123 | Lisa Feldman Barrett on Emotions, Actions, and the Brain
What this covers
Sean Carroll and Lisa Feldman Barrett examine the architecture and function of emotion at the level of neuroscience, psychology, and culture. Barrett, a leading researcher in affective science, argues that emotions are not triggered by hardwired circuits but constructed by the brain from bodily sensations and learned concepts. The conversation works through the evidence for this view—how neuroanatomy has been misread, how the predictive brain works to regulate the body, and why facial expressions and emotional responses are far more variable across contexts and cultures than popular wisdom suggests. The core claim is that the brain's primary job is not reasoning or perceiving but regulating the body through allostasis, and emotions serve that regulatory function as situated prescriptions for action rather than universal reactions.
The discussion spans several layers of analysis. Barrett dismisses the "triune brain" myth—the idea that human brains evolved in distinct layers, an ancient reptilian core beneath emotional and rational layers—showing instead that larger brains reuse the same neuron types rather than adding new ones. She draws distinctions between affect, the brain's simple barometer of bodily state, and emotion, which is constructed from affect plus learned concepts applied to sensory input. A notable case emerges around emotion-recognition technology: software can detect a scowl but cannot determine its meaning, because facial movements have low reliability as emotional indicators—people scowl when concentrating, confused, or amused, not just angry. This gap between observation and interpretation matters when such systems inform hiring, school admission, or criminal sentencing. Barrett also sketches how social reality works—how collective agreement imposes function and meaning on physical things—and connects this to how shared emotional meanings reduce metabolic burden by making people more predictable to each other. Throughout, the framework integrates molecular biology and neuroscience with anthropology and sociology, dissolving the old divide between evolutionary and social-constructivist accounts of mind.
Barrett argues emotions are not hardwired circuits but constructions the brain assembles from bodily affect and learned concepts to predict and guide action; the brain's core job is predictive regulation of the body (allostasis), and emotions are situated, variable, and partly socially constructed rather than universal reactions.
- The 'triune brain' (lizard/limbic/neocortex layers) is a myth; brains share the same neuron types and reorganize as they scale.
- Brains run by prediction, not reaction, to regulate the body efficiently (allostasis).
- Emotions are constructed concepts applied to affective sensory data, so expressions like scowls have low reliability and specificity.
The brain operates by prediction and allostasis, anticipating bodily needs before they arise rather than reacting passively.
- The brain is trapped in a dark, silent skull receiving only the effects (sense data) of unknown causes in the world and body; this is the inverse inference problem—knowing effects but having to guess causes—and the brain's only additional resource for solving it is what happened in the past under similar circumstances.
“your brain is trapped in a dark silent box for your whole life, called your skull... these are the effects of some set of causes, but your brain doesn't know the causes, it only knows the effects, so that's a classic problem that philosophers call a reverse inference problem”
- Brains and many biological systems run by prediction rather than reaction; a brain waiting passively to react to stimuli would be metabolically inefficient, so the brain anticipates the body's needs and meets them before they arise—for example changing blood pressure as it prepares to stand you up rather than after.
“brains and in fact, many biological systems run by prediction, not by reaction, so your brain is not sitting around dormant, waiting for stuff to happen and then reacting to it that is incredibly metabolically inefficient”
- Because the brain is the most metabolically expensive organ, it functions like a budget for the body—budgeting resources such as salt, glucose, oxygen, and water rather than money—deciding which movements and learning are worth the energy investment with the expectation of a return.
“you can think of your brain as running a budget for your body, it's not budgeting money, it's budgeting resources like salt and glucose and oxygen and water”
- The brain's most important job is to regulate the body—not to think rationally, see, or feel; thinking, seeing, and feeling are all in the service of regulating the body, which is the most metabolically expensive organ's primary function.
“your brain's most important job is to regulate your body, brains didn't evolve to think rationally, they didn't evolve to see or feel”
- Because the brain allostatically adjusts to repeated exercise to make it more efficient (muscle memory), repeating the same exercise undermines cardiovascular benefit; interval training works precisely because it changes the exercise the moment the brain begins to adjust, keeping the heart maximally exercised.
“the reason why interval training works so great is that the minute the brain is starting to allosterically adjust your trainer changes your exercise program.”
Emotions are constructed concepts the brain builds from affect and interoception, not discrete reactions triggered by facial expressions or innate circuits.
- Affect—simple feelings of pleasantness/unpleasantness and calm/worked-up—is the brain's simple barometer of how the body is doing, derived from constant interoceptive sense data; affect is always present from birth to death and is a property of consciousness, but affect is not emotion: emotions are complex constructions the brain builds out of affect.
“you experience what's happening in your body as very simple feelings of pleasant-ness, unpleasantness, feeling calm, feeling worked up... and that simple barometer is what you would call feeling or mood, or what a scientist like me would call affect.”
- There is nothing physically meaningful about any facial expression per se—a wide-eyed gasping stare is the Western stereotype of fear but is a threat display in many parts of the world; meaning is imposed psychologically, and when two people share that meaning in context, each becomes more predictable to the other, reducing the metabolic burden on both brains.
“there's nothing specifically meaningful in a physical sense about any facial expression, per se, it's that we have, in a psychological sense, we have imposed meaning on it”
- Data show people in Western (and urban Eastern) cultures scowl in anger only about 30% of the time—better than chance but meaning their face does something else 70% of the time—and people also scowl when confused, concentrating, hearing a bad joke, or having gas; so the scowl has low reliability and low specificity as an indicator of anger, which is just a Western stereotype of anger.
“the data show that about 30% of the time people do scowl in the West, people do scowl in anger... But what it means is that 70% of the time your face is doing something else other than scowling”
- Emotions are made when the brain, asking what past experiences are similar to the present sensory array, constructs an emotion concept—a set of predictions about the meaning of incoming sense data—rather than a fear circuit in a lizard brain being triggered; the brain weighs probabilities, prepares the body and anticipated sense data, and if predictions match, the response executes, feeling like a simple reaction despite extensive preparation.
“it's really more complicated than the idea that you have a fear circuit in your lizard brain that gets triggered and then poof, you're fearful”
Interoception is the brain's dynamic model of bodily states, not passive sensation, and shapes all perception through multimodal integration.
- Interoception is not the awareness of bodily sensations but the brain's modeling of the body's sensory environment—its guesses about what the heart, lungs, and resource distribution are doing—used to figure out how to control the body and decide what to do next.
“sometimes people mistake interoception for the awareness of those sense data, but really interoception is the brain's modeling of the sensory environment of the body”
- Because the senses are not isolated but speak to one another (multimodality), and interoception is intertwined with vision, hearing, and touch, what you literally see is influenced by the state of your body even though you may not realize it.
“literally what you see is influenced by the state of your body, you may not realise that, but it is actually right”
- Convergent evidence from neuroanatomy, signal processing, electrophysiology, and functional brain studies—literatures that don't talk to each other—all point to the same idea: the brain re-implements past experiences similar to current conditions to predict what happens next, controlling the body and creating present experience using the past.
“your brain is basically using the past to predict the future as a way of controlling your body and creating your experience in the present.”
- Constructing a concept (e.g., imagining biting a McIntosh apple) changes the firing of your own neurons in early visual cortex, motor areas, and eating-preparation regions even with eyes closed and nothing eaten; dreaming, imagining, and actually perceiving the world involve very similar neural processes, and communicating concepts via words plants those neural changes in another person's brain like a kind of mental telepathy.
“If I had your head in a brain scanner with your eyes closed, we would see changes in early visual cortex... we'd see all of these changes, even though your eyes are closed, you're not eating anything”
Repeated experiences reshape the brain's predictions and concepts, allowing deliberate cultivation of future behavior and social influence.
- The experiences you have repeatedly today become the predictions your brain makes tomorrow—you continually cultivate your past as a means of controlling your future, so repeated experiences (like watching a TV show) actually change which concepts your brain constructs.
“the experiences that you have today become the predictions of tomorrow, basically, you are cultivating continually your past as a means of controlling your future”
- A concept is not a static laundry list of features but a situated construction: when asked for the concept of a bird, people imagine an instance whose features depend on the function it serves in context—a crow as a generic bird, a budgie as a pet, a turkey or duck as food—and the brain makes the prediction that fits the situation.
“your brain is making a situated concept, it's like whatever the function of a bird is in a given situation, that's the prediction that your brain is making”
- Because your brain has a continuous conversation with your body and other brains, and because today's experiences become tomorrow's predictions, understanding how the brain works lets you make more deliberate decisions about who you want to be and what impact you want to have on others—science and philosophy are tools for living, not merely academic pursuits.
“understanding a little bit more about brains helps you make decisions that are a little more deliberate about who you wanna be and what kind of impact you wanna have on other people.”
Constructed-emotion theory unifies evolutionary biology with social constructivism and opens new scientific questions about experience.
- When you examine the molecular makeup and genes of brain cells, there are essentially no new neuron types in larger brains—a lizard, rat, human, and bird brain share the same complement of neurons; brains reorganize as bodies and brains grow larger over evolutionary time rather than adding new layers, so humans have the same complement of neurons as a blood-sucking lamprey.
“when you look into the interior of the cells, you can see that humans have the same complement of neurons as a blood-sucking lamprey.”
- Humans uniquely create social reality—imposing on a physical thing a function it does not have by its physical nature, real only by collective agreement; money is the perfect example because diverse things (paper, salt, diamonds, big immovable rocks, Bitcoin) gain value purely by collective agreement and lose it when consent is withdrawn, yet such constructed realities still exert tremendous causal influence in the world.
“they only derive their value by virtue of what's called a collective agreement... money is the perfect example of social reality because it's so malleable in this way, it's like a collective illusion that we all create together”
- Progress in science is sometimes not about answering old questions but about asking newer, better questions, and the constructed-emotion perspective is valuable precisely because it opens up many new questions that did not exist before.
“Sometimes progress in science isn't answering old questions, it's asking newer better questions, and that's really what this perspective allows you to do.”
- Actual executive control and the feeling of being in control are not synonymous: a control network of neurons selects which inputs and predictions to complete and operates all the time (relaxed in sleep or under psilocybin/ketamine, enhanced by amphetamines/dopamine) largely without awareness, while the subjective feeling of agency is not particularly related to the function of these neurons.
“your experience of control, of being in control, it is not related particularly to the function of these neurons. All of this stuff that's happening, prediction and selection... is happening without your awareness and without any sense of agency or control”
- Scholars who question the traditional notion of free will are probably correct, but that does not mean you have no control over your behavior or that you are not responsible for it; we do have control and are responsible, but the manner of that control looks different from what common sense leads us to believe.
“the traditional notion of free will is, I think questionable, but you do have... we do have control over our behaviour and we are responsible for our behaviour, but the manner of that control looks a little different than common sense might lead us to believe.”
- A theory, as defined by the National Academies of Sciences, is a set of hypotheses or ideas backed by a great deal of evidence—so the theory of constructed emotion is not mere personal opinion but a considered position based on decades of research on brain anatomy, function, and metabolics.
“The National Academies of Sciences defines a theory as a set of hypotheses or ideas that are backed up by a heck of a lot of evidence”
- The theory of constructed emotion gives an evolutionary and biological foundation to how the brain constructs every experience and action, providing a single framework that integrates evidence from molecular genetics and biology up through cultural inheritance, anthropology, and sociology—dissolving the old divide between evolutionary/biological theories and social-constructivist theories of the mind.
“it's providing a framework that can integrate evidence from molecular genetics and other forms of biology all the way up to cultural inheritance and anthropology and questions in sociology.”
- Instances of anger are not one thing with one pattern but a population of instances: anger can raise, lower, or leave blood pressure unchanged; sometimes it harms an offender, but sometimes it brings people together, makes you feel part of a group, helps you compete or win a debate, and can even be pleasant—so emotions are the brain's situated prescription for action rather than uniform reactions.
“There is a population of instances, your brain can make many concepts for anger, and that means that it can control your body in ways that are tied very closely to the situation”
Emotion-recognition technology fundamentally fails because reading emotions requires high-dimensional context impossible for current algorithms to compute.
- Companies building emotion-recognition technology confuse the observation of a facial movement (e.g., detecting a scowl) with its emotional meaning; software can detect scowling but cannot tell you what it means, because reading a face requires high-dimensional spatial and temporal context including the observer's own internal body state—so deploying such tech for consequential decisions like school admission, criminal charges, or hiring is a serious error.
“what companies are doing right now is they're confusing the observation of a movement with its meaning, with its emotional meaning.”
- When the brain infers what a facial movement means, it does not use one or two extra signals like body movement or heart rate; it uses high-dimensional data—many signals creating an elaborate spatial and temporal context, including the observer's internal body state—and no current research approximates this, so emergence is a fitting metaphor.
“your brain is using high dimensional data. Lots and lots and lots of signals that are spatially and temporally creating this really elaborate context, so emergence is not a bad metaphor for what is happening here”
The layered-brain myth and the distinction between rational and emotional decision-making are scientifically false.
- The pervasive belief that the human brain evolved in distinct layers—an inner 'lizard brain' for instincts, a limbic system for emotion in mammals, and a rational neocortex on top—is a myth; the metaphor of the brain as a battlefield between an ancient beast and a rational evolved self is scientifically false even though it remains potent.
“the metaphor is that your brain is basically a battlefield between your inner ancient beast and your more rational evolved self and the whole thing is basically a myth, but it's a really, really potent, powerful method that even scientists sometimes have trouble stepping away from.”
- Because affect is always present given the anatomy of how brains are structured, moments of rational decision are not moments without feeling—it is not possible to make a feelingless decision unless one has a non-neurotypical brain or something is seriously wrong.
“rational moments, moments of rational decision are not moments without feeling, that's just not possible based on the anatomy of how our brains are structured.”