnaires to multi-channel observation, largely because of the work he did with Gottman.

The criticisms are real and he has engaged with them directly.

The first concerns autonomic specificity itself. Meta-analyses by John Cacioppo and colleagues, published in two editions of the Handbook of Emotion, reviewed studies comparing at least two discrete emotions on at least two autonomic measures and concluded that the evidence is inconsistent. Of 37 autonomic measures examined, only a small set reliably differentiated discrete emotions, and the replicable findings were specific to particular comparisons — finger temperature decreases less in anger than in fear, but does not separate other emotions. The meta-analyses also found that the method used to induce an emotion has more reliable effects on physiological measures than the emotion itself does, which is the opposite of what the specificity hypothesis predicts. Multivariate approaches have improved the picture without settling it: pattern-classification studies report correct classification of emotional states from autonomic measures at roughly 45 percent in one case and 58 percent in another, which is better than chance but far from the discrete signatures the strong version of the thesis requires.

Levenson's 2014 reply is worth taking seriously, because it identifies genuine limitations in the literature rather than defending the finding. He argues that ethical constraints force researchers to use weak inductions — remembering an emotion or watching a clip rather than having one — and that weak inductions will inevitably produce weak differentiation; that most studies measure only heart rate and skin conductance because they are easy, ignoring gastrointestinal, pre-ejection period, and respiratory sinus arrhythmia measures; and that emotions are typically intense for only a few seconds, so averaging physiological response over an induction period destroys the signal. All three points are correct, and all three are also reasons the existing evidence cannot settle the question.

A second limitation applies to the ageing work. Much of it is cross-sectional: comparing people of different ages at one point in time confounds age with cohort, and the differences observed may reflect the historical circumstances in which people grew up rather than ageing itself. The longitudinal marriage study addresses this for its own sample, but that sample is narrow — predominantly white, middle-class, heterosexual couples in long-term first marriages — and the marital findings should not be read as claims about relationships generally.

A third concerns the causal claims in the marriage research. The finding that physiological arousal and specific negative behaviours predict relationship dissolution is robust within the design, but the design is observational. Showing that couples who divorce show a particular pattern does not establish that the pattern caused the divorce, and the intervention implications that have been drawn from the work by others sometimes outrun what the data support.

Finally, the cross-cultural extension of the specificity work, the 1992 Minangkabau study with Ekman, rests on a single non-Western sample and assumes that the emotion categories induced in West Sumatra map onto those induced in Berkeley. That assumption is the one the study was designed to test, and it cannot also be the premise.

Learn More

Related figures in this library: Paul Ekman, collaborator on the autonomic specificity and cross-cultural work.

References

Levenson, R. W. (1994). Human emotion: A functional view. In The Nature of Emotion. Oxford University Press.