The Power of Emotions

Some years ago, a man I had known for a long time came to me in the week he received a terminal diagnosis. He wanted to start exercising. I had spent most of my working life helping people use exercise to live with illness, and I thought I knew what to offer him: a program, a progression, a few numbers to watch.

What happened over the next year was not in any program I had written. We went on long bike rides. He tried kinds of training he had never done before. He got into the best shape of his adult life, until the illness took over. And on those rides, and in the hours around them, we talked: long conversations about what mattered, and long stretches when neither of us said anything at all.

I could measure what the training did for his body. I could not measure the rest of it. I never knew exactly what he was feeling on those rides, and I did not ask him to name it. What I saw was a man spending his last year with his family, outdoors, doing the things a working body lets you do. My own account of it, then and now, is that what helped him most was movement together with deep conversation and quiet presence. I believe that. I cannot prove it. This piece is, in part, my attempt to understand it.

The question that year left me with is not mainly about performance. It is about freedom: how our feelings widen or narrow our ability to act on purpose. Every section that follows answers some part of that question. Along the way I will ask whether the capacity to bring feeling, body, attention and the demands of a moment together is something that could be measured. I raise it here as a question and return to it at the end, after the evidence, as a hypothesis.

Here is the frame I will keep returning to, and it is the closest thing in this piece to a contribution of my own. Most of how we judge people asks one question: did it work? I think every performance, at work, in sport or in a life, deserves three. Did it work? What did it cost, in effort, sleep, health and relationships? And was it chosen, or was the person being driven by fear, habit or someone else’s expectations? A person can score well on the first for years while failing the other two, and look successful the whole time. Emotions bear on all three. I will argue that they bear on the third most of all, though that is my judgment, not something the research settles.

These are not three independent scores, and none is entirely within our control. Effectiveness can come at an unsustainable cost; agency can exist even when circumstances leave us few options. And agency is a matter of degree, not a switch. Human action is rarely completely free or completely compelled. By agency I mean the capacity to recognize what is influencing us and to respond in ways consistent with our values, within real biological, psychological and social limits. I offer the three questions not as a validated measure, but as three ways of asking whether a performance has served the person who produced it.

In an earlier piece, What Feelings Are For, I argued that emotions are signals with jobs rather than interference to be removed. If feelings carry information, then reading that information well is a skill. Is that skill a kind of intelligence? And what does it change: how well we perform, or how freely we live?

Aristotle saw the shape of the problem twenty-three centuries ago, in the line above. Anger is easy. Getting it right, toward the right person, in the right amount and at the right moment, is rare. He was writing about character, not intelligence, and he would not recognize the psychology below. But his point survives the translation: what counts is not whether we feel. It is how well the feeling fits the moment and what we do with it.

Here is where I come out, so you can argue with it as you go. The evidence supports a modest version of the claim, not the grand one. Skill with feeling predicts good things, though by less than the popular books suggest. Some parts of the skill have better evidence than the umbrella term that holds them. Much of it can be learned. And the strongest case for it is not about performance at all, but about how much of what we do is chosen, and how much is a feeling choosing for us.

The piece moves in five movements. The first asks what emotions tell us. The second asks how they help and hinder thinking. The third asks how they shape our freedom to respond, and what that means for happiness. The fourth follows the same question into the lives we share: trauma, love and work. The fifth asks what can be learned. I am a reader here, same as you, and fallible. Where the evidence is thin I will say so. Where I am guessing, I will label the guess.

First, what is an emotion?

Everything in this piece depends on two words, so they deserve more than a glossary line. Neither has a single agreed definition, and the arguments over them are real. But there is enough common ground to work with.

The most widely used definition comes from appraisal theory. Klaus Scherer describes an emotion as “an episode of interrelated, synchronized changes” across several of the body’s and mind’s systems, in response to an event judged relevant to something we care about.² He names five components: the appraisal itself, changes in the body, an urge to act, expression in face and voice, and the subjective feeling. In plain words, an emotion is the whole of you reacting, briefly and in a coordinated way, to something that matters.

Two things follow from that picture. First, a feeling, the part we are conscious of, is only one component. You can have the racing heart and the urge to flee before you have a word for what is happening. Second, every component is a possible entry point for skill. You can work on the appraisal (what does this mean?), the body (slow the breath), the expression, the urge, or the label you put on the feeling.

Emotion, mood, affect, feeling

These words get used interchangeably, and the confusion causes real arguments. Here is how I use them.

Not everyone agrees that the components travel together as neat packages. Lisa Feldman Barrett’s theory of constructed emotion proposes that the brain builds each emotional experience in the moment, from body signals and learned concepts.⁴ A meta-analysis of more than two hundred laboratory studies found that heart rate, breathing and sweating did not clearly distinguish one emotion category from another.⁵ For this piece the debate has a practical upshot. A racing heart does not announce which emotion it belongs to, so the name we give it is a judgment, and judgments can be learned and revised. That is part of why skill with feeling is possible at all.

One more distinction matters for everything that follows. Emotion is not the same as emotional expression, and a calm face is not evidence of a calm interior. People can feel a great deal and show little, or show a great deal and feel less than it seems. Skill with feeling begins inside, with noticing, and only then concerns what we show.

Five families of feeling

If emotional skill starts with telling feelings apart, it helps to have names that mean something. Here are the nineteen emotions I wrote about in the emotion series for SportsFlow, grouped in the same five families. The families are a teaching framework of my own, not a recognized scientific taxonomy, and real feelings often blur across them. Storms are the hot, unpleasant states that mobilize. Mirrors are the emotions of comparison and self-judgment. Low skies are the slow, heavy ones. Rising airs are the pleasant states that lift and move us. Open weather is the pleasant states that widen us toward the world and other people.

Calm is missing on purpose. In that series I treated calm not as one more weather but as the sky the weather moves through. That image needs one clarification. It does not mean physiological calm, a slow heart and loose shoulders, which is simply another state, and often a useful one. It means something closer to what the contemplative traditions call equanimity: a steadiness of awareness that can sit alongside a racing heart, grief or anger. You can be shaking and still be the sky. More on that in the third movement.

The full field guide, with a definition of each of the nineteen, what it seems to be for and what happens when it takes over, is in the appendix at the end, for reference.

Three things stand out when you look at the nineteen together. First, the pairs. Fear and anxiety differ by time: one is about now, the other about later. Shame and guilt differ by target: the whole self or a single act. Envy and jealousy differ by what is at stake: something you lack, or something you could lose. Regret and disappointment differ by whom you blame: yourself or the situation. Learning to distinguish these pairs is one useful beginning of emotional differentiation: recognizing not merely that something is being felt, but what kind of feeling it is, what may have brought it about, and what it asks of us. Intensity, mixed feelings and change over time are the rest of the work.

Second, every unpleasant emotion on the list has a job, and every pleasant one has a way of going wrong. That is the case of the earlier piece in one table. No emotion is illegitimate to feel. But a feeling can be legitimate and still be misread, and a feeling can be accurate and still be unwise to act on. The questions are whether it fits the moment, whether we have read it rightly, and what is worth doing about it.

Third, the families move differently. Storms mobilize the body for action. Mirrors are about standing: mine, yours, ours. Low skies slow us down. Rising airs pull us toward a goal. Open weather opens us outward to the world and to other people. The rest of this piece is about learning to read all five.

What a feeling knows, and when to believe it

The strongest version of the claim that feeling is part of intelligence comes from neurology. In 1985 Paul Eslinger and Antonio Damasio described a patient, known as EVR, whose frontal lobe surgery left his measured intelligence superior. Yet his life fell apart. He could no longer meet personal or professional responsibilities, to the point that he was suspected of malingering.⁶ Damasio’s explanation became the somatic marker hypothesis: bodily signals tag options as good or bad, and without them, reasoning loses its bearings.

The famous test came in 1997. In a card game where two decks paid well early but lost money over time, ten healthy people began to show small anticipatory sweat responses before choosing from the risky decks, and shifted toward the good ones before they could say why. Six patients with damage to the ventromedial prefrontal cortex never developed those responses, and some kept choosing badly even after they could describe which decks were risky.⁷

It is a beautiful story, and it has not held up as cleanly as it is usually told. When researchers asked players more sensitive questions, nearly everyone who was choosing well already knew enough, consciously, to explain the choice. They found no support for the idea that a gut signal arrived before knowledge.⁸ A major review concluded that the anticipatory signals replicate inconsistently and that the hypothesis needs more evidence to remain tenable.⁹ Another study found that poor performance on the same card game was not specific to the brain region the theory named.¹⁰

Here is what survives. Emotion is woven into how people value options. A 2015 review of decades of work called its effects on decisions powerful, pervasive and predictable, sometimes helpful and sometimes harmful.¹¹ The narrower claim, that a body signal routinely tells us the answer before our minds do, is contested.

Feelings that wandered in from somewhere else

Decision researchers distinguish feelings that belong to the choice in front of you from feelings carried in from somewhere unrelated. The first kind is information about the decision. The second is noise that feels like information.

The carried-in kind is real but often smaller than its reputation. In a well-known 2004 experiment, people shown a sad film clip were willing to pay more to acquire an ordinary object.¹² A 2018 replication found the sadness effect again (d = 0.40), and pooled with the original it came to d = 0.43, with a confidence interval running from 0.12 to 0.77.¹³ Real, modest, uncertain. The disgust effect from the same 2004 paper was not part of the replication.

The most famous example has fared worse. In a 1983 study, people phoned on sunny days reported more satisfaction with their whole lives than people phoned on rainy days, unless the interviewer first mentioned the weather.¹⁴ It became a textbook demonstration that mood is used as information. Thirty years later, an analysis of more than a million survey respondents found most daily weather conditions unrelated to life satisfaction, and the few effects that appeared were tiny.¹⁵ The idea that we sometimes read our mood as a verdict on our lives is still reasonable. The sunny-day study is not good evidence for it.

When to trust a gut feeling

The most useful rule I know on this comes from an unusual collaboration. Daniel Kahneman spent a career documenting how intuition misleads. Gary Klein spent one documenting how expert intuition saves lives. In 2009 they wrote a paper together to find where they disagreed, and mostly failed to. They concluded that intuitive judgments deserve trust under two conditions: the environment must offer valid cues, and the person must have had the opportunity to learn them.¹⁶ They added that how confident an intuition feels is not a reliable sign that it is right.

Notice what the rule does not say. It does not say feelings are wise, or that they are foolish. It says that an intuition about a learned pattern is only as good as the learning behind it. Emotions do other work as well, such as telling us what we need and what we value, and that work does not depend on expertise. A nurse’s unease about a patient, earned over thousands of shifts, is different from my unease about a stock.

Listening to the body

The body offers signals of its own, and some people seem to read them better than others. In a small study, London traders who counted their own heartbeats more accurately made more money and lasted longer on the trading floor, though the study could not establish cause.¹⁷ The measurement itself is under question: heartbeat-counting scores mostly reflect how much people under-count,¹⁸ and they track simply knowing what a typical resting heart rate is.¹⁹ The most useful result is the most honest one. Better heartbeat perception sometimes helped people decide and sometimes hurt them, depending on whether their bodily signals happened to favor the good choice or the bad one.²⁰ Hearing the body more clearly is not the same as the body being right. The skill lies in the hearing and then the weighing.

What emotional intelligence is, and is not

Peter Salovey and John Mayer introduced the term into academic psychology in 1990. They defined emotional intelligence as the ability to monitor your own and other people’s feelings, to tell them apart, and to use that information to guide thinking and action.²¹ They later organized it into four branches, from the most basic to the most complex.²² In a later paper they described it as “the ability to engage in sophisticated information processing about one’s own and others’ emotions and the ability to use this information as a guide to thinking and behavior.”²³

Three different things share the name

Part of the confusion about emotional intelligence is that the phrase covers three different kinds of measure, which correlate only modestly with each other. Researchers sort them in slightly different ways. One common scheme separates ability, self-report and mixed approaches,²⁴ and many writers distinguish ability EI, measured by performance tests, from trait EI, measured by questionnaires.²⁵

The popular version arrived in 1995 with Daniel Goleman’s bestseller, subtitled Why It Can Matter More Than IQ.²⁸ Its broad definitions and big claims did a great deal to spread the idea and a great deal to muddy it. Mayer, Salovey and Caruso, who built the ability model, later called such popular claims “misleading and unsupported by research” and described emotional intelligence as “just one variable among many.”²³ The evidence later in this piece agrees with them.

So when I use the term in this piece, I mean the ability version unless I say otherwise: the capacity to perceive, use, understand and manage emotion. Where the evidence comes from self-report questionnaires, I say so, because those numbers mean something different.

Just as important is what emotional intelligence is not. It is not being calm. It is not being nice, agreeable or positive. It is not low emotional intensity: a person can feel everything strongly and handle it skillfully. It is not the same as empathy, which is one part of it. And it is not a moral virtue. As we will see, the same skill can serve kindness or manipulation.

Is it really intelligence?

Now to the bigger question. Before asking whether skill with feeling is a kind of intelligence, it is worth asking what we would gain by calling it one. Psychologists have a name for a recurring mistake in their own field: the jingle-jangle fallacy. The jingle is assuming two things are the same because they share a name. The jangle is assuming they differ because their names do.²⁹ “Intelligence” is a magnet for both. Once a word carries prestige, everything wants to wear it.

The best-known attempt to multiply intelligences, Howard Gardner’s, has never been validated, and a critical review was equally skeptical of emotional intelligence.³⁰ Others have argued that emotional intelligence is not a form of intelligence at all, and is defined so broadly it loses meaning.²⁷ Those objections deserve to stand next to the enthusiasm.

Here are the terms as I use them. They are working definitions for this piece, and other writers draw the lines elsewhere.

Sternberg’s idea is the one most often borrowed for arguments like mine, so it is worth being careful with it. His 2019 paper is a theory. It proposes measuring creative, practical and wisdom-based skills, and says plainly that the construct validity of those assessments still needs to be demonstrated.³¹ If I wrote that skill with feeling “is adaptive intelligence,” I would be borrowing a name, not a finding.

So I will hold the word loosely. What the research can actually answer are three narrower questions. Does skill with feeling predict anything that matters, beyond what intelligence and personality already predict? Can it be learned? And does it change how freely people live? The rest of this piece takes them in that order.

Why being smart is not enough

The simplest way to put the relationship is this: cognitive intelligence sets part of what a person could do, and emotional functioning shapes part of what they actually do with it, especially when the stakes are personal.

Two lines of evidence give the picture some weight. The first concerns what strong feeling does to the machinery of thinking. Attentional control theory, one of the better-supported accounts of anxiety and cognition, separates effectiveness, the quality of the result, from efficiency, the effort and resources it took to get there. Anxiety tends to hurt efficiency first. People compensate with extra effort, so the result can hold up while the cost rises.¹ Pressure can also take away the very resource the most capable people rely on. In a 2005 study, math pressure hurt only people with high working-memory capacity, and only on the problems that demanded the most of it.³⁵

The second line is more uncomfortable. Intelligence does not reliably protect anyone from bias. Keith Stanovich and Richard West found that a range of thinking biases were largely independent of cognitive ability.³⁶ A later study, whose title states its finding, showed that cognitive sophistication does not reduce the bias blind spot, our habit of seeing bias in others more readily than in ourselves.³⁷

And sometimes skill makes motivated reasoning stronger. In one well-known experiment, people good with numbers solved a data problem accurately when it was framed as a skin cream trial. When the same numbers were framed as a gun-control study, answers split along political lines, and the split did not shrink among the most numerate. It grew.³⁸ A later replication on immigration attitudes in a European sample found the motivated reasoning again but not the extra polarization among the most numerate.³⁹ So the strongest form of the finding is unsettled. The weaker form is not: being good at reasoning does not stop us from reasoning toward what we want to be true.

The pattern is easy to recognize once you see it. Here are three constructed sketches, not real people. A scientist so attached to a theory that contrary data feels like an attack, who then uses real brilliance to explain the data away. An executive whose forecasts are excellent and whose read on colleagues is reliably wrong. Someone who can explain attachment theory in detail and still pursues, withdraws or retaliates the moment they feel rejected, then explains it beautifully afterward. None of them lacks intelligence. Each is thinking hard from a premise a feeling chose.

What the emotional intelligence research shows

If skill with feeling matters, people who score higher on it should do somewhat better in work, school, health and relationships, and the scores should predict something beyond what intelligence and personality already predict. That second test is the one that matters. A measure that only repackages conscientiousness and IQ is a new name, not a new finding.

On the first test, the answer is yes, modestly. Across large meta-analyses, emotional intelligence correlates with job and academic performance at roughly .20 to .30,^40,41 and with health, well-being and relationship satisfaction at similar levels, higher for self-report measures.^42,43,44 The figure below shows the detail.

The pattern in the colors is worth noticing. The biggest numbers belong to questionnaires in which people rate their own emotional skill. That fits a worry critics have raised for twenty years: self-report emotional intelligence overlaps heavily with personality and current mood, so a happy, agreeable person may score high on both the questionnaire and the outcome without emotional skill doing the work. The pattern is consistent with that worry. It does not prove it.

On the second test, beyond intelligence and personality, the answer is: a little. In the academic meta-analysis, ability-based emotional intelligence explained an extra 1.7 percent of the variance in performance after intelligence and the Big Five personality traits were accounted for. Within it, the understanding and managing branches did better, at 3.9 and 3.6 percent.⁴¹ An earlier meta-analysis found ability emotional intelligence helped predict job performance in jobs with heavy emotional labor and hurt prediction in jobs without it.⁴⁵

Measurement is the other sore spot. The best-known ability test, the MSCEIT, often scores answers by consensus, crediting whatever most people choose. Critics point out that this may capture common lay theories about emotion rather than skill, and an item-level analysis found many items too easy to separate people at the high end.²⁶

Then there is the question of what the skill is for. In a pair of studies, people who knew more about regulating emotion were more prosocial if they also had a strong moral identity, and more interpersonally deviant at work if they were also Machiavellian.⁴⁶ Skill with feeling is a capacity, not a virtue. It can soothe a friend or manipulate a colleague. That is worth remembering every time someone treats emotional intelligence as a synonym for goodness.

Taken together: real, modest, best measured by tests rather than self-ratings, and with no guarantee about how the skill will be used. Enough to take seriously; not enough to crown a new intelligence.

The skills under the label

An umbrella term can be weaker than the things it covers. When you look past “emotional intelligence” at the specific skills underneath, some of them have clearer evidence than the umbrella does.

Telling feelings apart

Some people experience a bad day as a single gray mass: stressed, upset, bad. Others can tell disappointment from resentment from fatigue. Psychologists call this emotion differentiation, or granularity. In an early diary study of 53 students, those who drew finer distinctions among their negative feelings reported regulating them more, especially when the feelings were intense.⁴⁷ Later reviews tied finer distinctions to less binge drinking and less severe anxiety and depression, while noting that the mechanisms still await testing.⁴⁸

The size of the effect is the honest part. A 2022 meta-analysis of seventeen studies found a correlation of −.15 between negative emotion differentiation and problem behaviors such as binge drinking and self-injury, falling to −.09 once overall negative feeling was taken into account.⁴⁹ Small, consistent, and correlational. Knowing what you feel seems to help a little, or people who are doing better find it easier to know what they feel, or both.

Putting a feeling into words also seems to soften it. A review of affect labeling describes effects similar to reappraisal, across self-report, body and brain measures, while warning that the mechanisms have not been directly investigated.⁵⁰ In one imaging study, labeling reduced how unpleasant images felt compared with simply looking at them, but less than deliberate reappraisal did.⁵¹ Naming is a start, not the whole job.

What people do with feelings

A large 2010 meta-analysis, 241 effect sizes from 114 studies, asked how habitual ways of handling emotion relate to anxiety, depression, eating and substance problems. Rumination had the largest association. Avoidance and suppression were also strongly linked. Reappraisal and acceptance, the strategies most treatments teach, showed smaller associations, which the authors called surprising.⁵² The studies were mostly questionnaires at a single time point, so they describe habits that travel with distress, not proof of what causes it.

Suppression deserves a word of its own because it is so often mistaken for composure. Across several samples of students, habitually hiding feelings went with more negative emotion, less positive emotion, lower well-being and less closeness to others.⁵³ Trying not to think about something can also bring it back, although the rebound in controlled studies is small to moderate and depends on what is being suppressed.⁵⁴ A calm face is not the same as a settled mind.

Flexibility, not one right strategy

The most interesting finding here may be that no single strategy is the right one. George Bonanno calls the belief otherwise the “fallacy of uniform efficacy.” His alternative, regulatory flexibility, has three parts: reading what a situation calls for, having a range of responses available, and noticing whether the chosen response is working.⁵⁵ He describes the evidence as preliminary.

One study gives the idea real texture. Shortly after the September 11 attacks, 101 New York undergraduates were tested on their ability both to amplify and to hide their emotional expression on cue. Two years later, among the 87 still in the study, those who could do both, rather than only one, had lower distress, even after accounting for how distressed they were at the start.⁵⁶ The skill was not calm. It was range.

The wider idea of psychological flexibility, staying in contact with the present and acting on your values even when what you feel is unpleasant, has been described as a fundamental aspect of health. Its proponents also note that it is unclear whether inflexibility causes problems or results from them.⁵⁷ Acceptance and commitment therapy, built around that idea, has been tested widely. A review of twenty meta-analyses, covering 12,477 participants, found it effective across the conditions examined and better than inactive controls and usual care, but not better than cognitive behavioral therapy.⁵⁸ Its developers note that the evidence for its mechanisms rests mostly on self-report,⁵⁹ and its main questionnaire may partly measure distress rather than flexibility.^60,61

Limited in different ways

One practical consequence follows. “Low emotional intelligence” is not a single deficit. Someone can be finely attuned to other people and still be swept away by their own anger. Someone else can regulate their own feelings coolly and miss everyone else’s. A third can name exactly what is happening and repeat the same behavior anyway. The table below is a way of sorting that, not a diagnostic tool.

What it costs not to know what you feel

The usual picture of low emotional skill is someone too emotional. I think the bigger cost is quieter: being steered by feelings you have not noticed while believing every choice is deliberate. The table lists the costs most often claimed, with my judgment of the evidence for each. They are possibilities, not destinies.

Mistaking the force of a feeling for the strength of an argument

Feelings do not only push conclusions. They also change how certain the conclusions feel. In one study, fearful people judged risks to be higher and preferred safer options, while angry people judged risks lower, more like happy people did.⁶³ A set of four experiments looked at certainty directly. Emotions that carry a sense of certainty led people to rely on shortcuts: they leaned more on a speaker’s credentials, stereotyped more and paid less attention to whether an argument was any good. Emotions that carry uncertainty led to more careful thinking. What mattered was the certainty in the feeling, not whether it was pleasant.⁶⁴

That is the philosophical heart of the problem. Shame can make a harsh judgment of yourself feel simply true. Fear can make a danger feel probable. Anger can make you feel sure. The intensity of a conviction is easy to mistake for evidence for it. The capacity to catch that, to notice the processes shaping your own thinking, has a name in psychology, metacognition, and it does not come free with a high IQ.

Two people, one cycle

In close relationships the costs compound, because each person’s reaction is the other’s situation. A familiar pattern: one partner feels uncertain and reaches for reassurance. The other feels pressured and pulls back. The pulling back raises the first partner’s anxiety, and the reaching intensifies. Both believe they are responding reasonably to the other. Researchers call this the demand-withdraw pattern. A meta-analysis of 74 studies with more than 14,000 participants linked it to lower relationship satisfaction and intimacy, poorer communication, and more anxiety and depression in the partners themselves, whichever partner did the demanding.⁶² The data are correlational, so the pattern may feed distress, follow from it, or both.

Neither person in that cycle necessarily lacks love or intelligence. What is often missing is the ability to see your own part while it is happening. Attachment history, communication skills and circumstances matter too, and emotional skill cannot explain a relationship on its own.

Small choices, long consequences

A single avoidant decision costs little. A pattern of them can shape a life. Skipping the hard conversation preserves comfort today and weakens the relationship over years. Avoiding failure protects self-esteem now and limits growth later. Suppressing sadness keeps you functioning and hides the loss that needs attention. I know of no study that follows this exact chain over decades, so read it as a plausible pathway, not a finding. It is also the place where emotional skill matters most, because no single moment looks important enough to examine.

Feeling better and living better

Here the argument turns. So far the question has been whether emotional skill helps us function. The more important question, I think, is whether it helps us live well, and that turns out to be a different target.

Psychologists separate two broad families of well-being. Hedonic well-being is feeling good: pleasure, comfort, satisfaction. Eudaimonic well-being, a term borrowed from Aristotle, is living well: purpose, growth, acting from your own values.³³ The split is useful and imperfect. A review of the field found eudaimonia defined so inconsistently that findings often fail to line up, partly because some researchers treat it as a way of living and others as a feeling.³⁴

The two overlap, but they come apart often enough to matter. A person can grieve while living a deeply meaningful life. A person can be anxious while doing the most important thing they have ever done. And a person can be comfortable and adrift.

Chasing the feeling

Some of the clearest evidence that feeling good is a slippery target comes from people who want it most. In two small studies of women, those who placed a very high value on being happy reported lower well-being when life was not especially stressful, and felt less happy after watching a happy film clip, apparently because the moment fell short of what they hoped.⁶⁵ A larger follow-up across five samples narrowed the finding. Aspiring to happiness was not the problem. Being anxiously concerned about whether you are happy enough was.⁶⁶ A review of the wider literature argued that happiness is not beneficial at every level, at every time, for every reason and in every form.⁶⁷

The opposite stance seems to help. Across a laboratory study, a diary study and a six-month follow-up, people who habitually accepted their unpleasant thoughts and feelings, rather than judging them, felt less negative emotion during stress, and that in turn went with better psychological health later.⁶⁸ Accepting the feeling did not mean accepting the situation. Accepting circumstances was a separate habit, and it did not carry the same benefit. All three studies were observational, so the benefit is a strong association, not a proven effect.

I find that a relief. The aim is not to feel good all the time. It is to stop treating each unpleasant feeling as an emergency.

The space between feeling and action

Agency is the word I keep coming back to. It does not mean control over your circumstances, or even over what you feel. It means some say in what happens next. Two people can feel the same rejection with the same force. One takes the feeling as a verdict, “I am not enough,” and acts on it at once: reassurance, withdrawal, attack. The other feels it just as much, recognizes it as a feeling and not a fact, and has a little room to look, wait and choose.

Self-determination theory, one of the most widely tested frameworks in motivation research, gives this a home. It proposes three basic psychological needs: autonomy, meaning the sense that your actions are your own, not independence from others; competence; and relatedness.⁷⁰ Satisfying those needs goes with greater well-being. In a meta-analysis of studies with older adults, for example, overall need satisfaction correlated −.48 with depression, though that estimate rests on just three studies.⁷¹ In a study across South Korea, Russia, Turkey and the United States, a sense that one’s actions were self-endorsed predicted well-being in all four cultures.⁷² The evidence is consistent and mostly correlational, and most of it comes from the theory’s own research community.

Emotional skill plausibly supports all three needs, and the influence probably runs both ways: people who feel safe and supported have more room to regulate.

The three questions, applied

This is where the three questions from the opening earn their keep. Effectiveness and cost are separate: anxiety, as we saw, tends to raise the cost of a result before it lowers the result itself.¹ Agency is separate again. The pattern I see most often in capable people is high on the first, high on the second and low on the third: excellent work, paid for with sleep and relationships, done because stopping feels impossible rather than because it was chosen. Nothing on a results sheet shows that pattern. The person living it often cannot see it either, until a feeling they have long been managing finally refuses to be managed.

The band at the bottom of that figure is not decoration. Agency is never only internal. In rich countries, depression and anxiety are one and a half to two times as common in the poorest fifth of the population as in the richest, and in a cash-transfer experiment in Kenya, transfers improved happiness and large ones lowered stress hormones.⁷³ Illness, discrimination and caregiving narrow real options regardless of anyone’s inner skill. No amount of emotional regulation replaces money, safety, opportunity or people who have your back. Any argument for emotional skill that forgets this becomes a way of telling people with hard lives that their suffering is a skills deficit. I do not believe that, and the evidence does not support it.

Nothing stays: happiness, equanimity and connection

If there is one fact about feelings I wish I had learned earlier, it is this: none of them lasts. Joy rises and falls. So does dread. Even grief, the feeling most likely to seem as if it will never end, changes shape over time, and so does the memory of the best afternoon of a life. Contemplatives noticed this long before anyone measured it. The Buddhist teaching of impermanence puts it at the center of things, and Rumi’s guest house, which I paraphrased in What Feelings Are For, imagines every feeling as a visitor who arrives and leaves.

Measurement agrees with that small point, though measurement and teaching are different kinds of claim, and one does not prove the other. When 233 high-school students reconstructed recent emotional episodes, the typical episode faded back to baseline within a couple of hours. The variation was enormous: sadness lasted longest, up to 240 times longer than shame, the briefest. What stretched a feeling out was not only how strong it was, but how important the event was and how much the person kept turning it over in their mind.⁷⁴ Rumination, in other words, is partly the act of refusing to let a feeling pass.

That has a plain implication for happiness. If happiness means a feeling, it cannot be held, because no feeling can. Chasing it as a state to be kept is chasing a wave. This may be part of why anxiously monitoring your own happiness backfires.⁶⁶

What passes, and what does not

It would be dishonest to say everything passes. Studies that follow the same people for years find that we adapt to a great deal, but not to everything. After marriage, happiness rises briefly and then drifts back. Widowhood hits hard, and on average people recover most of the way over about eight years, though not quite all of it. Unemployment leaves a lasting dip that can persist even after a new job. Severe disability shows little adaptation at all. And people differ widely in how far and how fast they adapt to the same event.^75,76 Some losses leave a mark. Equanimity is not pretending they don’t.

Underneath the moment-to-moment weather, there is a steadier layer. In a German panel of 3,608 people rated every year from 1984 to 2000, average life satisfaction in the first five years correlated .51 with the last five. Steadier than mood, then, but not fixed: about a quarter of people changed significantly over those years, more of them downward than up.⁷⁷ The steadier layer is real, and it can be built or worn away.

Equanimity is not flatness

Equanimity is easy to misunderstand. One group of contemplative researchers defines it as an even-minded state or disposition toward all experiences, pleasant, unpleasant or neutral.⁷⁸ Even-minded does not mean feeling less. It means meeting a feeling without immediately grasping it or pushing it away. The waves still come. You are less thrown by them. Nor is equanimity the same as physiological calm. A person can be even-minded with a pounding heart, in the middle of grief or in anger. It describes the relationship to the state, not the state itself.

I should be honest about how young this science is. The questionnaires that measure equanimity are only a few years old, rely on self-report, and disagree about how many parts it has.⁷⁹ One of them overlaps heavily with low neuroticism, so “trait equanimity” may partly be ordinary emotional stability under a gentler name.⁷⁹ The best trial I found was short: four sessions of mindfulness training reduced reactivity to unpleasant experience, but only in people whose mindfulness actually rose, and it did not increase acceptance as the authors expected.⁸⁰ Promising, not established.

Connection is the steadier ground

The evidence on connection is far larger. A meta-analysis of 148 studies following more than 300,000 people for an average of seven and a half years found that people with stronger social relationships had about 50 percent higher odds of survival (odds ratio 1.50, 95% CI 1.42 to 1.59). Being integrated into a web of relationships mattered more than simply not living alone. The authors caution that causation is hard to establish.⁸¹

Among 222 students, the happiest tenth were distinguished above all by strong relationships. They still had bad moods. The researchers’ summary is one of my favorite sentences in psychology: “No variable was sufficient for happiness, but good social relations were necessary.”⁸² And in a small daily study of couples in their eighties, marital satisfaction, though not simply time spent with other people, softened the link between bad health days and bad moods.⁸³ It was the quality of the bond, not the quantity of company.

Connection is not permanent either. People move, change and die. But connection is less a feeling than a practice, something renewed by what we do. I have argued elsewhere that love is closer to steady, present action than to an emotion, and I think the same holds here. A feeling of closeness passes like any other. The habit of showing up, listening and repairing is what remains.

So here is the version of the happiness question I find most useful. Not “how do I feel good and stay there?” That question has no answer, because nothing stays. Rather: “What steadier ground can I build under the weather?” The evidence points at two candidates, a way of meeting feelings without being thrown by them, and the people we share our lives with. Neither guarantees anything. Both last longer than a mood.

The self that feels threatened: emotion and the contemplative path

Emotional skill concerns how we relate to what we feel. Contemplative spirituality, in its many forms, concerns how we relate to experience itself: who we take ourselves to be, what we cling to, what we belong to. The two are not the same, and I want to keep straight what psychology can show and what the traditions propose. But they overlap more than either side usually admits.

Three ways to hold the same anger

Take anger. At its most reactive it says: I am angry, and I am right. With more awareness it becomes: I am feeling anger, and something I care about seems threatened. Many contemplative traditions point at a third possibility: anger is arising, it does not define me, and I do not have to act on it.

The third stance has a psychological counterpart with some evidence behind it. Decentering, observing thoughts and feelings as passing events rather than facts, is among the better-supported ways mindfulness programs seem to work.^85,86 A related technique from acceptance and commitment therapy, cognitive defusion, asks people to repeat a painful self-critical word aloud until it becomes just a sound. In a small early study of eight students, that reduced both the discomfort of the thought and how believable it felt, more than distraction or trying to control the thought did.⁸⁴ A tiny study, but a vivid one: the thought did not change, the grip did.

The parallels are real, and so are the differences. A person can have great emotional skill and no spiritual practice, or be devout and struggle badly with their own feelings. And spirituality includes things emotional skill does not touch: the sacred, ultimate meaning, for many people the divine.

What the research can and cannot say

Four lines of evidence bear on this, at very different strengths.

Spiritual community and health. In a study that followed 74,534 American nurses for sixteen years, women who attended religious services more than once a week had a 33 percent lower rate of death over the period than those who never attended (hazard ratio 0.67). Social support was the largest of the measured pathways, at 23 percent of the association, and no single pathway explained more than about a quarter. The authors stress that the data are observational, and that the finding is not a reason for doctors to prescribe attendance.⁸⁷

The form spirituality takes matters. A meta-analysis of 49 studies found that positive religious coping, such as seeking comfort and meaning, went with better adjustment to stress (r = .33), while negative religious coping, such as feeling punished or abandoned by God, went with more distress (r = .22).⁸⁸ A later meta-analysis of longitudinal studies found that religious and spiritual struggles predicted increases in distress over time, though the effect was small.⁸⁹ Spirituality can be a refuge or a source of fear and guilt. It depends on the form.

Experiences that shrink the self. Researchers describe self-transcendent experiences as passing states of reduced self-focus and increased connectedness, while noting that some forms of losing the sense of self belong to mental illness.⁹⁰ Awe is the most studied. In five studies, awe made people more generous and helpful; in one, just a minute looking up at towering trees reduced entitlement and increased helping.⁹¹ A 2025 meta-analysis of 221 studies reported a correlation of .37 between awe and prosocial behavior.⁹² Encouraging, still settling.

Meditation and the brain’s self-talk. One much-cited study compared twelve experienced meditators with thirteen non-meditators and found less activity in brain regions tied to self-referential thought and mind-wandering, the default mode network.⁹³ It is suggestive and small, and a major critical review of meditation research warns that imaging studies in this area are often small, cross-sectional and confounded by differences in movement and breathing.⁹⁴ None of it shows that meditation dissolves the ego or reveals a truth about consciousness.

So the scientific case is strongest for the psychological mechanisms: attention, decentering, regulation, meaning and community. It says nothing about which worldview is true.

Agency and surrender

Here is a paradox worth sitting with. I have spent this piece arguing for agency. Many spiritual traditions counsel surrender, acceptance, letting go. They seem to point in opposite directions.

I think the contradiction mostly dissolves once you separate controlling experience from choosing your response to it. Facing an uncertain future, you may control very little of what happens and little of whether fear arises. You can still have some freedom in how you relate to the fear and what you do next. Acceptance in this sense is not passivity, and accepting an internal experience never requires accepting injustice, mistreatment or harm. It is seeing what is happening instead of spending all your energy resisting the fact that it is happening, and then acting within reality rather than against it. The psychological research draws the same line: accepting feelings was linked with better health, while accepting circumstances was a separate habit without the same benefit.⁶⁸ The Stoics and the Buddhists would frame this very differently, and their underlying commitments differ a great deal. But they share this practical point.

Perhaps freedom is not control over our experience. Perhaps it is the capacity to take part in experience without being entirely controlled by it. This adds something to the space between feeling and action described earlier. The contemplative traditions point to a second space, between a feeling and the self that is having it. The first space lets you choose what to do. The second lets you hold the feeling more lightly while you choose.

When spirituality becomes avoidance

There is a shadow here, and it deserves a name. In the 1980s the psychotherapist John Welwood coined the term spiritual bypassing for using spiritual ideas to avoid unresolved emotional work.⁹⁵ It can look like reading your own anger as spiritual immaturity and suppressing it, even when it is signaling a real boundary violation. Or invoking acceptance to avoid confronting mistreatment. Or using nonattachment to justify emotional withdrawal. Or meditating to make an uncomfortable feeling go away rather than to understand it.

The evidence on spiritual bypassing is thin; the term comes from clinical writing and its measures are not yet well validated. But the pattern fits the rest of the evidence here. Suppression has costs whatever language it wears.⁵³ Sometimes maturity means accepting discomfort, and sometimes it means acting to change something harmful. The skill is telling the two apart.

The self that feels threatened

Underneath all of this is identity. Many of our strongest emotional reactions come when something central to our sense of self is threatened. We want to be loved, respected, safe, competent, understood. None of that is a problem in itself. But when identity depends rigidly on particular outcomes, every mistake becomes a threat to worth, every rejection a verdict, every uncertainty intolerable.

Emotional skill helps us see those patterns. Contemplative inquiry asks: do I have to keep organizing my identity around them? Not just “how can I respond better when I feel threatened?” but “why does this threaten me so much?” And eventually, “what is left of me when I no longer need every experience to confirm who I think I am?” That question belongs as much to philosophy and the traditions as to psychology, and I do not claim the research answers it. Saying our ordinary sense of self is partly built from learned stories that can be examined is not the same as saying the self is an illusion.

The research on wisdom offers a modest bridge. A broad review of the field concluded that its different empirical approaches converge on the morally grounded application of metacognition, thinking about your own thinking, in the service of good ends.⁹⁶ Emotional skill supplies some of the raw material. Wisdom also needs knowledge, experience, ethics and care for others. Skill with feeling alone is not enough, and the traditions have always said so.

Emotional skill gives us more freedom in how we respond to life. Contemplative insight may show how much of our suffering comes from insisting that life unfold a particular way.

What trauma teaches, and what it costs

Trauma complicates every simple story about emotional skill, because it can sharpen some abilities while weakening others. Someone who grew up in an unpredictable home may become remarkably good at reading a room: the shift in a voice, the set of a jaw, the conflict three moves away. The same person may find it hard to know what they themselves feel or need, to trust their own reading, or to act in any way other than to stay safe. Trauma can make a person extraordinarily skilled at reading their surroundings and less able to feel safe in them.

Sensitivity is not the same as accuracy

The research bears out the first half of that. In a well-known study of 47 children, those who had been physically abused identified angry faces from less visual information than other children did. The advantage was specific to anger. For sadness it ran the other way.⁹⁷ A review of seventeen studies of maltreated children found greater reactivity and a response bias toward negative faces, especially angry ones, but less accuracy on facial tasks overall, and the authors called the evidence exploratory.⁹⁸ A nervous system trained by danger gets faster at spotting a threat. It also starts seeing threat where there may be none. Vigilance is not the same thing as accuracy.

Attention trained outward

The second half, losing touch with your own inner life, has support too. A 2026 meta-analysis of 36 studies estimated that about 53 percent of people with post-traumatic stress disorder also meet criteria for alexithymia, difficulty identifying and describing their own feelings (95% CI 42 to 65 percent).⁹⁹ An earlier meta-analysis found a large association between the two.¹⁰⁰ That is an association, not proof that trauma causes alexithymia. But it fits a familiar pattern. Someone who spent years tracking what everyone else needed may have had little practice asking: what am I feeling, and what do I want?

Regulation, too, can organize itself around avoiding distress. Across 75 studies, post-traumatic stress symptoms correlated with rumination (r = .52), experiential avoidance (r = .48) and thought suppression (r = .29), most of it (64 of 75 studies) from studies measuring both at a single point in time.¹⁰¹ These strategies can make things bearable in the short run and narrow life in the long run.

So trauma rarely produces “low emotional intelligence” across the board. It tends to produce uneven development: fine attunement to others alongside limited access to yourself. That is one more reason to treat emotional skill as a set of capacities rather than a single score.

What I learned to read

I know this paradox from the inside. For some years as a boy I lived with an adult whose anger came in long tirades, often aimed at me. Even then I could see that the person delivering them was in pain. What I could never work out was what I had done to cause it. I did have an answer to those tirades. It was not one a child was allowed to give, so I gave none.

What it taught me, as I put it much later, was to walk very gently, and to be hypervigilant about every emotion and every inflection of a voice, right down to the way someone walked into a room. I became very good at reading people, and I still am. But it was not accuracy. It was an alarm set to go off early, and it went off at plenty of things that were not fires.

The sentence I needed was a simple one, and I did not hear it until I was a grown man: nobody had ever told me that I was not responsible for other people’s emotions. Without it, the only move left was to try harder. Decades later, in a marriage, I caught myself doing the same arithmetic, performing more and more with the sense that no amount would ever be enough. I understood the pattern long before I could step out of it. The insight had arrived. The freedom had not.

I have spent much of my working life building ways to read what people are carrying: questionnaires, instruments, a trained kind of attention. I will not draw that line more firmly than that. But I know where I learned to look.

Insight without freedom

This is where trauma meets agency, and where my own story above stops being only mine. Imagine someone who correctly senses that a partner is becoming distant. Their emotional perception is excellent. But if distance once meant abandonment, the feeling arrives with the force of the past. They seek reassurance, withdraw first, or work frantically to repair the connection. The problem is not recognizing the emotion. It is keeping any flexibility in how they respond to it. Trauma can strengthen the ability to predict what others will do while weakening the felt freedom to choose what you will do. That is why insight alone so often fails to change behavior. The space between feeling and action is not only a matter of skill. The past can close it, and what reopens it is less often a new understanding than repeated experiences of safety.

Growth is possible, not promised

Rumi’s image of the wound as the place where light enters is one I have carried for a long time, and I still believe it describes something real. Many people do report growth after adversity: deeper empathy, clearer priorities, closer relationships. The careful evidence asks us to hold the image as a possibility rather than a promise. A meta-analysis of 122 longitudinal studies with almost 100,000 participants measured actual change over time rather than people’s own sense of having grown. It found no general evidence that negative life events produce more growth than positive ones, no genuine growth in meaning or spirituality, and, in most controlled studies, no difference between people who went through an event and people who did not.¹⁰²

Three different outcomes get confused here: survival adaptation (becoming better at detecting and avoiding threat), recovery (regaining safety and flexibility) and growth (new perspectives or capacities from working through what happened). Growth can happen. It is not required, recovery does not depend on finding a silver lining, and nobody owes their suffering a lesson.

When contemplative practice meets trauma

Contemplative practice can help a survivor notice that a present reaction is being shaped by the past, and observe fear without treating it as proof of danger. It can also hurt. In a study of 60 Western Buddhist meditators who had run into difficulties, 43 percent reported re-experiencing traumatic memories during practice, and a trauma history was among the factors associated with how intense and lasting the difficulties were. The sample was recruited for its difficulties, so this says nothing about how common such problems are, only that they are real.¹⁰³ Long silent sitting with close attention to bodily sensation is not the place to start for everyone.

Rebuilding after trauma

Four capacities seem central, and they develop best somewhere safe. First, telling present conditions from past expectations: noticing when a reaction is shaped by what used to be true. Second, restoring awareness of your own needs, preferences and limits after years of tracking everyone else’s. Third, regulation without suppression: not removing fear or sadness, but gaining enough room to respond to what is actually here. Fourth, practicing agency in small, repeated choices: a boundary held, an uncertainty tolerated, a different response tried and survived.

Insight alone rarely does this. People usually need to experience safety, try a different response, and find that what happens is not what they learned to expect. When trauma symptoms are persistent or disabling, trauma-focused therapies have good evidence. A meta-analysis of 114 randomized trials found robust evidence for trauma-focused cognitive behavioral therapies and EMDR, with cognitive processing therapy, cognitive therapy and prolonged exposure among the strongest.¹⁰⁴ This is work to do with a trained clinician, not alone with an article.

Most of us carry some of both. The first is excellent in danger. The second travels better. If I had to name the single most meaningful expression of emotional skill after everything in this piece, it would be this: the capacity to tell what the past has taught us to expect from what the present actually requires, and to respond to the present.

Love: a feeling, an understanding, a verb

If any emotion tests everything in this piece, it is love, because love is more than an emotion. It involves feeling, but also attention, attachment, judgment and behavior. It can bring out remarkable generosity and resilience, and it can bend our reasoning and quietly take our agency. Is love something we experience, something we understand, or something we do? I think it is all three, and most of what it does to us depends on how the three relate.

Not one thing

We talk about love as if it were a single feeling, like anger or fear. It is closer to several systems that usually, but not always, move together.

The strands come apart all the time. You can be intensely attached to someone and treat them badly, or go on loving someone after the relationship ends. Most important, you can act lovingly while angry, disappointed or exhausted. Loving feelings do not guarantee loving behavior, and a passing absence of the feeling does not mean the love has gone.

Does love make us smarter?

In one sense, yes. Emotion decides what is worth attention, and love decides it powerfully. A parent notices a change in a child that a stranger would miss. A long-time partner remembers preferences and vulnerabilities that would otherwise seem trivial. Caring about someone makes us willing to spend real thought on their problems.

In another sense, love is the greatest distorter we have. Attachment makes rejection feel like danger, desire reads the evidence selectively, and fear of loss turns uncertainty into rumination.⁷⁴ An intelligent person can build convincing explanations for someone’s behavior because the alternative, that the relationship cannot give them what they need, is too painful to consider.³⁸

Three questions help separate the strands:

• What do I feel toward this person?
• What do I know about this person and what is actually happening between us?
• What would a genuinely loving response look like here?

The answers often disagree. Emotional skill helps you notice that they disagree. Wisdom is acting well when they do. Sometimes love calls for closeness. Sometimes it calls for a boundary. Sometimes it means letting someone make a choice you never would. At that point love looks less like a state and more like an exercise of judgment.

Being loved, feeling loved, being able to love

Close relationships are among the most robust predictors of well-being in all of psychology. Relatedness is one of the three basic needs in self-determination theory,⁷⁰ and the large studies of connection and survival, and of the happiest people, point the same way.^81,82 But three experiences hide inside the word. Being loved, feeling loved and being able to love are related, not identical. Someone can be surrounded by people who care for them and be unable to take it in. Someone else can be endlessly generous and never let themselves be known. A third person can feel intensely loved in a relationship that erodes their autonomy.

So what matters is not how much love, but its quality, and whether it lives alongside security, respect and freedom. In the study of couples in their eighties, it was satisfaction with the marriage, not time spent with people, that softened bad days.⁸³ Across 126 studies and more than 72,000 people, better marital quality went with better health, with small correlations of about .07 to .21, similar in size to the links between health and habits such as diet.¹⁰⁵

Love as a verb

I have come to think love shows itself less in grand gestures than in consistency: showing up, paying attention, doing the next caring thing when you are tired. Grand gestures have always felt to me more like performance than love. I am bad at them and reasonably good at consistency, steadiness and warmth, which served me well raising children and dogs. I still get seduced by intensity. Think of a parent caring for a child through exhaustion and worry. The warm feeling is not there every minute. The care is. Love lives in the behavior even when the feeling is hard to find.

Love as action looks something like this:

• Paying attention instead of assuming you already understand.
• Being honest when honesty is uncomfortable.
• Respecting the other person’s autonomy.
• Keeping boundaries that protect you both.
• Staying responsive in conflict rather than defending yourself on reflex.
• Accepting that caring does not give you control of the outcome.

None of these runs on affection alone. Each takes regulation, perspective-taking, judgment and agency, which is to say the emotional skills this whole piece has been about. Love as a feeling tells us someone matters. Love as action shows what we do with that.

The research on what loving support does for the person receiving it is encouraging. Brooke Feeney and Nancy Collins describe two jobs close relationships do for us: being a source of strength in hard times, and being a catalyst that encourages us to take up opportunities in good times. They also note that few studies test these predictions directly.¹⁰⁶ In a brain-imaging study, sixteen married women facing the threat of a mild shock showed broad reductions in threat-related brain activity while holding their husband’s hand, more so the happier the marriage, and only a limited reduction while holding a stranger’s.¹⁰⁷ A larger follow-up with 110 people found that a familiar partner’s hand reduced distress and threat-related activity, while a stranger’s did not.¹⁰⁸ And in research on what is called the Michelangelo phenomenon, partners who affirmed each other’s ideal selves helped each other move toward them.¹⁰⁹ At its best, love does not only comfort. It makes us braver.

When love learned to protect

Trauma, as the previous section showed, teaches us what love means. One person learns that loving means anticipating everyone else’s needs; another, that it means never depending on anyone. Both are protective, and both get in the way of intimacy. Neither means love is absent. But the capacity to love and the capacity to sustain a mutually healthy relationship are not the same, and loving feelings never excuse harmful behavior.

Love that expands agency, love that shrinks it

Love asks us to care about something beyond ourselves, and that makes us vulnerable. If your well-being matters to me, I cannot be indifferent to what happens to you. Does that mean love costs agency? I don’t think so. The line runs between freely caring for someone and being governed by the fear of losing them.

The goal of emotional development is not to love less intensely. It is to love without letting fear decide every response. Love also changes where the space between feeling and action lives. Between two people it is shared: a partner’s steadiness can widen it for you, as the hand-holding studies suggest, and a partner’s fear can close it for both of you.

The contemplative traditions call this the difference between love and attachment. Nonattachment is often heard as caring less. It can instead mean seeing that caring deeply for someone gives us no ownership of their choices or future. That is not a case for accepting mistreatment or staying where you do not fit. It is the difference between loving someone and requiring them to play a particular part in your life.

Self-love, carefully defined

Self-love is easy to shrink into self-esteem or affirmations. A more useful definition is treating your own well-being as worth the same thoughtful consideration you would give someone you love: noticing your needs, respecting your limits and owning your mistakes without treating each one as a verdict on your worth. Self-compassion, treating yourself with the kindness you would offer a friend, has a reasonable evidence base. Across twenty samples, higher self-compassion went with less psychopathology (r = −.54), though all of them used the same questionnaire.¹¹⁰ A meta-analysis of 27 randomized trials found that self-compassion training reduced rumination, stress, depression, anxiety and self-criticism.¹¹¹

Self-love in this sense is not constant self-approval. Sometimes it means admitting you behaved badly, accepting help, or ending a comfortable pattern that is quietly costing you. Plenty of people are endlessly compassionate toward everyone except themselves.

Where love shows the limits of intelligence

Love may reveal the gap between intelligence and living better than anything else. We know the conversation has to happen, and we avoid it. We see the pattern, and we repeat it. We understand the other person’s view and cannot bear what it makes us feel. We know we cannot control someone else’s choices, and we keep trying. None of this is a shortage of intelligence. It is the distance between knowing something and being able to live by it, and love makes that distance visible because the stakes are so personal.

Love as a feeling makes someone matter to us. Love as action asks us to behave as though they matter, even when our feelings in the moment pull another way. And perhaps mature love goes one step further: caring for another person without losing the capacity to care for ourselves.

Love at work: teams, care and performance

We rarely use the word love to explain why a group achieves what none of its members could alone. Exceptional teams depend first on skill, resources, preparation, leadership and the structure of the task. But trust, mutual commitment and real concern for one another often appear alongside them. My argument is conditional: care and trust may help a team use the abilities it already has, through trust, coordination and commitment, rather than creating superior performance on their own. Love here is a name for those mechanisms, not an established causal variable.

Which love?

I do not mean romance, or loving your employer. I mean what researchers call companionate love: warmth, care and concern for one another’s success and well-being. It has the same three faces as love anywhere, a feeling, an understanding and an action, and the action matters most. A team can enjoy each other’s company and get nothing done. Teammates who do not much like each other can still show deep commitment through what they do.

Two crews

Rowing makes this concrete. Picture two eights with the same fitness and technique, a constructed comparison rather than real crews. In the first, each rower is mostly concerned with their own performance and how it looks. In the second, each feels responsible to the other seven. In the first boat some attention always goes to self-monitoring and fear of being the one who makes the mistake. In the second, more is free for the shared rhythm and for recovering together. That does not guarantee the second boat is faster. But eight strong rowers do not automatically make one strong crew, and relationships help create the conditions for coordination.

What the research shows

The most direct study of love at work comes from Sigal Barsade and Olivia O’Neill. Over sixteen months in a long-term care facility, units with a stronger culture of companionate love, meaning everyday affection, caring, compassion and tenderness among staff, had less absenteeism and burnout and more satisfaction and teamwork. Residents in those units had better mood and quality of life and fewer trips to the emergency room. A second survey of 3,201 employees in seven industries linked the same kind of culture to satisfaction, commitment and accountability.¹¹² It is one facility and the design is correlational, but the outcomes include the residents’ health, not only staff opinion.

Amy Edmondson’s study of 51 teams in a manufacturing company found that psychological safety, a shared sense that it is safe to take interpersonal risks such as admitting a mistake, went with learning behavior, and learning behavior with team performance.¹¹³ A later meta-analysis of 136 samples found psychological safety correlated .43 with individual task performance and .29 with team performance. But when individual performance was measured objectively rather than rated by someone, the correlation fell to .07, and most of the data came from the same source.¹¹⁴ Safety clearly helps people learn and speak up. Its effect on hard output is less certain.

In sport, a meta-analysis of 118 effect sizes found that team cohesion correlated .34 with performance. Task cohesion, commitment to the shared goal, correlated .45. Social cohesion, simply liking one another, correlated only .11.¹¹⁵ What predicts performance is commitment to the shared work, more than friendship, and cohesion seems to predict later performance more strongly than the reverse.¹¹⁶ Believing in the team matters too. Across 96 studies, collective efficacy, a group’s shared confidence that it can succeed, correlated .35 with group performance.¹¹⁷

Support from organizations and leaders shows a similar pattern. Employees who feel valued report much less exhaustion and perform somewhat better,¹¹⁸ the quality of the relationship with one’s manager tracks performance,¹¹⁹ and compassion from colleagues goes with commitment.¹²⁰

Care for the people the work serves matters as well. In a field experiment with university fundraising callers, those who spent a few minutes meeting a student whose scholarship their calls had helped fund spent 142 percent more time on the phone a month later and raised 171 percent more money, while the comparison groups did not change. Only 39 callers took part, so the numbers are vivid and fragile.¹²¹

None of this shows that love, as such, causes better performance. It shows that the relational texture of a group is part of the system in which performance happens, not background to it.

The hidden cost of fear

Recall the distinction from attentional control theory between how good a result is and what it cost to get there.¹ A team can produce excellent results while burning large amounts of attention and emotion on internal conflict, insecurity and mistrust. An athlete afraid of being blamed spends some attention protecting themselves rather than executing. Put the same athlete in a place where standards are high but mistakes can be discussed without humiliation, and more of that attention can go to the task. The result may not change at first. What may change is the effort required to sustain it, and over a season that could show up in learning, consistency, fatigue and adaptability.

In the language of the opening, this is the second question, cost, at the level of a group. And it suggests one more thing about the space between feeling and action. In a frightened team, that space closes for everyone at once, because each person’s threat becomes part of everyone else’s situation.

So here is a hypothesis, labeled as one: a caring, trusting team may improve collective efficiency by reducing the resources spent managing interpersonal threat, and the benefit may show up in cost before it shows up in output. It is a plausible extension of attentional control theory and the psychological safety research. It has not been tested in this form.

Intelligence a group can actually use

Earlier I argued that emotional skill shapes how much of a person’s intelligence becomes usable. The same may hold for groups. A team holds a great deal of knowledge, skill and perception, but it is only useful if people share information, listen and adjust. Fear of looking incompetent makes people withhold. Status competition gets good ideas dismissed. Distrust produces duplicated work and refused delegation.

The research on collective intelligence is suggestive and contested. A 2010 study of 699 people in 192 groups found a general factor in group performance that was not strongly predicted by members’ average or highest intelligence. It was predicted by members’ social sensitivity, by how evenly they took turns speaking, and by the proportion of women in the group.¹²² An independent replication found the opposite on the central point: average member IQ explained about 80 percent of the differences between groups, and social perceptiveness and gender did not help.¹²³ A later analysis of 22 studies by the original team still found the collaboration process more important than individual skill.¹²⁴ My honest summary is that a group’s intelligence depends both on who is in it and on how they work together, and the balance between the two is unsettled. But the possibility is worth naming: the intelligence a group contains is not necessarily the intelligence it can use.

Care with accountability

None of this means loving teams are gentle or conflict-free. Caring about someone’s development sometimes requires uncomfortable honesty. A teammate who cares about you may tell you your preparation is not good enough. A coach who respects an athlete holds them to demanding standards. A colleague challenges your idea because the shared work matters. The difference is whether the criticism aims to improve the work and support the person, or to humiliate, dominate or protect someone’s status.

One experiment captures the combination. Seventh graders got critical feedback on an essay with one of two notes attached. The ordinary note said the comments were feedback on the paper. The “wise” note said, in effect: I am giving you these comments because I have very high expectations and I know you can reach them. In the first study, more students revised their essays after the wise note, especially Black students, whose revision rate rose from 27 to 64 percent. In a second group the next year, when every student had to revise, the wise note improved the quality of Black students’ revisions, though not White students’.¹²⁵ High standards plus belief in the person seems to matter, in some settings, for some people. It is not a magic sentence.

When love is used against people

Love does not make hard things easy. It often makes people willing to do harder things: the athlete who finishes a brutal session because the boat needs them, the nurse who gives exceptional care through exhaustion, the engineer who gives a project extraordinary attention because they believe in the people it serves. Feeling connected to others and to a shared purpose supports the kind of motivation people experience as their own.⁷⁰

That is also exactly where the danger lies. The commitment that fuels effort can fuel overtraining, burnout and self-sacrifice past healthy limits. A coach can exploit an athlete’s loyalty. An employer can call the company a family to normalize demands no family should make. Love may increase what people are willing to give. Emotional skill and good leadership decide when giving more is right, and caring for a team includes protecting it from its own devotion. No amount of belonging replaces fair pay, reasonable hours or the right to leave.

What success means

Most performance models ask how to get the most out of an individual. Love asks what becomes possible when people care about something beyond their own performance, and it changes how we judge results. A team that wins while damaging its people, or a workplace that runs on chronic fear, scores well on the first of the three questions and badly on the other two. If love can make us capable of more, the question I will keep asking is how to tell when it is helping us flourish and when it is asking us to give more than we should.

Can it be learned?

In this movement each claim carries a label. Established means consistent evidence from meta-analyses or large trials. Plausible means supported by pieces of evidence but not tested directly. Proposed means my own idea.

If emotional skill were fixed at birth, everything above would explain differences between people and change nothing. If it can be learned, it becomes a question of education.

What we start with Established

We do not start equal. Temperament differs from the cradle: some people react strongly and recover slowly, others the reverse. In a twin study, about 42 percent of the variation in self-reported trait emotional intelligence was attributable to genetic differences.¹²⁶ That is a typical figure for personality-like traits, and it is a statement about variation within a population, not a ceiling on any individual. Heritable does not mean fixed. Height is highly heritable, and nutrition still changes it.

What training does Established

The training evidence is encouraging and smaller than its advocates claim. A meta-analysis of 58 studies found a moderate positive effect of emotional intelligence training, with published studies showing larger effects than unpublished dissertations, a sign that some of the published effect may be optimism.¹²⁷ A second meta-analysis, of 24 studies in healthy adults, found a moderate improvement after training that was still present at follow-up.¹²⁸ Many of these studies measure outcomes with self-report questionnaires, which are easier to move than behavior.

School programs have been studied on a much larger scale. A landmark 2011 meta-analysis of 213 school-based social and emotional learning programs, covering 270,034 students, found improved social and emotional skills, attitudes and behavior, and an 11-percentile-point gain in academic achievement.¹²⁹ A follow-up meta-analysis of 82 programs and 97,406 students, with follow-ups ranging from six months to eighteen years, found that benefits persisted, and that the size of the initial skill gains predicted later well-being.¹³⁰

Then the sobering part. When a later meta-analysis of 89 studies compared tightly run efficacy trials with real-world effectiveness trials, the real-world results were smaller on six of seven outcomes.¹³¹ And the largest school mindfulness trial yet, MYRIAD, randomized 84 schools and 8,376 students in the United Kingdom. Mindfulness training was no better than ordinary schooling for depression risk, social-emotional functioning or well-being.¹³² Good intentions, a good curriculum and good teachers were not enough on their own.

Therapy that teaches people to catch and test their interpretations has the deepest evidence. A 2025 series of meta-analyses covering 375 trials and nearly 33,000 adults found cognitive behavioral therapy effective across depression and anxiety disorders, with large average effects. The same review graded the certainty of the evidence for depression as very low, because only one study in ten was at low risk of bias, and the effect shrank when the weaker studies were removed.¹³³ Effective, probably, and less dramatically so than the headline numbers suggest.

What training does not do Established

None of this makes anyone more intelligent in the IQ sense. Mindfulness training improves some specific attention and working-memory tasks, with effects that shrink against active control groups.^134,135 On anxiety it can be a serious tool: in one randomized trial of 276 adults with anxiety disorders, an eight-week course did about as well as a standard antidepressant, with far fewer side effects.¹³⁶

Be wary of brain claims, too. In two combined randomized trials with brain scans of 218 people, eight weeks of mindfulness-based stress reduction produced no detectable structural brain changes compared with control groups.¹³⁷ Learning changes the nervous system; that is what learning is. But “rewire your brain in eight weeks” is a slogan, not a finding. The defensible claim is plainer: people can learn better ways of noticing and responding, through instruction, practice and feedback.

One mechanism has decent support. Researchers describe decentering as the ability to observe thoughts and feelings as passing events in the mind rather than as facts about the world.⁸⁵ A meta-analysis of mediation studies found strong, consistent evidence that mindfulness programs work partly by reducing cognitive and emotional reactivity, and moderate evidence for reduced rumination and worry, while warning that most of the studies had methodological weaknesses.⁸⁶

How I would teach it Proposed

I would teach it the way coaches teach a physical skill, not the way schools teach a subject. You can explain a rowing stroke in detail, and the explanation will not survive the third minute of a race. Emotional skill is the same. Understanding regulation in a quiet room is a different thing from using it in a fight with someone you love. What closes the gap is practice with feedback, in gradually harder conditions.

Here is the sequence I would use. It is a proposed teaching order, not a validated program. The first four steps map onto established research; the fifth makes agency an explicit outcome, because noticing and regulating are means, not ends.

Two principles matter more than the sequence. The first is progression. Awareness is learned first in easy conditions and only then tried when the stakes are higher. For anyone with a trauma history or severe distress, that step should be individualized and, where it fits, supported by a clinician.

The second is the one I would underline: train flexibility, not calm. Teaching someone to stay calm in every situation would be like teaching an athlete only one pace. Comforting a grieving friend, holding a boundary, getting out of danger and making a hard decision each need a different emotional response. A well-trained emotional system does not produce one preferred state. It gets better at finding and holding the state the moment needs.^56,55 That makes emotional education closer to skill acquisition than to stress reduction.

That raises the largest possibility in this piece. What if emotional development were treated as a basic part of education, alongside reading, mathematics and reasoning? Not to make people calmer, happier or more compliant, but to help them understand what they feel, think clearly under pressure, handle relationships and run more of their own lives. The MYRIAD result is a warning that doing this well is hard. It is not an argument against trying.

Practices that steady, practices that harness

Some practices help when a feeling is too big or too tangled to use. Others help put a feeling’s energy to work. You need both, because the goal is neither calm nor intensity but fit. None of these needs equipment or a course. Each comes with its evidence so you can decide how much weight to give it. Pick one and try it for a week.

To steady

To harness

If you live with trauma or severe distress, start with the steadying practices, go gently, and involve a clinician when it helps. If sitting with your feelings becomes distressing, stop.¹⁴⁵

For people who teach, coach, lead or care

• Ask what a feeling is about before asking how to reduce it. Anxiety may be accurate information that someone is underprepared.
• Do not confuse composure with skill. A quiet person may be regulating well, or suppressing, which carries costs.⁵³
• Teach range, not calm. Help people find the state the task needs, not one ideal state.
• Find where the difficulty actually sits. Awareness, understanding, regulation, reading others and acting on what you know are different problems with different remedies.
• Do not make circumstances a skills problem. When someone is overloaded, poor or unsafe, help with the load first.⁷³
• Remember the dark side. Emotional skill can be used to manipulate. Teach it alongside ethics, not instead of it.⁴⁶

Where this goes wrong

Most of the cautions in this piece have appeared where they belong. Gathered in one place, they are these. Calling everything an intelligence empties the word.³⁰ Emotional skill is a capacity, not a virtue, and can serve manipulation as easily as kindness.⁴⁶ Telling people with hard lives that they need better regulation mistakes circumstances for a skills deficit.⁷³ Respecting a feeling as information is not the same as obeying it.^20,64 Accepting a feeling is not accepting a harmful situation, and reframing suits what cannot be changed better than what can.^68,146 Mindfulness helps some people with some problems; the largest school trial found no benefit, and a minority of meditators have adverse experiences.^132,145 If sitting with your feelings becomes distressing, stop, and talk to a teacher or a clinician.

Some popular ideas sound established and are not. One holds that anger often masks hurt or fear. It comes from emotion-focused therapy, where clinicians distinguish primary emotions from secondary ones that cover them, and it may be useful in practice, but I found no test of the distinction itself.¹⁴⁷ The somatic marker story, the sunny-day study and the trading-floor study belong on the same cautious list.

A measurement idea, clearly labeled

This section is a proposal, not a finding. In the earlier piece I suggested that performance might be best predicted by how well feeling, arousal, attention and task demands line up, rather than by any one of them. Here I widen the claim and say how it could be shown wrong.

The hypothesis: the ability to integrate emotional information with reasoning and action is a distinct capacity that predicts good outcomes across life, in relationships, decisions, psychological flexibility and well-being over years, beyond what cognitive ability, existing emotional intelligence tests, personality and baseline mental health already predict. If it predicts only performance, it is a performance model. If it predicts adaptation across many areas, the case for treating it as a broader human capacity gets much stronger.

There is a sharper version worth testing. One idea, which I find intuitive, is that emotional skill amplifies intelligence: the smarter you are, the more emotional skill matters, because there is more capacity to waste or to use. The opposite idea is that emotional skill compensates: it matters most for people with less cognitive ability to fall back on. The one study I found that tested this directly, with supervisors rating employees’ job performance, supported compensation.¹⁴⁸ My intuition may simply be wrong. It is also possible the answer differs by domain, and that emotional skill matters most in interpersonal judgment and least in technical problem-solving.

How it could fail

The hardest test is not whether the idea predicts anything, but whether it is anything new. If a measure of integration were interchangeable with an existing one, such as emotional intelligence, emotion regulation or psychological flexibility, it would be a new name for an old thing. If existing measures together predicted as well, it would be redundant. And if it showed up only in what people say about themselves, never in their bodies or behavior, it would be a questionnaire, not a capacity. The methodological detail is in a separate research note.

If integration adds nothing beyond what we already measure, the hypothesis is wrong, and that would be worth knowing.

What this leaves you with

I began with Aristotle’s observation that anger is easy and getting it right is rare. The research adds texture rather than a verdict: the skill is real but modest as a predictor, its parts are better supported than its name, it can be partly learned, and it is no substitute for character or circumstance.

Where I have ended up is somewhere I did not expect to go when I started. I came to this as a question about performance. I leave it as a question about freedom. The greatest cost of not knowing what we feel, in my judgment, is not that we feel too much. It is that we can be governed by feelings we never noticed while believing every choice was our own. And the greatest benefit of emotional development is not permanent happiness, which nothing provides, or the end of anxiety, anger, sadness or fear, which would be a loss. It is the space between what we feel and what we do, widened a little at a time, held more lightly, shared with other people, and stood on while the weather passes.

The three questions I began with now read differently to me. Effectiveness is what we know how to measure. Cost is what we are learning to measure. Agency, how far what we do is chosen, is what we measure least, and I have come to believe it matters most. That is a philosophical position, not a research finding, and I hold it in my own voice. Emotional skill touches all three, but its deepest work is on the third.

I think often about that year with the man who came to me after his diagnosis. He did not get to choose what was happening in his body. Every measure I knew how to take that year was about the first question. What I saw, week after week, belonged to the third: what he did with the time, the rides, the conversations, the quiet, his family, the outdoors. I still cannot measure what helped him. But I understand it better than I did, and most of what I understand is in this piece.

Perhaps one measure of intelligence, if we must use the word, is not how well we control our inner life but how well we respond to it. I would put it more simply. We cannot always choose what we feel. We can learn to understand it, weigh it against what is true, and find more room to choose what we do next. That can be learned. Start small, start where you are, and be suspicious of anyone, me included, who tells you it is simple.

Appendix: a field guide to nineteen emotions

For each emotion, the table gives what it registers, meaning the situation it is a reading of; what it seems to be for when it is doing its job; and what happens when it takes over. Where a researcher’s definition is quoted, it is the one I found most useful, not the only one. Many of the “registers” lines come from Richard Lazarus’s core relational themes, a compact statement of what each emotion is about.¹⁴⁹ The “for” and “takes over” columns are my synthesis of the research in this piece and its companion. Read them as working descriptions, not laws.